Proteasome-mediated Degradation of Apolipoprotein B Targets Both Nascent Peptides Cotranslationally Before Translocation and Full-length Apolipoprotein B After Translocation into the Endoplasmic Reticulum*

A major portion of newly synthesized apolipoprotein B (apoB) is degraded intracellularly. This degradation has been demonstrated to be mediated largely by the ubiquitin-proteasome pathway. We examined whether nascent apoB polypeptides or full-length apoB is selectively retrotranslocated from the endoplasmic reticulum into the cytosol for degradation. Herein, we found that full-length apoB as well as partial-length apoB peptides are ubiquitinated in HepG2 cells, and ubiquitination is an exclusively cytosolic process. Calnexin, which binds specifically to glycoproteins, has been postulated to promote apoB folding and complete translocation; we found that ubiquitinated apoB is bound to calnexin, suggesting that ubiquitinated apoB is glycosylated. In addition to calnexin binding, we have other pieces of evidence that the full-length intracellular ubiquitinated apoB is glycosylated, because (i) it binds to concanavalin A, and (ii) glycan can be demonstrated in the full-length ubiquitinated apoB by a chemical detection method involving oxidation of adjacent hydroxyl groups in the glycan moiety. Because glycosylation occurs inside the endoplasmic reticulum, the full-length glycosylated apoB must have been retrotranslocated into the cytosol for ubiquitination and proteasome-mediated degradation. Next we synchronized translation in HepG2 cells by puromycin treatment. A pulse-chase experiment using [35S]methionine labeling of intracellular apoB in these synchronized cells demonstrated that nascent partial-length apoB peptides are also ubiquitinated cotranslationally. We conclude that the ubiquitin proteasome-mediated degradation of apoB targets both nascent peptides cotranslationally before translocation as well as full-length apoB after its translocation into the endoplasmic reticulum.

As the sole protein component in low density lipoproteins, apoB-100 1 is an important determinant of atherosclerosis susceptibility (1,2). The plasma concentration of apoB-100 is a balance between its production rate in the liver and its removal from the circulation by receptor-and nonreceptor-mediated pathways (1). It has been known for over a decade that apoB-100 production in the liver is regulated almost exclusively at the posttranscriptional level (3). ApoB mRNA levels do not change substantially in response to various stimuli. However, there appears to be significant variation in the proportion of the newly translated apoB-100 that is degraded intracellularly before secretion (3)(4)(5)(6). The intracellular degradation of apoB is thus an important factor determining apoB production in the liver. Recently, the ubiquitin-proteasome pathway has been identified as a major mechanism for the intracellular degradation of apoB-100 (7)(8)(9). In the last few years, it has become evident that the proteasome is responsible for the degradation of a number of intracellular proteins, usually resulting in the complete degradation of the target proteins (10 -12). The site of proteasome-mediated degradation appears to be cytosolic, requiring the retrograde translocation of the proteins from the endoplasmic reticulum (ER) into the cytosol for degradation (10,12). ApoB is one of the largest secretory proteins known, and it has a complex structure (2,13,14). The proteolytic removal of apoB by proteasomes appears to be a mechanism by which the cell removes apoB molecules that are misfolded, especially when lipid supply is limiting. However, it is presently not known whether only incomplete nascent apoB peptides or full-length apoB molecules that are completely translocated into the ER are targeted for proteasome-mediated degradation. In this study, we have examined this mechanistic issue in the biogenesis of apoB-100 in a human hepatoma cell line, HepG2. We found that both incomplete nascent apoB peptides and the full-length apoB that has undergone glycosylation in the ER can be retrotranslocated into the cytosol for ubiquitination and proteasome degradation.

MATERIALS AND METHODS
Reagents-Trypsin, soybean trypsin inhibitor, N-acetyl-L-leucinyl-Lleucinyl-L-norleucinal (ALLN), N-ethylmaleimide (NEM), and rabbit anti-bovine ubiquitin polyclonal antibody were from Sigma. Gamma-Bind G (protein G-Sepharose), concanavalin A-Sepharose, and cyanogen bromide-activated Sepharose were from Amersham Pharmacia Biotech. Lactacystin was obtained from Dr. E. J. Corey (Harvard Medical School). Mouse anti-bovine ubiquitin monoclonal antibody and goat anti-human apoB polyclonal antibody were from Chemicon International, Inc. Rabbit anti-human albumin polyclonal antibody was from Research Diagnostic, Inc. Tunicamycin, L-[ 35 S]methionine, L-methionine, RPMI 1640 tissue culture medium, methionine-free RPMI 1640, and peroxidase-conjugated goat anti-human albumin polyclonal antibody (IgG fraction) were from ICN. Mouse anti-human apoB monoclonal antibodies (1D1, Bsol 22) were provided by Dr. R. W. Milne * This work was supported in part by National Institutes of Health Grant HL-16512 (to L. C.) and Fellowship Grant GM 17159 (to S. J. Y.). 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.
