Insulin Inhibits Hepatocellular Glucose Production by Utilizing Liver-enriched Transcriptional Inhibitory Protein to Disrupt the Association of CREB-binding Protein and RNA Polymerase II with the Phosphoenolpyruvate Carboxykinase Gene Promoter*

Hormones regulate glucose homeostasis, in part, by controlling the expression of gluconeogenic enzymes, such as phosphoenolpyruvate carboxykinase (PEPCK). Insulin and glucocorticoids reciprocally regulate PEPCK expression primarily at the level of gene transcription. We demonstrate here that glucocorticoids promote, whereas insulin disrupts, the association of CREB-binding protein (CBP) and RNA polymerase II with the hepatic PEPCK gene promoter in vivo. We also show that accessory factors, such as CCAAT/enhancer-binding protein β (C/EBPβ), can recruit CBP to drive transcription. Insulin increases protein levels of liver-enriched transcriptional inhibitory protein (LIP), an inhibitory form of C/EBPβ, in a phosphatidylinositol 3-kinase-dependent manner. LIP concomitantly replaces liver-enriched transcriptional activator protein on the PEPCK gene promoter, which can abrogate the recruitment of CBP and polymerase II, culminating in the repression of PEPCK expression and the attenuation of hepatocellular glucose production.

Glucose homeostasis is maintained, in part, by the hormonal regulation of hepatic gluconeogenesis. In response to hypoglycemia, glucagon-induced elevations in cAMP, as well as glucocorticoids, induce glucose production in the liver by increasing the expression of genes encoding hepatic enzymes that drive gluconeogenesis, such as phosphoenolpyruvate carboxykinase (PEPCK). 1 Conversely, in response to rising blood glucose levels, insulin rapidly inhibits PEPCK gene expression in a dominant fashion. Thus, tight control over hepatic gluconeogenesis is maintained by the opposing actions of hormones in response to constantly fluctuating plasma glucose concentrations.
These hormones regulate PEPCK expression primarily at the level of gene transcription. Meticulous study of the PEPCK gene promoter has revealed an overlapping set of hormone response units, each one consisting of a specific array of hormone response elements. Glucocorticoids, for example, induce PEPCK gene transcription through the glucocorticoid response unit, which is comprised of two glucocorticoid receptor binding sites (GR1 and GR2), three glucocorticoid accessory factor binding sites (gAF1, gAF2, and gAF3), and a cAMP response element (CRE) (1)(2)(3). Hepatic nuclear factor 4␣ (HNF4␣) and chicken ovalbumin upstream promoter transcription factor bind to gAF1, hepatic nuclear factor 3␤ (HNF3␤) binds to gAF2, chicken ovalbumin upstream promoter transcription factor binds to gAF3, and CCAAT/enhancer-binding protein ␤ (C/ EBP␤) binds to the CRE to mediate the glucocorticoid response (2,4,5). Likewise, cAMP signaling increases PEPCK gene transcription through the cAMP response unit, which is comprised of a CRE, two C/EBP binding sites (P3I and P4), and an AP-1 binding site (P3II) (6).
In contrast, the mechanism by which insulin dominantly represses PEPCK gene transcription has not been fully elucidated. Insulin signals through PI3K to inhibit PEPCK expression (7). Also, an insulin response sequence on the PEPCK gene promoter, which coincides with the location of gAF2, has been identified (8). However, deletion of this site in stably transfected cell lines blocks only part of the ability of insulin to repress PEPCK expression, suggesting that there might be a more proximal insulin response sequence through which insulin could act (9). Extensive analysis of the promoter has yet to reveal such a site. Alternatively, insulin might negatively regulate PEPCK gene transcription by disrupting the interactions of proteins involved in the communication among transcription factors, coactivators, and the general transcription machinery. To explore this possibility, we tested the effect of insulin on the assembly of transcription factors, coactivators, and the general transcription machinery on the endogenous PEPCK gene promoter. Our results demonstrate that glucocorticoids and cAMP increase, whereas insulin markedly decreases, the occupancy of CREB-binding protein (CBP) and RNA polymerase II (pol II) on the PEPCK gene promoter in vivo. In addition, the accessory factor C/EBP␤ can help to recruit CBP to drive transcription. Our results also show that insulin increases the protein levels of LIP, a natural inhibitory form of C/EBP␤, in a PI3K-dependent manner. Remarkably, insulin causes the concomitant, specific dissociation of LAP, the activating form of C/EBP␤, from the endogenous PEPCK gene promoter, suggesting that LIP successfully competes away LAP in response to insulin. Finally, the expression of exogenous LIP and its subsequent binding to the promoter disrupt the recruitment of CBP and pol II in response to activating hormones, resulting in the inhibition of PEPCK expression and the reduction of hepatocellular glucose production.

MATERIALS AND METHODS
Adenovirus Construction-The Ad-LIP recombinant adenovirus was prepared by excising the LIP cDNA from the REC-LIP plasmid by digestion with HindIII and EcoRV. The excised LIP cDNA was then ligated into the HindIII/EcoRV sites in the AdTrack-CMV plasmid, which was used subsequently to construct the recombinant adenovirus as described (10). The recombinant Ad-LIP virus was purified by freon extraction and cesium banding as described previously (11).
