CRTC2 Is a Coactivator of GR and Couples GR and CREB in the Regulation of Hepatic Gluconeogenesis

Abstract

Glucocorticoid hormones play essential roles in the regulation of gluconeogenesis in the liver, an adaptive response that is required for the maintenance of circulating glucose levels during fasting. Glucocorticoids do this by cooperating with glucagon, which is secreted from pancreatic islets to activate the cAMP-signaling pathway in hepatocytes. The cAMP-response element-binding protein (CREB)-regulated transcription coactivator 2 (CRTC2) is a coactivator known to be specific to CREB and plays a central role in the glucagon-mediated activation of gluconeogenesis in the early phase of fasting. We show here that CRTC2 also functions as a coactivator for the glucocorticoid receptor (GR). CRTC2 strongly enhances GR-induced transcriptional activity of glucocorticoid-responsive genes. CRTC2 physically interacts with the ligand-binding domain of the GR through a region spanning amino acids 561–693. Further, CRTC2 is required for the glucocorticoid-associated cooperative mRNA expression of the glucose-6-phosphatase, a rate-limiting enzyme for hepatic gluconeogenesis, by facilitating the attraction of GR and itself to its promoter region already occupied by CREB. CRTC2 is required for the maintenance of blood glucose levels during fasting in mice by enhancing the GR transcriptional activity on both the G6p and phosphoenolpyruvate carboxykinase (Pepck) genes. Finally, CRTC2 modulates the transcriptional activity of the progesterone receptor, indicating that it may influence the transcriptional activity of other steroid/nuclear receptors. Taken together, these results reveal that CRTC2 plays an essential role in the regulation of hepatic gluconeogenesis through coordinated regulation of the glucocorticoid/GR- and glucagon/CREB-signaling pathways on the key genes G6P and PEPCK.

Glucocorticoids, steroid hormones secreted from the zona fasciculata of the adrenal cortex, have diverse and strong regulatory activities on virtually all aspects of human biology, functioning as end effectors of the hypothalamic-pituitary-adrenal axis, the major regulatory system for organizing the adaptive response to environmental changes called stressors (1). These steroid hormones are essential for the maintenance of glucose supply to the brain and red blood cells during fasting by stimulating the gluconeogenesis in the liver, which generates glucose from glucogenic amino acids and free fatty acids liberated from muscles and adipose tissues (2). These actions of glucocorticoids are mediated by a single intracellular receptor molecule, the glucocorticoid receptor (GR), which is a member of the steroid/thyroid/nuclear receptor superfamily and functions as a ligand-dependent transcription factor (1, 3). Upon binding to glucocorticoid, GR translocates from the cytoplasm into the nucleus, and modulates the transcriptional activity of glucocorticoid-responsive genes by binding to the specific DNA sequences called glucocorticoid-response elements (GREs) located in the promoter region of these genes. GR consists of 3 subdomains, the N-terminal domain (NTD), middle DNA-binding domain (DBD), and C-terminal ligand-binding domain (LBD) (1). Binding of glucocorticoid to the GR LBD causes a conformational change in this domain and creates on LBD a ligand-dependent transactivation domain (activation function-2 [AF-2]), which activates GR transcriptional activity by cooperating with another transactivation domain AF-1 located in NTD (1). GR AF-2 physically interacts with cofactor molecules including p160-type histone acetyltransferase coactivators (or nuclear receptor coactivators [NCoAs]) through the latter's LxxLL motifs located in their nuclear receptor-binding domain (1).

In addition to glucocorticoids, glucagon strongly stimulates gluconeogenesis in the liver (4). This peptide hormone consists of 29 amino acids and is massively secreted upon fasting from the pancreatic islet α-cells into the hepatic portal vein. Secreted glucagon activates its cell surface G-protein-coupled receptor on hepatocytes and downstream cAMP-response element (CRE)-binding protein (CREB) to mediate its biologic effects into the nucleus of these cells (4–7). During fasting, glucocorticoids and glucagon cooperate with one another to stimulate gluconeogenesis in the liver by increasing the expression of the glucose-6-phosphatase (G6P) and the phosphoenolpyruvate carboxykinase (PEPCK), which are the major rate-limiting enzymes for this glucose-generating pathway (8, 9). These enzymes catalyze the conversion of glucose-6-phosphate and oxaloacetate to glucose and phosphoenolpyruvate, respectively. These 2 genes possess GREs and CRE in their promoter regions and binding of GR and CREB to these elements not only increases the transcriptional activity of these genes (10–12) but also causes a strong synergism for their induction (13, 14). Massive elevation of serum glucose concentrations is observed upon simultaneous treatment with glucocorticoid and glucagon in animals, whereas fasting significantly elevates the serum concentration of these hormones (7, 13–15).

The CREB-regulated transcription coactivator 2 (CRTC2) or the transducer of regulated CREB protein 2 functions as a coactivator specific to CREB (9). The human CRTC2 consists of 693 amino acids, physically interacts with CREB via its N-terminal portion, and coactivates the transcriptional activity of CREB through its C-terminally located transactivation domain (9, 16). In addition, the human CRTC2 has 4 serine residues (serines at amino acid positions 136, 171, 275, and 307) located in the central regulatory region. Several serine/threonine kinases, such as the AMP-activated protein kinase, salt-inducible kinase 2, and the microtubule affinity-regulating kinase 2, phosphorylate serine 171, 275, and/or 307, which creates binding sites for 14–3-3 proteins and subsequently causes segregation of CRTC2 into the cytoplasm; thus, these kinases are negative regulators for the transcriptional activity of this coactivator (9). Recently, the mammalian target of rapamycin is shown to phosphorylate serine 136 of the human CRTC2 as well (17).

In this article, we examined the hypothesis that CRTC2 acts as a coactivator of GR in addition to CREB. We found that CRTC2 enhances the transcriptional activity of glucocorticoid-responsive genes by physically interacting with GR LBD. Importantly, CRTC2 enhances the transcriptional activity of GR on the G6P and PEPCK genes and is necessary for the positive cooperation between GR and CREB on these genes. In addition, CRTC2 modulates the transcriptional activity of the progesterone receptor (PR), suggesting that its effect is not limited to GR but also functional to other steroid/nuclear hormone receptors.

