Differential Effects of Sphingomyelin Hydrolysis and Cholesterol Transport on Oxysterol-binding Protein Phosphorylation and Golgi Localization*

The deposition of de novo synthesized and lipoprotein-derived cholesterol at the plasma membrane and transport to the endoplasmic reticulum is dependent on sphingomyelin (SM) content. Here we show that hydrolysis of plasma membrane SM in Chinese hamster ovary cells by exogenous bacterial sphingomyelinase resulted in enhanced cholesterol esterification at the endoplasmic reticulum and rapid dephosphorylation of the oxysterol-binding protein (OSBP), a cytosolic/Golgi receptor for oxysterols such as 25-hydroxycholesterol. After sphingomyelinase treatment, restoration of OSBP phosphorylation closely paralleled resynthesis of SM and down-regulation of cholesterol ester synthesis. SM hydrolysis activated an okadaic acid-sensitive phosphatase that was not stimulated in Chinese hamster ovary cells by short chain ceramides. Agents that specifically blocked sphingomyelinase-mediated delivery of cholesterol to acyl-CoA:cholesterol acyltransferase (U18666A) or promoted cholesterol efflux to the medium (cyclodextrin) did not inhibit OSBP dephosphorylation. SM hydrolysis also promoted OSBP translocation from a vesicular compartment to the Golgi apparatus. Cyclodextrin and U18666A also caused OSBP translocation to the Golgi apparatus, suggesting that OSBP movement is coupled to changes in the cholesterol content of the plasma membrane or Golgi apparatus. These results identify OSBP as a potential target of SM turnover and cholesterol mobilization at the plasma membrane and/or Golgi apparatus.

made in the ER, as well as that released in the lysosome by lipoprotein catabolism, rapidly moves to the plasma membrane (1)(2)(3). Once the capacity of the plasma membrane to absorb cholesterol is exceeded, cholesterol is transported to the ER, where it is esterified, regulates 3-hydroxy-3-methylglutaryl-CoA reductase proteolysis, and inhibits proteolytic processing of sterol-regulatory element-binding proteins required for expression of sterol-regulated genes (4). This distribution of cholesterol between sites of regulation, synthesis, and deposition provides for efficient control of cellular cholesterol levels.
Sphingomyelin appears to play an important role in this process as demonstrated by the capacity of exogenous sphingomyelinase to degrade plasma membrane SM and stimulate cholesterol esterification (5) and sterol-regulatory elementbinding protein processing (6) in the ER. Cholesterol and SM are concentrated in the plasma membrane (7) and are associated with caveolar membrane structures (8). Change in the stoichiometry of cholesterol and SM, either by SM depletion or cholesterol loading, is accompanied by alterations in cholesterol homeostasis. For example, varying the SM content of macrophages altered the stimulation of ACAT activity by acetyl-LDL (9). The SM content of macrophages was also increased by acetyl-LDL cholesterol loading (10) and by oxysterols in CHO cells (11). While there is evidence that SM and cholesterol levels are coordinately regulated and this is important in cholesterol homeostasis, precise mechanisms remain to be determined (9 -13).
The mechanism(s) for cholesterol transport from the plasma membrane in response to SM depletion or influx of cholesterol from the lysosomes is poorly understood. Delivery of ␤-very low density lipoprotein or LDL cholesterol from the lysosomes to ACAT has energy-dependent and -independent components (1,14) and could involve a vesicle transport step (14), and the majority (Ͼ70%) transits through the plasma membrane (15,16). Stimulation of cholesterol esterification by SMase is energy-independent and involves vesiculation of the plasma membrane (14,17). SMase-stimulated cholesterol esterification is also insensitive to protease inhibitors that blocked ␤-very low density lipoprotein activation of ACAT (18). These differences could simply reflect the lysosome-plasma membrane transport component for lipoprotein-derived cholesterol delivery to the ER. However, stimulation of cholesterol esterification by lipoproteins and SMase was inhibited by U18666A, suggesting a common component to both pathways (19,20).
