Sterol carrier protein-2 overexpression enhances sterol cycling and inhibits cholesterol ester synthesis and high density lipoprotein cholesterol secretion.

Recent data indicate that sterol carrier protein-2 (SCP-2) functions in the rapid movement of newly synthesized cholesterol to the plasma membrane (Puglielli, L., Rigotti, A., Greco, A. V., Santos, M. J., and Nervi, F. (1995) J. Biol. Chem 270, 18723-18726). In order to further characterize the cellular function of SCP-2, we transfected McA-RH7777 rat hepatoma cells with a pre-SCP-2 cDNA expression construct. In stable transfectants, pre-SCP-2 processing resulted in an 8-fold increase in peroxisomal levels of SCP-2. SCP-2 overexpression increased the rates of newly synthesized cholesterol transfer to the plasma membrane and plasma membrane cholesterol internalization by 4-fold. There was no effect of SCP-2 overexpression on the microsomal levels of acyl-CoA:cholesterol acyltransferase and neutral cholesterol ester (CE) hydrolase; however, in the intact cell, CE synthesis and mass were reduced by 50%. SCP-2 overexpression also reduced high density lipoprotein-cholesterol secretion and apoA-I gene expression by 70% and doubled the rate of plasma membrane desmosterol conversion to cholesterol. We conclude that SCP-2 overexpression enhances the rate of cholesterol cycling, which reduces the availability of cholesterol for CE synthesis and alters the activity of a cellular cholesterol pool involved in regulating apoA-I-mediated high density lipoprotein cholesterol secretion. The net result of these changes in cholesterol metabolism is a 46% increase in plasma membrane cholesterol content, the implications of which are discussed.

Cellular free cholesterol is predominantly located in the plasma membrane (reviewed in Ref. 1). Cellular cholesterol content, however, is determined by the concerted action of intracellular enzymes and regulatory proteins as follows: ACAT 1 which catalyzes the synthesis of CE, sterol regulatory element binding proteins, which regulate the transcription of a number of genes involved in cholesterol metabolism (2,3), and various cell type-specific metabolic reactions, e.g. lipoprotein secretion, steroidogenesis, and bile acid synthesis. Since cholesterol is highly insoluble in an aqueous environment, it has been postulated that sterol carrier protein-2 (SCP-2) regulates the movement and thus the availability of cholesterol for different cellular processes (4 -6). The evidence supporting this contention was initially derived in large part from studies demonstrating that the addition of purified SCP-2 stimulated the in vitro conversion of sterol intermediates to cholesterol (7) and cholesterol to 7␣-hydroxycholesterol (8) steroid hormones (9 -13) and cholesterol ester (14). More recent studies indicate that SCP-2 gene expression is regulated by changes in cellular cholesterol content (15)(16)(17); however, in these studies, a direct role in cholesterol trafficking and esterification was not demonstrated. The strongest support for a role in cellular cholesterol metabolism comes from studies on the role of SCP-2 in steroidogenesis. SCP-2 gene expression is coordinately regulated with steroid hormone synthesis (18,19); moreover, Yamamoto et al. (20) have demonstrated that the delivery of cholesterol to mitochondria for steroidogenesis is enhanced by SCP-2 overexpression, and Chanderbahn et al. (21) have shown that steroidogenesis is inhibited by intracellular delivery of anti-SCP-2 antibody.
Although it was originally thought that SCP-2 acted as a soluble cytosolic sterol carrier protein, it is now clear that SCP-2 is a peroxisomal protein (22)(23)(24)(25)(26). In peroxisome-deficient cells, derived from patients with Zellweger's syndrome, pre-SCP-2 is not processed to SCP-2 and is rapidly degraded (27). Furthermore, peroxisome-deficient cell lines have a reduced capacity for cholesterol biosynthesis (28,29) and alterations in cholesterol metabolism (30,31); however, it is currently unclear if any of these abnormalities can be attributed to SCP-2 deficiency. In this regard, Puglielli et al. (32) recently demonstrated that Zellweger's fibroblasts have a delay in the rate of nascent cholesterol transfer to the plasma membrane and that this defect can be reproduced in normal fibroblasts by antisense SCP-2 oligonucleotide treatment. In these studies, it was also noted that antisense oligonucleotide treatment resulted in a decrease in the level of another SCP-2 gene product, referred to as SCPx (33). Although a recent report seems to indicate that it is more likely that SCPx functions as a peroxisomal 3-oxoacyl-coenzyme A thiolase (34), our laboratory and others have demonstrated that SCPx has in vitro sterol transfer activity (34,35).
