Tyrosine-containing sequence motifs of the human immunoglobulin G receptors FcRIIb1 and FcRIIb2 essential for endocytosis and regulation of calcium flux in B cells.

Human B cells express two closely related immunoglobulin G receptors, FcRIIb1 and FcRIIb2, which differ by a 19 amino acid insertion in the cytoplasmic tail of FcRIIb1. The cytoplasmic tails of both isoforms contain a conserved sequence motif (AENTITYSLL) essential for mediating endocytosis via FcRIIb2. Truncation of this motif abolished endocytosis, while replacement of tyrosine (Tyr273) in FcRIIb2 by phenylalanine had no effect on the amount and kinetics of ligand uptake. Co-cross-linking of FcRIIb1 or FcRIIb2 with the antigen receptor on B cells led to an abortive calcium signal. Neither isoform interfered with the early intracellular calcium mobilization, but both prevented the opening of a plasma membrane calcium channel essential for a sustained elevated intracellular calcium level. Modulation of calcium channel activity is mediated by the same sequence motif essential for endocytosis but requires the presence of Tyr292 in FcRIIb1 and Tyr273 in FcRIIb2. Co-cross-linking of FcRIIb1 with surface IgG is associated with tyrosine phosphorylation of Tyr292, whereas Tyr272 in FcRIIb2 was not phosphorylated. Thus, FcRIIb phosphorylation is probably not directly involved in the modulation of the calcium signal but may be essential for further diversification of signals transduced via the coexpressed isoforms FcRIIb1 and FcRIIb2.

Human B cells express two closely related immunoglobulin G receptors, FcRIIbl and FcRIIb2, which differ by a 19 amino acid insertion in the cytoplasmic tail of FcRIIbl. The cytoplasmic tails of both isoforms contain a conserved sequence motif (AENTITYSLL) essential for mediating endocytosis via FcRIIb2. Truncation of this motif abolished endocytosis, while replacement of tyrosine ('&P3) in FcRIIb2 by phenylalanine had no effect on the amount and kinetics of ligand uptake. Co-crosslinking of FcRIIbl or FcRIIb2 with the antigen receptor on B cells led to an abortive calcium signal. Neither isoform interfered with the early intracellular calcium mobilization, but both prevented the opening of a plasma membrane calcium channel essential for a sustained elevated intracellular calcium level. Modulation of calcium channel activity is mediated by the same sequence motif essential for endocytosis but requires the presence of wez in FcRIIbl and T y F 2 7 ' in FcRIIb2. Co-crosslinking of FcRIIbl with surface IgG is associated with tyrosine phosphorylation of w2, whereas w7' in FcRIIb2 was not phosphorylated. Thus, FcRIIb phosphorylation is probably not directly involved in the modulation of the calcium signal but may be essential for further diversification of signals transduced via the coexpressed isoforms FcRIIbl and FcRIIb2. ~~ Immunoglobulin G receptors (FcR)' exert a suppressive signal on the antigen-induced B cell differentiation when co-crosslinked with sIg on B cells (1). Engagement of the FcR prevents the sustained elevated intracellular calcium level in response to sIg cross-linking (2,3). Therefore, the plasma membrane calcium channel has been implicated to be the target of the FcR-derived negative signal. The mechanism by which the FcR interferes with the opening of the calcium channel is elusive.
Signal transduction events that mediate calcium signaling in B cells involve the activation of protein tyrosine kinases to the antigen receptor complex (4-7). Antigen receptor clustering leads to tyrosine phosphorylation of Ig-a and Ig-p as well as PLC-ylly2 (4, €9, which in turn generates inositol trisphos-Grant SFB223. The costs of publication of this article were defrayed in * This work was supported by the Deutsche Forschungsgemeinschaft part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $To whom correspondence should be addressed: Universitat Bielefeld, Fakultat fur Chemie, Biochemie 11, Universitatsstr. 25, D-33615 Bielefeld, Germany. Tel.: 49-521-106-2079; Fax: 49-521-106-6146. The abbreviations used are: FcR, immunoglobulin G receptor; sIg, surface Ig; ahIgG, heat-aggregated human IgG; ARH1, antigen receptor phosphate-buffered saline; PLC, phospholipase C; R A M , intact rabbit homology 1 motif; Ig, immunoglobulin; mAb, monoclonal antibody; PBS, anti-mouse IgG antibodies; wt, wild type(s); FITC, fluorescein isothiocyanate.

5).
Recent studies revealed that Ig-cu mainly accounts for the calcium signal elicited after antigen receptor cross-linking although clustering of both chains led to enhanced tyrosine phosphorylation (9).
