Translocation-arrested Apolipoprotein B Evades Proteasome Degradation via a Sterol-sensitive Block in Ubiquitin Conjugation*

In this study, we explored how sterol metabolism altered by the expression of cholesterol-7α-hydroxylase NADPH:oxygen oxidoreductase (7α-hydroxylase) affects the ubiquitin-dependent proteasome degradation of translocation-arrested apoB53 in Chinese hamster ovary cells. Stable expression of two different plasmids that encode either rat or human 7α-hydroxylase inhibited the ubiquitin conjugation of apoB and its subsequent degradation by the proteasome. Oxysterols (25-hydroxycholesterol and 7-ketocholesterol) reversed the inhibition of apoB degradation caused by 7α-hydroxylase. The combined results suggest that the normally rapid proteasome degradation of translocation-arrested apoB can be regulated by a sterol-sensitive polyubiquitin conjugation step in the endoplasmic reticulum. Blocked ubiquitin-dependent proteasome degradation caused translocation-arrested apoB to become sequestered in segregated membrane domains. Our results described for the first time a novel mechanism through which the “quality control” proteasome endoplasmic reticulum degradative pathway of translocation-arrested apoB is linked to sterol metabolism. Sterol-sensitive blocked ubiquitin conjugation appears to selectively inhibit the proteasome degradation of apoB, but not 7α-hydroxylase protein, with no impairment of cell vitality or function. Our findings may help to explain why the hepatic production of lipoproteins is increased when familial hypertriglyceridemic patients are treated with drugs that activate 7α-hydroxylase (e.g. bile acid-binding resins).

In this study, we explored how sterol metabolism altered by the expression of cholesterol-7␣-hydroxylase NADPH:oxygen oxidoreductase (7␣-hydroxylase) affects the ubiquitin-dependent proteasome degradation of translocation-arrested apoB53 in Chinese hamster ovary cells. Stable expression of two different plasmids that encode either rat or human 7␣-hydroxylase inhibited the ubiquitin conjugation of apoB and its subsequent degradation by the proteasome. Oxysterols (25hydroxycholesterol and 7-ketocholesterol) reversed the inhibition of apoB degradation caused by 7␣-hydroxylase. The combined results suggest that the normally rapid proteasome degradation of translocation-arrested apoB can be regulated by a sterol-sensitive polyubiquitin conjugation step in the endoplasmic reticulum. Blocked ubiquitin-dependent proteasome degradation caused translocation-arrested apoB to become sequestered in segregated membrane domains. Our results described for the first time a novel mechanism through which the "quality control" proteasome endoplasmic reticulum degradative pathway of translocation-arrested apoB is linked to sterol metabolism. Sterol-sensitive blocked ubiquitin conjugation appears to selectively inhibit the proteasome degradation of apoB, but not 7␣hydroxylase protein, with no impairment of cell vitality or function. Our findings may help to explain why the hepatic production of lipoproteins is increased when familial hypertriglyceridemic patients are treated with drugs that activate 7␣-hydroxylase (e.g. bile acid-binding resins).
ApoB is the major structural protein responsible for the assembly of lipoproteins by the liver and intestine. Multiple forms of apoB, designated as the percentage of the N terminus of the largest secretory product apoB100 (4536 amino acids), are produced from a single gene transcript by mRNA editing and proteolytic cleavage (reviewed in Refs. [1][2][3]. Overproduction of apoB-containing lipoproteins by the liver is responsible for familial combined hyperlipidemia (4). In addition, overproduction of triglyceride-rich lipoproteins is responsible for the human disease familial hypertriglyceridemia (5). In these patients, the secretion of triglyceride-rich lipoproteins varies in parallel with the rate of bile acid synthesis (6 -8). These findings suggest that the secretion of very low density lipoprotein triglyceride is linked to hepatic sterol metabolism via an as yet undefined mechanism that is dependent upon genes that contribute to hypertriglyceridemia.
The rate of hepatic secretion of apoB is regulated post-transcriptionally. Only a portion of de novo synthesized apoB is secreted; the remaining portion is degraded intracellularly (9). Interruption of apoB translocation is one of several criteria that lead to increased intracellular degradation (reviewed in Ref. 10). Both translocation and lipid addition require the presence of microsomal triglyceride transfer protein (MTP) 1 in the ER (11)(12)(13). MTP exists in the ER lumen as a heterodimer with protein-disulfide isomerase (reviewed in Ref. 14). In the absence of either sufficient lipid (15)(16)(17) or MTP lipid transfer activity (11)(12)(13), apoB translocation and lipoprotein assembly are blocked. The C terminus of resulting translocation-arrested apoB, which resides in the cytoplasm (18), is rapidly degraded by a ubiquitin-dependent proteasome process (19 -21).
The essential requirement of MTP for apoB translocation and lipoprotein assembly is exemplified by the finding that genetic loss of the expression of the MTP gene is responsible for the human recessive disorder abetalipoproteinemia (22,23). Similar to the CHO cells used in the studies reported here, the liver of abetalipoproteinemics lacks the ability to fully translocate apoB into the ER (24) and to secrete apoB-containing lipoproteins (25,26). Recent studies in mice having one MTP allele inactivated show significant reductions in MTP-lipid transfer activity and the ability to secrete apoB-containing lipoproteins by the liver (27). These findings raise the possibility that MTP expression may contribute to the rate-limiting step in the lipoprotein assembly/secretion pathway.
