acetylation in the regulation of gluconeogenesis - Nature

6 downloads 0 Views 536KB Size Report
CITED2 links hormonal signaling to PGC-1α acetylation in the regulation of gluconeogenesis. Mashito Sakai1,2, Michihiro Matsumoto1, Tomoko Tujimura1, Cao ...
letters

CITED2 links hormonal signaling to PGC-1α acetylation in the regulation of gluconeogenesis

npg

© 2012 Nature America, Inc. All rights reserved.

Mashito Sakai1,2, Michihiro Matsumoto1, Tomoko Tujimura1, Cao Yongheng1, Tetsuya Noguchi2, Kenjiro Inagaki1, Hiroshi Inoue3, Tetsuya Hosooka2, Kazuo Takazawa2, Yoshiaki Kido2,4, Kazuki Yasuda5, Ryuji Hiramatsu6, Yasushi Matsuki6 & Masato Kasuga7 During fasting, induction of hepatic gluconeogenesis is crucial to ensure proper energy homeostasis1. Such induction is dysregulated in type 2 diabetes, resulting in the development of fasting hyperglycemia2. Hormonal and nutrient regulation of metabolic adaptation during fasting is mediated predominantly by the transcriptional coactivator peroxisome proliferative activated receptor g coactivator 1a (PGC-1a) in concert with various other transcriptional regulators3–8. Although CITED2 (CBP- and p300-interacting transactivator with glutamic acid– and aspartic acid–rich COOH-terminal domain 2) interacts with many of these molecules9–11, the role of this protein in the regulation of hepatic gluconeogenesis was previously unknown. Here we show that CITED2 is required for the regulation of hepatic gluconeogenesis through PGC-1a. The abundance of CITED2 was increased in the livers of mice by fasting and in cultured hepatocytes by glucagon-cAMP–protein kinase A (PKA) signaling, and the amount of CITED2 in liver was higher in mice with type 2 diabetes than in non-diabetic mice. CITED2 inhibited the acetylation of PGC-1a by blocking its interaction with the acetyltransferase general control of amino acid synthesis 5–like 2 (GCN5). The consequent downregulation of PGC-1a acetylation resulted in an increase in its transcriptional coactivation activity and an increased expression of gluconeogenic genes. The interaction of CITED2 with GCN5 was disrupted by insulin in a manner that was dependent on phosphoinositide 3-kinase (PI3K)– thymoma viral proto-oncogene (Akt) signaling. Our results show that CITED2 functions as a transducer of glucagon and insulin signaling in the regulation of PGC-1a activity that is associated with the transcriptional control of gluconeogenesis and that this function is mediated through the modulation of GCN5-dependent PGC-1a acetylation. We also found that loss of hepatic CITED2 function suppresses gluconeogenesis in diabetic mice, suggesting it as a therapeutic target for hyperglycemia.

CITED2 is a transcriptional co-regulator that interacts with CREB binding protein (CBP) and the related protein p300 through their C termini and regulates gene transcription through modulation of the histone acetyltransferase activities of these proteins11. Loss-offunction studies in vivo revealed that CITED2 is required for normal embryonic development9,12–14 and adult hematopoiesis15. CITED2 also interacts with a wide variety of transcription factors, including hepatocyte nuclear factor 4a (HNF-4α)9, peroxisome proliferator– activated receptor α (PPAR-α)10, PPAR-γ and hypoxia-inducible factor 1α (HIF-1α)11 and thereby modulates gene transcription. Although some of the molecules targeted by CITED2 contribute to the regulation of hepatic gluconeogenesis5,16, the role of CITED2 in such regulation was previously unclear. As hepatic gluconeogenesis is increased by glucagon signaling in states of fasting, we deprived lean mice of food or administered glucagon by intraperitoneal injection. We found that the amount of CITED2 protein (Fig. 1a), but not of CITED2 mRNA (Supplementary Fig. 1a), increased in the livers of the fasted mice. We also found an increase in the amount of CITED2 protein in the livers of the glucagon-injected mice (Supplementary Fig. 1b). In addition, we also found that the hepatic abundance of both CITED2 protein and CITED2 mRNA was higher in two mouse models of type 2 diabetes (db/db mice, which are deficient for the leptin receptor, and obese mice fed a high-fat diet) compared to control mice (Fig. 1b). The amount of exogenous CITED2 in primary cultured mouse hepatocytes also increased as a result of exposure to glucagon or a cell-permeable analog of cAMP (pCPT-cAMP), which is a second messenger of glucagon signaling (Fig. 1c). This induction of CITED2 was PKA dependent and was mediated, at least in part, by inhibition of ubiquitination-dependent proteasomal degradation (Supplementary Fig. 1c,d). These results suggest that the abundance of CITED2 in the liver is increased by glucagon-cAMP-PKA signaling during fasting,

1Department

of Molecular Metabolic Regulation, Diabetes Research Center, Research Institute, National Center for Global Health and Medicine, Tokyo, Japan. of Internal Medicine, Division of Diabetes, Metabolism and Endocrinology, Kobe University Graduate School of Medicine, Kobe, Hyogo, Japan. 3Frontier Science Organization, Kanazawa University, Kanazawa, Ishikawa, Japan. 4Department of Biophysics, Division of Medical Chemistry, Kobe University Graduate School of Health Sciences, Kobe, Hyogo, Japan. 5Department of Metabolic Disorder, Diabetes Research Center, Research Institute, National Center for Global Health and Medicine, Tokyo, Japan. 6Pharmacology Research Laboratories, Dainippon Sumitomo Pharmaceuticals, Suita, Osaka, Japan. 7Diabetes Research Center, Research Institute, National Center for Global Health and Medicine, Tokyo, Japan. Correspondence should be addressed to M.M. ([email protected]) or M.K. ([email protected]). 2Department

Received 10 November 2011; accepted 27 January 2012; published online 18 March 2012; doi:10.1038/nm.2691

612

VOLUME 18 | NUMBER 4 | APRIL 2012  nature medicine

letters

npg

© 2012 Nature America, Inc. All rights reserved.

