Regulation of PGC-1 Promoter Activity by Protein Kinase B and the Forkhead Transcription Factor FKHR

HepG2 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS). Transfections were performed by the calcium phosphate method. pRSV-β-galactosidase plasmid was included in each transfection experiment to control for the efficiency of transfection. To ensure equal DNA amounts, empty plasmids were added in each transfection. After transfection, cells were incubated in DMEM supplemented with 10% FBS for 24 h. Thereafter, the medium was replaced with serum-free DMEM containing 0.1% BSA and incubation was continued for 18 h. When necessary, insulin or wortmannin were added. The luciferase activity was measured with AutoLumat (Berthold) and normalized for β-galactocidase activity in the same sample.

pcDNA3-FLAG-FKHR was generated by RT-PCR–based cloning of mouse FKHR cDNA into pcDNA3-FLAG vector as described previously (24). The putative PKB phosphorylation sites at Thr-24, Ser-253, and Ser-316 were mutated to Ala by PCR mutagenesis. A genomic clone encompassing human PGC-1 promoter extended from −992 to +90 bp relative to the transcription initiation site was subcloned into the BglII site of the promoterless pRL-basic luciferase reporter gene vector to generate PGC-1–promoter luc. The deletion mutants, various PGC-1 promoter (ΔIRS)-luc and (ΔCRE)-luc reporter plasmids, were generated by PCR mutagenesis. Constitutively active or dominant-negative PKB expression plasmids were obtained from Upstate Biotechnology.

A rabbit polyclonal antibody specific for mouse FKHR was raised against the bacterially expressed GST-FKHR (541–652 a.a.) fusion protein. The purified protein was injected into a rabbit and boosted five times. Serum from immunized rabbits was affinity purified by use of the antigen (Tanpaku Seisei Kougyou, Gunma, Japan).

Cells were grown to 95% confluence in DMEM supplemented with 10% FBS. Thereafter, medium was replaced with serum-free DMEM containing 0.1% BSA and incubation was continued for 18 h. Following the treatment with 100 nmol/l insulin for 30 min or not, cells were washed twice with PBS and cross-linked with 1% formaldehyde at 37°C for 10 min. Cells then were rinsed with ice-cold PBS twice and collected into PBS containing protease inhibitors and centrifuged for 4 min. Cells were resuspended in SDS lysis buffer (1% SDS, 10 mmol/l EDTA, 50 mmol/l Tris-HCl, pH 8.1 and protease inhibitors) and incubated for 10 min on ice. The lysate was sonicated four times for 10 s each (MISONIX, microson) followed by centrifugation for 10 min. Supernatants were collected and diluted in chromatin immunoprecipitation (ChIP) dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mmol/l EDTA, 167 mmol/l NaCl, 16.7 mmol/l Tris-HCl, pH 8.1, and protease inhibitors) followed by preclean with 30 μg sheared salmon sperm DNA, 5 μg normal rabbit IgG, and 60 μl protein G-Sepharose-50% slurry for 2 h at 4°C. Immunoprecipitation was performed overnight at 4°C with anti-FKHR-C3 antibody. After immunoprecipitation, 60 μl protein G-Sepharose-50% slurry and 20 μg salmon sperm DNA were added and the incubation was continued for another 1 h. Precipitates were washed sequentially for 5 min each in low salt (0.1% SDS, 1% Triton X-100, 2 mmol/l EDTA, 20 mmol/l Tris-HCl, pH 8.1, and 150 mmol/l NaCl), high salt (0.1% SDS, 1% Triton X-100, 2 mmol/l EDTA, 20 mmol/l Tris-HCl, pH 8.1, and 500 mmol/l NaCl), and LiCl wash buffer (0.25 mol/l LiCl, 1% NP-40, 1% deoxycholate, 1 mmol/l EDTA, and 10 mmol/l Tris-HCl, pH 8.1). Precipitates were then washed twice with TE buffer and eluted twice with 1% SDS and 0.1 mol/l NaHCO₃. Eluates were pooled and heated at 65°C for at least 4 h to reverse the formaldehyde cross-linking. Following the addition of Protease K and incubated for 1 h at 45°C, DNA fragments were purified by phenol/chloroform extraction and ethanol precipitation. Precipitated DNA fragments were analyzed by PCR amplification using primers detected against the human PGC-1 promoter region or the β-actin coding region as negative control. The PCR products were analyzed by electrophoresis using 0.7% agarose gel and visualized with ethidium bromide staining.