Therefore, in order to further examine the cellular role of SCP-2, in the context of intact peroxisomes and without confounding changes in SCPx, we prepared stable transfectants of McA-RH7777 rat hepatoma cells that overexpress SCP-2. Our studies demonstrate that cellular levels of SCP-2 determine the rate of bi-directional cholesterol movement, between the plasma membrane and intracellular cholesterol pools, and the availability of cholesterol for CE synthesis and HDL cholesterol secretion.  (36). Cholesterol oxidase (EC 1.1.3.6; Brevibacterium sp.) was from Sigma. Cholesterol and 25-hydroxycholesterol were from Steraloids (Wilton, NH). Zymosterol was purified from McA-RH7777 cells lipid extracts as described previously (37). All other sterols were from Sigma. Compound 58 -035 was provided by Dr. John Heider (Sandoz Inc., East Hanover, NJ). Oxidized sterol standards were prepared by treatment with cholesterol oxidase as described previously (37). HPLC grade solvents were from Fisher. DNA restriction and modification enzymes were purchased from New England BioLabs. All other chemicals were of reagent grade or better.

Materials-[
Plasmid Construction-A full-length cDNA of rat pre-SCP-2 (33) was provided by Dr. Udo Seedorf (Institut fur Arterioskleroseforschung an der Universitat Munster, Munster D-4400, Germany). The EcoRI-digested 800-base pair cDNA containing the pre-SCP-2 coding region extending from base 72 to 503 was cloned into the pCB6ϩ vector producing a plasmid referred to as pPre-SCP-2. pCB6ϩ (provided by Dr. V. Sukhatme, Department of Medicine, Beth Israel Hospital, Boston, MA) was developed from pCMV4 (38) by inserting EcoRI, KpnI, NotI, ClaI, and EcoRV sites between the BglII and HindIII sites of the polylinker and cloning the neomycin resistance gene into the BclI sites at 1951 and 3290 (39). The cloned cDNA fragment is under the transcriptional control of the immediate early gene of the human cytomegalovirus promotor, and the vector contains polyadenylation signals and the ampicillin resistance gene. pPre-SCP-2 plasmids were isolated and sequenced to ascertain that no mutations were introduced during replication.
At least 1 month prior to all experiments, G418 was removed from the medium. In the absence of G418, SCP-2 expression was monitored by Western blotting and found to be stable for at least 3 months or up to passage number 35. Twenty-four to forty-eight hours prior to the initiation of experiments, medium was removed and fresh medium containing either 10% FBS or 5% (v/v) lipoprotein-deficient fetal bovine serum, prepared as described (42), was added. All experiments were performed in 100-mm dishes with 15 ml of medium unless otherwise stated.
SDS-Gel Electrophoresis and Immunoblotting Techniques-Transfectants were harvested by scraping into a protease inhibitor mixture consisting of 20 mM Tris (pH 7.4), 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 100 M leupeptin, and 5 mM EDTA and passed through a 26-gauge needle. Protein concentrations were measured using the Bio-Rad protein assay (43). Samples (100 g) were separated by 12.5% SDS-PAGE and electroblotted to polyvinylidene difluoride membranes (Immobilon, Millipore) as described previously (41). A polyclonal rabbit anti-rat SCP-2 antiserum was incubated with membranes as described previously and detected with an anti-rabbit horseradish peroxidase-labeled secondary antibody-catalyzed chemiluminescence reaction (Amersham Corp.) (41). Protein bands were quantitated by scanning laser densitometry.
Immunoelectron Microscopy-Immunoelectron microscopy was used to demonstrate the subcellular location of SCP-2 and was performed by Dr. John Lewis (MICROMED microscopy facility, Bowman Gray School of Medicine, Wake Forest University, Winston-Salem, NC) as described (44). Briefly, samples for ImmunoGold analysis were embedded in LR white acrylic resin. Sections with thickness in the range of 70 -100 nm were collected on copper grids and stained with anti-SCP-2 antisera (2 g/ml). A mixture of secondary antisera containing equal parts of goat anti-mouse antibodies conjugated to a 15-nm gold bead and goat antirabbit antibodies conjugated to a 30-nm gold bead were used. Representative micrographs were taken, and a morphometric analysis was performed to determine the numbers of 30-nm gold beads (specific immunoreactivity) and 15-nm gold beads (nonspecific, background immunoreactivity) in peroxisomes and extraperoxisomal areas.
Lipid Analysis-Unless otherwise stated, lipids were extracted (45) and saponified using ethanolic KOH as described previously (41). [ 14 C]Cholesterol was added prior to lipid extraction in order to determine recovery, which typically averaged 80 -85%.
