The Group B Streptococcus–Secreted Protein CIP Interacts with C4, Preventing C3b Deposition via the Lectin and Classical Complement Pathways Open Access

Streptococcus agalactiae (group B Streptococcus [GBS]) colonizes the lower gastrointestinal and vaginal mucosae of about one third of women and can cause neonatal pneumonia, sepsis, and meningitis (1, 2). It is also an important etiological agent of morbidity in immunocompromised adults and of bovine mastitis (3). Both during colonization and in the infection stage, GBS bacteria are faced with the host innate immune defense, and one of the first barriers they encounter is the complement system. Several complement effector molecules can indeed sense and opsonize Gram-positive bacteria such as GBS and promote their phagocytic killing by neutrophils and macrophages (4).

The process of complement fixation can occur by three activation routes, the classical pathway (CP), the lectin pathway (LP), and the alternative pathway (AP), differing in their target recognition mechanisms and effector molecules. All three proteolytic cascades lead to cleavage of C3 and subsequent formation of the C3a anaphylatoxin and the C3b opsonin. C3a attracts and activates granulocytes, whereas C3b attaches covalently to the bacterial surface, amplifies complement activation, and labels cells for phagocytosis. Activation of the CP is initiated after C1q molecules are deposited on the bacterial surface via direct recognition, Ig binding, or pentraxins bridging and interact with C1r and C1s proteases to form the C1 proteolytic complex. Through the LP pathway, mannan-binding lectin or other lectins bind to microbial surface polysaccharides resulting in activation of mannan-binding lectin–associated serine protease (MASP). Both of the CP and LP proteolytic complexes can split surface-bound C4 into C4a plus C4b, and C2 into C2b plus the C2a protease. C4b and C2a directly interact to form the C3 convertase C4bC2a that cleaves native C3 into C3b. Surface-bound C3b is in turn the precursor of C3bBb, the AP C3 convertase that transforms new C3 molecules into C3b and C3a, thus greatly amplifying the number of C3b molecules opsonizing the bacteria and consequently phagocytic killing.

Three important host regulators controlling complement homeostasis are C3b-cleaving factor I, factor H, which acts as a cofactor of factor I and can also compete with factor B to displace Bb from the AP C3bBb convertase, and C4b-binding protein that interferes with the assembly of the CP/LP C4bC2a convertase.

Bacterial pathogens have evolved a series of innate defense evasion molecules that can block the complement proteolytic cascades or divert them to overcome immune clearance by the host. In the case of GBS, a prominent role in complement evasion is played by the thick capsular polysaccharide that surrounds the bacterial cell wall. Almost all GBS strains associated with human disease are encapsulated, belonging to 1 of 10 capsular types recognized by specific Abs: Ia, Ib, and II–IX. The 10 GBS capsular polysaccharide structures are created by diverse arrangements of galactose, glucose, N-acetylglucosamine, and sialic acid into unique repeating units that invariably contain sialic acid on their branching terminus (5). Type III GBS, frequently found in neonatal invasive infections, expresses a large amount of capsular polysaccharide that was shown to inhibit activation of the complement AP in adult sera deficient in specific Abs (6), whereas AP inhibition could be overcome by anti–type III polysaccharide IgG (7). Furthermore, a mutant strain expressing a capsule devoid of sialic acid showed a markedly decreased virulence (6, 8). Marques et al. (9) confirmed the critical role of the sialic acid–containing capsule in preventing C3b deposition, presumably through acquisition of factor H from host plasma, resulting in cleavage of C3b to iC3b by factor I and interruption of the AP C3b amplification loop.

Some of the GBS cell wall surface–anchored proteins containing a leucine-proline-any-threonine-glycine motif contribute to inhibition of bacterial phagocytic killing by blocking complement activation. These include the β protein, which interacts with the factor H complement regulator (10) and sialic acid–binding Siglec-5 lectin expressed on leukocyte surfaces (11), the histidine triad protein STH, which also binds factor H promoting complement degradation (12), and BibA, a group B Streptococcus immunogenic bacterial adhesin that specifically binds to human C4b-binding protein (13).

