Identification of OprF as a Complement Component C3 Binding Acceptor Molecule on the Surface of Pseudomonas aeruginosa

Abstract

Pseudomonas aeruginosa is a versatile opportunistic pathogen that can cause devastating persistent infections. Complement is a highly conserved pathway of the innate immune system, and its role in the first line of defense against pathogens is widely appreciated. One of the earliest events in the complement cascade is the conversion of C3 to C3a and C3b, the latter typically binds to one or more acceptor molecules on the pathogen surface. We previously demonstrated that complement C3b binding acceptors exist on the P. aeruginosa surface. In the current study, we utilized either C3 polyclonal or C3b monoclonal antibodies in a far-Western technique followed by mass spectroscopy to identify the C3b acceptor molecule(s) on the P. aeruginosa surface. Our data provide evidence that OprF (an outer membrane porin, highly conserved in the Pseudomonadaceae) binds C3b. An oprF-deficient P. aeruginosa strain exhibits reduced C3 deposition compared to the wild type. We observed reduced internalization of oprF-deficient bacteria by neutrophils after opsonization compared with wild-type P. aeruginosa. Heterologous expression of OprF significantly enhanced C3b binding and increased serum-mediated bactericidal effects in complement-susceptible Escherichia coli. Furthermore, the predicted secondary structure of the C-terminal, surface-exposed region of OprF has high structural identity to the OmpA domain of several other Gram-negative bacteria, one of which is known to bind C3b. Therefore, these findings provide new insights into the biology of complement interactions with P. aeruginosa and other Gram-negative bacteria.

INTRODUCTION

Pseudomonas aeruginosa is an opportunistic nosocomial human pathogen that causes a wide range of clinical symptoms and infections in immunocompromised patients. This bacterium is often the causative agent of acute and chronic infections of the airway, sepsis, burn wounds, and skin infections. Excessive infiltration of neutrophils at the site of infection and the presence of complement have been observed in inflamed tissue and blood (1, 2). Serum sensitivity is directly dependent upon recognition of the bacterial surface by antibodies and complement (3). Prior studies also suggested that complement binds to the P. aeruginosa surface and enhances phagocytosis by acting as an opsonin (4). In this study, we sought to understand the consequences of complement interaction with P. aeruginosa. C3 is the central component of the complement system, playing a crucial role in the three major complement component activation processes. C3b, a product of C3 activation, binds covalently to bacterial surfaces through either a reactive intramolecular ester or an amide bond (5). C3b also participates in the formation of C5 convertase that cleaves C5 and initiates the sequential deposition of the remaining complement components to form the C5b-9 membrane attack complex (MAC), which may result in direct killing of microorganisms (6).

The human complement system plays an important role in the clearance of early pulmonary P. aeruginosa infections (7). Compared to wild-type mice, complement-deficient mice, when challenged with P. aeruginosa, showed an amplified inflammatory response, had more robust colonization, and were unable to clear the infection (8, 9). Previous results showed that neutropenic mice are highly susceptible to P. aeruginosa infection (10, 11). Resistance of P. aeruginosa to killing by complement and evasion from neutrophils are major virulence traits that allow this organism to survive within the bloodstream and inflamed lung. To date, the actual mechanism for complement interaction on the P. aeruginosa surface is not well understood.

Our previous study showed that normal human serum significantly enhanced the oxidative burst response of neutrophils and that C3b is a critical opsonin for P. aeruginosa (4). We also found that the presence of the Psl polysaccharide partially masks the interaction of C3b with the bacterial surface (4). C3b has the potential to interact with several soluble and membrane bound bacterial proteins. Thus, identification of P. aeruginosa binding acceptor(s) for C3b is a goal of this study. We provide evidence that OprF serves as an acceptor for human complement C3b. Our data demonstrate that (i) OprF-mediated binding to C3b increases P. aeruginosa interactions with human neutrophils and (ii) interactions between C3b and Escherichia coli-expressing P. aeruginosa OprF trigger downstream events of the complement cascade, which enhance complement-mediated killing.

