SREBP-1c and Sp1 interact to regulate transcription of the gene for phosphoenolpyruvate carboxykinase (GTP) in the liver.

The sterol regulatory element-binding protein-1c (SREBP-1c), as well as SREBP-1a and SREBP-2, inhibit transcription of the gene encoding the cytosolic form of phosphoenolpyruvate carboxykinase (GTP) (PEPCK-C). There are two SREBP regulatory elements (SREs) in the PEPCK-C gene promoter (-322 to -313 and -590 to -581). The SRE at -590 overlaps an Sp1 site on the opposite strand of the DNA. These SREs bound SREBP-1a and SREBP-1c with low affinity but the addition of purified upstream stimulatory activity enhanced the binding of SREBP-1 to both of these sites. Mutating these SREs increased both unstimulated (5-fold) and protein kinase A-stimulated transcription (8-27-fold) from the PEPCK-C gene promoter; this was lost when both SREs were mutated. The SRE at -590 differs by a single base pair from the SRE in the low density lipoprotein (LDL) receptor gene (T in the PEPCK-C gene promoter at -582, compared with an A in the SRE of the gene for the LDL receptor promoter). Introduction of the LDL receptor SRE into the PEPCK-C gene promoter increased SREBP-1c binding and caused a 10-fold enhancement of basal transcription from the promoter, rather than an inhibition as observed with the SRE in the PEPCK-C gene promoter. The T/A change does not alter the binding of Sp1 to its site on the opposite strand of the DNA. Sp1 bound to the promoter independently of SREBP-1c but competed with SREBP-1c for binding. Sp1 does not bind to the SRE at -322. Chromatin immunoprecipitation analysis, using rat hepatocytes, demonstrated that SREBP-1 and Sp1 were associated in vivo with putative regulatory regions corresponding to the SREs in the PEPCK-C gene promoter. We propose that insulin represses transcription of the gene for PEPCK-C by inducing SREBP-1c production in the liver, which interferes with the stimulatory effect of Sp1 at -590 of the PEPCK-C gene promoter.

involved in the inhibitory effect of insulin on transcription from the PEPCK-C gene promoter. There is considerable evidence that the SREBP family of transcription factors also control the expression of genes that code for proteins important in lipid metabolism. These include the LDL receptor (3,4), HMG-CoA reductase (5), fatty acid synthase (6), and acetyl-CoA carboxylase (7). SREBP-1c thus regulates both lipid and carbohydrate metabolism, most likely via a reciprocal response to the hormones that control the genes involved in lipogenesis and gluconeogenesis.
SREBPs are members of the basic helix-loop-helix leucine zipper family of transcriptional regulatory proteins, having a unique dual binding specificity. They can bind to SREBP regulatory elements (SRE) (3) or to an E-box motif (8). There are two isoforms of SREBP-1, SREBP-1a and -1c, which are produced from a single gene on human chromosome 17p11.2, and differ only in their first exon (9). SREBP-2 is produced from a separate gene on human chromosome 22q13 (see Ref. 10 for review). SREBPs are synthesized as precursors and inserted into the membranes of the endoplasmic reticulum and nuclear envelope through a two-pass transmembrane segment, resulting in the amino and carboxyl tails facing the cytoplasm (11). The mature form of SREBP-1 is generated by proteolytic cleavage within the plane of the bilayer; this cleavage responds to the concentration of sterols or insulin (11,12) and results in the nuclear translocation of the mature transcription factor. The two isoforms of SREBP-1 are produced from two promoters on the SREBP-1 gene, each of which generates an mRNA with a different first exon that codes for one of the specific amino termini of 1a or 1c. These alternate exons are attached to the rest of the mRNA during splicing, yielding two isoforms of SREBP-1.
Our previous findings demonstrated an interaction between SREBP-1c and CBP. The latter is a transcriptional co-activator involved in coordinating the stimulatory effect of cAMP on PEPCK-C gene transcription (1). SREBP-1c blocks the stimulation of PKA-induced transcription from the PEPCK-C gene promoter, whereas the dominant negative form of SREBP-1c overcomes this inhibition. Both E1A and NF1c, which interact with CBP, have been shown to inhibit PEPCK-C gene transcription; dominant-negative SREBP-1c also overcomes this inhibition (1). Previous studies have shown that CBP is required for the sterol/SREBP-1-regulated increase in transcription from the promoters of the HMG-CoA synthase and HMG-CoA reductase genes in transient transfection assays (13,14). E1A blocked this induction of transcription by SREBP-1c, but it could be overcome by the coexpression of CBP and SREBP-1c. SREBP-1c binds to the promoter of the target gene and then to the amino-terminal domain (amino acids 1-451) of CBP (13). Our findings suggested a similar interaction between SREBP-1c, the PEPCK-C gene promoter, and CBP (1) but one that results in repression of transcription from the PEPCK-C gene promoter. In the current paper we have identified the binding sites of SREBP-1 in the PEPCK-C gene promoter and have shown that SREBP-1c requires a protein(s) present in USA for binding to the SRE in the gene promoter. Sp1, a ubiquitous transcription factor, prevents SREBP-1 from binding to the SRE in the PEPCK-C gene promoter at Ϫ590. When the levels of SREBP-1 increases in the liver, as would be expected in the presence of insulin, it displaces Sp1 from its site on the promoter, thereby inhibiting transcription. However, when the SRE in the PEPCK-C gene promoter is mutated to the SRE sequence present in the LDL receptor promoter, both Sp1 and SREBP-1c bind to the SRE resulting in a stimulation of transcription, not an inhibition.

