Different Sterol Regulatory Element-binding Protein-1 Isoforms Utilize Distinct Co-regulatory Factors to Activate the Promoter for Fatty Acid Synthase*

Sterol regulatory element-binding proteins (SREBPs) activate genes of cholesterol and fatty acid metabolism. In each case, a ubiquitous co-regulatory factor that binds to a neighboring recognition site is also required for efficient promoter activation. It is likely that gene-and pathway-specific regulation by the separate SREBP isoforms is dependent on subtle differences in how the individual proteins function with specific co-regulators to activate gene expression. In the studies reported here we extend these observations significantly by demon-strating that SREBPs are involved in both sterol regulation and carbohydrate activation of the FAS promoter. We also demonstrate that the previously implicated Sp1 site is largely dispensable for sterol regulation in established cultured cells, whereas a CCAAT-binding factor/ nuclear factor Y is critically important. In contrast, carbohydrate activation of the FAS promoter in primary hepatocytes is dependent upon SREBP and both the Sp1 and CCAAT-binding factor/nuclear factor Y sites. Because 1c is the predominant SREBP isoform expressed in hepatocytes and 1a is more abundant in sterol depleted established cell lines, this suggests that the different SREBP isoforms utilize distinct co-regulatory factors to activate target gene expression. When cells are in medium sufficient cholesterol the sterol regulatory element-binding proteins (SREBPs) are sequestered in the ER by two membrane span-ning domains. When cellular sterol levels drop, the amino-terminal portions of the SREBPs are released from the ER by

When cells are cultured in medium containing sufficient cholesterol the sterol regulatory element-binding proteins (SREBPs) 1 are sequestered in the ER by two membrane spanning domains. When cellular sterol levels drop, the aminoterminal portions of the SREBPs are released from the ER by two ordered proteolytic events (1). The liberated mature SREBPs then enter the nucleus where they activate transcription of various genes in the fatty acid and cholesterol metabolic pathways (2)(3)(4)(5)(6)(7)(8). Additionally, fatty acids have also been shown to affect the regulated processing of the SREBPs from their membrane bound precursor state (9,10); thus, the SREBPs are important regulators of both cholesterol and fatty acid metabolism.
SREBPs are weak activators of transcription by themselves, and for efficient promoter activation they require co-regulatory transcription factors that bind nearby DNA sequences. In the promoter for the low density lipoprotein (LDL) receptor this co-regulator is Sp1 (11). In contrast, recent studies have demonstrated that CCAAT-binding factor/nuclear factor-Y (CBF/ NF-Y) interacts with SREBP and is a key co-regulatory factor for 3-hydroxy-3-methylglutaryl-coenzyme A synthase, farnesyl-diphosphate synthase, and squalene synthase promoters (12)(13)(14).
In our previous report, we determined that two closely spaced SREBP-binding sites are important for SREBP-mediated sterol regulation of the rat fatty acid synthase promoter (15). We also noted the presence of an Sp1-binding site in the sterol regulatory region of the FAS promoter, and we showed that Sp1 functioned together with SREBP to synergistically activate the FAS promoter. These earlier studies were performed using a Drosophila SL2 transfection assay system where both SREBP and Sp1 were expressed from exogenously supplied plasmid templates (3). However, the synergistic effect of SREBP and Sp1 on the FAS promoter was about an order of magnitude lower than observed for either the LDL receptor or acetyl coenzyme A carboxylase promoters. Thus, we reasoned there may be another transcription factor that could work together with SREBPs to activate the FAS promoter to high levels.
In the region of the rat FAS promoter that is required for sterol regulation (3) there is an inverted CCAAT sequence (16) that has been shown to bind several proteins including the CBF/NF-Y factor (17,18). As mentioned above, this heterotrimeric factor has been shown to function together with SREBP to activate some of the genes of cholesterol metabolism. Thus, it was important to further evaluate the roles of Sp1 and CBF/NF-Y as potential co-regulators for SREBP-mediated activation of the FAS promoter. In the present report we show that the CBF/NF-Y site is critical for sterol regulation and for activation by ectopically expressed SREBP-1a and -2 in mammalian cells, whereas the Sp1 site is dispensable.