OprF is a major outer membrane porin that is involved in several crucial functions, including maintenance of cell structure, outer membrane permeability, environmental sensing, adhesion, biofilm formation, and virulence (12, 13). OprF also binds to interferon gamma, leading to stimulation of the quorum-sensing network (14). OprF is a highly conserved protein in the Pseudomonadaceae family with several surface antigenic epitopes (15), but its role in complement binding has not been described. OprF is also highly immunogenic and has been used extensively in vaccine production alone or in conjunction with flagellin (16, 17). Our current study evaluated how C3b binding to OprF promotes phagocyte interactions. This may be exploited to enhance opsonophagocytosis and complement-mediated killing.

MATERIALS AND METHODS

Strains, serum complement proteins, and growth media.

An in-frame nonpolar deletion of the oprF gene was constructed utilizing allelic exchange as described previously (18, 19). Wild-type (WT) P. aeruginosa (PAO1) and oprF-, oprD-, oprQ-, oprJ-, and oprI-deficient P. aeruginosa strains (20) were grown in Luria-Bertani broth with no NaCl (LBNS) without antibiotics. Low-copy-number plasmid pHSG576 harboring oprF and vector control (pHSG576 without oprF) were transformed in E. coli K-12 (21). E. coli/pHSG576 (21) was grown at 37°C in Luria-Bertani (LB) medium containing 30 μg/ml chloramphenicol with aeration by rotary shaking. IPTG (isopropyl-β-d-thiogalactopyranoside; 0.1 M) was used to induce OprF expression. No growth defect was observed due to expression of OprF in E. coli. Normal human serum, C1q-depleted serum, and purified human C1q protein were purchased from Complement Technology. Heat-inactivated serum (HIS) was prepared by incubating normal human serum at 56°C for 40 min.

Complement deposition and bacterial killing.

Bacteria grown to mid-log phase were opsonized with 20% fresh human serum, C1q-depleted human serum, C1q-depleted human serum complemented with 150 μg/ml of purified C1q protein, or heat-inactivated human serum at 37°C for 5 or 20 min. Reactions were stopped by the addition of 10 mM EDTA and incubation on ice. Bacteria were washed, fixed, and stained using anti-human C3, C3b, or C5 polyclonal antibodies (Complement Technology) at a 1:1,000 dilution. Following incubation, samples were washed and detected using Alexa Fluor 488-conjugated secondary antibody (Invitrogen) at a 1:500 dilution. After final washes, the pelleted bacteria were analyzed by flow cytometry (FACSCalibur; BD Biosciences) for quantification of complement deposition. Nonopsonized bacteria were used in each experiment to set the gate for analysis. Data from 30,000 gated events were collected, and the mean fluorescence intensity (MFI) was calculated for each sample. For proteinase K treatment, mid-log-phase bacteria were incubated with 1 mg/ml proteinase K (45 mAU/mg, Qiagen) in phosphate-buffered saline (PBS) for 3 h at 37°C. After extensive washing with PBS, C3 deposition was evaluated as above.

For the bacterial killing assays, mid-log-phase wild-type strain PAO1, P. aeruginosa PAO1ΔoprF, OprF-expressing E. coli, and vector control were opsonized with PBS containing 20% pooled normal human serum (Complement Technology), C5-depleted serum, or heat-inactivated serum for 30 min at 37°C with slow agitation. The tubes were placed on ice for 5 min to stop the reaction, bacteria were washed extensively, and 10-fold serial dilutions were plated to determine surviving CFU.

OprF accessibility.

