Unsaturated fatty acid-mediated decreases in sterol regulatory element-mediated gene transcription are linked to cellular sphingolipid metabolism.

A major physiological feedback mechanism of cholesterol in transcription of a number of lipid metabolism-related genes is mediated by sterol regulatory elements (SREs) and their binding proteins (SREBPs). Polyunsaturated free fatty acids alone, as well as synergistically with sterols, decrease SRE-mediated gene expression up to 80% in a dose-dependent manner by decreasing levels of the active transcription factor SREBP. We investigated potential mechanisms for this effect. We hypothesized that free fatty acids reduce SREBP-mediated gene transcription by increasing intracellular cholesterol content through the hydrolysis of cellular sphingomyelin, which has a high affinity for free cholesterol. We also questioned whether the lipid second messenger ceramide, a product of sphingomyelin hydrolysis, can decrease SRE-mediated gene transcription. First we investigated the effect of fatty acids on sphingomyelin hydrolysis. Incubation of [(3)H]choline-labeled cells with unsaturated (but not saturated) fatty acids induced hydrolysis of [(3)H]choline-labeled sphingomyelin. Also, incubation of cell extracts from fatty acid-treated cells with [(3)H]sphingomyelin increased generation of [(3)H]ceramide compared with control cells in vitro. We found that addition of ceramide analogs alone and additively with fatty acids decreased SRE expression and that ceramide analogs reduced levels of the transcriptionally active forms of SREBP-1 and SREBP-2. Increasing intracellular ceramide levels by exogenous sphingomyelinase or inhibition of ceramidase decreased SRE-mediated gene expression. None of the above conditions induced apoptosis. Incubation with U18666A, a compound that inhibits intracellular cholesterol movement, increased SRE-mediated gene transcription. C(2)-ceramide abrogated the effect of U18666A on SRE-mediated gene transcription, suggesting cholesterol-independent regulation of SREBP. We provide evidence that sphingomyelin hydrolysis and intermediates of sphingomyelin metabolism (in addition to cholesterol and fatty acids) contribute to regulation of SRE-mediated gene transcription.

Fatty acids and cholesterol interact at a number of metabolic levels. Both contribute to the composition of circulating lipoproteins and cellular membranes. Clinically, higher intakes of polyunsaturated free fatty acids are linked to lower levels of low density lipoprotein cholesterol, whereas higher intakes of saturated and trans-fatty acids correlate with increased low density lipoprotein cholesterol and decreased high density lipoprotein cholesterol as well as increased risk for cardiovascular disease (1). At the cellular level, cholesterol and fatty acids are linked to the regulation of intracellular cholesterol levels by activation of acyl-CoA:cholesterol acyltransferase (2,3). Of relevant interest, genes regulating both cholesterol and fatty acid metabolism contain sterol regulatory elements (SREs) 1 in their promoter regions. Still, little information is available on possible interactions between fatty acids and cholesterol in the regulation of gene transcription.
SRE-binding proteins (SREBPs) are transcription factors that are post-transcriptionally regulated. SREBP in its precursor form is located in the endoplasmic reticulum, where it is bound at the C-terminal end to the SREBP cleavage-activating protein. In sterol depletion, both proteins are translocated to the Golgi apparatus (4,5). Sequential cleavage by two proteases, site-1 protease and site-2 protease, releases transcriptionally active mature SREBP (mSREBP). In the nucleus, SREBP binds to SREs, cis-acting elements in the promoters of genes of cholesterol and fatty acid synthesis (6). Cholesterol exerts a negative feedback mechanism on its own synthesis by inhibiting the proteolytic cleavage of SREBP by site-1 protease.
SREBP-mediated regulation has been demonstrated for a number of key enzymes in the synthesis of polyunsaturated fatty acids (7). We (9) and others (8,10) have shown that monoand polyunsaturated fatty acids decrease SRE-mediated gene transcription by decreasing the amount of transcriptionally active mSREBP and that fatty acids exert this effect alone and synergistically with cholesterol. Fatty acid-mediated regulation of SREBP has also been demonstrated to occur at the transcriptional (8,11,12) and post-transcriptional (10) levels.
