Article

• The EMBO Journal (1998) 17, 3850 - 3857
• doi:10.1093/emboj/17.14.3850

Gain-of-function mutations in FcRI of NOD mice: implications for the evolution of the Ig superfamily

Amanda L. Gavin¹, Peck Szee Tan¹ and P.Mark Hogarth¹

1. The Austin Research Institute, Austin and Repatriation Medical Centre, Studley Road, Heidelberg, Victoria 3084, Australia

Correspondence to:

P.Mark Hogarth, E-mail: pm.hogarth@ari.unimelb.edu.au

Received 16 March 1998; Accepted 15 May 1998; Revised 15 May 1998

• Keywords:

  • diabetes,
  • Fc receptor,
  • IgG,
  • Ig-superfamily,
  • macrophage

Introduction

Most of the leukocyte Fc receptor (FcR) family are comprised of immunoglobulin (Ig)-like domains, and therefore belong to the Ig superfamily (Ig-SF) (Williams and Barclay, 1988). Structurally related to each other, and to their ligand, Ig, the FcR family represent examples of highly evolved structures that are able to function distinctly. Within the FcR family, the two 'low-affinity' IgG receptor subfamily, FcRII and FcRIII, consist of two structurally conserved Ig-like domains (Hulett and Hogarth, 1994). These receptors exhibit relatively low affinity for IgG (i.e. K_d <10^-7 M) and display a wide specificity for ligand, binding IgG2b, IgG2a and IgG1 immune complexes (Hulett and Hogarth, 1994).

FcRI, the high-affinity receptor for IgG, however, is unique. Although the first two extracellular domains are structurally related to both FcRII and FcRIII, FcRI contains a unique third Ig-like domain (Allen and Seed, 1989; Sears et al., 1990). The most widely distributed allele of FcRI in mice (e.g. BALB/c), binds only a single IgG subclass, IgG2a, with high affinity (FcRI-BALB, K_a = 110^-8 M), but will bind complexes of mouse IgG3 (IC₅₀ 510^-7 M) (Gavin et al., 1998). The interaction between the first two domains and the third domain is critical in determining the unique high affinity and restricted IgG specificity observed. Functional analysis of mouse FcRI structure has revealed that when the third Ig-like domain is removed, the first two Ig-like domains are able to bind IgG with broad specificity, being able to interact with mouse IgG2a, IgG1 and IgG2b but only with a low affinity similar to the low-affinity Fc receptors; the binding of IgG3 was not tested (Hulett et al., 1991).

Another allele of mouse FcRI has been identified (Prins et al., 1993). The non-obese diabetic (NOD) mouse has an allele (FcRI-NOD) that differs from BALB/c strain (FcRI-BALB) by 24 nucleotides in the coding region, 17 of which encode different amino acids, notably including a four-amino acid insertion between domains 2 and 3, and a frame shift that leads to the premature truncation of the cytoplasmic tail (Prins et al., 1993). Ligand binding analysis of FcRI-NOD revealed that this receptor bound mouse IgG2a with a 10-fold increase in affinity (K_a 110^-9 M) over that of FcRI-BALB (K_a 110^-8 M), although less was expressed on the cell surface, and both the affinity and expression differences could be attributed to the amino acid differences in the extracellular domains (Gavin et al., 1996).

This study identifies that, in addition to increased affinity for mouse IgG2a, the FcRI-NOD molecule can also bind monomeric mouse IgG3, but most surprising is the absolute gain in function by binding IgG2b, which is unable to bind to other FcRI allelomorphs. We have identified the extracellular domain interfaces as regions crucial to FcRI-NOD function and expression. Moreover, the identification of these regions, together with an analysis of its gene, indicates that this IgG receptor system has implications for our understanding of function and evolution of one of the largest superfamilies, the Ig superfamily.

NOD/Lt macrophages bind monomeric mouse IgG2b and IgG3

The binding of different IgG subclasses to NOD/Lt and BALB/c bone marrow macrophages (BMM) was determined by immunofluorescence (Table I). In two experiments, macrophages derived from BALB/c mice bound monomeric IgG2a but not IgG2b, IgG3 or IgG1. In contrast, NOD/Lt BMM bound IgG2b, IgG3 and IgG2a, but not IgG1. This finding was unexpected, as mouse FcRI, the only IgG receptor that detectably binds monomeric IgG on macrophages, binds only IgG2a with high affinity. The observation of IgG2b and IgG3 binding by NOD/Lt macrophages suggests that either NOD/Lt macrophages express a novel Fc Receptor, or the FcRI allele present on NOD/Lt macrophages displays unusual specificity for IgG.

Table 1: The binding of monomeric IgG by bone marrow macrophages of either BALB/c or NOD/Lt origin

View full table

FcRI-NOD binds monomeric mouse IgG3 and IgG2b

To address whether the unique FcRI allele, FcRI-NOD, was contributing the unusual IgG2b and IgG3 binding pattern observed of the NOD/Lt macrophages, the binding specificity of the two FcRI alleles, FcRI-NOD and FcRI-BALB, was examined using transfected CHO-K1 cell lines expressing FcRI-BALB (1N3-2) and FcRI-NOD (2B5-13C) (Figure 1). Immunofluorescence analysis revealed that transfectants expressing FcRI-NOD bound monomeric IgG2b, IgG3 and IgG2a, but not IgG1 (Figure 1A, B, C and D, respectively). In contrast, FcRI-BALB-expressing cells bound only monomeric IgG2a (Figure 1G). Thus, the binding of monomeric mouse IgG2b and IgG3 by NOD/Lt macrophages is due to the unusual specificity of FcRI-NOD expressed by these cells.

