Role of CCAAT Enhancer-binding Protein β in the Thyroid Hormone and cAMP Induction of Phosphoenolpyruvate Carboxykinase Gene Transcription*

Transcription of the gene for phosphoenolpyruvate carboxykinase (PEPCK) is stimulated by thyroid hormone (T3) and cAMP. Two DNA elements in the PEPCK promoter are required for T3 responsiveness including a thyroid hormone response element and a binding site called P3(I) for the CCAAT enhancer-binding protein (C/EBP). Both the α and β isoforms of C/EBP are highly expressed in the liver. C/EBPα contributes to the liver-specific expression and cAMP responsiveness of the PEPCK gene. In this study, we examined the ability of C/EBPβ when bound to the P3(I) site to regulate PEPCK gene expression. We report that C/EBPβ can stimulate basal expression and participate in the induction of PEPCK gene transcription by T3 and cAMP. The cAMP-responsive element-binding protein and AP1 proteins that contribute to the induction by cAMP are not involved in the stimulation by T3. A small region of the transactivation domain of C/EBPβ is sufficient for the stimulation of basal expression and cAMP responsiveness. Our results suggest that C/EBPα and C/EBPβ are functionally interchangeable when bound to the P3(I) site of the PEPCK promoter.

Transcription of the gene for phosphoenolpyruvate carboxykinase (PEPCK) is stimulated by thyroid hormone (T 3 ) and cAMP. Two DNA elements in the PEPCK promoter are required for T 3 responsiveness including a thyroid hormone response element and a binding site called P3(I) for the CCAAT enhancer-binding protein (C/EBP). Both the ␣ and ␤ isoforms of C/EBP are highly expressed in the liver. C/EBP␣ contributes to the liverspecific expression and cAMP responsiveness of the PEPCK gene. In this study, we examined the ability of C/EBP␤ when bound to the P3(I) site to regulate PEPCK gene expression. We report that C/EBP␤ can stimulate basal expression and participate in the induction of PEPCK gene transcription by T 3

and cAMP. The cAMPresponsive element-binding protein and AP1 proteins that contribute to the induction by cAMP are not involved in the stimulation by T 3 . A small region of the transactivation domain of C/EBP␤ is sufficient for the stimulation of basal expression and cAMP responsiveness. Our results suggest that C/EBP␣ and C/EBP␤ are functionally interchangeable when bound to the P3(I) site of the PEPCK promoter.
Expression of the phosphoenolpyruvate carboxykinase (PEPCK) 1 gene is controlled by a variety of hormones including glucagon (via cAMP), thyroid hormone (T 3 ), glucocorticoids, retinoic acid (RA), and insulin (1). Our studies have been directed at elucidating the mechanisms by which T 3 and cAMP stimulate the expression of the gene for PEPCK in the liver. Regulation of gene transcription by T 3 is mediated through the binding of liganded T 3 receptors (TR) to T 3 response elements (TRE) in the promoters of genes (2,3). The TR binds to DNA primarily as a heterodimer with the retinoid X receptor (RXR) (3,4). The ligand binding, dimerization, and transactivation domains of TR are contained within a broad region between amino acids 160 and 456 (3,5). Binding of T 3 by the TR results in conformational changes in the receptor and alteration of its transactivation properties (5). There are two isoforms ␣ and ␤ of the TR, and the TR␤ is highly expressed in the liver (2,6). Two elements in the PEPCK promoter are required for the full induction of PEPCK transcription by T 3 (see Fig. 1A). The first is the PEPCK-T 3 -responsive element (PTRE) contained within nucleotides Ϫ330 to Ϫ319, and the second is a binding site for the CCAAT enhancer-binding protein (C/EBP) called P3(I) between Ϫ250 and Ϫ234 (7,8). The PTRE has an unusual architecture consisting of two direct repeats separated by zero nucleotides (7). Mutation of either the PTRE or the P3(I) site eliminates the T 3 response (8).
