Sterol Regulatory Element-binding Protein Negatively Regulates Microsomal Triglyceride Transfer Protein Gene Transcription*

We herein report that mRNA expression of microsomal triglyceride transfer protein (MTP) and its protein synthesis decline in response to sterol depletion in HepG2 cells, and we functionally characterized the MTP gene promoter in an effort to investigate the molecular mechanisms by which MTP gene transcription is regulated. Luciferase assays using truncated versions of the reporter gene revealed that the region at −124 to +33 base pairs of the human promoter contains the elements required for the suppression of transcription by sterol depletion. Enforced expression of an active form of sterol regulatory element-binding protein (SREBP)-1 (amino acids 1–487) or -2 (amino acids 1–481), both of which are activated under sterol-depleted conditions, is able to mimic sterol-mediated down-regulation. Either further truncation of the promoter region or mutation of the putative SREBP-binding sequence (5′-GCAGCCCAC-3′, −124 to −116 base pairs) abolishes the sterol- and SREBP-dependent transcriptional regulation. Gel mobility shift assay showed that recombinant SREBP-2-(1–481) is able to bind the sequence. Enforced expression of a truncated form of SREBP-2 (amino acids 31–481), which acts as an inhibitor of transcription of the low density lipoprotein receptor gene because it lacks the transcriptional activation domain, also diminishes the luciferase activity, suggesting that direct binding to the promoter region might be sufficient and that the mechanism by which SREBPs inhibit MTP gene expression is distinct from that for the transcriptional stimulation of sterol-regulated genes. Although the SREBP-binding site overlaps a negative insulin-responsive element, insulin negatively regulates MTP gene expression even when the amount of the active form of SREBPs is quite low under the sterol-loaded conditions, indicating that SREBPs only slightly mediate, if at all, the insulin effects. Overall, we conclude that SREBPs are responsible for regulation of lipoprotein secretion via their control of MTP gene expression. Moreover, our results describe for the first time a novel mechanism by which SREBPs negatively regulate expression of the gene encoding the protein involved in lipid metabolism.

Microsomal triglyceride transfer protein (MTP) 1 plays a critical role in the assembly and secretion of very low density lipoproteins in the liver and chylomicrons in the intestine. MTP exists in the lumen of the endoplasmic reticulum as a heterodimer with protein-disulfide isomerase and is involved in the transfer of triglycerides, cholesterol esters, and phospholipids to newly synthesized apoB (1,2). In human patients with abetalipoproteinemia, the absence of functional MTP results in a defect in the assembly and secretion of plasma lipoproteins containing apoB (3,4). In the absence of either MTP lipid transfer activity or sufficient lipid, apoB translocation and lipoprotein assembly are blocked, and apoB is rapidly degraded by a ubiquitin-dependent proteasome process. Under physiological conditions, only a portion of de novo synthesized apoB is secreted; the remaining portion is degraded (5)(6)(7). These findings raise the possibility that changes in MTP activities under various physiological conditions may modulate lipoprotein production and secretion in the liver and intestine.
Recent studies have demonstrated that a high-fat diet fed to hamsters causes an increase in the hepatic MTP mRNA levels (8,9) and that insulin or high concentrations of glucose decrease MTP mRNA levels in HepG2 cells (10). Insulin treatment has also been reported to decrease very low density lipoprotein secretion from hepatocytes (11). On the other hand, oleate stimulates apoB-containing lipoprotein secretion by preventing the intracellular degradation of apoB (12,13), not by altering MTP mRNA levels (10). It has recently been shown that oxysterols regulate the production of lipoproteins by modulating the ubiquitin conjugation of apoB and its subsequent degradation by the proteasome (14). However, little is known about the effect of sterols on MTP mRNA and protein levels.
