Calnexin and other factors that alter translocation affect the rapid binding of ubiquitin to apoB in the Sec61 complex.

Several secretory proteins, including apolipoprotein B, have been shown to undergo degradation by proteasomes. We found that the rapid degradation of nascent apolipoprotein B in HepG2 cells was diminished but not abolished by the addition of any of three different inhibitors of proteasomes. Ubiquitin is conjugated to apolipoprotein B that is not assembled with sufficient lipids either during or soon after synthesis. In addition, we found that apolipoprotein B that has entered the endoplasmic reticulum sufficiently to become glycosylated can be degraded by proteasomes. Furthermore, we detected ubiquitin-apolipoprotein B that is associated with the Sec61 complex, the major constituent of the translocational channel. Treatment of cells with monomethylethanolamine or dithiothreitol decreased the translocation of apolipoprotein B and increased the proportion of ubiquitin-conjugated molecules associated with Sec61. Conversely, treatment of cells with oleic acid, which increased the proportion of translocated apolipoprotein B, decreased the amount of ubiquitin-apolipoprotein B in the Sec61 complex. Finally, we found that inhibition of the interaction between calnexin and apolipoprotein B decreases the translocation of apolipoprotein B, increases the ubiquitin-apolipoprotein B in the Sec61 complex, and increases the proteasomal degradation of glycosylated apolipoprotein B. Thus, ubiquitin can be attached to unassembled apolipoprotein B in the Sec61 complex, and this process is affected by factors including calnexin that alter the translocation of apolipoprotein B.

Several secretory proteins, including apolipoprotein B, have been shown to undergo degradation by proteasomes. We found that the rapid degradation of nascent apolipoprotein B in HepG2 cells was diminished but not abolished by the addition of any of three different inhibitors of proteasomes. Ubiquitin is conjugated to apolipoprotein B that is not assembled with sufficient lipids either during or soon after synthesis. In addition, we found that apolipoprotein B that has entered the endoplasmic reticulum sufficiently to become glycosylated can be degraded by proteasomes. Furthermore, we detected ubiquitin-apolipoprotein B that is associated with the Sec61 complex, the major constituent of the translocational channel. Treatment of cells with monomethylethanolamine or dithiothreitol decreased the translocation of apolipoprotein B and increased the proportion of ubiquitin-conjugated molecules associated with Sec61. Conversely, treatment of cells with oleic acid, which increased the proportion of translocated apolipoprotein B, decreased the amount of ubiquitin-apolipoprotein B in the Sec61 complex. Finally, we found that inhibition of the interaction between calnexin and apolipoprotein B decreases the translocation of apolipoprotein B, increases the ubiquitin-apolipoprotein B in the Sec61 complex, and increases the proteasomal degradation of glycosylated apolipoprotein B. Thus, ubiquitin can be attached to unassembled apolipoprotein B in the Sec61 complex, and this process is affected by factors including calnexin that alter the translocation of apolipoprotein B.
Apolipoprotein B (apoB) 1 is the large protein (Ͼ500 kDa) assembled with cholesterol and triglycerides into lipoprotein particles in hepatic and intestinal cells (1)(2)(3)(4). Full-length apoB100 is secreted on very low density lipoproteins from hepatic cells, and apoB serves as a ligand on low density lipoproteins for the low density lipoprotein receptor. Increased levels of apoB and low density lipoprotein-associated cholesterol in human plasma correlate with increased risks of coronary artery disease.
