Regulation of phosphatidylcholine metabolism in Chinese hamster ovary cells by the sterol regulatory element-binding protein (SREBP)/SREBP cleavage-activating protein pathway.

Sterol regulation-defective (SRD) 4 cells expressing a mutant sterol regulatory element-binding protein (SREBP) cleavage-activating protein (SCAP D443N) and Chinese hamster ovary (CHO) cells stably expressing SCAP (CHO-SCAP) and SCAP D443N (CHO-SCAP-D443N) have increased cholesterol and fatty acid synthesis because of constitutive processing of SREBPs. We assessed whether constitutive activation of SREBPs also influenced the CDP-choline pathway for phosphatidylcholine (PtdCho) biosynthesis. Relative to control CHO 7 cells, SRD 4 cells displayed increased PtdCho synthesis and degradation as indicated by a 4-6-fold increase in [(3)H]choline incorporation into PtdCho and 10-15-fold increase in intracellular [(3)H]glycerophosphocholine. [(3)H]Phosphocholine levels in SRD 4 cells were reduced by over 10-fold, suggesting enhanced activity of CTP:phosphocholine cytidylyltransferase alpha (CCTalpha). CHO-SCAP and CHO-SCAP D443N cells displayed modest increases in [(3)H]choline incorporation into PtdCho (2-fold) and only a 2-fold reduction in [(3)H]phosphocholine. Elevated PtdCho metabolism in SRD 4, compared with SCAP-overexpressing cells, was correlated with fatty acid synthesis. Inhibition of fatty acid synthesis by cerulenin resulted in almost complete normalization of PtdCho synthesis and choline metabolite profiles in SRD 4 cells, indicating that fatty acids or a fatty acid-derived metabolite was responsible for up-regulation of PtdCho synthesis. In contrast to apparent activation in vivo, CCTalpha protein, mRNA, and in vitro activity were reduced in SRD 4 cells and unchanged in SCAP transfected cells. Unlike control and SCAP transfected cells, CCTalpha in SRD 4 cells was localized by immunofluorescence to the nuclear envelope, suggesting that residual enzyme activity in these cells was in an active membrane-associated form. Translocation of CCTalpha to the nuclear envelope was reproduced by treatment of CHO 7 cells with exogenous oleate. We conclude that the SREBP/SCAP pathway regulates PtdCho synthesis via post-transcriptional activation of nuclear CCTalpha by fatty acids or a fatty acid-derived signal.

of proteins, phospholipids, sphingolipids, and cholesterol. The relative proportions and fatty acyl composition of these components dictate the physical properties of membranes, such as fluidity, surface potential, microdomain structure, and permeability (1). This in turn regulates the localization and activity of membrane-associated proteins (2). Assembly of membranes necessitates the coordinate synthesis and catabolism of phospholipids, sterols, and sphingolipids to create the unique properties of a given cellular membrane. This must be an extremely complex process that requires coordination of multiple biosynthetic and degradative enzymes and lipid transport activities. There is evidence from different experimental models showing coordinate synthesis of sphingomyelin, phosphatidylcholine (PtdCho), 1 and cholesterol (reviewed in Refs. 3 and 4). In several cases the affected enzymes in each pathway have been identified, but the mediators and mechanism for co-regulation remain unknown.
A potentially important pathway for coordinate regulation of membrane composition involves the sterol regulatory elementbinding proteins (SREBPs); transcription factors with a dual role in controlling fatty acid and cholesterol biosynthesis (5)(6)(7). SREBPs are tethered to nuclear/endoplasmic reticulum membranes by two transmembrane segments and undergo two proteolytic reactions that release the soluble, nuclear-localized N-terminal transcription factor domain. The first proteolysis step is sterol-regulated, requires SREBP cleavage-activating protein (SCAP), and is catalyzed by a unique membrane-bound protease (5). The second proteolytic step is constitutive and occurs at a site within the first transmembrane domain (5). Results from transgenic and cell models demonstrated that SREBP-2 and -1a stimulated transcription of numerous cholesterol biosynthetic enzymes and the low density lipoprotein receptor, whereas SREBP-1c/ADD-1 (adipocyte determination and differentation factor-1) isoforms are exclusively involved in the transcription of several key genes in fatty acid biosynthesis and desaturation (8,9). This specificity is not exclusive because both SREBP-1a and -2 appear to stimulate expression of both cholesterol and fatty acid biosynthetic genes. SREBPs also regulate expression of a series of genes involved in the production of NADPH and acetyl-CoA required for fatty acid biosynthesis (10). Recently, SREBP-1c/ADD-1 expression was demonstrated to be insulin-dependent (11) and required for insulin regulation of gene expression (12,13).