Several elements in the promoter are involved in the cAMP induction of PEPCK transcription including a cAMP-responsive element (CRE) at Ϫ90 to Ϫ82, the P3(I) site, and an AP-1 binding site at Ϫ260 to Ϫ250 (Fig. 1A) (9,10). Three families of transcription factors that contribute to the cAMP stimulation include the cAMP-responsive element-binding protein (CREB), AP-1 proteins, and C/EBP (11). The P3(I) site is central to both the T 3 and cAMP induction of PEPCK transcription. In vivo studies have confirmed the importance of the P3(I) site in regulation of PEPCK gene expression. Transgenic mice carrying a PEPCK-bovine growth hormone fusion gene with a mutation in the P3(I) site have reduced hepatic expression of the bovine growth hormone transgene and reduced cAMP responsiveness of the PEPCK-bovine growth hormone reporter vector (12). C/EBP proteins constitute a family of transcription factors of which the ␣ and ␤ isoforms are highly expressed in the liver (13)(14)(15). C/EBP␣ is expressed primarily in liver and adipose tissue and regulates the expression of hepatic genes involved in energy metabolism (16,17). In addition to participating in hormonal responsiveness, C/EBP proteins play an important role in PEPCK gene transcription by contributing to the induction of the PEPCK gene at birth and directing liver-specific expression (11,17,18). Our early studies showed that C/EBP␣ and ␤ could bind to the CRE, P3(I), and P4(I) elements in the PEPCK promoter (15,19). C/EBP␣ can participate in the T 3 and cAMP induction of PEPCK transcription, and recent studies have shown that C/EBP␣ can mediate a cAMP induction without CREB being present (8,11,20).
Since C/EBP␤ is highly expressed in the liver and binds to identical sites in the PEPCK promoter, we examined whether C/EBP␤ when bound to the P3(I) site could participate in the hormonal induction of PEPCK transcription. Our data demon-strate that C/EBP␤ can stimulate PEPCK gene expression through the P3(I) site and modulate hormone responsiveness. The results indicate that the ␣ or ␤ isoforms of C/EBP have similar functions when bound to this element.

MATERIALS AND METHODS
Gel Mobility Assays-Gel mobility assays were conducted on 5% non-denaturing polyacrylamide gels (80:1 acrylamide:bisacrylamide) in 22 mM Tris and 190 mM glycine at 4°C (8). Double-stranded oligomers were labeled with Klenow enzyme and [␣-32 P]dCTP (8). The binding reactions were performed at room temperature in 10% glycerol, 80 mM KCl, 25 mM Tris (pH 7.4), 1 mM EDTA, and 1 mM dithiothreitol. Each binding reaction contained 1.0 g of poly(dI-dC) as nonspecific competitor and proteins as indicated. Nuclear proteins were prepared from HepG2 cells by the method of Shapiro et al. (21). Antibodies to C/EBP␣ and RXR␣ were obtained from Santa Cruz Biochemicals. The antibody to C/EBP␤ was a generous gift from Anna May Diehl.
Construction of CAT and Luciferase Vectors-The ligation of the PEPCK promoter from Ϫ490 to ϩ73 to the CAT reporter gene (Ϫ490-PCAT) has been described (7). To introduce the Gal4 binding site into the P3(I) site of the PEPCK promoter (Ϫ490-P3G4), the Ϫ490-PCAT vector containing a BstIIE site in the P3(I) element (Ϫ490P3(I)-PCAT) (8) was digested with BstIIE and SauI and the double-stranded oligomer containing the sequence ggttacctcggagtactgtcctccgt was inserted into nucleotides Ϫ248 to Ϫ223. The Ϫ490 to ϩ73 region of the PEPCK promoter was ligated in front of the luciferase reporter gene by removing the PEPCK promoter fragment from Ϫ490-PCAT by digestion with KpnI and BglII and ligating into the polylinker of pGL3 basic (Promega).
The Gal4 site was introduced into the TRE region of Ϫ330-PTRE/G4 CAT by PCR amplification with the 5Ј primer, ccctctagatcggaggtactgtcctccgtctgac, containing the altered nucleotides and a 3Ј primer, ttagatctcagagcgtctcgcc (ϩ73 to ϩ52), which included the BglII site at ϩ73. The 5Ј primer introduced a XbaI site. The amplified promoter fragment was digested with XbaI and BglII and ligated in front of the CAT reporter gene. The sequence was confirmed by sequence analysis at the St. Jude Center for Biotechnology.
