Polyunsaturated Fatty Acids Decrease Expression of Promoters with Sterol Regulatory Elements by Decreasing Levels of Mature Sterol Regulatory Element-binding Protein*

Membrane physiology, plasma lipid levels, and intracellular sterol homeostasis are regulated by both fatty acids and cholesterol. Sterols regulate gene expression of key enzymes of cholesterol and fatty acid metabolism through proteolysis of the sterol regulatory element-binding protein (SREBP), which binds to sterol regulatory elements (SRE) contained in promoters of these genes. We investigated the effect of fatty acids on SRE-dependent gene expression and SREBP. Consistent results were obtained in three different cell lines (HepG2, Chinese hamster ovary, and CV-1) transfected with SRE-containing promoters linked to the luciferase expression vector. We show that micromolar concentrations of oleate and other polyunsaturated fatty acids (C18:2–C22:6) dose-dependently (0.075–0.6 mmol) decreased transcription of SRE-regulated genes by 20–75%. Few or no effects were seen with saturated free fatty acids. Fatty acid effects on SRE-dependent gene expression were independent and additive to those of exogenous sterols. Oleate decreased levels of the mature sterol regulatory element-binding proteins SREBP-1 and -2 and HMG-CoA synthase mRNA. Oleate had no effect in sterol regulation defective Chinese hamster ovary cells or in cells transfected with mutant SRE-containing promoters. We hypothesize that unsaturated fatty acids increase intracellular regulatory pools of cholesterol and thus affect mature SREBP levels and expression of SRE-dependent genes.

sis and modulate plasma and intracellular lipid levels. Both fatty acids and cholesterol stimulate cholesterol esterification via activation of acyl-coenzyme A:cholesterol acyltransferase (ACAT), 1 a reaction that is important in regulating intracellular cholesterol pools (1)(2)(3). Fatty acids and cholesterol also exert negative feedback control on their respective biosynthetic pathways (4 -6).
Sterol regulatory elements (SREs) are cis-acting promoter elements present in genes that perform key steps in lipid related metabolic pathways (7). In sterol depletion, the sterol regulatory element-binding proteins (SREBPs) are cleaved and activate SREs in the nucleus. Recent studies show that sterols regulate important genes of fatty acid as well as sterol metabolism (8 -10). Overexpression of SREBP-1 promotes fatty acid and triglyceride synthesis-related pathways as opposed to cholesterol pathways, whereas SREBP-2 overexpression affects cholesterol over triglyceride pathways (11).
Although fatty acids have been demonstrated to be important regulators of plasma LDL cholesterol levels due to alterations in LDL receptor activity (12,13). Precise molecular mechanisms of regulation are not defined. Like cholesterol, fatty acids also modulate mechanisms important to cell function and cholesterol homeostasis (1-3, 14, 15).
Because of the close interaction between fatty acid and cholesterol metabolism, we questioned whether fatty acids directly affect SREBP maturation and expression of genes that are controlled by SREs. Our experiments demonstrate that oleate and polyunsaturated fatty acids inhibit SREBP-regulated transcription of SRE reporters. This effect is additive to the inhibitory effects of cholesterol and is mediated via the SREBP proteolytic maturation cascade. The data provide new evidence for direct molecular interactions between fatty acid and cholesterol metabolism. 32 -P]Deoxycytidine 5Ј-triphosphate (Ͼ 3000 Ci/mmol) was obtained from ICN Biochemicals, Inc. (Irvine, CA). Enzymes for reverse transcription-polymerase chain reactions, Trizol, Lipofectin, and cell culture reagents were purchased from Life Technologies, Inc. Ethanol, oleate, and fatty acid free bovine serum albumin (BSA) were from Sigma. Monoclonal antibodies SREBP-1 (IgG-2A4) and SREBP-2 (IgG-7D4) were obtained from the supernatant of the hybridoma cell lines CRL 2198 and CRL 2121, respectively, American Tissue Culture Collection (ATCC, Rockville, MD).