Mid-log-phase grown WT strain PAO1, Δpsl mutant strain, and Psl⁺ strain (Psl-overexpressing PAO1) were incubated with anti-OprF antibody on ice for 1 h and counterstained with anti-mouse antibody conjugated with Alexa Fluor 488. After washing, the pelleted bacteria were resuspended in the PBS and analyzed by flow cytometry (FACS Calibur; BD Biosciences) for accessibility of OprF. Exopolysaccharide extracts were prepared according to the method of Byrd et al., and expression of Psl was determined by immunoblotting as described previously (4, 22).

Far-Western blot analysis.

P. aeruginosa was grown to late log phase, and a membrane preparation was generated as described previously (23). Twenty micrograms of this membrane preparation was suspended in 100 μl of SDS-PAGE sample buffer (50 mM Tris-HCl [pH 6.8], 2% SDS, 100 μM 2-β mercaptoethanol, 10% glycerol, 0.1% bromphenol blue). A far-Western blot analysis was performed as described previously (24). Briefly, proteins were separated under reducing conditions by SDS–10% PAGE and transferred to nitrocellulose membranes (Bio-Rad). After nonspecific binding was blocked with PBS containing 5% milk, the blots were incubated with normal human serum (final dilution of 20% in PBS) for 2 h, washed with PBS, and incubated with anti-human C3b antibody (1:5,000 dilution) in Tris-buffered saline-Tween 20 (TBS-T) supplemented with 0.5% milk. Binding of C3b antibody was detected using α anti-mouse IgG horseradish peroxidase (HRP)-conjugated secondary antibody (Invitrogen Life Technologies). A similar procedure was applied for detection of C3b binding with OprF-expressing E. coli, except that a whole-cell lysate was utilized. E. coli lysates were electrophoresed by the method of Laemmli (25) in 10% SDS–polyacrylamide gels and then either stained with Coomassie blue G-250 (Thermo Scientific) or transferred to a nitrocellulose membrane and processed for far-Western analysis using the anti-C3b antibody as described above.

LC/MS.

The protein band obtained from a stained Coomassie gel was washed and dehydrated with water and 200 mM NH₄HCO₃ in 50% acetonitrile. After dehydration, proteins were digested using trypsin at 37°C overnight. Peptides were extracted from the polyacrylamide with 0.5% formic acid and 50% acetonitrile solution and concentrated for liquid chromatography tandem mass spectrometry (LC/MS) analysis, which was performed on a Thermo Scientific linear trap quadrupole (LTQ) mass spectrometer equipped with a CaptiveSpray source (Bruker Daltonics, Billerica, MA) operated in positive ion mode (according to The Ohio State University [OSU] proteomics core facility). Samples were separated on a capillary column (0.2 by 150 mm; Magic C₁₈ AQ, 3 μm, 200 Å; Michrom Bioresources Inc., Auburn, CA). Protein identifications were based on the Mascot score of 75 or higher with a minimum of two unique peptides from one protein.

Neutrophil isolation and bacterial uptake by neutrophils.

Neutrophil isolation was performed as described previously using an approved institutional review board (IRB) protocol (2009H0314) at The Ohio State University (4). Neutrophils (2 × 10⁵/well) were seeded on poly-l-lysine-coated coverslips, and mid-log-phase P. aeruginosa, opsonized with 20% normal human serum or heat-inactivated serum, was added at a multiplicity of infection (MOI) (bacteria to neutrophil) of 5:1 to the neutrophil monolayers for 30 min at 37°C in 5% CO₂. Cells were washed extensively with Hanks' balanced salt solution (HBSS) to remove nonadherent bacteria, fixed, and stained with anti-Pseudomonas antibody (11) (diluted to 1:1,000) before and after cell membrane permeabilization with methanol (26). The anti-Pseudomonas antibody was raised in rabbit against whole surface antigens of strain PAO1. Visualization of bacteria-phagocyte association with or without permeabilization by confocal microscopy (Olympus FV 1000; 60× oil objective) allowed us to differentiate attached P. aeruginosa cells from internalized bacteria. Attached bacteria exposed to the antibodies before permeabilization stain green due to Alexa Fluor 488, while internalized bacteria exposed to the antibodies after permeabilization stain red due to Alexa Fluor 647 (each diluted 1:500). For quantification, 100 neutrophils were chosen at random, and the cell-associated versus internalized bacteria were counted and averaged by two readers blinded to the treatment conditions. E. coli expressing OprF was also studied for attachment to neutrophils. Bacteria were opsonized with 20% C5-depleted serum and stained with anti-C3 and anti-OprF antibody (27) conjugated with Alexa Fluor 488 and 647, respectively. Blue DAPI (4′,6-diamidino-2-phenylindole) fluorescence was used to stain the nucleus of phagocytes (magnification, ×60). Quantitation of bacterial association with neutrophils was done as described above.

