Binding of Streptococcus pneumoniae Endopeptidase O (PepO) to Complement Component C1q Modulates the Complement Attack and Promotes Host Cell Adherence*

Background: A detailed understanding of how pneumococci interact with the human host aids development of novel therapeutics. Results: PepO binds complement protein C1q and C4BP mediating pneumococcal-host cell adherence and evasion of the complement-mediated attack. Conclusion: PepO contributes to pneumococcal virulence. Significance: Pneumococci have multiple interrelated invasion and survival strategies. The Gram-positive species Streptococcus pneumoniae is a human pathogen causing severe local and life-threatening invasive diseases associated with high mortality rates and death. We demonstrated recently that pneumococcal endopeptidase O (PepO) is a ubiquitously expressed, multifunctional plasminogen and fibronectin-binding protein facilitating host cell invasion and evasion of innate immunity. In this study, we found that PepO interacts directly with the complement C1q protein, thereby attenuating the classical complement pathway and facilitating pneumococcal complement escape. PepO binds both free C1q and C1 complex in a dose-dependent manner based on ionic interactions. Our results indicate that recombinant PepO specifically inhibits the classical pathway of complement activation in both hemolytic and complement deposition assays. This inhibition is due to direct interaction of PepO with C1q, leading to a strong activation of the classical complement pathway, and results in consumption of complement components. In addition, PepO binds the classical complement pathway inhibitor C4BP, thereby regulating downstream complement activation. Importantly, pneumococcal surface-exposed PepO-C1q interaction mediates bacterial adherence to host epithelial cells. Taken together, PepO facilitates C1q-mediated bacterial adherence, whereas its localized release consumes complement as a result of its activation following binding of C1q, thus representing an additional mechanism of human complement escape by this versatile pathogen.

Diseases caused by the Gram-positive species Streptococcus pneumoniae are still among the leading cause of a global public health care problem (1,2). Despite the use of antibiotics and the availability of vaccines, the mortality rate due to pneumococcal infection remains alarming. Although pneumococci are commensals of the human nasopharynx, they are also a major pathogen causing infections ranging from local and mild, such as acute otitis media and sinusitis, to systemic and life-threatening conditions such as lobar pneumonia, sepsis, and meningitis. It has been estimated that there were ϳ14.5 million episodes of pneumococcal disease in 2000 among children less than 5 years of age with 826,000 deaths globally (2). To combat pneumococcal diseases, the development of new and improved therapies and alternative pneumococcal vaccines is needed. Consequently, the identification of essential factors facilitating colonization and subsequent dissemination and the elucidation of the molecular mechanism of pneumococcal-host interactions are important.
The complement system consists of over 35 serum and membrane-bound proteins and is a major arm of the innate immune system. It plays an important role not only in the host defense against invading pathogens but also in the removal of apoptotic and necrotic cells and in regulation of the antibody (Ab) 3 response of the adaptive immune system (3). Depending on the initiating agent, complement can be activated via three pathways, namely the classical, the lectin, and the alternative pathway. Binding of the C1 complex to immunoglobulins (Igs) recognizing invading pathogens usually triggers the classical pathway, whereas the lectin pathway is initiated by specific carbohydrates. The alternative pathway, which is continuously activated at low levels by autoactivation of the complement component C3, is amplified upon its deposition on the surface of the pathogen. All three pathways result in the deposition of opsonin C3b, release of proinflammatory anaphylatoxins C3a and C5a, and formation of a membrane attack complex, which causes lysis. However, the process is tightly regulated via soluble and membrane-bound complement inhibitors. This is important so as to restrict the action of complement at the site of foreign or modified surfaces and to avoid overconsumption and complement-mediated tissue injury.
Almost all microbes infecting an immunocompetent host are immediately confronted by the host complement system. However, many invasive pathogens, to survive and multiply within the human host, employ different mechanisms to inactivate or evade such attacks (4 -7). Pneumococci are highly versatile pathogens that take advantage of multiple host proteins and mechanisms to evade the immune attack and to colonize the host (8 -17). These mechanisms are essential for their dissemination in both the lungs and the bloodstream, resulting in an enhanced survival in the host. Pneumococci express several surface proteins that recruit the host complement inhibitors such as Factor H, vitronectin, and C4b-binding protein (C4BP) to the bacterial surface to control complement-mediated attack (10 -12, 14, 15, 18 -21). Additionally, these virulence factors function as adhesins that either interact directly with the host cell surface receptors or utilize host proteins as a molecular bridge between the bacteria and the host to facilitate colonization. Host proteins include extracellular matrix and serum proteins such as fibronectin, thrombospondin, vitronectin, plasminogen, Factor H, C4BP, and C1q (12,18,(22)(23)(24)(25)(26)(27).
We recently reported that C1q directly interacts with protein(s) at the pneumococcal surface (22). In turn, C1q functions as a molecular bridge between pneumococci and the host facilitating pneumococcal adherence and invasion (22). In the present study, we identified the multifunctional pneumococcal protein endopeptidase O (PepO) as the C1q ligand at the bacterial surface. We investigated the functional aspect of PepO-C1q interactions, its complement-modulating activities, and furthermore the ability to promote pneumococcal adherence to host cells.
