A two-site model for ApoB degradation in HepG2 cells.

Newly synthesized apolipoprotein B (apoB) undergoes rapid degradation in a pre-Golgi compartment in HepG2 cells. A major site of this early degradation seems to be on the cytosolic side of the endoplasmic reticulum (ER) membrane and is sensitive to N-acetyl-leucinyl-leucinyl-norleucinal (ALLN), which can inhibit neutral cysteine proteases and/or proteasome activity. Oleate (OA) treatment, which facilitates translocation of nascent apoB across the ER membrane, also reduces early degradation. In the present studies, we have used brefeldin A (BFA), which inhibits vesicular transport from the ER to the Golgi, to demonstrate that apoB can also be degraded by an ER luminal proteolytic activity that is distinct from the ALLN-sensitive proteases. Thus, when BFA-treated HepG2 cells were co-treated with ALLN, which protects apoB but does not facilitate its translocation into the ER lumen, degradation of newly synthesized apoB was significantly reduced compared with cells incubated with BFA alone. However, apoB degradation was rapid and complete when OA was added to media containing either BFA or ALLN/BFA. These results suggested that OA, by increasing translocation of nascent apoB into the ER lumen, exposed apoB to an ALLN-resistant proteolytic pathway. When we incubated HepG2 cells with dithiothreitol (DTT)/OA/BFA or DTT/OA/ALLN/BFA, degradation of apoB was inhibited. Furthermore, addition of DTT resulted in the accumulation of a 70-kDa amino-terminal fragment of apoB. Both full-length and amino-terminal apoB were degraded if DTT was removed from the incubation media; both were secreted if only BFA was removed. Thus, even after apoB is translocated into the ER lumen (thereby avoiding the initial proteolytic pathway), it can potentially be degraded by a lumenal proteolytic process that is ALLN-resistant but DTT-sensitive. The present results, together with previous studies, suggest that at least two distinct steps may be involved in the posttranslational degradation of apoB: 1) the first occurs while apoB is partially translocated and is ALLN-sensitive; and 2) the second occurs in the ER lumen and is DTT-sensitive. Finally, our results support the hypothesis that degradation of partially translocated apoB generates a 70-kDa amino-terminal fragment that is mainly degraded in the ER lumen by a DTT-sensitive pathway.

ApoB 1 secretion from cultured liver cells is regulated mainly at the posttranslational level. Thus, apoB mRNA levels are relatively stable under many conditions, whereas secretion of apoB-containing lipoproteins is altered (1)(2)(3)(4)(5). The impact of this posttranslational regulation is demonstrated by the observations that only a small to moderate proportion of the newly synthesized apoB is eventually secreted from primary rat hepatocytes (6,7), McArdle cells (8), and HepG2 cells (9). A major portion of newly synthesized apoB undergoes rapid intracellular degradation in a pre-Golgi or ER compartment (10 -12) in HepG2 cells as well as in apoB cDNA-transfected Chinese hamster ovary cells (13). ApoB can be protected from early intracellular degradation by incubation of cells with OA (9) or with the aldehydic tripeptide ALLN (13)(14)(15). OA treatment seems to stimulate apoB secretion by providing triglyceride to newly synthesized apoB and facilitating apoB translocation across the ER membrane (14). OA-induced translocation is probably the result of enhanced interaction between apoB and the microsomal triglyceride transfer protein (16). ALLN does not seem to facilitate the translocation of apoB across the ER membrane (14,15). Direct protection of apoB, either by inhibition of a neutral cystine protease or proteasome activity (17), is the likely mechanism underlying the effect of ALLN.