Several proteins have been implicated in the regulation of cellular cholesterol trafficking. Caveolin, a scaffold protein found in caveolae, binds cholesterol and facilitates its movement between the plasma membrane and ER (8). The NPC-1 protein, which is defective in the Niemann-Pick C lysosomal cholesterol storage disorder, plays an undefined role in cholesterol egress from lysosomes (21). Sterol carrier protein 2 has also been implicated in cholesterol transport from the ER to plasma membrane (22). Still other unidentified proteins are involved in the delivery of LDL-and SMase-derived cholesterol to the ER (23). Another protein that could be involved in cholesterol trafficking and regulation is the oxysterol-binding protein (OSBP). OSBP was identified as a high affinity receptor for oxysterols such as 25-hydroxycholesterol (24) that localized to the Golgi apparatus in the presence of its oxysterol ligand (25). The involvement of OSBP in cholesterol regulation was suggested by the positive correlation between oxysterol affinity for OSBP and suppression of cholesterol synthesis (26). More recently, overexpression of OSBP in CHO cells was shown to have pleotropic effects on cholesterol synthesis, expression of sterolregulated gene expression, and ACAT activity (27). Deletion of the OSBP pleckstrin homology domain resulted in loss of this phenotype and lack of association with the Golgi apparatus. OSBP is a phosphoprotein that can be identified on SDS-PAGE by a characteristic increase in apparent molecular mass of 2-3 kDa over the dephosphorylated form (28). While it is presently uncertain how OSBP affects cholesterol homeostasis, its position in the Golgi/vesicular compartment is suggestive of a role in cellular sterol or lipid trafficking.
In this study, we investigated whether SM hydrolysis at the plasma membrane and perturbation of cholesterol trafficking affected OSBP. As expected, bacterial SMase digested 60 -70% of cellular SM and promoted cholesterol esterification. This was accompanied by OSBP dephosphorylation, which was reversible upon SMase removal, and translocation to the Golgi apparatus. Furthermore, dephosphorylation of OSBP was directly linked to SM hydrolysis, but translocation to the Golgi apparatus was mediated by cholesterol depletion at the plasma membrane. . Goat anti-rabbit and goat anti-mouse antibodies conjugated to horseradish peroxidase were from Bio-Rad. C 2 -and C 6ceramide were prepared by acylation of sphingosine and complexed with BSA prior to addition to cells (29). Enhanced chemiluminescence kits were from Amersham Pharmacia Biotech. Tissue culture reagents were from Life Technologies, Inc. Protein was determined with a micro-BCA kit according to the manufacturer's instructions (Pierce).

Materials-Sphingomyelinase
Cell Culture-CHO-K1 cells were grown at 37°C in a humidified incubator in an atmosphere of 5% CO 2 . CHO-K1 cells overexpressing rabbit OSBP by 15-20-fold were prepared and cultured as described previously (27). Both control and overexpressing cells were seeded at 250,000 cells/60-mm dish in DMEM supplemented with 5% fetal bovine serum and 34 g of proline/ml. Twenty-four hours prior to the start of experiments, cells received fresh DMEM containing either 5% FCS or 5% lipoprotein-deficient serum (LPDS). Stock solutions of reagents were prepared in the following manner. Cyclodextrin (250 mM) was dissolved in distilled water, U18666A (2 mg/ml) was dissolved in ethanol, okadaic acid (1 mM) was prepared in Me 2 SO, and SMase and PLC were diluted in phosphate-buffered saline to 10 milliunits/l and 1 unit/l, respectively. LPDS was prepared by ultracentrifugation at a density of 1.21 g/ml and dialyzed extensively against phosphate-buffered saline (30). Human LDL (1.018 -1.063 g/ml) was isolated by ultracentrifugation (30).
Antibodies-Rabbit OSBP overexpressed in CHO-K1 cells was detected by immunoblotting and immunoprecipitation using a monoclonal antibody 11H9 (25,27). Since 11H9 is specific for the rabbit OSBP, a polyclonal antibody was prepared that recognized the protein from CHO-K1 cells. Antibody 104 was raised in rabbits against a glutathione S-transferase fusion protein expressing amino acids 201-309 of rabbit OSBP. The antibody was subsequently affinity-purified using the glutathione S-transferase-OSBP fusion protein coupled to Sepharose.
