Brefeldin A reversibly inhibits the assembly of apoB containing lipoproteins in McA-RH7777 cells.

BFA inhibited in a dose dependent way the assembly of apoB-48 very low density lipoprotein (VLDL) but allowed a normal rate of biosynthesis of the apolipoprotein and of the assembly of the dense (“high density lipoprotein (HDL)-like”) apoB-48 particle (apoB-48 HDL). The inhibition of the assembly of apoB-48 VLDL occurred at BFA levels that allowed a major secretion of both transferrin and apoB-48 HDL. The assembly of apoB-100 containing lipoproteins was also inhibited by BFA but could be reactivated by a 30-60 min chase in the absence of BFA, which agreed with the time that was estimated to be needed to restore the secretory pathway (approximately 60 min). Also the assembly of apoB-48 VLDL was reversible. Both apoB-48 and apoB-100 that was labeled in the presence of BFA assembled VLDL after removal of the BFA. Both apoB-100 and apoB-48 were associated with the membrane pellet of the microsomes. Virtually all (122 ± 30%) of the membrane associated pulse-labeled apoB-48 remained in the membrane after a 180-min chase in the presence of BFA, compared to only 21 ± 2% in normal cells (mean ± S.D., n = 4). The corresponding figures for apoB-100 was 40 ± 7% in BFA-treated cells and 9 ± 7% in normal cells (mean ± S.D., n = 4). Pulse-chase experiments with BFA offered conditions to selectively follow the turnover of membrane-associated apoB-100. Such experiments indicated that this apoB-100 pool is a precursor to VLDL.

Apolipoprotein B (apoB) 1 is the major protein component of the triacylglycerol-and cholesterol ester-rich plasma lipoproteins: chylomicrons, very low density lipoproteins (VLDL), intermediate density lipoproteins, and low density lipoproteins (LDL).
Two forms of apoB exist, referred to as apoB-100 and apoB-48 (1). ApoB-100 consists of 4536 amino acids, while apoB-48 corresponds to the N-terminal 48% of apoB-100. Both proteins are coded for by the same gene (2). The apoB-100 mRNA is converted to mRNA for apoB-48 by an enzymedependent deamination of cytidine in codon 2153, which converts this codon into the signal for translational stop (3)(4)(5).
In humans, apoB-48 is expressed mainly in the intestine, where it assembles chylomicrons, while apoB-100 is synthesized in the liver and is necessary for the assembly of VLDL particles (6). In contrast, the rat hepatocytes, as well as the rat hepatoma cells McA-RH7777 (7-9) express substantial amounts of both apoB-100 and apoB-48 (7, 10 -12).
Recent results (7) from studies in McA-RH7777 cells indicate that during translation/translocation apoB-48 forms a small dense particle (100 Å in diameter), floating as HDL. This small particle appears to acquire the major amount of lipids in a second step and is in this way converted into a VLDL particle. The second step is dependent upon ongoing biosynthesis of proteins other than apoB; moreover, it is stimulated by oleic acid (7).
The assembly of apoB-100 VLDL is more complicated (7), and our results indicate that the VLDL particles are assembled at least to a certain extent in relatively close connection with the translation/translocation of the protein.
Brefeldin A (BFA) is a fungal metabolite that has been shown to block the intracellular transfer and secretion of proteins (13), resulting in a reorganization of the secretory pathway (14,15). Both the reorganization of the pathway and the inhibition of the intracellular transport are reversible (16). The reorganization is due to an inhibition of the budding of transport vesicles in the pathway concurrent with continued activity of the retrograde transport. Of importance for the budding of transport vesicles in the secretory pathway is the binding of a 21-kDa protein, ADP-ribosylation factor (ARF), to the membrane, thus initiating the formation of coatomers (17,18). ARF needs to bind GTP to be able to bind to the membrane, and this GTP is hydrolyzed when the coat of the transport vesicle is depolymerized. The GDP is then exchanged for GTP by a guanidine nucleotide exchange protein. It is believed that BFA inhibits this exchange and thus prevents ARF from initiating the budding of transport vesicles, which in turn leads to the inhibition of the vesicular transport through the secretory pathway (19 -21). Recent results suggest that BFA inhibits an unknown target protein of importance for the function of the guanidine nucleotide exchange protein (22).
BFA has frequently been used to block the intracellular transport of proteins, allowing investigators to dissect important sorting and folding processes in the early part of the secretory pathway. For example, results using BFA-treated cells have demonstrated a post-translational degradation of apoB-100 (23). Since the assembly of apoB-containing lipoproteins is a complex process involving at least two steps (7) that may take place in different parts of the endoplasmic reticulum (24,25), it cannot be ruled out that BFA could have fundamental influences on this assembly process.
In this paper we have investigated the effect of BFA on the assembly of apoB-100 VLDL and apoB-48 VLDL.