The ALLN-sensitive degradation of nascent apoB is thought to occur on the cytosolic side of the ER membrane or in the cytoplasm. This is based on the finding of 70 -85-kDa aminoterminal fragments of apoB in the ER lumen and in the media of HepG2 cells (12,18,19). A cytosolic site for the protease(s) involved in early degradation of apoB is also compatible with the ability of OA to prevent degradation by facilitating translocation of nascent apoB across the ER membrane (14). On the other hand, later, alternative sites of degradation have been postulated from studies in rat hepatocytes and McArdle cells (20,21). Our previous observation (11) that OA treatment did not prevent degradation of apoB in cells that were co-treated with BFA, a drug that prevents vesicular transport from the ER to the Golgi, also suggested a posttranslocation ER-luminal site for apoB degradation. The aim of the present study was to further investigate this potential second site of apoB degradation pathway in HepG2 cells. Two specific questions were addressed: (a) Can apoB undergo degradation after translocation into the ER lumen? and (b) Is this second degradative process sensitive to ALLN?

EXPERIMENTAL PROCEDURES
Materials-L- [4, H]Leucine (135 Ci/mmol; catalog number TRK.683) was purchased from Amersham Corp. Monospecific antihuman apoB antiserum was raised in a rabbit. Monoclonal apoB antibodies (M19 and M47) were kindly provided by Dr. Linda Curtiss (Scripps Research Institute, La Jolla, CA). Protein A-Sepharose CL 4B was obtained from Pharmacia Biotech Inc. Minimum essential medium (MEM), nonessential amino acids, sodium pyruvate, and penicillin/streptomycin were from Life Technologies, Inc. laboratories. Fetal bovine serum was from Integen (Purchase, NY). Leucinefree medium was generated from a minimum essential selection kit (Life Technologies, Inc., catalog number 300 9050AV). Leupeptin and pepstatin A were from Peninsula Laboratories, Inc. (Belmont, CA). Bovine serum albumin (BSA) (essentially fatty acid-free), DTT, and OA (sodium salt) (catalog number 07501) were from Sigma. ALLN was from Boehringer Mannheim. All other chemicals were of the highest purity available.
Growth of Cells-HepG2 cells, obtained from ATCC, were grown in 35-mm dishes that had been coated with collagen. The cells were maintained in MEM containing 0.1 mM nonessential amino acids, 1 mM sodium pyruvate, penicillin (100 units/ml), streptomycin (100 g/ml), and 10% fetal bovine serum for 4 days with medium replenishment at day 3. The medium was then changed to serum-free MEM experimental medium containing 1.5% BSA as described below.
Immunoprecipitation-Immunoprecipitation of apoB in medium and cell homogenates was carried out exactly according to the method of Dixon et al. (9). Briefly, samples were mixed with NET buffer (150 mM NaCl, 5 mM EDTA, 50 mM Tris-HCl, pH 7.4, 0.5% Triton X-100, and 0.1% SDS) and an excess amount of anti-apoB antiserum, and the mixture was incubated on a shaker for 10 h at 4°C. Protein A-Sepharose CL 4B was added to the mixture, the incubation was continued for an additional 3 h, and the beads were washed extensively. ApoB was extracted from the protein A pellet with sample buffer (0.125 M Tris-HCl, pH 6.8, 4% SDS, 20% glycerol, and 10% mercaptoethanol) by boiling. An aliquot of each sample was run on SDS-polyacrylamide gel electrophoresis (3-15% gradient gel). The gel was treated with autofluor (National Diagnostics, Atlanta, GA) and, after drying, was exposed to a film (Kodak X-Omat AR) at Ϫ80°C.

RESULTS AND DISCUSSION
In BFA-treated Cells, OA Treatment Abolishes the Protective Effect of ALLN on Newly Synthesized apoB-We previously reported that OA protects apoB from intracellular degradation and stimulates apoB secretion from HepG2 cells (9). However, when exit from the ER compartment was blocked by BFA, OA lost its ability to protect newly synthesized apoB (11). Because we had also demonstrated that ALLN protects apoB directly from intracellular degradation without affecting the translocation of apoB, we sought to examine whether ALLN would protect apoB if, after the nascent protein was translocated across the ER membrane, its exit from the ER compartment was blocked.
