Streptococcal M5 Protein Prevents Neutrophil Phagocytosis by Interfering with CD11b/CD18 Receptor-Mediated Association and Signaling¹

In humans, group A streptococci (GAS)³ can cause a number of common infections, including tonsillitis and impetigo, but they can also lead to life-threatening conditions, such as necrotizing fasciitis and toxic shock (1). The surface of pathogenic GAS is covered with M proteins, which are major virulence factors and exist in >100 different variants, each of which defines a specific serotype (2, 3, 4). Experimental findings have suggested that the M proteins are involved in several stages of infection, such as adhesion to and invasion of epithelial cells (5, 6, 7, 8) and penetration of tissues (9, 10). There is also clear evidence that the M proteins are involved in evasion of innate (11) and adaptive (12) immunity. Resistance to the innate immune system is compatible with one of the most striking features of GAS, namely their ability to avoid being killed when incubated in human blood (3). Although the mechanism(s) underlying this property has not yet been satisfactorily explained, evidence suggests that an important role of M proteins is to limit deposition and generation of complement factor C3-derived opsonins, thereby limiting the GAS-phagocyte association (11, 13, 14). It should, however, be noted that recent findings have challenged this view, suggesting that both M protein-expressing and M-negative GAS are efficiently taken up by human neutrophils (15).

In blood, the dominating professional phagocyte is the neutrophil (16). In a nonimmune host, bacteria must be opsonized with the C3b or C3bi fragments to be identified and subsequently engulfed by neutrophils. This recognition is mediated by binding of C3b and C3bi to specific complement receptors (CRs) on the neutrophil surface, which include CD35 (CR1), and the integrins CD11b/CD18 (CR3, β₂α_M, Mac-1) and CD11c/CD18 (CR4, β₂α_X) (17). It is known that neutrophils express much more CD35 and CD11b/CD18 than CD11c/CD18, but the relative contribution of these receptors to complement-mediated phagocytosis of bacteria has not been determined.

The adhesion of complement-opsonized bacteria to neutrophils and the activation/intracellular signaling of CRs are crucial steps in initiating the phagocytic process. Nonreceptor tyrosine kinases have been found to be essential for the initiation of CD11/CD18 integrin-induced downstream signaling (18, 19, 20, 21, 22). Engagement of such integrins has been shown to activate tyrosine kinases of the Src family, as well as p72Syk, FakB (23, 24), and Pyk2 (25, 26). Upon activation, these tyrosine kinases phosphorylate and activate several proteins of potential importance for the phagocytic process, including those regulating Rho family GTPases (27). The CD11/CD18 integrin has previously been shown to trigger activation of the guanine nucleotide exchange factor Vav (28), the GTPase-activating protein p190RhoGAP, and the small GTPases RhoA (29) and Cdc42 (30, 31). Furthermore, RhoA activity has been shown to be necessary for CD11b/CD18-mediated phagocytosis of opsonized RBC in macrophages (27) and activation of Cdc42 for neutrophil phagocytosis of Candida albicans (32).

In this study, we used the GAS strain M5 Manfredo and an M5 protein-negative deletion mutant thereof (ΔM5) (33) to study the mechanisms of streptococcal resistance to neutrophil killing. By incubating these bacteria with isolated human neutrophils in the presence of plasma from nonimmune individuals, we could specifically study adhesion of bacteria to the neutrophils and the receptors and intracellular signals involved in the subsequent engulfment and killing. In contrast to a recent study (15), we find that the crucial mechanism whereby M protein blocks GAS killing is inhibition of bacteria-phagocyte association and the resulting CR signaling that initiates the phagocytic process.

M5 Manfredo is a GAS strain expressing the M5 protein. The strain ΔM5 is an isogenic derivative of M5 Manfredo, which does not express the M5 protein (33).

The polyclonal goat anti-GAS Ab was from Biogenesis (Poole, U.K.). FITC and Rhodamine Red-X-conjugated F(ab′)₂ of donkey anti-goat IgG were purchased from Jackson ImmunoResearch Laboratories (West Grove, PA). The mAb 4G10 (mouse anti-phosphotyrosine) was from Upstate Biotechnology (Lake Placid, NY). The mouse anti-Cdc42 mAb was from BD Transduction Laboratories (Lexington, KY). The rabbit anti-Rac2 polyclonal Ab was from Santa Cruz Biotechnology (Santa Cruz, CA). The HRP-conjugated goat anti-mouse (GAM) and anti-rabbit Abs were purchased from Dakopatts (Glostrup, Denmark). The receptor-blocking mouse mAbs are described in Table I.

