cDNA cloning of mouse and human cholesterol 25-hydroxylases, polytopic membrane proteins that synthesize a potent oxysterol regulator of lipid metabolism.

Oxysterols regulate the expression of genes involved in cholesterol and lipid metabolism and serve as intermediates in cholesterol catabolism. Among the most potent of regulatory oxysterols is 25-hydroxycholesterol, whose biosynthetic enzyme has not yet been isolated. Here, we report the cloning of cholesterol 25-hydroxylase cDNAs from the mouse and human. The encoded enzymes are polytopic membrane proteins of 298 and 272 amino acids, respectively, which contain clusters of histidine residues that are essential for catalytic activity. Unlike most other sterol hydroxylases, cholesterol 25-hydroxylase is not a cytochrome P450, but rather it is a member of a small family of enzymes that utilize diiron cofactors to catalyze the hydroxylation of hydrophobic substrates. The cholesterol 25-hydroxylase gene lacks introns, and in the human it is located on chromosome 10q23. The murine gene is expressed at low levels in multiple tissues. Expression of cholesterol 25-hydroxylase in transfected cells reduces the biosynthesis of cholesterol from acetate and suppresses the cleavage of sterol regulatory element binding protein-1 and -2. The data suggest that cholesterol 25-hydroxylase has the capacity to play an important role in regulating lipid metabolism by synthesizing a co-repressor that blocks sterol regulatory element binding protein processing and ultimately leads to inhibition of gene transcription.

Oxysterols regulate the expression of genes involved in cholesterol and lipid metabolism and serve as intermediates in cholesterol catabolism. Among the most potent of regulatory oxysterols is 25-hydroxycholesterol, whose biosynthetic enzyme has not yet been isolated. Here, we report the cloning of cholesterol 25-hydroxylase cDNAs from the mouse and human. The encoded enzymes are polytopic membrane proteins of 298 and 272 amino acids, respectively, which contain clusters of histidine residues that are essential for catalytic activity. Unlike most other sterol hydroxylases, cholesterol 25-hydroxylase is not a cytochrome P450, but rather it is a member of a small family of enzymes that utilize diiron cofactors to catalyze the hydroxylation of hydrophobic substrates. The cholesterol 25-hydroxylase gene lacks introns, and in the human it is located on chromosome 10q23. The murine gene is expressed at low levels in multiple tissues. Expression of cholesterol 25-hydroxylase in transfected cells reduces the biosynthesis of cholesterol from acetate and suppresses the cleavage of sterol regulatory element binding protein-1 and -2. The data suggest that cholesterol 25-hydroxylase has the capacity to play an important role in regulating lipid metabolism by synthesizing a co-repressor that blocks sterol regulatory element binding protein processing and ultimately leads to inhibition of gene transcription.
Oxysterols are formed by the hydroxylation of the side chain of cholesterol. 1 This modification renders the sterol more hy-drophilic and confers two important biological properties. First, the increased hydrophilicity enhances the ability of the oxysterol to cross membranes and thereby facilitates its movement between intracellular compartments, cells, and tissues. Second, oxysterols delivered in ethanol to cultured cells are potent regulators of the expression of genes involved in sterol and fatty acid metabolism (1,2).
The enhanced solubility of oxysterols is exploited by the body to maintain cholesterol homeostasis. In several tissues and cell types, including the brain, kidney, endothelium, and macrophages, cholesterol is converted into oxysterols that subsequently traverse the plasma membrane and are transported to the liver (3)(4)(5). In the liver, they are converted into bile acids by a newly described biosynthetic pathway (6). These bile acids are essential for normal lipid and fat-soluble vitamin metabolism (7).
Oxysterols are both positive and negative regulators of gene expression. As positive effectors, they bind to and activate the nuclear receptor LXR (8), which in turn increases transcription of the cholesterol 7␣-hydroxylase gene (9). This activation stimulates the conversion of cholesterol into bile acids (10). Mutation of the LXR gene in mice causes a loss of 7␣-hydroxylase gene induction and a build up of cholesterol in the liver (11). As negative regulators, oxysterols suppress the cleavage of two transcription factors known as sterol regulatory element binding proteins-1 and -2 (SREBP-1 and -2) (12). These proteins are synthesized as inactive precursors in the membrane compartment of the cell. When intracellular cholesterol levels decline, SREBPs are cleaved to release amino-terminal fragments that migrate to the nucleus and activate the transcription of a network of genes involved in cholesterol synthesis and supply (12). This activation in turn restores intracellular cholesterol levels.
The objectives of the current study were to obtain in vivo evidence for 25-hydroxylation of cholesterol and to identify the gene responsible for this putative activity. We here report the isolation of cDNAs encoding murine and human cholesterol 25-hydroxylases, and we demonstrate that expression of these cDNAs in cultured cells down-regulates cholesterol synthesis and SREBP processing.

EXPERIMENTAL PROCEDURES
Expression Cloning-Total RNA was prepared from 400 mg of an SREBP-1a transgenic mouse liver (14) using RNA-Stat 60 (Tel-Test, Inc. Friendswood, TX). Poly(A) ϩ RNA was prepared from total RNA by two cycles of chromatography on oligo(dT) (mRNA Purification Kit, Amersham Pharmacia Biotech). A size-fractionated, directional cDNA library with SalI and NotI cohesive ends at the 5Ј and 3Ј termini, respectively, was constructed from 4 g of poly(A) ϩ RNA using a Superscript Plasmid Kit (Life Technologies, Inc.). Size-fractionated cDNA (Ͼ1.0 kb, 10 ng) was ligated with 50 ng of pCMV6 expression vector (a derivative of pCMV4 (15) containing a NotI site in the polylinker) using a protocol and reagents supplied with the Superscript kit. Prior to ligation, the pCMV6 vector (1.2 g) was digested for 2 h with 10 units of NotI and SalI, respectively, in 30 l of 1ϫ SalI restriction buffer (New England Biolabs, Beverly, MA). The digested plasmid was purified by phenol/chloroform (1:1, v/v) extraction, electrophoresed on a 0.8% agarose gel, and recovered from the gel using a QIAquick Gel Extraction Kit (Qiagen GmbH, Germany).
Plasmid DNA was purified from the ligation reaction by precipitation with ammonium acetate/ethanol and resuspended in 4 l of water, of which 1 l was used to transform 40 l of Electromax Escherichia coli DH10B cells (Life Technologies, Inc.). The transformed bacteria were diluted into 1000 ml of LB medium containing ampicillin. Aliquots of cells were plated on LB ampicillin plates for calculation of the total number of recombinants. The remainder was divided into 400 pools of 2.5 ml each that were grown to saturation overnight at 37°C. The total number of independent recombinants in the cDNA library was 1.5 ϫ 10 6 , and each pool contained an average of 3800 recombinants. DNA was prepared from individual pools using a Wizard Miniprep Kit (Promega Inc, Madison, WI). The yield of plasmid DNA from each pool was approximately 25 g.
