Demonstration of a Physical Interaction between Microsomal Triglyceride Transfer Protein and Apolipoprotein B during the Assembly of ApoB-containing Lipoproteins*

Microsomal triglyceride (TG) transfer protein (MTP) is an endoplasmic reticulum lumenal protein consisting of a 97-kDa subunit and protein disulfide isomerase. It is believed that MTP delivers TG to nascent apoB molecules during the assembly of lipoprotein particles in the secretory pathway. Although in vitro studies have estab-lished the mechanism of TG transfer between donor and acceptor membranes, the mechanism of action of MTP in vivo remains unknown. The present studies were un- dertaken to examine whether or not the transfer of TG to nascent apoB in the endoplasmic reticulum involves the physical interaction between MTP and apoB. HepG2 cells were labeled with [ 3 H]leucine, lysed in a nondenaturing homogenizing buffer, and immunoprecipitated with anti-MTP antiserum. We found that labeled apoB and protein disulfide isomerase were co-immunopre- cipitated by this procedure. In addition, we were able to detect the 97-kDa subunit of MTP in these immunopre- cipitates by immunoblot. The association of MTP and apoB, as assessed in pulse-labeled cells by co-immuno- precipitation, was transient; apoB was prominent on fluorgraphy at 10 min of chase but minimal thereafter. Oleic acid treatment, which protects apoB from rapid intracellular degradation by increasing TG availability, increased both the degree and the duration of association between MTP and apoB dramatically. Inhibition of TG synthesis by Triacsin D, on


Microsomal triglyceride (TG) transfer protein (MTP)
is an endoplasmic reticulum lumenal protein consisting of a 97-kDa subunit and protein disulfide isomerase. It is believed that MTP delivers TG to nascent apoB molecules during the assembly of lipoprotein particles in the secretory pathway. Although in vitro studies have established the mechanism of TG transfer between donor and acceptor membranes, the mechanism of action of MTP in vivo remains unknown. The present studies were undertaken to examine whether or not the transfer of TG to nascent apoB in the endoplasmic reticulum involves the physical interaction between MTP and apoB. HepG2 cells were labeled with [ 3 H]leucine, lysed in a nondenaturing homogenizing buffer, and immunoprecipitated with anti-MTP antiserum. We found that labeled apoB and protein disulfide isomerase were co-immunoprecipitated by this procedure. In addition, we were able to detect the 97-kDa subunit of MTP in these immunoprecipitates by immunoblot. The association of MTP and apoB, as assessed in pulse-labeled cells by co-immunoprecipitation, was transient; apoB was prominent on fluorgraphy at 10 min of chase but minimal thereafter. Oleic acid treatment, which protects apoB from rapid intracellular degradation by increasing TG availability, increased both the degree and the duration of association between MTP and apoB dramatically. Inhibition of TG synthesis by Triacsin D, on the other hand, significantly decreased the MTP-apoB binding. N-Acetylleucyl-leucyl-norleucinal, a cysteine protease inhibitor, which directly protects apoB from rapid intracellular degradation but does not affect TG synthesis, increased the interaction between MTP and apoB only slightly, although it did prolong it. Our results suggest that direct interaction between MTP and apoB occurs during the assembly of apoB-containing lipoproteins in HepG2 cells.
The increased secretion rate of apoB-containing lipoprotein particles from liver results in elevated plasma levels of low density lipoproteins (1-4), a major risk factor for the development of atherosclerotic diseases. Studies have demonstrated that apoB secretion is regulated posttranslationally (5)(6)(7)(8)(9)(10)(11)(12). Among factors believed to affect the secretion of apoB-containing lipoproteins from HepG2 cells, the availability of newly synthesized triglyceride (TG) 1 may be the most important (13)(14)(15)(16)(17). Thus, when lipid availability is increased, rapid intracellular degradation of newly synthesized apoB is inhibited; inhibition of lipid synthesis, on the other hand, results in enhanced degradation of apoB.
Despite the large body of evidence demonstrating the importance of hepatic lipids in the assembly and secretion of apoBcontaining lipoproteins, TG synthesis appears to be normal, and lipid droplets accumulate, in livers of patients with abetalipoproteinemia (18), a condition characterized by the absence of apoB secretion. Recently, defective assembly and secretion of apoB-containing lipoproteins in affected patients was found to be associated with mutations in the gene encoding a 97-kDa protein, microsomal TG transfer protein (MTP) large subunit (19,20). In vitro studies (21,22) showed that MTP efficiently catalyzes the transfer of TG and other lipids from donor membranes to acceptor membranes. MTP large subunit forms a heterodimer with protein disulfide isomerase in the lumen of the endoplasmic reticulum (ER) in hepatocytes and enterocytes. MTP large subunit appears to be expressed normally only in hepatocytes and enterocytes; protein disulfide isomerase is ubiquitously expressed. Two recent studies (23,24) convincingly demonstrated that coordinate expression of large apoB truncations and MTP large subunit in cells that normally do not express either of the two proteins resulted in the efficient secretion of apoB-containing lipoproteins. ApoB was not secreted from these cells before MTP large subunit was expressed.
