Stimulatory Function of Paired Immunoglobulin-like Receptor-A in Mast Cell Line by Associating with Subunits Common to Fc Receptors*

Paired Ig-like receptors (PIR) are polymorphic type I transmembrane proteins belonging to an Ig superfamily encoded by multiple isotypic genes. They are expressed on immune cells such as mast cells, macrophages, and B lymphocytes. Two subtypes of PIR have been classified according to the difference in the primary structure of the PIR transmembrane and cytoplasmic regions. These subtypes are designated as PIR-A and PIR-B. In this study, the transmembrane and cytoplasmic regions of the PIR-A subtype were shown to mediate activation signal events such as cytoplasmic calcium mobilization, protein tyrosine phosphorylations, and degranulation in rat mast cell line RBL-2H3. The association of the Fc receptor γ and β subunits with PIR-A was shown to be responsible for PIR-A function but not required for membrane expression of PIR-A on COS-7 cells. We further revealed the role of two charged amino acid residues in the transmembrane region, namely arginine and glutamic acid, in PIR-A function and its association with the above subunits. In contrast to the inhibitory nature of the PIR-B subtype, present findings reveal that PIR-A potentially acts as a stimulatory receptor in mast cells, suggesting a mechanism for regulation of mast cell functions by the PIR family.

Paired Ig-like receptor (PIR) 1 (1,2) has recently been found to be a murine receptor analogous to human Fc receptor for IgA (Fc␣R), although its binding capacity for murine IgA has not been shown. Analyses of a number of cDNA sequences and genomic clones for PIR revealed a gene family consisting of at least three isotypic genes (3). Its structural features have been determined to consist of type I transmembrane glycoprotein with six conserved Ig-like domains followed by two distinct amino acid sequences encompassing the transmembrane to cytoplasmic region. These amino acid sequences serve as the basis for classification of PIR into two subtypes, PIR-A and PIR-B (2,3). mRNA expression for both subtypes has been detected in B cells, interleukin-3-induced bone marrow mast cells, and myelomonocytic lineage cells (2,3). PIR is currently thought to be a murine receptor homologous to the human receptor ILT/LIR because of the similarity of their primary structures (3,4), their expression patterns in immune cell types except for NK cells (5,6), the polymorphic nature of their isotypes (4 -7), and chromosomal locations (3,8,9). Recent studies have demonstrated inhibitory function and recognition for human major histocompatibility complex class I and virusrelated major histocompatibility complex class I-like proteins by some isotypes of the ILT/LIR family, suggesting a regulatory function of ILT/LIR for immune responses in the context of major histocompatibility complex class I recognition as in the case of killer cell inhibitory receptor (6,10). PIR-B was shown to function as an inhibitory receptor, whereas the functions of PIR-A and ligands of the entire PIR family remain unknown.
The main feature of PIR-B subtype is to harbor the conserved amino acid motifs in a cytoplasmic region denoted as immunoreceptor tyrosine-based inhibitory motif. Inhibitory function of PIR-B has been shown in splenic B cells (11), a B cell line (12), and a mast cell line (13), and the two immunoreceptor tyrosinebased inhibitory motifs of the PIR-B cytoplasmic region have been proven to exert inhibitory signaling by recruiting proteintyrosine phosphatase, SHP-1 or SHP-2, which commonly functions as the signal transducer of immunoreceptor tyrosinebased inhibitory motif-based receptors including killer cell inhibitory receptor (14 -17), Ly-49 (18), NKG2 (19), CD22 (20 -23), and ILT/LIR (5,6,10). The inhibitory nature of PIR-B led us to postulate a role of PIR signaling in regulation of immune responses involving mast cells, B cells, and macrophages.