‡ These authors contributed equally to this work. § Supported by the Karolinska Institute/Baylor College of Medicine Exchange Program and by the Henning and Johan Throne-Holsts Foundation.
¶ To whom correspondence should be addressed: Dept. of Cell Biology, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030. 1 The abbreviations used are: apoB, apolipoprotein B; ALLN, N-(Ottawa Heart Institute). The peroxidase labeling kit was from Pierce. The anti-goat IgG peroxidase, anti-rabbit IgG peroxidase, and oxidative glycan detection kit were from Boehringer Mannheim. The polyvinylidene difluoride membrane was from Bio-Rad. Prestained molecular weight standards and ECL chemiluminescence development solution were from Amersham Pharmacia Biotech. Rabbit anti-calnexin polyclonal antibody was kindly provided by Dr. Ari Helenius (Yale University Medical School). Cell Culture and Microsome Preparation-HepG2 cells from American Type Culture Collection were maintained at 37°C in an atmosphere with 5% CO 2 in RPMI 1640 medium containing 10% fetal calf serum, penicillin (100 units/ml), and streptomycin (100 g/ml). Cells were grown in 162-cm 2 flasks to 80 -90% confluence. Microsomes were prepared as described by Cosgrove et al. (15). In brief, the cells were scraped in phosphate-buffered saline after washing once with ice-cold phosphate-buffered saline. They were pelleted by centrifugation at 500 ϫ g for 5 min. The cells were then homogenized with a glass homogenizer in 0.25 M sucrose and 10 mM HEPES (pH 7.4) on ice. The cell homogenate was first centrifuged at 10,000 ϫ g for 15 min at 4°C, and the resulting supernatant was centrifuged at 100,000 ϫ g for 60 min at 4°C to pellet the microsomes.
Trypsin Digestion-Trypsin digestion of microsomes was done according to Bonnardel and Davis (16). Microsomes were suspended in 0.25 M sucrose and 10 mM HEPES, pH 7.4. Trypsin was then added to a final concentration of 200 g/ml. The mixture was incubated at room temperature for 20 min. After incubation, soybean trypsin inhibitor (final concentration, 5 mg/ml) was added to stop digestion. The microsomes were then lysed 2.5 h at 4°C by adding an equal volume of 2% sodium cholate in HEPES-buffered saline (50 mM HEPES, 200 mM NaCl, pH 7.4) containing 2 mM phenylmethylsulfonyl fluoride, 0.2 mM ALLN, and 10 mM NEM, and a protease inhibitor mixture (1 tablet for 50 ml, complete protease inhibitor tablet, Boehringer Mannheim). The lysate was then centrifuged at 10,000 ϫ g for 5 min. The supernatant was used for immunoprecipitation.
Immunoprecipitation and Immunoblot Analysis-Single immunoprecipitation and sequential immunoprecipitation were done as described previously (7). Cells were lysed for 2.5 h at 4°C in 2% sodium cholate in HEPES-buffered saline (50 mM HEPES, 200 mM NaCl, pH 7.4) containing 1 mM phenylmethylsulfonyl fluoride, 0.1 mM ALLN, 5 mM NEM, and complete protease inhibitors (Boehringer Mannheim). The cell lysate was then centrifuged at 10,000 ϫ g for 10 min. The supernatant was incubated with anti-apoB antibody at 4°C on a gyrating platform for 1 h. Gamma-Bind G beads were then added and the mixture was incubated for 5 h. After 3 washes with 2% cholate, for single immunoprecipitation, the immunoprecipitate was heated to 100°C for 5 min in 2ϫ SDS-polyacrylamide gel electrophoresis (PAGE) buffer containing 5% 2-mercaptoethanol and analyzed by SDS-PAGE. For sequential immunoprecipitation, the first antibody immunoprecipitation was done as described above. The Immunobead-antigen complex was washed 3 times and then incubated in 1% SDS in HEPES-buffered saline (0.1 ml) at 95°C for 3 min to dissociate the antigen from the beads. After centrifugation, the supernatant was removed and diluted with 1 ml of 1% Triton X-100 in HEPES-buffered saline containing 0.1 mM ALLN, 5 mM NEM, 1 mM phenylmethylsulfonyl fluoride, and complete protease inhibitors for the second antibody immunoprecipitation. The immunoprecipitates were separated by SDS-PAGE on 4 -15% gradient gels or 6% gels. Proteins from single or sequential immunoprecipitation on the SDS-PAGE were transferred overnight onto nitrocellulose or polyvinylidene difluoride membranes and probed either with peroxidase-conjugated primary antibody or with unconjugated primary antibody followed by appropriate peroxidase-conjugated secondary anti-IgG as indicated and detected by enhanced chemiluminescence (ECL kit, Amersham Pharmacia Biotech).