Human embryonic kidney 293 cells (ATCC CRL 1573) were plated on day 0 at a density of 7 ϫ 10 5 cells/60-mm dish in Medium A (Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum, 100 units/ml penicillin, and 100 g/ml streptomycin sulfate). On day 1, individual dishes were transfected with a mixture of plasmid DNAs that included pool DNA (5 g), pCMV-StAR (2 g), phct1 (2 g), and pVA-1 (1 g). The expression plasmid pCMV-StAR contains a fulllength cDNA encoding the murine steroidogenic acute regulatory protein (StAR, Ref. 16) and was a kind gift of Dr. Douglas Stocco, Texas Tech University Medical School, Lubbock, TX. The expression plasmid phct1 contains a full-length cDNA encoding the murine oxysterol 7␣hydroxylase enzyme (17,18). The original hct1 cDNA was a kind gift of Dr. Richard Lathe, University of Edinburgh, Edinburgh, Scotland. The plasmid pVA1 contains the adenovirus type 5 VAI gene (19). A positive control, in which cells were transfected with 1 ng of a murine sterol 27-hydroxylase expression plasmid diluted into 5 g of pCMV6 vector alone, was included in every experiment. Aliquots (40 l) of the transfection lipid pfx-8 (Invitrogen, Carlsbad, CA) dissolved in DMEM containing 15 mM HEPES were added to each plasmid mixture. Cells were incubated with the resulting lipid/DNA mixture in 5 ml of DMEM for 4 h at 37°C, in an atmosphere of 8.8% CO 2 . Assay of cholesterol 25-hydroxylase activity was thereafter carried out as described below.
To subdivide the positive pool containing a cholesterol 25-hydroxylase cDNA, an aliquot (40 l) of Electromax DH10B cells was transformed with 0.5 ng of DNA from the positive primary pool. A portion (0.5%) of the transformation mixture was diluted into 50 ml of LB medium containing ampicillin and divided into 20 pools, each containing ϳ490 recombinants for secondary screening. DNA was prepared from individual pools, and 293 cells were transfected as above, except that 6-well plates were used in place of 60-mm dishes, and amounts of reagents were scaled down accordingly. Tertiary screening was similarly performed except subpools of 50 recombinants were transfected. Quaternary screening was carried out with pools of 10 cDNA isolates derived from a matrix array of individual cDNAs to identify a single cholesterol 25-hydroxylase cDNA.
Measurement of Cholesterol 25-Hydroxylase Activity in Whole Cells-The transfection medium containing the cationic lipid was aspirated and replaced with 3 ml of Medium B (DMEM containing 10% newborn calf lipoprotein-poor serum, 100 units/ml penicillin, and 100 g/ml streptomycin sulfate) supplemented with 5 l of [4-14 C]cholesterol (56.6 mCi/mmol; 0.040 Ci/l; NEN Life Science Products). Cells were incubated for a further 60 h at 37°C in an atmosphere of 8.8% CO 2 .
Media from the transfected cells were collected and extracted with 8 ml of chloroform/methanol (2:1, v/v). The organic phase from each sample was taken to dryness under a stream of nitrogen, and residues were dissolved in 40-l aliquots of chloroform/methanol (2:1, v/v) and applied to 20 ϫ 20 cm prescored LK5DF silica gel TLC plates (Whatman, Hillsboro, OR) with preadsorbent layers. The plates were developed in ethyl acetate/toluene (4:6, v/v) and exposed to a Fuji BAS-MP phosphorimager plate overnight. Phosphorimage analysis was then performed on a Fuji BAS1000 apparatus.
Isolation of Human Cholesterol 25-Hydroxylase cDNA-A 372-base pair expressed sequence tag (EST, GenBank TM accession number W01328) with high sequence identity to a portion of the murine 25hydroxylase cDNA was identified by BLAST search. A bacterial stab transformed with a plasmid containing the EST sequence cloned into the pT3T7 vector was obtained from Research Genetics, Inc., Huntsville, AL. Plasmid DNA was prepared, and a 245-base pair fragment was amplified by the polymerase chain reaction using oligonucleotide primers of the following sequence: 5Ј-CTGGGACCACCTGAGGAGC-T-3Ј (forward primer) and 5Ј-GCCCAATGCAGCAGCGTCAC-3Ј (reverse primer). The thermocycler program consisted of 35 cycles of 94°C for 30 s, 60°C for 30 s, and 72°C for 30 s. The amplified cDNA fragment was cloned into pGEM-T Easy (Promega Corp., Madison, WI). The insert was excised from the plasmid with EcoRI and used for the preparation of a radiolabeled probe by random octamer priming (Megaprime Labeling Kit, Amersham Pharmacia Biotech). The probe was used to screen 200,000 plaques of a human lung cDNA library in bacteriophage gt10 (catalog number HL3004a, CLONTECH, Palo Alto, CA) using standard hybridization procedures (20). One positive clone was isolated whose cDNA insert was subcloned into the EcoRI sites of pBluescript SK ϩ (Stratagene Corp., La Jolla, CA) and pCMV6, yielding plasmids pBS-h25 and pCMV-h25, respectively.
Gene Mapping-Cholesterol 25-hydroxylase gene sequences were isolated from a murine genomic library prepared from 129SvEv DNA (a kind gift of Dr. Alan Bradley, Baylor College of Medicine, Houston, TX) and a human genomic library (catalog number 946204, Stratagene), both in bacteriophage FIX II, by screening with full-length cDNA probes corresponding to the murine and human cholesterol 25-hydroxylase cDNAs, respectively, using standard protocols (20). Approximately 600,000 murine and 400,000 human recombinants were screened, and one positive clone from each library was identified and purified to homogeneity. The corresponding genomic DNA inserts were excised from the bacteriophage vectors and ligated into the NotI site of pBluescript SK ϩ , yielding plasmids pBS-mg25 and pBS-hg25, respectively.