Compstatin, a 13-residue cyclic peptide that binds to complement component C3 and inhibits complement activation (34), was kindly provided by J. Lambris, Department of Pathology and Laboratory Medicine, University of Pennsylvania (Philadelphia, PA). Soluble rCR1 was provided by C. Pettey (Avant Immunotherapeutics, Needham, MA) (35). rM5 was purified from Escherichia coli lysates by affinity chromatography on human serum albumin-agarose, as described previously (33). The chemicals and their sources were as follows: dextran and Ficoll-Hypaque, Pharmacia Fine Chemicals (Uppsala, Sweden); Polymorphprep, Axis-Shield PoC AS (Oslo, Norway); Todd Hewitt broth and Bacto yeast extract, Difco (Stockholm, Sweden); the tyrosine kinase inhibitor genistein and its inactive analog genistin, Sigma-Aldrich (Stockholm, Sweden); the tyrosine kinase inhibitor erbstatin, Calbiochem (Stockholm, Sweden); RPMI 1640 supplemented with l-glutamine (300 mg/liter), Life Technologies (Paisley, U.K.); fluorescent mounting medium, DAKO (Carpenteria, CA); the protease inhibitor Pefabloc, Roche (Mannheim, Germany); the protease inhibitor aprotinin, Bayer (Leverkusen, Germany); isopropyl-β-d-thiogalactopyranoside, MBI Fermentas (Vilnius, Lithuania); glutathione Sepharose, Amersham Pharmacia Biotech AB (Uppsala, Sweden); Re-Blot Plus Strong Ab Stripping Solution, Chemicon International (Temecula, CA); all electrophoresis reagents, Bio-Rad (Richmond, CA); luminol, ICN Pharmacueticals (Costa Mesa, CA); and the fluorescent dye bis-carboxymethyl-carboxy-fluorescein-pentaacetoxy-methylester (BCECF-AM), Molecular Probes (Eugene, OR). All other chemicals were of analytical grade and purchased from Sigma-Aldrich.

An E. coli strain expressing the Rac and Cdc42 binding domain from PAK1B (PAKcrib; aa 56–267) as a fusion protein with GST (36) was obtained from Pontus Aspenström, Ludwig Institute for Cancer Research, Biomedical Centre (Uppsala, Sweden). Expression of the fusion protein was induced by 0.2 mM isopropyl-β-d-thiogalactopyranoside at 30°C. The bacteria were collected, washed (50 mM Tris-HCl, pH 7.5, 5 mM MgCl₂, 50 mM NaCl, 1 mM Pefabloc, 1.75 μg/ml aprotinin, 2 μg/ml leupeptin, 600 nM DTT), and subsequently sonicated. Triton X-100 was added to a final concentration of 1% (v/v), and the lysates were centrifuged. The cleared lysates were aliquoted and stored at −80°C with 10% glycerol and used as such.

The cells were isolated from the blood of healthy donors under endotoxin-free conditions, as previously described (37). The neutrophils were isolated by dextran sedimentation, followed by a brief hypotonic lysis of erythrocytes. The lysis was stopped by adding 3 ml of a buffer containing 565 mM NaCl, 2.7 mM KCl, 6.7 mM Na₂HPO₄·2H₂O, and 1.5 mM KH₂PO₄, pH 7.3, along with 3 ml of Krebs-Ringer’s modified phosphate buffer (120 mM NaCl, 4.9 mM KCl, 1.7 mM KH₂PO₄, 1.2 mM MgSO₄·7H₂O, 8.3 mM Na₂HPO₄·2H₂O, 10 mM glucose, pH 7.3). The cell suspension was then centrifuged on Ficoll-Hypaque and washed twice with Krebs-Ringer’s modified phosphate buffer. Finally, the cells were resuspended in a calcium-containing medium (136 mM NaCl, 4.7 mM KCl, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 5.0 mM NaHCO₃, 1.1 mM CaCl₂, 0.1 mM EGTA, 5.5 mM glucose, 20 mM HEPES, pH 7.4), resulting in a suspension consisting of ∼97% neutrophils.