In HepG2 cells, a large amount of newly synthesized apoB is degraded (5)(6)(7)(8). Intracellular disposal appears to be the principal means of regulating the secretion of apoB. Since an inhibitor of calpains and cysteine proteases, N-acetyl-leucylleucyl-norleucinal (ALLN), can protect apoB from degradation, an uncharacterized cysteine protease was proposed to be responsible for the apoB degradation (6,7). ALLN also acts on proteasomes (9), and other chemicals that more specifically inhibit proteasomes also decrease the degradation of apoB (10 -12). In addition, ubiquitin has been found on apoB (10,11). Several proteins that enter the endoplasmic reticulum (ER) are degraded by cytosolic proteasomes (13)(14)(15)(16)(17)(18)(19). Recently, it has been shown that most if not all of these proteins undergo retrograde transport from the lumen of the ER back into the cytosol through a protein-conducting channel containing the Sec61 complex (20 -22). The Sec61 complex, which consists of ␣, ␤, and ␥ subunits, is the major constituent of the channel used for translocation into the ER. However, apoB has not been shown to be transported retrograde prior to degradation. Several experimental interventions have been shown to affect translocation, secretion, and degradation. Stimulating lipid synthesis by the addition of oleic acid (OA) increases the proportion of apoB that is fully translocated and secreted and decreases the amount that is degraded (7,23). In contrast, treatment of hepatocytes with monomethylethanolamine (MME) (24), dithiothreitol (DTT) (25,26), or inhibitors of microsomal triglyceride-transfer protein (MTP) (27) decreases translocation and secretion of apoB. The effect of calnexin, another chaperone in the ER, on the translocation of apoB has not been examined previously. Molecules of apoB that are not completely translocated into the ER can be degraded. Thrift et al. (6) first showed in a non-hepatic cell line that molecules of apoB that are not fully translocated into the ER are degraded by an ALLN-sensitive pathway. More recent data suggest that the ubiquitin-proteasome pathway could be responsible for degradation of incompletely translocated apoB (11,12).
Unlike typical secretory proteins, not all of the apoB molecules are completely translocated into the ER. Several investigators have found newly synthesized apoB with large domains exposed to the cytosol. These domains are accessible to proteolytic cleavage (26,28,29) and binding by antibodies (30) or cytosolic proteins such as Hsp70 (31). The cytosolic domains of such molecules of apoB would be accessible to ubiquitin-conjugating enzymes and proteasomes. The location of this incompletely translocated apoB that could be bound by ubiquitin is not known. Such molecules could reside entirely on the cytosolic side of the ER membrane, whether due to aborted forward translocation or retrograde transfer. Alternatively, these ubiquitin-conjugated apoB proteins (ub-apoB) might be spanning the ER membrane either in the lipid bilayer itself or within the translocational apparatus.
In this study, we examine the kinetics and action of ubiquitin and proteasomes in the degradation of different subsets of apoB. We also demonstrate a role for calnexin in the proteasomal degradation of glycosylated apoB. Furthermore, we investigate the intertwined roles of translocation and proteasomal degradation in regulating the secretion of apoB.

EXPERIMENTAL PROCEDURES
Reagents-␣ modification of Eagle's medium (␣MEM) and Dulbecco's modified Eagle's medium (DMEM) were purchased from Mediatech. Minimum essential medium (MEM) without methionine/cysteine, water-soluble oleic acid, 2-(methylamino)ethanol (monomethylethanolamine, MME), ALLN, and brefeldin A (BFA) were obtained from Sigma. Castanospermine (CST) was purchased from Calbiochem. DTT was obtained from Fisher. Protein A-and protein G-agarose beads were purchased from Life Technologies, Inc. Trans-label [ 35 S]methionine/ cysteine and [ 3 H]D-mannose were obtained from ICN Biochemicals. Carboxylbenzyl-leucyl-leucyl-leucinal (ZL 3 H, which is the same compound as MG135) was the generous gift of Hidde Ploegh (15). Lactacystin was obtained from E. J. Corey (32). The antibody that was used for immunoprecipitation of ubiquitin was the kind gift of Arthur L. Haas (33). Antibodies against Sec61␣ and Sec61␤, and the peptide that was used to raise the antibody against Sec61␤, were generous gifts from Walther Mothes and Tom Rapoport (34,35). Antibody against ubiquitin that was used for Western blotting was purchased from Boehringer Mannheim, antisera against human apoB were purchased from Calbiochem and Boehringer Mannheim, and antiserum against calnexin was obtained from StressGen.