EXPERIMENTAL PROCEDURES
Materials-Eagle's minimum essential medium, nonessential amino acids, glutamine, penicillin, and streptomycin were from ICN Biomedicals (Costa Mesa, Ca). Fetal calf serum was from JRH Biosciences (United Kingdom), and brefeldin A from Epicentre Technologies (Madison, WI). Methionine, fatty acid-free bovine serum albumin, sodium pyruvate, disodium carbonate, sodium hydrogen carbonate, phenylmethylsulfonyl fluoride, pepstatin A, and leupeptin were from Sigma. Rabbit immunoglobulin was from DAKO (Glostrup, Denmark) and rabbit IgG fraction to rat transferrin from Organon Teknika Corp. (West Chester, PA). Trasylol ® (aprotinin) was from Bayer Leverkusen (Leverkusen, Germany). Immunoprecipitin and Eagle's minimal essential medium without methionine were from Life Technologies, Inc. (Paisley, Scotland). N-Acetyl-Leu-Leu-norleucinal was from Boehringer Mannheim (Mannheim, Germany). Amplify ® , [ 35 S]methionine-cysteine mixture, and Rainbow colored protein molecular weight markers were from Amersham (Amersham, UK), and Ready Safe ® was from Beckman (Fullerton, CA). All chemicals used for SDS-PAGE were from Bio-Rad.
Cell Culture-McA-RH7777 cells were cultured as described earlier (7) in Eagle's minimal essential medium, containing 20% fetal calf serum, 1.6 mM glutamine, 8.0 mM NaHCO 3 , 1.6 mM sodium pyruvate, 140 mg of streptomycin/ml, 140 IU of penicillin/ml, and 60 mg of nonessential amino acids/ml, in 5% CO 2 , at 37°C. The cells were split twice a week and fed every day.
Metabolic Labeling-The cells were pulse-labeled and chased as described (7) with the exception that [ 35 S]methionine-cysteine mixture was used instead of [ 35 S]methionine. Unless otherwise stated, the cells were incubated with 360 M oleic acid (26) for 2 h prior to, and during the experiments. Brefeldin A was dissolved in ethanol and added to the culture medium. Isolation of cells and the microsomal fraction was carried out as described (26). The luminal content of the vesicles was separated from the vesicle membranes by the sodium carbonate method (27), with some modifications as described (7,28). The following protease inhibitors were used: 0.1 mM leupeptin, 1 mM phenylmethylsulfonyl fluoride, 1 mM pepstatin A, 5 mM N-acetyl-Leu-Leu-norleucinal, and 100 KIU of aprotinin/ml.
Sucrose Gradient Ultracentrifugation of Lipoproteins-The lipoproteins present in the microsomal lumen or in the medium were separated by sucrose gradient ultracentrifugation. The gradient (7) was formed by layering the following from the bottom of the tube: 2 ml of 49% sucrose, 2 ml of 25% sucrose, 5 ml of the sample in 12.5% sucrose, and 3 ml of phosphate-buffered saline. All solutions contained 0.1 mM leupeptin, 1 mM phenylmethylsulfonyl fluoride, 1 mM pepstatin A, 5 mM N-acetyl-Leu-Leu-norleucinal, 100 KIU of aprotinin/ml, and 0.5 mM EDTA. The gradients were centrifuged in a Beckman SW40 rotor at 35,000 rpm, at 12°C, for 65 h. The gradients were unloaded from the bottom of the tube into 12-13 fractions.
Immunoprecipitation of ApoB and Electrophoresis-ApoB was precipitated from cells, medium as well as from fractions of the gradient, as described (7). Electrophoresis in SDS-polyacrylamide gels, autoradiography, and determination of the radioactivity in the proteins separated in the gels was carried out as described (29).

RESULTS
The Effect of Brefeldin A on the Assembly of ApoB Lipoproteins-McA-RH7777 cells were incubated with different amounts of BFA for 15 min, pulsed with [ 35 S]methionine-cysteine mixture for 10 min (in the presence of the indicated concentration of BFA) and then chased for periods of 30 or 120 min (also in the presence of the indicated amount of BFA).
