The Roles of γ1 Heavy Chain Membrane Expression and Cytoplasmic Tail in IgG1 Responses

See allHide authors and affiliations

Science  18 Apr 1997:
Vol. 276, Issue 5311, pp. 412-415
DOI: 10.1126/science.276.5311.412


In antibody responses, B cells switch from the expression of immunoglobulin (Ig) μ and δ heavy (H) chains to that of other Ig classes (α, γ, or ɛ), each with a distinct effector function. Membrane-bound forms of α, γ, and ɛ, but not μ and δ, have highly conserved cytoplasmic tails. Mutant mice unable to express membrane γ1 H chains or producing tailless γ1 H chains failed to generate efficient IgG1 responses and IgG1 memory. H chain membrane expression after class switching is thus required for these functions, and class switching equips the B cell antigen receptor with a regulatory cytoplasmic tail that naı̈ve B cells lack.

The B cell antigen receptor (BCR) initially expressed on B lymphocytes consists of membrane Ig (mIg) made up of μ heavy (H) chains plus light chains, associated with the Ig-α–Ig-β heterodimer (1). Because μ H chains have a cytoplasmic tail of only three residues, signaling through the BCR depends on the cytoplasmic tails of Ig-α and Ig-β. On stimulation with antigen, B cells often undergo isotype switching (2) leading to the expression of H chains of other classes. These other H chains are not only present in secreted antibodies but can also be expressed as a component of the BCR on the cell surface. Indeed, memory B cells, produced in T cell–dependent antibody responses, classically express BCRs that contain H chains of classes other than μ or δ. The membrane forms of γ, ɛ, and α H chains differ conspicuously from those of μ and δ in that they possess cytoplasmic tails of 28 (γ, ɛ) and 14 (α) amino acids, which are highly conserved in evolution (1). This raises the possibility that the BCRs expressed on naı̈ve and memory B cells exhibit distinct signaling properties.

Memory B cells generated in response to protein antigens typically express γ1 chains on the cell surface. We generated two types of mouse mutants: one unable to express γ1 chains on the membrane (IgG1Δ m) and the other able to express mIgG1 but lacking the γ1 cytoplasmic tail (IgG1Δ tail). The generation of the IgG1Δ m and IgG1Δ tail mutations by gene targeting is depicted in Fig. 1. A single vector was used to target the γ1 locus of embryonic stem (ES) cells bearing the IgHa allele (3). The IgG1Δ tail mutation was generated by Cre recombinase–mediated deletion of the neomycin resistance (neo r) gene from the targeted locus (4). The IgG1Δ tail mutation leads to the production of γ1 chains with a truncated cytoplasmic tail of three amino acids, identical to that of μ chains. The IgG1Δ tail mutation was transmitted into the mouse germ line, and the IgG1Δ m mutation was derived from it by a second step of Cre-mediated deletion in vivo, using the deleter strain (5).

Figure 1

Generation of IgG1Δ tail and IgG1Δ m mice (3). Genomic structure of (A) the murine IgG1 constant region locus, (B) the targeting vector, (C) the targeted allele, (D) the IgG1Δ tail allele, and (E) the IgG1Δ m allele are shown. The exons (coding parts depicted in open boxes, noncoding parts in shaded boxes) are marked. The exons M1 and M2 encode the transmembrane and cytoplasmic regions of IgG1. The loxPsites are indicated as solid triangles. The introduced stop codon (stop), the lengths of diagnostic restriction fragments (arrows), and the probe used for Southern (DNA) blot analysis (black bar) are shown. B, Bam HI. (F) Southern blot analysis. Bam HI digestion of genomic DNA yields fragments of 12.0, 9.8, and 8.0 kb corresponding to the wild-type (+), IgG1Δ tail (Δtail), and IgG1Δ m (Δm) allele, respectively.

The effects of the two mutations on B cell development and function were assessed in mice heterozygous or homozygous for either mutation. Heterozygous animals carried a wild-type allele of ballotype, the products of which can be distinguished from those of the mutant a alleles by anti-allotypic antibodies. All animals generated IgM- and IgD-bearing B cells in normal numbers. When B cells from homozygous mutants of either type were activated in vitro by bacterial lipopolysaccharide (LPS) in the presence of interleukin 4 (IL-4) (6), they switched to IgG1 expression with equal efficiency as the wild type. This was determined by intracellular staining of the activated cells after fixation and incubation with antibodies to IgG1 (Fig. 2A). Cells from the same cultures were also stained for IgG1 surface expression (Fig.2B). No staining was observed in the case of the IgG1Δ m mutant, as expected. In the case of IgG1Δ tail, surface IgG1-positive cells were seen at the expected frequency. However, the average staining intensity was roughly threefold less than that of wild-type cells, for reasons that remain to be explored.

