J Biol Chem, Vol. 274, Issue 46, 32725-32732, November 12, 1999

Sterol Response Element-binding Protein 1c (SREBP1c) Is Involved in the Polyunsaturated Fatty Acid Suppression of Hepatic S14 Gene Transcription^*

Michelle K. Mater, Annette P. Thelen, David A. Pan^§, and Donald B. Jump^¶

From the Departments of Physiology and Biochemistry, Michigan State University, East Lansing, Michigan 48824

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Polyunsaturated fatty acids (PUFA) suppress hepatic lipogenic gene transcription through a peroxisome proliferator activated receptor  (PPAR)- and cyclooxygenase-independent mechanism. Recently, the sterol response element-binding protein 1 (SREBP1) was implicated in the nutrient control of lipogenic gene expression. In this report, we have assessed the role SREBP1 plays in the PUFA control of three hepatic genes, fatty acid synthase, L-pyruvate kinase (LPK), and the S14 protein (S14). PUFA suppressed both the hepatic mRNA_SREBP1 through a PPAR-independent mechanism as well as SREBP1c nuclear content (nSREBP1c, 65 kDa). Co-transfection of primary hepatocytes revealed a differential sensitivity of the fatty acid synthase, S14, and LPK promoters to nSREBP1c overexpression. Of the three promoters examined, LPK was the least sensitive to overexpressed nSREBP1c. Promoter deletion and gel shift analyses of the S14 promoter localized a functional SREBP1c cis-regulatory element to an E-box-like sequence (¹³⁹TCGCCTGAT¹³¹) within the S14 PUFA response region. Although overexpression of nSREBP1c significantly reduced PUFA inhibition of S14CAT, overexpression of other factors that induced S14CAT activity, such as steroid receptor co-activator 1 or retinoid X receptor , had no effect on S14CAT PUFA sensitivity. These results suggest that PUFA regulates hepatic nSREBP1c, a factor that functionally interacts with the S14 PUFA response region. PUFA regulation of nSREBP1c may account for the PUFA-mediated suppression of hepatic S14 gene transcription.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Polyunsaturated fatty acids (PUFA)¹ (both n-3 and n-6) are potent inhibitors of hepatic de novo lipogenesis. PUFA suppression of de novo lipogenesis is due to inhibition of transcription of key genes involved in lipid synthesis, including acetyl-CoA carboxylase, fatty acid synthase, stearoyl-CoA desaturase 1, malic enzyme, and L-pyruvate kinase. Long chain n-3 fatty acids like eicosapentaenoic acid (20:5,n-3) also enhance fatty acid oxidation, decrease triglyceride synthesis, and reduce serum triglycerides. By regulating these pathways, PUFA promote a shift from fatty acid synthesis and storage to oxidation (for review see Ref. 1).

The molecular basis for PUFA regulation of hepatic gene expression is complex and involves several mechanisms. For example, PPAR is a ligand-activated nuclear receptor that regulates the expression of mitochondrial, peroxisomal, and microsomal enzymes involved in fatty acid metabolism. Certain hypolipemic drugs like fenofibrate and gemfibrozil as well as fatty acids activate PPAR (2). Studies using the PPAR null mouse indicate that both PPAR-dependent and PPAR-independent pathways are operative in the liver (3). For example, PPAR is required for n-3 PUFA activation of acyl- CoA oxidase (AOX) and cytochrome P450-4A (CPY4A) gene expression, but PPAR is not required for the PUFA-mediated suppression of lipogenic gene expression.

Other studies have shown that hepatic cyclooxygenase expression is restricted to non-parenchymal cells, e.g. Kupffer and endothelial cells (4). Interestingly, hepatic parenchymal cells possess G-protein-linked receptors that respond to prostanoids like prostaglandin E₂ and prostaglandin F₂. Prostaglandin E₂ treatment of these cells suppresses mRNAs encoding fatty acid synthase (FAS) and the S14 protein (S14). However, n-3 and n-6 PUFA suppression of hepatic lipogenic gene expression is independent of cyclooxygenase (4). The fact that PUFA regulation of lipogenic gene expression is independent of both cyclooxygenase and PPAR implicates other routes for control.

Recent studies by Worgall et al. (5) provide an important clue to another factor involved in PUFA regulation of gene expression. Their studies suggest that unsaturated fatty acids regulate the nuclear abundance of the sterol response element binding protein-1 (SREBP1). Previous studies show that SREBP1 subtypes played a major role in the hormonal/nutrient control of lipogenic gene expression (6-13). SREBPs are ~130-kDa proteins attached to the endoplasmic reticulum and nuclear membranes through transmembrane-spanning domains. SREBP1 and SREBP2 are encoded by separate genes, and SREBP1 is expressed as two subtypes, 1a and 1c, which arise from differential promoter and exon usage (13). In liver, SREBP1c is the predominant subtype, whereas SREBP1a is the predominant subtype expressed in nonhepatic tissues and most cell lines (7). Cholesterol regulates proteolytic cleavage of SREBP1a and -2 to release a mature (or nuclear) 65-kDa protein, i.e. nSREBP. nSREBP is a basic helix-loop-helix transcription factor that binds sterol response elements (SRE) and some E-boxes in sterol-responsive genes, e.g. hydroxymethylglutaryl-CoA reductase, low density lipoprotein receptor, and fatty acid synthase. Studies in transgenic mice overexpressing different SREBP subtypes suggest that SREBP1 may play a major role in the control of lipogenic gene expression, whereas SREBP2 functions in the regulation of cholesterol synthesis (7-10).

