Previous Article | Next Article ![]()
Molecular and Cellular Biology, February 2006, p. 1373-1385, Vol. 26, No. 4
0270-7306/06/$08.00+0 doi:10.1128/MCB.26.4.1373-1385.2006
Copyright © 2006, American Society for Microbiology. All Rights Reserved.
Division of Oncology, Department of Medicine, Stanford University Medical Center, Stanford, California,1 Department of Human Microbiology, Tel Aviv University Medical School, Tel Aviv, Israel2
Received 10 September 2005/ Returned for modification 26 October 2005/ Accepted 15 November 2005
|
|
|---|
|
|
|---|
Other members of the tetraspanin family are also involved in numerous cellular responses (19). These activities are due to the association of different tetraspanins with each other and with other tissue type-specific proteins in the cell membrane in the form of tetraspanin-enriched microdomains (TEM), also known as the tetraspanin web (39). The concept of the web was suggested because antibodies that reacted with different tetraspanin molecules induced similar biological effects and coimmunoprecipitated one another (27). Additional biochemical and microscopic studies have shown that the web is composed of several tetraspanin molecules that associate laterally with each other and with nontetraspanin partners (1, 39, 41, 48, 60). The composition of the web is different in each cell type, where partnerships are formed with cell surface receptors, adhesion molecules, and transmembrane signaling proteins (19). More recently, a growing list of intracellular signaling components, such as Rho-GTP, (30, 59, 61), were shown to be linked to the tetraspanin web. This suggests that the tetraspanin web mediates the cross talk between cell surface stimuli and intracellular signaling pathways, allowing cells to respond to a constantly changing environment in a specific and a highly regulated manner (28). However, it is not known how tetraspanins form such multiple, dynamic, yet specific interactions. To address this question we studied the interaction between CD81 and its B-cell-specific signaling partner, CD19 (3, 21, 26).
CD19 is a membrane glycoprotein and a member of the immunoglobulin superfamily. It is coexpressed with CD81 in B cells from the earliest stages of development (43) and is lost only upon terminal differentiation to plasma cells (42). In mature B cells, CD19 associates with complement receptor 2 (CR2/CD21) (33) and is pivotal for signal transduction induced upon coligation of the B-cell receptor (BCR) and the CD21/CD19 complexes by complement-opsonized antigen (10). CD19 consists of two N-terminal extracellular immunoglobulin domains, a transmembrane domain, and a long cytoplasmic tail, which contains conserved tyrosine residues that are required for signal transduction (6, 40).
CD81 is expressed on all primary B-lineage cells, from the early stages of B-cell development in the bone marrow (43); it associates with CD19 on the cell surface (3, 21, 26). Moreover, it is required for normal expression of CD19, as demonstrated with three independently derived knockout lines (31, 35, 50). The effect of CD81 on CD19 expression is specific to CD81, because B cells from other tetraspanin knockout mice, such as CD9 (43), CD37 (24), and Tssc6 (49), express wild-type (WT) levels of CD19. A detailed analysis of CD19 expression in Cd81/ primary B cells revealed that the most affected subset is the small pre-BII, which is almost devoid of CD19 (43). The fact that the levels of CD19 mRNA are normal, even in Cd81/ pre-BII cells, indicates that CD81 exerts its effect on CD19 at a posttranscriptional level (43).
The introduction of the human CD81 (hCD81) gene into Cd81/ primary B cells completely restores CD19 cell surface expression, indicating that the impairment in CD19 expression is caused by the lack of CD81 in these cells (43). Here we used this "add-in" approach to follow CD19 during biosynthesis and trafficking to the cell surface. We also used this approach to probe the function of the CD81/CD19 complex on the cell surface. The concept that emerges from these studies is that distinct CD81 domains act in different cellular compartments. Defining how CD81 regulates CD19 is a model aimed at understanding how tetraspanins and their partners assemble and function in the tetraspanin web.
|
|
|---|
Mouse Cd81/ and WT precursor B-cell lines. Early B-cell lines were generated by retroviral transduction of the BCR-ABL gene (45) into early B-cell colonies developed in bone marrow lymphopoietic cultures, which were generated from primary bone marrow cells of Cd81/ or WT littermate BALB/c mice as previously described (54). Transformed cells survived without stromal support and were subcloned at limiting dilutions, from which the single-cell-cloned 1C8 (Cd81/) and 2F3 (WT) cell lines were derived (see Fig. S1 in the supplemental material).