Studies in the last decade have revealed that another Gram-positive pathogen, Staphylococcus aureus, produces soluble proteins interfering with the activation of the complement system such as staphylococcal complement inhibitor (14), extracellular fibrinogen-binding protein (Efb) (15, 16), and extracellular adherence protein (Eap) (17). We hypothesized that S. agalactiae might secrete similar yet unidentified complement regulators assisting the bacteria to escape phagocytic killing. With this in mind, we screened the GBS genome for the presence of potentially secreted proteins that could display inhibitory activities on one or more complement pathways. A low molecular mass protein interfering with the CP and the LP was identified, and we could demonstrate its capacity to bind the C4 complement factor and prevent the formation of the C4bC2a CP/LP proconvertase and GBS phagocytic killing in the absence of anti-GBS Abs.

Human serum proteins C3b, C4b, and C2 were obtained from Calbiochem (Merck, Darmstadt, Germany). Factor B was from CompTech (Tyler, TX). BSA was purchased from Sigma-Aldrich (St. Louis, MO). Blood for serum preparation and for phagocytosis assays was drawn from healthy adult volunteers after informed consent and approval of the protocol by the Medical-Ethical Committee of the University of Pavia were obtained. Blood was allowed to clot for 15 min at 22°C, and serum was collected after centrifugation for 10 min at 1500 × g at 4°C and subsequently stored at −80°C. Complement factor–depleted serum was obtained from CompTech.

Escherichia coli BL21(DE3) (Stratagene, La Jolla, CA) was used as host for expression of recombinant proteins and was grown in Luria broth containing ampicillin (100 μg/ml). GBS clinical isolates COH1 (serotype III) (18), 515 (serotype Ia) (19), 2603 (serotype V) (20), 6313 (serotype III) (21), 383728 (NT), ES-NI-010 (serotype III), and SH0248 (NT) (22) were used in this study. Bacteria were grown at 37°C in Todd-Hewitt broth (Becton Dickinson, Sparks, MD) or in modified M9 medium (28 mM Na₂HPO₄, 22 mM KH₂PO₄ [pH 7.4] containing 8.5 mM NaCl, 18.7 mM NH₄Cl, 2 mM MgSO₄, 1 mM CaCl₂, 0.2% glucose, 0.3% yeast extract, and 1% casamino acids).

DNA encoding complement interfering protein (CIP) was amplified by PCR using chromosomal DNA isolated from GBS COH1 as template. To amplify DNA encoding CIP, forward (5′-GGAATTCCTAGCTAGCAAGAGTGATGGCATCTC-3′) and reverse (5′-GGAATTCCCG CTCGAGTCTAAAACTATCTTTTATTACTTT-3′) primers were used. Restriction enzyme cleavage sites NheI and XhoI were incorporated at the 5′ ends of the primers to facilitate cloning into the pET21b(+) expression plasmid (Novagen, Podenzano, Italy).

An overnight starter culture of the recombinant E. coli expressing CIP was diluted 1:50 in Luria broth and incubated at 37°C with shaking until the culture reached OD₆₀₀ of 0.6–0.8. Recombinant protein expression was induced by addition of isopropyl 1-thio-β-d-galactopyranoside (0.2 mM) and continued for 4 h. Bacteria were harvested by centrifugation at 1700 × g for 20 min and lysed by passage through a French press. The cell debris was removed by centrifugation (20,000 × g), and the filtered supernatant (0.45-μm membrane) was applied to a 1 ml Ni²⁺-Sepharose HisTrap HP column (GE Healthcare, Buckinghamshire, U.K.). The protein was eluted with 29 column vol 0.00–500 mM imidazole (Sigma-Aldrich) gradient in 0.5 M NaCl, 20 mM sodium phosphate (pH 7.4). Fractions corresponding to the recombinant protein were pooled and dialyzed against PBS. Protein concentrations were determined with a bicinchoninic acid protein assay kit (Pierce, Rockford, IL). Recombinant fibrinogen-binding protein 3 (Fib3) was expressed and purified as previously described (23).

Goat polyclonal anti-human C4 and rabbit polyclonal anti-human C3 Abs were purchased from Abcam (Cambridge, U.K.). HRP-conjugated secondary Abs were from Dako (Glostrup, Denmark). Polyclonal rabbit anti-CIP Ab and mouse anti–group B streptococcal surface immunogenic protein (Sip) Ab were obtained by immunizing New Zealand rabbits and CD1 mice with three doses of recombinant protein (20 μg). Protocols were approved by the Italian Ministry of Health (authorization 110/2012-B) and by the local Novartis Animal Welfare Body (authorization AWB 201114).