Relative CITED2 mRNA abundance

Relative CITED2 mRNA abundance

2.5 2.0 1.5 1.0 0.5 0 2.0

*

+

+

db/db IB: CITED2 IB: Histone H1

db/m db/db

*

1.5 1.0

Normal chow

0.5 0

db/m

Flag-CITED2: +

0 CITED2 shRNA: cAMP:

e

High fat

HF

IB: Histone H1

*

600 400 200 0

CITED2: cAMP: IB: CITED2

NC

800

–+ –

–+ –

100 0

–+ +

–+ – 400

*

200 100 0 –+ –

f

–+ +

300

–+ + 2,500 2,000 1,500

Glucose production (AU)

b

100

200

Glucose production (AU)

IB: Histone H1

IB: α-tubulin

200

*

300

4

*

*

3 2

*

1 0 –+ –

–+ +

*

15 10 5 0

–+ +

–+ – 1,000

Relative Pck1 mRNA abundance

IB: Flag

IB: CITED2

*

300

Relative Pck1 mRNA abundance

GLU cAMP

Relative Pck1 mRNA abundance

d –

Re-fed

Relative G6pc mRNA abundance

Fasted

Relative G6pc mRNA abundance

c Ad lib

Relative G6pc mRNA abundance

a

800

–+ +

*

600

Figure 1  CITED2 is upregulated by fasting and glucagon-cAMP and promotes 1,000 400 gluconeogenic gene expression. (a) Immunoblot analysis (IB) of CITED2 in nuclear 500 200 extracts prepared from the livers of C57BL/6J mice that had been allowed to feed 0 0 freely (ad lib), those deprived of food overnight (fasted) or those allowed to re-feed CITED2 (∆CR2): –+ –+ –+ –+ for 6 h after fasting. (b) Quantitative RT-PCR analysis of CITED2 mRNA and immunoblot cAMP: – + – + analysis of nuclear CITED2 in the livers of db/db or control db/m mice or of C57BL/6J mice maintained on normal chow (NC) or a high-fat diet (HF) and deprived of food for 3 h. (c) Immunoblot analysis of Flag-epitope–tagged CITED2 (Flag-CITED2) in primary cultured mouse hepatocytes exposed to glucagon (GLU), pCPT-cAMP (cAMP) or vehicle (−) for 6 h. (d,e) Effects of shRNAmediated depletion (d) or overexpression (e) of CITED2 on gluconeogenic gene expression and glucose production in primary hepatocytes incubated in the absence or presence of pCPT-cAMP for 6 h. (f) Effect of the ∆CR2 mutant of CITED2 on the gluconeogenic gene expression induced by pCPT-cAMP. All quantitative data are means ± s.e.m. n = 4 (d–f) or n = 6 (b). *P < 0.01 (Student’s unpaired t test).

­ rompting us to explore the role of CITED2 in the regulation of p hepatic gluconeogenesis in vitro. Depletion of CITED2 in primary mouse hepatocytes by infection with an adenoviral vector encoding an shRNA specific for mouse CITED2 mRNA attenuated the expression of the gluconeogenic genes G6pc (encoding the catalytic subunit of glucose-6-phosphatase) and Pck1 (encoding phosphoenolpyruvate carboxykinase), as well as the glucose production induced by pCPT-cAMP (Fig. 1d and Supplementary Fig. 1e). Conversely, forced expression of CITED2 augmented the effects of cAMP on gluconeogenic gene expression and glucose production (Fig. 1e and Supplementary Fig. 1f). CITED2 was previously shown to interact with CBP/p300 and HNF-4α through its conserved region 2 (CR2) and thereby modulate gene transcription9,11. We found that a mutant of CITED2 that lacks CR2 (∆CR2) and is unable to bind to CBP/p300 (ref. 11) or to HNF-4α9 also enhanced cAMP-dependent gluconeogenic gene expression to an extent similar to that observed with wild-type CITED2 (Fig. 1f and Supplementary Fig. 1g). This finding suggests that CITED2 regulates cAMP-induced gluconeogenic gene expression through a mechanism that is independent of its interaction with CBP, p300 or HNF-4α. We next examined the effect of CITED2 on hepatic gluconeogenesis in vivo. First we confirmed that the abundance of CITED2 in a nuclear extract from liver of lean mice increased gradually for up to 24 h during fasting, with the induction of G6pc and Pck1 expression showing a similar time course (Supplementary Fig. 2a). In lean mice, shRNA-mediated depletion of CITED2 in the liver resulted in reduced gluconeogenic gene expression in the liver and reduced blood glucose concentrations after fasting for 24 h and 6 h, respectively (Fig. 2a and Supplementary Fig. 2b). Consistent with these findings, mice infected with the adenovirus encoding CITED2 shRNA had lower blood glucose concentrations after administration of pyruvate than did control mice (Fig. 2b), indicative of reduced hepatic gluconeogenesis.

nature medicine  VOLUME 18 | NUMBER 4 | APRIL 2012

Knockdown of CITED2 in the livers of obese diabetic db/db mice significantly reduced gluconeogenic gene expression in the liver under fasting (24 h) and fed conditions as well as blood glucose concentrations under fasting (6 h) and fed conditions (Fig. 2c and Supplementary Fig. 2c,d). These results implicate CITED2 in the regu­lation of hepatic gluconeogenesis in vivo under both physio­ logical and pathological conditions. In contrast, overexpression of CITED2 in the livers of lean mice increased the hepatic expression of G6pc and Pck1 and increased blood glucose concentrations after fasting for 24 h (Fig. 2d and Supplementary Fig. 2e,f), and CITED2 overexpression also increased the glucose levels produced by the livers of lean mice after pyruvate administration (Fig. 2e). Together with the in vitro data shown in Figure 1, these results suggest that CITED2 has a key role in the induction of gluconeogenic gene expression in the liver of fasted mice. We then investigated the mechanism by which CITED2 upregulates gluconeogenic gene expression in the liver. Given that PGC-1α expression in the liver is induced by glucagon-cAMP signaling6,17 (in a manner that is dependent on CREB regulated transcription coactivator 2 (CRTC2, also known as TORC2)18) and that PGC-1α is a key transcriptional coactivator in the upregulation of gluconeogenic gene expression19, we examined the possible role of PGC-1α in the upregulation of hepatic gluconeogenesis mediated by CITED2. Consistent with such a role, we found that shRNA-mediated depletion of PGC-1α abolished the effect of CITED2 overexpression on the cAMP-dependent induction of gluconeogenic gene expression in primary hepatocytes isolated from the livers of lean C57BL/6J mice (Online Methods, Fig. 3a and Supplementary Fig. 3a,b). We therefore examined the effect of CITED2 on PGC-1α–dependent induction of gluconeogenic gene expression. Consistent with previous observations6, adenovirus-mediated overexpression of PGC-1α in primary cultured hepatocytes induced G6pc and Pck1 expression in the absence