Cell Culture and Pulse-Chase Labeling-HepG2 cells (obtained from American Type Culture Collection) were maintained at 37°C with 5% CO 2 in ␣MEM with 10% fetal bovine serum and 2 mM glutamine. Cells were used at about 90% confluence. In some experiments, proteasomal inhibitors, BFA, MME, OA, CST, or DTT were added to the preincubation, labeling, and chase media as indicated in the figure legends. HepG2 cells were preincubated in cysteine-, methionine-, and serumfree MEM for 1 h and then pulse-labeled with 50 -200 Ci/ml translabel [ 35 S]methionine/cysteine for 10 min or 1 h. After washing with phosphate-buffered saline (PBS) once, labeled cells were chased for various times in serum-free ␣MEM. For labeling with [ 3 H]mannose, glucose-and serum-free DMEM were used for the preincubation and pulse labeling. After 1 h preincubation, HepG2 cells were pulse-labeled with 50 Ci/ml [ 3 H]mannose for up to 1 h. The cells were washed with PBS once and then chased for different times in medium containing a 1,000-fold excess of unlabeled mannose.
For sequential immunoprecipitation, the immunoprecipitates were boiled for 10 min in elution buffer containing 2% SDS. The supernatants were diluted with 40-fold TXSWB, and a second immunoprecipitation was performed.
Electrophoresis and Immunoblotting-The immunoprecipitates were solubilized in sample buffer (4% SDS, 125 mM Tris-HCl, pH 6.8, 20% glycerol, 500 mM DTT) at 37°C, boiled for 10 min, and separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE). For labeling studies, the gel was treated with ENHANCE™ (DuPont), dried, and exposed to film. For Western blotting, proteins were transferred overnight onto nitrocellulose membranes, probed with antibody as indicated in the figure legends, and revealed by peroxidase conjugated to antirabbit or anti-goat IgG antibodies using the LumiGLO substrate kit (Kirkegaard & Perry Laboratories) according to the manufacturer's instructions. Some blots were stripped in buffer (2% SDS, 62.5 mM Tris-HCl, pH 6.8, 100 mM ␤-mercaptoethanol) at 70°C for 30 min prior to a second immunodetection.
Subcellular Fractionation and Flotation-HepG2 cells were suspended in homogenization buffer (10 mM Tris-HCl, pH 7.4, 250 mM sucrose, 1 mM PMSF, 50 M ALLN). Cell homogenization was carried out using a motorized Dounce homogenizer (30 strokes). The homogenate was centrifuged at 10,000 ϫ g for 10 min to pellet cellular debris and nuclei. Microsomes were isolated from the supernatant by centrifugation at 170,000 ϫ g for 2 h at 4°C. The microsomes were extracted with sodium carbonate, pH 11.5, for 1 h at 4°C (37), and the supernatant was subjected to sucrose gradient ultracentrifugation according to Boren et al. (38). (No apoB was detected in the pelleted membranes.) The gradient was unloaded into 12 fractions; the densities of each fraction were very similar to those previously reported with this method (38). ApoB was recovered from each fraction by immunoprecipitation and visualized by SDS-PAGE and Western blotting.
Sec61␤ Peptide Competition-HepG2 cells were preincubated for 2 h in cysteine-, methionine-, and serum-free MEM and then pulse-labeled with 200 Ci/ml trans-label [ 35 S]methionine/cysteine for 10 min. ALLN was included in the preincubation and labeling media. After washing with PBS, the labeled cells were lysed in digitonin buffer containing 150 mM NaCl. The peptide used to raise antibody against Sec61␤ was added to 1 aliquot of the cell lysate at a concentration of 50 g/ml. Immunoprecipitation with Sec61␤ antibody was carried out as described above.

RESULTS
ApoB Is Degraded Rapidly by Proteasomes-To define the kinetics and extent to which proteasomes degrade apoB, we performed pulse-chase analyses using HepG2 cells treated with ALLN, ZL 3 H, or lactacystin. HepG2 cells were pulse-labeled with [ 35 S]methionine/cysteine for 10 min in the presence or absence of different proteasome inhibitors and followed by various times of chase. ApoB was immunoprecipitated and analyzed by SDS-PAGE.
The three inhibitors had similar protective effects on apoB. Nascent apoB in untreated cells was dramatically degraded starting at a very early chase time; 40% of the apoB was degraded after 10 min of chase ( Fig. 1). Since it has been estimated that 14 -20 min are necessary to synthesize a fulllength apoB molecule (12,39), degradation must begin during or within a very short time after synthesis. Treatment of cells with lactacystin, ALLN, or ZL 3 H resulted in increased levels of apoB. After 30 min of chase, more than 75% of apoB molecules in the untreated cells were degraded, whereas about 70% of apoB molecules still remained intact in inhibitor-treated cells (Fig. 1). Analysis of the kinetics of degradation demonstrated a significant difference (p Ͻ 0.01) in rates of removal (slopes in Fig. 1B) in the presence or absence of an inhibitor. These data indicate that proteasomes degrade apoB molecules rapidly within a short time after synthesis.
