| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Lipid Research, Vol. 44, 22-32, January 2003
Copyright © 2003 by Lipid Research, Inc.
Review |
* Departments of Anatomy and Cell Biology, SUNY Downstate Medical Center, 450 Clarkson Avenue, , Brooklyn, NY 11203
Pediatrics, SUNY Downstate Medical Center, 450 Clarkson Avenue, , Brooklyn, NY 11203
Medicine Biophysics Division, SUNY Downstate Medical Center, 450 Clarkson Avenue, , Brooklyn, NY 11203
Published, JLR Papers in Press, September 16, 2002. DOI 10.1194/jlr.R200014-JLR200
1 To whom correspondence should be addressed. e-mail: mahmood.hussain{at}downstate.edu
 |
ABSTRACT |
|---|
1-ß1-2-ß2-3. We propose that MTP contains three structural motifs (N-terminal ß-barrel, central -helix, and C-terminal lipid cavity) and three functional domains (lipid transfer, membrane associating, and apoB binding). MTP's lipid transfer activity is required for the assembly of lipoproteins. This activity renders nascent apoB secretion-competent and may be involved in the import of triglycerides into the lumen of endoplasmic reticulum. In addition, MTP binds to apoB with high affinity involving ionic interactions. MTP interacts at multiple sites in the N-terminal ß1 structural domain of apoB. A novel antagonist that inhibits apoB-MTP binding decreases apoB secretion. Furthermore, site-directed mutagenesis and deletion analyses that inhibit apoB-MTP binding decrease apoB secretion. Lipids modulate protein-protein interactions between apoB and MTP. Lipids associated with MTP increase apoB-MTP binding whereas lipids associated with apoB decrease this binding. Thus, specific antagonist, site-directed mutagenesis, deletion analyses, and modulation studies support the notion that apoB-MTP binding plays a role in lipoprotein biogenesis. However, specific steps in lipoprotein assembly that require apoB-MTP binding have not been identified. ApoB-MTP binding may be important for the prevention of degradation and lipidation of nascent apoB.
Abbreviations: ER, endoplasmic reticulum; MTP, microsomal tri-glyceride transfer protein; M subunit, 97-kDa subunit of the MTP complex; P subunit, the 55-kDa PDI subunit of the MTP complex; PDI, protein disulfide isomerase
Supplementary key words protein-protein interactions • protein motifs • domains • abetalipoproteinemia • hypobetalipoproteinemia • apolipoprotein B • MTP
|
INTRODUCTION |
|---|
|
Structural and functional domains in apoB |
|---|
45 kb. In the liver, it is transcribed into a single mRNA of 15 kb and is translated into a single polypeptide of 4536 amino acids called apoB-100. In the intestine, the apoB mRNA is post-transcriptionally edited, resulting in the conversion of a glutamine codon into a stop codon. The edited mRNA is translated into a single polypeptide of 2,152 amino acids called apoB48. By comparing mean hydrophobic moments per amino acid residue and the average hydrophobicity of the hydrophobic face of the helices, Segrest et al. have proposed a pentapartite secondary structure for apoB-100 (3, 15–17). According to this model, apoB100 is comprised of three amphiphatic -helical domains alternating with two amphiphatic ß-sheet domains in an NH2-1-ß1-2-ß2-3-COOH configuration.
The 1 domain (B:58–795) is an independent globular domain. It associates with lipids but is incapable of forming a lipoprotein (18, 19). Nonetheless, it is required for lipoprotein assembly because its absence ablates lipoprotein assembly (20). It contains 12 cysteine residues that form six disulfide linkages (21). Proper disulfide bond formation between some cysteine residues is essential for the assembly of apoB-containing lipoproteins (22, 23). Based on sequence homology with lipovitellin, this region has been predicted to consist of a ß-barrel (B:1–263) and an -helical (B:294–592) domains (24) and has subsequently be called ß1 domain (3). As discussed below, this domain contains MTP binding site.