Although the above studies have clearly indicated a role for MTP in the assembly and secretion of apoB-containing lipoproteins from hepatocytes, the in vivo mechanism underlying this activity remains unknown. An unanswered question is how MTP transfers TG molecules to apoB during the assembly of lipoprotein particles. More specifically, does a physical interaction between MTP and apoB play a role in this process? The present studies were conducted to answer this question.

EXPERIMENTAL PROCEDURES
Materials-L- [4, H]Leucine (135 Ci/mmol, catalog number TRK683) and [2-3 H]glycerol (1.0 Ci/mmol, catalog number TRA.118) were purchased from Amersham Corp. Monospecific anti-human apoB antiserum was raised in rabbits. Anti-bovine MTP antiserum (raised in a rabbit) and anti-bovine MTP large subunit antibody (raised in a goat) were generated by one of the authors (J. R. W). These antibodies were characterized and utilized in previous studies (19,23). Protein A-Sepharose CL 4B was from Pharmacia LKB Biotechnology Inc. (Uppsala, Sweden). Bovine serum albumin (BSA) and oleic acid (sodium salt) were * This work was supported by National Institutes of Health Grants HL36000 and HL21006. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Growth of HepG2 Cells-HepG2 cells, obtained from ATCC, were cultured in 35-or 100-mm dishes in minimum essential medium containing 0.1 mM nonessential amino acids, 1.0 mM sodium pyruvate, 100 units/ml of penicillin, 100 g/ml of streptomycin, and 10% fetal bovine serum. The growth medium was changed to experimental medium when cells were about 90% confluent.
Labeling of HepG2 Cells-HepG2 cells were incubated with serumfree, leucine-free minimum essential medium containing 1.5% of BSA, 200 Ci/ml [ 3 H]leucine, and various additions as described in the figure legends.
Immunoprecipitation-After labeling, the cells were washed with cold phosphate-buffered saline and subsequently lysed in a nondenaturing lysis buffer (62.5 mM sucrose, 0.5% sodium deoxycholate, 0.5% Triton X-100, 50 g/ml pepstatin A, 50 g/ml leupeptin, 150 g/ml phenylmethylsulfonyl fluoride, 5 mM EDTA, 50 mM Tris-HCl, pH 7.4, and 150 mM NaCl). Immunoprecipitation of proteins was carried out according to the method of Dixon et al. (14). Briefly, medium or cell lysate samples were mixed with NET buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 5 mM EDTA, 0.5% Triton X-100, 0.1% SDS) and excess amounts of various antiserum. The mixture was incubated at 4°C on a shaker for 5 h. Protein A-Sepharose CL 4B was added to the mixture, and the incubation was continued for an additional 2 h. The beads were extensively washed with NET. Proteins were extracted from the beads with sample buffer by boiling for 4 min. An aliquot of the sample was run on SDS-PAGE followed by autography or immunoblot.
Immunoblot-Cell lysates or samples from the immunoprecipitation were run on SDS-PAGE, transferred to a nitrocellulose membrane, and blocked with 3% BSA in phosphate-buffered saline. Either anti-human apoB antiserum or anti-bovine MTP large subunit antiserum was incubated with the membrane for 2 h. After washing with phosphatebuffered saline, secondary antibody conjugated with horseradish peroxidase was then incubated with the membrane for 2 h. 4-Chloro-1naphthol and H 2 O2 were used as substrates to develop the membrane.

MTP Is Associated with ApoB in HepG2
Cells-Although the importance of MTP in the assembly of apoB-containing lipoproteins has been convincingly indicated (19,20,23,24), it is not clear how MTP acts. MTP was shown in vitro to transfer TG molecules from donor membranes to acceptor membranes without the participation of apoB (21,22). This raises the possibility that MTP facilitates movement of TG from the cytosolic to the lumenal side of the ER membranes or from the cytosolic side of the ER membrane to the surface of an apoB-containing lipoprotein in the ER lumen. However, experiments in cells lacking MTP indicate that apoB translocation across the ER membrane requires that protein; the importance of MTP at such an early step in apoB transport would suggest a physical interaction between the two proteins. Therefore, we first examined this possibility.
