Transcriptional Activation of the Stearoyl-CoA Desaturase 2 Gene by Sterol Regulatory Element-binding Protein/Adipocyte Determination and Differentiation Factor 1*

To identify genes that are transcriptionally activated by sterol regulatory element-binding proteins (SREBPs), we utilized mRNA differential display and mutant cells that express either high or low levels of transcriptionally active SREBP. This approach identified stearoyl-CoA desaturase 2 (SCD2) as a new SREBP-regulated gene. Cells were transiently transfected with reporter genes under the control of different fragments of the mouse SCD2 promoter. Constructs containing >199 base pairs of the SCD2 proximal promoter were activated following incubation of cells in sterol-depleted medium or as a result of co-expression of SREBP-1a, SREBP-2, or rat adipocyte determination and differentiation factor 1 (ADD1). Electromobility shift assays and DNase I footprint analysis demonstrated that recombinant SREBP-1a bound to a novel cis element (5′-AGCAGATTGTG-3′) in the proximal promoter of the SCD2 gene. The finding that the endogenous SCD2 mRNA levels were induced when wild-type Chinese hamster ovary fibroblasts were incubated in sterol-deficient medium is consistent with a role for SREBP in regulating transcription of the gene. These studies identify SCD2 as a new member of the family of genes that are transcriptionally regulated in response to changing levels of nuclear SREBP/ADD1. In addition, the sterol regulatory element in the SCD2 promoter is distinct from all previously characterized motifs that confer SREBP- and ADD1-dependent transcriptional activation.

derived from two SREBP genes (1). SREBP-1a and -1c mRNAs result from alternative splicing of exon 1 of the SREBP-1 gene (1). Adipocyte determination and differentiation factor 1 (ADD1), the rat homologue of SREBP-1c, was cloned independently as a result of studies aimed at identifying a nuclear protein that bound the E-box motif in the promoter of the fatty acid synthase gene (2). ADD1 expression was shown to be required for adipogenesis and for the induction of a number of mRNAs, including fatty acid synthase and glycerol-3-phosphate acyltransferase (2)(3)(4), that occurs during this differentiation process.
SREBPs/ADD1 are synthesized as 125-kDa proteins that contain two transmembrane domains that anchor the proteins to the endoplasmic reticulum (1). When cellular sterol levels are low, two distinct proteolytic events release a mature 68-kDa NH 2 -terminal fragment of SREBPs/ADD1, which then translocates to the nucleus, binds to the promoters of target genes, and activates transcription (1). Conversely, when levels of cellular sterols are high, proteolytic processing of SREBPs/ ADD1 is diminished, nuclear levels of the mature proteins decline, and transcription of target genes is low (1).
The mature, transcriptionally active SREBPs/ADD1 contain basic-helix-loop-helix-leucine zipper (bHLH-Zip) domains that are required for DNA binding and transactivation (1,2). bHLH-Zip domains are found in a number of other transcription factors, including Myc and Max, that bind to E-box motifs (CANNTG) (12). In contrast, ADD1/SREBPs bind to both E-box and non-E-box motifs (13). The ability of ADD1/SREBPs to bind to non-E-box motifs is a result of an atypical tyrosine replacing an arginine that is conserved in the basic domain of all other bHLH-containing proteins (13).
Promoters from a number of SREBP/ADD1-responsive genes have been analyzed to identify the nucleotide sequences necessary for SREBP/ADD1-DNA interaction. The results indicate that the nucleotide sequence of the different sterol response elements (SREs), vary considerably; many, but not all SREs, contain two half-sites (direct repeats of C/TCAC) separated by two nucleotides (1,7). However, a number of SREs do not conform to this sequence motif (5)(6)(7).
A number of mutant cell lines have been isolated that do not regulate the expression/transcription of the HMG-CoA reductase, HMG-CoA synthase, and the LDL receptor genes in response to altered cholesterol levels (14 -19). Such mutants include sterol regulatory defective (SRD) cell lines that constitutively express high (SRD-1, SRD-2) or low (SRD-6) levels of nuclear SREBPs irrespective of the sterol status of the medium or of the cells (14 -16). Subsequently, it was shown that the 5Ј part of the SREBP-2 gene, that contains the bHLH-Zip domains, had undergone a translocation event in SRD-1 and SRD-2 cells. This resulted in the synthesis of distinct fusion proteins that contain no transmembrane sequences and were translocated to the nucleus where they transactivate target genes (16). In contrast, SRD-6 cells produce little or no mature SREBPs as a result of a defect at the site-2 cleavage (20). As a result, the mRNA levels of a number of SREBPresponsive genes are extremely low, even when SRD-6 cells are incubated in the absence of sterols (15,21).