The other four domains (ß1-2-ß2-3) are also comprised of several short amphiphatic ß-strands and -helices. The ß-sheet domains (B:827–2001 and B:2571–4032) are essential for lipoprotein assembly and bind lipids non-reversibly. The assembly of these ß-sheets into lipoproteins requires 1 domain (20). The LDL receptor binding and heparin binding sites are in the ß2 domain. ApoB-48 contains ß1 domain only, whereas apoB-100 contains both the ß1 and ß2 domains. These two proteins are used for the assembly of two different lipoproteins, chylomicrons and VLDLs. The 2 (B:2045–2587) and 3 (B:4017–4515) domains consist of several amphiphatic helices that can reversibly associate with lipids, a characteristic property of exchangeable apolipoproteins. Their role in lipoprotein assembly is unknown.
|
Microsomal triglyceride transfer protein |
|---|
MTP enhances the rate of lipid transfer between vesicles (31, 32). Kinetic studies with model membranes suggest that MTP transfers lipids by a shuttle mechanism (33). In this mechanism, each MTP molecule is proposed to interact transiently with a membrane, extract lipid molecules, dissociate from the membrane, bind transiently with another membrane, deliver lipids rapidly to the second membrane, and become available for another cycle of lipid transfer. The lipid transfer activity was shown to be optimum with neutrally charged membranes and decreased in the presence of negatively charged lipids in vesicles (33). Kinetic studies suggest that MTP has two, one fast and one slow, lipid binding sites (34, 35). The fast site is implicated in lipid transfer (34).
|
Functional domains in the heterodimeric complex |
|---|
-helical domain, and C-terminal lipid-binding cavity (Fig. 1)
. We propose that MTP may contain at least three (lipid transfer, membrane associating, and apoB binding) functionally independent domains (Fig. 1). Kinetic studies indicate for the presence of two, one low and one high affinity, lipid-binding domains in MTP (34). The high affinity domain binds few molecules of neutral lipids and phospholipids and may represent the lipid transfer domain. The lipid transfer activity antagonists probably bind at this site and inhibit lipid transfer activity. Precise information about the lipid transfer domain in MTP is not available but based on the homology with lipovitellin (39), Read et al. (38) have suggested that MTP contains a C-terminal lipid binding cavity. The walls of the lipid binding cavity in MTP are formed by the A and C ß-sheets present in the M subunit (Fig. 1). The -helical domain holds these sheets together. The back of the cavity is probably covered by the P subunit (not shown in the figure). The lipid transfer domain may be involved in the loading and unloading of lipid molecules, a step necessary for their transfer. A nonsense mutation in the A sheet, Asn780Tyr, does not affect its binding to the P subunit but abolishes MTP's lipid transfer activity (40). Thus, the C-terminal 1/3rd of the M subunit and the P subunit may form a lipid transfer domain in MTP.
|
We showed that apoB-binding domain in MTP is different from the lipid transfer domain because the lipid transfer activity inhibitors do not inhibit apoB-MTP binding, and inhibitors that inhibit apoB-MTP binding have no effect on lipid transfer activity of MTP (42). Furthermore, immobilization of MTP results in partial loss of MTP's lipid transfer activity but has no effect on apoB-MTP binding (43, 44). Similarly, apoB-binding and membrane associating domains in MTP appear to be dissimilar. Evidence for the independent membrane associating and apoB binding domains also comes from the modulation of apoB-MTP interactions by lipids (41). If apoB and membrane binding domains were the same, then apoB-MTP binding would have decreased in the presence of lipids. Contrary to this expectation, association of MTP with lipids resulted in increased binding to apoB. Thus, apoB-binding domain in MTP appears to be different from both the lipid transfer and membrane associating domains in MTP.
