Studies on intracellular translocation of apolipoprotein B in a permeabilized HepG2 system.

Recent evidence suggests that the rate of apolipoprotein B-100 (apoB) translocation may be a key regulatory point in the production of apoB-containing lipoproteins. We have developed an in vitro system to measure the translocation rate of apoB in HepG2 cells. Intact cells were initially pretreated with oleate and N-acetyl-Leu-Leu-norleucinal to maximize the translocation rate while minimizing degradation. Cells were pulsed with [35S]methionine, chased (5-30 min), and then permeabilized with digitonin (75 μg/ml). Permeabilized cells were incubated with or without trypsin (200 μg/ml) for 10 min, and digestion was halted with soybean trypsin inhibitor (2 mg/ml). The rate of translocation was determined by comparing the amount of immunoprecipitable intact apoB in trypsin-treated cells with that in control cells at each time point. Under these conditions, two control proteins, α1-antitrypsin and transferrin, were fully protected from trypsin digestion, confirming the integrity of the secretory pathway in permeabilized cells. The percentage of apoB translocated steadily increased from 36% after 5 min to 71% after a 30-min chase (mean percentage, n = 3). A characteristic apoB fragmentation pattern resulted from trypsin digestion, and protected fragments of various size including N-terminal 60-70-kDa fragments were identified. Subcellular fractionation of the cells confirmed that the apoB pool protected from trypsin digestion was luminal in nature, confirming its translocation. ApoB translocation was significantly increased in oleate-treated cells compared with untreated cells. Inhibition of peptidylprolyl isomerase through the use of cyclosporin A and disruption of disulfide bond formation using dithiothreitol reduced the percentage of translocated apoB by 37 and 63%, respectively. Dithiothreitol induced specific changes in the pattern of protected apoB fragments, suggesting a conformational change in apoB that may hinder its translocation. Inhibition of N-linked glycosylation with tunicamycin did not significantly alter the rate of apoB translocation but appeared to stimulate its degradation. Together, the data suggest that the rate of apoB translocation across the membrane of the ER is determined by both lipid availability as well as the correct conformation of nascent apoB molecules.

Apolipoprotein B-100 (apoB) 1 is a 4536-residue polypeptide that comprises the major protein component of very low density lipoproteins (VLDLs) and low density lipoproteins (LDLs). Evidence to date appears to suggest that acute modulation of apoB is post-transcriptionally regulated (1)(2)(3)(4)(5) and that the rate of apoB production and secretion is regulated by co-translational and post-translational processes. Translocation efficiency and the rate of transport out of the ER may determine whether apoB is secreted or shunted into degradative pathways (6,7). Any apoB that is not translocated across the ER membrane and assembled into a lipoprotein particle is subsequently diverted for intracellular degradation (7)(8)(9). ApoB translocation is blocked in Chinese hamster ovary cells, suggesting that apoB requires a unique process for complete translocation that is not expressed in nonhepatic cells (10). More recently, Vance and co-workers (11,12) showed that apoB translocation into the lumen of microsomes of rat hepatocytes is disrupted in membranes enriched in phosphatidylmonomethylethanolamine and demonstrated that impaired translocation leads to increased degradation of the protein (12).
The exact mechanism as well as the relative importance of apoB translocation with respect to its production have yet to be agreed upon. A very recent study in HepG2 cells appears to confirm that the rate of translocation but not degradation is the key determinant of apoB secretion from the cells (13). Several studies have provided evidence to suggest that the translocation of apoB is such that the protein is not exposed to the cytosol (14,15). However, others have proposed that incomplete translocation results in a significant percentage of the apoB molecule remaining exposed on the cytosolic side of the ER (16 -18). The ability of apoB to assume a transmembrane form (despite lacking classical transmembrane sequences) has been suggested to be the result of specific "pause transfer sequences" in apoB that cause the transient translocation arrest of the protein (19). Recently, an alternative explanation was put forth that suggests that transmembrane forms of apoB may be the result of "ribosomal pausing" caused by tRNA persistence on the emerging apoB polypeptide (20).
A number of other factors have been identified as possible modulators of apoB translocation. The presence of oleate has been suggested to increase the translocation of apoB in HepG2 (21,22). As mentioned above, alteration of the phospholipid composition of the ER membrane through the incorporation of phosphatidylmonomethylethanolamine also blocks translocation of apoB (12). In addition, the cytosolic chaperone protein Hsp70 has been suggested to be a necessary component for apoB translocation (23). ApoB translocation has been previously studied by preparing isolated microsomes and proteolytic digestion of apoB or apoB variants exposed on the cytosolic side of these membranes. Here, we report the development of a novel translocation protocol that utilizes permeabilized HepG2 cells. It is important to note that in the current protocol intact cells were initially used to achieve intracellular apoB translocation; permeabilization was only carried out to allow the delivery of trypsin to the cytosolic side of the ER membrane (the rate of translocation measured therefore reflects that in intact cells). Permeabilized cells provide a more intact ER-Golgi organelle system for the study of apoB translocation when compared with isolated vesicles. Using permeabilized cells, we directly demonstrate that lipid enrichment of the cells with oleate increased the percentage of apoB translocated. Inhibition of proper folding through inhibition of peptidylprolyl isomerase or disruption of disulfide bond formation reduced apoB translocation. The data suggest that translocation is both a function of lipid availability and the conformational status of apoB.

