FcαRI (CD89) Alleles Determine the Proinflammatory Potential of Serum IgA1

The human IgA FcR (FcαRI; CD89) mediates a variety of immune system functions including degranulation, endocytosis, phagocytosis, cytokine synthesis, and cytokine release. We have identified a common, nonsynonymous, single nucleotide polymorphism (SNP) in the coding region of CD89 (844A→G) (rs16986050), which changes codon 248 from AGC (Ser248) to GGC (Gly248) in the cytoplasmic domain of the receptor. The two different alleles demonstrate significantly different FcαRI-mediated intracellular calcium mobilization and degranulation in rat basophilic leukemia cells and cytokine production (IL-6 and TNF-α) in murine macrophage P388D1 cells. In the absence of FcR γ-chain association in P388D1 cells, the Ser248-FcαRI allele does not mediate cytokine production, but the Gly248-FcαRI allele retains the capacity to mediate a robust production of proinflammatory cytokine. This allele-dependent difference is also seen with FcαRI-mediated IL-6 cytokine release by human neutrophils ex vivo. These findings and the enrichment of the proinflammatory Gly248-FcαRI allele in systemic lupus erythematosus populations in two ethnic groups compared with their respective non-systemic lupus erythematosus controls suggest that FcαRI (CD89) α-chain alleles may affect receptor-mediated signaling and play an important role in the modulation of immune responses in inflammatory diseases.

I mmunoglobulin A has a central role in host defense at mucosal surfaces, but its role in systemic immunity is less clear. In the intestinal lamina propria the B lineage cells produce polymeric IgA, which exists almost exclusively as a dimer joined by a J chain polypeptide and linked to the secretory component to form a complex called secretory IgA. Secretory IgA can bind Fc␣RI (CD89) but cannot trigger phagocytosis (1,2). Nevertheless, through Fc␣RI secretory IgA is able to initiate certain inflammatory responses such as a respiratory burst in polymorphonuclear leukocytes and degranulation of eosinophils (2,3). In human serum, IgA, the second most abundant Ig after IgG, makes up about one-fifth of the total Ig, is primarily of the IgA1 subclass, and exists mainly in monomeric form (Ͼ95%) (2). The two human IgA subclasses, IgA1 and IgA2, differ by an additional 13-aa sequence with O-linked glycosylation sites in the hinge region of IgA1 (4). The lack of 13 aa in the hinge confers on IgA2 its resistance to digestion by the bacterial proteases produced by microorganisms such as Streptococcus mutans, Neisseria meningitidis, and Haemophilus influenzae and may be the rationale for the predominance of IgA2 in mucosal secretions (5).
IgA-mediated immune effector responses such as phagocytosis, Ab-dependent cell-mediated cytotoxicity, respiratory burst, and cytokine release are primarily mediated through Fc␣RI (CD89), an IgA-specific receptor. Fc␣RI is a type I transmembrane (TM) 3 receptor expressed on cells of the myeloid lineage including neutrophils, monocytes, tissue macrophages, eosinophils, and subpopulations of dendritic cells (6). Both monomeric and dimeric IgA can bind to Fc␣RI, and IgA immune complexes can activate phagocytosis and other immune responses through the clustering of Fc␣RI (7).
The gene encoding human Fc␣RI is located in the leukocyte receptor cluster on human chromosome 19q13.4 along with NK cell inhibitory receptors and leukocyte Ig-like receptors (8,9). The Fc␣RI gene consists of five exons and spans ϳ12 kilobases. The signal peptide is encoded by first two exons, whereas the extracellular (EC) domains (EC1 and EC2) are encoded by exons 3 and 4, respectively. The last exon (exon 5) encodes the TM region and the short cytoplasmic (CYT) tail (10). Several alternatively spliced variants of cDNAs have been described, and a shorter version lacking 66 bp encoding EC2 is expressed on alveolar macrophages in vivo (11)(12)(13)(14)(15)(16). A significant decrease in the ratio of the EC2 deletion splice variant to the full-length Fc␣RI has been found in neutrophils from pneumonia patients compared with those of healthy individuals, suggesting that the alternatively spliced Fc␣RI isoform might have physiologic relevance in IgA-mediated host defense (16,17).
Fc␣RI shares sequence similarity with the leukocyte Ig-like receptors, killer inhibitory receptors, and other members of the leukocyte receptor cluster (8,9). Unlike the classical low affinity Fc␥RIII and high affinity FcRI, which bind their ligands via the membrane proximal EC Ig-like domain (EC2) and form 1:1 receptor:Fc complexes through interactions with hinge-proximal Fc regions (18,19), Fc␣RI has its ligand binding site on the membrane distal EC1 domain and forms a 2:1 complex with an Fc␣ dimer by binding to each C H 2-C H 3 domain interface (20 -25). Although Fc␣RI is only distantly related to Fc␥R and FcRI proteins, it does share some structural similarities and can associate with the FcR ␥-chain through its TM domain. Unlike Fc␥RIII and FcRI, however, Fc␣RI is often expressed in the absence of ␥-chain pairing (26,27) and the function of this nonpaired Fc␣RI has not been defined. Indeed, the net function of Fc␣RI per se is unresolved, with some investigators reporting a capacity to initiate a proinflammatory program (28 -30) while others report a noninflammatory profile (31)(32)(33)(34)(35)(36).
Functionally important single nucleotide polymorphisms (SNPs) in the coding regions of classical Fc␥Rs have provided significant insight into their biological functions (37,38). As part of an effort to understand the role of naturally occurring polymorphisms in Ig receptors, we sequenced the entire coding region of Fc␣RI from a large number of donors and confirmed a common nonsynonymous SNP (SNP 844A3 G) within the Fc␣RI CYT domain (39). Using this natural sequence variation as a biological probe, we have demonstrated that the two different alleles of the CYT domain of CD89 ␣-chain alter receptor signaling and its biological functions. In addition, we found that, in contrast to the Ser 248 -Fc␣RI allele, the Gly 248 ␣-chain allele of Fc␣RI is able to mediate certain biological functions in the absence of the FcR ␥-chain. Thus, this functionally important, nonsynonymous SNP in the CYT domain not only plays an important role in the regulation of Fc␣RI function but may also explain some of the divergent reports on the proinflammatory vs anti-inflammatory potential of Fc␣RI (6).