Construction of Gal4-C/EBP Expression Vectors-Gal4-C/EBP␤ 1-108 and 1-66 were constructed by PCR amplification. The forward primer, which contained an EcoRI site and the first 18 nucleotides of the rat C/EBP␤ cDNA, tccgaattcatgcaccgcctgctggcctgggac, was used in PCR amplification. Additionally, two reverse primers containing PstI sites and the nucleotides representing amino acids 108 -102 (gagctgcaggtaaccgtagtcggccggcttc) or 66 -59 (gagctgcagcagctaggggctgaagtcgatg) were used. Following PCR amplification with the forward and reverse primers and the rat C/EBP␤ cDNA as template (14), the appropriate DNA fragment was isolated from an agarose gel and digested with EcoRI and PstI. This DNA fragment was ligated into the mammalian Gal4 DNA expression vector called pM (CLONTECH). To create the Gal4-C/EBP␤ 1-25, the Gal4-C/EBP␤ 1-108 vector was digested with BalI and HindIII. The restriction site overhangs were filled in with Klenow enzyme and the four deoxynucloetide triphosphates, and then the plasmid was religated. The Gal4-C/EBP␤ 1-264 was created by digestion of 1-108 with BalI and PstI. The BalI to PstI fragment of C/EBP␤ cDNA containing amino acids 25-264 was ligated into this site. The Gal4-C/EBP␤ 1-138 were generated by digestion of 1-264 with HindIII and SfuI. These restriction site overhangs were filled in with Klenow enzyme and religated. Construction of the Gal4-C/EBP␣ vectors is given elsewhere (11,20). The sequence of all Gal4-C/EBP␤ vectors was confirmed by sequence analysis.
Cell Transfections, Luciferase Assays, and CAT Assays-HepG2 cells were transfected by calcium phosphate precipitation as described previously (8). Each transfection contained 5.0 g of PEPCK-CAT vector, 5.0 g of the expression vector RSV-TR␤, and 2.0 g of SV40-␤-gal as a transfection control. Total DNA concentration was kept constant by the addition of pBluescript (Stratagene). CAT assays were conducted with [ 3 H]chloramphenicol and n-butyryl-coenzyme A using the xylene phase extraction method (8). All transfections were performed in duplicate and repeated three to six times. For the transfections with the luciferase vectors, the calcium phosphate precipitate included 3.0 g of PEPCK-luciferase, 1.0 g of an expression vector for the catalytic subunit of protein kinase A, and 1.0 g of SV40-␤-gal. Luciferase assays were conducted with the luciferin reagent as outlined by the manufacturer (Promega).

RESULTS
The goal of these studies was to evaluate the role of C/EBP␤ proteins bound to the P3(I) site in modulating PEPCK gene expression. To demonstrate that C/EBP proteins are involved in the T 3 induction of PEPCK transcription, we tested a set of specific dominant negative vectors. A-C/EBP is a dominant negative C/EBP vector, which has the C/EBP leucine zipper attached to an acidic amphipathic helix containing 4 heptad repeats (22). The amphipathic helix interacts with the basic region of C/EBP proteins. A-C/EBP forms a strong non-DNA binding heterodimer and will inhibit all C/EBP proteins (22). A-CREB is a dominant negative CREB protein with a CREB leucine zipper and an amphipathic helix (23). A-Fos is a dominant negative Jun vector with the Fos leucine zipper and an amphipathic helix (24). To examine the effect of these dominant negative proteins on the T 3 responsiveness of the PEPCK gene, a PEPCK-CAT vector containing 490 base pairs of the PEPCK promoter driving the CAT reporter gene (Ϫ490-PCAT) was cotransfected with three dominant negative vectors (Table  I). The addition of T 3 stimulated the Ϫ490-PCAT vector 3.5fold. Cotransfection with A-C/EBP inhibited the 3-4-fold induction of Ϫ490-PCAT by T 3 . The dominant negative CREB did not affect the induction by T 3 , and cotransfection with the dominant negative Jun actually improved the T 3 response from 3.5-fold to 9-fold. Although the reason for the enhanced responsiveness is not clear, it has been shown that C/EBP and Fos/ Jun can bind to the P3(II) and P4(I) sites in the PEPCK promoter (15). Inhibition of the Fos/Jun binding may allow increased binding of C/EBP proteins. When the PTRE was converted to an optimal TRE containing a direct repeat of the AGGTCA motif separated by four nucleotides (DR4) in the context of the PEPCK promoter, the dominant negative C/EBP vector still inhibited the T 3 induction (data not shown). Therefore, the inhibition of the T 3 stimulation by A-C/EBP is dependent not on the specific sequence of the PTRE but on its context within the PEPCK promoter.