Immunoblotting, Immunoprecipitation, and Immunofluorescence-At the completion of experiments, CHO cells were harvested in ice-cold PBS and collected by centrifugation (2000 ϫ g for 5 min). Cell pellets were solubilized in 10 mM sodium phosphate (pH 7.4), 150 mM NaCl, 2 mM EDTA, 2 mM EGTA, 10 mM NaF, 1 mM sodium pyrophosphate, 1 mM ␤-glycerophosphate, 100 M phenylmethanesulfonyl fluoride, 2 g of aprotinin/ml, 2.5 g of leupeptin/ml, and 0.3% (w/v) Triton X-100 (buffer B) on ice for 15 min followed by centrifugation for 15 min at 10,000 ϫ g in a microcentrifuge. The supernatant, which contained all immunoreactive OSBP, was collected and analyzed by immunoblotting or immunoprecipitation. Triton X-100 extracts of CHO cells were resolved by SDS-PAGE on 6% gels and transferred to nitrocellulose, and OSBP was detected as described previously (27,28).
Following metabolic labeling of cells with [ 32 P]phosphate (refer to figure legends for specific details), OSBP was immunoprecipitated from Triton X-100 cell extracts by incubation with a 1:100 dilution of antibody 104 (whole serum) at 4°C for 2 h in 200 l of buffer B. A 50% slurry of protein A-Sepharose was added and incubated at 20°C for an additional 30 min. Sepharose beads were collected by centrifugation; washed 8 -10 times with 500 l of PBS, 1% (w/v) Triton X-100; and separated by SDS-6% polyacrylamide gel electrophoresis.
Immunofluorescence detection of endogenous OSBP in CHO-K1 cells was with affinity-purified antibody 104 and fluorescein isothiocyanatelabeled goat anti-rabbit secondary antibody. Cell manipulations, anti- Results are the mean and S.D. of triplicate determinations from a representative experiment. B, following SMase treatment, cells were harvested as described under "Experimental Procedures," the Triton X-100-soluble fraction (25 g of protein) was resolved by SDS-6% polyacrylamide gel electrophoresis, and OSBP was visualized by immunoblotting with monoclonal 11H9. C, cells were prelabeled with [ 32 P]phosphate (25 Ci/ml) in phosphate-free DMEM, 5% LPDS for 1 h. SMase was added directly to cells, and after 30 min cells were harvested and isotope incorporation into OSBP was assessed by immunoprecipitation with monoclonal 11H9 followed by SDS-PAGE. Autoradiography was for 5 days at Ϫ70°C using Amersham Pharmacia Biotech Hyperfilm. The above results are from a representative experiment that was repeated twice with similar results. body treatments, and microscopy were essentially as described for detection of overexpressed OSBP (25,27).
Lipid Analysis-Cholesterol esterification in cultured cells was measured by the incorporation of [ 3 H]oleate into cholesteryl ester (30). Cells were incubated in medium containing 100 M [ 3 H]oleate complexed to BSA at 37°C. The reaction was terminated by extraction of cellular lipids with hexane/isopropyl alcohol (3:2, v/v), and radiolabeled cholesteryl ester and triglyceride were separated by thin layer chromatography and quantitated by liquid scintillation counting.
To assess the extent of hydrolysis of plasma membrane SM and PtdCho by exogenous SMase and PLC, cells were incubated in DMEM with 5% FCS and 2 Ci of [methyl-3 H]choline/ml for 24 h. Exogenous SMase or PLC was added directly to labeled cells, and at the indicated times medium was removed and cells were rinsed once with cold PBS and scraped in 2 ml of methanol/water (5:4, v/v). Lipids were extracted with 6 ml of chloroform/methanol (2:1, v/v) and 4 ml of 0.58% (w/v) NaCl and dried under nitrogen (11). [methyl-3 H]Choline-labeled SM and Pt-dCho were separated by thin layer chromatography in chloroform/ methanol/water (65:25:4, v/v/v) and quantitated by liquid scintillation counting.
[ 3 H]Serine incorporation into SM and ceramide was quantitated by harvesting cells in methanol/water (5:4, v/v) and extracting lipids with chloroform/methanol (1:2, v/v) as described previously (11). SM and ceramide were resolved by thin layer chromatography in a solvent system of chloroform/methanol/water (65:25:4, v/v/v), identified by fluorography of the thin layer plate, and quantitated by liquid scintillation counting.