We hypothesized the existence of novel genes and/or known genes not previously ascribed to be under the transcriptional control of SREBPs and cellular sterols. To identify such genes, we utilized mRNA differential display (22), using RNAs isolated from SRD-2 and SRD-6 cells.
This approach identified a number of transcripts expressed at high levels in SRD-2 cells as compared with SRD-6 cells. Among these, we identified stearoyl-CoA desaturase 2 (SCD2), an enzyme involved in the synthesis of unsaturated fatty acids. SCD2 catalyzes the ⌬ 9 -cis desaturation of fatty acyl-CoAs such as stearoyl-CoA and palmitoyl-CoA to produce oleic and palmitoleic acid, respectively (23,24). The current studies demonstrate that a gene (SCD2) encoding an enzyme involved in fatty acid desaturation is transcriptionally regulated by SREBPs/ ADD1 or sterols. The finding provides a plausible explanation as to why certain mutant cells that are defective in SREBP processing require not only cholesterol for growth but also unsaturated fatty acids (17).

EXPERIMENTAL PROCEDURES
Materials-DNA modification and restriction enzymes were obtained from Life Technologies, Inc. 32 P-and 35 S-labeled nucleotide triphosphates were obtained from Andotok Life Sciences Co. pCI-SREBP-1a encoding amino acids 1-490 of SREBP-1a and pCI-SREBP-2 encoding amino acids 1-485 of SREBP-2 were kindly provided by Dr. J. Ericsson. The 403 amino acids corresponding to the amino-terminal of ADD1 were cloned into pSV-SPORT (Life Technologies, Inc.) (11). The tyrosine at amino acid position 320 of ADD1 was mutated to an arginine to produce ADD1-R. ADD1-R was also cloned into pSV-SPORT (11,13). Lipoprotein-deficient serum (LPDS) was purchased from PerImmune. The HMG-CoA synthase reporter gene (pSYNSRE) and all other reagents have been described elsewhere (4,5,21).
Oligonucleotides-Oligonucleotides were synthesized by Life Technologies, Inc.
mRNA Differential Display-Total RNA was isolated from SRD-2 and SRD-6 cells as described (21). Fifty g of total RNA was DNase I-treated to remove any contaminating DNA (MessageClean, Genhunter Corp.). DNA-free RNA (0.2 g) was reverse-transcribed utilizing a T 11 G primer. PCR was then performed using the T 11 G primer, a second primer, and [ 33 P]dATP as described (RNAimage, Genhunter Corp.). The products were separated on a 6% denaturing polyacrylamide gel and exposed to x-ray film for 20 h. Differentially displayed bands were excised from the gel, eluted, and reamplified by PCR.
DNA Sequencing-DNA fragments generated by PCR were subcloned into the pCR2.1-TOPO cloning vector (Invitrogen) as described by the supplier. Escherichia coli, strain DH5␣, were transformed and selected for growth on ampicillin-containing medium. ␤-Galactosidasedeficient colonies were screened for the presence of inserts following EcoRI digestion of the plasmid DNA and electrophoretic analysis on 1% agarose gels. Inserts were sequenced by the Sanger chain-termination method using the T7 Sequenase v2.0 kit (Amersham Pharmacia Biotech).
SCD2 Promoter-Reporter Gene Constructs-The 5Ј end of the SCD2 gene (nt Ϫ588 to ϩ81) was obtained by polymerase chain reaction using C57Bl/6J mouse genomic DNA as a template. The 5Ј primer contained nucleotides that corresponded to Ϫ588 to Ϫ567 of the published SCD2 genomic sequence (24) and additional nucleotides (5Ј-TACCCTCGAG-3Ј) containing an XhoI restriction site. The 3Ј primer contained nucleotides that corresponded to ϩ81 to ϩ60 of the published genomic sequence (24) and additional nucleotides (5Ј-ACTTAAGCTT-3Ј) containing an HindIII restriction site. The PCR product was cloned into the pGL2 basic vector, utilizing XhoI and HindIII restriction sites to produce pSCD2-588. This 670-bp sequence of the SCD2 gene (Ϫ588 to ϩ81) was identical to that reported by Kaestner et al. (24). This DNA was used as template for additional PCRs that utilized 5Ј primers corresponding to Ϫ199 to Ϫ167, Ϫ150 to Ϫ129, and Ϫ110 to Ϫ89 of the SCD2 genomic sequence in combination with the common 3Ј primer. These reactions generated a series of SCD2 promoter fragments that were cloned into the pGL2 basic vector to produce pSCD2-588, pSCD2-199, pSCD2-150, and pSCD2-110, respectively. Oligonucleotides (28mers) were used to introduce mutations into pSCD2-199 using QuikChange (Stratagene). The resulting reporter genes each contained three A to C mutations at nt Ϫ151, Ϫ148 and Ϫ146 (pSCD2-199mut A) or at nt Ϫ157 to Ϫ155 (pSCD2-199mut B), respectively. The promoter of each construct was sequenced to confirm sequence.