Statistics.

Statistical analysis was carried out using the unpaired two-tailed Student's t test. A P value of ≤0.05 was considered to be statistically significant. The asterisks in figures indicate P values as follows: *, P ≤ 0.05; **, P ≤ 0.005; ***, P ≤ 0.0005.

RESULTS

Identification of acceptor molecule(s) on the P. aeruginosa surface that binds C3b.

Our previous work indicated that there was a C3b acceptor molecule on the P. aeruginosa surface, which facilitates C3-mediated bacterial phagocytosis (4). To investigate the potential acceptor molecule(s), we first determined the biochemical nature of the C3b acceptor(s). Whole cells of wild-type P. aeruginosa PAO1 were treated with proteinase K, washed, and opsonized with normal human serum. Deposited C3b was stained using anti-human C3 antibody and analyzed by flow cytometry. Compared with buffer control, C3b binding was significantly reduced but not eliminated when PAO1 was treated with proteinase K (Fig. 1A). This suggested that the acceptor molecule for C3b binding on the bacterial surface involves one or more protein(s).

FIG 1.

OprF is a potential C3 binding acceptor on the P. aeruginosa surface. (A) Mid-log-phase P. aeruginosa was incubated with proteinase K (PK) at 37°C, washed with PBS, and incubated with 20% normal human serum. C3 binding to the bacterial surface was analyzed by flow cytometry. PK-treated bacteria showed reduced ability to bind to human complement C3b compared to untreated bacteria. (**, P ≤ 0.005, unpaired Student's t test). (B) Far-Western analysis showing C3b binding to P. aeruginosa OprF. P. aeruginosa membrane preparations were separated by SDS gel, transferred to nitrocellulose, renatured, and incubated with buffer (i) or normal human serum as a source of C3 (ii). Both blots were then washed and probed with anti-C3b antibody. ⭑, binding of C3 to OprF in the serum-treated blot at a molecular mass of ∼37 kDa.

To determine which P. aeruginosa proteins bind C3b, a membrane preparation was subjected to far-Western blotting. Whole-membrane fractions of P. aeruginosa PAO1 were loaded onto two parallel gels, transferred onto nitrocellulose, and incubated with PBS (Fig. 1B) or human serum (Fig. 1C). Both blots were probed with anti-human C3b antibody. C3b recognized a polypeptide of ∼37 kDa (Fig. 1C, asterisk), which was excised from a stained gel and processed by mass spectroscopy (see Materials and Methods). No reactive bands were observed when the blot was incubated with PBS (Fig. 1B). Of the peptide sequences retrieved, several matched the outer membrane porin OprF sequence and were highly represented (probability-based Mascot score, >200; a protein score of >75 is significant; P < 0.05). This suggested that OprF might be an acceptor for binding C3b. This has precedence, since Klebsiella pneumoniae (37) and Legionella pneumophila (38) porins also bind C3b.

OprF promotes binding to human C3b.