RESULTS
Hydrolysis of Plasma Membrane SM Promotes Dephosphorylation of OSBP-A well characterized response to the deple-tion of SM in the plasma membrane is the internalization of cholesterol to the ER, where it is esterified by ACAT (6), downregulates 3-hydroxy-3-methylglutaryl-CoA reductase activity (31), and inhibits proteolysis of the precursor forms of sterolregulatory element-binding proteins (7). However, the sequence of events for delivery of cholesterol from the plasma membrane to the ER events following SM depletion is unknown. The localization of OSBP to a Golgi/vesicular compartment and its putative role in sterol and sphingomyelin regulation (11,27) prompted an investigation of its role in SMasemediated cholesterol mobilization. Initially, CHO-K1 cells overexpressing rabbit OSBP were treated with increasing amounts of bacterial SMase for 30 min, and SM hydrolysis, cholesterol esterification, and OSBP expression were examined ( Fig. 1). As expected, cholesterol esterification was stimulated 3-fold in parallel with the hydrolysis of 50 -70% of cellular SM by SMase concentrations Ͼ 5 milliunits/ml (Fig. 1A). In untreated cells, OSBP migrated as a doublet of 97 and 101 kDa (Fig. 1B); the higher molecular weight isoform had decreased mobility in SDS-PAGE as the result of extensive phosphorylation of serine residues (29). With increasing SMase, total OSBP expression did not change, but there was a pronounced shift to the lower molecular weight, dephosphorylated form (Fig. 1B). When the phosphorylation status of OSBP was determined by [ 32 P]phosphate incorporation and immunoprecipitation (Fig.  1C), it was clear that SMase promoted rapid dephosphorylation of OSBP, which coincided with SM hydrolysis and stimulation of cholesterol esterification.
Similar experiments were performed to determine the temporal relationship between SMase-mediated dephosphoryl- At the indicated times, cells were harvested, and the level of OSBP phosphorylation was assessed as described in the legend to Fig. 1 ation of OSBP, SM hydrolysis, and cholesterol esterification (Fig. 2). SM hydrolysis was complete by 20 min, as was OSBP dephosphorylation, measured by immunoblotting (Fig. 2B) and [ 32 P]phosphate incorporation (Fig. 2C). Cholesterol esterification was still increasing by 60 min following SMase addition. In the experiments shown in Figs. 1 and 2, SMase treatment inhibited [ 32 P]phosphate incorporation into total cellular protein by Ͻ15% but blocked [ 32 P]phosphate incorporation into OSBP by 70 -90%.
It could be argued that effects of SMase shown in Figs. 1 and 2 are the result of nonspecific changes in plasma membrane structure due to excessive loss of phospholipid. To rule out this possibility, we tested whether plasma membrane PtdCho hydrolysis by exogenous PLC in overexpressing CHO-K1 cells affected OSBP phosphorylation and cholesterol esterification. At the highest concentration of PLC tested, 60% of cellular PtdCho was hydrolyzed compared with only 15% of cellular SM (Fig. 3A). However, PtdCho hydrolysis did not increase cholesterol esterification, nor did it stimulate OSBP dephosphorylation (as measured by immunoprecipitation of [ 32 P]phosphatelabeled OSBP; Fig. 3B).
When SMase was removed from cells, restoration of SM to normal levels occurred within 2-6 h (32), and cholesterol distribution was reestablished by 24 -48 h (33). In wild type CHO-K1 cells, we examined whether SMase promoted dephosphorylation of endogenous hamster OSBP and whether rephosphorylation occurred concomitantly with resynthesis of SM and down-regulation of ACAT activity (Fig. 4). After a 30-min SMase treatment (25 milliunits/ml) and removal of enzyme, [ 3 H]serine-labeled ceramide was gradually converted to SM over a 6-h period (Fig. 4A). Resynthesis of SM was paralleled by a decline in cholesterol esterification that reached a minimum by 4 h (Fig. 4B). Similar to the overexpressed rabbit protein, endogenous hamster OSBP was converted from the predominant high molecular weight phosphorylated form to the lower molecular weight dephosphorylated form by exogenous SMase (Fig. 4C). The proportion of phosphorylated OSBP slowly increased following removal of SMase and by 6 h was similar to the pretreatment distribution.