Northern Blot Analysis-Total RNA was isolated from SRD-1, SRD-2, SRD-6, and CHO cells using Trizol reagent (Life Technologies, Inc). Ten g of each RNA sample was fractionated by agarose/formaldehyde gel electrophoresis, transferred to nylon membranes, and crosslinked by exposure to UV light as described (26). [␣-32 P]dCTP-radiolabeled DNA probes were generated by the random priming labeling method (Amersham Pharmacia Biotech). Hybridization and quantification, using a PhosphorImager (Molecular Dynamics), were as described (5). To correct for differences in RNA loading in each lane, blots were also hybridized to a probe, 36B4, that hybridized to a constitutively expressed mRNA.
Transient Transfections and Reporter Gene Assays-Transient transfections were performed using minor modifications of the transfection MBS kit (Stratagene). Details of the transient transfection of HepG2 cells with luciferase reporter genes, a plasmid encoding ␤-galactosidase, pCI-SREBP-1a (encoding mature SREBP-1a) pCI-SREBP-2 (encoding mature SREBP-2), appear in previous publications (5,25). Following transfection, cells were incubated for 20 h in medium containing 10% LPDS supplemented with either 5 M mevinolin (inducing medium) or sterols (10 g/ml cholesterol and 1 g/ml 25-hydroxycholesterol) (repressing medium). The cells were lysed and the luciferase activities determined and normalized to the ␤-galactosidase activity to correct for minor differences in transfection efficiency (5).
Electromobility Shift Assays-Double-stranded DNA fragments corresponding to Ϫ588 to Ϫ278, Ϫ299 to Ϫ239, Ϫ259 to Ϫ178, Ϫ199 to Ϫ130, Ϫ150 to Ϫ91, Ϫ110 to ϩ1, and Ϫ259 to Ϫ91 of the murine SCD2 proximal promoter were generated by PCR. The primers used to generate these SCD2 DNA fragments were flanked by the same 10 nucleotides, containing either an XhoI or HindIII restriction site, as described above. Single-stranded DNA (Ϫ164 to Ϫ137) containing wildtype or mutated sequences were annealed and the double-stranded DNA isolated and used in EMSAs (5). Recombinant SREBP-1a was purified from E. coli extracts as described previously (5). The DNA was end-labeled with 32 P (20,000 cpm; 1.5 fmol) and used in EMSAs as described (5).
DNase I Footprint Analysis-A DNA fragment corresponding to nucleotides Ϫ259 to Ϫ91 of the proximal SCD2 promoter (24) was used for DNase I footprinting. The fragment was cloned into the pCR2.1-TOPO cloning vector (Invitrogen). Purified plasmid DNA was cut with either BamHI or NotI and then end-labeled with [␥-32 P]dATP. The labeled, linearized plasmid was digested with the reciprocal restriction enzyme, either NotI or BamHI, to liberate a fragment containing nucleotides Ϫ259 to Ϫ91, end-labeled at either end. The gel-purified, end-labeled fragments were incubated with purified SREBP-1a and treated with DNase I as described (5).

mRNA Differential Display Identifies a Transcript That Is Expressed at Higher Levels in SRD-2 Cells as Compared with SRD-6 Cells-
The mRNAs for a number of SREBP-responsive genes, including the LDL receptor, HMG-CoA synthase, HMG-CoA reductase, and farnesyl diphosphate synthase genes are expressed at 5-10-fold higher levels in SRD-1 and SRD-2 cells as compared with SRD-6 cells (14,15,21). We hypothesized that the mRNA levels of other SREBP-responsive genes would also be differentially expressed in these mutant cells. Fig. 1 shows the results obtained using the technique of mRNA differential display utilizing the arbitrary upstream primer AP9, the downstream primer T 11 G, and RNA isolated from SRD-2 and SRD-6 cells. The portion of the gel shown contains a 700-bp band that was present when SRD-2 cells were the source of RNA but was absent when SRD-6-derived RNA was used (Fig. 1, arrow). This band was excised from the gel, reamplified by PCR, subcloned, and sequenced. The sequence revealed a Ͼ97% identity with the 3Ј-untranslated region of mouse stearoyl-CoA desaturase 2 (SCD2). The 700-bp insert was radiolabeled and used as a probe in a Northern blot analysis (Fig. 2). The probe hybridized to a mRNA of approximately 5 kilobases that was expressed at higher levels in SRD-1 and SRD-2 cells as compared with SRD-6 cells (Fig. 2). These latter studies confirm that the band identified by mRNA differential display corresponds to a mRNA that is more abundant in SRD-2 than in SRD-6 cells. Using the same methodology we have identified a number of other differentially regulated genes. 2 Some, but not all, of these additional mRNAs are also expressed at higher levels in SRD1 cells as compared with SRD2 cells. 2 The reason for this differential expression of mRNAs in the two cell types (SRD-1 and SRD-2) that constitutively overexpress different SREBP fusion proteins is unknown.