In addition to regulation of cholesterol and fatty acid biosynthesis, SREBP/SCAP could potentially regulate the synthesis of phospholipids and sphingolipids by direct or indirect mechanisms. For example, SREBP 1 was shown to regulate the transcription of glycerol-3-phosphate acyltransferase in cultured cells (14) and transgenic mouse models (10). Glycerol-3phosphate acyltransferase is the initial enzyme in the synthesis of phosphatidic acid and diglyceride, two important precursors of phospholipids. Other evidence suggests that SREBPs influence phospholipid synthesis by indirect mechanisms. PtdCho synthesis was reduced in sterol-regulatory-defective (SRD) cells 6 (15) that have reduced cholesterol and fatty acid synthesis as a consequence of defective SREBP processing (5,16). The absence of active SREBPs in SRD 6 cells did not affect the activity of the first and last enzymes in the CDP-choline pathway, choline phosphotransferase (CPT) and choline kinase. Instead, the activity of CCT␣, the rate-limiting enzyme in the CDP-choline pathway (17), was reduced as the result of insufficient activation by fatty acids or a related derivative (15). Sphingomyelin synthesis was also decreased in SRD 6 cells, but again the in vitro activity of biosynthetic enzymes in the pathway was unaffected (18). Reduced sphingomyelin synthesis in SRD 6 cells could have been secondary to a relative deficiency in PtdCho, which provides the phosphocholine headgroup to sphingomyelin. Collectively, these results suggest that SREBPs regulate PtdCho and sphingomyelin synthesis indirectly by modifying the supply of precursors or lipid activators for CCT.
We have now examined PtdCho synthesis in SRD 4 cells (19,20), and CHO cell lines expressing wild-type or a sterol-resistant SCAP mutant (D443N). SRD 4 cells express a SCAP allele with a point mutation (D443N) in the putative "sterol sensing" domain that renders it insensitive to suppression by sterols, resulting in constitutive proteolysis of SREBP 1 and SREBP-2 to the mature transcription factors (20,23). As expected, these cell lines had elevated cholesterol and fatty acid synthesis but also displayed a 2-6-fold increase in PtdCho synthesis because of increased CCT␣ activity in vivo. In SRD 4 cells, activation of PtdCho synthesis (6-fold) was correlated with increased fatty acid synthesis and CCT␣ localization to the nuclear envelope. The modest increase in PtdCho and fatty acid synthesis in SCAP transfected cells (2-fold) was not accompanied by changes in CCT␣ expression or localization. Nuclear envelope localization of CCT␣ in control cells was reproduced by exogenous oleate, suggesting that elevated synthesis of this CCT␣ activator (or a derivative thereof) stimulated PtdCho synthesis in sterol regulation-defective cells. . Lipoprotein-deficient serum was prepared from fetal calf serum by centrifugation at 150,000 ϫ g for 32 h, followed by extensive dialysis against 10 mM phosphate (pH 7.4) and 150 mM NaCl (PBS) (22). Cell culture medium, fetal calf serum, S1 nuclease (from Aspergillus oryzae), and G418 were from Life Technologies, Inc.