Flow cytometry and MAbs. Cells were incubated with various combinations of fluorochrome- or biotin-conjugated monoclonal antibodies (MAbs). Biotin-conjugated reagents were counterstained with streptavidin-phycoerythrin (PE) (BD Immunocytometry Systems, San Jose, CA), and stained cells were analyzed using FACSCalibur (BD Immunocytometry Systems). Unless indicated otherwise, MAbs used for immunofluorescent staining were purchased from BD PharMingen (San Diego, CA): PE-anti-CD45R/B220 (RA3-6B2), PE-anti-CD19 (1D3), fluorescein isothiocyanate-anti-CD43 (S7), PE-anti-CD25 (7D4), PE-anti-hCD81 (JS-81), PE-anti-hCD9 (M-L13), and unconjugated anti-hCD9 (M-L13). The EZ-link sulfo-NHS biotinylation kit (Pierce Biotechnology, Inc., Rockford, IL) was used to biotinylate anti-hCD9 (clone 50H.19) (a generous gift from Andrew R. Shaw, University of Alberta, Alberta, Canada) and anti-hCD81 (clone 5A6) MAbs.
Metabolic labeling. Cells were seeded at 1.8 x 107 cells/ml in cysteine- and methionine-free RPMI 1640 medium (Amersham Life Science, Piscataway, NJ) supplemented with 10% fetal calf serum, 50 µM 2-mercaptoethanol, 20 mM HEPES, sodium pyruvate, and nonessential amino acids and incubated at 37°C for 1.5 h. Cells were pulsed for 30 min with 100 µCi/ml of L-[35S]methionine and L-[35S]cysteine (>1,000 Ci/mmol; Amersham Life Science), washed, and incubated (at 1.8 x 106 cells/ml) for increasing chase periods as indicated. Harvested cells were lysed in calcium- and magnesium-free phosphate-buffered saline (PBS) containing 1% NP-40 (BDH Laboratory, Poole, United Kingdom) and complete ETDA-free protease inhibitor mixture (Roche Diagnostics, Mannheim, Germany). Lysates were preabsorbed with 60 µl of a 50% (vol/vol) slurry of Sepharose-protein G (Sigma-Aldrich) overnight at 4°C and then immunoprecipitated with 20 µl of a 50% slurry of protein-G Sepharose beads prebound with 10 µg of anti-CD19 (1D3) or a control MAbs overnight at 4°C. The precipitated material was washed extensively, eluted in loading buffer, and then analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), followed by autoradiography of dried gels on a phosphorimager (Molecular Imager FX; Bio-Rad Laboratories, CA).
Chimeric hCD81/hCD9. hCD81/hCD9 chimeric molecules were constructed using a PCR-based method. PCR was performed on the hCD9 cDNA template (a gift from C. Boucheix and E. Rubinstein, Villejuif, France) or on the hCD81 cDNA template in separate tubes. This reaction incorporated an EcoRI site at the 5' end of one template and an XhoI site at the 3' end of the other template. The second primer in each reaction incorporated a 12-bp sequence of the paired cDNA at the junction site. These products were mixed, extended via the overlapping region, and amplified with primers containing the restriction sites. PCR products were cloned into a pCR-Blunt or pCR-Blunt II-TOPO vector (Invitrogen, Carlsbad, CA), and inserts were cleaved with EcoRI and XhoI restriction enzymes and ligated into the multicloning site of the retroviral vector.
Ecotropic viral production and cells infection. The cDNAs encoding hCD81, hCD9, or chimeric hCD81/hCD9 were cloned into the retroviral plasmid pBabeMN IRES-GFP (http://www.stanford.edu/group/nolan/retroviral_systems/phx.html). The vector contains an internal ribosome entry site (IRES) and was provided by G. Nolan (Stanford University). The tetraspanin cDNAs were inserted 5' of the IRES, followed by green fluorescent protein (GFP) cDNA. Ecotropic retroviruses were produced in the Phoenix (fNX)-Eco packaging cell line transfected with pBabe encoding tetraspanin cDNA/GFP or pBabe GFP, using the Gene PORTER 2 transfection reagent (Gene Therapy Systems, San Diego, CA) or the FuGENE 6 transfection reagent (Roche Diagnostics Corp, Indianapolis, IN) to infect the Cd81/ or the WT cell lines. Cells were plated at 0.5 x 106 cells/200 µl per well in 24-well plates, and 800 µl of retrovirus-containing supernatant with Polybrene at a final concentration of 1 µg/ml was added and spin infected as described previously (43). Following 72 h of incubation, a fraction of the infected cells was analyzed by flow cytometry. The rest of the cells were expanded, sorted by flow cytometry, and subcloned at limiting dilutions.