Normal human serum (NHS) was prepared from freshly drawn blood obtained from healthy volunteers with informed consent and according to the recommendations of the local ethical committee in Lund (permit 418/2008). The pooled blood was allowed to clot for 30 min at room temperature and then incubated for 1 h on ice. After two centrifugations, the serum fraction was frozen in aliquots and stored at Ϫ80°C.
Hemolytic Assay-The activity of the classical pathway of complement activation was assessed using sheep erythrocytes (Håtunalab). The erythrocytes were washed three times with ice-cold DGVB ϩϩ buffer (2.5 mM Veronal buffer (pH 7.3), 70 mM NaCl, 140 mM glucose, 0.1% gelatin, 1 mM MgCl 2 , and 0.15 mM CaCl 2 ), resuspended to a concentration of 10 9 cells/ml, and incubated, with gentle shaking for 20 min at 37°C, with an equal volume of amboceptor (Dade Behring) diluted 1:3,000 in DGVB ϩϩ buffer. Amboceptor consists of anti-sheep erythrocyte Abs, which activate the classical pathway of complement. After two washes, 8 ϫ 10 7 cells/ml were incubated for 1 h at 37°C with 0.2% NHS and increasing concentrations of PepO, diluted in DGVB ϩϩ buffer, in a total volume of 150 l. BSA and MID were used as negative controls. To assess the activity of the alternative pathway, rabbit erythrocytes (obtained from a local farm) were washed three times with Mg 2ϩ EGTA buffer (2.5 mM Veronal buffer (pH 7.3) containing 70 mM NaCl, 140 mM glucose, 0.1% gelatin, 7 mM MgCl 2 , and 10 mM EGTA). Erythrocytes at a concentration of 6 ϫ 10 7 cells/ml were incubated for 1 h at 37°C with 1.25% NHS together with increasing concentration of PepO or BSA diluted in Mg 2ϩ EGTA buffer in a total volume of 150 l. Following the incubation the samples were centrifuged at 800 ϫ g to pellet intact cells, and the hemolytic activity (i.e. the number of lysed erythrocytes and released hemoglobin) was determined by spectrophotometric measurement of absorbance at 405 nm (Varian Cary 50 MPR microplate reader). The lysis obtained in the absence of inhibitor was defined as 100% hemolytic activity.
Complement Activation Assays-To assess the ability of PepO to activate the complement pathway, microtiter plates (MaxiSorp; Nunc) were coated with 50 l of PepO (5 g/ml), aggregated human IgG (2.5 g/ml for the classical pathway), mannan (100 g/ml (Sigma) for the lectin pathway), or zymosan (20 g/ml (Sigma) for the alternative pathway) in PBS overnight at 4°C. Wells coated with BSA were used as a control. The plates were washed with 50 mM Tris-HCl (pH 8.0), 150 mM NaCl, and 0.1% Tween 20 between each step. The plates were blocked with 250 l of blocking solution (50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 0.1% Tween 20, 3% fish gelatin (Norland)) for 2 h at room temperature. After blocking, the plates were washed, and dilutions of NHS or serum deficient in C1q in GVB ϩϩ (5 mM Veronal buffer pH 7.4, 144 mM NaCl, 1 mM MgCl 2 , 0.15 mM CaCl 2 , and 1% gelatin) for classical and lectin pathway, respectively, were added to the plates and incubated for 15 min (for detection of C3b) or 45 min (for C1q detection) at 37°C. For the alternative pathway, the plates were incubated with NHS diluted in Mg 2ϩ EGTA buffer and incubated for 20 min (C3b) at 37°C. Following the incubation deposited complement proteins were detected using specific rabbit polyclonal Abs against C1q or C3d diluted 1:2,000 in blocking solution and horseradish peroxidase-labeled secondary Abs against rabbit (1:2,000 in blocking solution) for 1 h each at room temperature. The plates were developed with o-phenylenediamine (Dako) substrate and H 2 O 2 , and the absorbance at 490 nm was measured. For inhibition of classical complement pathway deposition by PepO, different concentrations of NHS (0.25% for C1q, 0.125% for C3b, and 0.5% for C9) diluted in GVB ϩϩ buffer were incubated with increasing concentrations of PepO or BSA at 37°C for 15 min. Later the mixtures were added to the human IgG-coated plates and incubated at 37°C, and deposited complement components were detected.