Materials-[methyl-
Cell Culture-SRD 4 cells were cultured in Dulbecco's modified Eagle's medium with 5% lipoprotein-deficient serum, 33 g/ml proline (medium A), and 0.3 g/ml 25-hydroxycholesterol at 37°C in CO 2 /air (1:19). CHO 7 cells were maintained in medium A without 25-hydroxycholesterol. For experiments, SRD 4 and CHO 7 cells were subcultured in 60-mm dishes in medium A without 25-hydroxycholesterol (refer to figure legends for specific details). CHO 7 cells overexpressing SCAP and SCAP D443N were prepared by calcium phosphate transfection with pTK-HSV-SCAP-T7, pTK-HSV-SCAP-T7 (D443N), or empty vector (23). Cells were grown in medium A with 600 g/ml G418 until individual colonies were evident at 12-14 days. SCAP transfected cells were trypsinized and cultured in medium A containing 600 g/ml G418 and 0.5 g/ml 25-hydroxycholesterol. After 10 days, thirty G418 and 25-hydroxycholesterol resistant colonies were harvested, expanded in culture for 10 -14 days, and screened for expression of SCAP and SCAP D443N by immunoblotting of total cell extracts with monoclonal antibodies specific for T7 or HSV epitope tags. Mock transfected cells were selected after growth in G418. Four transfected cell lines expressing epitope-tagged SCAP, SCAP D443N, or empty vector were selected for further characterization. Stock cultures of SCAP transfected cells were maintained in medium A with 300 g/ml G418 and 0.3 g/ml 25-hydroxycholesterol and were subcultured for experiments in medium A without G418 or 25-hydroxycholesterol. Analysis of Labeled Phospholipids, Choline Metabolites, and Sterols--After labeling with [ 3 H]choline (see figure legends for specific conditions), cells were rinsed once with cold PBS and scraped in 1 ml of methanol-water (5:4, v/v). The culture dish was rinsed with 1 ml of methanol-water, and the extracts were combined in a glass screw cap tube. [ 3 H]PtdCho and aqueous [ 3 H]choline metabolites were separated by extraction with chloroform as described previously (15). Labeled PtdCho was resolved by TLC in chloroform-methanol-water (65:25:4, v/v/v), whereas aqueous metabolites were separated in a solvent system of ethanol-water-ammonia (48:95:6, v/v/v). In some experiments, [ 3 H]PtdCho was measured by scintillation counting of an aliquot of the chloroform phase (Ͼ98% of the radioactivity was in PtdCho).
Sterol and fatty acid synthesis was measured by [ 14 C]acetate labeling (24). Briefly, cells were incubated with 5 Ci/ml [ 14 C]acetate for 2 h, cell monolayers were dissolved in 1 ml of 0.5 N NaOH, transferred to screwcap tubes, and saponified in 3 ml of ethanol and 0.5 ml of 50% (w/v) KOH for 1 h at 60°C. The sterol fraction was extracted with 4 ml of hexane and resolved by TLC in petroleum ether-diethyl ether-acetic acid (60:40:1, v/v/v). The zone corresponding to cholesterol was scraped into vials, and radioactivity was measured by scintillation counting. Fatty acids were extracted from the hydrolysate with 4 ml of hexane after acidification (pH Ͻ3) with HCl. Radioactivity in an aliquot was measured by scintillation counting (Ͼ99% of the radioactivity in this fraction were in fatty acids). Total cellular phosphocholine levels were determined by thin layer chromatography and phosphate analysis (15).
Enzyme Assays-Cells were harvested by scraping in cold PBS, sedimented at 2,000 ϫ g for 5 min, and homogenized in 20 mM Tris-HCl (pH 7.4), 10 mM EDTA, 5 mM dithiothreitol, and 0.1 mM phenylmethylsulfonyl fluoride by 20 passages through a 23-gauge needle. The homogenates were centrifuged for 60 min at 100,000 ϫ g, the soluble fraction was collected, and the particulate (total membrane) fraction was resuspended in the same buffer containing 250 mM sucrose. CCT activity in the membrane and soluble fractions was measured by conversion of [ 3 H]phosphocholine to CDP-[ 3 H]choline in the presence or absence of PtdCho-oleate (1:1, mol/mol) vesicles as described previously (15,25). CPT activity in membranes was measured by conversion of CDP-[ 3 H]choline to [ 3 H]PtdCho in the presence of 1 mM dioleoylglycerol/ 0.015% Tween-20 (26). Protein was measured by the method of Lowry et al. (27).