Ubiquitin Binds to ApoB within a Short Time after Its Synthesis-Protein ubiquitination is a critical step in the degradation of most proteins by the proteasome pathway (40). Ubiquitin has been detected on apoB (10,11), although it is unclear how soon after synthesis this conjugation occurs. To determine when ubiquitination of apoB takes place, HepG2 cells were pulse-labeled with [ 35 S]methionine/cysteine and then chased for 0, 10, 30, and 60 min. Sequential immunoprecipitation was performed using antiserum against apoB followed by antiserum against ubiquitin. Ub-apoB appears at a very early time even in the sample without chase which suggests that ubiquitin is conjugated to apoB during or immediately after synthesis (Fig. 2).
Only Non-floating ApoB Is Associated with Ubiquitin-Association with lipids is a prerequisite for apoB secretion, and some early steps in particle formation can occur cotranslationally (38). We wondered whether apoB that is assembled with lipids into nascent particles would escape conjugation with ubiquitin. To investigate the correlation between ubiquitin binding and the amount of lipid bound to apoB, microsomes were prepared from HepG2 cells, extracted with sodium carbonate, and the proteins that were released were fractionated by sucrose gradient ultracentrifugation according to Boren et al. (38). ApoB molecules were recovered from each fraction of the gradient by immunoprecipitation and analyzed by Western blotting using antibodies either against apoB or ubiquitin.
As described previously (38), some of the apoB was found to float in the gradient with a density of high density lipoprotein (HDL)-like particles (Fig. 3A, fractions 3-9). When the blot was reprobed with anti-ubiquitin antiserum, ubiquitin was found only on the lipid-poor apoB recovered from the bottom of the gradient (Fig. 3B, fractions 1 and 2). The anti-ubiquitin immunoreactive smear with apparent higher molecular weight reflects the covalent attachment of multiple ubiquitin chains to apoB molecules. These results indicate that only apoB that is not assembled with sufficient lipids can be marked with ubiquitin.
Glycosylated ApoB Is Degraded by Proteasomes-It has been shown that incompletely translocated apoB53 molecules are degraded by an ALLN-sensitive pathway in non-hepatic cells (6). However, it is not known to what extent full-length apoB100 enters the ER of hepatic cells prior to its degradation by proteasomes. To determine whether apoB with at least some domains translocated is degraded by proteasomes, we investigated the degradation of glycosylated apoB. Furthermore, we tested whether N-linked high-mannose oligosaccharides protect apoB or become attached to those molecules of apoB that have escaped proteasomal degradation. We pulse-labeled HepG2 cells with [ 3 H]mannose for 10 min, washed the cells with PBS, and then incubated the cells in medium with a 1,000-fold excess of unlabeled mannose for various times. Glycosylated apoB was degraded in a manner similar to 35 S-labeled apoB (Fig. 4). When HepG2 cells were treated with 10 M lactacystin, about three times as much glycosylated apoB remained after 240 min indicating that glycosylated apoB can be degraded by proteasomes.
Calnexin Protects ApoB from Degradation-Calnexin is a molecular chaperone that is localized in the ER (41). By associating transiently with nascent influenza hemagglutinin, calnexin can regulate protein folding, oligomerization, and degradation (42). Calnexin can bind to a number of glycoproteins including apoB, although the role of calnexin in apoB biogenesis is not defined (41,43). When cells are preincubated with castanospermine (CST), the trimming of glucose residues is prevented and the binding of glycoproteins by calnexin is inhibited (42). Since glycosylated apoB is degraded by proteasomes, we investigated the role of calnexin in the degradation of apoB.
First, we confirmed the effect of CST on the binding of calnexin to apoB. HepG2 cells were labeled for 1 h with [ 35 S]methionine/cysteine in the presence or absence of 1 mM CST, and the association of calnexin with apoB was measured by co-immunoprecipitation. As shown in Fig. 5A, treatment with CST decreased the co-immunoprecipitation of apoB by about 45%. However, CST had no effect on the immunoprecipitation of calnexin.