OSBP is phosphorylated on at least five distinct sites on serine residues, and approximately 70% of the phosphates at these sites turnover within 20 -30 min (28). Given this rapid phosphorylation cycle for OSBP, SMase hydrolysis could either be inhibiting an OSBP kinase or stimulating dephosphoryl- To assess resynthesis of SM after its depletion by exogenous bacterial SMase, cells were prelabeled with [ 3 H]serine (7.5 Ci/ml) in DMEM, 5% FCS for 24 h. Cells were then treated with SMase (25 milliunits/ml) for 30 min, medium was removed by washing the cells twice with warm PBS, and 2 ml of DMEM, 5% FCS was added to each dish at t ϭ 0. At the indicated times, duplicate dishes were harvested and [ 3 H]serine incorporation into SM (q) and ceramide (f) was measured as described under "Experimental Procedures." Results are the mean of duplicates from a representative experiment. B, for the measurement of cholesterol ester synthesis, cells were treated essentially as described for A. During the last 30 min of each time point, cells were pulse-labeled with 100 M [ 3 H]oleate/BSA, and incorporation into cholesteryl ester was measured. Results are expressed relative to the activity in a matched set of dishes that underwent identical manipulations but did not receive SMase. C, phosphorylation status of OSBP was assessed by SDS-6% polyacrylamide gel electrophoresis and immunoblotting of 25 g of protein from the Triton X-100-soluble fraction of cells using affinity-purified antibody 104 as described under "Experimental Procedures." Results are from representative experiments that were reproduced three times. ation via an okadaic acid-sensitive phosphatase (28), thus resulting in net dephosphorylation. Given the previous reports of protein phosphatase 2A activation by ceramide in vitro (34 -36), we chose to test whether the okadaic acid-sensitive OSBP phosphatase was stimulated by SMase treatment (Fig. 5). In these experiments, OSBP from overexpressing cells was measured by immunoblotting after SMase treatment and either preor post-treatment with okadaic acid. As shown in previous experiments, as little as 10 milliunits/ml SMase promoted dephosphorylation of OSBP, as determined by a shift to the low molecular weight isoform. When cells were pretreated with okadaic acid (500 nM), dephosphorylation was blocked, and OSBP was predominately in the high molecular weight isoform. More importantly, after OSBP was first dephosphorylated with SMase for 60 min, okadaic acid (500 nM) treatment for 30 min was effective in reversing dephosphorylation.
We tested if the okadaic acid-sensitive phosphatase could be stimulated to dephosphorylate OSBP in response to short chain analogues of ceramide in CHO cells (Fig. 6). In these experiments, SMase (50 milliunits/ml for 60 min) inhibited [ 32 P]phosphate incorporation into OSBP from overexpressing cells by 80%. However, this effect could not be recapitulated with C 2and C 6 -ceramides, even at concentrations up to 25 M for 1 h. Phosphorylation of endogenous OSBP in CHO-K1 cells was also unaffected by treatment with these short chain ceramides under similar conditions (results not shown).
Intracellular Localization of OSBP-We previously reported that OSBP overexpressed in CHO-K1 cells translocated to the Golgi apparatus in response to oxysterols but not LDL or other agents that alter cholesterol homeostasis (25). However, this was not confirmed for the endogenous protein due to the lack of a suitable antibody. We developed an affinity-purified antibody (antibody 104) that is capable of detecting endogenous OSBP in CHO-K1 cells and used this antibody to assess changes in localization of OSBP in response to SMase treatment by indirect immunofluorescence (Fig. 7). These experiments were performed on CHO-K1 cells cultured in DMEM with 5% LPDS and supplemented with human LDL. In untreated cells (NO ADDI-TION) the distribution of OSBP was diffuse and appeared localized to small vesicles, which in some instances were clustered around the nucleus. Following SMase treatment for 45 min, OSBP was associated with a structure that appeared at one pole of the nucleus. BFA disrupted this staining pattern, confirming it as the Golgi apparatus (Ref. 25; results not shown). Interestingly, when cells were treated with PLC, which did not promote ACAT activation or hydrolyze SM, OSBP strongly localized to the Golgi apparatus. Similarly, the protein kinase C activator TPA (100 nM) caused OSBP localization to the Golgi complex.