SCD2 mRNA Levels Are Regulated in Response to Sterols-Increased expression of SCD2 mRNA in SRD-1 and SRD-2 cells (Fig. 2) might be a direct result of the elevated levels of transcriptionally active SREBP in these cells (16). Alternatively, the effect may either be secondary, resulting from a cascade of effects initiated by elevated nuclear SREBP, or nonphysiological, because of the supraphysiological levels of transcription-ally active SREBP-2 fusion proteins. To distinguish between these possibilities, wild-type CHO cells were incubated in medium supplemented with 10% LPDS and either mevinolin (inducing conditions) or sterols (repressing conditions). These conditions are known to regulate the nuclear localization of SREBPs and to result in changes in the mRNA levels of SREBP-regulated genes (1). The results in Fig. 3 demonstrate that the SCD2 mRNA levels were regulated 4.2-fold by the sterol status of the cells; the levels were elevated when cells were sterol-deprived and repressed when sterols were added to the medium.
Regulation of SCD2 Promoter-Reporter Genes by Sterols and Co-expressed SREBPs-We next sought to determine whether expression of the SCD2 promoter-reporter genes were regulated in response to changes in the levels of either nuclear SREBPs or cellular sterols. HepG2 cells were utilized, because this cell type exhibits sterol-regulated expression of transiently transfected SREBP-responsive promoter-reporter genes (25). In contrast, sterol-regulated expression of many, if not all, such reporter genes is not observed in transiently transfected CHO cells (data not shown). For reasons that have yet to be determined, stably transfected CHO cells exhibit normal sterol-dependent regulation of these same SREBP-responsive reporter constructs (25).
Four different murine SCD2 promoter fragments were generated and cloned into a luciferase reporter vector (Fig. 4). Each SCD2 promoter-reporter construct was transiently transfected into HepG2 cells together with a plasmid encoding ␤-galactosidase under the control of a nonregulated, cytomegalovirusderived promoter. The cells were then incubated for 20 h in the absence or presence of sterols, lysed, and the luciferase activities determined. The ␤-galactosidase activity was used to correct for differences in transfection efficiency. Fig. 4 shows that incubation of cells in sterol-depleted medium resulted in a 5-12-fold increase in the activities of the reporter genes under the control of the proximal 588 or 199 bp of the SCD2 promoter. In contrast, reporter genes containing 150 or 110 bp of the proximal SCD2 promoter were unregulated and expressed at low levels (Fig. 4).
Taken together, the studies shown in Figs. 1-4 suggest that the transcription of the SCD2 gene is regulated by sterols and possibly by SREBPs. To determine the direct effect of SREBPs on transcription of the SCD2 promoter-reporter genes, the experiments illustrated in Fig. 5 were performed; HepG2 cells were transiently transfected with the indicated SCD2 promoter-reporter gene and plasmids that constitutively express either mature SREBP-1a or SREBP-2. The cells were then incubated for 20 h in the presence of excess sterols to prevent cleavage of endogenous SREBPs. Fig. 5A shows that pSCD2-588 was activated by co-expressed SREBP-1a or SREBP-2, in a dose-dependent manner. Maximal luciferase activities (Ͼ400fold increase) were obtained when the cells were transfected with 50 ng of plasmid encoding either SREBP-1a or SREBP-2 2 D. E. Tabor and P. A. Edwards, unpublished data.
FIG. 1. mRNA differential display identifies a mRNA that is expressed at a higher level in SRD-2 than in SRD-6 cells. DNAfree RNA (0.2 g) isolated from SRD-2 and SRD-6 cells was reversetranscribed in duplicate, utilizing a T 11 G primer. PCR was performed, using the 5Ј AP-9 primer and the T 11 G primer in the presence of [ 33 P]dATP, as described under "Experimental Procedures." PCR products were separated on a 6% polyacrylamide gel under denaturing conditions and the gel exposed to x-ray film for 20 h. The arrow indicates a band that corresponds to a transcript expressed at a higher level in SRD-2 than in SRD-6 cells.