We next examined the role of calnexin in the degradation of glycosylated apoB by proteasomes. HepG2 cells were pulse- ]methionine/cysteine for 10 min and then incubated in chase medium for 0, 10, 30, or 60 min. BFA (1 g/ml) was present in the preincubation, labeling, and chase media. ApoB was immunoprecipitated. Sequential immunoprecipitation was performed by mixing immunoprecipitates with elution buffer and boiling for 10 min. The supernatant was diluted with an excess of TXSWB, and immunoprecipitation was carried out with anti-ubiquitin antiserum, followed by SDS-PAGE analysis.
FIG. 3. Non-floating apoB is associated with ubiquitin. Microsomes were prepared from HepG2 cells and extracted with sodium carbonate, pH 11.5, and the supernatant was subjected to sucrose gradient ultracentrifugation (38) as described under "Experimental Procedures." ApoB was recovered from each fraction by immunoprecipitation and resolved by 8% SDS-PAGE. Western blotting analysis was carried out with anti-apoB antiserum. The blot was stripped prior to detection using anti-ubiquitin antiserum. Top panel, Western blot analysis with anti-apoB antiserum. Bottom panel, the same blot probed with anti-ubiquitin antiserum. Ub-apoB, ubiquitin-conjugated apoB. labeled in the presence or absence of 1 mM CST with [ 3 H]mannose and then chased with medium containing 1,000-fold unlabeled mannose with or without CST for 1 h. In the presence of CST, over twice as much apoB was degraded in 1 h compared with untreated cells (Fig. 5B). To ascertain the pathway that leads to the increased loss of apoB caused by the treatment of CST, we repeated the above experiment in the presence of lactacystin. The addition of lactacystin completely inhibited the increased degradation of apoB caused by CST (Fig. 5C). These results suggest that the interaction between apoB and calnexin protects apoB from degradation by proteasomes.
Ub-ApoB Is Associated with the Sec61 Complex-Not all molecules of apoB are fully translocated into the ER upon completion of translation; many have large domains exposed to the cytosol (26, 28 -31) where they could become bound by ubiquitin. We wondered if ubiquitin binds to this partially translocated apoB while it is still in the Sec61-containing translocation channel.
First, we established the specificity of co-immunoprecipitating apoB with antibody against Sec61␤. Digitonin extracts were prepared from HepG2 cells treated with ALLN and labeled with [ 35 S]methionine/cysteine. Under these conditions, the Sec61 complex (consisting of ␣, ␤, and ␥ subunits) remains intact (35). Upon the addition of antibody against Sec61␤, apoB was co-immunoprecipitated (Fig. 6A, second lane). To demonstrate that this material represented Sec61␤-apoB conjugates, we added Sec61␤ peptide to samples before the addition of antibody against Sec61␤. The co-immunoprecipitation of apoB with Sec61␤ was competitively blocked by an excess of peptide (Fig. 6A, 3rd lane). In addition, if non-immune serum was substituted for antibody against either Sec61␤ or apoB, no signal was detected (data not shown). Furthermore, apoB is still associated with Sec61␤ in the presence of up to 250 mM NaCl (data not shown). These data indicate that the antibody against Sec61␤ specifically co-immunoprecipitates apoB that is associated with the Sec61 complex.
By using these conditions to co-immunoprecipitate apoB with Sec61␤, we next determined whether some of the apoB associated with Sec61␤ is marked with ubiquitin and degraded by proteasomes. HepG2 cells were preincubated in the presence or absence of ALLN for 2 h and then lysed in digitonin buffer. Antibodies against Sec61␤ were added to the cleared lysates. The washed Sec61␤ immunoprecipitates were dissociated in 2% SDS and diluted in TXSWB, and the supernatants were incubated with antibody against apoB. The apoB immunoprecipitates were resolved by SDS-PAGE, and Western blotting was performed using antibody against ubiquitin (Fig. 6B). We detected an accumulation of ub-apoB associated with Sec61␤ in cells treated with ALLN. Similar ub-apoB conjugates can also be recovered with antibody against the ␣ subunit of the Sec61 complex (Fig. 6C). These data suggest that conjugation of ubiquitin with apoB can occur when it is associated with the Sec61 complex. Moreover, the marked decrease in ub-apoB in the Sec61 complex in cells that were not treated with ALLN (Fig.  6B, first lane) implies that proteasomes degrade at least some of these ub-apoB conjugates.