Plasmids-The pSyn SRE plasmid contains a generic TATA and three SRE elements (Ϫ325 to Ϫ225) of the hamster HMG-CoA synthase promoter fused into the luciferase pGL2 Basic vector (Promega, Madison, WI) (17,18). The fatty acid synthase promoter plasmid linked to luciferase (FAS-150) was described before (10). The JS-15 plasmid is * This work was supported by National Institutes of Health Grants HL 40404, DK 07715, and HL48044. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
* identical to pSyn SRE, except for a double point mutation resulting in loss of sterol regulated transcription (18). The ␤-galactosidase plasmid (␤-gal) used for transfection control, consists of the lacZ gene driven by a CMV promoter.
Transient DNA Transfection-Cells plated in 6-well plates to obtain 80% confluency after 24 h were transfected with pSyn SRE and CMV-␤-galactosidase (1 g each/well) using Lipofectin (10 l/well) as described by the manufacturer. Cells were transfected for 6 h and incubated for 24 h (CV-1 and CHO) or 36 h (HepG2). Cells were then washed twice with phosphate-buffered saline and lyzed with 400 l of cell culture lysis reagent (Promega, Madison, WI). After a further 15-min incubation, cells were scraped, transferred to microfuge tubes, vortexed for 15 s, and centrifuged at top speed for 2 min at 4°C, and the supernatant was analyzed.
Enzyme Assays-Luciferase activities were measured in a luminometer (Berthold LB 9501, Wallac Inc, Gaithersburg, MD) with a luciferin reagent from Promega (Madison, WI). ␤-Galactosidase assays were performed as described (19). Luciferase activity in relative light units was divided by ␤-galactosidase activity (activity/h) for each extract.
Free Fatty Acid Addition-Fatty acids were dissolved in 100% ethanol and complexed to BSA as described (3). Equal amounts of ethanol (Ͻ0.1% of medium volume) were added under control conditions and had no effects on reporter gene assay.
Immunoblots-Cells plated in regular growth medium were incubated on day 2 in DMEM 1% BSA alone (control), with oleate (0.15 mM or 0.3 mM), or with cholesterol (10 g/ml) plus 25-OH cholesterol (1 g/ml). After 24 h cells were scraped and pelleted at 1000 ϫ g. The pellet was resuspended in 250 mM Tris-Cl, pH 7.8. Cells were lyzed/ homogenized by passing 20 times through a 22 gauge needle. Protein was determined with the Bio-Rad reagent. Aliquots of each sample (100 g of protein) were electrophoresed on a denaturing 7.5% SDS-polyacrylamide gel electrophoresis. Monoclonal antibodies against SREBP-1 and -2 and peroxidase labeled anti-mouse IgG (Amersham Pharmacia Biotech NIF 824) were used for Western blot analysis. Protein mobilities were compared with molecular weight standards (Bio-Rad).
Northern Blots-Cells were treated as above. After 24 h total RNA was isolated using Trizol. 30 g of total RNA was separated by 1.2% denaturing agarose/formaldehyde electrophoresis and transferred by capillary transfer to Duralon UV-membranes (Stratagene, La Jolla, CA). Blots were hybridized in Quick-Hyb (Stratgene) with cDNA probes for HMG-CoA synthase and glyceraldehyde-3-phosphate dehydrogenase. Probes were labeled by random priming (Stratagene Prime-It® Random priming labeling kit) using 50 Ci of [␣ 32 -P]deoxycytidine 5Ј-triphosphate (3000 Ci/mmol) and 50 ng of DNA fragment.
cDNA Probe-The cDNA probe for Northern hybridization of HMG-CoA synthase was obtained by reverse transcription-polymerase chain reaction from human THP-1 macrophages mRNA using previously described primers (20).
Data Analysis-Statistical significance was calculated by paired t tests.

