Effect of Transmembrane and Cytoplasmic Domains of IgE on the IgE Response

See allHide authors and affiliations

Science  18 Apr 1997:
Vol. 276, Issue 5311, pp. 409-411
DOI: 10.1126/science.276.5311.409


B cells use immunoglobulin M (IgM) and IgD as antigen receptors, but after contact with antigen they can switch and use IgG, IgA, or IgE. In mice lacking the transmembrane and cytoplasmic domains of IgE, serum IgE is reduced by more than 95 percent and, after immunization, specific responses are negligible. In mice lacking most of the cytoplasmic tail of IgE, serum IgE levels are reduced by 50 percent and specific responses are reduced by 40 to 80 percent, without a clear secondary response. Thus, membrane expression is indispensable for IgE secretion in vivo, and the cytoplasmic tail influences the degree and quality of the response.

Immunoglobulin E contributes least to the serum immunoglobulins (Igs). Its specific function is not understood, although it is well known as the cause of allergic reactions (1). IgE, like other Igs, is also expressed as an integral membrane protein (mIgE) on B cells. The transmembrane segments of mIgs are 25 amino acids long, whereas the cytoplasmic domains vary in size from three residues [Lys-Val-Lys (KVK)] for mIgM and mIgD to 14 to 28 residues for other isotypes (2). The nature and effects of the signals generated by mIgs other than IgM and IgD are mostly unknown, but they may control affinity maturation, memory induction, and differentiation into plasma cells (3). To study the role of the transmembrane domain and cytoplasmic tail of mIgE, we made mouse lines that carried mutations in these domains in the germline ɛ gene, using the gene-targeting technique in embryonic stem (ES) cells (4, 5) (Fig. 1). The ΔM1M2 line lacks the transmembrane and cytoplasmic domains of IgE, whereas the KVKΔtail line can only express a cytoplasmic tail of three amino acid residues (KVK), which is identical to the cytoplasmic domain of mIgM and mIgD.

Figure 1

Construction of the mutant mouse lines ΔM1M2 and KVKΔtail. (A) Organization of theCɛ gene. The four constant-region exons are marked as CH1 to CH4. Membrane exons M1 and M2 are marked as indicated in inset at right. Selected restriction enzyme sites and the probes used for Southern blot analysis are shown. (B) Linearized targeting vector. (C) The Cɛ allele after primary targeting. (D) Cre-mediated recombination between the two loxP sites flanking the neo and tk genes results in the generation of the KVKΔtail allele. (E) Cre-mediated recombination between the most 5′ and 3′ loxP sites creates the ΔM1M2 allele. (F and G) Southern blot analysis of the primary targeting event. (F) Hind III–Eco RI–digested DNA was hybridized with the external probe. The sizes of wild-type and target fragments are indicated. The smaller (target) band is indicative of the presence of the singular loxP site (marked by the Eco RI site). Both clones 11/5/3 and 15/6/5 show equal intensities of the wild-type and target bands. (G) Hybridization of the Xba I–digested DNA with the neo probe shows singular integration events of clones 11/5/3 and 15/6/5. (H and I) Southern blot analysis of heterozygous and homozygous mice. (H) ΔM1M2 was confirmed by Hind III digestion and hybridization with the external probe. The sizes of the wild-type and targeted fragments are indicated. (I) KVKΔtail was confirmed by a Hind III–Sac I double digest and hybridization with a probe spanning the membrane exons. The mutant allele increases by the size of one singular loxP site (100 bp).

Serum Igs in 7-week-old unprimed mice were measured (Fig.2A). There was no difference in IgM, IgG1, IgG2a, IgG2b, IgG3, and IgA titers between age- and sex-matched wild-type and mutant mice; however, serum IgE was reduced by 94 to 98% in ΔM1M2 mice and by 50% in KVKΔtail mice. Similar reductions were found in 3- and 6-month-old animals (Fig. 2B). Thus, the mutations have no effect on serum concentrations of Igs other than IgE.

Figure 2

Serum levels of immunoglobulins in wild-type (black bars), KVKΔtail (white bars), and ΔM1M2 (gray bars) mice. (A) Seven-week-old mice of each group were bled, and the sera of individual mice of each group were pooled (6) (B).The mice used for immunization experiments with DNP-OVA (see Fig. 3A) were bled before and after immunization. w7, 7-week-old mice, preimmune sera; w14 and w24, 14- and 24-week-old mice, bled 1 week and 11 weeks after the second booster immunization. Serum levels of total IgE were determined and are shown above bars. Error bars indicate standard errors.

The immune response of the mutant mice was assessed with two different immunization protocols. First, antibody titers were measured after mice were immunized with the T cell–dependent antigen 2,4-dinitrophenyl-ovalbumin (DNP-OVA) (6). Serum levels of specific IgG1 were comparable in wild-type and mutant mice (Fig.3A). In ΔM1M2 mice, DNP-specific IgE antibodies were barely detectable, and in KVKΔtail mice titers were 50 to 80% lower than in control mice (Fig. 3B). Titers increased after the first booster, but a clear secondary response, as characterized by a strong and fast rise in specific antibody titer, was absent (Fig. 3B). Therefore, the transmembrane domain of IgE is indispensable for T cell–dependent IgE secretion, and the cytoplasmic tail influences the degree and quality of the response.