GAS were grown overnight in Todd Hewitt broth supplemented with 2 g/liter yeast extract. The supplementation with yeast extract was added, because it has a beneficial effect on bacterial growth. Ten milliliters of fresh, warm (37°C) culture medium was inoculated with 250 μl of the overnight culture, and the bacteria were grown to log phase (OD₆₂₀ = 0.150). Neutrophils (8 × 10⁵/ml) were incubated with bacteria (initial ratio 10⁴:6) in RPMI 1640 supplemented with l-glutamine, 20% (v/v) plasma from nonimmune individuals, and 10% (v/v) Todd Hewitt broth supplemented with 2 g/liter yeast extract at 37°C for 4 h, with end-over-end rotation. For experiments with soluble rM5 protein, neutrophils were preincubated with 200 μg/ml (final concentration) M5 protein for 30 min at 37°C in RPMI 1640 supplemented with l-glutamine. The neutrophils were then washed and resuspended in RPMI 1640 supplemented with l-glutamine before use in the bactericidal assay. Samples were withdrawn at different time points and plated on blood agar. The ratios given in the figures were calculated by dividing the number of CFUs at a specific time point by the number of CFUs at the beginning of the experiment.

Neutrophils (8 × 10⁵/ml) were incubated with bacteria (initial ratio 100:6) in RPMI 1640 supplemented with l-glutamine, 20% (v/v) plasma from nonimmune individuals, and 10% (v/v) Todd Hewitt broth supplemented with 2 g/liter yeast extract at 37°C for 3 h, with end-over-end rotation. The incubation was ended by placing the tubes on ice, and thereafter 5 × 100 μl of each sample (i.e., a 1/2 dilution in Todd Hewitt broth supplemented with 2 g/liter yeast extract) was prepared by cytospinning (500 rpm, 5 min) onto five separate glass slides. The sample on each slide was fixed with paraformaldehyde (2% in H₂O) on ice for 30 min and then washed twice in PBS. The slides were then incubated with a goat anti-GAS Ab (2 μl/ml in PBS) for 30 min at room temperature (RT), after which the neutrophils were permeabilized by incubating with 0.5% Triton X-100 in PBS for 3 min at RT. To stain extracellular bacteria associating with the neutrophils, the slides were overlaid with rhodamine-conjugated donkey anti-goat Ab (5 μg/ml in PBS) for 30 min at RT. The slides were then incubated at 37°C in a CO₂ incubator for 60 min with goat anti-GAS Ab (2 μl/ml in PBS) and then, to stain bacteria extra- and intracellularly associated with the neutrophils, for an additional 30 min with FITC-conjugated donkey anti-goat Ab (6 μg/ml in PBS). After each incubation, the slides were washed twice in PBS. Cover glasses were applied using a fluorescent mounting medium. A minimum of 120 neutrophils on each slide was examined in a fluorescence microscope, the proportion of neutrophils associated with bacteria was recorded, and the number of bacteria on each neutrophil was counted (bacteria not identified). Furthermore, for each sample, intracellular (green fluorescence) and extracellular (red and green fluorescence) bacteria were distinguished by their color and counted.

GAS were grown overnight in Todd Hewitt broth supplemented with 2 g/liter yeast extract. Ten milliliters of fresh, warm (37°C) culture medium was inoculated with 500 μl of the overnight culture, and the bacteria were grown to log phase (OD₆₂₀ = 0.4), washed twice in PBS, and subsequently stained with 1 mM BCECF-AM for 30 min at 37°C. Stained bacteria were washed twice in PBS and resuspended in 1 ml of PBS. Immediately before the addition to neutrophils, streptococcal chains were disrupted by sonication for 2 min in a Branson 1510 ultrasonic bath and subsequently diluted 10 times in PBS. Neutrophils were isolated from heparinized (17 U/ml) blood of healthy donors under endotoxin-free conditions. Whole blood was layered on Polymorphprep and centrifuged (450 × g, 40 min, 18°C). After centrifugation, the following fractions were apparent: mononuclear cells, neutrophils, and an erythrocyte pellet. The neutrophil layer was collected and suspended in 50 ml of RPMI 1640 supplemented with l-glutamine and centrifuged (350 × g, 10 min, 18°C). Residual contaminating erythrocytes were removed by hypotonic lysis. The cells (1 × 10⁶/ml) were finally resuspended in RPMI 1640 supplemented with l-glutamine and incubated with stained bacteria (initial ratio 1:15) in the presence of 20% (v/v) plasma from nonimmune individuals at 37°C for 30 min, with end-over-end rotation. For inhibition studies, neutrophils were preincubated with the indicated Abs at 37°C for 30 min before the addition of bacteria. Placing the tubes on ice and addition of ice-cold PBS ended the incubation, and the neutrophils were collected by centrifugation (200 × g, 10 min, 4°C). The neutrophils were resuspended in PBS and kept on ice until analyzed in a FACSCalibur flow cytometer from BD Biosciences (Heidelberg, Germany), at 488 nm. Granulocytes were selectively identified by appropriate settings of sideward scatter and forward scatter and were measured for green fluorescence from associated BCECF-AM-labeled bacteria.