Oleate Inhibits the Transcription of SRE-containing Promoters-
To determine effects of oleate on SRE-dependent gene expression, HepG2, CV-1, and CHO cells, cells differing in cell cholesterol homeostasis, were transiently transfected with a plasmid originating from hamster HMG-CoA synthase, containing three SRE linked to the luciferase reporter gene (pSyn SRE). HepG2 cells secrete lipoproteins and show high levels of cholesterol esterification through ACAT (21), CHO cells do not secrete lipoproteins but display substantial ACAT activity and storage of cholesteryl esters (22,23). CV-1 cells do not secrete lipoproteins and display low levels of ACAT activity. 2 As reported (24,25) all three cell lines responded to incubation with sterols by repression of SRE transcription (Fig. 1). In each cell line, addition of increasing amounts of oleate caused progressive and significant (p Ͻ 0.005) decreases of pSyn SRE expression (Fig. 1). Adding increasing amounts of sterols to oleate further decreased pSyn SRE expression. The effect was most marked in CV-1 cells compared with CHO and HepG2 cells. In CV-1 cells the combination of cholesterol and oleate was additive and resulted in almost complete inhibition of pSyn SRE expression (Fig. 1C).
Similar effects of oleate on SRE-mediated gene expression were observed with the rat fatty acid synthase promoter, FAS-150 (data not shown). Importantly, oleate had no detectable effect on cell growth, DNA synthesis, ␤-gal expression, or luciferase expression in luciferase control vectors (data not shown).
Oleate Decreases HMG-CoA Synthase mRNA Levels and the Mature SREBP Protein-We tested whether effects of oleate on SRE-dependent gene expression correspond to decreased mRNA levels of SRE-regulated genes and decreased levels of mature SREBP. After 24 h of incubation with oleate, HMG-CoA synthase mRNA levels in CV-1 cells were reduced to 57%, compared with a 61% decrease when incubated with sterols 2 T. Seo, T. S. Worgall, and R. J. Deckelbaum, unpublished data. ( Fig. 2). Using sterols as a control, Western analysis of CV-1 cells incubated with oleate alone showed a substantial decrease in the amount of mature SREBP-1 (Fig. 3). In five separate experiments, oleate alone (0.15-0.3 mM) markedly decreased levels of SREBP-2, as did cholesterol (data not shown).

Effect of Fatty Acid Chain Length and Saturation on SREdependent Gene Expression-pSyn SRE transfected CV-1 cells
were incubated with fatty acids of increasing chain length and unsaturation (Fig. 4). The results demonstrate that saturated fatty acids had no significant effect on reporter gene expression. However, unsaturated fatty acids caused significant (p Ͻ 0.005) down-regulation of reporter gene expression (Fig. 4), and inhibition increased with fatty acid length and number of double bonds. Because linoleate had a wider range of inhibitory effects on SRE-dependent gene expression than did oleate, we assessed effects of increasing free fatty acid (FFA) concentrations using linoleate. Fig. 4 (inset) shows that the effect of increases in FFA levels occurs in a graded, concentration-dependent manner.
Effects of Oleate on SRE-dependent Gene Regulation Are Not Dependent on ACAT Activity-To characterize potential interactions of ACAT and oleate with SRE-dependent gene expression, pSyn SRE-transfected CHO cells were incubated with two ACAT inhibitors, SANDOZ 58 -035 and DUP 128 (gift of Dr. J. Billheimer, DuPont Merck Pharmaceutical Co.). In the absence of fatty acids ACAT inhibitors alone caused a ϳ40% decrease of SRE-dependent gene expression, likely reflecting higher levels of free cholesterol in cellular regulatory pools. However, in the presence or absence of ACAT inhibitors, fatty acids and cholesterol resulted in very similar decreases of pSyn SRE gene expression (data not shown). Thus, effects of oleate on SRE-dependent gene expression are not linked to different levels of ACAT activity.
Effect of Oleate in Cells with Defects in Sterol Regulation-Different SRD cells, transfected with pSyn SRE were used to determine whether the effects of fatty acids on SRE-dependent gene expression are dependent on changes in the production of mature SREBP. SRD-2 cells produce a truncated, transcriptionally active form of SREBP-2 that is not suppressed by sterols (7,25). If oleate would inhibit binding of mature SREBPs to SRE or effects of oleate would occur in a SREBPindependent manner, reporter gene expression would diminish with oleate in these cells. Oleate had no effect on pSyn SRE expression in SRD-2 cells (Table I), suggesting that in the presence of excess mature SREBP, oleate does not inhibit SREdependent gene expression.
