Unusual biochemical features and follicular dendritic cell expression of human Fcα/μ receptor



The Fc receptor for IgA and IgM (Fcα/μR) is of particular interest because it can bind antibodies of both IgM and IgA isotypes and thus may play a pivotal role in systemic and mucosal immunity. Using IgM and IgA ligands and newly generated Fcα/μR specific monoclonal antibodies we have defined biochemical features and cellular distribution of the human Fcα/μR. Both recombinant and native forms of human Fcα/μR are expressed on the cell surface as remarkably stable homodimeric transmembrane glycoproteins that can bind specifically polymeric IgM or IgA. The only human B cells to express Fcα/μR, albeit at very low levels, are found in the pre-germinal center subpopulation defined by the IgD+/CD38+ phenotype. Hence the expression pattern differs from that of the mouse wherein Fcα/μR is expressed by both circulating and resident B cell populations. Significantly, the predominant cell type expressing the Fcα/μR in humans is the follicular dendritic cell of germinal centers. The Fcα/μR may thus function in antigen presentation and B cell selection in the germinal center response.


adenosine diphosphate


complement receptor


Fc receptor for IgA and IgM ; FDC: follicular dendritic cells ; huFcα/μR: human Fcα/μR ; MNC: mononuclear cells ; SA: streptavidin ; tTG: tissue transglutaminase


Fc receptors (FcR) for the different antibody isotypes are expressed by many cell types in the immune system. The interaction of antibodies with the FcR can initiate a broad spectrum of effector functions that are important in host defense 14. These functions include phagocytosis of antibody-coated microbes, lysosomal degradation of endocytosed immune complexes, antibody-dependent cell-mediated cytotoxicity, secretion of cytokines and chemokines, release of potent inflammatory mediators, regulation of antibody production by B cells, and enhancement of antigen presentation. These diverse regulatory roles depend upon the antibody isotype and cellular distribution of the corresponding FcR. Molecular characterization at both protein and genetic levels has been limited to FcR for IgG (FcγRI/CD64, FcγRII/CD32, FcγRIII/CD16, FcγRIV), IgE (FcϵRI), and IgA (FcαR/CD89) 15.

Although the existence of an IgM FcR (FcμR) on subpopulations of B, T, NK cells, macrophages and granulocytes was suggested more than two decades ago by many investigators including us 617, the gene encoding an FcμR defied identification before the discovery of a murine cDNA that encodes a protein able to bind the Fc portion of both IgA and IgM, hence its designation as Fcα/μR 18, 19. The human (hu) homolog has since been isolated from a lymph node-derived cDNA library 20.

The huFcα/μR is a type I transmembrane protein predicted to encode a mature core peptide with an Mr of ∼55 kDa, an isoelectric point of 9.4, and two potential sites for N-linked glycosylation. A single Ig-like domain occupying the N-terminal third part of the extracellular region is followed by an uncharacteristic stalk region, a transmembrane segment, and a cytoplasmic tail. The Ig-like domain contains a sequence motif (GxxVxIxCxYxxxSVNRHxRKYW) that is conserved in the poly-IgR of the human, bovine, rat, and mouse species, and is predicted to be the binding site for polymeric IgM and IgA 18, 21, 22. Cells transfected with the mouse Fcα/μR cDNA bind IgM with high affinity (∼3×109 M–1), both polymeric and monomeric IgA with intermediate affinity (∼3×108 M–1), but do not bind IgG 18.

Fcα/μR is a single copy gene in both mice and humans 23, and its expression has been reported for B cells, monocytes/macrophages, and unknown cell types in kidney and small intestine 18, 20. Immunofluorescence analysis using a receptor-specific mAb has confirmed its expression on circulating and splenic B cells and monocytes/macrophages in mice 18. While primary cultures of mesangial cells from human kidney were shown by RT-PCR to express Fcα/μR mRNA 24, the cellular distribution and functional properties of huFcα/μR have not been examined previously. Furthermore, the biochemical nature of functional cell surface Fcα/μR has never been characterized in either mice or humans. The present studies address the biochemical properties and cellular distribution of huFcα/μR.


Immunoglobulin binding specificity of human Fcα/μR

In order to confirm that huFcα/μR, like its mouse counterpart, binds IgM and IgA, a huFcα/μR cDNA construct was expressed in cell lines using both a retrovirus-mediated transduction strategy and a lipofectin-mediated transfection method. (In this manuscript, both transduced and transfected cells are referred to as transfectants for simplicity.) Three murine cell lines that are negative for Fcα/μR transcripts were chosen for this analysis, the BW5147 T lymphoma, the Ag8.653 plasmacytoma, and the BaF3 pro-B cell line.

As anticipated, huFcα/μR-transfected T cells bound IgM of both human and mouse origin in a dose-dependent fashion, whereas control T cells did not (Fig. 1A). This binding was not observed with monomeric IgM molecules. While this result was obtained using chemical reduction of pentameric IgM to produce IgM monomers, essentially the same results were obtained with pentameric and monomeric IgM anti-mouse RBC hybridoma antibodies (Fig. 1B). The huFcα/μR also bound IgA polymers, but not IgA monomers (Fig. 1A). No IgG binding irrespective of subclasses and aggregation was observed with the huFcα/μR-transfected cells. The same binding characteristics were demonstrable with huFcα/μR-transfected plasmacytoma and pro-B cell lines (not shown). These findings indicate that the Fcα/μR and the poly-IgR have a similar ligand binding specificity 21, 22.

Figure 1.