Immunoprecipitation. Cells (2 x 107 cells/ml) were lysed in calcium- and magnesium-free PBS containing 1% NP-40 or in 10 mM Tris (pH 7.4)-150 mM NaCl containing 1% Brij 97 (Sigma-Aldrich), both containing complete ETDA-free protease inhibitor mixture (Roche Diagnostics). Lysis in Brij 97 was supplemented with 1 mM CaCl2 plus 1 mM MgCl2 or with 1 mM EDTA, as indicated. Immunoprecipitation by anti-CD19 (1D3) MAb (BD PharMingen) or anti-hCD81 MAb (5A6) was performed as described above.
Enzymatic digestion. Immunoprecipitated molecules were eluted in denaturation buffer and divided into three fractions: untreated, digested with peptide N-glycosidase F (endo-F), or digested with endoglycosidase H (endo-H). Endo-F digestion of total N-glycan chains was performed using the 704S kit (New England Biolabs, Beverly, MA), and endo-H cleavage of high-mannose structures was performed using the PO703S kit (New England Biolabs). Sample buffer was added, and the digested proteins were analyzed by SDS-PAGE followed by immunoblotting.
Gel electrophoresis and immunoblotting. Proteins were separated by 10% or 12% SDS-PAGE and electrotransferred onto polyvinylidene difluoride membranes. Membranes were incubated with rabbit polyclonal anti-mouse CD19 serum (a generous gift from J. Cambier, Denver, CO) diluted 1:1,000, with anti-hCD81 (5A6) at 3.5 µg/ml, with anti-hCD9 (M-L13) (BD PharMingen) at 3 µg/ml, or with rabbit antiactin (I-19) (Santa Cruz Biotechnology, CA). Secondary antibodies were donkey anti-rabbit horseradish peroxidase-linked secondary Abs (1/10,000; Amersham Life Science) and goat anti-mouse immunoglobulin G (IgG) (Fc specific, horseradish peroxidase linked) (1/5,000; Sigma-Aldrich). Blots were visualized by chemiluminescence detection (ECL; Amersham, Little Chalfont, United Kingdom). Densitometry analysis was performed using the IPLab image analysis software (Scanalytics, Fairfax, VA).
Surface molecule cross-linking and annexin V binding. Purified 1C8 cells expressing recombinant tetraspanin were washed in cold calcium- and magnesium-free PBS and incubated with 1 µg/106 cells of either biotinylated anti-hCD9 MAb (50H.19), biotinylated anti-hCD81 (5A6) MAb, or matched biotinylated isotype controls for 15 min on ice. Cells were washed in cold calcium- and magnesium-free PBS for 5 min at 4°C, cross-linked with 1 µg/106 cells unconjugated streptavidin (Pierce Biotechnology, Inc.) for 15 min on ice, washed again, and analyzed for annexin V binding using the annexin V-PE apoptosis detection kit (BD PharMingen).