Figure 1.

FcRI-NOD but not FcRI-BALB binds mouse IgG2b and IgG3. FACScan analysis of IgG binding by CHO cells transfected with either FcRI-NOD cDNA (2B5-13C) (A–D) or FcRI-BALB cDNA (1N3-2) (E–H) was assessed. The binding of the mouse IgG isotypes IgG2b (A and E), IgG3 (B and F), IgG2a (C and G) or IgG1 (D and H) to FcRI-transfected CHO cells was detected with FITC-conjugated sheep (Fab) anti-mouse Ig (solid line). Background fluorescence (dotted line) was determined by incubation with FITC-conjugated sheep (Fab) anti-mouse Ig only.

View full figure (87 KB)

Mouse IgG2b and IgG3 compete with IgG2a for the FcRI-NOD binding site

To determine whether the unique binding properties of FcRI-NOD were either due to the creation of a new binding site, or to alteration of the existing IgG2a binding site, competition experiments were performed (Figure 2). Both monomeric IgG2b and IgG3 inhibited the binding of radiolabelled IgG2a to FcRI-NOD at high concentrations, i.e. mIgG2b IC₅₀ 13.5 g/ml (or 8.610^-8 M) and mIgG3 IC₅₀ 5.8 g/ml (or 3.910^-8 M). The implication of this experiment is that the IgG2b and IgG3 binding sites are identical or in proximity to the binding site for IgG2a. As expected, non-radiolabelled IgG2a inhibited radiolabelled IgG2a binding and gave an IC₅₀ value that correlates with the known affinity of FcRI-NOD for IgG2a (i.e. IC₅₀ 0.22 g/ml or 1.410^-9 M) (Gavin et al., 1996). Mouse IgG1, however, was unable to inhibit the binding of IgG2a at concentrations up to 150 g/ml, confirming that this IgG isotype binds poorly, if at all, to FcRI-NOD. The surprising finding that FcRI-NOD binds monomeric IgG2b and IgG3 represent clearly defined gains of function and the functional differences between FcRI-NOD and FcRI-BALB alleles provide useful tools for defining regions involved in ligand binding.

Figure 2.

IgG2a, IgG2b and IgG3 compete for similar binding sites on FcRI-NOD. The inhibition of radiolabelled IgG2a (7-20.6/30) binding to FcRI-NOD on 2B5-13C cells by mouse IgG1 (UM1, ), IgG2a (UM2a, ) IgG2b (49.2, ) or IgG3 (UM3, ) was determined. The data are expressed as amount of ¹²⁵I-IgG2a bound (ng) against the concentration of competitor added (g/ml).

View full figure (64 KB)

Interdomain junctions influence allele-specific affinity and expression

Although the FcRI-NOD and FcRI-BALB alleles are 93% identical in the extracellular domains, they exhibit remarkable functional differences in Ig binding. The FcRI-NOD phenotype is one of high affinity for multiple isotypes, and low surface expression, while the FcRI-BALB phenotype is one of restricted high affinity for IgG2a, with surface expression that is relatively high. The major sequence differences between FcRI-NOD and FcRI-BALB occur in the interdomain junctions where the FcRI-NOD allele contains a change (Asn88Lys88) and an insertion (Glu89) between domains 1 and 2 (D1/D2) and a four-amino acid insertion (Ala173-Phe174-Pro175-Leu176) between domains 2 and 3 (D2/D3) (Prins et al., 1993; Gavin et al., 1996). Utilizing the extensive homology of the two alleles, and their dramatic functional differences, mutant receptors were generated by exchanging interdomain regions between FcRI-NOD and FcRI-BALB to determine to what extent these regions contributed to the unique phenotype of each allele (Figure 3). Mutants were constructed wherein the sequence of FcRI-NOD between D1/D2 (residues HKED termed N1) and D2/D3 (residues KAFPLE termed N2) were replaced with the corresponding regions of FcRI-BALB (residues HND termed B1) and (residues KE termed B2), respectively (Figure 3). Reciprocal mutants were also generated where the D1/D2 and D2/D3 regions of FcRI-BALB were replaced with the appropriate amino acids of FcRI-NOD (Figure 3). The wild-type alleles of FcRI-NOD and FcRI-BALB were presented as NOD.N1.N2 and BALB.B1.B2, respectively. All of the mutant FcR generated could be translated in vitro and when transiently expressed by COS cells, all bound erythrocytes sensitized with mouse IgG2a (EA-IgG2a), indicating that they were expressed on the cell surface (Figure 4A–H). These cells did not bind unsensitized erythrocytes (not shown) and mock-transfected COS cells did not bind EA-IgG2a (Figure 4I), indicating that the immune complex binding by the mutant FcR was specific. Following transfection of COS cells with the mutant cDNA, cells were assayed for the capacity to bind monomeric IgG2a, and IgG3 (Figure 5). A summary of the mutant receptors' affinity for IgG2a and IgG3, and also IgG2b, and the levels of functional surface expression determined by IgG2a binding (B_max values presented as sites/cell) are presented in Table II.

Figure 3.

Mutants at exon/intron boundaries map regions involved in Ig binding. Schematic representation of cDNAs used are shown. The regions of FcRI are indicated as D1, D2, D3, extracellular domains 1, 2 and 3, respectively. The origin of the appropriate template sequence is indicated by various shading (NOD/Lt, white; BALB/c, black). The sequence involved in the exchanges are shown in detail between the domain junctions, with the mouse strain of origin shown (see Materials and methods for details). The mutant name indicates origin of template, either NOD or BALB, and then the origin of the interdomain region 1 (D1/D2) or region 2 (D2/D3) is indicated by either N (NOD/Lt) or B (BALB/c). The wild-type alleles of FcRI-NOD and FcRI-BALB are presented as NOD.N1.N2 and BALB.B1.B2, respectively.