|
Domains involved in subunit interactions |
|---|
-helical region of the M (M:297–603) subunit. Within this region, M:520–598 showed maximum binding to the P subunit. The binding between the N-terminus of the P subunit and the middle region of the M subunit may constitute a nucleation site for the heterodimerization of the two subunits. Subsequent binding of other regions in the P subunit with the C-terminal region in the M subunit are probably important for the formation of a soluble and biochemically active heterodimeric MTP complex. |
Protein-protein interactions between apoB and MTP |
|---|
We published a solid-liquid inter-phase binding assay in 1997 to study apoB-MTP binding. First, we immobilized different lipoproteins to microtiter plates and incubated them with 125I-labeled heterodimeric MTP. The amounts of MTP bound to LDL and VLDL were significantly higher than those bound to HDL (43). Next, the binding of MTP to different lipoproteins was compared with the binding of PDI. Lipoproteins bound to MTP but not to PDI. These studies indicated that the M subunit plays an important role in lipoprotein binding. Subsequently, we showed that immobilized heterodimeric MTP also interacted with lipoproteins present in solution. Kinetic studies demonstrated that protein-protein interactions between these proteins were of high affinity (Kd 10–30 nM).
In 1999, Shoulders and associates used baculoviral expression system and yeast two-hybrid system to study protein-protein interactions (24, 37). They expressed apoB-17 with M subunit, P subunit, or both M and P subunits in Sf 9 cells, immunoprecipitated apoB, and looked for the co-precipitation of the M and P subunits (24). PDI was not precipitated with apoB. However, the M subunit was immunoprecipitated with apoB-17 when expressed with the P subunit. These studies reinforced the notion that apoB interacts with the M subunit and the P subunit is probably not required for apoB binding.
|
Nature of interactions |
|---|
The importance of ionic interactions was substantiated by chemical modification of functional groups in apoB using group specific reagents (48). Modification of 38–44% of the aspartic and glutamic acid residues in LDL by glycine methyl ester had no effect on apoB-MTP binding. Alteration of all histidine residues by diethyl pyrocarbonate in LDL decreased, but did not abolish, apoB-MTP binding. Treatment with 1,2-cyclohexanedione resulted in the modification of 54% of the arginine residues in LDL and completely abolished its binding to MTP. Similarly, modification of 74% of lysine 
-amino groups by acetoacetylation abolished LDL-MTP binding. More importantly, hydroxylamine treatment of the modified LDL regenerated all modified arginine and lysine residues and completely restored MTP binding. Modification of the -amino groups by reductive methylation and acetylation further substantiated the involvement of positive charges. Reductive methylation does not alter the positive charges, whereas acetylation neutralizes the positive charges. Reductive methylation of 88% of lysine residues did not affect apoB-MTP binding. In contrast, acetylation of LDL completely abolished apoB-MTP binding. These studies showed that positive charges on the functional side groups of arginine and lysine in apoB are critical for apoB-MTP binding. Most likely, they interact with negatively charged residues in MTP.
Positively charged amino acid residues in apoB are known to interact with heparin and LDL receptors (49, 50). To determine whether the MTP binding site overlaps with the heparin binding site, Bakillah et al. studies the inhibition of LDL-MTP binding by heparin, chondroitin sulfate, and suramin (48). Heparin and chondroitin sulfate had no significant inhibitory effect on LDL-MTP binding. However, suramin, a highly charged polysulfated polycyclic hydrocarbon, inhibited apoB-MTP binding. Inhibition by suramin indicates that apoB-MTP binding involves ionic interactions. Lack of inhibition of apoB-MTP binding by heparin indicates that heparin binding and MTP binding sites are independent and different from each other. Thus, lysine and arginine residues crucial for MTP binding are different from those involved in heparin and LDL receptor binding.
|
Domains in apoB that interact with MTP |
|---|
1-globular domain of apoB.
