Staphylococcus aureus Formyl Peptide Receptor–like 1 Inhibitor (FLIPr) and Its Homologue FLIPr-like Are Potent FcγR Antagonists That Inhibit IgG-Mediated Effector Functions

To evade opsonophagocytosis, Staphylococcus aureus secretes various immunomodulatory molecules that interfere with effective opsonization by complement and/or IgG. Immune-evasion molecules targeting the phagocyte receptors for these opsonins have not been described. In this study, we demonstrate that S. aureus escapes from FcγR-mediated immunity by secreting a potent FcγR antagonist, FLIPr, or its homolog FLIPr-like. Both proteins were previously reported to function as formyl peptide receptor inhibitors. Binding of FLIPr was mainly restricted to FcγRII receptors, whereas FLIPr-like bound to different FcγR subclasses, and both competitively blocked IgG-ligand binding. They fully inhibited FcγR-mediated effector functions, including opsonophagocytosis and subsequent intracellular killing of S. aureus by neutrophils and Ab-dependent cellular cytotoxicity of tumor cells by both neutrophils and NK cells. In vivo, treatment of mice with FLIPr-like prevented the development of an immune complex–mediated FcγR-dependent Arthus reaction. This study reveals a novel immune-escape function for S. aureus–secreted proteins that may lead to the development of new therapeutic agents in FcγR-mediated diseases.

T he commensal bacterium Staphylococcus aureus is a leading cause of a wide variety of infections in humans and animals worldwide, ranging from mild superficial skin infections to fatal deep-seated infections that entail spread through the bloodstream and body (1). Its great success with regard to infection is explained by the numerous strategies that this commensal has developed to breach host immune responses (2,3).
The principal line of defense against extracellular bacterial infection relies on phagocytes, including residential macrophages and immigrating neutrophils. These leukocytes recognize, remove, and destroy invading bacteria through phagocytosis. Bacterial recognition can occur directly via membrane-bound pattern recognition receptors, although many pathogens, including S. aureus, resist direct engulfment by phagocytes and require opsonization by complement (C3b, C3bi) or Igs (IgA, IgG) for efficient phagocytosis. The phagocyte receptors for these opsonins, complement receptors CR1 and CR3, FcaR, and leukocyte FcgRs, respectively, become activated upon ligand binding. They induce the initial uptake of the pathogen into a phagosome that subsequently fuses with lysosomes, resulting in the intracellular killing of the pathogen. An intracellular IgG receptor, the neonatal FcR (FcRn) (4), contributes to this process as well; however, mechanistic details remain to be elucidated.
To evade phagocytosis, S. aureus expresses a variety of both secreted and cell surface-associated immunomodulatory proteins. These proteins bind to key elements in this process and inhibit their proper function. Several of these abrogate effective bacterial opsonization by complement and/or IgG (5). SCIN (6), Efb, and its C-terminal homolog Ecb (7) are secreted proteins that target complement deposition on the bacterial surface by direct and indirect inhibition of C3 convertases. Sbi is an excreted and cell wall-anchored protein with a dual function, disturbing both complement and IgG opsonization (8,9). Sbi binds to fragment C3d of complement factor C3 and to the constant domain (Fc) of IgG, thereby blocking interactions with FcgR and complement factor C1q. A similar inhibition of IgG Fc tail function, although targeting IgG1 exclusively, is provided by the secreted SSL10 (10). Staphylococcal protein A (SpA) is a cell wall-anchored IgG-binding (11) and antiphagocytic protein expressed by most strains of S. aureus. Its inhibitory effect on phagocytosis is explained by the concept that SpA captures IgG onto the bacterial surface in a maloriented position, thereby preventing recognition by neutrophil FcgR. In addition to these blocking proteins, S. aureus also expresses an enzymatic inhibitor of bacterial opsonization, staphylokinase, which degrades the bacterium-bound opsonins, C3b and IgG, by activation of plasminogen into plasmin, an active serine protease (12). Staphylococcal immunomodulatory molecules that directly target the phagocyte opsonin receptors have not been described. In this study, we focus on FcgRs, the receptors for IgG.
Through binding of the Fc domain of IgG, FcgRs activate immune effector cells on which they are broadly expressed (13)(14)(15). FcgRs are members of the Ig superfamily. They possess an extracellular ligand-recognizing a-chain consisting of two or three Iglike domains, a transmembrane region, and an intracellular tail. Cross-linking of FcgRs by IgG-opsonized particles or immune complexes induces several cell type-dependent immune responses, ranging from phagocytosis, respiratory burst, and Ab-dependent cellular cytotoxicity (ADCC) to secretion of inflammatory mediators, enhancement of Ag presentation, and regulation of Ab production. Based on structural and biochemical differences they are divided in three distinct, but closely related, classes of FcgRs-FcgRI (CD64), FcgRII (CD32), and FcgRIII (CD16)-containing 12 isoforms that differ in cellular distribution and affinity for IgG. Monocytes and macrophages express members of all three FcgR classes: FcgRIa, FcgRIIa/b/c, and FcgRIIIa (on monocyte subset). The most abundant phagocyte, the neutrophil, constitutively expresses only two members of the FcgR classes: FcgRIIa and FcgRIIIb. FcgRIIa is the predominant receptor involved in phagocytosis and binds IgG only in complexed or polymeric form. The FcgRIIa gene demonstrates allelic variation (16), resulting in two allotypes with differential affinity for IgG: FcgRIIa-H131(E27) and FcgRIIa-R131(Y27). Human neutrophils also express limited amounts of the inhibitory FcgRIIb (17,18). FcgR class III contains two isoforms: the medium-affinity receptor FcgRIIIa, which is expressed on NK cells, macrophages, and some monocytes, and the GPI-anchored FcgRIIIb, with a low affinity for IgG, which is constitutively expressed on neutrophils. For FcgRIIIa, two functionally different allelic variants are known that differ at position 158 (V158 and F158) (16). FcgRIIIb bears a polymorphism (16) at four positions in the extracellular region, resulting in two allotypes, neutrophil Ag (NA) 1 (R18 N47 D64 V88) and NA2 (S18 S47 N64 I88), which have different glycosylation patterns and interaction with IgG. The high-affinity IgGR FcgRIa, which is expressed on monocytes, macrophages, and activated neutrophils, is structurally unique in that it contains three extracellular Ig-like domains, in contrast to the two found in FcgRII and FcgRIII.
In this study, we investigated the hypothesis that S. aureus evades opsonophagocytosis via the secretion of an FcgRIIa inhibitory molecule. We describe the identification of two immunomodulatory homolog proteins secreted by S. aureus that target FcgR directly by recognizing epitopes in or near their IgG-binding domain. Both proteins were reported to be antagonists for the formyl peptide receptor (FPR) family. In this study, we demonstrate that these proteins are potent antagonists of FcgR-mediated effector functions in vitro, including opsonophagocytosis and ADCC, and in vivo.