FIG. 2. SCD2 mRNA levels are elevated in SRD-1 and SRD-2 as compared with SRD-6 cells. Ten g of total RNA, isolated from duplicate dishes of SRD-1, SRD-2, and SRD-6 cells, was separated on a 1% agarose/formaldehyde gel and transferred to a nylon membrane as described under "Experimental Procedures." The blot was hybridized with a radiolabeled probe generated from the cloned band indicated in Fig. 1 and then to a radiolabeled 36B4 probe. The SCD2 mRNA of approximately 5.0 kilobases and the control 36B4 RNA are indicated. (Fig. 5A). pSYNSRE, a luciferase reporter gene under the control of a promoter fragment of the HMG-CoA synthase gene and known to be regulated by SREBPs, was induced over 900-fold when cells were co-transfected with 50 ng of plasmid encoding either SREBP-1a or SREBP-2 (data not shown). The 400 -900fold increase in the activities of the SCD2 and HMG-CoA synthase promoter-reporter genes is much greater than that observed in response to changes in cellular sterol levels (Fig. 4). This is most likely the result of the supraphysiological levels of nuclear SREBP in cells that were co-transfected with plasmids encoding SREBP-1a or SREBP-2.
HepG2 cells were also transfected with each of the four SCD2 promoter-reporter genes in the absence or presence of lower levels (25 ng) of plasmid encoding SREBP-1a or SREBP-2 (Fig.  5B). The results show that the reporter genes, containing either 588 or 199 bp of the SCD2 promoter, were induced to similar high levels of expression by co-transfected SREBP-1a or SREBP-2 (Fig. 5B). The pSCD2-110 construct was unresponsive to co-expressed SREBPs (Fig. 5B). The pSCD2-150 construct was unresponsive to SREBP-1a but was stimulated by SREBP-2, although to a lower degree than that obtained with pSCD2-588 or pSCD2-199 (Fig. 5B). The reason for the different responsiveness of pSCD2-150 to co-expressed SREBP-1a and SREBP-2 is unknown.
Effect of Co-expression of a Dominant-negative SREBP-1a-To further investigate the requirement for SREBPs in the sterol-mediated induction of the SCD2 promoter-reporter constructs, we utilized a dominant-negative form of SREBP-1a (DN-SREBP-1a). As a result of deletion of the amino-terminal 90 amino acids, the DN-SREBP-1a is capable of binding DNA (data not shown), but is incapable of transcriptional activation. The data in Fig. 6 demonstrate that the increased expression of the pSCD2-588 construct, in response to depletion of media sterols, was prevented by co-expression of this dominant-negative form of SREBP-1a. The normal induction of pSYNSRE in response to sterol deprivation was also attenuated by co-expression of DN-SREBP-1a (Fig. 6), consistent with a functional role for SREBP in the transcriptional induction of the HMG-CoA synthase gene (1).
Electromobility Shift Assays and DNase I Footprint Analyses-Taken together, the studies illustrated in Figs. 1-6 demonstrate that the transcription of the SCD2 gene is regulated in response to changing levels of either cellular sterols or nuclear SREBPs. However, examination of the SCD2 proximal promoter sequence did not reveal a sequence with high identity to any previously characterized SRE.
To identify a potential SRE in the SCD2 promoter, EMSAs were performed with different fragments of the proximal SCD2 promoter and recombinant SREBP-1a. DNA fragments Ϫ588 to ϩ81, Ϫ299 to ϩ81, Ϫ259 to ϩ81, Ϫ199 to ϩ81 and Ϫ199 to Ϫ130 all gave a single shifted complex with recombinant SREBP-1a (data not shown). In contrast, fragments Ϫ299 to Ϫ239, Ϫ259 to Ϫ178 and Ϫ110 to ϩ1 gave no shifted complex (data not shown). The results of these EMSAs indicated that SREBP-1a likely bound to nucleotides located between Ϫ178 and Ϫ110 of the SCD2 promoter. FIG. 3. SCD2 mRNA levels are repressed by sterols. RNA was isolated from duplicate dishes of Chinese hamster ovary cells incubated in media containing 10% LPDS supplemented with either sterols (ϩ) (10 g/ml cholesterol and 1 g/ml 25-hydroxycholesterol) or 5 M mevinolin in the absence of sterols (Ϫ) as described under "Experimental Procedures." RNA (10 g) was fractionated on a 1% agarose/formaldehyde gel, transferred to a nylon membrane and hybridized to the radiolabeled probes specific for stearoyl-CoA desaturase 2 or 36B4 cDNA.

FIG. 4. Regulation of SCD2 promoter-luciferase reporter genes by sterols.