Lipids That Alter the Translocation of ApoB Affect the Amount of Ub-ApoB Found with the Sec61 Complex-Two lipids have been found to affect the proportion of newly synthesized apoB that is translocated fully into the ER: MME inhibits whereas OA increases the translocation and secretion of apoB (7,23,24). We examined the effect of OA and MME on ub-apoB associated with Sec61␤. HepG2 cells were incubated with OA (0.8 mM, 2 h) or MME (0.4 mM, 16 h) in the presence of ALLN (50 g/ml, 2 h). Western blotting using antibodies against ubiquitin or apoB were performed after sequentially immunoprecipitating Sec61␤ and apoB (Fig. 7). More apoB is found with Sec61␤ in cells treated with MME (Fig. 7A, compare lanes  1 and 2). Treatment with MME also increased the amount of ub-apoB associated with Sec61 (Fig. 7B, compare lanes 1 and  2). In contrast, treatment with OA increased apoB transloca- tion so that fewer molecules of apoB were associated with Sec61 (Fig. 7A, compare lanes 1 and 3). This effect on translocation corresponded with a decrease in ub-apoB in the Sec61 complex (Fig. 7B, compare lanes 1 and 3). The effects of these two lipids suggest that alterations of apoB translocation affect the amount of ub-apoB associated with Sec61.
Inhibition of Calnexin Binding Increases the Amount of ApoB and Ub-ApoB That Co-precipitates with the Sec61 Complex-We next investigated the effect of CST on apoB translocation and ub-apoB associated with the Sec61 complex. The Sec61␤-apoB immunoprecipitates from CST-and ALLNtreated HepG2 cells were compared with those from HepG2 cells treated with ALLN only. CST increased the amount of apoB in the Sec61 complex (Fig. 8A). This effect correlated with an increase in ub-apoB associated with Sec61 (Fig. 8B). Thus, inhibition of the interaction between calnexin and apoB increased the ub-apoB in the Sec61 complex.
DTT, Which Results in Misfolded ApoB, Increases Ub-ApoB Associated with the Sec61 Complex-Degradation of apoB is greatly stimulated if the proper formation of disulfide bonds is disrupted by DTT. Treatment of HepG2 cells with DTT also has been correlated with a decrease in the translocation of apoB into the ER (25,26). We examined the effect of DTT on the amount of apoB and ub-apoB associated with Sec61␤. ALLNtreated HepG2 cells were incubated in 2 mM DTT for 0, 2.5, or 5 min. Increasing the length of incubation in DTT caused a reduction in the total amount of intracellular apoB (Fig. 9A). In addition, increasing exposure to DTT was associated with little change in the quantity of apoB associated with the Sec61 complex (Fig. 9B). In view of the overall decrease in intracellular apoB, however, the proportion of intracellular apoB associated with Sec61 increased with increasing exposure to DTT. Even more notable was the dramatic increase in ub-apoB associated with Sec61␤ seen with longer exposure to DTT (Fig.  9C). Hence, the disruption of proper disulfide bond formation and folding caused by DTT increases the ub-apoB associated with the Sec61 complex. DISCUSSION Yeung et al. (10) first demonstrated that apoB is conjugated to ubiquitin and degraded by proteasomes in an ATP-dependent manner. Two other groups have provided additional evidence for the role of the ubiquitin-proteasome pathway in the degradation of apoB (11,12). Likewise, we have found that  2nd and 3rd lanes) were used for immunoprecipitation (Ip); in one sample, the peptide used to raise antibody against Sec61␤ was added (50 g/ml) prior to the immunoprecipitation. The washed immunoprecipitates were resolved by SDS-PAGE and fluorography. B, HepG2 cells were treated with 50 g/ml ALLN for 0 or 2 h and then lysed in digitonin buffer. The samples were subjected to sequential immunoprecipitation with antibodies against first Sec61␤ and then apoB. The immunoprecipitates were resolved by SDS-PAGE and visualized by Western blotting (WB) using antibodies against ubiquitin. The blot was stripped and reprobed with antibody against apoB (data not shown); the position of apoB is marked. C, ALLN-treated HepG2 cells were lysed in TX2 buffer. The sample was subjected to sequential immunoprecipitation with antibodies against first Sec61␣ and then apoB. The immunoprecipitates were resolved by SDS-PAGE and visualized by WB using antibodies against apoB or ubiquitin. proteasomes rapidly degrade a significant proportion of newly synthesized apoB. The kinetics of this proteasomal degradation is rapid, proteolysis begins during or within a very short time after synthesis and continues for over 1 h. This early degradation is consistent with recent evidence that proteolysis of incompletely elongated apoB can occur (12). We found that apoB is bound by ubiquitin either during or very soon after translation and disappears rapidly during the chase period. Thus, degradation by the proteasome, rather than ubiquitin conjugation, is the rate-limiting step.
