Tetraspanins have been hypothesized to facilitate the organization of functional multimolecular membrane complexes. In B cells the tetraspanin CD81 is a component of the CD19/CD21 complex. When coligated to the B cell Ag receptor (BCR), the CD19/CD21 complex significantly enhances BCR signaling in part by prolonging the association of the BCR with signaling-active lipid rafts. In this study CD81 is shown to associate with lipid rafts upon coligation of the BCR and the CD19/CD21 complex. Using B cells from CD81-deficient mice we demonstrate that in the absence of CD81, coligated BCR and CD19/CD21 complexes fail to partition into lipid rafts and enhance BCR signaling from rafts. Furthermore, a chimeric CD19 protein that associates only weakly if at all with CD81 fails to promote the association of coligated BCR with lipid rafts. The requirement for CD81 to promote lipid raft association may define a novel mechanism by which tetraspanins function as molecular facilitators of signaling receptors.
The tetraspanins are a highly conserved family of proteins containing four membrane-spanning domains, two extracellular loop domains, and short N- and C-terminal cytoplasmic domains (1, 2). The tetraspanin family is distinguished from other proteins that contain four transmembrane domains by highly conserved residues in the membrane-spanning domain and a cysteine containing motif in the large extracellular loop (LEL) 3 domain (3). Members of the tetraspanin family associate with a variety of different membrane protein complexes and with other tetraspanins (2). Tetraspanins have been implicated in diverse cellular functions, including adhesion, motility, metastasis, activation, proliferation, and differentiation. How the members of the tetraspanin family function in such a remarkable array of different activities remains to be elucidated.
CD81 is expressed on a wide variety of cells, including B cells, where it is a component of a complex composed of CD19, a transmembrane protein that serves as a signaling/adaptor protein; CD21, the receptor for the C3d component of complement; and Leu 13 (CD225), a small transmembrane IFN-α- and IFN-γ-inducible protein (4, 5, 6, 7). When coligated to the B cell Ag receptor (BCR) through the binding of complement-tagged Ags, the CD19/CD21 complex functions to enhance BCR signaling, thus lowering the threshold for B cell activation (8, 9). The critical role of the CD19/CD21 complex in Ab responses to T cell-dependent Ags is demonstrated by the phenotype of mice that are CD19 or CD21 deficient, which show significantly reduced Ab responses and/or germinal center formation (10, 11). Furthermore, normal mice, immunized with a recombinant protein that contains the Ag hen egg lysozyme (HEL) fused to three tandemly arranged copies of C3d (HEL-C3d) require 1/1000th the Ag dose of HEL alone to elicit the same Ab response (12).
Recently, evidence was provided that the CD19/CD21 complex functions to enhance BCR signaling in part by prolonging the association of the BCR with lipid rafts (13). Lipid rafts are submicroscopic, dynamic, cholesterol- and sphingolipid-rich membrane microdomains that in B cells serve to selectively concentrate a small number of membrane proteins, including the Src family kinase Lyn, and exclude the majority of membrane proteins, including negative regulators of BCR signaling, such as CD45 and CD22 (14, 15). In resting B cells the lipid rafts also exclude the BCR and the CD19/CD21 complex (13). Upon coligation, induced by the binding of complement-tagged Ags, both the BCR and the CD19/CD21 complex associate with lipid rafts. Signaling is initiated from within the rafts, as evidenced by phosphorylation of the Igα component of the BCR and CD19 and the recruitment of Vav, the guanine-nucleotide exchange factor for the Rho family of small GTPases (13). Significantly, the coligation of the BCR and the CD19/CD21 complex results in the prolonged association of the BCR with lipid rafts and prolonged signaling from the rafts compared with that achieved by BCR cross-linking alone (13). The ability of the CD19/CD21 complex to enhance BCR signaling by stabilizing its association with rafts appears to be a novel mechanism by which coreceptors can function.
To determine the role of CD81 in the enhancement of BCR signaling by the CD19/CD21 coreceptor complex, we studied CD81-deficient mice. Previous analyses of cd81−/− mice provided evidence that CD81 is necessary for normal responses to T-dependent Ags (16, 17, 18, 19). In CD81-deficient mice the primary Ab responses to weak T-dependent Ags are reduced compared with those in wild-type mice (16, 18). However, responses to type I T-independent Ags in CD81-deficient mice appear normal, and responses to type II T-independent Ags are normal or exaggerated (17, 18). BCR cross-linking in vitro resulted in normal Ca2+ fluxes and patterns of tyrosine-phosphorylated proteins in CD81-deficient B cells (19). The effect of coligation of the BCR and the CD81-deficient CD19/CD21 complex on BCR signaling was not directly examined, but impaired immune responses to T-dependent Ags in vivo in CD81-deficient mice may reflect a failure of the CD19/CD21 complex to function normally in response to the immunizing Ag. CD81 was also shown to be important for the expression of CD19 (16, 17, 19, 46), suggesting that CD19, which directly associates with CD81, is dependent on CD81 for its stable expression on the plasma membrane. In contrast to CD19, CD21 levels are normal on CD81-deficient B cells (16, 17, 19).