Duplicate dishes of HepG2 cells were transiently transfected with 1 g of the indicated SCD2 promoter-reporter gene and a plasmid encoding ␤-galactosidase under the control of a cytomegalovirus promoter. The cells were incubated for 20 h in medium containing 10% LPDS with or without supplemented sterols, as described in the legend to Fig. 3. The normalized luciferase values, obtained from duplicate dishes, varied less than 10%. A value of 1 was given to the luciferase activity obtained with pSCD2-110 following incubation of cells in the sterol-supplemented repressing medium. The results are representative of three separate experiments.

FIG. 5. Stimulation of SCD2 promoter-reporter genes by coexpression of SREBP-1a or SREBP-2.
SCD2 promoter-reporter genes were used as described under "Experimental Procedures." One g of the SCD2 promoter-reporter construct was transfected into HepG2 cells in duplicate along with a plasmid encoding ␤-galactosidase. In A, cells were co-transfected with the indicated amount of an expression plasmid encoding mature SREBP-1a (Ⅺ) or SREBP-2 (q). In B, cells were co-transfected with plasmids encoding either SREBP-1a or SREBP-2 or carrier plasmid (Ϫ), as indicated. The cells were incubated for 20 h in medium supplemented with 10% fetal bovine serum and sterols. The cells were lysed, and the normalized luciferase activities were determined as described. The variation in the normalized luciferase activities between duplicate dishes was less than 10%. The results are representative of three separate experiments.
Consequently, a 169-bp fragment (Ϫ259 to Ϫ91) was endlabeled and utilized either in an EMSA (Fig. 7A) or subjected to DNase I digestion in the absence or presence of recombinant SREBP-1a (Fig. 7B). This 169-bp fragment gave a single shifted DNA⅐SREBP-1a complex (Fig. 7A) consistent with the presence of a single SREBP-1a binding site. The results of Fig.  7B show that SREBP-1a protected nucleotides Ϫ151 to Ϫ141 from DNase I digestion. Examination of the nucleotide sequence in the footprinted region does not reveal any identity to known sterol regulatory elements or to a consensus E-box motif, and we concluded that SREBP-1a bound to a novel and distinct sequence. The presence of a DNase I hypersensitive site at Ϫ140 (Fig. 7B) resulted in a decrease in the radioactivity associated with all faster migrating DNA species and made it difficult to delineate the 5Ј border of the DNase I protected region. To better define the nucleotides that are necessary for formation of the SREBP-1a⅐DNA complex, additional EMSAs were performed. These studies utilized a fragment (nt Ϫ164 to Ϫ137) of the SCD2 promoter that overlapped the footprinted sequence and contained either wild-type or mutant sequences (Fig. 7C). The results indicate that both the wild-type and mutant B (mutations at Ϫ155 to Ϫ157) oligonucleotides formed a single shifted complex with recombinant SREBP-1a (Fig. 7C). In contrast, no shifted complex was formed with the mutant A oligonucleotide (Fig. 7C). These results demonstrate that SREBP-1a binds to nucleotides that lie between Ϫ151 and Ϫ141 (5Ј-AGCAGATTGTG-3Ј) and that mutation of the three adenines prevents SREBP-1a-DNA interactions in vitro.
A Functional Role for the Novel SRE in the SCD2 Promoter- Fig. 8 shows the results obtained when HepG2 cells were transiently transfected with reporter constructs pSCD2-199, pSCD2-199mut A (mutations at Ϫ151, Ϫ148, and Ϫ146) or pSCD2-199mut B (mutations at Ϫ155 to Ϫ157).
The expression of pSCD2-199 was increased 3-6-fold when cells were either incubated in lipid-depleted medium or were co-transfected with ADD1 and then incubated with excess sterols to inhibit the proteolytic release of endogenous SREBPs (Fig. 8A). ADD1 activation of pSCD2-199 was dependent on the amount of pADD1 co-transfected into the cells (Fig. 8A). In contrast, the activity of pSCD2-199 was not increased following co-transfection of ADD1-R (Fig. 8A), a protein that binds to an E-box motif but is unable to bind to SRE1 (13).