In this study, the rapid intracellular degradation of apoB is incompletely inhibited by any of three inhibitors of proteasomes. The protection of apoB by ALLN in this study is similar to previous studies in intact HepG2 cells (44,45). Possibly these inhibitors cannot sufficiently block the activity of proteasomes. However, no greater protection was detected even when 10 times the concentration of lactacystin was used (data not shown). In contrast, the degradation of T-cell receptor ␣ chains by proteasomes is nearly completely inhibited by concentrations of ALLN or lactacystin similar to those used in this study (19). An alternative explanation is that a non-proteasomal pathway might proteolyze apoB in the presence of these inhibitors. In particular, a pathway that is insensitive to ALLN has been shown to degrade apoB (45,46).
In HepG2 cells, assembly of lipoprotein particles of density similar to HDL appears to begin cotranslationally (38). Therefore, those molecules of apoB that are not assembled with sufficient lipids early in their biogenesis could be targeted for degradation. Indeed, we did not find ubiquitin associated with apoB assembled into HDL-like particles; instead, ubiquitin was found only on lipid-poor apoB. It is unknown what proportion of the lipid-deficient apoB is conjugated to ubiquitin since the Western blots cannot be compared quantitatively.
Glycosylated apoB proteins (that is molecules of apoB with some or all of their domains translocated into the lumen of the ER) also are degraded by proteasomes. Glycosylation of apoB in the ER does not correlate with assembly with sufficient lipid for secretion since glycosylated but not floating apoB can be degraded by the ubiquitin-proteasome pathway. Furthermore, our data demonstrate that apoB that has entered the ER far enough to be recognized by the glycosylation machinery still can undergo proteasomal degradation. This finding raises new questions such as whether the glycosylated region of apoB is degraded in the ER lumen or transferred into the cytosol and where the sugar residues are removed. Glycosylated regions of apoB might be deglycosylated prior to retrograde transfer just as major histocompatibility complex class I heavy chain molecules (20).
What is the topology of the ubiquitin-conjugated apoB in the bottom fraction of the sucrose gradient? Since these proteins were isolated by carbonate extraction of microsomes, the apoB could reside in the following: 1) the lumen, 2) the cytoplasmic side of the ER membrane, or 3) spanning the membrane in a proteinaceous channel. In this study, we demonstrate that ubiquitin can be conjugated to apoB while it still is associated with the Sec61 complex. It is not known whether the ub-apoB that co-immunoprecipitates with the Sec61 complex resides FIG. 8. Castanospermine increases the amount of apoB and ub-apoB associated with the Sec61 complex. HepG2 cells were incubated with 50 g/ml ALLN and 1 mM CST for 2 h. The samples were subjected to sequential immunoprecipitation (Ip) with Sec61␤ antibody followed by antibody against apoB. A, Western blot (WB) using antibody against apoB. B, the same blot was stripped and probed with antibody against ubiquitin. FIG. 9. Treatment of cells with DTT markedly increases the amount of ub-apoB associated with the Sec61 complex. HepG2 cells were treated with 50 g/ml ALLN for 2 h and 2 mM DTT for 0, 2.5, or 5 min. A, apoB immunoprecipitates (Ip) were analyzed by Western blotting (WB) using antibody against apoB. B and C, samples were subjected to sequential immunoprecipitation with Sec61␤ antiserum followed by antibody against apoB. Western blot analyses were performed with antibodies against apoB (B) or ubiquitin (C). within a translocation-competent channel. Although the ub-apoB is physically in contact with Sec61␣ and Sec61␤, other channel proteins that might be necessary for anterograde or retrograde transport of apoB could be missing. In addition, our data do not indicate directly whether ub-apoB associated with Sec61 is on its way into or out of the ER. Proteins such as major histocompatibility complex class I heavy chains (15,20) and carboxypeptidase Y (16,22) that completely enter the ER also are degraded by proteasomes. These