Our earlier data showed that the effect of fatty acids is dependent on changing SREBP cleavage. Fatty acids do not affect the mSREBP protein once it has been cleaved from the precursor protein. For example, fatty acids did not reduce SRE-mediated gene expression in mutant cells overexpressing the mSREBP protein (9). * This work was supported in part by National Institutes of Health Grant HL40404. 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  In this study, we tested the hypothesis that one contributing pathway whereby fatty acids modulate intracellular cholesterol movement is by affecting cellular sphingomyelin. Sphingomyelin, a major plasma membrane component, has a higher affinity for free cholesterol than for other phospholipids (13)(14)(15). Sphingomyelinase, an enzyme that hydrolyzes sphingomyelin, has been shown to be activated by fatty acids (16,17). We anticipated, as shown by others (18 -20), that decreasing plasma membrane sphingomyelin content would induce more free cholesterol to partition to other intracellular membranes (e.g. to the endoplasmic reticulum), resulting in decreased SREBP cleavage. Our data support this hypothesis. Furthermore, we found that a product of sphingomyelin hydrolysis, ceramide, independently and reversibly decreases SRE expression. Our data provide new evidence for sterol-and non-sterolrelated effects of fatty acids on the maturation cascade of SREBP.
Cell Culture and Stable Transfections-CHO cells were grown in F-12 nutrient mixture containing 10% fetal bovine serum, 1% (v/v) glutamine, 1% (v/v) penicillin/streptomycin, and 10% (v/v) fetal bovine serum at 37°C in humidified CO 2 (5%). To obtain stable transfectants, CHO cells were plated in 12-well plates at 50% confluency and transfected for 5 h in the presence of serum-free Dulbecco's modified Eagle's medium with pSyn-SRE (1 g/well) and pWLNeo (0.25 g/well) using LipofectAMINE (1.5 l/well). Cells were then incubated for 2 days in growth medium. On day 3, neomycin-containing medium (400 g/ml) was added. Selection for neomycin-resistant colonies was continued for 3 weeks. Pooled clones were analyzed for luciferase expression. Experiments were performed with pooled clones as well as with cells derived from a single clone. Cells were grown in the presence of 400 g/ml neomycin. For experimental use, cells were plated in the absence of neomycin at least 24 h ahead in regular growth medium.
Enzyme Assays-Cells to be analyzed for luciferase activity were lysed in lysis buffer containing 0.1% Triton X-100, 50 mM Hepes, and 10 mM MgSO 4 , pH 7.7. Cells were frozen once at Ϫ70°C, thawed, scraped, collected, vortexed, and briefly centrifuged to pellet cell debris. An aliquot was used to measure luciferase activities in a Berthold LB 9501 luminometer (Wallac, Gaithersburg, MD) with a luciferin reagent (Promega). Luciferase activity in relative light units was divided by protein content (mg/ml) for each extract. ␤-Galactosidase activities were determined in a 30-l aliquot using a high sensitivity ␤-galactosidase assay kit (Stratagene).
Free Fatty Acid Addition-Fatty acids were dissolved in 100% ethanol and complexed to BSA at a molar ratio of 2:1 as described (2). Equal amounts of ethanol (Ͻ0.1% of the medium volume) were added under control conditions and had no effect on reporter gene and other assays.
Protein Determination and Cell Viability-The amount of cell protein was determined by the Bio-Rad method, and BSA was used as a standard. Cell viability was assessed by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide thiazolyl blue test (23). In this assay, the ability of cells to convert 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide to an insoluble formazan by cleavage of the tetrazolium ring by dehydrogenase enzymes is evaluated and used to measure cell viability, cell number, and cytotoxicity. The percentage of apoptotic cells was determined by flow cytometric analysis of isolated nuclei stained with propidium iodide using a method previously described (24). Cells were pelleted at 200 ϫ g and fixed in 2 ml of cold 70% ethanol at 4°C for 60 min. The cells were then centrifuged, washed with 1 ml of phosphate-buffered saline, and resuspended in 0.5 ml of phosphatebuffered saline. To a 0.5-ml cell sample was added 0.5 ml of RNase (1 mg/ml). After mixing, 10 g/ml propidium iodide solution was added. The solution was kept at 4°C in the dark until measured by FACScan flow cytometry (Becton Dickinson, San Jose, CA).