The culture medium was recovered after the 120-min chase and the secretion of pulse-labeled transferrin was determined. The culture medium was also analyzed by gradient ultracentrifugation and apoB-48 and apoB-100 was recovered by immunoprecipitation and SDS-PAGE from each of the fractions of the gradient and the radioactivity was determined. As reported earlier (7), normal McA-RH7777 cells secreted apoB-100 on particles with the density of VLDL while apoB-48 is secreted both on VLDL and HDL-like particles (the major amount though on VLDL). In these experiments we therefore followed the radioactivity of apoB-100 as well as that of apoB-48 in the VLDL fraction (referred to as apoB-100 VLDL and apoB-48 VLDL, respectively). We also followed the apoB-48 radioactivity in the HDL density regions. The dense particle with apoB-48 that bands in the HDL density region will be referred to as apoB-48 "HDL." The results (Fig. 1A) indicated that the secretion of transferrin differed from that of apoB-48 VLDL with regard to the sensitivity to BFA. Thus while 70% of the original secretion (i.e. FIG. 1 A, dose-dependent inhibition of the secretion of transferrin, apoB-100 VLDL, apoB-48 VLDL, and the apoB-48 containing dense particle (B48 HDL) by BFA. McA-RH7777 cells, cultured in the presence of oleic acid, were preincubated with different concentrations of BFA (0, 0.1, 0.2, 0.5, 1, 2, 5, or 10 g/ml culture medium as indicated on the x axis) for 15 min, pulse-labeled for 10 min (in the presence of BFA as indicated), and finally chased for 120 min (in the presence of BFA as indicated). Transferrin was isolated from the culture medium by immunoprecipitation and PAGE in the presence of SDS. The band corresponding to transferrin was identified by autoradiography, cut out of the gel, and the radioactivity was counted. The culture medium was also subjected to gradient ultracentrifugation, apoB was immunoprecipitated from each fraction of the gradient, and apoB-100 and apoB-48 was separated by SDS-PAGE, followed by autoradiography. The bands corresponding to apoB-100 and apoB-48 in the VLDL fraction as well as apoB-48 in apoB-48 HDL were cut out of the gel, and the radioactivity was determined. The results are expressed as percent of the secretion obtained in the absence of BFA (mean Ϯ S.D., n ϭ 4). B, dose-dependent inhibition of the assembly of apoB-100 and apoB-48 containing lipoproteins by BFA. McA-RH7777 cells, cultured in the presence of oleic acid, were preincubated with different concentrations of BFA (0, 0.1, 0.2, 0.5, 1, 2, 5, or 10 g/ml culture medium as indicated in the figure) for 15 min, pulse-labeled for 10 min (in the presence of BFA as indicated), and finally chased for 30 min (in the presence of BFA as indicated). The total microsomal fractions were then recovered and the luminal content extracted by sodium carbonate and analyzed by gradient ultracentrifugation. ApoB was recovered from each fraction by immunoprecipitation and analyzed by SDS-PAGE followed by autoradiography. the secretion seen in the absence of BFA) of transferrin was retained at a BFA concentration of 0.1 g/ml and 60% at 0.2 g/ml, there was a rapid drop in the secretion of apoB-48 VLDL with increasing concentration of BFA. Thus only traces (6%) of the original secretion of apoB-48 VLDL were seen at a BFA concentration of 0.2 g/ml. Interestingly enough the secretion of apoB-48 HDL closely followed that of tranferrin but differed from that of apoB-48 VLDL.
Also the rate of secretion of apoB-100 VLDL decreased with increasing concentration of BFA, although the effect of BFA was not as pronounced as on the secretion of apoB-48 VLDL. However, in the presence of 0.2 g/ml BFA the remaining secretion of apoB-100 VLDL was significantly lower than that of transferrin and apoB-48 HDL.
After the 30-min chase, the luminal content of the total microsomal fraction was recovered and analyzed by gradient ultracentrifugation. ApoB-48 and apoB-100 were recovered by immunoprecipitation from each of the fractions of the gradient and analyzed by SDS-PAGE.
The results (Fig. 1B) indicated that the appearance of apoB-48 VLDL in the microsomal lumen closely followed the secretion of the lipoprotein into the medium. Thus, in agreement with the almost total loss of secretion of apoB-48 VLDL from cells treated with 0.2 g/ml BFA, no assembly of the lipoprotein could be detected in these cells (compare Fig. 1, A and B). On the contrary, a substantial amount of apoB-48 HDL was assembled even at a BFA concentration of 10 g/ml. Thus the assembly of this dense particle did not seem to be influenced by BFA but its rate of secretion was rather depending on the availability of an open secretory pathway.
Also the appearance of apoB-100 VLDL in the secretory pathway followed the secretion of the lipoprotein. Thus there was a significant reduction in the assembly of the lipoprotein at a BFA concentration of 0.2 g/ml and an almost complete failure to detect apoB-100 in the microsomal lumen at BFA concentrations above 0.5 g/ml.
The effect of BFA on the assembly of the apoB-containing lipoproteins was also investigated by pulse-chase studies. In these experiments the cells were treated with 10 g/ml BFA. The BFA-treated cells were preincubated for 15 min and pulsed for 10 min in the presence of BFA, as described above, and then chased (in the presence of BFA) for periods between 0 and 90 min. The normal McA-RH7777 cells were treated in the same way but in the absence of BFA.
In accordance with our previous results (7), the normal cells assembled apoB-100 VLDL and apoB-48 VLDL. The results from the entire pulse-chase experiment are not shown but the relevant information could be illustrated by Fig. 1B (0 g/ml BFA) which shows normal cells that have been pulsed for 10 min and chased for 30 min, i.e. the chase period after which the maximal amount of apoB radioactivity was recovered with the two VLDL particles from the microsomal lumen (7). As expected, apoB-48 formed, in addition to VLDL, also apoB-48 HDL. Also in agreement with our previous results (7) we demonstrated the assembly of apoB-100 particles with higher density which were not secreted from the cells.
In agreement with the experiments shown in Fig. 1B (10 g/ml BFA), no apoB-100 VLDL could be detected in the microsomal lumen of the BFA-treated cells after any of the chase periods (Data not shown). However, trace amounts of apoB-100 radioactivity could be detected in the denser fractions of the gradient during these experiments.