Direct Binding Assays-Microtiter plates were coated with PepO (5 g/ml), aggregated human IgG (2.5 g/ml), or C1q (5 g/ml) in PBS overnight at 4°C. Wells coated with BSA were used as control. After blocking, the plate was incubated with increasing concentrations of plasma-purified C1q or C1 complex for binding to PepO and aggregated human IgG, with plasma-purified C4BP for binding to PepO or with PepO for binding to immobilized C1q in binding buffer (50 mM HEPES (pH 7.4), 150 mM NaCl, 2 mM CaCl 2 , and 50 g/ml BSA) and incubated for 1 h 30 min at room temperature. To assess the effect of calcium on C1q binding to PepO, the assay was performed in binding buffer with and without 2 mM CaCl 2 . For investigation of the effect of ionic strength on C1q-PepO or C4BP-PepO interactions, the binding buffer was supplemented with additional NaCl (0 -800 mM). In inhibition assays, aggregated human IgG-coated plates were incubated with purified C1q (5 g/ml) in the absence or presence of PepO protein that was used as an inhibitor. To assess the binding of various forms and fragments of C4BP (rC4BP wild-type and mutants lacking individual domains) to PepO, 30 g/ml of each C4BP variant was incubated overnight at 4°C with PepO immobilized at 10 g/ml. Bound C1q, C4BP, or PepO was detected using specific polyclonal Abs. The plates were developed with o-phenylenediamine substrate and H 2 O 2 , and the absorbance at 490 nm was measured.
Biacore Analysis-To measure the kinetics of PepO binding to C1q and C4BP, surface plasmon resonance analysis was performed using Biacore 2000. C1q and C4BP, both purified from human plasma, were diluted to 50 g/ml in 10 mM sodium acetate (pH 4.0) and immobilized on the surface of a CM5 sensor chip to reach 1835 and 3016 response units, respectively. All experiments were performed at a continuous flow rate of 30 l/min using Biacore buffer (150 mM NaCl, 10 mM Hepes, 0.002% Tween 20, pH 7.4). The analyte, PepO, was injected in a concentration gradient from 22 to 1500 nM followed by two consecutive injections of 2 M NaCl for regeneration. The obtained sensorgrams were analyzed using Bio-evaluation software 3.0 using a 1:1 Langmuir binding model of interaction with drifting baseline.
Binding from Serum-Microtiter plates coated with PepO or BSA were blocked, and NHS diluted in GVB ϩϩ buffer or heat-inactivated serum (56°C for 30 min) diluted in binding buffer was added at the concentrations indicated in the figures. After incubation for 1 h and 30 min at 37°C for NHS and at room temperature for heat-inactivated serum, bound C4BP was detected. The plates were developed with o-phenylenediamine substrate and H 2 O 2 , and the absorbance at 490 nm was measured.
Bacterial Strains, Cell Lines, and Culture Conditions-S. pneumoniae NCTC10319 (serotype 35A) and D39 (serotype 2) were cultured on blood agar plates at 37°C and 5% CO 2 . Isogenic mutants that do not express PepO were constructed for the encapsulated D39 strain as described (10). A pepO-deficient mutant of D39 was grown in the presence of spectinomycin (50 g/ml). The strain Escherichia coli DH5␣ was cultured in Luria Bertani (LB) broth or LB agar plates at 37°C. The human A549 cell line (lung alveolar epithelial cells, type II pneumocytes; American Type Culture Collection) was cultured in DMEM (PAA Laboratories) supplemented with 10% heat-inactivated FCS (Invitrogen), 2 mM glutamine (PAA Laboratories), penicillin G (100 Units/ml), and streptomycin (0.1 mg/ml) (both from HyClone) at 37°C under 5% CO 2 atmosphere.
Direct Binding Assay with 125 I-C4BP-Plasma-purified C4BP (30 g) was labeled with 125 I using the chloramine-T method. Bacteria were grown overnight on blood agar plates and were washed once and resuspended to A 600 nm of 0.5 in PBS with 1% BSA. Bacteria (2 ϫ 10 8 cfu) were incubated for 1 h at room temperature with 125 I-labeled C4BP (100 kcpm/reaction) in PBS with 1% BSA (where kcpm is kilo counts per minute). The bacteria-bound 125 I-C4BP was separated from unbound 125 I-C4BP by centrifugation through 20% sucrose columns for 3 min at 10,000 rpm. Both bound and unbound 125 I-C4BP was measured in a gamma counter (Gamma Master 1277, LKB Wallac).
Flow Cytometric Analysis of C1q Binding to Pneumococci-Binding of plasma-purified C1q to S. pneumoniae D39 and its corresponding isogenic pepO mutant was measured using flow cytometry. Pneumococci were cultured overnight on blood agar plates, washed in PBS, and adjusted to 10 9 cfu/ml. Bacteria (5 ϫ 10 7 cfu) were incubated with purified C1q, at the indicated concentrations, in PBS for 1 h at 37°C. In inhibition assays, pneumococci NCTC10319 (5 ϫ 10 7 cfu in 100 l) were incubated with purified C1q (5 g/ml) or NHS (0.5%) in the absence or presence of PepO that was used as inhibitor. Finally, bacteria were fixed using 1% paraformaldehyde (Sigma), and the flow cytometry analysis was performed using CyFlow space (Partec) to detect the binding of C1q. Bacteria were detected using logforward and log-side scatter dot-plot, and a gating region was set to exclude debris and larger aggregates of bacteria. 15,000 bacteria/events were analyzed for fluorescence using log-scale amplifications. The geometric mean fluorescence intensity (GMFI) was used as a measure for binding activity.