613

letters

50 0

Fed

Fasted

80 60

** ** ** **

40

Control CITED2 KD

20 0

0 30 60 90 120 Time (min)

1.0 0.5 0

**

–1

–1

0.5 0

**

300

* *

200 100 0

Fed

Fasted

of pCPT-cAMP (Fig. 3b and Supplementary Fig. 3c). Furthermore, the effect of PGC-1α on gluconeogenic gene expression was enhanced by CITED2 overexpression (Fig. 3b and Supplementary Fig. 3c). Forced expression of CITED2 by adenovirus vector in the absence of PGC-1α resulted in a small but significant increase in gluconeogenic gene expression in a manner that was dependent on the concentration of virus-derived CITED2 (Fig. 3b and Supplementary Fig. 3c), an effect that might be mediated by activation of endogenous PGC-1α. We also examined the effect of CITED2 on PGC-1α–dependent transcriptional coactivation. Consistent with previous observations5, PGC-1α increased HNF-4α–dependent G6PC promoter activity in HEK293 cells (a human cell line) (Fig. 3c). This effect of PGC-1α was enhanced by the coexpression of PGC-1α with either wild-type or ∆CR2 mutant forms of CITED2 (Fig. 3c), which suggests a direct effect of CITED2 on PGC-1α rather than on HNF-4α in this setting. A direct effect of CITED2 on PGC-1α–dependent gene expression was further suggested by the observation that shRNA-mediated knockdown of CITED2 markedly attenuated the PGC-1α–mediated induction of G6pc and Pck1 expression in primary hepatocytes (Fig. 3d and Supplementary Fig. 3d). shRNA-mediated knockdown of CITED2 in hepatocytes also attenuated the expression of Ppara (encoding PPAR-α) and Cpt1a (encoding carnitine palmitoyltransferase 1A), two PGC-1α target genes that contribute to fatty acid oxidation, under basal, PGC-1α– induced (Supplementary Fig. 3e) and pCPT-cAMP–induced conditions (Supplementary Fig. 3f). Conversely, CITED2 overexpression enhanced the induction of these genes by cAMP (Supplementary Fig. 3g). Overexpression of CITED2 also enhanced the expression of another class of PGC-1α target genes that are involved in mitochondrial oxidative phosphorylation (Supplementary Fig. 3h). Together, these results indicate that CITED2 activates PGC-1α and thereby promotes hepatic gene expression related to gluconeogenesis, fatty acid oxidation and oxidative phosphorylation. We then investigated how CITED2 enhances PGC-1α activity. The activity of PGC-1α is regulated by various post-translational modifications, including acetylation, phosphorylation, methylation and small ubiquitin-like modifier (SUMO)ylation20. In particular,

614

Glucose (mg dl–1)

Glucose (mg dl–1)

e

Relative Pck1 mRNA abundance

Relative G6pc mRNA abundance

© 2012 Nature America, Inc. All rights reserved.

1.0

400

Figure 2  CITED2 increases hepatic glucose production in vivo β-Gal β-Gal through the upregulation of gluconeogenic gene expression in the 150 CITED2 CITED2 ** ** liver. (a) Effect of shRNA-mediated knockdown (KD) of CITED2 in 4 80 4 ** ** 100 the livers of C57BL/6J mice on gluconeogenic gene expression in ** 3 60 3 the liver under the fasted condition (24 h) as well as on blood ** 50 2 40 β-Gal 2 glucose concentration under fasted (6 h) and fed conditions (n = 7 CITED2 for each group). (b) Pyruvate tolerance test in C57BL/6J mice 1 20 1 0 expressing CITED2 (n = 11) or control shRNAs (n = 10) in the liver. 0 30 60 90 120 0 0 0 (c) Effect of shRNA-mediated knockdown of CITED2 on hepatic Time (min) G6pc and Pck1 expression under the fasted condition (24 h) as well as on blood glucose concentration under fasted (6 h) and fed conditions in insulin-resistant db/db mice (n = 7 for each group). (d) Effect of forced expression of CITED2 or β-galactosidase (β-Gal) in the liver on hepatic G6pc and Pck1 expression as well as on blood glucose concentration in C57BL/6J mice under the fasted condition for 24 h (n = 6 for each group). (e) Pyruvate tolerance test in C57BL/6J mice expressing ectopic CITED2 or β-galactosidase (n = 8 for each group). All data are means ± s.e.m. *P < 0.05, **P < 0.01 compared to the corresponding control value or for the indicated comparisons (Student’s unpaired t test).

d

npg

1.5 Glucose (mg dl )

0

**

100

1.5

Relative Pck1 mRNA abundance

0

0.5

*

100

Control CITED2 KD

Control CITED2 KD

Relative G6pc mRNA abundance

0.5

**

1.0

c 120

150

Glucose (mg dl )