Antibodies-To generate the LAP-specific 2952 antibody, a 19-amino acid peptide from the carboxyl terminus of rat C/EBP␤ encompassing amino acids 278 -296 was synthesized by Bio-Synthesis, Inc. as a keyhole limpet hemocyanin conjugate. This material was then sent to East Acres Biologicals where antibodies were raised in rabbits according to the company's recommended protocol. Antiserum from production bleeds of one rabbit (rabbit 2952) was incubated with purified, bacterially expressed rat p34-C/EBP␤ (LAP) that had been immobilized on Immobilon-P filter strips overnight at 4°C. The antigen-antibody-containing strips were washed at room temperature four times in Trisbuffered saline plus 0.05% Tween 20 (TBS-T), once with TBS-T containing 1 M NaCl, and then a final wash with TBS-T. Antibody was eluted from the antigen-containing strips by incubation in 0.2 M glycine, pH 2.2, containing 0.2% phenol red. Eluted antibody was neutralized immediately with the addition of sufficient 1.5 M Tris, pH 9, to bring the pH to 7.0, and bovine serum albumin was added to a final concentration of 1 mg/ml.
Cell Culture, Hormone Treatment, and Adenovirus Infection-H4IIE rat hepatoma cells were maintained in ␣-minimum essential medium supplemented with 10% (v/v) serum (2% newborn, 3% calf, 5% fetal bovine) at 37°C, 5% CO 2 . For hormone treatments, cells were incubated in serum-free Dulbecco's modified Eagle's medium containing the following: 500 nM dexamethasone (Sigma), 100 M 8-CPT-cAMP (Sigma) and 500 nM dexamethasone, or 500 nM dexamethasone and 10 nM insulin (Sigma). For adenovirus infections, cells were washed with phosphate-buffered saline and incubated for 36 -48 h in cell culture medium containing the appropriate amount of viral particles. Optical particle units were measured with a SmartSpec 3000 (Bio-Rad) spectrophotometer (extinction coefficient ϭ 1.1 ϫ 10 12 virus particles/A 260 unit). Infection efficiency was evaluated by ascertaining the proportion of cells coexpressing GFP, using a C5810 series CCD camera (Hamamatsu) attached to a DMIBR-E inverted microscope (Leica).
Glucose Production Assay-Glucose production from H4IIE cells treated with hormones for 8 h was measured essentially as described by Wang et al. (12) using a glucose assay kit (Sigma 510-A). The glucose production buffer was modified slightly to include glucose-free Dulbecco's modified Eagle's medium containing 20 mM sodium lactate, 1 mM sodium pyruvate, and 15 mM HEPES, pH 7.4, without phenol red.
Primer Extension Analysis-RNA was isolated using TRI Reagent (Molecular Research Center), as detailed in the manufacturer's protocol. The oligonucleotides PC28 and ACT25, designed to be complementary to positions 102-129 and 42-67 in the rat PEPCK and ␤-actin mRNAs, respectively (7), were radiolabeled with [␥-32 P]ATP (Amersham Biosciences) before use in primer extension assays. To measure the accumulation of PEPCK and ␤-actin mRNA, primer extension analysis was performed as described by Forest et al. (9).
Chromatin Immunoprecipitation Assay-The ChIP assay protocol was adapted from methods described previously (13)(14)(15)(16). H4IIE cells (1 ϫ 10 8 cells/condition) were washed with phosphate-buffered saline and preincubated in serum-free Dulbecco's modified Eagle's medium at 37°C for Ն48 h. After hormone treatments, cells were cross-linked with 1% formaldehyde (Fisher Scientific) in serum-free Dulbecco's modified Eagle's medium at 37°C for 5 min. To arrest cross-linking, glycine was added directly to the medium at a final concentration of 125 mM, and the cells were rinsed with ice-cold phosphate-buffered saline. Cells were harvested with cell scraping buffer (1 ml/plate of ice-cold phosphatebuffered saline with protease and phosphatase inhibitors: 1 mM phenylmethylsulfonyl fluoride, 1 g/ml aprotinin, 1 g/ml leupeptin, 1 g/ml pepstatin A, 20 mM NaF, 1 mM Na 3 VO 4 , 10 mM Na 4 P 2 O 7 , pH 8, 0.4 mM Na 2 MoO 4 , 125 nM okadaic acid). Cells were then pelleted by centrifugation at 700 ϫ g for 4 min at 4°C, resuspended in SDS lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris-HCl, pH 8, with protease and phosphatase inhibitors), and incubated on ice for 10 min. The lysates were transferred to prechilled Eppendorf tubes containing 25-m diameter glass beads (ratio of lysate to bead volume was 3:1), prewashed with cell scraping buffer. To shear chromatin, the lysate/bead mixture was sonicated (VirSonic, 2.5 mm tip) on ice for 12 10-s pulses at a setting of 3 (output of 4 -5 watts), yielding chromatin fragments of 100 -600 bp in size. Samples were centrifuged at 14,000 rpm for 10 min at 4°C to remove detritus, and the supernatant was divided