Immunoblotting and Immunoprecipitation of CCT␣-Soluble and total membrane fractions from cells were prepared as described above. Membranes were treated with 1% (v/v) Nonidet P-40 on ice for 15 min, and the detergent-soluble fraction was isolated by centrifugation at 15,000 ϫ g for 15 min. A total cell Nonidet P-40 soluble fraction was prepared in a similar manner. Equivalent amounts of protein from Nonidet P-40-solubilized cells, Nonidet P-40 solublilized membranes, and cytosol were resolved by SDS-10% PAGE and transferred to nitrocellulose. The filter was incubated with a polyclonal antibody against 45 amino acids of the C-terminal phosphorylation domain of CCT␣ (provided by Martin Post, Hospital for Sick Children, Toronto, Canada; Ref. 28), followed by a goat anti-rabbit antibody conjugated to horseradish peroxidase, and developed by the chemiluminescence method according to the manufacturer's instructions (Amersham Pharmacia Biotech). All antibody incubations were in 20 mM Tris-HCl (pH 7.4), 150 mM NaCl, 5% (w/v) skim milk powder, and 0.1% (v/v) Tween-20. This antibody does not cross-react with the CCT␤ isoforms because of sequence divergence in the C-terminal phosphorylation domain (29,30).
Phosphorylation of CCT␣ was measured by labeling cells for 15 h in phosphate-free medium A containing 25 Ci/ml [ 32 P]phosphate. Soluble and total membrane fractions were prepared in buffer A as described previously (15). The membrane fraction was treated with 0.3% Triton X-100 on ice for 15 min, and the detergent-soluble fraction was recovered after centrifugation at 400,000 ϫ g for 20 min at 4°C. [ 32 P]CCT␣ was immunoprecipitated from cytosol and solubilized membranes (50 g) in buffer A containing 1% Triton X-100 with a 1:400 dilution of an antibody against the membrane binding region of CCT␣ (kindly provided by Rosemary Cornell, Simon Fraser University, Vancouver, Canada; Ref. 31) for 1 h at 4°C. Protein A-Sepharose was added for 45 min at 20°C, followed by 6 -8 washes with 0.5 ml of PBS containing 1% Triton X-100. Samples were boiled in SDS-PAGE sample buffer and separated by SDS-PAGE in 10% gels. Dried gels were exposed to film at Ϫ70°C.
mRNA Quantitation-CCT␣ mRNA was quantitated by S1 nuclease protection assays using a [␣ 32 P]dATP-labeled antisense probe corresponding to a 86-base pair HindIII-EcoRI fragment of the rat cDNA as described previously (18). The CCT␣ S1 probe is against the 5Ј end of the mRNA and will not hybridize to CCT␤ mRNA because of limited sequence similarity in that region (29,30). Glyceraldehyde-3-phosphate dehydrogenase mRNA was used as an internal load control.

PtdCho Synthesis in SRD 4-and SCAP-overexpressing
Cells-Our previous results showing reduced PtdCho synthesis in cholesterol auxotrophic SRD 6 cells prompted an analysis of PtdCho synthesis in SRD 4 cells, which display elevated cholesterol synthesis that is resistant to down-regulation by 25hydroxycholesterol (19, 20, 23). PtdCho synthesis and turnover was measured in SRD 4 cells, as well as parental CHO 7 cells, by pulse labeling with ). The mass of phosphatidylethanolamine, phosphatidylserine, and sphingomyelin in SRD 4 cells was also similar to controls. In addition to the SCAP D443N mutation, SRD 4 cells have a single point mutation in the ACAT gene that renders the enzyme inactive (32). Although chronic ACAT inhibition by 58 -035 in CHO 7 or SCAP transfected CHO cells (see below) did not affect PtdCho synthesis (results not shown), we could not rule out the possibility that reduced ACAT activity or another mutation in the SRD 4 cells contributed to the results shown in Fig. 1. To address this question, CHO 7 cells were stably transfected with the cDNA for wild-type SCAP or the SCAP D443N mutant and PtdCho synthesis was examined. Initially, four clones resistant to killing by 0.5 g of 25-hydroxycholesterol/ml and expressing epitope-tagged SCAP or SCAP D443N (hereafter