View full figure (92 KB)

Figure 4.

Mutated FcRI are expressed on the cell surface and are functional. Monolayers of COS cells transfected with either NOD.N1.N2 (A), NOD.B1.B2 (B), NOD.N1.B2 (C), NOD.B1.N2 (D), BALB.B1.B2 (E), BALB.N1.N2 (F), BALB.B1.N2 (G) and BALB.N1.B2 (H) cDNA (see Figure 3 for mutant FcR construction), or expression vector only (I) rosetted with IgG2a anti-TNP-sensitized sheep erythrocytes.

View full figure (73 KB)

Figure 5.

Titration of radiolabelled monomeric mouse IgG2a and IgG3 binding to mutant FcRI. Shown is the binding of mouse IgG2a () and mouse IgG3 () by COS cells expressing NOD.N1.N2 (A), NOD.B1.B2 (B), NOD.N1.B2 (C), NOD.B1.N2 (D), BALB.B1.B2 (E), BALB.N1.N2 (F), BALB.B1.N2 (G) and BALB.N1.B2 (H) (see Figure 3 for mutant FcR construction).

View full figure (81 KB)

Table 2: Exchange of amino acids at the interdomain junctions of FcRI modify both affinity for IgG and cellular expression

View full table

FcRI-NOD-transfected cells (termed NOD.N1.N2) bound IgG2a, IgG2b and IgG3 and expressed 1210³ sites on the surface (Figure 5A and Table II) while FcRI-BALB-transfected cells (termed BALB.B1.B2) bound IgG2a with approximately one-tenth the affinity of FcRI-NOD while being expressed at higher levels on the cell surface (510⁵ sites/cell) (Figure 5E and Table II).

The analysis of the mutant receptors revealed that modification of the interdomain junctions can have allele-specific effects on receptor function. Replacement of FcRI-NOD D1/D2 and D2/D3 regions (N1 and N2 respectively) with sequence of FcRI-BALB (B1 and B2) produces a receptor (NOD.B1.B2) with a 'BALB-like' phenotype; this mutant only detectably binds IgG2a (not IgG2b or IgG3) with affinity similar to that of FcRI-BALB (18.410^-9 M versus 20.910^-9 M) and has relatively high surface expression (15-fold increase over NOD) (Figure 5B and Table II).

Mutant receptors, where only single interdomain junctions were modified, were analysed to dissect the effects individual junctions have on receptor function. The D2/D3 junction has a major impact on FcRI-NOD expression and function. Cell surface expression of FcRI-NOD is modified by the D2/D3 region, as replacement of this junction (N2) with sequence of BALB (B2) produces a mutant FcRI (NOD.N1.B2) which is expressed at levels 30-fold higher than NOD (37010³ sites/cell versus 1210³ sites/cell). It is clear that this region also contributes to the unique affinity and specificity of NOD, as replacement of the N2 sequence in the NOD.N1.B2 receptor leads to a loss in receptor affinity for both IgG2a and IgG3 with the K_ds for this interaction increased 10-fold (Figure 5C and Table II). The affinity of NOD.N1.B2 for IgG2a closely matches that described for the BALB/c allele (16.510^-9 M versus 20.910^-9 M, respectively). It is interesting that the decrease in affinity caused by change of the D2/D3 region in the receptor is not observed for IgG2b (Table II). This would indicate that the binding interaction of FcRI-NOD with IgG2b is distinct from that of IgG2a and IgG3, although all compete with each other, indicating that they all bind in close proximity but the focus of the interaction is different. Interestingly, although the affinity of FcRI-NOD for IgG2b is not affected by the replacement of the D2/D3 region, IgG2b binding is markedly affected by the origin of the D1/D2 junction (compare NOD.N1.B2 with NOD.B1.B2).

The D1/D2 junction has a distinct role in the function of FcRI-NOD. Its replacement (NOD.B1.N2) results in complete loss of monomeric Ig binding by this mutant receptor (Figure 5D and Table II). The lack of monomeric Ig binding by NOD.B1.N2 implies that the D1/D2 region either directly interacts with IgG at the binding site, or this region contributes indirectly by interacting with other residues to influence binding.

Analysis of the role the interdomain junctions have on BALB allele function revealed similar effects. The inclusion of both D1/D2 and D2/D3 sequence of NOD into the BALB sequence produces a mutant receptor (BALB.N1.N2) able to bind IgG2a with a 7-fold increase in affinity, and this molecule exhibits very low surface expression (100-fold decrease when compared with BALB) (Figure 5F and Table II). This mutant confirms that the interdomain regions have the capacity to influence markedly ligand binding and surface expression. The BALB.N1.N2 mutant displays a partial NOD phenotype, i.e. the affinity for IgG2a is remarkably enhanced, while the cell surface expression is dramatically reduced (compare Figure 5F and A). In contrast to the NOD allele, however, the inclusion of the NOD interdomain regions in FcRI-BALB did not create the capacity to bind IgG2b and IgG3, implying that other regions distinct in NOD must be additionally involved in this unique interaction.