The assembly of apoB-48 VLDL was also inhibited, however, apoB-48 HDL could be isolated from the lumen of the total microsomal fraction after the different periods of chase (c.f. To obtain an estimate of the assembly of the different apoB-48 and apoB-100 containing lipoproteins in the BFAtreated cells, we compared the formation of the particles in these cells with that in normal cells. The cells were pulsed for 10 min and chased for 30 min (see above). The luminal content of the total microsomal fraction was recovered and analyzed by gradient ultracentrifugation. ApoB-100 and apoB-48 was recovered from each fraction of the gradient by immunoprecipitation and SDS-PAGE. The apoB-48 radioactivity in apoB-48 VLDL and apoB-48 HDL was determined. We also determined the apoB-100 radioactivity present in apoB-100 VLDL, as well as in the more dense particles (present in the HDL and LDL density range). The results were expressed as the relation (percent) between the radioactivity of the lipoprotein recovered from BFA-treated cells and the corresponding lipoprotein recovered from normal cells. As could be expected from the appearance of the autoradiograms (Fig. 1B, compare 0 and 10 g/ml BFA) we could not detect any significant apoB-48 radioactivity in apoB-48 VLDL recovered from the microsomal lumen of the BFA-treated cells. On the contrary, the apoB-48 radioactivity associated with apoB-48 HDL recovered from the BFA-treated cells amounted to 120 Ϯ 27% (mean Ϯ S.D., n ϭ 4) of the corresponding radioactivity found in normal cells. Only traces of apoB-100 containing lipoproteins was detected in the microsomal content of the BFA-treated cells, thus in the HDL ϩ LDL density range we found 7 Ϯ 7% (mean Ϯ S.D., n ϭ 5) of the radioactivity present in the same density range in normal cells. The corresponding recovery of radioactive apoB-100 in apoB-100 VLDL was 2 Ϯ 3% (mean Ϯ S.D., n ϭ 5).
These results indicate that the only apoB containing lipoprotein that was assembled in the BFA-treated McA-RH7777 cells was apoB-48 HDL. The rate of assembly of this lipoprotein appeared to be in the same range as that in normal cells.
The rate of biosynthesis of apoB-48 and apoB-100 in normal and BFA-treated cells were estimated as the incorporation of radioactivity during a 10-min pulse with [ 35 S]methionine-cysteine mixture. There were no differences between normal and BFA-treated cells as to the radioactivity recovered in apoB-48 (at the highest BFA concentration used (10 g/ml): 36,690 Ϯ 6,065 dpm (normal) versus 36,963 Ϯ 2,808 dpm (BFA treated); mean Ϯ S.D., n ϭ 5). There was (at 10 g/ml BFA), however, a slight decrease in the total incorporation of radioactivity into apoB-100 (19,534 Ϯ 4,076 dpm versus 27,193 Ϯ 4,992; mean Ϯ S.D., n ϭ 5).
Effect of Brefeldin A on the Turnover of ApoB in the Microsomal Membrane-Contrary to the failure to detect apoB-100 in the microsomal lumen of the BFA-treated cells, we found a substantial amount of the protein associated with the membrane pellet. Also a significant amount of the apoB-48 radioactivity was recovered with the membrane pellet after the carbonate treatment (Fig. 2).
In the experiment shown in Fig. 2 we investigated the turnover of the membrane-associated form of apoB-100 and apoB-48. The BFA-treated cells (10 g/ml culture medium) were pulse-labeled for 10 min with [ 35 S]methionine-cysteine mixture and chased for periods between 0 and 180 min. The results indicated that there was a slow turnover of the membraneassociated apoB-48 in BFA-treated cells. To further investigate this we pulse labeled the cells for 10 min and chased for 10 and 180 min and determined the proportion of the radioactivity of the membrane-associated apoB, present after the 10-min chase, that remained in the cell after the 180-min chase. In normal cells only 21 Ϯ 2% of the radioactivity of the membraneassociate apoB-48 could be recovered after 180 min chase while in the BFA-treated cells we recovered 122 Ϯ 30% (data are mean Ϯ S.D., n ϭ 4). There was also a decrease in the turnover of membrane-associated apoB-100 in BFA-treated cells, however, not that pronounced as for apoB-48. Thus 40 Ϯ 7% of the membrane-associated apoB-100 remained after 180 min chase of the BFA-treated cells while 9 Ϯ 7% remained in normal cells (data are mean Ϯ S.D., n ϭ 4).
Reactivation of the Assembly Process-Since the effect of BFA on the intracellular transport is reversible, we investigated whether the assembly of the apoB-containing lipoproteins could be reactivated by chasing the cells in the absence of BFA. In the first experiments the cells were preincubated with BFA for 15 min and chased for 40 min in the presence of BFA. The cells were then chased for periods of 30, 60, and 120 min in the absence of BFA. After each chase period, the cells were incubated with [ 35 S]methionine-cysteine mixture for 90 min, and the appearance of apoB-containing VLDL was followed in the medium and in the luminal content of the microsomes. The results clearly indicated that already after a 30-min chase in the absence of BFA (ϩ90-min incubation with [ 35 S]methioninecysteine mixture), the cells assembled and secreted radiolabeled apoB-100 VLDL as well as apoB-48 VLDL (data not shown).
