Transcriptional Regulation of Squalene Epoxidase by Sterols and Inhibitors in HeLa Cells*

Regulation of squalene epoxidase (SE) gene expres- sion was studied in comparison with those of 3-hydroxy-3-methylglutaryl-CoA (HMG-CoA) reductase and low density lipoprotein (LDL) receptor. An increased expression of SE mRNA and protein content in mouse L929 cells grown in 10% lipoprotein-deficient fetal bovine serum (LPDS) for 48 h was found by performing immuno- blot and Northern blot analyses when compared with the culture in the presence of fetal bovine serum (FBS). The same results in mRNA levels were seen using human cell lines HepG2, HeLa, and Chang liver cells. The in- crease of SE mRNA in HeLa cells grown in LPDS was preventable in a dose-dependent manner by feeding cells with 25-hydroxycholesterol or cholesterol. When an SE inhibitor, NB-598, was fed to HeLa cells grown in LPDS, it caused further increases in mRNA levels of SE, HMG-CoA reductase, and LDL receptor. In contrast, NB- 598 had no effect on the message levels of these genes when fed to HeLa cells grown in FBS. These results suggest that sterol produced endogenously can also regulate SE expression at the level of transcription. g separated on a 10% SDS-polyacrylamide and transferred electrophoretically onto a nitrocellulose membrane and analyzed immunochemically with anti-recombinant rat SE antiserum. previously immunoblot analysis, were visualized with horseradish perox- idase-conjugated goat anti-rabbit IgG (Life Inc.), using Western blot chemiluminescence reagent (DuPont NEN). Molecular comparison with prestained high range molecular weight markers (Bio-Rad). isolated by guanidine thiocyanate-phenol-chloroform extraction using ISOGEN Kit (Nippon Gene Co.) according to instructions of manufacturer. RNA (cid:109) g) was electrophoresed in a 1% agarose gel containing 2.2 M formaldehyde and transferred to a nytran membrane Schuell, with 3.0 M sodium 0.3 M sodium citrate (20 (cid:51) SSC) transfer buffer. baked (80 °C for h) and incubated for min at 65 °C 0.5 M sodium phosphate (pH 7.2), 7% SDS, 1 m M EDTA (prehybridization [ 32 P]dCTP by random primer method MegaPrime fragments of human SE cDNA C-terminal part number D78129), human LDL recep- tor (1.2-kilobase Eco RI- Eco RV fragment of pLDLR3) human HMG-CoA reductase Bgl II fragment pHRED chain

Since cholesterol is an essential structural component of cytoplasmic membranes, it is crucial for cells to maintain intracellular cholesterol homeostasis. Cells acquire cholesterol both from the LDL 1 receptor-mediated pathway (1) and the biosynthetic pathway from acetyl-CoA (2,3). Brown and Goldstein (4) demonstrated that both pathways are controlled by end product repression by showing that sterol depletion resulted in increased levels of mRNA for the LDL receptor and two sequential enzymes in de novo cholesterol biosynthesis, HMG-CoA synthase, and HMG-CoA reductase. Furthermore, restoration of sterols resulted in decreased mRNA for these genes.
Although HMG-CoA reductase is considered to be the major regulatory enzyme in cholesterol biosynthesis, recent studies revealed that other enzymes involved in cholesterol biosynthe-sis, such as HMG-CoA synthase, farnesyl diphosphate synthase, and squalene synthase, are also regulated by sterols (5,6). HMG-CoA reductase inhibitors are widely used as agents for lowering plasma cholesterol levels. However, recent studies have revealed that mevalonate derived non-sterol metabolite(s), which play important roles in the regulation of normal cellular processes, are synthesized in a post-mevalonate pathway and that HMG-CoA reductase inhibitors cause the depletion of both mevalonate-derived non-sterol metabolite(s) as well as sterols (7,8). Since SE is situated after this branch point in the mevalonate pathway, cholesterol is the only end product for SE. Therefore, SE is considered to be a potential new target enzyme for anti-hyperlipidemic drugs (9,10).