Cellular sterol mass, sterol secretion, and total sterol synthesis were determined as follows. On day 0, cells were plated at a density of 2 ϫ 10 6 cells per 100-mm dish in medium with 10% FBS. On day 1, cell monolayers were rinsed with PBS and changed to fresh medium containing 5% LPDS. On day 2, plates to be used for sterol synthesis were pulsed with 100 Ci of [ 3 H]acetate and incubated for 2 h. At the end of 2 h, all cells and medium were harvested, and lipid extracts were prepared and saponified. Samples were analyzed by HPLC essentially as described (37) with the following modifications. Reversed-phase HPLC analysis was carried out using a Varian 5000 LC system with a Hewlett-Packard model 3390A programmable integrator connected to a Phenomenex IB-Sil 5-m C18 reversed phase column (4.6 ϫ 250 mm). The mobile phase consisted of acetonitrile/water (99:1, v/v) at a flow rate of 1.5 ml/min, and the effluent was monitored at 205 nm. In this system zymosterol eluted at the same time as desmosterol, and lathosterol eluted at the same time as cholesterol (Table I). In order to resolve these compounds lipid extracts were cholesterol oxidase-treated prior to HPLC analysis as described previously (37). Squalene and sterols were identified by their co-migration with known standards on HPLC. Radiolabeled squalene and sterols were quantitated by collecting the corresponding peak fractions and subsequent analysis by liquid scintillation counting. The sum of the radioactivity in the void volume and lipid peaks routinely accounted for Ͼ90% of the injected radioactivity. Lipid mass was calculated from the A 205 /ng lipid (zymosterol, 0.5 pg/A; desmosterol, 0.5 pg/A; lanosterol, 0.5 pg/A; cholesterol, 1.4 pg/A; squalene, 0.1 pg/A). The limits of detection were experimentally determined to be 0.01-0.1 g sterol/mg protein.

Microsomal Assay of the Conversion of [ 3 H]Desmosterol to [ 3 H]
Cholesterol-Microsomes (100 g) were prepared and incubated at 37°C for 1 and 2 h with 1 mM NADPH and 0.5 Ci of [ 3 H]desmosterol as described (46). Incubations were terminated by chilling on ice, and cells and medium were harvested and lipid extracts prepared and saponified. The conversion of [ 3 H]desmosterol to [ 3 H]cholesterol was determined by HPLC analysis and liquid scintillation counting.
Pulse-Chase Analysis of Cholesterol Biosynthesis-Cells (5 ϫ 10 6 ) were incubated in 10 ml of medium containing 5% LPDS and [ 3 H]acetate (100 Ci) at 37°C. After 1 h, the medium was removed and fresh chase medium containing 10 mM sodium acetate was added as described previously (42). At the indicated times, cells were extensively rinsed with PBS, dissociated with trypsin, extracted, and analyzed by HPLC for the radioactivity in cholesterol and sterol intermediates.

TABLE I
Separation of squalene and sterols by HPLC Reverse phase HPLC was carried out as described under "Experimental Procedures" using acetonitrile/water (99:1, v/v) at a flow rate of 1.5 ml/min. Oxidized sterols were prepared by dissolving sterols (1-10 g) in isopropyl alcohol, PBS, and cholesterol oxidase (5 IU/ml), followed by incubating for 30 min at 37°C. At the end of the incubation, samples were extracted with 5 volumes of chloroform/methanol (2:1, v/v) and assayed by HPLC. Both absolute retention times and retention times relative to cholesterol (ϰ c ) are given. Measurement of Cholesterol Transfer to and from the Plasma Membrane-The rate of newly synthesized cholesterol transfer to the plasma membrane was determined essentially as described (36). Briefly, on day 1, 1 ϫ 10 6 cells were plated in medium with 10% FBS. On day 2, the medium was changed to medium with 5% LPDS. On day 3, cells were incubated for the indicated times in medium with 5% LPDS and [ 3 H]acetate, 10 Ci/ml. Cells were dissociated with trypsin, washed with PBS, and fixed with glutaraldehyde prior to treatment with cholesterol oxidase as described (37). Briefly, cells were suspended in 1 ml of PBS, and duplicate aliquots were taken for protein determination. Glutaraldehyde was added to a final concentration of 1% and incubated on ice for 15 min, followed by two washes in 0.5 mM NaP i , 310 mM sucrose (4°C). After the final wash, cells were resuspended in 250 l of 0.1 mM EDTA, 310 mM sucrose and incubated at 37°C for 5 min. Cholesterol oxidase was added to a final concentration of 10 IU/ml, and the cells were incubated for 30 min at 37°C in a shaking water bath. The mixture was placed on ice and the volume brought to 0.5 ml with water prior to lipid extraction. Lipids were extracted and labeled cholesterol, desmosterol, cholestenone, and desmostenone, analyzed by HPLC, and quantitated by liquid scintillation counting.