As a control, two copies of a consensus TRE containing a DR4 motif were ligated in front of the enhancerless SV40 promoter driving the luciferase reporter gene and transfected into HepG2 cells with the three dominant negative proteins. Transcription of the SV40-luciferase vector was stimulated 11-fold by T 3 ( Table I). None of the dominant negative vectors decreased the T 3 response. In fact, the T 3 induction of the SV40luciferase vector was increased in the presence of A-C/EBP. induction of PEPCK gene transcription HepG2 cells were transfected with 5 g of PEPCK-CAT vector (Ϫ490-PCAT), 5 g of RSV-TR␤, 5 g of dominant negative expression vector for either C/EBP (A-C/EBP), CREB (A-CREB), or Jun (A-Fos), and 2.0 g of SV40-␤-gal. For the transfections with luciferase vectors, HepG2 cells were transfected with 3 g of SV40-luciferase (DR4X2 SV40-Luc), 3 g of RSV-TR␤, 3 g of dominant negative vector, and 1 g of SV40-␤-gal. Following transfection, the cells were placed in DMEM containing 10% charcoal-stripped fetal calf serum. Thyroid hormone (T 3 ) was added at a concentration of 100 nM for 48 h prior to harvesting for CAT assays. All transfections were repeated three to five times in duplicate. The results are presented as the -fold induction by T 3 Ϯ S.E.

Reporter vector
Dominant negative vector Induction by T 3 These observations indicate that the role of C/EBP in T 3 responsiveness is specific to the PEPCK gene and that the inhibition of PEPCK-CAT by A-C/EBP is not an inhibition of the TR. We have used both the CAT and luciferase reporter genes in these studies. Our results are similar with either reporter gene. Generally, we have used the CAT gene for our studies with T 3 and PEPCK promoter so that we can easily compare our results with our previous work. Our recent experiments with cAMP and the PEPCK promoter have been conducted with the luciferase gene. Previous studies had shown that C/EBP, CREB, and AP-1 proteins were involved in the stimulation of PEPCK transcription by cAMP (9). To determine whether dominant negative vectors for C/EBP, CREB, or Jun would affect the cAMP response, HepG2 cells were cotransfected with a mammalian expression vector for the catalytic subunit of protein kinase A (PKA), Ϫ490 PEPCK-luciferase (Ϫ490-PLuc), and the dominant negative vectors. Overexpression of PKA increased the luciferase activity 6-fold, and this stimulation was reduced by each dominant negative protein (Table II). These results confirm that the cAMP induction requires all three proteins and support the mutational studies of the PEPCK gene in which disruption of CRE, P3(I), P3(II), or P4(I) decreased the induction by cAMP (9,10). The A-C/EBP and A-Fos vectors decreased the basal expression of the PEPCK gene, while A-CREB did not reduce basal expression.
To test whether the DNA binding domain of the TR␤ was required for the T 3 induction of PEPCK transcription, the PTRE was converted into a Gal4 binding site to create Ϫ330 PTRE/G4-PCAT (see Fig. 1A). This CAT vector was cotransfected with a Gal4-TR␤ mammalian expression vector, which has the ligand binding/transactivation domain of the TR␤ (amino acids 168 -456) fused to the Gal4 DNA binding domain (25). The Gal4-TR␤ will bind to the Gal4 site in the PEPCK-CAT vector. An 80-fold stimulation of PEPCK-CAT activity was obtained (Table III). Even with the robust T 3 induction obtained with the Gal4-TR␤ vector, mutation of the P3(I) site (Ϫ330 PTRE/G4-P3 M PCAT) eliminated the T 3 response. Cotransfection of Ϫ330 PTRE/G4-PCAT with A-C/EBP also severely decreased the T 3 induction (data not shown).
To determine if the T 3 effect could be mediated by TR-RXR heterodimers, we used a Gal4-RXR␣ mammalian expression vector (Gal4-RXR␣) and a mammalian expression vector for the ligand binding domain of TR␤ (CMV-⌬⌬TR␤) (26). ⌬⌬TR␤ cannot bind DNA. When cotransfected with Gal4-RXR␣, ⌬⌬TR␤ associates with the promoter through heterodimerization with Gal4-RXR␣. Co-transfection of Ϫ330 PTRE/G4 PCAT with the combination of Gal4-RXR␣ and ⌬⌬TR␤ stimulated transcription 56-fold in the presence of T 3 (Table III). ⌬⌬TR␤ alone did not mediate a T 3 response, and mutation of the P3(I) site eliminated the T 3 induction. These observations indicate that the TR␤ ligand binding domain is sufficient for the interaction with C/EBP proteins if it is tethered to the PEPCK promoter. In addition, these results suggest that the T 3 effect could be mediated by either TR homodimers or TR-RXR heterodimers.