The promoter for FAS, like those of other key lipogenic genes, is also activated by a nutritional and hormonal process that is dependent on both insulin action and carbohydrate metabolism (19). In previous studies, mutations that disrupt the E-Box element that lies within and overlaps the tandem SREBP sites mentioned above were shown to be defective for insulin-dependent activation of FAS (20,21). In the present report we show that mutations that disrupt the tandem SREBP sites but leave the E-box intact are defective for stimulation by glucose and insulin treatment in primary hepatocytes where SREBP-1c is the major isoform. Additionally, both NF-Y and Sp1 sites are absolutely required as well. Thus, the two * This work was supported in part by National Institutes of Health Grant HL48044 and American Heart Association Grant 96008190. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Recipient of an undergraduate fellowship from the National Institutes of Health and a President's Undergraduate Fellowship from the Undergraduate Research Opportunities Program at University of California, Irvine.
ʈ To whom correspondence should be addressed: Dept. of Molecular Biology & Biochemistry, University of California, Irvine, CA 92697-3900. E-mail: tfosborn@ucl.edu. 1 The abbreviations used are: SREBP, sterol regulatory element-binding protein; LDL, low density lipoprotein; FAS, fatty acid synthase; CBF, CAATbinding factor; NF-Y, nuclear factor-Y; CMV, cytomegalovirus. SREBP-1 isoforms appear to have distinct co-regulator requirements to efficiently stimulate the FAS promoter in response to different nutritional challenge.

MATERIALS AND METHODS
Cells and Media-The CV-1, HepG2, and SL2 cell lines were used in transient transfections conditions as described before (11,15). Lipoprotein-deficient serum was prepared by ultracentrifugation of newborn bovine serum as described previously (22). Cholesterol and 25-hydroxycholesterol were obtained from Steraloids Inc., and stock solutions were dissolved in absolute ethanol.
Cell Culture and Transient Transfection Assay-CV-1 cells were grown in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and were plated at 125,000 cells/60-mm dish on day 0. On day 1 cells were transfected by the calcium phosphate co-precipitation method as described previously (11). Each dish received 5 g of each test reporter construct plasmid and the non-sterol-regulated CMV2 ␤-galactosidase plasmid, which contains the cytomegalovirus (CMV) promoter fused to the Escherichia coli ␤-galactosidase gene as an internal control for transfection efficiency. In addition, dishes received 2.5 g of salmon sperm DNA. Where indicated, cells were also transfected with the indicated amount of plasmid DNA corresponding to pNF-YA29, which we received from Dr. Mantovani (23). Dishes were incubated at 37°C and 7% CO 2. On day 2, 12-16 h after transfection dishes were washed three times with 1ϫ phosphate-buffered saline and refed either induced (Dulbecco's modified Eagle's medium and 10% lipiddeficient serum) or suppressed medium (same as induced medium but also containing 12 g/ml cholesterol and 1 g/ml 25-hydroxycholesterol), and dishes were incubated for another 24 h. On day 3 cells were harvested by scraping, and duplicate dishes were pooled prior to extraction of soluble protein extracts by three freeze thaw cycles. Luciferase and ␤-galactosidase activity were measured as described under "Enzyme Assays" below.
HepG2 cells were cultured in modified essential medium containing 10% fetal bovine serum and seeded for experiments at 175,00 cells/ 60-mm dish on day 0. On day 1 cells were transfected by the calcium phosphate co-precipitation method with the same amount of test plasmid, CMV2 ␤-galactosidase control plasmid, and salmon sperm DNA as described above for CV-1 cells. In addition, half of the dishes received a CMV expression clone encoding either amino acids 1-490 of the human SREBP-1a or 1-481 of human SREBP-2 proteins at the levels indicated in the figures, whereas the other half of the dishes only received salmon sperm DNA as a control (11) . 5 h after transfection cells were treated with 10% glycerol in 1ϫ phosphate-buffered saline for 1 min, washed three times with 1ϫ PBS, and refed the same medium as mentioned above. Cells were cultured for an additional 24 h at 37°C and 5% CO 2 , after which they were harvested and enzyme assays were performed on cell extracts as described above for CV-1 cells.