The roles that the individual junctions have on FcRI-BALB expression and function were also examined. Modification of the D2/D3 junction of FcRI-BALB however, is not well tolerated, as the insertion of the NOD-derived AFPL residues (N2) into FcRI-BALB to generate the BALB.B1.N2 mutant results in complete loss of monomeric Ig binding (Figure 5G and Table II). This result indicates that the presence of the AFPL sequence markedly reduces affinity for ligand when presented in the context of the BALB protein. In contrast to the dramatic effects of the D2/D3 junction on BALB function, the introduction of NOD-derived sequence to the interface between the D1/D2 of the FcRI-BALB (BALB.N1.B2) did not alter the affinity for IgG2a, IgG2b or IgG3, nor did it alter the levels of receptor expression (Figure 5H and Table II). By comparing BALB.N1.N2 with BALB.B1.N2, it is clear that the presence of NOD-derived D1/D2 sequence is required for high-affinity IgG2a binding. This observation implies that: (i) each interdomain region is dependent upon the presence of the other to maintain a functional high-affinity receptor; and (ii) each can contribute different aspects of function, e.g. N1 is essential for IgG2b binding (compare NOD.B1.B2 with NOD.N1.B2).

Intron/exon boundaries affect receptor function

Analysis of the intron/exon boundaries of the FcRI genes from NOD/Lt and BALB/c mice revealed that several nucleotide changes have occurred in the FcRI-NOD gene to alter exons and splice acceptor sites at the intron/exon boundaries (Figure 6). Within the D1/D2 junction, one nucleotide change (TG) and a three-nucleotide insertion at the 3' end of the exon encoding domain 1, leads to the amino acid change AsnLys88, and the insertion of Glu89 in the NOD/Lt allele. Three nucleotide changes were identified proximal to the D2/D3 junction, most notably an AG change creates a new splice acceptor site 12 nucleotides 5' of the BALB/c acceptor site, leading to the translation of four extra hydrophobic residues (Ala173-Phe174-Pro175-Leu176) in the NOD/Lt allele. Also, an AT change 3' of the donor site, and a TC change six nucleotides downstream from the newly created splice acceptor site were observed. The analysis of genomic sequence of FcRI has revealed that both the insertion and alteration of nucleotides at intron/exon junctions has occurred to encode the amino acids changed within the interdomain regions, and in light of the mutant receptor-binding data, these alterations lead to dramatic changes in function and expression of the surface IgG receptor.

Figure 6.

Alteration of intron/exon boundaries alters amino acid content of FcRI-NOD. Comparison of the sequence and encoded amino acids at the intron/exon junctions encoding the extracellular domains (D1,D2 and D3) of FcRI-NOD and FcRI-BALB are illustrated. Coding region sequences are shown in upper case and intron sequences in lower case. Splice donor and acceptor sites are underlined. Regions of sequence alteration or insertion in the FcRI-NOD gene are indicated by boxes.

View full figure (60 KB)

The novel finding that NOD/Lt bone marrow macrophages can bind monomeric IgG2b, IgG3, as well as IgG2a, led us to identify gains in function of the FcRI-NOD allele. In terms of the biological function of FcRI, the gains in affinity for IgG isotypes identified herein suggest that FcRI-NOD would be able to sample a broad range of IgG isotypes and participate in triggering effector functions at lower IgG concentrations than mice displaying the FcRI-BALB allele. Moreover, as IgG3 is a T-cell-independent antibody isotype (Perlmutter et al., 1978; Slack et al., 1980; Mongini et al., 1981), then FcRI-NOD would provide cellular effector function in the developing phases of adaptive immunity (Gavin et al., 1998). Although FcRI-NOD can associate with the FcR signalling molecule, the FcRI- subunit, the cytoplasmic tail of FcRI-NOD is truncated and the signalling potential of FcRI-NOD has not been determined (Gavin et al., 1996). The pathological significance of NOD/Lt macrophages binding monomeric IgG3, an isotype known to form pathogenic complexes (Abdelmoula et al., 1989; Lemoine et al., 1992), and IgG2b, an isotype present at high levels in NOD/Lt mice and known to be specific for autoantigens including GAD65 and insulin (Serreze et al., 1988; Hanson et al., 1996; Luan et al., 1996), remains unclear. Recent congenic mapping studies in mice, however, have demonstrated that although FcRI-NOD is genetically linked to a diabetes susceptibility locus, it is clearly distinct (Podolin et al., 1997).

The comparison of two highly related alleles of FcRI displaying different specificities for IgG led to the identification of regions that contribute to the function of mouse FcRI. Previous functional studies have demonstrated that in the absence of domain 3 (D3), the first two domains of mouse FcRI can bind IgG2a, IgG2b and IgG1, but with a dramatic loss in affinity (IgG3 was not tested) (Hulett et al., 1991). A similar decrease in affinity for IgG is observed in human FcRIb2, a two-Ig domain splice variant (Porges et al., 1992). Furthermore, exchange of D1 of mouse FcRI with D1 of mouse FcRII, does not alter the specificity of Ig binding, or have a major influence on the high-affinity binding of IgG2a (M.Hulett and P.M.Hogarth, in preparation). These studies provide evidence that D3 interacts with D2 to mask the low-affinity IgG2b and IgG1 binding site(s) and forms a high-affinity site for IgG2a, although the presence of D1 is required to maintain the structural conformation of the receptor.

The experiments presented herein demonstrate that the affinity and specificity of FcRI can also be modified by alteration of the domain junctions. Clearly, the D2/D3 junction is crucial for high-affinity binding as replacement of this region in both FcRI-NOD and FcRI-BALB leads to a decrease in affinity for IgG. As both NOD.N1.B2 and BALB.B1.N2 mutants can still bind IgG (BALB.B1.N2 binds IgG2a immune complexes only), then the D2/D3 junction itself is probably not the principal binding site. What becomes clear is that in the presence of BALB/c-derived D2/D3 sequence (B2), the affinity for IgG2a is 16–2010^-9 M, when present in either a NOD/Lt- or BALB/c-derived scaffold. However, the presence of NOD/Lt-derived D2/D3 sequence (N2) can, in different FcRI mutants, produce two contrasting effects. If N2 is co-expressed with N1, then very high affinity for IgG2a is achieved (1–310^-9 M, see NOD.N1.N2 and BALB.N1.N2). If N2 is mismatched with BALB/c-derived sequence (B1), then a dramatic loss of monomeric Ig binding results. These data demonstrate that D1/D2 and D2/D3 junctions must interact or cooperate to generate high-affinity IgG binding, in the NOD/Lt allele at least. Moreover, the presence of the AFPL (N2) sequence at the D2/D3 junction can only be functionally tolerated in the presence of the appropriate D1/D2 junction sequence (compare BALB.B1.N2 and BALB.N1.N2).