Association of Ubiquitinated ApoB with Calnexin-Cultured HepG2 cells were incubated in the absence or presence of 0.1 mM tunicamycin for 3 h. One flask (162 cm 2 ) each of treated and untreated HepG2 cells was processed for sequential immunoprecipitation with anti-apoB and then anti-ubiquitin antibodies as described above. The controls were immunoprecipitated with protein G-Sepharose alone in the second immunoprecipitation step. The samples were electrophoresed and immunoblotted with anti-calnexin antibody as described above.
Concanavalin A-Sepharose Binding-HepG2 cell lysate supernatant was prepared as described above and concanavalin A-Sepharose was added (100 l/ml of cell lysate). The mixture was incubated at 4°C for 5 h, and the concanavalin A-Sepharose was separated by centrifugation. The non-concanavalin A-bound apoB was recovered from the supernatant by immunoprecipitation using a goat anti-apoB polyclonal antibody. To dissociate the concanavalin A-bound apoB, the concanava-lin A-Sepharose beads were washed 3 times and incubated in 1% SDS in HEPES-buffered saline (0.1 ml) at 95°C for 3 min. After centrifugation, the supernatant was diluted to 1 ml with 1% Triton X-100 in HEPES-buffered saline containing 0.1 mM ALLN, 5 mM NEM, 1 mM phenylmethylsulfonyl fluoride, and complete protease inhibitors, and immunoprecipitation was performed using anti-apoB antibody. Immunoprecipitated concanavalin A-bound and unbound apoB were then boiled in 2ϫ SDS-PAGE buffer as described above for the immunoblot assay.
Oxidative Glycan Detection-Three 162-cm 2 flasks of HepG2 cells were solubilized in lysis buffer and homogenized in a glass Dounce homogenizer. The lysate was centrifuged at 100 ϫ g at 4°C for 10 min to remove cellular debris. The lysate was then passed through a column (4 ml) of anti-apoB-Sepharose beads. The column was prepared by conjugating 100 l of goat anti-human apoB antiserum with 1 g, dry weight, of cyanogen bromide-activated Sepharose according to the manufacturer's procedure (Amersham Pharmacia Biotech). The bound apoB was eluted by 2 bed volumes of 50 mM sodium acetate, pH 4.0. The eluate was allowed to drip directly into a tube containing 0.1 volume of 1 M Tris, pH 8.0. The eluate was mixed and then divided into two equal parts. One-half was loaded onto a column of anti-ubiquitin-Sepharose and the other half onto control Sepharose. The anti-ubiquitin-Sepharose was prepared by conjugating rabbit polyclonal anti-ubiquitin antiserum with cyanogen bromide-activated Sepharose. The control Sepharose was prepared by quenching the reactive sites on cyanogen bromide-activated Sepharose with excess Tris. The bound ubiquitinated apoB was eluted the same way. The eluate was then concentrated down to 500 l by using Centricon (molecular mass cutoff, 20 kDa). 30 l of each sample in SDS sample buffer was loaded onto the gels. After electrophoresis, the proteins were blotted onto nitrocellulose membranes, and oxidative glycan detection was performed using a kit (Boehringer Mannheim). Briefly, the membrane was incubated with an oxidizing agent, which oxidized the adjacent hydroxyl groups in carbohydrates. The oxidized group was then conjugated to digoxigenin. Detection was then performed with anti-digoxigenin-alkaline phosphatase with color development using 4-nitro blue tetrazolium chloride/5bromo-4-chloro-3-indolyl phosphate (NBT/x-phosphate solution). Carboxypeptidase Y, a glycoprotein, and albumin, a nonglycoprotein, were used as positive and negative controls, respectively.
