Synthesis, Assembly, and Intracellular Transport of the Platelet Glycoprotein Ib-IX-V Complex*

The platelet glycoprotein Ib-IX-V complex plays critical roles in adhering platelets to sites of blood vessel injury and in platelet aggregation under high fluid shear stress. The complex is composed of four membrane-spanning polypeptides: glycoprotein (GP) Ibα, GP Ibβ, GP IX, and GP V. Glycoprotein Ibα contains a binding site for von Willebrand factor through which it mediates platelet adhesion; GP V is required for the complex to bind thrombin with high affinity; and both GP Ibβ and GP IX are necessary for efficient plasma membrane expression of the complex. To further define the roles of the individual polypeptide subunits in the biosynthesis and intracellular transport of the GP Ib-IX-V complex, we studied full and partial complexes expressed in heterologous mammalian cells. We found that the full complex was formed within minutes in the endoplasmic reticulum before being transported into the Golgi cisternae. Approximately 160 min were required for the complex to be fully processed and to appear on the plasma membrane. About 25% of GP Ibα expressed as part of either a GP Ib-IX complex or a GP Ib-IX-V complex was degraded through a nonlysosomal pathway. Over 60% of GP Ibα, however, was degraded when it was expressed in partial complexes with only GP Ibβ or GP IX. The increased degradation was blocked by treating cells either with brefeldin A to prevent the transport of proteins from the endoplasmic reticulum to the Golgi or with lysosomal inhibitors, indicating that GP Ibα expressed in partial complexes was targeted to the lysosomes for degradation. These results indicate that the presence of both GP Ibβ and GP IX, but not the presence of GP V, is required for efficient processing and targeting of GP Ibα to the plasma membrane. Absence of either GP Ibβ or GP IX increased the rate of GP Ibα degradation, providing an explanation for why mutation of their genes leads to deficient GP Ibα expression and platelet adhesion in Bernard-Soulier syndrome, the deficiency disorder of the complex.

When the blood vessel is injured, platelets adhere to the subendothelial matrix exposed at the site of injury. This adhesion is initiated by an interaction between von Willebrand factor exposed on the subendothelium and the glycoprotein (GP) 1 Ib-IX-V complex on the platelet surface (1,2). The same interaction also precipitates pathological platelet aggregation induced by high shear stresses, which occurs at sites of arterial stenosis (3)(4)(5).
The GP Ib-IX-V complex is composed of four polypeptide subunits, GP Ib␣, GP Ib␤, GP IX, and GP V (1,6). Glycoprotein Ib␣ is disulfide-linked to GP Ib␤; GP IX and GP V associate with the complex by noncovalent means (7)(8)(9). Although the four polypeptides are encoded by different genes (10 -13), they are homologous to each other, all belonging to a phylogenetically widespread protein family defined by the presence of a motif containing tandemly repeated leucine-rich sequences (1,14). Among the four subunits, GP Ib␣ is the largest and so far the only subunit shown to bind von Willebrand factor and thrombin. This subunit also contains the sequence in its cytoplasmic domain that attaches the complex to the platelet cytoskeleton (15,16). Glycoprotein V has recently been shown to play a role in forming a high affinity thrombin-binding site within the complex (17). The roles of GP Ib␤ and GP IX in the functions of the complex have not been determined, with the exception that phosphorylation of GP Ib␤ by protein kinase A appears to be responsible for inhibiting the collagen-induced actin polymerization observed in platelets treated with agents that elevate cytoplasmic cyclic AMP (18). These smaller polypeptides do influence complex synthesis, however, because both are required for efficient expression of GP Ib␣ on the plasma membrane (19). In vitro studies demonstrating these requirements are supported by the finding that mutations of both GP Ib␤ and GP IX can cause Bernard-Soulier syndrome (the hereditary bleeding disorder caused by absence of the GP Ib-IX-V complex) by decreasing the amount of GP Ib␣ appearing on the cell surface (20 -23). Studies in transfected cell lines of the role of GP V in complex expression on the plasma membrane have yielded conflicting results (24 -26), and as yet no case of Bernard-Soulier syndrome associated with GP V mutations has been described.