referred to as CHO-SCAP or CHO-SCAP D443N) were isolated, and PtdCho synthesis was measured by a 1-h [ 3 H]choline pulse. The average rate of PtdCho synthesis in four mock transfected cell lines was 122.5 Ϯ 32.4 dpm/g protein, compared with 221.1 Ϯ 22.8 and 206.2 Ϯ 22.0 dpm/g protein in four CHO-SCAP and CHO-SCAP D443N cells, respectively. Increased PtdCho synthesis in SCAP transfected cells was also accompanied by a 2-3-fold reduction in Because all four SCAP-and SCAP D443N-overexpressing cell lines appeared to have a similar phenotype with respect to increased PtdCho synthesis, one cell line from each group was examined in detail. The expression of epitope-tagged and endogenous SCAP and SCAP D443N in the two cell lines is shown in Fig. 2. Immunoblotting with a HSV monoclonal antibody detected a protein doublet of approximately 140 -150 kDa in the Nonidet P-40 extracts of CHO-SCAP and CHO-SCAP D443N cell membranes but not in mock transfected controls ( Fig. 2A). It appeared that higher expression of wild-type SCAP compared with the D443N mutant was required to maintain 25-hydroxycholesterol resistance. The cell extracts from Fig.  2A, as well as those from CHO 7 and SRD 4, were also probed with antibody R-139 to detect both endogenous and overexpressed SCAP (21). This antibody also detected a 140 -150-kDa protein doublet in all cells. As expected, expression was highest in the CHO-SCAP and CHO-SCAP D443N cells relative to mock, CHO 7, and SRD 4 cells (Fig. 2B). Our previous results in SRD 6 cells showed that decreased PtdCho synthesis was not correlated with changes in cellular cholesterol levels but rather with the availability of fatty acids (15). However, these studies did not rule out the possibility that de novo synthesis of cholesterol and PtdCho are somehow coupled because cholesterol synthesis in SRD 6 cells is virtually absent because of aberrant SREBP processing (16,33). The relationship between cholesterol synthesis and PtdCho metabolism was examined in SRD 4 and SCAP-transfected cells using lovastatin to inhibit HMG-CoA reductase activity and cholesterol synthesis (Table I) 2. Expression of epitope-tagged and endogenous SCAP and SCAP D443N. The indicated cell lines were homogenized in Tris-HCl (pH 7.4), 5 mM EDTA, and 0.1 mM phenylmethylsulfonyl fluoride, and a total membrane fraction was isolated by centrifugation at 100,000 ϫ g for 1 h. The membrane fraction was then treated with 1% (v/v) Nonidet P-40 and subjected to centrifugation at 15,000 rpm in a microfuge. 40 g of the detergent-soluble supernatant was resolved by SDS-%6 PAGE and transferred to nitrocellulose. A, overexpressed, epitope-tagged SCAP was detected using a 1:500 dilution of HSV-tag monoclonal (Novagen) followed by 1:10,000 dilution of goat antimousehorseradish peroxidase conjugate. B, endogenous SCAP and SCAP D443N was detected using polyclonal antibody R-139 (5 g/ml) followed by a 1:10,000 dilution of goat antirabbit-horseradish peroxidase conjugate. Nitrocellulose filters were developed by the chemiluminescence method according the manufacturer's instructions (Amersham Pharmacia Biotech).  SRD cells, respectively). However, cerulenin suppressed choline incorporation into PtdCho and GPC by 70 and 85%, respectively, relative to controls. Cerulenin also dramatically increased radiolabeled phosphocholine (9-fold) in SRD 4 cells, indicative of reduced flux through the CDP-choline pathway as a consequence of decreased CCT activity. CCT␣ and CPT Activity in SRD 4 and CHO-SCAP Cells-We examined various parameters of CCT␣ expression and regulation to determine the cause of increased CCT␣ activity in SRD 4-and SCAP-overexpressing cells. Initially, we measured CCT␣ mRNA levels by S1 nuclease protection assays in SRD 4 and CHO-SCAP or CHO-SCAP D443N cells. As shown in Fig.  6, CCT␣ mRNA was reduced by 30% in SRD 4 cells and was unaffected in SCAP-and SCAP D443N-overexpressing cells.