In this study evidence is presented using B cells from CD81-deficient mice that CD81 is required for the association of the BCR and the CD19/CD21 complex with lipid rafts and for enhanced signaling from rafts upon coligation. In addition, a chimeric CD19 protein expressed in human Daudi B cells that associates only poorly with CD81 does not facilitate partitioning of the coligated BCR into lipid rafts. The ability of CD81 to enable association of BCR/coreceptor complexes with lipid rafts may define a novel mechanism by which members of the tetraspanin family stabilize their associated signaling receptors in functional membrane microdomains.
Materials and Methods
Mice and cell lines
CD81-deficient female and male mice (3–6 mo of age) bred on the BALB/c background and CD81-expressing wild-type littermates were generated and characterized as previously described (20). CD81-expressing MD4 × CBA/J female mice (aged 10 mo) transgenic for the HEL-specific BCR were obtained from The Jackson Laboratory (Bar Harbor, ME). CD81-deficient female BALB/c mice (aged 3–6 mo) transgenic for the anti-HEL-specific BCR were previously described (21). The mature human B lymphoblastoid cell line Daudi-expressing cell surface IgM, CD19, CD21, and CD81 was stably transfected with a chimeric CD4/19 construct in which the ectodomain of CD19 was replaced with that of CD4. The transfection, expression, and sequencing were previously described (22). The cells were cultured in IMDM (American Type Culture Collection, Manassas, VA) supplemented with 10% FCS (Life Technologies, Grand Island, NY), penicillin (100 U/ml) and streptomycin (100 μg/ml) both from Sigma-Aldrich (St. Louis, MO), and G-418 (2 μg/ml) (Life Technologies).
Ags and Abs
The recombinant protein HEL-C3d was expressed and purified as previously described (13). mAbs specific for human CD81 (5A6) and mouse CD81 (Eat-2) were generated and characterized as previously described (23, 24). The rat mAb cell lines, 7E9 and 7G6, specific for mouse CD21/CD35, were provided by Dr. M. Holers (University of Colorado Health Sciences Center, Denver, CO). Rabbit polyclonal Abs specific for Lyn kinase, phospholipase Cγ2 (PLCγ2), or Vav and goat polyclonal Abs specific for mouse CD21 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The following Abs were purchased from BD PharMingen (San Diego, CA): rat mAb specific for mouse CD19 (1D3), rat mAb specific for mouse CD45R/B220, and purified rat IgG2a and IgG2b. Biotinylated Fab of mouse anti-human IgM (DA4.4), biotinylated (Fab′)2 mouse anti-human CD19 IgG (ADF4.2), and polyclonal rabbit Abs specific for the cytoplasmic tail of human CD19 were previously described (25). A mAb specific to the ectodomain of human CD19 (BU12) was purchased from Ancell Immunology Research Products (Bayport, MN). Biotinylated anti-human CD4 IgG was purchased from Caltag Laboratories (Burlingame, CA), and the rabbit anti-human CD4 antiserum (T4-4) used for immunoblotting was obtained from the National Institutes of Health AIDS Reagents Program (Bethesda, MD). Mouse mAb specific for the extracellular domain of human CD21, mAb 171 (26), was provided by Dr. M. Holers (University of Colorado Health Sciences Center). A mAb specific for human CD45 was purchased from BD Transduction Laboratories (Lexington, KY). HRP-conjugated goat secondary Abs specific for rabbit, mouse, rat and hamster IgG and egg-white avidin were purchased from Jackson ImmunoResearch Laboratories (West Grove, PA). Rabbit polyclonal Abs specific for mouse Igα were generated and characterized in this laboratory (27). HM57 is a mouse mAb specific for human Igα (28). The HRP-conjugated phosphotyrosine-specific mAbs, PY20H and RC20H, were purchased from Transduction Laboratories. The biotin-conjugated, phosphotyrosine-specific mAb, 4G10, was purchased from Upstate Biotechnology (Lake Placid, NY). BSA, paraformaldehyde, and saponin were obtained from Sigma-Aldrich. ImmunoPure immobilized streptavidin (streptavidin-agarose) was purchased from Pierce (Rockford, IL). Affinity-purified (Fab′)2 goat Ab specific for mouse IgM and IgG and Cy2-conjugated Fab goat Ab specific for mouse Igμ were purchased from Jackson ImmunoResearch Laboratories. AlexaFluor 647-conjugated streptavidin, DiIC16, and Fast DiI were purchased from Molecular Probes (Eugene, OR). HRP-conjugated streptavidin was purchased from Amersham Pharmacia Biotech (Arlington Heights, IL). The HRP-based chemiluminescence ECL kit was purchased from Amersham Pharmacia Biotech (Little Chalfont, U.K.).