Donors
Anti-coagulated peripheral blood was obtained from healthy normal volunteers and from SLE patients fulfilling the revised American College of Rheumatology criteria for SLE (40). SLE patients and controls were recruited as part of the University of Alabama at Birmingham DISCOVERY cohort. The human studies were reviewed and approved by the Institutional Review Board of the University of Alabama at Birmingham, and all donors provided written informed consent.

Reagents
All mAbs used in this study were murine in origin. mAbs A59 and A77 are murine IgG1 (mIgG1) specific for Fc␣RI (CD89; gift from Dr. H. Kubagawa, University of Alabama at Birmingham, Birmingham, AL), and the 32.2 (mIgG1) mAb is specific for the Fc␥RI (CD64; American Type Culture Collection). The mAbs specific for CD89 and CD64 were produced and purified from the hybridoma cultures at the University of Alabama at Birmingham Epitope Recognition and Immunoreagent Core Facility. The F(abЈ) 2 of the mAbs A59 and 32.2 were generated at Rockland Immunochemicals, and the purity of these Ab F(abЈ) 2 preparations was monitored by SDS gel staining and Western blotting assays. Anti-human Fc␣RI mAb MIP8a (mIgG1) was purchased from Serotec. Anti-Fc␥RI-FITC (mAb 197 (mIgG2a) and mAb 22.2 (mIgG1)) were from Medarex. F(abЈ) 2 of mIgG and FITC-conjugated and unconjugated goat anti-mouse IgG (HϩL) were obtained from Jackson ImmunoResearch. Rabbit anti-Lyn polyclonal Ab was purchased from Santa Cruz Biotechnology. Rabbit anti ␥-chain polyclonal Ab was obtained from Upstate Biotech. An AminoLink kit (cyanogen bromide-activated Sepharose 4B coupling gel and reagents) was purchased from Pierce. Protein G-agarose and a protease inhibitor mixture were from Roche Diagnostics. Cytokine detection ELISA kits and substrates were from BD Pharmingen. Human IgA was obtained from Sigma-Aldrich (catalog no. I 1010).

Nucleic acid isolation
Genomic DNA was isolated using the Puregene DNA isolation kit (Gentra Systems). Briefly, 300 l of blood was lysed in 900 l of the red blood lysis solution. The PBLs were pelleted and lysed with 300 l of cell lysis solution. The RNase A solution was added to the cell lysate and incubated at 37°C for 15 min. Proteins were precipitated by adding 100 l of Protein Precipitation solution (Gentra Systems). DNA in the supernatant was precipitated with 300 l of 100% isopropanol and washed with 70% ethanol once. Total RNA was isolated from 10 7 PBL by using TRIzol total RNA isolation reagent (Invitrogen Life Technologies). Five micrograms of total RNA was used to synthesize cDNA with the SuperScript preamplification system (Invitrogen Life Technologies).

RT-PCR and cDNA sequencing
To facilitate heterozygote detection, a dye primer strategy was adopted for fluorescence-based automated cycle sequencing of PCR products on an ABI 377 sequencer (ABI PRISM dye primer cycle sequencing 21M13 FS and M13REV FS ready reaction kits; Applied Biosystems). Two sets of overlapping primers with either M13 universal or reverse primer sequence tags (underlined letters in the sequences below) at the appropriate 5Ј ends were used to amplify the whole coding region starting from nt 40 and ending at nt 903 of the Fc␣RI cDNA (GenBank accession no. X54150). The first set of primers amplifies from nt 34 to nt 563, yielding a PCR product 585 bp in length including the underlined M13 sequences: 5Ј-TG TAAAACGACGGCAGTAGCACGATGGACCCCAAACAG-3Ј (annealing between nt 34 and nt 54) and 5Ј-CAGGAAACAGCTATGACCGGTG TTCCCCACTTTGGTGC-3Ј (annealing between nt 542 and nt 563). The second set of primers amplifies from nt 458 to nt 934, yielding a the PCR product 513 bp in length including the underlined M13 sequences: 5Ј-TG TAAAACGACGGCCAGTAGAATATTTCCCTCACGTGC-3Ј (annealing between nt 458 and nt 497) and 5Ј-CAGGAAACAGCTATGACCCTGG CTCCTCTCTGCCTTCAC-3Ј (annealing between nt 914 and nt 934). The PCR was performed in a 9600 PCR system with 2 l of cDNA synthesized with the SuperScript preamplification system, 200 nM each primer, 200 M dNTPs, 1.5 mM MgCl 2 , and 2.5 U of TaqDNA polymerase in a 50-l reaction volume starting at 95°C for 5 min, 35 cycles of denaturing at 94°C for 30 s, annealing at 56°C for 45 s, and extension at 72°C for 1 min with a final extension at 72°C for 7 min. The PCR product was purified with the QIAquick gel extraction kit (Qiagen). All PCR products were sequenced in both directions.

Genotyping Fc␣RI CYT domain SNPs by direct sequencing
To accelerate the genotyping of Fc␣RI SNPs, a PCR of genomic DNA was designed for the fifth exon encoding for the TM segment and the CYT tail of Fc␣RI (Fig. 1A). The forward primer 5Ј-CCATCCACCAAGATTACA CGA-3Ј anneals between nt 692 and nt 712 and the reverse primer 5Ј-CT GGCTCCTCTCTGCCTTCAC-3Ј anneals between nt 914 and nt 934 (GenBank accession no. X54150). The PCR was performed with 200 ng of DNA, 200 nM each primer, 200 M dNTPs, 1.5 mM MgCl 2 , and 2.5 U of TaqDNA polymerase in a 50 l reaction volume starting at 95°C for 5 min, 35 cycles of denaturing at 94°C for 30 s, annealing at 58°C for 45 s, and extension at 72°C for 1 min with a final extension at 72°C for 7 min. All the PCR products were purified from a 2.5% agarose gel with the QIAquick gel extraction kit. The purified genomic PCR products from different donors were sequenced from both directions using the BigDye terminator sequencing kits on an ABI 377 (ABI PRISM BigDye terminator cycle sequencing ready reaction kits; Applied Biosystems).