Drosophila SL2 cells were cultured in Shields and Sang insect medium (Sigma) containing 10% heat-inactivated fetal bovine serum and were seeded at 1,200,000 cells/60-mm dish on day 0. On day 1 cells were transfected by the calcium phosphate co-precipitation method with each dish receiving 2 g of each test plasmid, 10.75 g of salmon sperm DNA, and 1 g of the control plasmid pPAC ␤-galactosidase containing the coding region of the E. coli ␤-galactosidase gene under the control of the Drosophila actin 5C promoter. The pPAC Sp1 and pPAC SREBP-1a constructs used for activation studies in SL2 cells contain the coding regions of the Sp1 or SREBP-1a (amino acids 1-490) genes under the control of the Drosophila actin 5C promoter. The pPAC Sp1 construct was obtained from Dr. Al Courey (UCLA), and the pPAC SREBP-1a construct was described previously (11). The coding regions for the three individual CBF/NF-Y subunits (A, B, and C) were each cloned into the pPAC backbone separately as BamHI/XhoI fragments.
Primary hepatocytes were isolated from male Harlan Sprague-Dawley rats (180 -260 g) using the collagenase perfusion method as described previously (24). After a 3-6-h attachment period, cells were transfected using the liposome reagent Targetfect (Targeting Systems, San Diego, CA) in modified Williams E medium with 23 mM HEPES, 0.01 M dexamethasone, 0.1 unit/ml insulin, 1 unit/ml penicillin, 1 g/ml streptomycin, and 5.5 mM glucose for 12-14 h. Subsequently, cells were cultured in medium containing either 5.5 mM glucose in the absence of insulin or 27.5 mM glucose in the presence of 0.1 unit/ml insulin. Matrigel (Collaborative Biomedical Products, Bedford, MA) was added to the plates at a concentration of 0.166 mg/ml after transfection. This treatment has been shown to enhance the maintenance of the differentiated phenotype in cultured hepatocytes (25). After a 48-h treatment, cells were harvested for luciferase assay. Results are ex-pressed as relative light units normalized by protein concentration of each sample.
FAS Promoter Plasmids-The rat FAS promoter construct designated as the wild type contains the FAS promoter fragment from Ϫ150 to Ϫ43 fused to a generic TATA box positioned upstream of the luciferase gene coding sequence and was described previously in (3). The mutant FAS constructs were generated by polymerase chain reaction using the wild type FAS construct as template and mutagenic oligonucleotides to introduce the desired multi-base point mutations. The 3Ј polymerase chain reaction oligonucleotide used to introduce the Sp1binding site mutation contained the following sequence: 5Ј-GCATA GCTAG CTCCC CGCGC GGCCA CGCCA CATGG GCTGA CAGCT TGGCT TTGTT GTTTAG GC-3Ј. The 5Ј polymerase chain reaction oligonucleotide used to introduce the NF-Y-binding site mutation contained the following sequence: 5Ј-GCATG GAGCT CATCA CCCCA CCGAC GGCGG CGCGC CGGGT CCCGG GGCGC AGCCC CGACG CTCCG AGGCC T-3Ј. The mutant bases in each oligonucleotide are those represented in bold type. The precise introduction of the desired mutations was confirmed by DNA sequencing prior to transfection studies.
Enzyme Assays-The luciferase activities were measured in a luminometer with a luciferin reagent from Promega Biotec. ␤-Galactosidase assays were performed by a standard colorimetric procedure with 2-nitrophenyl-␤-D-galactopyranoside as substrate (26). The normalized luciferase values were determined by dividing the luciferase activity in relative light units by the ␤-galactosidase activity (A 420 /h). The data presented here are from several experiments performed in duplicate for each plasmid (see figure legends for exact number of individual experiments).