into aliquots for subsequent 10-fold dilution in ChIP dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, pH 8, 16.7 mM Tris-HCl, pH 8 with protease and phosphatase inhibitors). To provide a positive control (input) for each condition, one undiluted aliquot was retained for further processing in parallel with all other samples at the reversal of cross-linking step. To reduce nonspecific background, each 1-ml chromatin sample was precleared with 10 l of protein A/G-agarose slurry (Santa Cruz), supplemented with 100 g/ml sonicated salmon sperm DNA (Stratagene), for 1 h at 4°C on a rotating wheel, after which the beads were pelleted, and the supernatant was transferred to a new tube. Chromatin complexes were immunoprecipitated for 12-18 h at 4°C while rotating with amounts (5-10 g) of primary antibody optimized for selective immunoprecipitation of signal, or without antibody (control) to provide negative controls. Immune complexes were collected with 40 l of protein A/G-agarose ϩ 100 g/ml salmon sperm DNA, while rotating for 3 h at 4°C, followed by centrifugation at 1,000 ϫ g for 1 min at 4°C. Secondary goat IgG anti-mouse IgM (Santa Cruz) antibody was added to samples immunoprecipitated with BAbCO mouse ascites fluid. The beads were washed for 5 min at 4°C with 1 ml of each of the following buffers in succession: low salt wash buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, pH 8, 20 mM Tris-HCl, pH 8, 150 mM NaCl), high salt wash buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, pH 8, 20 mM Tris-HCl, pH 8, 500 mM NaCl), LiCl wash buffer (0.25 M LiCl, 1% Nonidet P-40, 1% sodium deoxycholate, 1 mM EDTA pH 8, 10 mM Tris-HCl pH 8), and twice with TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8). Chromatin complexes were eluted from the beads in three consecutive 30-min rotating incubations with 200 l of elution buffer (1% SDS, 0.1 M NaHCO 3 ) at 22°C. To reverse cross-linking and digest RNA present in the samples, NaCl (200 mM final concentration) and RNase mixture (Ambion) were added, and the samples were incubated at 65°C for Ն 6 h. To digest proteins, samples were incubated at 45°C for 90 min after the addition of the following at their final concentrations: 10 mM EDTA, 40 mM Tris-HCl pH 6.5, and 50 g/ml proteinase K. Samples were extracted twice with phenol:chloroform:isoamyl alcohol (25:24:1) and once with chloroform:isoamyl alcohol (24:1). DNA was precipitated with 5 g of glycogen azure (Sigma) and 2ϫ volume of 95% ethanol, and pellets were collected by microcentrifugation for 30 min at 4°C. Samples were resuspended in 100 l of nuclease-free water (Promega) and stored at -80°C for subsequent PCR analysis.
Western Blot Analysis-H4IIE cell nuclear extracts were prepared as described by Waltner-Law et al. (21), subjected to electrophoresis through a 12% acrylamide gel, and transferred to a 0.45-m nitrocellulose membrane using a Mini-PROTEAN II apparatus (Bio-Rad). After thorough washing, the membrane was exposed first to primary antibody and then to alkaline phosphatase-conjugated secondary antibody. Protein bands were resolved by chemiluminescence using the Western Breeze immunodetection system (Novex). Individual bands were analyzed with a Fluor-S MultiImager using Quantity One software.

Hormonal Regulation of Glucose Production and PEPCK
Gene Expression-The production of glucose from gluconeogenic precursors in H4IIE hepatoma cells, subjected to the indicated hormone treatments (Fig. 1A), was assayed by incubating the cells in glucose-free medium and measuring the amount of glucose released into the medium. The synthetic glucocorticoid, dexamethasone, stimulates glucose production more than 5-fold over basal conditions. Additionally, the cAMP analog, 8-CPT-cAMP, along with dexamethasone, is able to induce glucose production ϳ8-fold over basal conditions. In contrast, when the cells are treated with dexamethasone and insulin together, glucose production drops precipitously to levels below that observed in basal conditions. Thus, hormones are able to regulate glucose production from non-carbohydrate precursors, lactate and pyruvate, in these hepatoma cells in a manner that reflects normal physiology.
To measure PEPCK gene expression under similar hormone conditions, primer extension assays were performed on RNA preparations from these cells (Fig. 1B). As shown previously (9,12), glucocorticoids stimulate the production of PEPCK mRNA, and cAMP is able to induce this stimulation further. In contrast, insulin inhibits PEPCK gene expression below even basal levels. Hence, the hormonal regulation of PEPCK gene expression parallels that of hepatocellular glucose production.
