|Home | About | Journals | Submit | Contact Us | Français|
Successful host defense against bacteria such as Staphylococcus aureus (SA) depends on a prompt response by circulating polymorphonuclear leukocytes (PMN). Stimulated PMN create in their phagosomes an environment inhospitable to most ingested bacteria. Granules that fuse with the phagosome deliver an array of catalytic and noncatalytic antimicrobial peptides, while activation of the NADPH oxidase at the phagosomal membrane generates reactive oxygen species within the phagosome, including hypochlorous acid (HOCl), formed by the oxidation of chloride by the granule protein myeloperoxidase in the presence of H2O2. In this study, we used SA-expressing cytosolic GFP to provide a novel probe of the fate of SA in human PMN. PMN bleaching of GFP in SA required phagocytosis, active myeloperoxidase, H2O2 from the NADPH oxidase, and chloride. Not all ingested SA were bleached, and the number of cocci within PMN-retaining fluorescent GFP closely correlated with the number of viable bacteria remaining intracellularly. The percent of intracellular fluorescent and viable SA increased at higher multiplicity of infection and when SA presented to PMN had been harvested from the stationary phase of growth. These studies demonstrate that the loss of GFP fluorescence in ingested SA provides a sensitive experimental probe for monitoring biochemical events within individual phagosomes and for identifying subpopulations of SA that resist intracellular PMN cytotoxicity. Defining the molecular basis of SA survival within PMN should provide important insights into bacterial and host properties that limit PMN antistaphylococcal action and thus contribute to the pathogenesis of staphylococcal infection.
Polymorphonuclear leukocytes (PMN)3 provide an important component of innate host defense against microbes and are especially important in response to pyogenic bacterial infections, such as those caused by Staphylococcus aureus. Beginning with phagocytosis, antimicrobial responses of PMN are tightly coupled with activation of the phagocyte NADPH oxidase and release of granule contents into the nascent phagosome (1). Oxidants, generated de novo by the NADPH oxidase under aerobic conditions, exert direct toxic effects on microbial targets, act as substrates to support generation of other reactive species, and promote and modify the intrinsic activity of other agents present in the phagosome. The fusion of preformed granules with the phagosome delivers a wide array of biologically active agents that likewise exert both direct and indirect effects on the ingested microbe. As such, the phagosome provides a specialized microenvironment where many different cytotoxic compounds operate alone and synergistically to equip the PMN to respond to many different types of microbes (2). The complex synergy in the phagosome is illustrated by the capacity of the granule protein myeloperoxidase (MPO) in the presence of H2O2 to catalyze the oxidation of chloride and generate hypochlorous acid (HOCl), a potent microbicidal agent (reviewed in Ref. 3).
Ironically, the same chemical and biochemical complexity that invests the PMN intraphagosomal microenvironment with such broad and potent antimicrobial cytotoxicity has often created a challenge to the identification of specific contributions of individual components to antibacterial action. Where the principal indicator of antimicrobial action is bacterial death, quantitated as diminished colony forming units (CFU), the difficulty in distinguishing the contribution of specific components is especially difficult. Even in the case of PMN-S. aureus interactions, where a functioning phagocyte NADPH oxidase is prerequisite for optimal bacterial killing, debate persists as to the nature of the potentiating effects of the oxidase and the specific nonoxidase components in the phagosome that contribute synergistically to amplify bacterial damage.
In this report, we take advantage of the specific and selective ability of the MPO-H2O2-chloride system to bleach GFP (4), which can be expressed within the cytosol of a wide variety of bacterial species, to study the fate of S. aureus in normal human PMN. Considered in the context of clinical studies of interactions between PMN and S. aureus that indicate that some ingested bacteria survive (5-7), the assessment of GFP bleaching during and after phagocytosis directly demonstrates differences in the effects of PMN cytotoxins on individual cocci under a variety of conditions. These studies provide novel insights into host and bacterial factors that contribute to the fate of ingested staphylococci within human PMN.
Clinical grade dextran T500 (m.w. 500,000 daltons) was purchased from Pharmacosmos A/S. Sterile endotoxin-free H2O and 0.9% sterile endotoxin-free sodium chloride for patient use were purchased from Baxter Healthcare. Ficoll-Paque PLUS was purchased from GE Healthcare (formally Amersham Biosciences); HEPES, HBSS, and Dulbecco’s PBS (DPBS), with or without divalent cations were obtained from Mediatech. PROTOCOL HEMA-3 staining kit was purchased from Fisher Diagnostics, whereas human serum albumin 25% was obtained from Talecris Biotherapeutics. Tryptic soy broth (TSB) and bacto agar were purchased from BD Biosciences. Chloramphenicol, diphenylene iodonium (DPI), poly-l-lysine coated Poly-Prep slides, and sodium gluconate were obtained from Sigma-Aldrich. Na125I (specific activity 17.4 Ci/mg) was purchased from PerkinElmer, whereas unlabelled NaI was purchased from Mallinckrodt. Saponin and glucose oxidase were obtained from Fluka Biochemika. Affinity-purified rabbit polyclonal Ab against GFP was purchased from Eusera. All other reagents were purchased from Fisher Scientific.
