Title: Mechanism for Fatty Acid “Sparing” Effect on Glucose-induced Transcription
Abstract: Carbohydrate-responsive element-binding protein (ChREBP) is a new transcription factor that binds to the carbohydrate-responsive element of the l-type pyruvate kinase gene (l-PK). The aim of this study was to investigate the mechanism by which feeding high fat diets results in decreased activity of ChREBP in the liver (Yamashita, H., Takenoshita, M., Sakurai, M., Bruick, R. K., Henzel, W. J., Shillinglaw, W., Arnot, D., and Uyeda, K. (2001) Proc. Natl. Acad. Sci. U. S. A. 98, 9116–9121). We cloned the rat liver ChREBP gene for use throughout this study. Acetate, octanoate, and palmitate inhibited the glucose-induced activation of l-PK transcription in ChREBP-overexpressed hepatocytes. In these hepatocytes, the cytosolic AMP concentration increased 30-fold and AMP-activated protein kinase activity was activated 2-fold. Similarly to the fatty acids, 5-amino-4-imidazolecarboxamide ribotide, a specific activator of AMP-activated protein kinase (AMPK) also inhibited the l-PK transcription activity in ChREBP-overexpressed hepatocytes. Using as a substrate a truncated ChREBP consisting of the C-terminal region, we demonstrated that phosphorylation by AMPK resulted in inactivation of the DNA binding activity. AMPK specifically phosphorylated Ser568 of ChREBP. A S568A mutant of the ChREBP gene showed tight DNA binding and lost its fatty acid sensitivity, whereas a S568D mutant showed weak DNA binding and inhibited l-PK transcription activity even in the absence of fatty acid. These results strongly suggested that the fatty acid inhibition of glucose-induced l-PK transcription resulted from AMPK phosphorylation of ChREBP at Ser568, which inactivated the DNA binding activity. AMPK was activated by the increased AMP that was generated by the fatty acid activation. Carbohydrate-responsive element-binding protein (ChREBP) is a new transcription factor that binds to the carbohydrate-responsive element of the l-type pyruvate kinase gene (l-PK). The aim of this study was to investigate the mechanism by which feeding high fat diets results in decreased activity of ChREBP in the liver (Yamashita, H., Takenoshita, M., Sakurai, M., Bruick, R. K., Henzel, W. J., Shillinglaw, W., Arnot, D., and Uyeda, K. (2001) Proc. Natl. Acad. Sci. U. S. A. 98, 9116–9121). We cloned the rat liver ChREBP gene for use throughout this study. Acetate, octanoate, and palmitate inhibited the glucose-induced activation of l-PK transcription in ChREBP-overexpressed hepatocytes. In these hepatocytes, the cytosolic AMP concentration increased 30-fold and AMP-activated protein kinase activity was activated 2-fold. Similarly to the fatty acids, 5-amino-4-imidazolecarboxamide ribotide, a specific activator of AMP-activated protein kinase (AMPK) also inhibited the l-PK transcription activity in ChREBP-overexpressed hepatocytes. Using as a substrate a truncated ChREBP consisting of the C-terminal region, we demonstrated that phosphorylation by AMPK resulted in inactivation of the DNA binding activity. AMPK specifically phosphorylated Ser568 of ChREBP. A S568A mutant of the ChREBP gene showed tight DNA binding and lost its fatty acid sensitivity, whereas a S568D mutant showed weak DNA binding and inhibited l-PK transcription activity even in the absence of fatty acid. These results strongly suggested that the fatty acid inhibition of glucose-induced l-PK transcription resulted from AMPK phosphorylation of ChREBP at Ser568, which inactivated the DNA binding activity. AMPK was activated by the increased AMP that was generated by the fatty acid activation. l-type pyruvate kinase carbohydrate-responsive element-binding protein AMP-activated protein kinase 3-hydroxy-3-methylglutaryl-CoA 5′- and 3′-rapid amplification of cDNA ends green fluorescent protein 5-amino-4-imidazolecarboxamide ribotide cAMP-dependent protein kinase nuclear localization signal proline-rich stretch basic helix-loop-helix, leucine zipper leucine zipper-like Glucose metabolism in liver is inhibited by administration of fatty acids, the so-called “glucose sparing” effect (1Williamson J.R. Krebs H.A. Biochem. J. 1961; 80: 540-547Crossref PubMed Scopus (118) Google Scholar, 2Ross B.D. Hems R. Krebs H.A. Biochem. J. 1967; 102: 942-951Crossref PubMed Google Scholar, 3Struck E. Ashmore J. Wieland O.H. Biochem. J. 1965; 343: 107-110Google Scholar). Fatty acids inhibit genes of key enzymes of glycolysis and lipogenesis such as l-pyruvate kinase (l-PK),1acetyl-CoA carboxylase, and fatty acid synthetase. l-PK, regulating the flux of metabolites through the pyruvate-phosphoenolpyruvate cycle (4Liimatta M. Towle H.C. Clarke S. Jump D.B. Mol. Endocrinol. 