PIR-A is defined as a group of noninhibitory type of PIR family receptors characterized by a short cytoplasmic region that is free of any consensus amino acid sequence for activation. Instead, the transmembrane region of PIR-A harbors positively and negatively charged amino acid residues (see Fig. 7). Transmembrane-charged residues can typically be seen in stimulatory receptors mediating a variety of immune responses, such as T cell receptor, the ligand binding ␣ chains of type I and type III Fc receptors for IgG (Fc␥RI␣ and Fc␥RIII␣, respectively), Fc␣R, killer cell inhibitory receptor-2DS/3DS (alternatively called KAR), and NKR-P1 (CD161). All of these themselves have no amino acid motif for activation but associate with signaling subunits such as CD3 complex, ␥ and ␤ chains (FcR␥ and FcR␤, respectively) of type I FcR for IgE (Fc⑀RI), and DAP12 (24 -28) to generate an activation signal in response to receptor aggregation. Previous studies on T cell receptor ␣ chain and Fc␣R have demonstrated the requirement of a positively charged amino acid residue in the transmem-brane region for their function and subunit association (29,30). The presence of charged amino acid residues in the transmembrane region of PIR-A suggests the possibility that PIR-A associates with activation subunits to deliver an activation signal into the cell. Our recent observations have suggested that one of the PIR-A isotypes, previously denoted by p91D, mediates the activation signal revealed by cytoplasmic calcium mobilization and degranulation in mast cell line (13).
The present study focuses on the following two points. The first point is the mechanism of PIR-A function, and the second point is the evaluation of the role in PIR-A function of two charged amino acid residues present in the PIR-A transmembrane region. We have shown that the association of homodimeric FcR␥ chains and FcR␤ enable PIR-A to generate an activation signal in a mast cell line and that the charged amino acid residues contribute to the subunit association and stimulatory function of PIR-A.

EXPERIMENTAL PROCEDURES
Cells and Antibodies-A rat cell line, RBL-2H3, was obtained from the Japanese Cancer Research Resources Bank (JCRB, Tokyo). This cell line has been shown to undergo a mutation causing constitutively active signaling of c-kit receptor and exhibit mast cell function (31). RBL-2H3 and its transfectants were maintained in Dulbecco's modified Eagle's medium supplemented with 8% fetal calf serum, 2 mM L-glutamine, antibiotics, and 20 M 2-mercaptoethanol at 37°C in a humidified CO 2 incubator. COS-7 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 5% fetal calf serum, antibiotics, and 2 mM L-glutamine. Pervanadate was prepared from 10 ml of 5 mM sodium orthovanadate solution incubated with 56.7 l of 30% hydrogen peroxide solution for several min at 25°C (32). COS-7 cells were stimulated by 50 M pervanadate in medium for 10 min under the culture condition. The F(abЈ) 2 fragments of rat monoclonal antibody specific for mouse Fc␥RII/III (␣FcR, 2.4G2, PharMingen, San Diego, CA) were prepared with pepsin cleavage (Immobilized pepsin, Pierce) at 37°C for 4 h followed by purification with gel filtration chromatography (Amersham Pharmacia Biotech) as described previously (33). Mouse IgE and IgG1 antibodies specific for trinitrophenyl hapten (anti-TNP IgE) was prepared with DEAE-cellulose column chromatography from supernatant of hybridoma.
DNA Constructions and Vectors-Mouse Fc␥RII, Fc␥RIII␣ (donated by Dr. J. V. Ravetch, The Rockefeller University, New York, NY) and mouse FcR␥ (donated by Dr. T. Kurosaki, Kansai Medical University, Osaka, Japan) in pcEXV-3 vector and mouse FcR␤ in pSVl vector were used for the stable and transient expression studies. The cDNA fragment coding for transmembrane and cytoplasmic regions of PIR-A was prepared from spleen RNA of 129/SvJ mouse by polymerase chain reaction (PCR) using a primer pair of PAF-1 and PAR-1: PAF-1, 5Ј-GAGGGCCCCACACAATGGAGAATCTCAT-3Ј (sense primer) and PAR-1, 5Ј-AAGGGCCCATCAGCTTTATTTCCCAGCG-3Ј (antisense primer). The PCR fragment, in which the nucleotide sequence corresponding to PIR-A cDNA matched that previously reported as p91B (available from EMBL/GenBank/DDBJ under accession number AF041035; Ref. 3), was digested with ApaI and then ligated into the ApaI restriction site of mouse Fc␥RII cDNA, which locates in the pretransmembrane, in the sense orientation. Mutations corresponding to ARM and AEQ1 (see Fig. 1A) were generated by PCR as well using PAF-2 and PAF-3, respectively, instead of PAF-1: PAF-2, 5Ј-GGGC-CCCACACAATGGAGAATCTCATCATGATG-3Ј (sense primer) and PAF-3, 5Ј-GGGCCCCACACAATGCAGAATCTC-3Ј (sense primer). The replaced bases are underlined. The mutation of AEQ2 was introduced by two rounds of PCR using connective primers of PAF-4 and PAR-4: PAF-4, 5Ј-TTCTAGCCACTCAGGCTT-3Ј (sense primer) and PAR-4, 5Ј-TCGCCAAGCCTGAGTGGC-3Ј (antisense primer). The replaced bases are underlined. These primers overlap each other surrounding the residue to be changed. The first round of PCR with PAF-1 plus PAR-4 and PAF-4 plus PAR-1 generated two mutant fragments that were subsequently connected by the second round of PCR with PAF-1 and PAR-1.