MATERIALS AND METHODS
Cell Culture-Monolayer cell cultures of HepG2 (ATCC HB8065) were maintained in ␣-modification of Eagle's minimal essential medium in culture flasks or multiwell dishes containing 10% fetal calf serum. Cells were grown in 35-or 60-mm dishes at 37°C, 5% CO 2 in complete medium (␣-minimal essential medium, 10% fetal bovine serum) (24). Cultures were allowed to reach 75-80% confluence before any experiment was carried out.
Metabolic Labeling, Permeabilization, and Trypsinization of HepG2 Cells-Cells were incubated with methionine-free minimal essential medium supplemented with 360 M oleic acid for 2 h prior to the start of the experiment. Following 1 h of incubation in the oleate-supplemented medium, ALLN was added (10 M) for the remaining 1 h. Cells were then pulsed with 80 -100 Ci/ml [ 35 S]methionine for 3 min, washed in Earle's balanced salt-solution three times, and chased in complete medium containing 10 mM methionine, 10 M ALLN for 5-30 min. The cells were then washed and incubated in CSK buffer (0.3 M sucrose, 0.1 M KCl, 2.5 mM MgCl 2 , 1 mM sodium-free EDTA, 10 mM PIPES, pH 6.8) containing 75 g/ml digitonin, 10 mM methionine, 150 M puromycin, 50 g/ml cycloheximide, 1.25 M ALLN for 5 min at room temperature. Digitonized cells were washed once in CSK buffer and were then incubated in the presence and absence of trypsin (200 g/ml) prepared in a CSK buffer supplemented with 10 mM methionine, 150 M puromycin, 50 g/ml cycloheximide, 1.25 M ALLN for 10 min at room temperature. An equal volume of CSK containing 2 mg/ml soybean trypsin inhibitor, 1 mM PMSF, 1.25 M ALLN, and 100 kallikrein-inactivating units/ml Trasylol was added to all dishes for 10 min at room temperature. The cells were then incubated an additional 10 min on ice and collected. The collected cells were then centrifuged briefly for 2 min at 10,000 rpm in a microcentrifuge. The supernatant was removed, and the cells were solubilized in a solubilization buffer (phosphate-buffered saline containing 1% Nonidet P-40, 1% deoxycholate, 5 mM EDTA, 1 mM EGTA, 1 mM PMSF, 100 kallikrein-inactivating units/ml Trasylol, 0.1 mM leupeptin, 0.5 M ALLN, 1 mg/ml soybean trypsin inhibitor). Cell extracts were centrifuged in a microcentrifuge at 14,000 rpm for 10 min, and the supernatant was subjected to immunoprecipitation. Subcellular Fractionation of ApoB-containing Lipoproteins-Isolation of the microsomal fraction and the separation of the luminal and membrane components were performed as described (25)(26)(27)(28). Intact cells that had been incubated for 5-30 min were permeabilized and then homogenized with a glass Dounce homogenizer as described (27,28). The homogenate was fractionated and then subjected to treatment with sodium carbonate pH 11 (27,28). The luminal component released by carbonate extraction was isolated from the membrane fraction by ultracentrifugation at 37,000 rpm for 90 min in a SW41 rotor. Both the membrane and the luminal fractions were then diluted with 800 l of a solubilization buffer containing 360 l of 5XC buffer (250 mM Tris-HCl, pH 7.4, 750 mM NaCl, 25 mM EDTA, 5 mM PMSF, 5% Triton X-100), 410 l of phosphate-buffered saline (450 kallikrein-inactivating units/ml Trasylol, 5 mM PMSF) and subjected to immunoprecipitation.
Immunoprecipitation-Immunoprecipitation was performed as described previously (29). In procedures where monoclonal antibodies to apoB were employed, they were first bound to Affi-gel beads then incubated with the samples overnight at 4°C and then centrifuged to pellet the beads. Immunoprecipitates were washed three times with the wash buffer (10 mM Tris-HCl, pH 7.4, 2 mM EDTA, 0.1% SDS, 1% Triton X-100). Finally, the immunoprecipitates were prepared for SDS-PAGE by suspending and boiling in 75 l of electrophoresis sample buffer (see below).
SDS-PAGE and Fluorography-SDS-PAGE was performed essentially as described (30). The gels were fixed and stained, and then they were fluorographed by incubating in Enhance (DuPont) or Amplify (Amersham Corp.). The gels were dried and exposed to Kodak X-Omat AR5 film at Ϫ80°C for 1-4 days.
To determine the radioactivity in apoB, apoB fragments, and ␣1antitrypsin, the bands corresponding to the proteins or protein fragments were visualized by fluorography, excised from the gel, and digested, and the radioactivity was counted. In some cases, the apoB bands were scanned using a Bio-Rad imaging densitometer.

