Cleavage of Complement C3b to iC3b on the Surface of Staphylococcus aureus Is Mediated by Serum Complement Factor I

Abstract

Complement-mediated opsonization of Staphylococcus aureus bearing the dominant capsule serotypes, serotypes 5 and 8, remains incompletely understood. We have previously shown that complement plays a vital role in the efficient phagocytosis of a serotype 5 S. aureus strain and that the opsonic fragments of the central complement protein C3, C3b and iC3b, were present on the bacterial surface after incubation in human serum. In the present studies, C3b and iC3b were found on several serotype 5 and 8 S. aureus strains after incubation in human serum. Using purified classical activation pathway complement proteins and the Western blot assay, we showed that when C3b was generated on the S. aureus surface no iC3b fragments were found, suggesting that other serum proteins may be required for cleaving C3b to iC3b. When C3b-coated S. aureus was incubated with serum factor I, a complement regulatory protein, iC3b was generated. Purified factor H, a serum protein cofactor for factor I, did not enhance factor I-mediated cleavage of C3b. These findings suggest that C3b cleavage to iC3b on S. aureus is mediated by serum factor I and does not require factor H.

Staphylococcus aureus remains one of the most common bacterial causes of severe human infections in both hospitals and communities (5, 9). Of considerable concern is this pathogen's ever-increasing resistance to antibiotics (4, 12, 18, 23). New antimicrobial therapies beyond the scope of traditional antibiotics may need to play a role in the treatment of S. aureus infections in the future. Therefore, a better understanding of the interactions between S. aureus and host defenses may help to identify novel therapies for infections caused by this dangerous pathogen.

The complement system is crucial for the opsonization of many bacterial pathogens to optimize their ingestion and subsequent destruction by phagocytes (reviewed in reference 10). Most S. aureus infections are caused by polysaccharide-encapsulated organisms of type 5 or type 8 (2, 20, 24). We have previously shown that complement depletion in mice increases the lethality of bacteremia caused by a capsule polysaccharide type 5 S. aureus strain (6). We have also shown that complement-mediated opsonization greatly increases the phagocytosis efficiency of a capsule polysaccharide type 5 S. aureus isolate compared to that of an isolate opsonized in normal human serum without complement activation (7).

It has been previously shown that S. aureus isolates opsonized in normal human serum will have considerable quantities of C3b and iC3b, fragments of the central opsonic complement protein C3, bound to their surfaces (6, 13). Although both C3b and iC3b function as opsonins, they act through different complement receptors, complement receptor 1 (CD35) and complement receptor 3 (CD11b/CD18), respectively (reviewed in reference 1). C3b can initiate the alternative complement pathway, depositing more opsonic C3 fragments on the target, and can activate the terminal complement cascade. However, once C3b has been cleaved to iC3b, neither alternative pathway initiation nor terminal complement cascade activation occurs (reviewed in reference 19). CD35/CD21-deficient mice, lacking complement receptors 1 and 2, experienced greatly increased lethality due to S. aureus bacteremia compared to isogenic mice with sufficient CD35/CD21 (unpublished data). This finding suggests that the interaction between S. aureus-bound C3b and CD35 may be critical for immune protection. We considered whether the cleavage of C3b to the iC3b form might inhibit optimal host defense against S. aureus by preventing the initiation of the alternative complement pathway and decreasing binding with CD35. If this process is a mechanism for S. aureus virulence, the process underlying C3b cleavage to iC3b on the S. aureus surface might be a target for therapeutic intervention.

Because of the potentially deleterious effects of uncontrolled complement activation, several complement down-regulatory proteins, including complement factor H and factor I, are ubiquitously present in serum (reviewed in reference 17). Factor H can bind to C3b and serve as a cofactor with factor I to cleave C3b to iC3b (11). Factor H can also accelerate the decay of the C3Bb convertase of the alternative complement pathway, resulting in decreased activation of this pathway (reviewed in reference 16). Notably, two other gram-positive bacteria, Streptococcus pyogenes and Streptococcus pneumoniae, are both able to bind complement factor H (8, 14, 15, 21, 22). We hypothesized that S. aureus might convert bound C3b to iC3b via similar interactions with factor H and factor I.

MATERIALS AND METHODS

Bacterial strains and growth.