Gel Mobility Shift Analysis/Protein Purification-Purified human Sp1 protein (from Promega) was used at 1 footprint unit/reaction. The binding and electrophoresis conditions for the electrophoretic gel mobility shift assays were described by Sanchez et al. (11). The probes used contained the FAS promoter sequence from Ϫ105 to Ϫ76 with either the wild type, mutant Sp1-binding site or mutant NF-Y-binding site sequences. The multi-base mutations in each of the two mutant probes are the same as the mutations introduced into the reporter plasmids described under "FAS Promoter Plasmids."

RESULTS
The cis-acting elements necessary for sterol regulation of the rat FAS promoter were previously shown to reside between Ϫ150 to Ϫ43 relative to the transcription initiation site (ϩ1) (3). In a follow up study we showed that two adjacent binding sites for SREBPs located between Ϫ72 and Ϫ53 are crucial for this effect (15). We also demonstrated that a GC-rich sequence at Ϫ80 bound the ubiquitous Sp1 factor and that ectopic expression of both SREBP and Sp1 would synergistically activate the FAS promoter in Drosophila SL2 cells.
However, the degree of synergistic activation of the FAS promoter by SREBP and Sp1 was an order of magnitude lower than for SREBP-Sp1 synergistic activation of the LDL receptor and acetyl coenzyme A carboxylase promoters (3,5,27). In addition to the Sp1-binding site at Ϫ80 there is a recognition site for CBF/NF-Y within the region of the FAS promoter that is required for sterol regulation (Fig. 1, bottom panel, and Refs. 15 and 16) To determine whether Sp1, CBF/NF-Y, or both were important co-regulators for SREBP-dependent and sterol-regulated activation of the FAS promoter, we first evaluated the effects of mutations in the DNA-binding sites for Sp1 or NF-Y on sterol regulation (Fig. 1). Multi-base substitution mutations were introduced into the Sp1 or CBF/NF-Y sites, and the resultant mutant promoters were analyzed in transient transfection assays for sterol regulation in CV-1 cells as described under "Materials and Methods." The Sp1 mutant was regulated by sterols in a similar fashion as the wild type promoter, whereas the CBF/NF-Y mutant was totally defective for sterol regulation. These data demonstrate that the CCAAT element is required for sterol regulation of the FAS promoter, whereas the Sp1 site is not.
We also compared these mutants with the wild type pro-moter in transient transfection assays where the mature form of either SREBP-1a or SREBP-2 was supplied exogenously from an expression plasmid (Fig. 2). The results extend the sterol regulation experiments and specifically demonstrate that the CBF/NF-Y site is important for activation by both SREBPs (Fig. 2, A and B, compare lanes 1-3 with lanes 7-9). However, the mutant with an altered Sp1 site was efficiently activated by both SREBPs (Fig. 2, A and B, compare lanes 1-3 with lanes 4 -6).
To confirm that Sp1 binding was actually disrupted by the base changes that were introduced, we evaluated Sp1 binding to the wild type and mutant DNAs by an electrophoretic gel mobility shift assay. The results demonstrated that the wild type and NF-Y mutant DNA fragments bound Sp1 efficiently, whereas the mutant probe designed to disrupt the Sp1 site did not (Fig. 3).
The experiments presented thus far demonstrate that the CBF/NF-Y consensus DNA site is important for sterol regulation of the FAS promoter. To evaluate the role of the heterotrimeric CBF/NF-Y protein directly, we performed transient DNA transfection assays in the presence of increasing amounts of a plasmid that expresses a mutant form of the A subunit of NF-Y. The mutant A subunit protein can associate with the other two subunits normally; however, the resultant complex is unable to bind specifically to DNA (23). Thus, when overexpressed it functions in a dominant negative fashion over endogenously expressed CBF/NF-Y and it has been used by several investigators including us to evaluate the role of CBF/NF-Y in the activation of other specific promoters (15,23,28). When the wild type FAS promoter was transfected into CV-1 cells along with increasing amounts of the NF-YA29 mutant plasmid, the level of expression of the FAS promoter under sterol depleted conditions was decreased in a dose-dependent manner (Fig.  4A). Similar results were obtained for the FAS reporter plasmid with the mutations in the Sp1 recognition site (Fig. 4B). That this effect is really specific for NF-Y is indicated by the fact that the NF-YA29 plasmid had no effect on SREBP-dependent activation of the LDL receptor promoter where only SREBP and Sp1 are involved (28).