Neutrophils (5 × 10⁶/ml) were incubated with bacteria (initial ratio 1:3) in RPMI 1640 supplemented with l-glutamine, 20% (v/v) plasma from nonimmune individuals, and 10% (v/v) Todd Hewitt broth supplemented with 2 g/L yeast extract at 37°C for 60 min, with end-over-end rotation. For inhibition studies, neutrophils were preincubated with the indicated inhibitor at 37°C for 30 min before adding bacteria. Genistein was solubilized in DMSO and used at a final concentration of 0–185 μM; the final concentration of DMSO never exceeded 0.1% (v/v). The incubation was ended at different time points by placing the tubes on ice. The neutrophils were collected by centrifugation (300 × g, 5 min, 4°C) and then lysed and boiled under reducing conditions in 100 μl of sample buffer (125 mM Tris-HCl, pH 6.8, 8% (v/v) glycerol, 2.5% (w/v) SDS, 0.05% (w/v) bromophenol blue) for 10 min. The samples were subjected to electrophoresis on 10% SDS-PAGE and transferred to polyscreen polyvinylidene difluoride (PVDF) membranes. The membranes were blocked in PBS supplemented with 0.05% (v/v) Tween 20 and 3% (w/v) skimmed milk and then incubated with a mouse anti-phosphotyrosine Ab (1/5000 dilution) for 60 min at RT. Thereafter, the blots were washed for 30 min in PBS supplemented with 0.05% (v/v) Tween 20 (PBST), incubated with HRP-conjugated GAM (1/10,000 dilution) for 60 min at RT, and then again washed for 30 min in PBST. The Abs were diluted in PBST supplemented with 3% (w/v) skimmed milk. Ab binding was visualized by ECL. The blots were subsequently scanned, and densitometric analyses were performed. In the present study, we were primarily interested in intracellular signals that could be implicated in the regulation of phagocytosis, and consequently, we would in particular look for tyrosine-phosphorylated exchange factors for small Rho GTPases or their upstream regulators. However, there exists a multitude of possible candidates, and consequently, we could not predict the specific proteins that were tyrosine phosphorylated, and the data presented for each sample therefore represent scanning of an entire lane.

Neutrophils and bacteria were incubated, as described above, and then neutrophils were collected by centrifugation (335 × g, 4 min, 4°C) and lysed in 1 ml of lysis buffer (50 mM Tris-HCl, pH 7.5, 10 mM MgCl₂, 100 mM NaCl, 1% (v/v) Triton X-100, 5% (v/v) glycerol, 800 μM Pefabloc, 5.6 μg/ml aprotinin, 4 μg/ml leupeptin, 800 nM Na₃VO₄), and 250 μl of GST-PAKcrib protein extract was added. The samples were kept on ice for 5 min and then centrifuged (14,000 × g, 10 min, 4°C). The supernatants were transferred to new tubes containing PBS-washed glutathione Sepharose beads (corresponding to 40 μl of stock solution) and incubated for 60 min at 4°C with end-over-end rotation. The beads were washed three times with ice-cold washing buffer (25 mM Tris-HCl, pH 7.5, 30 mM MgCl₂, 500 mM NaCl, 1% (v/v) Triton X-100, 1 mM DTT, 2 mM Na₃VO₄). A total of 40 μl of sample buffer with 130 mM DTT was added to the washed beads. After boiling for 10 min, solubilized proteins were subjected to electrophoresis on 12% SDS-PAGE and transferred to PVDF membranes. The membranes were blocked in PBST supplemented with 3% (w/v) skimmed milk and then incubated with a mouse anti-Cdc42 Ab (0.5 μg/ml) for 60 min at RT. Thereafter, the blots were washed for 10 min in PBST and then incubated with HRP-conjugated GAM (1/3000 dilution) for 60 min at RT. The Abs were diluted in PBST supplemented with 3% (w/v) skimmed milk. The blots were washed for 10 min in PBST, and Ab binding was visualized by ECL. The membranes were stripped and subsequently blotted for Rac2 (0.5 μg/ml). The blots were scanned for densitometric analysis.