proteins undergo retrograde transport back into the cytosol via a channel of which the Sec61 complex is a constituent. However, apoB is an atypical secretory protein since not all newly synthesized molecules are fully translocated into the ER (26, 28 -31). Those chains of apoB that are not translocated completely into the ER have cytosolic domains that can be bound by ubiquitin. In the presence of an inhibitor of MTP, apoB can be degraded before it is completely synthesized (12). Hence, ub-apoB probably forms while apoB is being translocated into the ER. This mechanism of degradation would circumvent the need to unfold full-length apoB in the ER for retrograde transport and prevent the accumulation of poorly lipidated apoB that might be insoluble in the aqueous environment of the ER. Thus, apoB might act more like a membrane-spanning protein such as the cystic fibrosis transmembrane conductance regulator (13,14) rather than a secretory protein en route to degradation by the ubiquitinproteasome pathway.
In this study, factors that altered the translocation of apoB into the ER had inverse effects on ub-apoB associated with Sec61␤ that parallel their effects on degradation. Treatment of HepG2 cells with MME decreased the translocation of apoB and increased the ub-apoB associated with Sec61. In contrast, OA increased the translocation of apoB and decreased the ub-apoB that co-precipitates with the Sec61 complex. Finally, DTT caused an increase in apoB with the Sec61 complex that was especially significant in view of the concomitant decrease in total apoB. This effect of DTT on the translocation of apoB was accompanied by a marked increase in ub-apoB associated with Sec61. Thus, our data indicate that alterations in the translocation of apoB lead to inverse changes in ub-apoB in the Sec61 complex. Other studies can be viewed as supporting this relationship between translocation and degradation. MTP is necessary for the translocation and secretion of apoB, and chemical inhibitors of MTP induce cotranslational degradation of apoB (12). Similarly, Hsp70 has been shown to increase the proteasomal degradation of apoB (11) perhaps by retarding the translocation of cytosolic domains of apoB.
Calnexin interacts with newly synthesized glycoproteins including apoB to facilitate proper trimming of glucose residues, folding, and assembly (41). By using castanospermine, we found that interfering with trimming of glucose residues, and thereby inhibiting binding to calnexin, was associated with over twice as much degradation of glycosylated apoB by proteasomes. Inhibition of the interaction between calnexin and apoB also increased the apoB and ub-apoB associated with Sec61. Since calnexin only binds glycosylated apoB, the effect of CST on total apoB is less dramatic than the effects of OA, MME, and DTT. Thus, although core glycosylation by itself does not protect apoB from degradation, interacting with calnexin to ensure proper trimming of glucose residues and, presumably, to promote correct folding and complete translocation, does safeguard glycosylated apoB from degradation. A similar role of calnexin in maturation and protection from degradation has been reported for other nascent proteins including subunits of T-cell receptors (47), major histocompatibility complex class I and class II complexes (48,49), influenza hemagglutinin (42), and nicotinic acetylcholine receptors (50).
Calnexin probably acts in concert with other molecular chaperones to facilitate the translocation, folding, and assembly of nascent apoB. Since apoB appears to be cotranslationally assembled into lipoprotein particles and the apoB that is unassembled with significant lipids can be bound by ubiquitin, the glycosylated chains of apoB that interact with calnexin might achieve a conformation that promotes their assembly with lipids catalyzed by microsomal triglyceride transfer protein (MTP) (27). Furthermore, calnexin may directly facilitate the forward translocation of apoB or prevent its retrograde transport into the cytosol. Calnexin could act cooperatively or successively with MTP, protein disulfide isomerase, BiP, and other molecular chaperones to promote forward translocation and assembly of secretion-competent apoB.