1994; 8: 1147-1153PubMed Google Scholar), is known to play an important role in hepatic glucose and lipid metabolism. The activity ofl-PK is subject to acute control by covalent modification and allosteric effectors (5Foufelle F. Girard J. Ferre P. Adv. Enzyme Regul. 1996; 36: 199-226Crossref PubMed Scopus (71) Google Scholar). On the other hand, long term control ofl-PK is achieved by regulating l-PK gene transcription (5Foufelle F. Girard J. Ferre P. Adv. Enzyme Regul. 1996; 36: 199-226Crossref PubMed Scopus (71) Google Scholar). Fatty acids inhibit transcription ofl-PK and other enzymes in glycolysis and lipogenesis pathways, whereas excess glucose induces expression of these genes (6Duplus E. Glorian M. Forest C. J. Biol. Chem. 2000; 275: 30749-30752Abstract Full Text Full Text PDF PubMed Scopus (303) Google Scholar). However, the mechanism of fatty acid inhibition of transcription is not understood. We recently identified a new transcription factor, which binds specifically to the carbohydrate-responsive element of thel-PK gene, and we have termed this new transcription factor, “carbohydrate-responsive element-binding protein” (ChREBP) (7Yamashita H. Takenoshita M. Sakurai M. Bruick R.K. Henzel W.J. Shillinglaw W. Arnot D. Uyeda K. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 9116-9121Crossref PubMed Scopus (543) Google Scholar). ChREBP is expressed specifically in liver and is responsive to diet. ChREBP is activated by a high carbohydrate diet and inhibited by a high fat diet (7Yamashita H. Takenoshita M. Sakurai M. Bruick R.K. Henzel W.J. Shillinglaw W. Arnot D. Uyeda K. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 9116-9121Crossref PubMed Scopus (543) Google Scholar). We showed that the DNA binding activity of ChREBP in nuclear extract of livers from rats fed a high fat diet is inhibited compared with those of control rats fed laboratory chow or a high carbohydrate diet, suggesting that ChREBP may be intimately involved in fatty acid inhibition of glucose metabolism. AMP-activated protein kinase (AMPK) is a multisubunit protein kinase, which appears to play a central role in lipid metabolism (8Hardie D.G. Biochim. Biophys. Acta. 1992; 1123: 231-238Crossref PubMed Scopus (166) Google Scholar). AMPK was first shown to catalyze phosphorylation and inactivation of 3-hydroxy-3-methylglutaryl-CoA (HMG-CoA) reductase and acetyl-CoA carboxylase, the rate-limiting enzymes of cholesterol and fatty acid synthesis, respectively (9Hardie D.G. Carling D. Eur. J. Biochem. 1997; 246: 259-273Crossref PubMed Scopus (1155) Google Scholar). AMPK is activated by a high AMP/ATP ratio in the cytosol, which occurs under stressed conditions such as heat shock, hypoxia, arsenite treatment, and starvation (10Moore F. Weekes J. Hardie D.G. Eur. J. Biochem. 1991; 199: 691-697Crossref PubMed Scopus (200) Google Scholar). The increase in AMP and the concomitant decrease in ATP, activating AMPK, suggests that this protein kinase inhibits activation of biosynthetic enzymes such as acetyl-CoA carboxylase to conserve ATP (11Hardie D.G. Carling D. Carlson M. Annu. Rev. Biochem. 1998; 67: 821-855Crossref PubMed Scopus (1290) Google Scholar). The objectives of this investigation were to determine: 1) whether our previous observation that the fatty acid inhibition of ChREBP in intact rats by fatty acid feeding was a result of AMPK action; 2) the target site of the phosphorylation; and 3) which processes, the nuclear localization or the DNA binding activity, are affected by fatty acids. All reagents were purchased from Sigma unless otherwise indicated. Results of mass spectrometry analysis of tryptic peptides of the purified ChREBP indicated that ChREBP matched with sequences in the mouse putative hepatic transcription factor, WBSCR14 (accession no. AF156604) (12de Luis O. Valero M.C. Jurado L.A. Eur. J. Hum. Genet. 2000; 8: 215-222Crossref PubMed Scopus (60) Google Scholar), with variable degrees of identity. To obtain a portion of rat ChREBP gene, two degenerate primers were designed based on the sequences of human and mouse WBSCR14. The degenerate oligonucleotide primer sequences were forward primer (5′-ATHCAYWSNGGNCAYTTYATGG-3′) and reverse primer (5′-GTNCCYTCNGTNACNGCNCKNG-3′). Total RNA was extracted from rat liver by Isogen (Nippon Gene, Toyama, Japan). The first stranded cDNA was synthesized from poly(A)+ RNA prepared with a mRNA Purification Kit (Amersham Biosciences, Inc.) using SuperScript II reverse transcriptase (Invitrogen). The PCR reaction mixture, in a final volume of 50 μl, contained 10 mm Tris-HCl, pH 8.3, 50 mm KCl, 1.5 mm MgCl2, 200 μm dNTPs, 1 μm amounts of each primer, 5 μl of