Wielisa assays (24) were conducted according to the manufacturer’s instructions (Wieslab, Malmö, Sweden). Briefly, normal human serum (1–6 μl) was preincubated for 30 min at 22°C in 100 μl assay buffer containing 0–200 nM of CIP or Fib3 and added to wells coated with IgM, LPS, or mannan. Complement activation was detected by using either alkaline phosphatase–labeled anti–C5-9 complex or anti-C3 (1:2000)– or anti-C4 (1:2000)–specific sera, followed by incubation with HRP-labeled secondary Abs (1:1000). Heat-inactivated serum was used as a negative control.

The EZ Complement CH50 clinical diagnostic assay kit (Diamedix, Miami, FL) was used as recommended by the manufacturer. Briefly, 5 μl human serum was preincubated for 1 h with increasing doses of CIP (0–350 nM) or Fib3 (350 nM) and then incubated at 22°C for 1 h with Ab-coated sheep RBCs (3 ml). Cells were then centrifuged (800 × g for 10 min) and the absorbance of the supernatants was measured at 405 nm to determine the percentage lysis in each sample. The data are expressed as percentage RBC lysis in CIP preincubated compared with nonpreincubated samples.

Microtiter wells were coated with 100 ng CIP and incubated overnight at 4°C in 50 mM carbonate buffer (pH 9.5). The wells were washed three times with PBS supplemented with 0.1% (v/v) Tween 20 (PBST), blocked with 2% BSA in PBST for 1 h at 22°C, and then probed with serial dilutions of normal human serum in PBS, followed by incubation with goat anti-C4 Ab (1:2000) and HRP-conjugated rabbit anti-goat IgG (1:1000) and detection of HRP enzymatic activity.

SDS-PAGE was performed on 12.5% polyacrylamide gels stained with Coomassie brilliant blue (Bio-Rad, Hercules, CA). For the Western blot assay, CIP was subjected to SDS-PAGE and then electroblotted onto a nitrocellulose membrane (GE Healthcare). The membrane was treated with a solution containing 5% (w/v) dried milk in PBS, washed, and incubated with 1% human serum for 1 h at 22°C. Following additional washings with PBST, the membrane was incubated for 1 h with an anti-C4 goat polyclonal Ab (1:5000). After several washings, the membrane was incubated with HRP-conjugated rabbit anti-goat IgG (1:10000), followed by detection of HRP enzymatic activity.

Surface plasmon resonance (SPR) to estimate the affinity of the interaction between C4b and CIP was conducted using a BIAcore X100 instrument (GE Life Sciences). Purified human C4b was covalently immobilized on dextran matrix CM5 sensor chip surface by using a C4b solution (30 μg/ml in 50 mM sodium acetate buffer [pH 5]) in a 1:1 dilution with N-hydroxysuccinimide and N-ethyl-N′-(3-dimethylaminopropyl)carbodiimide hydrochloride. The excess of active groups on the dextran matrix was blocked using 1 M ethanolamine (pH 8.5). On another flow cell, the dextran matrix was treated as described above but without any ligand to provide an uncoated reference flow cell. The running buffer used was PBS containing 0.005% (v/v) Tween 20. Two-fold linear dilution series (0.078–2.5 μM) of CIP in running buffer were passed over the ligand at the flow rate of 45 μl/min and all the sensorgrams were recorded at 22°C. Surface regeneration was achieved by injecting a solution of 25 mM NaOH. Sensorgrams from three sets of data for each concentration were collected. Association and dissociation kinetics parameters (K_a and K_d) and the equilibrium dissociation constant K_D were estimated with a 1:1 interaction model (Langmuir model) by nonlinear fitting, using BIAevaluation 1.0 software.