To investigate the potential role of OprF in C3b binding, we constructed an oprF nonpolar deletion (PAO1ΔoprF) and evaluated C3b binding. PAO1 and PAO1ΔoprF were incubated with normal human serum, washed, stained with anti-C3 or anti-C5 antibodies conjugated to fluorophores, and analyzed by flow cytometry. The results showed a significant reduction, but not elimination, of C3 binding to PAO1ΔoprF compared with that of the wild type (Fig. 2A). This suggests that OprF is a major but likely not the only C3b binding acceptor on the bacterial surface. Deposition of C5 exhibited a more drastic reduction in oprF-deficient P. aeruginosa than in PAO1 (Fig. 2B). This is likely due to amplification of complement activation, since C5 binding occurs downstream of C3 (29, 30). To evaluate the specificity of C3b binding, P. aeruginosa mutants defective in other porins were tested along with wild-type PAO1, and none of them showed a significant difference in C3b deposition (Fig. 2D).

FIG 2.

OprF promotes binding to human C3b. (A, B) Mid-log-phase bacteria were opsonized with 20% human serum for 5 or 20 min, and the reactions were stopped. Bacteria were fixed and stained with anti-human C3 antibody (A) or anti-human C5 antibody (B) as outlined in Materials and Methods. (C) OprF-expressing E. coli and vector control strains were incubated with C5-depleted human serum, washed, fixed, and stained with anti-C3 antibody. (D) To evaluate the specificity of the interaction of OprF and C3b, other porin mutants, ΔoprD, ΔoprQ, ΔoprI, and ΔoprJ mutants, along with wild-type P. aeruginosa, were opsonized with human serum and stained with anti-C3 antibody. C3b binding efficiency was analyzed by flow cytometry using Flow Jo software. Means ± standard errors of the means (SEM) are given (n = 3). Nonopsonized bacteria were used in each experiment to set the gate for analysis. Data from 30,000 gated events were collected, and mean fluorescence intensity (MFI) was calculated for each sample. **, P ≤ 0.005; ***, P ≤ 0.0005.

Furthermore, we sought to determine the role of the classical pathway in OprF-mediated C3 deposition on the bacterial surface. We assessed the ability of PAO1 and P. aeruginosa ΔoprF to bind C3 during incubation with C1q-depleted serum (which is devoid of classical pathway activity), C1q-repleted serum (C1q-depleted serum reconstituted with purified human C1q), or heat-inactivated serum (HIS). Our results showed limited C3 binding in HIS or C1q-depleted serum and significantly increased C3 binding in C1q-repleted serum (Fig. 3). These data provide evidence for activation of the classical complement pathway in mediating C3 binding to P. aeruginosa.

FIG 3.

The classical complement pathway mediates C3 binding to P. aeruginosa. Mid-log-phase cultures of wild-type PAO1 and oprF-deficient P. aeruginosa were incubated in 20% C1q-depleted human serum, C1q-repleted serum or heat-inactivated serum for 20 min at 37°C. The bacterial pellets were washed, and C3 binding was analyzed by flow cytometry. Means ± SEM are given (n = 3). **, P ≤ 0.005; ***, P ≤ 0.0005.

Psl restricts OprF accessibility.

We previously published that Psl polysaccharide limits the binding of C3b to the surface of P. aeruginosa (4). We postulated that Psl was masking C3b binding acceptors on the bacterial surface. To test this, we measured accessibility of anti-OprF antibodies in wild-type PAO1, Δpsl, and Psl-overproducing (Psl⁺⁺) strains of P. aeruginosa. There was significantly higher binding of anti-OprF antibodies on psl-deficient P. aeruginosa than of PAO1 and Psl-overproducing strains (Fig. 4A). Expression of Psl in these strains was confirmed by dot blot assays using anti-Psl antibody (Fig. 4B) (4). These data support our previous hypothesis that Psl masks OprF and limits C3b deposition.

FIG 4.