PMN from normal healthy volunteers and myeloperoxidase-deficient individuals were purified from peripheral blood as described in Ref. 8. Donor consent was obtained from each individual following the protocol approved by the Institutional Review Board for Human Subjects at the University of Iowa. Isolated PMN were suspended in sterile endotoxin-free HBSS without divalent cations, counted, and diluted to a final density no greater then 20 × 106/ml. From 95–98% of the isolated leukocytes were PMN, as determined by microscopic examination of the cell suspension after staining with the HEMA-3 kit. PMN were held on ice until use in subsequent assays.
We used S. aureus ALC 1435, a derivative of RN6390 containing the plasmid pALC1420 encoding GFP whose constitutive expression is controlled through the sar P1 promoter (9). S. aureus ALC 1435 was provided by Dr. Ambrose Cheung, Department of Microbiology, Dartmouth Medical School Hanover, NH.
Unless otherwise stated, we used mid-logarithmic growth staphylococci in our studies. Bacteria were grown overnight at 37°C shaking at 180 rpm in TSB supplemented with 10 μg/ml chloramphenicol to maintain the plasmid. Before use in the assay, the overnight culture was diluted to a density of 0.05 at an OD550 in fresh TSB medium containing 0.01% HSA and 10 μg/ml chloramphenicol. The bacteria were subsequently subcultured 2.5 h to reach mid-logarithmic phase (OD550 ~0.1–0.2) before harvesting by centrifugation at 5,000 × g for 5 min in a tabletop microcentrifuge, and resuspending in 20 mM HEPES buffered HBSS with divalent cations. The collected cells were then used immediately or placed on ice until opsonization.
In selected experiments, indicated in the text, staphylococci were serially subcultured to obtain a population of bacteria that was more uniformly in the early exponential phase, using a modification of a previously described method (10). Initially, staphylococci in tryptic soy broth supplemented with 0.01% HSA and 10 μg/ml chloramphenicol were cultured at 37°C shaking at 180rpm for 7 h. Bacteria were pelleted (2,000× g for 4 min at 4°C in Beckman Coulter J6-MI), washed twice with vortexing in sterile endotoxin-free saline, and stored overnight on ice at 4°C in TSB supplemented with 0.01% HSA and 10 μg/ml chloramphenicol. The following day, the stored bacteria were inoculated into fresh broth supplemented as above and the subculturing process was repeated three times (once for 3 h, then twice for 2 h each), with each subculture starting with a 1/100 dilution of the previous culture. After the final 2-h subculture, the pelleted bacteria were washed twice in saline and resuspended in TSB supplemented with 0.01% HSA, 10 μg/ml chloramphenicol, and 10% glycerol for storage at −80°C. Either freshly prepared or thawed stored samples were washed in DPBS, resuspended in 20 mM HEPES buffered HBSS with divalent cations, and held on ice until opsonization.
Bacteria were suspended at 107/ml in 20 mM HEPES-buffered HBSS with divalent cations, 1% HSA, and 10% pooled human serum and placed in a 37°C shaking water bath for 20 min to opsonize before being fed to PMN. After opsonization, the bacteria were pelleted (2,000 × g, 5 min), washed, and immediately used in phagocytosis assays (see below).
Before challenge with S. aureus, PMN were diluted to 107/ml and incubated at 37°C for 10 min with or without the addition of 10 μM DPI, or 0.5, 1, 2, or 5 mM NaN3 in HBSS containing divalent cations, 1% HSA, and 10% pooled human serum. Subsequently, the opsonized S. aureus and PMN samples were mixed at a ratio of one CFU of S. aureus (equivalent to four cocci) to one PMN (i.e., multiplicity of infection (MOI) = 1:1, both PMN and CFU bacteria at 5 × 106/ml) in 5 ml polypropylene roundbottom tubes, and incubated with shaking (180 rpm) in a 37°C water bath. After 5 min, samples were gently spun (500 × g) for 5 min in a Savant high speed tabletop centrifuge to pellet PMN and associated bacteria. The supernatant containing extracellular bacteria was discarded, and the pellet resuspended in the original volume of HBSS containing divalent cations supplemented with 1% HSA and 10% pooled human serum and returned to the 37°C water bath for specified intervals. To confirm that PMN-associated GFP-SA were sequestered within PMN, DPI-PMN challenged with opsonized GFP-SA were recovered after 10 min, stained with 0.1% crystal violet for 15 min at room temperature, and examined by fluorescence microscopy. Whereas fluorescence of GFP-SA directly exposed to crystal violet was immediately quenched, PMN-associated GFP-SA after 10 min of phagocytosis by DPI-PMN remained fluorescent (data not shown). These data indicate that the GFP-SA were sequestered within closed phagosomes by 10 min of phagocytosis under these experimental conditions.
For experiments in which opsonized GFP-SA were coated with purified MPO before being fed to PMN, the opsonized GFP-SA were pelleted and resuspended in PBS ± purified MPO for 1 h. Bacteria were pelleted, resuspended and fed to PMN as described above. It should be noted that MPO bound to GFP-SA very inefficiently; exposure of GFP-SA to 1 μM MPO resulted in retention of ~1.4 nM or 14 fmoles (0.2% of the initial amount) on the bacterial pellet containing 108 cocci (n = 3). For experiments in which extracellular oxidants were provided by glucose-glucose oxidase, the buffer was supplemented with 10 mM glucose and a final concentration of 2 μg/ml glucose oxidase was added to the resuspended DPI-treated samples.