In Vivo Assembly and Disassembly of Factors of the PEPCK Gene Promoter in Response to Hormones-The ChIP assay was used to determine whether the endogenous PEPCK gene promoter is occupied by a variety of factors under the above hormone conditions. Antibodies specific for pol II, CBP, GR, HNF4␣, HNF3␤, C/EBP␤, C/EBP␣, acetylated histone H3, and acetylated histone H4 were used to immunoprecipitate chromatin fragments ( Fig. 2A). PCR amplification was conducted using primer pairs specific for the proximal, regulatory region (1) and a distal, nonregulatory region (2) of the PEPCK gene promoter. The latter primer set was included as a control to ensure that the signals observed were specific for the more proximal, regulatory region of the PEPCK gene promoter. Each input sample was examined to ensure efficient chromatin shearing (Fig. 2G), yielding fragment sizes in the range of 100 -600 bp. The sensitivity range of the assay was assessed in parallel PCRs using increasing amounts of a control template (Fig. 2F). For some factors (e.g. pol II), we tested the ability of different antibodies to immunoprecipitate chromatin fragments and found results similar to those presented, thereby diminishing the possibility that the disappearance of a particular band might be the result of epitope masking during hormone treatment and cross-linking (data not shown).
Upon stimulation with either glucocorticoids alone or glucocorticoids and cAMP together, there is an increased association of pol II, CBP, and GR with the PEPCK gene promoter, presumably to drive transcription. In contrast, insulin markedly diminishes the association of pol II and CBP with the promoter (Fig. 2A). This may be one of the mechanisms through which insulin exerts its inhibitory effect on PEPCK expression and hepatic glucose production. Interestingly, GR occupancy remains constant as long as receptor ligand is present. Thus, insulin does not seem to inhibit PEPCK gene transcription by disrupting the binding of GR to the promoter, although this does not rule out the possibility of insulin-induced post-translational modifications of GR. The association of accessory factors, such as HNF4␣, HNF3␤, C/EBP␤, and C/EBP␣, does not change across the hormone conditions tested, suggesting that these factors are bound constitutively to the PEPCK gene promoter. Likewise, the acetylation state of local histone proteins does not appear to be affected by different hormone treatments. This is consistent with previous work demonstrating the presence of hypersensitive sites and in vivo footprinting patterns in the proximal region of the PEPCK gene promoter which are not altered either by induction of transcription with cAMP or glucocorticoids or by inhibition with insulin (22,23).
A recent study using synchronized MCF-7 cells showed that activating factors cycle on and off estrogen-responsive promoters on a time scale of 30 -60 min to drive transcription (24). To test this model of hormone activation, the ChIP assay was used to examine the assembly of pol II, C/EBP␣, and C/EBP␤ on the PEPCK gene promoter during a glucocorticoid time course (Fig.   FIG. 1. Glucocorticoids and cAMP stimulate, whereas insulin represses, glucose production and PEPCK gene expression. A, glucose production was measured by incubating hepatoma cells in glucose-free medium supplemented with lactate and pyruvate, after 8-h hormone treatments: no hormone (N), 500 nM dexamethasone (D), 100 M 8-CPT-cAMP ϩ 500 nM dexamethasone (CD), and 500 nM dexamethasone ϩ 10 nM insulin (DI). The bars depict the mean Ϯ S.E. of at least three independent experiments. B, PEPCK mRNA levels were assayed by primer extension of RNA preparations from cells treated for 2 h under the above hormone conditions. ␤-Actin mRNA levels were assayed for internal controls. 2C). Polymerase II occupancy on the PEPCK gene promoter increases in the presence of dexamethasone but is not altered with prolonged hormone exposure. In addition, the binding of both C/EBP family members does not change over time. Thus, we do not observe a cyclic assembly of factors on the PEPCK gene promoter in response to glucocorticoids.
To contrast the pattern of occupancy found with the PEPCK gene promoter in response to glucocorticoids, cAMP, and insulin, we examined gene 33 and ␤-actin using PCR primers specific for those gene promoters. Gene 33 was chosen because its expression in H4IIE cells is stimulated by all of the aforementioned hormones, including insulin (25,26). Consistent with this pattern of regulation, and unlike that of PEPCK, the association of pol II and CBP with the gene 33 promoter is enhanced in the presence of glucocorticoids, cAMP, and also insulin (Fig. 2D). ␤-Actin was chosen as a control; accordingly, the occupancy of pol II on the ␤-actin promoter does not significantly change across the different hormone conditions (Fig.  2E). Therefore, insulin does not disrupt the association of pol II at either the gene 33 promoter or the ␤-actin promoter, as it does at the PEPCK gene promoter. A, association of RNA pol II, CBP, GR, HNF4␣, HNF3␤, C/EBP␤, C/EBP␣, acetylated histone H3 (AcH3), and acetylated histone H4 (AcH4) with the endogenous PEPCK gene promoter was measured by ChIP assay across the following hormone conditions: no hormone (N), 500 nM dexamethasone (D), 100 M 8-CPT-cAMP ϩ 500 nM dexamethasone (CD), and 500 nM dexamethasone ϩ 10 nM insulin (DI). PCR amplification of immunoprecipitated chromatin fragments was conducted using primer pairs specific for the proximal, regulatory region (1) and a distal, nonregulatory region (2) of the PEPCK gene promoter. Control lanes show the results of immunoprecipitations performed in parallel without the application of primary antibodies. Input lanes show the results from samples not subjected to immunoprecipitation. All results depicted are representative of at least three independent experiments. B, linear diagram depicting the array of hormone response elements identified on the PEPCK gene