Materials
Luciferase assay reagents and the cell culture lysis reagent were purchased from Promega Corp. (Madison, WI). The QuikChange TM site-directed mutagenesis kit was from Stratagene (La Jolla, CA). The Slid-A-Lyser dialysis cassette was from Pierce. HepG2 cells were originally purchased from ATCC (Manassas, VA) and the WT-IR cells were a generous gift from Dr. Domenico Accili, Columbia University, New York. Dulbecco's modified Eagle's medium/Ham's F-12, ␣-minimal essential medium cell culture medium, and fetal calf serum were from Invitrogen, Inc. (Rockville, MD) and the restriction enzymes, T 4 DNA ligase, DNA polymerase I (large fragment), and Klenow fragment were purchased from New England Biolabs, Inc. (Beverly, MA). Anti-SREBP-1 antibody and Protein G PLUS-agarose beads were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The Ni-NTA-agarose beads, the Plasmid Mini and Qiafilter Midi and Maxi Kits were from Qiagen, Inc. (Valencia, CA), whereas poly(dI-dC) was from Amersham Biosciences. FuGENE 6, kanamycin, and chloramphenicol were purchased from Roche Applied Science. Fluorescent double-tagged and cold oligonucleotide primers (with 56-FAM at the 5Ј end and 36-FAM at the 3Ј end of the fragment being labeled) were synthesized by Integrated DNA Technologies Inc. (Coralville, IA) and purified further by high pressure liquid chromatography or PAGE. The dc-Bio-Rad reagent was purchased from Bio-Rad and the protease inhibitor mixture was from Sigma. All other reagents used in this study were of the highest quality obtainable. The expression vectors encoding the full-length CBP and the fragments of CBP were generous gifts from Dr. Richard Goodman, Vollum Institute, Portland, OR, and the expression vector encoding the catalytic subunit of protein kinase A was kindly provided by Dr. Masa-Aki Muramatsu. The cDNA encoding both normal and the dominant negative forms of SREBP-1c were generous gifts from Dr. Bruce Spiegelman, Harvard University, Boston, MA. The expression vector encoding SREBP-2 was purchased from ATCC. The histidine (His)tagged recombinant vector encoding SREBP-1a was a gift from Dr. Timothy Osborne, University of California, Irvine, CA, whereas the His-tagged recombinant vector encoding SREBP-1c was kindly provided by Dr. Pascal Ferre, INSERM, Paris, France. Dr. Robert Tjian generously provided the Sp1 used in these studies. USA was purified as described by Meisterernst et al. (15).

Methods
Vector Construction-The plasmid p2000luc was generated by ligating the XbaI/BglII fragment of the PEPCK-C gene promoter into pGL2basic (Promega) (which contains the luciferase structural gene) that had been digested with NheI and BglII. The plasmid was isolated with Plasmid Mini or Qiafilter Midi Kits (Qiagen, Inc.), and checked for purity by electrophoresis using an 0.8% agarose gel, followed by ethidium bromide staining and visualization with UV light. The cDNA encoding the mature form of SREBP-1a (first 490 amino acids of the NH 2 terminus) was cloned into the pRSET-B expression vector (16), whereas the cDNA encoding the mature form of SREBP-1c (first 446 amino acids of the NH 2 terminus) was cloned into the pQE30 expression vector.
Cell Culture and Transient Transfection Studies-HepG2 hepatoma cells were grown to 60% confluence in Dulbecco's minimal essential medium, supplemented with 50% F-12 nutrient mixture (Dulbecco's modified Eagle's medium/Ham's F-12 and 10% fetal calf serum) and antibiotics (50 units of penicillin and 50 g of streptomycin per ml) at 37°C in an atmosphere of 95% air, 5% CO 2 in 6-well plates for 24 h before transfection. WT-IR cells were passaged at 33°C in ␣-minimal essential medium supplemented with 4% fetal calf serum, 2 mM glutamine, and 10 M dexamethasone. After transfection, the cells were transferred to 37°C, with culture conditions as described above. The cells (1 ϫ 10 5 cells in 2 ml of medium/well) were then transfected with plasmid DNA (1 g of plasmid DNA to 4.5 l of FuGENE-6 transfection reagent per 35-mm well). After 24 h, the cells were washed with ice-cold 1ϫ phosphate-buffered saline, pH 7.4, and lysed by addition of 300 l of 1ϫ lysis buffer. The insoluble material was pelleted by centrifugation in 1.5-ml Eppendorf tubes for 6 min at 12,000 ϫ g and the supernatant was separated from the pellet for the measurement of the protein concentration and luciferase activity. For luciferase activity, 10 l of the cell lysate was used to measure the integrated light units over 10 s, using the luciferase assay system (Promega) and a luminometer (Tropix, Inc., Bedford, MA), as recommended by the manufacturer. This assay was routinely carried out in triplicate. The protein content of the extracts was determined by the dc-Bio-Rad protein assay method, using several concentrations of bovine serum albumin in cell lysis buffer as a standard.
Site-directed Mutagenesis-The SREs at Ϫ322 and Ϫ590 were mutated using the following oligonucleotides: for Ϫ322 the sequence was 5Ј-GGCGTCCCGGCCAGCCCTGTCCTTGACCTTTGTTGTGTAATT-AAGGCAAGAGCCTATAGTTTGCATCAGC-3Ј; for Ϫ590 the sequence was 5Ј-GGAATGAAGCTTACTTTTTTTTTTTCCTCTGC-3Ј (the mutated bases are in bold). These sites were mutated in combination or individually to generate single mutant vectors or double mutant vectors using the QuikChange TM mutagenesis kit. To generate Ϫ322sre and Ϫ590sre the PEPCK-C core SRE sequence at Ϫ322 and Ϫ590 was replaced by the consensus SRE sequence (5Ј-ATCACCCCAC-3Ј) found in the LDL receptor gene promoter. The plasmids resulting from the above mutations were isolated, and their sequence confirmed by restriction digestion and DNA sequencing, performed by the Case Western Reserve University Molecular Biology Core Laboratory.