It has been reported that sterols affect the transcription of a number of genes encoding enzymes and proteins involved in cholesterol and fatty acid metabolism through the actions of sterol regulatory element-binding proteins (SREBPs) (15). SREBPs activate the transcription of genes encoding enzymes involved in cholesterol synthesis, including 3-hydroxy-3-methylglutaryl (HMG)-CoA synthase, HMG-CoA reductase, farnesyl-diphosphate synthase, and squalene synthase, and the low density lipoprotein (LDL) receptor. SREBPs also stimulate the transcription of genes encoding fatty acid synthesis enzymes (acetyl-CoA carboxylase, fatty-acid synthase, glycerol-3-phosphate acyltransferase, and stearoyl-CoA desaturase).
Unlike other members of the basic helix-loop-helix leucine zipper family, SREBPs are synthesized as precursors bound to the endoplasmic reticulum membrane and nuclear envelope (16,17). The transcriptionally active NH 2 -terminal portion including the basic helix-loop-helix leucine zipper domain is released from the membrane by two-step proteolysis (18). In sterol-depleted cells, SREBPs are cleaved at site 1 in the endoplasmic reticulum luminal loop by site-1 protease, which allows site-2 protease to cleave at site 2 within the first membrane-spanning region. The activity of site-1 protease is subject to negative feedback regulation by cholesterol. When cells are overloaded with sterols, the proteolytic process is blocked; the NH 2 -terminal domains are not released; and transcription of the target genes declines.
In this report, we examine whether MTP mRNA and protein levels are affected by sterols and SREBPs, testing our hypothesis that SREBPs might control lipoprotein production and secretion by regulating MTP gene expression. We also characterize the human MTP promoter and define the region responsible for the sterol-dependent regulation. Furthermore, we investigate how SREBPs regulate MTP gene expression, compared with insulin-negative effects.

Construction of Reporter Genes for Luciferase Assay-
The luciferase reporter plasmids were constructed by cloning the BglII-HindIII polymerase chain reaction fragments coding the 5Ј-untranslated region of the human MTP gene into the same restriction sites of a pGL3 basic vector (Promega). To generate pMTPϪ204, pMTPϪ124, pMTPϪ109, and pMTPϪ100, polymerase chain reaction primers were designed to hybridize at the corresponding position (19) and were coupled with the common downstream primer from nucleotide ϩ33 (Fig. 1). To disrupt the SRE, the megaprimers were amplified with the upstream primer with six mutations (AGCCCA 3 GAATTC at the SRE, GCAGCCCAC) and the common downstream primer and further extended in the second polymerase chain reaction as described previously (20). The LDL receptor luciferase plasmid, designated pLDLR, contains the 1.5-kilobase human LDL receptor promoter (21) in a pGL3 basic vector. Expression plasmids, e.g. pSREBP1-(1-487), were constructed as described previously (21), and pSREBP2-(31-481) was constructed by replacing the XhoI-BstEII fragment (600 bp) of pSREBP2-(1-481) with the XhoI-BstEII polymerase chain reaction fragment (510 bp) amplified with the upstream primer (amino acids 31-36) utilizing methionine at the position corresponding to amino acid 31 as an initiator codon.
Tissue Cultures and Cell Transfection-Monolayers of human HepG2 cells were set up on day 0 (2.5 ϫ 10 5 cells/35-mm dish) in medium A (Dulbecco's modified Eagle's medium, 100 units/ml penicillin, 100 g/ml streptomycin, and 1 g/ml Fungizone) supplemented with 7% fetal calf serum (FCS). On day 1, the cells were transfected by the calcium phosphate method with 2 g of one of the reporter luciferase plasmids and 1 g of the pSV-␤-galactosidase control vector (Promega). After 4 h, the medium was removed, and the cells were then washed with phosphate-buffered saline (PBS) and refed with medium A containing 5% LPDS supplemented either with 1 g/ml 25-hydroxycholesterol plus 10 g/ml cholesterol or with a 50 M concentration of a HMG-CoA reductase inhibitor (pravastatin) plus 50 M sodium mevalonate. In the cotransfection experiments, 10 ng of an expression plasmid for SREBP-1 or -2 was added. After 4 h, the cells were refed with medium A containing 7% FCS. After 48 h of culture, the cells were processed, and the luciferase and ␤-galactosidase activities were measured as described previously (21). The ratio of luciferase activity in relative light units was divided by the ␤-galactosidase activity to give a normalized luciferase value.