Receptor cross-linking and lipid raft isolation
Mouse splenic B cells were purified as previously described (29), surface-biotinylated using sulfo-N-hydroxysuccinimide long-chain biotin (Pierce) as previously described (30), and incubated in the presence or the absence of (Fab′)2 rat Ab specific for mouse IgM and rat mAb specific for CD19 (1D3) or rat mAb specific for CD21 (7G6) for 30 min at 4°C, followed by co-cross-linking with goat Abs specific for rat Ig. Alternatively, biotinylated CD81-expressing and -deficient cells transgenic for the HEL-specific BCR were coligated using the recombinant Ag, HEL-C3d (10 μg/ml), as previously described (13). The BCR on Daudi cells was cross-linked by incubating the cells (1 × 108 in 1 ml) at 4°C for 30 min with biotinylated Fab of mouse Abs specific for human Igμ DA4.4 (10 μg/ml), followed by avidin (5 μg/ml) cross-linking. Coligation of the BCR to the CD19/CD21 complex was performed using biotinylated Fab-DA4.4 and biotinylated (Fab′)2 ADF4.2 mouse Abs specific for human CD19 at 4°C for 30 min, followed by avidin co-cross-linking. Coligation of the BCR to the chimeric CD4/19 receptor was similarly performed using biotinylated DA4.4 and biotinylated anti-human CD4, followed by avidin co-cross-linking. All cells were lysed in 1% Triton X-100 (Sigma-Aldrich) containing lysis buffer, and lipid rafts were isolated from the lysates by sucrose density centrifugation as previously described (30).
Immunoprecipitation and immunoblotting
Mouse and human Igμ, Igα, CD81, Lyn, mouse CD21, and CD45 and human CD19 and CD4/19 were detected in fractions from sucrose density gradients by SDS-PAGE and immunoblotting probing with specific Abs detected by HRP-conjugated goat Abs specific for the primary Abs. HRP-conjugated PY20 was used to detect tyrosine-phosphorylated proteins. When indicated, blots were stripped of the detecting Abs by incubation in stripping buffer (100 mM 2-ME, 2% SDS, and 62.5 mM Tris-HCl, pH 6.7; Sigma-Aldrich) at 50°C for 30 min and were reprobed using different Abs. Mouse CD19 was detected by immunoprecipitation from solubilized pooled sucrose gradient fractions prepared from biotinylated cells using mAb 1D3 or isotype control rat IgG2a and analysis by SDS-PAGE and immunoblot probing with HRP-conjugated streptavidin. Phosphorylated Vav, Igα, mouse CD19, and PLCγ2 were detected by immunoprecipitation from pooled rafts and non-rafts from sucrose gradient fractions solubilized in 0.5 ml of 5× RIPA lysis buffer (50 mM Tris-HCl, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, and 0.1% SDS, pH 7.5). Lysates were precleared with protein A-Sepharose (Amersham Pharmacia Biotech, Uppsala, Sweden) and normal rabbit serum (Sigma-Aldrich), followed by immunoprecipitation with 5 μg/ml specific rabbit Ab or rabbit IgG isotype control at 4°C with rotation overnight. Immunoprecipitates were analyzed by SDS-PAGE and immunoblot probing for tyrosine-phosphorylated proteins using HRP-conjugated PY20. CD81 and CD21 were analyzed by SDS-PAGE under nonreducing conditions to preserve the antigenic epitopes. Blots were developed using ECL (Amersham Pharmacia Biotech), and all films were quantified at the same exposure time by densitometry using Scion Image software (National Institutes of Health, Bethesda, MD).
Immunofluorescence and confocal microscopy
Purified splenic B cells from cd81+/+ and cd81−/− mice were incubated with Cy2-conjugated Fab goat Abs specific for mouse Igμ for 20 min at room temperature, followed by staining with DiIC16 for 5 min at room temperature before washing and transferring to poly-l-lysine-coated slide chambers (Labtek/Nunc, Naperville, IL). Cells were allowed to adhere to the slide for 60 min on ice before warming to room temperature. Cells (1.25 × 106) were incubated with 50 μg/ml (Fab′)2 goat Abs specific for mouse IgM and IgG alone or together with 50 μg/ml goat Abs specific for mouse CD21 at room temperature for various intervals (2–30 min). Primary Abs were cross-linked with rabbit Abs specific for goat Ig. Cells were fixed with 4% paraformaldehyde, washed and permeabilized in 0.05% saponin containing PBS buffer, and blocked with 1% BSA for 30 min at room temperature. Samples were incubated with the biotin-labeled, phosphotyrosine-specific mAb, 4G10, for 30 min at room temperature, followed by incubation with streptavidin conjugated to AlexaFluor 647 for 30 min.