Generation of the Fc␣RI expression constructs
The human Fc␣RI (CD89) expression constructs were generated by cloning KpnI/EcoRI-flanked RT-PCR products with 864 bp of the Fc␣RI coding region into the eukaryotic expression vector pcDNA3.1 (Invitrogen Life Technologies). The KpnI/EcoRI-flanked RT-PCR products from a heterozygous SNP 844G3 A donor were generated from human mixed mononuclear cell cDNA synthesized using the SuperScript preamplification system (Invitrogen Life Technologies), with the upper primer 5Ј-TAATAGGG TACCATGGACCCCAAACAGACCACCCTC-3Ј (underlined and bold nucleotides are the KpnI cutting site) annealing between nucleotide positions 40 and 62 and lower primer 5Ј-TTATTAGAATTCTTACTTGCAG ACACTTGGTGTTCGT-3Ј (underlined and bold nucleotides are EcoRI cutting site) annealing between nucleotide positions 879 and 904. RT-PCR was performed with KOD HiFi DNA polymerase (EMD Biosciences; distributed by Novagen). A point mutation at nucleotide position 728, which changes amino acid Arg 209 to Leu 209 in a TM segment, was made with the QuikChange site-directed mutagenesis kit (Stratagene) following the vendor's instruction. For Arg 209 to Leu 209 constructs (nt 728 changes), the sense primer 5Ј-AGAACTTGATCCTCATGGCCGTGGC-3Ј and antisense primer 5Ј-GCCACGGCCATGAGGATCAAGTTCT-3Ј were used on the template plasmids from Fc␣RI expression constructs with either the 844A (Ser 248 -Fc␣RI) allele or the 844G (Gly 248 -Fc␣RI) allele (underlined, bold, and italicized letters are intentional mutations). The sequences of all the cloned constructs were confirmed by direct sequencing from both directions on an ABI 377 sequencer with the ABI BigDye terminator cycle sequencing kit (Applied Biosystems).

Generation of the CD64EC-TM/CD89CYT chimeric receptor expression constructs
The expression constructs for a chimeric receptor comprised of human Fc␥RI (CD64) EC domains and a TM segment (CD64EC-TM) fused with Fc␣RI (CD89) CYT domain (CD89CYT) were generated by overlapping PCR to create KpnI/BamHI-flanked RT-PCR products of chimeric cDNA. The resultant CD64EC-TM/CD89CYT coding region was inserted into pcDNA3.1 (Invitrogen Life Technologies). The CD64EC-TM cDNA fragment was amplified from human cDNA with the upper primer 5Ј-TAA CGGGGTACCGGAGACAACATGTGGTTCTTGACAACT-3Ј (underlined and bold nucleotides are the KpnI cutting site) annealing between nucleotide positions 48 and 54 and the lower primer 5Ј-ATGGCTGTGCCAATTTTCTA TTGTCACCCAGAGAAC-3Ј annealing between nucleotide positions 958 and 975 (CD64; GenBank accession no. X14356). The CD89CYT cDNA fragment was amplified with the upper primer 5Ј-GTTCTCTGGGTGACAA TAGAAAATTGGCACAGCCAT-3Ј annealing between nucleotide positions 778 and 795 and the lower primer 5Ј-TTACTAGGATCCTTACTTGCAG ACACTTGGTGTTCGT-3Ј (underlined and bold nucleotides are the BamHI cutting site) annealing between nucleotide positions 879 and 904 (Fc␣RI GenBank accession no. X54150). The PCR fragments of CD64EC-TM and CD89CYT were purified with QIAquick gel extraction kit and mixed together with the upper primer 5Ј-TAACGGGGTACCGGAGACAACATGTGGTTC TTGACAACT-3Ј and the lower primer 5Ј-TTACTAGGATCCTTACTTGC AGACACTTGGTGTTCGT-3Ј (underlined and bold nucleotides are the KpnI and BamHI cutting sites) to generate RT-PCR products for the chimeric receptor. RT-PCR was performed with KOD HiFi DNA polymerase (EMD Biosciences; distributed by Novagen). The sequences of all cloned constructs were confirmed by direct sequencing from both directions on an ABI 377 sequencer with ABI BigDye terminator cycle sequencing kit (Applied Biosystems).

Generation of expression constructs for the GST-CD89CYT fusion protein
The GST/Fc␣RI (CD89) CYT domain fusion protein (GST-CD89CYT) expression constructs were generated through cloning BamHI/EcoRIflanked RT-PCR products with 126 bp of the CD89 CYT domain coding region (41 aa residues) into pGEX2T vector (Amersham Biosciences). The BamHI/EcoRI-flanked RT-PCR products were amplified with PBL cDNA from a heterozygous SNP 844 donor using the upper primer 5Ј-CCGC GTGGATCCATGGACCCCAAACAGACCACCCTC-3Ј (underlined and bold nucleotides are the BamHI cutting site) annealing between nucleotide positions 778 and 797 and the lower primer 5Ј-TTATTAGAATTCTTAC TTGCAGACACTTGGTGTTCGT-3Ј (underlined and bold nucleotides are the EcoRI cutting site) annealing between nucleotide positions 879 and 9044. RT-PCR was performed with TaqDNA polymerase (Invitrogen Life Technologies). The GST fusion constructs were transformed into DH5␣ Escherichia coli and the fusion proteins were produced and purified with glutathione-Sepharose beads (Amersham Biosciences) according to the vendor's instructions. The sequences of all cloned constructs were confirmed by sequencing from both directions on an ABI 377 sequencer with an ABI BigDye terminator cycle sequencing kit (Applied Biosystems).