The next experiments examined whether sequences adjacent to the C/EBP binding site were required for T 3 and cAMP responsiveness. A series of mutations was introduced through the P3(I) binding region in the context of the Ϫ490-PLuc and Ϫ490-PCAT vectors. The C/EBP binding site (TTGTGTAAG) is contained within nucleotides Ϫ243 to Ϫ235. The mutations that were introduced into the Ϫ490-PCAT vector of nucleotides Ϫ248 to Ϫ245, Ϫ243 to Ϫ238, and Ϫ237 to Ϫ235 are shown in Fig. 1B. Alteration of nucleotides Ϫ243/Ϫ238 and Ϫ237/Ϫ235 eliminated C/EBP binding, whereas mutation of nucleotides Ϫ248/Ϫ245 did not affect C/EBP binding (data not shown). Mutation of nucleotides Ϫ243/Ϫ238 and Ϫ237/Ϫ235 diminished PKA responsiveness and basal expression (Table IV). The mutation of nucleotides Ϫ248/45 caused a slight reduction in the PKA responsiveness and a small increase in basal transcription. Alteration of the base pairs Ϫ243/Ϫ238 and Ϫ237/ Ϫ235 eliminated the T 3 response. Mutation of nucleotides Ϫ248 to Ϫ245 reduced the T 3 stimulation modestly. In conjunction with the data from the dominant negative vectors, these results indicate that the protein bound to the P3(I) site that contributes to basal expression and hormone responsiveness is a member of the C/EBP family.
Both the ␣ and ␤ isoforms of C/EBP are present in the liver (13,15). To evaluate the relative abundance of C/EBP isoforms binding to P3(I) in rat liver or HepG2 cell nuclei, supershift gel mobility assays were conducted using nuclear proteins isolated from rat liver or HepG2 cells, a labeled P3(I) oligomer and antibodies to C/EBP␣ or C/EBP␤. Antibodies to C/EBP␣ and C/EBP␤ generated supershifted complexes, indicating that both C/EBP isoforms are present in rat liver nuclear extract and can bind P3(I) (data not shown). Since our transfection experiments were conducted in HepG2 cells, the binding of proteins from HepG2 cell nuclear extract to the P3(I) element was examined (Fig. 2). The pattern of binding of HepG2 cell nuclear extract to P3(I) was similar to that observed with rat liver nuclear extract, but one significant difference was observed. The antibody to C/EBP␣ caused only a slight disruption of binding, while the antibody to C/EBP␤ supershifted most of the binding activity. C/EBP␣ has been shown to be present in HepG2 cells (13), but C/EBP␤ is the predominant isoform. Adding competitor nucleotides with mutations in Ϫ248/Ϫ245 did not affect the gel shift pattern, but adding those with mutations in Ϫ243/Ϫ238 did disrupt most of the protein DNA interactions. These data suggest that C/EBP␤ is the major protein binding to this site in HepG2 cells.
Previous studies had established that C/EBP␣ could restore cAMP and T 3 responsiveness to the PEPCK promoter (8,11). Since C/EBP␤ is the major C/EBP isoform in HepG2 cells, our next experiments examined the ability of the C/EBP␤ to modulate the basal transcription and cAMP responsiveness of the PEPCK promoter. To conduct these experiments, a series of Gal4-C/EBP␤ mammalian expression vectors was created (shown in Fig. 1C). A Gal4 DNA binding site was introduced into the P3(I) site in the Ϫ490 PEPCK-CAT vector to generate Ϫ490-P3G4-PLuc (see Fig. 1A). The Ϫ490 P3G4-PLuc vectors were cotransfected with the expression vectors Gal4-C/EBP␤ into HepG2 cells. The results from these experiments are

TABLE III T 3 responsiveness of the PEPCK promoter can be mediated by TR
homodimers or heterodimers HepG2 cells were transfected with 5 g of the PEPCK-CAT vector and either 0.5 g of Gal4-TR␤, 0.5 g of ⌬⌬TR␤, 0.5 g of Gal-RXR␣, or 5 g of RSV-TR␤. All transfections contained 2 g of SV40-␤-gal. Following transfection, the cells were placed in medium containing 10% charcoal-stripped fetal bovine serum with 100 nM T 3 . Cells were harvested after 48 h and assayed for CAT activity. All transfections were done in duplicate and repeated at least three times. The data are presented as -fold induction Ϯ S.E.  Table II. To measure the effects of T 3 , HepG2 cells were transfected with 5 g of PCAT vector, 5 g of RSV-TR␤, and 2.5 g of SV40-␤-gal as described in the legend to Table I were able to stimulate the basal expression 8 -10-fold above the Ϫ490-P3G4-PLuc vector (Table V). Smaller regions of C/EBP␤ were unable to increase the basal expression of the PEPCK-Luc gene. These results indicate that peptides within the first 108 amino acids of the transactivation domain of C/EBP␤ will stimulate basal transcription.