The characterization o f the biochemical events leading to the inactivation of the plasma membrane calcium channel has been further complicated by the finding that human B cells, in contrast to mouse B cells, can express all FcRII (CD32) isoforms (FcRIIa, c, b l a n d b2; 10, 11, for a recent review see Ref. 12).
The signals mediated by ARHl motif containing FcRIIa/FcRIIc isoforms in B cells have been studied recently in our laboratory (13). The FcRIIb isoforms which are the predominantly expressed FcR in human B cells residue from alternative splicing a n d differ by a 19 amino acid insertion in the cytoplasmic tail of FcRIIbl. Since human B cells in contrast to mouse B cells (FcRIIbl+/FcRIIb2-) express both isoforms we aimed to determine whether both human FcRIIb isoforms generate redundant or divergent intracellular signals when expressed as a sole FcR in the FcR-mouse B cell line IIA1.6. By generating cytoplasmic truncated FcRIIb2 mutants and by site-directed mutagenesis of structurally related tyrosine residues present in both isoforms, we found that a sequence motif (AENTITYSLL) conserved between mouse and human FcRIIb isoforms mediates calcium flwr regulation and endocytosis of small IgG complexes. Here we show that the homologous tyrosine residues within the conserved sequence motif are essential for the downregulation of the plasma membrane calcium channel. Co-crosslinking of FcRIIbl, but not of FcRIIb2, with sIgG is associated with FcR tyrosine phosphorylation.

MATERWS AND METHODS
Generation ofReceptor Mutants-The cDNA of FcRIIb2 was isolated from a placental cDNA library as described recently (14). The cDNA of FcRIIbl was generated by reverse transcriptase-polymerase chain reaction using the FcRIIb2 cDNA as a template. Briefly, a 392-base pair BamHIIBgZII fragment of PL17 cloned in pUC18 was chosen as a template for PCR. Two overlapping fragments containing the insertion were generated using the M13 reverse and sequencing primers (Pharmacia, Freiburg, Germany) and FcRIIbl-specific 5'45"GGTATCCTGGGAG-AGCTGAAATCCGC'MTTTCCTG-3') and 3' primers (5"CTC'TCCCA-ACCAGCCAATCCCACTAATCCTG-3'). Both oligonucleotides contained a BstNI restriction site, which allowed subcloning of the PCR products and generation of the complete FcRIIbl cDNA. The FcRIIa-HR cDNA has been cloned via reverse transcriptase-polymerase chain reaction from total RNA prepared from leukocytes of a systemic lupus erythematosus patient (15). The FcRIIb2 cytoplasmic deletion mutants M1 and M13 have already been described (16). The numbering refers to the number of amino acids within the cytoplasmic region of the mutants. FcRIIb2 mutants M25 and M35 were obtained by introducing stop codons at the respective positions of the cytoplasmic tail (16). Briefly, a 926-base pair EcoRVBgZII fragment cloned in M13mp18 was used as a template for PCR. The DNA was amplified using M13 reverse primer and the mutagenic oligonucleotide containing either a BamHI GGATACCCTGAGTGCAGGGAAATGGGAGAGACCCTCCCTGAGAA-(M35,5'-GCTCTGTAGGATCCTGATGACC-3') or a SphI (M25,5'-CAC-CTAATCACTTCGCATGCACCCG-3') restriction site, which allowed further subcloning of the mutant. The tyrosine residues in the cytoplasmic region of FcRIIbl and FcRIIb2 were modified into a phenylalanine using the U.S.E. in vitro mutagenesis system as recommended by the supplier (Pharmacia). Mutagenic clones were generated using the U.S.E. MluVScaI selection primer converting a Sea1 restriction enzyme site in pUC18 into a MZuI site allowing selection against wild type plasmids. In addition the mutagenic oligonucleotides (FcRIIbl-F292/ and FcRIIbl-F258; 5'-GCTCTCCCGGGA'J."I'CCCTG-3') contained a newly introduced restriction enzyme site to facilitate screening of the clones. All receptor mutants were verified by sequencing both strands by the dideoxy chain termination method (17).