For two-color confocal and differential interference contrast image, a confocal laser scanning microscope (Axiovert 200M LSM 510 META; Zeiss, Jena, Germany) fitted with a 1.4 oil Plan-Apochromat ×63 objective was used, and images were acquired using channel mode, single-track acquisition with the main beam splitters HFT UV/488/543/633 for the excitation of Cy2, DiIC16, and AlexaFluor 647. Filter sets of BP 505–530 for Cy2, BP 560–615 for DiIC16, and LP 650 for AlexaFluor 647 were used. Parameters were adjusted to yield scan control of fixed pixel density at 512 × 512 pixels, 8-bit pixel depth, linear contrast of grayscale, 45-ms scan time, and pinhole size of 109 μm (DiIC16), 96 μm (BCR), or 109 μm (PTyr). Cell populations were scanned first for BCR and DiIC16, then the same fields were scanned a second time for BCR and PTyr. No significant signal saturation was noted in any of the images used for analysis. For colocalization scatter graphs between BCR-DiIC16 and BCR-PTyr, LSM 510 imaging examiner software was used. To compare the intensity of BCR colocalized with DiIC16 or PTyr, the images from three randomly chosen fields containing at least 100 cells were acquired as grayscale with linear contrast. The number of BCR pixels of a mean fluorescence intensity above a threshold of 1500 U colocalizing with either DiIC16 or PTyr was derived from the examiner software and expressed as colocalization coefficients, with values ranging from 0 to 1 (0 = no colocalization, 1 = all pixels colocalize).
Coligated CD19/CD21-BCR complexes fail to partition into lipid rafts in cd81−/− B cells
The ability of the CD19/CD21 complex to promote BCR raft association was analyzed in cd81−/− B cells and wild-type littermates (cd81+/+). The BCR and the CD19/CD21 complex were coligated on B cells from cd81+/+ and cd81−/− mice using rat Abs specific for the BCR and for CD19 cross-linked with goat Abs specific for rat Ig (Fig. 1,A). Cells were incubated with the primary Abs at 4°C, followed by incubation with the cross-linking Ab at 4°C, and the rafts were isolated as previously described (13) by lysing cells in 1% Triton X-100 detergent at 4°C, conditions under which lipid rafts are insoluble. The time course was previously shown to result in maximal partitioning of the BCR into rafts and signaling from rafts (13, 31). The lysates were applied to sucrose density gradients, and fractions from the sucrose gradients were analyzed by immunoblotting for the presence of CD19, CD21, CD81, Igμ, Igα, the Src kinase Lyn, and the phosphatase CD45. The insoluble membranes float at the top of the sucrose gradient in fractions 3–6, and the soluble material sediments at the bottom of the gradients in fractions 10–12. B cells from cd81−/− and cd81+/+ mice contain insoluble membranes that concentrate Lyn and exclude CD45 that operationally define lipid rafts (Fig. 1). Thus, CD81 does not appear to be required for the integrity of lipid rafts in B cells. In addition, in B cells from cd81+/+ mice, CD81 as well as CD19, CD21, and the BCR were contained in the soluble fraction (data not shown), indicating that CD81 is not constitutively present in lipid rafts. In B cells from cd81+/+ mice, coligation of CD19 with the BCR resulted in the association of the majority of CD19, CD81, CD21, Igμ, and Igα with lipid rafts (Fig. 1,A). In contrast, coligation of CD19 and the BCR on B cells from cd81−/− mice had little effect on the position of CD19, CD21, or the BCR Igμ and Igα chains in the plasma membrane (Fig. 1 A). CD19, CD21, and Igα were undetectable in the raft fractions, and only a small amount of the Igμ chain was detected in the rafts of cd81−/− cells. Thus, the coligation of CD81-deficient CD19/CD21 complexes to the BCR appeared to inhibit the BCR’s association with rafts.
B cells from cd81−/− mice express ∼30–50% of the wild-type levels of CD19, but express wild-type levels of CD21 and the BCR (16, 17, 19). To ensure that the failure to induce the association of the BCR and the CD19/CD21 complex with lipid rafts was not due to a failure to adequately engage the CD19/CD21 complex, the CD19/CD21 complex and the BCR were coligated using rat Abs specific for CD21 and the BCR, cross-linked with goat Abs specific for rat Ig (Fig. 1,B). Coligating the BCR and CD21 resulted in the translocation of the majority of CD19, CD21, CD81, Igμ, and Igα into the insoluble, Lyn-containing fractions in B cells from cd81+/+ mice (Fig. 1 B). In contrast, coligation of CD21 and the BCR on B cells from cd81−/− mice had little effect on the position of either the CD19/CD21 complex or the BCR in the plasma membrane. Indeed, nearly all the CD19/CD21 complex and the BCR remained in the soluble fractions after coligation in B cells from cd81−/− mice.