Psl limits accessibility of OprF for antibody binding. (A) Mid-log-phase wild-type PAO1, Δpsl mutant, and Psl-overproducing (PAO1 Psl⁺⁺) P. aeruginosa strains were incubated with anti-OprF antibody conjugated with Alexa Fluor 488, and binding of OprF on the bacterial surface was determined by flow cytometry. (B) Cell extracts were blotted on a nitrocellulose membrane, probed with anti-Psl antiserum, and detected by chemiluminescence using an HRP-conjugated secondary antibody. **, P ≤ 0.005; ***, P ≤ 0.0005.

Complement C3b binds to recombinant OprF expressed in E. coli.

Deletion of bacterial porins can result in perturbations in the cell wall structure (31, 32). Thus, it was a formal possibility that loss of C3b binding in the P. aeruginosa oprF mutant was an indirect effect due to changes in cell wall structure. This did not appear to be the case based on transmission electron microscopy (TEM) analysis of these cells (data not shown). Nonetheless, we examined if C3b would bind OprF when expressed in E. coli. For this, we transformed E. coli with pHSG576, a plasmid that harbors oprF (12). Analyses of stained proteins from whole-cell lysates revealed clear expression of OprF in E. coli compared with the vector control (Fig. 5A) or by Western blotting (Fig. 5B). To evaluate C3b binding to recombinantly expressed OprF, E. coli strains were subjected to far-Western analysis (see Materials and Methods). While C3b appeared to bind to several E. coli proteins, including those with the vector control, a unique band corresponding to OprF was recognized in OprF-expressing E. coli compared with vector control (Fig. 5C). We also repeated the C3 binding studies in C5-deficient serum because E. coli exhibits some level of serum-mediated killing (see below). As expected, we found a significant increase in deposition of C3 on OprF-expressing E. coli compared to the vector control (Fig. 2C).

FIG 5.

Recombinant OprF binds C3b. Lysates from recombinant OprF E. coli or vector control were stained with Coomassie blue G-250 (A) or subjected to Western blot with anti-OprF antibody (B) or incubated with normal human serum, washed, and probed with anti C3b antibody (C). Arrows represent the positions of OprF.

OprF promotes bacterial internalization by human neutrophils.

Since OprF on the P. aeruginosa surface binds C3b and C3b serves as a potent opsonin, we determined whether OprF-C3b interactions promote the internalization of P. aeruginosa by human neutrophils. Serum-opsonized strains PAO1 or PAO1ΔoprF were incubated with human neutrophils. Unattached bacteria were washed extensively, and bacterial association and internalization were assessed by differential immunofluorescence staining using confocal microscopy (see Materials and Methods). The number of oprF-defective bacteria phagocytosed by human neutrophils was reduced approximately 2.5-fold compared with the number of wild-type P. aeruginosa bacteria (Fig. 6A). Since OprF is highly immunogenic and extensively used in vaccine preparations, it was important to evaluate the potential role of serum immunoglobulins in this phagocytosis. We incubated wild-type PAO1 and oprF-deficient bacteria with 20% heat-inactivated serum (which contains functional antibody but is devoid of complement activity) and investigated bacterial uptake by neutrophils. We found a significant reduction in internalized WT PAO1 when treated with heat-inactivated serum compared to treatment with fresh serum, indicating that OprF contributes to complement-mediated phagocytosis and serum immunoglobulins play a limited role. As expected, oprF-deficient P. aeruginosa remained unaffected (Fig. 6A).

FIG 6.

OprF promotes complement-mediated internalization by human neutrophils. (A) Neutrophils were infected with normal human serum (NHS) or heat-inactivated serum (HIS) opsonized wild-type and oprF-deficient P. aeruginosa strains as described in Materials and Methods. A representative image shows internalization of NHS-opsonized P. aeruginosa by neutrophils. External bacteria exposed to the antibodies before permeabilization stained green. Bacteria exposed to the antibodies after permeabilization stained red (arrows indicate internalized bacteria). DAPI was used to stain the phagocyte nucleus (blue; magnification, ×60). Quantitation of internalized bacteria is denoted in parentheses below each image and represents an average of 100 infected neutrophils that were examined from triplicate coverslips in each test group. (B) C5-depleted-serum-opsonized OprF-expressing E. coli and vector control were incubated with neutrophils and stained with C3 and OprF antibody. Vector control E. coli bound only with C3 antibody and stained green (i), while OprF-expressing E. coli bound with both C3 and OprF antibodies and stained yellow (ii); blue DAPI fluorescence is used to stain the phagocyte nucleus (as above). Quantitation of bacterial internalization with neutrophils is denoted in parentheses below each image and represents an average for 100 infected neutrophils examined from triplicate coverslips in each test group. Shown are results representative of three independent experiments.