ApoB Translation Following Puromycin Synchronization-Puromycin synchronization was performed as described by Benoist and Grand-Perret (17). HepG2 cells were preincubated in the methionine-free medium for 20 min, and puromycin was added (10 M) for another 10 min. The cells were then washed 3 times with ice-cold medium. A 5-min pulse label was initiated by the addition of [ 35 S]methionine (100 Ci/ml) in methionine-free medium followed by a 10-min chase by the addition of cold methionine (15 mg/liter). The cells were lysed and ubiquitinated apoB was immunoprecipitated by goat polyclonal anti-apoB then by anti-ubiquitin antibody as described above. The immunoprecipitates were run on 6% SDS-PAGE, the gels were dried, and the autoradiographic image was captured using a storage phosphor system (Cyclone, Packard). In some parallel puromycin synchronization experiments, lactacystin (10 M) was included in the culture medium.

Polyubiquitination of ApoB in the Absence and Presence of
Lactacystin-In the first experiments, we studied the effect of a specific proteasome inhibitor, lactacystin, on the fate of intracellular apoB in cultured HepG2 cells (Fig. 1). In agreement with experiments using other less specific proteasome inhibitors (ALLN and carbobenzoxyl-leucyl-leucyl-norvalinal-H (MG115)) in our laboratory (7), lactacystin (10 M) treatment led to an accumulation of total intracellular apoB (compare lanes 3 and 4), as well as ubiquitinated apoB (compare lanes 1 and 2). Both in the absence and presence of lactacystin, the ubiquitin-immunoreactive material co-precipitated by the apoB antiserum is quite heterogeneous in size. The largest ubiquitinated apoB molecules have an apparent molecular mass significantly larger than 550 kDa, the size of native full-length unubiquitinated apoB, reflecting the presence of multiple 8.5-kDa ubiquitin conjugates covalently linked to the apoB molecule. Importantly, there are substantial amounts of ubiquitinated apoB immunoreactive material that are significantly smaller than full-length apoB-100. These short mole-cules in the form of a smear on the SDS-PAGE are detected in the absence of lactacystin but are much more abundant in its presence. They may represent short nascent apoB peptides that have been targeted for proteasome degradation by ubiquitination, or they may consist of proteolytic fragments of completed apoB chains that have undergone polyubiquitination.
Polyubiquitination Is a Cytosolic Process-To further characterize the ubiquitination process, we isolated microsomes from HepG2 cells and treated them with trypsin. As shown in Fig. 2, trypsin treatment results in a small decrease in the amount of full-length apoB-100, indicating that some of the apoB is on the cytosolic side of the ER. However, the persistence of a substantial amount of apoB following trypsin treatment indicates that only a portion of the full-length apoB under steady-state conditions is accessible to digestion by exogenously added trypsin, and the rest is within the lumen of the ER and not susceptible to the enzyme digestion. In agreement with the observations of Du et al. (18), upon trypsin treatment of HepG2 microsomes, we observed the accumulation of an 85-kDa N-terminal fragment, which is not ubiquitinated. No C-terminal fragment was detected when we used a C-terminal specific monoclonal antibody (Bsol 22) (19) for detection. Furthermore, trypsinization completely removed the ubiquitinated apoB, whereas under the identical experimental conditions, albumin was completely resistant to the proteolytic action of trypsin. Therefore, trypsin is excluded from the ER lumen and cannot access proteins within the lumen. These results indicate that polyubiquitination is an exclusively cytosolic process.
Ubiquitinated ApoB in HepG2 Cells Is Bound to Calnexin-Calnexin is a ubiquitous resident ER protein that binds to newly synthesized N-linked glycoproteins (20 -22). Chen et al. (9) demonstrated that inhibition of calnexin binding of glycosylated apoB by castanospermine was associated with the proteasome-mediated degradation of apoB. They speculated that calnexin safeguards glycosylated apoB from degradation by promoting correct folding and complete translocation. We investigated whether ubiquitinated apoB, which is tagged for proteasome degradation, would still be bound to calnexin in HepG2 cells. Ubiquitinated apoB was isolated from HepG2 cells by sequential purification using apoB and ubiquitin antisera. The ubiquitinated apoB was found to be bound to calnexin, which was released from the ubiquitinated apoB by SDS and detected by immunoblot analysis (Fig. 3). Therefore, cal-nexin binding per se does not provide absolute protection of apoB against ubiquitination. Furthermore, because calnexin only binds to glycoproteins (22), this experiment indicates that ubiquitinated apoB is glycosylated. As a control, inhibition of apoB glycosylation by tunicamycin was found to completely abolish calnexin binding to the unglycosylated apoB in HepG2 cells.