After the 5 min incubation and resuspension (see above), aliquots were removed to quantitate bacterial viability at the experimentally designated zero time. Each sample was diluted 1/50 into a standard 96-well plate containing 1% saponin prepared in H2O, incubated at room temperature for 15 min, mixed, and serially diluted with 0.9% saline. Several dilutions of each sample were plated drop-wise on TSB agar and placed overnight at 37°C for enumeration of colonies. The CFU averaged from replicate plates were calculated and data represented as a percent of viable organisms at zero time.
At each time point, samples were vortexed, and ~7.5 × 104 PMN containing S. aureus were spun onto primed (20 mM HEPES buffered HBSS supplemented with 0.2% HSA) poly-l-lysine coated poly-prep slides using the cytocentrifuge (Shandon Cytospin 2; Shandon Southern Products) at 18 × g for 10 min. Each slide was fixed in 10% formalin for 15 min at room temperature, washed once in DPBS, and covered with a coverslip using mounting medium to preserve fluorescence. A Zeiss Axioplan 2 photomicroscope was used to analyze cocci:PMN and to visualize loss of GFP fluorescence. For each sample, 50–100 PMN were examined. Experiments examining the effect of chloride concentration on loss of GFP fluorescence by ingested staphylococci used a chloride-free buffer to which specified amounts of a 122 mM chloride buffer (with identical cations) were added to achieve the desired final concentration of chloride. The chloride-free buffer had chloride salts in 10 mM PBS (pH 7.4) replaced with gluconate salts.
GFP-SA (5 × 107/ml) were suspended in a 3 ml cuvette and scanned (λex = 395 nm, λem = 480 – 540 nm) in a thermostated 37°C spectrofluorometer (Fluoromax 3, Jobin Yvon). Scanning was then performed immediately after each aliquot of H2O2, HOCl, NH2Cl (monochloramine) over the indicated concentration ranges was added to the cuvette to assess the effects of different concentrations of oxidants or the cell-free MPO-H2O2-chloride system (see below) on GFP-SA fluorescence.
In addition to assessing the impact of reagent-grade HOCl on fluorescence of GFP-SA, the effect of a HOCl-generating system, composed of purified MPO (2–80 nM), H2O2 (100 – 450 μM), and chloride (130 mM), was assessed by serial scanning. Purified MPO (Reinheit Zahl ~0.80) and H2O2 in PBS (pH 7.4) comprised the cell-free MPO-H2O2-chloride system. The stock concentrations of MPO and H2O2 were determined spectrophotometrically, using ε473 = 75 mM−1 cm−1 for reduced minus oxidized MPO and ε240 = 43.6 M−1 cm−1 for H2O2. GFP-SA (5 × 107) were suspended in 3 ml of PBS containing the complete system, MPO alone, or H2O2 alone and scanned serially every 2 min. In parallel, the chlorinating capacity of the MPO-H2O2-chloride system was assessed using a previously described method to quantitate taurine chlorination (11).
In experiments in which exogenous H2O2 was generated by glucose-glucose oxidase, the amount of H2O2 was quantitated using a H2O2 electrode (Apollo 4000, World Provision Instruments). The DPI-treated samples were prepared as described in the phagocytosis assay and resuspended in buffer supplemented with 10 mM glucose. The cell suspension was added to the prewarmed 37°C chamber, stirred for 5 min until the signal stabilized, and glucose oxidase was added to a final concentration of 50 μg/ml. The initial velocity of the glucose-glucose oxidase activity was calculated from an H2O2 standard curve generated daily. Data are expressed as nmoles/min.
MPO-dependent iodination was quantitated as previously described (12, 13). PMN were suspended at 107/ml and incubated at 37°C for 10 min in 20 mM HEPES buffered HBSS with divalent cations, 1% HSA, and 10% pooled human serum, with or without 100 μM sodium azide, in the presence of a mixed stock of radiolabeled Na125I and unlabeled iodide at a final concentration of 10 μM. The pretreated cells were then mixed with the opsonized S. aureus at an MOI of 1:1, 2:1, and 5:1, and placed into a 37°C water bath for 30 min. The reaction was stopped using a mixture of sodium thiosulfate and cold NaI, and the proteins and lipids precipitated with 20% trichloroacetic acid. The samples were washed twice in 10% TCA, sedimented, and the final pellet was resuspended in 10% SDS and counted for radioactivity.