Northern Blot Analysis-HepG2 cells were set up on day 0 (1.8 ϫ 10 6 cells/100-mm dish) in medium A supplemented with 7% FCS. On day 1, the medium was removed, and the cells were then washed with PBS and refed with medium A containing 5% LPDS supplemented with either sterols or the inhibitor, as described above. After 48 h of culture, total RNA was extracted and fractionated on formaldehyde-agarose gels and then transferred to nylon membranes (Roche Molecular Biochemicals). Riboprobes were prepared using human MTP cDNA (890 -1560 bp downstream of the initiator codon) and human glyceraldehyde-3-phosphate dehydrogenase cDNA (963-1245 bp) with a DIG RNA labeling kit (Roche Molecular Biochemicals). Hybridization signals were quantified with AttoPhos (Amersham Pharmacia Biotech) using a FluorImager 595 (Molecular Dynamics, Inc.).
Western Blot Analysis-HepG2 cells were set up on day 0 (5 ϫ 10 5 cells/100-mm dish) in medium A supplemented with 7% FCS. On day 1, the medium was removed, and the cells were then washed with PBS and refed with medium A containing 5% LPDS with or without 5 g/ml cholesterol. After 14 days of culture, the cells were refed with medium A containing 5% LPDS supplemented with either sterols or the inhibitor, as described above. After 2 days of culture, the cells were harvested, and Western blot analysis was carried out using a polyclonal antibody against human MTP with the Vistra fluorescence Western blotting kit (Amersham Pharmacia Biotech). Signals for 97-kDa MTP were quantified using a FluorImager 595.
Gel Mobility Shift Assay-A double-stranded DNA fragment corresponding to nucleotides Ϫ128 to Ϫ112 was 3Ј-end-labeled with digoxigenin-11-ddUTP using a DIG gel shift kit (Roche Molecular Biochemicals). The reaction mixture (20 l) contained 100 ng of recombinant SREBP-2-(1-481), 30 fmol of the end-labeled probe, 20 mM Hepes-KOH (pH 7.6), 1 mM EDTA, 10 mM (NH 4 ) 2 SO 4 , 1 mM dithiothreitol, 0.2% (w/v) Tween 20, and 30 mM KCl. Each reaction mixture was incubated at room temperature for 20 min. Following the addition of 0.5 g of antibodies, the reaction mixture was placed on ice for 30 min and then loaded directly onto a 6% polyacrylamide gel in 0.5ϫ buffer containing 45 mM Tris borate and 1 mM EDTA. In competition assays, an excess amount of an unlabeled 17-bp fragment was added prior to addition of the labeled probe. The bands were detected by an anti-digoxigenin antibody (Roche Molecular Biochemicals).
Insulin Effect-HepG2 cells were set up on day 0 in medium A supplemented with 2.5% FCS. On day 1, after 4 h of transfection, the medium was removed, and the cells were then washed with PBS and refed with medium A containing 2.5% LPDS supplemented with the indicated concentration of insulin in the presence of either sterols or the inhibitor, as described above. After 48 h of culture, the cells were harvested, and luciferase assays were carried out. For Western blot analysis of SREBP-2, the nuclear extracts of HepG2 cells without transfection were prepared as described previously (16).
Antibodies-Polyclonal antibodies (RS001 against human MTP and RS004 against human SREBP-2) were produced by immunizing rabbits with a fusion protein encoding six consecutive histidines followed by amino acids 300 -507 of human MTP and amino acids 1-481 of human SREBP-2, respectively. The fusion protein constructs were cloned into a pET28(a) vector (Novagen), expressed in Escherichia coli, and purified by Ni 2ϩ -Sepharose affinity chromatography.