GBS 2603 (5 × 10⁷ CFU) grown to stationary phase were coated onto microtiter wells overnight at 37°C in 50 mM carbonate buffer (pH 9.5). The wells were washed, blocked with 2% BSA for 1 h at 22°C, and incubated for 10 min at 22°C with 100 μl of normal or C1q-depleted human sera diluted to 5% with 20 mM HEPES (pH 7.4) containing 140 mM NaCl, 2 mM CaCl₂, 2 mM MgCl₂, and 0.1% (w/v) BSA, or with 20 mM HEPES (pH 7.4) containing 140 mM NaCl, 4 mM MgCl₂, 16 mM EGTA, and 0.1% (w/v) BSA and preincubated with increasing amounts (0–6 μM) of CIP. After washing, C3b deposition was detected by rabbit anti-C3 polyclonal Ab (1:2000) followed by HRP-conjugated goat anti-rabbit IgG (1:1000).

GBS 2603 (5 × 10⁷ CFU) was immobilized as above. Wells were washed, blocked with 2% BSA for 1 h at 22°C, and mixed with purified C3b or C4b (1 μg) preincubated with increasing concentrations of CIP (0–6 μM) in 20 mM HEPES (pH 7.4) with 140 mM NaCl, 2 mM CaCl₂, 2 mM MgCl₂, and 0.1% (w/v) BSA (assay buffer) in a total volume of 100 μl. After incubation for 10 min, unbound components were washed away with assay buffer, and deposited C3b or C4b was detected by rabbit anti-C3 or goat anti-C4 polyclonal Abs (1:2000) and HRP-conjugated secondary Abs (1:1000).

GBS bacteria were grown to stationary phase in M9 medium and centrifuged at 1500 × g. One milliliter culture medium was directly adsorbed onto a nitrocellulose membrane (dot blot) or 5% TCA precipitated and loaded onto SDS-PAGE followed by Western blot. Detection of CIP was performed as indicated in the SDS-PAGE and Western blotting section.

To evaluate the presence of CIP on the surface of GBS, 5 × 10⁷ CFU GBS grown to stationary phase were coated overnight onto microtiter wells at 37°C in 50 mM carbonate buffer (pH 9.5). Immobilized GBS was analyzed for the presence of CIP by addition of rabbit anti-CIP polyclonal sera (1:2000) followed by HRP-conjugated goat anti-rabbit IgG (1:1000). To examine binding of externally added CIP to the surface of GBS, bacteria were immobilized onto microtiter wells and then incubated with 1 μg purified recombinant CIP for 1 h at 22°C. The wells were washed and the total amount of CIP bound to plates determined as reported above.

The ability of GBS to survive in human blood in the presence of 2.5 or 5 μM CIP was measured as previously described (25). Briefly, S. agalactiae strains 6313 and 2603 grown to exponential phase (OD₆₀₀ of 0.3) were diluted in Todd–Hewitt broth, and 100 μl (∼5 × 10³ CFU) was added to 0.9 ml fresh blood obtained from human healthy volunteers and treated with 50 μg/ml of the anticoagulant hirudin (Refludan), (Pharmion, Rome, Italy). In control experiments, cytochalasin (10 μg/ml) was added to the reaction. Tubes were incubated at 37°C with gentle rocking and after 3 h, serial dilutions were plated to determine the number of surviving CFU. In a second set of control experiments, the growth of each strain in human plasma was checked in the presence of 5 μM CIP.

GBS bacteria (5 × 10⁷ CFU) grown to stationary phase or GBS polysaccharides V or III conjugated to HSA (100 ng), prepared as described by Nilo et al. (26), were coated overnight onto microtiter wells at 37°C in 50 mM carbonate buffer and then washed three times in washing buffer (0.05% Tween 20 in PBS). Plates were added with human sera diluted in PBS, 2% BSA, 0.05% Tween 20, incubated at 37°C for 1 h, washed with 0.05% Tween 20 in PBS, and then incubated for 90 min at 37°C with alkaline phosphatase–conjugated anti-human IgG (Sigma-Aldrich) in PBS/2% BSA/0.05% Tween 20. After washing, the plates were developed with para-Nitrophenyl Phosphate (4 mg/ml) and the absorbance was measured at 405 nm. Standard sera with known anti-polysaccharides III and V IgG concentrations were obtained from Baylor College (Houston, TX) (27). Based on these assays, the low titer sera used in our experiments contained <0.5 μg/ml anti-capsular Abs and yielded OD values <0.3 in ELISA with whole bacteria.