Bacterial supernatants and proteins
Clinical and laboratory strains of S. aureus were grown overnight at 37˚C in IMDM without Phenol Red (Invitrogen), and cell-free supernatant was harvested by centrifugation and filter sterilized. One liter of supernatant from S. aureus subsp. aureus N315 (a sequenced strain; GenBank BA000018) was passed over a 25-ml "Reactive Red 120" ligand dye cross-linked 4% beaded agarose column (Sigma-Aldrich) attached to an Akta-FPLC system (GE Healthcare Life Sciences). After washing with PBS, the column was eluted into 2.5-ml fractions using 1 M NaCl. PMSF (1 mM) was added, and fractions were dialyzed in PBS for 18 h. Fractions were screened for activity by anti-FcgRII mAb 7.3 staining on fresh human neutrophils. Active fractions were pooled, concentrated with a 10-kDa Centriprep (Amicon; Millipore), and separated on a GE Healthcare Superdex-75 gel filtration column into 2.5ml fractions that were again screened for activity. Active fractions were pooled and concentrated using a 10-kDa Centriprep and stored at 220˚C in small aliquots. Different fractions were precipitated with 20% TCA for 30 min on ice and analyzed on 15% SDS-PAGE (Mini-Protean II; Bio-Rad) by silver staining.
For affinity isolation, magnetic Cobalt-chelating beads (TALON Dynabeads; Invitrogen) were coated with recombinant His-tagged human CD32a (the extracellular domain Ala 36-Ile 218 of human FcgRIIa; #1330-CD; R&D Systems). Fifty microliters of beads was washed twice with PBS containing 0.1% Triton X-100 (PBS-Triton) and incubated for 30 min with 100 ml 200 mg/ml His-tagged CD32a in PBS. Beads were washed three times with PBS-Triton and incubated with purified supernatant for 18 h at 4˚C under gentle rotation in a total volume of 400 ml. Supernatant was discarded, and beads were washed three times with PBS-Triton, suspended in 30 ml SDS-PAGE sample buffer for 15 min, and heated for 2 min at 100˚C. The sample was centrifuged briefly (10 s at 10,000 3 g), and the supernatant was analyzed on 15% SDS-PAGE by silver staining. Bands were excised and sent for protein identification at the Department of Biomolecular Mass Spectrometry, Utrecht Institute for Pharmaceutical Sciences, Utrecht, The Netherlands.