Although the requirements for surface expression of a functional complex are fairly well understood, the mechanisms by which GP Ib␤ and GP IX participate in this process are not. Defining these mechanisms was the focus of the studies reported here. We used mammalian cells expressing full or partial GP Ib-IX-V complexes to study assembly and transport of the complex and to determine how the individual components influence complex transport, degradation, and targeting to the plasma membrane.

MATERIALS AND METHODS
Cell Lines-Transfected mammalian cell lines used in the studies are summarized in Table I. L cell lines were grown in Dulbecco's modified * This work was supported by a grant-in-aid from the American Heart Association, Texas Affiliate (to J. F. D.) and National Institutes of Health Grant HL46416 and a grant-in-aid from the American Heart Association (to J. A. L.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Eagle's medium (Life Technologies, Inc.), and Chinese hamster ovary (CHO) cell lines were grown in ␣-minimal essential medium (Life Technologies Inc.). Both media were supplemented with 10% heat-inactivated fetal bovine serum and selection drugs as indicated in Table I. Metabolic Labeling of Cells-Cells grown in 100-mm culture dishes were first incubated with a L-cystine-free Dulbecco's modified Eagle's medium (Tissue Culture Central Facility, University of California, San Francisco, CA) containing 5% dialyzed fetal bovine serum for 30 min at 37°C and then pulse-labeled with 100 Ci of [ 35 S]cysteine (Amersham Pharmacia Biotech) at 37°C for the times indicated under "Results" for each experiment. At the end of the pulse labeling, the cells were washed once with phosphate-buffered saline and incubated at 37°C for the indicated times with complete Dulbecco's modified Eagle's medium containing 50 times the normal concentrations of L-cystine. The labeled cells were then washed twice with ice-cold phosphate-buffered saline and lysed in 1 ml of lysis buffer containing 20 mM Tris, pH 7.4, 100 mM NaCl, and 1% digitonin. To inhibit proteolysis, the lysis buffer contained a mixture of protease inhibitors including 0.1 mg/ml soybean trypsin inhibitor (Sigma), 160 g/ml benzamidine (Sigma), 10 g/ml of leupeptin (Sigma), 10 mM phenylmethanesulfonylfluoride (Sigma), and 5 mM EDTA.
Immunoprecipitation and SDS-Polyacrylamide Gel Electrophoresis-Lysates from the metabolically labeled cells were first incubated with 40 l of a suspension of fixed Staphylococcus aureus beads (Pansorbin beads, Calbiochem Corp, La Jolla, CA) for 60 min at 4°C to adsorb proteins bound nonspecifically. After the beads were removed by centrifugation, the cell lysate was incubated overnight at 4°C with the GP Ib␣ antibody AK2 (1.2 g/ml, kindly provided by Dr. Michael C. Berndt, Baker Medical Research Institute, Prahran, Victoria, Australia). The lysate was then incubated with a rabbit anti-mouse IgG secondary antibody at 4 g/ml for an additional 60 min at 4°C. To precipitate the complex, 40 l of Pansorbin beads was added to the cell lysate and incubated for 60 min at 4°C. The beads with bound proteins were separated from the supernatant by centrifugation at 10,000 ϫ g for 5 min at 4°C. The bead pellets were washed three times with ice-cold lysis buffer and resuspended in SDS sample buffer to a final concentration of 2% ␤-mercaptoethanol in the lysate. The protein samples were then denatured by boiling for 5 min. For the partial complex expressing only GP Ib␤ and GP IX, the GP IX antibody SZ1 (2.5 g/ml, Zymed Laboratories Inc., South San Francisco, CA) was used for the precipitation. To specifically precipitate only the mature form of GP Ib␣ (27), the lysates were incubated with wheat germ agglutinin linked to Sepharose 4B beads (40 l at 1 mg/ml, Sigma) overnight at 4°C, and the beads were then washed several times with ice-cold phosphate-buffered saline. The samples were then resuspended in SDS sample buffer and denatured as described above.