Transfection and Assay for Membrane Expression-20 g of linearized DNA construct plus 1 g of linearized pSV2-Neo vector were transfected into 5 ϫ 10 6 of RBL-2H3 cells by electroporation with single pulse conditions of 250 V and 975 F (Gene Pulser II, Bio-Rad). The selection and cloning for neomycin-resistant cells were performed for 2 weeks in the presence of 100 g/ml geneticin (Life Technologies, Inc.).
For transient expression of the receptor of interest, 3 g for a single construct or a total of 6 g for two constructs were transfected into approximately 2 ϫ 10 6 COS-7 cells by a procedure with DEAE-dextran. In short, cells were incubated with DNA and DEAE-dextran (0.4 mg/ml) in serum-free Dulbecco's modified Eagle's medium buffered with 50 mM Tris-HCl, pH 7.4, at 37°C for a few hours and then additionally treated with 0.1 mM chloroquine (Sigma) for 3 h in serum-free Dulbecco's modified Eagle's medium. Cells were harvested at 48 h after transfection. Membrane expression of the receptor of interest was monitored for live cells with flow cytometric apparatus (FACSCalibur®, Becton Dickinson, San Jose, CA) by immunostaining of R-phycoerythrin-conjugated anti-FcR (2.4G2, PharMingen), which recognized a common epitope in extracellular regions of mouse Fc␥RII and Fc␥RIII␣ (34,35). Dead cells were eliminated from the data as the positive cells for propidium iodide staining. Base line for the positive expression was determined with ␣FcR staining for mock-transfected cells or with isotype-matched control antibody (rat IgG2b).
Measurement of Cytoplasmic Calcium Mobilization-Exponentially growing 10 6 cells in 1 ml of culture medium were labeled with 2 M of Fura-2AM (Molecular Probes, Eugene, OR) for 30 min at 35°C and sensitized with 1 g/ml of biotinylated ␣FcR or 1 g/ml of biotinylated mouse IgE for 10 min at 25°C. After unbound antibody was washed out, cells in 2 ml of phosphate-buffered saline supplemented with 1 mM CaCl 2 and 1 mM MgCl 2 were stimulated with 10 g of streptavidin while agitating gently. Cytoplasmic calcium mobilization was monitored at 510 nm emission wavelength excited by 340 and 360 nm with a fluorescence spectrophotometer (Hitachi model F-4500, Hitachi Ltd., Tokyo). Calibration and calculation of calcium concentration were performed as described (36).

Transmembrane and Cytoplasmic Regions of PIR-A Sufficiently Function for Triggering Cellular Activation in RBL-
2H3-To analyze PIR-A function in the rat mast cell line, RBL-2H3, we took advantage of the chimeric receptor consisting of the extracellular region of mouse Fc␥RIIB and the Cterminal portion of PIR-A encompassing the transmembrane to cytoplasmic regions, denoted by Fc␥RII-PIR-A (Fig. 1A). This portion is highly conserved in amino acid level among presently identified PIR-A isotypes and expresses a striking difference from the corresponding portion of PIR-B. According to the designation of PIR isotypes by Kubagawa et al. (2), the isotype of PIR-A used in this study matches PIR-A6 except for one amino acid mismatch at the second serine residue in the cytoplasmic region. The strategy of chimeric receptor enables us to perform functional analyses of PIR-A-derived signaling without the ligand or antibody to PIR, both of which are not available presently, and to analyze biochemical changes upon receptor aggregation in comparison with the established positive (mouse Fc␥RIII) and negative (mouse Fc␥RIIB) control by using the same monoclonal antibody (2.4G2, denoted as ␣FcR in this report).