cDNA, and 5 units of AmpliTaq Gold DNA polymerase (Applied Biosystems, Foster, CA). The PCR product was purified from an agarose gel and cloned into a pGEM-T Easy vector (Promega, Madison, WI). The DNA sequence of the cloned fragment was confirmed using the dideoxy chain termination method, with the BigDye terminator cycle sequencing kit and an ABI 377 DNA sequencer (Applied Biosystems). Based on the sequence of this cDNA fragment obtained using the degenerate primers, two additional pairs of primers were designed and used in 5′- and 3′-rapid amplification of cDNA ends (5′- and 3′-RACE) in an effort to generate a cDNA containing the entire length of rat the ChREBP gene. In these reactions, it was necessary to use a nested primer to generate a fragment of the correct size. The antisense primer and its nested primer were 5′-CCCAAGCAGCACAGGCACCAC-3′ and 5′-GCTCTTCCTCCGTTGCACATACTG-3′ to generate the fragment of cDNA corresponding to the 5′ end. The sense primer and its nested primer used in 3′-RACE were 5′-ACCCGCACGCTGCACAACTGGAAG-3′ and 5′-CTGAGGGATGAAATAGAGGAGCTC. The 5′-RACE product (450 bp) and 3′-RACE product (980 bp) were cloned into the pGEM-T Easy vector; the sequences of three independent clones of each were verified by DNA sequencing as described above. The full-length rat wild type ChREBP cDNA (accession no. AB074517) was ligated into the mammalian expression vector pcDNA3 (ChREBP/pcDNA3; Invitrogen, Carlsbad, CA) (7Yamashita H. Takenoshita M. Sakurai M. Bruick R.K. Henzel W.J. Shillinglaw W. Arnot D. Uyeda K. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 9116-9121Crossref PubMed Scopus (543) Google Scholar) or pEGFP-N3 (ChREBP/pEGFP; CLONTECH, Palo Alto, CA) encoding enhanced green fluorescent protein (GFP). Thel-PK gene promoter region, between positions −206 and −7, was ligated into the luciferase expression plasmid, pGL-3 Basic vector (Promega), as described previously (7Yamashita H. Takenoshita M. Sakurai M. Bruick R.K. Henzel W.J. Shillinglaw W. Arnot D. Uyeda K. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 9116-9121Crossref PubMed Scopus (543) Google Scholar). A truncated derivative of the C terminus of ChREBP (amino acids 540–804) was synthesized with PCR and inserted at the BamHI and HindIII sites of the T7-based expression vector, pGEX-5 (Amersham Biosciences, Inc.) to produce an N-terminal fusion with glutathione reductase. Point mutations in the ChREBP clones were created in the putative phosphorylation sites for AMPK using QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA) according to the manufacturer's instruction. The recombinant clone carrying the C terminus of ChREBP was expressed in BL21(DE3)pLysS cells induced with isopropyl-β-d-galactopyranoside. The expressed protein was partially purified from bacterial cell extracts using a glutathione-Sepharose 4B column according to the manufacturer's instructions. During affinity chromatography, the fusion protein lost DNA binding activity, presumably by rapid proteolysis. Therefore, crude extracts of the bacterial cell were used. Primary hepatocytes were prepared from male Sprague-Dawley rats (250–300 g) using the collagenase perfusion method (13Kawaguchi T. Sakisaka S. Harada M. Hanada S. Taniguchi E. Koga H. Sasatomi K. Tanikawa K. Sata M. Hepatol. Res. 2001; 20: 144-154Crossref PubMed Scopus (8) Google Scholar) and plated in collagen-coated 35-mm tissue culture plates (Primaria Falcon, Franklin Lakes, NJ) at a density of 1.0 × 106 cells/well in glucose-free Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 100 nm dexamethasone, 10 nminsulin, 100 units/ml penicillin, 100 μg/ml streptomycin, 10% dialyzed fetal bovine serum (Invitrogen), and 27.5 mmglucose. After a 6-h attachment period, hepatocytes were transfected using synthetic liposome (LipofectAMINE 2000; Invitrogen) in Opti-MEM I reduced serum medium as described (7Yamashita H. Takenoshita M. Sakurai M. Bruick R.K. Henzel W.J. Shillinglaw W. Arnot D. Uyeda K. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 9116-9121Crossref PubMed Scopus (543) Google Scholar) and incubated for 12 h. As a control, hepatocytes were exposed to synthetic liposome in the same manner as each experimental group. The medium containing the liposome-DNA complex was removed and replaced with glucose-free Dulbecco's modified Eagle's medium supplemented with 27.5 mm glucose plus either 150 μm albumin-bound specific fatty acids at a fatty acid/albumin ratio of 4:1 or 5-amino-4-imidazolecarboxamide ribotide (AICAR). The source of albumin for all studies was essentially fatty acid-free bovine serum albumin. Hepatocytes were scraped and collected directly into 0.5 ml of ice-cold 6% (v/v) HClO4. The