The time course of transfer of exogenous radiolabeled cholesterol to a cholesterol oxidase-insensitive pool was determined as described (47). Briefly, on day 1, 1 ϫ 10 6 cells were plated in medium with 10% FBS and 12 h later changed to medium with 5% LPDS. Twenty-four hours later, the medium was replaced with unsupplemented medium containing 0.05% bovine serum albumin. After a 5-min preincubation at 37°C, 5 Ci of [ 3 H]cholesterol was added in ethanol to each dish (ethanol concentration Ͻ1%), and the cells were labeled for 30 min at 37°C. The medium was then changed to unlabeled medium with 5% LPDS and the incubation continued for the times indicated. The cells were then dissociated with trypsin, washed with PBS, and fixed with glutaraldehyde prior to treatment with cholesterol oxidase as described above. Labeled cholesterol, cholestenone, and CE were analyzed by TLC using hexane/ ethyl ether/glacial acetic acid 130:30:2 as described (48) and assayed for radioactivity by liquid scintillation counting.
Radiolabeled LDL particles containing [ 3 H]cholesteryl oleate were prepared as described (49). Briefly, freshly isolated human plasma was incubated with 10 Ci of [ 3 H]cholesteryl oleate per ml of plasma for 24 h at room temperature. LDL and radiolabeled human LDL were collected by preparative ultracentrifugation (density ϭ 1.019 -1.063) as described (50). The incorporation of label into radiolabeled LDL typically averaged ϳ3 Ci of [ 3 H]cholesteryl oleate/mg of LDL protein.
Radiolabeled LDL was used in pulse-chase studies to determine the rate of transfer of LDL-derived cholesterol to the plasma membrane as described previously with minor modifications (51). Cells were seeded into 60-mm dishes at 10 6 cells/dish. After 48 h, the medium was changed to DMEM with 2% LPDS, and cells were incubated for 24 h. Cells were then cooled by rinsing with PBS at 20°C followed by the addition of fresh DMEM medium containing 2% LPDS, radiolabeled LDL (100 g/ml), and compound 58-035 (2 g/ml). After 2 h at 20°C, monolayers were extensively rinsed with PBS at 4°C followed by addition of 5 ml of DMEM medium with 2% LPDS, LDL (100 g/ml), and compound 58-035 (2 g/ml) at 37°C. At various times, cells were harvested, treated with cholesterol oxidase, and lipid extracts analyzed by HPLC as described above.
Characterization of CE Metabolism-Cellular CE content was estimated from the difference between free and total cholesterol content as described previously (52) except that [ 14 C]cholesterol was used as an internal recovery standard, and cholesterol mass was determined by HPLC analysis. CE and triglyceride synthesis were determined by measuring the rate of [ 3 H]oleate incorporation. Sodium [ 3 H]oleate-albumin complexes with a specific activity of 49,000 dpm/nmol and an albumin concentration of 120 mg/ml were prepared as described (53). Briefly, on day 0 cells were seeded into six-well plates (ϳ35,000 cells per well) in 2.0 ml of medium with 10% FBS. On day 1, cell monolayers were washed with DMEM and refed with 2.0 ml of medium with 5% LPDS or 10% FBS and LDL (100 g/ml). On day 2, half of the cells in 5% LPDS were either directly assayed for CE and triglyceride synthesis or were treated with 25-hydroxycholesterol at the doses indicated for 5 h. CE and triglyceride synthesis were assayed by the addition of sodium [ 3 H]oleate-albumin complexes (final oleate concentration 100 M) followed by incubation for 2 h at 37°C. Lipids were extracted from the plate in isopropyl alcohol, and monolayers were dissolved in 0.1 N NaOH, 1% SDS and assayed for protein using a modified Lowry procedure (54,55). Unlabeled cholesterol oleate and triolein (100 g each) were added as a carrier prior to analysis by TLC as described previously (41). CE and triglyceride bands were visualized by brief exposure to iodine, scraped from the plate, and quantitated by scintillation count-ing. Microsomal ACAT activity was assayed as described previously (56). Neutral CE hydrolase activity was assayed in cell homogenates as described previously (57).