The Drosophila SL2 cell line does not express functional equivalents of several mammalian transcriptional regulatory proteins, including the transcription factor Sp1 (29). However, activation of mammalian promoters can be reconstituted when expression constructs encoding the missing regulatory proteins are co-introduced by DNA transfection. Therefore, it is a useful cell based assay system for the analysis of transcription factor requirements for promoter activation because it provides a negative background for such studies (29). For example, this assay system was used to directly show that SREBP and Sp1 function as co-activators of the LDL receptor promoter (11). Based on our previous experiments and the data presented thus far, we hypothesized that SL2 cells may lack functional CBF/NF-Y and that a more efficient activation of the FAS promoter might be achieved by introducing expression constructs for CBF/NF-Y along with those for SREBP and Sp1.
To directly test this hypothesis we developed a co-stimulation assay for SREBP and CBF/NF-Y in Drosophila SL2 cells. We prepared Drosophila -based expression constructs for all three CBF/NF-Y subunits and first added them individually or in different combinations to determine which if any were functionally missing (Fig. 5). The results demonstrated that significant activation of the FAS promoter was achieved only when all three CBF/NF-Y expression constructs were added along with SREBP whereas the combination of SREBP and Sp1 was much less active (Fig. 5, compare lanes 4 and 12). The inclusion of Sp1, SREBP, and all three NF-Y subunits resulted in an even higher level of activation (lane 11).
The requirement for CBF/NF-Y in activation of the FAS promoter was evaluated more precisely by performing a concentration curve with the CBF/NF-Y expression plasmids in the presence or absence of a fixed amount of the SREBP-1a and Sp1 expressing constructs (Fig. 6). From this series of experiments it is clear that CBF/NF-Y alone (triangles) or in combination with Sp1 (diamonds) is not very effective, whereas SREBP-1a addition on top of NF-Y resulted in a significant stimulation of FAS promoter activity at all concentrations of CBF/NF-Y plasmids (squares). When Sp1 was included in addition to SREBP-1a, there was a higher level of activation observed at every concentration of CBF/NF-Y (circles). FAS expression is activated by insulin signaling and carbohydrate metabolism, and the FAS promoter has been studied as a model for lipogenic gene activation (19). A region encompassing the E-box and tandem SREBP sites discussed here is required for activation of the FAS promoter by insulin and carbohydrate signaling (20). To explore the potential roles of the SREBP, NF-Y, and Sp1 sites in activation of the FAS promoter by insulin and glucose, we evaluated the effects of the Sp1 and NF-Y site mutations described in Fig. 1 and the SREBP-binding site mutants we analyzed previously (15) on activation of the FAS promoter in primary rat hepatocytes in response to insulin and glucose (Fig. 7). The wild type promoter was activated over 6-fold by the addition of insulin and a high concentration of glucose, whereas the mutations in either the NF-Y or Sp1 sites resulted in a loss in activation. Additionally, we analyzed three other mutations that we made for our previous study that helped us determine that the tandem SREBP sites (and not the E-box element) are the critical cis-acting sites for sterol regulation. One of the mutations alters the E-box sequence and downstream SREBP site simultaneously (mutant B); another mutation alters the upstream SREBP site without affecting the E-box sequence (mutant D), and one mutant alters both SREBP sites simultaneously (mutant A/B/C). All three of these additional mutations disrupted SREBP binding to one (mutants B and D) or both (mutants A/B/C) SREBP sites and were defective for sterol regulation (15). The data in Fig. 7 demonstrate that all three mutants are also defective for acti-vation by insulin/glucose. Taken together these experiments document that the two tandem SREBP sites as well as both the Sp1 and NF-Y site are important for insulin/glucose activation of the FAS promoter.