Cholesterol-7␣-hydroxylase NADPH:oxygen oxidoreductase (7␣-hydroxylase) is a liver-specific gene product that controls bile acid synthesis, the major pathway responsible for eliminating cholesterol from the body (reviewed in Ref. 28). The level of expression of 7␣-hydroxylase, which is highly variable in response to diet, metabolic state, and diurnal cycle, has a marked influence on the assembly and secretion of apoB-containing lipoproteins by cultured rat hepatoma cells (29). Oxysterols, which are hydroxylated sterol derivatives, antagonize 7␣-hydroxylase-induced changes in the assembly and secretion of apoB-containing lipoproteins (29). In this study, we examined how expression of 7␣-hydroxylase affected the ER degradation of translocation-arrested apoB in CHO cells that lack MTP. Due to a lack of MTP, CHO cells can not fully translocate apoB across the ER, resulting in its rapid degradation (18,30). Our results show that the normally rapid proteasome degrada-tion of translocation-arrested apoB can be reversibly blocked by 7␣-hydroxylase via a sterol-sensitive ubiquitin conjugation step in the ER.

EXPERIMENTAL PROCEDURES
Materials-An expression plasmid encoding rat 7␣-hydroxylase driven by the CMV promoter was constructed and transfected into CHO cells, as described (29). Another pcDNA3 plasmid containing the coding region of human 7␣-hydroxylase driven by the CMV promoter was a gift from Alan McClelland (Genetic Therapy, Inc.). Plasmid pCW8 encoding dominant negative ubiquitin, Ub (K48R), was generously provided by Ron Kopito. An affinity-purified rabbit antibody against ubiquitin was generously provided by Arthur Haas. Lactacystin was a generous gift from Ardythe McCracken.
Cell Culture-All cells were cultured in modified Eagle's medium (MEM; Life Technologies) containing 5% fetal bovine serum (Gemni) and antibiotics (100 units/ml penicillin, 100 units/ml streptomycin, and 500 g/ml fungizone) and other antibiotics for selection as indicated below. Within each cell type, there was no observed difference in cell viability or growth rate between individual clones used for the experiments. For most experiments, cells were grown to 80% confluence in the absence of antibiotics unless indicated.
JF7 Cells-B53 cells (CHO cells expressing apoB53 (18,30) were plated at 10 6 cells/100-mm plate and were co-transfected with an expression plasmid encoding rat 7␣-hydroxylase driven by the CMV promoter and a plasmid conferring resistance to hygromycin (31) at an 18:1 molar ratio. Cells were selected in 500 g/ml hygromycin (for 7␣hydroxylase) and 400 g/ml G418 (for apoB). Resistant cells were characterized as described below.
B53-7␣h Cells-B53 cells were co-transfected with a pcDNA3 plasmid containing the coding region of human 7␣-hydroxylase driven by the CMV promoter and a plasmid conferring resistance to hygromycin, as described above. Resistant cells were characterized as described below.
B53 Cells Transfected with Dominant Negative Ubiquitin, Ub (K48R), Plasmid-The coding sequence for the dominant negative form of ubiquitin Ub (K48R) (with a Lys 3 Arg mutation at amino acid 48 (32) was released from plasmid pCW8 using restriction enzyme BamHI and cloned into expression vector pIND at the BamHI site (Ecdysoneinducible expression kit, Invitrogen). The pIND-K48R plasmid is then co-transfected with the pVgRXR plasmid at a 1:19 molar ratio into B53 cells. Cells were selected in 500 g/ml Zeocin for Ub (K48R) and then maintained in 250 g/ml Zeocin. Single cell clones were obtained.
Northern Blotting-Poly(A ϩ ) RNA was isolated from cells using a modification of the guanidinium isothiocyanate method, as described (33). The resulting mRNA (2-5 g) was separated on a 0.8% agarose gel, transferred to a nylon membrane, and probed with nick-translated 32 P-cDNA probes for 7␣-hydroxylase, and ␤-actin was prepared from gel-purified inserts as described (33).
Western Blotting-Cells were harvested in phosphate-buffered saline (PBS) containing a mixture of protease inhibitors (100 g/ml aprotinin, 100 g/ml leupeptin, and 2 mM phenylmethylsulfonyl fluoride). Western blotting was performed as described (18). Following SDS-PAGE, the gels were electroblotted onto nitrocellulose membranes. The nonspecific binding sites of the membranes were blocked using 10% defatted dry milk, followed by the addition of primary antibody. The relative amount of primary antibody bound to the proteins in the nitrocellulose membranes was detected with species-specific horseradish peroxidase-conjugated IgG. Blots were developed using the ECL detection kit (Amersham Pharmacia Biotech).