Generation of stable cell lines expressing Fc␣RI
The murine macrophage cell line P388D1 and the rat mast cell line rat basophilic leukemia (RBL)-2H3 were obtained from American Type Culture Collection and maintained in DMEM supplemented with 10% FCS (FBS) and L-glutamine (2 mM) in 5% CO 2 . Individual transfections were conducted on a 60-mm cell culture dish with the cell density at 80 -100% confluence. Fc␣RI expression construct plasmid DNA (8 g), purified with the Wizard PureFection plasmid DNA purification system (Promega) and the Lipofectamine 2000 reagent (12 l), were used for transfection according to the vendor's instructions (Invitrogen Life Technologies). Transfected cells were cultured in DMEM supplemented with 10% FCS and G418 (1 mg/ml) for the selection of stable transfectants. The G418-resistant cells were sorted on a FACSVantage system (BD Biosciences) for the equal expression of all constructs in both the P388D1 and RBL-2H3 cells. At least two independent cell lines derived from independent transfections were prepared to assess receptor function for each of the individual constructs.

Measurement of change in intracellular Ca 2ϩ concentration ([Ca 2ϩ ] i )
Changes in [Ca 2ϩ ] i , induced by the cross-linking of Fc␣RI with anti-Fc␣RI mAb (clone A59) F(abЈ) 2 were determined with Indo-1-acetoxymethylester (AM)-loaded (Molecular Probes), stably transfected cells expressing Fc␣RI and an SLM 8000 spectrofluorometer. Briefly, cells were washed once on the plate, lifted gently from plate, and washed again with 10 ml of Ca 2ϩ -free and Mg 2ϩ -free Hank's buffer. Washed cells were re-suspended in Iscove's medium plus 5% FBS at a concentration of 10 7 cells/ml and incubated with 5 M Indo-1-AM at 37°C for 40 min. Cell preparations to be opsonized with mAb A59 F(abЈ) 2 were washed, resuspended in Ca 2ϩand Mg 2ϩ -supplemented Hank's buffer at concentration of 10 7 cells/ml, and incubated with saturating concentrations of A59 F(abЈ) 2 (5 g/ml) at 4°C for 30 min. All cells were washed and re-suspended at 5 ϫ 10 6 cells/ml in 1 mM Ca 2ϩ , 1 mM Mg 2ϩ , 1 mg/ml BSA, 10 mM HEPES Hank's buffer, and then 0.4 ml of cells (2 ϫ 10 6 cells) were immediately transferred to a continuously stirring 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 baseline for 60 s, goat F(abЈ) 2 anti-mouse IgG was added at final concentrations of 20 g/ml, 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 as described elsewhere (37). Degranulation assays RBL-2H3 cells (10 5 cells/well) expressing either CD89 (Fc␣RI) or CD64EC-TM/CD89CYT chimeric receptors were cultured overnight in 24-well culture plates (Corning). The culture medium were removed from plates and the cells were incubated with DMEM containing anti-human Fc␣RI mAb F(abЈ) 2 (clone A59; final concentration of 5 g/ml) or antihuman CD64 mAb F(ab-) 2 (clone 32.2; final concentration of 2.5 g/ml) for 45 min at 4°C. The cells were washed with Tyrode's buffer (130 mM NaCl, 5 mM KCl, 1.4 mM CaCl 2 , 1 mM MgCl 2 , 5.6 mM glucose, 10 mM HEPES, and 0.1% BSA (pH 7.4)) and then stimulated with goat anti-mouse F(abЈ) 2 (Jackson ImmunoResearch Laboratories) at a final concentration of 20 g/ml in the Tyrode's buffer. The supernatants were collected at 0, 15, 30, 45, and 60 min for the measurement of ␤-hexosaminidase activity. Cells from control wells were lysed with the same volume of 0.1% Triton X-100 in Tyrode's buffer for evaluation of the total ␤-hexosaminidase activity for each RBL-2H3 stable cell line. Supernatants and cell lysates were incubated with substrate (1.3 mg/ml p-nitrophenyl-N-acetyl ␤-D-glucosamine) (Sigma-Aldrich) in 0.1 M sodium citrate (pH 4.5) for 1 h at 37°C. The reaction was stopped by 0.2 M sodium carbonate buffer (pH 10.0) and the enzyme reactivity was evaluated by measuring OD at 405 nm. The percentage of specific ␤-hexosaminidase activity released was calculated as follows: percentage release ϭ 100 ϫ average supernatant activity from 4 wells/average cell lysate activity from four control wells.