We successfully isolated stable clones of RBL-2H3 expressing the chimeric receptor, mouse Fc␥RIIB or Fc␥RIII, by cell surface immunostaining with ␣FcR (Fig. 1B). Immunostainings with isotype-matched control antibody or untransfected The amino acid sequence corresponding to PIR-A is boxed. Charged amino acid residues are indicated with circled symbols on above them. The predicted transmembrane region is determined according to the computed algorithm (SOSUI system) and highlighted by black background. The amino acid residue is denoted by a one-letter code. B, membrane expression of the transfected receptor on RBL-2H3 cells. Filled histogram represents the level of membrane expression of the receptor revealed by flow cytometry with R-phycoerythrin-conjugated ␣FcR (2.4G2) staining. The negative reference (shadowed) is given by staining with control antibody (R-phycoerythrin-rat IgG2b). C, degranulation response of RBL-2H3 cells on receptor aggregation. The percentage of degranulation denotes the percentage of serotonin released into medium. Receptors on the RBL-2H3 transfectant were aggregated by intact ␣FcR (5 g/ml in sensitization) plus F(abЈ) 2 goat anti-rat IgG (GAR) (shaded), F(abЈ) 2 ␣FcR (5 g/ml in sensitization) plus GAR (filled), biotinylated ␣FcR (5 g/ml in sensitization) plus streptavidin (horizontally striped), or biotinylated anti-TNP IgE (1 g/ml in sensitization) plus streptavidin (vertically striped). Negative control was given by GAR (open) or streptavidin (dotted) treatment. The value of more than 3% of the standard error of triplicate samples is indicated on each column. D, cytoplasmic calcium mobilization ([Ca 2ϩ ] i ) on receptor aggregation. The cells (10 6 ) labeled with Fura-2 were stimulated by streptavidin (the time point indicated by arrow) after sensitization with 1 g of biotinylated anti-TNP IgE (thin trace) or 5 g of biotinylated ␣FcR (thick trace) in 1 ml of culture medium. E, tyrosine phosphorylation of total cellular protein in response to the receptor aggregation. Protein phosphorylations were terminated at the time points indicated on each lane. The whole cell lysate from 5 ϫ 10 4 cells was separated on 7.5% (upper) and 15% (lower) SDS-polyacrylamide gels, and the blots were probed with anti-phosphotyrosine antibody (anti-pTyr). The clone induced and the antibodies used for sensitization are indicated on and below the blot, respectively. cells assure the specificity of 2.4G2 staining for RBL-2H3 cells. The effects of receptor aggregation on cellular activation were evaluated with degranulation revealed by serotonin release (Fig. 1C), cytoplasmic calcium mobilization (Fig. 1D), and activation-induced tyrosine phosphorylation of total cellular proteins (Fig. 1E). Every mode of aggregation of Fc␥RII-PIR-A receptor using ␣FcR induced degranulation to an extent comparable with that of Fc␥RIII, which is a well characterized stimulatory receptor in mast cells. Comparable induction of degranulation by intact and F(abЈ) 2 fragment of ␣FcR ruled out any additive stimulatory effects from recognition of ␣FcR-bearing Fc portion (rat IgG2b) by unknown receptor on RBL-2H3 cells. No degranulation was detected in wild-type RBL and Fc␥RIIB clone in response to ␣FcR stimulation, eliminating any possibility for nonspecific stimulatory effects of the reagents and methods on cellular activation. To qualify the earlier traits for cellular activation induced by the aggregation of transfected receptors, we observed the time-dependent kinetics of cytoplasmic calcium mobilization and tyrosine phosphorylation of total cellular proteins after the saturated stimulation. Fc␥RII-PIR-A aggregation elicited calcium mobilization in a manner essentially similar to that of Fc␥RIII in respect of the rapid increment comparable with that with Fc⑀RI, slow but substantial retraction, and level of calcium concentration (Fig.  1D). Fc␥RII-PIR-A aggregation also induced tyrosine phosphorylation in total cellular proteins as revealed by anti-phosphotyrosine blot (Fig. 1E). Proteins migrating around 150, 100, 70, and 30 -40 kDa were extensively phosphorylated in response to Fc␥RII-PIR-A aggregation. This induced pattern was not grossly different from those by Fc␥RIII and Fc⑀RI aggregation. These results consistently support that the PIR-A moiety corresponding to transmembrane and cytoplasmic region is sensitive to aggregation and capable of generating an activation signal in the manner similar to Fc␥RIII.