Surface-enhanced laser desorption ionization time-of-flight mass spectrometry
For identification by mass, a Ciphergen (Bio-Rad) IMAC30 ProteinChip Array was used that incorporates nitrilotriacetic acid groups forming stable complexes with metal ions. After loading the array with nickel sulfate, 10 mg/ml His-tagged soluble FcgRIIa was added and subsequently incubated with semipurified concentrated staphylococcal supernatant. The array was washed with PBS, rinsed with water, air dried, and treated with a saturated solution of sinapinic acid, an energy-absorbing molecule that assists in desorption and ionization. After air drying, the array was analyzed using the Ciphergen ProteinChip System Series 4000 read at a setting optimized for low molecular mass range. Spectra were externally calibrated, baseline corrected, and normalized to total ion current within a mass/charge (m/z) range of 1,500-50,000 Da.

Cells and flow cytometry staining
The mouse P388D1 macrophage cell line was maintained in RPMI 1640 medium supplemented with 10% (v/v) FCS. Human neutrophils and mononuclear cells were isolated from heparinized blood, as described (21). Cells were incubated with staphylococcal supernatants (30% v/v) or purified recombinant proteins for 0-10 min on ice, washed, and subsequently stained with fluorescent mAb. Fluorescent intensities were measured by flow cytometry and expressed relative to control cells not exposed to supernatant or recombinant protein. FLIPr and FLIPr-like were labeled with FITC, as described (21,22), and incubated with leukocytes in the presence of specific markers for monocytes (PE-labeled anti-CD14), T cells (allophycocyanin-labeled anti-CD3), B cells (PE-labeled anti-CD19), or NK cells (PE-labeled CD56 and allophycocyanin-labeled anti-CD3). Cells were gated by their scatter and marker profile.

Evaluation of FcgR binding and Ig inhibition by ELISA
To detect binding of staphylococcal proteins to FcgR subclasses, MaxiSorp plates (Nunc) were coated overnight at 4˚C with 100 ml FLIPr, FLIPr-like, or chemotaxis inhibitory protein of Staphylococcus aureus (CHIPS) at 1 mg/ml in PBS. Plates were washed three times with PBS supplemented with 0.05% v/v Tween 20 and incubated with a serial dilution of different soluble FcgR proteins containing a polyhistidine tag in PBS supplemented with 0.05% v/v Tween 20 and 0.2% w/v BSA. After 2 h of shaking (300 rpm) at room temperature, plates were washed and incubated for 90 min with HRP-conjugated anti-His mAb to detect FcgR binding. Inhibition of Ig binding to the different FcgR subclasses by staphylococcal proteins was measured with another ELISA. Plates were coated overnight with 4 mg/ml anti-polyHistidine mAb, washed, and incubated with 2 mg/ml soluble Histagged FcgR for 90 min. The plates were washed and incubated with a serial dilution of the staphylococcal proteins. After washing, the different subclasses of FcgR were probed with an optimal concentration of human monoclonal IgG1 (HuMab-KLH) that was detected with HRP-conjugated anti-human Fc. Statistical analysis was performed using a two-tailed Student t test.

Phagocytosis and killing
Phagocytosis was measured with FITC-labeled staphylococci (6). In a 96well plate, bacteria were mixed with a concentration range of human serum (pooled from 15 donors), complement-inactivated serum (30 min at 56˚C), or purified IgG (by Protein G affinity chromatography from the same serum pool) for 15 min at 37˚C for opsonization. Subsequently, neutrophils or macrophages, with or without inhibitor, were added at a 10:1 bacterium/ cell ratio and incubated for 15 min at 37˚C on a shaker (750 rpm). The reaction was stopped with ice-cold paraformaldehyde (1%), and cellassociated fluorescent bacteria were analyzed by flow cytometry. Phagocytosis is defined as the percentage of cells with a positive fluorescent signal. To compare the experiments, the values are expressed relative to the second highest opsonin concentration of control cells without inhibitory protein. Killing of staphylococci was determined by colony plate counting. Bacteria were opsonized for 15 min with 1% complement-inactivated serum, mixed with neutrophils at a 1:10 ratio, and incubated for 1 h at 37˚C while shaking. A 50-ml sample was diluted into 2.5 ml ice-cold water [pH 11], vortexed, and incubated for 5 min on ice to lyse the neutrophils. The remaining viable bacteria in each reaction mixture were counted by plating 50 ml onto nutrient agar plates, followed by overnight incubation at 37˚C. Controls consisted of nonopsonized bacteria and bacteria without neutrophils.