The denatured protein samples were resolved on 4 -20% gradient SDS-polyacrylamide gels. To ensure equal loading, the total protein concentration of each sample was first determined using a commercial protein assay kit (BCA protein assay kit, Pierce), and 10 g of total protein was loaded for each sample. After electrophoresis, the gels were fixed in a solution of 50% (v/v) methanol and 10% acetic acid in water for 30 min at room temperature and dried at 80°C for 2 h. The dried gels were then exposed to a PhosphorImager plate for 72 h at room temperature. Incorporated [ 35 S]cysteine was detected by scanning the plate on a Fuji BAS-1000 Bio-Imaging Analyzer (Fuji, Tokyo, Japan), and the data were analyzed with MacBAS software. All the experiments were performed three to six times in duplicate sets of CHO and L cell lines Endoglycosidase Treatment-Endoglycosidase D (Endo-D, Boehringer Mannheim) and endoglycosidase H (Endo-H, Genzyme, Cambridge, MA) were used to determine the compartmental locations of the polypeptides during transport. Polypeptides are resistant to the actions of Endo-D until they reach the cis compartment of the Golgi; in contrast, the polypeptides are sensitive to the action of Endo-H until they enter the medial compartment of the Golgi. For Endo-D treatment, immunoprecipitated polypeptides were resuspended in 20 l of sodium phosphate buffer, pH 5.5, and boiled for 5 min. The denatured samples were then treated with Endo-D (5 microunits/mg of protein) for 24 h at 37°C. For Endo-H treatment, immunoprecipitated proteins were resuspended in 20 l of 50 mM sodium citrate, pH 6.0, and denatured. The samples were then incubated with Endo-H at a final concentration of 10 M/mg of protein for 12 h at 37°C. Following the treatments, the protein samples were boiled again to inactivate the enzyme and resolved on reducing 4Ϫ20% gradient SDS-polyacrylamide gels.
Blockade of ER to Golgi Transport and Inhibition of Lysosomal Activity-To block transport of the polypeptides from the ER to the Golgi apparatus, brefeldin A (Sigma) was added to the medium at a final concentration of 5 M during the pulse-chase experiments.
To specifically inhibit lysosomal activity, cells were treated with a combination of NH 4 Cl and leupeptin (Sigma) at final concentrations of 50 M and 20 g/ml, respectively, during the pulse-chase experiments. Neither inhibitor has been shown to affect the efficiency of assembly of protein complexes or the rate of peptide glycosylation (28).
Western Blot-To identify the polypeptides immunoprecipitated by the GP Ib␣ antibody AK2, portions of metabolically labeled cell lysates from L␣␤IX cells were separated on 4 -20% gradient SDS-polyacrylamide gels under reduced and nonreduced conditions in parallel with pulse-chase experiments. Separated polypeptides were electrophoretically transferred to the nitrocellulose membrane. To block nonspecific binding, the membranes were first incubated with 5% nonfat milk in Tris-buffered saline (50 mM Tris-HCl, pH 7.4, and 100 mM NaCl) for 60 min at room temperature. The membrane was then incubated for 60 min at room temperature in Tris-buffered saline containing the anti-GP Ib␣ monoclonal antibody, WM23 (kindly provided by Dr. Michael C. Berndt) at a final concentration of 1.0 g/ml, 1% nonfat milk, and 0.2% Tween 20. After the membrane was washed to remove unbound antibody, specific antibody binding was visualized using horseradish peroxidase-conjugated sheep anti-mouse IgG (Amersham Pharmacia Biotech) and luminol substrate with the ECL detection kit (Amersham Pharmacia Biotech) according to the manufacturer's instructions.