SE is located in the endoplasmic reticulum and catalyzes the conversion of squalene to 2,3(S)-oxidosqualene, when coupled with a component of microsomal electron transport chain, NADPH-cytochrome P-450 reductase. SE seems to be an important rate-limiting enzyme in cholesterol biosynthesis, since it has an extremely low specific activity in comparison with HMG-CoA reductase or squalene synthase in HepG2 cells (11,12) and since supplementation of exogenous cholesterol resulted in the accumulation of labeled squalene from precursor mevalonate in human renal carcinoma cells (13). The activity of rat or human hepatic SE was shown to be regulated by dietary cholesterol or HMG-CoA reductase inhibitors (11,12). However, the regulation of SE protein and mRNA levels has not been directly investigated. We reported previously the isolation of rat and mouse SE cDNAs (14,15). In this report, we examine the regulation of SE transcription by sterols as well as inhibitors of SE and HMG-CoA reductase, and compare this with the regulation, by these agents, of the HMG-CoA reductase and LDL receptor genes.

EXPERIMENTAL PROCEDURES
Materials-[␣-32 P]dCTP (3000 Ci/mmol) was purchased from DuPont NEN. Cholesterol and 25-hydroxycholesterol were purchased from Nakarai Tesque Co. (Tokyo, Japan) and Sigma, respectively. Lovastatin was kindly provided by Merck, and NB-598 was a gift from Dr. T. Kamei of Banyu Pharmaceutical Co. (Tokyo, Japan). Fetal bovine serum (FBS) was purchased from Bioserum (Canterbury, Australia). Lipoproteindeficient fetal bovine serum (LPDS) (d Ͼ 1.215 g/ml) was prepared by ultracentrifugation as described by Goldstein et al. (16). HepG2 and Chang liver cells were gifts from Dr. J. Tashiro of Chiba University School of Medicine and Banyu Pharmaceutical Co., respectively. Cell culture medium was purchased from Nissui Pharmaceutical Co. (Tokyo, Japan).
Cell Culture and Induction-Cells were grown in Dulbecco's modified essential medium (DMEM) containing 10% fetal bovine serum (FBS medium). 5.0 ϫ 10 6 cells were seeded in 100-mm dishes containing 10 ml of FBS medium. L929, HeLa, and Chang liver cells reached confluent monolayer in 24 h, while HepG2 cells reached confluent monolayer in 5 or 6 days. The cells were washed with phosphate-buffered saline twice, and the medium was replaced with 10 ml of FBS medium or DMEM medium containing 10% LPDS medium. As indicated, the medium was supplemented with cholesterol, 25-hydroxycholesterol, lovastatin, NB-598, or ethanol vehicle alone (1%) and cultured for 48 h. * This work was supported by Grants-in-aid 07044234 and 06557135 for Science Research from the Ministry of Education, Science and Culture, Japan. Financial support was also provided by the Mitsui Life Social Welfare Foundation of Japan. 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.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EMBL Data Bank with accession number(s) D78129.
Immunoblot Analysis of SE-Crude extracts from L929 cell microsome fractions were prepared as described previously (14). Twenty g of the extracts were separated on a 10% SDS-polyacrylamide gel and transferred electrophoretically onto a nitrocellulose membrane and analyzed immunochemically with anti-recombinant rat SE antiserum. The antiserum was prepared as described previously (14). For immunoblot analysis, the antibodies were visualized with horseradish peroxidase-conjugated goat anti-rabbit IgG (Life Technologies, Inc.), using Renaissance Western blot chemiluminescence reagent (DuPont NEN). Molecular mass was estimated by comparison with prestained high range molecular weight markers (Bio-Rad).
RNA Preparation and Northern Blot Analysis-RNA was isolated by guanidine thiocyanate-phenol-chloroform extraction using ISOGEN Kit (Nippon Gene Co.) according to instructions of the manufacturer. Total RNA (20 g) was electrophoresed in a 1% agarose gel containing 2.2 M formaldehyde and transferred to a nytran membrane (Schleicher and Schuell, Dassel, Germany) with 3.0 M sodium chloride, 0.3 M sodium citrate (20 ϫ SSC) as the transfer buffer. The membrane was baked (80°C for 2 h) and incubated for 10 min at 65°C in 0.5 M sodium phosphate buffer (pH 7.2), 7% SDS, 1 mM EDTA (prehybridization buffer). Probes were made with [␣-32 P]dCTP by random primer method using a MegaPrime Kit (Amersham Corp.). The fragments of human SE cDNA C-terminal part (accession number D78129), human LDL receptor (1.2-kilobase EcoRI-EcoRV fragment of pLDLR3) (17), human HMG-CoA reductase (2.8-kilobase BglII fragment of pHRED 102) (18), and polymerase chain reaction product (572-1092) of human ␤-actin pseudogene (204pBV2.1) (19) were used as templates. Hybridization was carried out at 65°C in the same solution as prehybridization, except for the addition of probe. The membrane was washed with 40 mM sodium phosphate buffer (pH 7.2), 0.1% SDS at room temperature for 5 min three times, and at 65°C for 60 min. After washing, the membrane was exposed to x-ray film (Fuji RX) with an intensifying screen. The levels of mRNAs were quantified by BAS 1000 system (Fuji Photo Film Co.). Rehybridization of the same blot for ␤-actin probe was carried out, and all mRNA expression levels were normalized by the intensity of the ␤-actin mRNA band.