The chromosomal location of the human cholesterol 25-hydroxylase gene was determined by fluorescent in situ hybridization (FISH) and by polymerase chain reaction amplification of somatic cell and radiation hybrid panel DNAs. FISH mapping was performed by See DNA Biotech, Inc. (Downsview, Ontario, Canada). The bacteriophage clone harboring the human 25-hydroxylase gene described above was labeled with biotinylated dATP for use as a FISH probe. Of 100 mitotic figures analyzed, 91 showed hybridization signals on paired sister chromatids corresponding to chromosome 10. Comparison of the signal positions with bands generated by staining with 4,6-diamidino-2-phenylindole indicated that hybridization occurred at band q23. Radiation and somatic cell hybrid mapping were performed using DNAs in the Somatic Cell Hybrid Mapping Panel 2 (Coriell Institute of Medical Research, Camden, NJ) and the Stanford G-3 radiation hybrid panel (Research Genetics, Huntsville, AL). The primer pair used for amplification was 5Ј-CTGGGACCACCTGAGGAGCT-3Ј (forward primer) and 5Ј-GCCCA-ATGCAGCAGCGTCAC-3Ј (reverse primer), which, respectively, correspond to nucleotides 79 -98 and 333-314 of the human gene sequence (Fig. 6A). The thermocycler program consisted of 35 cycles of 94°C for 15 s and 68°C for 30 s on a Perkin-Elmer GeneAmp 9600 machine. Only somatic cell hybrid DNAs containing human chromosome 10 produced a positive amplification signal. Analysis of the radiation hybrid data through the Stanford Genome Center server 2 indicated linkage of the cholesterol 25-hydroxylase gene to the SHGC-15188 marker (LOD score ϭ 4.4, cR_1000 ϭ 45.76) on chromosome 10 in the vicinity of band q23.
DNA Sequencing and RNA Blotting-DNA sequencing was performed on an ABI Prism 377 sequencer using thermocycler sequencing protocols and fluorescent dye terminators. Contiguous DNA sequences were assembled using MacVector software (IBI-Kodak Corp., New Haven, CT), and sequence alignments were generated using a Lasergene software package (DNASTAR, Inc., Madison, WI).
For RNA blotting, a murine multiple tissue RNA blot (CLONTECH, catalog number 7762-1) was hybridized overnight in 50% formamide hybridization buffer at 42°C with a full-length murine 25-hydroxylase cDNA probe using standard procedures (20). The probe was radioactively labeled by random nonamer priming with [ 32 P]CTP. The blot was washed stringently at 65°C, in 0.1ϫ SSC containing 0.1% (w/v) SDS before exposure for 5 days to Kodak X-OMAT AR film at Ϫ80°C using an intensifying screen.
Antibodies-An antipeptide antibody against the sequence RRYKI-HPDFSPSVKQ, representing amino acids 69 -83 of the murine cholesterol 25-hydroxylase (Fig. 3A), was raised in rabbits. This sequence was synthesized as a multiple antigen peptide by Bio-Synthesis, Inc. (Lewisville, TX). For the initial immunization, 100 g of peptide was administered intramuscularly as a dispersion in Freund's complete adjuvant to two New Zealand White male rabbits, 3 months of age. Boosts of 100 g of antigen in Freund's incomplete adjuvant were given on average every 5 weeks, and bleeds were drawn 7 days after each boost. One of the two resulting antisera, U104, was used here after affinity purification on peptide antigen columns (21).
Epitope Tagging-To construct an epitope-tagged version of the murine cholesterol 25-hydroxylase enzyme, a cDNA fragment spanning the coding region and having BspDI and XbaI restriction sites at the 5Ј and 3Ј ends, respectively, was amplified by the polymerase chain reaction using the oligonucleotide primers 5Ј-AAAATCGATGGCTGCTACAAC-GGTTCGGA-3Ј (forward primer) and 5Ј-AAATCTAGATCAAGTCTGT-TTCTTCTTCTGGATCA-3Ј (reverse primer). The template was the plasmid pCMV-m25, and the thermocycler program consisted of 35 cycles of 94°C at 30 s, 57°C at 15 s, and 72°C at 60 s. The amplified cDNA fragment was purified on a Centricon-100 column (Amicon Corp., Beverly, MA), digested with BspDI and XbaI, repurified by chromatography on a Centricon-100 column, and ligated into a modified pcDNA3 vector containing a sequence for a c-Myc epitope in front of the BspDI site. The desired recombinant was termed pcDNA3-NH 2 -myc-m25. This plasmid encodes a fusion protein in which the sequence MEQKLI-SEEDLNEQKLISEEDLNID containing two tandem copies of a c-Myc epitope is linked to the amino terminus of the murine cholesterol 25-hydroxylase protein lacking the initial methionine residue. A similar strategy was used to place tandem c-Myc epitopes at the amino terminus of the human cholesterol 25-hydroxylase cDNA, producing the plasmid pcDNA3-NH 2 -myc-h25.
A double c-Myc epitope (EQKLISEEDLNEQKLISEEDLN) was placed at the carboxyl terminus of the murine cholesterol 25-hydroxylase cDNA as follows. An oligonucleotide encoding these epitopes with an RsrII cohesive 5Ј-end and a blunt 3Ј-end was formed by annealing the phosphorylated oligonucleotide primers 5Ј-GTCCGCGGAGCAAAA-GCTCATTTCTGAAGAGGACTTGAATGAGCAAAAGCTCATTTCTGA-AGAGGACTTGAATTAG-3Ј and 5Ј-CTAATTCAAGTCCTCTTCAGAA-ATGAGCTTTTGCTCATTCAAGTCCTCTTCAGAAATGAGCTTTTGC-TCCGCGG-3Ј as described (22). The annealed duplex was ligated into pCMV-m25 that had been digested with Eco47III and RsrII and purified by agarose gel electrophoresis. The resulting plasmid, pCMV-m25-COOH-myc, encodes a protein comprising amino acids 1-267 of the murine cholesterol 25-hydroxylase fused to the two c-Myc epitope sequences.
Mutagenesis-Site-directed mutagenesis (23) was carried out using a polymerase chain reaction-based kit (Quik-Change, Stratagene) on the plasmid pCMV-m25. The mutagenic oligonucleotide primers (5Ј-GGCT-CACCACGACATGCAACAATCTCAGTTTAACTGC-3Ј and 5Ј-GCAGT-TAAACTGAGATTGTTGCATGTCGTGGTGAGCC-3Ј) were designed to convert histidine codons at positions 242 and 243 of the murine protein to glutamine codons. Mutagenesis was carried out according to instructions provided by the manufacturer. Plasmid DNA products were subjected to DNA sequence analysis to confirm the presence of the substitution mutations and the absence of spurious mutations. The plasmid containing the desired mutations was named pCMV-m25-HH242QQ.