The association of the BCR and CD19/CD21 complex with lipid rafts was also investigated in B cells from cd81−/− mice that expressed the transgene for an HEL-specific BCR (21) and from mice that expressed the HEL-specific BCR transgene but are cd81+/+ (Fig. 1,C). The CD19/CD21 complex was coligated to the HEL-specific BCR using the recombinant Ag, HEL-C3d, containing HEL fused to three tandemly arranged copies of C3d, the ligand for CD21, as previously described (13). HEL-C3d has been demonstrated previously to be a highly potent immunogen in vivo (12) and to induce raft association of the BCR, CD19, and CD21 in vitro (13). Incubation with HEL-C3d resulted in the association of CD19, CD21, CD81 and the BCR Igμ and Igα with lipid rafts in B cells from cd81+/+ mice expressing the HEL-specific BCR (Fig. 1 C). In contrast, B cells from cd81−/− mice expressing the HEL-specific BCR failed to translocate CD19 and CD21 into lipid rafts. Compared with B cells from cd81+/+ mice, B cells from cd81−/− mice also showed significantly reduced association of BCR Igμ and Igα with the rafts following incubation with HEL-C3d. Taken together, these results provide evidence that CD81 plays an essential role in the stable association of the coligated CD19/CD21 complex and the BCR with lipid rafts.
CD81 deficiency did not appear to influence the behavior of the BCR when simply cross-linked to itself (Fig. 2). Cross-linking the BCR using Ig-specific Abs resulted in the partitioning of a similar portion of BCR Igμ and Igα into lipid rafts in cd81+/+ and cd81−/− B cells (Fig. 2). As previously reported, BCR cross-linking did not influence the position of CD21 (Fig. 2) or CD19 (data not shown) in the membrane, both of which remained in the soluble fractions. More BCR appeared to partition into lipid rafts in cd81−/− B cells following BCR cross-linking alone (Fig. 2) vs coligation of the BCR to the CD19/CD21 complex (Fig. 1), suggesting that coligation of the CD81-deficient CD19/CD21 complex to the BCR not only failed to promote raft partitioning of the BCR, but was detrimental to partitioning.
Failure of CD19/CD21-BCR complexes to partition into rafts results in reduced signaling from rafts
Our previous studies showed that the partitioning of coligated BCR-CD19/CD21 complexes into rafts resulted in enhanced signaling from rafts (13). The signaling activity of the raft and soluble fractions from B cells from cd81+/+ and cd81−/− mice was assessed. After coligation of the BCR and the CD19/CD21 complex,the cells were solubilized, and the raft and soluble fractions were separated by sucrose density gradient centrifugation. The fractions were subjected to SDS-PAGE, immunoblotted, and probed for tyrosine-phosphorylated proteins. After coligation of the CD19/CD21 complex and the BCR, the raft fractions from cd81+/+ B cells contained a large number of tyrosine-phosphorylated proteins (Fig. 3). A number of tyrosine-phosphorylated proteins were also isolated in the soluble fraction. In contrast, the raft fractions from cd81−/− B cells showed no detectable tyrosine-phosphorylated proteins in the raft fractions, and fewer tyrosine-phosphorylated proteins in the soluble fraction compared with cd81+/+ B cells. To identify specific BCR/CD19 signaling components, the sucrose gradient fractions were subjected to immunoprecipitation using Abs to Igα, CD19, Vav, and PLCγ2. The immunoprecipitates were analyzed by SDS-PAGE, immunoblotted, and probed for phosphotyrosine. The rafts obtained from cd81+/+ B cells after coligation of BCR and the CD19/CD21 complex contained tyrosine-phosphorylated Igα, CD19, Vav, and PLCγ2, indicating that these signaling components were recruited to the lipid rafts after coligation (Fig. 3). In contrast, the B cells from cd81−/− mice showed no detectable tyrosine phosphorylation of Igα or CD19 in either the raft or soluble fractions. Both Vav and PLCγ2 were detectably phosphorylated in cd81−/− B cells, but the phosphoproteins were not recruited to the lipid rafts (Fig. 3). Thus, the CD81-deficient CD19/CD21 complex, when coligated to the BCR, fails to promote BCR signaling from lipid rafts.
BCR cross-linking alone in cd81−/− B cells resulted in the appearance of phosphorylated Igα in the raft fractions that was only slightly reduced by comparison with that observed in cd81+/+ B cells (Fig. 2). In both cd81−/− and cd81+/+ B cells, CD19, Vav, and PLCγ2 were phosphorylated, but were not recruited to rafts. The phosphorylation of CD19 after BCR cross-linking was weaker in cd81−/− compared with cd81+/+ B cells, which may reflect the reduced expression of CD19 by cd81−/− B cells. After BCR cross-linking alone, Igα was phosphorylated in cd81−/− B cells, whereas phosphorylated Igα was not detected after coligation of the BCR to the CD19/CD21 complex (Fig. 3), suggesting that coligation of the CD81-deficient CD19/CD21 complex was detrimental to BCR signaling.