Pulse-Chase Analysis and Immunoprecipitation-B53 and JF7 cells were grown to 80% confluence on 60-mm plates, after which the culture medium was changed to methionine-free MEM (Sigma). One hour later, cells were pulsed with [ 35 S]methionine (100 Ci/ml; DuPont) for 10 min, after which cells were chased with culture medium containing a 1000fold excess of unlabeled methionine. At indicated chase time points, cells were lysed in 1 ml of TETN buffer (25 mM Tris, pH 7.5, 5 mM EDTA, 250 mM NaCl, and 1% Triton X-100) containing a mixture of protease inhibitors (100 g/ml aprotinin, 100 g/ml leupeptin, and 2 mM phenylmethylsulfonyl fluoride). Immunoprecipitation was carried out as described (18). Solubilized proteins were precleared with Sepharose CL-4B. 5 l of a rabbit antiserum specific for human apoB was preincubated with 20 l of protein A-Sepharose beads (dry volume) in 1 ml of TETN buffer at 4°C overnight. The beads were then washed with TETN buffer three times. The antibody-bound protein A-Sepharose conjugates were incubated overnight at 4°C with the cellular protein samples in an amount determined empirically to completely bind the apoB present in each sample. Beads were recovered by centrifugation in a microcentrifuge and were washed three times with the TETN buffer. The immunoprecipitates were dissolved in sample buffer containing SDS and ␤-mercaptoethanol and separated on a 1-20% gradient SDS-PAGE. The protein content of the cell lysate was determined by the Bradford assay (Bio-Rad).
Microsome Isolation and Digestion with Trypsin-JF7 cells cultured to 85% confluence were disrupted by nitrogen cavitation, and microsomes were isolated by subsequent ultracentrifugation (9). Protein content was analyzed by the Bradford protein assay (Bio-Rad). Trypsin digestion was carried out as described (18). Each microsomal sample (containing 50 g of protein) was incubated with 15 g/ml trypsin in ST buffer (0.25 M sucrose, 10 mM Tris-HCl buffer, pH 7.4) for 30 min on ice. The digestion was stopped by adding protease inhibitors (300 g/ml soybean trypsin inhibitor, 2 mM phenylmethylsulfonyl fluoride, 100 g/ml aprotinin, and 100 g/ml leupeptin each). Microsomes were recovered by centrifugation at 45,000 rpm in a Beckman TLA 45 rotor for 2 h at 4°C. The microsomal pellet was resuspended in ST buffer and solubilized in sample buffer containing ␤-mercaptoethanol. The samples were separated on a linear 1-20% polyacrylamide gradient SDS-PAGE, Western blotted, and detected by using anti-human apoB monoclonal Ab 1D1 and a rabbit antiserum against protein-disulfide isomerase (a generous gift from Steve Fuller).
Indirect Immunofluorescent Microscopy-Cells were grown on coverslips in 100-mm plates to 60% confluence. Media were removed from the plates, and the cells were washed three times with PBS before they were fixed with 3% paraformaldehyde for 15 min. Cells were then permeabilized with 1% Triton X-100 for 7 min. Nonspecific binding sites in the cells were blocked by incubation with 3% bovine serum albumin in PBS for 30 min. The cells were then incubated for 45 min with 20 l of affinity-purified antibodies. At the end of the incubation, cells were washed four times with PBST (PBS containing 0.1% Tween 20) and then incubated with the appropriate species-specific Texas Red-conjugated IgG. After washing with PBST, the coverslips were examined using a Nikon microscope with a camera attached.
Detection of Ubiquitinated ApoB-Cells were washed once with icecold PBS. Cellular proteins were then solubilized in 1 ml of TETN buffer containing a mixture of protease inhibitors (see "Pulse-Chase Analysis and Immunoprecipitation"). ApoB was immunoprecipitated using 5 l of a rabbit antiserum specific for human apoB preincubated with 20 l of protein A-Sepharose beads. The immunoprecipitates were separated on a 1-20% gradient SDS-PAGE and then electroblotted onto nitrocellulose membranes. The ubiquitin conjugated on apoB was detected by an affinity-purified rabbit antibody against ubiquitin, followed by chemiluminescence (Amersham Pharmacia Biotech). Western blots were scanned by a densitometer and analyzed by the ImageQuant program.

Expression of 7␣-Hydroxylase mRNA and Protein in B53
Cells-Previous results showed that when apoB53 is stably expressed in CHO cells (B53 cells), it is rapidly degraded, producing an N-terminal fragment that is secreted without a lipid core (18). The proteolytic inhibitor ALLN blocks the formation of the N-terminal apoB fragment and causes intact apoB53 to accumulate as a transmembrane protein in isolated microsomes. These cells, stably transfected with a plasmid expressing rat 7␣-hydroxylase, are hereafter referred to as JF7 cells. The rate of growth of JF7 cells was indistinguishable from B53 and wild-type CHO-K1 cells. In JF7 cells, a single band for 7␣-hydroxylase mRNA was detected by Northern blot, whereas none was detected in B53 cells (data not shown). Analysis of the content of 7␣-hydroxylase protein by Western blot showed a single protein band, identical in size to the native protein (data shown below). The enzyme activity of 7␣-hydroxylase in CHO cells transfected with the CMV-driven plasmid containing the rat cDNA was ϳ10 pmol/mg protein/min.