Measurement of Cholesterol 25-Hydroxylase Activity in Cell Lysates-On day 0, a derivative of Chinese hamster ovarian cells expressing the polyoma virus middle T antigen (31) (CHOP cells) was plated at a density of 750,000 cells/100-mm dish in Medium C (1:1 (v/v) DMEM/ Ham's F12 medium containing 5% fetal calf serum, 100 units/ml penicillin, and 100 g/ml streptomycin sulfate). On day 1, the cells were transfected with 1.5 g of pVA-1 and 13.5 g of pCMV6, pCMV-m25-HH242QQ, or pCMV-m25 per dish for vector, mutant, and wild type 25-hydroxylase transfections, respectively. 60 l of pfx-8 lipid was used as a transfection reagent as described above. On day 2, cells were incubated for 1 h with 2% (w/v) 2-hydroxypropyl-␤-cyclodextrin dissolved in a 1:1 solution of DMEM/Ham's F12 medium. The cells were washed once with ice-cold PBS and then harvested in the same buffer using a rubber policeman. After centrifugation at 1000 ϫ g for 5 min, the buffer was aspirated, and the cell pellet was resuspended in 1 ml of 50 mM potassium phosphate buffer, pH 7.4, containing protease inhibitors (Boehringer Mannheim Complete Mini, EDTA-free, at the concentration recommended by the supplier). A cell lysate was prepared using a Polytron set at 10,000 rpm, with three bursts of 3 s each with 30-s intervals between bursts. Incubations were performed at 37°C with 140 g of cell lysate protein in 50 mM potassium phosphate buffer, pH 7.4, containing 5 mM NADPH. [4-14 C]Cholesterol was added in 4 l of 45% (w/v) 2-hydroxypropylcyclodextrin in water to a final concentration of 5 M. The total volume of the incubation was adjusted to 200 l. After 2 h, reactions were extracted with chloroform/methanol (2:1, v/v) and analyzed by thin layer chromatography.
SREBP Cleavage Assay-Stock cultures of CHO-7 cells, a subline of CHO-K1 cells selected for growth in lipoprotein-deficient serum (25), were maintained in Medium D (1:1 (v/v) DMEM/Ham's F12 medium containing 5% newborn calf lipoprotein-poor serum, 100 units/ml penicillin, and 100 g/ml streptomycin sulfate). On day 0, cells were plated at a density of 7 ϫ 10 5 cells/60-mm dish. On day 1, transfections were carried out using 4 g/dish of the indicated plasmid DNA and Lipo-fectAMINE reagent (Life Technologies, Inc.) according to the manufacturer's instructions and with modifications as described (26). After transfection, fresh media supplemented with 50 M mevalonate, 50 M compactin, and 0.2% ethanol containing either no sterol or a mixture of sterols (final concentrations of 1 g/ml 25-hydroxycholesterol and 10 g/ml cholesterol) were added. The cells were returned for 20 h to a 37°C incubator, harvested, and then fractionated into a nuclear extract and a 10 5 ϫ g membrane pellet as described (27). Immunoblot analyses of SREBP-1 and -2 proteins were performed using a SuperSignal Substrate kit (Pierce) and the murine monoclonal antibodies IgG-2A4 and IgG-7D4 (28,29).
Gas Chromatography-Mass Spectrometry-Six-well plates of CHOP cells were transfected with pCMV-m25 or with vector alone as described above and incubated for 48 h in Medium D supplemented with 10 g/ml cholesterol. Thereafter, media were extracted with chloroform/methanol (2:1, v/v; 5 ml/well), and the organic phase was separated and taken to dryness under a stream of nitrogen. Extracts from 6 wells were combined for subsequent procedures. The samples were purified on Isolute Silica columns (International Sorbent Technology, Mid Glamorgan, UK), and hydroxyl groups were converted to trimethylsilyl ethers as described previously (13).
Gas chromatography-mass spectrometry was performed on a Varian 3400 gas chromatograph equipped with an HP-5MS capillary column (30 m ϫ 0.25 mm, 0.25-m phase thickness) connected to a Finnigan SSQ700 mass spectrometer. The gas chromatography temperature program was 180°C for 1 min, followed by a temperature gradient of 10°C/min to 300°C, and a final elution at 300°C for 15 min. Helium was used as the carrier gas at an injector valve pressure of 6 pounds/ square inch. Injector and transfer line temperatures were set to 280°C, and the injector was operated in the splitless mode. The machine was operated in the electron ionization mode with electron energy set to 70 eV, and the quadrupole was scanned between m/z 100 and 500 at a rate of 1 scan/1.5 s.
Analysis of N-Linked Carbohydrates-To examine the sensitivity of N-linked carbohydrates on cholesterol 25-hydroxylase to endoglycosidase digestion, COS M6 cells were initially plated at a density of 5 ϫ 10 5 cells/60-mm dish in Medium A on Day 0 of the experiment. On Day 1, one dish each was transfected with 4.5 g of pCMV-m25, pcDNA3-NH 2myc-m25, pCMV-m25-COOH-myc, pCMV-h25, or pcDNA3-NH 2 -myc-h25 together with 0.5 g of pVA1 and 20 l of pfx-8 lipid as described above. After transfection, cells were cultured in Medium B. On day 2, cells were harvested using a rubber policeman, pelleted at 1000 ϫ g, washed with 1 ml of phosphate-buffered saline, pH 7.4, and resuspended in 0.2 ml of a buffer containing 10 mM Tris-Cl, pH 8.0, and 1 mM EDTA. Cells were lysed by 20 passages through a 22-gauge needle, and aliquots of the lysates (10 l, ϳ40 g of protein) were treated with endoglycosidase H or peptide N-glycosidase F overnight in a volume of 30 l according to the instructions given by the supplier (New England Biolabs, Beverly, MA). One volume of 2ϫ Laemmli gel loading buffer was then added to each sample, followed by incubation at 100°C for 10 min and electrophoresis through a 12% SDS-polyacrylamide gel for 16 h at constant current (10 mA). Separated proteins were electroblotted to polyvinylidene difluoride membranes (30), which were incubated with affinity purified (21) antipeptide antibody directed against cholesterol 25-hydroxylase (U-104, see above) at 0.8 g/ml. A goat anti-rabbit horseradish peroxidase-conjugated antibody (Amersham Pharmacia Biotech) was used as secondary antibody, and visualization was via an ECL Plus kit (Amersham Pharmacia Biotech).