The structural integrity of cytoplasmic GFP in GFP-SA after bleaching by HOCl or 120 min after phagocytosis by PMN, both conditions under which the fluorescence of GFP-SA was completely bleached, was assessed by probing electroblots with a affinity-purified polyclonal rabbit Ab against GFP. GFP-SA (5 × 107) treated with buffer or with reagent grade HOCl were solubilized by vigorous vortexing in hot SDS sample buffer (14) for 5–10 min. Proteins were separated by SDS-PAGE in 5–20% gradient polyacrylamide gels and electroblotted to nitrocellulose filter. The electroblot was blocked, probed with rabbit anti-GFP and a secondary donkey antirabbit Ab conjugated with HRP, and visualized with chemilumiscence reagents (SuperSignal West Detection Kit, Thermo Fisher Scientific). In experiments assessing the integrity of GFP in GFP-SA phagocytosed by PMN, a subcellular fraction enriched for granules and phagosomes was analyzed. PMN or PMN that had phagocytosed GFP-SA were pelleted, resuspended in relaxation buffer (100 mM KCl, 3 mM NaCl, 1 mM Na2 ATP, 3.5 mM MgCl2 (pH 7.3)), and disrupted by N2 bomb cavitation, as previously performed (15). Unbroken PMN and nuclei were removed by low speed centrifugation (500 × g), and granules and phagosomes were pelleted (20,000 × g, 20 min, 4°C). Pellets of PMN containing GFP-SA, as well as a pellet of GFP-SA alone, were solubilized for SDS-PAGE and immunoblotting as described above.
During our studies of the synergistic activity of the PMN NADPH oxidase and extracellular Group IIA phospholipase A2 on ingested S. aureus (16), we noted that GFP-expressing S. aureus (GFP-SA) ingested by normal PMN lost fluorescence over time (Fig. 1A). The loss of fluorescence was not strain-specific and was seen in each of several different strains of S. aureus, including both hospital-acquired and community-associated methicillin-resistant strains (data not shown). The loss of GFP fluorescence required ingestion of the bacteria by PMN, as extracellular SA remained fluorescent when incubated with PMN in which phagocytosis was inhibited by treatment with dihydrocytochalasin B (data not shown). Reasoning that the loss of GFP fluorescence reflected damage to GFP-SA in the phagosome, we compared the relative kinetics of the loss of fluorescence vs loss of viability of ingested bacteria (Fig. 1B). There was a relative delay in the onset of loss of GFP-SA fluorescence relative to the loss of viability in PMN-associated bacteria (Fig. 1B). At early time points, PMN killing of SA exceeded the loss of fluorescence, although values were virtually identical at 120 min, with ~80% of the ingested bacteria dead and nonfluorescent at that time. These data indicate that PMN cytotoxins significantly damaged or killed ingested SA in the phagosome before GFP fluorescence was lost and, conversely, that ~20% of ingested GFP-SA resisted PMN cytotoxic effects and remained fluorescent. Phagosomes containing GFP-SA did not communicate with the extracellular space, as demonstrated by the failure of crystal violet to quench intraphagosomal GFP-SA at 120 min (data not shown), excluding leakage of cytotoxins from the phagosome as an explanation for the subpopulation of surviving staphylococci.
In contrast to the loss of GFP fluorescence in normal PMN (Fig. 1A), internalized GFP-SA remained fluorescent after ingestion by PMN treated with diphenylene iodonium (DPI) (Fig. 2A), a potent inhibitor of flavoproteins, including the phagocyte NADPH oxidase (17). Despite complete inhibition of NADPH oxidase activity, phagocytosis is normal in DPI-treated PMN (16, 18). To determine whether the dependence on an active NADPH oxidase reflected a need for the oxidase as an oxidant source, we provided an exogenous H2O2-generating system to PMN with endogenous oxidase activity inhibited by DPI. DPI-treated PMN were challenged with GFP-SA in the absence or presence of glucose-glucose oxidase (GGO), and fluorescence and viability of ingested GFP-SA were assessed. The GGO system generated 90–110 nmoles/ml H2O2, as measured directly with a H2O2 electrode. Whereas GFP-SA within DPI-treated PMN remained fluorescent at 120 min after ingestion (Fig. 2B), significantly fewer organisms were persistently fluorescent in DPI-PMN exposed to the H2O2 source, although exogenous H2O2 did not restore levels to that seen in normal PMN. In parallel, DPI-PMN without supplemental H2O2 were less able to kill GFP-SA than were control PMN, but the addition of GGO partially restored bacterial killing to the DPI-treated PMN (data not shown). The augmentation of activity in DPI-PMN in the presence of the exogenous oxidant source required all components of the GGO system and was inhibited by the addition of catalase (Fig. 2B). These data indicate that the loss of GFP-SA fluorescence required oxidants and that the NADPH oxidase was the source of ROS in the phagosome.
To determine whether MPO contributed to or was required for the loss of GFP-SA fluorescence, normal PMN were challenged with GFP-SA in the presence of sodium azide (NaN3). Whereas NaN3 inhibits MPO activity in intact PMN (3), the NADPH oxidase is NaN3-resistant (13, 19-21), and NaN3-treated PMN generate more oxidants than do normal PMN stimulated in the absence of NaN3 (13, 20, 22, 23). The fluorescence of GFP-SA persisted in NaN3-treated PMN in a manner very similar to that seen in DPI-treated PMN (Fig. 2C). In parallel, PMN from a MPO-deficient patient (MPO-D) (24) were challenged with opsonized GFP-SA and ingested GFP-SA remained fluorescent in MPO-D PMN (Fig. 2C and data not shown). Exposing GFP-SA to 50 nM MPO before ingestion by MPO-D PMN partially reconstituted the PMN-mediated loss of GFP-SA fluorescence, but not to values seen with normal PMN (Fig. 2C), perhaps a reflection of the inefficiency with which MPO associated with GFP-SA during the exposure step (vide supra). Taken together, these data confirmed that enzymatically active MPO was required for the loss of fluorescence by GFP-SA phagocytosed by PMN.