S. aureus strains bearing capsule polysaccharide, including type 5 capsule strains Reynolds, Newman, and Lowenstein and type 8 capsule strains Becker and Wright, were used. Bacteria were grown on Columbia 2% NaCl agar plates overnight at 37°C. On the day of the experiment, two colonies from the overnight cultures were inoculated into Columbia 2% NaCl broth and incubated at 37°C with shaking to the mid-logarithmic phase of growth, achieved after incubation in broth for approximately 2 h and confirmed by optical densitometry at 600 nm. In the mid-logarithmic phase of growth, minimal capsule is expressed by strains Reynolds and Lowenstein (6).

Complement buffers.

Complement activation and C3 binding experiments were performed with GVBS⁺⁺ buffer (Veronal-buffered saline [VBS] with 0.1% gelatin, 0.15 mM CaCl₂, and 1.0 mM MgCl₂). Complement activation was inhibited by EDTA-GVBS⁻⁻ buffer (VBS with 0.1% gelatin and 0.01 M EDTA). Purified complement component activation incubations were performed with 60% DGVBS⁺⁺ buffer (60% VBS with 3% dextrose, 0.1% gelatin, 0.15 mM CaCl₂, and 1.0 mM MgCl₂).

Complement and immunoglobulin sources.

Healthy human volunteers donated blood under an Institutional Review Board-approved protocol (Eastern Virginia Medical School IRB no. 02-06-EX-0216). Donated blood was obtained in Vacutainer (Becton Dickinson, Franklin Lakes, N.J.) tubes without additives. The blood was allowed to clot at room temperature for 1 h and on ice for 2 h. The clot was then sedimented to produce normal human serum (NHS), which was recovered and stored at −80°C. All sera used had normal total hemolytic complement levels (CH50). C8-depleted human serum was obtained commercially (Advanced Research Technologies Inc., San Diego, Calif.). The immunoglobulin G (IgG) used for sensitizing S. aureus for classical complement pathway activation was Gamimune N (Miles Inc., Elkhart, Ind.). Purified human complement proteins C1, C4, C2, and C3; factor H; factor I; and polyclonal antibodies directed against human C3, human factor H, and human factor I were purchased from a commercial source (Advanced Research Technologies). The purity of factor H and factor I was confirmed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and total protein staining with Sypro Ruby (Molecular Probes Inc., Eugene, Oreg.). The functional activity of factor H as a cofactor for factor I-mediated cleavage of C3b to iC3b was demonstrated by using C3b-coated bovine serum albumin (BSA) beads.

Opsonization with serum.

Bacteria grown to the mid-logarithmic phase were washed with GVBS⁺⁺ buffer and brought to a standard concentration by using optical densitometry at 600 nm. Bacteria (5 × 10⁷ organisms) were incubated in 10% NHS in GVBS⁺⁺ buffer in a total volume of 0.5 ml at 37°C for 30 min unless otherwise indicated. Washed antibody-sensitized sheep erythrocytes (EA) were incubated with 10% C8-depleted serum in GVBS⁺⁺ at 37°C for the times indicated to allow opsonization with C3 without lysis.

Opsonization with purified complement components.

Washed bacteria (1 × 10⁹ organisms in GVBS⁺⁺ buffer) were incubated with 0.5% intravenous immune globulin for 30 min at 4°C. These antibody-incubated bacteria were then incubated with the purified human complement proteins C1, C4, C2, and C3. C1 incubation was performed with 2.5 μg of C1/ml in 60% DGVBS⁺⁺ buffer at 30°C for 15 min. C4 incubation was performed with 6.6 μg of C4/ml in 60% DGVBS⁺⁺ buffer at 37°C for 45 min. C2 and C3 incubation was performed with C2 and C3 together, with 1.8 μg of C2/ml and 2 μg of C3/ml in 60% DGVBS⁺⁺ buffer at 30°C for the times indicated.

C3b cleavage to iC3b.