concentrations of ADP and ATP were determined enzymatically in the acid extract with methods described (14Kawaguchi T. Veech R.L. Uyeda K. J. Biol. Chem. 2001; 276: 28554-28561Abstract Full Text Full Text PDF PubMed Scopus (64) Google Scholar), and the cytosolic AMP was calculated by the method of Veech et al. (15Veech R.L. Lawson J.W. Cornell N.W. Krebs H.A. J. Biol. Chem. 1979; 254: 6538-6547Abstract Full Text PDF PubMed Google Scholar). Hepatocytes were directly lysed in the culture medium by adding 1.5 ml of buffer A (50 mm Tris-HCl, pH 7.5, 50 mm NaF, 5 mm sodium pyrophosphate, 1 mm EDTA, 10% glycerol, 1 mm dithiothreitol) + 1% Triton X-100. The extract was centrifuged at 15,800 × g for 15 min, and the resulting supernatant solution was transferred to a fresh tube, adjusted to 10% with polyethylene glycol 8000 (Appligene, Illkirich, France), and incubated on ice for 20 min. Following centrifugation at 15,800 × g for 15 min, the pellet was suspended in 100 μl of buffer A. Aliquots were used to assay the AMPK activity with the SAMS peptide (synthetic peptide HMRSAMSGLHLVKRR) as a substrate and [γ-32P]ATP (specific activity, 3000 Ci/mmol) in the presence of 200 μm 5′-AMP as described previously (16da Silva Xavier G. Leclerc I. Salt I.P. Doiron B. Hardie D.G. Kahn A. Rutter G.A. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 4023-4028Crossref PubMed Scopus (187) Google Scholar). Transfected cells were cultured for 12 h in culture medium containing 27.5 mmglucose plus 150 μm albumin-bound specific fatty acids or AICAR, washed twice with 2 ml Ca2+- and Mg2+-free phosphate-buffered saline, and lysed with Passive lysis buffer (Promega). Firefly luciferase and Renilla (sea pansy) luciferase activities were measured sequentially using a Dual-Luciferase Reporter assay system (Promega) and a model TD-20E Luminometer (Turner Design, Sunnyvale, CA). After measuring the firefly luciferase signal and the Renilla luciferase signal, the index of relative luciferase activity, indicating l-PK transcription activity, was calculated according to manufacturer's instructions (Promega). Subcellular localization of GFP-fused wild type ChREBP or its mutants were determined using a confocal laser scanning microscope (Bio-Rad). To quantitate nuclear localization of ChREBP, for each condition, 100 transfected hepatocytes from five independent experiments were scored in a blinded fashion for whether the GFP-fused ChREBP was predominantly nuclear or cytoplasmic. The identity of the nucleus was verified by comparative phase-contrast microscopy. Gel mobility shift assays were performed as described previously (7Yamashita H. Takenoshita M. Sakurai M. Bruick R.K. Henzel W.J. Shillinglaw W. Arnot D. Uyeda K. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 9116-9121Crossref PubMed Scopus (543) Google Scholar). Double-stranded oligonucleotides (major late transcription factor) were prepared by mixing equal amounts of the complementary single-stranded DNAs in 50 mm NaCl, heated to 70 °C for 15 min, and cooled to room temperature. The annealed oligonucleotides were labeled with 32P in the presence of [γ-32P]ATP and polynucleotide kinase. DNA-binding reactions were carried out in a 20-μl reaction mixture containing 20 mm Hepes-KOH, pH 7.9, 50 mm KCl, 5 mm dithiothreitol, 0.2 mm EDTA, 0.5 mm phenylmethylsulfonyl fluoride, 10% glycerol, and 2 μg of poly(dI-dC). The reaction mixture contained 40,000 cpm (1 nm) of the labeled DNA and 3 μg of recombinant ChREBP. The reaction mixture was incubated at room temperature for 30 min, applied to a 4.5% nondenaturing polyacrylamide gel for electrophoresis at 10 V/cm at 4 °C in a Tris/glycine buffer system for 2 h. The gel was dried and exposed to Hyperfilm-MP (Amersham Biosciences, Inc.) at −80 °C. To quantify the DNA binding, the bands corresponding to ChREBP-DNA complex were cut from the gel and counted in scintillation fluid. The reaction mixture contained in a final volume of 100 μl 20 mm HEPES-NaOH, pH 7.0, 0.4 mmdithiothreitol, 0.01% Brji-35, 300 μm AMP, 5 mm MgCl2, 0.2 mm[γ-32P]ATP (3000 Ci/mmol), and 50 milliunits of AMPK (Upstate Biotechnology, Inc., Lake Placid, NY). The following synthetic peptides at 25 μm were used as substrates: SAMS peptide (HMRSAMSGLHLVKRR), a peptide containing a putative AMPK phosphorylation site (Ser568) of ChREBP (STVPSTLLRPPESPDAVP; amino acids 556–573 of rat ChREBP), and a peptide containing a cAMP-dependent protein kinase (PKA) phosphorylation site (Thr666) (VDNNKMENRRITTHISAEQKR; amino acids 656–675 of the rat ChREBP). Each peptide was added to a tube containing the reaction mixture, and the tube was incubated at 30 °C. At the times indicated, aliquots were removed and