To gain insight into the mechanism(s) whereby GAS avoid being killed in whole blood, we incubated isolated human neutrophils with M5 or ΔM5 bacteria in the presence of plasma from nonimmune individuals for up to 4 h. Samples were taken at different time points and plated on blood agar, and ratios (CFUs at a given time point divided by CFUs at onset of the experiment) were calculated for each time point. The results clearly show that M5 had increased by a factor >300 after 3 h of incubation (Fig. 1,A), whereas the number of ΔM5 had started to decrease already within the first hour (Fig. 1,B). To explore why more ΔM5 than M5 were killed, the interactions between neutrophils and the two GAS strains were studied by fluorescence microscopy. Neutrophils were incubated with M5 or ΔM5 for 3 h in the presence of plasma from nonimmune individuals and were subsequently centrifuged onto glass slides. The samples were double stained and examined in a fluorescence microscope to distinguish between GAS that had and had not been internalized by neutrophils. The examination revealed a striking difference: only 7% of neutrophils were associated with M5 bacteria, whereas 21% of neutrophils incubated with ΔM5 were associated with bacteria (Fig. 1,C), despite the presence of at least 300 times more bacteria in the M5 samples after 3 h (Fig. 1,A). Moreover, examination of at least 120 neutrophils per sample showed that 4 times more ΔM5 than M5 bacteria had become associated with the neutrophils (Fig. 1,D), despite the presence of many more bacteria in the M5 samples. This shows that M protein expression very strongly limits adhesion to neutrophils. There was a similar difference between intracellular ΔM5 and M5 bacteria (Fig. 1,E). However, the proportion of ingested M5 and ΔM5 bacteria relative to the number of associated bacteria did not exhibit a statistical significant difference (Fig. 1,F). We can therefore conclude that once attached, the streptococci become ingested regardless of M protein expression. It is known that M proteins can be released to the medium during streptococcal growth. To investigate whether soluble M protein has an effect on the killing capacity of neutrophils, we performed experiments in which neutrophils were preincubated with purified and soluble rM5 protein before incubation with ΔM5 streptococci. In the bactericidal assay, the ratios between CFU after 3 h and CFU at onset of the experiment for ΔM5 bacteria incubated with neutrophils preincubated in medium alone were 3.6, 4.3, 0.8, and 1.6, whereas for ΔM5 bacteria incubated with neutrophils that had been preincubated with medium supplemented with soluble rM5 protein the ratios were 6.2, 5.3, 2.4, and 2.2. Although this indicates a slight inhibitory effect of soluble M5 protein, the outcome is very small when compared with that of bacteria-bound M5 protein. The CFU ratios obtained for M5 bacteria were normally above 300 (Fig. 1 A).

Neutrophil adhesion is largely mediated by CRs. To investigate the relationship between CR activation and the elimination of GAS, we studied the effects of complement impairment on the killing of ΔM5 bacteria. Soluble CR1 (sCR1) blocks the generation of all opsonic fragments of C3 by inhibiting the C3 convertase (38, 39). However, if fragments of C3 are added to sCR1, it can also serve as a cofactor for factor I-mediated proteolytic degradation of C3b (a CD35 ligand), which yields C3bi (a CD11b/CD18 ligand) and ultimately C3d (a ligand for CD21 (CR2) and possibly also CD11c/CD18). In the present study, the sCR1-induced inhibition of the C3 convertase will inhibit any formation of C3 fragments, and thus exclude an impact of its cofactor function. Consequently, sCR1 is used to explore the involvement of CRs in general. The importance of complement activation for elimination of ΔM5 by neutrophils was clearly indicated by the ability of sCR1 to inhibit phagocytic killing of these bacteria (Fig. 2). The finding that the synthetic peptide compstatin, which blocks the cleavage of C3 (34), also inhibited killing of ΔM5 (Fig. 2) further emphasized the significant role for complement activation. These observations suggest that a CR on neutrophils is responsible for the adhesion-mediated elimination of these bacteria. Therefore, we used a panel of Abs that block putative CRs. We found that the addition of either of two different Abs against the CD11b/CD18 receptor inhibited the killing of ΔM5 (Fig. 3). In contrast, incubation with Abs against CD35, CD11a/CD18, or CD11c/CD18 had no effect on phagocytic killing of these mutant bacteria (Fig. 3). Moreover, although most individuals can be expected to have high levels of Abs against surface components present on all GAS strains, a blocking Ab directed against CD32 (FcγRIIa), which is the most important Fc receptor on neutrophils, had no effect on killing of ΔM5 (Fig. 3). These results suggest that the CD11b/CD18 receptor initiates killing via its function as an adhesion receptor on neutrophils.