The results of Fig. 8B demonstrate that all three reporter genes were expressed at similar levels in sterol-treated cells. Incubation of the cells in lipid-depleted medium resulted in a   FIG. 7. SREBP-1a binds to sequences in the proximal SCD2 promoter. In A, a 32 P-end-labeled probe (Ϫ259 to Ϫ91) was incubated in the absence or presence of recombinant SREBP-1a as indicated and analyzed in an EMSA as described under "Experimental Procedures." The volume of SREBP-1a was 0.5 l (lane 2) or 1 l (lane 3). The shifted SREBP-1a⅐DNA complex and the free probe are indicated. In B, the end-labeled probe (Ϫ259 to Ϫ91) was incubated in the absence or presence of SREBP-1a. DNase I digestion was performed, and the resulting products were separated by denaturing polyacrylamide gel electrophoresis. The presence of a DNase I protected region and corresponding nucleotide sequence are indicated. In C, end-labeled probes (Ϫ164 to Ϫ137) corresponding to wild-type (wt) sequence of the SCD2 promoter or containing three G to A mutations (mut B) or three A to C mutations (mut A) were used in EMSAs in the absence or presence of recombinant SREBP-1a (1 l). The shifted SREBP-1a⅐DNA complex and the free probe are indicated. The footprinted region, identified in Fig. 7B, is underlined.   FIG. 6. Co-expression of a dominant-negative SREBP-1a  3.6-fold increase in the expression of pSCD2-199 and pSCD2-199mut B (Fig. 8B). In contrast, pSCD2-199mut A was induced less than 1.4-fold under these conditions, and the maximal reporter gene activity was Ͻ15% of that observed with the control, pSCD2-199 (Fig. 8B). These results demonstrate that the three mutations introduced into the promoter to produce pSCD2-199mut A interfere with the normal sterol-regulated expression of the SCD2 promoter-reporter gene, whereas mutation of nucleotides at Ϫ155 to Ϫ157 (mutant B) have little or no effect on sterol-regulated expression of the reporter gene. DISCUSSION The current experiments demonstrate that transcription of the mouse stearoyl-CoA desaturase 2 gene is regulated in response to alterations in the levels of either exogenous sterols or the levels of nuclear SREBPs/ADD1. Our studies suggest that the increased expression/nuclear localization of SREBPs/ ADD1, following incubation of cells in sterol-depleted medium, is sufficient to enhance transcription of the SCD2 gene and results in increased SCD2 mRNA levels.
The current studies identify a novel sequence in the SCD2 proximal promoter that functions as an SRE. The nucleotide sequence (5Ј-AGCAGATTGTG-3Ј; Ϫ151 to Ϫ141) is distinct from previously described SREs. Specifically, this sequence does not contain direct repeats of C/TCAC that have been identified in some (1, 7), but not all (5-7), functional SREs, nor does it contain an E-box motif. However, two nonconsensus E-box motifs are located between Ϫ152 and Ϫ143 of the SCD2 promoter, viz. 5Ј-CAGCAG-3Ј and 5Ј-CAGATTG-3Ј. The nucleotides that vary from the classic E-box motif (CANNTG) are underlined. A number of bHLH-containing proteins, including Max/c-Myc, Max, c-Myc, and USF have been shown to bind to both canonical and noncanonical E-box motifs in vitro (27). The noncanonical motifs, identified by EMSAs in this latter study, include CACGAG and CACGTTG. However, it was not established whether these noncanonical motifs function to bind bHLH-containing proteins in vivo and consequently activate transcription. To determine whether the SCD2 promoter con-tained a functional noncanonical E-box between nt Ϫ151 and Ϫ141, we compared the effects of ADD1 and ADD1-R co-expression on the activation of pSCD2-199 (Fig. 8A). Mature ADD1 contains a tyrosine at amino acid position 320 and can bind to both SREs and E-box motifs (2,13). All other proteins that contain a bHLH and bind to E-box motifs have an arginine at the corresponding position (position 13 of the basic domain) (28). Conversion of the tyrosine at position 320 of ADD1 to an arginine results in the production of a protein (ADD1-R) that binds to an E-box motif but is unable to bind to SRE1 (13). Thus, the activation of a reporter gene by co-expressed ADD1 or ADD1-R can be used to determine, indirectly, whether the promoter contains a functional SRE and/or E-box motif. The studies of Fig. 8 demonstrate that co-expression of ADD1, but not of ADD1-R, results in transcriptional activation of pSCD2-199. Thus, we conclude that the sequence between Ϫ151 and Ϫ141 of the SCD2 promoter contains a novel SRE but that it does not contain a functional E-box.
In a recent study Athanikar and Osborne (29) demonstrated that replacement of the SRE1, in an LDL receptor promoterreporter gene, with an E-box motif resulted in the loss of sterol-regulated transcription. The mutant promoter was promiscuously activated without regard to the sterol content of the cell, presumably as the result of interaction of other nuclear proteins, that contain bHLH domains, with the E-box motif (29).