Purification of Recombinant Proteins-Bacterial stocks containing the plasmids encoding the desired recombinant protein were inoculated in 20 ml of LB broth containing 100 g/ml ampicillin and 25 g/ml kanamycin or 100 g/ml ampicillin and 34 g/ml chloramphenicol, and shaken vigorously overnight at 37°C. One liter of LB media was then inoculated with the 20 ml of bacterial culture and grown at 37°C until A 600 reached 0.6. Isopropyl-1-thio-␤-D-galactopyranoside was added to culture and it was further shaken for 4 h. The cells were centrifuged at 4,000 ϫ g for 20 min, and then resuspended in lysis buffer (50 mM NaH 2 PO 4 , pH 8.0, 300 mM NaCl, 10 mM imidazole, 30% glycerol, 20 mM ␤-mercaptoethanol, 0.2 M phenylmethylsulfonyl fluoride, and 1 ml/20 g wet weight Escherichia coli cells of protease inhibitor mixture for Histagged protein). Lysozyme, 1 mg/ml, was added to the resuspended cells and the sample was placed on ice for 30 min. The cells were then sonicated using 10-s bursts at 200 -300 W with a 10-s cooling period between bursts. The lysate was centrifuged at 10,000 ϫ g for 30 min. To the supernatant, a 1:4 ratio of 50% Ni-NTA slurry was added and the mixture incubated overnight at 4°C. After incubation, the Ni-NTA/ protein slurry was centrifuged at 750 -800 ϫ g for 5 min, a wash buffer (50 mM NaH 2 PO 4 , pH 8.0, 300 mM NaCl, 20 mM imidazole, 30% glycerol, and 20 mM ␤-mercaptoethanol) was added, and this step was repeated 8 times using the wash buffer. The pellet was re-suspended in elution buffer (50 mM NaH 2 PO 4 , pH 8.0, 300 mM NaCl, 250 mM imidazole, 30% glycerol, and 20 mM ␤-mercaptoethanol) and incubated at room temperature for 25 min. The beads were pelleted by centrifugation at 750 -800 ϫ g at 4°C for 15 min and the supernatant was collected and dialyzed overnight against 3 liters of the dialysis buffer (50 mM NaH 2 PO 4 , pH 8.0, 300 mM NaCl, 30% glycerol, and 20 mM ␤-mercaptoethanol); the buffer was changed at least once during this period. The purity of the protein was determined by electrophoresis through a 12% SDS-PAGE gel. The proteins were stored at Ϫ70°C until use. consisting of 5Ј-56-FAM/GGGGTGTGTTACCCCACTAGGTGTC/36-F-AM-3Ј; and Ϫ590 (Ϫ597/Ϫ573) consisting of 5Ј-56-FAM/GCTTACAAT-CACCCCTCCCTCTGCA/36-FAM-3Ј. The 590sre oligonucleotide had the same sequence as Ϫ590 except for a change in one base pair (bold A) and consisted of 5Ј-56-FAM/GCTTACAATCACCCCACCCTCTGCA/36-FAM-3Ј; this segment of DNA corresponded to the SRE-1 in the LDL receptor. An unlabeled oligonucleotide, which corresponds to an Sp1 binding site, had the sequence 5Ј-ACCCCCGGCCCGCCCCGCCGTCG-3Ј. The reaction mixture (10 l) for the EMSA contained 5 l of a 5ϫ binding buffer (20% glycerol, 5 mM MgCl 2 , 2.5 mM EDTA, 2.5 mM dithiothreitol, 250 mM NaCl, 50 mM Tris-HCl, pH 7.5, and 0.25 mg/ml of poly(dI-dC)⅐(dI-dC)), 0.2 g of poly(dI-dC)⅐(dI-dC), and 1 pmol of the doubly tagged fluorescent oligonucleotide with or without competitor. The reaction was carried out at 25°C for 20 min and the product of the reaction was resolved by electrophoresis using a 4% non-denaturing polyacrylamide gel. After electrophoresis, the gel was scanned and resolved wet, using a Typhoon 9200 PhosphorImager® scanner (Amersham Biosciences). The image was analyzed by ImageQuant® Solutions software for Windows 2000.
Chromatin Immunoprecipitation Assay (ChIP Assay)-Rats (200 g body weight) were fasted overnight and then fed a diet high in carbohydrates for 7 days. The animals were then killed, their livers removed, and primary hepatocytes prepared as described by Berry and Friend (17) and modified by Leffert (18). The ChIP assay was performed essentially as described by Boyd et al. (19), as modified by Massillon et al. (20). After plating for 1 day, formaldehyde was added to the culture medium to a final concentration of 1% and the cells were incubated at room temperature for an additional 15 min. To quench the formaldehyde-induced cross-linking, glycine was then added to a final concentration of 125 mM. The cells were incubated for an additional 5 min at room temperature, the medium was decanted, and cells were rinsed with cold 1ϫ phosphate-buffered saline. The rest of the procedure was conducted on ice. The cells were centrifuged (5 min at 3,000 rpm) and then washed with 10 ml of 1ϫ TBS (150 mM NaCl, 20 mM Tris-HCl, pH 7.6), centrifuged, and the supernatant discarded. At this point, the cells can be kept on ice for several hours; alternatively they may be frozen in liquid nitrogen and stored at Ϫ80°C for subsequent analysis. The cell pellet was resuspended in 1 ml of RIPA buffer (20 mM Tris-HCl, pH 8.0, 140 mM NaCl, 1% Triton X-100, 0.1% SDS, 1% deoxycholic acid, 2 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, and protease inhibitor mixture) and the cells were lysed by 5-8 passages through a 21-gauge needle to shear the DNA; they were then incubated on ice for 30 min. The lysate was transferred to a 15-ml Falcon conical tube and chromatin was further sheared by sonication (power settings of 7 and a 60% duty cycle, three 10-s pulses, while cooling on ice between rounds of sonication). The sonicated extracts were first centrifuged at 3,000 rpm for 5 min at 4°C, then the supernatant was centrifuged again at 14,000 rpm for 15 min at 4°C. Fifty l of the supernatant fraction was saved for the measurement of total input of chromatin (chromatin input fraction). Thirty l of Protein G PLUS-agarose (Santa Cruz, Sc-2002) slurry, 2 g of sheared salmon sperm DNA, and 2.5 g of preimmune serum was added to the supernatant and incubated for 50 min at 4°C while mixing by rotation. The samples were then centrifuged at 7,500 rpm for 2 min, primary antibodies against SREBP-1 or Sp1 were added to the supernatant, and the mixture was incubated overnight, with rotation, at a temperature of 4°C. The immune complexes were collected by centrifugation and Protein G PLUS-agarose beads (30-l bed volume) and 2 g of sheared salmon sperm DNA was added. The samples were then incubated while rotating for 60 min at 4°C and centrifuged at 7,500 rpm for 2 min at 4°C. A 50-l aliquot of the supernatant was kept for sizing the DNA and the rest was discarded. The immunoprecipitated complex (beads) was resuspended in 1 ml of RIPA buffer and incubated for 5 min at 4°C and then centrifuged for 2 min at 7,500 rpm at 4°C. This washing procedure was repeated twice. The supernatant from the last wash was discarded and the wash was repeated twice with 1 ml of RIPA-500 buffer (20 mM Tris-HCl, pH 8.0, 500 mM NaCl, 1% Triton X-100, 0.1% SDS, 1% deoxycholic acid, 1 mM phenylmethylsulfonyl fluoride, and protease inhibitor mixture). The supernatant was discarded and 1 ml of a LiCl detergent solution (0.5% deoxycholic acid, 1 mM EDTA, 250 mM LiCl, 0.5% Nonidet P-40, 20 mM Tris-HCl, pH 8.0, 1 mM phenylmethylsulfonyl fluoride, and protease inhibitor mixture) was added to the beads. The beads were washed twice using the same conditions as described for the RIPA buffer wash. The supernatant was discarded and 1 ml of 1ϫ TBS was added and the beads were washed as described previously for the RIPA buffer wash. After the washes were completed, the immunoprecipitated complexes were eluted from the beads by the addition of 100 l of 1% SDS, 1ϫ TE solution (1 mM Tris-HCl, pH 8.0, and 1 mM EDTA) and incubated at 65°C for 10 min. The complexes were then centrifuged for 3 min and the eluate was transferred to a fresh tube. The beads were washed with 150 l of 0.67% SDS, 1ϫ TE solution, centrifuged for 3 min, and the supernatant was added to the previous eluate. The protein-DNA crosslinking was reversed by incubating the eluate fraction overnight at 65°C. The eluate was centrifuged briefly and treated with Proteinase K solution (0.5 M EDTA, 20 mg/ml glycogen, 10 mg/ml Proteinase K, 1 M Tris-HCl, pH 6.5) and incubated for 1 h at 45°C. The DNA from the sample was purified using a QIAquick PCR purification kit (Qiagen, catalog number 28104), and stored in 35 l of water at Ϫ20°C. The DNA isolated by immunoprecipitation was analyzed by PCR using primers to the two SREs of the PEPCK-C gene promoter at Ϫ322 and Ϫ590. These were Ϫ322 forward, 5Ј-GACAACCAGCAGCCACTGGCA-CAC-3Ј; Ϫ322 reverse, 5Ј-GACCGTGACTGTTGCTGATGC-3Ј; Ϫ590 forward, 5Ј-GAGCAGGGGTCAGTATGT-3Ј; and Ϫ590 reverse, 5Ј-GG-GAATTCAGCTCTGAGACGTG-3Ј. These primers were designed to amplify regions of the PEPCK-C gene promoter known to bind SREBP-1 and Sp1. In addition, primers were constructed for a region of the ␤-actin structural gene (forward, 5Ј-GAGTGAGCGAGCCGGAGC-CAAT-3Ј; and reverse, 5Ј-GACGTCCCTGCTTACCTGGT-3Ј) and to a region of the PEPCK-C gene promoter that does not bind SREBP-1 or Sp1 (forward, 5Ј-GATCATGGACCTCCAGGTCATTTCG-3Ј; and reverse, 5Ј-GACGTCCCTGCTTACCTGGT-3Ј). These were used as negative controls. The starting chromatin input fraction was diluted 1/1000 and analyzed by PCR simultaneously with the samples. The amplified DNA fragments were separated by electrophoresis using 2% agarose gels and stained with ethidium bromide.