Cytokine analysis
Cells were stimulated in 24-well tissue culture plates (Corning) with surface-bound mAb A59 F(abЈ) 2 . Wells were coated with either anti-Fc␣RI FIGURE 2. The Ser 248 /Gly 248 polymorphism alters calcium mobilization and degranulation in RBL cells. A, There was no significant difference in expression of the Ser 248 -Fc␣RI and Gly 248 -Fc␣RI alleles in stably transfected RBL-2H3 cells (mean fluorescent intensity, 86.5 ϩ 21.6 for the Ser 248 -Fc␣RI allele and 91.2 ϩ 11 for the Gly 248 -Fc␣RI allele, p ϭ not significant; n ϭ 4). No detectable degranulation and intracellular calcium influx were induced by F(abЈ) 2 of mIgG, A59, and 32.2 plus goat anti-mouse F(abЈ) 2 in the parental RBL-2H3 cells. B, Anti-CD89 mAb A59 F(abЈ) 2 induces a rise in intracellular Ca 2ϩ levels in RBL cells expressing the Gly 248 -Fc␣RI and Ser 248 -Fc␣RI alleles. RBL cells were loaded with indo-1-AM, opsonized with mAb A59 F(abЈ) 2 (5 g/ml), and stimulated at 60 s with F(abЈ) 2 goat anti-mouse (20 g/ml). In all experiments, the rise in [Ca 2ϩ ] i induced by Gly 248 -Fc␣RI was greater than that by Ser 248 -Fc␣RI (mean peak of [Ca 2ϩ ] i (nM) above baseline: 280.4 Ϯ 51.68 and 114.0 Ϯ 20.8 for the Gly 248 -Fc␣RI and Ser 248 -Fc␣RI alleles, respectively; p Ͻ 0.01 from t test; n ϭ 5). No calcium influx was seen when RBL cells expressing either the Gly 248 -Fc␣RI and Ser 248 -Fc␣RI alleles were stimulated with mIgG F(abЈ) 2 plus goat anti-mouse F(abЈ) 2 . C, Anti-CD89 mAb A59 F(abЈ) 2 induced degranulation in RBL cells expressing Gly 248 -Fc␣RI and Ser 248 -Fc␣RI alleles. RBL cells were prelabeled with mAb A59 F(abЈ) 2 (5 g/ml) in the tissue culture wells and stimulated with F(abЈ) 2 goat anti-mouse (20 g/ml). In all experiments the degranulation measured as the release of ␤-hexosaminidase activity was significantly higher in the RBL cells with Gly 248 -Fc␣RI than with Ser 248 -Fc␣RI at all time points ‫,ء(‬ p Ͻ 0.01 from t test; n ϭ 4). No degranulation was induced by the irrelevant mIgG F(abЈ) 2 plus goat anti-mouse F(abЈ) 2 in the stable RBL cells expressing either Gly 248 -Fc␣RI and Ser 248 -Fc␣RI alleles. D, Anti-CD64 mAb 32.2 F(abЈ) 2 induced degranulation in RBL cells expressing chimeric receptors of CD64EC-TM/CD89CYT with the Gly 248 allele or Ser 248 alleles. RBL cells were opsonized with mAb 32.2 F(abЈ) 2 (2.5 g/ml) in the tissue culture wells and stimulated with F(abЈ) 2 goat anti-mouse (2.5 g/ml). In all experiments degranulation measured by the release of ␤-hexosaminidase activity through CD64 cross-linking was significantly higher in the RBL cells with the Gly 248 allele than the Ser 248 allele at all time points ‫,ء(‬ p Ͻ 0.01 from t test; n ϭ 4). No degranulation was induced by the irrelevant mIgG F(abЈ) 2 plus goat anti-mouse F(abЈ) 2 in the stable RBL cells expressing either Gly 248 allele and Ser 248 allele chimeric receptors.  Fig. 1). n, The number of donors sequenced. mAb A59 F(abЈ) 2 (20 g/ml) or control mIgG F(abЈ) 2 (20 g/ml) overnight at room temperature. The Fc␣RI-expressing cells were added to coated wells and the culture media were collected after 24 h. The levels of murine IL-6 or TNF-␣ in culture medium were quantified by ELISA (BD Pharmingen).

Immunoprecipitation of Fc␣RI and detection of ␥-chain
Fc␣RI was immunoprecipitated from the transfected cell lines using either mAb MIP8a (Serotec) or mAb A77 prebound to protein G-agarose. The ␥-Chain was detected with polyclonal rabbit anti-␥-chain Abs (Upstate Biotech). Cells (2 ϫ 10 7 cells/ml) were lysed in PBS containing 1% Nonidet P-40 with a protease inhibitor mixture and used for immunoprecipitation. Immunoprecipitates were analyzed with SDS-PAGE and Western blotting.

Column immunoprecipitation of Fc␣RI and detection of Lyn
Affinity columns were generated through the coupling of the A77 mAb to the cyanogen bromide-activated Sepharose 4B beads. P388D1 cells (10 8

Immunoprecipitation of Lyn and detection of Fc␣RI
A rabbit anti-Lyn Ab was used to immunoprecipitate Lyn from P388D1 cells expressing different forms of Fc␣RI. Coimmunoprecipitated Fc␣RI was detected by Western blotting with mAb MIP8a (Serotec). Cells (2 ϫ 10 7 cells/ml) were lysed in PBS containing 1% Nonidet P-40 with protease inhibitor mixture and used for immunoprecipitation and Western blotting.
residual RBCs were lysed with Puregene red blood lysis solution (Gentra Systems). The resulting neutrophils were washed twice with HBSS. The cells were then counted and resuspended in RPMI 1640 medium supplemented with 10% FBS and 10 mM HEPES (pH 7.2). The neutrophils were stimulated in 12-well tissue culture plates (Corning) previously coated overnight with anti-Fc␣RI mAb A59 F(abЈ) 2 (20 g/ml), mIgG1 F(abЈ) 2 (20 g/ml), or human IgA (20 g/ml). The wells were washed twice with HBSS before use and 2 ϫ 10 6 neutrophils in 1 ml of culture medium were used for each well. The cells were incubated at 37°C in 5% CO 2 for 24 h and the culture supernatants were collected for cytokine determination. The levels of human IL-6 in culture medium were quantified by ELISA (BD Pharmingen).

Flow cytometry
Aliquots of cells of 5 ϫ 10 5 cells in 0.1 ml PBS were incubated with saturating concentrations of primary mAb for 30 min at 4°C followed by two washes. For indirect immunofluorescence, the cells were then incubated with saturating concentrations of FITC-conjugated goat anti-mouse IgG F(abЈ) 2 at 4°C for another 30 min. After washing, the cells were analyzed immediately for immunofluorescence intensity using a FACScan (BD Biosciences).

Data analysis
Student's t test, unpaired or paired according to experimental design, was used to analyze the data for degranulation, calcium influx, and cytokine production. The paired t test was used to compare means between conditions in matched pair design. The unpaired t test was used to compare means between samples of donors homozygous for the different alleles. The Mantel-Haenszel 2 test was used to test for differences in the genotype and allele frequencies in affected donors and controls both in African Americans and in the combined genetic data from two ethnic groups (Caucasian and African American). The association between genotype and the human lupus phenotype in Caucasians was tested with Fisher's exact test because of the low frequency number in one cell.

SNP detection in Fc␣RI coding region
Two sets of overlapping RT-PCRs were used to amplify the full length of the Fc␣RI coding region and the purified PCR products were directly sequenced. As shown in Fig. 1A, we identified five SNPs located at nucleotide positions 363 (rs1865096), 376 (rs11666735), 679, 836, and 844 in the coding region of Fc␣RI. The SNPs 363G3 A and 679T3 C are silent polymorphisms that change an arginine codon from AGA to AGG and a leucine codon from TTG to CTG, respectively (Table I). Nonsynonymous SNPs at 836 and at 376, which change a proline codon (CCG) to a leucine (CTG) and an aspartate (GAC) to asparagine (AAC), respectively, both have minor allele frequencies Ͻ0.10. More importantly for the purposes of this study, we confirmed that the SNP 844A3 G transition that changes amino acid codon 248 from serine (AGC) to glycine (GGC) is a more common polymorphism located in the CYT domain of CD89 ( Fig. 1) (39).

The Gly 248 and Ser 248 alleles of Fc␣RI differ in their capacity for receptor-mediated calcium mobilization
Fc␣RI can associate with the FcR ␥-chain that contains an ITAM (41). Upon Ab-mediated cross-linking of Fc␣RI or binding of IgA immune complexes to Fc␣RI, the tyrosines of the ␥-chain ITAM are phosphorylated by Src family kinases and the tyrosine-based signaling cascade is initiated (42). Because the CYT domain of Fc␥RIII (CD16) influences receptor-mediated intracellular calcium mobilization (43), we hypothesized that the two Fc␣RI CYT domain alleles might also affect the receptor signaling. Using stably transfected RBL cells sorted for the equivalent expression of each of the two Fc␣RI alleles on the cell surface to ensure equivalent binding of the stimulant ( Fig. 2A), we tested comparative receptor-initiated calcium signaling. RBL cells expressing Fc␣RI were opsonized with anti-Fc␣RI mAb A59 F(abЈ) 2 . When the goat anti-mouse IgG (HϩL) F(abЈ) 2 was added to cross-link the Fc␣RI, cells expressing the Gly 248 -Fc␣RI allele were more efficient in mediating intracellular calcium release than Ser 248 -Fc␣RI allele (Fig. 2B).

The Gly 248 and Ser 248 alleles of Fc␣RI differ in their capacity for receptor-mediated degranulation
Fc␣RI is expressed on myeloid cells including neutrophils, eosinophils, monocytes, dendritic cells, and tissue macrophages. Fc␣RI can initiate potent immune responses such as degranulation, endocytosis, phagocytosis, Ab-dependent cellular cytotoxicity, Ag presentation, and the release of inflammatory mediators. Because of the importance of receptor-mediated degranulation, we investigated whether the two alleles of Fc␣RI mediate comparable degranulation in RBL cells. Despite comparable expression of Ser 248 -Fc␣RI and Gly 248 -Fc␣RI, cross-linking of the Fc␣RI on the RBL cell induced significantly more degranulation with the Gly 248 -Fc␣RI allele than with the Ser 248 allele (Fig. 2C)  CD64EC-TM/CD89CYT chimeric receptors on the cell surface (data not shown). Remarkably, similar to native Fc␣RI receptors, cross-linking of the chimeric receptors induced significantly more degranulation in RBL cells expressing the CD64EC-TM/ CD89CYT receptor carrying a Gly 248 -Fc␣RI allele than in cells expressing CD64EC-TM/CD89CYT carrying a Ser 248 -Fc␣RI allele (Fig. 2D). These data demonstrate that the Fc␣RI (CD89) ␣-chain CYT domain influences the receptor signaling capacity. Furthermore, the two naturally occurring alleles of the Fc␣RI ␣-chain CYT domain significantly alter the receptor's degranulation function.

The Gly 248 and Ser 248 alleles of Fc␣RI alter receptor-mediated cytokine release
Receptor-mediated cytokine release plays an important role in modulation of the immune system. To compare the effects of the Fc␣RI alleles on cytokine release, we generated stable P388D1 cell lines with comparable expression of the different Fc␣RI alleles (Ser 248 -Fc␣RI or Gly 248 -Fc␣RI) with or without mutations within the TM domain to alter ␥-chain association (Arg 209 or Leu 209 ; Fig.  3A). In contrast to transfectants with the wild-type Arg 209 TM sequence of Fc␣RI, the TM mutants carrying Leu 209 lost their association with the endogenous FcR ␥-chain (Fig. 3B). Crosslinking of the Fc␣RI on cells expressing the Gly 248 -Fc␣RI allele induced significantly more IL-6 and TNF-␣ release than that induced with cells expressing the Ser 248 -Fc␣RI allele (Fig. 3C, upper panel for IL-6 and lower panel for TNF-␣). Because Fc␣RI not paired with an FcR ␥-chain represents the major fraction of Fc␣RI molecules expressed on the cell surfaces of human monocytes and neutrophils (26), we also determined whether the "␥-chainless" Fc␣RI might be able to mediate cytokine production and whether there might be a difference in cytokine release between the Ser 248 -Fc␣RI and Gly 248 -Fc␣RI alleles in the absence of the FcR ␥-chain. As shown in Fig. 3, A and B, cells expressing the Ser 248 -Fc␣RI and Gly 248 -Fc␣RI alleles with the TM Leu 209 mutation showed comparable receptor expression. In the absence of FcR ␥-chain pairing, stimulation of the Ser 248 -Fc␣RI allele receptor was unable to induce IL-6 and TNF-␣ release (two left bars in Fig. 3D, upper panel for IL-6 and lower panel for TNF-␣, respectively). Surprisingly, however, stimulation of the Gly 248 -Fc␣RI allele induces significant release of IL-6 and TNF-␣ in the absence of the FcR ␥-chain pairing (two right bars in Fig. 3D, upper panel for IL-6 and lower panel for TNF-␣, respectively). These data demonstrate that the CD89 ␣-chain, and specifically the Gly 248 -Fc␣RI allele of the receptor, maintains signaling capacity in the absence of the ␥-chain.
The data further demonstrate that the alleles of the Fc␣RI ␣-chain CYT domain do not require the presence of the FcR ␥-chain to significantly influence receptor functions and raise the possibility of another molecular partner mediating this function.

Lyn association with Fc␣RI is allele sensitive
The cross-linking of Fc␣R activates Src family kinase activity and induces tyrosine phosphorylation of several cellular proteins, including p72 Syk , a major target of early protein tyrosine kinase activity. The Src family member Lyn coprecipitates with aggregated Fc␣RI complexes and is one of the protein tyrosine kinases involved in receptor-mediated signaling (44). These observations prompted us to investigate whether Lyn might associate with Fc␣R independent of the ␥-chain and provide an explanation for the activity of Gly 248 -Fc␣RI. We used three independent assays to assess the association of Fc␣RI with Lyn. First, we generated a GST-CD89 CYT tail fusion protein to confirm that Lyn is associated with the CYT tail of Fc␣RI. Fig. 4A demonstrates that the Fc␣RI CYT domain, fused with GST, is able to bind Lyn in P388D1 cell lysates. GST-CD89CYT was also able to pull down Lyn from 293 cells and U937 cells (data not shown), demonstrating that Lyn associates with Fc␣RI through the Fc␣RI CYT tail. Second, we used a column-based immunoprecipitation assay to establish the relationship between Lyn and Fc␣RI. Lysates from cell lines expressing four different constructs of Fc␣RI at similar levels ( Fig. 3A) were precleared on irrelevant mIgG1 isotype control columns. The precleared cell lysates were then passed over  columns packed with agarose beads conjugated with the anti-Fc␣RI mAb A77. The proteins eluted from these affinity columns were probed with rabbit anti-Lyn Abs. The bottom panel of Fig. 4B shows that the levels of CD89 (Fc␣RI) eluted from the columns were similar among four different cell lines. However, more Lyn was isolated from lysates with the Gly 248 -Fc␣RI allele than with the Ser 248 -Fc␣RI allele in the presence of ␥-chain association (upper panel of Fig. 4B, lanes 1 and 2). This difference was also pronounced in the TM mutants carrying Leu 209 that are unable to pair with the FcR ␥-chain (upper panel of Fig. 4B, lanes 3 and 4).
No CD89 or Lyn could be detected by Western blotting in eluates from the mIgG1 control columns (data not shown). These data establish the specificity of interaction between Lyn and Fc␣RI and show the quantitative differences in the association between Lyn and two alleles of Fc␣RI. Finally, we immunoprecipitated Lyn using rabbit anti-Lyn Abs and performed anti-Fc␣RI immunoblot analysis to examine the amount of receptor associated with Lyn. Although a comparable amount of Lyn was immunoprecipitated from each of the different cell lines expressing Fc␣RI constructs (Fig. 4C, bottom panel), Lyn was able to pull down significantly more Gly 248 -Fc␣RI protein (Fig. 4C, lane 2) than Ser 248 -Fc␣RI protein (lane 1) in the presence of ␥-chain association. To further demonstrate the dependence of the association between Fc␣RI and Lyn on the Fc␣RI ␣-chain, cell lines expressing similar levels of Leu 209 TM mutant Fc␣RI were used for coimmunoprecipitation. Although the Leu 209 TM mutant receptor of both Fc␣RI alleles did not associate with endogenous ␥-chains (Fig. 3B, upper panel; Fig. 4B, middle panel), the Gly 248 -Fc␣RI readily coimmunoprecipitated with Lyn (Fig. 4C, upper panel, lane 4). In contrast, very little if any of the Ser 248 -Fc␣RI allele receptor (Fig. 4C, upper panel, lane 3) was evident.

The Ser 248 and Gly 248 alleles of Fc␣RI affect receptor mediated IL-6 cytokine release by human neutrophils
Because human neutrophils are the primary cell population constitutively expressing high levels of Fc␣RI, we compared the effects of the Fc␣RI alleles on IL-6 release by neutrophils from the genotyped homozygous donors. Neutrophils from donors of both genotypes expressed comparable levels of Fc␣RI (data not shown). Consistent with the results obtained from the transfected P388D1 cell lines, cross-linking of the Fc␣RI with A59 F(abЈ) 2 induced significantly more IL-6 release by neutrophils from Gly 248 -Fc␣RI homozygous donors compared with the Ser 248 -Fc␣RI homozygous donors ( Fig. 5A; p Ͻ 0.05). Human IgA-mediated cross-linking of the receptors also induced significantly more IL-6 release by neu-trophils from Gly 248 -Fc␣RI allele homozygous donors compared with the Ser 248 -Fc␣RI allele homozygous donors ( Fig. 5B; p Ͻ 0.05). Interestingly, the cross-linking of Fc␣RI in the Ser 248 -Fc␣RI homozygous donors significantly inhibited IL-6 release in human IgA-stimulated neutrophils ( p ϭ 0.015; Fig. 5B). Taken together, our data demonstrate that the alleles of Fc␣RI CYT domain significantly affect receptor mediated cytokine release by primary human neutrophils ex vivo.

The Gly 248 -Fc␣RI allele and SLE
Because Fc␣RI can activate a variety of cell programs in myeloid cells and IgA autoantibodies have received increasing attention (45,46), we speculated that the proinflammatory Fc␣RI alleles with their impact on receptor signaling might be associated with inflammatory diseases. We compared the genotype frequencies in African American and European American controls and found a significant enrichment of Gly 248 -Fc␣RI in African Americans compared with European Americans (Table II; p Ͻ 0.01 from Fisher's exact test). Furthermore, in African Americans the Gly 248 -Fc␣RI allele was significantly enriched in SLE patients compared with corresponding normal controls (Mantel-Haenszel 2 ; p ϭ 0.0334). A similar tendency toward an increased Gly 248 -Fc␣RI allele frequency was evident in European American SLE patients compared with controls and was most evident in a dominant genetic model with enrichment of homozygosity of the Gly 248 -Fc␣RI genotype ( p Ͻ 0.05 from Fisher's exact test). In a combined analysis to test the significance of enrichment of the Gly 248 -Fc␣RI allele in both populations, we found this allele significantly enriched in SLE populations compared with the control populations (Mantel-Haenszel 2 ; p ϭ 0.0194). These data suggest that the Gly 248 -Fc␣RI allele may be considered a candidate risk factor for autoimmune disease, such as SLE.