Dependence of intraphagosomal GFP-SA fluorescence decay on oxidants and on active MPO implicated a likely role for the PMN MPO-H2O2-Cl system in mediating the loss of fluorescence by ingested bacteria (3). Consistent with this prediction, loss of intraphagosomal GFP-SA fluorescence correlated inversely with the concentration of chloride in the incubation buffer (Fig. 2D). In the absence of chloride, ~90% ingested GFP-SA remained fluorescent at 120 min, in contrast to the ~25% value at 122 mM chloride. As phagocytosis, degranulation, and NADPH oxidase activity in PMN were normal at the varied concentrations of chloride (data not shown) and gluconate (up to 150 mM) did not inhibit chlorination by the cell-free MPO-H2O2 system (data not shown), the persistence of fluorescence by ingested GFP-SA in the absence of chloride reflected the consequences of an incomplete HOCl-generating system.
To assess the susceptibility of GFP-SA to the direct effects of HOCl, GFP-SA were incubated in varied concentrations of reagent grade HOCl while GFP fluorescence was simultaneously monitored in a spectrofluorometer (Fig. 3A). Fluorescence of GFP-SA decayed in the presence of HOCl in a dose-dependent fashion, with fluorescence of 5 × 107 GFP-SA decreased to 50% of control in the presence of 46.2 ± 9.9 μM (mean ± SEM, n = 5 separate experiments, each in triplicate). The HOCl-generating system was likewise effective; in the presence of 50 nM purified MPO, 150 μM H2O2, and 130 mM chloride, fluorescence of 5 × 107 GFP-SA decreased 50% within 12.3 min at room temperature (n = 5) (Fig. 3B). For both HOCl and the MPO-H2O2chloride system, there was an initial delay in loss of GFP-SA fluorescence; at low concentrations of HOCl (Fig. 3A) or at early time points in MPO dependent generation of HOCl (Fig. 3B), there was little change in fluorescence, followed by more rapid loss of fluorescence. For example, the initial additions of HOCl had little, if any, influence on fluorescence of GFP-SA; once 40 μM HOCl was added, loss of GFP-SA fluorescence was dose-dependent (Fig. 3A). Likewise, in the first 6 min, the HOCl-generating system had little effect on GFP-SA fluorescence, whereas subsequently the loss of fluorescence became nearly linear (Fig. 3B). Despite loss of fluorescence, the GFP in GFP-SA remained structurally intact. Immunoblots of GFP-SA exposed to millimolar concentrations of HOCl (Fig. 3C), ~300-fold greater than those mediating loss of fluorescence, or of GFP-SA after 120 min in PMN (Fig. 3D), a time when >90% of the organisms have lost fluorescence, demonstrated no decrease in size or amount of GFP Thus, the MPO-dependent HOCl-generating system, both in vitro and within human PMN, reduced the fluorescence of GFP-SA without degrading GFP.
H2O2 accounts for nearly all the oxygen consumed by the NADPH oxidase of activated PMN (25, 26) and serves as the substrate for the production of HOCl in the phagosome. To assess the susceptibility of GFP-SA fluorescence to more proximal oxidants generated by activated PMN, GFP-SA were exposed to varied amounts of reagent-grade H2O2 while simultaneously monitoring fluorescence spectrofluorometrically (Fig. 3E). There was no loss of fluorescence at concentrations of H2O2 as high as 10 mM. This failure of H2O2 treatment to compromise GFP-SA fluorescence is consistent with the data from MPO-deficient PMN, where relatively larger amounts of oxidants from the NADPH oxidase in the absence of MPO (13, 20, 22, 23) did not diminish the fluorescence of ingested GFP-SA (Fig. 2C and data not shown). These data demonstrate that only a subset of oxidants modified by reaction with MPO abrogated the fluorescence of GFP-SA.
In addition to HOCl, the MPO-H2O2-Cl system reacts with both bacterial and host proteins in the phagosome to generate an array of biologically active derivatives, including long-lived chloramines (reviewed in Ref. 3). Reasoning that chloramines might also contribute to the loss of GFP-SA fluorescence, we investigated whether monochloramine treatment altered the fluorescence of GFP-SA. However, GFP-SA fluorescence was unaffected by concentrations of monochloramine (NH2Cl) as high as 7 mM (Fig. 3E). Taken together, these data strongly implicate HOCl produced in the phagosome by the MPO-H2O2-chloride system as the agent responsible for the loss of GFP-SA fluorescence. Given that peroxidase-dependent chlorination of tyrosine 66 in GFP renders GFP no longer fluorescent (27), the loss of fluorescence by GFP-SA likely reflects bleaching of GFP by HOCl in the phagosome.