Cytochemistry-For indirect immunocytochemistry, COS M6 cells were plated at a density of 4 ϫ 10 4 cells per well on glass coverslips placed in 6-well dishes containing Medium B. On Day 1, cells were transfected with either a vector alone control (pCMV6) or with pCMV-m25-COOH-myc-m25 or pcDNA3-NH 2 -myc-m25. Three g of plasmid DNA and 12 l of pfx-8 lipid were used per well. After transfection, cells were cultured in Medium B. Indirect immunocytochemistry was then performed with the indicated antibody and lectin probes as follows. Cells were fixed for 30 min with 3% (w/v) paraformaldehyde in Hanks' balanced salt solution, pH 7.4. Following fixation, the coverslips were briefly rinsed with PBS (0.1 M phosphate buffer, pH 7.4, 0.15 M NaCl), and free aldehyde groups were quenched by incubation in PBS containing 50 mM NH 4 Cl for 30 min. Permeabilization was accomplished by incubation in 0.1% (v/v) Triton X-100 in H 2 O for 7 min on ice, followed by rinsing in PBS containing 1% (w/v) bovine serum albumin (blocking buffer) for 30 min at room temperature. Coverslips were incubated with rabbit anti-c-Myc IgG (Upstate Biotechnology Inc., 10 g/ml in blocking buffer) for 2 h at room temperature. Finally, coverslips were incubated with fluorescein isothiocyanate goat anti-rabbit IgG (Zymed Laboratories Inc.; 20 g/ml in blocking buffer) for 1 h at room temperature. For Golgi compartment staining, rhodamine-labeled wheat germ agglutinin was added during the second antibody incubation at a concentration of 1.25 g/ml. Coverslips were washed three times with PBS containing 0.1% bovine serum albumin after each antibody or lectin incubation. Cells were photographed using a Zeiss Photomicroscope.
Inhibitor Studies-Transfection of CHOP cells was carried out as described above with the following exceptions. Cells were plated on Day 0 at a density of 150,000 cells/well in 6-well plates containing Medium C. On Day 1, cells were transfected with 2. Stable Cell Lines-EcR-CHO cells (Invitrogen), a cell line stably expressing the subunits of the Drosophila ecdysone receptor, RXR and VgECR (31), were plated on day 0 at a density of 500,000 cells/100-mm dish in Medium C supplemented with 250 g/ml Zeocin. On day 1, cells were transfected with 5 g of pIND-m25 using 15 l of Fugene 6 (Boehringer Mannheim) according to the instructions of the manufacturer. The plasmid pIND-m25 was constructed by insertion of a fulllength mouse 25-hydroxylase cDNA fragment from plasmid pCMV-m25 into the pIND vector (Invitrogen). On day 2, cells were split 1:15 and plated in fresh Medium C supplemented with 250 g/ml Zeocin and 700 g/ml geneticin. Cells were refed this medium every 2nd day, and on day 10, groups of five geneticin-resistant colonies were replated in individual wells. Following expansion, cells were tested for cholesterol 25-hydroxylase expression by addition of ponasterone (Invitrogen) to a final concentration of 5 M (16 h), followed by immunoblot analysis of total cell protein. Antibody U104 (affinity purified) was used to detect expression of the enzyme as described above. One positive group of cells was selected and subcloned through one additional round to ensure clonality. One of the resulting cell strains, designated TR3102a, which manifests high level expression of cholesterol 25-hydroxylase upon ponasterone induction, and another, designated TR3102g, with no detectable inducible enzyme expression, were selected and maintained as lines. For routine induction experiments, 10 M ponasterone was used for the indicated periods.
Cholesterol Biosynthesis-TR3102a and TR3102g cells were plated on day 0 at a density of 250,000 cells/60-mm dish in Medium C containing Zeocin and geneticin as above. On day 1, the medium was changed to Medium F (Medium D containing 250 g/ml Zeocin, 700 g/ml geneticin, and 10 M ponasterone). On day 2, 20 l of an aqueous solution containing 15 Ci of [1,2-14 C]acetate (American Radiolabeled Chemicals), adjusted with cold acetate to a final mass of 1 mol, was added to each dish. The additions were made in a staggered fashion so that all cells were harvested at the same time, corresponding to incubation times of 2, 4, and 6 h, respectively. The total time of induction with ponasterone was 27 h, including the acetate labeling period. Nonsaponifiable lipids were isolated and analyzed by thin layer chromatography as described (32) except that 5 Ci of [26, H]25-hydroxycholesterol (NEN Life Science Products) was used as a standard. Quantification of acetate incorporation into cholesterol was via phosphorimage analysis.

RESULTS
The livers of transgenic mice overexpressing the transcription factor SREBP-1a accumulate large quantities of cholesterol and triglycerides, owing to the overproduction of lipid synthesizing enzymes (14). An analysis of stool lipids by mass spectrometry revealed that these animals also excrete high levels of several oxysterols, including 25-hydroxycholesterol. To isolate cDNAs that encode putative oxysterol synthesizing enzymes from the livers of these transgenic mice, an expression cloning strategy in cultured mammalian cells was conceived and optimized. The basic premise of the screen was to transfect cells with pools of hepatic cDNAs cloned into an expression vector, add [ 14 C]cholesterol to the medium, and then measure the conversion of this substrate into oxysterols by thin layer chromatography assay. Initially, we used a previously isolated sterol 27-hydroxylase cDNA, whose encoded enzyme converts cholesterol into the oxysterol 27-hydroxycholesterol (15), to optimize assay parameters.
Chromatography studies with oxysterol standards revealed that the separation between cholesterol and some oxysterols was poor on silica gel plates. Furthermore, in control transfection studies with the sterol 27-hydroxylase expression vector, the strong phosphorimage signal from the substrate often obscured a weaker product signal. To overcome these problems, a cDNA encoding a murine oxysterol 7␣-hydroxylase (18)  conversion of oxysterol products to their 7␣-hydroxylated forms. The oxysterol 7␣-hydroxylase also possesses a minor 2-hydroxylase activity against 7␣-hydroxylated sterols (18); thus the formation of 2,7␣-hydroxylated oxysterols was expected. Both of these classes of hydroxylated oxysterols were readily separated from cholesterol by thin layer chromatography. The optimum transfection host (293 cells), transfection method (lipofection with pfx-8 lipid), and transient expression time (60 h) were determined. Further experiments confirmed that the sterol 27-hydroxylase enzyme, which is located in the mitochondria (15), was stimulated 2-3-fold when a cDNA encoding the murine steroidogenic acute regulatory protein was cotransfected into cells (16,33). Finally, addition of the adenovirus VA1 gene, which enhances the translation of mRNAs transcribed from transfected plasmids (19), to the DNA mixture stimulated expression levels another 1.5-fold. Under these optimized conditions, sterol 27-hydroxylase enzyme activity could be detected over background when the cDNA expression vector was diluted 3,000 -5,000-fold.