I)-ansfection of IIA1.6 Cells-The A2O-derived mouse B cell line IM1.6 bearing a somatic defect in the endogenous FcR gene (18) was propagated in Click's medium containing 10% fetal calf serum (Life 'Ibchnologies, Inc., Eggenstein-Leopoldshafen, Germany) as described recently (13). For transfection the cDNAs of the wt FcRIIa, FcRIIbl, FcRIIb2, and the receptor mutants were cloned in the eucaryotic expression vector pBEHpACl8 (19) containing the puromycin resistance gene and linearized with ScaI. Transfection, selection, and isolation of stable cell clones has been recently described in detail (131. Briefly, 5 x 10' cells were transfected with 25 pg of linearized DNA by electroporation using a Bio-Rad Gene Pulser (Bio-Rad, Miinchen, Germany), applying 260 V and 960 microfarads. Selection was started 3 d after transfection in Click's medium supplemented with 3 pg/ml puromycin (Sigma, Deisenhofen, Germany). Batch cultures were tested after 2 weeks by flow cytometry for the expression of FcRII using the FcRII specific mAbATl0 (20) generously provided by Dr. M. J. Glennie. FcRII' cell clones were isolated by limited dilution and checked routinely for stable expression. In some cases FcRII expressing batch cultures were enriched before single cell cloning by panning the cells preloaded with ahIgG on Petri dishes coated with goat anti-human IgG (Dianova, Hamburg, Germany) (21). The adherent cells were recovered after trypsin (Bioehrom KG, Berlin, Germany) treatment.
Qualita~zve and Quantitative ~e a s u r e m e~t of FcR~I-med~ted ahIgG Internalization-% determine the uptake of ligand by immunofluorescence microscopy FcRII expressing IIA1.6 cells were processed as described in detail recently (13). Cells were incubated for 30 min at 4 "C with 20 pg/ml ahIgG (14). After extensive washing the cells were attached to multiwell coverslips. Endocytosis of bound ligand was initiated by incubating the cells for 30 min at 37 "C. Uptake was stopped by chilling the cells on ice, while the control cells were constantly kept on ice. After fixation using PBS containing 3.7% formaldehyde and 2% sucrose for 15 min at 4 "C, the cells were permeabilized with PBS containing 0.2% Triton X-100 for 3 min at 25 OC. Internalized ahIgG was visualized using an Fcab'), fragment of FITC-conjugated goat antihuman IgG (Dianova, Hamburg, Germany). Cells were embedded in Mowiol4-88 (Hoechst AG, FRG) and analyzed with a Zeiss Axiophot microscope. To quantify the amount of internalized ligand by flow cytometry, cells were treated as described above, omitting permeabilization and fixation of the cells. Cell surface-bound ligand was determined by flow cytometry (FACScan, Becton Dickinson, Heidelberg). The amount of internalized ligand was calculated as percent of initially bound ahIgG (ratio of the mean fluorescence intensity of cells allowed to internalize immune complexes to cells at 4 "C).
Analysis of Intracellular Calcium Mobilization-Cells were loaded with the membrane permeable acetoxymethylesters of Fluo-3 and the pH indicator SNARF-1 (both purchased from Molecular Probes, Eugene, OR) to correct for possible heterogeneity in Fluo-3 loading as described (13). The ratio of Fluo-3/SNARF-l fluorescence intensities represents an appropriate parameter for measuring calcium signals that are independent of cell volume or granularity (22). IIA1.6 cells (1 x lo6 celldml) were resuspended in Click's medium supplemented with 5% fetal calf serum and incubated with a final concentration of 1 pm of Fko-3 AM, 0.2 pm of SNARF-1, and 0.02% Pluronic (~olecular Probes, Eugene, OR) for 30 rnin at 30 "C. After dilution with the same volume of Ciick's medium supplemented with 5% fetal calf serum and further incubation for 10 min at 37 "C, cells were washed twice with Immunodetection of Phosphotyrosine-containing FcRII-Immunoprecipitation and induction of tyrosine phosphorylation was performed as described in detail recently (13). Cultured cells were washed once with serum-free Click's medium, adjusted to a concentration of 4 x 107/ml in the same medium, and prewarmed at 37 "C. Surface IgG were cross-linked by the addition of 50 pg/ml of IU"F(ab'),, while sIgG and FcR were co-cross-linked with 75 pg/ml of the intact antibody. K. Pulford (24). Antibody-FcR complexes were adsorbed to 25 pl of Protein G-Sepharose (Sigma, Deisenhofen, Germany). Control immunoprecipitates received 5 pg of mIgGl IMOPC 21, Sigma, Deisenhofen, Germany). Samples were extensively washed with RIPA buffer, and bound material was eluted by boiling with 2-fold concentrated Laemmli buffer for 5 min. Immunoprecipitates were separated by electrophoresis on 10% SDS-polyacrylamide gels (25) and were then transferred to nitrocellulose (0.2 pm) filters (Schleicher and Schull, Dassel, Germany). Unspecific binding sites were blocked by incubation in PBS, 3% skimmed milk for 1 h. Filters were probed with 1 pgiml anti-phosphotyrosine antibody 4G10 (Biomol, Hamburg, Germany) for 2 h in PBS, 3% skimmed milk and subsequently with horseradish peroxidaselabeled goat anti-mouse IgG + IgM (Dianova, Hamburg, Germany) for an additional 90 min. To exclude that the lack of detectable phosphorylation of FcRIIb2 is a result of the mAb 4G10 used, we additionally used rabbit anti-phosphotyrosine antibodies (Biomol, Hamburg, Germany). After extensive washing bound antibodies were detected using the Enhanced Chemiluminiscence detection system and Hyperfilm ECL films (Amersham Corp., Braunschweig, Germany). Filters were reprobed with the pan-FcRII mAb II1A5 raised in our laboratoryZ after the removal of bound antibodies by incubation for 30 min at 50 "C in stripping buffer (62.5 m~ Tris-HCl, pH 6.7, 100 m~ 2-mercaptoethanol, 2% (wiv) SDS). ECL films were assayed densitometrically at 350 nm (Desaga, CD60, Heidelberg) to determine the extent of FcRIIb phosphorylation. Relative phosphorylation of FcRIIbl was calculated from the ratio of the peak integral of phosphorylated to precipitated receptor.