Sphingomyelinase Assay-The sphingomyelinase activity of cell extracts was measured as described (21). CHO cells were plated in 35-mm plates and incubated for 18 h with 0.3 mM oleate, 10 g/ml cholesterol plus 1 g/ml 25-hydroxycholesterol, the control conditions containing 1% BSA, or 10 microunits/ml sphingomyelinase. After the incubation, cells were washed two times with ice-cold 0.  (25).
Sphingolipid Radiolabeling and Quantitation-To label cellular sphingomyelin, 0.5 Ci/ml [ 3 H]choline was added to the culture medium, and labeling was performed for 48 h to allow for equilibrium distribution (26,27). Cells were grown in 6-well plates for the different experiments. To assess the effect of fatty acids on sphingomyelin hydrolysis, cells were treated for 1 h in the presence of 0.3 mM stearate, oleate, or arachidonate. Controls were incubated with serum-free medium containing 1% BSA, 10 g/ml cholesterol plus 1 g/ml 25-hydroxycholesterol, or 10 microunits/ml sphingomyelinase. Because membrane phospholipids may turnover rapidly, we analyzed the effect of fatty acids over the first hour (28,29). After the incubation, cells were lysed in 400 l of lysis buffer, scraped, and transferred to glass tubes. An aliquot was used for luciferase and protein determination. Then, 1 ml of ice-cold methanol, 2 ml of chloroform, and 0.5 ml of 0.1 N HCl were added; vortexed; and spun at 800 ϫ g for 10 min. The upper phase was discarded, and the organic phase was washed with 3ϫ 2 ml of 0.001 N HCl. Lipids were then dried under N 2 . Alkaline hydrolysis was performed to remove [ 3 H]phosphatidylcholine. Dried lipid was incubated in 2 ml of 0.1 N KOH in methanol at 37°C for 1 h. Lipids were then re-extracted by adding 2 ml of chloroform and 1.2 ml of balanced salt solution (135 mM NaCl, 4.5 mM KCl, 1.5 mM CaCl 2 , 0.5 mM MgCl 2 , 5.6 mM glucose, and 10 mM Hepes, pH 7.2) containing 100 mM EDTA (1.08 ml/0.12 ml). After vortexing and centrifugation at 800 ϫ g for 5 min, the lower phase was dried under N 2 (27,30). The extracted lipids were then dissolved in 50 l of chloroform/methanol (1:1), spotted onto TLC plates (Silica Gel 60, Merck, Darmstadt, Germany), and chromatographed with chloroform, methanol, and 0.22% aqueous CaCl 2 (60:35:8, v/v) (31). Ceramide and sphingomyelin (dissolved at 1 g/l) were run as standards. The lipids were identified according to their R F values after visualization in an iodine vapor tank. The TLC plate was cut at the corresponding lipid spots, mixed with scintillation fluid (Ultima Gold, Packard Instrument Co.), and analyzed in a Wallac 1409 scintillation counter (Perkin Elmer Wallac, Gaithersburg, MD). Results are expressed in dpm/mg of protein as a percentage of total counts.
Western Blot Analysis-Cells were plated on day 1 in regular growth medium. On day 2, cells were incubated in control medium (1% fatty acid-free BSA) or under the respective conditions. Two hours before harvesting, all cells received 25 g/ml N-acetyl-leucyl-leucyl-norleucinal to inhibit proteolysis of SREBP by the proteasome. After 8 h, cells were scraped and pelleted at 1000 ϫ g. The pellet was resuspended in Western lysis buffer (10 mM Tris-Cl, 100 mM NaCl, and 1% SDS, pH 7.6) containing protease inhibitors (Complete TM , Roche Molecular Biochemicals). Cells were lysed and homogenized by passing 20 times though a 22-gauge needle. An aliquot of each sample (30 g of protein) was subjected to electrophoresis on a 7.5% denaturing SDS-polyacrylamide gel. The monoclonal antibodies against SREBP-1 and SREBP-2 (BD Biosciences, San Jose, CA) and peroxidase-labeled anti-mouse IgG (NIF 824; Amersham Biosciences, Inc.) were used for Western blot analysis according to the manufacturers' instructions. Detection was performed with the ECL method (Amersham Biosciences, Inc.). Protein mobilities were compared with prestained broad-range molecular mass standards (Bio-Rad).