HepG2 cells were labeled for 4 h with [ 3 H]leucine and lysed in a nondenaturing buffer. Immunoprecipitation was carried out with anti-human apoB, anti-MTP, or anti-MTP large sub-unit antisera under nondenaturing conditions. With anti-apoB antiserum, apoB100 was precipitated from cell lysates (Fig. 1). ApoB, however, was also precipitated by either anti-MTP or anti-MTP large subunit antibodies (Fig. 1). The latter results suggested interaction between apoB and MTP. Anti-MTP antibody appeared to precipitate a small proportion of the labeled apoB pool. This is consistent with the finding that only 5-10% of newly synthesized apoB is secreted from HepG2 cells under basal conditions (14 -17).
Since other proteins (with the exception of protein disulfide isomerase; see below) in the cell lysate were not precipitated by the anti-MTP antibodies, it appeared that the MTP-apoB interaction was specific. We were concerned at first that neither the apoB antiserum nor the two anti-MTP antibodies precipitated MTP large subunit (Fig. 1). This is not surprising, however, when one considers that the MTP large subunit has a half-life reported to be more than 100 h (26). Thus, our protocol would radiolabel a very small proportion of the cellular MTP large subunit pool. On the other hand, the anti-MTP antibody did immunoprecipitate the small subunit of MTP (protein disulfide isomerase) together with apoB (Fig. 1).
To confirm the physical interaction between apoB and MTP, immunoblotting experiments were carried out. HepG2 cells were lysed with the nondenaturing buffer, and immunoprecipitation was carried out with either nonimmune serum or MTP large subunit antiserum. The immunoprecipitates and an aliquot of HepG2 cell whole lysate were run on SDS-PAGE and transferred to a nitrocellulose membrane. The membrane was blotted with anti-apoB antibodies. As shown in Fig. 2A, apoB was not detected with nonimmune serum; apoB was detected, however, with anti-MTP large subunit antiserum. In another experiment, MTP large subunit was detected in an anti-MTP large subunit immunoprecipitate by immunoblotting with anti-MTP large subunit antiserum (Fig. 2B).
HepG2 cell lysates were also immunoprecipitated with nonimmune serum or anti-human apoB antiserum. The immunoprecipitate was run on SDS-PAGE and transferred to a nitrocellulose membrane. The membrane was blotted with anti-MTP large subunit antiserum. MTP large subunit (97 kDa) was detected in the anti-apoB immunoprecipitate but not in the nonimmune precipitate (Fig. 2C).
The MTP-ApoB Interaction Is Coordinated with the Synthe-sis of TG-We previously reported that proteolysis and lipidfacilitated apoB translocation are two competitive processes that determine the rate of apoB secretion (15). To further confirm the association between MTP-apoB during the assembly of apoB-containing lipoproteins, we examined the MTP-apoB interaction under various conditions that would alter apoB secretion by affecting either proteolytic activity or lipid availability. HepG2 cells were preincubated with one of the following agents for 1 h: 1) 1.5% BSA alone, 2) 1. cipitated. In cells treated with OA, MTP-apoB binding was increased 3-5-fold (Fig. 3B). This effect was eliminated when TGI was added to the media together with OA. ALLN, although protecting intracellular apoB from rapid degradation (15), only minimally affected the degree of binding of apoB to MTP. In a parallel experiment, OA treatment increased the synthesis of TG 5-fold (Fig. 3C); TGI significantly blocked the stimulation of TG synthesis by OA. In contrast, ALLN did not change TG synthesis. None of the agents used affected the amount of MTP large subunit in HepG2 cells as determined by immunoblotting (Fig. 3D).
The results of these two experiments indicated that the MTP-apoB interaction was closely coupled with, or dependent upon, TG synthesis. These results are consistent with a pathway in which MTP is abundant but only binds to apoB in the presence of newly synthesized triglyceride. Additionally, our findings support the idea that a MTP-apoB physical interaction may be important in both the TG transfer process and lipoprotein assembly. On the other hand, the accumulation of apoB, in the absence of increased TG availability (as would occur in the presence of ALLN), does not appear to significantly modulate the MTP-apoB interaction.