Expression of 7␣-Hydroxylase Causes Intact ApoB53 to Accumulate by Blocking a Pathway That Is Also Inhibited by ALLN-In the absence of proteolytic inhibitors, intact apoB53 was nearly absent in B53 cells (Fig. 1, lane 1). Treating B53 cells with ALLN caused intact apoB53 to accumulate (Fig. 1,  lane 2). Remarkably, 7␣-hydroxylase expression caused intact apoB53 to accumulate (JF7 cells; Fig. 1, lane 3). The amount of apoB53 in JF7 and B53 cells treated with ALLN was similar (Fig. 1, compare lane 3 with lane 2). Moreover, treating JF7 cells with ALLN did not cause further accumulation of apoB53 (Fig. 1, lane 4). This experiment was performed at least five times (using different cell preparations), and there were no significant differences in the content of apoB53 in JF7 cells treated with or without ALLN. Additional studies using [ 35 S]methionine pulse labeling and immunoprecipitation also showed that ALLN did not cause the accumulation of intact apoB53 in JF7 cells (data not shown). Interestingly, treating JF7 cells with ALLN did increase the amount of smaller apoB fragments (Fig. 1, lane 4). These data suggest that, in contrast to apoB53, the degradation of small apoB peptides is blocked by ALLN in JF7 cells. Thus, JF7 cells are sensitive to proteolytic block by ALLN. These combined data suggest that the expression of 7␣-hydroxylase caused intact apoB53 to accumulate by blocking a proteolytic process that is blocked by ALLN in cells that do not express 7␣-hydroxylase (i.e. B53 cells; Fig. 1

, lane 2).
To exclude the possibility that the phenotype we observed was due to an artifact of the individual plasmid, the transfection procedure, or the site of genomic integration, we isolated three individual single cell clones of B53 cells expressing rat 7␣-hydroxylase. Each clone showed a comparable level of intact apoB53 in the absence of ALLN. Moreover, to rule out the possibility that the JF7 phenotype was due to a phenomenon unique to the transfected plasmid, we used an entirely different plasmid encoding human 7␣-hydroxylase (B53-7␣h cells). Cells stably expressing human 7␣-hydroxylase displayed an accumulation of apoB53 that was similar to that of JF7 cells (Fig. 1,  lane 5). Additional experiments showed that B53 cells transfected with different vectors not expressing 7␣-hydroxylase (e.g. luciferase and MTP) together with a hygromycin resistance plasmid did not cause the JF7 phenotype (i.e. the accumulation of intact apoB53; data not shown). These data indicate that the accumulation of apoB53 in JF7 cells is due to the expression of 7␣-hydroxylase (rat or human) and cannot be ascribed to a phenomenon caused by the transfection procedure, metabolic selection, or a single type of expression plasmid. The finding that inhibition of proteolysis by ALLN caused equivalent amounts of intact apoB53 to accumulate in B53 cells and JF7 cells (Fig. 1, compare lanes 2 and 4) suggests that both cells types synthesize similar amounts of apoB53.
Expression of 7␣-Hydroxylase in CHO Cells Blocks the Degradation of ApoB53-The turnover of newly synthesized apoB53 was determined in JF7 and B53 cells using pulse-chase analysis. In both groups of cells, maximal accumulation of [ 35 S]methionine-labeled apoB53 was detected after 30 min of chase (Fig. 2). Similar pulse-chase labeling of apoB has been observed using rat hepatocytes (9) and human hepatoma cells (16). The amount of maximally labeled apoB53 in JF7 cells at the 30-min chase time was 14-fold higher than that in B53 cells. These data are consistent with the proposal that apoB53 was degraded in a manner that appeared to be co-translational and that the expression of 7␣-hydroxylase in JF7 cells blocked this process.
Oxysterols Reverse the Inhibition of ApoB53 Degradation Caused by 7␣-Hydroxylase-We examined if oxysterols could reverse the block in apoB53 degradation exhibited by JF7 cells. Adding 25-hydroxycholesterol and 7-ketocholesterol to JF7 cells caused a marked increased in the rate of apoB53 degradation (Fig. 3). The oxysterol effect was observed during the 30-min pulse period and the 30-min chase period. After the 30-min chase period, the rate of decay of apoB53 was similar to that of untreated cells (Fig. 3). Additional data show that total protein synthesis was indistinguishable between JF7, B53, and JF7 cells treated with oxysterols (i.e. the incorporation of [ 35 S]methionine into trichloroacetic acid-precipitable protein was similar in all groups; data not shown). The pulse-chase experiments were performed three separate times (using different preparations of B53 and JF7 cells). In all experiments, oxysterols reversed the blocked degradation of intact apoB53. In two additional experiments, JF7 cells were incubated with and without oxysterols for 24 h. Western blot analysis of the cell extracts showed a 70 -95% decrease in the cellular content of apoB53 in cells treated with oxysterols (data not shown). The combined data indicate that the rapid degradation of apoB that is blocked by expression of 7␣-hydroxylase can be reversed by oxysterols.