Development of a Translocation Protocol Using
Permeabilized Cells-The procedure employed for the preparation of permeabilized HepG2 cells was a modification of previous work performed in our laboratory, which used permeabilized cells to study apoB degradation (29). Parameters were chosen in order to maximize the rate of translocation while simultaneously minimizing possible confounding factors such as degradation and secretion. All cells unless otherwise indicated were preincubated with oleate to maximize apoB translocation and increase the sensitivity of the measurements. Throughout the course of the translocation assay and in a 1-h preincubation period, ALLN was present (10 M in intact cells and 1.25 M in permeabilized cells) to minimize the degradation of apoB. Finally, to ensure optimal trypsin accessibility to proteins on the cytosolic face of the ER membrane, 150 M puromycin was included in the permeabilization and trypsinization steps. Puromycin treatment results in the release of ribosomes from the ER and should therefore ensure maximal exposure of any cytosolic apoB polypeptide chains. Conditions were optimized with respect to digitonin and trypsin concentrations. Cells at 80 -90% confluence were treated with a range of digitonin concentration in order to determine the concentration that allowed for maximal trypsin accessibility with minimal damage to the ER and Golgi. Fig. 1 shows the percentage of trypsin- Cells were permeabilized with digitonin (20 -75 g/ml) for 5 min, and permeabilized cells were incubated in the presence or absence of trypsin for 10 min. Trypsin digestion was halted by the addition of protease inhibitors, and cells were subsequently subjected to immunoprecipitation with a specific anti-apoB antibody. Immunoprecipitates were analyzed by SDS-PAGE and fluorography. ApoB radioactivity was quantitated by cutting and scintillation counting of the apoB band. resistant apoB as a function of digitonin concentration. At the early time point (5-min pulse, 5-min chase), a high percentage of total apoB was digested at all digitonin concentrations. The high sensitivity of apoB to trypsin at this early time point is expected, since a high percentage of labeled apoB at this stage is still associated with the ribosomes and therefore mainly cytosolic. At a later chase period (5-min pulse, 20-min chase), more of the intact apoB was protected from trypsin digestion compared with the earlier time point. The decrease in trypsin sensitivity of apoB is evidence of translocation of apoB into the lumen of the ER. The percentage of trypsin-resistant apoB at both time points was also observed to be a function of digitonin concentration and decreased with increasing digitonin concentration. This variability was suspected to be the result of differential permeabilization of the HepG2 cells, and in order to maintain adequate and consistent trypsin accessibility, the highest possible digitonin concentration would subsequently be required. However in order to verify the integrity of intracellular organelles at high concentrations of digitonin, the trypsin resistance of the luminal protein ␣1-antitrypsin was investigated. In the same experiment shown in Fig. 1, ␣1-antitrypsin was also immunoprecipitated from the cells. Essentially all the ␣1-antitrypsin molecules were resistant to trypsin digestion both after a 5-min chase (mean percentage of trypsin-resistant ␣1-antitrypsin ϭ 103.7 Ϯ 6) and after a 20-min chase (mean percentage of trypsin-resistant ␣1-antitrypsin ϭ 117.3 Ϯ 13). The percentage of ␣1-antitrypsin protected from trypsin digestion remained high irrespective of digitonin concentration. Because ␣1-antitrypsin was protected from trypsin digestion even at the highest digitonin concentration, it can be suggested that the organelles of the semipermeable cells were relatively undamaged. Therefore, the highest digitonin concentration (75 g/ml) was used in order to ensure efficient permeabilization and consistent trypsin accessibility. Fig. 2 shows the immunoprecipitable trypsin-generated apoB fragments occurring at a range of trypsin concentrations (0 -300 g/ml). Increasing trypsin concentrations resulted in the formation of a set of distinct apoB fragments. The intensity of these fragments was greatest at 200 g/ml, and this concentration was therefore used in all subsequent translocation assays.
Measurement of ApoB Translocation in HepG2 Cells-In this and all subsequent experiments, intact HepG2 cells were initially pulse-chased to achieve biosynthesis and translocation of apoB across the membrane of the ER. Permeabilization was only carried out to allow the delivery of trypsin to the cytosolic side of the ER membrane. The rate of translocation measured thus reflects the rate in intact cells and not permeabilized cells. Fig. 3 shows the immunoprecipitable apoB recovered from permeabilized cells at different times of chase in the presence and absence of trypsin. A gradual increase in the amount of trypsin-resistant intact apoB (approximate size, 550 kDa) was detected over time in the translocation assay. The amount of trypsin-resistant apoB was measured as a percentage of the intact apoB immunoprecipitated from control cells not subjected to trypsin treatment at similar time points. It should be noted that at the later time points a slight decrease was observed in the intensity of the intact apoB band when compared with earlier time points. Based on the recovery of apoB from the media of cells at the later time points (data not shown), this decrease was determined to be the result of secretion of apoB that is expected to occur after 30 min of pulse and chase. Fig. 4a demonstrates the rate of apoB translocation in intact cells measured by permeabilization and trypsinization of the cells. The percentage of intact apoB remaining was 36 Ϯ 8% after 5 min, 40 Ϯ 15% after 10 min, 51 Ϯ 14% after 20 min, and 71 Ϯ 14% after 30 min. Due to the short pulse time and the fact that apoB is co-translationally translocated, a large proportion of the labeled apoB molecules immunoprecipitated from control cells at early time points were in the form of shorter nascent chains. As time progressed, there was a disappearance of these smaller nascent chains coupled with an increase in the amount of intact apoB. These events suggest the elongation and completion of apoB molecules. However, in order to discount the possibility that the increasing amount of intact apoB was due to residual 35 S labeling occurring during the chase period, 10-l aliquots of the total immunoprecipitated apoB from both control and trypsin-treated cells were removed and subjected to scintillation counting. Fig. 4b shows that the highest amount of total radioactivity occurred at the earliest time points, suggesting that there was no further radiolabeling of any newly translated apoB molecules following the initial pulse.