To evaluate the ability of these constructs to mediate a cAMP induction, the Gal4-C/EBP␤ expression vectors were cotransfected with the catalytic subunit of PKA and Ϫ490-PLuc (Table  V). The Ϫ490-PLuc vector was stimulated 9.4-fold by the overexpression of PKA, while mutation of the P3(I) site decreased this induction to 2.3-fold. The Gal-C/EBP␤ 1-264 was not able to increase the cAMP responsiveness of the PEPCK promoter. Previously, it had been shown that a Gal4-C/EBP␤ vector with full-length C/EBP␤ was not able to restore cAMP responsiveness (11). However, cotransfection with Gal4-C/EBP␤ 1-138 and 1-108 completely restored cAMP responsiveness. The Gal4-C/EBP␤ 1-66 increased the cAMP stimulation from 2.2-7-fold, and the Gal4-C/EBP␤ 1-25 vector did not mediate a cAMP response. As a control, the Ϫ490-P3G4-PLuc was cotransfected with Gal4-C/EBP␣ 6 -217, which contains amino acids 6 -217 of C/EBP␣ ligated to the Gal4 DNA binding domain (11). The Gal4-C/EBP␣ 6 -217 vector stimulated both basal expression and cAMP responsiveness, as had been reported previously (20). These results demonstrate that either isoform of C/EBP can regulate PEPCK gene expression through the P3(I) site.
We tested whether the Gal4-C/EBP␤ vectors could restore T 3 responsiveness to the Ϫ490-P3G4 reporter gene. The Ϫ490 P3G4-PCAT vector was cotransfected with Gal4-C/EBP␤ expression vectors and the transfected cells exposed to T 3 . As is shown in Table VI, introduction of the Gal4 site into P3(I) eliminated T 3 responsiveness. Cotransfection with Gal4-C/ EBP␤ 1-138 and 1-108 increased the T 3 response 2.5-and 2.8-fold, respectively. The Gal4-C/EBP␤ 1-66 vector was less effective in restoring T 3 responsiveness, while the vector with amino acids 1-25 had no effect. These data suggest that additional regions of the protein such as the basic region or leucine zipper are needed for the T 3 induction. Gal4 expression vectors expressing full-length C/EBP␣ or ␤ were tested, but these vectors did not increase the T 3 response (data not shown). Since the Gal4-C/EBP vectors with full-length C/EBP did not restore T 3 responsiveness to the PEPCK gene, the basic region or the leucine zipper may need to be associated with the P3 site to mediate this response.
Previously, we reported that cotransfection of Gal4-C/EBP␣ 6 -217 with Ϫ490-P3G4-PCAT partially restored T 3 inducibility (8). Since the DNA binding and dimerization domains of C/EBP are on the carboxyl terminus, it is possible that having the Gal4 domain on the amino terminus of the fusion protein was preventing the full restoration of T 3 responsiveness. The Gal4-C/EBP␣ 6 -217 "flipped" vector, which has the Gal4 DNA binding domain on the carboxyl terminus of the fusion protein rather than the amino terminus, was tested (Table VII). Like Gal4-C/EBP␣ 6 -217, this vector increased the T 3 response to 2.4-fold. Serial deletions of Gal4-C/EBP␣ were tested, and they were less effective than Gal4-C/EBP␣ 6 -217 in providing a T 3 induction (Table VII). These observations suggest that either the C/EBP␣ or ␤ isoforms will be able to mediate a T 3 induction of PEPCK gene expression. Importantly, these results demonstrate that different regions of the C/EBP proteins participate in the T 3 and cAMP response. The transactivation domains of C/EBP␣ or ␤ are sufficient to restore cAMP responsiveness, but other domains of C/EBP␣ or ␤ appear to be required for the full T 3 response.  VI Gal4-C/EBP␤ can restore T 3 responsiveness to a PEPCK promoter HepG2 cells were transfected with 5 g of Ϫ490-PCAT, 5 g of RSV-TR␤, 1 g of Gal4-C/EBP␤, and 2 g of SV40-␤-gal as described in Table I. Cells were treated with 100 nM T 3 for 48 h. The transfections were done in duplicate and repeated four times. The results are presented as -fold induction by T 3 Ϯ S.E.