RESULTS
The FcRIIbl and FcRIIb2 Possess Different Capacities to Mediate Endocytosis in B Cells Distinct from FcRIIa-In this study we determined the endocytosis efficiency of the FcRIIb isofonns b l and b2 expressed in B cells in comparison to the FcRIIa isoform. We extended our previous results (13) obtained for the low responder allelic variant of FcRIIa (FcRIIa-LR) with the high responder variant of FcRIIa (FcRIIa-HR). This polymorphism has been originally defined by their different affinity for murine IgGl immune complexes, which is a result of an amino acid substitution a t position 131 (arginine in FcRIIa-HR instead of histidine in FcRIIa-LR) (26). As determined by immunofluorescence microscopy, FcRIIb2 expressing cells internalize most of the bound ahIgG within 30 min (Fig. lB), while FcRIIbl-expressing cells clearly showed no intracellular staining (Fig. 1G). In contrast to FcRIIb2 we observed most of the ligand clustered in polar caps, which were undetectable at 4 "C. -. -.
V. Weinrich, manuscript in preparation. Their appearance therefore reveals the expected temperature dependence (Fig. IT).
Cell clones expressing equal numbers of FcRIIa-HR, FcRIIb2, and FcRIIbl were selected to evaluate the amount and kinetics of FcRIIb2-mediated uptake (Fig. 2). Endocytosis of bound ahIgG was assessed by flow cytometry and expressed as the difference between maximum binding at 4 "C and remaining cell surface-bound ligand a t 37 "C. As summarized in Fig. 3A, FcRIIa-HR internalizes ahIgG very rapidly, comparable to FcRIIa-LR (13). Within 5 min a plateau was reached corresponding to 50% endocytosis of the originally bound ligand. Compared to FcRIIa, FcRIIb2 mediated a slow uptake of ahIgG, with no clearly detectable maximum (Fig. 3A). After 5 min at 37 "C, only 13% of the FcRIIb2-bound ahIgG had been internalized, which is about one-third of that observed for FcRIIa. Further incubation of FcRIIb2 expressing IIA1.6 cells for up to 30 min increased the amount of internalized ligand to 33%, which is significantly lower than for FcRIIa (50%). Although both FcRIIa and FcRIIb2 mediated the endocytosis of ahIgG in B cells, the type of FcR expressed has a pronounced effect on the rate and the amount of internalized ligand which is in contrast to our previous results obtained for FcRIIa and FcRIIb2 expressed in the FcR-fibroblastic cell line BHK-21 (16). In these cells FcRIIa and FcRIIb2 revealed no difference in the amount of internalized ligand, which might be due to cell type-specific associated effector proteins. FcRIIbl, which differs from FcRIIb2 by a 19 amino acid insertion within the cytoplasmic tail, failed to mediate the internalization of bound ligand (ahIgG) (Fig. 3A).
Endocytosis of ahZgG by FcRIZb2-expressing B Cells Is Dependent on a Region between Amino Acids 25 and 35 of the Cytoplasmic Domain-By introducing stop codons into the cDNA of FcRIIb2, we generated a series of receptors with truncated cytoplasmic tails (Fig. 4). Two of the deletion mutants, FcRIIb2-M25 and FcRIIb2-M35, were designed to analyze the function of a motif present in the FcRIIb isoforms (AENTI-TYSLL). This motif nearly resembles that found in other endocytotic receptors as the mannose 6-phosphate receptor, the transferrin receptor, or the low density lipoprotein receptor, consisting of polar and positively charged amino acids proximal to a tyrosine residue (27). To analyze the functional importance of the tyrosine residue within this motif, we introduced a phenylalanine residue at position 273 in FcRIIb2 and position 292 in FcRIIbl (Fig. 4). The tyrosine located a t position 258 in the intervening 19 amino acids of the cytoplasmic tail of FcRIIbl was replaced by a phenylalanine to analyze its possible role in blocking immune complex internalization (Fig. 4).