A second set of experiments was carried out to determine the time needed for the reactivation of the assembly process. Cells were preincubated with BFA for 15 min and then chased for 0, 5, 15, 30, and 60 min in the absence of BFA. After each chase period the cells were labeled with [ 35 S]methionine-cysteine mixture for 10 min. The total microsomal fraction was then recovered and its luminal content analyzed by gradient ultracentrifugation. The results (Fig. 3A) indicated that the assembly of apoB-100 VLDL had been reactivated after a 30-min chase in the absence of BFA. In two additional experiments we could detect the assembly of apoB-100 VLDL after the 45 and 60 min chase, respectively. Thus the reactivation of the assembly of apoB-100 VLDL in cells treated with 10 g/ml BFA required a BFA-free chase of 30 -60 min. To determine how this reactivation time compared with the time needed to reconstitute the secretory pathway, we followed the secretion of transferrin. BFA-treated cells ( The cells were then chased in the absence of BFA for periods between 0 and 120 min. The medium was recovered after each chase period and transferrin was immunoprecipitated and analyzed by SDS-PAGE, followed by autoradiography. The band corresponding to transferrin was cut out of the gel and the radioactivity determined. The normal cells were treated in the same way except that BFA was omitted. were then chased in the absence of BFA for periods between 0 and 120 min. After each period we followed the secretion of radioactive transferrin in the medium. The results indicated that in the normal cells, transferrin started to be secreted after a chase of 20 -30 min (Fig. 3B). In the BFA-treated cells the secretion of transferrin was delayed; it started to appear in the medium after 90 min chase. It should be pointed out that the total biosynthesis of transferrin and the final amount of radioactive transferrin secreted was comparable under the two conditions (not shown). We also followed the secretion of the high molecular mass immunoglobulin-binding protein (the tentative unprocessed complement factor C3) (30). This protein was secreted after a 30-min chase from untreated cells, while BFAtreated cells started to secrete it after 90 min (not shown). These results indicate that it takes approximately 60 min for the secretory pathway to be reorganized.
No attempts were made to estimate the time needed to reactivate the assembly process for apoB-48 VLDL. The reason for this is elaborated under "Discussion." To evaluate whether apoB-48 that was pulse-labeled in the presence of BFA could be utilized for the assembly of VLDL when the BFA had been removed and the assembly process reactivated, we repeated one of our earlier published experiments (7) comparing normal and BFA-treated cells (Fig. 4). The cells were preincubated with or without BFA for 15 min and pulse-labeled with [ 35 S]methionine-cysteine mixture for 30 min in the presence or absence of BFA. The cells were then chased for 120 min in the absence of BFA (both BFA-treated and normal cells). The preincubation, labeling, and 120-min chase were all carried out in the absence of oleic acid. After the 120-min chase, the luminal content of the cells contained apoB-48 HDL (Fig. 4A). ApoB-48 was also associated with the membrane of the microsomal fraction (not shown). There were no obvious differences between normal and BFA-treated cells. The culture medium was then changed and oleic acid added, and the cells were chased for 180 min. Analyses of the medium after the 180-min chase revealed that apoB-48 VLDL was the quantitatively dominating form of apoB-48 secreted in both normal and BFA-treated cells (Fig. 4B). The observation in the normal cells agreed with our previous observation (7). These data indicated that the assembly of apoB-48 VLDL could be reactivated and that full-length apoB-48 synthesized in the presence of BFA could be used for the biosynthesis of VLDL.
Only small amounts of apoB-100 VLDL could be detected in the medium after the 180-min chase, provided that the cells were extensively washed after the preceding 120-min chase. These results are in agreement with our earlier observations (7).