Reporter vector
Gal4 expression vector Induction by T 3

DISCUSSION
Expression of the PEPCK gene is controlled by hormonal signals as well as by developmental and tissue-specific factors. Each hormonal response of the PEPCK gene is mediated through a combination of hormone response elements and accessory factor sites. Our studies have defined various transcription factors involved in regulation by T 3 and cAMP. Previous work has focused on the role of C/EBP␣ in the T 3 and cAMP effects because C/EBP␣ is highly expressed in the liver and clearly contributes to the induction of PEPCK expression at birth (8,18,20). In this study, we have examined the contributions of C/EBP␤ to the basal expression and hormonal responsiveness of the PEPCK gene. Our results indicate that C/EBP␤, along with the TR, can mediate the T 3 induction of PEPCK transcription. C/EBP␤ can participate in cAMP responsiveness, and 108 amino acids of the transactivation domain will impart full cAMP responsiveness to the PEPCK promoter.
In several genes, the transcriptional stimulation by T 3 is dependent on both the TR and other transcription factors, often called accessory factors. Accessory factors can be positioned in close proximity or widely separated from the TR. In the liver, NF-Y is required for the T 3 stimulation of S14 gene transcription (28). The TREs of the S14 gene are located between Ϫ2,700 and Ϫ2,500, while the NF-Y binding site is near the start site of transcription (28,29). In the heart, myocyte-specific enhancer factor 2 (MEF2) is involved in the T 3 induction of the ␣-cardiac myosin heavy chain gene (30). MEF2 and the TR bind to adjacent sites in the ␣-myosin heavy chain promoter and can physically interact in solution (30). The stimulation of the human placental lactogen B gene by T 3 is dependent on the binding of the pituitary factor GHF1 (also called Pit-1) (31). Our studies have identified C/EBP proteins as an additional family of accessory factors involved in T 3 action.
C/EBP proteins possess three functional regions including a transactivation domain on the amino terminus, a basic region that binds DNA and leucine zipper that dimerizes with other C/EBP proteins. Within the first 100 amino acids of C/EBP␣ and ␤ are three conserved peptide sequences, which contribute to the transactivation capabilities of these proteins (27,32,33). In C/EBP␤, these peptides regions were called activation domain modules 1, 2, and 3 (ADM1, 2, and 3) and removal of any of these peptides reduced the ability of C/EBP␤ to stimulate basal transcription (27). The ADM3 (amino acids 83-92) is homologous to the homology box 2 (HOB2) domain found in Fos and Jun (34). The Gal4-C/EBP␤ vector expressing the first 108 amino acids of C/EBP␤ stimulated PEPCK basal transcription and restored cAMP responsiveness. In the Gal4-C/EBP␤ vector expressing amino acids 1-66, the ADM3 is deleted and a portion of ADM2 (amino acids 56 -72) is removed. Since the Gal4-C/EBP␤ 1-66 vector did not stimulate basal expression of Ϫ490-P3G4-PLuc, it suggests that ADM3 is crucial for the stimulation of basal transcription. However, amino acids 1-66 did partially restore the cAMP response, indicating that additional regions of C/EBP␤ are involved in mediating the cAMP induction.
The regions of C/EBP␣ that are involved in cAMP responsiveness have been narrowly defined (20). Amino acids 1-124 were sufficient for a cAMP effect (20). In these studies, three Gal4 sites were ligated in front of a CAT reporter gene driven by a neutral promoter and this reporter vector was cotransfected with Gal4-C/EBP␣ vectors. Further deletion beyond amino acid 124 from the carboxyl terminus of C/EBP␣ resulted in a loss of cAMP responsiveness. In addition, the first 50 amino acids were essential for the cAMP induction. Most interestingly, mutation of amino acids 67, 77, and 78 in C/EBP␣ that interact with TBP did not affect the ability of this Gal4-C/EBP␣ protein to mediate a cAMP response (20,32). However, the Gal4-C/EBP␣ vector with these mutations was unable to stimulate basal transcription. Therefore, different amino acids within C/EBP␣ are involved in the regulation of basal transcription and cAMP responsiveness. Our results suggest that different domains of C/EBP␣ and ␤ mediate cAMP responsiveness. Amino acids 96 -124 are essential for the cAMP induction by C/EBP␣. There are no homologous domains within C/EBP␤. These experiments were conducted in a slightly different manner since the transfections in this study were all conducted with a Gal4 binding site in the context of the PEPCK gene, while those of the previous study were conducted using multimerized Gal4 sites. Nonetheless, there does not appear to be a single homologous peptide region of C/EBP␣ and ␤ that is responsible for the cAMP induction.