Materials and Methods

Plasmids

pCDNA3.1 His/B-CRTC2 and pCDNA3.1 His/B-CRTC2(1–561), which, respectively, express full-length (FL) and the amino acids 1–561 of the human CRTC2 were constructed by subcloning the corresponding CRTC2 cDNA fragments from the pCMV-SPORT6-hCRTC2 (Thermo Fisher Scientific, Inc) into pCDNA3.1His/B (Invitrogen). pGFP-CRTC2 wild-type, S171A, S275A, and S307A, which, respectively, express the FL human CRTC2 fused with green fluorescent protein (GFP) and its mutants harboring the indicated amino acid replacements, were gifts from Dr H. Takemori (National Institute of Biomedical Innovation, Osaka, Japan) (18). pVP16-CRTC2 FL, (1–55), (56–143), (143–320), and (316–693), which, respectively, express FL and corresponding portions of the human CRTC2 fused with the activation domain (AD) of the herpes simplex virus VP16 protein were constructed by inserting the corresponding cDNAs of the human CRTC2 into pVP16 (Invitrogen). pGEX-4T3-hCRTC2 FL, (1–55), (56–144), (143–320), (316–693), (316–450), (445–561), (561–693), (316–561), (445–693), and (316–693), which express corresponding portions of the human CRTC2 fused with the glutathione S-transferase (GST) protein were constructed by inserting their coding cDNAs into pGEX-4T3 (GE Healthcare Bio-Sciences Corp). The human GRα-expressing pRShGRα and its control pRSerbA⁻¹, and pGR107 that expresses the human GRα under the control of the SP6 promoter were gifts from R. M. Evans (Salk Institute, La Jolla, CA). pM-GR FL and LBD, pGEX-4T3-GRα FL, NTD, DBD, and LBD, which express the corresponding portions of the human GRalphα fused with galactose-responsive transcription factor GAL4 (GAL4) DBD or GST, were previously reported (19). pM-GR ΔAF2, which expresses the GAL4-DBD-fused GRalphα mutant that has an amino acid replacement from glutamic acid to alanine at position 755 (E755A) of the FL human GRα, was created by introducing the corresponding nucleotide change into pM-GR FL. The E755A mutation is located inside the helix-12 of GRalphα LBD, and destroy the AF-2 surface of GRalphα and completely abolish physical association of this receptor to p160-type histone acetyltransferase NCoAs (20). pSVPRA, which expresses the human PR-A, was a gift from S. S. Simons Jr (National Institutes of Health, Bethesda, MD) (21). pMMTV-Luc and p17mer-TK-luc, which express the luciferase, respectively, under the control of the glucocorticoid-responsive mouse mammary tumor virus (MMTV) promoter and 4 GAL4 response elements, were gifts from G. L. Hager (National Institutes of Health, Bethesda, MD) and M. J. Tsai (Baylor College of Medicine, Houston, TX), respectively (22). pRc/RSV-CREB341 and RSV-protein kinase A (PKA), which, respectively, express CREB and a constitutively active form of the protein kinase A (PKA) were gifts from Dr R. H. Goodman (Vollum Institute, Portland, OR). pGL4.29[Luc2P/CRE/Hygro] and pGL4.73[hRluc/SV40], which, respectively, express the firefly and Renilla luciferase under the control of the synthetic CRE or the simian virus 40 promoter, were purchased from Promega Corp.

Cell culture, transfection, and reporter assay

Human colon cancer HCT116 cells and cervical carcinoma HeLa cells were previously reported (23). Human breast ductal carcinoma T47D and rat hepatoma H4IIE cells were purchased from the American Type Cell Collection. Cells were maintained in McCoy's 5A or DMEM supplemented with 10% fetal bovine serum, 100 U/mL of penicillin, and 100 μg/mL of streptomycin. HCT116 cells do not express endogenous GR, whereas HeLa and H4IIE cells have the endogenous and functional GR (23, 24). For reporter assays, these cells were transfected with the indicated amounts (0–0.2 μg/well) of CRTC2- and/or GR-related plasmids together with 0.3 μg/well of pMMTV-Luc or p17mer-TK-luc and 0.01 μg/well of pGL4.73[hRluc/SV40] using Lipofectamine 2000 (Invitrogen) or the polyethylenimine (Polysciences, Inc) in 24-well plates. For HCT116 cells, 0.05 μg/well of GR-expressing plasmid were included. Control plasmids were used to account for the same amounts of plasmid DNA. In some experiments, 5 nmol of control or CRTC2-directing small interfering RNA (siRNA) (Santa Cruz Biotechnology, Inc) was also transfected. Six hours after the transfection, media were changed and cells were treated with 10⁻⁶M or indicated concentrations of dexamethasone or progesterone, 10⁻⁵M RU 486 and/or 10⁻⁵M forskolin for an additional 24 hours. Lysates were then analyzed for firefly and Renilla luciferase activities by using the Dual-Luciferase Assay kit and the GloMax Luminometer (Promega).

SYBR Green real-time PCR analysis

Total RNA was purified from cellular or animal samples by using the RNeasy Mini kit (QIAGEN Sciences, Inc). cDNA was synthesized using the TaqMan Reverse Transcription Reagents and oligo-dT as a primer (Applied Biosystems). PCR was then performed in the 7500 Real-time PCR System (Applied Biosystems), as previously described (25, 26). Amplification was performed in a 2-step cycle: 95°C for 10 minutes, then 45 cycles of the reaction consisting of denaturing at 95°C for 15 seconds and annealing/extension at 60°C for 1 minute. Primer pairs for quantifying mRNA levels of CRTC2, G6P, glucocorticoid-induced leucine zipper (GILZ), PEPCK, and ribosomal protein large P0 (RPLP0) used in the SYBR Green real-time PCR were designed so that the sequence spanning between a forward and a reverse primer contains at least one intron. Their sequences are listed in Supplemental Table 1. Threshold cycle (C_t) values of the examined genes were normalized with those of the RPLP0, and fold changes were obtained by using the comparative C_t method (27).