ApoB Accumulates as a Transmembrane Protein in JF7 Cells-We examined the localization of apoB53 that accumulated in JF7 cells following subcellular membrane fractionation and isolation. Following cell disruption and ultracentrifugation, Ͼ90% of the apoB53 in JF7 cells was isolated in the 100,000 ϫ g pellet (i.e. microsomes). Moreover, in microsomes prepared from JF7 cells essentially all apoB53 was susceptible to digestion with exogenous trypsin (Fig. 4). The major and smallest molecular weight proteolytic fragment of apoB produced by trypsin digestion had a molecular mass of 69 kDa, which is identical to the results obtained using B53 cells treated with ALLN (18). There were two additional immunoreactive bands (of about 128 and 116 kDa) that were detected in JF7 microsomes following trypsin treatment. These immunoreactive bands showed markedly less chemiluminescence compared with the 69-kDa band. The 69-kDa fragment contained a defined apoB N-terminal epitope (residues 474 -539) as demonstrated by its recognition by monoclonal antibody 1D1 (34) (Fig. 4). In contrast to the complete degradation of apoB53, the ER luminal protein, protein-disulfide isomerase, was resistant to trypsin digestion (Fig. 4), indicating that the microsomes remained intact. The combined data suggest that the majority of apoB53 that accumulates in JF7 cells assumes a transmembrane orientation in which 69 kDa of the N terminus resides within the ER lumen and the remaining C terminus is exposed to the cytoplasm.
In JF7 cells, ApoB53 Accumulates in Segregated Membrane Domains of the Secretory Pathway-Indirect immunofluorescence was used to examine the accumulation of apoB in JF7 cells. We used two epitope-specific antibodies: monoclonal an-tibody 1D1, which recognizes the N-terminal residues 474 -539 (34), and a rabbit antiserum, which recognizes a C-terminal epitope of apoB53 at residue 2140 (35). The N-terminal epitopespecific antibody will recognize both intact apoB53 and Nterminal apoB peptides, whereas the C-terminal specific antiserum will recognize only intact apoB53. In B53 cells, the N-terminal specific antibody resulted in a diffuse reticular immunofluorescence pattern (Fig. 5). In JF7 cells, the N-terminal epitope-specific antibody showed a unique punctate immunofluorescence pattern that overlaid a more diffuse reticular pattern (Fig. 5). In contrast, the antiserum recognizing the C-terminal epitope showed no specific immunofluorescence in B53 cells, while there was a defined punctate immunofluorescence pattern in JF7 cells (Fig. 5). The punctate distribution of apoB53 detected by the C-terminal epitope antibody in JF7 cells is similar to the pattern observed for apoB100 in HepG2 cells treated with lactacystin (20). These data suggest that in JF7 cells, intact apoB53 accumulates in isolated domains of the secretory pathway (e.g. "Russell Bodies") in a manner similar to that described for a mutant form of IgM that can be neither degraded nor secreted (36). Sequestration of translocation-arrested apoB may allow normal cell function and viability to be maintained when the "quality control" proteasome degradation pathway is evaded.
ApoB Ubiquitination Is Blocked in JF7 Cells-To examine if the expression of 7␣-hydroxylase might have inhibited the proteasome degradation of apoB53, cells were cultured with or without the proteasome inhibitor lactacystin. Cells were harvested, and apoB was immunoprecipitated using a polyclonal rabbit antibody against human apoB. The immunoprecipitates were separated by an SDS-PAGE gel, blotted onto membranes, and then reacted with an antibody against ubiquitin and subsequently with monoclonal antibody 1D1, which recognizes the N terminus of apoB. In B53 cells not treated with lactacystin, almost no intact apoB53 was detected using the antibody against apoB (Fig. 6A, lane 1). There were a limited number of protein bands produced with the anti-ubiquitin antiserum (Fig.  6B, lane 1). In marked contrast, treating B53 cells with lactacystin caused immunoreactive apoB53 to accumulate (Fig. 6A,  lane 2). Moreover, the accumulated apoB53 contained polyubiquitin conjugates as shown by the molecular weight "ladder"   (18). Microsomes (containing 50 g of protein) were incubated with 15 g/ml of trypsin for 30 min on ice. The digestion was stopped by adding a mixture of protease inhibitors. Microsomes were recovered by ultracentrifugation and subjected to SDS-PAGE and Western blotting. ApoB (monoclonal antibody 1D1 recognizing the N-terminal epitope of apoB (34)) and proteindisulfide isomerase were detected using an ECL kit (Amersham Pharmacia Biotech). The migration of molecular weight standards is indicated.