Unsaturated fatty acids reduced the nuclear content of SREBP1a in cultured CV1, CHO, and HepG2 cells (5). The linkage of SREBP1 to lipogenic gene expression and the fact that unsaturated fatty acids affect the nSREBP1 levels suggest fatty acid control of lipogenic gene expression might involve the regulation of nSREBP1 abundance. In this report, we have examined the role hepatic SREBP1c plays in the PUFA regulation of rat hepatic lipogenic gene expression. Our studies show that PUFA down-regulates mRNA_SREBP1 in liver and cultured primary hepatocytes but not in 3T3-L1 adipocytes. Like other lipogenic genes (5), PUFA suppress SREBP1 gene expression through a PPAR-independent mechanism. PUFA also suppresses the nuclear content of SREBP1 in cultured primary hepatocytes. We have used the S14 gene as a model for lipogenic gene expression because the hormonal/nutrient control of S14 gene transcription parallels that described for several lipogenic genes (1). Moreover, a PUFA-regulatory region has been identified within the S14 proximal promoter (14). With this model we tested the hypothesis that SREBP1c functionally interacts with the PUFA response region (PUFA RR). Our studies have localized a cis-regulatory target for SREBP1c action within the S14 promoter. Moreover, overexpression of SREBP1c significantly affects PUFA regulation of S14 gene expression. These studies provide compelling evidence for SREBP1c as a principal target for PUFA control of S14 gene transcription and possibly other hepatic lipogenic genes.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animals-- Male Charles River (CD1) rats were meal-fed a high carbohydrate diet (ICN) supplemented with various fats. The composition of the fats used is shown in Table I (14-16). Gemfibrozil and WY14,643 were added to some of the high carbohydrate diets at 0.2 and 0.1%, respectively. Feeding of wild type and PPAR-null mice was described previously (3).

Primary Hepatocyte Preparation, Transfection, and Treatments-- Primary hepatocytes prepared by collagenase (Liberase, Roche Molecular Biochemicals) perfusion of rat liver were plated at 3 × 10⁶/60-mm Primaria plate for chloramphenicol acetyltransferase (CAT) assays or 10⁷/100-mm Primaria plate for RNA analysis (3, 4). Cells were plated in Williams E medium containing 10 mM glucose, 10% fetal bovine serum, 200 nM insulin, and 10 nM dexamethasone. After a 4-6-h cell attachment period, medium was removed and replaced with Williams E medium containing 10 mM lactate, 200 nM insulin, 10 nM dexamethasone but without serum.

Cells were transfected with CAT reporter vectors (2 µg/plate) and thyroid hormone receptor 1 expression vector (murine leukemia virus TR1, 1 µg/plate) and/or SREBP1 expression vectors (see the legend of Figs. 5, 6, and 9 for details) using Lipofectin^TM (6.6 µl/µg DNA) (Life Technologies, Inc.). Sixteen h later, cells were treated for up to 48 h with Williams E medium containing 25 mM glucose, 1 µM insulin, 1 µM triiodothyronine, and 10 nM dexamethasone and 250 µM fatty acids (Nu-Chek Prep, Elysian, MN) plus 50 µM albumin (3, 4).

RNA Analysis-- Total RNA isolated from rat or mouse liver and primary hepatocytes or 3T3-L1 adipocytes was transferred to nitrocellulose for hybridization with ³²P[cDNA] for S14, FAS, LPK, SREBP1, or CYP4A2 (3, 4). Hybridization was quantified by PhosphorImager analysis (Molecular Dynamics). Our CYP4A2 probe also detected CYP4A3 (93% homologous to CYP4A2). A full-length cDNA for SREBP1/ADD1 (17-19) was obtained from B. Spiegelman (Dana Farber Cancer Center, Boston, MA).

Plasmid Construction-- The construction of S14CAT200, S14CAT124,S14CAT155, S14CAT156, S14CAT158, TKCAT222 was described previously (3, 4, 14). SV-Sport (ADD1/SREBP1c) obtained from B. Spiegelman contained the full-length SREBP1c and generated a ~130-kDa protein in an in vitro transcription/translation reaction (Promega). The mature (nuclear) form of SREBP1c was prepared by polymerase chain reaction amplification of the full-length SV-Sport-SREBP1c plasmid:oligonucleotide primers (sense, 5'-atataagaattcGGAGCCATGGATTGCACATTT; and antisense, 5'-atatatagatctTCAAGGTTTCATGCCCTCCATAGAC) (17, 18). The polymerase chain reaction product encoding the first 403 amino acids was subcloned into both SV-Sport (Life Technologies) and SG5 (Stratagene) expression vectors. DNA sequence and orientation was verified by DNA sequencing. In vitro transcription/translation of these plasmids yielded a protein of ~65 kDa, characteristic of the nSREBP1c.

Assays-- Serum triglycerides were measured using the enzyme-linked kit from Sigma. Cells were assayed for CAT activity and protein concentration (4). CAT units: cpm of [¹⁴C]butylated chloramphenicol/100 µg of protein/h. Each experiment was done in triplicate and repeated at least three times.

Western Analysis for SREBP1-- Measurement of SREBP1 protein was performed using 100 µg of primary hepatocyte nuclear protein prepared in the absence of detergents (20) and separated by SDS-polyacrylamide gel electrophoresis. After electrophoretic transfer to nitrocellulose, blots were exposed to anti-SREBP1 antibody (IgG-2A4) obtained from the supernatant of the hybridoma cell line CRL 2121 (American Tissue Culture Collection, Manassas, VA). The detection system employed the Renaissance Western blot chemiluminescence Plus kit (NEN Life Science Products).