Statistical analysis was carried out using GraphPad Prism statistical analysis software. Differences between groups were analyzed by ANOVA with the appropriate posttest and by using repeated measures where required. A p value <0.05 was considered statistically significant.

In the attempt to identify GBS-secreted virulence effectors mediating host complement evasion, we interrogated a library of surface-predicted Ags (28) for the possible presence of proteins devoid of transmembrane or leucine-proline-any-threonine-glycine motifs that could interfere with the activation of one or more complement pathways. We focused our attention on a 153-residue polypeptide showing a partial homology with the staphylococcal secreted complement inhibitors Efb and Eap (∼15% identity and 35% similarity to both proteins) (Fig. 1). The corresponding gene was first annotated as san_2130 (now COH1_1804) in the genome sequence of the GBS serotype III ST-17 strain COH1. The COH1 open reading frame lacking the predicted hydrophobic leader region was cloned into a E. coli high copy number plasmid under a strong promoter. A soluble His-tagged protein of ∼15 kDa was purified with high yields from a positive clone.

The SAN_2130 recombinant protein was tested for its ability to inhibit the human complement pathways, using the Wielisa kit that measures complement activation in human serum by an Ab directed against the C5b-9 terminal complex (24). In the assay, specific activation of the CP, AP, or LP is achieved by coating ELISA wells either with IgM, LPS, or mannan, respectively. As shown in Fig. 2A, preincubation of human serum with increasing amounts of SAN_2130 resulted in a dose-dependent inhibition of the CP and LP, as detected by a decreased deposition of the C5b-9 complex. Inhibition levels reached those obtained when the assay was performed with heat-inactivated serum. None of the complement pathways was affected by the presence of the unrelated GBS protein Fib3, used as negative control (23).

To assess whether any intermediate steps of the complement activation process could be inhibited by SAN_2130, a variant of the Wielisa assay where the presence of C3b or C4b on the plate surface was detected with appropriate specific antisera was set up. In the new conditions, preincubation of human serum with SAN_2130 resulted in a strong reduction of C3b detection in the wells where the CP and LP were tested, whereas no effect on C4b deposition was observed for any of the pathways. As expected, SAN_2130 failed to block deposition of C3b via AP (Fig. 2B).

The complement inhibitory effect of SAN_2130 was further investigated using the CH50 assay, where activation of the classical complement pathway can be examined by measuring hemoglobin release by a standardized suspension of sheep erythrocytes sensitized with specific Abs. Upon addition of SAN_2130, a dose-dependent reduction of erythrocyte lysis was observed, whereas the control protein Fib3 did not show any effect (Fig. 2C).

The obtained data revealed an inhibitory effect of SAN_2130 on human complement activation via CP and LP, and for this reason the protein was named CIP.

The functional studies presented above suggested that CIP inhibited an event mediating activation of C3 via the CP and the LP. We predicted that CIP could act on either the fully assembled CP/LP C3 convertase (C4b2a) or an isolated component thereof. In a Western blot assay, CIP bound to C4 present in EDTA-treated serum, whereas no signal was visualized when a C4-containing membrane was overlaid with a C4-depleted serum (Fig. 3A). This result was confirmed by an ELISA in which C4 from human serum was demonstrated to bind dose-dependently to surface-coated CIP (Fig. 3B). A similar saturable binding was obtained when purified C4 or C4b was immobilized on microtiter wells and incubated with soluble CIP, whereas no significant interaction was detected when increasing concentrations of CIP were incubated with immobilized purified C2 (Fig. 4A). To determine the affinity of the CIP interaction with C4b, a SPR study was conducted. Human C4b immobilized on the surface of a dextran chip was incubated with soluble CIP concentrations ranging from 0.078 to 2.5 μM. The GBS protein bound to C4b in a dose-dependent manner, with a measured apparent K_D of 95.3 ± 5.2 nM (Fig. 4B).

Formation of the CP/LP C3 convertase is a stepwise process that starts with the deposition of surface-bound C4b. Although C4b has no enzymatic activity on its own, it serves as a molecular scaffold binding to C2 to yield the C4bC2 proconvertase and for C1s/MASP-dependent cleavage of C2 to generate the fully active C4b2a convertase. Therefore, we examined the effect of CIP on the interaction between purified C4b and C2 in an ELISA format.