promoter, as well as transcription factors that bind to these elements in vitro. The factors listed above the promoter participate in the response to glucocorticoids; factors listed below the promoter participate in the response to cAMP. Promoter position relative to the transcription start site is indicated numerically. C, the presence of pol II, C/EBP␣, and C/EBP␤ on the PEPCK gene promoter was determined by ChIP assay under conditions in which the cells were incubated with dexamethasone (Dex) for increasing amounts of time, as indicated. D, occupancy of pol II and CBP on the gene 33 promoter was determined by ChIP assay after the hepatoma cells were treated with hormone conditions similar to those described in A. E, likewise, occupancy of pol II on the ␤-actin promoter was determined by ChIP assay. F, the sensitivity range of the ChIP assay was determined by subjecting increasing amounts of immunoprecipitated genomic DNA to identical PCR conditions. G, sonication efficiency was measured by subjecting sheared chromatin to agarose gel electrophoresis. On the right, the relative amount of chromatin is plotted against fragment size (bp). cessory elements abolishes the response altogether (2,3,27). But because their binding does not change in response to hormones, how might they help to mediate hormone-induced gene transcription? One possible role might be to assemble a platform upon which coactivators and the preinitiation complex may be recruited to drive transcription. Among the array of accessory elements on the PEPCK gene promoter, we chose to examine further the CRE because it is a pleiotropic element important for several aspects of PEPCK gene regulation, including basal, cAMP-, and glucocorticoid-induced gene transcription (28 -30). Furthermore, we know from previous work that C/EBP␤ functions well at this element to help mediate a glucocorticoid response (19).

Role of Accessory Factors and Coactivators in
To test whether C/EBP␤ can serve as an accessory factor in PEPCK gene regulation under different hormone conditions when tethered to the CRE, H4IIE hepatoma cells were transiently transfected with either a wild-type PEPCK gene promoter-luciferase reporter construct (PEPCK-LUC) or a mutant construct in which the CRE is replaced by an upstream activating sequence to which cotransfected, chimeric GAL4 proteins bind (C/U-PEPCK-LUC) (Fig. 3A). After transfection, the cells were treated with hormones and harvested for luciferase assays. The wild-type PEPCK gene promoter construct exhibits a strong 16-fold induction by glucocorticoids and a 28-fold induction by glucocorticoids and cAMP together. Insulin suppresses the glucocorticoid induction by 77%. These results reflect the hormonal regulation patterns observed with endogenous PEPCK mRNA levels and hepatocellular glucose production (compare with Fig. 1). When GAL4-C/EBP␤ is cotransfected with C/U-PEPCK-LUC, the magnitude of reporter gene expression is very similar that of the wild-type construct. There is a 17-fold induction by glucocorticoids and a 25-fold induction by glucocorticoids and cAMP together when C/EBP␤ is tethered to the PEPCK CRE. Likewise, insulin suppresses glucocorticoid induction by 80%. However, when only the GAL4 DBD is cotransfected with C/U-PEPCK-LUC, the magnitude of hormone-induced reporter expression drops by about 50%. Similar results are observed when the empty parent vector is used for cotransfection (data not shown). This is consistent with previous studies in which approximately half of the basal, glucocorticoid, and cAMP induction is lost when the CRE is deleted or mutated (28,30).
C/EBP␤ and p300 interact both in vitro and in vivo, a functional collaboration that results in enhanced gene transcription (31). The region of C/EBP␤ which interacts with p300 coincides with both its amino-terminal transactivation domain and an adjacent accessory factor domain, essential for glucocorticoid induction of PEPCK gene expression (19). To determine whether C/EBP␤ can help to recruit the coactivator CBP (closely related to p300) to the PEPCK gene promoter, transient transfection assays similar to those described above were performed with the addition of an expression plasmid for CBP (Fig. 3B). For direct comparison, the horizontal line inside each bar denotes the reporter strength exhibited in the absence of exogenous CBP (results from Fig. 3A). The presence of exogenous CBP generally induces reporter expression a further 2-fold. Once again, the magnitude of the hormone responses when C/EBP␤, but not the GAL4 DBD, is tethered to the CRE mimics that of the wild-type reporter construct. Thus, C/EBP␤ is able to function as an accessory factor not only to help bring about these various hormone responses from the PEPCK CRE, but also to help recruit CBP to the promoter to induce these responses further.
If indeed one of the roles of accessory factors is to recruit coactivators to the promoter, then it should be possible to circumvent this task by tethering coactivators directly to the promoter using the GAL4 fusion system. To test this hypothesis, GAL4-CBP was transfected into cells along with the C/U-PEPCK-LUC construct (Fig. 3B, far right). Remarkably, basal transcription increases 4-fold over that seen with the wild-type construct. Thus, when tethered directly to the CRE, CBP can induce hormone-driven gene transcription, thereby lending credence to the idea that accessory factors function, at least in part, to help recruit coactivators to the promoter. On top of that, glucocorticoid induction rises to 22-fold over basal, and glucocorticoid/cAMP induction is 25-fold over basal. This suggests that hormones affect not only the recruitment of coactivators to gene promoters but also their ability to interact with other proteins to affect transcription. In fact, when we tether the coactivator SRC-1 (via the GAL4 DBD) to other elements within the glucocorticoid response unit, we observe a similar induction with the addition of dexamethasone (32).