Trypsin Treatment of Semipermeable Cells Generates a Specific ApoB Fragmentation Pattern-Analysis of Fig. 3 reveals that trypsin treatment of semipermeable cells generates a number of distinct sets of immunoprecipitable apoB fragments (fragments with sizes of 335, 200, 150, and 60 -70 kDa). The intensity of all immunoprecipitable apoB fragments was highest at the early time points of the translocation assay and continually declined over time. The decrease in intensity of the trypsin-generated fragments coincided with an increase in the intensity of the intact apoB molecule over the same time period. Fig. 4c demonstrates the gradual increase in the amount of protected intact apoB coupled with the simultaneous decrease in the amount of the 60 -70-kDa fragments. The accumulation of the intact apoB molecule (550 kDa) and the steady decline of the trypsin-generated fragments appears to suggest that over time nascent apoB chains move away from cytosolic,

FIG. 2. Effect of trypsin concentration on apoB digestion.
Nearly confluent cells were treated in the presence of 360 M oleate (2 h) and 10 M ALLN (1 h). Cells were pulsed for 5 min with [ 35 S]methionine and were chased for 5 min with excess cold methionine. Cells were permeabilized with digitonin (75 g/ml) for 5 min, and permeabilized cells were incubated in the presence or absence of trypsin (0 -300 g/ml) for 10 min. Trypsin digestion was halted by the addition of protease inhibitors, and cells were subsequently subjected to immunoprecipitation with a specific anti-apoB antibody. Immunoprecipitates were analyzed by SDS-PAGE and fluorography.
trypsin-accessible side of the ER into the lumen, where they are protected from proteolysis.
The Membranes of the ER and Golgi as Well as the ER-Golgi Transport System Are Preserved in Trypsin-treated Semipermeable Cells-In order to demonstrate that the membranes of the ER and Golgi were not compromised during the process of permeabilization, the translocation assay was performed using ␣1-antitrypsin as a control protein. ␣1-Antitrypsin forms a number of distinct intracellular species during its transport from the ER to the Golgi, particularly a precursor/ER form (P), which lacks complex N-linked sugars, and a mature/Golgi form (M), which is secreted from the cell (31). A similar experiment was performed as in Fig. 4, and ␣1-antitrypsin was immunoprecipitated from both control and trypsin-treated cells at all time points. Fig. 5A shows that essentially all ␣1-antitrypsin molecules (both P and M forms) were protected from digestion at every time point in the translocation assay in both intact and permeabilized cells. Trypsin resistance at the earliest time points of the assay is evidence that permeabilization does not disrupt the ER membrane and that proteins have been completely translocated into the ER. Analysis of the autoradiogram in Fig. 5A also demonstrates through the decrease in gel mobility of ␣1-antitrypsin that the protein acquires complex Nlinked oligosaccharides. The resistance of this mature form of ␣1-antitrypsin to trypsin digestion indicates that the Golgi membranes were also not damaged by permeabilization of the cells. To address the possibility that the trypsin resistance of ␣1-antitrypsin was not due to an inherent resistance to the protease, both intact and permeabilized cells were subject to trypsin treatment in the presence and absence of the detergent Triton X-100. In both intact and permeabilized cells, disruption of the intracellular organelles with Triton X-100 results in essentially complete digestion of immunoprecipitated ␣1-antitrypsin in those cells treated with trypsin.
As a final control, the translocation protocol was conducted using a larger secretory protein in order to ensure that the accessibility of apoB to trypsin was not an artifact of the large size of the protein. Fig. 5B shows the immunoprecipitable transferrin from intact and permeabilized cells in the presence and absence of trypsin. Transferrin is an 80-kDa secretory protein that is known not to associate with the ER membrane and therefore should be resistant to trypsin digestion during its transport and secretion if the intracellular organelles are undamaged. The percentage of transferrin resistant to trypsindigestion in permeabilized cells was determined to increase from 77 Ϯ 9% after 10 min of chase to 96 Ϯ 3% based on comparison with untreated control cells at similar time points. At the later time points, the insensitivity to trypsin digestion in permeabilized cells was similar to that observed in intact cells. The results of the data suggest that proteins that are translocated into the lumen of the ER are protected from digestion by exogenous trypsin under the parameters of the translocation protocol.
Confirmation of the Use of Trypsin-resistant ApoB as a Measure of Translocation-In order to validate the use of trypsinresistant apoB as a measure of its translocation, we compared the method utilized in the present report with an alternative method to measure apoB translocation, namely the percentage of luminal apoB extracted from isolated microsomes. Identical experiments were conducted with respect to pulse-chase protocols and subsequently subjected to either trypsin digestion or microsomal isolation and luminal extraction. Fig. 6 depicts the increase in both trypsin-resistant apoB and luminal apoB over time. After 30 min of chase, the percentages of trypsin-resistant apoB and microsomal luminal apoB were approximately equal. This suggests that the percentage of trypsin-resistant apoB is a reflection of the percentage of intact apoB that has been translocated into the lumen. However, at earlier time points there was a higher percentage of luminal apoB present compared with trypsin-resistant apoB. The higher percentage might represent partially translocated apoB that is still membrane-associated and trypsin-accessible but mostly luminal in nature. This fraction may be extracted with the luminal content upon carbonate extraction and therefore detected in the luminal fraction.