SREBP-1 and SREBP-2 Are Equally Effective in Inhibiting Transcription from the PEPCK-C Gene Promoter in Hepatoma
Cells-SREBP-1c and SREBP-2 are the predominant isoforms of SREBP in the liver, whereas SREBP-1a is present at low levels in this tissue (21). To demonstrate that SREBP-1a and SREBP-2 influence transcription from the PEPCK-C gene promoter, expression vectors encoding SREBP-1a and SREBP-2 were transfected individually into HepG2 cells, together with the PEPCK-C gene promoter (Ϫ2000 to ϩ73) linked to the luciferase reporter gene; their effect on gene transcription was compared with the effect of SREBP-1c. SREBP-1a reduced PKA-stimulated induction of transcription from the PEPCK-C gene promoter by 95%, whereas SREBP-2 reduced transcription by about 90% from the level of transcription noted in the absence of transfected SREBP ( Fig. 1, top panel). In the presence of PKA, SREBP-1c repressed transcription from the PEPCK-C gene promoter by 70% from the level of transcription noted in the absence of transfected SREBP-1c ( Fig. 1, top panel). As noted previously with SREBP-1c (1), SREBP-1a and SREBP-2 did not inhibit basal transcription from the PEPCK-C gene promoter (Fig. 1, bottom panel). It is important to note that basal transcription from the PEPCK-C gene promoter is ϳ1% of the PKA stimulated value and is thus already very low. Increasing the concentration of SREBP-1c, SREBP-1a, or SREBP-2 inhibited PKA-stimulated PEPCK-C gene transcription in a dose-dependant manner, without altering basal transcription (data not shown). It is thus likely that all three isoforms of SREBP can effectively inhibit PKA-stimulated transcription from the PEPCK-C gene in the liver.
Identifying Regions of the PEPCK-C Gene Promoter Involved in the Transcriptional Effect of SREBP-1c-An analysis of the PEPCK-C gene promoter using MatInspector Professional® (Genomatix) indicated the presence of several putative sites that resemble the consensus SRE present in the promoter of the LDL receptor gene (5Ј-ATCACCCCAC-3Ј). These sequences are present in the PEPCK-C gene promoter at Ϫ167 to Ϫ158, Ϫ322 to Ϫ313, Ϫ345 to Ϫ336, Ϫ535 to Ϫ526, and Ϫ590 to Ϫ581 on the template strand. Of these five sites, only two (Ϫ322 to Ϫ313 and Ϫ590 to Ϫ581) were identified as strong SREBP-1 binding sites in subsequent experiments (see below); these sites are shown in Fig. 2 and their position in the PEPCK-C gene promoter is indicated.

SREBP-1c and PEPCK-C Gene Transcription
A detailed analysis of the binding affinity of SREBP-1c and SREBP-1a to the putative SREs in the PEPCK-C gene promoter was performed using fluorescent EMSA, with recombinant proteins. The purity of the recombinant proteins used in these experiments was determined by electrophoresis using a 12% SDS-PAGE gel (Fig. 3A). The SREBP-1a generated by our plasmid encodes the first 490 amino acids of the protein; it migrates at about 80 kDa on an SDS gel. Recombinant SREBP-1c generated by our plasmid contains 446 amino acids and migrates at about 64 kDa in the SDS gel; it has an actual molecular weight is 56 kDa. SREBPs are known to run aberrantly on SDS gels relative to their molecular weights. This phenomenon is well described (16) and is a property of the amino acid composition of the extreme amino terminus of the protein.
There was weak binding of SREBP-1c to two of the SREs at Ϫ322 and Ϫ590 (Fig. 3B, lanes 5 and 11) and no detectable binding to the SREs at Ϫ167, Ϫ345, and Ϫ535 (data not shown); a similar pattern was noted with SREBP-1a (Fig. 3B,  lanes 3 and 9). Our fluorescent EMSA analysis thus focused on the putative SREs at Ϫ322 and Ϫ590. When USA, a general transcription cofactor fraction purified from of HeLa nuclei (22), was added to the binding reaction, there was a marked increase in affinity of both SREBP-1a and SREBP-1c for Ϫ322 and Ϫ590 (Fig. 3B, compare lanes 3 versus 4, 5 versus 6, 9 versus 10, and 11 versus 12). There was no binding of either SREBP-1a or SREBP-1c to a 25-bp, doubly labeled fragment of 18 S ribosomal RNA that served as a negative control (Fig. 3B,  lanes 13 through 18). In addition, an increasing concentration of the unlabeled oligonucleotides for the SREs at Ϫ322 and Ϫ590 totally competed out the binding of SREBP-1c, in the presence of USA, to the labeled SREs (data not presented). The recombinant form of SREBP-1a is about 20 kDa larger than SREBP-1c (see Fig. 3A), however, the complex formed between SREBP-1a and both putative SREs of the PEPCK-C gene promoter migrates faster than the complex between the SRE and SREBP-1c after electrophoresis in the non-denaturing polyacrylamide gel (Fig. 3B, compare lanes 3 versus 5 and 9 versus 11). This is presumably because of a higher number of negatively charged amino acids that are present in SREBP-1a. The binding affinity of SREBP-1c in the presence of USA was considerably greater than SREBP-1a for both of the SREs in the PEPCK-C gene promoter. Interestingly, the presence of USA did not alter the migration of the band formed with either SREBP-1a or -1c, suggesting that a protein(s) present in USA is facilitating binding of SREBP-1 to the SRE, without itself binding to the DNA or that the ternary complex is unstable during migration through the electric field. Because SREBP-1c is the major isoform of SREBP-1 in the liver (23), further work focused on its effect on PEPCK-C gene transcription.