Bacteria incubated with C1, C4, C2, and C3 were washed and then incubated with factor H (40 μg/ml), factor I (4 μg/ml), or both in EDTA-GVBS⁻⁻ buffer for 60 min at 37°C. The concentrations of factor H and factor I used were 10% of the concentrations found in undiluted serum. After they were washed twice in EDTA-GVBS⁻⁻ buffer, the bacteria (1 × 10⁹ CFU) were incubated in 0.25 ml of 25 mM methylamine for 60 min at 37°C to release C3 fragments bound by ester bonds and centrifuged to remove solid particles. In some experiments, washed opsonized bacteria (1 × 10⁹ CFU) were allowed to spontaneously release C3 fragments from their surfaces during a 60-min incubation at 37°C in 0.25 ml of EDTA-GVBS⁻⁻ buffer. Samples were analyzed by SDS-PAGE and Western blot analysis or by staining of the SDS-PAGE gel with Sypro Ruby (Molecular Probes). Sypro Ruby is a semiquantitative total protein stain (3). Optical densitometry measurements of the gel bands were performed by calculating the area under the intensity profile curve with Quantity One software (Bio-Rad Laboratories, Hercules, Calif.). Western blot analysis was performed with polyclonal goat anti-human C3 antibody (Advanced Research Technologies) and horseradish peroxidase-labeled rabbit anti-goat antibody (Sigma-Aldrich, St. Louis, Mo.). Bound antibody was then detected by using enhanced chemiluminescence (Amersham Co., Arlington Heights, Ill.).

Factor I binding to S. aureus.

Factor I binding to S. aureus Reynolds was tested in an environment of excess quantities of BSA. Factor I (0.01 mg) was incubated in 1 × 10⁸ CFU of S. aureus Reynolds in EDTA-GVBS⁻⁻ buffer for 60 min at 37°C with 0 to 10 mg of BSA. Bacteria were then washed three times in EDTA-GVBS⁻⁻ buffer, boiled for 5 min in SDS sample buffer, and sedimented to remove solid particles. Supernatants were analyzed by SDS-PAGE and Western blot analysis with anti-factor I antibody and by total protein staining of the SDS-PAGE gel with Sypro Ruby.

Factor H binding to S. aureus.

Factor H binding to mid-logarithmic-phase S. aureus Reynolds was tested with various amounts of factor H to evaluate the dose-response relationship. Various quantities of factor H were incubated with 2.5 × 10⁸ CFU of bacteria for 60 min at 37°C in EDTA-GVBS⁻⁻ buffer. Bacteria were washed three times in EDTA-GVBS⁻⁻ buffer and then boiled for 5 min in SDS sample buffer and centrifuged to remove solid particles. Supernatants were analyzed by SDS-PAGE and Western blot analysis and by factor H enzyme-linked immunosorbent assay (ELISA).

Factor H binding to S. aureus Reynolds was also tested in an environment of excess quantities of BSA. Factor H (0.005 mg) was incubated in 1 × 10⁸ CFU of S. aureus Reynolds in EDTA-GVBS⁻⁻ buffer for 60 min at 37°C with 0 to 5 mg of BSA. Bacteria were washed three times in EDTA-GVBS⁻⁻ buffer, boiled for 5 min in SDS sample buffer, and sedimented to remove solid particles. Supernatants were analyzed by SDS-PAGE and Western blot analysis and by factor H ELISA.

Factor H measurement by ELISA.

Flat-bottom Immulon 2 plates were coated with mouse anti-human factor H (Serotec Ltd., Oxford, United Kingdom) at 20 μg/ml in a carbonate coating buffer overnight at 4°C. At the time of use, plates were washed three times with phosphate-buffered saline (PBS)-Tween (0.5%) and blocked with 3% BSA in PBS-Tween overnight at 4°C. Plates were washed three times with PBS-Tween, and then factor H and the samples were diluted in 3% BSA in PBS-Tween and allowed to incubate in the plates for 2 h at 25°C. Plates were washed three times with PBS-Tween, and then goat anti-factor H antibody was diluted to 200 μg/ml and allowed to incubate in the plates for 1 h at 25°C. Plates were washed three times with PBS-Tween, and then rabbit anti-goat horseradish peroxidase antibody (15 μg/ml) was added to each well and allowed to incubate for 1 h at 25°C. Plates were washed three times with PBS-Tween and developed with TMB Plus (Accurate Chemical, Westbury, N.Y.), reactions were stopped with 2.5 N H₂SO₄, and results were read at 450 nm.

Statistical analysis.

Results of replicate experiments were averaged, and means ± standard errors were calculated. Comparisons were made by Student's t test with P values of <0.05 being considered statistically significant.