The ChIP assay results ( Fig. 2A) demonstrate that insulin blocks the association of CBP with the endogenous PEPCK gene promoter. The transfection assay results (Fig. 3B) show that accessory factors, such as C/EBP␤, can act to help recruit CBP to the PEPCK gene promoter and thereby drive transcription. Together, this would suggest a potential mechanism: perhaps insulin inhibits PEPCK gene transcription, at least in part, by somehow disrupting the ability of accessory factors to recruit coactivators to the promoter.
Effect of Insulin on LIP-In H4IIE cells, C/EBP␤ can be found in two predominant forms: p34-LAP, liver-enriched transcriptional activator protein, and p20-LIP, liver-enriched transcriptional inhibitory protein. Synthesized from an internal translation initiation site (33), LIP has the same amino acid sequence as the corresponding region of LAP. Hence, LIP contains the carboxyl-terminal dimerization and DNA binding domains but lacks the amino-terminal activation domain of LAP, and it is thought to function as a natural negative regulator of transcription (34). To examine the effect of insulin on endogenous C/EBP␤ proteins, the levels of p34-LAP and p20-LIP were measured by Western blot of nuclear extracts from cells treated with either activating hormones in the presence or absence of insulin (Fig. 4A) or insulin alone under conditions of increasing duration (Fig. 4B). Almost immediately, insulin increased protein levels of LIP with respect to those of LAP, reaching a new steady-state level after 90 min. Furthermore, when insulin and the PI3K inhibitor LY 294002 are added together, the observed increase in the LIP:LAP ratio is abrogated (Fig. 4C). This is particularly interesting because the inhibition of PEPCK gene transcription by insulin also occurs in a PI3K-dependent manner (7). Insulin does not affect the levels of two other CREbinding proteins, CREB and C/EBP␣, under similar conditions (data not shown).
Exchange of LAP by LIP in Response to Insulin-An increased LIP:LAP ratio could provide a mechanism for insulin to disrupt LAP function. Prior work has shown that LIP can compete successfully with LAP for binding to DNA recognition sequences at substoichiometric levels (34). Perhaps by increasing the amount of available LIP protein, insulin causes the displacement of LAP from the PEPCK gene promoter, thereby disabling its accessory factor function. To test this hypothesis, we examined the association of LAP with the PEPCK gene promoter in response to insulin. To distinguish between the actions of endogenous LIP and LAP, something that has heretofore not been possible, we developed a unique antibody (2952) that recognizes only LAP. The peptide used to generate the LAP-specific 2952 antibody contains a 19-amino acid sequence that is shared by both LIP and LAP. However, the antibody recognizes only a specific conformation of this epitope, one that is found in LAP, but not LIP, a fortuitous and very useful happenstance. The Western blot results show that in contrast to the sc-150 antibody, which recognizes both forms of C/EBP␤, the 2952 antibody recognizes LAP specifically and does not recognize LIP (Fig. 5A). As before (see Fig. 2A), the ChIP assay results show that CBP and pol II are recruited to the promoter upon glucocorticoid induction and are not associated with the promoter in the presence of insulin, this time at 30 min (Fig.  5B). The dissociation of pol II from the promoter with the insulin time course starkly contrasts the recruitment of pol II observed with a glucocorticoid time course (compare with Fig.  2B). Use of the sc-150 antibody shows that C/EBP␤ is present at the promoter regardless of hormone treatment; however, this result does not specify which form of C/EBP␤ is associated under these different hormone conditions. In contrast, use of the 2952 antibody demonstrates that LAP binding is disrupted in vivo with insulin treatment. Importantly, insulin-induced dissociation of pol II, CBP, and LAP is reversed in cells treated with the PI3K inhibitor LY 294002 (Fig. 5C). Taken together, these results clearly demonstrate that LIP replaces LAP at the endogenous PEPCK gene promoter in response to insulin and that this process depends on an intact PI3K signaling pathway. Unfortunately, a LIP-specific antibody is currently not available, so we are unable to perform the converse experiment.
Effects of LIP on Recruitment of Activators, PEPCK Gene Expression, and Glucose Production-To test whether LIP can disrupt the ability of hormones to recruit coactivators, activate gene transcription, and drive hepatocellular glucose production, an adenovirus construct was created to express T7-tagged LIP. Applying 4 ϫ 10 10 optical particle units of either Ad-GFP or Ad-LIP virus to each plate of H4IIE cells resulted in the successful infection of more than 90% of the cell population as indicated by GFP expression (Fig. 6A). Cells infected with the Ad-LIP virus exhibited dose-dependent expression of the T7tagged LIP construct (p24-T7-LIP is the p20-LIP protein with the addition of a 4-kDa T7 tag) as measured by Western blot FIG. 4. Insulin increases the LIP:LAP ratio in a PI3K-dependent manner. A, the presence of p34-LAP and p20-LIP in nuclear extracts from H4IIE cells treated with the following array of hormones for 2 h was detected by Western blot using C/EBP␤ antiserum (sc-150): no hormone (N), 100 M 8-CPT-cAMP (C), 500 nM dexamethasone (D), and 100 M 8-CPT-cAMP ϩ 500 nM dexamethasone (CD), all with or without 10 nM insulin. All results shown are representative of at least three independent experiments. B, the presence of p34-LAP and p20-LIP in nuclear extracts from cells treated with 10 nM insulin for the indicated times was detected by Western blot using C/EBP␤ antiserum (sc-150). C, the chemiluminescent bands were analyzed using a Fluor-S MultiImager with Quantity One software. Average LIP:LAP ratios from three independent experiments, such as those illustrated in B, were plotted against the duration of insulin treatment (gray boxes). The same method was used to measure the LIP:LAP ratio from cells treated simultaneously with insulin and the PI3K inhibitor LY 294002 (20 M) for 2 h (black box). (Fig. 6B). The in vivo association of T7-LIP, pol II, CBP, LAP, GR, and C/EBP␣ with the PEPCK gene promoter was measured using the ChIP assay on cells infected with Ad-LIP and treated with hormones (Fig. 6C). Using an antibody specific for the T7 tag, the results show that the T7-LIP protein binds the PEPCK gene promoter constitutively, much like the endogenous accessory factors. However, the increase in occupancy of pol II and CBP observed before upon glucocorticoid or glucocorticoid/cAMP induction (see Fig. 2A) is diminished significantly. Thus, the expression of exogenous LIP blocks the recruitment of CBP and pol II to the endogenous PEPCK gene promoter. At least in part this is probably because LIP lacks the aminoterminal activation domain of LAP, the same region that was shown to be important for the ability of C/EBP␤ to interact with p300/CBP (31). The decrease in PEPCK promoter occupancy by CBP was not the result of decreased CBP levels, because endogenous CBP was unaffected by LIP expression (data not shown). In addition, the expression of LIP blocks the association of LAP with the PEPCK gene promoter, reminiscent of that occurring after insulin treatment (Fig. 5B); however, it does not affect the occupancy of GR or C/EBP␣ on the promoter.
To examine the functional consequences of expressing an exogenous, dominant-negative construct, such as LIP, we measured PEPCK mRNA levels and glucose production across the same hormone conditions (Fig. 6, D and E). Ad-LIP infection, but not Ad-GFP infection, results in decreased PEPCK mRNA levels. Furthermore, the expression of exogenous LIP represses glucose production in a dose-dependent manner. Ad-LIP did not affect basal levels of PEPCK expression and glucose production (data not shown). Taken together, these results demonstrate that the expression of LIP is able to 1) disrupt the recruitment of activating factors to the PEPCK gene promoter in vivo; 2), as a functional consequence, repress endogenous PEPCK gene expression; and 3), as a physiological outcome, attenuate hepatocellular glucose production.

DISCUSSION
The data presented in this study demonstrate that at least one of the functions of hormone response unit accessory factors is to recruit coactivators to the promoter to drive transcription in response to hormones. Our results offer a model by which glucocorticoids induce, whereas insulin inhibits, PEPCK gene transcription (Fig. 7). In this model, the accessory factors bind constitutively to the promoter and assemble a platform that can serve to recruit activating factors in response to hormonal signals. Upon stimulation by glucocorticoids (dexamethasone), several factors assemble on the promoter to drive transcription, including GR, CBP, and the pol II holoenzyme. When glucocorticoids and the dominant inhibitor insulin are present simultaneously, the occupancy of CBP and pol II is lost, even though ligand-bound GR remains associated with the promoter.
In addition, our results show that insulin not only increases the protein levels of LIP, but also that it does so in a manner that relies on an intact PI3K signaling pathway. This correlates well with previous work in which insulin was shown to repress PEPCK gene expression in a PI3K-dependent manner (7). Furthermore, this increase in LIP leads to the concomitant replacement of LAP with LIP on the PEPCK gene promoter in vivo (Fig. 7). Finally, we show that the expression of LIP and its subsequent binding to the PEPCK gene promoter disrupt the recruitment of CBP and pol II by glucocorticoids, which culminates in the inhibition of PEPCK gene expression and hepatocellular glucose production.
Over the past few years, several different groups, including ours, have proposed a variety of putative insulin response factors, many of which are able to mimic the hormonal affects on PEPCK gene expression (20,21,(35)(36)(37)(38). To date, however, none affords the global regulation so characteristic of the ability of insulin to repress PEPCK gene transcription dominantly, along with a defined mechanism of action specific to a particular region of the PEPCK gene promoter. Others have recently alluded to, though not proven, a role for C/EBP␤ in repressing PEPCK expression by insulin (39). In fact, we cannot claim C/EBP␤ to be the one and only factor that acutely inhibits PEPCK gene transcription in response to insulin because new protein synthesis is not necessary for the immediate actions of insulin (40). Nevertheless, our results reveal a novel mechanism by which insulin does act, that is, by raising the protein FIG. 5. Insulin replaces LAP with LIP on the PEPCK gene promoter. A, parallel Western blots were performed on H4IIE cell nuclear extracts using 2952 (an antibody that recognizes p34-LAP but not p20-LIP) and sc-150 (an antibody that recognizes both forms of C/EBP␤). B, the association of pol II, CBP, C/EBP␤, and LAP with the endogenous PEPCK gene promoter was examined using the ChIP assay on H4IIE cells treated with the indicated hormones: no hormone (N), 500 nM dexamethasone (D), and 500 nM dexamethasone ϩ 10 nM insulin (DI). The cells were treated with hormones for 30 min, with the addition of an extended time course to determine the occupancy of the promoter by the above factors with prolonged exposure to insulin. All results depicted are representative of at least three independent experiments. C, likewise, the ChIP assay was used to assess the in vivo association of the above factors with the PEPCK gene promoter in cells treated with dexamethasone and insulin, with and without LY 294002. All results shown are representative of at least three independent experiments. levels of LIP (a natural inhibitory transcription factor) to compete away LAP (its activating counterpart) and thereby abrogating the ability of LAP to support the assembly of a CBP⅐pol II complex on the promoter. Loss-of-function experiments would be decisive, but these are not yet possible. To disrupt the function of LIP specifically, transgenic animals would need to be designed in such a way as to substitute the internal methionine, which serves as an alternative translation initiation site (33), with a carefully chosen amino acid. Additionally, a cell line would need to be established from such animals to conduct experiments in which glucose and counterregulatory hormones, all of which also affect PEPCK gene transcription, could be carefully controlled. Unfortunately, there is no guarantee that such a mutation would not alter the function of LAP; and, even if such a design were fruitful, LAP conversion to LIP by a post-translational processing enzyme has been proposed as an alternative mechanism for LIP expression (41,42). When a LIP-specific antibody becomes available, it would be interesting to continue this work by testing the association of LIP itself with the PEPCK gene promoter in vivo. FIG. 6. LIP abrogates the recruitment of CBP and pol II, represses PEPCK gene expression, and attenuates hepatocellular glucose production. A, the efficiency with which the adenovirus constructs (Ad-LIP or Ad-GFP) infected H4IIE cells was determined by measuring GFP expression levels. B, expression of p24-T7-LIP was confirmed by Western blot (using C/EBP␤ antibody sc-150) of nuclear extracts from cells infected with increasing amounts of the Ad-LIP virus. C, in vivo association of T7-LIP, pol II, CBP, LAP, GR, and C/EBP␣ with the PEPCK gene promoter was measured using the ChIP assay on cells infected with Ad-LIP and treated with the following hormones: no hormone (N), 500 nM dexamethasone (D), 100 M 8-CPT-cAMP ϩ 500 nM dexamethasone (CD), and 500 nM dexamethasone ϩ 10 nM insulin (DI). All results shown are representative of at least three independent experiments. D, PEPCK gene expression levels were measured by primer extension of RNA preparations from cells treated with dexamethasone (Dex) and infected with increasing amounts of Ad-GFP or Ad-LIP. ␤-Actin mRNA levels were measured for internal controls. E, glucose production was measured by incubating cells with glucose-free medium supplemented with lactate and pyruvate after dexamethasone treatment and adenovirus infection. The bars depict the mean Ϯ S.E. of at least three individual experiments.
FIG. 7. Glucocorticoids promote, whereas insulin disrupts, the assembly of a transcriptionally active PEPCK gene promoter. Across different hormonal environments, accessory factors bind to the endogenous PEPCK gene promoter. C/EBP␤ is shown in green, whereas other factors that bind constitutively to the PEPCK gene promoter are shown in gray. Basal transcription is maintained by weak or transient recruitment of the pol II holoenzyme. Glucocorticoids induce transcription by recruiting GR, CBP, and pol II to the promoter. Insulin represses transcription by promoting the replacement of LAP (ovals) with LIP (half-ovals) on the PEPCK gene promoter, thereby disrupting the recruitment of CBP and pol II. Meanwhile, GR remains bound in the presence of insulin as long as glucocorticoids are present. GR is shown in orange, CBP is shown in yellow, and the pol II holoenzyme is shown in blue.
Insulin is a central regulator of metabolism, opposing the effects of many different hormones (e.g. glucagon, glucocorticoids, retinoic acid, thyroid hormone) that can induce the expression of gluconeogenic enzymes, such as PEPCK. Each of the activating hormones up-regulates PEPCK gene transcription through separate, albeit overlapping, hormone response units assembled on the promoter. Therefore, insulin must possess either the capacity to counteract each hormone response (and basal transcription) individually through different means or the ability to disrupt a common mechanism of gene transcription, shared by all activating signals. The results from this study support the latter and provide evidence to suggest that insulin inhibits gene transcription, at least in part, by disturbing the interactions between promoter-bound transcription factors and the coactivator-general transcription machinery complex recruited to drive transcription. This novel mechanism could account for the action of insulin at the PEPCK gene promoter and other genes involved in the control of glucose homeostasis.