One important issue is why the Gal4-C/EBP␣ and ␤ vectors do not restore full T 3 responsiveness. One possibility is that either the leucine zipper or the basic region of C/EBP is required for the T 3 induction, as has been shown in the regulation of other genes. For example, the leucine zipper of C/EBP␤ synergizes with Sp1 in the regulation of the rat CYP2D5 P450 gene (35). We cannot address this question with the Gal4-C/ EBP expression vectors because, when the full-length C/EBP␣ or ␤ protein is attached to the Gal4 DNA binding domain, the resulting protein inhibits basal and hormone-regulated transcription (20,27). This observation suggests that the DNA binding domain must be in contact with the C/EBP binding site. Another possibility is that the Gal4 DNA binding domain altered the configuration of the C/EBP transactivation domain, making the necessary protein interactions for T 3 responsiveness impossible. While both the ␣ and ␤ isoforms of C/EBP can have a role in the regulation of PEPCK expression, in vivo they are not functionally redundant. In C/EBP␣ knock-out mice, the PEPCK gene was not induced at birth although PEPCK expression did rise after several hours (17,18). In addition, the PEPCK gene was poorly responsive to cAMP (18). C/EBP␣ knock-out mice die at birth of hypoglycemia and other complications (17,18). In the liver of adult mice in which the C/EBP␣ was conditionally knocked out, PEPCK expression was decreased (36). C/EBP␤ knock-out mice displayed a mixed phenotype with 50% of the mice dying at birth and showing reduced PEPCK gene expression (18). The remaining mice were viable and normal with respect to PEPCK expression. Other data indicate that the binding of C/EBP␤ to the PEPCK promoter may be increased by cAMP. When animals are injected with cAMP, the abundance of C/EBP␤ rises rapidly in the liver and the binding of C/EBP␤ to the P3(I) site increases (37). In physiologic situations such as with increased exercise, the abundance of C/EBP␤ increases in the liver (37). Our data TABLE VII Gal4-C/EBP␣ can restore T 3 responsiveness to a PEPCK promoter HepG2 cells were transfected with 5 g of Ϫ490-PCAT, 5 g of RSV-TR␤, 5 g of Gal4-C/EBP␣ fusion plasmid, and 5 g of SV40-␤-gal as described in Table I. Cells were treated with 100 NM T 3 for 48 h. The transfections were done in duplicate and were repeated four times. The results are presented as -fold induction by T 3 Ϯ S.E. demonstrate that, when C/EBP␤ is bound to the P3(I) site, it can stimulate basal expression and contribute to cAMP responsiveness. While our studies have focused on the PTRE as T 3 response element, this site is involved in several other hormone responses. Accessory factors are essential for the glucocorticoid induction of PEPCK transcription. Two weak glucocorticoidresponsive elements (GRE) have been identified (38). Like the TRE in the PEPCK gene, these GREs function well only in the context of the PEPCK promoter and confer very modest glucocorticoid responsiveness to a neutral promoter (38). Three accessory factor sites (AF1, AF2, and AF3) that are needed for glucocorticoid responsiveness have been defined (40,41,42). Mutation of any of these sites reduces the stimulation by glucocorticoids. AF3 overlaps the PTRE, and both chicken ovalbumin upstream promoter-transcription factor and the retinoic acid receptor (RAR) can bind the AF3/PTRE site (39,42). The stimulation of transcription by retinoic acid (RA) is mediated in part through the AF3/PTRE site in the promoter (8,42,43). Interestingly, even though the RAR binds the PTRE, the RA induction is independent of C/EBP proteins further demonstrating the specificity of the involvement of C/EBP proteins in the T 3 response (8).
Coactivator proteins provide a potential mechanism by which TR and C/EBP proteins could interact in the regulation of the PEPCK gene (44). Many of the transcription factors involved in the hormonal control of PEPCK expression are capable of interacting with various coactivators. The steroid receptor coactivator (SRC-1) can enhance the transactivation capability of a variety of steroid hormone receptors including the TR, RAR, glucocorticoid receptor, progesterone receptor, and estrogen receptor (45). SRC-1 has widespread tissue-specific distribution and is part of a family of nuclear receptor coactivators including SRC-2 (also cloned as GRIP1, TIF2) and SRC-3 (called RAC3, pCIP, and AIB1) (44,46,47). The CREBbinding protein was initially described as a coactivator for CREB, which enhanced cAMP signaling (48). Subsequently, CREB-binding protein has been demonstrated to interact with a wide variety of proteins including steroid hormone receptors, SRC proteins, and C/EBP␤ (44,49,50). Coactivators may help to coordinate the interactions of these various transcription factors on the PEPCK promoter in the absence of direct physical contact between the factors themselves.