Identification of the Individual Polypeptides of the Complex-
The identities of the individual polypeptides immunoprecipitated by the GP Ib␣ monoclonal antibody, AK2, were determined by two experimental means: by Western blotting using the monoclonal GP Ib␣ antibody WM23 and by comparing the sizes of the polypeptides separated on a polyacrylamide gel under reducing and nonreducing conditions. On a nonreducing gel, the disulfide link between GP Ib␣ and GP Ib␤ remained intact, and a corresponding band with the combined molecular mass of GP Ib␣ and GP Ib␤ was detected by WM23 (Fig. 1A). The molecular mass of the band dropped by the expected amount in the reducing gel (Fig. 1A). On the autoradiograph, the GP Ib␤ band disappeared in the nonreducing gel with a corresponding increase in the molecular mass of the GP Ib␣ band (Fig. 1B). Three polypeptides, GP Ib␣, GP Ib␤, and GP IX, were seen on the autoradiograph of the AK2 immunoprecipitate under reducing conditions (Fig. 1B), with GP Ib␣ being detected by WM23 on Western blot (Fig. 1A).
The GP Ib-IX-V Complex Is Formed Rapidly in the Endoplasmic Reticulum-To study the rate and sequence of complex assembly, L cells expressing the full GP Ib-IX-V complex (L␣␤IXV cells) were metabolically labeled with [ 35 S]cysteine for 5 min and chased for 0, 0.5, 2, 5, and 10 min at 37°C. The cells were then lysed with 1% digitonin, and the complex was immunoprecipitated with AK2. All four polypeptides of the complex, GP Ib␣, GP Ib␤, GP IX, and GP V, were precipitated by the antibody after 5 min of pulse labeling, with or without chase ( Fig. 2A). The apparent molecular mass of GP Ib␣ was approximately 70 kDa, overlapping with GP V. This low molecular mass form of GP Ib␣ corresponded to an intermediate form of GP Ib␣ that was partially N-glycosylated (10) but had not undergone O-glycosylation (10,29). Because O-glycosylation occurs only in the Golgi cisternae, this result suggested that the four polypeptides became associated very rapidly after their synthesis and transport into the ER. To confirm the ER location of complex formation, we treated the cell lysates with Endo-D and detected no molecular mass decrease of GP Ib␣ (Fig. 2B), indicating that the complex had indeed formed in the ER. A glycoprotein becomes Endo-D-sensitive only when its high mannose sugars have been processed to Man 5 GlcNAc 2 oligosaccharide intermediates by ␣-1,2-mannosidase I, which resides in the cis-Golgi; thus Endo-D resistance indicates that a glycoprotein is located in the ER or in pre-cis Golgi cisternae (30).

Transport of the GP Ib-IX-V Complex to the Cell Surface-
Next, we determined the time required for transport of the newly synthesized GP Ib-IX-V complex to the cell surface. Cells that express the full complex (␣␤IXV cells) were first pulselabeled with [ 35 S]cysteine for 10 min and then chased for 0, 30, 90, and 150 min. They were then lysed with 1% digitonin, and GP Ib␣ was immunoprecipitated with AK2. To follow the intracellular location of GP Ib␣, half of the samples from each time point were treated with Endo-H, an endoglycosidase that cleaves only N-glycan substrates before they are trimmed by mannosidase II in the medial Golgi. Fully mature GP Ib␣ was recognized by precipitation with wheat germ agglutinin, a lectin that binds to the terminal sialic acids of GP Ib␣ carbohydrate chains (31,32). Sialic acid is only added in the late Golgi compartments. When ␣␤IXV cells were pulse-labeled with [ 35 S]cysteine for 10 min, a low molecular mass, Endo-Hsensitive form of GP Ib␣ can be detected after 0, 30, and 90 min chases (Fig. 3). In parallel, a high molecular mass, fully glycosylated form of GP Ib␣ (125 kDa) became apparent after 30 min of chase. This high molecular mass form of GP Ib␣ was resist- Under reducing conditions, three specific bands were immunoprecipitated in L␣␤IX cells corresponding to GP Ib␣, GP Ib␤, and GP IX in order of decreasing molecular mass. Under nonreducing conditions, the band corresponding to GP Ib␤ disappeared, and the higher molecular mass band migrated more slowly, indicating that GP Ib␣ and GP Ib␤ are disulfide-linked under these conditions.