The functional role of the 5 putative SREs (Ϫ167, Ϫ322, Ϫ345, Ϫ535, and Ϫ590) was tested by site-directed mutagenesis of these regions, either alone or in combination, within the intact PEPCK-C gene promoter. The modified promoter, linked to the luciferase reporter gene, was co-transfected, together with an SREBP-1c expression vector, into HepG2 cells in the presence or absence of an expression vector for PKA. Only the results of experiments demonstrating the effect of mutations in the SREs at Ϫ322 and Ϫ590 are shown in Fig. 4, because mutating the putative SREs at Ϫ167, Ϫ345, and Ϫ535 had no effect on the inhibition of transcription from the PEPCK-C gene promoter by SREBP-1 (data not shown). Mutation of the SRE at Ϫ590 resulted in a 5-fold increase in basal, unstimulated transcription from the PEPCK-C gene promoter, whereas a mutation at Ϫ322 had no effect (Fig. 4, bottom panel, open  bars). In addition, when the SREs at both Ϫ590 and Ϫ322 were mutated, there was also a 4-fold increase in unstimulated transcription from the PEPCK-C gene promoter (Fig. 4, bottom  panel, open bars). Co-transfection with SREBP-1c resulted in a 6.5-fold stimulation of transcription when the SRE at Ϫ590 was mutated, but there was no change when a mutation was introduced into the SRE at Ϫ322 (Fig. 4, bottom panel, black bars). In the presence of an overexpressed catalytic subunit of PKA, a mutation at Ϫ322 caused a 4-fold increase in transcription from the PEPCK-C gene promoter (Fig. 4, top panel, open  bars), whereas there was an approximate 13-fold increase in transcription when a mutation was introduced into the SRE at Ϫ590. SREBP-1c inhibited PKA-induced transcription from the native PEPCK-C gene promoter by about 80% (Fig. 4, top  panel); it had the same negative effect when mutations were introduced at Ϫ322 and Ϫ590. However, a mutation at both Ϫ322 and Ϫ590 resulted in a total loss of inhibition of transcription from the PEPCK-C gene promoter by SREBP-1c (Fig.  4, top panel, black bars). These data strongly support the conclusion that SREBP-1c acts to inhibit PEPCK-C gene transcription by interacting with the SREs at both Ϫ322 and Ϫ590.
A Consensus SRE Introduced into the PEPCK-C Gene Promoter Causes an Induction of Transcription by SREBP-1c, Not an Inhibition-The SRE at Ϫ590 differs from the consensus SRE sequence in the LDL receptor gene promoter by a T to A difference at nucleotide Ϫ582. Interestingly, there is a putative Sp1 binding site on the opposite strand of the PEPCK-C gene promoter at Ϫ585 to Ϫ572 (Fig. 2). In the consensus SRE found in the LDL receptor gene promoter, there is no Sp1 binding site on the opposite strand of the DNA; the Sp1 site is adjacent to the SRE and on the same strand (3,16,24). Creating a consensus SRE in the PEPCK-C gene promoter also mutates the putative Sp1 site on the opposite strand of the DNA of the PEPCK-C gene promoter. Using a fluorescent EMSA, we tested the relative binding of SREBP-1c, in the presence or absence of USA, to an SRE in which the T at position Ϫ582 was mutated to an A (590sre); all other flanking nucleotides in the DNA probe were the same as in the PEPCK-C gene promoter (Fig. 5). Binding of SREBP-1c to the modified SRE (590sre) was greater than to the native SRE at Ϫ590 (Fig. 5, lane 3 versus 7) and the addition of USA did not increase binding to 590sre (Fig. 5, lane  7 versus 8). It is important to note that a protein(s) in USA binds to both the native and modified PEPCK-C SRE (Fig. 5, see arrows) in the absence but not the presence of SREBP-1c. This same binding was noted in the gel shown in Fig. 6B. The nature of the protein(s) present in USA that enhance specific binding to 590sre is not clear at the present time. The results in Fig. 5 demonstrate that the one base pair mutation, as occurs in 590sre, abolishes the requirement for USA, because its presence did not enhance SREBP-1c binding to the SRE found in the gene for the LDL receptor.
Relative Binding of Sp1 to the 590 and 590sre-We next explored the relative binding affinity of Sp1 to the 590 and 590sre sites (Figs. 6 and 7). Sp1 bound to the SRE at Ϫ590 with a slightly higher affinity than it did to 590sre (Fig. 6A, lane 2, and Fig. 7A, lane 2). When SREBP-1c and Sp1 were present together with USA, Sp1 competes with SREBP-1c for binding to the SRE of the PEPCK-C gene promoter (Fig. 6A, lanes  6 -10). Even the lowest concentration of Sp1 (15 ng) caused a significantly reduced binding of SREBP-1c (Fig. 6A, lane 6). Thus, in the presence of even low concentrations of Sp1, SREBP-1c does not bind well to the PEPCK-C gene promoter, perhaps because of the overlapping binding sequences these proteins share on the promoter (see Fig. 2). In Fig. 6B, the reverse experiment was performed. Increasing concentrations of SREBP-1c markedly inhibited the binding of Sp1 to the SRE at Ϫ590 (Fig. 6B, lanes 6 -8). Different results were obtained with the 590sre oligonucleotide. Sp1 failed to efficiently compete SREBP-1c binding to 590sre, even at the highest concentrations of Sp1 used (75 ng) (Fig. 7A, lane 8). Sp1 was competed from binding to 590sre, even in the presence of SREBP-1c, by increasing concentrations of a double-stranded oligonucleotide that corresponds to a consensus Sp1 binding site from the G␣ (i2) gene promoter (25) (Fig. 7B, lanes 5-9). This indicates that Sp1 binds to the 590sre independently of SREBP-1.