|
|
|---|
To determine how CD81 regulates CD19 expression, we studied the involvement of distinct structural domains of the CD81 molecule in this function. For this, we generated a panel of chimeric hCD81/hCD9 molecules. CD9 was chosen to complement the tetraspanin backbone because among all tetraspanins it is most structurally related to CD81 (32) and because CD9 has no effect of its own on CD19 expression (43). The chimeric molecules, which express the indicated domains of CD81, including the N' terminus (N'), TM, small extracellular loop (SEL), large extracellular loop (LEL), and C' terminus (C'), were transduced into the Cd81/ cell line, and their effect on CD19 expression was determined. All the chimeric molecules were similarly expressed on the cell surface (see Fig. S3 in the supplemental material). We first analyzed chimeras in which the CD81 domains were progressively replaced from the C' with the corresponding sequences of CD9. Cells were transduced with the tetraspanin-IRES-GFP retroviruses and, in parallel, with a control GFP retrovirus; the transduced cells were then compared for CD19 expression (Fig. 1). Chimeras I(N'-LEL), II(N'-TM3), III(N'-SEL), and IV(N'-TM1) supported the expression of CD19, equally to that of the entire hCD81 molecule (Fig. 1a), demonstrating that the N'-TM1 region of CD81 is sufficient. When this region was further divided, chimera X(N') had no activity, whereas chimera XIII(TM1) retained the full ability to support normal surface expression of CD19 (Fig. 1). This demonstrated that CD81 TM1 is sufficient for normal cell surface expression of CD19. These differences in the abilities of the chimeric molecules to effect CD19 expression were not due to variation in their own levels of surface expression (see Fig. S3 in the supplemental material). In a second panel of chimeras, we replaced CD81 domains from the N' region (Fig. 1b). Only chimeras that contain CD81 TM1 showed a two- to threefold increase in CD19 expression over cells transduced with GFP alone. This criterion is based on primary B cells, in which the absence of CD81 results in a two- to threefold reduction of cell surface expression of CD19 (43).
![]() View larger version (38K): [in a new window] |
FIG. 1. The first transmembrane domain (TM1) of hCD81 is sufficient for normal surface expression of CD19. A panel of retroviruses encoding hCD81 (boldface lines), hCD9 (dotted lines), or chimeric hCD81/hCD9 molecules were used to infect the Cd81/ early B-cell line 1C8. Roman numerals represent laboratory designations for the chimeric molecules, and the CD81 region in each chimera is in parentheses. Cells were infected with the indicated tetraspanin-IRES-GFP retroviruses and stained with an anti-CD19 MAb (filled histograms). In parallel, CD19 expression was analyzed in cells infected at the same time with a control GFP retrovirus (boldface-line histograms). Each of the panels compares CD19 expression in the transduced cells (gated GFP positive). Cells were also stained with an isotype control MAbs (dotted-line histograms). The infected cells express equivalent levels of the transduced molecules (see Fig. S3 in the supplemental material). Only chimeras that contain the CD81 TM1 domain express normal levels of CD19, which are two to threefold higher than the levels expressed in 1C8 cells. (a) Replacement of CD81 domains from the C terminus with the corresponding CD9 sequence; (b) replacement of CD81 domains from the N terminus with the corresponding CD9 sequence. The data represent two or three independent flow cytometry analyses for each chimera.
|
![]() View larger version (30K): [in a new window] |
FIG. 2. CD81 affects intracellular processing of CD19. Analysis of CD19 glycoforms expressed in the Cd81/ cell line (1C8) and in subclones expressing hCD81 or chimeric tetraspanin molecules. The mature (mCD19) (solid arrows) and the precursor (pCD19) (dashed arrow) CD19 glycoforms were detected by immunoblotting with rabbit anti-CD19 Ab. (a) Cells lacking CD81 (1C8) accumulate smaller amounts of slow-migrating mCD19 and larger amounts of pCD19 than CD81-expressing cells (1C8/hCD81); the data represent one of three analyses. (b) Sensitivity of the CD19 glycoforms to digestion by endo-F and endo-H. Immunoprecipitated (IP) CD19 molecules were digested with endo-H or endo-F, as indicated, and detected by immunoblotting (IB). The CD19 molecules in both cell lines are equally sensitive to endo-F digestion, and the deglycosylated CD19 molecules migrate similarly. The difference in the migration of undigested mCD19 molecules in 1C8 and 1C8/hCD81 cells is therefore due to a variation in glycosylation. These data represent one of two analyses. (c) Analysis of subclones expressing hCD81 and chimeric molecules. The mCD19 molecules are more abundant in cells expressing the TM1 domains of CD81 [hCD81, II(N'-TM3), IV(N'-TM1), and XIII(TM1)], whereas pCD19 molecules are more abundant in cells that do not contain TM1 [VII(LEL) and VI(TM2-C')]. The membrane was reblotted with an antiactin Ab (lower panel). (d) The ratios between pCD19 and mCD19 in each lane were determined by densitometry. Subclones that support optimal expression of CD19 (Fig. 1) have low pCD19/mCD19 ratios (white bars), whereas subclones that do not support CD19 expression have higher pCD19/mCD19 ratio (gray bars). The data represent a summary of four independent experiments; error bars indicate SD.