Immunofluorescence analysis of Ig binding by huFcα/μR. (A) BW5147 cells transfected with the vector containing huFcα/μR cDNA in incorrect (upper histograms) or correct (lower histograms) transcriptional orientations were incubated with PBS (open histograms) or the indicated Ig ligands (shaded) of either human (hu) or mouse (mo) at a protein concentration of 30 μg/mL. The bound Ig were revealed by addition of a mixture of biotin-labeled anti-κ and anti-λ mAb plus allophycocyanin-SA (columns 1–5) or PE-labeled goat anti-mouse Ig antibody (colums 6, 7) before flow cytometric analysis. The numbers indicate the mean fluorescence intensity of each Ig ligand. Note that while mouse Ig ligands appear to bind more strongly to the huFcα/μR-transfected cells than human Ig ligands, this is due to use of different developing reagents. (B) An equal mixture of untransduced BW5147 cells (GFP) and BW5147 cells transduced with the bicistronic pMXsIG construct containing huFcα/μR and GFP cDNA were incubated with culture supernatants containing the indicated concentrations of pentameric (bottom) and monomeric (top) IgM anti-mouse RBC mAb before developing with biotin-labeled anti-mouse κ mAb and allophycocyanin-SA.

Anti-Fcα/μR-specific mAb

To generate receptor-specific antibodies, mice were immunized with a recombinant protein corresponding to the Fcα/μR Ig-like domain produced in Escherichia coli. One mAb (AM9, γ2bκ isotype) was shown by immunofluorescence analysis to react with recombinant Fcα/μR on the cell surface of BW5147 T lymphoma, Ag8.653 plasmacytoma, and BaF3 pro-B transfectants (Fig. 2). The epitope recognized by this mAb appears to be located outside the ligand-binding site, since pre-incubation with the AM9 mAb did not inhibit subsequent binding of IgM and IgA ligands, and vice versa. Another mAb (AM26, μκ isotype) was also found to react with huFcα/μR-transfected, but not control, cells which were either viable or pre-fixed with paraformaldehyde, suggesting that it has an anti-Fcα/μR specificity. However, the μκ isotype of this mAb precluded its usage in most experiments.

Figure 2.

Specificity of the AM9 anti-Fcα/μR mAb. huFcα/μR cDNA was introduced into BW5147 T lymphoma (upper panel), Ag8.653 plasmacytoma (middle panel) and BaF3 pro-B (lower panel) cells by either transfection or transduction. The resultant stable transfectants (shaded histograms) as well as the control cells (open histograms) were incubated with biotin-labeled AM9 anti-Fcα/μR mAb, before developing with PE-SA. The stained cells were analyzed by flow cytometry. The staining patterns with a biotin-labeled isotype-matched control (δED5-11 anti-Id mAb) were identical between the transfectants and the controls (not shown) and were equivalent to those of control cells with the AM9 mAb. Note that the AM9 mAb reacts with the transfectants with huFcα/μR-expressing cells, but not with the control cells.

Biochemical characteristics of cell surface Fcα/μR

The receptor structure was initially assessed using the huFcα/μR-transfected BW5147 T cells, because no human hematopoietic cell lines expressing endogenous Fcα/μR had been identified. Cell surface proteins were iodinated and solubilized in 1% NP-40 before immunoprecipitation with the receptor-specific mAb and the Ig ligands. A major glycoprotein with an estimated Mr of 125 kDa was precipitated with the AM9 anti-Fcα/μR mAb (but not with the isotype-matched control mAb) from the membrane lysates of the Fcα/μR transfectants, but not the control cells (Fig. 3A left). The same results were also obtained with the AM26 mAb and polyclonal anti-Fcα/μR antisera (not shown). This 125-kDa glycoprotein was capable of binding an IgM ligand and, to much lesser extent, an IgA ligand, but not IgG ligands, thereby confirming its ligand binding specificity.

Figure 3.

SDS-PAGE analysis of cell surface Fcα/μR. Fcα/μR-transfected (A–D) and tonsillar MNC (E) were examined for cell surface proteins by iodination (A–C, E) and for total cellular proteins by Western blot analysis (D). (A) Plasma membrane proteins on BW5147 cells transfected with huFcα/μR cDNA or control vector DNA were iodinated and lysed prior to incubation in plates pre-coated with anti-Fcα/μR mAb (AM9) or with representative mouse myeloma Ig (IgG, IgA, IgM). The bound materials were resolved by SDS-7.5% PAGE under reducing conditions before autoradiography. In the right panel, the isolated AM9-reactive proteins were left undigested (–) or digested (+) with neuraminidase plus O-glycanase (Neur/O-Glyc) and/or N-glycanase (N-Glyc) prior to SDS-PAGE analysis. Essentially the same results were obtained with another anti-Fcα/μR mAb, AM26 (not shown). (B) Membrane lysates of surface-iodinated huFcα/μR cDNA-transfected cells were incubated with beads pre-coated with AM9 or control mAb. Water (–) or urea to a final concentration of 6 M (+) was added to the isolated materials before SDS-PAGE in gels containing 4 M urea. (C) The indicated stable transfectants expressing huFcα/μR or mouse Fcα/μR were subjected to surface iodination and immunoprecipitation as in (B). (D) Non-radiolabeled, huFcα/μR cDNA-transfected and control BW5147 cells were lysed and subjected to immunoprecipitation as in (A). The bound materials were resolved by SDS-PAGE, transferred onto membranes and blotted with biotin-labeled AM9 mAb before developing with HRP-SA and visualization by enhanced chemiluminescence. Note the additional band with an Mr of ∼65 kD. (E) Lysates of surface-iodinated tonsillar MNC were subjected to immunoprecipitation and analysis as in (B).