Regulation of MTP mRNA and Protein Levels-HepG2 cells
were cultured with LPDS in the presence of either sterols (sterol-loaded conditions) or a HMG-CoA reductase inhibitor (pravastatin) plus mevalonic acid (sterol-depleted conditions), and their total RNA was prepared. In the absence of sterols, the MTP mRNA level was reduced by 53% ( Fig. 2A). To see if such reduced mRNA levels lead to an decrease in MTP protein levels, Western blot analysis was carried out. Because the half-life of the MTP protein has been reported to be relatively long (4.4 days), HepG2 cells were cultured for a longer than usual period either in the presence or absence of cholesterol (10). The level of a 97-kDa band detected by a polyclonal antibody against human MTP declined by 49% in the absence of sterols (Fig. 2B). These results indicate that intracellular cholesterol affects MTP protein levels in the long term through the regulation of its mRNA levels.
MTP Gene Expression Is Regulated by Sterols-It is of interest to determine whether the above phenomenon is due to transcriptional regulation of the MTP gene. Thus, we isolated the 5Ј-flanking region of the human MTP gene and searched for the sequence motifs potentially responsible for such regulation. It has been demonstrated using luciferase assays that the MTP promoter activity is positively regulated by cholesterol (22). We constructed various deletion versions of reporter genes ( Fig. 1) and carried out luciferase assays. HepG2 cells were transfected with these reporter genes and cultured under sterol-loaded or -depleted conditions for 2 days. Fig. 3 shows that a significant decrease in luciferase activity was observed under the steroldepleted conditions with the Ϫ204 and Ϫ124 bp reporter genes (pMTPϪ204 and pMTPϪ124, respectively), but there was no alteration with the Ϫ109 and Ϫ100 bp reporter genes (pMTPϪ109 and pMTPϪ100, respectively). The magnitude of suppression of luciferase activity by sterol depletion (40-50%) was similar to that of endogenous MTP mRNA (Fig. 2). Deletion of the region between Ϫ109 and Ϫ100 bp containing a putative AP-1-binding site ( Fig. 1) reduced luciferase activity dramatically, suggesting that this region is critical for transcription.
SREBPs Can Regulate MTP Gene Expression-To determine whether the above phenomenon is due to effects of SREBPs that are activated by proteolysis under sterol-depleted conditions, HepG2 cells were cotransfected with one of the reporter genes and an expression construct of an active form of SREBP-2 (amino acids 1-481) or SREBP-1 (amino acids 1-487). Enforced expression of SREBP-2-(1-481) or SREBP-1-(1-487) diminished the luciferase activity of pMTPϪ204 or pMTPϪ124, respectively (Table I), mimicking the sterol-depleted conditions observed in Fig. 2. Further deletions from Ϫ124 to Ϫ109 or Ϫ100 bp abolished the inhibitory effects of SREBP-1 and -2. These results strongly suggest that the human MTP promoter contains the elements regulated by sterols and that the region between Ϫ124 and Ϫ109 bp is necessary for SREBP-mediated transcriptional regulation.
SREBP-2 Can Bind the Sterol Regulatory Element-The region between Ϫ124 and Ϫ109 bp contains a GCAGCCCAC sequence (Ϫ124 to Ϫ116 bp), resembling the SRE sequence in the LDL receptor gene (TCACCCCAC). To determine whether the putative SRE is able to bind SREBP-2, we performed gel mobility shift assays with recombinant human SREBP-2-(1-481) and a digoxigenin-labeled DNA fragment. As shown in Fig. 4, a single-shifted DNA-protein complex was observed in the presence of recombinant SREBP-2 (lane 2). The band almost completely disappeared in the presence of an excess amount of an unlabeled wild-type probe, but not a mutant probe (lanes 4 and 5), and was supershifted by antibodies FIG. 2. Northern and Western blot analyses of the MTP gene and protein in HepG2 cells. A, HepG2 cells were cultured with medium containing 5% LPDS supplemented either with 1 g/ml 25hydroxycholesterol plus 10 g/ml cholesterol (sterol-loaded conditions) or with a 50 M concentration of a HMG-CoA reductase inhibitor (pravastatin) plus 50 M sodium mevalonate (sterol-depleted conditions) for 48 h. Twenty-g total RNA samples were fractionated on 1% agarose gel, transferred to nylon membrane, and hybridized with a digoxigenin-labeled riboprobe for either MTP or glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The -fold change in MTP mRNA, relative to that under sterol-loaded conditions, was calculated after correction for loading differences with glyceraldehyde-3-phosphate dehydrogenase. In three separate experiments, the same relative mRNA levels were obtained. B, HepG2 cells were cultured with medium containing 5% LPDS with or without 5 g/ml cholesterol for 14 days. The cells were further cultured under either sterol-loaded or -depleted conditions for 48 h. Fifty-g protein samples were fractionated on 7% polyacrylamide gel and blotted with antibodies against human MTP. In three separate experiments, the same results were obtained. Signals in both Northern and Western blots were quantified with a FluorImager 595. against human SREBP-2 (lane 3). These results clearly show that SREBP-2 is capable of binding the GCAGCCCAC sequence in the MTP promoter.
The SRE Is Involved in Transcriptional Regulation-To verify that the SRE is involved in the SREBP-mediated transcriptional regulation of the human MTP gene, luciferase assays using the SRE mutant reporter gene were carried out (Fig. 5). HepG2 cells were cultured under sterol-loaded or -depleted conditions or were cotransfected with an expression construct of an active form of SREBP-2. In support of the involvement of the SRE site, suppression of luciferase activities by both sterol depletion and SREBP-2 expression was abolished when the SRE was mutated (pMTPϪ124SREKO) (Fig. 5, A and B) or deleted (pMTPϪ109) (Fig. 6A), and the level of luciferase expression increased ϳ1.8-fold by mutation of the SRE (Fig. 5). These results demonstrate that the SRE site is essential for the SREBP-mediated transcriptional regulation of the MTP gene.
The Transactivation Domain of SREBP-2 Is Not Required for Regulation-SREBPs are activators of transcription of the genes encoding the LDL receptor, HMG-CoA synthase, and reductase. The acidic NH 2 -terminal region of SREBPs acts as the transcriptional activation domain, and deletion of this sequence prevents transcriptional activation of the LDL receptor gene (16). To determine whether a truncated version of SREBP-2 (amino acids 31-481) lacking 30 amino acids in the transactivation domain coded in exon 1 inhibits MTP transcription, luciferase assays using reporter genes and expression constructs of wild-type or truncated SREBP-2 were carried out. As long as the SRE was intact, both wild-type and truncated SREBP-2 similarly diminished luciferase activity by ϳ40% (Fig. 6A). Once the SRE was deleted (pMTPϪ109) or mutated (pMTPϪ124SREKO), neither wild-type nor truncated SREBP-2 altered luciferase activity. On the contrary, SREBP-2-(31-481) inhibited transcription of the LDL receptor, whereas SREBP-2-(1-481) stimulated it (Fig. 6B). These results show that the mechanism by which MTP transcription is regulated by SREBPs is distinct from that for the LDL receptor gene and further support the direct involvement of SREBPs through binding to the SRE.
Insulin and SREBPs Negatively Regulate MTP Gene Expression Independently-It has been reported that insulin negatively regulates MTP gene expression and that the region between Ϫ123 and Ϫ112 bp is essential for the insulin response (22). The fact that the SRE in the MTP promoter is found in this region raises the possibility that SREBPs might mediate certain gene regulatory effects of insulin or that SREBPs might simply mimic the functions of an unknown DNA-binding protein activated by insulin. To see if the insulin effect is dependent on SREBPs, luciferase assays using various reporter genes in the presence or absence of insulin were performed. HepG2 cells were transfected with one of the reporter genes and cultured under either sterol-loaded or -depleted conditions in the presence of the indicated concentration of insulin for 2 days. When cells were transfected with the Ϫ124 bp reporter gene, MTP gene expression was suppressed under sterol-loaded conditions (Fig. 7A, white bars) by 38 and 53% at insulin concentrations of 0.1 and 1.0 M, respectively. However, insulin did not affect the reduced luciferase activity that occurred under sterol-depleted conditions (Fig. 7A, shaded bars). Luciferase activity with the LDL receptor reporter gene was increased under sterol-depleted conditions in the presence or absence of insulin (Fig. 7B). Fig. 7C shows that the amounts of active forms of nuclear SREBP-2 under sterol-loaded conditions were significantly low in the presence or absence of insulin as determined by Western blotting. Therefore, the suppression of MTP gene expression by insulin under sterol-loaded conditions might imply that SREBPs only slightly mediate, if at all, the insulin-negative effect. When cells were transfected with the Ϫ109 bp reporter gene without both the SRE and the putative insulin-responsive element, luciferase activity was not changed by the addition of sterols or insulin to the medium, in accordance with the results in Fig. 3 as well as observations in a previous report (22) (data not shown). DISCUSSION The purpose of this study was to gain insight into the effects of sterols on the MTP activity that controls the secretion of apoB-containing lipoproteins, chylomicrons, and very low density lipoproteins from the small intestine and liver. We have demonstrated that sterols regulate MTP gene expression and the amount of the protein in HepG2 cells (Fig. 2). We performed a series of luciferase assays to identify the sequence motifs responsible for transcriptional regulation of the MTP gene. Since it has been reported that all the putative positive and negative response elements for the liver-specific MTP gene expression are localized within the human MTP promoter Ϫ142 bp region (22), we focused on the promoter activity of the first 200 bp 5Ј to the transcription start site. The following line of evidence indicates that SREBPs can negatively regulate MTP gene expression through binding to the SRE site in the promoter. First, the reporter gene containing at least the Ϫ124

FIG. 4. SREBP-2 binds to the GCAGCCCAC sequence in the MTP promoter.
A double-stranded DNA fragment corresponding to nucleotides Ϫ128 to Ϫ112 was 3Ј-end-labeled with digoxigenin-11-ddUTP. In lanes 1 and 2, the reaction mixture was incubated without or with recombinant SREBP-2-(1-481), respectively. Following the addition of 0.5 g of antibodies, the reaction mixture was placed on ice for 30 min (lane 3). In competition assays, a 1000-fold molar excess of an unlabeled 17-bp fragment (a wild-type (WT) fragment and a mutant (Mut) fragment with GAATTC instead of AGCCCA in the SRE) was added prior to addition of the labeled probe. DNA-protein complexes transferred to nitrocellulose membrane were detected with anti-digoxigenin antibodies.
bp region was positively regulated by sterols, i.e. was negatively regulated by sterol deprivation (Fig. 3). When either human embryonic kidney 293 or HeLa cells were transfected with the reporter genes plus an expression plasmid for hepatocyte nuclear factor-4 (HNF-4), luciferase activities were detectable, and the sterol-mediated regulation was also observed (data not shown), suggesting that HNF-4 is necessary for cell type-specific expression of the MTP gene and that the regulation is not liver-specific. Second, the Ϫ124 to Ϫ109 bp region was involved in the sterol-mediated regulation of MTP gene expression (Fig. 3). Third, enforced expression of an active form of either SREBP-1 or -2 suppressed MTP gene expression, mimicking the effect of sterol depletion (Table I). Indeed, we found a putative SREBP-binding site, a GCAGCCCAC sequence resembling the original SRE (TCACCCAC) in the human LDL receptor promoter, in the Ϫ124 to Ϫ109 bp region and confirmed by gel mobility shift assays that this sequence is able to bind SREBP-2 (Fig. 4). Fourth, deletion or disruption of the SRE abolished the response to both sterols and SREBP-2 ( Figs. 5 and 6). We show that the MTP gene is up-regulated by sterols, as well as the gene encoding cholesterol 7␣-hydroxylase, the rate-limiting enzyme in the bile acid synthesis pathway, regulated by liver X receptors (LXRs), nuclear receptors activated by oxidized derivatives of cholesterol (23). Although this fact raises the possibility that MTP is another target gene for LXRs, we failed to find a putative LXR-binding element with a direct repeat of the half-site sequence AGGTCA separated by 4 nucleotides in the Ϫ124 to Ϫ109 bp region.
It has been demonstrated that SREBPs stimulate the expression of target genes in cooperation with either of the general transcription factors Sp1 (24) and NF-Y (21). Recent studies show that cAMP response element-binding protein-binding protein is capable of binding the NH 2 -terminal transactivation domain of SREBPs and is required for SREBP-regulated tran- scription (25,26). The fact that neither the Sp1-nor NF-Ybinding site exists in the vicinity of the SRE in the MTP promoter indicates that these cooperative factors are not required for the inhibitory effect of SREBPs on MTP gene expression. As shown in Fig. 6, a dominant-negative form of SREBP-2 (amino acids 31-481) lacking the transactivation domain, which is unlikely to bind to cAMP response element-binding protein-binding protein, suppressed MTP gene expression in the same manner as wild-type SREBP-2-(1-481). Although the stimulation of LDL receptor gene expression required the transactivation domain, the suppression of MTP gene expression did not. Taken together, we conclude that the mechanism by which SREBPs suppress MTP gene expression is novel and distinct from the known mechanisms through which the transcription of numerous genes, including the LDL receptor and HMG-CoA synthase genes, is up-regulated by the combination of SREBPs with cooperative factors.
In this study, we found that the SRE (Ϫ124 to Ϫ116 bp) overlaps the putative insulin-responsive element (Ϫ123 to Ϫ112 bp) previously identified (22). We demonstrate that insulin negatively regulates MTP gene expression and that the approximate Ϫ124 to Ϫ109 bp region is involved in regulation in accordance with previous in vivo and in vitro observations (10,22). We had questioned whether SREBPs mediate the gene regulatory effects of insulin, and the current results suggest that SREBPs regulate MTP gene expression independently of the insulin-negative effects. First, MTP gene expression was significantly repressed by insulin under sterol-loaded conditions even though the amount of active forms of nuclear SREBP-2 was quite low (Fig. 7, A and C). Second, increased amounts of active forms of nuclear SREBPs under sterol-depleted conditions suppressed the transcription of the MTP gene even at lower concentrations of insulin, suggesting that the inhibitory effects of SREBPs do not require the signaling induced by insulin (Fig. 7A). Although a previous report (27) demonstrated that SREBP-1 mediates activation of the LDL receptor by insulin and that the effect might be linked to the mitogen-activated protein kinase cascade, in the current study, the sterol-mediated transcriptional regulation of the LDL receptor was not significantly altered in the presence or absence of insulin (Fig. 7B). However, we cannot rule out the possibility that a trace amount of the active form of nuclear SREBPs under sterol-loaded conditions mediates the insulin effect.
It has been demonstrated that the insulin response sequence on the fatty-acid synthase promoter contains an E-box CANNTG sequence that is able to interact with the ubiquitous basic helix-loop-helix leucine zipper transcription factors, upstream stimulatory factor (USF)-1 and USF2 (28). Two adjacent SRE sequences on the fatty-acid synthase promoter flank the E-box. Unlike MTP gene expression, expression of the fatty-acid synthase gene is stimulated by insulin and SREBPs (28,29). Analysis of endogenous fatty-acid synthase mRNA expression in USF1 and USF2 knockout mice revealed that USFs and SREBPs independently activate fatty-acid synthase gene expression (30). It is therefore possible that SREBPs might be able to substitute for the unidentified factor(s) bound to the MTP insulin-responsive element by binding an overlapping site, thereby down-regulating MTP gene expression. In FIG. 7. Effect of insulin and sterols on the expression of MTP and the LDL receptor genes. A and B, HepG2 cells were cultured with medium containing 2.5% FCS for 24 h. The cells were transfected with the indicated reporter plasmid for 4 h, washed with PBS, and then refed with medium containing 2.5% LPDS supplemented with either sterols (sterol-loaded conditions; white bars) or the inhibitor (sterol-depleted conditions; shaded bars) plus the indicated concentration of insulin. After 48 h of culture, the cells were processed for the luciferase and ␤-galactosidase assays. Luciferase values were normalized to ␤-galactosidase activity. Promoter activities of the respective constructs under sterol-loaded conditions without addition of insulin are considered as 100% (pMTPϪ124, 1500ϳ2300 relative light units/unit; and pLDLR, 5500ϳ8400 relative light units/unit). The values given are the average of data from more than three experiments. Data are expressed as means Ϯ S.D. C, HepG2 cells were cultured with medium containing 2.5% FCS for 24 h and further cultured with medium containing 2.5% LPDS supplemented with either sterols (sterol-loaded conditions (ϩ)) or the inhibitor (sterol-depleted conditions (Ϫ)) plus the indicated concentration of insulin for 48 h. Thirty g of protein of the nuclear extracts was subjected to Western blot analysis using polyclonal antibodies directed against the NH 2 -terminal 481 amino acids of human SREBP-2. The arrow denotes the NH 2 -terminal active forms of SREBP-2.
our experiments, we found that further deletion of the NH 2terminal portion of SREBP-2 from amino acids 32 to 180 abolished the inhibitory effect (data not shown). Because only 29% of the amino acid sequence for the NH 2 -terminal region (amino acids 32-180) of SREBP-1 and -2 is identical, it appears that the entire region might be required for sufficient inhibition of the transcriptional activation of the MTP gene rather than in association with unidentified corepressor. As shown in Fig. 2, deletion of the region between Ϫ109 and Ϫ100 bp of the MTP promoter, which is located close to the SRE and contains a putative AP-1-binding site (19,22), diminished the luciferase activity dramatically. It is possible that insulin and SREBPs may disrupt the interaction between AP-1 and this region, thereby diminishing MTP gene expression. Further studies will be required to elucidate the precise mechanism.
Horton and co-workers (31)(32)(33) reported that MTP mRNA was elevated 3-6-fold in the livers of transgenic mice overexpressing SREBP-2, but was not significantly increased in mice either overexpressing mutant SREBP cleavage-activating protein, which stimulates the proteolytic processing of endogenous SREBPs in a sterol-independent manner (1.7-fold), or overexpressing SREBP-1a (1.2-fold). We are aware that these findings are apparently not in accordance with our results, but as the authors themselves stated in these reports, it remains to be elucidated whether the elevation of MTP mRNA in these animals is a direct effect of altered SREBP activity or whether it is secondary to increased lipid content in the liver. It is possible that overexpression of SREBPs in these transgenic mice might activate certain crucial genes such as AP-1 and/or HNF-4, which might be involved in MTP gene expression (22). Furthermore, a profound elevation of fatty acid synthesis in these mice might modulate the level of HNF-4 transcription by the production of an increased amount of fatty acyl-CoA ligands for HNF-4 (34). In the absence of resolution of these competing explanations, the results obtained from these transgenic mice studies cannot therefore be taken to be in clear contradiction to the findings that we report. Further investigation will be necessary to determine the source of the ambiguity.
In summary, sterols positively regulate MTP gene expression and protein synthesis in HepG2 cells. Unlike acute insulin effects, stimulated transcription of the MTP gene by higher intracellular sterol levels over the longer term is able to bring about an increase in MTP protein levels, despite the slow turnover rate of the MTP protein. It is likely that elevated MTP activity augments lipoprotein production and secretion by facilitating the assembly of apoB and lipids, preventing the intracellular degradation of apoB. It is noteworthy that cholesterol is one of the determinants of lipoprotein production and secretion through its effect on MTP activities. The sterol-dependent transcriptional regulation of the MTP gene is mediated by SREBPs through binding to the SRE in the MTP promoter. Here we report for the first time a novel mechanism by which SREBPs are able to inhibit MTP gene expression, a mechanism distinct from that of the transcriptional stimulation of the sterol-regulated genes related to cholesterol and fatty acid metabolism. In this regard, it is important to note that SREBPs play a central role in lipid metabolism, regulating not only the synthesis and uptake of cholesterol and fatty acids, but also lipoprotein production and secretion.