FIG. 2. The GP Ib-IX-V complex is formed in the endoplasmic reticulum.
A, L cells expressing the full complex (L␣␤IXV cells) were pulsed for 5 min with [ 35 S]cysteine and chased for up to 10 min. At different chase times the cells were lysed in 1% digitonin, and the complex was immunoprecipitated with the GP Ib␣ antibody AK2. GP V migrates at the same molecular mass as immature GP Ib␣ (24). B, L␣␤IXV cells were pulsed for 5 min without subsequent chase. The AK2 immunoprecipitate was treated with Endo-D as described under "Materials and Methods." GP Ib␣ was in a low molecular mass form that was resistant to Endo-D treatment. The figure is representative of three individual experiments. ant to Endo-H treatment, indicating that the polypeptide had passed the medial Golgi compartment (Fig. 3). The ratio of high to low molecular mass forms of GP Ib␣ increased with increasing chase time. At 150 min, almost all of the newly synthesized GP Ib␣ had been converted to the high molecular mass form that could be precipitated by wheat germ agglutinin (data not shown), indicating that the complex requires approximately 160 min from the time of its synthesis to reach the cell surface.
Intracellular Degradation of the Complex-In addition to the processing of GP Ib␣ we observed in the pulse-chase experiments, we also noted an overall decrease in the density of GP Ib␣ over time. After 160 min of chase, only 75% of the original labeled amount of GP Ib␣ can be detected in pulse-chase experiments with L␣␤IXV cells, indicating that a portion is being degraded intracellularly. The degradation could not be inhibited by treating the cells with the lysosomal inhibitors leupeptin and ammonium chloride but was prevented by treating the cells with brefeldin A (data not shown). Having quantitated GP Ib␣ degradation in the full complex, we investigated whether lack of the other complex polypeptides lead to further degradation of GP Ib␣. First, we studied whether GP V had any influence on GP Ib␣ degradation. Our previous studies had shown that the polypeptide has very little influence on the surface level of GP Ib␣ (24). We compared GP Ib␣ degradation in a cell line expressing the full complex (␣␤IXV cells) to that in cells lacking only GP V (␣␤IX cells). Both cell lines were pulselabeled for 30 min and chased for up to 150 min. GP Ib␣ was immunoprecipitated with AK2 and electrophoresed. The densities of GP Ib␣ bands and the ratio of the high to low molecular mass forms were determined as an index of GP Ib␣ maturation. The extent of GP Ib␣ maturation was similar whether GP V was present or not (Fig. 4, A and B).
We next studied the roles of GP Ib␤ and GP IX in the synthesis and transport of GP Ib␣. For these studies, we immunoprecipitated GP Ib␣ with the monoclonal antibody AK2.