Figure 2

Flow cytometric analysis of spleen cells derived from wild-type (thin lines), IgG1Δ tail/ Δ tail(bold lines), and IgG1Δ m/ Δ m mice (dotted lines) after in vitro stimulation with LPS and IL-4 (17). Shown are (A) staining of intracellular IgG1 and (B) IgG1 surface expression.

Marked differences were seen between mutants and the wild type at the level of IgG1 responses. When heterozygous mutants were immunized with a T cell–dependent antigen, chicken γ-globulin (CG) coupled to 4-hydroxy-3-nitro-phenylacetyl (NP), they produced far less NP-specific IgG1 from the mutant than from the wild-type alleles. The differences were more pronounced for the IgG1Δ m than for the IgG1Δ tail mutant and most severe in the secondary response, where in both cases, IgG1 from the mutant alleles was hardly detectable (7). In homozygous mutants (Fig.3), the serum IgG1 concentrations were reduced by factors of 24 for IgG1Δ tail and 71 for IgG1Δ m, compared with control mice of strain 129 (Fig. 3A). Other IgG classes were unaffected. Upon immunization with NP-CG, both mutant strains produced strongly impaired primary and secondary IgG1 responses, which were about two orders of magnitude below the control in the case of the IgG1Δ m and about one order of magnitude in that of the IgG1Δ tail mutant (Fig. 3, C and D). Again, other Ig classes were unaffected (Fig. 3, E and F). Affinity maturation of IgG1 antibodies in the course of the response was less efficient in the IgG1Δ tail mutant than in the controls (Fig. 3B). For the IgG1Δ m mutant, the low antibody titers did not allow us such an analysis.

Figure 3

Immune response in wild-type (open dots), IgG1Δ tail/ Δ tail(black dots), and IgG1Δ m/ Δ m (gray dots) mice (18). (A) Serum levels of Ig isotypes are shown for unimmunized wild-type, IgG1Δ tail/ Δ tail, and IgG1Δ m/ Δ m mice. (B) Affinity maturation of NP-specific IgG1 serum antibody in wild-type and IgG1 mice is shown at various days after the first immunization. (C to F) NP-specific IgG1 (C andD) and IgG2b (E and F) serum levels in wild-type, IgG1Δ tail/ Δ tail , and IgG1Δ m/ Δ m mice. Immunizations at day 0 and day 42 are indicated by arrows. Bars indicate mean values.

Our results demonstrate that B cells expressing either the IgG1Δ m or the IgG1Δ tail mutation are unable to mount efficient primary and secondary IgG1 responses in vivo, either in competition with B cells expressing wild-type IgG1 (the heterozygous mutant animals) or in the absence of such cells (Fig.3). Assuming that these responses depend on the expansion of cells expressing IgG1 in their BCR as earlier work suggests (8), this result is expected for the IgG1Δ m mutant, which is unable to express IgG1 at the surface. In the case of the IgG1Δ tail mutant, the impaired response could be because of inefficient expansion or persistence of surface IgG1-bearing cells, impaired terminal differentiation into antibody secreting cells, or both. We have approached this matter by determining the numbers of surface IgG1-positive splenic B cells in homozygous mutant mice, either disregarding their antigen-binding specificity or as IgG1+NP+B cells during the anti-NP response. A well-defined subset of IgG1-positive cells was identified 4 weeks after immunization with NP-CG in both mutant and wild-type mice (Fig.4) (9). In contrast to what was seen in the case of LPS-activated blasts (Fig. 2), the intensity of staining for mIgG1 did not apparently differ between the wild type and the mutant for most of the cells of this subset, although there may be a small fraction of bright cells in the wild type that are missing in the mutant (Fig. 4, A and B). A minority of the IgG1-positive cells specifically bind NP-carrier conjugates (Fig. 4, C and D), in agreement with earlier studies (10). These cells represent typical memory B cells selected in the germinal center reaction upon primary immunization with T cell–dependent antigens (11). The mutant animal analyzed in Fig. 4 harbored significantly (approximately 25 times) less IgG1-positive cells than did its wild-type counterpart. As a rule and irrespective of intentional immunization, the mutants displayed roughly ten times fewer IgG1-positive cells in their spleens than did the wild type (Table 1).