1.0

1.5 Glucose (mg dl–1)

1.5

b

Control CITED2 KD

Control CITED2 KD Relative Pck1 mRNA abundance

Relative G6pc mRNA abundance

a

acetylation of PGC-1α is regulated in response to energy status by the NAD+-dependent protein deacetylase sirtuin 1 (SIRT1), which deacytelates PGC-1α and thus activates it, and by the acetyltransferase GCN5, which acetylates PGC-1α and thus inactivates it21–23. We found that overexpression of CITED2 inhibited the acetylation of PGC-1α, whereas CITED2 knockdown promoted this modification both in cultured hepatocytes (Fig. 4a) and in the livers of lean C57BL/6J (Fig. 4b) or obese db/db (Supplementary Fig. 4a) mice. In livers of lean mice, the amounts of CITED2 and PGC-1α and the amount of acetylation of PGC-1α underwent opposite changes in response to a feeding-fasting–re-feeding regimen (Fig. 4c), consistent with the notion that CITED2 attenuates PGC-1α acetylation. These observations suggest that CITED2 suppresses the acetylation of PGC-1α and thereby upregulates its activity via an effect on SIRT1 or on GCN5 or on both. Thus, we tested whether CITED2 is capable of direct interaction with PGC-1α, SIRT1 or GCN5. A cellbased coimmunoprecipitation analysis revealed that CITED2 interacted with neither PGC-1α nor SIRT1 (Supplementary Fig. 4b,c). In addition, overexpression of CITED2 did not affect the interaction of PGC-1α with SIRT1, the cellular ratio of NAD+ to NADH or the acetylation of FoxO1, another deacetylation target of SIRT1, in vitro or in vivo (Supplementary Fig. 4d–g). These results indicate that CITED2 does not affect the SIRT1-mediated deacetylation of PGC-1α. However, immunoprecipitation of Flag-tagged CITED2 from transfected HEK293 cells resulted in the co-precipitation of cotransfected GCN5 containing a v-myc myelocytomatosis viral oncogene homolog (Myc)-epitope tag and vice versa (Fig. 4d). We also detected such coimmunoprecipitation for endogenous GCN5 and Flag-tagged CITED2 in AML12 hepatoma cells (a murine cell line) (Fig. 4e) and in lean mouse livers expressing exogenous CITED2 at physiological levels (Fig. 4e and Supplementary Fig. 4h). These data indicate that CITED2 interacts with GCN5. We also tested whether CITED2 affects the PGC-1α acetylation that is mediated by GCN5. Consistent with previous observations23, forced expression of GCN5 increased the acetylation of PGC-1α in HEK293 cells (Fig. 4f), and we further found that adenoviral coexpression of CITED2 with PGC-1α and GCN5 both inhibited

VOLUME 18 | NUMBER 4 | APRIL 2012  nature medicine

letters

100

0 CITED2:

PGC-1α shRNA: cAMP:

300 200 100

150 100 50

–+

–+

–+

–+

–+

–+

–+

–+

0 CITED2:



+



+



+



+

PGC-1α:



+

0



**

**

15

*

10 5

**

Luciferase activity (AU)

200

**

400

Relative G6pc mRNA abundance

Relative Pck1 mRNA abundance

**

300

c Relative Pck1 mRNA abundance

b Relative G6pc mRNA abundance

a

hG6PC-luc

** **

20 15 10

**

5

– + ++

– +

– + ++

– +

0 HNF-4α:

– – – + + + +



+



+

PGC-1α:

– + – – + + +

CITED2:

– – + – – + –

CITED2 (∆CR2):

– – – – – – +

0

+

d

npg

© 2012 Nature America, Inc. All rights reserved.

GCN5-dependent acetylation of PGC-1α and reduced the amount of GCN5 that coimmunoprecipitated with PGC-1α in a manner dependent on the dose of virally expressed CITED2 (Fig. 4f). In primary hepatocytes, the enhancement of PGC-1α–dependent gluconeogenic gene expression by exogenous CITED2 was abolished by GCN5 over­ expression (Fig. 4g), whereas increased PGC-1α acetylation as a result of CITED2 knockdown was attenuated by shRNA-mediated depletion of GCN5 (Supplementary Fig. 4i). These data suggest that GCN5 is targeted by CITED2 in the reduction of PGC-1α acetylation followed by the enhancement of PGC-1α–dependent gluconeogenic gene expression. However, the activity of GCN5, as assessed by a histone acetyltransferase assay in vitro with histone H3 as substrate, was not inhibited but, rather, was enhanced by CITED2 overexpression (Supplementary Fig. 4j), supporting the notion that CITED2 binds to GCN5 and thereby interferes with the physical interaction between GCN5 and PGC-1α. We also investigated whether CITED2 affects the subcellular localization of PGC-1α and GCN5. Consistent with previous observations23, PGC-1α and GCN5 colocalized in a distinct punctate pattern in the nuclei of transfected HEK293 cells (Supplementary Fig. 5a). Coexpression of CITED2 with both PGC-1α and GCN5 interrupted the colocalization of the latter two proteins to nuclear foci (Supplementary Fig. 5a), reflecting an impaired action of GCN5 (ref. 23) and further supporting the notion that CITED2 interacts with and interferes with the action of GCN5. To identify the protein domains that are required for the interaction between CITED2 and GCN5, we performed a coimmunoprecipitation assay with a series of deletion mutants of proteins (Supplementary Fig. 5b,c). We found both the CR1 and serine-glycine–rich junction (SRJ) domains of CITED2 and the region spanning amino acids 333–542 of GCN5, which is immediately downstream of the p300/ CBP-associated factor (PCAF) homology domain, to be required for the CITED2-GCN5 interaction (Fig. 4h and Supplementary Fig. 5b,c). Consistent with this finding, expression of the ∆SRJ mutant of CITED2, which lacks the SRJ domain, neither enhanced the cAMPinduced expression of gluconeogenic genes nor inhibited the GCN5dependent acetylation of PGC-1α (Supplementary Fig. 5d,e). We next investigated whether insulin affects the functional interaction between CITED2 and GCN5. In HEK293 and AML12 cells