As reported in Fig. 5A, CIP inhibited binding of C4b to surface-coated C2 in a dose-dependent fashion. Conversely, no interfering effect by CIP was observed on the binding of Bb to immobilized C3b (Fig. 5B). Taken together, these data suggested that binding of CIP to C4b constitutes the molecular basis of the specific inhibition of both CP and LP activity and confirmed that the streptococcal protein does not affect the formation of the C3bBb convertase.

We subsequently examined the expression and localization of the CIP protein across eight sequenced GBS strains of human or bovine origin (22), six of which contain the cip open reading frame in their genome.

To assess secretion of the CIP protein, bacteria were grown to stationary phase and culture supernatants were spotted onto a nitrocellulose membrane, followed by incubation with anti-CIP sera. As shown in Fig. 6A, different amounts of protein were secreted by all the strains carrying the gene, whereas no signal was detected on the spots corresponding to the two cip⁻ isolates. Conversely, the unrelated and highly conserved GBS protein Sip was uniformly expressed in all the strains tested. The specificity of the dot blot was further confirmed by a Western immunoblotting assay, where a protein with a molecular mass corresponding to CIP was detected in the supernatants of cip⁺ but not in cip⁻ supernatants (Fig. 6B). Remarkably, all the above reported results were obtained when a chemically defined medium added with 0.3% yeast extract was used for GBS growth, whereas no expression of CIP was detected using Todd–Hewitt broth or other rich media (unpublished results), suggesting tight regulation of CIP expression.

To investigate whether CIP could also be found on the GBS surface, bacteria grown to stationary phase were directly spotted onto ELISA microtiter wells, followed by incubation with anti-CIP and HRP-conjugated secondary Abs. As shown in Fig. 6C, the strains containing the cip gene responded positively to the Abs, whereas no signal was detected on the cip⁻ bacteria. Notably, externally added CIP to immobilized strains resulted in highly increased signals, indicative of the protein binding to the surface of the bacteria. CIP bound to all GBS strains irrespective of the presence of the corresponding gene in their genome, suggesting that capturing of CIP is an intrinsic property of the GBS surface. Overall, the data suggested that CIP is secreted by the bacteria and can rebind to the GBS surface.

We sought to determine whether the interaction of CIP with C4 could have an impact on the deposition of C3b from human serum on the GBS cell surface, possibly by inhibiting the formation of the CP/LP C4b2a convertase. This was examined incubating the type V GBS 2603 strain lacking the cip gene, with human serum in the presence of increasing concentrations of CIP. For these experiments we used a normal human serum containing low levels of Abs against GBS 2603, as assessed by ELISA.

The serum was diluted with Mg²⁺/Ca²⁺- or Mg²⁺/EGTA-containing buffers, preincubated with increasing amounts of CIP or with buffer alone, and the mixtures were added to microtiter wells coated with GBS 2603. After washing, the plates were examined for the presence of C3b bound to bacteria with anti-C3 IgG. As shown in Fig. 7A, in the absence of CIP, OD values of ∼0.7 were detected when only the AP was operational (buffer containing Mg²⁺/EGTA), whereas OD saturating values of 2.5 were measured with serum containing Mg²⁺/Ca²⁺ ions. This finding was consistent with previous reports indicating that, in the absence of specific Abs, GBS is mainly susceptible to the classical and/or the lectin pathways, whereas its sialic acid–containing capsule renders the organism resistant to the complement AP (6). Remarkably, addition of increasing concentrations of CIP in Mg²⁺/Ca²⁺ buffer progressively reduced C3b deposition to levels comparable to those observed when serum was incubated with Mg²⁺/EGTA, either in the absence or presence of CIP (Fig. 7A). In summary, CIP appeared to inhibit C3b deposition via the classical and/or the lectin complement pathways without affecting the AP.

To understand which of the two pathways (CP or LP) was more relevant in the explored experimental setting, we tested the potential activity of CIP in the same assay using a C1q-depleted low anti-GBS Ab titer serum. The results were equivalent to those reported in Fig. 7A in terms both of C3b deposition in Mg²⁺/EGTA- or Mg²⁺/Ca²⁺-containing buffers, and of CIP-promoted inhibition levels (Fig. 7B). These data pointed toward a key role of the LP in C3b deposition on the GBS surface in the absence of specific Abs, and confirmed the interfering effect of CIP on the LP activation.