Pneumococcal Host Cell Infection Assay-A549 cells were seeded at a density of 5 ϫ 10 4 cells/well in plain medium either on 24-well tissue culture plates (Nunc) or on glass coverslips (diameter, 12 mm) when assayed by immunofluorescence and cultivated for 48 h. Confluent monolayers were washed thoroughly and infected with pneumococci in DMEM medium supplemented with 1% FCS (infection medium) for 3 h at 37°C using a multiplicity of infection of 25. Bacteria were incubated for 20 min with or without plasma-purified C1q (10 g) in the absence or the presence of PepO (10 g) in a total volume of 100 l of infection medium at 37°C prior to infections. The infection assays were carried out in a total volume of 500 l after adding the bacteria. After infection, cells were washed extensively with infection medium to remove unbound bacteria. The total number of adherent and intracellular bacteria was monitored after detachment and lysis of cells with saponin (Sigma) (1.0% w/v) followed by plating the bacteria on blood agar.
Fluorescence Microscopy-Pneumococci attached to host epithelial cells were stained using polyclonal anti-pneumococcal Abs (IgG) in combination with a secondary goat anti-rabbit IgG coupled with Alexa Fluor 488 (green) or Alexa Fluor 546 (red) (10,33). After infection, the glass coverslips were blocked for 1 h at room temperature with 10% FCS in PBS followed by incubation with anti-pneumococcal Abs (1:100) for 1 h at room temperature. After washing, bound Abs were detected with an Alexa Fluor-labeled goat anti-rabbit Ig conjugate. For staining of intracellular bacteria, infected cells were permeabilized with 1% Triton X-100 followed by staining with the Abs. The glass coverslips were embedded "upside down" in mounting medium (DAKO), sealed with nail polish, and stored at 4°C. A confocal laser scanning microscope (Zeiss LSM 510 META) and the appropriate software were used for the image acquisition.
Deposition of C3b on Pneumococci-The deposition of C3b on viable pneumococci was measured using flow cytometry. A bacterial suspension of D39 and its isogenic pepO mutant (containing 5 ϫ 10 7 cfu) was incubated with increasing concentration of NHS in sterile GVB ϩϩ buffer in a total volume of 0.1 ml for 15 min at 37°C. In the inhibition assay, bacteria were incubated with or without PepO (10 g) pretreated with 0.25% NHS. Human serum albumin was used as a negative control. After incubation, bacteria were washed and stained for deposition of C3b using FITC labeled anti-C3c Abs (Dako) and analyzed by flow cytometry.
Bactericidal Assay-E. coli DH5␣ was cultured in LB broth until the exponential growth phase. After harvesting, the cells were washed once in GVB ϩϩ buffer and adjusted to 2 ϫ 10 4 cfu/ml. NHS was diluted in GVB ϩϩ to a concentration of 0.2% and incubated with various concentrations of PepO or A1AT for 10 min at 37°C. Thereafter, 1000 bacterial cells were added and incubated with serum for 30 min at 37°C in a total volume of 100 l. Following the incubation, viable bacterial counts were determined by plating samples onto LB agar plates.

RESULTS
PepO Inhibits the Hemolytic Activity of Human Serum-To investigate whether the secreted pneumococcal PepO has any inhibitory effect on complement, hemolytic assays were performed to measure the activity of the classical and the alternative pathways. The addition of recombinant PepO but not MID or BSA to sheep erythrocytes and human serum resulted in significant dose-dependent inhibition of the lysis through the classical pathway (Fig. 1A). However, no effect of PepO was observed on erythrocytes lysis mediated by the alternative pathway (Fig. 1B).
To determine at which step the classical cascade is inhibited by PepO, microtiter plates were coated with aggregated IgG as a complement-activating ligand, and NHS preincubated with PepO was added. Deposition of individual complement components was measured with specific antibodies. The pretreatment of serum with PepO significantly inhibited the deposition of C1q, C3b, and C9 through the classical pathway in a dosedependent manner (Fig. 2). However, no such inhibition was observed for the alternative pathway (data not shown). Taken together, our data indicated that PepO specifically inhibits the classical complement cascade at the level of C1 deposition.
PepO Activates Complement-As PepO significantly inhibits the classical pathway, we asked whether the inhibition observed is due to the ability of PepO to activate and consequently consume complement. Microtiter plates coated with PepO, aggregated IgG as a positive control, and BSA as a negative control were incubated with NHS followed by detection of various deposited complement components with specific antibodies. PepO strongly activated the classical pathway, as was evident from strong deposition of C1q and C3b (Fig. 3, A and B). The To test activation of the lectin pathway, human serum depleted of C1q was used, but here negligible deposition of C3b was triggered by PepO (Fig. 3C). The C1q-depleted serum used had an intact lectin pathway as was evident by strong deposition of C3b on mannan-coated microtiter plates (Fig. 3C). Moreover, no classical pathway activation was detected with C1qdepleted serum, as measured by C4b deposition on human IgG or PepO (data not shown). In contrast, activation of the alternative complement pathway was observed, where PepO induced the deposition of C3b (Fig. 3D). Taken together, our data indicate that PepO strongly activates complement.