RESULTS

C3 fragments bound to multiple virulent-serotype S. aureus strains.

Three capsule serotype 5 and two capsule serotype 8 S. aureus strains were incubated in 10% NHS and analyzed by Western blot analysis with polyclonal anti-C3 antibody (Fig. 1). All strains bound C3b and iC3b on their surfaces. These findings show that virulent S. aureus serotypes 5 and 8 activated complement in NHS with the result that both C3b and iC3b bound to their surfaces.

FIG. 1.

Deposited C3 fragments on the surfaces of capsule type 5 S. aureus (strains Reynolds, Newman, and Lowenstein) and capsule type 8 S. aureus (strains Becker and Wright) incubated in 10% NHS. Western blot analysis with polyclonal anti-C3 antibody showed that C3b (104 and 75 kDa) and iC3b (75, 63, and 42 kDa) were present for all strains.

C3 fragments deposited on S. aureus vary over time.

S. aureus Reynolds was incubated in 10% NHS for intervals of 0 to 30 min and analyzed by Western blot analysis (Fig. 2). The classical complement model of EA was incubated in C8-depleted serum (to prevent lysis) for comparison. EA showed deposition of C3b and iC3b by 5 min, but by 20 min, iC3 was present and nearly all C3b had vanished. This result likely represented cleavage of the bound C3b to iC3b by factor I in the presence of a cofactor like CD35 or factor H. In comparison, S. aureus Reynolds also showed C3b and iC3b bound to the surface by 5 min, but both forms were still present on the S. aureus surface at 30 min. This suggested that C3b was cleaved rapidly to iC3b on the S. aureus surface but that the cleavage did not proceed to completion like it did for EA.

FIG. 2.

Comparison of C3 fragments deposited on the surfaces of EA incubated in C8-depleted serum and C3 fragments on S. aureus Reynolds after incubation in 10% NHS. Western blot analysis with polyclonal anti-C3 antibody showed that C3b (104 and 75 kDa) and iC3b (75, 63, and 42 kDa) were present on EA after 5 min and that the remaining C3b was cleaved to iC3b by 20 min. C3b and iC3b were present on S. aureus Reynolds at all time points after 5 min.

C3 fragments bound to S. aureus by purified classical pathway complement components.

S. aureus Reynolds was coated with pooled polyclonal human IgG, incubated with C1, C4, C2, and C3 to form the classical pathway C3 convertase and activate C3 fragment binding, and then analyzed by Western blot analysis (Fig. 3). C3b was present on the surface of S. aureus Reynolds at 30, 60, and 90 min, but no iC3b was observed. In comparison, EA incubated in C1, C2, C4, and C3 showed only C3b, as would be expected in the absence of factor I. These findings demonstrate that when the classical complement pathway was activated, C3b was deposited on the S. aureus surface and that, in buffer alone, no cleavage to iC3b forms occurred.

FIG. 3.

C3 fragments on the surfaces of EA and S. aureus Reynolds after incubation with antibody and purified complement components C1, C4, C2, and C3 (C1423). Western blot analysis with polyclonal anti-C3 antibody showed that in the absence of other serum factors only C3b (104 and 75 kDa) was bound to the surface of EA and S. aureus Reynolds. Even for incubations of up to 90 min, no iC3b was found on the S. aureus surface. S. aureus Reynolds incubated in 10% NHS is shown for comparison.

C3b is cleaved to iC3b on S. aureus in the presence of factor H and factor I.

S. aureus Reynolds was coated with IgG and then combined with purified complement components C1, C4, C2, and C3 to activate the classical complement pathway and bind C3b. The C3b-coated S. aureus Reynolds was then incubated with factor H, factor I, or both and analyzed by Western blotting for bound C3 fragments on the S. aureus surface (Fig. 4A) and C3 fragments released from the bacterial surface (Fig. 4B). The resulting blots suggest that incubating C3b-coated S. aureus in factor I or in factors H and I results in decreased C3b and increased iC3b generation. Total protein staining of the SDS-PAGE gel with Sypro Ruby (Fig. 5A) allowed quantitation by optical densitometry (Fig. 5B). Incubation in factor I decreased the relative amount of C3b (P = 0.002) and increased the relative amount of iC3b (P = 0.002) compared to the amounts found after incubation in neutral buffer. Incubation in factor H and factor I together decreased the relative amount of C3b (P = 0.021) and increased the amount of iC3b (P = 0.047) compared to the amounts found after incubation in neutral buffer. Incubation in factor H alone did not decrease the amount of C3b (P = 0.36) or increase the amount of iC3b (P = 0.07) compared to the amounts found after incubation in neutral buffer. The presence of iC3b in the absence of factor H and factor I represents the background staining of the gel by Sypro Ruby. No specific iC3b band was visible in the S. aureus-only lane.