|
Kinetics of CD19 processing in the presence and absence of CD81. To determine the effect of CD81 on the kinetics of CD19 processing, we used radioactive pulse-chase metabolic labeling. Cells (1C8 and 1C8/hCD81) were labeled for 30 min with [35S]methionine and [35S]cysteine and then chased for the indicated time periods (Fig. 3a). In both cell lines the magnitude of radioactive incorporation into CD19 was low, indicating a low rate of CD19 synthesis. Nevertheless, specific forms of newly synthesized CD19 molecules could be compared in the matched cell lines. In both cell lines, the low-molecular-weight pCD19 form was detected at the end of the pulse. It gradually decreased with chase periods, while the amount of the high-molecular-weight mCD19 form increased. This confirms the precursor-product relationship between the low-molecular-weight pCD19 form and the high-molecular-weight mCD19 form. In 1C8 cells (Fig. 3a, right panel), there was a significant delay in the transition from the pCD19 to the mCD19 molecules compared to that in 1C8/hCD81 cells (Fig. 3a, left panel). In the absence of CD81, even at the 120-min chase period, there were fewer mCD19 molecules and more pCD19 molecules than in the presence of CD81 (Fig. 3b), as was the case for the relative amounts of these glycoforms at steady state (Fig. 2a). In addition, the newly synthesized mCD19 molecules migrated slower in 1C8 cells, as seen for molecules accumulated in steady state. Thus, in the absence of CD81, the expression of CD19 on the cell surface is impaired because of a defect in the maturation of CD19 from the ER/pre-Golgi. In addition, CD81 has an effect on the modification of CD19 N-glycans, which are known to occur in the Golgi.
![]() View larger version (38K): [in a new window] |
FIG. 3. CD81 facilitates the exit of CD19 from the ER. (a) Pulse-chase analysis of CD19 in cells labeled for 30 min with [35S]methionine and [35S]cysteine and then chased for the indicated time periods. Labeled cells were lysed, followed by immunoprecipitation (IP) with an anti-CD19 MAb (1D3) or isotype control MAb (IC). In the presence of CD81 (1C8/hCD81 cells, left panel), mCD19 molecules (solid arrow) can already be seen at the end of the pulse period (0-min chase), and most of the pCD19 molecules (dashed arrow) are chased within 1 h. In the absence of CD81 (1C8 cells, right panel) mCD19 molecules can be detected only after 1 h of chase, while pCD19 molecules can still be seen even after 2 h of chase. The open arrowheads point to nonspecific bands, immunoprecipitated by the isotype control MAb. (b) The amounts of mCD19 and pCD19 in each lane were determined by densitometry and are presented as ratios of these glycoforms. These data represent three independent experiments.
|
![]() View larger version (41K): [in a new window] |
FIG. 4. CD81 associates with both the precursor and the mature CD19 glycoforms. 1C8/hCD81 cells were lysed in the indicated lysis buffer, immunoprecipitated (IP), and immunoblotted (IB) with the indicated MAbs. Lanes contain either 10 µl of cell lysates or immunoprecipitated molecules from 400 µl of lysate. (a) The anti-CD81 MAb coprecipitates both pCD19 and mCD19 under mild (Brij 97), but not harsh (NP-40) lysis conditions. The data represent two independent experiments. (b) Cell lysates of 1C8/hCD81 or 1C8/VI(TM2-C') cells, immunoprecipitated with the indicated MAbs, were digested by either endo-H or endo-F, and CD19 molecules were immunoblotted. The anti-hCD81 MAb coprecipitates pCD19 molecules, which are sensitive to endo-H digestion and migrate similarly to the endo-F-deglycosylated CD19. IC, isotype control.