Ubiquitinated ApoB Is Glycosylated-Calnexin binding to ubiquitinated apoB constitutes only indirect evidence for the presence of carbohydrate conjugates in the ubiquitinated protein. To obtain more direct evidence of the presence of carbohydrate, we used two other independent methods to analyze for the presence of carbohydrate conjugates in intracellular ubiquitinated apoB. In the first series of experiments, we used concanavalin A-Sepharose to affinity purify intracellular apoB from cultured HepG2 cells. Concanavalin A binds to apoB molecules that display covalently linked ␣-D-mannopyranosyl and/or ␣-D-glucopyranosyl on the surface that is available for interaction with the lectin; therefore, all concanavalin A-bound apoB molecules would be glycosylated. The apoB recovered in the unbound fraction would contain unglycosylated apoB as well as glycosylated apoB whose D-glucose or D-mannose moieties are inaccessible for concanavalin A binding. By immuno- blot analysis against a ubiquitin antiserum, the apoB in the unbound fraction was found to be heavily ubiquitinated (Fig.  4A). Most importantly, the concanavalin A-bound apoB, which contains only glycosylated apoB, also displays substantial polyubiquitination. The heterogeneous nature of the concanavalin A-bound and unbound ubiquitinated glycosylated and unglycosylated apoB is very similar to that shown in Fig. 1 and to similar blots on total intracellular apoB from our laboratory (7) and other laboratories (8,9). The presence of fulllength molecules in lane 2 of Fig. 4A suggests that completed and glycosylated apoB proteins are susceptible to ubiquitination. The presence of low molecular weight ubiquitinated apoB peptides that contain carbohydrate further suggests that some apoB chains that are partially in the lumen of the ER are also glycosylated and that glycosylation per se, like calnexin binding, does not protect against polyubiquitination and proteasome-mediated degradation.
In a second set of experiments to complement the concanavalin A binding studies, we used a chemical approach to document that ubiquitinated apoB contains glycan. Ubiquitinated apoB isolated from HepG2 cells was fractionated on SDS-PAGE, and the presence of glycan on the ubiquitinated apoB was determined by an oxidative glycan detection method as described under "Materials and Methods." An unglycosylated protein, albumin, and a glycosylated protein, carboxypeptidase Y, were used as negative and positive controls, respectively. By using this method, there is unequivocal evidence that intracellular ubiquitinated apoB is glycosylated, because clear glycan staining is seen in the ubiquitinated apoB lane (Fig. 4B, lane 3). The glycan detection method is specific for glycoproteins; another glycoprotein, carboxypeptidase Y, also shows positive staining, but an unglycosylated protein, albumin, remains unstained. We note that the carbohydrate conjugates are found almost exclusively on the full-length ubiquinated apoB protein, indicating that the full-length glycosylated apoB-100 is a substrate for retrotranslocation and polyubiquitination. A smaller ubiquitinated apoB fragment that is also glycosylated probably represents a proteolytic fragment of apoB.