To address the above hypothesis, we tested whether Abs toward CRs could inhibit the association between ΔM5 and neutrophils. The bacteria were labeled with a fluorescent dye and incubated with neutrophils that had been preincubated with Abs against CD11b and/or CD35 in the presence of plasma from nonimmune individuals. The association between bacteria and neutrophils was analyzed by flow cytometry. In agreement with the results presented in Fig. 1,D, ΔM5 associated more efficiently to neutrophils as compared with the wild-type strain (Fig. 4, A, top left, and B, top left). Increasing concentrations of the anti-CD11b Ab inhibited the association of ΔM5 to neutrophils in a dose-dependent manner (Fig. 4,A, top right and middle panels). Addition of an anti-CD35 Ab potentiated the blocking capability of anti-CD11b Ab (Fig. 4,A, bottom left), indicating that CD35 contributes to the complement-dependent association between ΔM5 and neutrophils. However, this receptor is not capable on its own to mediate the association between ΔM5 and neutrophils (Fig. 4,A, bottom right) or the killing of ΔM5 (Fig. 3) to a larger extent. In addition, an anti-CD11b Ab blocked the limited association of M5 bacteria to human neutrophils (Figs. 1, C and D, and 4,B, top panels), while an anti-CD35 Ab had no additional effect (Fig. 4 B, bottom left).

To gain additional support for our conclusions that adhesion and the CD11b/CD18 receptor are important for phagocytic killing of ΔM5, we investigated the signaling effects in neutrophils following their interaction with M5 and ΔM5 bacteria. Activation of CRs on neutrophils is known to trigger an increase in tyrosine phosphorylation of several different proteins, including some of which might well be involved in the phagocytic process (18, 19, 26). Because small Rho GTPases are implicated in regulation of phagocytosis, it would be particularly interesting to look for tyrosine phosphorylation of exchange factors for small Rho GTPases or their upstream regulators. However, because a multitude of possible exchange factors exists, we believe that, at this stage, antiphosphotyrosine analysis of whole cell lysates is a more adequate approach than antiphosphotyrosine analysis of immunoprecipitates of all possible upstream regulators of small Rho GTPases. Human neutrophils were incubated with bacteria (M5 and ΔM5 separately) in the presence of plasma from nonimmune individuals for different periods of time, after which the cells were lysed and protein tyrosine phosphorylations were analyzed by Western blotting. The accumulated results of densitometric scanning of whole lanes clearly show that a more pronounced phosphorylation response was triggered by the ΔM5 than by the M5 bacteria (Fig. 5,A), and that this difference was even more marked for some of the protein bands (Fig. 5,B). The specificity of the signaling response is reflected by the ability of the tyrosine kinase inhibitor genistein to inhibit ΔM5-induced phosphorylations (Fig. 5, C and D). The observation that neutrophils that had been incubated with ΔM5 in the presence of sCR1 showed a marked reduced tyrosine phosphorylation response (Fig. 5, E and F) provides further evidence that complement-CD11b/CD18 interactions are important for the association (Fig. 4) and phagocytic killing (Fig. 3) of ΔM5 by these phagocytes.

To directly test the hypothesis that adhesion-induced CR-mediated tyrosine kinase signaling is involved in the killing of ΔM5 bacteria, we tested the effects of genistein on the killing of ΔM5. The importance of ΔM5-induced tyrosine phosphorylations for elimination of this GAS strain by neutrophils was clearly indicated by the ability of genistein (but not its inactive analog genistin) to inhibit phagocytic killing of ΔM5 (Fig. 6). The finding that an erbstatin analog, another tyrosine kinase inhibitor, also blocked the killing of ΔM5 (data not shown) further supports the role of CR-mediated tyrosine kinase activation in neutrophil killing of GAS.

Tyrosine kinases are initial elements in the signaling cascade initiating phagocytosis. Recent data suggest that activation of Rho GTPases is intimately linked to phagocytosis (27). Rho GTPases are more distal signaling molecules in phagocytosis, and the activation of Rho GTPases has been shown to be dependent on tyrosine kinases (40); hence, the involvement of both tyrosine kinases and Rho GTPases is quite plausible in the present study. We used the GST-PAKcrib-binding assay (41), which uses the fact that the GST-PAKcrib fusion protein only binds active, GTP-bound forms of the Rho GTPases Cdc42 and Rac, to study the activation of Cdc42 and Rac2 (the totally dominating Rac protein in neutrophils) during neutrophil interactions with M5 and ΔM5 bacteria. Neutrophils and bacteria were incubated in the presence of plasma from nonimmune individuals for different periods of time, after which the cells were lysed and their contents of activated Cdc42 and Rac2 were analyzed. The accumulated results show an increase in the activation of Cdc42 for neutrophils that have been incubated with ΔM5 (Fig. 7,A), whereas no activation of Rac2 was observed (Fig. 7,B). In contrast, in neutrophils incubated with the wild-type strain (M5), no effect on the level of GTP-bound Cdc42 was observed (Fig. 7,A), while the level of active Rac2 decreased, but with a low level of statistical significance (Fig. 7,B). Finally, we performed identical experiments in the presence of the tyrosine kinase inhibitor genistein. This approach resulted in inhibition of the ΔM5-induced activation of Cdc42 (Fig. 7 C), thus establishing a link between ΔM5-induced tyrosine kinase activity, phosphorylation, and activation of a GDP/GTP exchange factor for Cdc42, and an increased amount of active GTP-bound Cdc42.