Figure 3

Serum IgE responses in immunized mice. (A and B) Mice (six animals per group) were immunized with DNP-OVA and received booster immunizations after 2 and 6 weeks. Mice were bled at the indicated times and the sera of each group were pooled. Results are expressed as the serum dilution where half-maximum absorbance was obtained for IgG1 (A) and IgE (B). (C) Mice (five animals per group) were infested withN. brasiliensis at d0 and d77 and bled at the indicated times. Results are given as arithmetic means of serum levels of IgE. Bars indicate standard errors.

To determine whether the reduction in IgE titers in the mutant lines was caused by reduced levels of IgE production per cell or by a smaller number of cells that produce IgE, we measured the increase in the number of IgE-secreting cells 1 week after a third booster immunization with DNP-OVA (7). There was an average increase of 786 IgE-secreting cells in the wild-type mice, an average of 334 in KVKΔtail mice, and an average of 0 in ΔM1M2 mice. Therefore, the reduced IgE levels in the mutant mice reflected smaller numbers of IgE-secreting cells.

The second immunization protocol involved infestation with the helminth Nippostrongylus brasiliensis (8).N. brasiliensis induces robust IgG1 and IgE production, both through a dominant activation of type 2 T helper (TH2) cells and a strong T cell–independent activation of B cells (9). Switch recombination to IgE and IgG1 is dependent on interleukin-4 (IL-4), whereas switch recombination to IgG2a is not induced by IL-4 (10, 11). Serum IgG1 and IgG2a levels showed the expected pattern both in wild-type and mutant mice: an 8- to 10-fold increase in IgG1 levels at day 14 after infestation, and no increase in IgG2a titer. In wild-type mice (Fig. 3C), serum IgE rose from 300 ng/ml to 16 μg/ml by day 14 after infestation; in ΔM1M2 mice it rose from 20 ng/ml to 1.4 μg/ml; and in KVKΔtail mice it rose from 130 ng/ml to 8 μg/ml. After secondary infestation 11 weeks after the first challenge with N. brasiliensis, a strong and fast IgE response was seen in the wild-type mice (Fig.3C), which is indicative of a memory response. The IgE response in the KVKΔtail mice was also substantial but was at 55% of the wild-type response. In the ΔM1M2 mice, IgE was now clearly measurable; however, the response was sluggish and was reduced (13% of the wild-type response). The results indicate that the IgE response to N. brasiliensis also needs a specific interaction with the IgE antigen receptor complex on the B cell, accompanied by strong TH2 cell activity.

To determine whether class switch to IgE was impaired by the targeting event, we stimulated isolated spleen cells of wild-type, ΔM1M2, and KVKΔtail mice in vitro with lipopolysaccharide (LPS) and IL-4 (10, 12). As shown in Fig. 4, the concentrations of IgE and IgG1 in the culture supernatants were comparable in wild-type and mutant mice. These results imply that the reduced IgE titers found in both mutant lines are solely a reflection of the loss of biological activities associated with the transmembrane and cytoplasmic domains of IgE.

Figure 4

(A) IgG1 and (B) IgE production of splenic B cells after in vitro stimulation with LPS and IL-4 at the indicated concentrations. Wild type, black bars; KVKΔtail, white bars; ΔM1M2, gray bars. Error bars indicate standard errors.

There are two possible explanations for these findings. First, signals generated via mIg are needed at all times, not only for the maturation process but also for the expansion of antigen-specific cells. Second, antigen presentation to TH cells is necessary during an antibody response, and only the antigen receptor is capable of effective antigen capture for presentation. The hypotheses are not necessarily mutually exclusive. All Ig classes can associate with the Ig-α–Ig-β heterodimer, the signal-transducing unit of the B cell receptor (2, 13). Further, both the Ig-α–Ig-β sheath and the cytoplasmic tail of mIg (14,15) have been implicated in guiding receptor-bound antigen via the receptor to the antigen-processing compartments. Key residues for internalization are present in the tails in the form of a Tyr-X-X-Ile/Met motif, where X is any amino acid (15,16). These facts predict that the results we obtained in the KVKΔtail and ΔM1M2 lines can be extended to the IgG isotypes and perhaps to IgA. Indeed, Kaisho et al. (17) reach very similar conclusions in studying mice carrying matching mutations in the γ1 gene.

  • * Present address: Institut für Genetik und Allgemeine Biologie, Hellbrunnerstraße 34, A-5020 Salzburg, Austria.

  • To whom correspondence should be addressed. E-mail: lamers{at}


Stay Connected to Science

Navigate This Article