PepO Binds C1q-Because PepO strongly activates the classical pathway of complement activation, it is very likely that the observed effect is a consequence of the direct interaction between PepO and C1q. To test this hypothesis, PepO was incubated with increasing concentrations of C1q purified from human plasma followed by detection using specific Abs. Interestingly, a dose-dependent binding of C1q to PepO was observed (Fig. 4A). Similar results were obtained using the C1 complex, consisting of C1q, the recognition molecule, and two Ca 2ϩ -dependent dimeric serine proteases, C1r and C1s (tetramer C1r 2 -C1s 2 ), as a source of C1q (Fig. 4B). In addition, the binding of pneumococcal PepO to immobilized C1q was dose-dependent (Fig. 4C). Interestingly, the presence or the absence of calcium had no significant effect on the PepO-C1q interaction (Fig. 4D). Further, SPR was used to analyze the interaction between PepO and C1q (Fig. 4E). Once again a concentration-dependent binding of PepO to immobilized C1q was detected (K D ϭ 0.22 M). Representative sensorgrams for the interaction are shown. Moreover, the presence of PepO but not MID or BSA significantly inhibited the binding of plasmapurified C1q to aggregated human IgG (Fig. 4F). Taken together, these data confirmed that PepO binds C1q.
To determine whether the PepO-C1q interaction is hydrophobic or ionic in character, the binding assays were conducted in the presence of varying NaCl concentrations. The binding of C1q to PepO decreased with increasing NaCl concentrations (Fig. 4G). At a NaCl concentration of 250 mM, the C1q binding to PepO was significantly reduced as compared with binding observed at the physiological concentration of 150 mM NaCl. Augmenting the ionic strength resulted in further reduction in binding of the protein to PepO. Similar results were obtained for the interaction between C1q and IgG (Fig. 4G). These results suggest that the binding is influenced by ionic strength.
PepO at the Pneumococcal Surface Functions as a C1q Ligand-Our data show that recombinant PepO interacts specifically with C1q. Therefore, we wondered whether this is also true for the bacterial surface-presented PepO. To address this, C1q binding to S. pneumoniae D39 serotype-2 strain and its isogenic-pepO mutant strain (10) was analyzed. A dose-dependent binding of C1q to the wild-type pneumococcus was observed (Fig. 4H). However, compared with the wild-type strain, a significant reduction in the binding of C1q to the isogenic-pepO knock-out mutant was observed (Fig. 4H). The results indicate that bacterial surface-presented PepO is indeed a C1q-binding protein.
The Complement Inhibitor C4BP Binds PepO-We showed that pneumococcal enolase interacts with C4BP and thereby mediates complement evasion (23). To test whether this was also the case for PepO, the protein was incubated with several dilutions of NHS in GVB ϩϩ buffer or with heat-inactivated serum in binding buffer, and C4BP binding was detected with specific Abs. A dose-dependent C4BP binding was observed when normal or heat-inactivated human serum was added (Fig.  5, A and B). To corroborate the above results, we analyzed the binding of C4BP purified from human plasma to PepO-coated plates followed by detection using specific Abs. A dose-dependent binding of C4BP to PepO was observed (Fig. 5C). In addition, the interaction between PepO and C4BP was analyzed by SPR (Fig. 5D). A concentration-dependent, high affinity binding of PepO to immobilized C4BP was detected (K D ϭ 0.17 M). PepO was preincubated with NHS in GVB ϩϩ buffer and added to microtiter plates coated with aggregated human IgG. BSA was used as a negative control. Deposition of C1q, C3b, and C9 was measured using polyclonal antibodies. The amount of deposited complement components in the absence of inhibitor was set as 100%. The data represent the means Ϯ S.D. of three independent experiments performed in duplicates. Statistical significance of differences was calculated using two-way analysis of variance and Bonferroni's post-test. *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001.
Representative sensorgrams for the interaction are shown. These data confirmed that PepO is a C4BP-binding protein.
Furthermore, the binding of C4BP to PepO decreased with increasing NaCl concentrations (Fig. 5E). At a NaCl concentration of 250 mM, the C4BP binding to PepO was reduced by almost 60% as compared with binding at the physiological concentration of 150 mM NaCl. These results suggested that the binding is based on ionic interactions between C4BP and PepO.
To further elucidate the role of PepO in pneumococcal binding of C4BP, the binding of radiolabeled plasma-purified C4BP to wild-type strain D39 and NCTC10319 and their respective isogenic pepO mutants was analyzed. We observed that both pneumococcal strains bound C4BP, albeit with varying binding percentage, which was expected as the D39 serotype-2 strain is highly encapsulated, whereas the NCTC10319 background has a relatively thin capsule. Nevertheless, a moderate but significant reduction in C4BP binding was observed for the respective pepO mutants (Fig. 5F). Taken together, these data showed that in addition to enolase (33), the PepO is a pneumococcal C4BPbinding protein.
Localization of the Binding Site for PepO on C4BP-We then characterized the interaction between C4BP and PepO using various forms of C4BP, which are represented schematically in Fig. 6, A-C. Binding experiments demonstrated marginal reduction in the binding of all C4BP mutants, lacking one CCP domain at the time, as compared with the rC4BP. However, the C4BP construct lacking CCP8 (⌬CCP8) showed significantly decreased binding capacity for PepO (20.28 Ϯ 2.81%) as compared with that of rC4BP (Fig. 6D). The result suggests that C4BP contacts PepO primarily via the binding sites, which are localized within CCP8 of the ␣-chain.