As determined by immunofluorescence microscopy, FcRIIb2-M35 lacking the carboxyl-terminal 9 amino acids internalizes ahIgG comparable to the FcRIIb2 wt (Fig. 1C). In contrast, FcRIIb2-M25, which lacks amino acids YSLL of the motif, only revealed cell surface staining after 30 min of incubation a t 37 "C ( Fig. 1D). Similarly, mutants M13 and M1 were ineffective in triggering the uptake of ahIgG (data not shown). Replacing T y P 3 in FcRIIb2 by phenylalanine (FcRIIb2-F273) did not influence the endocytosis efficiency (Fig. 1E). Additionally, the FcRIIbl-F292 and FcRIIbl-F258 mutants were as ineffective as the FcRIIbl wt in mediating endocytosis of small immune complexes (Fig. 1, G-Z). Obviously, amino acids other than w5* within the 19 amino acid insertion of FcRIIbl, which blocks endocytosis very efficiently, must be essential (Fig. 1G).
To evaluate whether the deletion or mutation of amino acids in the cytoplasmic tail of FcRIIb2 might influence both the amount and kinetics of endocytosis, the transfectants were analyzed by flow cytometry. As summarized in Fig. 3B, the amount and the kinetics of ahIgG uptake of M35 were nearly indistinguishable from that of FcRIIb2 wt. Additionally, replacement of w73 by a phenylalanine residue affected the endocytosis of bound ahIgG only marginally. The small difference observed varied up to 5% among three different selected cell clones analyzed. Using the more sensitive method of flow cytometry compared to immunofluorescence microscopy, we could not detect internalization of ahIgG either bound to FcRIIbl (Fig. 3A) wt or the (Y+F) mutants of FcRIIbl (Fig. 3B). Similar results were obtained for the mouse FcRIIb2 isoform, but alanine instead of tyrosine significantly reduced the amount of internalized ligand (28,29). Our results show that, although endocytosis via FcRIIb2 is dependent on the presence of the tyrosine containing sequence motif, the respective tyrosine residue is irrelevant for endocytosis. Therefore, our results strongly favor a di-leucine containing endocytosis motif which is part of this homologous sequence and deleted in FcRIIb2-M25.
Cross-linking of FcRIIbl or FcRIZb2 on B Cells Does Not Digger Calcium Signaling and Tyrosine Phosphorylation of FcR-We have recently shown that in FcRIIa-LR expressing IIA1.6 cells ahIgG binding induces a transient rise in the intracellular calcium concentration independent of extracellular calcium (13). This is due to a signal transducing motif within the cytoplasmic domain of both allelic FcRIIa variants (Fig. 4). In contrast, FcRIIbl or FcRIIb2 cross-linking did not lead to an increase in the intracellular calcium concentration (Fig. 5, A  and B). Furthermore, FcRIIbl or FcRIIb2 cross-linking by ahIgG did not cause tyrosine phosphorylation of FcRIIbl and b2 (Fig. 61, whereas FcRIIa became rapidly phosphorylated (13). Thus, the minimal protein tyrosine kinase recognition

I n h~b i~~o n of the s I g~-~~u c e d Calcium Influx by FcRIIbli FcRIIbZ Is Dependent on the Presence of the Carboxyl-terminal nrosine
Residues-Cross-linking of the antigen receptor on IIA1.6 cells causes the rapid release of calcium from intracellular stores followed by a prolonged elevated intracellular calcium level (Fig. 7B 1. Activation of B cells with F(ab'), fragments of rabbit anti-mouse IgG (RA"F(ab'1,) to cross-link sIgG in the absence of extracellular calcium prevented further calcium signaling. This indicates that the long-lasting enhanced intra- cellular calcium level is due to the opening of plasma membrane calcium channels (Fig. 7A). Both human FcRIIbl (Fig.  7 A ) and FcRIIb2 (Fig. &4) expressed on mouse B cells efficiently inhibit further calcium signaling by co-cross-linking FcR with sIgG via intact rabbit anti mouse IgG antibodies (RAM). The rapid decay observed in FcRIIbl and FcRIIb2 expressing IIA1.6 cells in response to RAM strongly resembles that measured in cells stimulated with F(ab'), fragments in the absence of extracellular calcium. Deletion of 9 amino acids at the carboxyl terminus of FcRIIb2 to 35 cytoplasmic amino acids (FcRIIb2-M35, Fig. 8B) had no influence on the inhibition of calcium signaling. Further truncation of the cytoplasmic tail to 25 amino acids (FcRIIb2-M25, Fig. 8C) either deleted or altered the region involved in the inhibition of the antigen receptorinduced calcium influx.