Effect of FBS and LPDS Media on SE in L929 Cells-To
establish whether LPDS and FBS have effects on the expression of SE, we analyzed the levels of SE protein and mRNA in L929 cells grown in 10% LPDS or FBS. Mouse L929 cells were harvested after incubation in medium with either LPDS or FBS for 48 h. An Western blot analysis of a microsomal extract of the cultured cells, using rabbit anti-rat SE antibodies, is shown in Fig. 1A. A single protein band with apparent molecular mass of 63.8 kDa was increased in the LPDS-cultured cells than in FBS cultured cells. To assess whether an increase of SE mRNA also occurs, Northern blot analysis of L929 cell extracts was performed with a rat SE cDNA and presented in Fig. 1B. To normalize the amount of loaded RNA, we also quantified ␤-actin mRNA. A single 2.8-kilobase SE mRNA was present under both conditions, but its level in LPDS medium was 4.2-fold higher than in FBS medium. These results strongly suggest that the SE gene is regulated by sterols at the transcriptional level.
Effect of FBS and LPDS Media on SE mRNA in Human Cell Lines-To confirm whether the increase of SE mRNA in LPDS was also observed in the human cell lines HepG2, HeLa, and Chang liver cells, Northern blot analysis was performed using a cDNA probe for human SE, as shown in Fig. 2. Human SE mRNA demonstrated one major and one minor transcript in HepG2 cells supplemented with FBS or LPDS and in HeLa cells supplemented with LPDS (data not shown). A pronounced increase of human SE mRNA was obtained in HeLa cells grown in LPDS (Fig. 2). The relative increase of SE mRNA, HepG2, HeLa, and Chang liver cells grown in LPDS was 2.8-, 5.5-, and 1.7-fold, respectively, when compared with growth in FBS medium.
Time Course of the LPDS Effect on SE, HMG-CoA Reductase, and LDL Receptor mRNAs in HeLa Cells-HeLa cells were incubated for the indicated times in LPDS or FBS medium, and SE mRNA levels were determined by Northern blot analysis. The levels of human SE mRNA linearly increased in LPDS medium, reaching 5.4-fold after 48 h of incubation, whereas there was no significant change in SE mRNA levels in FBS medium (data not shown). The same Northern blot was rehybridized with human HMG-CoA reductase and LDL receptor cDNA probes to examine the time course of their expression in LPDS medium. As shown in Fig. 3, the relative amount of mRNA for both HMG-CoA reductase and LDL receptor was coordinately elevated with SE mRNA, although SE mRNA showed the highest level of increase in LPDS medium.
Suppression of SE mRNA Expression by Sterols in LPDScultured HeLa Cells-To confirm whether expression of SE mRNA is regulated by sterol itself, we incubated HeLa cells in LPDS medium in the presence of either cholesterol or 25hydroxycholesterol and then performed Northern blot analysis with SE cDNA probe (Fig. 4A). As shown in Fig. 4B, both cholesterol and 25-hydroxycholesterol suppressed the expression of SE mRNA in a dose-dependent manner, although 25- After confluent monolayers of L929 cells were incubated for 48 h in either FBS or LPDS medium, cells were collected by centrifugation, and crude microsome extracts and total RNAs were prepared from the cells as described under "Experimental Procedures." A, 20 g of crude microsome extracts were subjected to electrophoresis on an SDS-polyacrylamide (10%) gel and transferred to a nitrocellulose filter. The filter was incubated with rabbit anti-recombinant rat SE serum (diluted 1:1000), followed by incubation with horseradish peroxidase-conjugated goat anti-rabbit IgG (Life Technologies, Inc.), and visualized using Renaissance Western blot chemiluminescence reagent (DuPont NEN). The filter was exposed to Fuji RX film for 20 min at room temperature. We used a high range prestained standard as a molecular weight marker (Bio-Rad). B, 20 g of total RNAs from the cells cultured in FBS or LPDS medium were electrophoresed on a 1% agarose gel and transferred to a nytran membrane. The filter was hybridized with 32 Plabeled rat SE cDNA probe and exposed to Fuji RX film with an intensifying screen at Ϫ70°C for 10 days. The same filter was subsequently hybridized with human ␤-actin probe and exposed to film for 4 h.