The following double stranded oligonucleotides were used, all biotinylated at the 3′-end of the sense strand, IRS1: AGTGTGTTGGTATTTTTCCCTCAGTTC, IRS2: ACATACAGGCTATTTTG-TTGATTAAAC, and IRS3: GCCACTTGCTTGTTTTGGAAGGAAAAT. The consensus insulin response sequences are italicized. 3×IRS consisted of three copies of the IRS derived from the IGFBP-1 promoter, while 3×IRS mut carried substitutions in IRS core residues of 3×IRS. An avidin-biotin conjugate DNA precipitation (ABCD) assay was performed by incubating cell extracts derived from HepG2 cells with double-stranded DNA immobilized on streptavidin agarose in binding buffer (20 mmol/l HEPES [pH 7.5], 150 mmol/l KCl, 0.5% Triton X-100, 0.5 mmol/l EDTA, 20 g/ml poly [dI-dC], and protease inhibitors). After 20 min of incubation at 4°C, the beads were washed four times with the same buffer and proteins were resolved by SDS-PAGE, followed by electrotransfer onto polyvinylidine fluoride membrane and probed with the anti-FKHR antibody. Chemiluminescent detection relied on horseradish peroxidase–conjugated secondary antibodies.

A recent study has shown that insulin reduces endogenous PGC-1 mRNA in FAO hepatoma cells (23). This observation raises the question of how insulin modulates PGC-1 expression. To reveal the molecular mechanism whereby insulin suppresses PGC-1 expression, a promoter sequence of the human PGC-1 gene was isolated and inserted into a promoterless luciferase reporter gene plasmid. Because we found that a fragment of the human PGC-1 promoter from −992 to +90, relative to the transcription start site, contains three consensus IRSs, we hypothesized that a trans-factor(s) modulated by insulin signaling binds to IRSs and regulates the PGC-1 promoter activity.

To investigate the effect of insulin on PGC-1 promoter, we conducted a series of transient transfection assays in HepG2 cells. First, the reporter plasmid containing the human PGC-1 promoter (PGC-1 promoter luc) was transfected into HepG2 cells, following treatment with or without 10% FBS or 100 nmol/l insulin. Serum stimulation substantially suppressed PGC-1 promoter activity, and particularly, insulin treatment was far more effective in reducing it by 50% than incubation with serum-free medium (Fig. 1A). This is in agreement with previous data in which insulin suppresses PGC-1 gene transcription in intact hepatoma cells (23), and thus we considered that the PGC-1 promoter luciferase (luc) reporter plasmid included the abilities to mimic the inhibitory effect of insulin on PGC-1 expression. Next, because expression of active PKB is sufficient to mimic insulin’s inhibitory effects on transcription through IRS-containing promoters (25–27), we tested whether the basal PGC-1 promoter activity is regulated by coexpression of a constitutively active (CA) or a dominant negative (DN) mutants of PKB. Figure 1B shows that transfected PKB (CA) mutant represses basal PGC-1 promoter activity to the same extent as insulin treatment (Fig. 1A), whereas no inhibition was detected with PKB (DN) mutant, suggesting that the inhibitory effect of insulin on the PGC-1 promoter are mediated by PKB. Furthermore, to confirm that the effect of insulin actually involves the PKB pathway, we investigated whether the insulin-inhibited PGC-1 promoter activity is recovered by coexpression of a PKB (DN). As expected, overexpressed PKB (DN) counteracted the inhibitory effect of insulin on the PGC-1 promoter (Fig. 1C). These results indicated that the PKB pathway acts downstream of insulin receptor signaling to modulate PGC-1 gene transcription in HepG2 cells.