Electrophoretic Mobility Shift Assays (EMSA)-- EMSA was carried out as described (21, 22) with the following modifications. SREBP1c-SG5 encoding the mature form (~65 kDa) of nSREBP1c was transcribed and translated in a TNT wheat germ translation system (Promega). Products in the TNT reticulocyte translation system interfered with binding of SREBP1 to some S14 and FAS promoter elements. The oligonucleotides used in this study were (only the sense strand is shown): FAS SRE: CATCCGGCATCACCCCACCGACGGCG (BamHI ends) (18); FAS E-box: CATCCGTCAGCCCATGTGGCGTGGCG (BamHI ends) (18); S14 80/120: GATCCAAAACGCTGGGATTGGCTCAAAACAAGGCG (BamHI ends) (23); S14 116/158: AGCTTGTCCCTGGGTAGATGGATCGCCTGATACGGACACTGGCGACTGCA (HindIII/PstI ends) (23); S14 159/188: AGCTTCTGAAGTGACA AGCAGAAGCCTGGCCAGGTCTGCA (HindIII/PstI ends) (23); S14 189/220: AGCTTGTTGGCCGCCCACTGAGGCAGTCATGCAGACTGA (HindIII/PstI ends) (23); S14 220/256: AGCTTCCTTCTAACTGGTTGAGCAGCTGCTAAGAAGACTGCA (HindIII/PstI ends) (23); LPK E-box: GATCCTGGGCGCACGGGGCACTCCCGTGGTA (BamHI ends) (24).

[³²P]DNA (~25,000 cpm or ~0.5 nM) were included in a 20-µl binding reaction for 20 min at room temperature. In competition EMSA, unlabeled DNA (1 pmol; 50 nM) was added 20 min before the addition of [³²P]DNA. At the conclusion of the reaction, samples were electrophoretically separated in 10% polyacrylamide gels with 9 mM Tris-Cl, pH 8.3, 9 mM borate, 0.25 mM EDTA as buffer. After separation, gels were dried and exposed to x-ray film.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

PUFA Down-regulates Hepatic SREBP1 Gene Expression in Vivo through a PPAR-independent Mechanism-- In CV1, COS, and HepG2 cells, unsaturated fatty acids suppress the nuclear content of SREBP1a (5), a transcription factor involved in the regulation of several lipogenic genes (6-13, 17-20, 25-27). Since PUFA suppress lipogenic gene expression in liver, we examined the effect of different fat diets on hepatic SREBP1 gene expression (Fig. 1).

View larger version (47K): [in this window] [in a new window]   Fig. 1.   Dietary PUFA coordinately suppresses mRNAs encoding rat hepatic FAS, LPK, S14, and SREBP1. Rats were meal-fed a high carbohydrate diet containing no fat or supplemented with lard, olive, corn, walnut, or fish oil (10% w/w) for 7 days (14). Panel A, livers were harvested, extracted for total RNA, and assayed for specific mRNAs by Northern analysis. Hybridization was quantified by PhosphorImager analysis. Panel B, plasma was assayed for serum triglycerides (TG) (see "Materials and Methods"). Results are represented as the % of no fat control (mean ± S.D., n > 8).

Rats were meal-fed a high carbohydrate diet containing no fat or supplemented with lard, olive, corn, walnut, or fish oil at 10% w/w for 5 days. Hepatic levels of mRNAs encoding SREBP1, FAS, S14, and LPK (panel A) and serum triglycerides (panel B) were measured. Although our SREBP1 cDNA cannot distinguish between SREBP1a and -1c, the predominant subtype expressed in liver is SREBP1c (7). Thus, Northern analysis of hepatic mRNA reflects measurement of SREBP1c gene expression. When compared with rats fed a fat-free diet, the addition of either saturated or monounsaturated fat to the high carbohydrate diet had no effect on the hepatic levels of any mRNA examined. However, hepatic levels of all four mRNAs were suppressed by diets containing PUFA. Fish oil was the most potent inhibitor of lipogenic gene expression.

The effects of dietary fat on mRNA_SREBP1c are specific because previous studies have shown that these dietary fats have no effect on the mRNA encoding -actin, glucokinase, or phosphoenolpyruvate carboxykinase (4, 16). Moreover, mRNAs encoding AOX and cytochrome P450 4A (CYP4A) were induced only by fish oil (3). SREBP1a has been linked to the expression of genes involved in triglyceride assembly and secretion in hepatoma cell lines (28). Our results show that of the dietary fats tested, only fish oil suppressed serum triglycerides. Thus, PUFA regulation of hepatic de novo lipogenesis and triglyceride synthesis may not be tightly correlated to serum triglycerides.

A major target for fatty acid regulation of gene expression in the liver is PPAR (2, 3). To access the contribution PPAR makes to the regulation of hepatic SREBP1 gene expression, rats were fed a high carbohydrate diet supplemented with the PPAR activators gemfibrozil (0.2% w/w) or WY14,643 (0.1% w/w) for 7 days. Using mRNA_CYP4A as a positive control for PPAR activation (3), gemfibrozil and WY14,643 induced mRNA_CYP4A 2.2- and 6-fold, respectively. In contrast, mRNA_SREBP1c was unaffected by these treatments (Fig. 2A).

View larger version (21K): [in this window] [in a new window]   Fig. 2.   PPAR is not required for the PUFA-mediated suppression of mRNASREBP1. Panel A, rats were meal-fed a high carbohydrate diet (HiCHO) without and with supplementation of gemfibrozil (GEM, 0.2% w/w) or WY14,643 (0.1% w/w). After 7 days on the diet, livers were harvested, extracted for total RNA, and assayed for specific mRNACYP4A and mRNASREBP1 by Northern analysis. Results are represented as % change in mRNA (mean ± S.D., n > 8). Panel B, wild type (+/+) and PPAR null (/) mice were meal fed a high carbohydrate diet supplemented with either 10% olive oil or 10% fish oil for 5 days. The effect of this treatment on mRNAs encoding S14, FAS, acyl-CoA oxidase, and CYP4A has been previously published (3). Using the same RNA preparations, mRNASREBP1c was measured by Northern analysis, and the results were quantified by phosphoimaging. Results are expressed as mRNASREBP1 abundance (mean ± S.D., n = 4). PI Units, PhosphorImager units.