Coligated BCR-CD19/CD21 complexes in cd81−/− B cells fail to polarize and colocalize with lipid rafts
The results presented above provided evidence that CD81 was essential for the partitioning of coligated BCR-CD19/CD21 complexes into lipid rafts and signaling from rafts by the criteria of detergent solubility. To determine whether intact cells require CD81 for the association of the coligated BCR-CD19/CD21 complexes with lipid rafts, laser scanning microscopy was used. B cells from cd81+/+ and cd81−/− mice were labeled with the raft-partitioning fluorescent lipid dye DiIC16 to visualize lipid rafts (32) and with Cy2-conjugated Fab anti-Igμ to label the BCR. The cells were either untreated or treated with Abs to BCR and CD21 to coligate the BCR and the CD19/CD21 complex, as described above, for various lengths of time, fixed, and analyzed by confocal microscopy using a Zeiss META 510 system. To visualize the phosphotyrosine-containing proteins, the fixed cells were permeabilized and incubated with a biotin-conjugated, phosphotyrosine-specific mAb detected using AlexaFluor 647-conjugated streptavidin. The cells were analyzed in a pairwise fashion for colocalization of the BCR with DiIC16 and of the BCR with phosphotyrosine-containing proteins. Shown is a field of cells scanned for BCR and DiIC16 and images of individual representative cells scanned for the BCR, DiIC16, and phosphotyrosine (Figs. 4 and 5). In unligated B cells from both cd81+/+ and cd81−/− mice, the lipid raft dye DiIC16 and the BCR appeared uniformly distributed in the plasma membrane (Figs. 4 and 5). The merged images showed no colocalization of the BCR and DiC16. Phosphotyrosine staining was weak, and a considerable portion of the phosphotyrosine-containing proteins was present in the cytoplasm. The merged images showed little colocalization of the BCR and phosphotyrosine. Upon cross-linking of the BCR alone, in B cells from both cd81+/+ and cd81−/− mice, the BCR formed patches that colocalized with the lipid raft DiIC16 marker (Figs. 4 and 5). The staining for phosphotyrosine-containing proteins increased, and these were almost exclusively plasma membrane associated. A portion of the phosphotyrosine-containing proteins colocalize with BCR patches. However, coligation of the BCR and the CD19/CD21 complex in cd81+/+, but not in cd81−/−, B cells resulted in the formation of polarized caps of the BCR that colocalize with the DiIC16 raft marker and with phosphotyrosine-containing proteins (Figs. 4 and 5). These signaling-active caps persist for >30 min. In cd81−/− B cells, coligating the BCR to the CD19/CD21 complex resulted in very little change in their plasma membrane expression compared with that of the unligated controls (Fig. 5).
The number of B cells that formed BCR patched or capped structures after BCR cross-linking was similar in cells from cd81−/− and cd81+/+ mice (Fig. 6 A). In contrast, significantly fewer cd81−/− B cells formed caps after coligation of the BCR and the CD19/CD21 complex compared with cd81+/+ B cells. Moreover, few cd81−/− B cells formed patched structures after BCR and CD19/CD21 coligation, suggesting that the coligation of the CD81-deficient CD19/CD21 complex was detrimental to an early step in the process.
The degree of colocalization of the BCR with the DiIC16 raft marker and with phosphotyrosine-containing proteins was quantified for cd81+/+ and cd81−/− B cells pixel-by-pixel for three randomly selected fields, each containing ∼100 cells (Fig. 6,B). For each pixel that contained BCR fluorescence intensity above a threshold of 1500 U, the amounts of DiC16 and phosphotyrosine were determined. After BCR cross-linking, the degree of colocalization of the BCR with DiIC16 and with phosphotyrosine-containing proteins was similar in cd81+/+ and cd81−/− B cells (Fig. 6,B). In contrast, after BCR and CD21 coligation, cd81−/− B cells showed significantly less colocalization of the BCR with either DiC16- or phosphotyrosine-containing proteins compared with cd81+/+ B cells (Fig. 6 B). These results show that the CD81-deficient CD19/CD21 coreceptor complex fails to function to stably partition the coengaged BCR into signaling lipid rafts. These results are in agreement with those from the experiments described above using detergent solubility to identify rafts.
A chimeric CD19 protein that does not associate with CD81 is a poor facilitator of raft association
The requirement of CD81 for the association of the CD19/CD21 complex with rafts was further explored using a Daudi B cell line that expresses a chimeric CD19 protein that fails to associate with CD81. The Daudi B cell line analyzed expressed, in addition to the wild-type endogenous CD19, a chimeric CD19 receptor in which the ectodomain of CD19 was replaced by CD4 (CD4/19) (22). Previous studies showed that CD81 associates with the CD19/CD21 complex primarily through interactions with the ectodomain of CD19 and that the CD4/19 chimeric protein does not coimmunoprecipitate with CD81 (22, 33). Thus, by the criteria of coimmunoprecipitation, CD81 associates poorly if at all with CD19; however, a weak association between the two cannot be ruled out.