Coimmunoprecipitation and Western blotting

HeLa cells were treated with 10⁻⁶M dexamethasone for 3 hours and were lysed in the buffer containing 50mM Tris-HCl (pH 7.4), 150mM NaCl, 0.1% sodium dodecyl sulfate, 1% Nonidet P-40, 0.5% sodium deoxycholate, and 1 Tablet/50-mL Complete Tablet (Roche). Immunoprecipitation was carried out with the anti-GR or anti-CRTC2 antibody, or the control IgG, and a protein complex was precipitated with protein A/G agarose (Santa Cruz Biotechnology, Inc), as previously reported (24). Samples were run on 4%–12% Bis-Tris gels together with 3% of input used for the coimmunoprecipitation reaction. Western blottings were then performed for GR or CRTC2 using their specific antibodies, and their protein bands were visualized using the horseradish-conjugated secondary antibody and the Amersham ECL Detection Reagents (GE Healthcare Bio-Science Corp). Western blottings were also performed for the mouse liver samples using anti-CRTC2 or anti-β-actin antibody.

Microscopic imaging of GFP-CRTC2

HCT116 cells were plated on coated 25-mm glass bottom dishes (MatTek Corp) in phenol red-free McCoy's 5A, and were transfected with the indicated wild-type- or mutant CRTC2-expressing plasmids. Expressed GFP-CRTC2s were detected under the inverted fluorescence microscope (Leica DM IRB) as described previously (28). Twelve-bit black-and-white images were captured using a digital charge-coupled device camera (Hamamatsu Photonics K.K.). Image analysis and presentation were performed using the Openlab software (Improvision).

GST pull-down assay

³⁵S-labeled human GR and CRTC2 were generated by the in vitro transcription/translation reactions (Promega) using pR107 and pCDNA3.1His/B-CRTC2 as templates, respectively, and were tested for interaction with GST-GRs or GST-CRTC2s immobilized on the glutathione-sepharose beads (GE Healthcare Bio-Science Corp) in the presence or absence of 10⁻⁵M dexamethasone in the buffer containing 50mM Tris-HCl (pH 8.0), 50mM NaCl, 1mM EDTA, 0.1% Nonidet P-40, 10% glycerol, and 0.1-mg/mL BSA at 4°C for 1.5 hour, as previously described (21). For the reactions employing GST-GRs, no dexamethasone was added, because these bacterially produced GR fusion proteins were known to be unresponsive to glucocorticoids. After vigorous washing with the buffer, proteins were eluted and separated on 8% or 4%–12% Bis-Tris gels together with 3% of inputs used for the pull-down reactions. Gels were fixed, treated with Enlighting (NEN Life Science Products, Inc), dried and exposed to film. The same amounts of GST-related proteins as used in GST pull-down assays were run on 4%–12% Bis-Tris gels together with the Mark 12 Unstained Standard (Invitrogen). Gels were subsequently stained with the Simple Blue Safe Stain (Invitrogen), dried, and their images were scanned.

Chromatin immunoprecipitation (ChIP) assay

ChIP assays were performed in H4IIE cells, as previously described (29). Cells were transfected with control or CRTC2 siRNA, or pCDNA3.1His/B-CRTC2(1–561) using the Nucleofector system (Lonza) with over 50% transfection efficiency and were treated with 10⁻⁶M dexamethasone and/or 10⁻⁵M forskolin for 8 hours. They were then fixed, DNA and bound proteins were cross-linked, and ChIP assays were performed by coprecipitating the DNA/protein complexes with anti-GR, anti-CRTC2, or anti-CREB antibody or rabbit control IgG (Santa Cruz Biotechnology, Inc). The promoter regions of the G6p and the period 1 (Per1) genes that contain GREs and/or CRE (10, 30) were amplified from the prepared DNA samples using the primer pairs shown in Supplemental Table 2 in the SYBR Green real-time PCR. The rat G6p gene promoter contains 3 tandem GREs (GRE-A, GRE-B, and GRE-C: in a region spanning from nucleotides −209 to −154 from the transcription start site) and 2 CREs (CRE-1 and CRE-2, in the area spanning from nucleotides −174 to −143), and they are all functional, as reported by Vander Kooi et al (10). The GREs primer pair used in this study covered GRE-A and GRE-B (−232 to −169), whereas that for CREs spanned over CRE-1 and CRE-2 (−187 to −121) (10). Binding of GR to the promoter region of the rat Per1 gene was accessed by using the primer pair, which covers the promoter area containing one tandem GREs located approximately 3.4 kbps upstream of the Per1 gene (31–33). The DNA region amplified by this primer pair does not include CRE. Obtained C_t values of ChIP samples were normalized for those of corresponding inputs and those with control IgG, and their relative precipitation were demonstrated as fold precipitation above the baseline.

Mouse study

All animal procedures used in this article were approved by the Eunice Kennedy Shriver National Institute of Child Health and Human Development Animal Care and Use Committee and are accordance with its study guideline. Female mice aged 12–16 weeks were maintained under standard conditions at least for 2 weeks before experiments. They were subjected to a 12-hour light, 12-hour dark cycle (lights-on time was 6 am) and provided with food and water ad libitum before the experiments. Mice (n = 5) were injected ip either with control or CRTC2 shRNA-expressing lentiviral particles (1 × 10⁶ infectious units of virus per animal) (Santa Cruz Biotechnology, Inc). Twenty-four hours after the injection of lentiviral particles, they were injected ip with dexamethasone (1.5 mg/kg) and/or RU 486 (10 mg/kg). These compounds dissolved in ethanol were diluted with PBS for injection into mice. Just after the injection of these steroids, mice were subjected to 3-hour fasting or normal chaw. Euthanasia was performed around 8 pm with CO₂ asphyxiation. Blood samples were obtained from the heart, whereas the livers were resected. Obtained sera and livers were then stored for measuring glucose levels and for purification of total RNA, respectively. Some of the liver samples were used for Western blottings to evaluate the CRTC2 protein expression. Serum glucose levels were determined using the Glucose Assay kit (Abcam K.K.).

Statistical analysis

All experiments except the animal studies were performed as triplicate and were repeated at least 3 times, and their representative results were shown in figures. Statistical analysis was performed by using the Student's t test with 2-tailed value or one-way ANOVA with Bonferroni correction in the GraphPad Prism 6 (GraphPad Software). Statistical significance was set at P < .05.