In parallel, OprF-expressing E. coli strains were opsonized with C5-depleted serum to avoid serum-mediated killing of E. coli (see below). Complement C3-mediated bacterial association with neutrophils was evaluated by dual-staining confocal microscopy (see Materials and Methods). OprF-expressing E. coli showed 5-fold-enhanced association with neutrophils compared with the vector control (Fig. 6B). The increased association of OprF-expressing bacteria with human neutrophils in the presence of serum provides evidence that C3b binding to OprF facilitates interaction of E. coli with neutrophils.

Expression of P. aeruginosa OprF facilitates complement-mediated killing of E. coli.

Finally, we evaluated if the presence of OprF in E. coli and the subsequent enhanced complement deposition promote serum-mediated killing. We used nonimmune fresh human serum, heat-inactivated serum, and C5-depleted serum. As expected, C5-depleted serum did not exhibit any complement-mediated bacterial killing in OprF-expressing E. coli and vector control since C5 is essential for deposition of the membrane attack complex (MAC). While the vector control E. coli strain was susceptible to fresh serum killing, expression of OprF further increased serum-mediated bactericidal effects by 2.5 log₁₀ units (Fig. 7). This finding is likely due to the fact that C3b binding is increased in OprF-expressing bacteria, leading to increased MAC deposition and complement-mediated killing. Heat-inactivated serum is devoid of complement activity and thus served as a negative control for complement-mediated reactions.

FIG 7.

Recombinant expression of OprF increases complement-mediated killing of E. coli. An OprF-expressing strain of E. coli and vector control were incubated with 20% normal human serum, heat-inactivated serum, and C5-depleted serum for 45 min at 37°C. Bacterial survival is expressed as CFU/ml. Data represent means ± SEM from a representative experiment (n = 3; ***, P ≤ 0.005, unpaired Student's t test).

DISCUSSION

The complement system plays an important role in the clearance of bacteria. The bactericidal effect of C3 is mediated by either opsonophagocytosis or complement-mediated killing. We have previously observed that serum-opsonized P. aeruginosa showed more-robust oxidative burst response generated by neutrophils than did unopsonized bacteria (4). The objective of the current study was to identify potential complement acceptor molecule(s) on the P. aeruginosa surface and investigate complement binding and complement-mediated phagocytosis. By utilizing the far-Western assay and mass spectroscopy, we identified an outer membrane protein, OprF porin, that binds C3b (Fig. 1). Collectively, our data demonstrate that (i) C3b binding is markedly reduced in an oprF-deficient P. aeruginosa strain, (ii) far-Western blot studies show that C3b binds to OprF, (iii) OprF-mediated binding to C3b increases P. aeruginosa phagocytosis by human neutrophils, and (iv) interaction of C3b with E. coli-expressing P. aeruginosa OprF triggers downstream events of the complement cascade, enhancing complement-mediated killing.