Despite the fact that the predicted mature core peptide of Fcα/μR has an Mr of ∼55 kDa, the size of this functional Fcα/μR molecule was the same under both reducing and non-reducing conditions, indicating that the 125-kDa Fcα/μR protein is resistant to boiling in SDS and to reduction by 2-ME. Removal of carbohydrates by glycosidase treatments resulted in reduction of the Mr by only ∼10 kDa, indicating that the unexpected increase in Fcα/μR molecular mass was not due to extensive glycosylation (Fig. 3A, right). The addition of urea during SDS-PAGE failed to dissociate the 125-kDa protein into smaller mobility species (Fig. 3B). A similar molecular mass estimate was obtained for the recombinant huFcα/μR expressed on the cell surface of other transfectants, Ag8.653 plasmacytoma (∼120 kDa) and BaF3 pro-B cells (∼135 kDa) (Fig. 3C). The observed variation in Mr likely represents cell-type specific glycosylation pattern.

Furthermore, the recombinant mouse Fcα/μR on the cell surface of BaF3 cells was found to have an even larger Mr of 150–210 kDa under reducing conditions (Fig. 3C), suggesting a much more extensive and complex glycosylation of the mouse receptor. This is consistent with the presence of four potential sites for N-linked glycosylation in the mouse Fcα/μR compared with two sites in the huFcα/μR 18. Notably, the protein blot analysis of huFcα/μR transfectants revealed the existence of both ∼125-kDa and ∼65-kDa proteins when the materials reactive with the AM9 and AM26 anti-Fcα/μR mAb were isolated from whole lysates of non-radiolabeled cells (Fig. 3D), suggesting that the expected 65-kDa protein exists intracellularly and is modified prior to cell surface display.

To confirm that the AM9-reactive cell surface glycoprotein of ∼125 kDa is indeed an Fcα/μR homodimer, the 125-kDa glycoprotein was highly purified from Fcα/μR-transfected BW5147 cells and was subjected to trypsin digestion and mass spectrometric analysis. The MASCOT search results revealed a protein with probability-based Mowse score of 91 (gi|18032042 Homo sapiens Fcα/μR) and with peptide coverage of 12%. This identification was confirmed by reverse-phase liquid chromatography-electrospray ionization tandem mass spectrometry analysis with peptide coverage of 21%. These findings thus identify the AM9-reactive glycoprotein of ∼125 kDa as the Fcα/μR and indicate that no other proteins are covalently associated with the Fcα/μR, consistent with its predicted homodimeric structure.

To determine whether the native Fcα/μR has a similar post-translational modification, cell surface proteins on viable tonsillar mononuclear cells (MNC) were iodinated and subjected to immunoprecipitation analysis. A major glycoprotein with an Mr of ∼115 kDa was specifically precipitated from the membrane lysates with AM9 anti-Fcα/μR, but not with the isotype-matched control mAb (Fig. 3E). A minor protein of ∼175 kDa was also precipitated with the AM9 mAb, raising the possibility that Fcα/μR may also be present as a trimer on the surface of tonsillar MNC. Again, the mobility of these molecules was the same under both reducing and non-reducing conditions. Collectively, these findings suggest that the functional cell surface Fcα/μR protein on tonsillar MNC and transfectants is a remarkably stable homodimeric glycoprotein with an Mr of ∼120 kDa.

Selective expression of Fcα/μR in secondary lymphoid tissues

The pattern of huFcα/μR expression in hematopoietic tissues was assessed by RT-PCR and flow cytometric analyses. Surprisingly, unlike the mouse receptor, huFcα/μR transcripts were not detectable in blood, spleen or bone marrow MNC. The huFcα/μR transcripts were identified in tonsillar cells (Fig. 4A) and their identity as huFcα/μR was confirmed by nucleotide sequence analysis. When tonsillar CD19+ B cells were separated into the follicular mantle naive (IgD+/CD38), pre-germinal center (IgD+/CD38+), germinal center (IgD/CD38+) and memory (IgD/CD38) subpopulations, Fcα/μR transcripts were identified only in the pre-germinal center B cell subpopulation, which comprises ∼5% of the tonsillar B cells (Fig. 4B). However, we noted that the Fcα/μR RT-PCR signals from tonsil samples were stronger when using RNA templates extracted directly from the minced tissue fragments versus those from the isolated MNC (see Fig. 4A). In contrast, the Fcα/μR RT-PCR signals from splenic samples were not detectable even using RNA templates extracted from the tissue fragments (not shown).

Figure 4.

RT-PCR analysis of Fcα/μR expression in hematopoietic tissues and tonsillar B cell subpopulations. (A) Three micrograms of total RNA extracted from bone marrow (BM), blood, spleen and tonsil MNC and from whole tonsil tissue were converted to first-strand cDNA, and the cDNA was amplified with Pfu DNA polymerase and a set of primers for huFcα/μR (top) and GAPDH (bottom). Amplified products were electrophoresed and stained with ethidium bromide. Note that the Fcα/μR RT-PCR signals from tonsils are stronger when using RNA extracted directly from the minced tissue fragments rather than the isolated MNC. (B) The tonsillar CD19+ B cells were separated into the follicular mantle/naive (IgD+/CD38), pre-germinal center (IgD+/CD38+), germinal center (IgD/CD38+), and memory (IgD/CD38) B cell subsets, and each population was isolated by a high-speed MoFlo cell sorter before RNA extraction. The frequencies (%) from three different experiments were 60.3 ± 10.1 for IgD+/CD38, 4.8 ± 1.2 for IgD+/CD38+, 11.2 ± 4.6 for IgD/CD38+ and 23.8 ± 14.6 for IgD/CD38 subsets. Three micrograms of total RNA of each sorted cell fraction were subjected to RT-PCR, and a representative RT-PCR result of the sorted cells is shown. Note that Fcα/μR transcripts were identified only in the pre-germinal center B cell population (IgD+/CD38+).