Because the major isoforms of SREBPs in sterol-depleted cultured cells are SREBP-1a and SREBP-2 and the major nuclear SREBP isoform in normal hepatocytes is SREBP-1c (30), the experiments presented so far suggest that SREBP-1c may require both Sp1 and NF-Y for efficient activation of the FAS promoter, whereas activation by SREBP-1a would be largely dependent only on NF-Y. To test this possibility we analyzed the co-regulatory factor requirements for each of SREBP-1a and -1c in Drosophila SL2 cells where we can specifically manipulate the functional levels of SREBPs, NF-Y, and Sp1 directly by co-transfection (Fig. 8). We evaluated the activation of the FAS promoter by increasing concentrations of NF-Y in the presence of SREBP-1a or SREBP-1c expressing constructs in the presence or absence of co-transfected Sp1 plasmid. Consistent with the mammalian cell studies and our other SL2 studies, efficient activation by SREBP-1a occurred when SREBP-1a and NF-Y were co-transfected (Fig. 8, open squares); however, high level of activation by SREBP-1c required that both NF-Yand Sp1-expressing plasmids were included (Fig. 8, compare  open and closed circles). DISCUSSION Cholesterol and fatty acid metabolism are coordinately regulated through the action of the SREBPs. However, when separate control of each pathway is required there has to be a mechanism to modulate one process independently from the other. Because the SREBPs require generic factors as transcriptional co-regulators, it is likely that part of the independent regulation results from the recruitment of distinct transcriptional co-regulators by SREBPs to activate key genes of one process or the other. Thus, the first step in understanding how the SREBPs differentially regulate gene expression is to identify the important SREBP co-regulatory proteins in the promoters for critical genes of each pathway such as FAS.
In our previous work we showed there are two critical SREBP sites and a putative co-regulatory Sp1 site in the region of the FAS promoter that is essential for SREBP activation and sterol regulation in cultured cells. We also demonstrated that SREBPs and Sp1 could synergistically activate the FAS promoter in Drosophila SL2 cells (13). However, we noted that the FIG. 3. Electrophoretic gel mobility shift assay analysis of Sp1 binding to FAS promoter. 34-base pair double-stranded oligonucleotide probes corresponding to Ϫ105 to Ϫ76 of either the wild type (WT) FAS sequence, the mutated CCAAT box motif sequence (NF-Y mut.), or the mutated Sp1-binding site sequence (Sp1 mut.) were 32 P-end-labeled and used in a gel mobility shift assay as described under "Materials and Methods." The probes were incubated in the absence (denoted by Ϫ) or in the presence of 1 footprint unit of human Sp1 protein (indicated by ϩ). The position of the shifted Sp1 complex is shown by the arrow at the right. The apparent difference in size of the two shifted complexes is due to an anomaly in the gel dye front.

FIG. 2. The inverted CCAAT motif at ؊100 is required for both SREBP-1a and 2 activation of the rat FAS promoter in HepG2 cells.
HepG2 cells were transfected with the wild type (WT) or mutant FAS constructs along with the CMV2 ␤-galactosidase construct as a control as described under "Materials and Methods." A, dishes received 0, 10, or 30 ng/60-mm dish of a CMV-driven expression vector encoding amino acids 1-490 of SREBP-1a. B, dishes received 0, 10, or 30 ng/60-mm dish of a CMV-driven expression vector encoding amino acids 1-481 of SREBP-2. Fold activation corresponds to the ratio of normalized luciferase activity obtained from dishes where SREBP-1a or-2 was co-transfected divided by the activity obtained from dishes transfected in the absence of the SREBP expression vector. The mean fold activation value for each plasmid and the standard error for three independent experiments performed in duplicate are presented. degree of activation by SREBP and Sp1 was modest compared with the robust activation that resulted from the interaction of these same two factors on the LDL receptor and acetyl coenzyme A carboxylase promoters (3,5,11). Also, there is a CBF/ NF-Y site close to the putative key Sp1 site (16). Based on these observations we designed a series of experiments to specifically evaluate the roles of CBF/NF-Y and Sp1 as co-regulators in SREBP activation of the FAS promoter.