It has been shown that PKB phosphorylates the Forkhead transcription factor FKHR, and consequently represses its ability of transactivaton, probably by disturbing the nuclear import of FKHR (28–30). To study a potential role of FKHR in regulating PGC-1 promoter activity, HepG2 cells were cotransfected with the PGC-1 promoter luc and an FKHR expression plasmid. Under a condition of serum-starvation, transfected FKHR in HepG2 cells stimulated PGC-1 promoter activity in a dose-dependent manner, indicating that the transactivation of PGC-1 gene is indeed mediated by FKHR (Fig. 2A). Furthermore, to examine the effects of insulin on FKHR-induced PGC-1 promoter activity, HepG2 cells were cotransfected with the PGC-1 promoter luc and FKHR expression plasmids and then incubated in serum-free medium with or without insulin treatment (Fig. 2B). Compared with the control insulin-free condition, insulin was able to suppress FKHR-stimulated PGC-1 promoter activity by ∼50% and to a similar extent for inhibiting basal activity (Fig. 2B). These results suggest that insulin inhibits induction of PGC-1 gene by specifically interfering with FKHR activity.

In response to insulin, PKB-mediated phosphorylation is shown to modulate the function of FKHR, which directly binds to the IRS in several gene promoters in vitro (7,10–13,24). We have identified the three consensus IRSs in the PGC-1 promoter: IRS1, IRS2, and IRS3 (Fig. 3A). To investigate whether the PGC-1 promoter is a direct target for FKHR in vivo, we performed ChIP assays in the presence or absence of insulin treatment. HepG2 cells were serum-starved for 18 h followed by treatment with or without insulin, and then soluble chromatin was prepared from cultured cells after cross-linking. For immunoprecipitation of FKHR-bound genomic DNA fragments, a specific antibody against FKHR was used and the immunoprecipitated DNAs were analyzed by PCR using the specific pairs of primers spanning the IRSs (Fig. 3A). Figure 3B shows that the PGC-1 promoter was efficiently recovered from immunoprecipitates of FKHR but not of control normal rabbit IgG. No PCR product was obtained from normal rabbit IgG or FKHR immunoprecipitates using control β-actin primers. In addition, FKHR recruitment to the PGC-1 promoter was significantly decreased after insulin treatment. These data show that endogenous FKHR can bind to the chromatinized PGC-1 promoter in the serum-starved condition in HepG2 cells.

Because the PGC-1 promoter is identified as an in vivo target for FKHR, we next determined whether FKHR could interact with three putative IRSs within the PGC-1 promoter in vitro. To this end, we performed an ABCD assay using biotin conjugate, double-stranded oligonucleotides that contains consensus insulin response sequences termed IRS1, IRS2, or IRS3 (Fig. 4A). Whole cell lysates extracted from intact HepG2 cells were mixed with each oligonucleotide immobilized on streptavidin agarose. As shown in Fig. 4B, endogenous FKHR strongly binds to IRS2 and IRS3, but weakly to IRS1 probe. In the control experiment, a biotinylated 3×IRS, which consists of the three copies of IRS derived from the IGFBP-1 promoter sequence, precipitated FKHR efficiently, whereas, a mutated IRS (3×IRS mut) did not. These findings indicate that FKHR binds to all putative IRSs in the PGC-1 promoter in vitro.

Furthermore, to confirm the FKHR response region in the PGC-1 promoter, we performed the reporter assay using a series of putative IRS- and CRE-deleted promoter constructs. Respective IRSs located from −976 to −970 (IRS1), −586 to −580 (IRS2), and −354 to −348 (IRS3) are depicted in Fig. 3A, and the transcriptional activities of these constructs were tested in the presence or absence of the FKHR expression plasmid. Compared with intact PGC-1 promoter luc, the deletion of each IRS1, IRS2, or IRS3 reduced the FKHR-mediated activation by 30%, whereas no substantial activation was observed using the ΔIRS1/2/3 promoter, which lacks all the IRSs (Fig. 4C). Remarkably, this construct reduced basal promoter activity by 40% (data not shown), indicating that the three IRSs are required for maintaining the basal promoter activity in HepG2 cells. In contrast, the deletion of CRE (from −132 to −125) in the PGC-1 promoter had no influence on the extent of fold activation by FKHR compared with the intact PGC-1 promoter activity. Taken together, these results demonstrate that the three consensus IRSs, characterized in the PGC-1 promoter (−992 to +90), are required for both the basal and FKHR-stimulated reporter activity.