Analysis of Lipoprotein Secretion and Apolipopoprotein Gene
Expression-To analyze the effect of SCP-2 overexpression on lipoprotein secretion, medium lipoproteins were fractionated by fast protein liquid chromatography (FPLC) using a method developed by Dr. Catherine Reardon (University of Chicago, Chicago IL). Briefly, 4.5 ϫ 10 6 cells were plated in 75-cm 2 flasks in medium with 10% FBS. Twenty-four hours later, the medium was changed to 1% LPDS with 1 Ci/ml (10 Ci/plate) [ 14 C]cholesterol and 1 Ci/ml (10 Ci/plate) [ 3 H]acetate. Cells were incubated in labeling medium for 24 h and then the medium was removed and concentrated to ϳ2 ml by centrifugation in a centriprep number 10 (Amicon, Beverly, MA) for 40 min at 3000 ϫ g. To the concentrated medium an equal volume of FPLC buffer (20 mM KH 2 PO 4 , 20 mM Na 2 HPO 4 , 50 mM NaCl, 0.03% NaEDTA, and 0.02% sodium azide (pH 7.2)) was added, and this was concentrated to 0.5 ml using a Centricon number 10 (Amicon) as described above. The concentrated medium was stored at 4°C in aprotinin, 2 mg/ml; phenylmethylsulfonyl fluoride, 1 mM; sodium azide, 0.002%; and EDTA, 0.01%; for no longer than 12 h before analysis. The entire sample (ϳ0.5 ml) was fractionated on a Superose 6 10/30 column (Pharmacia Biotech Inc.) at a flow rate of 0.4 ml/min, and 70 0.4-ml fractions were collected. Rat plasma was used to calibrate the elution profile of lipoproteins as follows: fractions 8 -18, very low density lipoprotein; 19 -32, low density lipoprotein (LDL); 33-50 high density lipoprotein (HDL); and 50 -70 lipoprotein-free samples. Every three fractions were combined, and carrier cholesterol (100 g) was added prior to extraction. Lipid extracts were analyzed by TLC (41) and cholesterol bands, visualized by exposure to iodine, were scraped into scintillation vials and counted.
The effect of SCP-2 overexpression on apolipoprotein A-I (apoA-I) and E (apoE) secretion was determined by immunoprecipitation of [ 35 S]methionine-labeled apolipoproteins (58). Briefly, 4 ϫ 10 6 cells in 100-mm dishes were incubated with 100 Ci/ml 35 S-Protein labeling mix and 40 mM methionine in methionine-free DMEM with 1% LPDS for 18 h. The medium was harvested and cellular debris removed by centrifugation. Equal amounts of trichloroacetic acid-precipitable counts (ϳ250,000 cpm) were immunoprecipitated with anti-rat apoA-I and apoE antisera (52), followed by separation on a 12.5% SDS-PAGE and autoradiography.
Apolipoprotein transcript abundance was determined in total RNA extracted as described (59). RNA was determined to be intact by following separation by 1% agarose-formaldehyde gel electrophoresis of 20-g aliquots. Northern blots were prepared by using nitrocellulose membranes and were hybridized to rat apoA-I and E cDNAs as described (52). The apoA-I cDNA (60) was provided by Dr. Jeffrey Gordon (Washington University, St. Louis, MO), and the apoE cDNA (61) was provided by Dr. John Taylor (Gladstone Foundation, San Francisco, CA). mRNA abundance was calculated by quantitative scanning densitometry.
Data Analysis-Data were analyzed by a paired t test or one-way analysis of variance and Dunnett's test for multiple comparisons (Minitab). Kinetic modeling was performed by computer using the SAAM II software for Windows (SAAM, Institute, University of Washington).

Establishment of Stable Pre-SCP-2 Transfectants-After
transfection with pPre-SCP-2, or the empty NEO vector, and selection with G418, multiple colonies were obtained and screened for expression by Western blotting. In SCP-2 overexpressing cells (SCPmed and SCPhigh cells), the level of pre-SCP-2 was Ͻ25% of the total SCP-2 immunoreactivity, and the level of SCP-2 was increased ϳ10-fold compared with NEO cells (Fig. 1). The level of SCP-2 in pre-SCP-2 transfectants, however, was within the physiologic range found in normal rat liver (compare lanes 3 and 4 with lane 5). SCP-2 overexpression had no effect on SCPx levels or the previously recognized 30and 35-kDa immunologically cross-reactive isoforms of SCP-2 (41).
Effect of SCP-2 Overexpression on Sterol Synthesis and Mass-Total sterol synthesis was reduced by 12%, from 156,200 Ϯ 3010 in NEO cells to 137,400 Ϯ 4100 dpm/h/mg protein in SCPhigh cells. Cellular cholesterol mass, however, was 38 and 46% higher in SCPmed and SCPhigh cells, in parallel with 80, 67, and 53% reductions in the levels of lanosterol, zymosterol, and desmosterol, respectively, compared with levels in NEO cells (Table II). There was no difference in the sterol content between wild-type McA-RH7777 cells and NEO transfectants (data not shown). Squalene content was unaffected by SCP-2 overexpression and was 0.5 Ϯ 0.2 g/mg protein. The lower levels of cellular desmosterol, in SCPhigh cells, could not be explained by an increase in the microsomal enzyme activity responsible for converting desmosterol to cholesterol (data not shown). In order to determine the effect of SCP-2 overexpression on cholesterol distribution, NEO and SCPhigh cells were fixed and treated with cholesterol oxidase. Consistent with previous reports (37) and in both cell lines, 90% of cholesterol and 80% of desmosterol were oxidized by cholesterol oxidase. These findings indicate that the plasma membrane contains the majority of cellular sterol and that the changes in sterol mass, induced by SCP-2 overexpression, predominantly affect plasma membrane sterol pools.