C1q Does Not Compete with C4BP for Binding to PepO-To
address the question whether C4BP and C1q bind simultaneously to pneumococci and whether binding sites for these two proteins on PepO are independent of each other, a constant concentration of C4BP (10 g/ml) was added together with increasing concentrations of C1q or a constant concentration of C1q (1 g/ml) along with increasing amounts of C4BP, using PepO as bait. C4BP and C1q were detected using specific Abs. Interestingly, C4BP binding to PepO was not affected by the presence of increasing concentrations of C1q (Fig. 6E). Similarly, C4BP at increasing amounts did not influence the binding of C1q to the immobilized PepO (Fig. 6F). Taken together, the two proteins do not compete with each other for binding to PepO, indicating that C4BP and C1q can bind simultaneously to PepO.
Inhibition of C1q-mediated Pneumococcal Adherence to Host Cells by PepO-We have recently shown that C1q facilitates pneumococcal adhesion to host cells (22). To confirm the role of PepO in this process, blocking experiments were performed. First, the binding of C1q to pneumococci was measured by flow cytometry in the presence of PepO. Flow cytometry analyses indicated a dose-dependent competitive inhibition of C1q binding to pneumococci by PepO (Fig. 7, A and B). In contrast, BSA showed no inhibitory effect (Fig. 7B). Similar inhibition was observed using serum as a source of C1q (Fig. 7, C and D). In cell culture infections, PepO was assessed for its ability to inhibit C1q-mediated adherence of pneumococci. Host cells were infected with pneumococci that had been pretreated with a mixture of C1q and PepO. The results showed that PepO inhibited C1q-mediated pneumococcal adhesion to host cells (Fig. 7, E and F). These inhibition experiments thus confirmed that PepO binds C1q and facilitates bacterial adherence to host cells.
Influence of PepO on Bacterial Opsonization-We have previously demonstrated that the survival of an isogenic pepO mutant strain in whole blood is significantly reduced as compared with the wild-type strain (10). To corroborate the results, C3b surface deposition on the wild-type and the mutant strain from human serum was determined. As expected, C3b deposition on the isogenic pepO mutant strain was significantly higher, at all the serum concentrations tested, as compared with the wild-type strain (Fig. 8A). However, the presence of added PepO significantly reduced the C3b deposition (Fig. 8B). Representative sensorgrams are presented. F, various concentrations of PepO were preincubated with 5 g/ml C1q and added to the plates with immobilized aggregated human IgG. BSA and MID were used as controls. Bound C1q was detected with specific Abs. The amount of bound C1q in the absence of inhibitor was set as 100%. G, microtiter plates were coated with PepO, and the effect of different concentrations of NaCl on binding of plasma-purified C1q (5 g/ml) was analyzed. Aggregated human IgG and BSA were used as positive and negative control, respectively. Bindings were detected with specific Abs. The data represent means Ϯ S.D. of three independent experiments performed in duplicates. H, pneumococcal surface-exposed PepO functions as a C1q receptor. Pneumococcal wild-type strain D39 and its isogenic pepO mutant D39⌬pepO were incubated with 0, 5, and 10 g/ml concentration of purified C1q for 60 min at 37°C. Binding of C1q was analyzed by flow cytometry. Results were expressed as GMFI (means Ϯ S.D.) from three independent experiments. Statistical significance was calculated using two-way ANOVA test. ns: not significant; *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001.
In addition, using an E. coli DH5␣ model system, the bactericidal activity assay of human serum was tested to assess the effect of PepO on bacterial survival. The pretreatment of human serum with PepO, but not with the control protein (A1AT), rescued the E. coli from the complement-mediated lytic activity of the serum (Fig. 8C). In summary, our data identified two new distinct functions of PepO, one as an adhesin promoting bacterial colonization by utilizing a host protein as a bridging molecule, and another as a complement modulator (Fig. 9).

DISCUSSION
The innate immune system, and in particular the complement system, plays a crucial role in maintaining tissue homeostasis and is an important first line of host defense mechanism against invading pathogens. However, during the course of evolution, many pathogens including the Gram-positive S. pneumoniae have evolved strategies to combat the host comple-ment-mediated attack. Such strategies are not only relevant for successful colonization but also for the survival and dissemination of the pathogen within the human host. Therefore, a detailed understanding of the underlying mechanism of host cell colonization and escape of immune attack is necessary to develop new treatments.