ApoB Fragments Resulting from Trypsin Digestion Are Localized in the Microsomal Lumen-Trypsin digestion of permeabilized HepG2 cells was shown to generate a specific immunoprecipitable apoB fragmentation pattern, with fragments ranging in size from 60 to 330 kDa. Protection of these fragments from trypsin digestion suggested that they have been moved away from the trypsin-accessible or cytosolic face of the ER. Subcellular fractionation experiments were conducted to determine whether these fragments were luminal in nature. Fig. 7 shows that the immunoprecipitable apoB fragments from trypsin-treated HepG2 cells were also recovered from the luminal fraction of the isolated microsomes. Only trace amounts of these fragments were detectable in the membrane fraction. The data confirm the notion that trypsin-generated apoB fragments in permeabilized cells are luminal in origin. The possibility exists that the fragments detected in the lumen are released from membrane-bound apoB chains following trypsin digestion of permeabilized cells. These fragments of apoB may represent regions of the molecule that have been translocated across the lipid bilayer and are exposed to the luminal side of ER membrane.
The 60 -70-kDa Trypsin-generated ApoB Fragments Originate from the N-terminal Region of the Intact ApoB Molecule-The origin of the 60 -70-kDa trypsin-generated apoB fragments was investigated through the use of a battery of monoclonal antibodies. Fig. 8 shows that the immunoprecipitable apoB fragments having an approximate molecular mass of 60 -70 kDa were selected predominantly by the polyclonal antibody and by the N-terminal monoclonal antibody 1D1 (epitope mapping 474 -539). Selection by 2D8 (epitope mapping 1438 -1480) was very weak, and no immunoprecipitable apoB fragments were detected using the C-terminal monoclonal antibodies Sol6 and Sol16 (epitope mappings 2488 -2453 and 4027-4081, respectively). The data suggest that the 60 -70-kDa fragments originate from the extreme N-terminal of apoB. The evidence is consistent with a protein being co-translationally translocated from the N terminus to C terminus.
ApoB Translocation Is Increased in Oleate-treated HepG2 Cells-The effect of oleate on the translocation of apoB was investigated using permeabilized HepG2 cells that were incubated in the presence and absence of trypsin. Fig. 9 shows the percentage of trypsin-resistant immunoprecipitable apoB from permeabilized cells that had been preincubated with and without oleate. There was a higher percentage of trypsin-resistant apoB in oleate-treated cells compared with untreated cells at all time points. Although the percentage of trypsin-resistant apoB increased over time in control cells, the amount of apoB translocated remained relatively constant after 20 min of chase. This suggests that lipid availability is a limiting factor in apoB translocation. After 10 min of chase there was a 2-fold increase in the percentage of trypsin-resistant apoB in oleatetreated cells compared with untreated HepG2 cells. The dramatic increase in the amount of apoB translocated in oleatetreated cells at the 10-min chase could reflect the increased amount of apoB that is being directly translocated into the lumen in lipid-enriched cells. It is suggested that in untreated HepG2 cells a higher proportion of the labeled apoB pool might initially become membrane-bound and move into the lumen at a slower rate.
Effect of N-Linked Glycosylation on ApoB Translocation-To determine the relationship between N-linked glycosylation and translocation, experiments were performed in cells preincubated with 5 g/ml tunicamycin for 3 h. Analysis of the autoradiogram in Fig. 10 demonstrates that the apoB fragmentation pattern in trypsin-treated cells is largely similar between control and tunicamycin-treated cells, the most notable difference being an increase in gel mobility of the 60 -70-kDa apoB fragments in tunicamycin-treated cells. This shift suggests that the N-terminal 60 -70-kDa fragments are N-linked glycosylated, which is further evidence for their luminal exposure. The intact apoB from tunicamycin-treated cells also migrated faster in the gel (approximate molecular mass of 510 kDa), confirming the inhibition of glycosylation.
Inhibition of N-linked glycosylation did not appear to significantly affect the intracellular movement of apoB (data not shown). There was a gradual increase in the trypsin-resistant ]methionine, chased for 7 min, and incubated for 5-30 min with excess cold methionine. Cells were permeabilized with digitonin (75 g/ml) for 5 min, and permeabilized cells were incubated in the presence or absence of trypsin (200 g/ml) for 10 min. Trypsin digestion was halted by the addition of protease inhibitors, and cells were subsequently subjected to immunoprecipitation with a specific anti-apoB antibody. Immunoprecipitates were analyzed by SDS-PAGE and fluorography. ApoB radioactivity was quantitated by cutting and scintillation counting of the apoB. The results in A are mean Ϯ S.D. (n ϭ 3). b, cells were treated, permeabi-lized, andtrypsinized according to the protocol described in A. ApoB was immunoprecipitated, and equal aliquots of total apoB immunoprecipitates from both control and trypsin-treated cells were suspended in SDS-PAGE buffer and subjected to scintillation counting. The figure shows the results from a representative experiment and are given as a mean Ϯ S.D. (n ϭ 2). c, cells were treated according to the pulse-chase protocol described for determination of the basal rate of apoB translocation. The radioactivity of apoB and the 70-kDa apoB fragment was quantitated by cutting and scintillation counting of the specific apoB band. The figure shows the results of a representative experiment. apoB over time. Following 30 min of chase, there was on average only a 9% decrease in the amount of trypsin-resistant apoB in tunicamycin-treated cells compared with control cells. Despite this observation, tunicamycin treatment of HepG2 cells significantly reduced both intracellular and secreted levels of apoB (data not shown), possibly due to increased degradation, as previously reported (29). Overall, the data suggest that inhibition of N-linked glycosylation does not significantly affect the rate of apoB translocation.