The results shown in Figs. 1 and 2 indicate that BFA inhibits the formation of apoB-100 lipoproteins and that the protein is associated with the membrane. To address the question if apoB-100 that had been labeled in the presence of BFA and was associated with the membrane could be used for the assembly of VLDL, the following experiment was carried out: the cells were cultured in the absence of oleic acid and preincubated with BFA (10 g/ml) for 15 min. After a 30-min pulse (in the presence of BFA) the cells were chased for 30 min (in the presence of BFA) to allow nascent polypeptides to be completed. As could be expected from the results shown in Figs. 1 and 2, only full-length apoB-100, associated with the membrane, could be detected after this chase and no apoB-100 was present in the lumen (not shown but compare Figs. 1 and 2). The culture medium was changed and the cells were chased for 90 min in the absence of BFA. The culture medium was again changed and the cells were chased for 180 min in the absence of BFA but in the presence of oleic acid. The luminal contents of the microsomes were recovered after the 90-min chase, we collected the culture medium after the 90 and 180 min chase. Both the luminal contents of the microsomes (Fig. 5A) and the culture medium (Fig. 5B) were analyzed by gradient ultracentrifugation. ApoB was recovered from each fraction of the gradient by immunoprecipitation and analyzed by autoradiography. The obtained results showed that after the 90-min chase the apoB-100 that originally only was associated with the membrane of the microsomes, now appeared on VLDL as well as on denser particles in the secretory pathway (Fig. 5A), moreover The total microsomal fraction was recovered and the luminal content was extracted by sodium carbonate and subjected to ultracentrifugation in a sucrose gradient. The gradient was unloaded from the bottom, and apoB-100 and apoB-48 were recovered from each fraction by immunoprecipitation and analyzed by SDS-PAGE followed by autoradiography. B, oleic acid induces the assembly of VLDL particles containing radioactive apoB-48 (open squares) that had been labeled in the absence (ϪBFA) or presence (ϩBFA) of BFA. (Open circles represent apoB-100.) McA-RH7777 cells, cultured in the absence of oleic acid, were preincubated, pulsed, and chased as in A. After the 120-min chase, the culture medium was changed and the cells were chased for an additional 180 min in the presence of oleic acid. The culture medium was recovered and subjected to ultracentrifugation in a sucrose gradient. The gradients were unloaded from the bottom, and apoB was recovered from each fraction by immunoprecipitation and analyzed by SDS-PAGE followed by autoradiography. The bands corresponding to apoB-48 and apoB-100 were also cut out of the dried gel and digested, and the radioactivity was counted. apoB-100 VLDL had accumulated in the medium and this accumulation continued during the 180-min chase (Fig. 5B). A substantial amount of the apoB-48 HDL was present in the lumen after 90 min chase while only a small amount of apoB-48 VLDL could be detected (Fig. 5A). The apoB-48 containing dense particle accumulated in the medium after 90 min chase (carried out in the absence of oleic acid) while only a small amount of apoB-48 VLDL was seen after this chase period (Fig.  5B). However, a substantial accumulation of apoB-48 VLDL was seen in the medium after the 180-min chase that was carried out in the presence of oleic acid (Fig. 5B).
These results indicate that full-length apoB-100 associated with the membrane could serve as a precursor for the assembly of VLDL particles as well as for the more dense apoB-100 containing particles. The results also support the results shown in Fig. 4.
Effect of BFA on the Biosynthesis of Triacylglycerol-One possibility is that BFA influences the rate of biosynthesis of lipids. To test this, the cells (normal cells incubated with or without oleic acid or cells that had been incubated with oleic acid and treated with BFA) were labeled with [ 14 C]glycerol for periods between 0 and 60 min and the incorporation of radioactivity into triacylglycerol was measured. The results (data not shown) indicated that BFA did not influence the biosynthesis rate of triacylglycerol. DISCUSSION In this paper we present results that indicate that BFA has a profound effect on the assembly of VLDL particles in McA-RH7777 cells. Thus virtually no apoB-48 VLDL was formed in the cells treated with BFA at concentrations of 0.2 g/ml and higher. On the contrary the biosynthesis of apoB-48 was not influenced by BFA. Moreover the protein was to a significant degree translocated to the dense (HDL-like) particle (referred to as apoB-48 HDL) that has been indicated to be converted to VLDL in a second step (7). The assembly of apoB-48 HDL appeared to be as large in the BFA-treated cells as in the normal cells. ApoB-48 was also associated with the membrane of the microsomes.
Our earlier results have indicated a two-step mechanism for the assembly of apoB-48 VLDL, the first step being the translation/translocation of apoB-48, forming apoB-48 HDL, a 100-Å particle which in a second step is converted into VLDL by the addition of the dominating amount of lipids (7). Taken together the obtained results indicate that BFA inhibits the assembly of apoB-48 VLDL by selectively interfering with the second step in the process, but allowing the first step to occur, i.e. the protein is translated and translocated to the lumen of the endoplasmic reticulum lumen. The observation that BFA inhibits the turnover of the membrane-associated apoB-48 cannot yet be explained. It could perhaps suggest also that this form of apoB-48 could be used as substrate for the assembly of VLDL. If this is the case it could be envisioned that an inhibition of the second step results in a reduced mobilization of apoB-48 from the membrane and thus a decrease in the turnover of this pool of the protein. However, it has been indicated (31-33) that the membrane-bound apoB, at least apoB-100, undergoes post-translational degradation. It is of course possible that a subset of the apoB molecules in the membrane has the capacity to form lipoproteins. Our observation that the membrane-associated apoB-100 is a precursor for the assembly of VLDL may argue in favor of this possibility.
The second step is highly dependent on the availability of fatty acids (7). Thus in the presence of oleic acid the McA-RH7777 cells secrete most of apoB-48 on VLDL particles while in the absence of the fatty acid the cells assemble and secrete almost exclusively apoB-48 HDL (these observations are confirmed in the present paper). Therefore it seems as if BFA induces, in cells cultured in the presence of oleic acid, an assembly of apoB-48 lipoproteins that very much resemble that of cells that had not been incubated with oleic acid. Thus BFA appears to dissociate the translation/translocation from the major addition of the core lipids, in particular triacylglycerol. Obviously, an inhibition of the biosynthesis of triacylglycerol by BFA could explain our observations. We have, however, demonstrated the same rate of biosynthesis of triacylglycerol in normal and BFA-treated cells.