FIG. 4.

C3 fragments present on the surface of S. aureus Reynolds after incubation with antibody and purified complement factors C1, C4, C2, and C3 (C1423). Some fractions were then incubated with factor H, factor I, or both (H&I). Surface-bound C3 fragments were released by incubation with methylamine (A) or allowed to spontaneously shed from the S. aureus surface (B). Western blot analysis with polyclonal anti-C3 antibody showed that incubation in factor I or in factors H and I together was associated with bound and shed C3b (104 and 75 kDa) and iC3b (75, 63, and 42 kDa). Purified C3b and iC3b are shown for comparison.

FIG. 5.

C3 fragments shed from the surface of S. aureus Reynolds after incubation with antibody and purified complement factors C1, C4, C2, and C3 (C1243 in panel A, C1423 in panel B). Some fractions were then incubated with factor H, factor I, or both (H&I). C3 fragments were allowed to spontaneously shed from the S. aureus surface. (A) Total protein staining of the SDS-PAGE gel showed that incubation in factor I or in factors H and I together was associated with decreased C3b (104 and 75 kDa) and increased iC3b (75, 63, and 42 kDa). (B) Optical densitometry measurements of C3b (104-kDa band) and iC3b (42-kDa band) were calculated as a ratio of the 75-kDa band. Less C3b and more iC3b were present after incubation in factor I (P = 0.002 and P = 0.002, respectively) or in factors H and I together (P = 0.022 and P = 0.047, respectively) than after incubation in buffer alone. Data are means of results from three independent experiments. Error bars denote standard errors of the means.

Binding of factor I to S. aureus.

Increasing concentrations of BSA were examined for their effects on factor I binding to S. aureus. In the absence of BSA, factor I was associated with S. aureus. With the addition of increasing amounts of BSA, the factor I signal diminished but was shown to persist both by Western blot analysis (Fig. 6A) and by total protein staining with Sypro Ruby (Fig. 6B). Optical densitometry performed with the total protein stained gel showed that the amount of factor I bound to S. aureus was 33% in the presence of a 1,000-fold excess of BSA compared to the amount of factor I bound in the absence of BSA.

FIG. 6.

Factor I (Fctr I) bound to mid-logarithmic-phase S. aureus Reynolds incubated in various concentrations of BSA (relative to factor I). Western blot analysis with anti-factor I antibody (A) and total protein staining of the SDS-PAGE gel (B) showed that the binding of factor I to S. aureus Reynolds was diminished by concentrations of BSA in vast excess of factor I. S. aureus Reynolds was incubated in factor I and BSA in EDTA-GVBS⁻⁻ buffer for 60 min at 37°C.

Binding of factor H to S. aureus.

Various concentrations of purified factor H were incubated with S. aureus Reynolds and examined for binding to the organisms by Western blotting (Fig. 7A) and by ELISA (Fig. 7B). Incubation in higher concentrations of factor H was associated with increased signal in Western blots and increased factor H binding in the ELISA. These findings indicate a dose-response relationship between the amount of factor H exposed to S. aureus and the amount of factor H bound to S. aureus.

FIG. 7.

Various concentrations of factor H bound to mid-logarithmic-phase S. aureus Reynolds. Western blot analysis with anti-factor H antibody (A) and factor H ELISA (B) showed that factor H binding to S. aureus Reynolds increased with incubation in higher concentrations of factor H in a dose-response fashion. S. aureus Reynolds was incubated in different quantities of factor H in EDTA-GVBS⁻⁻ buffer for 60 min at 37°C. Data are means of results from at least three independent experiments. Error bars denote standard errors of the means.