Cholesterol Mobilization and OSBP Phosphorylation and Localization-SM turnover at the plasma membrane results in the generation of ceramide and possibly other bioactive metabolites such as sphingosine and sphingosine 1-phosphate (37), as  Fig. 1 and expressed as percentage remaining relative to untreated cells. To assess OSBP phosphorylation, cells were prelabeled in phosphate-free DMEM, 5% dialyzed FCS containing [ 32 P]phosphate (200 Ci/ml) for 1 h prior to the direct addition of CD or SMase. Phosphorylated OSBP was resolved by immunoprecipitation with antibody 104 and SDS-6% polyacrylamide gel electrophoresis. The dried gel was exposed to Amersham Pharmacia Biotech Hyperfilm at Ϫ70°C for 30 h. B, experiments were carried out essentially as described above. U18666A (0.5 and 2 g/ml) was added to cells simultaneously with SMase (50 milliunits/ml), and SM hydrolysis, cholesterol esterification, and OSBP phosphorylation were measured. The autoradiogram was the result of a 24-h exposure to Amersham Pharmacia Biotech Hyperfilm at Ϫ70°C.
well as cholesterol transport to the ER. We endeavored to dissociate these two effects of SM hydrolysis and determine if cholesterol mobilization was required for dephosphorylation of OSBP by specifically blocking SMase-mediated cholesterol transport to the ER with cyclodextrins (CD) and the hydrophobic amine U18666A. The former compound has been shown to effectively inhibit SMase-mediated ACAT activation by promoting cholesterol efflux to the medium rather than to the ER (38). U18666A inhibited cholesterol desorption from the plasma membrane in SMase-treated cells by 60%, but its mechanism of action is unknown (19). Fig. 8A shows a representative experiment where CHO-K1 cells were treated with combinations of CD and SMase and the effect on OSBP phosphorylation, cholesterol esterification, and SM hydrolysis was assessed. The addition of 5 mM CD to cells in the absence of SMase caused 85% inhibition of cholesterol esterification and a 15% decrease in [ 3 H]choline-labeled SM. This did not result in significant changes in OSBP phosphorylation (as measured by [ 32 P]phosphate incorporation). When cells were simultaneously treated with SMase and 5 mM CD for 60 min, cholesterol esterification was again inhibited by Ͼ90%. This was accompanied by a slight increase in OSBP phosphorylation (compared with SMase alone), which could be accounted for by partial inhibition of SMase hydrolysis. The lower concentration of CD (1 mM) did not affect SM hydrolysis and reduced esterification rates to control levels (20.5 versus 17.5 nmol/min/mg) but did not reverse OSBP dephosphorylation by SMase.
In similar experiments, the effect of U18666A on cholesterol esterification and SMase-mediated dephosphorylation of OSBP was determined (Fig. 8B). Similar to a previous report (15), U18666A was found to block cholesterol esterification in both control and SMase-treated cells, but this was not accompanied by changes in phosphorylation of OSBP (as measured by [ 32 P]phosphate incorporation and immunoprecipitation). U18666A alone or in combination with SMase did not alter SM hydrolysis.
Results in Fig. 8 showed that inhibition of cholesterol transport to the ER or cholesterol depletion of the plasma membrane did not alter OSBP dephosphorylation by SMase. However, OSBP translocation to the Golgi apparatus could be dependent on alterations in cholesterol trafficking initiated by SM hydrolysis or by agents that themselves alter cholesterol movement to the ER. Thus, we tested whether cholesterol depletion at the plasma membrane or blockage of its internalization to the endoplasmic reticulum alone was sufficient to promote OSBP translocation to the Golgi apparatus. Compared with untreated controls, CHO-K1 cells treated with 1 mM CD for 1 h displayed a noticeable increase in OSBP fluorescence in the Golgi apparatus (Fig. 9). The effect was more dramatic with 5 mM CD, and OSBP appeared to be almost completely associated with a single highly compact structure at one pole of the nucleus. CHO-K1 cells treated simultaneously with CD and SMase displayed Golgi localization of OSBP that was indistinguishable from cells treated with CD or SMase alone (refer to Fig. 7). OSBP localization to the Golgi apparatus in CD-pretreated cells was confirmed by disruption of the staining pattern with BFA. Similar to results with overexpressing cells (25,27), endogenous OSBP translocated to the Golgi apparatus in response to 25-hydroxycholesterol. Similar experiments to those shown in Fig. 9 were performed using U18666A. Concentrations of U18666A that inhibited basal and SMase activated cholesterol esterification caused OSBP localization to the Golgi, albeit to a lesser extent than CD (Fig. 10). U18666A did not have an appreciable effect on SMase-mediated OSBP localization. DISCUSSION In this report, we identified OSBP as a potential downstream target following SM hydrolysis and cholesterol mobilization from the plasma membrane. OSBP was originally identified as a high affinity receptor for oxysterol suppressers of cholesterol synthesis and was thus hypothesized to regulate cholesterol homeostasis following binding of these ligands. Previous studies had exclusively examined the effect of exogenous oxysterols on OSBP. While these studies have provided insight into the mechanism of OSBP regulation, they did not address the critical question of whether OSBP responded to changes in cellular cholesterol homeostasis initiated by other mechanisms. In this study, we show that OSBP has a two-part response to hydrolysis of plasma membrane SM: rapid dephosphorylation and translocation to the Golgi apparatus. Golgi translocation was a more generalized response to inhibition of cholesterol trafficking and depletion of cholesterol at the plasma membrane.