As indicated above, SREBPs/ADD1 often bind to two direct repeats (C/TCAC) that are separated by one nucleotide (1, 7). However, both the current and previous studies (5)(6)(7) indicate that other sequences that do not conform to this consensus sequence function as SREs in the promoters of various genes. A difference in the affinity of SREBP-1a, SREBP-1c/ADD1, or SREBP-2 for the various SRE motifs that have been identified may provide an additional level of transcriptional regulation that results in activation of distinct genes/pathways.
The observation that overexpression of SREBP-1a or SREBP-2 resulted in a weak, but definite, stimulation of an FIG. 8. Transcriptional regulation of wild-type and mutant SCD2 promoter-reporter genes in response to sterol depletion or co-expression of either ADD1 or ADD1-R. In A, duplicate dishes of HepG2 cells were transiently transfected with pSCD2-199 (1 g), a ␤-galactosidase expression plasmid (1 g), and the indicated amount of either pADD1 or pADD1-R. The cells were then incubated for 20 h in medium containing 10% LPDS in the absence (Ϫ) or presence (ϩ) of sterols (10 g/ml cholesterol and 1 g/ml 25-hydroxycholesterol). The relative luciferase activities were determined after normalization for minor variations in the efficiency of transfection. In B, duplicate dishes of cells were transiently transfected with 1 g of pSCD2-199, pSCD2-199mut A, or pSCD2-199mut B as indicated and the cells then incubated for 20 h in medium containing 10% LPDS in the absence (Ϫ) or presence (ϩ) of sterols. Normalization of luciferase activities was as described above in A. The values shown are the means of duplicate dishes (variation Ͻ10%) and are representative of two or more experiments. The fold changes in luciferase activities are indicated. SCD2 promoter-reporter gene that terminated at Ϫ150 (pSCD2-150) (Fig. 5B) may indicate that additional upstream cis elements and transcription factors are necessary for full activation of the reporter gene. Under more physiological conditions, such as when cells are incubated in sterol-depleted medium, the nuclear levels of SREBPs are insufficient to activate pSCD2-150, although they are sufficient to activate pSCD2-588, pSCD2-199 (Fig. 4) and pSYNSRE (data not shown).
mRNA levels of SCD1 are also induced during differentiation of preadipocytes to adipocytes (24,30). Kaestner et al. reported that the proximal promoters of SCD2 and SCD1 contained 146 bp with 77% sequence identity (24). The SRE identified in the current study in the promoter of the SCD2 gene lies within this conserved sequence and shows 10/11 identity with nucleotides Ϫ423 to Ϫ413 in the SCD1 promoter (24). In addition, the mouse genes encoding SCD1 and SCD2 co-localize on chromosome 19, possibly as a result of gene duplication (31). Thus, the observation that hepatic SCD1 mRNA levels were induced in transgenic mice that overexpress SREBP-1a (32) is consistent with the direct transcriptional activation of the SCD1 gene by SREBP-1a. Further studies will be required to determine whether transcription of the mouse SCD1 gene, in response to increased levels of mature SREBPs, is dependent on the motif that has 10/11 identity with the novel SRE identified in the current study and/or on a more distal sequence (Ϫ495 to Ϫ486) (24) that shows 9/10 identity with SRE1.
More than 16 years ago Chin and Chang (33) demonstrated that there was a time-dependent increase in the enzymatic activity of stearoyl-CoA desaturase (presumably SCD1 and SCD2) when Chinese hamster ovary cells were incubated in media that contained lipid-depleted serum. They noted that the increase in stearoyl-CoA desaturase activity was prevented when the medium was supplemented with medium that contained cholesterol (33). Subsequently, Chang and co-workers (17) isolated a mutant Chinese hamster ovary cell that both failed to induce cholesterol biosynthetic enzymes and the LDL receptor in response to cellular sterol depletion and was auxotrophic for both cholesterol and unsaturated fatty acids. Interestingly, revertants regained the capacity to synthesize cholesterol and unsaturated fatty acids concomitantly (17). Recently, Rawson et al. (34) used these mutant cells and complementation cloning to identify a putative metalloprotease that cleaves SREBPs at site-2. The current demonstration that SCD2 is transcriptionally regulated by nuclear SREBP/ADD1 provides an explanation for the coordinate regulation of cholesterol synthesis and fatty acid desaturation that was noted in these earlier studies.
In summary, we have identified stearoyl-CoA desaturase 2 as a new member of the family of SREBP/ADD1-responsive genes. To date, these include genes involved in cholesterol biosynthesis and uptake, fatty acid synthesis, triglyceride synthesis, and now fatty acid desaturation. Sequence analysis of the SCD2 proximal promoter alone would not have predicted that it contained an SRE. The studies reported here demon-strate that the mRNA differential display technique may prove useful in identifying novel genes that are regulated by sterols and are important in the development of human disease.