A second approach to assess PPAR involvement in SREBP1 gene expression used RNAs derived from wild type (+/+) and PPAR null (/) mice fed either olive or fish oil for 5 days (3). In wild type mice, fish oil promoted a modest (~20%) decline in mRNA_SREBP1. However, in PPAR null (/) mice, fish oil suppressed hepatic mRNA_SREBP1 by 70% (Fig. 2B). Like other lipogenic genes (3), PUFA-mediated suppression of mRNA_SREBP1 does not require PPAR. However, a functional PPAR might influence hepatic PUFA metabolism, which in turn, may alter mRNA_SREBP1 sensitivity to PUFA.

PUFA Suppresses SREBP1 Gene Expression in Primary Hepatocytes but Not 3T3-L1 Adipocytes-- PUFA acts directly on hepatocytes and adipocytes to suppress lipogenic gene expression (3, 4, 14, 16, 29). The effects of specific fatty acids on SREBP1 gene expression were assessed in both primary hepatocytes (Fig. 3A) and 3T3-L1 adipocytes (Fig. 3B). Fatty acid treatment of primary hepatocytes for 48 h shows that 18:1,n-9 had no effect on the mRNAs encoding FAS, LPK, S14, or SREBP1 when compared with cells receiving no fatty acid. However, the addition of PUFA leads to suppression of mRNA encoding each protein. The extent of inhibition increases as the degree of unsaturation or chain length increases. Thus, cells treated with 20:5,n-3 display the greatest suppression of mRNA. mRNA_actin was unaffected by these treatments, whereas mRNA_AOX and mRNA_CYP4A were induced only by 20:5,n-3 (3).

View larger version (37K): [in this window] [in a new window]   Fig. 3.   PUFA suppress mRNASREBP1 in primary hepatocytes but not 3T3-L1 adipocytes. A, primary hepatocytes were fed Williams E medium supplemented with 25 mM glucose, 1 µM insulin, and 1 µM triiodothyronine and 50 µM albumin (Alb) without and with 250 µM 18:1(n-9), 18:2(n-6), 20:4(n-6), or 20:5(n-3) for 48 h with one medium change. At the conclusion of the treatment, mRNAs were extracted and assayed for the relative abundance of FAS, LPK, S14, and SREBP1. Northern analysis of other mRNAs like -actin, glucokinase, CYP2E1, and phosphoenolpyruvate carboxykinase revealed no effect of fat treatment mRNA abundance (14, 16). B, differentiated 3T3-L1 adipocytes (29) were maintained in Dulbecco's modified Eagle's medium without serum but supplemented with 1 µM insulin and albumin (50 µM). Specific fatty acid treatment, i.e. 18:1(n-9), 20:4(n-6), 20:5(n-3), was at 250 µM for 48 h. At the conclusion of the treatment, mRNAs were extracted and assayed for the relative abundance of FAS, S14, and SREBP1. Results are expressed as % albumin control, mean ± S.D., n > 4.

In 3T3-L1 adipocytes, SREBP1a is the predominant subtype (7). PUFA suppress mRNA_FAS and mRNA_S14 in 3T3-L1 adipocytes through a cyclooxygenase 1-dependent conversion of 20:4,n-6 to prostanoids and subsequent activation of a pertussis toxin-sensitive G_i-linked receptor-signaling cascade (4, 29). Treatment of 3T3-L1 adipocytes with 18:1,n-9; 20:4,n-6, or 20:5,n-3 had no effect on mRNA_SREBP1a. Although mRNA_SREBP1c is suppressed in hepatocytes, mRNA_SREBP1a is not affected by PUFA in 3T3-L1 adipocytes. Thus, the effects of PUFA on SREBP1 gene expression are tissue-specific. Since 20:4,n-4 inhibition of S14 and FAS gene expression is cyclooxygenase-dependent (29), this result would suggest that prostanoids may have little effect on SREBP1a gene expression.

PUFA Suppresses Nuclear Content of SREBP1 in Hepatocytes-- The nuclear content of the mature form of SREBP1a (65 kDa) is suppressed in CV1, CHO, and HepG2 cells following treatment with unsaturated fatty acids (5). As shown above, SREBP1c gene expression in liver and cultured primary hepatocytes is suppressed by PUFA but not by monounsaturated fatty acids (Figs. 1-3). To determine if the decline in SREBP1 gene expression paralleled the decline in either the precursor (~130 kDa) or mature form (65 kDa) of SREBP1c protein, Western analysis was preformed using crude nuclear extracts from cultured primary hepatocytes treated with either no fat or 250 µM 18:1,n-9 or 20:4,n-6 for 48 h. After fatty acid treatment, nuclear extracts were prepared, and proteins were electrophoretically separated (20). A representative Western blot is shown, and quantitative results from three separate experiments are illustrated (Fig. 4). SREBP1 precursor (~130 kDa) levels were suppressed 26 and 50% by 18:1,n-9 and 20:4,n-6, respectively. The mature (or nuclear, 65 kDa) form was not affected by 18:1,n-9 but was suppressed by 70% in nuclei of 20:4,n-6-treated cells.