In order to further characterize the reason for the reduced levels of sterol intermediates in SCPhigh cells, we examined the kinetics of lanosterol, zymosterol, and desmosterol conversion to cholesterol by acetate pulse-chase combined with the cholesterol oxidase method. At the initial time point, in both NEO and SCPhigh cells, the sum of labeled lanosterol, zymosterol, and zymostenone was less than 10% of total labeled sterol and fell to undetectable levels by 4 h (data not shown). Desmostenone, on the other hand, was detected in both cell lines up to the last time point (Fig. 3), indicating that the conversion of desmosterol to cholesterol is the rate-limiting step in sterol intermediate processing in these cells. In both NEO and SCPhigh cells, there was a strict precursor product relationship between plasma membrane desmosterol and total cholesterol (cholesterol ϩ cholestenone). SCP-2 overexpression facilitated this reaction, and after 6 h of acetate chase, over 90% of cellular desmosterol was converted to cholesterol, as compared with only 46% in NEO cells. A nonlinear regression analysis of the desmostenone disappearance data demonstrated that for both cell lines, the kinetics of plasma membrane desmosterol to cholesterol conversion best fit a two-compartment model. In NEO cells, 20% of plasma membrane desmosterol was in a rapidly turning over pool (t1 ⁄2 ϳ36 min), and 80% was in a slower pool with a t1 ⁄2 ϳ14 h. SCP-2 overexpression increased the proportion of desmosterol in the fast pool to 57% (t1 ⁄2 ϳ28 min) and doubled the rate of desmosterol flux through the slow pool (t1 ⁄2 ϳ7 h).
Effect of SCP-2 Overexpression on Sterol Trafficking-The rate of transfer of newly synthesized desmosterol and cholesterol to the plasma membrane was examined using cholesterol oxidase. The rate of appearance of labeled desmosterol in the plasma membrane was unchanged by SCP-2 overexpression and reached the steady state proportion of 80% by 120 min with a t1 ⁄2 ϳ42 min (Fig. 4A). The transfer of nascent cholesterol to the plasma membrane, however, was significantly increased by SCP-2 overexpression (Fig. 4B). In NEO cells, it took 120 min before the percentage of newly synthesized cholesterol in the plasma membrane reached 90% or steady state levels (t1 ⁄2 ϳ 32 min). In contrast, in SCPhigh cells, the appearance of newly synthesized cholesterol in the plasma membrane reached 90% or steady state by 45 min (t1 ⁄2 ϳ 8 min). This 4-fold increase in transfer rate could not be explained by an increase in the amount of labeled sterol secreted into the media (in both NEO and SCPhigh cells, Ͻ1% of the total labeled cholesterol and desmosterol was secreted into the media at 2 h).
In order to determine if SCP-2 overexpression altered the rate of plasma membrane cholesterol internalization, studies were performed on cells loaded with exogenously administered [ 3 H]cholesterol followed by cholesterol oxidase treatment at the indicated times (Fig. 5). The rate of appearance of [ 3 H]cholesterol within the cell was significantly increased by SCP-2 overexpression and reached steady state levels at 2-4 h with a t1 ⁄2 ϳ1 h. In NEO cells, however, it took between 8 and 16 h to reach steady state with a t1 ⁄2 ϳ4.1 h that was 4-fold slower. The difference in the rate of cholesterol influx could not be ex- H]cholesterol in the plasma membrane was also the same in NEO and SCPhigh cells and reached a steady state level between 1 and 1.5 h with a t1 ⁄2 ϳ18 min (Fig. 6). The time course of appearance of [ 3 H]cholesterol in the medium mirrored the time course of appearance in the plasma membrane and was never more than 2% of the total [ 3 H]cholesterol produced.
Effect of SCP-2 Overexpression on CE Metabolism-Free cholesterol and CE content were measured in NEO and SCPhigh cells grown in lipid-free medium with 5% LPDS and medium with 10% FBS and added LDL (100 g/ml LDL). In both NEO and SCPhigh cells, free cholesterol and CE content increased in parallel with medium lipid content (Table III). SCP-2 overexpression, however, reduced CE content by 53 and 70%, in LPDS and LDL medium, respectively. The decrease in CE content in SCPhigh cells resulted from a Ͼ50% reduction in the synthesis of CE in both LPDS and LDL medium, respectively. The effect of SCP-2 overexpression was specific for CE synthesis as there was no effect of SCP-2 overexpression on the incorporation of [ 3 H]oleate into triglyceride in either LPDS or LDL medium.