Inhibition of Peptidylprolyl Isomerase Decreases ApoB Translocation-To investigate whether inhibition of peptidylprolyl isomerase could modulate the amount of apoB translocated into the ER, intact HepG2 cells were preincubated in the presence of 10 M cyclosporin A. In the presence of cyclosporin A, there was on average a 38% decrease after 30 min of chase (n ϭ 2) in the amount of intact trypsin-resistant apoB in cyclosporin A-treated cells compared with control cells over a 10 -30-min chase period. Fig. 11 shows the pattern of immunoprecipitable apoB and its fragments from trypsin-treated cells incubated in the presence and absence of cyclosporin A. The trypsin-generated fragmentation pattern did not differ appreciably between control and treatment conditions. This would suggest that inhibition of peptidylprolyl isomerase decreases the rate of apoB translocation without a detectable modification in the conformation of the protein.
Effect of DTT on the Translocation of ApoB-The effect of proper folding on the translocation of apoB was investigated by preincubating intact HepG2 cells with 2 mM DTT for 3 min. DTT was also present in the pulse and throughout the course of the chase period. The presence of DTT resulted in an average decrease of 63% in intact trypsin-resistant apoB after 30 min of chase. This decrease in the percentage of apoB translocated occurred despite an apparent increase of 14.3% in the amount of apoB synthesized in DTT-treated HepG2 cells. Fig. 12 shows the immunoprecipitable apoB fragments from trypsin-treated HepG2 cells incubated in the presence and absence of DTT. Of note is the appearance of a band in the 60 -70-kDa region in trypsin-untreated DTT treated cells. The identity of this band is unclear, but a similar protein was observed in a recent study (32), which suggested it to be an apoB degradation fragment generated in intact cells. Utilizing a battery of monoclonal apoB antibodies, we could not confirm this finding (data not shown). We are currently attempting to identify the nature of this band. DISCUSSION The present study reports direct investigation of apoB translocation using a new protocol involving pulse-chase labeling of intact HepG2 cells, permeabilization, and trypsin digestion of permeabilized cells. We suggest that a permeabilized system offers distinct advantages to the currently employed method of trypsin treatment of microsomes isolated by ultracentrifugation. Perhaps the most beneficial advantage is that permeabilization allows for the exposure of the ER and Golgi to exogenous trypsin without dramatically altering the morphology or the integrity of the organelles (33). A permeabilized cell system has recently been used to investigate the translocation, folding, and assembly of secretory and membrane proteins in HT-1080

FIG. 5. Effect of trypsin digestion of permeabilized HepG2 cells on control, luminal secretory proteins.
Nearly confluent cells were treated in the presence of 360 M oleate (2 h) and 10 M ALLN (1 h). Cells were pulsed for 3 min with [ 35 S]methionine, chased for 7 min, and incubated for 10 -30 min with excess cold methionine. Cells were either directly treated with trypsin (lanes 1-3) or initially permeabilized with digitonin (75 g/ml) for 5 min, and permeabilized cells were incubated in the presence or absence of trypsin (200 g/ml) for 10 min (lanes 4 -12). Trypsin treatment of selected dishes was also conducted in the presence of 0.5% Triton X-100 (lanes 3, 6, 9, and 12). Trypsin digestion was halted by the addition of protease inhibitors, and cells were subsequently subjected to immunoprecipitation with a specific anti-␣1-antitrypsin antibody. Immunoprecipitates were analyzed by SDS-PAGE and fluorography. B, cells were treated according to the pulse-chase protocol described in A. Cells were either directly treated with trypsin (lanes 1-3) or initially permeabilized before being subjected to trypsin treatment (lanes 4 -12) conducted as described in A. Trypsin digestion was halted by the addition of protease inhibitors, and cells were subsequently immunoprecipitated with a specific anti-transferrin antibody and analyzed as described in A.
human fibrosarcoma cells (34). The present translocation protocol could be applied to a variety of experimental conditions to investigate the mechanisms of apoB translocation.
It should be noted that some variability was observed between experiments in percentage of apoB protected at different times of chases, particularly at the earlier time points. Such variability makes fully quantitative measurement of apoB translocation rather difficult. This difficulty stems from the inherent complexity of the biochemical mechanism under investigation as well as the technical challenges of conducting the protocol. The exact nature of the association of apoB with the ER membrane is still not fully resolved and remains controversial. Some of the variabilities observed in the trypsin sensitivity of apoB in permeabilized cells may be due to actual biological alterations in the conformation of apoB during its association with the ER membrane. The higher variability observed at the earlier chase times of 5 and 10 min is actually expected, considering that freshly translated apoB chains would be still associated with ribosomes at this stage and would be largely cytosolic in nature. Sensitivity of these apoB chains to trypsin digestion could vary depending on their state of association with the ER membrane. A recent study has demonstrated that specific sequences within the translocating apoB molecule function to separate the ribosome/ER junction such that apoB is exposed to the cytosol (35). The differential opening of this junction is another factor that could possibly contribute to the variability in accessibility of trypsin. This factor is likely to be most influential during the earlier stages of the protocol, when the translocation rate of labeled apoB is highest. Also the amount of intact apoB at these early time points (5 min) does not represent the total apoB pool that would be present in the cytosolic domain and luminal compartment of the ER. Many of the chains digested and some of those protected from digestion would be smaller than full-length apoB and therefore would not be accounted for by counting the intact apoB band.