Nascent Partial-length ApoB Peptides Are Also Ubiquitinated-The evidence presented thus far strongly indicates that full-length apoB that has undergone glycosylation, a luminal event in the ER, is the substrate for ubiquitination and proteasome degradation. We also addressed the question whether some of the nascent apoB chains are also targeted for degradation by ubiquitination during translation and prior to completing synthesis of the molecule. ApoB-100 translation in HepG2 cells was synchronized by treatment with puromycin prior to pulse labeling with [ 35 S]methionine. By this technique, only newly initiated apoB polypeptide chains would be labeled. As shown in Fig. 5, nascent apoB chains of sizes ϳ220 -400 kDa were labeled under these pulse-chase conditions. After purifying the ubiquitinated nascent chains with an anti-ubiquitin antibody and fractionating them on SDS-PAGE, we observed that a proportion of these 35 S-labeled nascent polypep- FIG. 4. Experiments on the glycosylation state of ubiquitinated apoB. A, concanavalin A binding of ubiquitinated apoB. ApoB was isolated from the culture medium and cellular lysate of HepG2 cells previously incubated in 0.1 mM ALLN for 2 h. Glycosylated apoB fractions were isolated by concanavalin A-Sepharose binding as described under "Materials and Methods." The bound apoB from the secreted (Sec.)and intracellular (Intracell.)compartments and the unbound intracellular apoB were analyzed by 4 -15% gradient SDS-PAGE and detected by Western blot analysis. Note the presence of ubiquitin immunoreactivity in the concanavalin A-bound and unbound apoB. ␣-Ub, anti-ubiquitin antibody; ␣-ApoB, anti-human apoB monoclonal antibody 1D1. B, chemical detection of glycan on intracellular ubiquitinated apoB. Ubiquitinated apoB was isolated from cellular lysates of cultured HepG2 cells by sequential immunoprecipitation using anti-apoB and anti-ubiquitin. The ubiquitinated apoB was analyzed by 4 -15% gradient SDS-PAGE. The presence of glycan in the apoB protein (lane 3) was detected by an oxidative glycan detection method described under "Materials and Methods." Carboxypeptidase Y, a glycoprotein, was used as a positive control (lane 1). Albumin, a nonglycoprotein (lane 2), and control beads (i.e. nonimmune beads instead of anti-ubiquitin beads in the second immunoprecipitation) (lane 4) were used as negative controls. Glycan staining was readily detected in ubiquitinated apoB in the full-length apoB-100 region and in a lower molecular mass region, which probably represents a proteolytic fragment of ubiquitinated apoB.  1 and 2) and analyzed by SDS-PAGE and fluorography. The 35 S-labeled ubiquitinated apoB consists of both full-length and partial-length apoB in these unsynchronized cells. Synchronization was performed in the presence of puromycin as described under "Materials and Methods." In these synchronized HepG2 cells, total and ubiquitinated apoB were isolated in the absence (lanes 3 and 4) and presence (lanes 6 and 7) of lactacystin (10 M). Ubiquitinated apoB was detected in puromycinsynchronized HepG2 cells in both cases but was more intense in the presence of lactacystin. Therefore newly initiated incomplete apoB polypeptides are substrates for polyubiquitination during translation. Control, use of control beads instead of anti-ubiquitin antibody in the second immunoprecipitation. ␣-ApoB, anti-apoB; ␣-Ub, anti-ubiquitin antibody.
tides is ubiquitinated. Therefore, significant ubiquitination occurs on nascent apoB chains during translation. Because we showed above that apoB ubiquitination occurs exclusively in the cytosolic compartment, ubiquitination of these apoB polypeptides must happen cotranslationally before the emerging nascent chain is translocated into the ER. DISCUSSION We first reported that the intracellular degradation of apoB-100 in HepG2 cells occurs largely via the ubiquitin-proteasome pathway, because it has the following characteristics: (i) it is ATP-dependent, (ii) the intracellular apoB destined for intracellular degradation is ubiquitinated, and (iii) proteasome inhibitors inhibit the degradation leading to the accumulation of ubiquitinated apoB (7). These observations were subsequently confirmed and extended by other laboratories (8,9,17). Fisher et al. (8) showed that induced overexpression of hsp70 in HepG2 cells promotes apoB degradation, suggesting a possible role of this chaperone in apoB degradation. Benoist and Grand-Perret (17) demonstrated that inhibition of lipid transfer by a microsomal triglyceride transfer protein inhibitor promotes the degradation of nascent apoB peptides by proteasomes. Chen et al. (9) found that calnexin protects against proteasome-mediated degradation. They also presented experiments that suggest that apoB that has entered the ER far enough to be recognized by the glycosylation machinery may still undergo proteasomal degradation.