Results

CRTC2 modulates the transcriptional activity of GR and PR

To examine our hypothesis that CRTC2 functions as a coactivator of GR, we employed a transient transfection-based reporter assay in HCT116 cells. Overexpression of CRTC2 dose-dependently enhanced GR transcriptional activity on the MMTV promoter in a dexamethasone-dependent fashion, whereas its overexpression strongly shifted dexamethasone-titration curve on this promoter upward, suggesting that CRTC2 functions as a typical coactivator for this receptor (Figure 1A). Interestingly, CRTC2 overexpression repressed PR-induced transcriptional activity on this promoter in a dose-dependent fashion, and shifted the progesterone-titration curve downward (Figure 1B). These results indicate that the effect of CRTC2 is not limited to GR but also functional to other steroid or nuclear hormone receptors, although the direction of its effect was opposite for PR in these cells under the conditions we employed.

Figure 1.

CRTC2 modulates GR- or PR-induced transcriptional activity. A, CRTC2 enhances GR transcriptional activity in a dose-dependent fashion and shifts the dexamethasone (Dex)-titration curve upward in HCT116 cells. HCT116 cells were transfected with the GR-expressing plasmid together with pMMTV-Luc and the control pGL4.73[hRluc/SV40]. Indicated amounts (left panel) or 0.2 μg/well (right panel) of CRTC2-expressing plasmid were also included. 10⁻⁶M (left panel) or indicated concentrations (right panel) of Dex were subsequently added to media. Bars or circles represent mean ± SE values of firefly luciferase activity normalized for Renilla luciferase activity. **, P < .01, compared with the baseline in the presence of Dex (left panel) or in the absence of the CRTC2 transfection with the same concentration of Dex (right panel). B, CRTC2 represses PR transcriptional activity in a dose-dependent fashion and shifts the progesterone (P4)-titration curve downward in HCT116 cells. HCT116 cells were transfected with the PR-expressing plasmid together with pMMTV-Luc and the control pGL4.73[hRluc/SV40]. Indicated amounts (left panel) or 0.2 μg/well (right panel) of CRTC2-expressing plasmid were also included. 10⁻⁶M (left panel) or indicated concentrations (right panel) of P4 were subsequently added to media. Bars or circles represent mean ± SE values of firefly luciferase activity normalized for Renilla luciferase activity. *, P < .05; **, P < .01, compared with the baseline in the presence of P4 (left panel) or in the absence of the CRTC2 transfection with the same concentration of P4 (right panel).

We next examined the effect of CRTC2 on the endogenous glucocorticoid- or progesterone-responsive genes by knocking down the expression of endogenous CRTC2 in HeLa or T47D cells. Transfection of CRTC2 siRNA significantly reduced mRNA expression of a well-known glucocorticoid-responsive gene the GILZ in a dexamethasone-dependent fashion in HeLa cells, whereas it strongly enhanced progesterone-induced mRNA expression of the progesterone-responsive Krueppel-like factor 5 (KLF5) gene in T47D cells (Figure 2, A and B) (25, 34). This siRNA efficiently reduced CRTC2 mRNA abundance in these cells. We interpret these results to indicate that endogenous CRTC2 acts to regulate GR- or PR-induced transcriptional activity of their endogenous, steroid-responsive genes.

Figure 2.

CRTC2 modulates mRNA expression of the endogenous glucocorticoid- or progesterone (P4)-responsive gene. A, CRTC2 knockdown attenuates glucocorticoid-responsive GILZ mRNA expression in HeLa cells. HeLa cells were transfected with CRTC2 or control siRNA, and were treated with 10⁻⁶M dexamethasone (Dex). mRNA expression of GILZ, CRTC2, and RPLP0 was determined with the SYBR Green real-time PCR using their specific primers. Bars represent mean ± SE values of fold GILZ (left panel) or CRTC2 (right panel) mRNA expression normalized for RPLP0 mRNA expression. Broken line indicates the level of fold expression as “1.” **, P < .01; n.s., not significant, compared with the baseline under the same treatment. B, CRTC2 knockdown enhances P4-responsive KLF5 mRNA expression in T47D cells. T47D cells were transfected with CRTC2 or control siRNA, and were treated with 10⁻⁶M P4. mRNA expression of KLF5, CRTC2, and RPLP0 was determined with the SYBR Green real-time PCR using their specific primers. Bars represent mean ± SE values of fold KLF5 (left panel) or CRTC2 (right panel) mRNA expression normalized for RPLP0 mRNA expression. Broken line indicates the level of fold expression as “1.” **, P < .01; n.s., not significant, compared with the baseline under the same treatment.

To further evaluate specificity of the CRTC2's effects on GR and PR, we tested in HCT116 cells the CRTC2 mutants defective in its serine residues (S171A, S275A, and S307A), which are known to be phosphorylated by several serine/threonine kinases (9). CRTC2 S275A and S307A developed an augmented enhancing effect on CREB-induced transcriptional activity on the synthetic CRE-driven promoter, whereas CRTC2 S171A caused the enhancement similar to its wild type, suggesting that these 3 CRTC2 mutants are functional, and the kinase(s) phosphorylating serine 275 and serine 307 appear(s) to be active in HCT116 cells, whereas that (those) for serine 171 is (are) not (Figure 3A, right panel). CRTC2 S275A and S307A demonstrated a stronger enhancement on GR-induced transcriptional activity of the MMTV promoter than the wild-type CRTC2, whereas they showed a stronger suppressive effect on PR-induced transcriptional activity of the same promoter (Figure 3A, left and middle panels). As expected, the CRTC2 S171A mutant showed the enhancement/repression of GR/PR transcriptional activity similar to the wild-type CRTC2. Wild-type CRTC2 and all these mutants were equally expressed (Figure 3B). We also examined subcellular localization of GFP-CRTC2 proteins in HCT116 cells. In accordance with our results observed in the reporter assays and a previous report (18), most of the GFP-CRTC2 S275A and S307A were found in the nucleus, whereas wild-type GFP-CRTC2 and the S171A mutant were localized moderately in the cytoplasm in addition to the nucleus (Figure 3C). These results suggest that the CRTC2 mutants tested are functional on GR- and PR-induced transcriptional activity, and these receptors appear to be under the regulation of several kinases through phosphorylation of the CRTC2.

Figure 3.