The primary mechanism for CCT␣ regulation is via posttranscriptional mechanisms involving phosphorylation and translocation to the endoplasmic reticulum or nuclear envelope (17). To assess whether in vivo activation of CCT␣ was reflected in changes in enzyme distribution or activation in vitro, activity was measured in the soluble and total membrane fraction of cells, either in the presence or absence of PtdCho:oleate vesicles. In vitro CCT␣ activity in SRD 4 cells was reduced 3-and 2-fold in the soluble and membrane fractions, respectively, relative to controls (Table II). However, residual soluble and membrane CCT␣ activity from SRD 4 cells was activated by PC:oleate vesicles to a similar extend as controls. CPT activity in SRD 4 cells was also reduced by 30%. In vitro CCT␣ activity in soluble and membrane fractions was also measured in CHO-SCAP and CHO-SCAP D443N cells (Table III). CCT␣ activity, measured with and without PC:oleate vesicles, from control and the two SCAP transfected cells was not significantly different. The one exception was significant increase (60%) in unstimulated soluble CCT␣ activity in CHO-SCAP D443N cells. CPT activity in CHO-SCAP cells was not significantly different from controls.
Next, we compared enzyme activity with the levels of CCT␣ protein in soluble, membrane, and total homogenates of SRD 4and SCAP-overexpressing cells by immunoblotting with polyclonal antibody against the N-terminal phosphorylation domain of CCT␣ (28). Consistent with enzyme activity measurements in Table II,  Cells were cultured in medium A for 3 days prior to harvesting mRNA. CCT␣ (CT) mRNA levels were assayed in SRD 4 and CHO-SCAP cells as described previously (15). Briefly, a single-stranded probe (labeled with [ 32 P]dATP) for CCT␣ or GAPDH was hybridized with 25 g of total RNA and then treated with S1 nuclease. Protected probe was separated on a 6% acrylamide/urea gel and exposed to film for 12-24 h at Ϫ70°C. The CCT␣ mRNA levels were determined relative to GAPDH expression. Values are expressed relative to control cells and are the means of three separate experiments. controls (Fig. 7). CCT␣ protein mass in total cell homogenates was also reduced by approximately 2-fold in SRD 4 cells. In CHO-SCAP and CHO-SCAP D443N cells, CCT␣ protein expression was similar to mock transfected cells in all fractions. Again, this finding agreed with enzyme activity measurements in Table III. Similar results to those shown in Fig. 7 were also obtained using a polyclonal antibody against the membranebinding domain of CCT␣ (results not shown).
Finally, we determined whether the phosphorylation status of CCT␣ was altered in SRD 4 or SCAP transfected cells (Fig.  8). In these experiments, CCT␣ was labeled with [ 32 P]phosphate in vivo, immunoprecipitated from soluble and membrane fractions and analyzed by SDS-PAGE. Phosphorylation of soluble and particulate CCT␣ from SRD 4 cells was reduced relative to CHO 7 controls, consistent with reduced protein expression and enzyme activity in both these fractions (Fig. 7). There were no apparent changes in the phosphorylation of CCT␣ in CHO-SCAP or CHO-SCAP D443N cells.
Localization of CCT␣ in SRD 4 and SCAP Transfected Cells-CCT␣ in CHO and HeLa cells was shown to strongly localize to the nucleus and translocate to the nuclear envelope in response to phospholipase C-mediated degradation of Ptd-Cho and exogenous oleate (35)(36)(37). This raised the possibility that CCT␣ activities shown in Table II and III did not accurately reflect the distribution of CCT␣ in intact cells because of trapping in the nucleus or disruption of CCT␣ association with membranes during fractionation. To circumvent these potential problems, we analyzed the intracellular distribution of CCT␣ by indirect immunofluorescence (Fig. 9). As reported for CHO-K1 cells (37), CCT␣ in CHO 7 cells was exclusively localized in the nucleus, but with no evidence of nuclear envelope staining (Fig. 9, A and D). In contrast, SRD 4 cells displayed reduced CCT␣ immunofluorescence in the interior of the nucleus and a prominent fluorescent ring around the periphery corresponding to the nuclear envelope (Fig. 9B). This staining pattern could be reproduced in CHO 7 cells treated with 0.1 mM oleate for 1 h (Fig. 9C). CCT␣ also localized to structures within the nucleus of SRD 4 and oleate-treated CHO 7 cells that did not appear to be associated with the nuclear envelope. CCT␣ in CHO-SCAP (Fig. 9E) and CHO-SCAP D443N (Fig. 9F) was primarily in the interior of the nucleus and was not localized to the nuclear envelope.