FIG. 5. Indirect immunofluorescence microscopy of apoB in B53 and JF7 cells. Cells were grown on coverslips to 60% confluence. After removing the media, cells were washed and fixed with paraformaldehyde. Cells were then permeabilized with 1% Triton X-100. Nonspecific binding sites in the cells were blocked by incubation with 3% bovine serum albumin. The cells were then incubated for 45 min with affinity-purified antibodies: N-terminal specific (N-t) monoclonal antibody 1D1, which recognizes an epitope residing between amino acids 474 and 539 of human apoB (34) and C-terminal specific (C-t) rabbit antibody made against a synthetic human apoB peptide containing residue 2140 (35). The cells were washed and then incubated with Texas Red-conjugated goat anti-mouse or anti-rabbit IgG. The coverslips were examined using a Nikon microscope with an attached camera.
detected by the anti-ubiquitin antibody (Fig. 6B, lane 2). Although there was a large amount of apoB53 accumulated in untreated JF7 cells (Fig. 6A, lane 3), the amount of ubiquitin conjugates of apoB was similar to that in B53 cells without lactacystin treatment (Fig. 6B, compare lane 3 with lane 1). Treating JF7 cells with lactacystin did not change the amount of apoB53 in the cell (Fig. 6A, lane 4) or the amount of apoB ubiquitin conjugates (Fig. 6B, lane 4). Additional studies show that when the same cell extracts were treated with preimmune control rabbit antiserum, there were no immunoreactive bands that contained either the anti-apoB or the anti-ubiquitin epitopes (data not shown). The lack of accumulation of ubiquitin-conjugated apoB53 in JF7 cells (with and without lactacystin treatment) suggests that 7␣-hydroxylase blocks the ubiquitin conjugation of apoB53.
The rate of turnover of 7␣-hydroxylase in vivo is rapid and is essential for the diurnal rise and fall in enzyme expression (37). We examined if the 7␣-hydroxylase expressed in JF7 cells was also degraded by the proteasome. The specific proteasome inhibitor lactacystin caused a marked increase in the accumulation of 7␣-hydroxylase protein in JF7 cells (Fig. 6C). These findings suggest that while expression of 7␣-hydroxylase blocks the proteasome degradation of apoB53, it does not block its own degradation by the proteasome.
Recapitulation of the JF7 Cell Phenotype by Expressing a Dominant Negative Form of Ubiquitin in B53 Cells-To examine the hypothesis that 7␣-hydroxylase blocks apoB53 degradation by blocking its conjugation with ubiquitin, we attempted to recapitulate the JF7 cell phenotype by blocking ubiquitin conjugation using a dominant-negative form of ubiquitin (32). The dominant-negative form of ubiquitin has a lysine to arginine mutation that prevents polyubiquitination (Ub (K48R)) (32). Two different cell lines expressing dominant negative ubiquitin were derived from B53 cells transfected with either a plasmid containing the constitutive CMV promoter or a plasmid system that is ecdysone-inducible (38).
B53 cells were co-transfected with a plasmid expressing Ub K48R, which contains a His tag and a c-myc epitope (32) and a plasmid conferring puromycin resistance (to allow for metabolic selection in the presence of puromycin 10 g/ml). A single colony of puromycin-resistant cells was obtained and charac-terized. These cells grew at a rate indistinguishable from B53 and wild-type CHO-K1 cells (data not shown). The following data indicate that these cells expressed Ub (K48R) mRNA and protein (Fig. 7A): 1) a Northern blot showed Ub (K48R) mRNA in Ub (K48R) cells but not from B53; 2) a protein having the expected molecular weight of Ub (K48R) could be affinityisolated by a nickel column from the Ub (K48R) cells but not from B53 cells. This protein displayed immunoreactivity to an anti-ubiquitin antibody and to an anti-c-myc antibody. Moreover, cells expressing Ub (K48R) showed an accumulation of intact apoB53 (Fig. 7B) that was similar to that displayed by JF7 cells (Fig. 1). Furthermore, like JF7 cells, the apoB53 that accumulated in cells expressing UB (K48R) showed no accumulation of ubiquitin-conjugated apoB53 when treated with either ALLN or lactacystin (data not shown).
To rule out the possibility that the selection for stable expression of Ub (K48R) might have caused the accumulation of apoB53 via a mechanism not directly related to impaired ubiquitin conjugation, we isolated three separate clones of B53 cells expressing Ub (K48R) in a form that can be reversibly induced by the insect hormone muristerone A. In the absence of muristerone A treatment, B53 cells transfected with the plasmid showed a phenotype similar to that of B53 cells (i.e. no intact apoB53 was detected; Fig. 7C). In contrast, following treatment with muristerone A, there was an accumulation of apoB53 in all three independent single cell clones (Fig. 7C). Analysis of the culture medium of these cells treated with muristerone A showed that no intact apoB53 was secreted (data not shown). Thus, the expression of the dominant-negative form of ubiquitin recapitulated the phenotype of JF7 cells.
7␣-Hydroxylase Increases the Cellular Content of Mature SREBP1 and Increases the Synthesis of Lipids-The cellular content of mature SREBP1 was greater in JF7 cells compared with B53 cells (Fig. 8A). Adding 25-hydroxycholesterol (1 g/ ml) with cholesterol (10 g/ml) caused the cellular content of mature SREBP1 to decrease in both JF7 and B53 cells. These data suggest that CHO cells, like McArdle rat hepatoma cells (29) respond to the expression of 7␣-hydroxylase (increasing) and oxysterols (decreasing) by changing the cellular content of mature SREBP1. The synthesis of all very low density lipoprotein lipids (phospholipids, triglycerides, cholesterol, and cholesterol esters) were all increased in JF7 cells compared with wild-type CHO and B53 cells (Fig. 8B).  (34)) and ubiquitin (B, using an affinity-purified rabbit antibody). In other experiments, JF7 cells were treated with and without 50 M lactacystin for 14 h. Cells were harvested and disrupted by nitrogen cavitation, and microsomes were isolated by ultracentrifugation. Microsomes from treated and untreated cells were Western blotted using an affinity-isolated antibody directed against rat 7␣-hydroxylase (C).