HepG2 cells were preincubated with either BSA or BSA/ ALLN for 1 h, pulse-labeled with [ 3 H]leucine for 10 min, and chased in the presence of BFA for 1 h. Cells that received only BSA during preincubation and labeling were chased in either BSA-or BSA/OA-containing medium. Cells that received BSA/ ALLN during preincubation and labeling were chased in either ALLN-or OA/ALLN-containing medium. As expected, no apoB was secreted into the medium from cells chased in the presence of BFA, as determined by immunoprecipitation (data not shown). The results (Fig. 1A) show that addition of BFA during the chase period in BSA-treated cells resulted in nearly complete intracellular degradation of newly synthesized apoB after a 60-min chase. Cells treated with ALLN alone showed a markedly reduced degradation of apoB during the BFA chase. On the other hand (and as we reported previously), BSA-treated cells to which OA was added together with BFA at the start of the chase did not differ from BSA-only-treated cells. Finally, when OA and BFA were added to the ALLN-treated cells at the start of the chase, nearly complete degradation of apoB was also observed, indicating a loss of ALLN-associated protection of apoB. OA does not affect the ability of ALLN to protect apoB in cells not treated with BFA; in fact, OA and ALLN have additive effects on apoB secretion (14). The results suggested, therefore, that (1) ALLN reduced apoB degradation in BFAtreated cells, possibly by inhibiting the initial degradation of partially translocated apoB (14,15) and (2) OA, by increasing apoB translocation across the ER membrane (see experiment below), exposed apoB to an ER luminal proteolytic process that was ALLN-resistant. The failure of OA to protect apoB in BFA-treated cells was not the result of failure to stimulate new triglyceride (11). Furthermore, albumin was not degraded in the presence of BFA under any co-incubation condition (Fig.  1B), suggesting that both of these degradative pathways were relatively specific for apoB.
Because BFA treatment also causes retrograde flow of Golgi contents into the ER, the proteolytic activity observed with ALLN/OA/BFA could have originated in the Golgi compartment. Several lines of evidence indicate that this was unlikely: 1) previous in vitro experiments have demonstrated that an isolated ER compartment contains apoB-specific proteolytic activity, which is not present in an isolated Golgi compartment (11); 2) when retrograde flow from Golgi to ER was blocked with nocodazole, BFA treatment still resulted in a nearly complete degradation of apoB (11); and 3) in the present study, when HepG2 cells were chased in the presence of OA/ALLN and Monensin (which blocks vesicular transport from the Golgi to the plasma membrane), newly synthesized apoB, which was trapped intracellularly under this condition, was not significantly degraded (data not shown).
In contrast to HepG2 cells, apoB seems to be stable in BFAtreated rat hepatocytes (21), suggesting that significant degradation of newly synthesized apoB in rat hepatocytes is not associated with the ER compartment. Indeed, Wang et al. (21) concluded that apoB degradation in rat hepatocytes might occur in a post-ER compartment. We have no data directly addressing these apparent differences between HepG2 cells and primary rat hepatocytes.
DTT Treatment Inhibits Degradation of Full-Length apoB and Is Associated with Accumulation of a 70-kDa Amino-terminal Fragment of apoB-In screening agents that might inhibit apoB degradation in BFA-treated HepG2 cells, we first found that DTT protected apoB from intracellular degradation in untreated cells. Although DTT seems to affect the secretion of proteins containing disulfide bonds (22), Shelness and Thornburg (23) demonstrated recently that once the aminoterminal disulfide bonds are formed, apoB is no longer sensitive to DTT treatment. Therefore, in our experiments, DTT was always added after the labeling period.