Previously studies have shown that this antibody recognizes GP Ib␣ in the absence of the other GP Ib-IX polypeptides. Its epitope has been localized to the amino-terminal 275 amino acid residues of GP Ib␣, and it has been shown to be capable of precipitating a proteolytic fragment containing only this region (33,47). AK2 also binds GP Ib␣ in full or partial complexes expressed in mammalian cells (19). These results indicated that binding of AK2 to GP Ib␣ did not require the presence of other subunits and that the epitope specificity of AK2 was not altered when GP Ib␣ was expressed in a partial complex. To evaluate intracellular degradation of GP Ib␣, cell lines that express the full complex (␣␤IXV cells), only GP Ib␣ and GP Ib␤ (␣␤ cells), and only GP Ib␣ and GP IX (␣IX cells) were pulselabeled for 10 min and chased for up to 150 min. The complexes were immunoprecipitated with AK2. In both of the cell lines with partial complexes, over 60% of GP Ib␣ was degraded  during the 150-min chase period compared with the 25% of GP Ib␣ degraded in the full complex over the same time period (Fig. 5). Glycoprotein Ib␣ was degraded more rapidly in ␣IX cells than in ␣␤ cells. Another obvious difference between the cells expressing the full complex and those with partial complexes was that only in the cells with a full complex was a significant proportion of GP Ib␣ processed to the mature high molecular mass form. We also found that AK2 coprecipitated both GP Ib␣ and GP Ib␤ in ␣␤ cells but failed to coprecipitate GP IX in ␣IX cells (Fig. 5), indicating that GP Ib␣ did not directly associate with GP IX. In contrast to the rapid degradation of GP Ib␣ in partial complexes, neither GP Ib␤ nor GP IX was significantly degraded in CHO␤IX cells over the 150min chase period (Fig. 5).
Inhibition of Intracellular Degradation-Because a greater amount of GP Ib␣ was degraded during transport in the partial complexes (␣␤ and ␣IX cells), it was of interest to determine the subcellular department where the degradation occurred. To locate the degradation site, L␣␤ cells were pulsed and chased either in the presence of brefeldin A (34) or in the presence of two lysosomal inhibitors, leupeptin and ammonium chloride. Brefeldin A essentially abrogated the degradation (Fig. 6), suggesting that degradation was a post-ER event. The bulk of the degradation was also inhibited by treating cells with leupeptin and ammonium chloride (Fig. 6), indicating that GP Ib␣ in the partial complexes was targeted to the lysosomes where it was rapidly degraded. DISCUSSION The GP Ib-IX-V complex is a large and complicated assembly of polypeptides on the plasma membrane. It contains four polypeptide subunits, likely present as multiple copies in the functional receptor (1). In the current studies, we sought to elucidate how this complex is assembled in the cell and transported to the plasma membrane and to understand how its nonligand-binding subunits, GP Ib␤, GP IX, and GP V, participate in this process.
The studies were carried out on transfected mammalian cell lines that express full or partial complexes. In cells expressing the full complex, we found that the complex was fully assembled in the endoplasmic reticulum at the earliest times after pulse labeling the cells with [S 35 ]cysteine ( Fig. 2A). Complex assembly in the ER was confirmed by two means. First, GP Ib␣ precipitated at the earliest time of pulsing (5 min) was present as a low molecular mass form lacking O-glycosylation ( Fig. 2A), a modification that only occurs in the Golgi compartments (35). Second, this immature form of GP Ib␣ was resistant to Endo-D treatment, indicating that it had yet to reach the cis-Golgi (Fig.  2B) (30,36). These findings put the GP Ib-IX-V complex into a large group of multiple subunit and multimeric membrane protein complexes that form in the ER, including the T cell receptor (37), acetylcholine receptor (38), influenza hemagglutinin trimer (39), and acetylcholinesterase (40).
Once the complex is formed, it is transported into the Golgi compartments for further modifications, most notably O-glycosylation, which causes a large increase in the molecular mass of GP Ib␣ (Fig. 3) as expected based on the determination that approximately half of the molecular mass of the extracellular domain of GP Ib␣ in platelets is from carbohydrate (31). The modifications took approximately 160 min to complete because wheat germ agglutinin could only precipitate GP Ib␣ after a 150-min chase. We chose wheat germ agglutinin because it recognizes the terminal sialic acids on O-linked oligosaccharides of GP Ib␣ (32,41,42). Sialylation is the last modification of O-linked carbohydrate chains to occur before glycoproteins are either secreted or expressed on the plasma membrane.
Early formation of the complex suggests that assembly is a prerequisite for transport. One apparent reason for such early assembly is to prevent the polypeptides from being degraded during transport, a possibility that prompted us to investigate the stability of the polypeptides when formation of a full complex was precluded.