Immunofluorescence analysis using the AM9 anti-Fcα/μR mAb revealed that none of the B cells, T cells, monocytes/macrophages, or NK cells in human blood and spleen samples expressed Fcα/μR on their cell surface irrespective of age, ethnic origin or gender of the blood donors (Fig. 5A). Activation of blood MNC with phorbol ester or LPS, which might partially recapitulate events in the tonsil, failed to induce Fcα/μR expression by blood MNC (not shown). Blood granulocytes, erythrocytes, and marrow MNC were also found to be negative for Fcα/μR. In tonsillar samples, T cells were non-reactive but a small subpopulation of B cells was weakly reactive with the AM9 anti-Fcα/μR mAb, consistent with the transcript analysis (Fig. 5B, upper panels).

Figure 5.

Two-color immunofluorescence analysis of Fcα/μR expression by blood and tonsillar MNC. MNC isolated from adult blood (A) and juvenile tonsils (B) by density gradient centrifugation were stained with a combination of biotin-labeled AM9 anti-Fcα/μR mAb (allophycocyanin-SA as a developing reagent) and FITC-labeled lineage-specific mAb. Viable cells in the PI population were analyzed for immunofluorescence staining by flow cytometry. Quadrants were determined based upon the background staining with isotype-matched control mAb labeled with the corresponding fluorochromes. Numbers indicate the frequency (%) of positively stained cells in each quadrant. For tonsillar CD14+ cells (B, lower panels), CD14+ cells were gated and examined for immunofluorescence staining with anti-Fcα/μR (solid line) and control (dotted line) mAb. Note that a small subpopulation of the tonsillar CD19+ B cells express Fcα/μR at relatively low cell surface levels and that half of the tonsillar CD14+ cells express Fcα/μR at relatively high levels.

Interestingly, half of the CD14+ cells, which represent ∼0.5% of the tonsillar MNC, were found to express relatively high levels of cell surface Fcα/μR (Fig. 5B, lower panels). It is presently unclear whether these Fcα/μR+/CD14+ cells are macrophages or follicular dendritic cells (FDC) as described 25, 26. Collectively, these findings indicate the selective expression of huFcα/μR by only a small subset of pre-germinal center B cells in secondary lymphoid tissues. This expression pattern thus differs significantly from mouse Fcα/μR, which is expressed by both circulating and resident B cell populations 18.

Expression of Fcα/μR by follicular dendritic cells

In the RT-PCR transcript analysis described above the Fcα/μR signal was stronger with tonsillar tissue fragments than with MNC isolated from this tissue (see Fig. 4A). When frozen tissue sections of tonsils were examined by an immunohistochemical analysis, the anti-Fcα/μR mAb decorated the germinal centers in a staining pattern suggestive of an interstitial reticular network (Fig. 6C, E). This staining pattern was similar to that observed with the FDC-specific mAb HJ2 (Fig. 6F) 27. Two-color immunofluorescence analysis of PE-labeled AM9 mAb (Fig. 6G) along with Alexa 555-labeled 7DC mAb specific for the CD21 long isoform (Fig. 6H), which is also known to be specifically expressed by FDC 25, revealed co-expression, but not complete co-localization, of the two molecules on FDC (Fig. 6I).

Figure 6.

Immunohistochemical analysis of Fcα/μR expression in tonsils. Serial frozen sections of tonsils were fixed in acetone, rehydrated in PBS, and incubated with biotin-labeled isotype-matched control mAb (δED5 anti-Id) (A), IA6-2 anti-IgD (B), or AM9 anti-Fcα/μR (C). Sections were also incubated sequentially with unlabeled AM9 mAb (E) or HJ2 anti-FDC mAb (F) and with biotin-labeled 187.5 rat anti-mouse κ mAb. The biotin-labeled mAb were developed by incubation with HRP-SA before developing with diaminobenzidine tetrahydrochloride and counter staining with methyl green. The image in (D) depicts the hematoxylin-eosin staining. For two-color immunofluorescence (G–I), sections were doubly stained with PE-labeled AM9 anti-Fcα/μR (G) and Alexa Fluor 555-labeled 7D6 anti-CD21L (H) mAbs. The merged fluorescence image is shown (I). Note the unique Fcα/μR staining with the interstitial dendrite-like architecture which is quite similar to the staining with the FDC-specific mAb.

The follicular mantle zone B cells, which were stained with anti-IgD mAb (Fig. 6B), and large germinal center lymphocytes were negative for Fcα/μR, as was anticipated from the lack of Fcα/μR expression by germinal center B cells. Fcα/μR+ cells were also not observed in the interfollicular regions or in the stratified squamous epithelium. Consistent with the RT-PCR results (Fig. 4A), Fcα/μR+ cells were not observed in most normal splenic samples, because no germinal centers were found in the white pulp of these tissues. These findings suggest that the predominant cell type expressing Fcα/μR in tonsils is the FDC.