PIR-A Constitutively Associates with Homodimeric FcR␥ and an FcR␤ in RBL-2H3-PIR-A in itself is free of any known amino acid motifs for activation in its cytoplasmic region. Then the question arises as to how PIR-A generated the activation signal. The similarity of biochemical traits in activation and the conservation of charged amino acid residues in the transmembrane region (see Fig. 7) over PIR-A and Fc␥RIII␣ led us to the hypothesis of the similar or shared receptor composition between these receptors in mast cells. To identify subunits constitutively associating with Fc␥RII-PIR-A, digitonintreated cell extracts from untreated cells were subjected to immunoprecipitation with ␣FcR or control antibodies. The samples in part were prepared both under reduced and nonreduced conditions. Immunoblots following the immunoprecipitations (Fig. 2) clearly revealed the reduced and nonreduced FcR␥ subunits at 8 kDa and mainly 16 kDa, respectively, and the reduced FcR␤ subunit was at 30 kDa for Fc␥RII-PIR-A and Fc␥RIII preparations. The nonreduced FcR␤ was detected at 30 kDa in the same samples (data not shown). Other signals near 16 kDa for FcR␥ under nonreducing conditions previously have been observed as well (26,38), assumed to be FcR␥ with unknown modification. The positive reference samples of RBL-2H3 whole cell extract and the immunoprecipitation with anti-FcR␥ and anti-FcR␤ antibodies followed these signals as expected. No signal was detected in anti-FcR␥ immunoprecipitates from wild-type RBL-2H3 and Fc␥RIIB-expressing cells as well as a negative reference sample of COS-7 whole cell extract (Fig. 2). These results confirmed the fact shown by the recent studies in which FcR␥ associates with PIR-A (39,40) and is also showing the new finding that FcR␤ is in the PIR-A complex, indicating that Fc␥RII-PIR-A receptor associates with homodimeric FcR␥ and single FcR␤ in RBL-2H3 to trigger the downstream events shared with Fc⑀RI and Fc␥RIII in mast cells.

PIR-A Does Not Require FcR␥ or FcR␤ for Its Membrane Expression in COS-7 Cells-
Because the previous study has demonstrated the necessity of FcR␥ for Fc␥RIII␣ expression on the cell surface (41,42), we questioned if PIR-A required FcR␥ or FcR␤ for its expression on the cell surface as well. Transient expressions of Fc␥RII-PIR-A, Fc␥RIIB, and Fc␥RIII␣ with or without subunit are examined in COS-7 cells, in which endogenous expressions of FcR␥ and FcR␤ are not detected (Fig. 2). Membrane expression of Fc␥RII-PIR-A was detected regardless of co-expression of subunit, followed by the pattern of Fc␥RIIB whose expression on cell membrane is known to be independent of subunit expression (Fig. 3A). In striking contrast, membrane expression of Fc␥RIII␣ was dependent upon co-expression of FcR␥ as shown previously. Utilizing the Fc binding property of extracellular regions of Fc␥RII-PIR-A and Fc␥RIII, membrane expressions of these two receptors were examined by rosetting formation with mouse IgG1-opsonized sheep red blood cells. Consistent with the data from ␣FcR detection, the transfectants expressing Fc␥RII-PIR-A and expressing Fc␥RII-PIR-A plus FcR␥ displayed rosetting for the opsonized sheep red blood cells to a similar extent of that expressing Fc␥RIII␣ plus FcR␥ (data not shown), suggesting topologically normal expression of Fc␥RII-PIR-A in the absence of FcR␥. These results indicate that PIR-A expresses a different requirement of FcR␥ or FcR␤ for its membrane expression from Fc␥RIII␣ in COS-7 cells. Then a question arises as to whether or not intrinsic subunits in COS-7 allowed Fc␥RII-PIR-A expression in place of FcR␥. To answer this question in part, we attempted to detect any asso- ciation of tyrosine-phosphorylated protein with Fc␥RII-PIR-A in COS-7 cells after treatment of pervanadate, which is known as the inhibitor for protein-tyrosine phosphatases to enforce tyrosine phosphorylations of cytoplasmic proteins. Digitonin cell lysate from COS-7 or RBL-2H3 transfectants was used for immunoprecipitation with ␣FcR followed by anti-phosphotyrosine blot (Fig. 3B). Tyrosine-phosphorylated proteins corresponding to FcR␥ in size were detected in samples from transfectants with Fc␥RII-PIR-A plus FcR␥ and Fc␥RIII␣ plus FcR␥, as well as in that from Fc␥RIII␣ in RBL-2H3. However, there was no detection for tyrosine-phosphorylated protein specifically observed in the sample from Fc␥RII-PIR-A single transfectant in the range lower than 25 kDa, where ordinal activation subunits are supposed to be detected. A shorter exposed film did not show any definitively specific signals over the entire range of separation (data not shown). It is suggested that PIR-A can exist on the cell surface as both active and inactive receptors in signal transduction.