nature medicine  VOLUME 18 | NUMBER 4 | APRIL 2012

Relative Pck1 mRNA abundance

Relative G6pc mRNA abundance

Figure 3  CITED2 enhances PGC-1α coactivation activity in primary hepatocytes. (a) Primary ** 30 ** 25 mouse hepatocytes infected with adenoviruses encoding PGC-1α shRNA or CITED2, as 20 20 indicated, were incubated in the absence or presence of pCPT-cAMP for 6 h and then assayed 15 for G6pc and Pck1 expression (n = 4 for each group). (b) Gluconeogenic gene expression 10 10 in primary mouse hepatocytes infected with adenoviruses encoding CITED2 (at two different 5 multiplicities of infection (MOI), indicated by ‘+’ (MOI = 3) and ‘++’ (MOI = 6)) or PGC-1α 0 0 (n = 5 for each group). (c) Luciferase reporter assay of PGC-1α transcriptional coactivation CITED2 shRNA: – + – + – + – + activity in HEK293 cells transfected with a G6PC-promoter luciferase reporter plasmid PGC-1α: – + – + (hG6PC-luc) and expression plasmids for the indicated proteins (n = 3 for each group). (d) Gluconeogenic gene expression in primary hepatocytes infected with adenoviruses encoding CITED2 shRNA or PGC-1α, as indicated (n = 5 for each group). All data are means ± s.e.m. *P < 0.05, **P < 0.01 (Student’s unpaired t test).

expressing CITED2 with and without exogenous GCN5, respectively, the amount of CITED2 that coimmunoprecipitated with GCN5 was reduced by 80% after exposure of the cells to insulin for 30 min, whereas exposure to insulin did not affect the overall abundance of CITED2 or GCN5 in either type of cell (Fig. 4i,e and Supplementary Fig. 5f). We also found this effect of insulin in lean mouse livers expressing physiological levels of Flag-tagged CITED2 (Fig. 4e). In vivo treatment with glucagon or exposure of AML12 cells to pCPTcAMP in vitro increased the amount of CITED2 that coimmuno­ precipitated with GCN5 (Fig. 4e). The effect of insulin on the interaction between CITED2 and GCN5 was abolished by exposure of the cells to the PI3K inhibitor LY294002 or an Akt inhibitor but not by treatment with the mammalian target of rapamycin (mTOR) inhibitor rapamycin (Fig. 4i). Furthermore, insulin inhibited the effects of CITED2 both on the GCN5-dependent acetylation of PGC-1α and on the amount of GCN5 that coimmunoprecipitated with PGC-1α (Fig. 4j). These data suggest that disruption of the CITED2-GCN5 complex by the insulin-PI3K-Akt signaling axis promotes the GCN5dependent acetylation and inactivation of PGC-1α. Two insulin-activated kinases, Akt and Cdc2-like kinase 2 (Clk2), the latter of which is induced by Akt activation, directly phosphorylate PGC-1α and suppress its coactivation activity24,25. We therefore examined whether the CITED2-GCN5 interaction is affected by the phosphorylation of PGC-1α by Akt or Clk2 using two mutant forms of PGC-1α: ∆SR, which lacks the serine-arginine (SR) domain containing serine and threonine residues that are phosphorylated by Akt or Clk2 (ref. 25), and S570A, in which Ser570 has been replaced with alanine and which is not phosphorylated by Akt24. Insulin failed to suppress the effects of CITED2 on the GCN5-dependent acetylation of either PGC-1α mutant or on the amount of GCN5 that coimmuno­ precipitated with the mutants (Supplementary Fig. 5g). These data indicate that phosphorylation of PGC-1α at Ser570 by Akt is required for the restoration of the GCN5–PGC-1α complex by insulin. Glucose homeostasis under conditions of a fluctuating food supply is maintained by strict regulation of hepatic gluconeogenesis through the counter-regulatory actions of the pancreatic hormones glucagon (released during fasting) and insulin (released during feeding). We have now shown that CITED2 functions as a transducer of signaling by these pancreatic hormones by regulating the activity of

615

IP: Flag IB: PGC-1α IP: Flag IB: Ac-Lys

b

IB: CITED2 Flag–PGC-1α: HA-CITED2:

– –

+ –

β-Gal

shRNA: Control

CITED2

IP: PGC-1α IB: PGC-1α IP: CITED2 IB: CITED2

IP: GCN5 IB: GCN5 IP: GCN5 IB: Flag

© 2012 Nature America, Inc. All rights reserved.

IB: Histone H1

d Input

Myc-GCN5: –

+



+

Flag-CITED2: +

+

+

+

Flag-CITED2: +

+

+

+

+

+

+

+

+

– –

+ –

+ +

– +

i

IB: GCN5 IB: Flag

Flag-CITED2: Glucagon:

+ –

+ –

+ +

+ +

Insulin:







+

+ + – –

+ + + –

+ + + +

+ + + –

Akt inhibitor: –







+

+

+

+

+

Myc-GCN5: +



+

+

+

Flag-CITED2: –





+ +++

IP: Myc IB: Myc IP: Myc IB: Flag IB: Flag

IB: Akt

IB: p70S6K +

+

+

+

+

+

Insulin: –



+

Rapamycin: –





+ – +

*

60

+ + +

*

*

40 20

CR3

+ – –

+ + –

+ – +

+ + +

CR1 SRJ CR2

GCN5 PCAF HD

AT Bromo

IP: T7 IB: PGC-1α IP: T7 IB: Myc IP: T7 IB: Ac-Lys IB: Myc

IB: Phospho-p70S6K Flag-CITED2: + Myc-GCN5: –

+ + –

CITED2

j

IP: Myc IB: Myc IP: Myc IB: Flag IB: Flag

+ – –

0 PGC-1α: – GCN5: – CITED2: –

h

*

*

10

80

IB: Flag T7–PGC-1α: –

20

IB: Myc

IP: T7 IB: PGC-1α IP: T7 IB: Myc IP: T7 IB: Ac-Lys IB: Myc

IP: GCN5

IP: IgG

IB: Phospho-Akt + – – –

f

*

30

IB: Flag

Input IgG Flag IB: Flag Myc-GCN5: +

40

0 PGC-1α: – GCN5: – CITED2: –

IP

IP: Myc IB: GCN5

IB: Histone H1

IB: CITED2

npg

IB: CITED2

+ +

e

Flag-CITED2: Myc-GCN5: Insulin: LY294002:

g

Re-fed

IP: PGC-1α IB: Ac-Lys

IP: PGC-1α IB: Ac-Lys

IP: Flag IB: Ac-Lys

Flag-CITED2: – cAMP: – Insulin: –

Fasted

IP: PGC-1α IB: PGC-1α

Flag-CITED2

IP: Flag IB: PGC-1α

Flag–PGC-1α: + CITED2 shRNA: –

Ad lib

IB: Flag

+ +

IB: CITED2

c

IP: PGC-1α IB: Ac-Lys IP: PGC-1α IB: PGC-1α

Relative Pck1 mRNA abundance

a

Relative G6pc mRNA abundance

letters

IB: Flag

+

T7–PGC-1α: – Myc-GCN5: + Flag-CITED2: –

+ – –

+ + –

+ + +

+ + +

+

Insulin: –







+

Figure 4  CITED2 inhibits GCN5-dependent PGC-1α acetylation by interfering with the PGC-1α–GCN5 interaction in an insulin-sensitive manner. (a) Immunoblot analysis with the indicated antibodies of the precipitates after immunoprecipitation (IP) with antibodies to Flag from AML12 cells (top) or primary mouse hepatocytes (bottom) expressing Flag-tagged PGC-1α (Flag–PGC-1α), hemagglutinin-epitope (HA)–tagged CITED2 (HA-CITED2) or CITED2 shRNA. A precipitate from control-virus–infected cells (left lane) was included as a control in the AML12 cell experiment. Ac-Lys, acetylated lysine. (b) PGC-1α acetylation in liver lysates from C57BL/6J mice injected with adenoviruses encoding either Flag-CITED2 or β-galactosidase (top) or control or CITED2 shRNAs (bottom). (c) PGC-1α acetylation in liver nuclear extracts from mice fed ad libitum (ad lib), mice fasted for 24 h or mice re-fed for 6 h after fasting. (d) Immunoblot analysis of precipitates (or cell lysates, Input) from HEK293 cells expressing Myc-epitope–tagged GCN5 (Myc-GCN5) and Flag-CITED2 after immunoprecipitation with antibodies to Myc or Flag (or with control immunoglobulin G (IgG)). (e) Immunoprecipitation-immunoblot analysis of AML12 cells expressing Flag-CITED2 and incubated with pCPT-cAMP (12 h) and then also with insulin (30 min) (left). Immunoprecipitation-immunoblot analysis of liver nuclear extracts from mice infected with adenovirus encoding Flag-CITED2 that were deprived of food and injected with glucagon, followed by insulin injection after immunoprecipitation with antibodies to GCN5 (or with control IgG) (right). (f) Immunoprecipitation-immunoblot analysis of HEK293 cells expressing T7-epitope–tagged PGC-1α (T7–PGC-1α), Myc-GCN5 or Flag-CITED2. (g) Gluconeogenic gene expression (means ± s.e.m., n = 5) in primary hepatocytes infected with adenoviruses encoding PGC-1α, GCN5 or CITED2. *P < 0.01 (Student’s unpaired t test). (h) Domain organization of CITED2 and GCN5 indicating the regions (arrows) that mediate their interaction. AT, acetyltransferase domain; HD, homology domain. (i) Immunoprecipitation-immunoblot analysis of HEK293 cells expressing Flag-CITED2 or Myc-GCN5 that had been incubated with LY294002, Akt inhibitor or rapamycin (30 min) and then also with insulin (30 min). Phospho-Akt, phosphorylated Akt; phospho-p70S6K, phosphorylated 70-kDa ribosomal protein S6 kinase. (j) Immunoprecipitation-immunoblot analysis of HEK293 cells expressing the indicated proteins and incubated with or without insulin (30 min).

PGC-1α through modulation of its GCN5-dependent acetylation. CITED2 exerts this action through a mechanism that is independent of its interaction with CBP/p300. Although abundant evidence suggests that GCN5 functions as an acetyltransferase for PGC-1α and thereby suppresses its activity, the signal that controls this process was unclear26,27. Our data now indicate that the opposing signaling pathways triggered by glucagon and insulin converge on CITED2 and thereby regulate the GCN5-dependent acetylation of PGC-1α. The upregulation of both CITED2 and PGC-1α expression by glucagon-cAMP-PKA signaling during fasting probably maximizes the increase in the amount of active PGC-1α in response to fasting, resulting in full activation of the gluconeogenic gene program. However, insulin-induced disruption of the CITED2-GCN5 complex

616

in collaboration with the Akt- and Clk2-mediated phosphorylation of PGC-1α that is elicited by this hormone24,25 seems to contribute to the acute suppression of PGC-1α activity and the consequent termination of the fasting-induced gluconeogenic program (Supplementary Fig. 6). In mice with obesity and type 2 diabetes, CITED2 expression in the liver was upregulated, and shRNA-mediated knockdown of CITED2 improved glycemia through the suppression of gluconeogenesis, suggesting that CITED2 may contribute to the pathophysio­ logy of diabetes in this setting. Investigation of whether dysregulation of CITED2 in the liver contributes to insulin resistance and type 2 diabetes in humans is warranted. If this is found to be the case, pharmacological targeting of CITED2 may represent a potential option to treat these conditions.

VOLUME 18 | NUMBER 4 | APRIL 2012  nature medicine

letters Methods Methods and any associated references are available in the online version of the paper at http://www.nature.com/naturemedicine/. Note: Supplementary information is available on the Nature Medicine website. Acknowledgments We thank D. Schmoll (Sanofi-Aventis Deutschland GmbH) for the G6PCpromoter reporter plasmid, H. Shimano (University of Tsukuba) for the pcDNA3.1-3 × Flag-HNF-4α plasmid and H. Takamoto for technical assistance. This work was supported by a Grant-in-Aid for Creative Scientific Research (to M.K.) and a Grant-in-Aid for Scientific Research (C) (21591155 to M.M.) from the Ministry of Education, Culture, Sports, Science and Technology of Japan, a grant from the National Center for Global Health and Medicine (21S116 to M.M.), a grant from Takeda Science Foundation (to M.M.) and a Novo Nordisk Pharma Insulin Award (to M.M.).

COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. Published online at http://www.nature.com/naturemedicine/. Reprints and permissions information is available online at http://www.nature.com/ reprints/index.html. 1. Cahill, G.F. Jr. Fuel metabolism in starvation. Annu. Rev. Nutr. 26, 1–22 (2006). 2. Biddinger, S.B. & Kahn, C.R. From mice to men: insights into the insulin resistance syndromes. Annu. Rev. Physiol. 68, 123–158 (2006). 3. Puigserver, P. et al. Insulin-regulated hepatic gluconeogenesis through FOXO1–PGC-1α interaction. Nature 423, 550–555 (2003). 4. Matsumoto, M., Pocai, A., Rossetti, L., Depinho, R.A. & Accili, D. Impaired regulation of hepatic glucose production in mice lacking the forkhead transcription factor Foxo1 in liver. Cell Metab. 6, 208–216 (2007). 5. Rhee, J. et al. Regulation of hepatic fasting response by PPARγ coactivator-1α (PGC-1): requirement for hepatocyte nuclear factor 4α in gluconeogenesis. Proc. Natl. Acad. Sci. USA 100, 4012–4017 (2003). 6. Yoon, J.C. et al. Control of hepatic gluconeogenesis through the transcriptional coactivator PGC-1. Nature 413, 131–138 (2001). 7. Puigserver, P. et al. Activation of PPARγ coactivator-1 through transcription factor docking. Science 286, 1368–1371 (1999).

npg

© 2012 Nature America, Inc. All rights reserved.

AUTHOR CONTRIBUTIONS M.M., T.N. and M.K. conceived of the experiments, and M.S., M.M. and M.K. designed the experiments. M.S., M.M., T.T., C.Y., K.I., H.I., T.H., K.T., Y.K., K.Y., R.H. and Y.M. performed the experiments. M.S., M.M. and M.K. interpreted the data and wrote the manuscript. M.M. and M.K. supervised the study.

8. Matsumoto, M. & Accili, D. The tangled path to glucose production. Nat. Med. 12, 33–34 (2006). 9. Qu, X. et al. Cited2, a coactivator of HNF4α, is essential for liver development. EMBO J. 26, 4445–4456 (2007). 10. Tien, E.S., Davis, J.W. & Vanden Heuvel, J.P. Identification of the CREB-binding protein/p300-interacting protein CITED2 as a peroxisome proliferator-activated receptor α coregulator. J. Biol. Chem. 279, 24053–24063 (2004). 11. Bhattacharya, S. et al. Functional role of p35srj, a novel p300/CBP binding protein, during transactivation by HIF-1. Genes Dev. 13, 64–75 (1999). 12. Bamforth, S.D. et al. Cardiac malformations, adrenal agenesis, neural crest defects and exencephaly in mice lacking Cited2, a new Tfap2 co-activator. Nat. Genet. 29, 469–474 (2001). 13. Yin, Z. et al. The essential role of Cited2, a negative regulator for HIF-1α, in heart development and neurulation. Proc. Natl. Acad. Sci. USA 99, 10488–10493 (2002). 14. Chen, Y., Haviernik, P., Bunting, K.D. & Yang, Y.C. Cited2 is required for normal hematopoiesis in the murine fetal liver. Blood 110, 2889–2898 (2007). 15. Kranc, K.R. et al. Cited2 is an essential regulator of adult hematopoietic stem cells. Cell Stem Cell 5, 659–665 (2009). 16. Zhou, X.Y. et al. Insulin regulation of hepatic gluconeogenesis through phosphorylation of CREB-binding protein. Nat. Med. 10, 633–637 (2004). 17. Herzig, S. et al. CREB regulates hepatic gluconeogenesis through the coactivator PGC-1. Nature 413, 179–183 (2001). 18. Koo, S.H. et al. The CREB coactivator TORC2 is a key regulator of fasting glucose metabolism. Nature 437, 1109–1111 (2005). 19. Lin, J., Handschin, C. & Spiegelman, B.M. Metabolic control through the PGC-1 family of transcription coactivators. Cell Metab. 1, 361–370 (2005). 20. Rodgers, J.T., Lerin, C., Gerhart-Hines, Z. & Puigserver, P. Metabolic adaptations through the PGC-1α and SIRT1 pathways. FEBS Lett. 582, 46–53 (2008). 21. Rodgers, J.T. et al. Nutrient control of glucose homeostasis through a complex of PGC-1α and SIRT1. Nature 434, 113–118 (2005). 22. Nemoto, S., Fergusson, M.M. & Finkel, T. SIRT1 functionally interacts with the metabolic regulator and transcriptional coactivator PGC-1α. J. Biol. Chem. 280, 16456–16460 (2005). 23. Lerin, C. et al. GCN5 acetyltransferase complex controls glucose metabolism through transcriptional repression of PGC-1α. Cell Metab. 3, 429–438 (2006). 24. Li, X., Monks, B., Ge, Q. & Birnbaum, M.J. Akt/PKB regulates hepatic metabolism by directly inhibiting PGC-1α transcription coactivator. Nature 447, 1012–1016 (2007). 25. Rodgers, J.T., Haas, W., Gygi, S.P. & Puigserver, P. Cdc2-like kinase 2 is an insulin-regulated suppressor of hepatic gluconeogenesis. Cell Metab. 11, 23–34 (2010). 26. Jeninga, E.H., Schoonjans, K. & Auwerx, J. Reversible acetylation of PGC-1: connecting energy sensors and effectors to guarantee metabolic flexibility. Oncogene 29, 4617–4624 (2010). 27. Dominy, J.E. Jr., Lee, Y., Gerhart-Hines, Z. & Puigserver, P. Nutrient-dependent regulation of PGC-1α’s acetylation state and metabolic function through the enzymatic activities of Sirt1/GCN5. Biochim. Biophys. Acta 1804, 1676–1683 (2010).

nature medicine  VOLUME 18 | NUMBER 4 | APRIL 2012

617

ONLINE METHODS

npg

© 2012 Nature America, Inc. All rights reserved.