Northern Blots-CHO cells were plated on day 1 at 80% confluency and treated under the respective conditions for 8 h. Total RNA was isolated by Trizol reagent (Invitrogen) as described by the manufacturer. RNA concentration was calculated from absorbance at 260 nm. Total RNA (30 g) was separated on formaldehyde-containing 1.2% denaturing agarose and capillary-transferred to Duralon UV membranes (Stratagene). The cDNA probe for Northern hybridization of HMG-CoA synthase was obtained by reverse transcription-PCR from human THP-1 macrophage mRNA using previously described primers (9). The blot was hybridized in Quick-Hyb (Stratagene) for 1 h with cDNA probes for HMG-CoA synthase and glyceraldehyde-3-phosphatedehydrogenase corresponding to bases 247-882 as a loading control. Probes were labeled by random priming (Stratagene Prime-It Random priming labeling kit) using 50 Ci of [␣-32 P]CTP (3000 Ci/mmol) and 50 ng of DNA fragment.
Data Analysis-Statistical significance was calculated by paired t tests. Unless otherwise indicated, results are given as means Ϯ S.D. All experiments were repeated on different days at least three times and each time in triplicate.

Fatty Acids Decrease Cellular Sphingomyelin in Cultured
Cells-Most cellular cholesterol is located within the plasma membrane (14), where it has a high affinity for sphingomyelin (13,32). We investigated the ability of different fatty acids to enhance hydrolysis of sphingomyelin, a mechanism that would promote the release of cholesterol from the plasma membrane. CHO cells were incubated for 1 h with myristic acid, palmitate, stearate, oleate, or arachidonate. Before addition of fatty acids, the CHO cells were incubated with [ 3 H]choline for 48 h to label cellular sphingomyelin. The results in Fig. 1 (Fig. 2).
Sphingomyelinase Decreases SRE-mediated Gene Transcription-It has been shown that short-term incubations (90 -150 min) with exogenous sphingomyelinase decrease the amount of mSREBP protein (33). To examine the effect of sphingomyelin hydrolysis over a longer period of time, cells that stably express the SRE linked to a luciferase reporter gene were incubated in the presence of sphingomyelinase for 8 h. Sphingomyelinase decreased SRE-mediated gene transcription to one-third after 8 h. Incubation of sphingomyelinase together with 0.3 mM oleate did not further decrease SRE-mediated luciferase expression (Fig. 3).
Ceramide Decreases SRE-mediated Gene Transcription Alone and Additively with Fatty Acids-Sphingomyelin hydrolysis decreases the capacity to bind cholesterol within the plasma membrane (34 -36) and results in the generation of phosphocholine and ceramide. Ceramide is a second messenger implicated in multiple cellular pathways such as cell growth, differentiation, regulation of transcription factors, and apoptosis (37,38). By several lines of evidence, we determined that fatty acids or ceramide analogs did not induce apoptosis or cell death under our experimental conditions. First, the 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide mitochondrial uptake assay (23) showed no differences between and radioactivity (dpm) associated with each lipid class was determined by scintillation counting. Data are expressed as the percentage of radioactivity in dpm associated with the ceramide or sphingomyelin fraction. Data represent the average of two experiments performed in triplicate. Cell extracts from oleate-treated cells significantly increased ceramide generation compared with controls (p Ͻ 0.05). Cell extracts from cholesterol-treated cells had no effect. control and treatment groups up to 24 h. Second, propidium iodide uptake of nuclei analyzed by FACScan (24) showed no difference between control and treatment groups (data not shown). Third, the effects of ceramide were reversible and did not affect expression of another reporter gene (see below). Fourth, protein levels were not decreased by ceramide treatment.