Positively Charged Arg 626 Residue in Transmembrane Region Is Necessary for PIR-A Function-A positively charged amino acid such as arginine, lysine, or histidine in the transmembrane region is frequently found in FcR␥-dependent stimulatory receptors, although this is not the case for human and mouse Fc⑀RI␣ and human Fc␥RIII␣ (see Fig. 7A). In addition to the positively charged residue, it could be pointed out as a secondary common feature that a negatively charged residue such as aspartic acid or glutamic acid exists in the C-terminal portion of the transmembrane region, suggesting a role of the negative charged residue in receptor function. Because PIR-A possesses both conserved charged residues in the transmembrane region, we accordingly evaluated the roles of Arg and Glu at positions 626 and 643 (Arg 626 and Glu 643 ), respectively, in PIR-A functions (Fig. 4). In addition, Glu at position 622 (Glu 622 ) located in the N-terminal peri-transmembrane region is found to be characteristic among the other FcR␥-dependent receptors and to be conserved at the corresponding residue of ILT1/LIR7, so that its role in PIR-A function was investigated. Point mutation at the position corresponding to each of three charged residues was introduced into the Fc␥RII-PIR-A to generate mutant chimeric receptor with single replacement of Glu 622 , Arg 626 , or Glu 643 by the noncharged residues, glutamine, methionine, or glutamine, respectively. These mutant DNA constructs and related products (protein and transfectant) are denoted by AEQ1, ARM, or AEQ2, respectively (Fig.  4A). All of the mutant constructs were successfully expressed on RBL-2H3 cells as well as the prototype Fc␥RII-PIR-A (Fig.  4B). Degranulation and calcium mobilization assay were performed to examine capacity for signal transduction of mutant receptors (Fig. 4, D and E). ARM mutation was found to totally remove the capacity for signal transduction from Fc␥RII-PIR-A receptor, and the unresponsiveness of the ARM clones could be reconfirmed with the other clones independently isolated (Fig.  4E). On the other hand, AEQ1 and AEQ2 mutation conserved PIR-A function, although delayed calcium response was observed in AEQ2 clones (Fig. 4D). These results indicate that Arg 626 is necessary and Glu 622 and Glu 643 is not critical for PIR-A function.
Both Arg 626 and Glu 643 in PIR-A Transmembrane Play an Important Role in Subunit Association with PIR-A-We examined whether the functional alteration by single mutation could be attributed to the difference in capacity of the mutant receptor to bind to subunit. Digitonin lysates from untreated cells were used for immunoprecipitation with the saturating amount of ␣FcR antibody followed by immunoblot with anti-FcR␥ or anti-FcR␤ antibody (Fig. 5A). ARM and AEQ2 mutation were found to attenuate the association of both FcR␥ and FcR␤ to the mutant receptors, although a small amount of FcR␥ and FcR␤ was still found to be associated. By densitometric analysis, the amount of subunits associating with ARM and AEQ2 mutant receptors was estimated to be 8 and 23% for FcR␥, respectively, and 7 and 16% for FcR␤, respectively, of that to wild-type Fc␥RII-PIR-A receptor. The mutation of AEQ1 did not significantly perturb the association of the subunits. The difference of subunit association in quantity did not reflect the variance of expression of subunits among the clones analyzed (Fig. 5, blots for whole cell lysates). Decrease of subunit association by ARM and AEQ2 mutations was supported in COS-7 cells co-expressing mutant receptors and FcR␥ (Fig. 5B). These results indicate that Arg 626 and Glu 643 in the PIR-A transmembrane region are respectively important for subunit association with PIR-A.