HepG2 cells were first labeled with [ 3 H]leucine for 10 min and chased in the presence of either BSA or DTT for 60 min. Fig. 2 shows that DTT did not affect the secretion of ␣-antitrypsin ( Fig. 2A), which does not contain any disulfide bonds, but did inhibit the secretion of albumin (Fig. 2B), which contains disulfide bonds. DTT, added after labeling was completed, increased the secretion of apoB 1.5 Ϯ 0.3-fold (mean of four separate experiments) (Fig. 2C). Increased apoB in the media was associated with protection of apoB from intracellular degradation in the DTT-treated cells (Fig. 2C); significantly more apoB100 was detected in cells chased in the presence of DTT (2.2 Ϯ 0.5-fold; n ϭ 4) compared with BSAtreated cells. In both the cell and the medium, DTT also increased a band at the top of the gel; this is probably aggregated apoB. Additionally, three other bands with M r of approximately 330,000, 170,000, and 70,000 were co-immunoprecipitated with full-length apoB by anti-apoB antibody from DTT-treated cells but not from BSA-treated cells. The nature of the two larger proteins with M r of 330,000 and 170,000 is not known, and they were not reproducibly detected. These two proteins were co-immunoprecipitated with albumin by anti-human albumin antibody in DTT-treated cells (Fig. 2B). Furthermore, the 330-kDa band has been observed in media immunoprecipitation in all of our previous studies and is precipitated by Sepharose beads alone; it is not apoB. The 70-kDa protein, as reported previously (12,18,19), was only precipitated by antibody to apoB. It was not precipitated by anti-albumin nor was it seen on Western blotting of anti-apoB immunoprecipitates with anti-albumin antibody (data not shown). The 70-kDa band was found to be an aminoterminal fragment of apoB in the following experiment.
HepG2 cells were incubated with either BSA or DTT for 2 h and lysed. Aliquots of the lysate were run on a SDS-polyacrylamide gel electrophoresis, transferred to a nitrocellulose membrane, and blotted with the following: 1) a polyclonal antibody to full-length apoB (Poly), 2) a monoclonal antibody against the amino-terminal fragment of apoB (Mab-N), or 3) a monoclonal antibody against carboxyl-terminal apoB (Mab-C). The bands were developed with secondary antibody linked to ECL reagents. As shown in Fig. 3, the antibody to full-length apoB detected mainly full-length apoB in BSA-treated cells; a very small amount of the 70-kDa protein was detected. With DTT treatment, however, detection of the 70-kDa fragment increased significantly. The results with the amino-terminal antibody were similar to those obtained with the full-length apoB antibody. In contrast, the carboxyl-terminal antibody detected only full-length apoB in both BSA-treated and DTT-treated cells. As discussed above, the two bands migrating just ahead of full-length apoB may be other apoB peptides or nonspecifically cross-reacting proteins.
Together, these experiments (Figs. 2C and 3) indicated that DTT treatment could protect both full-length apoB and its amino-terminal fragment from intracellular degradation in HepG2 cells incubated in BSA. These findings raised several questions: 1) Would DTT protect apoB from degradation in BFA-treated cells? 2) Was DTT inhibiting a proteolytic process that was different from the one affected by ALLN? 3) Was the DTT-sensitive degradation occurring in the lumen of the ER? and 4) What was the relationship of the 70-kDa amino-terminal fragment to the degradation of full-length apoB? Several experiments were performed to examine these questions. Fig. 1, the addition of OA to ALLN-treated cells in the presence of BFA resulted in a nearly complete degradation of apoB. This is in sharp contrast to the additive effects of OA and ALLN on the protection of apoB in native cells not treated with BFA (14). To determine if DTT inhibited a proteolytic pathway distinct from ALLN, HepG2 cells were pretreated with ALLN for 1 h, pulse-labeled for 10 min with pothesis that DTT inhibits a process that is distinct from the ALLN-sensitive proteolytic pathway. Furthermore, these results, together with those presented in Fig. 1, suggest that the DTT-sensitive proteolytic process is situated in the ER lumen and exerts its effects on apoB that has been translocated across the ER membrane. The latter conclusion is based on two observations: 1) the addition of OA, which facilitates translocation of nascent apoB into the ER lumen (14), abolished the ability of ALLN to prevent apoB degradation in BFA-treated cells but did not affect DTT-associated protection; and 2) DTT, but not ALLN, prevented degradation of the 70-kDa aminoterminal fragment of apoB, which is believed to efficiently translocate into the ER lumen (18).