spotted onto P-81 ion exchange chromatography paper (Whatman International, Maidstone, United Kingdom). The paper was washed with 0.75% phosphoric acid, and phosphate incorporated into each peptide was measured by counting in liquid scintillation fluid as described previously (17Kitamura K. Kangawa K. Matsuo H. Uyeda K. J. Biol. Chem. 1988; 263: 16796-16801Abstract Full Text PDF PubMed Google Scholar). All data were expressed as mean ± S.E. Differences between two groups were analyzed using the Mann-Whitney U test. Comparisons among multiple groups were analyzed using the Kruskal-Wallis analysis of variance. A pvalue less than 0.05 was considered statistically significant. Rat liver cDNA fragment encoding ChREBP was generated by PCR using the degenerate oligonucleotide primers, and the sequence of the cloned cDNA was determined. Based on the sequence of this fragment, additional gene-specific primers were synthesized and used in 5′- and 3′-RACE, which produced 450- and 980-bp cDNA fragments. The resulting sequences overlapped and allowed determination of the complete nucleotide sequence of the ChREBP cDNA. The rat ChREBP cDNA encoded 865 amino acids (a predicted Mr= 94,780) with a nuclear localization signal (NLS) domain, a proline-rich stretch (Pro-rich) domain, a basic helix-loop-helix and leucine-zipper (bHLH/ZIP) domain, and a leucine zipper-like (ZIP-like) domain (Fig. 1, A andB). The rat ChREBP showed identities of 93.8 and 82.0% with amino acid sequences of the mouse and human counterparts, respectively, with the C-terminal region (amino acids 652–865) highly conserved (>95% identity). In contrast to the other species, the rat ChREBP had an additional PKA phosphorylation site at Ser516. To determine the glucose response of rat ChREBP, primary cultured hepatocytes were cotransfected with ChREBP/pcDNA3 and the luciferase reporter plasmid, pGL-3, carrying the l-PK promoter region between positions −206 and −7. These transfected hepatocytes were used to investigate glucose activation of l-PK gene transcription by measuring luciferase activity. Preliminary experiments showed that LipofectAMINE 2000 alone did not act as a fatty acid (data not shown). Increased transcription activity of the l-PK gene in the control, i.e. empty vector-transfected hepatocytes, was the result of endogenous ChREBP (Fig. 2). The rat ChREBP showed ∼3.5-fold activation of l-PK transcriptional activity in high glucose compared with low glucose, which was comparable with that of the mouse ChREBP (7Yamashita H. Takenoshita M. Sakurai M. Bruick R.K. Henzel W.J. Shillinglaw W. Arnot D. Uyeda K. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 9116-9121Crossref PubMed Scopus (543) Google Scholar) (Fig. 2). The cloned rat ChREBP was used throughout this investigation. Hepatocytes transfected with ChREBP and maintained in medium containing high glucose (27.5 mm) expressed highl-PK transcriptional activity (Fig.3). Addition of 150 μmacetate, octanoate, or palmitate to the high glucose medium inhibited the glucose activation by at least 75% (Fig. 3). A higher concentration (1 mm) of fatty acids eliminated the glucose activation completely (data not shown). Thus, all these fatty acids inhibited equally the ChREBP-dependent glucose activation of l-PK transcription. To determine whether this inhibition of transcription was the result of inhibition of the DNA binding activity of ChREBP, a gel shift assay of extracts of those hepatocytes was performed (1 × 106cells), but no DNA binding activity was detectable in any of the hepatocyte extracts, including those grown in glucose alone, suggesting that the activity of ChREBP was too low to detect. In the activation of fatty acids to fatty acyl-CoA, AMP is generated in the cytoplasm. We have determined the adenine nucleotide concentrations in perchloric acid extracts of hepatocytes incubated in the presence of fatty acids. The cytosolic AMP increased ∼30-fold, whereas ADP and ATP remained constant in the presence of fatty acids (Table I). The 30-fold rise in AMP/ATP ratio is sufficient for activation of AMPK, which is known to play a central role in regulation of lipid metabolism. By addition of acetate, octanoate, or palmitate to the high glucose medium, AMPK activity also was increased ∼2-fold compared with those in the absence (Fig. 4). Thus, these results suggested that the fatty acids activated AMPK as a result of the increase in the AMP/ATP ratio in hepatocytes, and the fatty acid inhibition of the glucose-induced l-PK transcription resulted from the activation of AMPK, which inactivated ChREBP.Table IAMP, ADP, and ATP concentrations, and AMP/ATP ratio in cultured hepatocytes in the presence of fatty