To clarify a mechanism for signal transduction by AEQ2 receptor that substantially loses the capacity for subunit association, we examined whether FcR␥ was involved in AEQ2derived signal transduction. By taking advantage of Ig Fc binding capacity of extracellular region of chimeric receptors used in this study, cells were stimulated with mouse IgG1-containing immune complex; subsequently FcR␥ was immunoprecipitated and examined by tyrosine phosphorylation by anti-phosphotyrosine blot. FcR␥ was shown to be phosphorylated in consequence of AEQ2 receptor aggregation as well as Fc␥RII-PIR-A and AEQ1 receptor (Fig. 6A), indicating the involvement of FcR␥ in AEQ2-derived signal transduction. Then cells were stimulated with IgG1-containing immune complex as well as above, subsequently chimeric receptors were immunoprecipitated, and FcR␥ and phosphorylated FcR␥ were respectively detected by anti-FcR␥ and anti-phosphotyrosine antibodies (Fig. 6B). The results indicate that phosphorylation of FcR␥ indeed takes place in the fraction associated with AEQ2 receptor but that the amount of FcR␥ associated with AEQ2 receptor remains unchanged after stimulation, suggesting the mechanism by which the minor fraction of AEQ2-subunit complex sufficiently elicits signal transduction.

DISCUSSION
The present results demonstrate the potent stimulatory function of PIR-A in mast cell line RBL-2H3, and the association of FcR␤ as well as homodimeric FcR␥ with PIR-A to activate the signaling pathway shared with Fc⑀RI and Fc␥RIII. Our results also confirmed the recently reported FcR␥ association with PIR-A (39,40). Previous studies on FcR␥-deficient mice have demonstrated that mast cells were affected in effector functions but not in ontogeny (40,42), suggesting that PIR-A may not be involved in a developmental signal to support the differentiation of mast cells. Thus, physiological PIR-A functions are discussed in relation to effector functions of mast cells. Several lines of evidence based on recent experiments in vivo have indicated that Fc⑀RI and Fc␥RIII on mast cells play an important role in triggering distinct types of inflammatory responses such as anaphylaxis and Arthus reaction (42)(43)(44)(45)(46)(47). Mast cell activation by these FcRs may also contribute to the development of chronic allergic syndromes in humans, examples of which include atopic syndrome and bronchial hypersensitivity, by means of activating other cell types with mast cell-derived inflammatory cytokines (48). These allergic manifestations can presently be attributed, at least in part, to the result of up-regulation of signals by Fc⑀RI and/or Fc␥RIII in mast cells. The present findings lead to the tempting possibility that PIR-A aggregation exerts an additive effect on the signal by FcRs and, consequently, that PIR-A functions as an accelerator in developing mast cell-related pathological manifestations.
ILT/LIRs are thought to be the human homologue of PIR, and its mRNA expression in human lung mast cells has been reported by Arm et al. (9). The transmembranes of their noninhibitory types, ILT1/LIR7, and PIR-A express strikingly conserved primary structures (Fig. 7A), suggesting that noninhibitory types ILT1/LIR7 associate with FcR␥ and FcR␤. In fact, FcR␥ association with ILT1 has been reported very recently (49). Thus, the insight from our findings may be allowed to extend to human physiology.
We have shown that PIR-A potentially acts as a stimulatory receptor, and its function relates to the association of FcR␥ and FcR␤ subunits in RBL-2H3 cells. Not all cell types bearing PIR-A express FcR␤ subunits, i.e. monocytes and granulocytes. As in the case of Fc receptors (50, 51), FcR␤ may not be necessary but may act as an accelerator for signal transduction. The role of FcR␤ in PIR-A-derived signal transduction should be addressed by further investigation. The mRNA for PIR-A and PIR-B were also detected in mature B cells that are known to express neither FcR␥ nor FcR␤. Based on current information, PIR-A cannot exert any stimulatory function, so that PIR-B has a dominant function over PIR-A in mature B cells. To further understand the mechanism of positive and negative regulations by PIR-A and PIR-B receptors, we also examined whether PIR-A requires subunits for its membrane expression. In contrast to Fc␥RIII, Fc␥RII-PIR-A did not require FcR␥ for its membrane expression in COS-7 cells as well as human Fc␣R (26,30). Our results for PIR-A expression using COS-7 cells are similar to the results in transfected 293T cells (39) but different from the results in transfected LTK fibroblasts or splenocytes from FcR␥ deficient mice (40). Because the two cell lines permissive to expression of PIR-A in the absence of FcR␥ were those transformed with the gene encoding SV40 large T antigen, a high level of PIR-A translation could cause redundant accumulation of the receptor protein in these cells, resulting in membrane expression without FcR␥ association. It is also possible that the FcR␥ requirement for PIR-A expression might differ by cell type, although the mature cell population present in the spleen requires FcR␥ for PIR-A expression (40). In this sense, the physiological requirement of FcR␥ for PIR-A membrane expression still needs to be investigated using a highly sorted cell species.