The present study has focused on determination of the cellular distribution and biochemical nature of the huFcα/μR using receptor-specific PCR primers, mAb and the Ig ligands. Several intriguing features of the receptor have been discovered: (1) The cell surface Fcα/μR is a remarkably stable homodimeric glycoprotein with an estimated Mr of 115–135 kDa. (2) huFcα/μR is expressed by a small subpopulation of B cells that resides in tonsillar tissues and not in the spleen or in the circulation; hence there are significant species differences in Fcα/μR expression between humans and mice. (3) The germinal center FDC is the predominant cell type expressing Fcα/μR in tonsils.

The initial finding that the cell surface Fcα/μR on BW5147 T cell transfectants migrated as an ∼125-kDa glycoprotein on SDS-PAGE under both reducing and non-reducing conditions was unexpected, since the size of the huFcα/μR core protein is predicted to be ∼55 kDa. The remarkable increase in Fcα/μR Mr was not due to extensive glycosylation such as occurs with the FcαR/CD89 glycoprotein on activated eosinophils, where ∼60% of the ∼85-kDa Mr is contributed by carbohydrates 28. Despite many Ser/Thr residues and two potential sites for N-linked glycosylation in the predicted huFcα/μR protein, removal of both O- and N-linked carbohydrates from the 125-kDa glycoprotein resulted in reduction of its Mr by only ∼10 kDa. Mass spectrometric analysis of the 125-kDa glycoprotein has revealed that the functional cell surface Fcα/μR is an SDS-resistant and 2-ME-insensitive homodimer of the ∼55-kDa Fcα/μR chain.

These results differ from those reported previously for mouse Fcα/μR and huFcα/μR 18, 24. In both cases the flag-tagged Fcα/μR transiently expressed by COS cells was resolved as an ∼70-kDa species on SDS-PAGE under both reducing and non-reducing conditions as determined by protein blot analysis of whole-cell lysates using anti-flag antibodies. A simple explanation for this discrepancy is that the ∼70-kDa protein is the immature intracellular form rather than the functional Fcα/μR expressed on the plasma membrane, as our size estimation is based on the cell surface iodination. In agreement with this interpretation, our protein blot analysis of Fcα/μR transfectants identified both ∼125- and ∼65-kDa proteins reactive with the AM9 anti-Fcα/μR mAb in the unlabeled whole-cell lysates (see Fig. 3D).

There are several precedents for formation of SDS-resistant and 2-ME-insensitive oligomers by post-translational modifications. For example, a fatty-acylated Cys cluster at the interface between the transmembrane and spacer domains of synaptotagmin I, a potential major Ca2+ sensor in the central nervous system, is essential for stable SDS-resistant, 2-ME-insensitive self oligomerization 29. As five conserved Cys residues in the predicted mature Fcα/μR are not clustered but instead scattered in the N-terminal third part of the extracellular region, the acylation-mediated post-translational modification is an unlikely explanation for the observed features of the Fcα/μR.

A more likely modification may be the tissue transglutaminase (tTG)-mediated cross-linkage that occurs between the γ-carboxamide group of a Glu residue and the ϵ-amino group of a Lys residue to form an ϵ-(γ-glutamyl) lysine isopeptide bond 30. The cross-linked products with high Mr are highly resistant to mechanical challenges (e.g. boiling in SDS and 2-ME) and proteolytic degradation, and their accumulation is found in a number of tissues including skin and hair and in important biological processes (e.g. blood clotting and wound healing). Perhaps more relevant for the immune system is the example of CD38, a 45-kDa type II transmembrane protein with a bifunctional ectoenzyme activity, the synthesis of cyclic adenosine diphosphate (ADP)-ribose from NAD+ and the hydrolysis of cyclic ADP-ribose to ADP-ribose. The tTG-mediated post-translational modification of CD38 yields a higher Mr form (190 versus 45 kDa) and alters its catalytic properties (hydrolase to cyclase) 31. As six conserved Glu and Lys residues are present in the predicted extracellular region of human, mouse and orangutan Fcα/μR, the tTG-mediated post-translational modification is a likely mechanism for the 2-ME-resistant covalent cross-linkage of the cell surface Fcα/μR, a possibility currently being explored.

The designation of Fcα/μR was based on the ability of the receptor to bind the Fc portion of IgA and IgM 18. However, this receptor more closely resembles the poly-IgR than the classical FcR. The huFcα/μR gene is more closely linked to the poly-IgR gene on chromosome 1q32.1 than to the cluster of genes encoding the classical FcR (1q21–23) 23. The FcR and most of their relatives have a conserved feature of a signal peptide encoded by two separate exons, the second of which is either a 21-bp or 36-bp “mini-exon”; however, this feature is not found in the Fcα/μR gene (32, our unpublished observation). The exon/intron boundaries of the FcR and other Ig gene superfamily members are phase I (i.e. they occur after the first nucleotide of the triplet codon) with exception of the exons coding the cytoplasmic region, whereas those of the huFcα/μR gene are of all three phases.

The Ig-like domain of the Fcα/μR has high homology to the first N-terminal Ig-like domain of the poly-IgR 18, but the remaining extracellular, so-called stalk region, is quite distinct either from the poly-IgR or the FcR. Although the stalk region has no identifiable protein domain characteristics as determined by analysis employing available protein family databases, it is conceivable that this region may participate in the SDS-resistant 2-ME-insensitive interactions of the cell surface Fcα/μR. Finally, like poly-IgR, huFcα/μR binds IgM and IgA polymers, but not their monomeric forms.