The CRTC2 mutants defective in phosphorylation sites modulate GR- or PR-induced transcriptional activity. A, The CRTC2 mutants defective in phosphorylation sites differentially modulate the transcriptional activity of GR, PR, or CREB in HCT116 cells. HCT116 cells were transfected with GR-, PR-, or CREB-expressing plasmid together with pMMTV-Luc (for GR and PR) or pGL4.29[Luc2P/CRE/Hygro] (for CREB) and the control pGL4.73[hRluc/SV40]. 0.2 μg/well of wild-type GFP-fused CRTC2- or indicated mutant-expressing plasmid were also included. 10⁻⁶M dexamethasone (Dex) (left panel) or progesterone (P4) (middle panel) was subsequently added to media. For CREB, 0.05 μg/well of the plasmid expressing a constitutively active form of the PKA mutant was transfected (right panel). Bars represent mean ± SE values of firefly luciferase activity normalized for Renilla luciferase activity. *, P < .05; **, P < .01; n.s., not significant, compared with the conditions indicated or to the condition with the wild-type CRTC2 transfection under the same treatment. B, Wild-type CRTC2 and its mutants employed in A are equally expressed from their carrying plasmids in HCT116 cells in Western blottings. HCT116 cells grown on 6-well plates were transfected with wild-type GFP-fused CRTC2- or indicated mutant-expressing plasmids. Cells were lysed and whole homogenate was run on 4%–12% Bis-Tris gels. Western blottings were performed by using anti-CRTC2 or anti-β-actin antibody. C, GFP-fused wild-type or mutant CRTC2s demonstrate differential intracellular localization. HCT116 cells were plated on coated 25-mm glass bottom dishes in phenol red-free McCoy's 5A and were transfected with the indicated wild-type- or mutant CRTC2-expressing plasmids. Expressed GFP-CRTC2s were detected under an inverted fluorescence microscope.

CRTC2 physically interacts with GR through its C-terminal portion

We next examined the possibility that CRTC2 interacts with these steroid receptors. For this purpose, we employed GR as a model receptor. In the coimmunoprecipitation assay employing anti-GR or anti-CRTC2 antibody and precipitating their endogenous molecules in HeLa cells, GR and CRTC2 interacted with each other in a dexamethasone-dependent fashion (Figure 4, A and B). We therefore examined their physical interaction in the GST pull-down assay employing the GR and CRTC2 mutants expressed as fusions with GST (Figure 5). In the assay using in vitro translated and labeled CRTC2 and the GST-fused GRs, CRTC2 bound FL GR: it was physically associated strongly with LBD and weakly with DBD, but it did not interact with NTD at all (Figure 5A). Dexamethasone was not added in this assay, because bacterially produced GST-GR fusion proteins are not responsive to glucocorticoids. In the second assay using in vitro translated and labeled GR, this receptor bound the GST fusion of the FL or CRTC2(316–693) in a ligand-independent fashion (Figure 5B, left panel). It is generally known that the in vitro translated and labeled GR using the reticulocyte lysates is fully responsive to glucocorticoids (35, 36). In the third assay employing the labeled GR and the GST-fused CRTC2 fragments covering the amino acids 316–693, GR strongly interacted with the CRTC2 fragments, which harbor amino acids 516–693 (Figure 5B, right panel). Taken together, these results suggest that GR and CRTC2 physically interact with each other in a glucocorticoid-independent fashion through LBD of the former and the portion spanning from amino acid 561 to 693 of the latter in the GST pull-down assays. We further examined interaction of GR and CRTC2 in the mammalian 2-hybrid assay employing GAL4-DBD-fused GR LBD and VP16 AD-fused CRTC2 fragments (Figure 5C). In this assay, GR LBD interacted with FL CRTC2 as well as its fragment CRTC2(316–693) in a dexamethasone-dependent fashion (Figure 5C, left panel). In addition, GR LBD weakly interacted with CRTC2(56–143). Because GR and CRTC2 interacted physically with each other in a dexamethasone-independent fashion in the GST pull-down assays, the GR LBD surface outside of the AF-2 appears to be responsible for their interaction. To examine this possibility, we employed the mutant GAL4-DBD-fused GR defective in AF2 (ΔAF2), which has an amino acid replacement from glutamic acid to alanine at position 755 (E755A) of the FL human GR. The E755A mutation is located inside the helix-12 of GR LBD. It is known to destroy the AF-2 surface of the GR and abolishes physical association of GR to p160-type NCoAs, whereas it preserves the ligand-binding activity of the receptor (20). GAL4-DBD-GR ΔAF2 demonstrated significantly reduced basal transcriptional activity in the presence of dexamethasone compared with its wild type, whereas its binding activity to CRTC2 was highly conserved based on the high levels of its fold association to CRTC2 (Figure 5C, right panel). Further, we could not identify the LxxLL motif or its similar sequences in the C-terminal portion of CRTC2. These results indicate that CRTC2 binds the GR LBD outside of its ligand-dependent AF-2 surface.

Figure 4.

CRTC2 and GR are coprecipitated with each other in a dexamethasone (Dex)-dependent fashion. HeLa cells were treated with 10⁻⁶M Dex, and coimmunoprecipitation was carried out using anti-GR (A) or anti-CRTC2 (B) antibody or control IgG. Blotted GR and CRTC2 were visualized with their specific antibodies in Western blottings. Two percent of the samples used in the coimmunoprecipitation reactions was used as “input.” Ab, antibody.

Figure 5.