To determine whether SREBP-1 and Sp1 bind to the PEPCK-C gene promoter in hepatocytes in vivo, we carried out ChIP analysis (Fig. 8). Hepatocytes were isolated from rats fed a high carbohydrate diet for 1 week to maximize the inhibition of PEPCK-C gene transcription. The binding of both SREBP-1 and Sp1 to their sites on the PEPCK-C gene promoter was confirmed by ChIP analysis using specific antibodies to each of these transcription factors. As shown in Fig. 8, PCR amplification of the SREs at both Ϫ322 and Ϫ590 of the PEPCK-C gene promoter resulted in bands that bound SREBP-1 in vivo. Similarly, antibodies to Sp1 detected that protein bound to the putative site for Sp1 at Ϫ590 on the PEPCK-C gene promoter. No binding was detected to the negative control, a segment of DNA upstream of the putative binding sites of SREBP-1 in the PEPCK-C gene promoter (Ϫ1900 to Ϫ1700). In addition, primers to the ␤-actin structural gene did not detect any amplified bands, demonstrating the specificity of the antibodies. Although regions of the PEPCK-C gene promoter encompassing the putative SREs at both Ϫ322 and Ϫ590 were amplified in this experiment, it is not possible to entirely discriminate between these sites because the size of the sheared DNA is not less than 400 bp; on this basis, some of the sheared segments of DNA would be expected to contain both putative SREs. Despite this caveat, we conclude that in rat hepatocytes both SREBP-1 and Sp1 bind to the PEPCK-C gene promoter at the sites predicted by our gel shift analysis.
Quantitation of the binding data from the fluorescent EMSA is shown in Fig. 9. The pattern of binding of SREBP-1c to both 590 and 590sre was remarkably different. The affinity of SREBP-1c for 590sre was clearly higher than for 590 and an increasing concentration of SREBP-1c resulted in decreased binding of this transcription factor to 590 (Fig. 9A). The reason for this unusual pattern of binding of SREBP-1 to 590 is not immediately evident. The effect of Sp1 on the two SREs is also markedly different. Sp1 does not alter the binding of SREBP-1c to 590sre (Fig. 9B). In contrast, Sp1 at a concentration as low as 15 ng, maximally displaced SREBP-1c binding to 590, even at increasing concentrations of SREBP-1c (Fig. 9B). The effect of an increasing concentration of SREBP-1c on the binding of Sp1 to the SRE at 590 is shown in Fig. 9C. SREBP-1c competes with Sp1 for binding to 590. However, at higher concentrations of SREBP-1c the binding of this protein itself to the SRE is destabilized (Fig. 9C). The reason for this destabilization is also not clear.
Effect of the Consensus SRE on Transcription from the PEPCK-C Gene Promoter-A chimeric gene was constructed that included the PEPCK-C gene promoter (Ϫ2000 to ϩ73), modified to contain the consensus SRE (590sre), and linked to the structural gene for luciferase. The chimeric gene containing the consensus SRE (590sre) was co-transfected with SREBP-1c into HepG2 cells (Fig. 10) and incubated in the presence and absence of PKA. Transcription from this promoter was compared with the native PEPCK-C gene promoter linked to the luciferase gene and a promoter with the SRE consensus sequence placed at Ϫ322. Overexpression of Sp1 in the presence of overexpressed SREBP-1c stimulated basal transcription from the PEPCK-C gene promoter by about 4-fold (Fig. 10,  bottom panel). Neither transcription factor alone had an effect on unstimulated basal transcription. Surprisingly, the single nucleotide difference in the SRE at Ϫ590 (590sre) in the PEPCK-C gene promoter altered the transcriptional response of the promoter to SREBP-1c; unstimulated transcription was increased more than 10-fold. Co-transfection with Sp1 or the combination of Sp1 and SREBP-1c further increased transcription (Fig. 10, bottom panel). PKA is known to induce transcription from the PEPCK-C gene promoter and all of the constructs tested in Fig. 10 respond to overexpression of PKA. In the   FIG. 3. Binding of SREBP-1a and SREBP-1c to putative SREs in the PEPCK-C gene promoter in the presence or absence of USA. Panel A, Histagged recombinant SREBP-1a and SREBP-1c were purified as outlined under "Experimental Procedures" and 0.25 g of SREBP-1a and 0.8 g of SREBP-1c were loaded on a 12% SDS-polyacrylamide gel and the proteins were separated by electrophoresis to determine their purity. SREBP-1a migrated at its predicted size of 80 kDa and SREBP-1c migrated at 64 kDa. Panel B, labeled fragments representing the SREs at Ϫ322 and Ϫ590 of the PEPCK-C gene promoter and a negative control (a labeled scrambled sequence) were incubated with purified SREBP-1a (0.42 g) and SREBP-1c (0.45 g), in the presence or absence of 0.5 g of USA, to determine the relative binding of each protein. The DNA-protein complex was isolated by 4% non-denaturing PAGE. Bands were visualized using a PhosphorImager®, as described in detail under "Experimental Procedures." Lanes 1, 7, and 13, labeled DNA fragment alone; presence of PKA, overexpression of SREBP-1c inhibited transcription from the native PEPCK-C gene promoter by about 50 -60%. Transcription from 590sre was elevated over the native PEPCK-C gene promoter by 12-fold and co-transfection with Sp1 and/or SREBP-1c increased transcription by 15-20fold (Fig. 10, top panel). Co-transfection with the combination of Sp1 and SREBP-1c did not stimulate transcription from the PEPCK-C gene promoter in the presence of PKA above that observed with the 590sre vector alone. When the SRE from the LDL receptor gene promoter was introduced into the SRE at 322 in the PEPCK-C gene promoter, there was a low basal level of transcription and no effect of either Sp1 or SREBP-1c on transcription. From the experiments in Fig. 10, we conclude that the PEPCK-C gene promoter can be made to respond positively to SREBP-1c by a single nucleotide substitution in the SRE at Ϫ590.