ADCC
ADCC assays against 51 Cr-labeled A1207 target cells were performed as described (23,24). In short, isolated neutrophils or NK cells (from PBMCs using The EasySep Human NK cell Enrichment Kit; STEMCELL Technologies) were preincubated with 3 mg/ml inhibitory protein. Sensitizing Abs (1 mg/ml) were added to microtiter plates, and ADCC was induced by adding effector and target cells at an E:T ratio of 80:1 for neutrophils and 10:1 for NK cells. The plates with NK cells as effector cells were centrifuged for 5 min at 300 rpm. After 4 h at 37˚C, [ 51 Cr] release from triplicates was measured (cpm). The percentage of cellular cytotoxicity was calculated using the following formula: percentage of specific lysis = (experimental cpm 2 basal cpm)/(maximal cpm 2 basal cpm) 3 100. Maximal [ 51 Cr] release was determined in the presence of 5% Triton X-100, and basal lysis was determined in the absence of sensitizing Abs and effector cells. Data are expressed as relative lysis compared with buffertreated cells toward mAb-coated targets. Statistical analysis was performed using a two-tailed Student t test.
Peritoneal Arthus reaction BALB/c mice were injected i.v. with 100 ml OVA (20 mg/kg of body weight; Sigma-Aldrich), immediately followed by i.p. injection of 200 ml rabbit anti-OVA IgG (800 mg/mouse; MP Biomedicals) in sterile PBS, as described (25). For inhibition experiments, 100 ml FLIPr-like was administered i.v. and i.p. (600 mg/ml) 30 min before initiation of the Arthus reaction. Mice in different treatment groups were killed 6 h after onset of the reaction. The peritoneal cavity was lavaged with 6 ml ice-cold PBS/ 0.1% BSA. Peritoneal cells were washed once with PBS, and neutrophil numbers were calculated from cytospin slides stained with Diff-Quick. Neutrophil numbers per square millimeter were calculated from 20 microscopic fields. Animal care was provided in accordance with National Institutes of Health guidelines. Animal studies were approved by the Cincinnati Children's Hospital Medical Center Institutional Animal Care and Use Committee.

Statistical analysis
Significant differences (p , 0.05) were identified with Student t tests and one-way ANOVA using GraphPad Prism 5.0 software (GraphPad). Data are expressed as mean 6 SD or SEM.

Identification of a putative staphylococcal FcgRIIa inhibitor
To investigate potentially secreted staphylococcal FcgRIIa antagonists, culture supernatants of various clinical and laboratory S. aureus strains were screened for their ability to competitively inhibit specific neutrophil FcgRIIa staining by flow cytometry. The anti-FcgRII mAb clone 7.3 was used because of its specificity for an epitope in the IgG binding site of the receptor (26). Several staphylococcal supernatants inhibited mAb 7.3 binding to FcgRIIa on neutrophils (data not shown). Supernatant of sequenced S. aureus strain N315 was among the most potent (Fig. 1A) and was selected for purification of the inhibitory compound. Ligand dye chromatography, gel permeation, and anion exchange chromatography were performed, and activity of the fractions was monitored by mAb 7.3 neutrophil staining. Using a chip coated with soluble rFcgRIIa, combined with mass detection in a surface-enhanced laser desorption/ionization time-of-flight system, identified a 12.3-kDa mass peak in the semipurified staphylococcal supernatant (Fig. 1B). Affinity isolation on magnetic beads, coupled with soluble FcgRIIa, was used to obtain a final active preparation containing one single nearly homogeneous protein band ∼13 kDa on SDS-PAGE (Fig.  1C). Gel extraction and subsequent mass spectrometry identified the protein as FLIPr. FLIPr, a secreted staphylococcal protein of 105 aa with a mass of 12.3 kDa, is a potent inhibitor of the chemotaxis receptor FPR2 (formerly annotated as formyl peptide receptor-like 1 or FPRL1) (21). Initially, FLIPr was discovered based on its 49% sequence homology with CHIPS (chp), the chemotaxis inhibitory protein of S. aureus (20).
To confirm that FLIPr was the active FcgRIIa inhibitory compound present in the staphylococcal supernatant, recombinant FLIPr was tested for its effect on neutrophil FcgRIIa staining. Like the supernatant, recombinant FLIPr inhibited binding of mAb 7.3 to neutrophils, with an IC 50 of 48.4 ng/ml (range, 33.5-69.8 ng/ml) (Fig. 1D). The recombinant homolog of FLIPr, FLIPr-like (104 aa with ∼72% identity) (22), was ∼10-fold more active in FcgRIIa blocking compared with FLIPr, with an IC 50 of 4.9 ng/ml (range, 3.7-6.5 ng/ml). The related staphylococcal protein CHIPS did not affect FcgRII staining. Because of its structural and functional similarities, CHIPS was used as control protein in further experiments.

FLIPr and FLIPr-like bind different FcgR isoforms and block IgG binding
A capture ELISA was performed to confirm that FLIPr and FLIPrlike directly interact with FcgRIIa, as well as to analyze the specificity for other Fc(g)Rs. Different recombinant soluble FcgRs and their polymorphic isoforms containing a polyhistidine tag (Ia, IIa-H131 and IIa-R131, IIb, IIIa-V158 and IIIa-F158, and IIIb-NA1 and IIIb-NA2) were evaluated for their relative binding to the coated staphylococcal proteins FLIPr, FLIPr-like, and CHIPS. Both FLIPr and FLIPr-like bound to soluble FcgRIIa in ELISA (Fig. 2B, 2C). In contrast to FLIPr-like, FLIPr demonstrated differential binding to the two allelic variants of FcgRIIa that differ at position 131, which is located in the ligand-binding region and is crucial for IgG binding (15). FLIPr bound preferentially to the high-affinity isoform FcgRIIa-H131, whereas binding to the low-affinity R131 isoform was almost negligible. The highly homologous inhibitory receptor FcgRIIb, whose ectodomain differs at 10 positions from FcgRIIa (11 for the FcgRIIa-H131 variant) (27), including two (or three for FcgRIIa-H131) amino acids involved in IgG binding, was bound equally well by FLIPr and FLIPr-like (Fig. 2D). With regard to the other FcgR subclasses, FLIPr did not bind significantly to FcgR class I and III receptors ( Fig. 2A, 2E-H) and could be considered class II specific. In contrast, FLIPr-like bound to FcgRIa ( Fig. 2A) and differentially to FcgRIIIa-V158 and FcgRIIIa-F158, an amino acid also involved in IgG binding (Fig. 2E, 2F). The two isoforms of FcgRIIIb, differing by six ectodomain amino acids from FcgRIIIa-F158 (and 5 aa from FcgRIIIa-V158), of which two are located in IgG-binding regions (23,24,28), were both bound by FLIPr-like, but only minimally in the case of FcgRIIIb-NA1. CHIPS did not bind to any of the soluble FcgRs. Furthermore, none of the staphylococcal proteins bound to soluble FcaR or FcεR, the receptors for IgA and IgE, respectively, in ELISA (data not shown). Because all FcgRs are composed of Ig-like domains strongly resembling the domains of IgG, we used ELISA to test whether FLIPr and FLIPrlike bound to IgG molecules, as well. In contrast to Mac-1, a streptococcal FcgR-binding protein that does bind IgG (29), neither FLIPr nor FLIPr-like bound IgG in ELISA (data not shown).
To investigate whether FLIPr and FLIPr-like interfere with FcgR-IgG interactions, an inhibition ELISA was performed. FcgRs were incubated with staphylococcal proteins, and relative binding of recombinant human IgG1 to FcgR was determined. FLIPr and FLIPr-like efficiently prevented binding of human IgG1 to the FcgRs (Fig. 3), in accordance with their specific binding. The control protein CHIPS did not inhibit FcgR-IgG binding. In summary, FLIPr seems to be class II restricted and preferentially inhibiting IgG binding to FcgRIIa-H131 (Fig. 3B), the predominant receptor in phagocytosis. In contrast, FLIPr-like displays a broader profile and inhibits FcgR-IgG interactions for all FcgRs, with the exception of FcgRIIIb (Fig. 3G).
To further investigate the binding site of FLIPr and FLIPr-like on the FcgR, additional mAb-competition studies were performed. In addition to anti-FcgRII mAb 7.3, FLIPr and FLIPr-like inhibited neutrophil binding of the more FcgRIIa-specific mAb IV.3, which also recognizes D2 epitopes involved in IgG interactions (26) (Fig.  4A, 4B). In contrast, neutrophil staining with the IgG-blocking anti-FcgRII mAb FLI8.26, which binds critical, although distinct, epitopes in D2 (26), was not affected by FLIPr and FLIPrlike (Fig. 4C). Similar patterns of competitive inhibition of FcgRII mAb staining by FLIPr and FLIPr-like were observed on human   Fig. 2A-F). However, on B cells (expressing only FcgRIIb), binding of anti-FcgRII mAb FLI8.26 was inhibited by FLIPr (Supplemental Fig. 2C). Monocyte staining with the partially blocking anti-FcgRI mAb 10.1 (30,31) was not affected by FLIPr or FLIPr-like (Supplemental Fig. 2G). Because FLIPr-like was shown to bind and block soluble FcgRIIIa in ELISA, competitive inhibition of anti-FcgRIII mAb 3G8-mediated staining of human NK cells expressing FcgRIIIa (and FcgRIIc for some individuals) was analyzed. mAb clone 3G8, which does not distinguish between FcgRIIIa and FcgRIIIb, binds to the putative FG loop within D2 (32), the major binding site for IgG. FLIPr-like efficiently prevented the mAb 3G8-mediated FcgRIIIa staining on NK cells (Fig. 4D). In contrast to the ELISA results, FLIPr inhibited this mAb 3G8-mediated NK cell staining even more strongly. NK cell staining with the FcgRIIIspecific mAb clone B73.1, which recognizes epitopes distinct from the IgG binding site restricted to the membrane distal domain 1 (D1) (32), was not inhibited by the staphylococcal proteins (Fig.  4E). Both FLIPr and FLIPr-like demonstrated inhibition of anti-FcgRIII mAb 3G8-mediated staining of neutrophils (expressing FcgRIIIb and FcgRIIa) (Fig. 4F); this was in partial contrast to the ELISA results that indicated no binding of FLIPr to FcgRIIIb. Binding of an IgA-blocking anti-FcaR mAb to monocytes and neutrophils was not affected by FLIPr or FLIPr-like (Supplemental Fig. 2H, 2I). The control protein CHIPS did not interfere with leukocyte anti-FcgR mAb binding in any of the experiments. Altogether, FLIPr and FLIPr-like inhibited anti-FcgR mAb from binding to epitopes involved in FcgR-IgG interactions, indicating that the staphylococcal proteins bind FcgRs in or near their IgG Fc-binding regions within D2.

Evasion of FcgR-mediated phagocytosis and killing
To test whether FLIPr and FLIPr-like could inhibit FcgR-mediated effector functions, we performed phagocytosis experiments. Fluorescently labeled staphylococci were opsonized with purified human IgG, complement-inactivated human pooled serum, or untreated human pooled serum with intact complement activity.
Human neutrophils were incubated with the staphylococcal proteins at a concentration of 3 mg/ml, and phagocytosis was determined at increasing opsonin concentrations. Both FLIPr and FLIPr-like efficiently inhibited FcgR-mediated phagocytosis of staphylococci opsonized with purified IgG (Fig. 5A). Staphylococci opsonized with complement-inactivated serum were also fully protected from phagocytosis by FLIPr-like, whereas FLIPr did so less strongly (Fig 5B). In the presence of complement, the inhibitory effects of FLIPr and FLIPr-like on phagocytosis were only observed at the lower serum concentrations of 0.5 and 0.25% (Fig. 5C). The inhibitory proteins were both effective after a short preincubation with the neutrophils and when serum (or IgG), inhibitory protein, cells, and, finally, bacteria were immediately mixed together. Titration of staphylococcal protein concentrations demonstrated the 2-fold (2.3 6 0.4) higher potency of FLIPr-like, which up to 0.75 mg/ml completely inhibited FcgR-mediated neutrophil phagocytosis in the presence of complement-inactivated serum (Supplemental Fig. 3). The reduced neutrophil bacterial uptake observed in the phagocytosis experiments was reflected by an increased S. aureus survival in bacterial-killing experiments in the presence of FLIPr and FLIPr-like (Fig. 5D). FLIPr-like did not interfere with the function of FcRn, the structurally distinct, other IgGR involved in phagocytosis (Fig. 5E).

Inhibition of FcgR-mediated tumor cell killing
Next, we tested whether FLIPr and FLIPr-like also inhibited additional FcgR-dependent cellular responses, such as ADCC, using specific human mAbs against tumor targets. Neutrophil FcgRIIamediated ADCC through panitumumab (IgG2; Fig. 6A) and cetuximab (IgG1; Fig. 6B), both directed against the epidermal growth factor receptor, were almost completely prevented by 3 mg/ml FLIPr and FLIPr-like, reaching the level of lysis in the absence of sensitizing mAb (no Ab). In agreement with ELISA and competitive mAb-staining data, human IgA-mediated neutrophil FcaRI-dependent ADCC was not affected by the staphylococcal FcgR inhibitors (Fig. 6C). NK cell FcgRIIIa (and FcgRIIc in some individuals)-mediated ADCC via IgG1 was inhibited by FLIPr- like, but not by FLIPr (Fig. 6D), in accordance with their selective binding preferences to FcgRII and FcgRIII in ELISA (Fig. 2). CHIPS did not interfere with neutrophil-or NK cell-mediated ADCC. Taken together, FLIPr and FLIPr-like effectively inhibited phagocyte activation in vitro by preventing engagement and crosslinking of FcgR by IgG opsonized to target cells.

FLIPr-like prevents Arthus reaction in mice
To investigate the in vivo potency of the staphylococcal FcgR antagonists, the immune complex-mediated passive Arthus reaction model in mice was used (33). In this model, i.p. generation of immune complexes initiates a type III hypersensitivity reaction by FcgR-dependent influx of neutrophils. We confirmed that both proteins inhibited the murine FcgR-mediated phagocytosis using the mouse macrophage cell line P388D1. Similar to human neutrophils, mouse phagocytes were inhibited by FLIPr and FLIPrlike (Fig. 7A). Furthermore, FLIPr and FLIPr-like inhibited P388D1 staining with the ligand-blocking mAb 2.4G2 (34), which recognizes murine FcgRII and FcgRIII, and blocked cellular binding of rabbit IgG, which was used to generate immune complexes in the Arthus model (data not shown). Because FLIPrlike proved to be the most effective inhibitor for mouse FcgR, this protein was tested in the murine peritonitis model. Administration of FLIPr-like prior to the challenge with OVA (i.v.)-anti-OVA IgG (i.p.) completely prevented the influx of neutrophils into the peritoneal cavity (Fig. 7B). Therefore, FLIPr-like fully protected mice from an immune complex-based FcgR-dependent inflammatory disease.

Discussion
In this article, we describe a previously unrecognized role for the two homologous staphylococcal immune evasion proteins FLIPr, and FLIPr-like. They bind FcgR and abrogate their IgG-recognizing function. FcgR-bearing immune cells that are involved in crucial antibacterial defense mechanisms lose their capacity to specifically recognize pathogenic targets. S. aureus demonstrates a novel strategy to evade host immunity. Mac-2 is an FcgR-binding protein secreted by Streptococcus pyogenes. It bound soluble FcgRII and FcgRIII and competed with IgG binding (29), as well as prevented the generation of reactive oxygen species by opsonized latex beads. However, in a whole-blood survival assay of S. pyogenes, only Mac-2 with intact IgG endopeptidase activity facilitated bacterial survival (35). To our knowledge, no other bacterial FcgR antagonist has been described; therefore, this is the first study to demonstrate a bacterial FcgR-binding protein that inhibits FcgR-mediated effector functions in vitro, as well as in vivo.
FLIPr and FLIPr-like are secreted proteins encoded by genes that cluster with other immune-evasion molecules, including SCIN, Efb, and Ecb, in the immune evasion cluster II (7). The genomes of all sequenced S. aureus strains contain either the gene encoding FLIPr or FLIPr-like, suggesting that the function of these homologs is essential to S. aureus survival and replication (36). They both exhibit dual functionality, as is observed for other S. aureus immunomodulatory molecules, such as CHIPS, which interacts with the FPR and C5aR (20), and SSL7, which binds both C5 and IgA (37). We previously reported their first recognized function in immune evasion, which is directed at the escape from neutrophils and monocytes by blocking FPR-mediated cell activation and chemotaxis. In this report, we demonstrate that FLIPr and FLIPrlike also bind FcgR. Both receptor families are found on macrophages, peripheral blood neutrophils, and monocytes; however, in contrast to FcgRs, FPRs are not expressed on resting NK cells (38). It is likely that different regions within FLIPr and FLIPr-like are responsible for the inhibition of the FPR versus FcgR family of receptors that share no sequence or structural homology. Cross-talk between FPR and FcgR is described. Activation of FPR1 by fMLF, the archetype formylated peptide originating from growing bacteria, results in an increased surface expression of many cellular receptors, including FcgR (39,40). Furthermore, it was suggested that signaling of FcgRIIIb might involve FPR and that activation of FcgRIIIb on human neutrophils modulates fMLF binding and migration (41). It is not known whether FLIPr and FLIPr-like can bind to both FPR2 and FcgR on phagocytes simultaneously, thereby inducing colocalization.
Despite their high degree of homology, FLIPr was almost exclusively restricted to class II receptors, with a preference for the high-affinity polymorphic allotype of the receptor FcgRIIa-H131. FLIPr-like bound to most FcgR isoforms tested, including FcgRI, which is expressed on monocytes, macrophages, and transiently on neutrophils during infection (42). FcgRIIIa is the predominant receptor in ADCC by NK cells, although it is involved in phagocytosis by mononuclear cells, as well. FcgRIIIa is also solely bound by FLIPr-like.
All FcgRs have similar modes of binding to IgG via sites that are structurally conserved (15,43). X-ray crystallography, together with mutagenesis studies and homology modeling, demonstrated that the FcgR surface, which interacts with IgG, is primarily located in the extracellular D2 of all FcgRs, including FcgRI, which contains three (instead of two) extracellular domains. The FcgR binding site on IgG is primarily formed by residues in the lower hinge region and three segments of the CH2 domain. The fact that FLIPr and FLIPr-like bind to the FcgR, as well as antagonize FcgR-IgG interactions, was already indicated by the fact that both proteins competitively inhibited neutrophil staining with mAb clone 7.3, which was characterized as recognizing epitopes within the IgG binding site of FcgRII (26). Subsequently, other mAbs, binding to distinct epitopes in or near the IgG binding site of both  FcgRIIa (mAb IV.3) and FcgRIII (mAb 3G8), also were inhibited. mAb clone B73.1, which binds epitopes on D1 outside of the IgGligand binding domain of FcgRIII, was not inhibited by FLIPr or FLIPr-like on NK cells. The lack of inhibition of ligand-blocking anti-FcgRII mAb FLI8.26 can be explained by the fact that this mAb is a member of another cluster (II) of mAb that binds the receptor via distinct D2 epitopes. The inhibition of anti-FcgRIII mAb 3G8 from binding to neutrophils and NK cells by FLIPr, without any binding to class III receptors by ELISA, is difficult to explain. It may be that this mAb (mouse IgG1) binds with its Fc tail to FcgRIIc, also expressed by NK cells (donor dependently) (44). This discrepancy correlates with binding of fluorescent FLIPr to NK cells that do not express FPR2, the other FLIPr-binding surface receptor (21). A second indication that both proteins bind to or near the IgG-recognizing region of the FcgR was reflected in the fact that the binding of FLIPr and FLIPr-like was influenced by single polymorphic differences in the FcgR ectodomain IgG-binding amino acids (FcgRIIa-131 and FcgRIIIa-158). The ELISA data with regard to recombinant expressed soluble receptor variants show a good correlation between protein binding and functional inhibition of IgG1-ligand interactions. However, a disadvantage is the variation in affinity between the FcgRs for monomeric IgG (45), although the IgG1 concentration was optimized for each FcgR class.
The most important FcgR-mediated immune response in the battle against S. aureus infection is phagocytosis and subsequent killing. FcgRIIa-mediated phagocytosis of IgG-opsonized bacteria by human neutrophils was efficiently prevented by FLIPr and FLIPr-like. FLIPr-like inhibition was stronger than that of FLIPr, which could be due, in part, to the (unknown) variation in donor FcgRIIa genotype. FLIPr is a more efficient FcgRIIa-H131 inhibitor. FcgR polymorphisms have been associated with susceptibility to or severity of autoimmune and infectious disease (16). van Mirre et al. (46) showed that the ratio of activating FcRgIIa/ inhibitory FcgRIIb2 mRNA in neutrophils varies among individuals, is associated with FCgR2b promoter haplotypes, and is accompanied by differences in the responsiveness of these cells to IgG complexes. However, the same investigators showed that, in contrast to the much smaller IgG dimers and aggregates, no significant difference was observed for phagocytosis of IgG-opsonized FITC-labeled S. aureus.
The inhibition of phagocytosis in serum-containing active complement was only observed at low serum concentrations. This could indicate that, under conditions with sufficient serum (.1%), the phagocytic process operates mainly via the complement receptors, and only at low serum concentrations is the process dependent on FcgR. Therefore, one could argue that these proteins contribute minimally to the immune-evasion arsenal of staphylococci. However, the relative amounts of individual complement components and IgG levels in the interstitial tissue outside of the blood vessels are unknown and could be very different, consequently skewing toward FcgR dependency. In addition, S. aureus secretes several potent inhibitors of the complement system (5). In contrast to SpA, SSL10, and Sbi, which bind to IgG and inhibit phagocytosis, FLIPr and FLIPr-like directly target the relevant receptor of the phagocytosis process.
The ability of FLIPr and FLIPr-like to block FcgR-mediated functions was also shown in the ADCC experiments that used well-defined monoclonal human subclass IgG molecules to initiate tumor cell killing. Both FLIPr and FLIPr-like completely blocked neutrophil-mediated ADCC without interfering with the IgAdependent killing of the same tumor cell target. The NK cell FcgRIIIa-mediated ADCC was completely abolished by FLIPrlike, whereas FLIPr was ineffective. This correlates with the ELISA data demonstrating lack of binding of FLIPr to the soluble extracellular part of FcgRIIIa. In addition to its role in antitumor immunity, ADCC is an important immune response against virally infected cells. Although speculative, by inhibiting ADCC and consequently suppressing local antiviral immunity, FLIPr-like might play a pathogenic role in the observed coinfections of S. aureus and influenza virus (47).
In vitro, FLIPr-like proved to be an effective inhibitor of mouse FcgR-mediated phagocytosis. Mouse and human FcgR are structurally very similar and exhibit identical modes of interaction with IgG (15). FcgRs play a central role in the pathogenesis of immune complex-based inflammation and disease, such as Ab-induced arthritis, immune anemia, and immune complex glomerulonephritis (25,48). The reverse passive Arthus reaction in mice is a well-established in vivo model of immune complex-mediated inflammatory disease that relies on C5a/C5aR and the activating FcgR (33,49). FcR g-chain 2/2 mice, as well as C5aR 2/2 mice, exhibited abolished neutrophil influx in the peritoneum (25). Intervention with C5 activation by the staphylococcal protein Ecb or SSL7 also completely inhibited neutrophil accumulation in the peritoneum (7,50). In this study, we demonstrate that administration of FLIPr-like blocked neutrophil influx in the peritoneum by inhibition of the other arm of this inflammatory model.
S. aureus secretes a set of two homologous proteins that effectively bind and neutralize FcgR-mediated effector mechanisms. Because these proteins bind several FcgR isoforms expressed on different leukocytes, these proteins are a valuable tool in dissecting the general contribution of FcgR to inflammation and infection. Whether FLIPr and FLIPr-like contribute to the virulence of S. aureus needs to be addressed in an appropriate in vivo infection model. Effective IgG-mediated phagocytosis targeting FcgR in mice is hampered by the lack of sufficient levels of naturally occurring anti-S. aureus Abs, as shown for several surface expressed and secreted proteins by multiplex assay (51) and proteome array (52). However, the purified proteins proved to be effective as an FcgR inhibitor in mice.