Assay of microsomal ACAT and cytosolic neutral CE hydrolase activity revealed that the differences in CE content, induced by SCP-2 overexpression, could not be explained by changes in in vitro enzyme activity. ACAT activity was linear between 1 and 2 h and was not different between NEO and SCPhigh cells (4.7 Ϯ 0.3 versus 4.6 Ϯ 0.2 nmol/h/mg protein, respectively). The same was true of CE hydrolase activity  which was 90 Ϯ 12 and 84 Ϯ 6 pmol/h/mg protein in NEO and SCPhigh cells, respectively.
In order to determine if the effect of SCP-2 on CE synthesis could be reversed by treatment with oxysterol, we examined the effect of 25-hydroxycholesterol on CE synthesis in SCPhigh and NEO cells. Cells were preincubated for 24 h in medium with 5% LPDS, followed by treatment for 5 h with 25-hydroxycholesterol at the doses indicated (Fig. 7). In both NEO and SCPhigh cells, CE synthesis increased in a dose-dependent manner and plateaued at a 25-hydroxycholesterol dose of 4 g/ml. At the 4 g/ml dose of 25-hydroxycholesterol, CE synthesis was 2-fold higher in SCPhigh cells. The incorporation of [ 3 H]oleate into triglyceride was unaffected by SCP-2 overexpression as well as 25-hydroxycholesterol treatment.
Effect of SCP-2 Overexpression on Sterol and Lipoprotein Cholesterol Secretion-Incubation of cells in lipid-free medium for 24 h and assay of medium sterol levels revealed that SCP-2 overexpression reduced the appearance of desmosterol by 85% and cholesterol by 59 and 72% in SCPmed and SCPhigh cells, respectively (Table IV). In order to determine if the reduction in cholesterol appearance in the medium was due to a reduction in lipoprotein cholesterol secretion, we labeled cells with [ 3 H]acetate and [ 14 C]cholesterol and separated secreted lipoproteins by FPLC. Of the total labeled cholesterol secreted, Ͼ95% was unesterified. The major lipoprotein species found in the medium of both NEO and SCPhigh cells eluted with the same mobility as HDL (fractions 33-50) and contained Ͼ95% of the labeled cholesterol. SCP-2 overexpression reduced the secretion of [ 3 H]cholesterol by 70% (Fig. 8A) and [ 14 C]cholesterol by 52% (Fig. 8B). The 70% reduction in the secretion of newly synthesized HDL cholesterol parallels the 72% reduction in cellular cholesterol loss (see Table IV).
The final series of experiments examined the effect of SCP-2 overexpression on the gene expression of apolipoproteins (apo) A-I and E. The secretion of [ 35 S]methionine-labeled apoA-I and apoE was examined in the medium from wild-type McA-RH7777, NEO, SCPmed, and SCPhigh cells. SCP-2 overexpression reduced the secretion of apoA-I and E by 70 and 23%, respectively (Fig. 9A). This decrease was due to a 72 and 50% decrease in apoA-I and E transcript levels (Fig. 9B). Thus, the magnitude of the changes in HDL cholesterol and apoA-I secretion, observed in these studies, correlates with the decrease in total cholesterol mass secreted by SCP-2 transfectants (see Table IV and Fig. 8).

DISCUSSION
The transfer of nascent cholesterol to the plasma membrane is a rapid (62) energy (63) and SCP-2 (32) -dependent reaction that occurs against a large cholesterol gradient. The current studies, taken together with data from Puglielli et al. (32), indicate that the rate of this reaction is regulated and proportional to the peroxisomal level of SCP-2. Moreover, in SCPhigh cells the rate of cholesterol internalization paralleled the rate of cholesterol transfer to the plasma membrane, suggesting that the latter event determines the rate of bi-directional cholesterol flux. Lange et al. (64) has demonstrated that ACAT activity is substrate-dependent and limited by the supply of cholesterol from the plasma membrane. At steady state, however, only a fraction of the cholesterol that is internalized from the plasma membrane is converted to CE, whereas the remainder is rapidly transported back to the plasma membrane (64). Thus, the bi-directional movement of cholesterol represents a continuous cycle that may have regulatory significance for ACAT action. In this regard, SCP-2 overexpression reduced CE synthesis, regardless of whether cells were cholesterol-deprived or repleted with LDL in the medium. In hepatoma cells, exogenous and endogenously synthesized sources of cholesterol are equally available for esterification, indicating that cholesterol molecules from all sources rapidly mix at some point prior to ACAT action (65). Although the methodology used, in the current study, to determine the rate of cholesterol transfer to the plasma membrane measures newly synthesized cholesterol movement, the effect of SCP-2 overexpression on CE synthesis implies that SCP-2 enhances the trafficking of intracellular cholesterol, regardless of the source. Since lysosomally derived cholesterol transfers to the plasma membrane prior to ACAT action (51,66), and SCP-2 overexpression had no effect on this transfer reaction, it appears that LDL-derived cholesterol must complete a cycle before it becomes substrate for SCP-2 action.