To further delineate the MTP binding site in the ß1-domain, we expressed different apoB sequences as FLAG-apoB chimeras in COS cells (44). Two chimeras, B:1–300 and B:270–570 were secreted to similar extent, but their binding to MTP was significantly different. B:270–570 bound robustly to MTP but the binding of B:1–300 was considerably low (4–25% of B:270–570 in different experiments, unpublished observations). We concluded that B:270–570 contains a high affinity binding site for MTP. Attempts were then made to identify the minimum sequence that could bind to MTP. For this purpose, we subjected B:270–570 to N- and C-terminal truncations at proline residues (Table 1). C-terminal deletion to amino acid 430 resulted in significant loss of MTP binding. These studies indicated that amino acids 430–570 are critical for MTP binding. Progressive N-terminal truncations of amino acids from 270 to 341 resulted in progressive decrease in MTP binding, and truncations to amino acid 430 resulted in significant loss of MTP binding. More importantly, these studies showed that B:430–570 do not bind to MTP. Thus, it appears that amino acids 270–570 are required for optimum binding.
|
1-domain of apoB. Critical amino acids in this region that interact with MTP are not known. |
Domains in MTP that interact with apoB |
|---|
|
Factors modulating apoB-MTP binding |
|---|
Effect of PDI on apoB-MTP binding
In yeast two-hybrid system, Bradbury et al. observed that M:520–598 interact with the P subunit and apoB-17 (37). They expressed M subunit and apoB-17 in the presence and absence of P subunit in insect cells and observed that the P subunit decreased the binding of apoB to the M subunit. These studies indicate that the P subunit may affect apoB binding to M subunit of MTP most likely prior to the formation of heterodimeric MTP complex. It remains to be determined whether the P subunit can modulate apoB binding to the heterodimeric MTP. Most likely, it may not for the following reasons. There is excess of PDI in the ER lumen and yet apoB-MTP complexes have been isolated by immunoprecipitations. The M subunit probably does not exist in the ER unassociated with the P subunit and thus has a limited ability to interact with apoB in the absence of the P subunit.
|
Modulation of apoB-MTP binding by lipids |
|---|
10% of nascent apoB is associated with MTP in HepG2 cells. Inhibition of intracellular apoB degradation by proteosomal inhibitors prolonged the duration of apoB-MTP binding without increasing the amounts of apoB bound to MTP. In contrast, increases in triglyceride synthesis by the supplementation of oleic acid increased the amounts of apoB associated with MTP by 3–5-folds. Thus, these studies provided initial evidence that lipids may modulate apoB-MTP binding. We observed that lipids have a profound effect on apoB-MTP binding (41). Zwitterionic phospholipids, phosphotidylcholine, and sphingomyelin increased apoB-MTP binding and negatively charged phospholipids, phosphatidylinositol, and phosphatidylserine and decreased apoB-MTP binding (41). The negatively charged lipids, most likely, competed with ionic interactions between these proteins. However, the effect of zwitterionic lipids was intriguing. Kinetic analysis revealed that phosphatidylcholine vesicles increased affinity between proteins without affecting the number of binding sites. To understand the mechanisms for increased affinity between proteins, we pre-incubated apoB and MTP with lipids. Pre-incubation of apoB with lipids had no effect on MTP binding. In contrast, pre-incubation of MTP with phosphatidylcholine vesicles increased apoB binding by 4-fold (41). During pre-incubation, MTP stably associated with phospholipid vesicles and the MTP-lipid vesicles bound better to apoB. These studies clearly indicate that the association of MTP with phospholipid vesicles results in increased affinity for apoB. It is likely that apoB first interacts with MTP via ionic interactions. Next, lipids bound to MTP may interact either with lipid-binding domains in apoB or with lipids present in apoB-containing lipoproteins. Thus, protein-protein interactions may bring lipids into close proximity leading to additional hydrophobic interactions between lipids. Combination of ionic and hydrophobic interactions may result in increased affinity.