We have recently identified and characterized a ubiquitously expressed pneumococcal PepO. PepO is a multifunctional protein that interacts with plasminogen and fibronectin and facilitates pneumococcal adherence and invasion of host cells (10). Besides being present on the bacterial cell surface, PepO is also found in the culture supernatants. Indeed, we could estimate that 1 ml of early stationary phase culture of S. pneumoniae strain NCTC10319 and D39 contains about 0.33 Ϯ 0.05 and 0.45 Ϯ 0.15 g of PepO protein, respectively. These concentrations, for respective strains, are ϳ3.7-17.8-fold higher than the concentration present in early mid-log phase as estimated previously (10). Although the exact mechanism of PepO surface FIGURE 5. C4BP binding to PepO. A-C, PepO was immobilized, and increasing amounts of NHS (A), heat-inactivated serum (B), or plasma-purified C4BP (C) were added. BSA was used as control. Bound C4BP was detected using specific polyclonal Abs. Statistical significance was calculated using the two-way ANOVA test. ***, p Ͻ 0.001. D, binding of PepO to immobilized C4BP as analyzed by surface plasmon resonance. Increasing concentrations of PepO (22-1500 nM) were injected onto a C4BP-coated CM5 sensor chip. The amount of PepO associating with the C4BP was measured in response units. Representative sensorgrams are presented. E, the effect of different concentrations of NaCl on binding of plasma-purified C4BP (10 g/ml) to PepO was analyzed. Specific polyclonal Abs detected bound C4BP. BSA was used as negative control. One-way ANOVA test was performed to calculate statistical differences of the groups as compared with the binding at 150 mM NaCl. ***, p Ͻ 0.001. F, Pneumococcal wild-type strain D39 and NCTC10319 and their respective isogenic pepO mutants were incubated with 100 kcpm 125 I-C4BP for 1 h at 37°C (where kcpm is kilo counts per minute). Binding was detected using a Wizard 2 gamma counter. Statistical significance was calculated using Student's t test. *, p Ͻ 0.05; **, p Ͻ 0.01. Data are presented as means Ϯ S.D. of three independent experiments done in duplicates.
presentation, secretion, or presence into the culture supernatant is not clear, the localized concentration is expected to increase severalfold as pneumococci undergo autolysis. Importantly, PepO also plays an important role in pneumococcal survival in whole blood (10). In this study, we tested whether the secreted form of PepO has any additional role in pneumococcal infection, and in particular, whether it has any complement modulatory activity. Indeed, here we have attributed a new function to PepO, as a classical complement pathway inhibitor facilitating pneumococcal complement escape.
We observed that PepO specifically inhibits the classical pathway of complement in both hemolytic and complement deposition assays. This was attributed to the binding between PepO and C1q, the recognition subunit of the first component of the classical pathway, the C1 complex. PepO bound plasmapurified C1q in a dose-dependent manner, which was independent of the presence of calcium. A similar interaction was observed when using C1 complex as a source of C1q. In line with this, PepO inhibits the binding of plasma-purified C1q as well as deposition of C1 from NHS onto human IgG-coated plates. A similar inhibition in C1q and C3b deposition on the pneumococcal surface was observed from PepO-treated NHS, indicating an additional complement evasion strategy. Interestingly, the activation of the classical pathway has been suggested to be a major host defense mechanism during pneumococcal infections (34). Therefore, employing strategies to circumvent this host defense mechanism are critical for pneumococcal survival.
The mechanism by which soluble PepO inhibits the classical pathway is related to the PepO-C1q interaction, which triggers strong complement activation via the classical pathway, leading to complement consumption. This was evident by the fact that the presence of PepO significantly inhibited the hemolytic activity as well as the bactericidal activity of the human serum. However, no such activation was observed for the lectin pathway, but PepO activated the alternative pathway by a yet C4BP ␣-chain deletion mutants lack single CCP domains (represented by circles with X in each ␣-chain). D, the binding of C4BP variants to immobilized PepO was analyzed. Bound C4BP was detected with polyclonal Abs. The graph represents data from two independent experiments done in duplicates Ϯ S.D. E and F, simultaneous binding of C1q and C4BP to PepO immobilized on microtiter plates. A constant amount of C4BP (10 g/ml) together with increasing amounts of C1q (E) or constant amount of C1q (1 g/ml) with increasing amounts of C4BP (F) was added. Bound C4BP and C1q were detected using specific Abs. Data presented are from three independent duplicate experiments Ϯ S.D. Statistical significance was calculated using one-way ANOVA and Dunnett's post-test. ns: not significant; *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001. unknown mechanism, which warrants further investigation. The strong level of classical complement pathway activation suggested this interaction to be similar to that observed for C1q and antibodies clustered on their target. In addition, the PepO-C1q interaction is strongly influenced by the ionic strength, similar to the C1q-antibodies interaction. Taken together, the data suggest that C1q interacts primarily via its C-terminal globular heads with PepO. Indeed, this is in accordance with our recent finding, that C1q interacts directly in an antibodyindependent manner via its C-terminal globular heads with the pneumococcal surface-presented protein(s) and is functionally active (22). Apart from PepO, C1q will likely interact with other bacterial proteins as well as C-reactive protein and specific antibodies that both bind to the bacterial surface. NHS used in this study was obtained from healthy donors, who are likely to have encountered pneumococci and carry specific antibodies.