To determine whether PKB-dependent phosphorylation of FKHR is critical for the PGC-1 gene suppression, we used a FKHR (3A) mutant, which is mutated at the three PKB phosphorylation sites (T24, S253, and S316) to alanines. This mutant is known to be resistant to the inhibitory effect by phosphorylation, resulting in the promotion of nuclear retention and of access to the target sequences (7–10,28,29). Cotransfection assays showed that FKHR (3A) possessed the activity to induce the PGC-1 gene transcription about twofold more than wild-type FKHR, whereas that activity was not suppressed by coexpression of PKB (CA) compared with the suppression of basal or FKHR wild-type–stimulated activity (Fig. 5A). Consistent with the cotransfection assays, insulin was not able to inhibit FKHR (3A)-stimulated PGC-1 promoter activity (data not shown). These results indicate that phosphorylation of putative PKB sites in FKHR is required for PKB-dependent suppression of FKHR-stimulated PGC-1 promoter activity.

Furthermore, to examine whether the three IRSs in the PGC-1 promoter are sufficient for the inhibitory effect of insulin, HepG2 cells were transfected with intact or ΔIRS1/2/3 PGC-1 promoter luc and then incubated with insulin and the phosphatidylinositol (PI)-3 kinase inhibitor, wortmannin, as indicated in Fig. 5B. In the intact promoter, treatment with wortmannin abolished the inhibitory effect of insulin, revealing the involvement of PI3 kinase in the insulin-mediated PGC-1 inhibition. In contrast, although the inhibitory effect of insulin on the PGC-1 promoter was not completely diminished by eliminating all three IRSs, no substantial inhibition was observed by insulin and wortmannin treatments in the ΔIRS1/2/3 PGC-1 promoter (Fig. 5B). Remarkably, this construct reduced basal promoter activity by 50%, indicating that the three IRSs are required for maintaining the basal promoter activity in HepG2 cells. The attenuated inhibition in ΔIRS1/2/3 PGC-1 promoter implies that FKHR partially confers the inhibitory effect of insulin via the IRSs in the PGC-1 promoter.

Based on our results, we propose a model for the transcriptional regulation of the PGC-1 promoter in which both FKHR and CREB stimulate the promoter activity via each response element (Fig. 6). Glucagon is known to potentiate the cAMP signaling pathway, resulting in phosphorylation of CREB by PKA (protein kinase A), and consequently the phosphorylated CREB appears to recruit a coactivator CBP to the PGC-1 promoter through the CRE, and induce PGC-1 expression (23). In this study, we elucidated that FKHR activates the PGC-1 promoter through the three consensus IRSs identified, while insulin inhibits both basal and FKHR-stimulated PGC-1 promoter activities by PKB-dependent phosphorylation of FKHR (Figs. 1–5). These findings are reminiscent of a recent study analyzing transgenic mice with the cardiac-specific expression of activated PKB (31). The significant downregulation of PGC-1 is observed by overexpression of activated PKB, indicating that PKB signaling suppresses the basal PGC-1 expression in vivo.

Alternatively, we found that the PGC-1 promoter was partially suppressed by insulin (Fig. 5B) or transfected PKB (CA) (data not shown), even after deletion of the three IRSs (Fig. 5B). A possible explanation for this inhibition is that a ΔIRS1/2/3 PGC-1 promoter construct still contains the functional CRE. A previous study demonstrated that insulin-dependent activation of the Ras pathway represses CREB activity by inducing recruitment of the S6 kinase pp90_RSK to CBP (32). Therefore, binding of pp90_RSK to CBP might counteract basal cAMP signal transduction, and thereby suppresses the ΔIRS1/2/3 PGC-1 promoter activity. In addition to the PGC-1 promoter, similar observations are reported in the G-6-Pase and PEPCK genes, where disruption of IRS does not completely impair the inhibitory effect of insulin on their promoters (11,17). These findings suggest that other cis-activating elements or trans-acting factors also play roles in mediating the effect of insulin signaling.