It is possible that the failure to assemble apoB-48 VLDL is due to the reorganization of the secretory pathway induced by BFA. However, both the secretion of apoB-48 VLDL and the appearance of the lipoproteins in the microsomal lumen were almost completely inhibited by a BFA concentration that allowed a significant secretion of transferrin. Even more interesting was the observation that apoB-48 HDL was still secreted at a BFA concentration that almost completely inhibited the assembly and secretion of apoB-48 VLDL. Thus BFA inhibits the secretion of apoB-48 HDL and apoB-48 VLDL by different mechanisms.
The inhibition of the secretion of apoB-48 HDL followed that of transferrin, indicating that the influence of BFA on the secretion of apoB-48 HDL is due to the inhibition of the transport through the secretory pathway. This is in agreement with the observation that BFA did not interfere with the assembly of the particle. Together our results indicate that the inhibition of the assembly of apoB-48 VLDL (and the subsequent secretion of the lipoprotein) occurs at BFA levels that leave the secretory pathway largely intact and allows assembled apoB-48 containing lipoproteins to be secreted. This in turn points to the pres- ence of a specific mechanism by which BFA influences the assembly of apoB-48 VLDL, a mechanism that differs from the major influence of the molecule on the transport through the secretory pathway.
Another possibility is that BFA inhibits the transport of the precursor to the site of the second step. The presence of such subcompartments has been indicated (24,25). However, there are no results that support a vesicular transport between them. Arguing against an inhibition of the transport to the site of the second step is again the difference between the concentration of BFA needed to inhibit the assembly of apoB-48 VLDL and that needed to inhibit the transport of transferrin and apoB-48 HDL through the secretory pathway. The intepretation of our data as a support for the presence of a vesicular transport between the two steps in the assembly process implies that this transport is unique for apoB-48 that should be assembled into VLDL and is not utilized by other secretory proteins, such as transferrin or even by the other apoB-48 containing lipoproteins, apoB-48 HDL, assembled in the cell. Although this does not seems likely, we can at present not rule out the possibility of such a specific transport between the first and second step in the assembly process.
Finally, BFA may interact with a target protein that is needed for the second step in the conversion of apoB-48 to VLDL. Our previous results (7) have indicated that the second step in the assembly process depends upon ongoing biosynthesis of proteins other than apoB-48. There are several reports indicating that BFA inhibits the intracellular transfer by binding to a target protein (15, 19, 34 -36). BFA inhibits the intracellular transfer and induces the reorganization of the secretory pathway by preventing the budding of transport vesicles. Essential for this budding is the binding of GTP to ARF. The GTP-loaded ARF can bind to the membrane and initiate the polymerization of the coat-protein subunits to form the coatomer (17,18). BFA inhibits the nucleotide shift (GDP to GTP) on ARF, thus preventing ARF from initiating the budding of transport vesicles (19 -21). Recent data may suggest that the target for BFA is another factor than the guanidine nucleotide exchange protein but central for its activity (21). It could be suggested as a subject for future studies that the second step in the apoB-48 VLDL assembly process involves a GTP-binding protein with a GTP-GDP cycle that is essential for the lipid loading in this step. This could be a protein involved in a specific transfer of precursors to the VLDL assembly or a protein involved in the formation or addition of the major amount of lipids to the VLDL particle.
Like apoB-48, apoB-100 did not form VLDL particles in the presence of BFA, moreover it was not to any significant degree assembled into the denser particles that have been identified in the microsomal lumen (7,26). Thus only traces of apoB-100 could be recovered from the microsomal lumen of BFA-treated cells. BFA induced a decrease in the biosynthesis of apoB-100, however, more than 70% of the biosynthesis rate seen in normal cells remained after treatment with 10 g/ml BFA. Thus the decrease of the biosynthesis rate of apoB-100 could not explain the almost complete absence of apoB-100 lipoproteins in the secretory pathway. Instead we observed that the dominating amount of apoB-100 radioactivity was resistant to extraction with sodium carbonate, indicating that the protein is almost exclusively associated with the microsomal membrane. It should be pointed out that also other molecules have been shown to prevent the assembly of apoB-100 lipoproteins. Thus a moderate accumulation of monomethylethanolamine in the (endoplasmic reticulum) membrane (37) inhibits the translocation of apoB-100 to the lumen of the endoplasmic reticulum. Our results may in fact support the previous observation that the assembly of apoB-100 VLDL is closely linked to the translation/translocation of apoB-100 (7). Thus, if the addition of the major load of triacylglycerol is uncoupled from the translation/ translocation process, the protein would rather be associated with the membrane. An alternative explanation for these results could be that BFA induced a general inhibition of the translocation of proteins. The observation that other secretory proteins were completed in the presence of BFA argues against such a general inhibition of the translocation. Moreover, apoB-48 is translocated to the microsomal lumen.