View larger version (27K): [in this window] [in a new window]   Fig. 4.   PUFA suppress hepatocyte nuclear content of mature SREBP1. Primary hepatocytes were maintained in Williams E medium without and with specific fatty acids as described above. After a 48-h treatment with the fatty acids, cells were harvested for preparation of a crude nuclear fraction (20). Proteins were extracted and electrophoresed in a 4-20% polyacrylamide gel with SDS-Tris-glycine as the buffer. Proteins were transferred to nitrocellulose, and SREBP1 was detected using antibody derived from CRL-2121 hybridoma cells. These cells specifically express monoclonal antibodies against SREBP1. Antigen-antibody complexes were detected using the Renaissance Western blot Chemiluminence Reagent Plus Kit (NEN Life Science Products). A representative experiment is shown in panel A. Protein molecular weight markers are shown on the left, and SREBP1 precursor and mature form are shown on the right. The precursor region contains multiple bands and probably represents uncontrolled proteolytic action in the crude nuclear extracts. Western analysis with the SREBP2 antibody (CRL-2198) did not detect changes in precursor or mature forms of SREBP2 (not shown). Con, control. Panel B, the results of three separate experiments was quantified by video densitometry and expressed as -fold change in SREBP1c precursor or mature form. n = 6; mean ± S.D.

The decline in precursor 130-kDa SREBP1c after PUFA treatment is consistent with the decline in mRNA_SREBP1 levels found both in vivo and in primary hepatocytes (Figs. 1-3). The decline in the nuclear form, however, appears to exceed the change in precursor and might implicate either proteolytic processing of precursor SREBP1c to the nuclear form or the enhanced turnover of nSREBP1c. Others suggest that unsaturated fatty acids affect the conversion of the precursor to mature form of SREBP1 (5, 30). We find little or no effect of monounsaturated fatty acids on nSREBP1 levels but a profound effect of PUFA on nSREBP1 content. Clearly, additional studies will be required to define how PUFA regulates nSREBP1c levels.

Effects of nSREBP1c on FAS, LPK, and S14 Gene Transcription-- SREBP1c is a basic helix-loop-helix transcription factor that binds sterol response elements (SRE) and E-boxes (CANNTG) in certain promoters and induces the transcription of several genes involved in cholesterol and fatty acid synthesis (13, 25, 26). The FAS promoter contains both an SRE (at 150 bp) and E-box (at 64 bp) (25), whereas S14 and LPK promoters contain functional E-boxes at 1440 and 160 bp, respectively (31). The PUFA suppression of mRNA_SREBP1 and nSREBP1c (Figs. 1-4) coupled with its potential interaction with multiple lipogenic genes makes SREBP1c an attractive candidate to explain the apparent coordinate regulation of lipogenic genes by PUFA.

To assess how nuclear content of SREBP1c affects hepatic lipogenic gene expression, a co-transfection approach was used to overexpress nSREBP1c in the presence of three lipogenic gene promoters. Accordingly, primary hepatocytes were co-transfected with SG5-nSREBP1c (see "Materials and Methods") along with either FAS CAT, S14CAT, or LPK CAT. Results illustrated in Fig. 5 show that as little as 30 ng of co-transfected nSREBP1c expression vector induced both S14 and FAS CAT by ~5-fold but had no effect on LPK CAT activity. In fact, as little as 1 ng of co-transfected nSREBP1c was sufficient to induce S14CAT activity 2-fold (not shown). At 1 µg of co-transfected nSREBP1c, FAS, S14, and LPK CAT activity were induced 12-, 10-, and 4-fold, respectively. These results indicate a differential sensitivity of the three promoters for overexpressed nSREBP1c. Of the promoters tested, LPK is the least sensitive to elevated nSREBP1c expression.

View larger version (33K): [in this window] [in a new window]   Fig. 5.   Differential sensitivity of the FAS, S14, and LPK promoters to overexpressed nSREBP1c. Primary hepatocytes were transiently transfected with either FAS CAT, S14CAT, or LPK CAT at 2 µg/plate each. Cells were co-transfected with either 30 ng or 1 µg of SG5-nSREBP1c (nuclear form, nSREBP1). The total amount of co-transfecting plasmid was kept at 1 µg/plate by supplementing with SG5 empty vector. After 48 h in culture with 1 media change, cells were harvested and assayed for CAT activity. Results are expressed as fold induction in CAT activity, mean ± S.D., n  6. The FAS promoter extends from 265 to +16 bp, a region that contains both the SRE and E-box. The S14CAT promoter extends from 2.8 kilobases to +19 bp. The LPK promoter extends from 4.3 kilobases to +12 bp.

nSREBP1c functionally interacts with SRE elements within the FAS promoter (25). Although LPK CAT is weakly responsive to overexpressed nSREBP1 (Fig. 5), the target for this action is within the glucose response region, a region containing a direct repeat of E-boxes.² Kim et al. (32) report that SREBP1c binds E-boxes in the S14 CHO RR, i.e. between 1.6 and 1.4 kilobases upstream from the transcription start site. Here we use promoter deletion to identify the functional target for nSREBP1c action in the S14 promoter. The full-length S14CAT reporter gene activity (S170) was induced ~11-fold by overexpression of nSERBP1c (Fig. 6). Removal of the region extending from 2500 to 220 bp had no effect on this activity. Based on studies by others, this deletion should remove elements associated with the glucose/insulin induction of S14 gene transcription (31). Block deletion of the 1.6/1.4-kilobase region yielded essentially the same result, i.e. no change in response of S14CAT to overexpressed nSREBP1c (not shown). Thus, the putative S14 CHO RR is not essential for SREBP1c regulation of S14 gene transcription.