Replacement of the carboxyl-terminal tyrosine residues in FcRIIbl (Tyr"") (Fig. 7 B ) and FcRIIb2 (Tyr"73) (Fig. 8D) by phenylalanine abolished the FcR-mediated inhibition in calcium signaling. Thus, the same region in the cytoplasmic domain of FcRIIb2, but distinct amino acids within this motif, mediate endocytosis of small immune complexes and the downregulation of the antigen receptor-induced calcium response. Replacement of Q r 2 5 8 by a phenylalanine had no influence on the inhibition of the calcium influx (Fig. 7C). Comparison of our results with those obtained by Amigorena et al. (28) for the mouse FcRIIb isoforms reveals that a relatively short, conserved stretch of amino acids corresponding to amino acids 26-31 (referred to the cytoplasmic region of human FcRIIb2) in the cytoplasmic region of both mouse and human FcRIIb isoforms must be involved in the inhibition of calcium signaling.
In particular, the presence of the tyrosine residue within the homologous amino acid sequences of the human FcRIIb isoforms is essential for the inhibition of the antigen receptorinduced calcium influx. Co-cross-linking ofFcR with sIgG Leads to Qrosine Phosphorylation of FcRIIbl But Not ofFcRIIb2 in IIA1. 6 Cells-Having demonstrated the functional significance of Qr2" in FcRIIbl and T y 1 2 7 3 in FcRIIb2 for the inhibition of calcium signaling in B cells we questioned whether these amino acids are targets of protein tyrosine kinases involved in signal transduction via these FcR. FcRIIbl and FcRIIb2 expressed in IIA1.6 cells were cross-linked either with RAM-F(ab'), or the intact rabbit antibody. Stimulation of FcRIIbl-or FcRIIb2-expressing IIA1.6 cells by sIgG cross-linking for 1, 3, 5, 10, and 15 min did not induce tyrosine phosphorylation of FcRIIbl or FcRIIb2.  ing cells (Fig. 6, lanes 4 and 10) activated for 15 min with RA"F(ab'),.
In contrast to sIgG cross-linking, we found that co-crosslinking of FcRIIbl with the antigen receptor induces tyrosine phosphorylation of FcRIIbl (Fig. 6A, lane 3). Surprisingly, FcRIIb2 was not phosphorylated under the same conditions (Fig. 6A, lane 9) although, except for the 19 amino acid insertion, its cytoplasmic region is identical with FcRIIbl. This is not the result of different kinetics of FcRIIb2 phosphorylation compared to FcRIIbl since extensive analyses of the tyrosine phosphorylation at various time points from 30 s to 1 h revealed no tyrosine phosphorylation of FcRIIb2. Due to the insertion, FcRIIbl contains 1 additional tyrosine residue (w5*) compared to FcRIIb2. Replacement of w5" + Phe258 of FcRIIbl had no influence on the rapidly induced tyrosine phosphorylation in response to FcR-sIgG co-cross-linking with R A M . Since FcRIIbl-F292 was not phosphorylated when co-cross-linked with sIgG (Fig. 6, lane 8), we conclude that the Wg2 is the sole tyrosine residue phosphorylated under these conditions. The data presented here are representative for experiments where the FcRIIbl-F292-expressing IIA1.6 cells were stimulated for 30 s to 1 h, and the assays were carried out at least in triplicate.
Furthermore, we additionally used rabbit anti-phosphotyrosine antibodies, which also gave negative results of FcRIIbl-Ty?" and FcRIIb2 tyrosine phosphorylation (results not shown). Therefore, the same tyrosine residue (W"),, which is essential for the FcRIIblwt-mediated down-regulation of the antigen receptor induced calcium signal, becomes phosphorylated under co-cross-linking conditions (Fig. 6 A , lane 6). By reprobing the phosphotyrosine blots with the FcRII-specific mAb IIlA5, we verified in each experiment that the molecular mass of the phosphoprotein corresponds to the precipitated FcRIIbl. Although FcRIIb2 and FcRIIbl-F292 were precipitated from the cell lysates in amounts comparable to FcRIIbl (Fig. 6B), we could not detect tyrosine phosphorylation of FcRIIb2 or F C R I I~~-T~?~~ in Western blots. Thus, although superficially located in the same sequence motif, the tyrosine residues of both FcRIIb isoforms are not equal substrates of protein tyrosine kinases.