HOCl is highly reactive, with a clearly defined hierarchy in its propensity to react with biological substrates (28, 29). Although HOCl can rapidly diffuse across membranes, successful diffusion depends on the presence and relative susceptibility of potentially competing substrates that might consume HOCl before it could enter a target cell. We reasoned that the susceptibility of GFP to bleaching would depend on the presence and relative susceptibility of competing substrates on the surface and in the cytoplasm of GFP-SA that would compete with GFP for reaction with HOCl. The composition and properties of the staphylococcal envelope change extensively during the progression of bacteria from logarithmic to stationary phase (30). Accordingly, the array of competing substrates available to react with HOCl would likewise vary with the growth-phase of the GFP-SA. We therefore anticipated that the susceptibility of ingested GFP-SA to bleaching within the PMN would depend on the growth stage of the bacteria at the time that they were ingested. To test this hypothesis, we compared PMN bleaching of GFP-SA recovered from mid-logarithmic phase, our standard condition, stationary phase, or from cultures in which rapidly growing SA were selectively enriched (see Materials and Methods). The population that was enriched in rapidly growing GFP-SA (early exponential phase) was more susceptible to PMN killing and bleaching than were our standard mid-log grown bacteria (Fig. 4). Conversely, a significantly greater fraction of GFP-SA harvested after overnight culture (stationary phase) was more resistant to bleaching. In light of the relative kinetics of GFP-SA bleaching and killing (Fig. 1B), the data suggest that growth phase-related changes in the composition of surface structures, dictated the potential substrates available to compete with GFP for reacting with HOCl.
To examine in a different way the impact of competing substrates on bleaching of GFP-SA, we challenged PMN with GFP-SA at varied MOI, comparing our standard condition (1 CFU:1 PMN, equivalent to 4 cocci: 1 PMN) with higher ratios of 2 CFU:1 PMN and 5 CFU:1 PMN (equivalent to 8 cocci: 1 PMN and 20 cocci: 1 PMN, respectively). The number of cocci per PMN was enumerated and the fraction of intracellular organisms that remained fluorescent determined 120 min after ingestion by PMN (Fig. 5A). As the multiplicity of infection increased, PMN ingested more bacteria, as reflected by an increase in the average number of cocci per PMN. In parallel, the fraction of GFP-SA that remained fluorescent (Fig. 5A) and resisted killing increased at higher MOI. The viability of intracellular GFP-SA 120 min after phagocytosis was 30.9 ± 6.0, 89.0 ± 17.4, and 101.8 ± 8.8 (mean ± SEM, n = 6) at MOI of 1 CFU: 1 PMN, 2 CFU: 1 PMN, and 5 CFU: 1 PMN, respectively. Thus, survival of ingested GFP-SA paralleled the increases in the bacterial challenge presented to PMN, suggesting that the organism load at higher ratios of infection exceeded the capacity of intraphagosomal antimicrobial activity, the H2O2-MPO-chloride system, or both. To assess specifically the capacity of MPO-mediated biochemistry independent of GFP bleaching, we quantitated protein radioiodination by PMN stimulated by GFP-SA at varied MOI (Fig. 5B). As more bacteria were ingested at higher MOI (Fig. 5C), there were commensurate increases in protein radioiodinated and in the fraction of bacteria that survived.
To determine whether the changes in bleaching and killing of GFP-SA at higher MOI were due to limiting amounts of the components of the MPO-H2O2-chloride system, we supplemented PMN with MPO or with oxidants. Pre-exposing GFP-SA to 50 nM purified MPO before incubation with PMN did not significantly increase either radioiodination (Fig. 5B) or bacterial killing (data not shown) that occurred during and after phagocytosis. Even when GFP-SA were exposed to 1 μM MPO before feeding to PMN, conditions in which 1.4 ± 0.3 nM (n = 3) enzymatically MPO could be detected after incubation; neither killing nor bleaching of GFP-SA was increased. Thus, the ability of some ingested GFP-SA to escape killing and bleaching at the MOI tested did not reflect insufficient mobilization of MPO from granules during phagocytosis. However, supplemental H2O2, achieved by coincubating the PMN-GFP-SA with glucose-glucose oxidase, increased the percentage of ingested GFP-SA that were bleached, but the increase was statistically significant only at the highest MOI examined (Fig. 5C). Thus, insufficient production of oxidants, but not the release of MPO, accounted, at least in part, for the ability of GFP-SA in PMN phagosomes to escape killing and bleaching at the each of the MOI tested.
Innate host defense against infecting staphylococci depends in large part on optimal PMN function, an association most dramatically demonstrated by the frequency and severity of such infections in patients with neutropenia or with inherited PMN disorders such as chronic granulomatous disease (31). In addition, both oxidants and MPO figure prominently in efficient killing of staphylococci, as PMN from patients with chronic granulomatous disease or inherited MPO deficiency exhibit defective antimicrobial activity (32). However, it has long been recognized that a significant fraction of ingested S. aureus can survive in the phagosome of normal PMN, thereby creating a potential reservoir for protracted infection and inflammation (6). The recent international emergence of community-associated methicillin-resistant S. aureus as a serious infectious disease problem has reinvigorated efforts to unravel both the strategies of the organism and the PMN (33-41) to influence the outcome of the host-pathogen interaction. In this context, our observations of the fate of GFP-SA provide several important insights into events within the phagosome of normal PMN and identify variables in SA-PMN interactions that may contribute to the eventual fate of ingested staphylococci.