The inverse relationship between SCP-2 levels and CE synthesis, observed in the current studies, has been previously reported (67). Furthermore, Hirai et al. (16) have recently studied the time course of changes in SCP-2 gene expression and CE synthesis during macrophage foam cell formation and found that, initially, CE synthesis increased and then declined as levels of SCP-2 protein increased in parallel with free cholesterol levels. In the absence of effects of SCP-2 overexpression on the in vitro activity of ACAT and neutral CE hydrolase, we propose that SCP-2 inhibits CE synthesis by increasing the rate of cholesterol cycling, which prevents cholesterol from equilibrating with ACAT, and by shifting the distribution of cholesterol from the cell interior to the plasma membrane. Moreover, the transfer action of SCP-2 secondarily increases plasma membrane cholesterol capacity, which functions to decrease the mass of cycled cholesterol. The distribution of cellular cholesterol has important implications for the regulation of sterol-dependent genes that respond to changes in putative pools of regulatory sterol. In this regard, current knowledge suggests that the rate of apoA-I production is the main determinant of HDL production (68); hence, the decrease in apoA-I gene expression induced by SCP-2 overexpression decreases HDL cholesterol secretion. Changes in the gene expression of apoA-I and E have been correlated with changes in cellular cholesterol content (52, 69 -71), suggesting that SCP-2 overexpression alters the expression of these genes by reducing CE   The mechanism by which SCP-2 acts from within peroxisomes to regulate cholesterol transfer to the plasma membrane is unclear; however, insight into this process can be gleaned from ultrastructural studies of peroxisomes and biochemical studies of cholesterol trafficking. Previous studies have demonstrated an inverse relationship between peroxisomal SCP-2 content and peroxisomal levels of free cholesterol (67). This observation suggests that reduced SCP-2 results in slower clearance of intracellular cholesterol and peroxisomal cholesterol accumulation. Peroxisomal cholesterol may arise from peroxisomal cholesterol biosynthesis (28,29) or it may come from adjacent lipid droplets or vesicles (26), derived from plasma membrane pinocytosis (47,72). Under these conditions, the branching reticular structure of peroxisomes may allow these organelles to function as a central collecting and processing site for intracellular cholesterol. Thus when SCP-2 levels are low, intracellular cholesterol may equilibrate with other cellular membranes; however, a significant increase in intracellular cholesterol is prevented by ACAT activation and the feedback inhibition of LDL receptor and cholesterol biosynthetic gene expression (2,3,73). Some or all of the remaining intracellular cholesterol is transferred to the plasma membrane in a process that is dependent on intact peroxisomes and SCP-2 (32), yet mediated by shuttle vesicles that arise in the endoplasmic reticulum (36,63,74). Although the basis for this apparent discrepancy is unclear, ultrastructural studies have demonstrated an intimate relationship between peroxisomes and endoplasmic reticulum that occurs too frequently to be attributed to chance and is marked by the formation of limiting plates at the sites of membrane contact (75,76). The sterol transfer action of SCP-2 (77), therefore, may determine the peroxisomal disposition of cholesterol that is derived from adjacent donor vesicles by transferring it to adjacent acceptor membranes.
SCP-2 overexpression reduced total sterol synthesis and yet maintained cholesterol balance (see Tables III and IV) by increasing the rate of conversion of plasma membrane desmosterol to cholesterol. This observation suggests that plasma membrane sterol intermediates serve as a reserve that can be called upon to maintain cholesterol synthesis when the activity of cholesterologenic enzymes is suppressed. Recent work by Johnson et al. (78) demonstrated that extracellular acceptors substantially increased the molar efflux of cellular desmosterol, compared with cholesterol, suggesting that desmosterol may exist in a distinct plasma membrane domain. The effect of SCP-2 on desmosterol to cholesterol conversion does not result from a change in the level of microsomal 24-reductase activity, or the rate of desmosterol transfer to the plasma membrane but instead reflects an enhanced rate of desmosterol internalization, which is consistent with the observed shift from a slowly exchanging domain to a more rapidly exchanging plasma membrane domain.
In conclusion, increasing evidence indicates that SCP-2 gene expression is regulated in a tissue-specific manner (41) and in response to a growing number of stimuli (15-18, 41, 79). The current studies suggest that changes in SCP-2 gene expression may have a significant role in the regulation of sterol metabolism through changes in the rate of cholesterol cycling and distribution. In this regard, the role of SCP-2 in bile acid synthesis and biliary cholesterol enrichment needs to be examined, particularly in view of recent data demonstrating a strong correlation between hepatic SCP-2 levels, free cholesterol content, and the incidence of gallstones (80). SCP-2-mediated changes in plasma membrane cholesterol content may also be important in the regulation of integral membrane protein function (81,82), the regulation of cholesterol supply for plasma membrane production during cell growth (41), and HDL-mediated reverse cholesterol transport.