DISCUSSION
Identification of a dual role for SREBPs in transcriptional regulation of sterol and fatty acid biosynthesis and desaturation (5) indicates that these membrane-bound transcription factors are vital for the maintenance of membrane structure and function. Interestingly, the sensor that regulates sterol-dependent proteolysis of SREBPs, SCAP, is a transmembrane protein and appears to respond to or "sense" the membrane environment in terms of sterol and fatty acid composition (23,38). This provides an important feedback loop, in which SCAP   has a central role by virtue of association with the lipid bilayer, that regulates membrane sterol and lipid composition. In this study we demonstrate that uncoupling this regulatory loop by overexpression of SCAP or a sterol insensitive SCAP mutant (D443N) results not only in increased sterol and fatty acid synthesis but also increased synthesis and catabolism of Pt-dCho. This provides further evidence that sterol, fatty acid, and PtdCho synthesis are coordinately controlled via the activity of the SREBP/SCAP regulatory pathway. PtdCho synthesis was not regulated at the transcriptional level, as shown for other enzymes of sterol and fatty acid synthesis but rather by altering the availability of fatty acids or a related lipid activator of CCT␣, the rate-limiting enzyme in the CDPcholine pathway. PtdCho metabolism was examined in three cell lines that had increased sterol and fatty acid synthesis because of manipulation of SCAP expression. SRD 4 cells and CHO 7 cells stably transfected with epitope-tagged versions of SCAP and SCAP D443N had significantly elevated cholesterol synthesis. However, SRD 4 cells had a 3-fold increase in fatty acid synthesis, compared with only a 25-30% increase in CHO-SCAP and CHO-SCAP D443N cells. This is a significant variation in phenotype that may explain differences in Ptd-Cho metabolism and CCT␣ activity and localization. All three cell lines displayed alterations in PtdCho synthesis and metabolite profiles, but the changes in SRD 4 cells were much greater compared with CHO-SCAP and CHO-SCAP D443N cells. In SRD 4 cells, there was rapid flux of [ 3 H]choline through the biosynthetic pathway such that virtually no phosphocholine accumulated and PtdCho was the major biosynthetic product at the end of a 1 h pulse. In comparison, the majority of the [ 3 H]choline incorporated into control and SCAP-transfected cells was confined to phosphocholine at the end of the 1-h pulse. Taking into account the high level of GPC in SRD 4 cells at the end the pulse period, we conclude that as much as 30% of newly made PtdCho was degraded during this time. If degradation is factored in, PtdCho synthesis in SRD 4 cells is actually elevated by approximately 6-fold relative to controls at the end of the 1-h pulse.
Also consistent with a role of fatty acids in elevated CCT␣ activity was the localization of CCT␣ to the nuclear envelope in SRD 4 cells (Fig. 9). CCT␣ translocation to the nuclear envelope and membrane/particulate fraction was previously observed in response to exogenous fatty acids in numerous cell models (17,36,39). Exogenous oleate in the medium of CHO 7 cells also stimulated CCT␣ translocation to the nuclear envelope, suggesting that elevated fatty acid synthesis in SRD 4 cells is responsible for CCT␣ activation, membrane localization, and increased PtdCho synthesis. CCT␣ in CHO-SCAP and CHO-SCAP D443N cells was not localized to the nuclear envelope, consistent with poor stimulation of fatty acid synthesis relative to SRD 4 cells, possibly related to high level SCAP and SCAP D443N expression in a wild-type background.