To distinguish whether the observed CIP effect was a consequence of the inhibition of C3b formation by the LP C4bC2a convertase rather than the result of a direct interference on C3b deposition, purified C3b was preincubated with increasing amounts of soluble CIP and the mixture was added to wells coated with GBS 2603. As shown in Fig. 8A, the effect of soluble CIP on bacterial deposition of C3b was negligible even at the highest concentrations. A similar observation was made when the CIP effect on C4b deposition was examined.

Expanding this analysis, CIP was first preincubated with immobilized GBS 2603 to allow binding of the protein to the bacterial surface, followed by addition of C3b or C4b. No decrease in C3b/C4b deposition was detected either of these new experimental conditions (Fig. 8B).

Together with the former observation that CIP blocks the interaction of the purified C4b with C2, these results definitively confirm that CIP interferes with C3b formation via the C4b2a convertase, rather than preventing C3b or C4b deposition.

To gain additional insights on the involvement of CIP in GBS pathogenesis, experiments were designed aimed at assessing the effect of CIP on bacterial survival in human whole blood in the absence of anti-capsular Abs. Blood samples from selected donors with very low anti-GBS Ab titer were used for this analysis. Recombinant hirudin was used as anticoagulant to preserve complement activity. A dose-dependent inhibition of killing of the cip⁻ GBS strains 6313 (Fig. 9A) and 2603 (Fig. 9B) in the presence of 2.5 or 5 μM CIP was observed, whereas no effect was detected with the protein Fib3. In a control experiment, bacteria were incubated for 3 h in blood and added with cytochalasin D to block phagocytosis. The treatment almost completely abolished phagocytosis, demonstrating the involvement of actin cytoskeletal rearrangements in the process. To further exclude the possibility that the observed CIP effect could be somehow related to the alteration of GBS growth, streptococci were grown in plasma (absence of blood cells) in the presence/absence of CIP. In both conditions GBS proliferated and grew to the same extent (data not shown). These experiments indicated that CIP can play a role in the ability of GBS to survive in human blood by preventing phagocytosis.

The effective uptake and killing of GBS by host phagocytic cells require opsonization of the bacterium by specific Abs and/or the complement system. To date most of the work describing the interactions between S. agalactiae and complement components has primarily focused on the role of sialic acid–rich capsule (6) and factor H–binding proteins (12, 29) in the prevention of C3-mediated opsonization through inhibition of the AP. Low levels of maternal anti-capsular polysaccharide IgG correlate with an increased neonatal susceptibility to GBS infection (30), and a higher amount of these Abs increases the efficiency of GBS killing by polymorphonuclear phagocytes via a complement-dependent mechanism involving the AP (31, 32).

Serum deficient in anti-capsular IgG can still trigger significant phagocytic killing of GBS by a process involving complement activation that requires Ca²⁺ and Mg²⁺ and is hence AP-independent (32, 33). It is therefore presumable that these bacteria have evolved additional virulence factors counteracting the activation of the CP or LP.

In this study, we describe, to our knowledge for the first time, a proteinaceous factor secreted in the culture medium of GBS, which we named CIP, that forms a high-affinity complex with C4b and can block the assembly of the CP/LP C4bC2a proconvertase, resulting in impaired C3b deposition. In fact, CIP bound C4b in a saturable manner with high affinity. The C4bC2a inhibitory effect of CIP was highly specific, as there was no interference on the binding of Bb to immobilized C3b (AP convertase), nor on the deposition of purified C3b and C4b on the GBS surface.

Even though CIP could in principle hamper both CP and LP activation, when we tested its effect in sera with low titers of anti-GBS IgG, we observed inhibition of the formation/deposition of C3b irrespective of the presence of C1q. The data confirm previous observations indicating that C1q binding to the GBS surface (34) is per se not sufficient to mediate Ab-independent opsonophagocytic killing (35). Recent studies highlighted a role of the LP L-ficolin/MASP complexes in the formation of the C4bC2a convertase on the GBS surface, initiation of C3b deposition, and phagocytic killing in the absence of specific Abs (36), a process that can be further amplified by anti-capule polysaccaride IgG and the AP convertase. The same authors suggested that deficiencies in L-ficolin in cord serum could be a risk factor for neonatal GBS infection (37). We concluded that, in absence of specific Abs, CIP could interfere with LP complement amplification. Interestingly, Ali et al. (38) recently demonstrated that the LP L-ficolin is also a critical component of the innate immune response to pneumococcal infection.