The DTT-sensitive Proteolytic Pathway Is Active in the ER Lumen and Degrades Both Full-Length apoB and the 70-kDa
Amino-Terminal Fragment of apoB-To further support the proposal that DTT was inhibiting proteolysis of apoB that had already translocated across the ER membrane, the following experiment was carried out. HepG2 cells were labeled with [ 3 H]leucine for 10 min and chased in serum-free medium in the presence of OA, DTT, and BFA for 10 min or 2 h. To determine the topology of newly synthesized apoB under this experimental condition, ER was isolated at both the 10-min and 2-h time points by the ball-bearing homogenization/sucrose gradient ultracentrifuge method described in our previous study (11). Proteinase K sensitivity experiments were performed using the isolated ER. As shown in Fig. 5, at 10-min chase, approximately 70% of newly synthesized apoB was sensitive to proteinase K treatment. After 2-h chase in the presence of OA/ DTT/BFA, this apoB had lost its sensitivity to proteinase K treatment, indicating that translocation into the ER lumen had occurred. In addition, the 70-kDa fragment, presumably generated during the degradation of initially labeled apoB, was easily detected at this time and was also insensitive to proteinase K treatment. When the ER preparations were treated with proteinase K plus Triton X-100, apoB was completely sensitive to proteinase K at both times.
In a separate experiment using the same incubation and labeling protocol and a 2-h chase, cells in some dishes were lysed at the end of chase, and intracellular apoB was determined. Cells in the remaining dishes were divided into two groups. For the first group, the chased medium was changed to serum-free medium in the presence of OA and BFA; DTT was withdrawn from the incubation. For the second group, the chase medium was changed to serum-free medium in the presence of OA and DTT; BFA was withdrawn from the incubation. The cells were then chased for an additional hour. At the end of the chase, medium was collected, and cells were lysed for apoB immunoprecipitation. As shown in Fig. 6, after the initial labeling and 2-h chase in the presence of OA, BFA, and DTT, both full-length apoB and the 70-kDa fragment were detected in the cell lysate. When DTT was removed from the chase medium at the end of the 2-h chase, and the cells were further chased for 1 h (ϪDTT), both full-length apoB and the 70-kDa fragment were degraded with no secretion observed. The unidentified nonspecifically precipitated 330-kDa protein was not degraded under this condition. In contrast, when BFA was removed and OA/DTT were still present during the additional 1-h chase (ϪBFA), these proteins were all secreted into the medium. This experiment strongly supports the view that DTT protects both full-length apoB and its amino-terminal fragment from degradation within the lumen of the ER.
The 70-kDa Amino-terminal Fragment of apoB Is Generated by the ALLN-sensitive Proteolytic Pathway and Is Degraded by the DTT-sensitive Pathway-To better characterize the relationship between the 70-kDa amino-terminal fragment and apoB degradation, the following experiments were performed. HepG2 cells were preincubated with either BSA or ALLN for 1 h, labeled for 10 min, and chased for 10, 30, and 60 min in serum-free medium containing one of three agents (BSA, DTT, or ALLN) or the combination of DTT plus ALLN (Fig. 7A). The results demonstrate that in BSA-treated cells, full-length apoB was rapidly degraded (13.3 Ϯ 2.7% of the initial amount was detected at 60-min chase; n ϭ 4), and no visible 70-kDa fragment accumulated during the chase period. Addition of DTT to the chase medium protected full-length apoB (28.5 Ϯ 5% of the initial amount was detected at 60-min chase; n ϭ 4), and the 70-kDa fragment was now easily detected. When ALLN was present prelabeling and in the chase medium, full-length apoB was also protected (51.2 Ϯ 13.2% of the initial amount was detected at 60-min of chase; n ϭ 3). However, with only ALLN, Aliquots of ER preparation were incubated at 4°C for 1 h under one of the following three conditions: no addition, 220 g/ml proteinase K, or 220 g/ml proteinase K and 1.25% Triton X-100. After incubation, immunoprecipitation was carried out using anti-apoB antibody. The remaining cells were divided into two groups. For the first group, the chase medium was changed to serum-free medium in the presence of OA and BFA; DTT was withdrawn from the incubation (ϪDTT). For the second group, the chase medium was changed to