The role of the D1/D2 junction in affinity is difficult to evaluate in isolation. When B1 is expressed with B2 then moderate affinity for IgG2a is achieved. In contrast, and as mentioned above, the presence of B1 with N2 results in a dramatic loss of affinity, implying a functional mismatch. It is clear, however, that the NOD/Lt-derived D1/D2 sequence can influence the specificity of the FcRI-NOD allele. By comparing NOD.N1.B2 with NOD.B1.B2, it becomes clear that the N1 sequence must contribute to the broad specificity (IgG2b and IgG3 binding) exhibited by the FcRI-NOD allele. However, inclusion of the FcRI-NOD D1/D2 junction in the FcRI-BALB (BALB.N1.B2) does not dramatically alter the affinity for IgG2a, IgG2b or IgG3. This implies that N1 must interact with other residues novel in the FcRI-NOD allele to achieve the IgG2b and IgG3 binding pattern observed. It is interesting to note that one unique difference between FcRI-NOD and FcRI-BALB is an Asp135Gly136 alteration that corresponds to an Ig interactive region defined for other FcR, namely FcRII and FcRI (Warmerdam et al., 1991; Tate et al., 1992; Hulett et al., 1993, 1995).

Studies of the binding sites of other FcR family members, namely FcRII, FcRIII and FcRI, have identified loop and strand regions within the extracellular domains (principally domain 2) as being involved in direct Ig binding (Hulett et al., 1993, 1994, 1995; Mallamaci et al., 1993; Hibbs et al., 1994; McDonnell et al., 1996). This may be true for FcRI also; however, this paper demonstrates that changes to the interdomain regions, can dramatically alter function of FcRs, perhaps by altering the structural presentation of the Ig binding domains. Based on homology with other Ig-like structures including CD2 (Jones et al., 1992), CD4 (Wang et al., 1990), KIR (Fan et al., 1997) and FcRII (M.Powell and P.M.Hogarth, in preparation), it is interesting to note that the D1/D2 interface would be spatially close to the FG loop in D2, a region known to play a key role in ligand binding in FcRII and FcRI (Hulett et al., 1993, 1994). In the FcRI-NOD allele, the presence of two highly charged residues (Lys88, Glu89) in this region may impinge on the FG loop, thereby influencing binding. The spatial orientation of the unique third Ig-like domain of FcRI to D1 and D2 has not been determined and it is difficult to speculate how alteration to the D2/D3 junction may influence the binding site(s) contained in the first two Ig-like domains. The consequences of the interdomain alterations, in the case of the FcRI-NOD allele at least, led to dramatic gains or diversity of function.

The Ig-superfamily is comprised of a vast array of members exhibiting diverse functions including cell–cell recognition and adhesion, soluble ligand binding and antigen recognition, all of which depend on the versatility of the shared common Ig homology unit. Williams, Barclay and colleagues postulated that during the evolution of the Ig superfamily, modifications of the function of individual receptors might occur by the acquisition of exons and their subsequent modification (Williams and Barclay, 1988). Analysis of the high-affinity receptor for IgG (FcRI) from strains of mice described herein identifies a mechanism by which the postulated modification of newly acquired exons could take place to provide gains in function.

Sequence analysis of the FcRI-NOD gene demonstrated that the change of AG and subsequent generation of a new splice acceptor site encoding the D2/D3 junction of FcRI-NOD leads to the translation of four extra residues (Ala173-Phe174-Pro175-Leu176). As demonstrated by mutant BALB.B1.N2, this single event would result in a receptor with dramatically reduced affinity for IgG. A second event at the D1/D2 junction leading to the amino acid change AsnLys88, and insertion of Glu89, of itself does not appear to influence function (see BALB.N1.B2); however, in combination both events produce a receptor with enhanced affinity for IgG2a and low cell surface expression (see BALB.N1.N2). Finally, it is the combination of these events, and their interaction with other sequence alterations in the FcRI-NOD allele, that create an entirely new binding interaction with IgG2b and produce the broad specificity, high-affinity receptor that is encoded by the FcRI-NOD allele.

Analysis of the Ig-superfamily reveals many mechanisms for the diversification of function within this family. In contrast to BCR and TCR, FcR must rely solely on evolutionarily acquired variation. Previous studies by our laboratory have demonstrated that the third Ig-like domain unique to FcRI contributes markedly to the particular ligand affinity and specificity exhibited by this receptor class (Hulett et al., 1991). As the related low-affinity receptors have only two Ig-like domains, the evolutionary acquisition of the extra exon encoding domain 3 (either by duplication or acquisition) supports the theory of exon acquisition as a means of acquiring new or diverse functions. The characterization of two alleles of mouse FcRI further develop our ideas about the mechanisms of diversification of function. Indeed, these alleles provide examples where modification at intron/exon junctions can markedly alter function of Ig homology domains.