FIG. 2. Expression of SE mRNA in human cell lines.
Confluent monolayers of HepG2, HeLa, and Chang liver cells were cultured for 48 h in FBS medium or LPDS medium, and then total RNA was isolated from the cells. Twenty g of the total RNA was subjected to Northern blot analysis using a 32 P-labeled human SE cDNA probe. The same filter was subsequently hybridized with human ␤-actin probe. The signals were quantified using BAS 1000 system and normalized to ␤-actin mRNA levels. These relative intensities of SE mRNA signals were then plotted for both culture condition. hydroxycholesterol was much more effective than cholesterol. Fig. 5A (lanes 1 and 6), incubation of HeLa cells in LPDS medium resulted in increased SE mRNA levels. Addition of a specific inhibitor for SE, NB-598 (Fig. 5A, lanes 6 -10) produced a further increase in SE mRNA levels, with maximal increase at 1 M. The relative increase of mRNA for SE, HMG-CoA reductase, and LDL receptor by 1 M NB-598 in LPDS medium was 13.6-, 7.5-, and 3.1-fold, respectively (Fig. 5B). When the increase of these mRNAs in LPDS medium itself was sub-tracted from this effect, 57% of the maximal increase of SE and HMG-CoA reductase mRNAs and 16% of the maximal increase of LDL receptor mRNA could be attributed to NB-598. However, the levels of mRNA for SE, HMG-CoA reductase, and LDL receptor in FBS medium remained unchanged by NB-598. The increase of SE mRNA levels by a specific inhibitor of HMG-CoA reductase, lovastatin, is shown in Fig. 6. Maximal increase was observed at 1 M in LPDS medium (Fig. 6A). After incubation of HeLa cells for 48 h with 1 M lovastatin in LPDS medium, the relative increase of SE, HMG-CoA reductase, and LDL receptor mRNAs was 8.9-, 8.5-, and 2.5-fold, respectively (Fig.  6B). After subtracting the increase by LPDS alone, 62% of the maximal increase of HMG-CoA reductase and 35% of the maximal increase of SE mRNA could be attributed to lovastatin. Lovastatin had no effect on LDL receptor mRNA in LPDS medium. In FBS medium, however, 1 M lovastatin produced a 3-fold increase of HMG-CoA reductase mRNA, but had no effect on the levels of SE or LDL receptor mRNAs. DISCUSSION Intracellular cholesterol homeostasis is maintained primarily through regulation of cholesterol biosynthetic and LDL receptor mediated pathways. Many genes involved in the biosynthetic pathway are coordinately controlled by sterols (4 -6). Previously, Hidaka et al. (11) reported that SE activity is con- FIG. 3. Time course of SE, HMG-CoA reductase, and LDL receptor mRNA increase by LPDS in HeLa cells. After incubation of various duration in LPDS medium, total RNA was prepared from the cells, and Northern blot analysis was performed on 20 g. The filter was hybridized with SE probe. The same membrane was rehybridized with HMG-CoA reductase, LDL receptor, and ␤-actin probes. All mRNA levels in this figure were determined by the BAS 1000 system and normalized to ␤-actin mRNA levels. The average intensity (n ϭ 2 experiments) of the mRNA bands was plotted relative to the value for the mRNA at zero time. SE, squalene epoxidase; RED, HMG-CoA reductase; LDLR, LDL receptor.

FIG. 4. Effects of sterols on SE mRNA expression in HeLa cells.