The effect of ceramide on SRE-mediated gene expression was investigated next. Fig. 4 shows that the ceramide analog C 2ceramide decreased SRE-mediated gene transcription by 5 h and that this effect was synergistic with oleate. Incubation for 16 h with 0.3 mM oleate alone caused the previously described decrease in SRE-mediated gene expression (9). After 16 h, the effect of ceramide alone on SRE expression apparently disappeared; but in the presence of 0.3 mM oleate, a substantial decrease persisted. Addition of fresh C 2 -ceramide at 8 h sustained the decrease in SRE-mediated reporter gene expression, suggesting instability of the compound over an extended period of time. 2 Because C 2 -ceramide can serve as a substrate for the synthesis of sphingomyelin, we also tested the effects of the longer chain C 6 -and C 8 -ceramides, which are not substrates for sphingomyelin synthesis. Incubation for 8 h with C 6 -or C 8ceramide similarly decreased SRE-mediated gene transcription to 46 Ϯ 8 and 43 Ϯ 7%, respectively.
We determined next whether the effect of ceramide analogs on SRE-mediated gene transcription is reversible. Cells were incubated with C 2 -ceramide for 5 h, sufficient time to substantially decrease SRE-mediated reporter gene transcription, as shown in Fig. 4. The medium was then changed to control medium (1% BSA) for 3 h for a total incubation time of 8 h. Fig.  5 shows that SRE-mediated gene transcription returned to control levels, indicating that the effects of ceramide are reversible.
We then determined the protein levels of SREBP at 8 and 16 h in the presence of C 2 -ceramide. The levels of mSREBP-1 and SREBP-2 decreased at 8 h (Fig. 6A). Parallel to the reporter gene assays (Fig. 4), mSREBP levels returned to control values at 16 h, in keeping with the concept of instability of ceramide over an extended time period (Fig. 6B). This indicates that the effects of ceramide on decreasing mSREBP levels are also reversible.
It is noteworthy, that the effect of dihydroceramide analogs on SRE-mediated gene transcription were indistinguishable from those of ceramide analogs (Fig. 7A). Ceramide differs from dihydroceramide by the presence of a 4,5-trans-double bond in the sphingoid backbone; but unlike ceramides, which have a role in the induction of cell signaling of apoptosis, dihydroceramides do not induce apoptosis (39,40). Addition of C 2 -ceramide analogs did not affect the expression of a ␤-galactosidase control reporter gene (Fig. 7B). These results further indicate that decreased SRE expression in response to ceramides is not related to apoptotic pathways or other "toxic" or irreversible pathways.
Increased Endogenous Ceramide Decreases SRE Transcription-Ceramide analogs might have different metabolic properties compared with endogenous cellular ceramide. Thus, we investigated the effect of increased endogenous ceramide levels on SRE-mediated gene transcription. Cells were incubated for 16 h in the presence or absence of 0.3 mM oleate with a 30 M concentration of the alkaline ceramidase inhibitor D-MAPP (41,42). It has been shown that D-MAPP increases cellular ceramide levels (41,42). Incubation with D-MAPP reduced SREmediated gene transcription and HMG-CoA synthase mRNA levels by half compared with controls ( Figs. 8 and 10). This effect was further enhanced in the presence of oleate (Fig. 8). Thus, increasing endogenous ceramide levels by oleate (Figs. 1  and 2), addition of exogenous sphingomyelinase (Fig. 3), addition of exogenous ceramides (Figs. 4 -7), and inhibition of endogenous ceramide hydrolysis (Fig. 8) all have a similar effect on reducing SRE-mediated gene transcription.