The genes encoding limiting enzymes for gluconeogenesis are powerfully controlled at the transcriptional level by key hormones, particularly insulin, glucagon, and glucocorticoid. Although recent data have shown a critical role for PKB in insulin-dependent inhibition of hepatic gluconeogenesis, it has not been elucidated whether that component is involved in the regulation of PGC-1 expression (33). In this paper, we show that insulin can suppress PGC-1 promoter activity by mediating the function of FKHR via PKB signaling pathway. This finding provides an idea that FKHR can regulate PEPCK gene expression, not only through direct binding to that promoter, but also through indirect activation, in which the induced PGC-1 protein works as a coactivator with HNF-4 and glucocorticoid receptors in PEPCK gene transcription (22). This two-dimension model may explain a paradox that insulin inhibits PEPCK gene transcription independently of the IRS on that promoter (17).

The liver-specific insulin-receptor knockout (LIRKO) mice are unique models exhibiting dramatic insulin resistance in the liver, severe glucose intolerance, and the failure of insulin to suppress hepatic glucose production (34). This line of the mice also displays the transcriptional induction of G-6-Pase and PEPCK genes in the liver. Interestingly, recent data have shown that the expression of PGC-1 distinctively increases in the livers of the LIRKO mice (22). Although an indirect effect of insulin cannot be excluded, this in vivo observation implies that insulin is the dominant regulator of PGC-1 expression, possibly through FKHR-regulated transcription. Here, we propose a model in which FKHR is essential for maintaining basal PGC-1 expression, while CREB is necessary for stimulating further induction of PGC-1 by glucagon in the starved condition.

Our results argue for a three-stage model in the activation of PGC-1 gene by FKHR and CREB from feeding to fasting. First, in the feeding stage, glucagon secretion drops to the basal level while insulin increases to the maximal level, resulting in phosphorylation of sole FKHR by PKB. Because phosphorylated FKHR is excluded from the nucleus and unphosphorylated CREB cannot recruit CBP as a coactivator, neither of them can stimulate PGC-1 transcription. Hence, hepatic gluconeogenesis through the induced expression of PGC-1 would be almost eliminated in the feeding stage. Second, during early fasting stage (interphase), glucagon secretion gradually goes up while insulin level moderately drops, consequently increasing in the proportions of phosphorylated CREB and unphosphorylated FKHR, and then both activated transcription factors can induce PGC-1 expression. Third, under prolonged fasting stage (so called starved), glucagon secretion further increases to the maximal level while insulin decreases to the basal level, leading to CREB phosphorylation by PKA and preventing FKHR from PKB-dependent phosphorylation. Consequently, both phosphorylated CREB and unphosphorylated FKHR may synergistically induce PGC-1 expression, and hence the hepatic gluconeogenesis occurs rapidly to maintain basal blood glucose level for protecting the body against hypoglycemia.

In addition to the previous reports in which FKHR was shown to play an important role in mediating the expression of gluconeogenic genes including G-6-Pase and PEPCK, we have shown here that FKHR also regulates PGC-1 expression in the transcriptional level. This result suggests that the molecular mechanism of FKHR transcription is principally linked to regulating gluconeogenesis. Because excessive hepatic glucose production by the gluconeogenesis is known to be a major contributor to both the fasting hyperglycemia and the exaggerated postprandial hyperglycemia in diabetes (35), controlling FKHR functions in the liver could be a remarkable target for antidiabetic therapy. To conduct that strategy, however, further studies would be required to elucidate the tissue-specific physiological significance of FKHR functions.

This work was supported by grants from “Research for the Future” Program (The Japan Society for the Promotion of Science: JSPS—RFTF 97L00804), Grant-in-Aid for Scientific Research (A) and Grant-in Aid for Scientific Research on Priority Areas from the Ministry of Education, Science, Sports, and Culture of Japan, The Asahi Glass Foundation and the Research Grant for Cardiovascular Disease (11C-1) from the Ministry of Health, Labor, and Welfare.

We thank all members of Fukamizu laboratory for their helpful discussion.