View larger version (22K): [in this window] [in a new window]   Fig. 6.   SREBP1c functionally interacts with the S14 PUFA RR. Construction of S14 proximal promoter deletion fusion genes (S170, S155, S200, S156) and the S14 thyroid hormone response region fused upstream of the TK promoter were described previously (3, 4, 14). The proximal promoter has 3 functional regions, a TATA box at 28/22 bp, an NF1 site at 67/49 bp, and an NFY site at 104/99 bp. The location of the PUFA response region (PUFA RR) is between 220/80 bp and is marked by a heavy line. Primary hepatocytes were transiently transfected with the various CAT fusion genes (2 µg/plate) and co-transfected with SG5-nSREBP1 (1 µg/plate). After 48 h in culture, cells were harvested for CAT activity. Results are expressed as fold induction of CAT by nSREBP1; mean ± S.D., n  9.

Sequential deletion of the proximal promoter, from 220 to 170 bp and from 170 to 120 bp, led to a progressive decline in the response of S14CAT to overexpressed nSREBP1c. The S14CAT reporter gene extending to 120 bp (S156) and the reporter gene containing the thymidine kinase (T222) promoter were not induced by overexpressed nSREBP1c. Thus, the effect of nSREBP1c on S14CAT activity is promoter-specific. The functional target for this control is within the S14 PUFA RR, i.e. between 170/120 bp upstream from the transcription start site.

nSREBP1c Binds within the S14 PUFA RR-- The PUFA RR within the S14 promoter extends from 220 to 80 bp upstream of the transcription start site (14). nSREBP1c functionally interacts with the PUFA RR between 170 and 120 bp (Fig. 6). EMSA was used to determine if nSREBP1c binds directly to this region. nSREBP1c was prepared by in vitro transcription/translation (see "Materials and Methods") and extracts from SG5 programmed translations were used as a null control for protein translation. The FAS E-box and SRE were used as positive controls for SREBP1c binding.

nSREBP1c binds both the FAS-E-box and FAS-SRE (Fig. 7). nSREBP1c does not bind to the S14 120/80-bp element, an element previously reported to bind NFY and CAAT enhancer-binding protein  (21). nSREBP1c bound well to the 158/116-bp region, a region within the functional target for nSREBP1c action (Fig. 6). No direct binding of nSREBP1c to S14 promoter elements between 220 and 158 was detected (not shown). To confirm this specific association of nSREBP1c with the S14 promoter element (158/116 bp), competition gel shift analysis was used (Fig. 7B). [³²P]FAS-SRE was used as the labeled probe, and 1 pmol of various DNAs were used to challenge binding. At 1 pmol, both FAS SRE and FAS E-box are effective competitors for nSREBP1c binding. The S14 elements 120/80, 188/159, 219/189, 256/220 and the LPK E-box failed to compete for binding. Interestingly, the 256/220-bp element contains an E-box at ²³⁴CACGTG, but this DNA fails to compete for SREBP1c binding to the FAS-SRE. Of the S14 DNAs tested, only 158/116 competed well for binding.

View larger version (38K): [in this window] [in a new window]   Fig. 7.   nSREBP1c specifically binds the S14 PUFA RR. In vitro translated nSREBP1c was prepared and used in EMSA (3, 21). Control samples were prepared from translations programmed with SG5 (empty vector). Panel A, direct binding assay of SG5- or nSREBP1c-programmed extracts (3 µl/reaction). The 32P-labeled DNAs (FAS SRE, FAS E-box, S14 120/80, and S14 158/116) used for direct binding are identified below the corresponding lanes. The arrow marks the location of the nSREBP1c-DNA complex, and the asterisk marks the location of a nonspecific band. The nonspecific band generated only in the SG5-programmed extracts and was not competed with any DNA used in this study (not shown). IVTP, in vitro translated protein. Panel B, competition EMSA. The FAS SRE was used as labeled probe. One pmol of competing double-stranded DNA was added to the reaction before 32P-labeled DNA as described under "Materials and Methods." These results in both panels A and B are representative of three separate studies. The sequence of the competing DNAs is shown under "Materials and Methods."

Further definition of nSREBP1c binding to the 158/116-bp region involved competition gel shift analysis using mutant versions of the 158/116-bp element. Fig. 8A illustrates the sequence of the 158/116-bp element and the location of the 3-bp mutation scan. The results of the competition binding show that mutations between 139 and 131 lead to a loss of competition of SREBP1c for the native 158/116-bp element (Fig. 8B). The corresponding sequence for this regions is (¹³⁹TCGCCTGA¹³¹) (Fig. 8A). In Fig. 8C, we compare the S14 SREBP1c binding sequence with known binding sites found in other promoter (13, 26, 27, 33). SREs are a direct repeat (5'-YCCAY-3', where Y represents pyrimidine) separated by a nonessential base. The S14 sequence does not compare favorably with either the consensus SRE or SREs from the low density lipoprotein receptor or FAS promoters. A similar comparison of the S14 sequences with E-box elements reported to bind SREBP1c indicate a better alignment. Based on this comparison, we suggest that nSREBP1c binds an E-box-like element within the S14 proximal promoter at 139/131 bp.

View larger version (53K): [in this window] [in a new window]   Fig. 8.   SREBP1c binds an E-box-like sequence in the S14-PUFA RR. Competitive gel shift analysis was used to localize the binding site for SREBP1c in the 158/166-bp region. A, sequence of the 158/116-bp region. Bracketed sequences were mutated by converting purines to pyrimidines and vise versa. B, 12 oligonucleotide pairs were synthesized, annealed, and used at 1 pmol/reaction (50 nM) in a competitive gel shift binding assay ("Materials and Methods"). The [32P]DNA was the 158/116-bp fragment, and nSREBP1c was generated by in vitro transcription/translation. The lane at the far left represents the binding of SREBP1c with no competing DNA. Lanes labeled Self or FAS-SRE contained unlabeled 158/116 bp (Self) and FAS SRE as competing DNAs. The numbered lanes represent the binding of SREBP1c in the present of the competing mutant version of the 158/116 bp. The numbers correspond to the location of the mutations (bracketed in panel A). C, a comparison of the SREBP1c binding site (139/131 bp) to known SRE-like elements from the low density lipoprotein receptor (LDLR) (13), FAS (25), and acetyl CoA carboxylase (26) promoters and E-box-like elements from the FAS (25) and SCD1 and SCD2 (33) promoters.