Fatty Acids and Ceramide Decrease SRE Expression during Inhibition of Intracellular Cholesterol Transport-We previously reported that the effects of oleate on decreasing SREmediated gene transcription are dependent on cleavage of precursor SREBP within cells and thus are likely associated with changes in intracellular cholesterol trafficking (9). We next questioned whether the observed effects of ceramide are related to cholesterol transfer within the cell. Cholesterol regulates the activity of SREBP cleavage-activating protein and SREBP translocation to the Golgi apparatus, where a sterolregulated protease initiates the release of the transcriptionally active mature form (4,5). Because cholesterol partitions between the plasma membrane and other compartments within the cell, we tested the relationship of cholesterol movement and ceramide-mediated decreases in SRE expression in the presence of U18666A. U18666A decreases the cholesterol content of the endoplasmic reticulum (43) and impedes the movement of cholesterol between the lysosomal compartment and the plasma membrane and from the plasma membrane to other intracellular compartments (44,45). Incubation with U18666A (2.5 M) increases SRE-mediated reporter gene expression 2-fold and HMG-CoA synthase mRNA levels 1.7-fold, consistent with an inability of cellular cholesterol to reach and inhibit an SREBP cleavage-activating protein regulatory compartment. Addition of ceramide almost completely abolished the increase in SRE-mediated gene transcription induced by U18666A (Fig. 9). Therefore, ceramide-mediated reduction of SRE-mediated gene transcription is in part independent of intracellular cholesterol movement.
We next examined whether the results obtained with SRE reporter gene assays reflect changes in the regulation of HMG-CoA synthase mRNA (Fig. 10), a gene known to be sensitively regulated by the SRE (46). C 2 -ceramide and D-MAPP decreased the amount of HMG-CoA synthase mRNA by half. Exogenous sphingomyelinase reduced the HMG-CoA synthase mRNA levels to 30% compared with controls. Thus, the effects of ceramide on SRE-mediated gene expression are reflected in the transcription of HMG-CoA synthase, a gene recognized to be closely regulated by the SRE (46). DISCUSSION We investigated potential mechanisms by which fatty acids can decrease SRE-mediated gene transcription. Previously, we have shown that fatty acids regulate SRE-mediated gene transcription by decreasing mSREBP in the absence of cholesterol as well as synergistically with cholesterol (9). We now demonstrate that fatty acids can affect SRE-mediated gene transcription by two interdependent mechanisms. Fatty acids can induce sphingomyelin hydrolysis, which can result in the intracellular displacement of membrane cholesterol. Sphingomyelin hydrolysis also results in the generation of ceramide, which can decrease levels of transcriptionally active mSREBP and SRE-mediated gene transcription.
We first investigated the hypothesis that unsaturated fatty acids decrease SRE-mediated gene expression by pathways that would lead to increased intracellular cholesterol levels. Cholesterol and sphingomyelin contents of cellular membranes are unequally distributed within the cell. Most cholesterol and sphingomyelin are found in the plasma membrane, whereas the endoplasmic reticulum contains much smaller amounts (14,47,48). Cholesterol binds to sphingomyelin with higher affinity than to other membrane phospholipids (13)(14)(15). Sphingomyelin hydrolysis forces cholesterol to translocate extracellularly in the presence of cholesterol acceptors (35). In the absence of acceptors, cholesterol is displaced to intracellular compartments (18 -20). It has been shown that unsaturated fatty acids can induce sphingomyelin hydrolysis (16,17). We confirmed that physiological concentrations of unsaturated (but not saturated) fatty acids induce hydrolysis of cellular sphingomyelin (Fig. 1). We demonstrated that cell extracts from cells treated for up to 20 h with oleate induced sphingo- myelinase hydrolysis, whereas cell extracts from cholesteroltreated cells had no effect (Fig. 2). We then showed that exogenous sphingomyelinase resulted in persistently decreased SRE-mediated reporter gene expression and HMG-CoA reductase levels for up to 8 h (Fig. 3). High concentrations of bacterial sphingomyelinase have been demonstrated to reduce cellular levels of mSREBP within 30 min (33).