FIG. 7. Dominant negative ubiquitin blocks apoB degradation.
A, B53 cells and B53 cells stably expressing Ub (K48R) were examined for the presence of Ub (K48R) as follows: mRNA by Northern blots (top) and nickel affinity column-isolated protein recognized by anti-ubiquitin antibody following Western blotting (middle) and also recognized by an anti-c-myc antibody (bottom). B, B53 cells and B53 cells stably expressing Ub (K48R) were examined for the presence of apoB53 by Western blotting using monoclonal antibody 1D1 (34). C, three individual clones of B53 cells stably expressing dominant negative ubiquitin Ub (K48R) produced by an ecdysone-inducible plasmid were incubated with or without muristerone (1 M) for 3 days and then with 5 M muristerone for 1 day. Cellular proteins were separated on SDS-polyacrylamide gel and then Western blotted for apoB by monoclonal antibody 1D1 (34). The migration of molecular weight standards is indicated.

DISCUSSION
The signals responsible for initiating ubiquitin conjugation and entrance into the "quality control" ubiquitin-dependent proteasome pathway are not well understood. These signals may be mediated by changes in protein structure caused by folding, post-translational modifications, and/or association with other molecules (e.g. ligands and substrates). The results of our studies suggest that the signals regulating the entrance of apoB into the ubiquitin-dependent proteasome degradation pathway are intimately linked to sterol metabolism. Furthermore, sterol-sensitive ubiquitin conjugation determines the two possible fates of apoB in CHO cells: 1) rapid degradation by the proteasome or 2) sequestration as a transmembrane protein in segregated domains of the ER.
Treating B53 cells with lactacystin caused the accumulation of intact apoB53 and 7␣-hydroxylase protein in JF7 cells (Fig.  6). Since lactacystin is a specific inhibitor of the proteasome (39), it is reasonable to assume that the proteasome is responsible for degrading apoB and 7␣-hydroxylase protein. Moreover, ubiquitin-conjugates of apoB53 accumulated as a result of lactacystin-blocked proteasome function (Fig. 7). Thus, we conclude that in CHO cells ubiquitin-conjugated apoB is the substrate degraded by the proteasome. In human hepatoma cells, lactacystin also caused ubiquitin-conjugated apoB to accumulate as a translocation-arrested form (19 -21). These results suggest that the degradation pathway of translocation-arrested apoB is similar in hepatic and nonhepatic cells.
In marked contrast to the ubiquitin-conjugated intact apoB53 that accumulates in cells treated with lactacystin, the intact apoB that accumulates in JF7 cells is not ubiquitinconjugated (Fig. 7). These data provide compelling evidence showing that expression of 7␣-hydroxylase markedly reduces the proteasome degradation of apoB53 by inhibiting ubiquitin conjugation. Moreover, the additional finding that expression of a dominant negative ubiquitin plasmid in B53 cells recapitulates the JF7 cell phenotype further suggests that the ubiquitin conjugation step is the site where 7␣-hydroxylase acts to divert translocation-arrested apoB away from the proteasomedegradative pathway.
Our finding that B53 cells that stably express the dominant negative ubiquitin, Ub (K48R), accumulate apoB53 and show similar rates of viability and replication as CHO cells not expressing Ub (K48R) suggests that essential processes thought to involve ubiquitin-dependent proteasome degradation (e.g. those required for cell cycle (40)) still function or are replaced by other processes. In order for the dominant negative Ub (K48R) to block ubiquitin conjugation, it must compete with the endogenous ubiquitin pool for entering a particular ubiquitin conjugation system. It is possible that the ubiquitin-dependent proteasome degradation of individual proteins may have different susceptibilities to inhibition by the dominant negative ubiquitin. Thus, processes that are essential for survival may be resistant to Ub (K48R) inhibition. Alternatively, there may be other "ubiquitin-like" molecules that can replace essential ubiquitin-dependent proteasome processes (e.g. sentrin (41,42) and Smt3p (43)). The B53 cells stably expressing Ub (K48R) may be useful to examine these possibilities.
In JF7 cells, intact apoB53 accumulates in microsomal membranes in a form that is susceptible to degradation by exogenous trypsin (Fig. 4). These data suggest that in JF7 cells, intact apoB53 spans the membrane bilayer. Moreover, trypsin digestion of microsomes from JF7 cells produced a similar 69-kDa N-terminal fragment that was produced by digestion of microsomes obtained from B53 cells treated with ALLN (18). These data indicate that 7␣-hydroxylase does not overcome the block in apoB translocation exhibited by CHO cells. This conclusion is further supported by the finding that JF7 cells do not secrete any intact apoB53. We have proposed that the block in apoB translocation exhibited by CHO cells is due to the lack of MTP expression (18). Additional studies showing that expressing MTP in JF7 cells allows intact apoB to be secreted as a lipoprotein particle (data not shown) support this hypothesis.