GR and CRTC2 interact with each other through their LBD and C-terminal portion. A, GR LBD physically interacts with CRTC2 in GST pull-down assays. In vitro translated and labeled CRTC2 was incubated with bacterially produced FL GR (FL) or its subdomains (NTD, DBD, or LBD) fused with GST. The CRTC2 precipitated with GST beads was run on 4%–12% Bis-Tris gels, and radioactivity was detected with x-ray films (top gel). Three percent of the input CRTC2 was loaded as a positive control. The same amounts of GST-fused GR-related proteins as used in the top panel experiment were run on 4%–12% Bis-Tris gels and were visualized with the Simple Blue Safe Stain (bottom gel). 1, GST; 2, GST-GR FL; 3, GST-GR NTD; 4, GST-GR DBD; 5, GST-GR LBD. B, CRTC2(561–693) physically interacts with GR in GST pull-down assays. In vitro translated and labeled GR was incubated with bacterially produced FL CRTC2 (FL) or its indicated deletion mutants fused with GST in the presence or absence of 10⁻⁵M dexamethasone (Dex). The GR precipitated with GST beads was run on 4%–12% Bis-Tris gels, and radioactivity was detected with x-ray films (top gels). Three percent of the input GR was loaded as a positive control. The same amounts of GST-fused CRTC2-related proteins as used in the top panel experiments were run on 4%–12% Bis-Tris gels and were visualized with the Simple Blue Safe Stain (bottom gels). Left bottom gel: 1, GST; 2, GST-CRTC2 FL; 3, GST-CRTC2(1–55); 4, GST-CRTC2(56–144); 5, GST-CRTC2(143–320); 6, GST-CRTC2(316–693). Right bottom gel: 1, GST; 2, GST-CRTC2(316–450); 3, GST-CRTC2(445–561); 4, GST-CRTC2(561–693); 5, GST-CRTC2(316–561); 6, GST-CRTC2(445–693); 7, GST-CRTC2(316–693). C, GR LBD and CRTC2(316–693) interact with each other in an former's AF2-independent fashion in mammalian 2-hybrid assays. HCT116 cells were transfected with the GAL4-DBD-GR LBD, GAL4-DBD-GR wild type (WT), or its AF2-defective mutant (ΔAF2) fusion-expressing plasmid and the plasmid expressing the FL CRTC2 or its indicated fragments fused to VP16 AD, together with p17mer-TK-Luc and the control pGL4.73[hRluc/SV40]. 10⁻⁶M Dex or vehicle was subsequently added to media. Bars represent mean ± SE values of firefly luciferase activity normalized for Renilla luciferase activity. In right panel, numbers above the 2 bars indicate mean ± SE values of the fold association of the shown 2 molecules (calculated by dividing the values over the baseline obtained in the presence of CRTC2 and Dex by those obtained in the absence of CRTC2 but in the presence of Dex). **, P < .01, compared with the baseline in the presence of Dex; #, the statistical value was obtained by comparing with the condition in the presence of GAL4 DBD-GR WT and Dex and in the absence of VP16 AD-CRTC2. D, The CRTC2(1–561) does not enhance GR-induced transcriptional activity but acts as a dominant-negative mutant for the WT CRTC2 in HCT116 cells. HCT116 cells were transfected with the GR-expressing plasmid together with pMMTV-Luc and the control pGL4.73[hRluc/SV40]. 0.2 μg/well of the plasmid expressing WT CRTC2 and/or its mutant harboring amino acids 1–561 was also included in some conditions. 10⁻⁶M Dex were subsequently added to media. Bars represent mean ± SE values of firefly luciferase activity normalized for Renilla luciferase activity. **, P < .01; n.s., not significant, compared with the conditions indicated. E, Mutual binding sites and functional distribution of the human GR and CRTC2. GR and CRTC2 mutually interact with each other through the former's LBD (not including AF-2) and the latter's portion spanning amino acids 561–693. Their functional domains and serine residues phosphorylated by various kinases are shown in CRTC2 (9). CBD, CREB-binding domain; REG, regulatory domain; SD, splicing domain; TAD, transactivation domain.

We then verified the functional importance of the physical interaction we identified in the GST pull-down and the mammalian 2-hybrid assay (Figure 5D). In the reporter assay employing the MMTV promoter in HCT116 cells, wild-type CRTC2 strongly enhanced GR-induced transcriptional activity, whereas CRTC2(1–561), which is defective in its binding site for GR LBD, lost the enhancing effect. Indeed, this CRTC2 fragment behaved as a dominant negative mutant for wild-type CRTC2 on the enhancement of GR transcriptional activity. Thus, these results indicate that GR and CRTC2 physically interact with each other through LBD of the former and the portion spanning amino acids 561–693 of the latter (Figure 5E). Additionally, the ligand-dependent interaction of these molecules observed in the coimmunoprecipitation and mammalian 2-hybrid assays may in part be based on the ligand-induced translocation of GR to the nucleus and subsequent access of this receptor to DNA.

CRTC2 is required for the cooperative stimulation of G6p mRNA expression by the glucocorticoid- and cAMP-signaling pathways in vitro

It is known that GR and CREB cooperatively stimulate mRNA expression of the G6p gene upon their association to their respective response elements located in the promoter region of this gene (13, 14). Physical interaction of GR and CREB, however, is not clear (37, 38). We therefore examined possible contribution of CRTC2 to this association in rat hepatoma H4IIE cells. Dexamethasone activated G6p mRNA expression by approximately 3-fold, whereas the adenylyl cyclase activator forskolin significantly stimulated the expression, and addition of dexamethasone further increased it additively. Transfection of CRTC2 siRNA markedly reduced CRTC2 mRNA expression even in the presence of forskolin and abolished dexamethasone-dependent enhancement (Figure 6A). This CRTC2 siRNA markedly suppressed endogenous expression of the Crtc2 mRNA (Figure 6B). Transfection of CRTC2(1–561), which we found to function as a dominant negative mutant, also abolished dexamethasone-dependent enhancement of G6p mRNA expression, even though this mutant weakly enhanced forskolin-stimulated mRNA expression of this gene (Figure 6A). These results reveal that CRTC2 is required for the cooperative regulation of the G6p gene expression by GR and CREB, possibly by functioning as a bridging factor for these 2 molecules. We tested this possibility using a ChIP assay for the association of CRTC2, GR, and CREB on previously reported GREs and CRE of the G6p gene (10). CRTC2 was associated with GREs of this gene in a dexamethasone-dependent fashion, and addition of forskolin further increased its association (Figure 6C, top panel). This enhancement on the attraction of CRTC2 to G6p GREs was completely attenuated with RU 486, a known GR antagonist. Interestingly, GR's association to G6p GREs was further enhanced by forskolin in the presence of dexamethasone (Figure 6C, middle panel). CREB was constitutively associated with G6p CRE in the presence and absence of forskolin in accordance with previous reports (16, 39), whereas this CREB binding was weakly enhanced by dexamethasone, which was further abolished in the presence of RU 486 (Figure 6C, bottom panel). In contrast to the G6p gene, which harbors both GREs and CRE, forskolin did not change the attraction of CRTC2 or GR to GREs of the Per1 gene, which has GREs but no CRE in the ChIP target area (Figure 6D) (30). We chose Per1 GREs instead of Gilz GREs in this experiment, because the former is well verified in rodents, whereas the latter is not characterized as yet (30–33). These results are consistent with the cooperative attraction of CRTC2 to the G6p promoter by dexamethasone. Also, forskolin enhanced the G6p mRNA expression induced by dexamethasone, and both GREs and CRE were required for this cooperation. It appears that CRTC2 acts as a bridging factor for GR and CREB on the G6p promoter, facilitating the association of these receptor/transcription factor to their respective binding site when both signaling systems are activated.