Previous studies have shown that OprF is one of the most abundant proteins in the P. aeruginosa outer membrane (12, 13). OprF is highly conserved in pseudomonads, allowing the passage of small molecules across the outer membrane, and has been studied extensively as a vaccine candidate (15, 35, 36). Porins have also been described as acceptor molecules for C3b in Salmonella enterica serovar Typhimurium, Klebsiella pneumoniae, Neisseria meningitidis, and Legionella pneumophila (28, 37–39). Porin-independent C3b binding acceptor molecules have been identified in several other pathogens. These molecules include the heparin binding agglutinin (HbA) protein from Mycobacterium tuberculosis and Mycobacterium avium (40), a surface immunoglobulin binding protein (Sbi) and extracellular fibrogen binding protein (Efb) from Staphylococcus aureus (41, 42), and phenolic glycolipid-1 from Mycobacterium leprae (43). We also observed that oprF-deficient P. aeruginosa and E. coli were still able to bind some C3b, indicating the likelihood of other C3b binding molecules on these bacterial surfaces. For P. aeruginosa, these acceptors do not appear to be other porins such as OprQ, OprI, OprJ, or OprD.

We further observed that binding of OprF to C3b has a significant impact on phagocytosis by neutrophils. In E. coli, recombinant-expressing OprF showed increased interaction with neutrophils compared to vector control (Fig. 6). This also significantly promoted downstream events of the complement cascade (i.e., assembly of the MAC) and increased the bactericidal effect of serum in E. coli (Fig. 7). However, this was not the case for P. aeruginosa, in which we did not observe complement-mediated killing (data not shown). There are several known mechanisms for pathogens to evade complement-mediated killing (41, 42). P. aeruginosa binds factor H and factor H-related protein to its surface through elongation factor protein Tuf, leading to C3b degradation (44). It is also possible that one or more molecules on the P. aeruginosa surface inhibit formation of the MAC, resulting in protection from serum killing. Bacterial killing by complement is associated with dissolution of the outer membrane, and the molecular form of the C5-9 complex is crucial for optimal killing (45, 46). Trypanosoma cruzi amastigotes escape complement-mediated killing because MAC, while bound efficiently, is not inserted deep enough in the parasite lipid bilayer and fails to form pores in the membrane (47). Therefore, the correct stoichiometric ratio of MAC complement components may be suboptimal for lysis of P. aeruginosa. Previous literature has highlighted the role of lipopolysaccharide (LPS) in reducing P. aeruginosa serum bactericidal activity, but further studies are needed in this area of research (48, 49).

We demonstrated previously that the exopolysaccharide Psl reduces the host innate immune response by limiting complement deposition on P. aeruginosa (4). We also found that psl-deficient P. aeruginosa exhibits significantly increased complement binding and is thereby taken up by neutrophils more readily. These results led us to the current hypotheses that a C3b acceptor molecule exists on the P. aeruginosa surface, i.e., OprF, and that Psl masks the accessibility of OprF for interaction with C3b. Data in the current study support these hypotheses.

To better understand the interactions between OprF and C3b, the OprF predicted secondary structure was analyzed. The predicted surface-exposed C-terminal 240- to 350-amino-acid region is structurally conserved with OmpA domains of Acinetobacter, Salmonella, Legionella, Neisseria, and several other Gram-negative bacteria. Consensus secondary structure predictions indicate that the C-terminal regions of these porins have three α-helices and four parallel β sheets (with the exception of Neisseria [Fig. 8]). Moreover, our studies showing C3b binding to P. aeruginosa OprF as well as data with L. pneumophila major outer membrane protein (MOMP) (38) support our hypothesis that C3b has evolved to recognize patterns present in conserved C-terminal porin domains. Understanding the molecular mechanism of C3 binding to bacterial surfaces offers a potential means for preventing and/or treating diseases caused by these Gram-negative pathogens. Moreover, since C3b binding to proteins greatly facilitates antigen presentation by antigen-presenting cells (50), C3b binding to OprF may be one mechanism whereby OprF serves as a potent vaccine (36, 51).

FIG 8.

3D model of conserved C-terminal regions of porins. The predicted surface-exposed domain of P. aeruginosa OprF shows structural identity with the C-terminal domains of several other Gram-negative porins (conserved four yellow β sheets and three pink α helices).