Overexpression of nSREBP1c, but Not SRC-1, Abrogates PUFA Suppression of S14CAT Activity-- nSREBP1c binds to and functionally interacts with the S14 promoter at 139/131 bp. This region is within the cis-regulatory target for PUFA control. PUFA suppress hepatocellular mRNA_SREBP1c and the nuclear level of nSREBP1c. Taken together, these studies suggest that PUFA controls S14 gene transcription by regulating the nuclear content of a factor binding this region. If this is so, then overexpression of nSREBP1c should eliminate the PUFA suppression of S14CAT activity. To test this hypothesis, primary hepatocytes were co-transfected with nSREBP1c and S14CAT (S170) and treated with specific fatty acids (Fig. 9).

View larger version (29K): [in this window] [in a new window]   Fig. 9.   Overexpressed nSREBP1c, but not SRC-1, overrides PUFA suppression of S14CAT activity. Primary hepatocytes transfected with S14CAT (S170) and SG5-nSREBP1c (at 0, 30 ng, or 1 µg/plate, see Fig. 5 for details) or SG5-SRC1 (1 µg/plate). The total amount of co-transfecting plasmid was kept at 1 µg/plate by supplementing with SG5 empty vector. Cells were treated with either 18:1 (n-9) or 20:4 (n-6) for 48 h with one media change. Cells were harvested and assayed for CAT activity, and results are expressed as fold induction of S14CAT activity (open bars) and % inhibition of S14CAT Activity by 20:4,n-6 (solid bars). Mean ± S.D., n  9.

In the absence of co-transfected nSREBP1c, 20:4,n-6 suppressed S14CAT activity by ~65%. Transfecting primary hepatocytes with nSREBP1c at 30 ng/plate and 1 µg/plate induced S14CAT activity 5- and 13-fold, respectively (Fig. 9). In these same cells, 20:4,n-6 treatment suppressed S14CAT activity by 30 and 10%, respectively. Thus, increasing the hepatocellular content of nSREBP1c reduces 20:4,n-6-mediated inhibition of S14CAT activity from 65% to  20% inhibition. In some studies where nSREBP1c overexpression was more stimulatory to S14CAT activity, suppression of S14CAT by PUFA was completely lost.

To establish that the abrogation of PUFA inhibition was not simply due to induction of S14CAT activity, cells were co-transfected with SRC-1, a nuclear receptor co-activator. Unlike nSREBP1c, SRC-1 had no effect on S14CAT activity when co-transfected at <1 µg/plate. At 1 µg/plate, SRC-1 induced S14CAT activity ~5-fold, a level comparable to that seen with 30 ng/plate nSREBP1c. In contrast to nSREBP1c, elevated SRC-1 expression had no effect on PUFA sensitivity of the S14CAT reporter gene. Similar results were obtained when hepatocytes were co-transfected with RXR, a co-receptor required for TR1 binding to the upstream thyroid hormone response elements (not shown). Although overexpression of nSREBP1c, SRC-1, and RXR can induce S14CAT activity, only nSREBP1c was able to override PUFA suppression of S14CAT activity. These results suggest that the nSREBP1c is a limiting factor required for S14CAT activity in primary hepatocytes. PUFA regulation of mRNA_SREBP1c and nSREBP1c may explain the PUFA-mediated suppression of S14 gene transcription.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The association of SREBP1 with lipogenic gene expression (6-13, 17-20) and the finding that unsaturated fatty acids suppress SREBP1 protein levels in cell lines (5) prompted our analysis of the role this transcription factor played in the PUFA regulation of rat hepatic lipogenic gene expression. Our goals were to 1) to examine the effect of PUFA on hepatic SREBP1 gene expression and protein levels, 2) to determine if nSREBP1c regulates S14 gene transcription and specific lipogenic genes, e.g. FAS and LPK, 3) to determine if nSREBP1 overexpression abrogates PUFA suppression of S14 gene transcription.

The pattern of suppression of mRNA_SREBP1 by various fat diets (in vivo) and specific fatty acids (in culture) parallels results found for other lipogenic genes and suggests an apparent coordinate regulatory mechanism. Like other lipogenic genes (1, 3), PPAR is not involved in this regulatory mechanism (Figs. 1-3). The effects of PUFA on SREBP1 gene expression are tissue-specific where PUFA down-regulates mRNA_SREBP1c in hepatocytes but does not affect mRNA_SREBP1a in adipocytes (Fig. 3). Preliminary transcriptional run-on studies suggest that SREBP1c gene transcription in rat liver is low relative to FAS and S14. In contrast to FAS and S14, SREBP1c run-on transcription was not affected by PUFA, suggesting the mechanism of PUFA control may be post-transcriptional.³ Thus, the pretranslational regulation of SREBP1c by PUFA may be more like the PUFA-regulated post-transcriptional regulatory mechanism described for glucose-6-phosphate dehydrogenase (34).

The fall in mRNA_SREBP1c following PUFA treatment was accompanied by a corresponding fall in the precursor (pSREBP1c) and a more dramatic fall in the mature (nuclear) form of SREBP1c. Recent evidence suggests that SREBP1c proteolytic cleavage is not sensitive to cellular cholesterol levels (35). Others have suggested that fatty acids affected proteolytic cleavage of pSREBP1a in a fashion similar to cholesterol regulation of SREBP1a and -2. An alternative explanation for the changes in nSREBP1c levels might be more rapid turnover. nSREBP1c is phosphorylated through a mitogen-activated protein kinase pathway (36), and this form of covalent modification might be a signal to alter nSREBP1c activity or its turnover. Clearly, additional study will be required to understand how PUFA regulates nSREBP1c levels. Because PUFA also suppresses S14 and FAS gene expression in white adipose tissue and L1 adipocytes (29), PUFA effects on SREBP1a protein levels in these cell types will need to be assessed.