Figure 6.

CRTC2 acts as a bridging factor for the cooperation between GR and CREB on the G6p gene. A and B, CRTC2 is required for the cooperative enhancement of the G6p mRNA expression by forskolin (FSK) and dexamethasone (Dex) in H4IIE cells. H4IIE cells were transfected with CRTC2 or control siRNA, or CRTC2(1–561)-expressing plasmid. They were subsequently treated with 10⁻⁵M FSK and/or 10⁻⁶M Dex. mRNA expression of Crtc2, G6p, and Rplp0 was determined with the SYBR Green real-time PCR using their specific primers. Bars represent mean ± SE values of fold G6p (A) or Crtc2 (B) mRNA expression normalized for Rplp0 mRNA expression. Broken line indicates the level of fold expression as “1.” **, P < .01; n.s., not significant, compared with the conditions indicated (A), or to the conditions transfected with control siRNA with the same Dex treatment (B). C and D, CRTC2, GR, and CREB are differentially attracted to G6p and Per1 GREs in H4IIE cells. H4IIE cells were transfected and treated with 10⁻⁵M FSK, 10⁻⁶M Dex, and/or 10⁻⁵M RU 486. Association of CRTC2, GR, or CREB on G6p or Per1 GREs or G6p CRE was determined with the ChIP assay using anti-CRTC2, anti-GR, or anti-CREB antibody and the subsequent SYBR Green real-time PCR using specific primers for G6p or Per1 GREs, or G6p CRE to test for CRTC2, GR, or CREB's association to DNA. Y-axis, data were normalized for signal with control IgG, and the increase (fold association) was further calculated as the baseline without any treatment as “1.” Bars represent mean ± SE values of fold association. Broken line indicates the level of fold association as “1.” *, P < .05; **, P < .01; n.s., not significant, compared with the conditions indicated.

CRTC2 is required for the glucocorticoid-associated cooperative maintenance of serum glucose levels in fasting mice through enhancement of G6p and Pepck mRNA expression

To examine the physiologic importance of CRTC2 on glucocorticoid-mediated regulation of glucose metabolism, we studied the mice injected with either control or the CRTC2 shRNA-expressing lentiviral particles. Twenty-four hours after injection of lentiviral particles, mice were injected with dexamethasone, RU 486 and/or vehicle (PBS). Immediately after this treatment, they were fasted or kept under normal chaw for 3 hours (Figure 7). Dexamethasone injection increased serum glucose levels in both normal chaw and fasting mice (Figure 7A). Infection of these mice with CRTC2 shRNA-expressing lentiviral particles, which markedly attenuated the CRTC2 protein expression in the liver (Figure 7B), significantly reduced serum glucose levels and abolished their dexamethasone-mediated elevation. RU 486 injection also attenuated dexamethasone-induced increase of serum glucose levels observed under fasting. Dexamethasone and fasting increased mRNA expression of G6p and Pepck genes in the liver, and simultaneous application of these 2 stimuli markedly increased mRNA expression of these genes, indicating the presence of cooperation between dexamethasone and fasting (Figure 7, C and D). Again, CRTC2 shRNA-expressing lentiviral particles abolished this cooperative enhancement as well as fasting-mediated increase of mRNA expression of these genes. RU 486 abolished dexamethasone-induced increase of their mRNA expression, whereas it did not completely attenuate their mRNA elevation induced by fasting. Taken together, these results strongly suggest that CRTC2 is required for the maintenance of glucose levels under fasting by supporting the cooperative enhancement of the G6p and Pepck gene expression by glucocorticoid-activated GR and cAMP (glucagon)-stimulated CREB.

Figure 7.

CRTC2 is required for the cooperative maintenance of serum glucose levels and enhancement of G6p and Pepck mRNA expression by fasting and dexamethasone (Dex) in mouse. Female mice (5 for each condition) were injected with the control or CRTC2 shRNA-expressing lentiviral particles (ip injection with 1 × 10⁶ infectious units of virus per animal). Twenty-four hours after viral infection, they were injected ip with Dex and/or RU 486 (1.5 mg/kg of Dex or 10 mg/kg of RU 486) and were subjected to 3-hour fasting or normal chaw immediately after the steroid treatment. Their sera and livers were sampled for measuring serum glucose levels and G6p, Pepck, or Rplp0 mRNA expression, respectively. Bars represent mean ± SE values of serum glucose levels (A) or fold G6p (C) or Pepck (D) mRNA expression normalized for Rplp0 mRNA expression. Broken line indicates the level of fold expression as “1.” Expression of CRTC2 or control β-actin was examined using Western blottings in the liver of the mice under the treatment with CRTC2 shRNA-expressing or control lentiviral particles (B). *, P < .05; **, P < .01; n.s., not significant, compared with the conditions indicated.

Discussion

We examined the impact of CRTC2 on GR-induced transcriptional activity using several different approaches. We found that CRTC2 strongly enhanced the transcriptional activity of GR through the direct interaction between the C-terminal portion (amino acids 561–693) of the CRTC2 and LBD of the GR. CRTC2 was required for the functional cooperation between glucocorticoid- and cAMP-mediated signaling pathways on the expression of G6P and PEPCK genes by facilitating the attraction of GR and CREB to their respective response elements. CRTC2 also contributed to the maintenance of serum glucose levels during fasting. In addition to GR, CRTC2 modulated the transcriptional activity of PR, suggesting that its effect is not limited to GR but can alter function of other steroid/nuclear hormone receptors.