Because of the coincidence of FcRIIbl phosphorylation at position 292 after co-cross-linking with sIgG and the loss of function when Wg2 was changed into an F292, we were interested in whether the kinetics of FcRIIbl phosphorylation parallels the time course of the decay of the calcium signal. This   (Fig. 9A). To correct deviations in the amount of precipitated FcR, the phosphotyrosine-containing immunoblots were subsequently reprobed with the pan-FcRII mAb IIlA5 (Fig. 9B). Both blots were densitometrically analyzed, and the amount of phosphorylated FcRIIbl was calculated from the proportion of the peak integral of phosphorylated to precipitated receptor. We observed a very rapid tyrosine phosphorylation of FcRIIbl after co-cross-linking with sIgG, reaching maximum phosphorylation within 30 s after stimulation (Fig. 9, A and C). The phosphorylation of FcRIIbl did not continue to a plateau level as observed for the antigen receptor-associated Ig-a and Ig-@ after sIg cross-linking (8) but declined to 80% of the initial value within 1 min. After 3 min at 37 "C, only 10% of the FcRIIbl was phosphoryl-ated. Phosphorylation returned to the base line within 5 min.
To compare the kinetics of FcRIIbl phosphorylation with those of the inhibition of the calcium influx, we determined the half-decay time of the calcium response after FcR and sIgG co-cross-linking. This was calculated to be about 30-40 s from the time needed to reach the peak Fluo-S/SNARF-l proportion to the half-maximal value and was comparable to that estimated after sIgM cross-linking in calcium-free buffer (Fig. 7A).
Thus, the maximum of FcRIIbl phospho~lation precedes the inhibition of the calcium influx. However, our finding that FcRIIb2 in anti-phosphotyrosine Western blots was not detectably phosphorylated argues against a general role of tyrosine phosphorylation for the FcRIIbl/FcRIIb2-mediated regulation of the calcium channel opening. In addition, we could not observe any significant difference in the half-decay time of the calcium response in FcRIIbl-or FcRIIb2-expressing cells. It is more likely that the phosphorylation of FcRIIbl is part of a yet uncharacterized signaling cascade leading to the down-regula-  (10, 11). To analyze the signals transduced via the individual isoforms, FcRIIbl and FcRIIb2 were expressed in the mouse B cell line IIA1.6, which is devoid of the endogenous FcR. However, the expression pattern of key components of the signal transduction cascade, i.e. the protein tyrosine kinases fyn, blk, and yes, is similar in mouse and human B cells (9,30).

FcRIIb-mediated Signaling in B Cells 30643
We demonstrated that the human FcRIIb2 in contrast to FcRIIbl expressed in B cells mediates the endocytosis of small immune complexes, although the amount and time course of the internalization of receptor-bound ahIgG are significantly lower compared to the FcRIIa expressed in B cells (11,13).
Endocytosis of immune complexes via the murine FcRIIb2 expressed in IIA1.6 cells has been shown to enhance antigen presentation and the proliferation of an interleukin-2-dependent T cell line (28). Because mouse B cells do not naturally express FcRIIb2 it has been speculated that the mis-regulation of FcRIIb2 expression in B cells may lead to autoimmunity. Therefore, homologous expression of FcRIIb2 in human B cells provides a mechanism leading to the internalization, processing, and presentation of antigen independent of the specificity of the antigen receptor. Previous results have shown that the activation of CD4' T cells requires a co-stimulatory signal provided by the antigen-presenting cell. In the absence of a costimulatory signal, T cells were tolarized to the respective antigen (31,32). Thus, T cell activation may depend on the state of activation of the antigen-presenting B cell. Moreover, although both FcRIIa and FcRIIb2 mediate the uptake of immune complexes in B cells, FcRIIb2 might be involved in the regulation of cellular processes different from those involving FcRIIa: FcRIIa containing the tyrosine-dependent cell activation motif induces intracellular calcium release in contrast to FcRIIb2 (11, 13, 33). Furthermore, ligand binding to FcRIIb2, in contrast to FcRIIa (13), is not associated with receptor tyrosine phosphorylation. Thus, receptor clustering per se is insufficient to induce FcRIIb phosphorylation.
The sequence motif, which accounts for the coated pit localization of FcRIIb2 recently demonstrated by us (34), can be limited to the last 4 amino acids of AENTITYSLL deleted in FcRIIb2-M25. Since replacement of the tyrosine residue ( T~I ?~~) by phenylalanine only slightly affected endocytosis, similar to the mouse FcRIIb2 (29), Ty?73 does not contribute to a functional coated pit localization signal. Several integral membrane proteins such as the CD3-7 chain (35) take advantage of a leucine-leucine motif within the internalization signal. A very similar sequence motif occurs in FcRIIb2 and has been deleted in the M25 mutant (Fig. 9). Based on these data we suggest that the 2 leucines are essential for the FcRIIb2mediated endocytosis of small immune complexes.