To confirm the relationship between CD11b/CD18 and the activation of Cdc42, we preincubated neutrophils with either sCR1 or an anti-CD11b Ab before incubation with ΔM5 bacteria and then analyzed the activation status of Cdc42. Impairing complement activation by the addition of sCR1 clearly inhibited the ΔM5-induced activation of Cdc42 (Fig. 8,A). Similarly, Ab blocking of CD11b/CD18 resulted in a similar, but less pronounced effect (Fig. 8,B), and accumulated data revealed a clear inhibition of ΔM5-induced activation of Cdc42 also in this situation (Fig. 8 C).

The sequential results obtained in this study are summarized in a schematic model that outlines how phagocytosis of ΔM5 bacteria is regulated and where the M proteins interfere with this process (Fig. 9).

The first deductions that can be made from the present study is that neutrophils are indeed responsible for the killing of GAS in whole blood, and that phagocytosis is an important mechanism by which these leukocytes eliminate streptococci in whole blood. Furthermore, the capacity of the M5 protein to impair the recognition/adhesion step in the phagocytic process represents the most important difference between the M5-expressing and nonexpressing GAS, enabling M5-expressing streptococci to survive in the presence of neutrophils.

These findings are in contrast with recently published data by Staali et al. (15), who concluded that GAS are efficiently phagocytosed by human neutrophils regardless of their M protein expression. However, their conclusion is based on an experimental approach in which the degree of phagocytosis is studied in the absence of complement and after the bacteria and the neutrophils have been copelleted by centrifugation. Such an experimental approach will synchronize the phagocytic process, but it will unfortunately also exclude detection of any differences in terms of an initial complement-dependent adhesion between the bacteria and the neutrophils. However, the results of our study are in good agreement with those of DeMaster et al. (42). In their study, a M49⁺ strain of GAS became associated with ∼30% of the neutrophils in whole human blood, whereas, similar to our results, association was three times larger between the neutrophils and an isogenic strain of GAS that did not express any M proteins (Mrp, Emm, or Enn). It should be mentioned that these authors also demonstrated a role of complement and the CD11b/CD18 receptor in the killing of GAS.

In a nonimmune host, opsonization of bacteria with fragments of the complement component C3 is essential for neutrophils to be able to recognize, associate with, and subsequently ingest the bacteria (43). This finding and data showing that less C3 is bound by M protein-expressing GAS than by isogenic nonexpressing strains, and that lack of complement fixation leads to increased survival of the bacteria (11), suggest that association with and engulfment of non-M5-expressing GAS by neutrophils are related to activation of cell surface CRs on the phagocytes. Together with previous results, these data emphasize the role of M proteins in limiting deposition/generation of C3b/C3bi fragments on the streptococcal surface. The mechanism whereby this is accomplished remains unclear. An attractive hypothesis suggests that bound opsonins are degraded with the help of the complement regulators factor H, factor H-like protein 1, and C4b-binding protein, all of which specifically interact with various M proteins (13, 44, 45). However, GAS expressing M5 protein variants deleted for the regions responsible for the interaction with factor H and factor H-like protein 1 are still able to grow in the bactericidal assay (46). Furthermore, many M proteins lack the ability to bind any of the soluble complement regulators (47). It is therefore reasonable to suggest that other, and to some extent less specific, mechanisms are involved in limiting the deposition/generation of C3b/C3bi fragments. One option is that the unique ability of M proteins to interact with quantitatively dominating proteins in plasma, e.g., albumin, fibrinogen, and IgG, sterically blocks access of C3 to the bacterial surface. Indeed, an M5 variant specifically lacking the fibrinogen-binding region failed to provide GAS with resistance against killing by human neutrophils (48), and complement deposition on this strain was increased (46). Regardless of the mechanism responsible for impairing complement deposition on M protein-expressing GAS, it will reduce the capacity of these bacteria to adhere to human neutrophils, as demonstrated in the present investigation. In addition to an ability of M proteins to impair complement deposition on the bacterial surface, it has also been proposed that M proteins could directly, by virtue of their overall negative charge, repel the interaction with neutrophils, because this phagocytic cell also has an overall negative surface charge (49). Additional experiments are clearly required to decide which, if any, of these hypotheses is correct.