acidsAMPADPATPAMP/ATP rationmol/g cellsμmol/g cellsμmol/g cellsGlucose35 ± 61.28 ± 0.083.11 ± 0.320.011 ± 0.0007Glucose + acetate910 ± 871-ap < 0.05 compared with glucose alone.1.32 ± 0.112.99 ± 0.360.30 ± 0.011-ap < 0.05 compared with glucose alone.Glucose + octanoate942 ± 1061-ap < 0.05 compared with glucose alone.1.29 ± 0.153.09 ± 0.550.30 ± 0.031-ap < 0.05 compared with glucose alone.Glucose + palmitate956 ± 951-ap < 0.05 compared with glucose alone.1.31 ± 0.142.91 ± 0.410.33 ± 0.011-ap < 0.05 compared with glucose alone.Hepatocytes were incubated for 12 h with 150 μmacetate, octanoate, or palmitate. AMP, ADP, ATP, and AMP/ATP ratio were determined as described under “Experimental Procedures.” Values are expressed as mean ± S.E. (n = 4).1-a p < 0.05 compared with glucose alone. Open table in a new tab Hepatocytes were incubated for 12 h with 150 μmacetate, octanoate, or palmitate. AMP, ADP, ATP, and AMP/ATP ratio were determined as described under “Experimental Procedures.” Values are expressed as mean ± S.E. (n = 4). To confirm that the activated AMPK inhibits ChREBP-mediated transcriptional activation of the l-PK gene in cultured hepatocytes, the effect of AICAR was examined. AICAR is a specific activator of AMPK that mimics AMP in intact cells. AICAR at 50 and 200 μm inhibited the glucose activation ofl-PK gene transcription ∼80% and over 90%, respectively (Fig. 5), suggesting that the activated AMPK phosphorylated ChREBP, resulting in inhibition of the glucose activation of l-PK gene transcription. Nuclear translocation of ChREBP is one of the important processes in glucose activation ofl-PK gene transcription (18Kawaguchi T. Takenoshita M. Kabashima T. Uyeda K. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 13710-13715Crossref PubMed Scopus (315) Google Scholar). To investigate the effect of fatty acids on subcellular localization of ChREBP, we constructed GFP-fused ChREBP. GFP-fused ChREBP was detected mainly in the cytoplasm of hepatocytes under low glucose and migrated into the nucleus under high glucose, indicating that high glucose induced nuclear import of ChREBP (18Kawaguchi T. Takenoshita M. Kabashima T. Uyeda K. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 13710-13715Crossref PubMed Scopus (315) Google Scholar). Hepatocytes incubated in high glucose in the presence of palmitate also showed nuclear translocation of ChREBP (data not shown). Thus, the fatty acid did not affect the translocation of ChREBP into the nucleus, suggesting that fatty acid inactivation affected only the DNA binding activity. DNA binding is another important activity of ChREBP in the glucose activation ofl-PK gene transcription (18Kawaguchi T. Takenoshita M. Kabashima T. Uyeda K. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 13710-13715Crossref PubMed Scopus (315) Google Scholar). Using an expressed, truncated ChREBP containing the suspected AMPK phosphorylation site (Ser568), we determined the effect of AMPK-dependent phosphorylation of Ser568 on DNA binding activity. The DNA binding activity, as assayed in a gel-shift assay of the truncated ChREBP, was gradually inactivated with time by treatment with AMPK (Fig. 6 andinset). To investigate possible phosphorylation by AMPK directly, we determined32P incorporation into the synthetic peptide containing the putative AMPK phosphorylation site of ChREBP. To confirm the specificity of the phosphorylation by AMPK, a synthetic peptide, termed SAMS peptide, commonly used as a substrate for AMPK, and a synthetic peptide for a PKA phosphorylation site of ChREBP were compared. As shown in Fig. 7, the AMPK site peptide of ChREBP was phosphorylated as rapidly as was the SAMS peptide, whereas the PKA phosphorylation site peptide of ChREBP was not phosphorylated, indicating that AMPK showed specificity toward the phosphorylation site of ChREBP. To confirm the effect of the AMPK phosphorylation site of ChREBP, we constructed the S568A and S568D mutants (Fig.8A) of ChREBP by site-directed mutagenesis and transfected hepatocytes with these mutant ChREBP derivatives. The DNA binding was tight in the S568A mutant and weak in the S568D mutant compared with the wild-type ChREBP (Fig.8B). The S568A mutant showed high transcriptional activity of the l-PK gene in medium containing high glucose and palmitate, as expected, because the mutation to Ala presumably mimicked the dephospho- form of ChREBP. On the other hand, the S568D mutant showed low activity of l-PK gene transcription even in the presence of high glucose, comparable with that of the wild-type ChREBP (Fig. 8B). The Asp derivative of ChREBP presumably mimicked the phosphorylated ChREBP at the AMPK target Ser568. Thus, these results strongly supported the contention that Ser568was the AMPK phosphorylation site. ChREBP, a newly discovered transcription factor, plays an essential role in glucose-induced l-PK gene transcription by binding to the carbohydrate-responsive element of l-PK promoter (7Yamashita H. Takenoshita M. Sakurai M. Bruick R.K. Henzel W.J. Shillinglaw W. Arnot D. Uyeda K. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 9116-9121Crossref PubMed Scopus (543) Google Scholar, 18Kawaguchi T. Takenoshita M. Kabashima T. Uyeda K. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 13710-13715Crossref PubMed Scopus (315) Google Scholar). It is well known that glucose metabolism is inhibited by fatty acids, which serve as an alternative fuel source and thus conserve glucose. This phenomenon has been termed the fat sparing effect on glucose (1Williamson J.R. Krebs H.A. Biochem. J. 1961; 80: 540-547Crossref PubMed Scopus (118) Google Scholar, 2Ross B.D. Hems R. Krebs H.A. Biochem. J. 1967; 102: 942-951Crossref PubMed Google Scholar, 3Struck E. Ashmore J. Wieland O.H. Biochem. J. 1965; 343: 107-110Google Scholar). Previously, we demonstrated that rats fed a high fat diet showed inhibited l-PK transcription, and the DNA binding activity of ChREBP in the nuclear extract of liver was inhibited compared with that of rats fed a high carbohydrate diet (7Yamashita H. Takenoshita M. Sakurai M. Bruick R.K. Henzel W.J. Shillinglaw W. Arnot D. Uyeda K. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 9116-9121Crossref PubMed Scopus (543) Google Scholar). In this report we demonstrated that administration of a variety of fatty acids to hepatocytes overexpressing ChREBP resulted in inhibition of the glucose activation of l-PK gene expression (Fig.3). The mechanism of ChREBP inhibition of the glucose activation ofl-PK gene expression by fatty acids appeared to be mediated by phosphorylation of Ser568 by AMPK, which was activated by increased AMP that was produced by the activation of fatty acids catalyzed by acyl-CoA synthetase (EC 6.2.1.3).Fatty acids+ATP+CoA→acylCoAsynthetaseacyl−CoA+AMP+PPiREACTION1AMPK is activated by the increased AMP/ATP ratio in cytoplasm and acts as a key metabolic “master switch” by phosphorylating target proteins involved in lipid metabolism (11Hardie D.G. Carling D. Carlson M. Annu. Rev. Biochem. 1998; 67: 821-855Crossref PubMed Scopus (1290) Google Scholar). Evidence that fatty acid inhibition of the glucose activation of ChREBP and l-PK transcription is mediated by phosphorylation by AMPK is as follows: (a) rats fed a high fat diet showed inhibition of the DNA binding activity of ChREBP (7Yamashita H. Takenoshita M. Sakurai M. Bruick R.K. Henzel W.J. Shillinglaw W. Arnot D. Uyeda K. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 9116-9121Crossref PubMed Scopus (543) Google Scholar); (b) AMP/ATP ratio increased 30-fold in the cytosol of hepatocytes incubated in the presence of fatty acids (Table I), and under these conditions, AMPK activity in the cell extracts increased 2-fold (Fig. 4); (c) AICAR inhibited both the glucose-induced activation of l-PK transcription in the ChREBP overexpressed hepatocytes (Fig. 5) and the DNA binding activity of ChREBP (Fig. 6); (d) AMPK phosphorylated the synthetic AMPK site peptide specifically at the same rate as that of SAMS peptide in vitro (Fig. 7); and (e) the S568D mutant of ChREBP lost DNA binding activity (Fig. 8B) andl-PK transcription activity (Fig. 8C), even in the presence of high glucose, whereas the S568A mutant was fully active even in the presence of fatty acids. Fatty acid synthesis occurs only under conditions of excess carbohydrate intake, most rapidly during high carbohydrate feeding. To produce the malonyl-CoA required for de novo fatty acid synthesis, glucose must be metabolized in glycolysis to pyruvate, a process requiring the action of pyruvate kinase. After it is formed in cytosol by glycolysis, pyruvate enters the mitochondria to form citrate, which leaves the mitochondria to be converted to cytosolic acetyl-CoA, the necessary precursor of malonyl-CoA. Under these conditions, blood free fatty acids are low. During periods of high fat feeding or during starvation, when free fatty acids are high, fat synthesis is inhibited. It is therefore not surprising that the amount of pyruvate kinase should be decreased by the same processes which decrease the amount of the fatty acid synthesizing enzymes. There are a number of regulatory enzymes of fatty acid, sterols, and isoprenoid syntheses that are regulated by AMPK, including acetyl-CoA carboxylase, HMG-CoA reductase, glycogen synthase, and hormone-sensitive lipase (8Hardie D.G. Biochim. Biophys. Acta. 1992; 1123: 231-238Crossref PubMed Scopus (166) Google Scholar, 9Hardie D.G. Carling D. Eur. J. Biochem. 