serum-free medium in the presence of OA and DTT; BFA was withdrawn from the incubation (ϪBFA). The cells were chased for an additional hour. At the end of the chase, the medium was collected, and the cells were lysed for apoB immunoprecipitation. This experiment was performed three times with similar results. very little 70-kDa fragment was detected. In contrast, when both DTT and ALLN were added at the start of the chase to ALLN-pretreated cells, both full-length apoB (58.1 Ϯ 11% of the initial amount was detected at 60-min chase; n ϭ 3) and the 70-kDa fragment were protected. Of note was the observation that significantly less 70-kDa fragment was present when ALLN/DTT was added to the media than when DTT alone was added. Similar results were observed in studies with OA (Fig.  7B). This less 70-kDa fragment was detected during chase with OA/DTT compared with DTT alone, even though the combination of OA/DTT protected full-length apoB to a greater extent (DTT:28.5 Ϯ 5%, n ϭ 4; OA/DTT:37.7 Ϯ 6.4%, n ϭ 3) (Fig. 7B). These results indicated that the 70-kDa fragment of apoB is degraded by an ALLN-resistant proteolytic process and that the ALLN-sensitive pathway (which is also affected by OA) generates the 70-kDa amino-terminal fragment of apoB.
Although we have assumed that the ALLN-resistant proteolytic process is in the ER lumen, we cannot rule out the possibility that this process is associated with the luminal side of the ER membrane because a recent study by Rustaeus et al. (24) suggested that in BFA-treated McArdle cells, over 40% of apoB100 is associated with the ER membrane. However, in that study, apoB48 high density lipoprotein particles and some apoB100 high density lipoprotein/low density lipoprotein particles, as well as transferrin, were detected in the ER lumen. Furthermore, Rustaeus et al. found that BFA prevented the formation of apoB48 and apoB100 very low density lipoprotein (24). Thus it seems that in McArdle cells, BFA specifically inhibited the further addition of lipids needed to form very low density lipoprotein-size particles from more dense apoB-lipoproteins but did not affect apoB translocation across the ER membranes. Because only a small fraction of apoB is assembled into very low density lipoprotein in HepG2 cells (25), we believe that BFA did not affect the early events in apoB transport and lipidation in our experiments.
In summary, the present studies provide evidence demon-strating that apoB degradation in HepG2 cells occurs as a two-step process. The results presented in the present studies, together with previous studies (11,13,14,18), suggest the following model of apoB degradation in HepG2 cells. Under basal (triglyceride-poor) conditions, only a small fraction of the newly synthesized apoB undergoes lipid-facilitated translocation across the ER membrane. The majority of newly synthesized apoB is subject, therefore, to proteolysis by an ALLNsensitive process that is either in the cytosol or on the cytosolic side of the ER membranes. A recent publication (17) suggests that the proteasome is involved in apoB degradation and that ALLN inhibits that process. The ALLN-sensitive proteolysis generates a 70-kDa amino-terminal fragment of apoB, consistent with data that the extreme amino-terminal of apoB is efficiently translocated into the ER lumen independent of corelipid availability or the presence of microsomal triglyceride transfer protein (18,26,27). The remaining domains of apoB have not been detected in pulse-chase studies and are probably rapidly degraded by a cytosolic pathway. Under conditions in which triglyceride availability is increased (e.g. OA in the medium), a larger fraction of the newly synthesized apoB is targeted for translocation and secretion, and less 70-kDa fragment is generated. However, after translocation of nascent apoB is complete, a second proteolytic pathway present in the lumen of the ER can degrade apoB that is not efficiently transferred to the Golgi. This second-stage degradative process, which is ALLN-resistant but DTT-sensitive, probably controls against the secretion of apoB that has misfolded after translocation. It is also possible that inadequate lipidation of the core of the primordial apoB particle, possibly during a second stage of lipoprotein formation (28 -32), might target apoB for degradation by the DTT-sensitive pathway. Characterization of the DTT-sensitive proteolytic pathway could provide new approaches to reducing the secretion of apoB-containing lipoproteins in dyslipidemic patients.