|
![]() View larger version (32K): [in a new window] |
FIG. 5. CD81 LEL augments the association with CD19. The indicated subclones were lysed as indicated, and 400 µl of each lysate was immunoprecipitated (IP) with anti-CD19 (1D3) or directly loaded (10 µl per lane). The coprecipitated tetraspanins were immunoblotted (IB) with (a) anti-hCD81 [chimera VII(LEL)] or (b) anti-hCD9 [(chimera XIII(TM1) and 1C8/hCD9)] (upper panels). The anti-CD19 Ab coprecipitates a larger fraction of chimera VII(LEL) than chimera XIII(TM1) or the whole hCD9 molecules. The ratio of CD19-precipitated tetraspanin molecules to the total amount of CD19 is the highest in chimera VII(LEL), as shown by reblotting of the membranes with anti-CD19 Abs (lower panels). The data represent two independent experiments.
|
![]() View larger version (42K): [in a new window] |
FIG. 6. Engagement of CD81 on the cell surface increases annexin V binding and requires the CD81 LEL and TM2-TM3 domains. (a) 1C8/hCD81 cells were incubated with the biotinylated anti-hCD81 MAb 5A6 (left panel) or with a biotinylated isotype control MAb (middle panel) and then cross-linked by streptavidin. Cross-linking of CD81 (solid line) induced an increase in annexin V binding in live cells (7-amino-actinomycin D [7AAD] negative, R3-gated cells) in comparison to treatment with the control MAb (dotted line), as shown in the histograms (right panel). (b) 1C8 cells expressing the indicated chimera were stimulated as for panel a. Live cells (GFP positive) (see Fig. S5a in the supplemental material) were gated and analyzed for annexin V binding. All analyzed cells express the hCD81 LEL and bind the 5A6 MAb (see Fig. S3 in the supplemental material); however, only cells that contain the CD81 TM2 and TM3 domains show an increase in annexin V binding in response to cross-linking by the antibody. Similar results were obtained by cross-linking using unconjugated MAbs and a secondary anti-mouse Ig Ab (data not shown). These data represent six independent experiments using cells purified from three independent retroviral infections. (c) The Cd81+/+ early B-cell line 2F3 was infected by retroviruses encoding hCD81 or chimera VII(LEL) (see Fig. S5b in the supplemental material). Infected cells (GFP positive) were sorted and analyzed for annexin V binding following cross-linking with the antibody, as for panel a. Cells expressing hCD81 increased annexin V binding, while those expressing chimera VII(LEL) failed to do so, although 2F3 cells express normal levels of CD19. (d) 1C8/hCD9 or 1C8/II(N'-SEL) cells (see Fig. S5c in the supplemental material) were incubated with the biotinylated anti-hCD9 (50H.19) MAb or an isotype control MAb and cross-linked by streptavidin. Flow cytometry showed no increase annexin V binding. In histograms showing annexin V, specific MAbs are presented by solid lines and isotype controls by dotted lines. These data represent three independent experiments.
|
To determine whether the CD81 TM2 and TM3 domains are sufficient for the induction of annexin V binding, we used chimera II(N'-TM3), in which the only common regions with chimera VI(TM2-C') are the CD81 TM2 and TM3 domains. We used a biotinylated anti-hCD9 MAb to cross-link the hCD9 LEL domain in this chimera. There was no increase in annexin V binding in chimera II(N'-TM3) or in 1C8/hCD9 cells (Fig. 6d). This analysis was repeated using a different anti-hCD9 MAb, which also had no effect on annexin V binding (data not shown). Taken together, these results indicate that the TM2-TM3 region of CD81 can participate in cellular responses leading to an increase in annexin V binding only in the context of the CD81 LEL. It is therefore likely that the engagement of CD81 recruits CD19 via the LEL domains, which only in the context of TM2-TM3 induce membrane reorganization.
|
|
|---|
3 (23). Our findings support the proposition that tetraspanin web assembly is initiated in the ER, where the tetraspanin-partner building blocks are first formed, and then the web is further assembled with other tetraspanins and their partners during intracellular processing in the Golgi (25, 58). In the absence of CD81, cells express low levels of surface mCD19 and accumulate the precursor form, pCD19 (Fig. 2a). The higher pCD19/mCD19 ratio in the absence of CD81 in steady state (Fig. 2d) indicates that CD81 has a strong effect on the intracellular maturation of CD19. Indeed, kinetic analysis of CD19 processing demonstrates that, even after 2 h of chase, pCD19 molecules fail to mature and are retained in the ER, resulting in lower levels of mCD19 (Fig. 3). This is likely due to an ER quality control function in which incomplete complexes are retained in the ER by resident chaperones and are eventually degraded (4, 13). However, in contrast to the situation with other multiprotein membrane complexes (20), the maturation and cell surface appearance of CD19 in B cells are only partially dependent on CD81.