A library consisting of 1.5 ϫ 10 6 individual cDNAs was next constructed in a pCMV6 vector using poly(A) ϩ mRNA isolated from an SREBP-1a transgenic mouse liver. Two hundred fiftyfive aliquots of ϳ3,800 individual plasmids from the library were screened using the optimized parameters described above. The data of Fig. 1 show the results of a transfection experiment that revealed several positive cDNA pools. The pool analyzed in lane 1 did not contain a cDNA encoding a cholesterol-metabolizing enzyme. However, the pools analyzed in lanes 2-5 produced low to very low levels of two sterols that migrated more slowly and thus were more hydrophilic than the cholesterol substrate. Additional experiments revealed that the pools analyzed in lanes 2-4 contained sterol 27-hydroxylase cDNAs, whereas that analyzed in lane 5 contained a cholesterol 25-hydroxylase cDNA (data not shown).
The pool containing the 25-hydroxylase cDNA was progressively subdivided and expressed to isolate a single cDNA. As the purity of the cDNA increased, the level of product generated in the transfected cells also increased to the point that cotransfection of the oxysterol 7␣-hydroxylase cDNA was dispensable. Additional experiments revealed that the cDNA-encoded enzyme was not stimulated by inclusion of the steroidogenic acute activator cDNA, suggesting that it was not a mitochondrial protein. Transfection of the pure cDNA into CHOP cells produced abundant 25-hydroxylase enzyme activity that increased with time of incubation ( Fig. 2A). The activity was stimulated approximately 10-fold by treatment of transfected cells with 2-hydroxypropyl-␤-cyclodextrin ( Fig. 2A). This compound presumably removes endogenous cholesterol from the membranes of the transfected cells that otherwise competes with the exogenously added radiolabeled cholesterol substrate (34).
The chemical structure of the oxysterol produced by the isolated cDNA was determined by gas chromatography-electron ionization mass spectrometry (Fig. 2B). The media from cells transfected with the putative cholesterol 25-hydroxylase cDNA contained a prominent sterol eluting at 23.56 min from the gas chromatography column. This sterol was not present in the media of mock-transfected cells (Fig. 2B, left panels). The mass spectrum of the cDNA-generated product was virtually identical to that of an authentic 25-hydroxycholesterol standard (Fig. 2B, right panels).
A search of the DNA data bases revealed a human EST with identity to the murine 25-hydroxylase cDNA. This EST was used to isolate a near full-length cDNA encoding the human 25-hydroxylase as described under "Experimental Procedures." Transfection into 293 cells of the human cDNA cloned in a pCMV6 vector produced abundant 25-hydroxylase enzyme activity ( Fig. 2A). This activity was stimulated approximately 5-fold by treatment of cells with 2-hydroxypropyl-␤-cyclodextrin ( Fig. 2A). Fig. 3A shows an alignment of the deduced amino acid sequences of the murine and human cholesterol 25-hydroxylases. The two proteins share 78% sequence identity (Fig. 3A), whereas the encoding cDNAs are 82% identical in their translated regions. The predicated molecular weights of the murine and human enzymes are 34,700 and 31,700, respectively. The most notable difference between the two is a 26-amino acid extension at the carboxyl terminus of the murine enzyme that is not present in the human enzyme (Fig. 3A). Both proteins have three clusters of conserved histidine residues (amino acids 143-147, 157-161, and 238 -243 in the murine sequence, Fig. 3A). Similar clusters of histidine residues are present in a Pseudomonas alkane hydroxylase and xylene monooxygenase (35,36), the eukaryotic stearoyl-CoA desaturases (37), and the yeast and human C-4 sterol methyl oxidases (38,39). These enzymes are members of a family of proteins that utilize diiron cofactors to catalyze diverse reactions on hydrophobic substrates. Hydropathy analyses of the 25-hydroxylase sequences revealed four conserved regions of extended hydrophobicity that may constitute as many as eight transmembrane domains (Fig. 3B). The location of the first, second, and fourth of these hydrophobic regions coincided with similar sequences in the Pseudomonas alkane hydroxylase, which contains six transmembrane domains (40).
To determine if the clustered histidine residues in choles-terol 25-hydroxylase were important for enzyme activity, a pair of histidine codons at positions 242 and 243 in the murine protein were changed to glutamine codons by site-directed mutagenesis of the wild type cDNA. The resulting mutant cDNA was transfected into CHOP cells and assayed for expression of the protein and for cholesterol 25-hydroxylase enzyme activity (Fig. 4A). Mutation of the two histidine residues had no effect on steady state expression levels as judged by immunoblotting (Fig. 4A, left panel) but eliminated enzyme activity in transfected cells (Fig. 4A, right panel). Similar results were obtained in a second experiment in which cholesterol 25-hydroxylase enzyme activity was measured in cell lysates rather than in intact cells (Fig. 4B). The subcellular localization of cholesterol 25-hydroxylase was assessed in two ways. First, the presence and structure of asparagine-linked carbohydrates were analyzed by endoglycosidase digestion (Fig. 5A). Expression of murine or human cDNAs in COS cells produced two forms of the enzyme that differed in mass by approximately 3 kDa as judged by immunoblotting (lanes 1, 4, and 7). When total membrane proteins from transfected cells were digested with endoglycosidase H (lanes 2, 5 and 8) or peptide:N-glycosidase F (lanes 3, 6, and 9), only a single form of cholesterol 25-hydroxylase was detected that migrated with the lower molecular mass enzyme of untreated cells. These data suggested that cholesterol 25-hydroxylase was present in the membrane fraction of the cell and that some of the molecules were glycosylated with high mannose, asparagine-linked carbohydrates. In agreement with these results, the sequence of the murine enzyme contains two potential sites for N-linked glycosylation (amino acids 5 and 163, Fig.  3A), and the human enzyme sequence contains three such sites (amino acids 5, 163, and 189, Fig. 3A).
The second approach to examine the subcellular location of . Predicted hydrophobic regions of the protein sequences fall below the midlines, whereas hydrophilic sequences rise above the midlines. MacVector software was used to generate the data. The window size in the scan was 12 amino acids. cholesterol 25-hydroxylase employed immunocytochemistry. Simian COS cells were transfected with the plasmid pCMV-m25-COOH-myc, which encodes a carboxyl-terminal, c-Myc epitope-tagged version of the murine cholesterol 25-hydroxylase. After a transient expression period, cells were permeabilized and recombinant cholesterol 25-hydroxylase detected by indirect immunocytochemistry using a fluorescein-labeled secondary antibody. At the same time, the transfected cells were stained with rhodamine-labeled wheat germ agglutinin (Sigma, catalog number L5266), a lectin that binds to glycoproteins concentrated in the Golgi compartment. As shown in Fig. 5B, upper panel, fluorescein signal (green) representing cholesterol 25-hydroxylase was detected in the endoplasmic reticulum and a perinuclear compartment of transfected cells. The perinuclear compartment was tentatively identified as the Golgi apparatus based on colocalization with the wheat germ agglutinin lectin (lower panel, orange rhodamine signal). Similar results were obtained when the c-Myc epitope was placed at the amino terminus of the expressed murine enzyme (data not shown).