In studies of GP Ib␣ stability, we observed that about 25% of GP Ib␣ expressed even in the full complex was degraded by nonlysosomal mechanisms. We then investigated whether a higher proportion of GP Ib␣ would be degraded if any of the other polypeptides were excluded from the complex.
Glycoprotein V was our first target, based on its loose association with the other subunits and the observations that GP V is not required for efficient expression of the complex on the cell surface (19,24). We found that GP Ib␣ was degraded to a similar extent in the presence or absence of GP V (Fig. 4A). The polypeptide also did not influence the rate of GP Ib␣ intracellular transport as indicated by its lack of influence on the appearance of the mature higher molecular mass form of GP Ib␣ (Fig. 4B). These results are consistent with our previous findings that GP V does not enhance the surface expression of GP Ib␣ (19,24) although its presence may be important in inducing high affinity binding of thrombin to the complex (17). Also consistent with these findings, no cases of Bernard-Soulier syndrome caused by mutations of GP V gene have yet been reported.
Unlike GP V, however, both GP Ib␤ and GP IX are important for GP Ib␣ stability. Lack of either of these subunits led to degradation of over 60% of GP Ib␣ during the 150-min chase period (Fig. 5). Degradation above that observed with the full complex occurred before GP Ib␣ was fully O-glycosylated and was completely inhibited by the lysosomal inhibitors leupeptin and NH 4 Cl. This finding indicates that one role of the association of GP Ib␣ with GP Ib␤ and GP IX is to prevent lysosomal targeting. The protective effects appear to correlate with whether two subunits associate because degradation of GP Ib␣ was greater in ␣IX cells than in ␣␤ cells. We were able to demonstrate association of GP Ib␣ with GP Ib␤ but not with GP IX in these partial complexes (Fig. 5), consistent with our previously reported results (43). In cells containing both GP Ib␤ and GP IX, these polypeptides degraded to a much lesser extent than did GP Ib␣ (Fig. 5), suggesting that intracellular degradation may specifically target to the latter polypeptide. This could be because levels of GP Ib␣ have to be more tightly regulated, because it contains most of the sites through which the GP Ib-IX-V complex interacts with other proteins. Glycoprotein Ib␤ and GP IX are resistant to proteolysis when complexed with each other, although only the stability of GP Ib␤ has been studied when expressed alone (23). Stability of GP Ib␤ appears to be unaffected by the presence of GP IX. This selective degradation may serve as a checkpoint to ensure expression of only the correctly assembled and biologically active GP Ib-IX-V complexes on the plasma membrane (36,44). A similar situation exists for the well studied T cell receptor-CD3 complex. When expressed alone in COS cells, CD3␥, CD3⑀, and are stable, whereas TCR␣, TCR␤, and CD3␦ are rapidly degraded (45,46). A subcomplex is therefore stable if it contains one or more stable chains (CD3␥, CD3⑀, or ) but is not stable when it contains only the unstable chains (TCR␣, TCR␤, and CD3␦), suggesting that a stable polypeptide, by associating with an unstable one, can prevent the latter from being degraded.
In summary, we have demonstrated that formation the GP Ib-IX-V complex occurs immediately after synthesis of its constituent polypeptides. This newly assembled complex takes approximately 3 h to reach the cell surface, in transit undergoing a number of post-translational modifications. Efficient transport of the ligand-binding subunit of the complex, GP Ib␣, to the plasma membrane requires both GP Ib␤ and GP IX, which direct GP Ib␣ away from lysosomes and therefore protect it from proteolysis. Subcomplexes of different polypeptides form normally within the cell, but GP Ib␣ in complexes lacking GP Ib␤ or GP IX is more susceptible to lysosomal proteolysis. Our studies provide evidence that a decreased or absent expression of the GP Ib-IX-V complex caused by the mutations in GP Ib␤ or GP IX in Bernard-Soulier syndrome patients is due to intracellular degradation of the complex as well as defective membrane targeting.