A key finding from this study was that SM hydrolysis by exogenous SMase caused dephosphorylation of OSBP, an effect that was independent of cholesterol mobilization. Dephosphorylation occurred within 20 min of SMase addition to control and overexpressing cells, was temporally related to SM degradation and resynthesis, and involved stimulation of an okadaic acid-sensitive protein phosphatase. Our finding that okadaic acid reversed dephosphorylation of OSBP in SMase-pretreated cells and increased phosphorylation of OSBP in untreated CHO-K1 cells (28) is consistent with the involvement of protein phosphatase 2A. Hannun and co-workers identified an okadaic acid-sensitive, ceramide-activated protein phosphatase as protein phosphatase 2A (34). Ceramide appears to directly activate the catalytic subunit of the heterotrimeric enzyme (39). Ceramide-activated protein phosphatase is a target for ceramide produced via receptor-mediated activation of SM hydrolysis and has been implicated in the regulation of c-myc expression (40), c-jun and AP-1 activity (41,42), and ceramide-induced apoptosis (36).
While dephosphorylation of OSBP appeared to involve a ceramide-activated protein phosphatase-like activity, short chain ceramides added to intact CHO-K1 cells failed to mimic the effects of SMase. We previously showed that C 2 -, C 4 -, and C 6 -ceramide analogues are taken up by CHO-K1 cells and metabolized to short and long chain sphingolipids (29). However, short chain ceramide metabolism in CHO cells (5-20% in 1 h depending on the ceramide analogue) would not account for attenuation of a short chain ceramide signal. The inhibition of ACAT in CHO cells by micromolar concentrations of short chain ceramides also suggests that these analogues are taken up (43). It is feasible that ceramide-activated protein phosphatase activity in CHO-K1 cells is poorly activated by short chain ceramides and prefers the long chain species produced by SM hydrolysis. Alternatively, ceramide metabolites such as sphingosine, sphingosine 1-phosphate, or ceramide 1-phosphate could mediate OSBP dephosphorylation.
If SM hydrolysis promotes dephosphorylation of OSBP, what is the biological significance of this signal? We have recently made several observations that link OSBP and oxysterols to SM metabolism. A role for OSBP in regulation of SM synthesis is supported by our finding that 25-hydroxycholesterol, in addition to causing OSBP translocation to the Golgi apparatus, stimulated SM synthesis in CHO-K1 cells (11). In addition, overexpression of OSBP in CHO-K1 cells increased 25-hydroxy-cholesterol-stimulated SM synthesis by 2-3-fold. 2 These results suggest that OSBP and oxysterols are involved in the regulation SM synthesis, and in SM-depleted cells, OSBP dephosphorylation and Golgi localization are involved in the resynthesis of SM. While 25-hydroxycholesterol is required for stimulation of SM synthesis, it did not affect phosphorylation (28). This suggests a complex relationship where translocation of OSBP to the Golgi apparatus stimulates SM synthesis, and this response is enhanced by dephosphorylation that occurs in response to SM hydrolysis. The precise role phosphorylation plays in OSBP function will require identification, mutation, and functional analysis of the five or six serine phosphorylation sites in the protein (28).