Discussion
The two naturally occurring alleles of human Fc␣RI, Ser 248 -Fc␣RI and Gly 248 -Fc␣RI, differ in their quantitative functions and, remarkably, unlike Ser 248 -Fc␣RI, the Gly 248 allele is capable of inducing cytokine release even in the absence of FcR ␥-chain association. This capacity is presumably due, at least in part, to its ability to bind Lyn. This association of a Lyn kinase with Fc␣RI may be mediated by an unconventional Src homology 3 domain in the CYT domain of Fc␣RI comprised of the peptide sequence Trp 247 (Ser 248 /Gly 248 )Gln 249 Gln 250 , which includes the polymorphic site. This sequence is similar to the unconventional Src homology 3 domain binding motif (WxxQF/Y) frequently found in peroxisomal membrane proteins (47)(48)(49). Thus, it seems that Gly 248 -Fc␣RI may have a constant potential for an activating and perhaps proinflammatory role, whereas the Ser 248 -Fc␣RI allele requires Fc␣RI:FcR ␥-chain pairing for the induction of cytokine release. Given that the net biological activity of Fc␣RI, therefore, is dependent on both the genotype of the donor and the extent of Fc␣RI:FcR ␥-chain pairing, it is not surprising that investigators have drawn divergent conclusions about the activating potential of Fc␣RI-initiated activity (33,34).
The difference in the biological potential of the two alleles provides a tool for probing the idea that the balance of Fc␣RI-initiated activating/inhibitory activities might reflect these naturally occurring alleles and be important in setting thresholds for inflammatory responses (36). Accordingly, we reasoned that Fc␣RI alleles might play a role in the pathophysiology of disease. Somewhat to our surprise, the Gly 248 -Fc␣RI allele is over represented in systemic lupus. Although the mean IgA1 and total IgA in samples from SLE patients are significantly higher than those from the normal population (50), IgA Abs have received little attention in SLE. IgA anti-Ro/SSA, anti-La/SSB, anti-cardiolipin, and anti-␤ 2 -glycoprotein-I autoantibodies are present in the serum of patients with SLE, Sjogren's syndrome, and anti-cardiolipin syndrome (45,51,52) and, very recently, reports indicate that not only EBV seropositivity but also a specific IgA Ab against EBV are associated with SLE patients (46,53). The reports of IgA autoantibodies and IgA anti-EBV seropositivity suggest a provocative and perhaps underappreciated link in which a more activating host response to IgA might influence an autoimmune and inflammatory response. The differential biologies of the Ser 248 -Fc␣RI and Gly 248 -Fc␣RI alleles could also play a role in other phenotypes involving IgA, including dermatitis herpetiformis, celiac disease, periodontal disease, and possibly IgA nephropathy. Even the uptake and inactivation of virus, demonstrated in epithelial cells with anti-HIV IgA, might be impacted by CD89 alleles (54). Because mice do not have FCAR, the gene encoding CD89, clues will not be forthcoming from studies of spontaneous mouse phenotypes but will need to be done in humans or, possibly, in mice reconstituted with human immune systems (55,56).
Despite abundant mucosal and serum IgA, the overall role of IgA and Fc␣RI in host defense and immune regulation remains a matter of some controversy. As noted, some studies have suggested an anti-inflammatory function for IgA and Fc␣RI, as demonstrated in the suppression of neutrophil chemotaxis, down-regulation of the generation of reactive oxygen species, and the induction of the anti-inflammatory cytokine IL-1ra (32,34). Because IgA in the serum does not fix complement by the classical pathway, IgA is thought to act as a "discrete housekeeper" in which foreign Ags are incorporated into immune complexes by IgA and removed by the phagocytic system, with little or no resultant inflammation (57). However, Fc␣RI is a multichain immune recognition receptor that can associate with the ITAM-containing FcR ␥-chain subunit. The resultant trimeric Fc␣RI/(FcR ␥-chain) 2 initiates established phosphotyrosine-based signaling cascades, and this model is compatible with an activating, proinflammatory receptor (30,41). The observation that the majority of Fc␣RI expressed on human cells are not associated with ␥-chain (26), however, makes Fc␣RI distinct from Fc␥RI and Fc␥RIII, and the apparent association of Gly 248 -Fc␣RI with Lyn provides an alternative mechanism for initiating cell programs in an allelesensitive fashion. The intensity of cytokine production stimulated through Gly 248 -Fc␣RI alone is less than that through the Fc␣RI: FcR ␥-chain, and these two signaling assemblies may have qualitative as well as quantitative differences in their capacities. Our data also suggest that IgA may inhibit IL6 production by neutrophils from donors carry the Ser 248 allele of Fc␣RI. This inhibitory role may be related to the inhibitory functions of the FcR ␥-chain recently described by Pasquier et al. (36) and by Olas et al. (58). Although the positively charged arginine in the TM domain of Fc␣RI is required for the pairing with the FcR ␥-chain, little is known about the mechanisms regulating the pairing of Fc␣RI:FcR ␥-chain. This step may be an important element in the understanding of overall Fc␣RI biology.
Our genotyping data suggest that the proinflammatory 844G allele (Gly 248 -Fc␣RI) may be associated with SLE in African American and Caucasian populations, raising the possibility that IgA Ab production may play an important role in modulating the immune responses. This association needs to be extended in independent case control and family-based association studies, and linkage disequilibrium with other susceptibility gene(s) in this region must also be evaluated. However, based on the reports that Fc␥Rs contribute to the susceptibility to human autoimmune diseases beyond SLE, genetic associations of Fc␣RI variants with other human inflammatory diseases, perhaps conditioned on other genetic effects and stratified by different ethnic groups, are likely to be identified.