Antibody reagents

Antibodies used as either tissue culture supernatant or protein A-purified IgG from culture supernatant included UM1 (anti-TNP, mIgG1), UM2a (anti-TNP, mIgG2a) and UM3 (anti-TNP, mIgG3) (Tsujimura et al., 1990), 7-20.6/30 (anti-Ly1.1, mIgG2a) (Hogarth et al., 1980). The antibody 49.2 (anti-TNP, IgG2b) was purchased from PharMingen (San Diego, CA, USA). The isotype of each antibody was confirmed by ELISA using specific anti-mouse Ig isotype antibodies (PharMingen). The monoclonal antibodies that were used in FACS analysis were detected with FITC-conjugated sheep anti-mouse Ig (Amrad Pharmacia Biotech, Boronia, Australia). Purified antibodies were stored at -80°C and thawed once only for use. Analytical gel filtration (Superdex-200, Pharmacia, Uppsala, Sweden) routinely showed that the antibody preparations used were 90–95% monomeric. UM1, UM2a and UM3 were kind gifts from Dr Tohru Masuda (Kyoto University).

Transfection of FcR cDNA constructs

CHO-K1 cells stably expressing either FcRI of BALB/c (1N3-2) or NOD/Lt (2B5-13C) origin were generated by calcium phosphate transfection of the expression vector pCDNA3 (Invitrogen, Carlsbad, CA, USA) containing either mouse FcRI cDNA (Sears et al., 1990; Gavin et al., 1996). Because of initial low surface expression of FcRI-NOD, the stable line 2B5 was sorted twice for clones with high expression (2B5-13C). BALB/c, NOD/Lt and mutant FcRI were also transiently expressed by COS cells using Lipofectamine™ according to the manufacturer's instructions (Life Technologies, Gaithersburg, MD, USA). Binding assays were performed on transfected cells 48 h post transfection and the efficiency of transient transfection and expression was 50%.

IgG isotype binding and competition assays

In the FACS analysis, the binding of monoclonal antibodies UM1 (anti-TNP, mIgG1), UM2a (anti-TNP, mIgG2a), 49.2 (anti-TNP, IgG2b) UM3 (anti-TNP, mIgG3) to transfected CHO cells (1N3-2 and 2B5-13C) were detected using FITC-conjugated sheep anti-mouse Ig (Silenus, Melbourne, Australia). In the competition assays, various concentrations of unlabelled mouse IgG1 (UM1), IgG2a (UM2a), IgG2b (49.2) and IgG3 (UM3) were incubated with 2B5-13C cells at 110⁶ cells/ml for 15 min on ice prior to the addition of ¹²⁵I-labelled IgG2a (2 g/ml). After further incubation on ice for 2 h, the amount of bound ¹²⁵I-labelled IgG2a was determined as described (Gavin et al., 1996). The IC₅₀ value was determined to be the concentration of IgG at which half of the maximum level of radiolabelled IgG2a remained bound. The binding of radiolabelled monomeric IgG2a (7-20.6/30) and IgG3 (UM3) and IgG2b (49.2) to various mouse FcRI-transfected COS cells was assayed and K_d and B_max values determined by Scatchard plot analysis. The B_max value presented is the amount of IgG2a binding sites per cell. Non-specific Ig binding was determined by either the amount of ¹²⁵I-labelled IgG bound in the presence of 100-fold excess of non-radiolabelled IgG, or by non-specific binding to mock-transfected COS cells and the counts subtracted from total binding to give specific IgG bound.

Erythrocyte antibody rosetting

The expression of FcRI mutants was tested using erythrocyte antibodies (EAs), wherein sheep erythrocytes were coated with TNP before sensitization with IgG2a anti-TNP (UM2a) as described (Hulett et al., 1991). Monolayers of cells were directly tested for the expression of functional FcRI by incubation with 2% EA for 10 min at 37°C, then place on ice for 30 min, after which unbound erythrocytes were washed away and the monolayers observed. A cell was scored as binding positively when greater than six red blood cells bound to its surface.

Generation of mutant FcRI cDNA expression constructs

Mutant FcRs were generated using splice overlap extension PCR (SOE–PCR) (Horton et al., 1989) using either FcRI-NOD or FcRI-BALB cDNA. FcRI-NOD cDNA was used as template for the generation of NOD.B1.N2 and NOD.N1.B2 mutants. Briefly, NOD.B1.N2 was generated using oligonucleotide pairs MDH3 + NM1 and NM2 + TISM6 to generate two fragments which were spliced together using oligonucleotides MDH3 + TISM6. For NOD.N1.B2, oligonucleotide pairs MDH3 + NM3 and NM4 and TISM5, followed by MDH3 + TISM6 were used. Full-length mutant cDNAs (1.2kb) were ligated into the expression vector pCDNA3 (Invitrogen, Carlsbad, CA, USA). The NOD.B1.B2 mutant was generated by restriction enzyme digest of NOD.B1.N2 (KpnI and BstBI) and NOD.N1.B2 (BstBI and ApaI), and subsequent ligation of purified fragments into pCDNA3 digested with KpnI and ApaI. FcRI-BALB cDNA was used as a template to generate BALB.N1.B2 and BALB.B1.N2 mutants. Briefly, the generation of BALB.N1.B2 used the oligonucleotide pairs MDH3 + NOD-1 and NOD-2 + TISM6 to generate two fragments which were spliced together using the oligonucleotides MDH3 and TISM6. Similarly, BALB.B1.N2 was generated using MDH3 + NOD-7 and NOD-6 and TISM6, followed by MDH3 + TISM6. Following subcloning into the expression vector pCDNA3 (Invitrogen) the mutant BALB.N1.N2 was generated by ligating purified restriction enzyme fragments (BbsI) of BALB.N1.B2 and BALB.B1.N2. All constructs were confirmed by ABI Prism Dye terminator cycle sequencing, analysed by ABI PRISM 377 (Perkin-Elmer).