We introduced multi-base substitutions within the predicted recognition sites for both Sp1 and NF-Y and evaluated the effects on sterol regulation (Fig. 1) and activation by ectopically introduced and constitutively active forms of either SREBP-1a or -2 in cultured cells (Fig. 2). The results demonstrate that the CBF/NF-Y site is essential for sterol regulation and activation by both individual SREBPs, whereas the Sp1 site does not play an important role. We confirmed that the mutations we introduced did destroy Sp1 binding (Fig. 3) and that CBF/NF-Y activity is required for sterol regulation of FAS (Fig. 4). Thus, CBF/NF-Y is the critical co-regulator for SREBP activation of the FAS promoter in response to sterol deprivation.
We also evaluated the SREBP co-regulator specificity in a Drosophila SL2 transfection assay. For these studies we first had to determine whether the SL2 cells could be used to evaluate activation by exogenously supplied CBF/NF-Y and which if any of the three individual subunits were functionally missing in these cells. The experiments in Fig. 5 demonstrate that SL2 cells are indeed an efficient host-cell system to evaluate activation by mammalian CBF/NF-Y and that efficient activation required that expression vectors for all three CBF/NF-Y subunits were together with the SREBP-1a expressing plasmid. A further stimulation of FAS promoter activity was observed when the Sp1 plasmid was additionally added. To more carefully evaluate the roles of CBF/NF-Y and Sp1, we performed a series of experiments where the CBF/NF-Y expressing plasmids were added at increasing concentrations in the presence and absence of a fixed amount of the SREBP or Sp1 expressing constructs (Fig. 6). The results clearly document that CBF/NF-Y and SREBP synergistically activate the FAS promoter; however, a further stimulation was observed when Sp1 was added. Thus, Sp1 augments the stimulatory activity of SREBP and CBF/NF-Y.
We have used a similar co-transfection assay to evaluate the co-regulator requirements for SREBP activation and sterol regulation in the promoter for 3-hydroxy-3-methylglutaryl-coenzyme A synthase (31). In this promoter, three separate proteins: SREBP, CBF/NF-Y, and cAMP element-binding protein/ activating transcription factor are all simultaneously required for sterol regulation and appreciable activation in SL2 cells. This is similar to but distinct from the FAS promoter where the third unique third factor (in this case Sp1) is not essential for sterol regulation but does provide a further stimulation of promoter activity in the co-transfection assay.
The FAS proximal promoter harbors a compact arrangement of transcription factor-binding sites. The two key SREBP sites that we have identified overlap the two halves of a classic E-box at Ϫ65. This is the consensus site for classic bHLH DNAbinding factors such as upstream stimulatory factor (20). There are also the Sp1 and CBF/NF-Y sites that are analyzed here and another classic SREBP site at Ϫ150. The current results taken together with previous studies strongly suggest that the two tandem SREBP sites that flank the E-box and the CBF/ NF-Y site are the key elements for sterol regulation by the SREBPs.
USF binding to the E-box at Ϫ65 has been implicated in the activation of FAS by insulin signaling (20), but other studies suggest an involvement for SREBPs in this process (21,32). Because a mutation that alters SREBP binding and activation by sterol deprivation (mutant D) but leaves the E-box element unaltered is also defective for insulin/glucose activation (Fig.  7), our results are consistent with an involvement of SREBPs in this process. Additionally, the analysis of the NF-Y and Sp1 site mutants extend these observations and suggest that both of these co-regulatory factors are required for insulin/glucose activation. This is different for sterol regulation where the NF-Y site is crucial, whereas the Sp1 site was largely dispensable (Fig. 1). FIG. 7. Activation of the FAS promoter in primary hepatocytes by glucose and insulin. FAS promoter constructs containing sequences of either wild type (WT) or various mutations from Ϫ151 to Ϫ42 were linked to a basal promoter in the pGL2 construct as described above and in an earlier report (15). The sequence of the wild type FAS promoter is shown at the bottom, and the bases that were mutated in each construct are underlined and labeled similar to the labeling in Fig. 1. Each construct was tested for a response to glucose and insulin in primary hepatocytes as described under "Materials and Methods." Cells were cultured in 5.5 mM glucose in the absence of insulin (hatched bars) or 27.5 mM glucose in the presence of 0.1 unit/ml insulin (dotted bars) for 48 h. Luciferase activity is shown as relative activity with the value of the wild type FAS construct at 27.5 mM glucose with insulin as 100%. Values represent the means (Ϯ S.E.) of three independent experiments, each with either duplicate or triplicate transfections.