Our data extend previously reported observations of the susceptibility of purified GFP or GFP expressed in Escherichia coli to HOCl or the MPO-H2O2-Cl system of PMN (4). The loss of GFP fluorescence in the ingested SA depended on the presence of an intact MPO-H2O2-chloride system (Fig. 2), with partial reconstitution of GFP bleaching by DPI-treated PMN when an exogenous source of H2O2 was supplied (Fig. 2B), indicating an essential role for HOCl in this setting as well. PMN-mediated killing of SA exhibited the same features and prerequisites, thus demonstrating that the phagocyte oxidase provides oxidants that are essential for efficient killing of staphylococci and that operate synergistically with MPO to support the antimicrobial HOCl-generating system. The loss of GFP-SA fluorescence was produced only by HOCl and was not seen with H2O2 or NH2Cl, reactive species also generated in the PMN phagosome (42), even at ~1000-fold higher concentrations those that were effective when using HOCl. Furthermore, HOCl and PMN-mediated loss of GFP-SA fluorescence occurred without detectable degradation of GFP (Fig. 3, C and D). Thus, the modification that disrupts the chromophore of GFP reflects very specific HOCl-mediated chemistry within the PMN phagosome. Although either chlorination or nitration of tyrosine 66 in GFP ablates its fluorescence (27), nitration does not occur in the phagosome of human PMN (43, 44). Consequently the bleaching of ingested GFP-SA by PMN likely reflects chlorination of GFP and thus provides a probe of MPO-dependent chlorination in an individual phagosome in real time.
The application of GFP bleaching as a probe of the antimicrobial activity of PMN-derived HOCl requires recognition of HOCl chemical reactivity within the complex biological context of the neutrophil phagosome. In general, HOCl-mediated modifications of biological targets depend on 1) sufficient production of HOCl, 2) access of HOCl to potential targets, and 3) the relative susceptibility of substrates to modification. When exposed to HOCl or the MPO-H2O2-chloride system, recombinant GFP in solution is rapidly bleached in a dose-dependent fashion, with <20 μM HOCl causing complete loss of GFP fluorescence by direct oxidative bleaching of the chromophore (4). In contrast, our data (Fig. 3, A and B) and previous studies demonstrate that bleaching of GFP when expressed in bacterial cytoplasm required higher concentrations of HOCl or more time, when HOCl is provided by the MPO-H2O2-chloride system. Because phagocytosis rapidly triggers production of HOCl by human PMN (3) and bleaching of GFP-SA occurred in PMN, we concluded that sufficient HOCl was generated in phagosomes to support bleaching of GFP and that the observed delay in GFP bleaching more likely reflected the reaction properties of HOCl rather than insufficient HOCl.
In contrast to the situation with recombinant GFP, GFP-SA present a more complex and varied substrates for reaction with HOCl. HOCl readily reacts with surface structures when bacteria are exposed to HOCl extracellularly, as discussed in detail by Hurst (reviewed in Ref. 45). The initial focus of microbial attack by HOCl is on the bacterial envelope. Consequently, HOCl will encounter and react with many susceptible targets on the surface of GFP-SA, thus consuming some of the HOCl before it can breach the bacterial surface. Furthermore, any biomolecules in the cytoplasm that are intrinsically very susceptible to HOCl-mediated oxidation are affected later and only after concentrations of HOCl in great excess of that required for killing are added (45-47). The competition with GFP for reacting with HOCl is even more complex in the phagosomes of neutrophils, where the granule proteins dominate and provide ample substrate for reactions with HOCl and other oxidants. In fact, most of chlorotyrosine residues recovered from neutrophils that ingest S. aureus are of host origin (44).
Even if HOCl rapidly accessed the bacterial cytoplasm, biomolecules in the cytoplasm would compete with GFP for reacting for HOCl, depending on their relative intrinsic vulnerability for oxidation by HOCl. Sulfur-containing amino acids (notably, cysteine and methionine) are roughly a million-fold more susceptible to reaction with HOCl than is tyrosine (~3 × 107 M−1 s−1 for methionine or cysteine vs 44 M−1 s−1 for tyrosine) (28, 29) and it is thus likely that HOCl would react with such targets in S. aureus before bleaching GFP. As a consequence of these complexities and the competition between GFP and more vulnerable substrates, the relative kinetics of bleaching and killing of GFP-SA differed (Fig. 1B). Analogous findings were reported by Palazzo et al. (4), where a similar delay in the bleaching of GFP-E. coli after additions of HOCl were noted and where ~10-fold more HOCl was required to bleach cytosolic GFP than was needed to kill the GFP-E. coli (4). Very similar observations regarding the relative rates of bacterial killing vs oxidative damage by HOCl have been reported for several bacterial targets, including a variety of sulfhydryl groups, b-cytochromes, β-carotene, iron-sulfur centers in succinate dehydrogenase, β-galactosidase, aldolase, and F1-ATPase (45, 47-53). Collectively, the survival curves of bacteria exposed to HOCl or to the MPO-H2O2-chloride system reflect the sequential attack of HOCl on microbial targets. Vulnerable but nonessential surface sites undergo oxidation and N-chlorination first, consuming HOCl but not compromising microbial viability. Subsequently, HOCl attacks and inactivates membrane proteins essential for energy generation, thereby culminating in critical damage and microbial death (reviewed in Ref. 54). Ultimately, HOCl will react with susceptible cytoplasmic elements, as seen with the GFP in our GFP-SA; the close correlation between the percent of SA that retained GFP fluorescence and the number of surviving intracellular SA recovered as CFU strongly suggest that the bleaching of GFP after 120 min corresponded to dying or dead bacteria.