Phagocytosis of complement-opsonized bacteria may be mediated by different CRs expressed on the surface of the neutrophils. One of these, the β₂ integrin CD11b/CD18, has been found to be important both for the adhesion of neutrophils to different biological surfaces (19, 20) and for phagocytosis (43). The results of this study strongly suggest that the CD11b/CD18 integrin is the major receptor responsible for phagocytosis of ΔM5 bacteria, and that the other two CRs on neutrophils, CD35 and CD11c/CD18, are of limited importance in this process. The critical role of complement-CR interaction is further emphasized by our data showing that an Ab known to specifically block binding of IgG to the major FcR on neutrophils (FcγRIIa, CD32) had no effect on the phagocytic killing of ΔM5 bacteria. To our knowledge, we are the first to show that an increase in protein tyrosine phosphorylations occurs during CD11b/CD18-mediated phagocytosis of Gram-positive bacteria in neutrophils. Although a limited number of reports has investigated a role of protein tyrosine phosphorylations in complement-mediated phagocytosis of bacteria in macrophages, the conclusions from these reports have not been consistent. Kusner et al. (50) showed that CR-mediated phagocytosis resulted in an increase in tyrosine phosphorylation of multiple proteins and that phagocytosis was blocked by protein tyrosine kinase inhibitors. In contrast, Allen and Aderem (51) found that activation of protein tyrosine kinases was not a prerequisite for complement-mediated ingestion, but rather increased the efficiency of ingestion. In comparison with these findings in macrophages, we found that direct or indirect inhibition of tyrosine kinases in neutrophils inhibited the killing of ΔM5 bacteria, indicating that tyrosine kinases play a pivotal role both in the ingestion and in the phagocytic killing of ΔM5 by these leukocytes. To the best of our knowledge, this report provides the first direct link between increased tyrosine phosphorylations and ingestion and phagocytic killing of extracellular bacteria.

In human neutrophils, several different downstream targets have been reported for CD11b/CD18-induced activation of nonreceptor tyrosine kinases (52). In the present context, the family of Rho GTPases is of particular interest, due to their implication in the regulation of the phagocytic process (27). Investigations of the mechanisms involved in neutrophil phagocytosis of C. albicans have indicated a role for Cdc42 (32) and possibly also for Rac (53). Our data suggest an involvement of a tyrosine kinase-dependent activation of a GDP/GTP exchange factor that then will activate Cdc42, but not Rac2, in CD11b/CD18-mediated phagocytosis of living GAS. This is in good agreement with the observation by Dharmawardhane et al. (54), who demonstrated that p21-activated kinase, a downstream target of Cdc42 and Rac, plays a regulatory role in the initiation of CR-mediated phagocytosis in neutrophils.

In conclusion, by using a non-M5-expressing GAS strain, we found that phagocytic killing of GAS is mediated by activation of complement and opsonization of the bacteria with C3b/C3bi, enabling the interaction with CD11b/CD18 integrins. Adhesion of the bacteria to this integrin then results in the activation of a tyrosine kinase/Cdc42-signaling pathway that finally will enable the neutrophils to phagocytose these bacteria, schematically outlined in Fig. 9. Moreover, our comparison of M5-expressing and nonexpressing GAS revealed that the former strain resists phagocytic killing by reducing its association with neutrophils. Our results are in contrast to the recent suggestion that M proteins inhibit killing of streptococci by interfering with intracellular killing, rather than the recognition of the bacteria (15). Instead, our present analyses of the receptors and signaling pathway involved in the adhesion and phagocytosis of ΔM5 and M5 bacteria are compatible with the finding that M proteins can decrease the deposition and/or generation of opsonic complement fragments on the surface of GAS (11).

We thank Malgorzata Berlikowski and Maria Baumgarten for excellent technical assistance, and Patty Ödman for linguistic revision of the manuscript. We thank Dr. Nancy Hogg (Leukocyte Adhesion Laboratory, Imperial Cancer Research Fund, London, U.K.) and Dr. Eric J. Brown (University of California, School of Medicine, San Francisco, CA) for providing the mAbs, clones 3.9 and 3D9, respectively. We are also grateful to Dr. John D. Lambris (Department of Pathology and Laboratory Medicine, University of Pennsylvania) for compstatin, Dr. Carolyn Pettey (Avant Immunotherapeutics) for sCR1, and Dr. Pontus Aspenström (Ludwig Institute for Cancer Research, Biomedical Centre) for the GST-PAKcrib-expressing E. coli strain.