The surface expression of the transfected CD4/19 was similar to that of the endogenous wild-type CD19 and CD21 proteins as measured by flow cytometry using specific Abs (Fig. 7,A). Moreover, quantification of the number of CD4- vs CD19-specific Abs bound per cell using the BD Quantibrite system showed that ∼2-fold more CD4-specific Abs bound to the transfected Daudi cells compared with CD19-specific Abs (S. Brooks and R. Carter, unpublished observations). The CD4/19-expressing cells also expressed CD81 in an amount similar to untransfected Daudi cells (data not shown). To test the ability of the CD4/19 chimeric protein to facilitate the association of BCR with lipid rafts, the chimeric CD4/19 protein and the BCR were coligated using biotinylated CD4- and Ig-specific Abs cross-linked with avidin, and lipid rafts were isolated. In the absence of coligation, CD4/19, CD81, and BCR Igμ and Igα chains resided outside of lipid rafts that contained Lyn and excluded CD45 (Fig. 7,B). Coligating CD4/19 and the BCR resulted in the association of <20% of the BCR and CD4/19 with lipid rafts and an association of only 15% of CD81 (Fig. 7,B) as quantified by densitometry. In contrast, coligation of the endogenous wild-type CD19 to the BCR in CD4/19-transfected cells resulted in the association of 50% of CD19 and 45% of CD81 with rafts. The association of the BCR with lipid rafts when coligated to the chimeric CD4/19 protein (5–10% of Igμ and Igα) was less efficient than that achieved by cross-linking the BCR to itself (35% of Igμ and 80% of Igα; Fig. 7). This suggests that the CD4/19 chimeric protein when coligated to the BCR may negatively influence the stability of the BCR in rafts similar to the observations in the B cells from cd81−/− mice.
The reduced ability of the CD4/19-BCR complex to stably associate with rafts after coligation was reflected in a dramatic reduction in the tyrosine phosphorylation of Igα and Lyn in the lipid raft fractions compared with the coligation of the wild-type CD19 complex with the BCR in the same cells (Fig. 7,B). Significantly, coligation of the CD4/19 chimera and the BCR resulted in phosphorylation of CD19, Vav, and PLCγ2, but the phosphorylated proteins did not become raft associated as they did after coligation of wild-type CD19 and the BCR (Fig. 7 B). Thus, the chimeric CD4/19 protein that does not associate with CD81 failed to assemble a signaling complex and prolong signaling in lipid rafts.
Recent evidence indicates that lipid raft membrane microdomains play a key role in immune cell receptor signaling by providing a mechanism to laterally segregate membrane proteins on the cell surface (14, 15). In resting cells rafts are estimated to be small, highly dynamic structures containing only thousands of molecules (34, 35). In resting B cells, lipid rafts concentrate Lyn and exclude most membrane proteins, including the BCR and certain negative regulators of BCR signaling, such as CD45 and CD22. Upon Ag binding and oligomerization, the BCR partitions into lipid rafts by a mechanism that does not require a signaling-competent BCR or association of the BCR with an intact actin cytoskeleton (31, 36). However, for the BCR to stably reside in rafts it appears that both signaling and cytoskeleton attachment are necessary (P. C. Cheng and S. K. Pierce, unpublished observations). It has been proposed that signaling and attachment to the cytoskeleton may be required to promote the aggregation or clustering of rafts into larger, more stable structures that function as signaling platforms (14, 15). Recently, evidence was provided that the coreceptor CD19/CD21 complex, when coligated to the BCR through the binding of complement-coupled Ags, served to prolong the residency of the BCR in rafts and its signaling from rafts (13). The ability of the CD19/CD21 complex to facilitate BCR-raft association may represent a novel mechanism by which a coreceptor functions to regulate receptor signaling. In this study we asked whether the tetraspanin CD81 plays a role in the raft-stabilizing function of the CD19/CD21 complex.
CD81 is a member of the large tetraspanin superfamily (2). Because tetraspanins associate with a variety of protein complexes and are involved in a remarkable number of diverse cellular activities, they have been proposed to function as molecular facilitators of the protein complexes with which they associate (1). Tetraspanins in complex with other membrane proteins have previously been described in rafts in nonlymphoid cells, although the complexes themselves were not dependent on membrane cholesterol (37). The results presented in this study demonstrate that in wild-type B cells, CD81 partitions into lipid rafts upon coengagement of BCR and the CD19/CD21 complex. The association of the BCR and CD19/CD21 complexes with lipid rafts is CD81 dependent, and in the absence of CD81, there is a dramatic reduction in the association of the BCR and the CD19/CD21 complex with lipid rafts upon coligation. The results described in this study also provide evidence that CD81 is necessary for the assembly in rafts of a BCR-CD19/CD21 signaling complex, including Vav and PLCγ2. Coligating the BCR and the CD19/CD21 complex in cd81−/− B cells resulted in phosphorylation of Vav and PLCγ2, but these were not recruited to rafts. The overall magnitude of tyrosine phosphorylation of proteins was lower in the cd81−/− B cells.
Evidence for colocalization of the BCR with lipid rafts after cross-linking and coligation to the CD19/CD21 was also provided in this study by confocal microscopy of intact cells. The results of this analysis provided evidence that the BCR colocalized with lipid rafts in patches and caps after cross-linking and coligation in cd81+/+ B cells. The BCR-raft structures also appear to be signaling-active and to concentrate tyrosine-phosphorylated proteins. B cells from cd81−/− mice formed signaling-active patched structures that colocalized with the raft marker after BCR cross-linking alone, but formed neither patches nor caps after coligation of the BCR and the CD19/CD21 complex. These results provide important information concerning the requirement for CD81 in the patching and capping of the BCR and are also significant for providing an independent measure of the association of the BCR with lipid rafts that did not depend on disrupting the cells with detergents.
Although there is a clear phenotypic difference in the ability of coligated BCR-CD19/CD21 complexes to stably associate with lipid rafts in cd81+/+ vs cd81−/− B cells, the mechanism by which CD81 functions remains to be elucidated. The reduced expression of CD19 on cd81−/− B cells could contribute to the phenotype. However, even though cd81−/− B cells express as much as 50% of the levels of CD19 expressed by cd81+/+ B cells, virtually none of the CD19 in the cd81−/− B cells partitioned into rafts (Fig. 1). In addition, in Daudi B cells that express similar levels of both a wild-type CD19 that associates with CD81 and a chimeric CD19 receptor that contains the extracellular domain of CD4 (CD4/19) and does not associate with CD81 (22), coligation of the BCR to the wild-type CD19, but not to the CD4/19 chimera, resulted in stable association with rafts. It is also possible that CD81 is required for the proper assembly of CD19/CD21 complexes that allows their stable partitioning into lipid rafts. However, because CD19 and CD21 pair through their extracellular and transmembrane domains independently of CD81 pairing to CD19 (22), CD19/CD21 complexes would be predicted to form in the cd81−/− B cells.
The structural characteristics of the tetraspanin CD81 that allow it to facilitate the association of BCR and the CD19/CD21 complex with lipid rafts also remain to be determined. Although the characteristics of integral membrane proteins that allow their association with rafts is not known, several observations have highlighted the importance of the transmembrane regions (31, 38, 39). Indeed, the tetraspanins have highly conserved residues in their transmembrane domains that are likely to be important for their function (2, 3). In addition, for most membrane proteins that reside constitutively in rafts, acylation of the protein has been shown to be essential (40, 41). As recent studies have demonstrated that CD81 is palmitoylated (42, 43), we speculate that the association of CD81 with rafts may involve a reversible palmitoylation event following coreceptor ligation. Lastly, the crystal structure of the LEL of CD81 showed it to be mushroom-like, composed of five α helixes arranged into head and stalk domains (44). Sequence analysis of a large number of tetraspanins indicated that the key structural features of the LEL, including a hydrophobic interface, are conserved (45). This and the fact that the CD81 LEL crystallized as a dimer leads to the proposal that tetraspanins may assemble into dimers or higher multimers at the cell surface and guide the clustering of tetraspanin-associated proteins. The formation of such multimers may promote partitioning of the complexes into lipid rafts and clustering of the lipid rafts.
In summary, the results presented in this study point to a significant role for a member of the tetraspanin family, CD81, in the function of the coreceptor CD19/CD21 complex in amplifying and prolonging BCR signaling from lipid rafts. As lipid rafts have been implicated to play a role in signaling through a variety of cell surface protein complexes, it is possible that the ability to facilitate raft association is a common feature of members of the tetraspanin family. Thus, tetraspanins may function as molecular facilitators in part by promoting raft association.
We are grateful to Dr. Tian Jin (National Institute of Allergy and Infectious Diseases, National Institutes of Health) for his expert advice and assistance with the confocal microscopy analysis, and to Dr. Michael Holers (University of Colorado Health Sciences Center) for his generosity in providing the CD21-specific reagents.
This work was supported in part by National Institutes of Health Grants AI42265 (to R.C.) and CA34233 and AI45900 (to S.L.). R.C. was supported by the Medical Research Service, Department of Veterans Affairs, and S.B. was supported by National Institutes of Health Training Grant AI07051.
Abbreviations used in this paper: LEL, large extracellular loop; BCR, B cell Ag receptor; HEL, hen egg lysozyme; PLCγ2, phospholipase Cγ2.