The results from mutation analyses on Fc␥RII-PIR-A demonstrate the role of transmembrane-charged amino acids of PIR-A, Arg 626 and Glu 643 , in subunit association and PIR-Amediated signal transduction. Charged amino acids of the transmembrane region are commonly found in stimulatory receptors (Fig. 7). The hydrophobic nature of ␣ helix structure is thought to be a basic requirement for membrane integration by the transmembrane region (52). Accordingly, the presence of charged amino acids is unfavorable for stable membrane expression. However, membrane expression of these stimulatory receptors is presently rationalized by association of a subunit bearing counter-charged transmembrane region to achieve hydrophobicity by neutralizing transmembrane charges. As A, total tyrosine phosphorylation of FcR␥ (arrowhead with p-␥) was detected by anti-FcR␥ (anti-␥) immunoprecipitation (IP) followed by anti-phosphotyrosine (anti-pTyr) blot. Transfected RBL cells (10 7 ) in 1 ml of medium were stimulated with IgG1-immune complex consisting of 50 g of anti-TNP IgG1 and 2.5 g of TNP 7 -OVA and were solublized in RIPA buffer. Unstimulated samples were prepared in the same way except for the addition of immune complex. ϩ, stimulated; Ϫ, unstimulated. Total amount of FcR␥ (arrowhead with ␥) was shown by anti-FcR␥ blot. B, tyrosine phosphorylation of FcR␥ (arrowhead with p-␥) associated with PIR-A related receptor was detected by anti-FcR immunoprecipitation followed by anti-phosphotyrosine blot. Transfected RBL cells (5 ϫ 10 7 ) in 3 ml of medium were stimulated with IgG1 immune complex consisting of 150 g of anti-TNP IgG1 and 7.5 g of TNP 7 -OVA and were solublized in digitonin lysis buffer. Total amount of FcR␥ (arrowhead with ␥) associated with receptors was shown by anti-FcR␥ blot.
shown Fig. 7B, FcR␥ distributes two charged amino acids, aspartic acid and arginine, in the transmembrane region at seemingly parallel positions to those of PIR-A with the opposite charges. Our results indicate that both Arg 626 and Glu 643 of PIR-A each have an effect on the binding affinity of FcR␥ and FcR␤ to PIR-A, supporting the existence of a mechanism for subunit assembly and specificity based on electrostatic protein interaction at a membrane site. We unexpectedly observed that the requirement of Arg 626 and Glu 643 for PIR-A-derived signal transduction does not parallel that for subunit association. The loss of the negative charge of Glu 643 does result in a decrease of subunit association but does not affect the capacity for FcR␥ phosphorylation and its downstream PIR-A-mediated function. These findings brought us to assume the following two mechanisms for AEQ2-derived signal transduction. The first was that an increase in subunit association with AEQ2 receptor took place along with receptor aggregation, and the second was that efficient phosphorylation of FcR␥ was undertaken by the minor fraction of AEQ2 receptor where the subunit association was resistant to mutation. Stimulation of AEQ2 receptor was found to induce FcR␥ phosphorylation in both total and AEQ2associated FcR␥ fractions to the same extent as the intact receptor, despite the fact that the amount of subunit associated with AEQ2 receptor remained much smaller than the amount of intact receptor. These findings may support the latter mechanism mentioned above and suggest the presence of a functionally competent fraction of the receptor-subunit complex in the membrane. It is important to note that our discussions were based on experiments using detergent-soluble cell fractions. Recent findings have shown the importance of detergent-insoluble fractions in signal transduction for some receptors. We did not examine whether or not AEQ2 receptor functioned in detergent-insoluble fractions. Further investigation is therefore required to understand the mechanism of receptor function and its subunit association.