We next isolated the murine and human cholesterol 25hydroxylase genes. DNA sequence analysis of the isolated genomic DNAs and comparison to the respective cDNA sequences revealed that both genes lacked introns (Fig. 6A). This structure was confirmed by Southern blotting analyses of murine and human DNA and by direct amplification of the genes from genomic DNA (data not shown). Transfection into 293 cells of a plasmid containing the genomic DNA insert from bacteriophage clone that encompassed ϳ10 kb of 5Ј-flanking DNA and ϳ3 kb of 3Ј-flanking DNA of the murine 25-hydroxylase gene resulted in the expression of enzyme activity, which suggested that the isolated gene was not a pseudogene and that requisite regulatory sequences were located close to the coding region (data not shown). The human gene was localized to chromosome 10q23.3 by somatic and radiation hybrid DNA panel mapping and fluorescence in situ hybridization (Fig. 6B).
The tissue distribution of the murine cholesterol 25-hydroxylase mRNA was assessed by blot hybridization (Fig. 7). Low levels of a 1.5-kb mRNA were present in the heart, lung, and kidney. The mRNA was not detected in the livers of control mice (Fig. 7); however, it was present in RNA from the liver of the SREBP-1a transgenic mouse used to prepare the original cDNA expression library (data not shown). RNA blotting experiments using commercially available filters revealed only very low levels of human cholesterol 25-hydroxylase mRNA in 16 different tissues (data not shown).
The potent regulatory effects of 25-hydroxycholesterol were first observed in assays that measured the suppressive effects of oxysterols on cholesterol synthesis (1,2). To determine if the 25-hydroxycholesterol synthesized by the 25-hydroxylase enzyme could suppress cholesterol synthesis, a line of CHO cells containing an ecdysone-inducible 25-hydroxylase cDNA was isolated as described under "Experimental Procedures." These cells, and a control cell line that did not contain the 25-hydroxylase cDNA, were induced with the ecdysone analog ponasterone, and the incorporation of [ 14 C]acetate into cholesterol was measured as a function of time. As shown in Fig. 8A, induction with ecdysone led to a marked reduction of cholesterol synthesis in cells containing the 25-hydroxylase cDNA but had no effect on this parameter in the control cells.
The experiments shown in Fig. 8B were carried out to determine if expression of cholesterol 25-hydroxylase in transfected cells affected the processing of SREBP transcription factors. Cultured CHO-7 cells were transiently transfected with either vector alone or an expression vector containing the murine 25-hydroxylase cDNA. After 24 h, fractions enriched in membrane or nuclear proteins were prepared from the transfected cells. Equal amounts of protein from each subcellular compartment were separated by gel electrophoresis, and the levels of SREBP-1 and -2 were determined by immunoblotting. Mocktransfected cells grown in the absence of sterols to induce SREBP-1 cleavage contained intact, uncleaved SREBP-1 in the membrane fraction and cleaved SREBP-1 in the nuclear fraction (Fig. 8B, lane 1). Mock-transfected cells grown in the presence of sterols (cholesterol plus 25-hydroxycholesterol) contained a majority of the immunodetectable SREBP-1 in the membrane fraction (lane 2). Cells transfected with the 25hydroxylase cDNA and grown in the absence of sterols contained a majority of SREBP-1 in the membrane fraction even though no exogenous sterols were added (lane 3), presumably because the 25-hydroxycholesterol produced by the expressed 25-hydroxylase suppressed cleavage of the transcription factor. Similar results were obtained when the processing of SREBP-2 was followed by subcellular fractionation and immunoblotting (Fig. 8B, right panels, lanes 4 -6).
We next tested the ability of different sterols to inhibit cholesterol 25-hydroxylase (Fig. 9). In these experiments, 293 cells were transfected with a 25-hydroxylase cDNA, treated with 2-hydroxypropyl-␤-cyclodextrin to remove endogenous cholesterol, and then incubated with 3 M [ 14 C]cholesterol and the indicated concentrations of unlabeled inhibitor sterol (Fig. 9). The rank order of inhibition for the nine sterols tested was desmosterol Ͼ cholestanol Ͼ 25-hydroxycholesterol Ͼ epicholesterol Ͼ sitosterol Ͼ Ͼ coprostanol ϭ 25-oxo-27-nor-cholesterol (Fig. 9). When desmosterol and [ 14 C]cholesterol were present in equimolar amounts (3 M), enzyme activity was decreased by 30%, whereas coprostanol did not inhibit the enzyme at this concentration. The observed inhibition of 25-hydroxylase activ- ity could be due to individual sterols acting as either true inhibitors of the enzyme (i.e. not as substrates) or as competitors of the cholesterol substrate. In the case of desmosterol (5,24-cholestadien-3␤-ol), which cannot be 25-hydroxylated due to the ⌬ 24 bond, this sterol may be acting as a true inhibitor. DISCUSSION We report the isolation of cDNAs encoding enzymes that convert cholesterol to 25-hydroxycholesterol. An expression cloning assay was initially developed in 293 cells, and screening of Ͼ10 6 independent clones produced a murine cDNA specifying cholesterol 25-hydroxylase. A cDNA encoding the human homologue was subsequently isolated by hybridization screening. The encoded murine and human cholesterol 25-hydroxylases are hydrophobic enzymes that synthesize 25-hydroxycholesterol, share 78% sequence identity, and are predicted to span the membrane multiple times. In transfected cells, the enzyme is localized to the endoplasmic reticulum and the Golgi compartment. The cholesterol 25-hydroxylase gene is expressed at low levels in several murine tissues as judged by RNA blotting. The expression of a cholesterol 25-hydroxylase cDNA in transfected cells suppresses the incorporation of acetate into cholesterol and the proteolytic activation of SREBPs. Taken together, the data suggest that cholesterol 25-hydroxylase may play an important regulatory role in cholesterol metabolism by synthesizing 25-hydroxycholesterol, a known sterol inhibitor of SREBPdependent gene transcription.