Interestingly, trypsin treatment of permeabilized HepG2 cells resulted in generation of a characteristic set of apoB fragments. The pattern and intensity of these fragments appeared to provide additional information regarding the translocation of apoB. The gradual decline in intensity of these fragments over time appeared to parallel the steady increase in the amount of trypsin-resistant intact apoB observed over the same time period. Both processes were found to be distinct but complementary evidence for a gradual movement of apoB into the lumen of intracellular membranes. Further analysis of the trypsin-generated apoB fragments demonstrated that the fragments were luminal in nature. Evidence demonstrating that these fragments are luminal further supports the notion that the fragments as well as the trypsin-resistant intact apoB are representative of translocation into the microsomal lumen.
Our data also indicate that the specific set of fragments having a molecular mass of 60 -70 kDa were from the Nterminal region of apoB. The characterization of the 60 -70-kDa fragments as originating from the N terminus of apoB provides additional information that cannot be readily ob-  Fig. 3. Permeabilized cells were either incubated in the presence or absence of trypsin (200 g/ml) for 10 min followed by the addition of protease inhibitors or were subjected to subcellular fractionation to isolate microsomes. Luminal apoB was extracted from the microsomes by sodium carbonate treatment and was separated from the membrane fraction by centrifugation (SW55, 35,000 rpm, 93 min). All samples were immunoprecipitated with a specific anti-apoB antibody and were analyzed by SDS-PAGE and fluorography. ApoB radioactivity was quantitated by either cutting and scintillation counting of apoB or by densitometric scanning. The results are given as the percentage of trypsin-resistant apoB (mean Ϯ S.D. (n ϭ 3)) and the percentage of luminal apoB (mean Ϯ S.D. (n ϭ 3)).

FIG. 7. Subcellular fractionation of permeabilized cells and analysis of apoB fragments.
Nearly confluent cells were treated in the presence of 360 M oleate (2 h) and 10 M ALLN (1 h). Cells were pulsed for 5 min with [ 35 S]methionine and were chased for 20 min with excess cold methionine. Cells were permeabilized with digitonin (75 g/ml) for 5 min, and permeabilized cells were incubated in the presence or absence of trypsin (200 g/ml) for 10 min. Trypsin digestion was halted by the addition of protease inhibitors, and cells were subsequently subjected to subcellular fractionation to isolate microsomes. Luminal apoB (L) was extracted from the microsomes by carbonate treatment and was separated from the membrane fraction (M) by centrifugation (SW55, 35,000 rpm, 93 min). Due to the large volume, each luminal sample was subsequently aliquoted into three fractions in order to properly immunoprecipitate the luminal component. All samples were immunoprecipitated with a specific anti-apoB antibody and were analyzed by SDS-PAGE and fluorography. tained from determination of the amount of intact apoB translocated. Alterations in the pattern of trypsin-generated fragments can be used to investigate changes in the conformation of the apoB molecule during its translocation across the ER membrane. It has been previously shown using an 89-kDa truncated form of the N-terminal region of apoB (apoB17) that this region has some degree of flexibility with respect to conformation (36). It is therefore likely, as we have demonstrated in the present paper, that changes in the translocation of apoB may result from changes in the conformation of apoB in the N-terminal region. Identification and analysis of intracellularly generated apoB fragments has been previously used in our laboratory (29,37) as well as others (38) as a marker of apoB degradation in permeabilized and intact cells. We have also demonstrated an alteration in the apoB fragmentation pattern in response to inhibition of N-linked glycosylation (29). It was suggested that this change in the apoB fragments was the result of a conformational change in the apoB molecule.
Comparison of the amount of apoB translocated over time based on either percentage of trypsin-resistant apoB or percentage of luminal apoB yielded similar results. The general trend of an increasing amount of apoB translocated over time was found using both techniques. Based on these two methods, it was determined that 70% of apoB was translocated after the maximum chase period in oleate-treated cells. This percentage is consistent with a previous study that determined that 71% of the initial labeled pool of apoB was recovered from the media in oleate-treated HepG2 cells incubated in the presence of ALLN (39). It is apparent that in order to achieve secretion of 71% of labeled apoB, a comparable translocation rate is required. In both the presence and absence of oleate there was a gradual increase in the percentage of trypsin-resistant apoB over the initial time periods. However, while this percentage in oleatetreated cells continued to increase, the amount of trypsinresistant apoB in untreated cells plateaued.
Our data demonstrating an oleate-stimulated increase in the percentage of apoB translocated is in agreement with a previous study that initially demonstrated that apoB was less sus-ceptible to protease digestion in microsomes isolated from oleate-treated HepG2 cells (40). In the presence of oleate, HepG2 cells assemble secretion-competent LDL-VLDL-like particles, while in a lipid-poor environment cells predominately produce a secretion-incompetent high density lipoprotein-like particle (22,25,26). We postulate, based on our results, that the mode of apoB translocation may differ in the assembly of these two types of lipoprotein particles. Increased availability of core lipids may result in an increased amount of apoB translocated directly into the lumen and assembled into an LDL-VLDL-like particle. In the absence of a sufficient lipid pool, translocation of apoB is slowed such that apoB will initially have a greater degree of association with the ER membrane.