The sterol regulation experiments were performed in cultured cells where the 1a isoform is more prevalent, and the insulin/glucose activation studies were performed in hepatocytes where 1c is the most abundant SREBP isoform expressed. This suggested that activation of the FAS promoter by SREBP-1a may require NF-Y predominantly and not Sp1, whereas activation by SREBP-1c may require both NF-Y and Sp1. The two SREBP-1 isoforms differ as a result of alternative promoter usage and unique mRNA splicing (30). The effect is that the amino-terminal region of 1c lacks 29 amino acids that are present in 1a and has 5 additional unique residues. This region of SREBP-1a contains its transcription activation domain, which has been shown to interact with the CBP/p300 co-activator family (33) as well as with the more recently identified ARC/DRIP multi-subunit complex (34). Native SREBP-1c is a much weaker transcriptional activator than 1a (30). It does not efficiently interact with CBP/p300, 2 and it interacts with only a subset of the ARC/DRIP components. 3 Thus, efficient activation by SREBP-1c would be predicted to require additional factors relative to SREBP-1a, and our data are consistent with this view. To directly investigate this possibility we performed activation experiments in Drosophila SL2 cells where we can directly manipulate the identity and functional levels of the individual transcription factors (Fig. 8). The results of these experiments are consistent with the results from the mammalian cell transfection experiments and indicate that efficient activation of the FAS promoter by SREBP-1a is largely dependent on NF-Y and not Sp1, whereas SREBP-1c requires both co-regulatory factors to efficiently activate the FAS promoter.
Our data are also consistent with a model where the amino termini of the different SREBP-1 isoforms differentially affect protein-protein interactions between SREBP-1 and NF-Y or Sp1 that may alter DNA binding. However, we have previously shown that SREBP-1 and NF-Y interact directly in solution. This effect required prior assembly of the NF-Y factor into its heterotrimeric complex and only required a small region of the SREBP-1 protein including its DNA-binding domain (12). The amino-terminal region where SREBP-1a and -1c differ was totally dispensable for the interaction. In the LDL receptor promoter we have also demonstrated that SREBP stimulates Sp1 to bind the LDL receptor promoter DNA, but this effect only requires the SREBP DNA-binding domain as well (11). It also required only a small portion of the Sp1 protein surrounding its DNA-binding domain. However, for transcriptional synergy in cells additional domains of both SREBP and Sp1 were required (35). Thus, synergy occurs at two steps: one is at the level of DNA binding, and the other is after DNA binding where the transcriptional activation domains of each protein are essential. Based on the fact that the in vitro protein-protein interaction between NF-Y and SREBP and the in vitro interaction between Sp1 and SREBP only require the DNA-binding domain of SREBP, we think it is likely that a similar two-step model for transcriptional synergy is also likely in the FAS promoter.
The differential recruitment of co-regulators by the SREBPs and mechanisms that specifically target their interactions with SREBPs could provide at least a partial basis for selective activation of FAS and fatty acid metabolism independently from cholesterol synthesis. In this regard, we have recently demonstrated that the yin yang 1 protein (YY1) can selectively down-regulate the LDL receptor promoter through targeting the gene-specific interaction between Sp1 and SREBP that is required to activate the LDL receptor promoter (36).