Contrary to the immunofluorescence localization of CCT␣ to the nuclear envelope in SRD 4 cells, CCT␣ activity and/or mass was unchanged or reduced in the particulate/membrane fraction of these cells. CCT␣ translocation to membranes was previously shown to be required for activation and increased PtdCho synthesis (reviewed in Refs. 17 and 38). However, translocation in many of these studies was initiated by the acute addition of high concentrations of exogenous fatty acids or manipulation of Ptd-Cho content with phospholipase C (36,40). Why CCT␣ association with the nuclear envelope was not detected in the particulate fraction of SRD 4 cells in vitro is unclear. It is possible that CCT␣ in SRD 4 cells is weakly associated with the nuclear envelope and dissociated during cell homogenization or that the cell homogenization conditions caused significant trapping of CCT␣ in the nucleus that would confound measurements of the membranebound enzyme. The latter seems unlikely because we did not FIG. 8. Phosphorylation of CCT␣ was not influenced by SCAP or SCAP D443N overexpression. CCT␣ phosphorylation was assessed by [ 32 P]phosphate labeling and immunoprecipitation with a rabbit polyclonal antibody against the membrane binding region of CCT␣ (26) as described under "Experimental Procedures." [ 32 P]CCT␣ was separated by 8% SDS-PAGE, and the dried gel was exposed to film for 2 days at Ϫ70°C. Similar results were seen in two other experiments. (v/v) formaldehyde for 15 min at 20°C and permeabilized in 0.05% Triton X-100 for 10 min at Ϫ20°C. Coverslips were then incubated in PBS with 1% (w/v) bovine serum albumin and an antibody against the C terminus of CCT␣ (1/4,000 dilution) for 12 h at 4°C, followed by a secondary goat anti-rabbit fluorescein isothiocyanate-labeled antibody for 45 min at 20°C. CCT␣ was visualized by laser confocal microscopy (Ziess LSM 510), and images were processed using Adobe Photoshop software. observe increased particulate CCT␣ in SRD 4 cells using different cell fractionation conditions but have observed a shift of CCT␣ activity and mass to the particulate fraction in oleatetreated CHO 7. 2 The possibility that the CCT␤ isoform (29,30) is interfering with activity measurements seems unlikely. The expression of total CCT activity in cytosol and membranes closely paralleled the expression of CCT␣ protein in these fractions. As well, we have not detected expression in our CHO cells using a ␤-isoform-specific antibody.
Although exogenous fatty acids promote rapid membrane translocation and activation of CCT␣ in intact cells, resulting in elevated PtdCho synthesis (17,39), there was little evidence that fatty acids regulate PtdCho synthesis in vivo. Our finding that inhibition of fatty acid synthesis by cerulenin in SRD 4 cells resulted in decreased radiolabeling of PtdCho is compelling evidence of a regulatory role for in vivo fatty acid synthesis. Reduction of fatty acid synthesis by cerulenin appeared to down-regulate CCT␣ activity based on a 9-fold increase in [ 3 H]phosphocholine levels in treated SRD 4 cells. This relationship between fatty acid and PtdCho synthesis is supported by previous results with SRD 6 cells, which have a 60 -80% reduction in both fatty acid and PtdCho synthesis (5, 15) because of a defect in SREBP processing by the site two protease (33). Addition of exogenous oleate to SRD 6 cells restored PtdCho synthesis to control levels because of activation of CCT␣. Cholesterol was not deemed to have a direct regulatory role based on a lack of effect of cholesterol supplementation on PtdCho synthesis in SRD 6 cells (15) and no effect of lovastatin in SRD 4, SCAP-transfected, or control cells (this study).
An interesting feature of PtdCho metabolism in SRD 4 cells and to a lesser extent in SCAP transfected cells was the coordinate increase in both synthesis and degradation of radiolabeled PtdCho. The net result was that PtdCho mass in these cell was not significantly altered. A corresponding increase in PtdCho degradation in response to increase synthesis has been previously observed in other cell models where CCT␣ activity was increased by overexpression (41)(42)(43). Because of the large accumulation of GPC that accompanied PtdCho degradation in these cell models, phospholipase A activity is presumed to be involved. Recently, a calcium-independent phospholipase A 2 (iPLA 2 ; Ref. 44) was implicated in turnover of excess PtdCho generated by overexpression of CCT␣ in HeLa cells (42) and CHO cells (43). Based on these findings, iPLA 2 is likely involved in PtdCho degradation in SRD 4 cells. However, we do not know whether iPLA 2 activity is independently regulated by SREBPs/SCAP. If this were the case, release of fatty acids by iPLA 2 would stimulate CCT␣ activity and possibly contribute to increased PtdCho synthesis in SRD 4 cells.
Results presented here support the concept that PtdCho, fatty acid, and sterol synthesis are co-regulated by the SREBP/ SCAP pathway. The finding that PtdCho synthesis is controlled post-transcriptionally suggests that the influence of SREBPs is far reaching and could involve regulation of other lipid metabolic pathways by related mechanisms. Important questions that have yet to be addressed are whether PtdCho synthesis is regulated by both SREBP-1 and 2 and the nature of the fatty acid signal that stimulates CCT␣.