One of the most important features of assessing GFP bleaching during and after phagocytosis is the insight it provides on the fate of individual cocci within PMN. Our results provide compelling evidence for differences in the fate of individual bacteria, in some circumstances even within a single neutrophil (Fig. 1A). Several previous studies have strongly suggested that the killing of intraphagosomal staphylococci by PMN is incomplete (5-7, 33, 35, 55, 56) and clinical observations demonstrate the survival of a subpopulation of ingested staphylococci that persist in infected patients (57). The very close correlation between the number of SA surviving within PMN and the fraction of intraphagosomal bacteria retaining GFP fluorescence supports the finding that some ingested SA can survive within PMN and raises the possibility that the differential retention of GFP fluorescence could be used to sort and separately monitor the subsequent fates of living and dead SA within PMN.
The host and microbial variables that contribute to the eventual fate of ingested SA in PMN are likely extremely complex, but our studies point to three important factors: 1) the growth phase of the staphylococci, 2) the MOI and the resultant number of organisms ingested by each PMN, and 3) the amount of H2O2 generated in an individual phagosome. The relative resistance of ingested GFP-SA to killing and bleaching paralleled the duration bacteria were in culture before being fed to PMN. The rank order of resistance to GFP bleaching, overnight cultures > mid-log phase cultures > repetitive consecutive early mid-log cultures (Fig. 4), suggests that slowly or non growing SA were more resistant to PMN cytotoxicity than were those growing more rapidly. This differential susceptibility of the array of potential competitive substrates to cytotoxins in the phagosome due to growth rate-dependent changes in the bacterial envelope the profile of secreted microbial products, or both. Independent of other bacterial variables, higher MOI and the associated greater intracellular bacterial loads also compromised the effectiveness of PMN antimicrobial action. By monitoring GFP bleaching as a function of MOI and the corresponding number of ingested and bleached bacteria within individual PMN, we clearly demonstrated that it is the number of SA per PMN that determined the likelihood of GFP bleaching and microbial killing. This relationship was the same at each MOI tested (Fig. 5A); the higher MOI simply increased the probability of greater bacterial loads within individual PMN and thus favored conditions in which intraphagosomal cytotoxins were insufficient to more extensively damaged SA. Under these taxing conditions, it is possible that one or more antimicrobial component in the phagosome becomes limiting. The increase in iodination proportional to the more robust bacterial uptake at higher MOI (Fig. 5B) indicates the presence of sufficient MPO to support MPO-dependent cytotoxicity, a conclusion supported by the failure of added MPO to augment iodination, bleaching, or killing. In contrast, supplementation with exogenous H2O2, especially at higher MOI (Fig. 5C), reduced the percent of ingested GFP-SA that resisted bleaching, suggesting that the increased phagocytic burden taxed the capacity of the NADPH oxidase to contribute, directly or indirectly, to antimicrobial action in the phagosome independent of the MPO system. Furthermore, decreased bleaching of GFP-SA at higher MOI even in the presence of supplemental H2O2 suggests that additional factors in the phagosome may be somewhat limiting in the face of increased substrate load. Among the possible contributing factors are antibacterial peptides and proteins whose mobilization from preformed granules to individual phagosomes may be limiting at high bacterial loads. However, the failure of GFP-SA to lose fluorescence in the absence of oxidants, as in DPI-PMN, or active MPO, as in NaN3-PMN or MPO-deficient PMN, indicates that granule proteins alone were unable to kill and bleach GFP-SA. The presence of bleached and fluorescent GFP-SA in neighboring phagosomes within the same neutrophil underscores the variability in the potential fate of ingested SA, whether as a reflection of the individual organism’s metabolic state or due to stochastic differences in mobilization of granule proteins and oxidase assembly at a given phagosome. An important clinical implication of these findings is that the bacterial metabolic state and relatively small differences in bacterial inoculum in relation to the number of PMN recruited could dictate significant differences in the outcome of interactions between SA and PMN, possibly helping to explain the practical challenges often seen in resolving infections with S. aureus.
In summary, the specific bleaching of GFP expressed within staphylococci demonstrates that oxidants generated by the NADPH oxidase and that chlorination mediated by MPO collectively contribute to the potent antimicrobial activity within the PMN phagosome. The loss of GFP fluorescence by ingested SA provides a sensitive experimental probe for monitoring biochemical events within individual phagosomes and for identifying the subpopulation of organisms that resist PMN cytotoxins. Defining the molecular basis of SA survival within PMN will provide important insights into the pathophysiology of staphylococcal infection, the microbial attributes that contribute to the clinical challenges that staphylococci pose, and the development of novel therapeutic agents (58).
We thank Sally McCormick for technical assistance with some aspects of these studies.
1The work was supported by National Institutes of Health Grants AI 18571 (to J.P.W.) and AI 070958 (to W.M.N.), by a Merit Review Grant (to W.M.N.) from the Veterans Administration, and by facilities and resources at the Veterans Administration in Iowa City, IA (to W.M.N.).
The authors have no financial conflict of interest.
3Abbreviations used in this paper: