The Ca2+ Dependence of Human Fcγ Receptor-initiated Phagocytosis

Differing roles for [Ca2+]i transients in FcγR-mediated phagocytosis have been suggested based on the observations that antibody-opsonized erythrocyte phagocytosis by human neutrophils shows a [Ca2+]i dependence, while that by murine macrophages appears [Ca2+]i-independent. To explore whether this difference might reflect different receptor isoforms or different cell types, we studied the [Ca2+]i dependence of receptor-initiated phagocytosis by human FcγRIIa and a panel of FcγRIIa cytoplasmic domain mutants expressed in murine P388D1 cells and by human FcγR endogenously expressed on human neutrophils and monocytes. Wild-type and point mutants of huFcγRIIa stably transfected into murine P388D1 cells have different capacities to initiate a [Ca2+]i transient, which are closely correlated with quantitative phagocytosis (r = 0.94, p < 0.0001). Phagocytosis both by huFcγRIIa in P388D1 cells and by huFcγRIIa endogenously expressed on neutrophils and blood monocytes shows [Ca2+]i dependence. Phagocytosis of antibody-opsonized erythrocytes by neutrophils demonstrated greater susceptibility to [Ca2+]i quenching compared with FcγRIIa-specific internalization with E-IV.3, suggesting that the phagocytosis activating property of FcγRIIIb in neutrophils also engages a [Ca2+]i-dependent element. In contrast, phagocytosis by human FcγRIa, endogenously expressed on blood monocytes, is [Ca2+]i-independent. Despite the importance of a consensus tyrosine activation motif for both receptors, FcγRIa and FcγRIIa engage at least some distinct signaling elements to initiate phagocytosis. The recognition that both of the phagocytic receptors on murine macrophages and human FcγRIa associate with the FcϵRI γ-chain, which contains a tyrosine activation motif distinct from that in the FcγRIIa cytoplasmic domain, suggests that [Ca2+]i-independent phagocytosis is a property associated with the utilization of γ-chains by FcγR.

Receptors for the Fc region of IgG (Fc␥R) 1 provide a link between antibody-antigen complexes and cellular-based effector functions and are critical in the regulation of the inflammatory response (1,2). Significant structural diversity between the three gene families encoding Fc␥R is observed (1)(2)(3)(4)(5). Nonetheless, Fc␥R share certain intracellular signaling pathways. The common themes in Fc␥R signaling pathways involve the activation of protein tyrosine kinases followed by a transient rise in intracellular Ca 2ϩ levels. The [Ca 2ϩ ] i increase is essential for many cellular functions and is required for the phagocyte Fc␥R-induced oxidative burst (6,7).
Many lines of evidence in both human and murine systems indicate that tyrosine phosphorylation events are critical for phagocyte Fc␥R functions, including phagocytosis (8 -10). In addition, in many systems examined, tyrosine kinase activity is required for the receptor-induced rise in [Ca 2ϩ ] i (presumably through tyrosine phosphorylation of phospholipase C␥1 and generation of inositol 1,4,5-trisphosphate). However, the role of [Ca 2ϩ ] i in Fc␥ receptor-mediated phagocytosis has been controversial. For example, work in murine macrophage cell lines suggests that transients in [Ca 2ϩ ] i are not essential for phagocytosis of antibody-opsonized erythrocytes (EA) (11)(12)(13). In contrast, phagocytosis of EA by human neutrophils is significantly impaired by chelation of intracellular calcium and abrogation of [Ca 2ϩ ] i transients (14,15). The ability of [Ca 2ϩ ] idepleted neutrophils to mediate phagocytosis initiated by other cell surface receptors suggests that the [Ca 2ϩ ] i -dependent EA phagocytosis by human neutrophils may reflect a particular property of the Fc␥ receptors on these cells (16).
Indeed, each of the studies of the [Ca 2ϩ ] i -dependence of Fc␥ receptor mediated phagocytosis has used EA probes that engage all available Fc␥ receptor types and has not systematically distinguished between different cell types or cells derived from different species. Recent data indicate that important species differences do exist for Fc␥ receptors. For example, human Fc␥RIIA and Fc␥RIIIB, expressed on neutrophils, do not have murine homologues (1)(2)(3). Human Fc␥RIIA stably transfected into the murine macrophage cell line P388D1 mediates receptor-specific phagocytosis but in a [Ca 2ϩ ] i -dependent fashion (17). In contrast, murine Fc␥RII (the IIb isoform) does not have a tyrosine activation motif nor does it trigger a [Ca 2ϩ ] transient or protein tyrosine phosphorylation (18). Murine Fc␥RII is also unable to mediate phagocytosis in macrophages in either a [Ca 2ϩ ] i -dependent or independent fashion (19). These observations, coupled with the recent data of Stendahl and co-workers (20) that [Ca 2ϩ ] i storage organelles accumulate at contact sites during phagocytosis in human neu-trophils prompted us to reexamine the question of the [Ca 2ϩ ] i dependence of Fc␥ receptor-mediated phagocytosis by human Fc␥ receptors, endogenously expressed by human cells and stably transfected into the P388D1 murine macrophage cell line.

MATERIALS AND METHODS
Reagents-NHS-LC-biotin, sulfo-NHS-biotin, and streptavidin were obtained from Pierce. BAPTA-AM and Indo-1/AM were from Molecular Probes (Eugene, OR). A 10 mM stock of BAPTA-AM in dimethyl sulfoxide was prepared and stored at Ϫ20°C. Genistein was from Life Technologies, Inc. and was stored as a 20 mg/ml stock in dimethyl sulfoxide at Ϫ20°C. Anti-Fc␥R mAbs 32.2, 22 (Fc␥RI, CD64) and IV.3 (Fc␥RII, CD32) Fab or F(abЈ) 2 fragments were obtained from Medarex (Annandale, NJ); IgM anti-H-2D d was obtained from Pharmingen (San Diego, CA). Fab and F(abЈ) 2 fragments were free of intact IgG as detected by silver stain analysis after SDS-polyacrylamide gel electrophoresis and size exclusion high performance liquid chromatography. Unconjugated F(abЈ) 2 goat anti-mouse IgG (GAM) for mAb cross-linking and phycoerythrin-and fluorescein isothiocyanate-conjugated GAM for immunofluorescence flow cytometry studies were obtained from Jackson Immunoresearch (West Grove, PA) or Boehringer Mannheim. Fetal calf serum and RPMI 1640 were obtained from Life Technologies, Inc. All other reagents were from Sigma.
Cells and Cell Lines-Mutant Fc␥RIIA cDNAs were made by oligonucleotide primer-directed site-specific mutagenesis of a human Fc␥RIIA cDNA generously provided by J. Kochan (Hoffman-La Roche, Nutley, NJ) (31). Mutants, confirmed by sequencing (Sequenase 2.0, U. S. Biochemical Corp.), were subcloned into pcEXV-3 (32) and transfected by CaPO 4 precipitation in the presence of 25-100 M chloroquine into the murine macrophage cell line, P388D1. Stable transfectants were screened and selected by flow cytometry, and assessment of receptor expression was demonstrated between 1.1 and 2.5 ϫ 10 6 receptors/cell.
Fresh anti-coagulated human peripheral blood was separated by centrifugation through a discontinuous two-step Ficoll-Hypaque gradient (33). Mixed mononuclear cells were isolated from the upper interface and washed with Hanks' balanced salt solution. Neutrophils were isolated from the lower interface and washed with Hanks' balanced salt solution. Contaminating erythrocytes in the neutrophil harvest were lysed with hypotonic saline (0.2% NaCl) for 20 s followed by 1.6% NaCl and a final wash with Hanks' balanced salt solution. After final washes, cells were resuspended to 5 ϫ 10 6 cells/ml in PBS prior to immunofluorescent staining or in RPMI containing 10% fetal calf serum prior to phagocytosis.
Treatment of cells with BAPTA (to quench intracellular Ca 2ϩ levels) or genistein (to block protein tyrosine kinase activity) was performed as described previously (10,17,34). Briefly, cells were incubated with BAPTA-AM (1-100 M) in buffer without free Ca 2ϩ for 30 min at room temperature (neutrophils and monocytes) or 37°C (P388D1 transfectants) followed by one wash. Buffer containing Ca 2ϩ was then added and handled as described below. For genistein treatment, cells were preincubated with 100 g/ml genistein for 30 min, and then the genistein was maintained at the same concentration through the phagocytic assay. Controls included loading cells with the BAPTA-AM and genis-tein solvent (dimethyl sulfoxide) at appropriate concentrations for the same period of time. As an alternative to BAPTA-AM treatment for quenching [Ca 2ϩ ] i , cells were allowed to drain their intracellular Ca 2ϩ stores as described by Rosales et al. (35). Levels of intracellular Ca 2ϩ were directly determined in all cases as described below.
Measurements of [Ca 2ϩ ] i -Intracellular [Ca 2ϩ ] i was determined in Indo-1/AM-loaded cells using an SLM 8000 fluorimeter and the simultaneous 405/490 nm fluorescence emission ratio as described previously (17,34). Briefly, suspensions of cells at 10 7 /ml in Ca 2ϩ -and Mg 2ϩ -free phosphate-buffered saline, pH 7.4, were incubated with 5 M Indo-1/AM at 37°C for 15 min and washed in PBS. Cells preparations to be opsonized with receptor-specific mAb Fab were resuspended in Ca 2ϩand Mg 2ϩ -free PBS at 10 7 cells/ml, incubated with saturating concentrations of mAb IV.3 Fab (0.5 g/ml) or mAb 22 F(abЈ) 2 (2 g/ml) at 37°C for 5 min, and washed in PBS. All cell preparations were resuspended in 1.1 mM Ca 2ϩ , 1.6 mM Mg 2ϩ PBS at 37°C for 5 min and then immediately transferred to a continuously stirred cell cuvette maintained at 37°C in the SLM 8000. With excitation at 355 nm, the simultaneous fluorescence emission at 405 and 490 nm was measured, integrated, and recorded each second. After establishing a base line for 60 s, goat anti-mouse F(abЈ) 2 was added (35 g/ml final concentration), and data acquisition was continued for an additional 3.5 min. Each sample was individually calibrated by lysing cells in 1% Triton X-100 to determine the maximal emission ratio and by adding EDTA (20 mM final concentration) to determine the minimal ratio. The Indo-1 fluorescence emission ratio was converted to [Ca 2ϩ ] i by the method of Grynkiewicz et al. (36).
Measurement of Phagocytosis-Biotinylated bovine erythrocytes (E B ) and biotinylated anti-Fc␥R were prepared as described previously (37). Briefly, erythrocytes (at 1 ϫ 10 9 cells/ml in 0.1 M carbonate buffer (pH 8.6)) were incubated with 250 g/ml of sulfo-NHS-biotin for 20 min at 4°C with mixing. E B (1 ϫ 10 9 E/ml) were coated with an equal volume of streptavidin (250 g/ml) for 30 min at 4°C with mixing. The streptavidin-coated E B (E BA ) were then washed and resuspended to 1 ϫ 10 9 erythrocytes/ml for immediate use. mAb were biotinylated with NHS-LC-biotin in 0.1 M carbonate buffer (pH 8.6). Typically, 50 g/ml NHS-LC-biotin was used to biotinylate mAb (1 mg/ml) for 60 min at room temperature with occasional mixing. The free biotin was removed by extensive dialysis against PBS (pH 7.4). Small volume dialysis (ranging from 50 -100 l for the mAb) was performed in a dialysis chamber (Pierce). mAb-conjugated erythrocytes were prepared by incubating E BA with dilutions of biotinylated anti-Fc␥R mAb (37). mAb-coated E BA were resuspended in RPMI 1640-fetal calf serum, an aliquot was removed for analysis by indirect immunofluorescence, and the remaining cells were used immediately for the phagocytosis or attachment assays. EA were prepared by incubating bovine erythrocytes with a 1:4 dilution of the maximal subagglutinating titer of rabbit anti-bovine erythrocyte IgG as described previously (38).
For quantitation of mAb-coated E BA or EA phagocytosis by fresh human cells, erythrocytes were mixed with 100 l of fresh neutrophils (5 ϫ 10 6 cells/ml) or fresh mononuclear cells (5 ϫ 10 6 monocytes/ml, determined by myeloperoxidase staining) at a ratio of 25:1 (mAb-coated E BA ) or 50:1 (EA) (37). The cell mixture was pelleted for 5 min at room temperature at 44 ϫ g and then incubated at 37°C for 20 min (neutrophils) or 1 h (mononuclear cells). After the nonphagocytosed erythrocytes were lysed with hypotonic saline (0.2% NaCl for 20 s followed by the addition of an equal volume of 1.6% NaCl), phagocytosis was quantitated by light microscopy. The data are expressed as phagocytic index (PI, the number of ingested particles/100 neutrophils or monocytes).
Phagocytosis by transfected P388D1 cells was determined in an adherent assay system. P388D1 cells (5 ϫ 10 5 cells/ml) were allowed to adhere to round glass coverslips in culture dishes overnight at 37°C. Coverslips were then transferred to clean culture dishes and EA-or mAb-coated E BA were added (50 l at 5 ϫ 10 7 erythrocytes/ml) were added and incubated for 1 h at 37°C. Noninternalized erythrocytes were lysed by brief immersion of the coverslip in dH 2 O followed by immersion in buffer. Phagocytosis was quantitated by light microscopy and expressed as phagocytic index as described above.
Heat-treated and serum-treated zymosan were prepared as described previously (38). Briefly, heat-treated zymosan were prepared by boiling 10 mg of zymosan for 10 min. Serum-treated zymosan were prepared by incubating 2 mg of zymosan with 2 ml of normal human serum for 30 min at 37°C. Following washing, both heat-treated zymosan and serum-treated zymosan were resuspended to 2.5 ϫ 108/ml. For phagocytosis, heat-treated zymosan or serum-treated zymosan were mixed with neutrophils (5:1, zymosan/neutrophil ratio), pelleted, and incubated for 20 min at 37°C. Phagocytosis was assessed by light microscopy.
Determination of Fc␥ Receptor Alleles-Determination of Fc␥RIIIb alleles, NA1 and NA2, was performed by quantitative flow cytometry with mAbs CLB-FcR-gran 1, CLB-gran 11, and GRM1 (33,38). The assignment of NA type was confirmed by leukoagglutination as described previously (38) and by immunoprecipitation of selected donors (33). Phenotyping of donors for the LR-HR alleles of Fc␥RIIa was performed by quantitative flow cytometry using mAbs 41.H16 and IV.3 as described previously (39,40).
Data Analysis-Phagocytosis data are displayed as the mean Ϯ S.D. Ca 2ϩ data are representative experiments. Differences in phagocytosis between phagocytic probes were compared with a Student's t test and differences between probes over a range of BAPTA concentrations (see Fig. 4B) was determined using two-way analysis of variance.

RESULTS
Several recent observations including information on species differences in Fc␥ receptor isoforms and function (1,2,17,19) have prompted a reconsideration of the studies supporting [Ca 2ϩ ] i -dependent and [Ca 2ϩ ] i -independent Fc␥ receptor-mediated phagocytosis. For example, studies by Stendahl et al. (20) have shown that there is an accumulation of [Ca 2ϩ ] i storage organelles during phagocytosis in human neutrophils. These observations suggest the possibility that the Fc␥ receptors expressed in human neutrophils are functionally distinct from murine Fc␥ receptors in engaging [Ca 2ϩ ] i -dependent elements for phagocytosis. Indeed, initial studies of human Fc␥RIIA truncation mutants stably transfected into P388D1 cells have shown that all truncations unable to initiate a [Ca 2ϩ ] i transient are unable to mediate receptor-specific phagocytosis (17). The evidence for an essential role for [Ca 2ϩ ] i in Fc␥RIIa phagocytosis is strengthened by the ability of BAPTA, a chelator of [Ca 2ϩ ] i , to block phagocytosis by Fc␥RIIa wild-type receptor in P388D1 cells (17). Accordingly, we have examined these relationships in a series of Fc␥RIIa transfectants with point mutations in the region of the cytoplasmic domain containing the YXXL tyrosine activation motif. Mutations in this region (Fig.  1) can lead to altered binding and activation of p72 syk , which in turn phosphorylates phospholipase C␥-1 leading to the generation of inositol 1,4,5-trisphosphate and [Ca 2ϩ ] i transients (41). Studies of transfected Fc␥R mutants indicate that receptormediated phagocytosis is also altered (42,43) Fifteen mutants were constructed (Fig. 1), and stable transfectants expressing between 1.1 and 2.6 ϫ 10 6 receptors/cell were selected. The [Ca 2ϩ ] transient observed after cross-linking each mutant receptor was measured and ranged from no response for several mutants of tyrosine residues within the tyrosine activation motif to a flux of approximately 500 nM (Fig.  2). The measured [Ca 2ϩ ] i transients were abrogated by pretreatment of cells with BAPTA, were unaffected by 10 mM EGTA extracellularly, and therefore were due to mobilization of [Ca 2ϩ ] i from intracellular stores. Among the 15 cell lines expressing different mutant Fc␥RIIa, there was no significant relationship between quantitative receptor expression measured by flow cytometry and peak [Ca 2ϩ ] i flux (p Ͼ 0.10; not significant).
The same mutants were probed for Fc␥RIIa-specific phago-

FIG. 1. Comparison of the tyrosine activation consensus motif (46) with that for human Fc␥RIIA and human
␥-chain of Fc⑀RI shows that Fc␥RIIA has a 12-residue, rather than a 7-residue, sequence between the two YXXL motifs (A). Point mutations were designed to alter both the YXXL motifs and adjacent residues (B).

FIG. 2.
A , receptor-specific fluxes in [Ca 2ϩ ] i were initiated in Indo-1loaded transfected P388D1 cells by cross-linking human Fc␥RIIa with mAb IV.3 Fab and GAM F(abЈ) 2 as described previously in selected mutant Fc␥RIIa expressing cells. Each sample was calibrated as described by Grynkiewicz et al. (36). B, receptor-specific phagocytosis in selected mutated Fc␥RIIa transfected P388D1 cells was determined using mAb IV.3 Fab coated erythrocytes as described previously (37). The mean and the standard deviation are shown (n ϭ 6 for each mutant). cytosis using the receptor-specific mAb IV.3 Fab conjugated to erythrocytes via a streptavidin bridge (E-IV.3). The density of mAb IV.3 conjugation to erythrocytes was monitored by flow cytometry. Nonbiotinylated erythrocytes and biotinylated but unconjugated erythrocytes neither bound nor were internalized by transfected or nontransfected P388D1 cells. Furthermore, erythrocytes coupled to an IgM anti-H2-D d via the streptavidin bridge bound to P388D1 as expected but were not internalized (mean attachment index ϭ 133; phagocytic index ϭ 0; n ϭ 3). Quantitative phagocytosis of E-IV.3 ranged from no internalization for the same tyrosine mutants that failed to elicit a [Ca 2ϩ ] i transient to a maximum phagocytic index of 126.5 Ϯ 16 E-IV.3 ingested per 100 cells (Fig. 2B). Among the 15 different mutants, there was no significant relationship between quantitative receptor expression and E-IV.3 ingestion (p Ͼ 0.10; not significant). As with wild-type Fc␥RIIa (17), the protein tyrosine kinase inhibitor genistein inhibited by Ͼ95% phagocytosis of E-IV.3 by the phagocytic mutant forms of Fc␥RIIa and of EA by the native murine Fc␥R in parental P388D1.
There was, however, a striking relationship between peak [Ca 2ϩ ] i and Fc␥RIIa-specific phagocytosis (Fig. 3) with a correlation coefficient of 0.94 (p Ͻ 0.0001). Importantly, even the most phagocytic of the mutant receptors was sensitive to chelation of [Ca 2ϩ ] i by 50 M BAPTA (Fig. 3, inset).
The strong correlation between [Ca 2ϩ ] i and phagocytosis for human Fc␥RIIa even in the environment of a murine macrophage cell line and the dependence of Fc␥RIIa-mediated phagocytosis on [Ca 2ϩ ] i suggested that this property might reflect the characteristics of human Fc␥RIIa per se. Since previous studies in human and murine cells had not probed Fc␥ receptor function in a receptor-specific fashion (11)(12)(13)(14)(15), we sought to explore the properties of Fc␥RIIa expressed endogenously on neutrophils. Our previous studies with neutrophils indicated that Fc␥RIIa alone can mediate phagocytosis (10). Therefore, using receptor-specific engagement of Fc␥RIIa with mAb IV.3 Fab and cross-linking with GAM F(abЈ) 2 , we defined the ability of BAPTA pretreatment of neutrophils to blunt the [Ca 2ϩ ] i response (Fig. 4A). Correspondingly, we assessed quantitative receptor-specific phagocytosis (Fig. 4B). A BAPTA dose-dependent inhibition of the Fc␥RIIa-mediated rise in [Ca 2ϩ ] i was observed with complete inhibition achieved by 50 M BAPTA.
BAPTA also reduced phagocytosis in a dose-dependent fashion with ϳ50% reduction at a loading concentration of 50 M (p Ͻ 0.002). Neither phagocytosis of heat-treated zymosan nor of serum-treated zymosan was altered by 50 M BAPTA pretreatment (both 99 -100% of untreated control cells; p Ͼ 0.5, not significant) (Fig. 4B). Higher loading concentrations of BAPTA did not lead to further decrement of E-IV.3 internalization (PI ϭ 50.6% of control at 100 M). 50 M BAPTA was also sufficient to abrogate detectable [Ca 2ϩ ] i transients initiated by 10 Ϫ7 FMLP (⌬[Ca 2ϩ ] i indistinguishable from base line, n ϭ 7).
As an alternative technique to deplete intracellular free Ca 2ϩ levels, we allowed neutrophils to incubate in Ca 2ϩ /Mg 2ϩfree media to exhaust intracellular Ca 2ϩ stores (Fig. 5) as described by Rosales and Brown (35). Fc␥RIIa-specific phagocytosis by neutrophils treated in this manner was markedly blunted relative to control cells in the presence of physiologic levels of Ca 2ϩ /Mg 2ϩ (PI ϭ 48.0 Ϯ 13.1% of control, n ϭ 3). The extent of inhibition was comparable with that achieved by BAPTA pretreatment.
To compare our results in neutrophils with those of Lew and co-workers (14), we also probed neutrophils with EA for the effects of [Ca 2ϩ ] i chelation on phagocytosis engaging both Fc␥RIIa and Fc␥RIIIb. Although the GPI-anchored Fc␥RIIIb does not mediate phagocytosis itself, it does elicit a [Ca 2ϩ ] i transient and functions synergistically with Fc␥RIIa for an enhanced phagocytic response (10). As reported by Lew (14), EA phagocytosis was profoundly reduced by [Ca 2ϩ ] i chelation (Fig. 4B)  To test this hypothesis as the basis for the known difference in EA phagocytosis by individuals homozygous for the two different alleles of Fc␥RIIIb (38,40), we examined the relative ability of Fc␥RIIIb in NA1 and NA2 homozygotes to elicit [Ca 2ϩ ] i fluxes in neutrophils. Engagement of Fc␥RIIIb by antireceptor mAb 3G8 and cross-linking with either GAM or with streptavidin leads to a [Ca 2ϩ ] i flux derived from intracellular stores (34). When the anti-Fc␥RIII mAb 3G8 IgG, a murine IgG1, was used to initiate the [Ca 2ϩ ] i transient, consistent differences between donors were noted that were attributable to the His-131/Arg-131 polymorphism of Fc␥RIIa and presumably the ability of Fc␥RIIa to engage the Fc region of 3G8 IgG and form heterotypic Fc␥RIIa-Fc␥RIIIb receptor clusters (Fig.  6A). However, no consistent difference in the magnitude of the [Ca 2ϩ ] i flux could be attributed to the phenotype of Fc␥RIIIb engaged by 3G8 F(abЈ) 2 with subsequent cross-linking in five matched pairs of NA homozygous donors (Fig. 6B). Taken together, these observations suggest that a [Ca 2ϩ ] i -sensitive signaling element is engaged by Fc␥RIIa during EA phagocytosis but that the magnitude of the Fc␥RIIIb-induced [Ca 2ϩ ] i flux per se does not explain the difference in quantitative phagocytosis between NA1 and NA2 homozygous donors.
Demonstration of the [Ca 2ϩ ] i dependence of E-IV.3 phagocytosis in neutrophils and in murine P388D1 cells resolves the apparent controversy between DiVirgilio and others about the role of [Ca 2ϩ ] i in phagocytosis (11)(12)(13)(14)(15), but it does not address the issue of whether a human homologue of a "[Ca 2ϩ ] i -independent" murine receptor would also be [Ca 2ϩ ] i independent for phagocytosis. Accordingly, we examined the effects of BAPTA on Fc␥RIIa-specific (E-IV.3) and Fc␥RIa-specific (E-22) phagocytosis by human peripheral blood monocytes. Both receptors mediate [Ca 2ϩ ] i transients after receptor cross-linking that are blocked by 50 M BAPTA (Fig. 7A). As anticipated, E-IV.3 showed more than a 50% decrement in phagocytosis after pretreatment with 50 M BAPTA ( Fig. 7B; PI ϭ 40.5 Ϯ 13.5% of control, p Ͻ 0.002). In contrast E-22, the specific probe for Fc␥RIa, showed only a minimal change that was significantly less than E-IV.3 (p Ͻ 0.002) and not significantly different from control (Fig. 7B, p Ͼ 0.5). These data for human Fc␥RIa are similar to those for phagocytic murine Fc␥ receptors on elicited peritoneal macrophages (12,44). programs. Indeed, the two Fc␥ receptors constitutively expressed on human neutrophils do not have corresponding murine homologues. While human Fc␥RIIa is phagocytic, murine Fc␥RII (the Fc␥RIIb isoform) on macrophages is unable to mediate [Ca 2ϩ ] transients, induction of tyrosine phosphorylation, or phagocytosis (18,19). These observations, coupled with the [Ca 2ϩ ] i dependence of huFc␥RIIa-mediated phagocytosis in stably transfected cells, suggest that the apparent controversy might simply reflect structurally different receptors each engaging at least some distinct signaling elements.
Our data indicate that human Fc␥RIIa, both in a murine macrophage environment and endogenously expressed in human neutrophils, shows significant [Ca 2ϩ ] i dependence for phagocytosis. This is in contrast to human Fc␥RIa, which like its murine homologue, shows minimal or no [Ca 2ϩ ] dependence. These observations emphasize that Fc␥RIIa, which has a cytoplasmic tyrosine activation motif distinct from that in the ␥-chain associated with Fc␥RIa (due to 12 versus 7 amino acids separating the YXXL sequences, respectively), engages some signaling elements distinct from Fc␥RIa. The fact that Fc␥RIIa phagocytosis in native cells was not completely [Ca 2ϩ ] i -dependent suggests, however, that this receptor may engage several signaling pathways, a property now recognized for other Fc receptors (45).
The mechanism underlying the variation in [Ca 2ϩ ] i transients and phagocytosis for different Fc␥RIIa mutants may relate to variable efficiency in engaging the SH2 domain of the protein tyrosine kinase p72 syk . Fc␥RIIa does bind p72 syk (25,26,28,29), and the binding of the ZAP70 homologue to a similar YXXL tyrosine activation motif shows a high degree of sensitivity to mutations in the flanking sequences (41, 46 -48). In the Fc⑀RI model system, which incorporates the YXXL tyrosine activation motif in ␥-chain, p72 syk binding and phosphorylation leads to tyrosine phosphorylation of phospholipase C␥1 (either directly by p72 syk or through an intermediary kinase(s)), inositol lipid breakdown with generation of inositol 1,4,5-trisphosphate, and a [Ca 2ϩ ] i flux (49 -53). Disruption of this sequence by inhibition of p72 syk with piceatannol or specific inhibitory peptide abrogates the [Ca 2ϩ ] i flux (53,54). The essential role for [Ca 2ϩ ] i in Fc␥RIIa-mediated phagocytosis is strongly supported by the correlation of [Ca 2ϩ ] i with phagocytosis between the 15 different Fc␥RIIa mutants, by the ability of BAPTA to abrogate phagocytosis by both wild-type and mutant receptors (Fig. 3) and by the ability of BAPTA to abrogate phagocytosis by Fc␥RIIa in neutrophils without affecting ingestion of zymosan and in monocytes without affecting internalization by Fc␥RIa. However, the critical [Ca 2ϩ ] i -dependent element(s) for Fc␥RIIa signaling remain unidentified at present. Preliminary experiments with cyclosporin A, an inhibitor of calcineurin that is a [Ca 2ϩ ] i -dependent phosphatase, show little or no effect on human Fc␥RIIa phagocytosis in transfected P388D1 cells, although calcineurin activity is essential for neutrophil motility in some circumstances (55). Recent data suggest that L-plastin, a [Ca 2ϩ ] i -regulated actin-bundling protein, may be a candidate for a [Ca 2ϩ ] i -dependent element essential for Fc receptor-mediated phagocytosis in neutrophils (56).
The greater susceptibility of EA phagocytosis to Ca 2ϩ buffering compared with Fc␥RIIa-specific phagocytosis with E-IV.3 suggests that the activation of phagocytosis by Fc␥RIIIb in human neutrophils has Ca 2ϩ -dependent elements. Since Fc␥RIIIb elicits a [Ca 2ϩ ] i flux and functions synergistically with Fc␥RIIa for phagocytosis, an allele-specific variation in Fc␥RIIIb-initiated [Ca 2ϩ ] i would provide a straightforward mechanism for the quantitative difference in phagocytosis shown by donors homozygous for different Fc␥RIIIb alleles. Although we could define consistent differences in [Ca 2ϩ ] i transients between individuals, with those elicited by intact antireceptor IgG relating to the His-131/Arg-131 polymorphism of Fc␥RIIa (57)(58)(59), donors homozygous for the NA1/NA2 Fc␥RIIIb alleles were not different in their ability to generate Fc␥RIIIb-specific fluxes in [Ca 2ϩ ] i . Thus, some other mechanism, perhaps relating to the differences in glycosylation of the NA1 and NA2 alleles and potential carbohydrate-mediated interactions with other cell surface molecules such as CD11b/ CD18 (60, 61) must be involved. While we cannot explain the NA1/NA2 difference in phagocytosis on the basis of quantitative differences in [Ca 2ϩ ] i interacting with the partially [Ca 2ϩ ] i -dependent phagocytosis of Fc␥RIIa, the difference in [Ca 2ϩ ] i dependence between Fc␥RIIa and Fc␥RIa underscores the fact that different Fc␥ receptors can engage distinct signaling elements. Whether these distinct elements reflect primary sequence differences in the tyrosine activation motifs used by each of the receptors or some modifying contribution of the cytoplasmic domain of the ligand binding Fc␥RIa ␣ chain (compare Ref. 52) remains to be determined. These distinct elements may converge on some cell programs, but they also provide the foundation for differences in elicited cell programs and for selective therapeutic intervention.