In this study, we have demonstrated conclusively that the ubiquitin-proteasome-mediated degradation of apoB affects both incomplete nascent apoB polypeptides cotranslationally as well as the full-length glycosylated apoB protein, which could only have happened to molecules that have completely translocated into the ER. In puromycin-synchronized HepG2 cells, we showed that some nascent partial-length apoB polypeptides are targeted for degradation by ubiquitination during translation before they are translocated into the ER. The evidence for the involvement of proteasomes in the degradation of the incomplete nascent chains corroborates that presented by Benoist and Grand-Perret (17) on microsomal triglyceride transfer protein inhibitor-treated cells. Our experiments demonstrate that proteasome-mediated degradation of newly synthesized partial-length apoB peptides during translation occurs also in the absence of microsomal triglyceride transfer protein inhibitors. Our observations on the fate of the full-length glycosylated apoB are supported by the study of Chen et al. (9) using [ 3 H]mannose labeling, although they did not show that the mannose-labeled apoB was actually targeted for degradation by ubiquitination. Herein we showed directly that full-length ubiquitinated apoB is glycosylated. We have provided three pieces of evidence for the presence of carbohydrate conjugates on ubiquitinated apoB: (i) ubiquitinated apoB retains its binding affinity to calnexin, a glycoprotein-specific chaperone (20 -22), an interaction that can be abolished by tunicamycin treatment; (ii) ubiquitinated apoB displays binding to concanavalin A; and (iii) the presence of glycan on ubiquitinated apoB is demonstrated by glycan-specific oxidative staining. The fact that the full-length ubiquitinated apoB has covalently linked carbohydrate indicates that the protein has completely translocated into the ER before it is retrotranslocated back to the cytosol for degradation. Finally, we note that calnexin is a membrane protein of the ER that associates with newly synthesized N-linked glycoproteins (21,22). The luminal domain of calnexin binds to glycoproteins only if they are at least monoglucosylated (22), and this lectin-like chaperone is thought to assist the folding and oligomeric assembly of glycoproteins and possibly to prevent premature degradation. Here we showed that if calnexin offers protection of apoB from pro-teasome-mediated degradation, the protection is incomplete, and ubiquitinated apoB actually has affinity for the chaperone.
It is difficult from the available evidence to quantify how much of the apoB tagged for degradation by ubiquitination consists of incomplete nascent chains and how much of it consists of full-length glycosylated apoB protein. As discussed earlier, although concanavalin A only binds to and thus selectively purifies glycosylated apoB that possesses accessible Dglucose and/or D-mannose moieties on the surface, the unbound fraction may be a mixture of unglycosylated and glycosylated apoB polypeptides. Furthermore, glycoproteins in the ER are partially deglycosylated by the action of peptide:N-glycanase, prior to proteolytic degradation in the cytosol (23)(24)(25), which would further impair concanavalin A binding and oxidative glycan detection. Therefore, the substantially lower amount of glycosylated versus unglycosylated apoB among the ubiquitinated apoB pool observed in these experiments represents an underestimate of the actual amount of the full-length, once fully glycosylated apoB protein that is targeted for proteasome degradation.
The partially translated polypeptides and the fully translocated apoB could be transported to the cytosolic compartment for proteasome-mediated degradation via one or more of three possible pathways (Fig. 6): (i) the partially translated nascent polypeptides are mostly cytosolic in location when they are ubiquitinated and are presumably fed into the proteasomes by retrotranslocation through the same translocon in which the nascent chain is partially partitioned; (ii) the fully translocated and fully glycosylated apoB-100 proteins could be retrogradetranslocated, i.e. via the same translocon through which they have entered the ER; or (iii) these molecules could have been completely discharged into the lumen of the ER and be translocated back into the cytosol in a retrograde fashion via a different retrotranslocon. Our experiments do not allow us to There are three possible pathways for the retrotranslocation of apoB to the cytosol for proteasomal degradation. 1, nascent partially synthesized apoB polypeptides are ubiquitinated cotranslationally and are fed to the proteasomes in a retrograde fashion. 2, full-length glycosylated apoB is retrotranslocated from the ER lumen back into the cytosol for ubiquitination via the same translocon through which the apoB was initially translocated into the ER. 3, the full-length glycosylated apoB is discharged into the ER and is retrotranslocated into the cytosol for ubiquitination and degradation via a different retrotranslocon. In 2 and 3, it is not clear if the N-or C-terminal end of the full-length protein will be the leading end of the retrotranslocated molecule. For simplicity, other proteins and processes involved in the biogenesis of apoB, e.g. microsomal triglyceride transfer protein, hsp70, Bip, and calnexin, have been omitted from this figure. Ub, ubiquitin; (Ub) n , polyubiquitin; CHO, carbohydrate. make a distinction among these possibilities, which will be the subject for future investigations.