Lipids associated with apoB decrease apoB-MTP binding
Pre-incubation of apoB with phospholipids had no effect on apoB-MTP binding (41). However, apoB-associated lipids decrease MTP binding. Partial de-lipidation of VLDL with Tween-20 increased the binding of VLDL-apoB to MTP (43). Similarly, partial de-lipidation of LDL with taurocholate increased its binding to MTP. These studies suggest that the amounts of lipids associated with apoB have a negative effect on apoB-MTP binding. Hence, increased apoB lipidation may decrease apoB-MTP binding.
|
MTP and apoB-lipoprotein assembly |
|---|
-hydroxylase has been shown to be necessary for apoB secretion (54). Third, the importance of the lipid transfer activity in the lipidation of apoB polypeptides was reinforced by using specific inhibitors (55–60). MTP inhibitors that inhibited lipid transfer activity in vitro decreased apoB secretion in vivo. Thus, MTP's lipid transfer activity is essential for the assembly and secretion of apoB-containing lipoproteins. However, it is still not clear how this lipid transfer activity of MTP results in a net transfer of lipids to nascent apoB. So far, no in vitro assay has been reported that demonstrates net transfer of lipids to apoB-lipoproteins resulting in the assembly of larger lipoproteins. Substantial evidence exists to suggest that MTP's lipid transfer activity is essential for two processes in lipoprotein assembly. First, it is required for rendering apoB-secretion competent. Rusinol et al. have shown that nascent apoB can be assembled into lipoproteins in a cell-free system in the absence of MTP (61). The secretion competency is achieved by the addition of lipids to nascent apoB and inhibition of intracellular degradation. The lipidation process inhibits the degradation of nascent apoB and promotes lipoprotein assembly. The dependency of apoB on MTP with respect to its lipidation and secretion appears to depend on the length of apoB. Specific inhibitor and gene ablation studies have shown that the larger apoB peptides are more dependent on MTP compared to smaller apoB peptides (62–64). In fact, there are reports indicating that apoB-48 may be secreted in the absence of MTP (64). Nicodeme et al. (62) have suggested that a region between apoB-51 and apoB-53 has a high requirement for lipids and for the lipid transfer activity of MTP.
Evidence is accumulating to suggest that MTP's lipid transfer activity may be required for the accumulation of triglycerides in the lumen of the ER. It has been shown that the absence of MTP or inhibition of MTP leads to decreased triglycerides in the lumen of the ER (64, 65). It has been suggested that MTP may in involved in the formation and stabilization of lumenal lipid droplets (41). The accumulation of the triglycerides in the ER is probably required for the "core expansion" of the nascent primordial particle (see below).
Importance of apoB-MTP binding in lipoprotein assembly
We observed that MTP's lipid transfer activity antagonists had no effect on apoB-MTP binding (42). These studies encouraged us to look for compounds that might inhibit protein-protein interactions without affecting lipid transfer activity of MTP. Screening of several compounds from the Atherogenics Inc. resulted in the identification of a compound, AGI-S17, that inhibited apoB-MTP binding by 70% at 40 µM and had no effect on MTP's lipid transfer activity (42). Next, we studied the effect of AGI-S17 in HepG2 cells. Incubation of HepG2 cells with AGI-S17 had no effect on the lipid transfer activity of MTP but decreased its binding to apoB. These studies indicated that AGI-S17 inhibited intracellular apoB-MTP binding without affecting lipid-transfer activity. In addition, AGI-S17 inhibited the secretion of total and nascent apoB by 70%. Most likely, inhibition of apoB secretion by AGI-S17 is a consequence of the inhibition of intracellular apoB-MTP binding. These studies provide strong evidence that protein-protein interactions between apoB and MTP may be important for lipoprotein assembly and secretion.
Evidence for the importance of apoB-MTP binding in lipoprotein assembly also comes from the studies of Bradbury et al. (37). They showed that mutations in B:521–721 decrease MTP binding and apoB secretion. Moreover, Liang and Ginsberg (51) showed that deletion of B:1–210 that contains the first MTP binding site decreased the secretion of apoB-34 but had no effect on apoB-16 secretion. As discussed before, modulation of apoB-MTP binding by lipids also indicates that it may be a physiologically significant event. It should be pointed out that apoB has been shown to bind several other proteins (66), but modulation for the binding of these proteins to apoB has not yet been demonstrated. Thus, specific antagonists, site-directed mutagenesis, deletion analysis, and modulation studies lend strong support to the notion that apoB-MTP binding plays an important role in lipoprotein biogenesis.
We propose that apoB-MTP binding may play different roles during lipoprotein biosynthesis (Fig. 2) . For clarity, apoB-MTP binding is shown to play an important role in two steps, A) prevention of apoB degradation and B) lipidation of nascent apoB. We have suggested that the emerging nascent apoB polypeptide interacts with MTP prior to its lipidation due to the localization of its binding site in the N-terminal 17% of the molecule (42, 43). It is known that apoB undergoes dislocation from the ER and degradation by proteosomes if not assembled into lipoproteins (67, 68). It is proposed that MTP binding may facilitate the import of nascent apoB into the ER lumen, prevent dislocation from the ER, and inhibit proteosomal degradation. These processes can be facilitated further by the concomitant lipidation of the emerging apoB peptide by MTP (Fig. 2A, i). If inhibitors or mutations inhibit apoB-MTP binding, apoB is predicted to undergo degradation involving hsp70 (69) and hsp90 binding (70), ubiquitination, and proteosomal degradation (Fig. 2A, ii).
|
Very little is known about the lipid droplet formation. This process may require efficient lipid synthesis. In addition, it may require MTP because mice deficient in MTP expression do not accumulate lipid droplets in the ER lumen (64). Wang et al. suggested that MTP activity might be necessary for the accumulation of triglycerides into the ER lumen (65). We have shown that MTP exists associated with lipids in the ER lumen (41). Thus, MTP may play a role in the formation and stabilization of lipid droplets. Obviously, this property of the MTP does not require apoB binding.
The role of MTP in the fusion of lipids droplets with primordial lipoproteins to form nascent lipoproteins in a process called "core expansion" is not known. The apoB-MTP binding may bring lipids associated with apoB and MTP into close proximity and facilitate their fusion. It is also possible that MTP may initiate the fusion of lipids. MTP has been proposed to contain sequences that may have fusogenic properties (38). Thus, apoB-MTP binding may also play an important role during core expansion of primordial lipoproteins.
|
Concluding remarks |
|---|
It remains to be determined whether inhibition of apoB-MTP will increase intracellular triglyceride accumulation. This is important because MTP's lipid transfer activity inhibitors have not yet proven to be useful in controlling plasma lipid levels because they significantly decrease lipoprotein secretion and cause increased accumulation of lipids in the liver. The structural information may be useful in obtaining compounds that partially inhibit lipoprotein assembly and secretion. Sub-optimal inhibition of lipoprotein assembly may provide a more desirable phenotype of lowering plasma lipid levels coupled with low lipid accumulation in the liver.
ApoB has been shown to interact with several other chaperones in the ER (66). At present, no information is available about the modulation of the binding of different chaperones to apoB. Particularly, it is not known how the binding of these chaperones to apoB affects MTP binding and vice versa. Modulation of the binding of various chaperones to apoB may play an important role in the maturation of apoB into secretion-competent lipoprotein, and their understanding will provide new insights into lipoprotein biosynthesis. A proteomic approach has resulted in the identification of several apoB-binding proteins (71). It remains to be determined whether these proteins play a role in lipoprotein assembly.
The role of MTP in the import of triglycerides into the lumen of ER also needs further investigation. The acceptors involved in the import need to be identified. It remains to be determined how these acceptors compete with apoB and how MTP discriminates between different acceptors and avoids futile cycle of triglyceride transfer between membranes.
|
ACKNOWLEDGMENTS |
|---|
Manuscript received September 9, 2002 and in revised form September 12, 2002.
|
REFERENCES |
|---|
TOP
ABSTRACT
INTRODUCTION
/
angles, to deduce the structure of the M subunit. Various sheets and helices have been either truncated or deleted to accommodate the absence of homologous sequences in the M subunit. Side chains are replaced using dictionary definitions, keeping the main chain intact. Sequence gaps and insertions assigned to loops and ß-turns are rebuilt with the fewest residues needed based on loop searches of known PDB structures. Using the AMBER force field (