Our previous results (7) indicated that a substantial amount of apoB-100 VLDL could be formed after the completion of the nascent polypeptides. These conclusions were based on the observation that, during pulse-chase studies, a substantial accumulation of radioactive apoB-100 was seen in the VLDL fraction after chase periods (7) when the nascent polypeptides could be expected to be completed (28). A post-translational recruitment of apoB-100 for the assembly of lipoproteins is also indicated by the results of other authors (38). To reconcile the present data with these observations, it has to be assumed that the precursors for this post-translational assembly of VLDL are associated with the endoplasmic reticulum membrane. In fact, we have previously suggested (25,26,39) that the membraneassociated apoB-100 is a precursor for the assembly of lipoproteins. We now have the tools to experimentally address this question since the BFA-treated McA-RH7777 cells offer a unique system which allow studies of the fate of the membraneassociated apoB-100. Thus in pulse-chase studies BFA could be used to inhibit the assembly of the apoB-100 lipoproteins during the elongation of the labeled nascent chains, thus preventing them from forming lipoproteins. Instead all pulse-labeled apoB-100 will be associated with the membrane of the microsomes. After removing the BFA this membrane-associated apoB-100 was assembled into VLDL and secreted. These results clearly indicate that the apoB-100 that is associated with the membrane of the microsomes could be used as a precursor for the assembly of VLDL. As pointed out above, it has been indicated (31-33) that the membrane bound protein undergoes post-translational degradation. However, there may be processes in the membrane that sort apoB-100 to lipoprotein assembly or to degradation. Interestingly enough, recent results (40) suggest that there may be a step after the formation of the first 80 kDa of apoB-100 that is of importance for the sorting of apoB-100 between a degradational pathway and an assembly pathway. These results are supported by our previous observation that apoB-100 starts to interact with lipids (forming lipoproteins) after it has reached a size of 80 to 100 kDa (31). Also of interest is the observation (41) that the first 80-kDa domain of apoB-100 appears to be globular, with relatively little lipid-binding capacity, but highly stabilized by disulfides.
The N-terminal 80-kDa fragment has been suggested to be cleaved off and secreted from the cell (40). It could be suggested that this cleavage is a signal to post-translational degradation of the protein (40). A provocative hypothesis for future experimental testing is that apoB-100 retains its capacity to assemble VLDL as long as it keeps its N-terminal 80-kDa globular domain. Indeed recent observations (42) suggest that the N-terminal domain of apoB is of importance for the correct interaction between the protein and the microsomal transfer protein (regarding microsomal transfer protein, see for example, Ref. 43).
The assembly of VLDL could be reactivated by chasing away the BFA. The time needed to achieve such a reactivation of the assembly of apoB-100 VLDL was 30 -60 min. This time period is in the same range as that observed for the reorganization of the secretory pathway (16) and for the restarting of the intra-cellular transfer and secretion of transferrin, as determined in this paper. It most likely reflects the time needed to inactivate the BFA (44).
We were not able to give an estimate of the time needed to restart the assembly of apoB-48 VLDL. The reason for this is the difference in the kinetics for the assembly of apoB-48 VLDL and apoB-100 VLDL. ApoB-100 VLDL is assembled very close to the translation/translocation of the protein, and will therefore be formed and released to the secretory pathway during a short pulse (7). On the contrary there is a delay of 15-30 min before assembled apoB-48 VLDL could be detected in pulselabeled cell (7). The method used to estimate the time needed for the restart of the VLDL assembly gives, for obvious reasons, very approximate values and the delay in the appearance of the pulse-labeled apoB-48 VLDL will further complicate the evaluation of the results.
In summary the results presented in the present paper indicate that BFA selectively blocks the major addition of lipids to apoB-48, thus inhibiting the formation of VLDL, but leaves the translation and translocation of the protein intact. Thus the results support our model for the assembly of apoB-48 VLDL (7). This model states that the assembly process consists of two well separated steps, the translation/translocation of the protein forming apoB-48 HDL and the addition of the major amount of lipids. The dose-response experiments points to a unique mechanism, perhaps involving the interference with a GTP-GDP cycle, behind the effect of BFA on the assembly process, a mechanism that is separate from that by which BFA influences the secretory pathway.
Our previous results have indicated that the assembly of apoB-100 VLDL, at least to a certain degree, occurs in close connection with the translation/translocation of the apolipoprotein (7), even during the translation/translocation of the protein (7,31). We have also obtained results that support a post-translational recruitment of apoB-100 to the assembly of VLDL (7) but we have not been able to establish the form in which this VLDL precursor exists. In this paper we demonstrate that BFA inhibits the assembly of all apoB-100 lipoprotein, leaving the apolipoprotein in association with the microsomal membrane. We were able to demonstrate that this membrane-associated apoB-100 could be used as a precursor for the assembly of VLDL. As in the case of apoB-48, the obtained results may point to a GTP-GDP cycle involved in the addition of lipids to apoB-100 and/or the subsequent translocation of the protein to the lumen of the secretory pathway.