Gluconeogenesis is essential for the production of glucose during fasting to maintain circulating levels of glucose (40). Gluconeogenesis involves multiple stepwise chemical reactions in which G6P and PEPCK are the rate-limiting enzymes (8). Upon fasting, the hypothalamic-pituitary-adrenal axis and the pancreatic islet α-cells sense reduction of serum glucose levels, and secrete massive amounts of glucocorticoids and glucagon into circulation (14, 15). Secreted glucocorticoids and glucagon then cooperatively stimulate the expression of the G6P and PEPCK genes, strongly activating gluconeogenesis in the liver to maintain the serum levels of glucose. It was previously known that glucagon and its downstream cAMP/CREB-signaling pathway, including CRTC2, play a major role in the activation of gluconeogenesis especially in the early phase (∼several hours) of fasting (9, 13). On the other hand, glucocorticoids/GR-signaling system together with the forkhead transcription factor 1, hepatocyte nuclear factor 4 (HNF4), and their coactivator, the peroxisome proliferator-activated receptor-γ coactivator 1 maintain the gluconeogenesis in its later phase (∼24 h to days) (9, 13). This previous consensus on the regulation of gluconeogenesis during fasting is rather simple, because glucocorticoids and glucagon (at the transcriptional level, GR and CREB) are reported to activate gluconeogenesis cooperatively through changing the transcriptional rates of its key enzyme genes, G6P and PEPCK in vitro and in vivo (41). Our results also indicate that the GR-signaling system is important in the maintenance of gluconeogenesis from the early phase of fasting.

We found that CRTC2, the coactivator previously thought to be specific to CREB, functions as a coactivator of GR, binding to the latter's LBD and strongly enhancing its transcriptional activity on the simple glucocorticoid-responsive genes devoid of CRE. Further, CRTC2 is necessary for cooperative activation of the G6P and PEPCK genes by GR and CREB. The promoters of these genes contain both GREs and CRE (thus, they are composite promoters in regard to these consensus sequences), and CRTC2 facilitates the attraction of GR to their respective response elements. The exact mechanism(s) underlying this phenomenon is (are) not known, but it appears that CRTC2 stabilizes the transcriptional complex containing GR, CREB and itself, formed on these promoters. Although CREB was consistently associated with G6p CRE both in the presence and absence of forskolin, this association was weakly enhanced by dexamethasone in the presence of forskolin in this study. It is likely that stabilization of the transcriptional complex by CRTC2 may help causing this weak change.

Although we did not examine the possibility, it is feasible that CRTC2 also interacts with HNF4 and supports its transcriptional activity, because HNF4 is an orphan receptor and a member of the steroid/thyroid/nuclear receptor super family similar to GR and PR. Indeed, the promoter region of the G6P and PEPCK genes that confers binding of GR contains binding sites for HNF4 (10, 42). The promoter region of the rat Pepck gene also contains a binding site for another orphan receptor the chicken ovalbumin upstream-transcription factor, and this molecule cooperates with GR for stimulating the Pepck promoter activity (42, 43). Thus, CRTC2 might organize the transcriptional activity of multiple transcription factors including CREB, GR, and possibly, HNF4 and chicken ovalbumin upstream-transcription factor, on the G6P and/or PEPCK gene promoters and play a central role in the regulation of gluconeogenesis in the early phase of fasting. In addition, CRTC2 may also be functional for smooth transition of transcriptional regulation from the early phase maintained by CRTC2 to the later phase organized by peroxisome proliferator-activated receptor-γ coactivator 1 through interacting with the GR, which is functional in both phases (13). Because fasting induces the colocalization of several transcription factors including GR around the CREB sites constitutively occupied by CREB, CRTC2 may be one of the cofactors underlying and supporting this fasting-associated transcriptional reprogramming (39). Further, the upstream kinases phosphorylating key residues of the CRTC2 for its cytoplasmic localization might influence gluconeogenesis by modulating through CRTC2 the formation/activity of the transcription complex containing these nuclear receptors/transcription factors on the G6P and/or PEPCK gene promoters.

In addition to the cooperative regulation of gluconeogenesis, CRTC2 might also be important for the signal cross talk between the GR- and the catecholamine/adrenergic receptor-signaling system, because the β₂-adrenoreceptor agonists enhance the glucocorticoid actions and reverse the cellular resistance to these steroids observed in the patients with bronchial asthma (44).

CRTC2 modulates the transcriptional activity of PR in addition to GR. Its regulation on this steroid receptor was repressive in our experimental conditions, and the molecular mechanism(s) underlying the CRTC2-mediated differential regulation of PR- and GR-induced transcriptional activity is(are) not known. However, we speculate that CRTC2 disrupts the transcriptional complex formed with PR on its responsive promoters in our experimental conditions. Indeed, coactivators and corepressors can have varied activity in a context- and gene-dependent manner (21, 45). Crosstalk between progesterone and the cAMP-signaling pathway is well known in various aspects of female reproductive system (46). Thus, it is highly possible that CRTC2 is a key molecule for the coordinated regulation of PR and cAMP/CREB-signaling system in this important biologic activity.

Acknowledgments

We thank Dr N. P. Curthoys, Dr R. M. Evans, Dr R. Goodman, Dr G. L. Hager, Dr S. S. Simons Jr, Dr H. Takemori, Dr M. J. Tsai, and Dr N. Warriar for providing us their plasmids and Mr E. K. Zachman for technical support and proofreading the manuscript.

This work was supported by the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development, the National Institutes of Health, and the Sidra Medical and Research Center. S.S. was supported by the Asahikawa Medical University.

Disclosure Summary: The authors have nothing to disclose.

Abbreviations

 
• AD

  activation domain

 
• AF

  activation function

 
• ChIP

  chromatin immunoprecipitation

 
• CRE

  cAMP-response element

 
• CREB

  CRE-binding protein

 
• CRTC2

  CREB-regulated transcription coactivator 2

 
• DBD

  DNA-binding domain

 
• FL

  full length

 
• GAL4

  galactose-responsive transcription factor GAL4

 
• GILZ

  glucocorticoid-induced leucine zipper

 
• G6P

  glucose-6-phosphatase

 
• GR

  glucocorticoid receptor

 
• GRE

  glucocorticoid-response element

 
• GFP

  green fluorescent protein

 
• HNF4

  hepatocyte nuclear factor 4

 
• KLF5

  Krueppel-like factor 5

 
• LBD

  ligand-binding domain

 
• MMTV

  mouse mammary tumor virus

 
• NCoA

  nuclear receptor coactivator

 
• NTD

  N-terminal domain

 
• PEPCK

  phosphoenolpyruvate carboxykinase

 
• Per1

  period 1

 
• PKA

  protein kinase A

 
• PR

  progesterone receptor

 
• RPLP0

  ribosomal protein large P0.