The lack of expression of Fcα/μR by circulating human B cells and monocytes is unexpected, because mouse blood B cells and monocytes have been shown to constitutively express Fcα/μR on their cell surface 18. Given the fact that the mouse Fcα/μR binds IgM with high affinity (∼3×109 M–1) and IgA with intermediate affinity (∼3×108 M–1) 18, the receptor on mouse blood B cells and monocytes must always be occupied with the ligands in vivo, and the ligand-bound Fcα/μR should be internalized, rapidly retrieved from early endosomes and returned to the cell surface, otherwise the constitutive expression of Fcα/μR can not easily be explained. Such rapid receptor recycling has been observed with many other receptors including the FcαR/CD89 on blood phagocytic cells 33, 34.

If this scenario is indeed correct, then one of the functions of the Fcα/μR on circulating B cells and monocytes in mice may be to maintain the plasma concentrations of IgM and IgA at constant levels. This hypothesis is testable by disruption of the Fcα/μR gene in mice. The lack of Fcα/μR expression by human blood B cells is consistent with our previous findings that IgM binding was not observed with B cells and monocytes freshly isolated from normal individuals, but became evident when B cells were activated by stimulation with anti-μ, LPS or phorbol ester 9, 12. The receptor responsible for this IgM Fc binding, designated FcμR, is an ∼58-kDa glycoprotein that is linked to the membrane via a glycosylphosphatidyl inositol linkage 9, 12 and is likely encoded by an as yet unidentified gene. The previous finding that IgA binding was restricted to human blood monocytes and granulocytes 4, 35, 36 is also consistent with the lack of Fcα/μR expression by blood B cells.

The finding that the expression of Fcα/μR by tonsillar B cells is restricted to the CD38+/IgD+ pre-germinal center B cell subset is of interest as this population has been shown to be unique in terms of phenotype and somatic mutation 37, 38. These cells express germinal center markers (e.g. CD10, CD71 and CD77), the cell death-associated antigen CD95/Fas, and the proliferation-associated nuclear antigen Ki67. Interestingly, many of them bear very highly mutated IgVH region genes, but they do not mature into circulating memory B cells. B cells with such pre-germinal center phenotype are disproportionately expanded in blood from individuals with systemic lupus erythematosus or primary Sjögren's syndrome 39, 40.

It is presently unclear why Fcα/μR is selectively expressed, albeit at low levels, by this subset of B cells. One possible explanation is that FDC release immune complex-coated bodies (iccosomes) as the germinal center response is initiated, which can bind and be taken up by B cells in the germinal center. Iccosomes are expected to contain immune complexes and FDC membrane-bound molecules (e.g. FcγR and complement receptors) as well as mRNA. We have found that Fcα/μR is also expressed by Paneth cells in small intestinal crypts, the proximal tubular epithelial cells in kidneys, and the serous acini and small epithelial cells of salivary glands (manuscript in preparation). Thus it is likely that the expression of Fcα/μR gene is regulated by complex mechanisms with multiple tissue- or lineage-specific transcription factors.

An important clue to the in vivo function of the Fcα/μR may be that FDC are the predominant cell type to express the receptor in tonsils. Our immunohistochemical analysis reveals that the extensive Fcα/μR expression correlates with the appearance of germinal center FDC in secondary lymphoid tissues (manuscript in preparation). The origin of FDC has been controversial, as the cells appear not to originate from bone marrow progenitors, but may instead be derived from mesenchymal precursors 4143. FDC are dendroid non-phagocytic cells, and as their most notable function, they retain native antigens on their cell surface for long periods of time 43.

Complement receptor (CR)1/CD35 and CR2/CD21, and FcγRII/CD32 are clearly involved in antigen trapping. The C3 complement cleavage fragments, on covalent attachment to either antigen or immune complexes, can bind the CR1/CD35 and the CR2/CD21, which are highly expressed on FDC 41, 42. IgG immune complexes also bind the FcγRII/CD32 on FDC. It seems likely that Fcα/μR on FDC may play a similar role in trapping of IgM or IgA immune complexes and in presenting the intact antigens to B cells in germinal centers. FDC may thus be endowed with the ability to capture immune complexes of all three major Ig isotypes. However, Fcα/μR may have even more pronounced effects than FcγRII on the development of lymphoid follicles in secondary lymphoid tissues during early immune responses, when IgM is the predominant Ig isotype.

Materials and methods

Human Fcα/μR stable transfectants

The huFcα/μR cDNA was subcloned into the EcoRI sites of the pMX-Neo or pMXsIG retroviral vector which was kindly provided by Dr. Toshio Kitamura (Institute of Medical Sciences, University of Tokyo, Tokyo, Japan) and the resulting construct was transiently transfected into the BOSC23 cell line for packaging ecotropic retrovirus 18, 44, 45. The resultant culture supernatants were used immediately to transduce the murine BW5147 T lymphoma or BaF3 pro-B cell line. Two days after infection, the cells exhibiting IgM binding (for pMX-Neo) or GFP (for pMXsIG) were sorted by FACS.

In other experiments, the huFcα/μR cDNA was subcloned into the pCDNA3 vector and transfected into the mouse BW5147 T lymphoma and Ag8.653 plasmacytoma cell lines and stable transfectants were established as described 46. The nucleotide sequences of the insert DNA of all constructs in these experiments were extensively analyzed on both strands, and the expected sequences, reading frames, and orientations were confirmed.