Figure 4

Representative flow cytometric analyses of mIgM/mIgD spleen cells from wild-type (A and C) and IgG1Δ tail/ Δ tail (B and D) mice at day 28 after immunization with NP-CG. (A) and (B) Histograms showing surface IgG1 expression. Bars indicate the mIgG1-positive populations. (C) and (D) NP-binding versus mIgG1 expression. NP-binding and nonbinding mIgG1 positive cells are framed. In dot plots, 105 lymphocytes [as defined by forward and side scatters (19)] were collected, and mIgM/mIgD cells were analyzed. For the histograms, 2 million lymphocytes were collected, and the bulk of mIgG1-negative cells (as indicated by the leftmost vertical lines of dot plots) were excluded.

Table 1

NP-binding and mIgG1-expressing cell populations in the spleen of wild-type and IgG1Δ tail/ Δ tailmice before and after immunization with NP-CG. We analyzed 105 cells as described in Fig. 4. Results are expressed as percentages of total spleen cells. Mean values and standard deviations are shown, except for the results obtained at day 49.

View this table:

We conclude that surface IgG1 expression is essential for the generation of efficient primary and secondary IgG1 responses and that both the primary IgG1 response as well as the expansion or maintenance, or both, of IgG1-bearing memory B cells depend strongly on the cytoplasmic tail of the γ1 chain. Achatz et al.(12) have reached similar conclusions for the IgE response, suggesting that IgG, IgE, and perhaps IgA responses follow similar rules.

How could the cytoplasmic tail of the γ1 chain and perhaps other H chains exert its function? One possibility is that it stabilizes IgG1 surface expression (Fig. 2), allowing more efficient triggering of IgG1-expressing cells by antigen. However, as the level of IgG1 surface expression appeared to be close to normal in memory cells of IgG1Δ tail mice (Fig.4), we favor the view that the tail is directly involved in the mechanism by which IgG1-expressing B cells are triggered into the response. Weiser et al. (13) have shown that transformed B cells expressing an IgG2a BCR require the cytoplasmic tail of γ2a (and specifically a tyrosine-based motif in this structure, which is also present in γ1) for efficient presentation of antigen to T cells, following sIg-mediated internalization. As IgG1 responses are usually T cell–driven, inefficient presentation of antigen to T cells by B cells expressing tail-less γ1 chains could explain the observed phenotype of the IgG1Δ tail mutant. Thus, upon switching to IgG1 expression in the T cell–driven germinal center reaction, further expansion and mutation of the antigen-activated cells would become dependent on IgG1-BCR–mediated antigen presentation and, therefore, the cytoplasmic tail of the γ1 chain. This would explain the inefficiency of affinity maturation and memory cell generation in the mutant.

The cytoplasmic tails of γ chains may by themselves be unable to transduce a signal into the cell (14). However, if the triggering of IgG-expressing germinal center and memory B cells [which are known to be potent antigen-presenting cells (15)] depends exclusively on their interaction with activated T helper cells, then the activation of these B cells would be mediated by the cytoplasmic tails of the γ chains instead of the Ig-α–Ig-β heterodimer involved in the activation of IgM-expressing naı̈ve B cells. Consistent with this possibility is the finding that in cultured cells, IgG can be expressed at the cell surface in the absence of Ig-α and Ig-β (16). Whether this is the case in murine IgG1-positive memory B cells is presently unknown.

On the basis of our results and those of Achatz et al.(12) and of Weiser et al. (13), and given the structural differences between the cytoplasmic tails of antibody H chains of different classes (associated with different effector functions), these structures whose significance was previously elusive, emerge as important regulators of the class distribution of antibody responses and of immunological memory, and thus as potential targets for therapeutic intervention.

  • * To whom correspondence should be addressed.

  • Present address: Department of Biochemistry, Hyogo College of Medicine, Hyogo 663, Japan. E-mail: tkaisho{at}


View Abstract

Navigate This Article