Mice. All mouse experiments were performed according to procedures approved by the Institutional Animal Care and Use Committee of the National Center for Global Health and Medicine. C57BL/6J male mice (Charles River Laboratories) as well as db/db (C57BLKS/J Iar-+Leprdb/+Leprdb) and db/m male mice (Institute for Animal Reproduction) were studied at 8 weeks of age. Recombinant adenoviruses were injected into the tail veins of mice as described previously28 at 1.0 × 109 or 3.0 × 109 plaque-forming units for C57BL/6J or db/db mice, respectively. Experiments were performed 4 d after adenovirus injection. Blood glucose concentrations were measured with the use of a standard glucose sensor (Glutest Ace, Sanwa Kagaku Kenkyusho). Pyruvate tolerance tests were performed as described previously29. In experiments studying mice fed on a high-fat diet, C57BL/6J mice were fed chow containing 30% fat by weight (14% bovine fat, 14% porcine fat and 2% soybean oil) from 4–24 weeks of age. In experiments investigating the effects of pancreatic hormones, mice were deprived of food for 3 h, intraperitoneally injected with 100 µg kg−1 of glucagon and, after 30 min, injected with 0.75 U kg−1 of insulin. Liver extracts were prepared 30 min after insulin treatment. Plasmids. CITED2, PGC-1α, GCN5 and SIRT1 complementary DNAs (cDNAs) were cloned from liver cDNA from C57BL/6J mice into the mammalian expression vectors pcDNA3 or pcDNA3.1 (Invitrogen). The cDNAs for the deletion mutants of CITED2, GCN5 and PGC-1α, as well as PGC-1α S570A, were generated with the use of a KOD-plus Mutagenesis Kit (Toyobo). A luciferase reporter plasmid containing the human G6PC promoter (pGL3hG6PC-luc, including nucleotides −1,227 to +57 of G6PC relative to the transcription start site) was described previously30. Adenoviruses. All recombinant adenoviruses were constructed with the use of an Adenovirus Dual Expression Kit (Takara Bio). Flag-CITED2, HA-CITED2, Flag-CITED2 (∆CR2), Flag-CITED2 (∆SRJ), Flag–PGC-1α and β-galactosidase were expressed under the control of a CAG promoter (cytomegalovirus enhancer, chicken β-actin promoter and rabbit β-globin poly(A) signal), whereas shRNAs were expressed under the control of a U6 promoter. CITED2 and PGC-1α shRNAs were based on the sequences 5′-TGACGGACTTCGTGTGCA-3′ and 5′-GTATCTGACCACAAACGAT-3′, respectively. A negative control shRNA sequence (BD Biosciences) was used as a control. Primary hepatocytes or AML12 cells were infected with adenoviruses 2 d after plating. The analyses of gene and protein expression and the glucose production assay were performed 2 d after infection. Cell culture. Primary hepatocytes were isolated from 8- to 12-week-old male C57BL/6J mice as described previously28 and were incubated overnight in serum-free Medium 199 (Invitrogen) before the addition of either glucagon (100 nM) or pCPT-cAMP (100 µM)31. AML12 mouse hepatoma and HEK293 cells were obtained from American Type Culture Collection. AML12 cells

nature medicine

were cultured in a 1:1 (v/v) mixture of DMEM and Ham’s F12 medium that was supplemented with insulin (5 µg ml−1), transferrin (5 µg ml−1), selenium (5 ng ml−1), dexamethasone (40 ng ml−1) and 10% FBS, whereas HEK293 cells were cultured in DMEM supplemented with 10% FBS. Where indicated, HEK293 cells were exposed to 30 µM LY294002, 3 µM Akt1/2 kinase inhibitor or 0.2 µM rapamycin for 30 min before incubation in the additional presence of 100 nM insulin for 30 min. Protein interaction analysis. Protein-protein interactions in cells were examined with coimmunoprecipitation assays. Epitope-tagged proteins were expressed in HEK293 cells by transient transfection for 24 h with the corresponding plasmids and with the use of Lipofectamine 2000 (Invitrogen). The cells were then lysed in a lysis buffer containing 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.5% Nonidet P-40, 2 mM EDTA, 10% glycerol, 10 mM nicotinamide, 1 µM trichostatin A, 10 µM MG132 and protease and phosphatase inhibitors (Roche), and the lysates were subjected to immunoprecipitation with the indicated antibodies and protein G–Sepharose (GE Healthcare). The immunoprecipitates were fractionated by SDS-PAGE and subjected to immuno­ blot analysis with the indicated antibodies. Glucose production assay. Primary hepatocytes were cultured in serum-free Medium 199 in the absence or presence of 100 µM pCPT-cAMP for 16 h as previously described31. They were then incubated for 6 h in glucose-free and phenol-red–free DMEM (pH 7.4) supplemented with sodium lactate and pyruvate before the measurement of glucose released into the medium with the use of a colorimetric assay (Wako)4. Data are presented in AU. Statistical analyses. Data are presented as means ± s.e.m. and were analyzed with Student’s unpaired t test with the use of GraphPad Prism software. P < 0.05 was considered statistically significant. Additional methods. Detailed methodology is described in the Supplementary Methods. 28. Matsumoto, M., Han, S., Kitamura, T. & Accili, D. Dual role of transcription factor FoxO1 in controlling hepatic insulin sensitivity and lipid metabolism. J. Clin. Invest. 116, 2464–2472 (2006). 29. Miyake, K. et al. Hyperinsulinemia, glucose intolerance, and dyslipidemia induced by acute inhibition of phosphoinositide 3-kinase signaling in the liver. J. Clin. Invest. 110, 1483–1491 (2002). 30. Schmoll, D., Grempler, R., Barthel, A., Joost, H.G. & Walther, R. Phorbol esterinduced activation of mitogen-activated protein kinase/extracellular-signal-regulated kinase kinase and extracellular-signal-regulated protein kinase decreases glucose-6phosphatase gene expression. Biochem. J. 357, 867–873 (2001). 31. Matsumoto, M. et al. Role of the insulin receptor substrate 1 and phosphatidylinositol 3-kinase signaling pathway in insulin-induced expression of sterol regulatory element binding protein 1c and glucokinase genes in rat hepatocytes. Diabetes 51, 1672–1680 (2002).

doi:10.11038/nm.2691