Our results demonstrating that inhibition of N-linked glycosylation does not significantly affect the translocation of apoB was consistent with previous studies suggesting that glycosylation is not essential for the production and assembly of apoBcontaining lipoproteins (41). Although inhibition of glycosylation has been shown not to block the secretion of apoB (41), treatment with tunicamycin in the current study did result in a significant decrease in the amount of intracellular apoB. Previous work in our laboratory as well as others have also demonstrated a decrease in the total amount of apoB as a result of inhibition of glycosylation (29,41). This decrease has been suggested to result from an increased susceptibility to degradation rather than inhibition of protein synthesis (29,42). This degradation of apoB was largely ALLN-insensitive, suggesting the involvement of other proteases such as those involved in the degradation of misfolded or abnormal proteins in the ER.
The investigation of the effect of cyclosporin A (CsA) on the production of apoB demonstrated that inhibition of peptidylprolyl isomerase resulted in a decrease in the rate of translocation. Coupled to this effect was either a decrease in synthesis and/or an increase in the rate of degradation. Increased degradation of apoB in CsA-treated cells cannot be eliminated as a possibility despite the presence of ALLN. Agents that are suspected to induce a conformational change in apoB may result in degradation that is ALLN-insensitive. We have previously demonstrated ALLN-insensitive degradation of apoB through inhibition of glycosylation (29) and more recently through the disruption of disulfide bond formation. 2 However, it is important to note that any effect of CsA on apoB degradation should not influence the determination of its translocation rate (which is measured by determining the ratio of trypsin-resistant apoB to total apoB). Our results are in agreement with a recent study that reported a decrease in the synthesis and secretion of apoB in CsA-treated HepG2 cells (43). CsA has also been demonstrated to retard but not block the folding and secretion of transferrin from HepG2 cells (44). Although it is likely that CsA is retarding the folding of apoB, we suggest that the decreased apoB translocation in the presence of CsA may also result from a reduction in efficiency of lipid transfer, which is required for the assembly of apoB-containing lipoproteins. It has been demonstrated that peptidylprolyl isomerase activates protein-disulfide isomerase (45), a subunit of the microsomal triglyceride transfer protein (46). Reduction in lipid transfer as a result of alteration in the function of the microsomal triglyceride transfer protein may mimic a lipid-poor environment even in oleate-treated cells. The end result may be an increased percentage of apoB becoming membrane-bound and trypsin-accessible.
The disruption of disulfide bond formation through the use of DTT resulted in a decrease in the percentage of apoB translo-2 J. Macri and K. Adeli, submitted for publication. . Cells that had been initially treated with DTT were subsequently treated with DTT throughout the course of the experiment. Cells were treated according to the translocation protocol described in Fig. 3. Cells were subsequently subjected to immunoprecipitation with a specific anti-apoB antibody, and the immunoprecipitates were analyzed by SDS-PAGE and fluorography. ApoB radioactivity was quantitated by cutting and scintillation counting of the apoB band. The figure shows a representative experiment performed in the presence and absence of DTT.
cated. This reduction occurred despite an apparent increase in the synthesis of apoB. We suggest that DTT induced a conformational change to the extent that apoB was misfolded. Misfolding altered the translocation of apoB such that a higher percentage became trypsin-accessible. A number of studies have utilized DTT to investigate the folding of secretory proteins in the ER (46 -50). In the presence of DTT, cysteine residues on secretory proteins are prevented from forming proper disulfide linkages resulting in a misfolded protein. Such misfolded proteins can either be retained in the ER (47) or degraded (48).
Evidence supporting the suggestion that DTT induced the misfolding of apoB may be obtained from analysis of the apoB fragments resulting from trypsin digestion. There was a distinct difference in the pattern of apoB fragments between cells treated in the presence and absence of DTT. Most notable was an alteration in the size and number of N-terminal fragments occurring in the 60 -70-kDa region in DTT-treated cells. In the presence of DTT, there was an increased number of fragments occurring in the 60 -70-kDa region with molecular weights that were slightly different from those observed in control cells. The increased number of fragments suggests that treatment with DTT caused a conformational change in the N-terminal region of apoB such that additional trypsin cleavage sites became exposed. These results are consistent with a previous study that investigated the structural domains of apoB in LDL (51). These researchers demonstrated that treatment of LDL with DTT generated additional apoB fragments upon trypsin digestion when compared with untreated cells and suggested that this observation was the result of a DTT-induced conformational change in apoB (51). We suggest that the decrease in the percentage of translocated apoB in HepG2 cells treated with DTT could result from a disruption of disulfide bond formation occurring in the disulfide-abundant N-terminal region of apoB. This disruption could possibly impair the proper folding of apoB, resulting in an increased percentage of apoB becoming membrane-associated and trypsin-accessible. The possibility also exists that treatment of HepG2 cells with DTT may inhibit the proper transfer of lipids required for lipoprotein assembly.
Based on the data presented we suggest that translocation of apoB can be regulated not only by the availability of a sufficient lipid pool but also by the conformational status of the apoB nascent chain. The question now becomes, "Does the lack of lipid cause a conformational change in apoB that impairs its translocation, or do conformational changes in apoB impede the transfer of required lipids, affecting translocation and assembly?" These two events may be intimately linked, and further investigations will be needed to address these questions.