Sequencing intron/exon boundaries

Genomic DNA isolated from either NOD/Lt or BALB/c mice was used as template in PCR using Expand™ High Fidelity enzyme according to the manufacturer's recommendations (Boehringer Mannheim, Germany). Oligonucleotides EX1 and EX8, and EX5 and EX6 were used to generate fragments spanning exons encoding domains 1 and 2, and domains 2 and 3 respectively, which were purified and sequenced using ABI PRISM 377.

Oligonucleotides are listed:

NM1: 5'-CCAATCATTGTGGATTTGCAACTG-3';

NM2: 5'-ATCCACAATGATTGGCTGCTACTC-3';

NM3: 5'-AACAGCTCTTTCACCGTGATGGA-3';

NM4: 5'-ACGGTGAAAGAGCTGTTTACCAC-3';

NOD-1: 5'-CCAATCTTCCTTGTGGATTTGC-3';

NOD-2: 5'-CCACAAGGAAGATTGGCTGCTA-3';

NOD-6: 5'-GAAAGCTTTCCCTTTAGAGCTGTTTACCACGC-3';

NOD-7: 5'-CTCTAAAGGGAAAGCTTTCACCGTGATGGA-3';

MDH3: 5'-TTTGTCGACATGATTCTTACCAGCTT-3';

TISM6: 5'-TTTGAATTCCAGTCTGTATATTTGC-3';

EX1: 5'-caccagtggcttaggttctgc-3';

EX8: 5'-ggatgcttttgtagtggctcttc-3';

EX5: 5'-AAGAATAAACTGGTGTACAATGTGG-3';

EX6: 5'-AAGGAGAAGTGAAGCTGTAAGCC-3'.

The oligonucleotide name is in bold with coding sequence in upper case; the intronic sequence is presented in lower case.

Abdelmoula M, Spertini F, Shibata T, Gyotoku Y, Luzuy S, Lambert PH and Izui S (1989) IgG3 is the major source of cryoglobulins in mice. J Immunol, 143, 526–532. | PubMed | ISI | ChemPort |

Allen JM and Seed B (1989) Isolation and expression of functional high affinity Fc receptor complementary DNAs. Science, 243, 378–381. | Article | PubMed | ISI | ChemPort |

Fan QR, Mosyak L, Winter CC, Wagtmann N, Long EO and Wiley DC (1997) Structure of the inhibitory receptor for human natural killer cells resembles haematopoietic receptors. Nature, 389, 96–100. | Article | PubMed | ISI | ChemPort |

Gavin AL, Hamilton JA and Hogarth PM (1996) Extracellular mutations of non-obese diabetic mouse FcRI modify surface expression and ligand binding. J Biol Chem, 271, 17091–17099. | Article | PubMed | ISI | ChemPort |

Gavin AL, Barnes N, Dijstelbloem HM and Hogarth PM (1998) Identification of the mouse IgG3 receptor: implications for antibody effector function at the interface between innate and adaptive immunity. J Immunol, 160, 20–23. | PubMed | ISI | ChemPort |

Hanson MS, Cetkovic-Cvrlje M, Ramiya VK, Atkinson MA, Maclaren NK, Singh B, Elliott JF, Serreze DV and Leiter EH (1996) Quantitative thresholds of MHC class II I-E expressed on hemopoietically derived antigen-presenting cells in transgenic NOD/Lt mice determine level of diabetes resistance and indicate mechanism of protection. J Immunol, 157, 1279–1287. | PubMed | ISI | ChemPort |

Hibbs ML, Tolvanen M and Carpen O (1994) Membrane-proximal Ig-like domain of FcRIII (CD16) contains residues critical for ligand binding. J Immunol, 152, 4466–4474. | PubMed | ISI | ChemPort |

Hogarth PM, Potter TA, Cornell FN, McLachlan R and McKenzie IF (1980) Monoclonal antibodies to murine cell surface antigens. I. Lyt-1.1. J Immunol, 125, 1618–1624. | PubMed | ISI | ChemPort |

Horton RM, Hunt HD, Ho SN, Pullen JK and Pease LR (1989) Engineering hybrid genes without the use of restriction enzymes: gene splicing by overlap extension. Gene, 77, 61–68. | Article | PubMed | ISI | ChemPort |

Hulett MD and Hogarth PM (1994) Molecular basis of Fc receptor function. Adv Immunol, 57, 1–127. | PubMed | ISI | ChemPort |

Hulett MD, Osman N, McKenzie IF and Hogarth PM (1991) Chimeric Fc receptors identify functional domains of the murine high affinity receptor for IgG. J Immunol, 147, 1863–1868. | PubMed | ISI | ChemPort |

Hulett MD, McKenzie IF and Hogarth PM (1993) Chimeric Fc receptors identify immunoglobulin-binding regions in human FcRII and FcRI. Eur J Immunol, 23, 640–645. | PubMed | ISI | ChemPort |

Hulett MD, Witort E, Brinkworth RI, McKenzie IF and Hogarth PM (1994) Identification of the IgG binding site of the human low affinity receptor for IgG FcRII. Enhancement and ablation of binding by site-directed mutagenesis. J Biol Chem, 269, 15287–15293. | PubMed | ISI | ChemPort |

Hulett MD, Witort E, Brinkworth RI, McKenzie IF and Hogarth PM (1995) Multiple regions of human FcRII (CD32) contribute to the binding of IgG. J Biol Chem, 270, 21188–21194. | Article | PubMed | ISI | ChemPort |