Another key finding from this study was the association of OSBP with the Golgi apparatus following cholesterol depletion of the plasma membrane. This was evident both for SMasemediated cholesterol translocation to the ER and desorption of cholesterol to CD in the medium. U18666A also increased OSBP localization at the Golgi apparatus, albeit much less than CD or SMase. This was not unexpected, since U18666A was only 60% effective in blocking cholesterol efflux from the plasma membrane (19). The effect of U18666A could also be explained by inhibition of delivery of LDL cholesterol from the lysosome for replenishment of plasma membrane pools (1). Indeed, we have observed that prolonged treatment of CHO-K1 cells grown in LDL with U18666A results in more profound localization of OSBP in the Golgi apparatus. 3 Prior to these FIG. 10. Immunofluorescence localization of OSBP in CHO-K1 cells treated with U18666A. CHO-K1 cells were cultured with LDL as described in the legend to Fig. 9. Cells received no addition, 0.5 g of U18666A/ml, 2 g of U18666A/ml, or 2 g of U18666A/ml with 50 milliunits of SMase/ml. After 60 min, cells were processed for immunofluorescence as described in the legend to Fig. 7. studies, it was assumed that OSBP only translocated in response to 25-hydroxycholesterol (25). If one hypothesizes that oxysterols are made in response to increased cellular cholesterol, it seems paradoxical that depletion of cellular cholesterol (CD) or inhibition of movement from the plasma membrane (U18666A) would have the same effect. However, 25-hydroxycholesterol also caused depletion of a cholesterol oxidase-sensitive pool of cholesterol at the plasma membrane and promoted rapid delivery of cholesterol to ACAT (44). In this regard, 25-hydroxycholesterol binding to OSBP could initiate a pathway that promotes cholesterol transport from the plasma membrane to the ER. CD, SMase, and U18666A have the same effect on OSBP but apparently in the absence of generation of an oxysterol ligand. We cannot rule out the possibility that these treatments produced an oxysterol ligand that facilitated OSBP movement to the Golgi apparatus. While this could explain the effects of SMase, where cholesterol released to the interior of the cell could be converted to an appropriate ligand, it is difficult to reconcile with CD-mediated OSBP translocation, since the cell is depleted of cholesterol. Rather, the data are consistent with a model where OSBP is regulated by oxysterols, perhaps to initiate changes in cholesterol trafficking, as well as changes in cholesterol movement generated by other stimuli.
The protein kinase C activators TPA and diglyceride (produced in PLC-treated cells) also promoted OSBP localization to the Golgi apparatus. These effects are clearly independent of cholesterol mobilization, since neither resulted in increased cholesterol esterification at the ER ( Fig. 3; unpublished results). TPA and PLC treatment also did not affect phosphorylation of OSBP ( Fig. 3; Ref. 28). It is interesting to note that several protein kinase C isoforms have been localized to the Golgi apparatus (45)(46)(47). TPA treatment of HepG2 cells resulted in condensation of the Golgi apparatus and associated protein kinase C-immunofluorescence, but this was probably due to changes in cytoskeletal structure following protein kinase C activation (48). It is possible that TPA and PLC promote Golgi localization of OSBP by a similar mechanism.
The cholesterol-dependent translocation of OSBP between the Golgi and vesicular/cytoplasmic compartments has some interesting parallels to the intracellular movement of caveolin, a cholesterol-binding protein in caveolae (8). Similar to OSBP, modification of cholesterol pools at the plasma membrane, in this case using cholesterol oxidase, caused a rapid redistribution of caveolin to the Golgi apparatus (49). Caveolin movement between the ER, Golgi, and plasma membrane appears to be important for trafficking of cholesterol from the ER to the plasma membrane (50), perhaps by stabilizing cholesterol/ sphingolipid-rich domains (51). Unlike caveolin, which binds cholesterol and forms a hairpin-like structure in membranes (8,51,52), OSBP is soluble, weakly interacts with membranes (25), and binds a variety of soluble cholesterol derivatives with high affinity. OSBP does not appear to associate with caveolae or the plasma membrane and thus may function in a separate pathway or in one aspect of a common, shared pathway for cholesterol movement between the ER, Golgi, and plasma membrane.