A, HeLa cells were cultured in LPDS medium with the following additions: ethanol vehicle alone (lane 2); 1, 10, 100, 1000, or 10,000 ng/ml of cholesterol (lanes 3-7); or 25-hydroxycholesterol (lanes 8 -12), respectively. The cells in lane 1 were cultured in FBS medium. After 48-h incubation, total RNA was prepared from the cells, and 20 g of RNA was subjected to Northern blot analysis. The filter was exposed to Fuji RX film with an intensifying screen at Ϫ70°C for 10 days. B, the data in A were quantified and the average intensity (n ϭ 2 experiments) of the SE mRNA band was plotted relative to the value for the SE mRNA band in FBS-cultured cells. All mRNA levels in this figure were determined as in Fig. 3. 25-OH chol., 25-hydroxycholesterol.  5 and 10). After a 48-h incubation, total RNA was prepared from the cells, and 20 g was subjected to Northern blot analysis. The filter was exposed to Fuji RX film with an intensifying screen at Ϫ70°C for 10 days. B, HeLa cells were cultured in FBS medium, FBS medium with 1 M NB-598, or LPDS medium with 1 M NB-598. After a 48-h incubation, total RNA was isolated from the cells. Twenty g of the RNAs were subjected to Northern blot analysis using a human SE cDNA probe, and the same filter was rehybridized with human HMG-CoA reductase and LDL receptor cDNA probes. The data were quantified, and the average intensity (n ϭ 2 experiments) of the bands was plotted relative to the values in the same cells cultured in FBS medium. All mRNA levels in this figure were determined by BAS 1000 system and normalized to ␤-actin mRNA levels determined after rehybridization. SE, squalene epoxidase; RED, HMG-CoA reductase; LDLR, LDL receptor. trolled by endogenous and exogenous sterols. In this report, Western blot and Northern blot analyses of L929 cells (Fig. 1) clearly demonstrate that SE activity is regulated by changes in enzyme levels and that these changes occur mainly at the transcriptional level. In addition, we detected increased SE mRNA levels in the human cell lines HepG2, HeLa, and Chang liver cells when grown in LPDS medium, although the relative rates of increase are quite different in the three cell lines (Fig.  2). The increase of SE mRNA in LPDS medium is significantly greater than either HMG-CoA reductase or LDL receptor mRNAs (Fig. 3). On the basis of these and earlier studies (11)(12)(13) it seems likely that SE can serve as the rate-limiting enzyme in the post-mevalonate cholesterol biosynthetic pathway.
25-Hydroxycholesterol is known to down-regulate HMG-CoA reductase and LDL receptor transcription (4). We demonstrate in Fig. 4 that this same sterol, as well as cholesterol, suppresses the increase in SE mRNA normally observed in LPDS cultured HeLa cells. Although 25-hydroxycholesterol suppressed this increase more effectively than cholesterol, this difference may be due partly to the lower solubility of cholesterol. Since SE is involved in the pathway of synthesizing cholesterol alone after branch point, the fact that inhibition of SE activity by NB-598 causes an increase of SE, HMG-CoA reductase, and LDL receptor mRNA levels (Fig. 5) strongly suggests that the sterol produced endogenously can down-regulate these sterol-sensitive genes. We also observed that SE and LDL receptor mRNA levels are increased to a great extent by NB-598 than lovastatin, but the reverse is true for HMG-CoA reductase mRNA (Figs. 5 and 6). Since we recently demonstrated that purified recombinant SE catalyzes the conversion of squalene epoxide to squalene diepoxide (20), which is thought to be a precursor of physiological oxysterol, epoxycholesterol (21), it seems possible that NB-598 up-regulates transcription of the SE and LDL receptor genes higher than lovastatin by inhibiting oxysterol formation. Since lovastatin inhibits the formation of mevalonate-derived non-sterol metabolite(s) as well as cholesterol, the greater effect of lovastatin on HMG-CoA reductase mRNA levels can be explained if the HMG-CoA reductase gene is the only target of negative feedback regulation by the non-sterol metabolite(s) in the transcriptional level.
Recently, a cDNA for the sterol regulatory element-binding protein 1 was cloned, and this factor has been shown to upregulate both HMG-CoA synthase and LDL receptor genes via octanucleotide sequence termed sterol regulatory element-1 (22,23). Although the coordinated regulation of the mRNAs for LDL receptor, SE, HMG-CoA reductase, and other enzymes involved in cholesterol biosynthesis may be achieved by a common feedback mechanism, no sterol regulatory elements have yet been identified for the SE gene. Therefore, analysis of the SE gene promoter is in progress. The availability of SE expression systems provide a novel and important step toward elucidating the molecular mechanism(s) involved in the regulation of SE transcription by cholesterol.