Ig binding assay of human Fcα/μR

The huFcα/μR-transfected and control cells were sequentially incubated with various concentrations of human myeloma Ig, a mixture of biotin-labeled mouse mAb specific for a human κ or λ light chain and allophycocyanin-labeled streptavidin (SA) 35, 47. For production of IgM monomers, pentameric myeloma IgMκ was reduced with 0.2 M 2-ME and alkylated with 0.3 M iodoacetamide, followed by dialysis against PBS to generate non-covalent association of μ heavy and κ light chains. Alternatively, both pentameric and monomeric mouse IgM anti-mouse RBC hybridoma mAb 48 were used and their concentrations were determined by a sandwich ELISA using goat anti-mouse μ antibodies for capturing IgM and enzyme-labeled goat anti-mouse μ antibody for a developing reagent 49. The latter monomeric IgM was generated from the original IgM hybridoma clone by introducing mutations at position 414 and 575 (Cys to Ser).

Polymeric and monomeric forms of IgA1κ and IgA2λ myeloma proteins were kindly provided by Drs. Jiri Mestecky and Jan Novak (Dept. of Microbiology, University of Alabama at Birmingham). IgG subclasses were purchased from Sigma. In some experiments, mouse myeloma IgMκ, IgAκ and IgG1κ were also used as ligands along with PE-labeled goat anti-mouse Ig antibody (Southern Biotechnology Associates, Birmingham, AL). The purity and molecular configuration of these Ig samples were confirmed by both SDS-PAGE and non-denaturing PAGE. The stained cells were analyzed by flow cytometry as described 46, 50.

Preparation of recombinant human Fcα/μR protein and receptor-specific mAb

The Ig-like domain of the huFcα/μR was amplified by PCR using the Pfu DNA polymerase, the huFcα/μR cDNA as a template 20, and a set of primers (5′-GGATCCGCAGCTCCAAATTCATTGAA-3′ for the forward primer with a BamHI site and 5′-AAGCTTGCAGAGATGGTCAGATTC-3′ for the reverse primer with a HindIII site). The amplified PCR product was digested with BamHI and HindIII and ligated into the Pqe-30 prokaryotic expression vector as described 50. After verifying the insert DNA sequence, the Ig-like domain of Fcα/μR with six attached His residues was produced in E. coli strain, M15 [pREP4] cells, on induction with 2 mM isopropylthio-β-D-galactoside, and was purified from cell lysates by an affinity Ni column. After validation of the purity by SDS-PAGE, the isolated recombinant huFcα/μR protein was dialyzed against PBS and injected s.c. into BALB/c mice.

Hybridomas were generated by fusion of the regional lymph node cells with the Ag8.653 cell line, screened for their reactivity with the immunogen and other irrelevant recombinant proteins in ELISA, and with the huFcα/μR-transfected and control cells as described 50. Two Fcα/μR-specific mAb clones, AM9 (γ2bκ isotype) and AM26 (μκ isotype), were used in the present study. Sera were also collected from the immunized mice and used as polyclonal anti-Fcα/μR reagents.

Immunoprecipitation, SDS-PAGE and mass spectrometric analyses

Cell surface proteins on transfectants (107) and tonsillar MNC (∼3×108) were labeled with Na125I (1–3 mCi) and solubilized in 1% NP-40 lysis buffer containing protease inhibitors as described 12, 28. The cleared lysates were incubated with the 187.5 rat anti-mouse κ mAb-coupled beads that were pre-coated with either AM9 anti-Fcα/μR or δED5-11 control mAb. Alternatively, the cleared lysates were incubated in wells pre-coated sequentially with rat anti-mouse κ mAb, then with mouse myeloma IgMκ, IgAκ or IgG1κ. The bound materials were dissociated with 2% SDS (with or without 0.72 M 2-ME), boiled, and resolved by SDS-7.5% PAGE under both reducing and non-reducing conditions before autoradiography. In some experiments, urea was added to the dissociated materials to a final concentration of 6 M before separation by gels containing 4 M urea. For glycosidase treatments, the bound Fcα/μR proteins were digested with neuraminidase, O-glycanase (GLYKO, Novato, CA) and/or N-glycanase (New England BioLaboratories, Beverly, MA) according to the procedures recommended by the manufacturer.

For protein blot analysis, Fcα/μR and control transfectants were solubilized in the lysis buffer and the cleared lysates were transferred into wells pre-coated with the AM9 and AM26 anti-Fcα/μR or control mAb. The bound materials were dissociated, separated on SDS-PAGE, transferred onto membranes, and blotted with biotin-labeled AM9 mAb before developing with HRP-SA and visualization by enhanced chemiluminescence employing the oxidation of luminol in the presence of phenols (Amersham Life Science, Little Chalfont, UK) 46.

For mass spectrometric and micro-sequencing analyses, Fcα/μR-transfected cells (3×109) were incubated with the AM9 mAb prior to lysis and immunoprecipitation with protein-G-coupled beads. The AM9-reactive cell surface proteins were dissociated, resolved by SDS-PAGE, excised from gels, and subjected to tryptic digestion for mass spectrometric analysis using matrix-assisted laser desorption ionization time-of-flight mass spectrometry as described 51. The resulting de-isotoped peptide masses for a protein were analyzed via the MASCOT search engine (http://www.matrixscience.com) allowing for one tryptic cleavage and methionine oxidation. Protein records with Mowse score greater than 73 were considered significant. To confirm the putative polypeptide identification suggested by matrix-assisted laser desorption ionization time-of-flight mass spectrometry, tryptic digests were analyzed by reverse-phase liquid chromatography-electrospray ionization tandem mass spectrometry 51.