The findings showing that apoB can exist as a stable transmembrane ER protein are nevertheless paradoxical, since apoB has no amphipathic ␣-helices that are sufficiently long to act as canonical membrane-spanning domains (44 -46). A similar situation exists for prion protein, which like apoB contains no canonical membrane-spanning domains but can exist as a stable transmembrane protein (47). It is likely that in addition to amphipathic ␣-helices, alternative structures allow proteins to stably span membrane bilayers.
Our findings showing that oxysterols reverse the blocked degradation of apoB in JF7 cells (Fig. 3) provide further evidence that 7␣-hydroxylase blocks the ubiquitin conjugation of apoB by altering the sterol status of cells. There are other examples of proteins that are proteolytically cleaved in the ER by processes that are regulated by 7␣-hydroxylase and/or oxysterols. Degradation of the ER enzyme HMG-CoA reductase, which regulates the isoprenoid biosynthetic pathway, is intimately linked to the sterol status of cells (48). The rate of degradation of HMG-CoA reductase is accelerated by several nonsterol and sterol metabolites (49 -52). Similar to the results reported here for apoB, lactacystin also blocks the rapid degradation of HMG-CoA reductase (53). In yeast, disruption of a gene required for proteasome assembly causes the degradation of HMG-CoA reductase protein to be impaired (54). The identities of the signals and the mechanisms through which sterols affect the proteolytic processing and/or degradation of HMG-CoA reductase remain unknown.
Our findings show that 7␣-hydroxylase expression in CHO cells blocked the ubiquitin conjugation of apoB in a manner that could be reversed by oxysterols. These data suggest that sterols alter ubiquitin conjugation of apoB. It is interesting to note that JF7 cells contain significantly greater levels of mature SREBP1 than do B53 cells (Fig. 7A). In addition, similar to FIG. 8. Cellular content of SREBP1 and lipid synthesis in B53 and JF7 cells. B53 and JF7 cells were plated and grown to 85% confluence. A, both groups of cells were then cultured for 24 h in medium with or without 25-hydroxycholesterol (1 g/ml) with cholesterol (10 g/ml). Cells were harvested and fractionated into nuclei and membrane fractions, and the content of SREBP1 was determined by Western blot analysis. B, CHO-K1, B53, and JF7 cells were plated and grown to 85% confluence. [2-14 C]Acetate (3.3 Ci/ml) was added to the medium, and cells were incubated for 2 h. Cells and medium were harvested and extracted with chloroform/methanol, the radiolabeled lipids were separated by TLC, and the radioactivity was quantitated by ␤-scintillation analysis. Values represent the means ϮS.D. of three separate cell extracts in each group of cells. the reversal of the blocked degradation of apoB53, the increased content of mature SREBP1 displayed by JF7 cells is also blocked by oxysterols, albeit with one potentially important difference. The decrease in mature SREBP1 required only 25-hydroxycholesterol, whereas the reversal of apoB degradation required both 7-ketocholesterol and 25-hydroxycholesterol. 7-Ketocholesterol is unique in its ability to block the enzymatic activity of 7␣-hydroxylase (55). Based on these findings, it is reasonable to propose that the enzymatic activity of 7␣-hydroxylase is responsible for reducing the ubiquitin conjugation and subsequent degradation of translocation-arrested apoB53. The combined data are consistent with the proposal that 7␣-hydroxylase expression acts via changing sterol metabolism and that oxysterols antagonize this effect. As a result, the ubiquitin conjugation of apoB is impaired, as is its degradation by the proteasome. Concurrent with the blocked degradation of apoB, there is an increase in the content of mature SREBP1 and lipogenesis. Thus, expression of 7␣-hydroxylase in both CHO cells and McArdle rat hepatoma cells (29) causes similar phenotypic changes: increased content of SREBP1, increased lipogenesis, and decreased degradation of apoB. There is one important difference between these two cell types: the presence of MTP in McArdle rat hepatoma cells allows apoB to be translocated and assembled into a lipoprotein particle.
Overproduction of triglyceride-rich lipoproteins is responsible for the human disease familial hypertriglyceridemia (5). In these patients, the secretion of triglyceride-rich lipoproteins varies in parallel with the rate of bile acid synthesis (6 -8). The increased synthesis of bile acids in hypertriglyceridemic patients has been ascribed to decreased bile acid absorption (8). Interestingly, treatment of hypertriglyceridemic patients with chenodeoxycholic acid, which would be expected to decrease the activity of 7␣-hydroxylase, reduces fasting plasma triglyceride levels (56). Conversely, in many hypertriglyceridemic patients, bile acid binding resins increase bile acid synthesis and cause a transient exacerbation of their hypertriglyceridemia due to increased hepatic production of very low density lipoprotein (57). Our findings showing that 7␣-hydroxylase blocks the ubiquitin-dependent proteasome degradation of apoB may provide insights toward understanding the complex relationships between lipoprotein assembly/secretion and hepatic cholesterol/bile acid metabolism and how these relationships may contribute to the disordered lipid metabolism.