Interestingly, PepO can also recruit the classical and lectin complement pathway inhibitor C4BP, thereby keeping a check on their activation at the level of C3 convertases. We observed a dose-dependent interaction between PepO and C4BP at physiological ionic strength from NHS and heat-inactivated serum. Moreover, similar binding was observed with C4BP purified FIGURE 7. Blocking of C1q binding to pneumococci and C1q-mediated adherence to host cells by PepO. A-D, competitive inhibition experiments. Binding of plasma-purified C1q (A and B) to pneumococci (NCTC10319) or from NHS (0.5%) (C and D) was measured in the absence of exogenous added PepO proteins. BSA was used as negative control. Bound C1q was analyzed by flow cytometry using specific Abs. A representative flow cytometry histogram from three independent experiments is shown (A and C), and GMFI values (D) or normalized to the percentages (B) from three independent experiments performed in duplicates are presented. Statistical significance was calculated using two-way ANOVA test, **, p Ͻ 0.01; ***, p Ͻ 0.001. E and F, in cell culture blocking experiments, C1q-mediated adherence of pneumococci (NCTC10319) to A549 lung epithelial cells in the absence and presence of PepO (10 g) was analyzed. E, adherence was determined by counting the cfu per well obtained from sample aliquots plated onto blood agar plates after 3 h of infection. F, representative immunofluorescence microscopy image of adherent pneumococci to A549 cells. Bar is equal to 10 m. Two-way ANOVA test was performed to determine the statistical difference between the groups. Results present the means Ϯ S.D. of at least three independent experiments. ns: not significant; **, p Ͻ 0.01 relative to infections carried out in the absence of C1q. from plasma, mediated via CCP8 of the ␣-chain, suggesting PepO-C4BP binding to be a direct interaction that does not require any additional serum factor(s). Both C1q and C4BP are abundant plasma proteins, and it seemed highly possible that they may compete with each other for binding to PepO. However, our results indicate that binding of C4BP is not outcompeted by C1q and vice versa, suggesting that both proteins interact with different motifs on pneumococcal PepO.
Colonization of host cells is an important step in the pathogenesis of any infectious agents including pneumococci, which have developed multiple strategies such as expression of adhesins or acquirement of various host proteins to facilitate colonization (12,18,(22)(23)(24)(25)(26)(27). We have recently reported a novel mechanism of pneumococcal-host interaction whereby C1q, which is primarily involved in the initiation of the classical complement pathway, directly interacts with pneumococci and functions as a molecular bridge facilitating pneumococcal colonization of host cells (22). However, not much is known about C1q ligand(s) on the pneumococcal surface. A recent study suggests pneumococcal GAPDH as a potential C1q ligand (35). Although recombinant PepO binds C1q, we asked whether bacterial surface-presented PepO could function as a ligand for C1q. Indeed, pneumococcal surface-presented PepO interacts with C1q. This was evident by the fact that the pepO mutant of highly encapsulated D39 serotype-2 strain showed significant reduction in binding of C1q. Additionally, our data clearly demonstrated that, as a C1q ligand, PepO facilitates C1q-mediated pneumococcal adherence to host epithelial cells. However, it is intriguing whether C1/C1q bound directly to pneumococcal surface-presented PepO would favor complement activation or C1qmediated bacterial host cell adherence under in vivo conditions. It is reasonable to suggest that both mechanisms may occur.
In summary, we have demonstrated that PepO is indeed a multifunctional protein playing an important role in pneumococcal infections. Interestingly, these data suggest a differential effect of surface-presented and secreted PepO. Although the surface-presented PepO promotes C1q-mediated pneumococcal host cell adherence, the localized release consumes complement as a result of its activation following C1q interaction, thus Pneumococcal wild-type strain D39 and its isogenic pepO mutant D39⌬pepO were incubated with the indicated concentrations of NHS in GVB ϩϩ buffer for 15 min at 37°C. The bacteria were thereafter washed and incubated with a FITC-conjugated anti-C3c Ab followed by flow cytometry analysis. Less C3b was deposited on the wild-type strain as compared with the mutant. B, in inhibition experiments, C3b deposition was investigated on the surface of mutant bacteria in the absence or the presence of PepO. Human serum albumin was used as a negative control. Results were expressed as GMFI (means Ϯ S.D.) after subtracting the Ab background from two independent experiments performed in duplicates. C, PepO inhibits the bactericidal activity of human serum. E. coli DH5␣ (1000 cfu) were incubated for 30 min at 37°C with 0.2% NHS pretreated with increasing concentrations of PepO or control protein A1AT, and the surviving bacteria were enumerated after overnight culture on LB agar plates. The survival was expressed as the percentage of inoculum. The data represent the means Ϯ S.D. of three independent experiments performed in duplicates. Statistical significance was calculated using a two-way ANOVA test. ns: not significant; *, p Ͻ 0.05; ***, p Ͻ 0.001. The surface-presented PepO promotes C1q-mediated pneumococcal host cell colonization, and the secreted form consumes complement as a result of its activation following C1q interaction and thus helps in complement evasion. Furthermore, PepO binds the complement inhibitor C4BP, which when occurring at the bacterial surface further attenuates complement deposition. C4BP may also bind secreted PepO, but this does not compete out binding of C1q.
representing an additional complement escape mechanism employed by this versatile pathogen.