Flow cytometric analyses

MNC were isolated from long bone, tonsil and spleen tissues, which were obtained from the University of Alabama at Birmingham Tissue Procurement Service, and from peripheral blood of normal adult volunteers as described 28, 35. Granulocytes were isolated from the erythrocyte pellets of blood samples by differential sedimentation in 1.5% dextran 28. Approval for use of these human materials in this investigation was obtained from the University of Alabama at Birmingham Institutional Review Board. MNC and granulocytes were first incubated with aggregated human IgG to block FcγR and then with biotin-labeled AM9 anti-huFcα/μR mAb and FITC-labeled lineage-specific mAb against CD3, CD19, CD14, CD15, CD56, and CD10 (BD Biosciences, San Diego, CA). Isotype-matched controls included biotin-labeled δED5-11 anti-Id 47 and FITC-labeled irrelevant mAb (BD Biosciences). The binding of biotinylated mAb was revealed by addition of allophycocyanin-SA. Stained cells were analyzed by a FACSCalibur cytometer 50.

For tonsillar B cell subsets, MNC were stained with a combination of allophycocyanin-anti-CD38, biotinylated anti-CD19 (plus PE-SA) and FITC-anti-IgD mAb as described 32, 52. CD19+ B cells were separated into the follicular mantle/naive (IgD+/CD38), pre-germinal center (IgD+/CD38+), germinal center (IgD/CD38+) and memory (IgD/CD38) B cell subpopulations, and each population was isolated using a high-speed MoFlo cell sorter and subjected to RT-PCR analysis.

Immunohistochemical analysis

Tonsillar tissue fragments embedded in OCT compound were snap-frozen in liquid nitrogen, cut in 4-6 μm thickness with a cryostat, fixed in acetone, and rehydrated in PBS. Sections were incubated with biotin-labeled mAb specific for Fcα/μR (AM9), IgD (IA6–2; Pharmingen, San Diego) or an irrelevant myeloma Id (δED5-11) at a protein concentration of 5 μg/mL. After blocking endogenous peroxidase activity, bound antibodies were detected with HRP-SA (Vector Laboratories, Burlingame, CA) before developing with diaminobenzidine tetrahydrochloride (Sigma) and counter staining with methyl green 53. In some experiments, tissue sections were incubated with the FDC-specific mAb HJ2 27 and 7D6 25, which were kindly provided by Dr. Moon Nahm (Department of Pathology, University of Alabama at Birmingham) and Dr. Yong-Jun Liu (Department of Immunology, University of Texas MD Anderson Cancer Center, Houston, TX), respectively, before developing with biotin-labeled rat anti-mouse κ mAb (187.5) 54. For two-color immunofluorescence analysis, frozen tissue sections were doubly stained with PE-labeled AM9 anti-Fcα/μR mAb (1 μg/mL) and Alexa Fluor 555-labeled 7D6 FDC-specific mAb prior to examination by immunofluorescence microscopy.

RT-PCR analysis

Three micrograms of total RNA isolated from tissues and separated cells were converted to first-strand cDNA using an oligo(dT)18 primer and SuperScript II reverse transcriptase (Invitrogen) as described 46, 50. One twentieth of the cDNA was amplified with Pfu DNA polymerase and a set of primers (forward 5′-GGATCTTGTTTCTGGACTGAGA-3′ and reverse 5′-CTCAGGGTCCTGGATTTCTC-3′). These primers correspond to the 5′ untranslated region and the translation termination site of Fcα/μR, respectively, and yielded a 1757-bp fragment. Each amplification reaction included 35 cycles of denaturation at 95°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 5 min. A final extension was performed at 72°C for 7 min. Amplification of GAPDH transcripts with the primers 5′-GGTCGGAGTCAACGGATTTGG-3′ and 5′-CCTCCGACGCCTGCTTCACCA-3′) was performed as a control. Amplified products were analyzed by 0.7% agarose gel electrophoresis. In some experiments, the amplified PCR products of ∼1.8 kb were excised from gels and subcloned into the pCR-Blunt II-TOPO vector (Invitrogen) before nucleotide sequence analysis to confirm their identity as Fcα/μR.


The authors thank: J. Mestecky and J. Novak for IgA proteins; M. Nahm for HJ2 anti-FDC mAb; Y. J. Liu for 7D6 anti-CD21L mAb; G. L. Gartland and J. F. Kearney for FACS and immunofluorescence microscopic analyses; L. Wilson and M. Kirk for mass spectrometric and micro-sequencing analyses (PL: Stephen Barnes, CA13148); K. C. Sexton for providing human tissues (PL: William E. Grizzle, CA13148); Toshio Kitamura for vectors; Jacquelin B. Bennett for manuscript preparation; P. D. Burrows, G. V. W. Johnson, S. Izui, A. R. Lawton, M. D. Cooper, H. Kim and S. Barnes for helpful advice and criticism. K.K. is a scholar of the Seichi Sakai Foundation, Chiba, Japan. I.T. is an overseas Research Scholar of the Ministry of Education, Culture, Sports, Science and Technology in Japan. This work was supported in part by NIH/NIAID grants AI52243 and AI42127 (H.K.).


Conflict of interest: The authors declare no financial or commercial conflict of interest.


  1. 1


  2. 2


  3. 3


  4. 4


  5. 5


  6. 6