DISCUSSION
The detailed regulation of transcription of the gene for PEPCK-C by insulin is not well understood, despite considerable research. An insulin regulatory sequence in the promoter has been identified and shown in cultured cells to be responsible for about half of the inhibition of transcription from the PEPCK-C gene promoter typically observed after insulin exposure (26,27). However, when the insulin regulatory sequence was deleted from the PEPCK-C gene promoter and the gene introduced into transgenic mice, the induction of the gene in the liver was greatly reduced during diabetes, but insulin was still able to repress its transcription (28,29). This suggested that there was an alternative site that is responsible for the negative effect of insulin on PEPCK-C gene transcription. SREBP-1c, acting via the SREs identified in this paper, is thus an attractive candidate to mediate the inhibitory effect of in-sulin. In this regard, the activation of SREBPs by insulin in a wide variety of mammalian tissues provides a unifying mechanism to explain the broad specificity of this hormone in the transcriptional regulation of genes that code for proteins of importance in metabolism. It is interesting to note that the level of nuclear SREBP-1c in the livers of mice declined markedly after fasting for 24 h and increased 4-fold when fasted animals were re-fed a high carbohydrate diet (30). The mRNA for SREBP-1c virtually disappeared after as little as 6 h of fasting. These dietary conditions also caused a dramatic decrease in the level of transcription of the gene for PEPCK-C (31,32). Whereas the nuclear form of SREBP-1c varied with carbohydrate refeeding, there was no change in the precursor protein in the liver membranes of the mice, suggesting that the diet also altered the rate of precursor cleavage or altered the rate of SREBP-1c degradation (30). All this is consistent with an insulin-mediated effect of SREBP-1 on PEPCK-C gene transcription.
Insulin increases the transcription of genes involved in anabolic processes such as fatty acid and cholesterol synthesis (e.g. fatty acid synthase, acetyl-CoA carboxylase, squalene synthase, and HMG-CoA reductase), while inhibiting the expression of genes that are involved in the synthesis of carbohydrate, such as PEPCK-C and glucose-6-phosphatase (Glu-6-Pase) (1,12,33,34). The basis of this specificity resides in the complex interaction of SREBPs with regulatory elements in the promoters of the various genes whose transcription is being regulated by insulin. This involves an interaction between the individual SREBPs and other transcription factors and with co-regulators such as CBP/p300 (13) and PGC-1 (35). Recently, Puigserver et al. (36) have reported that PGC-1␣ inhibits the forkhead transcription factor, FOXO1, stimulation of PEPCK-C and Glu-6-Pase gene transcription in a Akt-dependent manner, suggest-ing that this coactivator is an important component in the insulin regulation of transcription of genes involved in gluconeogenesis. It is also clear that the three isoforms of SREBP regulate the transcription of distinct sets of genes, indicating that there is specificity in the interaction between promoter elements in the affected genes and the various transcription factors involved (37). In this regard, the nucleotide sequences of the SREs in defined genes that are regulated by SREBPs are variable, suggesting the specific sequence variation of the SRE in a target gene contributes to the unique transcriptional response of the gene.
Mechanism of SREBP Inhibition of PEPCK-C Gene Transcription-Any discussion of the mechanism of SREBP-1c inhibition of PEPCK-C gene transcription starts with a consideration of the models developed to explain the action of this transcription factor on genes involved in cholesterol synthesis. A decade ago, Wang et al. (4) reported that the promoter of the gene for the LDL receptor, a gene that has several SREs, termed repeats 1-3 (see Fig. 3 for the consensus sequence of the SREs in the LDL receptor gene promoter), also contains an Sp1 binding site adjacent to SRE repeat 3. Based on these findings, they suggested that SREBP-2 interacted with Sp1 to control transcription of the gene for the LDL receptor. Later studies by Sanchez et al. (16) demonstrated that these two transcription factors act synergistically and communicate with each other when bound to the DNA. They proposed a model, based on their studies, and those of Wang et al. (4), for the coordinated control of transcription of the LDL receptor gene by sterols. When cellular levels of sterol fall, the mature form of SREBP is translocated to the nucleus where it binds to SRE repeat 2 and interacts with Sp1. As a result, it increases the binding of Sp1 to its adjacent binding site, stimulating gene transcription. The interaction between SREBP and Sp1 on the LDL receptor gene promoter is unstable but is favored by the high levels of SREBP in the nucleus that occur when there is a chronically low concentration of sterols in the cell; this sustains the increased rates of gene transcription. When the intracellular concentration of sterols increases, because of high LDL receptor activity, the flux of SREBP into the nucleus is markedly decreased, resulting in a dissociation of DNA-bound SREBP and subsequent SREBP degradation. This would result in the predicted decrease in transcription of the LDL receptor gene.
In the present study we show that the PEPCK-C gene promoter, like the promoter for the LDL receptor gene, contains two functional SREs, one centered at Ϫ322 and the other at Ϫ590. In addition, ChIP assay demonstrated that SREBP-1 bound to the PEPCK-C gene promoter in the nuclei of rat hepatocytes in vivo, in the vicinity of the SREs (Fig. 8). Interestingly, the SRE in the PEPCK-C gene promoter at Ϫ590 differs from the consensus SRE in the LDL receptor gene promoter by a single base at Ϫ582. This seemingly subtle difference, when introduced into the native PEPCK-C gene promoter, converts the SRE into a positive rather than a negative regulatory element. It is likely that this is because of a critical alteration caused by this mutation in the Sp1 binding site that is present on the opposite strand of DNA to the SRE. This alteration also increases the binding of SREBP-1c to the SRE. In a cellular transfection assay using the consensus SRE from the LDL gene promoter, SREBP-1c markedly stimulated transcription from the PEPCK-C gene promoter. Our data show that the SRE in the PEPCK-C gene promoter does not bind SREBP-1c as firmly in the presence of low concentrations of Sp1 and it is thus not as effective an inhibitor of PEPCK-C gene transcription as SREBP-1c. We propose that as the concentration of SREBP-1c increases in the nucleus, because of the presence of an increase in the levels of insulin in the liver, SREBP-1c competes with Sp1 for binding to the SRE at Ϫ590 resulting in an inhibition of PEPCK-C gene transcription. The data shown in Fig. 10 demonstrate that, in the presence of PKA, SREBP-1c co-transfected with Sp1 resulted in an inhibition of transcription from the PEPCK-C gene promoter.
It is of interest to note the change caused by a single base pair modification in the SRE of the PEPCK-C gene promoter (590sre). The level of unstimulated transcription was increased to equal that noted for the native PEPCK-C gene promoter after the addition of PKA; Sp1 further stimulated transcription even in the presence of SREBP-1. Based on the model of Bennett and Osborne (38), it is assumed that with the mutated SRE, Sp1 interacts with SREBP-1c to increase the level of transcription from the PEPCK-C gene promoter.