MTP-ApoB Interaction Occurs at an Early Stage in the Assembly of ApoB-containing Lipoproteins-The assembly of apoB-containing lipoproteins in HepG2 cells begins cotranslationally in the ER compartment (13). Data from rat hepatocytes suggest, however, that the bulk of the core lipid is added after translation and translocation have been completed (27)(28)(29). This appears to be particularly true for apoB48-containing particles (28,29), where a two-step process of lipoprotein assembly seems to predominate. The two-step assembly model was essentially drawn from studies conducted in rat hepatocytes. The assembly of apoB-containing lipoprotein particles in HepG2 cells is probably distinct in that the second step observed in rat hepatocytes may not exist in HepG2 cells. Thus, even in the presence of OA, HepG2 cells secrete predominantly intermediate density lipoproteins and low density lipoproteins (30). These data suggest that in HepG2 cells, MTP might play a role in the early stages of lipoprotein assembly. To examine this possibility we carried out studies in which HepG2 cells were pulse-labeled for 10 min with [ 3 H]leucine and chased up to 180 min. At each time point, total apoB and MTP-bound apoB were determined (Fig. 4). As we have demonstrated previously (14), apoB was rapidly degraded during the chase. Thus, over 50% of initially labeled apoB was degraded within the first 20 min of chase. No secretion had occurred by that time (data not shown). After chase for 60 min, less than 20% of initially labeled apoB was found in the cells. At the same time, only about 5% of the labeled apoB was in the medium (data not shown). On the other hand, the maximum MTP-apoB binding was detected at 10 min of chase; the binding was rapidly decreased by 20 min of chase and minimal thereafter (Fig. 4). This result indicated that, in untreated HepG2 cells, the MTP-apoB interaction occurred transiently at an early stage of apoBlipoprotein assembly. As noted earlier, only a very small proportion of newly synthesized apoB interacts with MTP, a finding consistent with the very low level of apoB secretion from HepG2 cells under basal conditions. MTP-ApoB Interaction Parallels the Extent of ApoB Translocation across the ER Membrane-When the cells were treated with OA, the MTP-apoB interaction was increased significantly at 10 min of chase (Fig. 5A). The high level of interaction remained relatively constant during the next 10 min of chase, OA treatment markedly increased and also prolonged the association of MTP with apoB. ALLN treatment only increased the association slightly but prolonged it significantly. Similar results were obtained in three experiments (panels B and C). OA was added to some of the ALLN-treated cells after they had been chased for 30 min, and the cells were chased for an additional 10 -30 min in serum-free medium. At each time point, cells were lysed and immunoprecipitated with either anti-apoB or anti-MTP large subunit antiserum (panel B). The highest band represents the aggregated material on the top of the gel. The addition of OA after a 30-min chase, which would increase translocation of nascent apoB (15), sharply increased MTP-apoB binding (compare ALLN at the 60-min time point with ALLN at 30 min ϩ OA at the 20-or 30-min time point). Overall, the results suggest that increased translocation of apoB into the ER lumen (OA treatment) is associated with greater interaction of apoB with MTP. On the other hand, simply increasing apoB content without effective translocation (ALLN treatment) mainly prolongs the low, basal level of interaction. after which it decreased rapidly. This is compatible with the ability of OA to protect apoB from intracellular degradation by facilitating its translocation across the ER membranes and targeting it for secretion (15,31). By contrast, when the cells were treated with ALLN, the MTP-apoB interaction was only minimally increased compared with control cells at 10 min of chase (Fig. 5A). However, ALLN treatment markedly prolonged the association of MTP with apoB, a result compatible with the direct inhibition of apoB degradation by ALLN, which appears to allow more nascent apoB to slowly translocate across the ER membranes (15). Similar results were obtained in a second experiment (Fig. 5B). In addition, when OA was added to some of the ALLN-treated cells after they had been chased for 30 min, the MTP-apoB interaction was significantly increased, reaching the level observed at the 10 min point in the ALLN-treated cells (Fig. 5B). In Fig. 5, panels A and B are representative of three experiments that are summarized in panel C. This result mirrors our previously reported experiment where OA added to ALLN-treated cells late in the chase period resulted in a sharp increase in apoB secretion (15). Overall, the results of these experiments indicate that the MTP-apoB interaction occurs very early in the lifetime of apoB, probably while it still possesses a transmembrane topology. Indeed, our new results, together with prior results from this laboratory and from others, provide strong support for a scheme in which MTP participates in the completion of apoB translocation across the ER membrane. Studies relevant to the later involvement of MTP in core-lipid addition to primordial lipoprotein particles will require further studies, probably in rat hepatocytes.
In summary, the present studies demonstrate that the physical interaction between MTP and nascent apoB participates in the assembly of apoB-containing lipoproteins. This interaction is very closely coordinated with TG synthesis and appears to be linked to completion of apoB translocation and targeting for secretion. The molecular characteristics of this interaction remain to be determined.