In the absence of CD81, mCD19 molecules that reach the surface migrate differently (Fig. 2a). This is due to an alteration in the final N-glycan structure (Fig. 2b), suggesting that lack of CD81 also has an effect on posttranslational modification of CD19 in the Golgi. This is the first example in which a tetraspanin has been shown to facilitate the accurate processing of its partner protein in the Golgi. Understanding this function of CD81 will be important in light of a recent study that demonstrated cotrafficking of human immunodeficiency virus with CD81 in infected dendritic cells (16). Human immunodeficiency virus might therefore subvert an intracellular transport capability of CD81.
The engagement of CD81 on mature human B cells induces a strong antiproliferative effect (36). Here, cross-linking of CD81 on mouse early B cells induced the externalization of PS on the plasma membrane (Fig. 6), an early apoptotic event, as measured by annexin V binding (52). CD81 has been demonstrated to act as a costimulatory molecule in human and mouse B and T cells (reviewed in reference 28), suggesting that coengagement of CD81 facilitates the reorganization of the membrane, thereby reducing the threshold of cell activation. This costimulatory function of CD81 has also been demonstrated by the engagement of a natural ligand, the hepatitis C virus envelope proteins (37).
Distinct domains of CD81 are linked with different functions.
Previous studies have linked structural domains of specific tetraspanins with particular functions. For example, an epitope within the CD151 LEL mediates the interaction with
3 integrin (23), and CD82 TM1 is required for cell surface expression of CD82 (5). Other studies have demonstrated the contribution of the cytoplasmic regions of tetraspanins, via palmitoylated cysteines, to their clustering in TEM (47). In this study we analyzed domains within an individual tetraspanin molecule and found that distinct CD81 domains within the tetraspanin backbone function in different cellular compartments (Table 1; Fig. 1). It is important to note that each of the chimeric molecules retained at least one of the described functions of CD81 and was expressed at the cell surface at a level similar to that of the intact CD81 or CD9 molecules.
|
View this table: [in a new window] |
TABLE 1. Distinct CD81 domains mediate different functions
|
(ii) CD81 TM2-TM3 region. The CD81 TM2-TM3 region facilitates the transmission of signaling, which is initiated upon engagement of CD81 on the cell membrane. CD81 plays a role in membrane reorganization in response to coligation of CD19/CD21/CD81 together with the BCR (9). Additional studies suggested that the palmitoylation state of CD81 is involved in signal transduction (8). However, the cytoplasmic inner loop flanked by TM2 and TM3 (QESQC) contains a potential palmitoylation site, which is identical in CD81 and CD9. It is therefore probable that potential palmitoylation sites at the TM/inner loop interface, which differ between the two molecules, may be involved. Alternatively, the helical structure of the CD81 TM2 and/or TM3 could be required for the reorganization of the signaling complex in the membrane. The TM2 and TM3 domains of the molecule may also be necessary for the formation of CD81 homodimers and thereby contribute to membrane reorganization. We speculate that following direct engagement of CD81, the LEL domain recruits CD19 into TEM, which activates a cellular response only in conjunction with CD81 TM2-TM3. Indeed, cells expressing chimeras that respond (chimeras I and VI) or that do not respond (chimeras II and VII) demonstrate this requirement for both the LEL and the TM2-TM3 region of CD81 (Table 1).