Soluble FcR generated either by proteolytical cleavage or alternative splicing have been shown to regulate B cell activation (36,37). Shedding of FcRII has been observed after antigen-driven B cell activation and may involve FcRII phosphorylation (38). Ligation of sIgG on IIA1.6 with RA"F(ab'1, fragments as a surrogate for antigen did not induce any tyrosine phosphorylation of FcRIIbl or FcRIIb2. Furthermore, the phosphorylation of mouse FcRIIbl after antigen receptor crosslinking was on serine and threonine residues (39). Although fyn is associated with FcRII in resting and activated human B cells (30), FcRII was found to be predominantly phosphorylated on serine and threonine residues. Since we could not observe any tyrosine phosphorylation of both FcRIIb isoforms expressed in mouse B cells after antigen receptor cross-linking, we suggest that serinelthreonine rather than tyrosine phosphorylation of FcRIIb is probably involved in the shedding of FcRII from B cells. B cell proliferation can also be inhibited upon sIg-FcR ligation with antigen-antibody complexes (1). Co-cross-linking of either FcRIIbl or FcRIIb2 with sIg markedly influences the antigen receptor-induced calcium signaling by preventing the sustained, prolonged, elevated calcium level seen after crosslinking sIg with RA"F(ab'),. In contrast, the first rapid rise in the intracellular calcium concentration due to the opening of inositol trisphosphate-sensitive calcium stores remains unaffected. Thus, engagement of the FcRIIb isoforms interferes with the opening of a yet unknown plasma membrane calcium channel (3). Our data show that the region involved in the inhibition of further calcium signaling overlaps with the region essential for mediating endocytosis since truncation of the cytoplasmic tail to 25 amino acids (FcRIIb2-M25) abolished the negative effect of FcRIIb2 on the calcium influx in B cells. In addition, deletion of the entire region in the mouse FcRIIb2 impaired the negative influence on calcium signaling (28). However, in contrast to the sequence motif required for endo-cytosis, replacement of the tyrosine residues by phenylalanine affected the receptors' ability to modulate the plasma membrane calcium channel activity. Most notably, although w73 (FcRIIb2) and Wg2 (FcRIIbl~ are located within the same conserved sequence motif they are not equal substrates for protein tyrosine kinases. Surprisingly, only FcRIIbl and FcRIIbl-F258 were found to become detectably phosphorylated in anti-phosphotyrosine Western blots. Our results clearly demonstrate that FcRIIbl is phosphorylated on Wg2, which as has been demonstrated by amino acid sequence alignment i s the homologous tyrosine residue to w73 of FcRIIbZ. Thus, the 19 amino acid insert in FcRIIbl not only abolished the endocytosis capacity of FcRIIbl but positively influenced the accessibility of a possible target sequence for a protein tyrosine kinase yet to be characterized. Moreover, the homologous tyrosine residue in the mouse FcRIIbl is also essential for the inhibition of the calcium influx and becomes phosphorylated when FcRIIbl and sIgG are co-cross-linked (40). We suggest, that the inserted 19 amino acid sequence renders the overall folding of the cytoplasmic region of FcRIIbl. This might alter the tyrosine phosphorylation motif, which is part of both FcRIIbl and FcRIIb2, resulting in a different folding pattern of the cytoplasmic tail. Recent results obtained by Sarmay et al. (30) have shown that FcRIIb in human B cells is associated with the protein tyrosine kinase fyn, which supports our result that FcRIIbl expressed in mouse €3 cells is a phosphor acceptor. O u r finding, that the down-regulation of calcium signaling by both FcRIIb isoforms is dependent on the presence of a conserved tyrosine residue, which only becomes phosph0rylated in one receptor isoform (FcRIIbl), makes it difficult to decide whether tyrosine phospho~lation is involved in the FcRIIbmediated regulation of the plasma membrane calcium channel.
Whether the mouse FcRIIb2 isoform in the mouse B cell line IIA1.6 becomes tyrosine phosphorylated when co-cross-linked with sIgG remains to be determined. We suggest that due to the inserted sequence found in FcRIIbl the human FcRIIb isofoms associate with distinct sets of cytoplasmic effector proteins, which may initiate distinct signal transduction pathways. The characterization of isoform-specific associated cytoplasmic proteins will help to understand the precise role of the tyrosine residues within the cytoplasmic tails of FcRIIb2 and FcRIIbl.
Apart from the regulation of calcium signaling in B cells, tyrosine phosphorylation of FcRIIbl may contribute to the differential regulation of the activity of the two FcRIIb isoforms. Thus, tyrosine phosphorylation could positively or negatively regulate isoform-specific effector functions. erous gift of mAb KB61, Dr. M. J. Glennie for the AT10 hybridoma cell