Sphingomyelin hydrolysis results not only in the release of cholesterol bound to sphingomyelin, but also in the generation of ceramide. Ceramide has multiple cellular roles, from induction of apoptosis to cell growth and differentiation (49). Experimentally, short chain ceramide and dihydroceramide analogs both promote cell growth and differentiation, but only ceramide analogs (not dihydroceramide analogs) can induce apoptosis (39,40). Several lines of evidence indicate that the induction of apoptosis had no role in the regulation of SRE-mediated gene transcription investigated under our experimental conditions. 1) Ceramide analogs did not induce changes in the cell cycle or mitochondrial function under our experimental conditions. 2) The effect of ceramide analogs and dihydroceramide analogs were indistinguishable (Fig. 7) (of note, dihydroceramide analogs have no role in the induction of apoptosis (39)). 3) Contrary to our results, mSREBP levels have been reported to be increased in apoptosis (50), but we observed a decrease. 4) The effect of ceramide analogs on SRE-mediated gene transcription was reversible (Fig. 5). 5) Ceramide analogs had no adverse Cells were harvested, lysed, and analyzed for luciferase activity and protein content. Data are expressed as a percentage of the control and represent the average of two experiments, each performed in triplicate. U18666A significantly increased SRE-mediated luciferase expression compared with controls (p Ͻ 0.05). C 2 -ceramide significantly decreased SRE-mediated luciferase expression compared with controls and U18666A (p Ͻ 0.05). In the presence of C 2 -ceramide and U18666A, SRE-mediated luciferase expression was significantly reduced compared with U18666A alone (p Ͻ 0.05). RLU, relative light units. effect on the transcription of the unregulated ␤-galactosidase reporter gene (Fig. 7).
Scheek et al. (33) have reported that exogenous ceramides have no effect on SREBP proteolysis within a time period of 150 min. Perhaps longer exposure to ceramide may be required because we noted that the ceramide-mediated decrease in SRE-mediated luciferase expression appeared after 4 h (data not shown).
In our previous study (9), we have shown that oleate and other polyunsaturated fatty acids decrease SRE-mediated gene transcription alone as well as synergistically with exogenous cholesterol and concluded that the effect is not dependent on exogenous cholesterol, but is likely related to endogenous cellular cholesterol movement. Our present work links fatty acidinduced sphingomyelin hydrolysis to decreasing sphingomyelin levels coupled with the concomitant generation of ceramide from sphingomyelin hydrolysis. To investigate whether ceramide-mediated effects on SRE-mediated gene transcription depend on cholesterol movement, we used U18666A, a substance well characterized to decrease intracellular cholesterol trafficking and cholesterol content of the endoplasmic reticulum (43,51). We showed that U18666A increased SRE-mediated reporter gene expression and HMG-CoA synthase mRNA levels. Ceramide analogs decreased SRE-mediated gene transcription despite the presence of U18666A (Fig. 9), suggesting that the effects of ceramide on SRE-mediated gene transcription are partly independent of intracellular cholesterol trafficking. Although the exact mechanisms whereby ceramide inhibits SREBP cleavage are not yet defined, it is possible that ceramide affects SREBP proteolysis or intracellular trafficking (52).
Fatty acids are also known to decrease SREBP-1c mRNA transcription by antagonizing ligand-dependent activation of liver-x-receptors and by decreasing the half-lives of hepatic SREBP-1a and SREBP-1c though a cycloheximide-sensitive process (12,53). Both mechanisms result in decreased precur-sor protein levels. Because the levels of precursor SREBP-1 and SREBP-2 were not decreased in the presence of ceramide analogs, this potential regulatory mechanism was not further investigated (but cannot be excluded) in our study.
Sphingomyelin and cholesterol levels are coordinately regulated, and both are important metabolically in a large number of physiological and biochemical pathways. Both are constituents of lipid rafts, and sphingomyelin synthesis and levels determine the capacity of membranes to absorb cholesterol (54). Our results indicate that sphingomyelin hydrolysis contributes to these homeostatic changes by initiating two separate regulatory mechanisms of SREBP proteolysis: a cholesterol-mediated decrease in SRE-mediated gene transcription by decreasing plasma membrane cholesterol and a ceramide-mediated effect that occurs even in the presence of inhibitors of cellular cholesterol movement. Ceramide, a product of sphingomyelin hydrolysis, thus joins cholesterol and fatty acids as a regulator of SRE-mediated gene transcription.