(iii) CD81 TM1 domain. The CD81 TM1 domain alone is sufficient to support normal surface expression of CD19 (Fig. 1). It enables the pCD19 molecules to exit from the ER/pre-Golgi complex, thus increasing the level of mCD19 (Fig. 2d). However, this domain does not support the complete processing of N-glycans on mCD19 (Fig. 2c). In addition, under lysis conditions that maintain tetraspanin-partner associations, chimeric XIII(TM1) molecules are poorly precipitated by the anti-CD19 MAb, similarly to CD9 (Fig. 5b). It is likely that other detergent conditions could uncover the differences between CD81 TM1 and CD9 with respect to their association with CD19. Finally, CD81 TM1 is also not necessary for the externalization of PS in response to cross-linking, because chimera VI, which does not contain TM1, showed an increase in annexin V binding (Fig. 6). Taken together, this evidence suggests that the CD81 TM1 domain is crucial only for the maturation of CD19 from the ER to the cell surface (Table 1). This result complements the previous finding that the TM domain of CD19 is important for its cell surface expression. In that study, replacement of the CD19 TM and cytoplasmic domains (CD19-LAM) resulted in a dramatic loss of CD19 surface expression (2). The exit of CD19 from the ER might require the release of an ER resident chaperone, such as calnexin, and the CD81 TM1 region might be necessary for this release (38, 57). In this scenario, the absence of CD81 TM1 would prolong binding of pCD19 to the ER chaperon, marking the molecules and targeting them for degradation, as demonstrated for other multiprotein complexes, such as IgM and major histocompatibility complex class I. Interestingly, CD81 TM1, but not CD9 TM1, contains the GxxxG motif (glycine, any three residues, glycine), which is suggested to enable close proximity between interacting membrane helices (14). This motif is not present in the TM domain of CD19, but it could interact with another participating protein.
(iv) CD81 N' domain. The CD81 N' domain, in the context of TM1 in chimera IV(N'TM1), is likely to support the glycosylation of CD19 in the Golgi (Fig. 2c). This region contains potential lipid modification sites, a myristoylation site, and two palmitoylation sites in CD81 and a single palmitoylation site in CD9. Posttranslational lipid modifications occur in the Golgi, and we therefore hypothesize that an acylated CD81 N' domain could mediate the retention of the CD19/CD81 complex in the Golgi, which then permits the proper carbohydrate processing. Indeed, fatty acid acylation was shown to provide a mechanism for the targeting of proteins to selected subcellular membranes (53). Future studies using mutated acylation sites should determine whether modification by lipids is indeed involved.
In summary, CD81 is required for the proper processing, intracellular transport, and membrane functions of CD19, and these different roles are embodied in different domains of the CD81 molecule. The assembly of the CD81/CD19 partnership is initiated during protein biosynthesis in the ER, as observed for other oligomeric complexes. In the ER, CD81 LEL associates with CD19 and CD81 TM1 facilitates the escort of CD19 to the cell surface. In the Golgi, CD81 N' is required for the accurate modification of the N-glycan structure of CD19. At the cell surface, the association of the two molecules takes place via their extracellular domains. The signaling function of CD19 may be enabled by activation via CD81 TM2-TM3 (Table 1; Fig. 7). Altogether, this study focuses on the partnership with CD19 as a model with which to unravel how tetraspanin-partner building blocks are formed, are assembled, and function within the tetraspanin web. The participation of distinct CD81 domains in varied functions may explain the pleiotropic effects of CD81 within the tetraspanin web. CD81 is exploited by two major human pathogens, hepatitis C virus and Plasmodium falciparum, and by analogy these pathogens may interact with different CD81 domains during their respective life cycles.
![]() View larger version (43K): [in a new window] |
FIG. 7. Distinct structural domains of CD81 function in different cellular compartments in B cells. The association between the CD81 LEL and CD19 (symbolized by a black ring) is maintained throughout maturation and trafficking of the molecules to the cell surface, where it is required for the activation of both molecules. However, the association via the CD81 LEL is not sufficient to support surface expression of CD19. ER exit and cell surface expression is supported by CD81 TM1 (highlighted by a green square). ER-retained molecules are sensitive to endo-H digestion of their high-mannose core oligosaccharide (symbolized by - ). In the Golgi, the CD81 N' tail (highlighted by a blue square) is required for proper N glycosylation of CD19 (complexed oligosaccharides are indicated by ). At the cell surface, engagement of CD81 with biotinylated (blue circle) anti-CD81 MAb cross-linked by streptavidin (SA) stimulates externalization of phosphatidylserine, increasing annexin V binding (purple). This membrane reorganization (curved arrow) requires CD81 TM2-TM3 (highlighted by a brown square) only in the context of the CD81 LEL.
|
We thank Ronald Levy for his comments on the manuscript.
Supplemental material for this article may be found at http://mcb.asm.org/. ![]()
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»