1997; 246: 259-273Crossref PubMed Scopus (1155) Google Scholar, 10Moore F. Weekes J. Hardie D.G. Eur. J. Biochem. 1991; 199: 691-697Crossref PubMed Scopus (200) Google Scholar, 11Hardie D.G. Carling D. Carlson M. Annu. Rev. Biochem. 1998; 67: 821-855Crossref PubMed Scopus (1290) Google Scholar). Dyck et al. (19Dyck J.R. Kudo N. Barr A.J. Davies S.P. Hardie D.G. Lopaschuk G.D. Eur. J. Biochem. 1999; 262: 184-190Crossref PubMed Scopus (141) Google Scholar) reported that the activation of AMPK leads to inactivation of acetyl-CoA carboxylase by phosphorylation, blocking fatty acid synthesis, an energy-consuming process, and by decreasing intercellular malonyl-CoA, permitting activation of carnitine palmitoyl-CoA transferase I and fatty acid oxidation to generate energy. Furthermore, some transcription factors such as peroxisome proliferator-activated receptor γ appear to be regulated by AMPK (20Yang Y. Hong Y.H. Shen X. Frankowski C. Camp H.S. Leff T. J. Biol. Chem. 2001; 276: 38341-38344Abstract Full Text Full Text PDF PubMed Scopus (175) Google Scholar). Based on the observation that AICAR inhibits l-PK transcription, Vaulontet al. (21Vaulont S. Vasseur-Cognet M. Kahn A. J. Biol. Chem. 2000; 275: 31555-31558Abstract Full Text Full Text PDF PubMed Scopus (227) Google Scholar) proposed that AMPK is involved in the transcriptional regulation. More recently, they demonstrated that AICAR inhibits transcription of hepatocyte nuclear factor-4α in hepatocytes (22Leclerc I. Lenzner C. Gourdon L. Vaulont S. Kahn A. Viollet B. Diabetes. 2001; 50: 1515-1521Crossref PubMed Scopus (137) Google Scholar). Because hepatocyte nuclear factor-4α binds to the promoter adjacent to the glucose-responsive element of the l-PK gene (23Puzenat N. Vaulont S. Kahn A. Raymondjean M. Biochem. Biophys. Res. Commun. 1992; 189: 1119-1128Crossref PubMed Scopus (7) Google Scholar), it is regulated by AMPK. The results herein demonstrated that the phosphorylation by AMPK resulted in inhibition of the DNA binding activity of ChREBP. Thus, the AMPK inhibition of ChREBP is more direct regulation of l-PK gene transcription. Phosphorylation and dephosphorylation of transcription factors have wide-ranging effects such as DNA binding activity, transactivating capacity, and subcellular localization (22Leclerc I. Lenzner C. Gourdon L. Vaulont S. Kahn A. Viollet B. Diabetes. 2001; 50: 1515-1521Crossref PubMed Scopus (137) Google Scholar, 24Glineur C. Davioud C.E. Vandenbunder B. Biochem. J. 2000; 352: 583-591Crossref PubMed Scopus (33) Google Scholar, 25Katan K.Y. Shaul Y. Eur. J. Biochem. 2001; 268: 3108-3116Crossref PubMed Scopus (22) Google Scholar). ChREBP contains at least three phosphorylation sites for PKA, and two of the sites play important roles in cAMP-mediated inhibition of the glucose-activation of l-PK transcription (18Kawaguchi T. Takenoshita M. Kabashima T. Uyeda K. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 13710-13715Crossref PubMed Scopus (315) Google Scholar). These two PKA phosphorylation sites have different regulatory functions. The first site (Ser196), located near the NLS domain, is involved in regulation of nuclear localization of ChREBP, and phosphorylation of the site results in complete inhibition of the nuclear import (18Kawaguchi T. Takenoshita M. Kabashima T. Uyeda K. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 13710-13715Crossref PubMed Scopus (315) Google Scholar). The second site (Thr666) is located within the basic residues that are essential in DNA binding and phosphorylation of this site abolishes the DNA binding and the transcriptional activity (18Kawaguchi T. Takenoshita M. Kabashima T. Uyeda K. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 13710-13715Crossref PubMed Scopus (315) Google Scholar). The present study added an additional control site of ChREBP that is regulated by nutrients and mediated by AMPK. The phosphorylation of this site (Ser568) by AMPK resulted in inhibition of both the DNA binding activity and transcription of the l-PK gene. Thus, the present observation demonstrated that AMPK plays an important role in down-regulation of ChREBP by phosphorylation of Ser568 under hyperlipidemic conditions. However, because the S568A mutant was still able to mediate ∼45% inhibition of the glucose effect, whereas the wild-type ChREBP mediated 80% inhibition (Fig. 8C), there may be an additional regulatory mechanism of ChREBP under these conditions. Further studies will focus on these additional regulatory mechanisms. We thank Dr. Richard L. Veech and Dr. Sarah McIntire for critical reading.