Both isoforms of SREBP-1 bind poorly to the SRE in the PEPCK-C gene promoter in the absence of USA. A protein(s) in USA also binds to the SRE but the resulting band (see the arrows in Figs. 5 and 6) has a mobility greater than that noted with added SREBP-1c. When SREBP-1c is present with USA, this band is no longer evident. It is unlikely that a stable complex forms between protein(s) in USA and SREBP-1c because the size of the band shifted with SREBP-1c in the presence of USA is the same as the band noted with SREBP-1c alone. From these results, we speculate that USA modifies SREBP-1 in a way that increases its binding affinity to the SRE of the PEPCK-C gene promoter or that the ternary complex does not survive gel electrophoresis. It is also possible that protein(s) in USA bind to the SRE and alters its configuration, thus allowing SREBP-1 to bind more efficiently. USA does not have a similar effect on 590sre; SREBP-1c binding to 590sre is not affected by protein(s) in USA. Because the two SREs vary by only a single base pair, it is likely that a modification of SREBP-1c is not required for the high affinity binding to 590sre. The question of SREBP modification has been addressed by Bennett and Osborne (38) who have demonstrated that SREBP-1 and Sp1 interact in the absence of DNA to form a complex through the "buttonhead" domain of Sp1 and the DNA binding domain region of SREBP-1 (amino acid 321-490 of SREBP-1a). 2 This type of interaction alters the binding of the SREBP-1⅐Sp1 complex to the LDL receptor gene promoter. In the case of the PEPCK-C gene promoter, the SRE and the Sp1 binding sites are not adjacent, but are on opposite strands of the DNA. The binding of one transcription factor to this common site in the DNA precludes the binding of the other.
A more definitive understanding of the interactions involved in the effect of USA on SREBP-1c binding to the SRE in the PEPCK-C gene promoter will require further fractionation of proteins in USA. This complex is purified from HeLa nuclear extract and contains a number of proteins with distinct positive and negative transcriptional effects (15). The positive cofactors include PC1/poly(ADP-ribose) polymerase, PC2/Mediator, PC3/ DNA topoisomerase I, PC4, and PC5 (15, 39 -41); inhibitory components of this complex, such as NC1, have also been isolated and characterized (15). We could find no effect of PC2/ Mediator, PC3, PC4, and NC1 on the binding of SREBP-1c to the SRE of the PEPCK-C gene promoter (data not shown). It is interesting to note that USA stimulates transcription by USF, NFB, and Sp1 (15,22). However, the addition of USA did not alter the binding of Sp1 to the SRE at Ϫ590 of the PEPCK-C gene promoter. It is clear, from this initial research that whatever component(s) USA is/are involved in the effect, there is a high degree of specificity involved. A single base pair change of the T in the SRE at Ϫ582 an A, which creates the SRE present in the LDL receptor gene promoter, changes the SRE so that USA is no longer required for high affinity binding of SREBP-1c.
The presence of two functional SREs in the PEPCK-C gene promoter leads to the question of the role of each in the control of PEPCK-C gene transcription. Only one of the sites binds Sp1 (Ϫ590), whereas the other (Ϫ322) corresponds with the thyroid hormone regulatory element (42) (it is also an Accessory Factor 3 binding site) and to the P4 site that binds Fos/Jun heterodimers (43). Whether these sites interact with SREBP-1 binding at the SRE at Ϫ322 is not presently known. However, the consensus SRE of the LDL receptor gene promoter does not function in the same way when inserted into the PEPCK-C gene promoter at Ϫ322 as it does when it is present at Ϫ590. There must, therefore, be a very specific set of protein interactions on the promoter that are site-specific.
Mutation of the SRE at Ϫ590 results in a marked increase (14-fold) in PKA-induced transcription from the modified PEPCK-C gene promoter, as compared with the native promoter (Fig. 4, top panel). This suggests that the ablation of this site relieves the inhibition of transcription caused by the binding of SREBP-1. Mutation of the SRE at Ϫ590 does not prevent the inhibition of transcription by overexpressed SREBP-1c, supporting the important role of the SRE at Ϫ322. Mutation of the two sites results in a complete loss of the inhibitory effect of overexpressed SREBP-1c on PKA-stimulated transcription from the PEPCK-C gene promoter. The SRE mutation at Ϫ590 also caused SREBP-1c to stimulate basal transcription from the PEPCK-C gene promoter rather than inhibiting it. This suggests an interaction of SREBP-1 with other nuclear factors, such as those present in USA, that have yet to be identified. A complete model of SREBP-1 regulation of PEPCK-C gene transcription must take into account the regulation of transcription from both SREs in the promoter.
Physiological Role of SREBPs in the Control of Carbohydrate Metabolism-A central feature of the biological role of SREBP-1 is its acute activation by insulin. This activation suggests that SREBP-1 could play a critical role in coordinating the control of carbohydrate, as well as lipid metabolism. Thus, the demonstration that SREBP-1c, when introduced into hepatoma cells (1), hepatocytes (1), or the livers of mice with an adenoviral vector (2), inhibits transcription of the gene for PEPCK-C, further supports this concept of reciprocal regulation of the two opposing metabolic pathways. However, deletion of the individual genes for SREBP in mice has not provided a clear picture of whether individual members of this family of transcription factors are involved in the control of carbohydrate metabolism. For example, the deletion of the gene for SREBP-lc resulted in a 2.9-fold increase in the concentration of PEPCK-C mRNA in the livers of these mice (44), although this increase may be because of a variation in food intake by these animals. However, overexpression of the gene for SREBP-1 in the livers of mice, using the PEPCK-C gene promoter, does not result in the predicted negative feedback on transcription by SREBP-1 (23). Yamamato et al. (45) have recently reported that overexpressing the gene for SREBP-1a and SREBP-1c in the livers of mice does cause a dramatic reduction of both PEPCK-C and Glu-6-Pase mRNA in the liver. Thus, there are currently two different observations in the literature regarding the direct effect of overexpressed SREBP-1 on the regulation of PEPCK-C gene transcription in the livers of transgenic or knock-out mice.
Deleting the genes for either SREBP-1a or SREBP-2 is embryonic lethal, negating any further analysis of the metabolic phenotype of these mice. In the current paper, we demonstrate that all three isoforms of SREBP can inhibit transcription from the PEPCK-C gene promoter. It is thus possible that in the livers of mice with a deletion in the gene for a specific isoform of SREBP, one of the other members of the family can function to ensure the continued regulation of PEPCK-C gene transcription. In this regard, Liang et al. (44) reported that SREBP-1a and/or SREBP-2 could partially substitute for SREBP-1c and permit the insulin-mediated induction of acetyl-CoA carboxylase and fatty acid synthase in the livers of mice lacking the gene for SREBP-1c. This effect was gene specific, because there was no increase in the level of hepatic mRNA caused by insulin for other lipogenic enzymes normally regulated by SREBP-1c, such as glucose-6-phosphate dehydrogenase and NADP malate dehydrogenase. Finally, preliminary data on the control of transcription from the glucose-6-phosphatase gene promoter demonstrate that it is inhibited by co-transfection with SREBP-1c into hepatoma cells in culture. 3 Thus, the transcription of two genes that are critical for gluconeogenesis are strongly affected by SREBP-1.