Various concentrations of BSA in excess of that of factor H were incubated with S. aureus strain Reynolds to determine by ELISA whether BSA could competitively inhibit factor H binding (Fig. 8). The amount of factor H bound to S. aureus in the presence of a 1,000-fold excess of BSA was 3% that of factor H bound to S. aureus in the absence of BSA.

FIG. 8.

Purified factor H bound to mid-logarithmic-phase S. aureus Reynolds incubated in various concentrations of BSA. Factor H ELISA showed that factor H binding to S. aureus decreased in the presence of excess BSA. S. aureus Reynolds was incubated in factor H and BSA for 60 min at 37°C. Each datum point represents the mean of results from three independent experiments. Error bars denote standard errors of the means.

DISCUSSION

Complement is a critical element in host defense against many types of microorganisms, but the interactions between encapsulated S. aureus and complement proteins remain poorly understood. We hypothesized that C3b cleavage to iC3b might occur on the surface of S. aureus, thus providing the organism with a method of immune evasion due to the fact that these C3 fragments have different affinities for complement receptors. Additionally, iC3b cannot form the alternative pathway C3 convertase and iC3b cannot form the C5 convertase required to activate the terminal complement cascade (reviewed in reference 17).

Incubation of S. aureus in human serum led to C3b and iC3b binding to the surface of virulent capsule polysaccharide type 5 and type 8 S. aureus strains. These two capsule serotypes account for the majority of S. aureus infections in humans, indicating that cleavage of C3b to iC3b occurs on the surfaces of virulent S. aureus serotypes.

C3b cleavage to iC3b occurred quickly on the surface of S. aureus in human serum, similar to the iC3b generation observed on EA. Unlike the case for EA, in which all C3b had been cleaved to iC3b after 30 min in serum, iC3b and C3b were bound to the S. aureus surface after 30 min in serum. Whether this result is due to ongoing C3 deposition over time or a lower efficiency of the process by which C3b cleavage to iC3b occurs on the S. aureus surface remains unclear.

Unlike serum-mediated complement activation, the building of the classical activation pathway with purified complement proteins on the S. aureus surface via incubation with IgG, C1, C4, C2, and C3 did not cause any iC3b generation. This result suggests that serum proteins other than complement components are needed to cleave C3b to iC3b. We hypothesized that the serum complement regulatory proteins factor H and factor I could contribute to the cleavage of C3b to iC3b on the S. aureus surface because other gram-positive bacteria, Streptococcus pyogenes and Streptococcus pneumoniae, bind factor H (8, 14, 15, 21, 22), a cofactor for factor I-mediated cleavage of C3b to iC3b.

When the classical activation pathway was formed with purified complement proteins to generate C3b binding to S. aureus, factor I was able to cleave C3b to iC3b. In other complement models, factor I-mediated cleavage of C3b to iC3b requires the activity of a cofactor. In humans, these cofactors are the serum protein factor H and the cell-surface proteins CD35, membrane cofactor protein, and C4-binding protein (reviewed in reference 16). Cleavage of C3b to iC3b on the S. aureus surface in the presence of factor I but in the absence of a known cofactor for factor I, as shown in these experiments, is a novel observation suggesting that factor I is enzymatically active on the surface of S. aureus.

Next, we examined the specificity of purified factor I binding to S. aureus by using BSA as a nonspecific competitive inhibitor. Factor I binding to mid-logarithmic-phase S. aureus was diminished in the presence of a BSA excess, suggesting that factor I binding to S. aureus may be nonspecific.

We speculated that S. aureus might bind factor H, as was shown for Streptococcus pyogenes and Streptococcus pneumoniae. Dose-response testing for factor H by Western blot analysis and by ELISA confirmed that S. aureus incubated with increasing quantities of factor H bound larger amounts of factor H to their surfaces. However, factor H binding to S. aureus was much less efficient in the presence of excess BSA. Thus, S. aureus appears to bind factor H in a nonspecific manner.

In summary, the opsonic complement C3 fragments C3b and iC3b appear on the surfaces of several virulent S. aureus strains of capsule polysaccharide serotypes 5 and 8. The C3b deposited on the S. aureus surface is not cleaved to iC3b in the absence of serum proteins. Factor I mediates the cleavage of C3b to iC3b on the surface of S. aureus and appears to be able to function without the serum cofactor, factor H. Thus, the process of factor I binding is a potential target for therapeutic intervention to facilitate host defense against S. aureus.