Recent preliminary evidence indicates that CIP can also bind C3b (unpublished results). However, as reported in Figs. 2 and 5B, this interaction does not prevent the formation of the AP convertase, nor C3b deposition on the bacterial surface via AP (Fig. 6). The biochemical characterization of CIP binding to C3b and the biological implications of this new CIP interaction are currently under investigation in our laboratories.

Searching for CIP homologs in other bacterial species, we found that a subset of strains from another human streptococcal pathogen, the group A Streptococcus (GAS), expresses a protein displaying 46% identity with CIP. The GBS cip gene and its GAS homolog are present in a similar phage-derived genomic region named RD2 that was acquired by horizontal transfer and is integrated into a tRNA gene flanked by direct repeats. In addition to CIP, RD2 encodes several proteins with predicted secretion signal sequences, among which the R28 protein that has been implicated in host–pathogen interactions (39). Interestingly, in GAS this region is present in all serotype M28 strains and in strains of other serotypes associated with maternal–fetal urogenital infections (40, 41). In the case of GBS, analysis of the 373 genomes present in the NCBI database revealed 80 cip⁺ strains mainly belonging to the hypervirulent type III ST-17, ST-23, and to the bovine ST-61-67.

Although CIP is secreted, the purified soluble protein can bind to the bacterial surface. Rebinding of CIP to GBS occurs both on cip⁺ and cip⁻ strains, suggesting that the GBS cell wall peptidoglycan contains yet undetermined components that can recognize and bind to CIP. Thus, we speculate that in coculture conditions cip⁺ GBS strains can transfer to cip⁻ strains the ability to neutralize complement system.

The mechanism by which CIP interferes with CP and LP activation is reminiscent of the recently described effect of the Eap-secreted protein expressed by S. aureus. Indeed, Woehl et al. (17) demonstrated a direct nanomolar affinity interaction of Eap with C4b and consequent inhibition of C4bC2 assembly. The similarities between CIP and Eap go beyond the mechanism by which they disrupt the formation of the CP/LP proconvertase. In fact, CIP and Eap display partial homology, both are secreted in the culture medium and rebind to the bacterial surface, and they ultimately interfere with the deposition of C3b and phagocytic killing.

The contribution of CIP to GBS virulence by using isogenic mutant strains remains to be investigated. Woehl et al. (17) observed that a knockout mutant carrying an in-frame deletion of Eap did not show reduced levels of bacterial C3b deposition or phagocytosis compared with wild-type, whereas a protective effect was achieved when the protein was added exogenously, as we observed for GBS CIP. The data suggested that exogenous Eap, but not surface-retained Eap, could significantly contribute to S. aureus complement evasion.

In conclusion, the results obtained in our studies provide new insights into the mechanisms of GBS immune evasion and uncover a novel strategy of CP/LP regulation that may hold significant implications for prophylactic/therapeutic interventions against overwhelming infections caused by this pathogen.

We acknowledge Monica Fabbrini and Alessandra Acquaviva for ELISA experiments to determine anti-GBS titers in the human sera used for our study and Francesco Berti, Barbara Brogioni, and Pala Lo Surdo for support on SPR experiments. We thank the investigators who provided the GBS strains ES-NI-010 (Javier Rodriguez, DEVANI program, European Commission Seventh Framework no. 200481), 2603 (Graziella Orefici, ISS, Rome, Italy), COH1 (Dennis Kasper, Brigham and Women's Hospital, Boston, MA), 515 (Carol Baker, Baylor College of Medicine, Houston, TX), and SH0248 and 383728 (Mario Rodriguez, University of Lisbon, Lisbon, Portugal).

I.M., R.R., and S.B. are employees of GSK Vaccines S.r.l. The remaining authors have no financial conflicts of interest.