Unlike other enzymes that catalyze hydroxylation reactions on sterol and steroid substrates, the murine and human cholesterol 25-hydroxylases are not cytochrome P450s (Fig. 3A). Rather, they belong to a growing family of enzymes that utilize oxygen and a diiron cofactor to catalyze hydroxylation reactions on different substrates. The diiron cofactor can be either Fe-O-Fe or Fe-OH-Fe and is bound to the enzyme through interactions with clustered histidine or glutamate residues (37,41). Sequence comparisons reveal three classes of enzymes in this family, one of which is composed of membrane-bound stearoyl-CoA desaturases, alkane hydroxylase, and xylene monooxygenase (37). These proteins contain multiple membranespanning domains and an unique arrangement of histidine clusters that are postulated to bind the diiron cofactor and to have catalytic function. The murine and human cholesterol 25-hydroxylases contain both of these shared structural features (Fig. 3), and mutation of two of the conserved histidine residues in the murine enzyme eliminates activity (Fig. 4). Similar landmarks are also present in the C-4 sterol methyl oxidase isolated from Saccharomyces cerevisiae and man (38, 39), a rat protein of unknown function termed neurorep 1 (42), and in several cDNA sequences present in the data bases (e.g. GenBank TM accession number U40941 from Caenorhabditis elegans). The cholesterol 25-hydroxylases are 27% identical in sequence to the C-4 sterol methyl oxidases, and they share 25-30% sequence identity with the neurorep 1 and C. elegans proteins. These four proteins may thus have arisen from a common ancestor involved in sterol metabolism.
Cells transfected with the murine cholesterol 25-hydroxylase cDNA produce authentic 25-hydroxycholesterol (Fig. 2B), and the oxysterol synthesized in cells transfected with the human cDNA comigrates with the murine product on thin layer chromatography plates (Fig. 2A). No cDNA-dependent metabolism of 25-hydroxycholesterol, dehydroepiandrosterone, pregnenolone, 5␣-androstane-3␣, 17␤-diol, 5␣-androstane-3␤, 17␤-diol, FIG. 7. Tissue distribution of murine cholesterol 25-hydroxylase mRNA. RNA blot showing cholesterol 25-hydroxylase mRNA in the heart, lung, and kidney. A commercially obtained filter (Mouse Multiple Tissue Blot, CLONTECH) was probed with a radiolabeled, full-length cDNA as described under "Experimental Procedures." After washing, the filter was exposed to x-ray film for 5 days. were transiently transfected with the indicated plasmids as described under "Experimental Procedures." Transfected cells were harvested, and fractions containing membrane or nuclear proteins were prepared. Aliquots of protein (60 -80 g) were separated by SDS-polyacrylamide electrophoresis and blotted with antibodies directed against the hamster SREBP-1 and -2 proteins. Antibody-antigen complexes were visualized by enhanced chemiluminescence. An ϳ120-kDa protein corresponding to uncleaved SREBP precursor was detected in the membrane fraction of the cell (arrows on right of upper panels), whereas an ϳ68-kDa cleavage product was detected in the nuclear fraction (arrows on right of lower panels). Expression of the transfected cholesterol 25-hydroxylase cDNA was confirmed by thin layer chromatography assay of enzyme activity (data not shown). testosterone, dihydrotestosterone, progesterone, androsterone, estradiol, corticosterone, or hydrocortisone was detected when these radiolabeled compounds were added to the medium of transfected cells (data not shown). 24-Hydroxycholesterol was converted to 24,25-dihydroxycholesterol in cells expressing a 25-hydroxylase cDNA. These data suggest that the major enzymatic activity encoded by the isolated cDNAs involves 25hydroxylation of cholesterol and 24-hydroxycholesterol. However, because we have been unable to estimate kinetic constants for the cholesterol 25-hydroxylase enzymes due to the presence of endogenous cholesterol in the transfected cells and their lysates, and because an exhaustive list of potential substrates has not yet been tested, the possibility remains that the observed cholesterol 25-hydroxylase activity is a side reaction detectable only in transfected cells overexpressing these enzymes. Additional biochemical and genetic experiments will be performed to exclude formally this possibility.
Cellular cholesterol homeostasis is maintained by feedback repression of genes involved in cholesterol synthesis and supply (12). Conclusive evidence obtained in cultured cells establishes that repression is mediated through a membrane-bound complex consisting of transcriptional activators (SREBPs), proteases (S1P and S2P, see Ref. 26), and associated cofactors (SCAP, see Ref. 28). When cholesterol builds up in the cell, the activation of SREBPs by proteolysis is suppressed and target gene transcription decreases, thus preventing a build up of intracellular cholesterol. The active agent within the cell that signals repression has not been identified; however, it is known that the addition of cholesterol to cultured cells only weakly suppresses cholesterol synthesis (43), whereas the addition of 25-hydroxycholesterol brings about a rapid decrease in both cholesterol synthesis (1, 2) and the transcription of genes that mediate this process (44,45). Several indirect lines of evidence presented here suggest that cholesterol 25-hydroxylase may be involved in the synthesis of a co-repressor of gene transcription. First, expression of a cholesterol 25-hydroxylase cDNA in transfected cells suppresses cholesterol synthesis and the cleavage of endogenous SREBPs (Fig. 8). Second, the enzyme is localized to the endoplasmic reticulum and Golgi membranes (Fig. 5B), which are sites that would facilitate interaction with components of the cholesterol regulatory complex. Third, the mRNA encoding 25-hydroxylase appears to be present at low levels in a number of tissues and cell types (Fig. 7). This distribution would be expected of an enzyme that synthesizes a powerful repressor of cholesterol gene expression.
How is 25-hydroxycholesterol acting as an oxysterol regulator and what is the role of the 25-hydroxylase in its action? The regulation of cholesterol uptake and biosynthetic enzymes shares features in common with classic inducible and repressible systems of bacteria. On the one hand, oxysterols, which are intermediates of cholesterol catabolism, act as co-repressors to decrease transcription from responsive genes just as the build up of tryptophan within a bacterial cell decreases the synthesis of its biosynthetic enzymes (46). On the other hand, whereas tryptophan binds to and activates a transcriptional repressor, oxysterols act in the membrane compartment of a mammalian cell to decrease the production of an essential positive transcription factor. We show here that endogenously synthesized oxysterols decrease the cleavage of SREBP precursors located in the endoplasmic reticulum; however, the mechanism by which the sterol inhibits proteolysis remains to be determined. Oxysterol binding is thought to induce an altered conformation in SCAP that leads to a decrease in SREBP cleavage (12). We are currently testing this hypothesis to explore further the role of cholesterol 25-hydroxylase in lipid metabolism.