SRD-4 cells are characterized by unregulated high levels of proteolysis of precursor SREBP due to a dominant activating mutation of SREBP cleavage-activating protein and show no ACAT activity (7,26). If oleate decreased the effects of SREBP cleavage-activating protein on SREBP cleavage, luciferase expression should decrease when incubated with oleate. Incubation with oleate with and without sterols did not affect SREdependent gene expression in SRD-4 cells (Table I).
A second line of evidence that the fatty acid effect is mediated by SREBP interacting with a sterol-specific binding site is provided by experiments in which CHO cells were transiently transfected with JS-15, a pSyn SRE plasmid mutated in the SRE regions resulting in insensitivity to sterols (18). Incubation of these cells with fatty acids did not down-regulate SRE expression (Table I). DISCUSSION We studied the effects of fatty acids on SREBP-dependent gene expression. The regulatory effects of sterols on genes of fatty acid metabolism via SREBPs have been shown (8,9,(27)(28)(29). We now demonstrate that the effects of mono-and polyunsaturated fatty acids are also SREBP-mediated and result in decreased transcription of at least two SRE regulated  (Table I). Thus, the effect of oleate is independent of but additive to exogenous sterols and is mediated by SREBPs.
The promoter constructs used (10,18) differed with regard to the number of SRE elements and the coregulatory transcription factor that is required for SREBP to activate transcription (10). Thus, effects of fatty acids are not dependent on the type of coregulator or the number of SRE elements.
What are potential mechanisms whereby fatty acids modulate SREBP cleavage? Changes in membrane fluidity are not likely to play a major role because both cholesterol and FFA decrease SRE expression but have opposite effects on fluidity (14,15). Cholesterol increases and polyunsaturated fatty acids decrease membrane fluidity (14,15). Moreover, not only do fatty acids and cholesterol have similar effects, but they also act additively in our study.
Another potential model is that the effects of fatty acids on SRE expression are linked to increasing the intracellular cholesterol pool. The effects of oleate in the presence of cholesterol were most marked in CV-1 cells. The threshold for SREBP regulation might be reached faster in CV-1 cells by the additive effects of cholesterol and oleate due to their low capacity to esterify cholesterol. Effects of oleate were not linked to its ability to enhance ACAT activity. Studies in animal and cell models suggest that long chain fatty acids influence the distribution of sterols in regulatory intracellular cholesterol pools (2,3). Most cell cholesterol is located in the plasma membrane (31), where free cholesterol binds to sphingomyelin with higher affinity than to other membrane phospholipids. Depletion of sphingomyelin in cultured cells blocks proteolysis of SREBP at site 1 (30). Of interest, unsaturated FFAs are activators of neutral sphingomyelinases, which decrease membrane sphingomyelin (32,33). Thus, by decreasing sphingomyelin, fatty acids may displace membrane free cholesterol, making it available as a signal to decrease SREBP maturation.
A number of genes affecting cholesterol homeostasis and fatty acid metabolism contain the SREBP sensitive SREs in their promoter regions. Although the effects of increasing FFA on inhibiting SRE-dependent gene expression are in line with their ability to decrease fatty acid synthase enzyme activity, polyunsaturated FFAs up-regulate LDL receptor activity both in cultured cells (4) and in vivo (13). Results of Rumsey et al. (3) suggest that oleate and polyunsaturated fatty acids do not alter levels of LDL receptor mRNA. Because the promoter region of the LDL receptor gene also contains an SRE, it is likely that FFA can change LDL receptor activity through mechanisms that do not involve SREBPs, perhaps at post-translational or post-transcriptional levels.
In summary, our data showing FFA-mediated effects on SREBP and SRE-dependent gene expression provide another link between fatty acid and cholesterol metabolism.   and SRD-4 cells or in CHO-K1 cells transfected with mutated pSYN-SRE SRD-2 and SRD-4 cells were transfected with pSyn SRE and CMV-␤-gal (1 g each/well). CHO-K1 cells were transfected with JS-15 and CMV-␤-gal (1 g each/well). Each cell line was incubated with control medium (1% BSA), cholesterol (10 g/ml cholesterol plus 1 g/ml 25-OH cholesterol), or oleate (0.3 mM) for 24 h. Luciferase to ␤-gal expression ratios are expressed as the percentages Ϯ S.D. of control experiments performed in triplicate.