Studies with transgenic mice overexpressing SREBP1c (10) as well as primary hepatocyte studies using dominant negative versions of nSREBP1c (12) support the notion that nSREBP1c is a key factor in lipogenic gene expression. These observations coupled with the fact that several lipogenic genes have SREs and E-boxes that potentially interact with nSREBP1 (25-27) make it attractive to suggest that fatty acid regulation of hepatic SREBP1c can account for the coordinate regulation of multiple lipogenic and related genes. Two lines of evidence suggest this conclusion might be premature. First, although both FAS and S14 bind SREBP1 at either SRE and E-box-like elements and are strongly induced by overexpressed SREBP1c (Fig. 5), only S14CAT is consistently suppressed by PUFA (14).³ This indicates that the two promoters are sufficiently different and that changes in nSREBP1c levels alone may not be the sole determinant for PUFA regulation of these genes. A second argument against SREBP1c as a coordinator for fatty acid regulation of lipogenic gene expression deals with its role in LPK expression. The E-boxes in the LPK CHO RR (170/142 bp) bind SREBP1c weakly (Fig. 6 and Ref. 37), and a full-length LPK CAT reporter gene is weakly sensitive to nSREBP1 when compared with FAS CAT and S14CAT (Fig. 5). Finally, the target for PUFA control in the LPK promoter is located 3' of the E-boxes, elements that bind HNF4 and NF1 (24, 38). Thus, PUFA regulation of the nSREBP1c levels may account for only part of the PUFA regulation of lipogenic gene expression.

SREBP binding to promoters and its function within promoters is augmented by other transcription factors binding nearby elements (25, 26). In the FAS and acetyl CoA carboxylase promoters, Sp1 binds near the SREs, whereas NFY binds a Y-box near the SRE in the hydroxymethylglutaryl-CoA synthase and farnesyl diphosphate synthase promoters (39, 40). Although a prospective Sp1 site is located at 214 bp in the S14 promoter, its role in S14 gene transcription has not been evaluated. NFY binds the S14 promoter at 104/99 bp and plays an obligatory role in enhancer-mediated transactivation of S14 gene transcription (21). Any mutation that disrupts NFY binding or substitution of other CCAAT-box-binding proteins, like CAAT enhancer-binding protein , essentially inactivates the gene. nSREBP1c functionally interacts with and binds specifically to an E-box-like sequence ¹³⁹TCGCCTGT¹³¹ within the S14 PUFA RR. Linker-scanning mutation analysis across the 146 to 129-bp region leads to a loss of function.² Like NFY (21), nSREBP1c may also play an obligatory role in S14 gene transcription. Additional studies will be required to establish the interaction between nSREBP1 and NFY (at 104/99 bp) as well as other factors upstream of the SREBP1c binding site that facilitate S14 gene transcription.

In conclusion, nSREBP1c binds to and functionally interacts with the S14 PUFA RR at 139/131 bp upstream from the S14 transcription start site. PUFA regulate rat hepatic SREBP1c gene expression and its nuclear content. Together, these observations provide an explanation, at least in part, for the PUFA-mediated suppression of S14 gene expression. PUFA also regulates the activity of PPAR, a nuclear receptor that controls the expression of multiple mitochondrial, peroxisomal, and microsomal enzymes involved in fatty acid oxidation (2). Very long chain n-3 PUFA not only activate PPAR but also directly inhibit triglyceride assembly and secretion (41). The down-regulation of nSREBP1c coupled with the activation of PPAR by n-3 PUFA provides a molecular explanation for the well known shift in hepatic lipid metabolism from lipid synthesis and storage to oxidation associated with ingestion of n-3 PUFA (1). Disruption of this regulatory scheme might contribute to changes in hepatic lipid metabolism associated with obesity and insulin resistance (42, 43).

	FOOTNOTES

^* This research was supported by National Institutes of Health Grant DK 43220, U. S. Department of Agriculture Grant 98-35200-6064), and the Michigan Agriculture Experiment Station.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Contributed equally to this work.

^§ A Lilly Fellow of the Life Science Research Foundation.

^¶ To whom correspondence should be addressed: 115 Giltner Hall, Physiology Dept., Michigan State University, East Lansing, MI 48824. Tel.: 517-355-6475 (ext. 1246); E-mail: Jump@pilot.msu.edu.

² M. K. Mater, A. P. Thelen, D. A. Pan, and D. B. Jump, manuscript in preparation.

³ M. K. Mater, A. P. Thelen, D. A. Pan, and D. B. Jump, unpublished observation.

	ABBREVIATIONS

The abbreviations used are: PUFA, polyunsaturated fatty acids; PPAR, peroxisome proliferator-activated receptor; S14, S14 protein; FAS, fatty acid synthase; SREBP1, sterol response element-binding protein 1; nSREBP1, nuclear SREBP1; ADD1, adipocyte differentiation and determination factor 1; SRC1, steroid receptor co-activator 1; CAT, chloramphenicol acetyl transferase; EMSA, electrophoretic mobility shift assay; AOX, acyl-CoA oxidase; CPY4A, cytochrome P450-4A; CHO, Chinese hamster ovary; RR, response region; TK, thymidine kinase; LPK, L-pyruvate kinase; bp, base pair(s); NFY, Nuclear factor Y.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES