Regulation of Glucose-6-phosphatase Gene Expression by Protein Kinase Bα and the Forkhead Transcription Factor FKHR

Glucose-6-phosphatase plays an important role in the regulation of hepatic glucose production, and insulin suppresses glucose-6-phosphatase gene expression. Recent studies indicate that protein kinase B and Forkhead proteins contribute to insulin-regulated gene expression in the liver. Here, we examined the role of protein kinase B and Forkhead proteins in mediating effects of insulin on glucose-6-phosphatasepromoter activity. Transient transfection studies with reporter gene constructs demonstrate that insulin suppresses both basal and dexamethasone/cAMP-induced activity of the glucose-6-phosphatasepromoter in H4IIE hepatoma cells. Both effects are partially mimicked by coexpression of protein kinase Bα. Coexpression of the Forkhead transcription factor FKHR stimulates the glucose-6-phosphatase promoter activity via interaction with an insulin response unit (IRU), and this activation is suppressed by protein kinase B. Coexpression of a mutated form of FKHR that cannot be phosphorylated by protein kinase B abolishes the regulation of theglucose-6-phosphatase promoter by protein kinase B and disrupts the ability of insulin to regulate theglucose-6-phosphatase promoter via the IRU. Mutation of the insulin response unit of the glucose-6-phosphatase promoter also prevents the regulation of promoter activity by FKHR and protein kinase B but only partially impairs the ability of insulin to suppress both basal and dexamethasone/cAMP-stimulated promoter function. Taken together, these results indicate that signaling by protein kinase B to Forkhead proteins can account for the ability of insulin to regulateglucose-6-phosphatase promoter activity via the IRU and that other mechanisms that are independent of the IRU, protein kinase B, and Forkhead proteins also are important in mediating effects of in insulin on glucose-6-phosphatase gene expression.

Glucose-6-phosphatase (Glc-6-Pase) 1 catalyzes the hydrolysis of glucose 6-phosphate to glucose, which is the terminal step of both hepatic gluconeogenesis and glycogen breakdown. Glc-6-Pase is induced in starved and diabetic animals (1,2). In vitro models have shown that glucocorticoids and cAMP induce Glc-6-Pase gene expression. This effect is opposed by insulin, which also is able to reduce basal expression of the Glc-6-Pase gene (3)(4)(5)(6). Identification of the signaling events that connect the insulin receptor to the Glc-6-Pase promoter, leading to the subsequent repression of gene transcription, is of particular interest because Glc-6-Pase plays a key role in the regulation of hepatic glucose production and blood glucose homeostasis.
In H4IIE hepatoma cells, activation of class 1a phosphoinositide 3-kinase (PI 3-kinase), but not of the Ras/Raf/MAP kinase pathway, is necessary for the suppression of Glc-6-Pase promoter activity by insulin (6). The formation of PtdIns(3,4,5)P 3 catalyzed by PI 3-kinase has been shown to increase the activity of 3-phosphoinositide-dependent protein kinase-1 (PDK1) and to result in a conformational change in PKB which renders it susceptible to phosphorylation and activation by PDK1. PDK1 phosphorylates also activate other members of the AGC family of protein kinases, in addition to protein kinase B (PKB) (see Refs. 7-9 for review). The stimulation of PKB suppresses basal activity of the insulin-like growth factor-binding protein-1 (IGFBP-1) promoter via a conserved insulin response sequence (10) and represses the Bt 2 cAMP-and dexamethasone-induced phosphoenolpyruvate carboxykinase (PEPCK) gene transcription (11). PKB also stimulates the fatty-acid synthase gene promoter (12).
PKB is known to translocate to the nucleus of cells after insulin treatment (13), where it can phosphorylate members of the Forkhead/winged helix family of transcription factors, including FKHR, FKHRL1, and AFX (Refs. 14 -19 and reviewed in Ref. 20). Genetic studies in Caenorhabditis elegans indicate that signaling of PKB via Forkhead transcription factors plays an important role in mediating effects of insulin signaling, where PKB homologues (AKT1 and AKT2) are thought to suppress dauer formation via the phosphorylation of DAF-16, a member of the Forkhead/winged helix family of transcription factors. DAF-16 has a high homology to a subfamily of human Forkhead proteins, including FKHR, FKHRL1, and AFX (21,22). Recent studies have shown that FKHR binds to Forkheadbinding sites within the insulin response element of the IG-FBP-1 promoter and stimulates promoter activity in a sequence-specific fashion (16,17,19,23). Insulin suppresses this transactivation via phosphorylation of FKHR, leading to the exclusion of the transcription factor from the nucleus (24,25). After the administration of insulin, FKHR becomes phosphorylated in vivo at three PKB phosphorylation sites (Thr-24, Ser-256, and Ser-319) in a PI 3-kinase-dependent manner (16,17,19,25). PKB is able to phosphorylate all three sites in vitro and after the overexpression of the kinase in vivo (16,17,19). By using insulin receptor-deficient hepatocytes, evidence has been provided that one of these sites (Thr-24) also may be phosphorylated by a kinase that is distinct from PKB␣ (25).
An insulin response unit (IRU) has been identified within the Glc-6-Pase promoter. This IRU contains three potential Forkhead-binding sites with sequences similar to the insulin response sequences in the IGFBP-1 gene (5,17,26,27). This IRU is involved in the repression of basal and stimulated Glc-6-Pase gene transcription by insulin (5,26). The Glc-6-Pase IRU region is located between nucleotides Ϫ198 and Ϫ159 in the mouse promoter (5) and between nucleotides Ϫ196 and Ϫ156 in the human promoter (5,26,27). Recently, it has been shown that recombinant FKHR is able to bind to a double-stranded oligonucleotide probe containing this sequence (27). In the present paper, we report that FKHR is able to stimulate and that PKB is able to decrease basal and dexamethasone/cAMP-stimulated Glc-6-Pase promoter activity. These effects are mediated by the IRU and are blocked by a PKB-insensitive FKHR mutant.

Materials
Restriction endonucleases, modifying enzymes, and luciferase assay reagent were purchased from Promega. Plasmid purification systems were from Qiagen. N 6 ,2Ј-O-dibutyryl cyclic AMP (Bt 2 cAMP) was purchased from Roche Molecular Biochemicals. LY294002 was from Calbiochem. All other reagents were purchased in analytical grade from either Sigma or Merck.

Methods
Plasmid Construction and Site-directed Mutagenesis-The Glc-6-Pase reporter gene construct Glc-6-Pase(Ϫ1227/ϩ57) was created by cloning the human Glc-6-Pase promoter fragment Ϫ1227 to ϩ57 into the promoterless luciferase reporter gene vector pGL3 basic (Promega) (4). To generate the construct Glc-6-Pase(Ϫ1227/ϩ57/IRUmut) the IRU within the vector Glc-6-Pase(Ϫ1227/ϩ57) was mutated from Ϫ196 5Ј-CGATCAGGCTGTTTTTGTGTGCCTGTTTTTCTATTTTACG-3Ј Ϫ156 to 5Ј-CGATCAGGCTCGAGTTGTGTGCCTCTTTTTCTCTTTTACG-3Ј (mutated nucleotides are underlined). These mutations replace residues that are critical for mediating effects of insulin via related insulin response sequences (T(G/A)TTT) (11,14,15,20) and include mutations of nucleotides that have been shown to be critical for the binding of recombinant FKHR to the IRU (27). The construct pGL(PDH) was created by cloning nt Ϫ929 to ϩ79 of the E1␣-subunit of the human pyruvate dehydrogenase (E1␣-PDH) gene into the HindIII/XhoI sites of pGL3. The construct pGL(PDH-IRU) was created by cloning a doublestranded oligonucleotide with the sequence between nt Ϫ196 and nt Ϫ156 of the Glc-6-Pase promoter with flanking ApaI sites into the respective restriction site at nt Ϫ217 of the E1␣-PDH promoter fragment in construct pGL(PDH). The expression vectors for hemagglutinin epitope (HA)-tagged wild type PKB␣ (HA-PKB␣), the constitutively active HA-T308D/S473D-PKB␣ (HA-CA-PKB␣), and the "kinase-dead" mutant HA-K179A-PKB␣ (HA-KD-PKB␣) have been described elsewhere (28). The expression vectors for FKHR and the TSS-Ala FKHR mutant, in which the phosphorylation sites Thr-24, Ser-256, and Ser-319 are mutated to Ala, were described previously (16,17). The construct Helix3mut FKHR was generated by site-directed mutagenesis as previously reported (17) using a single-stranded oligonucleotide containing the following sequence: 5Ј-TAGGGACAGATTAAGACGAATTG-AATTGAATTCTTCCCGCCCGCCGAGCTGTT-3Ј. This results in the mutation of Trp-209 to Gly and His-215 to Pro. Both sites are highly conserved in Forkhead proteins. His-215 is thought to make direct contact with DNA-binding sites (29), and the mutation of His-215 alone has been shown to disrupt binding of FKHR to the IGFBP-1 insulin response element (19).
In some experiments the expression vectors were replaced by pCI-Neo (Promega) as a vector control. After the glycerol shock (15%, 2 min), the transfected cells were incubated for 1 h without serum and then in the presence or absence of dexamethasone (1 M), Bt 2 cAMP (500 M), and insulin for 18 -20 h. In some experiments transfected cells were preincubated for 10 min with LY294002 (100 M) before the other hormones were added. Cell extracts were prepared, and luciferase activities were determined using the Dual-luciferase Assay Reagent (Promega), according to the manufacturer's instructions. All experiments were performed at least three times each in triplicate with at least two different DNA preparations. Statistical analysis was performed using the InStat program. In order to overexpress FKHR in mammalian cells, 2 ϫ 10 6 HEK 293 cells were transiently transfected with 15 g of the expression vector for FKHR or pCI-Neo using the calcium phosphate/DNA coprecipitation method. After 16 h the medium was aspirated and replaced with fresh Dulbecco's modified Eagle's medium, 10% fetal calf serum. The cells were incubated for 12 h and then serum-starved for 12 h. The nuclear extracts were prepared as described below.
Preparation of Nuclear Extracts and Electromobility Shift Assay (EMSA)-7 ϫ 10 6 H4IIE or HepG2 cells were serum-starved for 24 h and then incubated with wortmannin (50 nM) for 15 min to reduce phosphorylation of Forkhead proteins that might interfere with binding. Nuclear extracts were prepared according to the method of Schreiber et al. (30), and protein content was measured by the Bio-Rad dye binding assay. Double-stranded oligonucleotides containing the wild type IRU 5Ј-CAGGCTGTTTTTGTGTGCCTGTTTTTCTATTTTACGTA-A-3Ј (IRU) or mutated IRU sequence 5Ј-CAGGCTCGAGTTGTGTGCC-TCTTTTTCTCTTTTACGTAA-3Ј (IRUmut) were end-labeled with [␥-32 P]ATP using T4 polynucleotide kinase. For EMSA, nuclear extracts (10 g of protein) were incubated in a total of 20 l with a 32 P-labeled double-stranded oligonucleotides (25,000 cpm) in binding buffer (20 mM HEPES/NaOH (pH 7.5), 100 mM NaCl, 12.5% (v/v) glycerol, 50 ng/l of poly(dI-dC)-poly(dI-dC), 500 ng/l bovine serum albumin, 500 M dithiothreitol) for 20 min at room temperature. Where indicated nuclear extracts were preincubated for 15 min with either 2 g of an anti-FKHR antibody (Santa Cruz Biotechnology) or the indicated molar excess of unlabeled oligonucleotide competitors. Bound and free probe were resolved on 6% nondenaturing polyacrylamide gels prior to autoradiography.
Western Blotting-20 g of nuclear proteins from H4IIE cells or HepG2 cells were resolved by 10% SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes (Schleicher & Schü ll). The filters were blocked using Blocking Reagent (Roche Molecular Biochemicals) and incubated with 0.2 g/ml anti-FKHR antibody (Santa Cruz Biotechnology) overnight at 4°C. The secondary antibody was alkaline phosphatase-conjugated anti-mouse Ig from goat (Dianova), diluted 1:40,000 and incubated with the blot for 2 h. Detection was performed using nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate stock solution (Roche Molecular Biochemicals).
Determination of PKB Isoenzyme Activities-Serum-starved H4IIE cells were incubated in the presence or absence of dexamethasone (1 M), Bt 2 cAMP (500 M), insulin (500 nM), and the PI 3-kinase inhibitor LY294002 (100 M). At the indicated time points, cell extracts were prepared, and PKB isoenzymes were immunoprecipitated using isoform-specific antibodies (31). The immunoprecipitates were assayed for PKB activity using Crosstide as substrate, essentially as described previously (31). One unit of protein kinase activity was that amount that catalyzed the phosphorylation of 1 nmol of substrate in 1 min.
Immunoprecipitation and Kinase Activity of HA-tagged PKB␣-5 ϫ 10 6 H4IIE cells were transfected in 14.5-cm dishes with 34 g of the indicated expression vector for HA-tagged PKB␣ as described above. After the glycerol shock, the cells were serum-starved for 18 h and then incubated in the presence or absence of 500 nM insulin for 10 min. The cells were lysed in 50 mM Tris-HCl, pH 7.5, 0.1% (m/v) Triton X-100, 1 mM EDTA, 1 mM EGTA, 50 mM NaF, 10 mM sodium ␤-glycerophosphate, 1 mM sodium orthovanadate, 0.1% (v/v) 2-mercaptoethanol, and 1 M microcystin-LR. The immunoprecipitation was carried out using 1.2 mg of protein of the cell lysates and 5 l of protein G-Sepharose coupled to 2 g of HA monoclonal antibody for 2 h as described (28). The immunoprecipitate was assayed for PKB activity using Crosstide as substrate (28).
RT-PCR of FKHR and RNase Protection Assay-Total RNA was prepared from rat H4IIE cells using the RNeasy mini-Kit (Qiagen) according to the manufacturer's instructions and quantified by absorption at 260 nm. cDNA was prepared by reverse transcription of 1 g of total RNA using oligo(dT) 15 primers and a reverse transcription kit (Promega). FKHR cDNA were amplified by 35 cycles of PCR using Taq polymerase (Amersham Pharmacia Biotech). The primers for the PCR were designed based on sequences that are conserved in mouse and human FKHR using H4IIE RNA, and sequences for primers were 5Ј-TCATCACCAAGGCCATCGAG-3Ј(sense) and 5Ј-GTTATGAGATGC-CTGGCTGC-3Ј(antisense). The RT-PCR product (1 kilobase pair) was cloned into pGEM-T Easy (Promega) and sequenced (ALF, Amersham Pharmacia Biotech). For RNase protection assay, a 302-bp NcoI/AccI fragment encoding nucleotides 245-547 of the RT-PCR product of rat FKHR was subcloned into pGEM5, and a cRNA complementary to FKHR RNA was synthesized in the presence of [␣-32 P]CTP using the Sp6 promoter. This probe (500,000 cpm) was hybridized for 18 h at 45°C with 50 or 100 g of total RNA from H4IIE cells and then digested with RNase T1 and RNase A (Roche Molecular Biochemicals). Protected fragments were separated by 8 M urea, 6% polyacrylamide gel electrophoresis. The probe and protected fragment were sized by comparison to X174 DNA standards (Promega) end-labeled with T4 polynucleotide kinase.

RESULTS
Several reports have shown that insulin leads to the activation of PKB (7-10, 28, 31). Three isoenzymes of PKB have been described, named ␣, ␤, and ␥. In order to study a potential role of PKB in the regulation of Glc-6-Pase gene expression, we first studied the activation of PKB isoenzymes in H4IIE hepatoma cells. In this study, insulin led to a strong activation of the PKB␣ isoform (Table I). The PKB␤ isoform was also activated but to a lesser degree. No PKB␥ activity was detectable (data not shown). 10 min after insulin administration half-maximal activation of the PKB␣ isoform was observed. The application of the PI 3-kinase inhibitor LY294002 completely blocked activation of PKB by insulin. The addition of Bt 2 cAMP and dexamethasone reduced PKB activity after 900 min but did not disrupt the ability of insulin to stimulate PKB␣ activity at 120 and 900 min. These data indicate that PKB␣ is the predominant form activated by insulin in H4IIE hepatoma cells, which corresponds to the situation in isolated hepatocytes (31). In addition, activation of PI 3-kinase, which is essential for the regulation of Glc-6-Pase gene expression by insulin (6), also is required for the activation of PKB in H4IIE cells.
In order to study a possible role of PKB␣ activation in the control of Glc-6-Pase gene expression, we overexpressed different HA-tagged PKB␣ mutants in H4IIE cells and immunoprecipitated the expressed proteins with an anti-HA antibody. After the overexpression of wild type HA-PKB␣, a specific activity of 0.1 milliunit/mg total protein was measured, which was increased over 10-fold by insulin stimulation. 2 The overexpression of the constitutively active construct HA-CA-PKB␣ yielded a 3-fold higher specific PKB activity, which was not increased further by insulin. No activity was detectable with the catalytically inactive mutant HA-KD-PKB␣. These results are in agreement with data obtained by the transfection of other cell lines with the HA-PKB␣ constructs (28).
Insulin treatment suppresses the basal as well as the Bt 2 cAMP/dexamethasone-induced expression of a luciferase reporter gene under control of the human Glc-6-Pase promoter fragment Ϫ1227/ϩ57 in transiently transfected H4IIE hepatoma cells (Fig. 1). Coexpression studies revealed that the expression of either the wild type HA-PKB␣ or the constitutively active form HA-CA-PKB␣ also reduces both basal (   either basal or induced promoter activity or the effect of insulin compared with a vector control. Although the expression of HA-CA-PKB␣ led to a higher PKB activity in the cells than that of the construct HA-PKB␣ (above), these constructs are equally effective in suppressing Glc-6-Pase promoter activity and cause only a partial suppression compared with the administration of insulin. Increasing the amount of expressed HA-CA-PKB␣ does not lead to a greater reduction of either basal or stimulated Glc-6-Pase promoter activation (data not shown). These results indicate that PKB␣ activity suppresses Glc-6-Pase promoter activity but only partially accounts for the effect of insulin.
PKB is known to phosphorylate the Forkhead transcription factor FKHR (16 -20, 23-25). In order to study a potential role of FKHR in the regulation of Glc-6-Pase promoter activity, we determined whether FKHR is expressed in H4IIE cells. We first performed RT-PCR of total RNA prepared from H4IIE cells RNA using primers based on sequences that are conserved in mouse and human FKHR genes. The RT-PCR product was cloned, and three clones were sequenced (GenBank TM accession number AF247812). The deduced amino acid sequence of this cDNA showed the highest homology to the mouse (83%) and to the human (80%) FKHR and a lower homology to mouse FKHRL1 (45%) and AFX (39%), indicating that it was derived from RNA coding for rat FKHR. The predicted amino acid sequence of this fragment revealed that the phosphorylation sites for PKB corresponding to Ser-256 and Ser-319 in the human FKHR are conserved. FKHR mRNA in H4IIE cells was also detected by RNase protection assay. We used a labeled RNA probe containing 302 bp of FKHR cRNA and 26 bp complementary to the vector. A fragment from this probe of approximately 300 bp was protected from digestion with RNase by hybridization with total H4IIE RNA but not by tRNA ( Fig.  2A). The presence of FKHR protein in H4IIE cells was studied by Western blotting of nuclear extracts prepared from serumstarved and wortmannin-treated H4IIE cells. An antibody against FKHR recognized a band of approximately 70 kDa (Fig.  2B). Nuclear extract of human HepG2 cells, which express FKHR (23), was used as a positive control. Together, these results demonstrate that FKHR mRNA and protein are expressed in H4IIe cells.
Previous EMSA studies have shown that recombinant FKHR is able to bind to a double-stranded oligonucleotide probe containing the sequence of the Glc-6-Pase IRU in a sequencespecific fashion (27). We have confirmed that FKHR also interacts with the Glc-6-Pase IRU in a sequence-specific fashion when it is overexpressed in 293 cells. As shown in Fig. 3A, an oligonucleotide probe containing the Glc-6-Pase IRU forms four complexes with nuclear extracts prepared from 293 cells transfected with the FKHR expression vector. The complexes a and b were supershifted by an antibody against FKHR. The formation of the complexes a and b was not observed when nuclear extracts are incubated with a probe containing a mutated form of the Glc-6-Pase IRU (IRUmut), which does not interact with recombinant FKHR in vitro (27) or mediate effects of insulin on promoter activity in reporter gene studies ( Ref. 27 and below). The complexes c and d, but not a and b, were as well formed using extracts from mock-transfected cells, which do not overexpress FKHR. By using this protein extract two additional complexes were observed. The results indicate that complexes a and b were formed by the interaction of FKHR or an immunologically related protein with the IRU of the Glc-6-Pase.
We also examined whether we could detect interactions between endogenous FKHR and the Glc-6-Pase IRU by EMSA. As shown in Fig. 3B, oligonucleotide probes containing either the wild type Glc-6-Pase IRU or the mutant (IRUmut) form the same nucleoprotein complexes after incubation with nuclear extracts prepared from H4IIE cells. Also, an excess of unlabeled oligonucleotide competitors containing either the wild type IRU or mutated sequence inhibits the formation of the complexes that are formed with either probe. Since the mutated IRU does not mediate effects of insulin on reporter gene function (Ref. 27 and below), these results indicate that these complexes are not likely to involve interaction with proteins that are critical for mediating effects of insulin on promoter function. Preincubation with anti-FKHR antibody does not supershift or disrupt the formation of these complexes. Together, these results indicate that FKHR is not likely to be present in these complexes and that this assay is not sufficiently sensitive to detect readily interactions between the Glc-6-Pase IRU and endogenous FKHR.
To determine whether FKHR may interact with the Glc-6-Pase IRU in vivo, we cotransfected H4IIE cells with FKHR expression vectors and the Glc-6-Pase reporter gene construct. As shown in the left panel of Fig. 4A, both wild type FKHR and the mutant TSS-Ala FKHR, which is not phosphorylated by PKB (17), strongly stimulate Glc-6-Pase promoter activity. This effect is dramatically reduced by transfection with the Helix3mut FKHR expression vector, which possesses two mutations within the DNA binding domain (19,29). Mutation of the IRU in the Glc-6-Pase promoter (Glc-6-Pase(Ϫ1227/ϩ57/ IRUmut)) also dramatically reduces the ability of FKHR and TSS-Ala FKHR to stimulate promoter function (Fig. 4A, right  panel). Together, these results indicate that FKHR stimulates Glc-6-Pase promoter activity gene via binding to the IRU. This was confirmed by cloning the IRU into the heterologous E1␣-PDH gene promoter, creating the pGL(PDH-IRU) construct. This led to an about 2-fold activation of the promoter after the coexpression of FKHR and TSS-Ala FKHR but not Helix3mut FKHR (Fig. 4B). In contrast, FKHR and TSS-Ala FKHR do not stimulate the activity of the pGL(PDH) construct, indicating that this effect is mediated via the IRU. In order to characterize the role of FKHR in the regulation of Glc-6-Pase gene expression by PKB, we coexpressed FKHR and the HA-PKB␣ proteins together with the Glc-6-Pase reporter gene construct in H4IIE cells. As shown in Fig. 5A, constitutively active HA-PKB␣ reduces the activation of basal Glc-6-Pase promoter activity by FKHR by about 80%, whereas catalytically inactive PKB does not disrupt the effect of FKHR on Glc-6-Pase promoter activity. Coexpression of TSS-Ala FKHR prevents the regulation by constitutively active PKB␣ (Fig.  5A), even after the coexpression of 5 times higher amounts of HA-CA-PKB␣ than shown in Fig. 5 (data not shown). Coexpression of the TSS-Ala FKHR mutant had no significant effect on the extent of the induction of the promoter by Bt 2 cAMP/dexamethasone but abolished the suppression of Bt 2 cAMP/dexamethasone-induced reporter gene expression by PKB␣ (Fig. 5B). This suggests that PKB␣ may suppress both basal and the Bt 2 cAMP/dexamethasone-induced Glc-6-Pase promoter activity primarily by disrupting transactivation by FKHR.
This hypothesis demands that the Forkhead-binding sites within the IRU of the Glc-6-Pase promoter (27) should be the cis-acting sequence for the PKB effect. In order to test this possibility, we coexpressed the Glc-6-Pase promoter construct (Ϫ1227/ϩ57/IRUmut) together with the PKB constructs. As shown in Fig. 6A, mutation of the IRU prevents the inhibition of basal activity by active forms of PKB␣ (Fig. 6A). The induction of luciferase expression by Bt 2 cAMP/dexamethasone was slightly lower in the IRU-mutated Glc-6-Pase promoter compared with wild type Glc-6-Pase promoter (5.4 Ϯ 0.5-fold versus 7.6 Ϯ 0.8-fold). As shown in Fig. 6B, the induction of Glc-6-Pase promoter activity is not opposed by the overexpression of either HA-PKB␣ or HA-CA-PKB␣ when the IRU is mutated. However, insulin can still inhibit the activity of mutated Glc-6-Pase promoter (Ϫ1227/ϩ57/IRUmut), although not as effectively as the wild type promoter (Fig. 1, see below).
Insulin was able to inhibit the basal expression of the construct pGL(PDH-IRU) significantly but not that of the construct pGL(PDH). This confirms previous reports demonstrating that the IRU of the Glc-6-Pase promoter is able to transfer insulin regulation to a heterologous promoter (5,26). In addition, we found that the overexpression of the constitutively active PKB was able to suppress the basal expression of the construct pGL(PDH-IRU) but not that of the construct pGL(PDH) (Fig. 7A). This PKB effect was blocked by the coexpression of TSS-Ala FKHR (Fig. 7B). Taken together, these results indicate that the IRU of the Glc-6-Pase promoter is critical for stimulation by FKHR and suppression by PKB.
Next, we wanted to assess the relative role of the PKB/ Forkhead pathway mediating the overall effect of insulin on promoter activity via the IRU. Insulin (0.5 nM) was able to inhibit basal and induced promoter activity by about 80 and 90%, respectively, after coexpression of the wild type Glc-6-Pase promoter together with the wild type FKHR mutant (Fig.  8, A and B). This effect of insulin was blocked by LY294002, indicating that it is mediated via a PI 3-kinase-dependent mechanism. Interestingly, coexpression of the TSS-Ala FKHR mutant, which completely blocked the PKB effect on the Glc-6-Pase gene expression, caused only a modest reduction in this effect of insulin, leading to an overall inhibition of about 60% by insulin, and this effect of insulin also was blocked by LY294002. In contrast, overexpression of TSS-Ala FKHR completely blocked the ability of insulin to suppress promoter activity in the pGL(PDH-IRU) construct. Together, these results suggest that signaling via PKB to Forkhead proteins may play a role in mediating effects of insulin on promoter activity via the Glc-6-Pase IRU and that other PI 3K-dependent mechanisms involving cis-acting sequences outside the IRU also may contribute to the ability of insulin to suppress activity of the Glc-6-Pase promoter.
To evaluate this possibility further, we examine the ability of insulin to suppress promoter activity in constructs where the IRU has been mutated. As shown in Fig. 9, the ability of insulin to inhibit basal (Fig. 9A) and stimulated (Fig. 9B) promoter activity in H4IIE cells is reduced by 20 -30% when the Glc-6-Pase IRU is mutated, compared with the wild type promoter. Interestingly, this IRU-independent effect of insulin also is blocked by treatment with LY294002, indicating that it is mediated via a PI 3-kinase dependent mechanism (Fig. 10, A and  B). The ability of insulin to inhibit promoter activity in the Glc-6-Pase(Ϫ1227/ϩ57/IRUmut) construct was not affected by the coexpression of either FKHR or TSS-Ala FKHR, which do not transactivate this promoter (Fig. 11), indicating that this effect of insulin does not involve signaling via PKB to Forkhead proteins. Taken together, these results indicate that signaling via PKB to Forkhead proteins and the IRU accounts for about 20 -30% of the effect of insulin on Glc-6-Pase promoter activity in H4IIE cells.

DISCUSSION
In this study, we have demonstrated that PKB␣ is able to suppress Glc-6-Pase promoter activity. This provides further evidence for a role of this protein kinase in the regulation of metabolism (8) and supports the concept that it can contribute to the control of net hepatic glucose production. The effect of PKB on basal activity of the Glc-6-Pase promoter requires Forkhead-binding sites within the IRU and could be mediated by disrupting transactivation by Forkhead transcription factors. In this respect, the regulation of basal Glc-6-Pase promoter activity in H4IIE cells by PKB is similar to that of the IGFBP-1 gene promoter in HepG2 cells (17,19,23). Our results suggest that the same mechanism also may account for the effect of PKB on Bt 2 cAMP/dexamethasone-induced promoter activity as well.
We detected FKHR expression in H4IIE cells, supporting the concept that endogenous Forkhead proteins may contribute to the regulation of Glc-6-Pase gene expression. Recombinant FKHR is able bind to the IRU of the Glc-6-Pase promoter, and we could show that the mutation of the Forkhead-binding sites within the IRU of the Glc-6-Pase promoter prevented both the transactivation of promoter activity by coexpressed FKHR as well as the regulation by PKB. However, we were not able to detect the binding of endogenous FKHR to an oligonucleotide probe with the sequence of the IRU by EMSA. Binding sites for HNF-3 and the glucocorticoid receptor have been described within the IRU (2), and interactions with these transcription factors might be responsible for the complexes we observed in the EMSA. However, the functions of these proteins have not been described to be regulated by PKB, and competitive binding studies with an unlabeled oligonucleotide probe containing a mutated IRU sequence indicate that the complexes we did observe are not likely to be critical for mediating effects of insulin on promoter function. These results are similar to those of Durham et al. (23) who showed that recombinant FKHR is able to bind to the insulin response element of the IGFBP-1 promoter and is able to transactivate the IGFBP-1 promoter, but did not detect the binding of endogenous FKHR to the promoter using extracts of HepG2 cells, presumably because levels of endogenous FKHR levels were too low to detect the formation of a protein complex under the experimental conditions of the EMSA. Since AFX and FKHRL-1 bind to similar DNA elements (13,20), are able to stimulate promoter activity via related insulin response sequences in cotransfection studies (13-15, 17, 18, 20, 23), and contain conserved PKB phosphorylation sites, we cannot exclude the possibility that FKHRL1 (and AFX), and possibly other factors also might contribute to the regulation of Glc-6-Pase promoter activity in H4IIE cells. HNF-3␤, which is also a Forkhead protein, acts as an accessory factor for the glucocorticoid stimulation of the PEPCK promoter (32). Our observation that mutation of the Forkheadbinding sites in the Glc-6-Pase IRU reduces the ability of dexamethasone/cAMP to stimulate promoter activity might be an indication for a similar role of Forkhead proteins in the induction of the Glc-6-Pase promoter by glucocorticoids. Nuclear exclusion of such an accessory protein after insulin treatment would directly contribute to the inhibition of the dexamethasone-induced Glc-6-Pase expression. In addition, we found that the overexpression of FKHR stimulates the Glc-6-Pase promoter more strongly than the E1␣-PDH promoter containing the IRU of the Glc-6-Pase gene. Therefore, it might be that the transactivation by FKHR is enhanced by additional proteins, which bind to cis-active elements located in the Glc-6-Pase promoter, but not in the pGL(PDH-IRU) construct. Previous studies suggest that HNF-1 acts as an accessory factor for the insulin effect mediated by the IRU of the Glc-6-Pase promoter (26). A recent report (33) indicates that HNF-1 and the Forkhead protein HNF-3␤ exhibit a synergistic effect on the GLUT2 gene promoter activity. It will be interesting to determine whether HNF-1 and FKHR also function cooperatively in stimulating the activity of the Glc-6-Pase promoter and in mediating the effects of insulin on Glc-6-Pase gene expression.
The IRU of the Glc-6-Pase promoter has been previously characterized as a cis-acting sequence involved in the regulation of basal promoter activity by insulin (5,26). Our results indicate that signaling via PKB to Forkhead proteins may mediate this effect of insulin. At the same time, we also found that insulin regulates basal and induced Glc-6-Pase promoter activity by insulin even after the mutation of the IRU. This result indicates that additional cis-activating elements and pathways also play a role in mediating the effect of insulin. This finding is similar to the situation in the PEPCK promoter, where insulin has been shown to regulate stimulated promoter activity even after disruption of a known insulin response element (34 -36). Mechanisms similar to those suggested for the PEPCK promoter (34 -36) may also contribute to the regulation of the Glc-6-Pase gene transcription by insulin.
The overall effect of insulin on the activity of the Glc-6-Pase promoter (before and after mutation of the IRU) that we observed is greater than the effect that has been reported in previous studies (5,26). A possible explanation for this discrep- ancy is that HepG2 cells were used in previous studies (5,26), whereas we performed our experiments in H4IIE cells. These cell lines show differences in their insulin responsiveness. For example, insulin can suppress cAMP-induced PEPCK expression in H4IIE cells but not in HepG2 cells (36). In addition, differences in the regulation of Glc-6-Pase gene expression between these two cell lines have been reported (4). Furthermore, we mutated the IRU within a human promoter fragment which extends from nt Ϫ1227 to ϩ57, whereas a shorter fragment of the mouse promoter (nt Ϫ 751 to Ϫ ϩ 66) was studied by others (26). It is possible that additional cis-active elements are present in our construct which could contribute to the regulation of Glc-6-Pase gene transcription by insulin.
In the present paper, we found that activated and wild type PKB were equally effective in suppressing the activity of the Glc-6-Pase promoter, although transfection with the constitutively active construct led to higher PKB activity than wild type PKB in unstimulated cells. It has previously been reported that low levels of PKB activity are sufficient to suppress cAMP/ dexamethasone-induced PEPCK expression (11). At the same time, it is possible that overexpression of PKB might suppress Glc-6-Pase gene expression by a mechanism that is not related to its enzymatic activity, for example by sequestering intracellular targets, including endogenous FKHR. However, we also found that the effect of PKB is blocked by the overexpression TSS-Ala FKHR, which is not phosphorylated by PKB. This result provides strong evidence that PKB suppresses Glc-6-Pase promoter activity by its enzymatic activity and the phosphorylation of Forkhead proteins.
We also found that basal activity of the Glc-6-Pase promoter and the effect of insulin are slightly reduced when the control vector for the PKB constructs is coexpressed with reporter gene constructs. The reason for this result is unclear, but might be caused by a competition for transcription factors and/or their coactivators. Nevertheless, we did not observe an inhibition of the insulin effect by the coexpression of the kinase-dead PKB compared with the control vector. This construct has been reported to have dominant negative effects on the regulation of IGFBP-1 gene promoter activity by insulin and PI 3-kinase in HepG2 cells (10). However, the failure of the catalytically inactive mutant to block the effect of insulin on Glc-6-Pase promoter activity does not exclude a participation of PKB in mediating this effect of insulin. As previously noted (11), very high levels of expression of dominant negative mutants are required to block the activation of endogenous PKB, and such a high level of expression cannot be achieved in transient expression experiments of H4IIE cells due to the low transfection efficiency of this cell type.
Due to this limitation in the use of dominant negative PKB constructs in H4IIE cells and the absence of a known specific inhibitor of PKB, we used two indirect approaches to address the contribution of the PKB/Forkhead pathway in mediating the overall effect of insulin on the activity of the Glc-6-Pase promoter. First, we mutated the IRU in the Glc-6-Pase promoter, which is known to bind Forkhead proteins. In addition, we overexpressed the TSS-Ala mutant of FKHR, which is not phosphorylated by PKB. Both procedures completely prevented the regulation of the gene promoter by PKB and reduced the effect of insulin on promoter activity by ϳ30%. These observations indicate that signaling via the PKB/Forkhead pathway plays a role in the overall regulation of the Glc-6-Pase promoter activity by insulin in H4IIE cells and that other signaling pathways and cis-and trans-acting factors also are likely to contribute to the effect of insulin on Glc-6-Pase gene expression.
In this context, it is important to note that signaling by pathways other than PKB may contribute to the regulation of Glc-6-Pase promoter activity via the IRU and Forkhead proteins. It has been suggested that FKHR can be phosphorylated at Thr-24 by some kinase(s) other than PKB␣ (25), and the effect of this other kinase could also be blocked by overexpression of TSS-Ala FKHR or the use of the Glc-6-Pase(Ϫ1227/ϩ57/ IRUmut) construct. However, the phosphorylation of Ser-256 in human FKHR (Ser-253 in mouse FKHR) by PKB is thought to be critical for the disruption of transactivation by insulin (17)(18)(19)25).
In this study, we found that PI 3-kinase plays a critical role in mediating the effect of insulin on the activity of the intact Glc-6-Pase promoter, consistent with a previous report (6). Interestingly, we found that the regulation of the Glc-6-Pase promoter by insulin also depends to a great extent on PI 3-kinase even after the IRU is mutated and the effects of PKB and FKHR on promoter activity have been disrupted. Previous studies have shown that insulin can suppress PEPCK promoter activity by a PI 3-kinase-dependent pathway independent of the activation of PKB␣, protein kinase C, and Rac (37,38). The results of the present study indicate that insulin may regulate Glc-6-Pase gene expression by multiple mechanisms, including signaling via PKB and Forkhead proteins to the Glc-6-Pase IRU and other PI 3-kinase-dependent pathways involving cis-acting elements outside the IRU.