Jones EY, Davis SJ, Williams AF, Harlos K and Stuart DI (1992) Crystal structure at 2.8 Å resolution of a soluble form of the cell adhesion molecule CD2. Nature, 360, 232–239. | Article | PubMed | ISI | ChemPort |

Lemoine R, Berney T, Shibata T, Fulpius T, Gyotoku Y, Shimada H, Sawada S and Izui S (1992) Induction of 'wire-loop' lesions by murine monoclonal IgG3 cryoglobulins. Kidney Int, 41, 65–72. | Article | PubMed | ISI | ChemPort |

Luan JJ, Monteiro RC, Sautes C, Fluteau G, Eloy L, Fridman WH, Bach JF and Garchon HJ (1996) Defective FcRII gene expression in macrophages of NOD mice: genetic linkage with up-regulation of IgG1 and IgG2b in serum. J Immunol, 157, 4707–4716. | PubMed | ISI | ChemPort |

Mallamaci MA, Chizzonite R, Griffin M, Nettleton M, Hakimi J, Tsien WH and Kochan JP (1993) Identification of sites on the human FcRI alpha subunit which are involved in binding human and rat IgE. J Biol Chem, 268, 22076–22083. | PubMed | ISI | ChemPort |

McDonnell JM, Beavil AJ, Mackay GA, Jameson BA, Korngold R, Gould HJ and Sutton BJ (1996) Structure based design and characterization of peptides that inhibit IgE binding to its high-affinity receptor. Nature Struct Biol, 3, 419–426. | Article

Mongini PK, Stein KE and Paul WE (1981) T cell regulation of IgG subclass antibody production in response to T- independent antigens. J Exp Med, 153, 1–12. | Article | PubMed | ISI | ChemPort |

Perlmutter RM, Hansburg D, Briles DE, Nicolotti RA and Davie JM (1978) Subclass restriction of murine anti-carbohydrate antibodies. J Immunol, 121, 566–572. | PubMed | ISI | ChemPort |

Podolin PL, Denny P, Lord CJ, Hill NJ, Todd JA, Peterson LB, Wicker LS and Lyons PA (1997) Congenic mapping of the insulin-dependent diabetes (Idd) gene, Idd10, localizes two genes mediating the Idd10 effect and eliminates the candidate Fcgr1. J Immunol, 159, 1835–1843. | PubMed | ISI | ChemPort |

Porges AJ, Redecha PB, Doebele R, Pan LC, Salmon JE and Kimberly RP (1992) Novel Fc receptor I family gene products in human mononuclear cells. J Clin Invest, 90, 2102–2109. | PubMed | ISI | ChemPort |

Prins JB et al. (1993) Linkage on chromosome 3 of autoimmune diabetes and defective Fc receptor for IgG in NOD mice. Science, 260, 695–698. | Article | PubMed | ISI | ChemPort |

Sears DW, Osman N, Tate B, McKenzie IF and Hogarth PM (1990) Molecular cloning and expression of the mouse high affinity Fc receptor for IgG. J Immunol, 144, 371–378. | PubMed | ISI | ChemPort |

Serreze DV, Leiter EH, Kuff EL, Jardieu P and Ishizaka K (1988) Molecular mimicry between insulin and retroviral antigen p73. Development of cross-reactive autoantibodies in sera of NOD and C57BL/KsJ db/db mice. Diabetes, 37, 351–358. | PubMed | ISI | ChemPort |

Slack J, Der-Balian GP, Nahm M and Davie JM (1980) Subclass restriction of murine antibodies. II. The IgG plaque-forming cell response to thymus-independent type 1 and type 2 antigens in normal mice and mice expressing an X-linked immunodeficiency. J Exp Med, 151, 853–862. | Article | PubMed | ISI | ChemPort |

Tate BJ, Witort E, McKenzie IF and Hogarth PM (1992) Expression of the high responder/non-responder human Fc gamma RII. Analysis by PCR and transfection into FcR-COS cells. Immunol Cell Biol, 70, 79–87. | Article | PubMed | ISI | ChemPort |

Tsujimura K, Park YH, Miyama-Inaba M, Meguro T, Ohno T, Ueda M and Masuda T (1990) Comparative studies on FcR (FcRII, FcRIII, and FcR ) functions of murine B cells. J Immunol, 144, 4571–4578. | PubMed | ISI | ChemPort |

Wang JH et al. (1990) Atomic structure of a fragment of human CD4 containing two immunoglobulin-like domains. Nature, 348, 411–418. | Article | PubMed | ISI | ChemPort |

Warmerdam PA, van de Winkel JG, Vlug A, Westerdaal NA and Capel PJ (1991) A single amino acid in the second Ig-like domain of the human Fc gamma receptor II is critical for human IgG2 binding. J Immunol, 147, 1338–1343. | PubMed | ISI | ChemPort |

Williams AF and Barclay AN (1988) The immunoglobulin superfamily–domains for cell surface recognition. Annu Rev Immunol, 6, 381–405. | Article | PubMed | ISI | ChemPort |

Main navigation

• Search EMBO Journal

  Search

• Advanced search

• The EMBO Journal

  • Home
  • Aims and scope
  • Current issue
  • Advance online publication
  • Web focuses
  • Archive:
  • Browse by issue
  • Browse by subject
  • Browse by category
  • Free online sample issue
  • Press releases

• Authors and Referees

  • Editorial process
  • Guide for authors
  • Submit an article
  • Guide for referees
  • Editors and Editorial Board
  • Contact editorial office

• Customer Service

  • Subscribe
  • Order sample copy
  • Purchase articles
  • Reprints and permissions
  • Contact NPG
  • Advertising