Activation of Basal Gluconeogenesis by Coactivator p300 Maintains Hepatic Glycogen Storage

Because hepatic glycogenolysis maintains euglycemia during early fasting, proper hepatic glycogen synthesis in the fed/postprandial states is critical. It has been known for decades that gluconeogenesis is essential for hepatic glycogen synthesis; however, the molecular mechanism remains unknown. In this report, we show that depletion of hepatic p300 reduces glycogen synthesis, decreases hepatic glycogen storage, and leads to relative hypoglycemia. We previously reported that insulin suppressed gluconeogenesis by phosphorylating cAMP response element binding protein-binding protein (CBP) at S436 and disassembling the cAMP response element-binding protein-CBP complex. However, p300, which is closely related to CBP, lacks the corresponding S436 phosphorylation site found on CBP. In a phosphorylation-competent p300G422S knock-in mouse model, we found that mutant mice exhibited reduced hepatic glycogen content and produced significantly less glycogen in a tracer incorporation assay in the postprandial state. Our study demonstrates the important and unique role of p300 in glycogen synthesis through maintaining basal gluconeogenesis.

The liver plays a critical role in maintaining blood glucose levels within the normal range throughout the fed-fast cycle. During early fasting, hepatic glycogenolysis maintains euglycemia. In contrast, gluconeogenesis plays a dominant role in maintaining blood glucose levels during prolonged fasting. During fed and postprandial states, elevated blood glucose levels promptly increase insulin and decrease glucagon secretion. These hormonal changes act in concert to decrease glucose production in the liver by suppressing glycogenolysis and gluconeogenesis and increase glucose utilization in peripheral tissues by activating glycolysis. Glycogen metabolism is under hormonal regulation by both glucagon and insulin (1–3). Glucagon stimulates the breakdown of glycogen through activation of glycogen phosphorylase by phosphorylating this enzyme at Ser14 (4). In contrast, insulin increases glycogen synthase (GS) activity through the activation of protein kinase B (Akt), which subsequently leads to phosphorylation and deactivation of glycogen synthase kinase 3 (5–7). The phosphorylation of GS by glycogen synthase kinase 3 at a cluster of COOH-terminal serine residues inhibits GS enzymatic activity. Insulin also regulates glycogen metabolism through activation of phosphoprotein phosphatase 1, which, in turn, mediates the dephosphorylation of GS and glycogen phosphorylase; these insulin effects lead to the further activation of GS and inhibition of glycogen phosphorylase (6). Conversely, glucose 6-phosphate plays a key role in regulating GS activity and is able to negate the inactivation of GS due to phosphorylation and fully restore enzymatic activity (6, 8). The increase in hepatic glucose 6-phosphate concentration leading to an elevation in glycogen synthesis can be observed in prolonged (72 h) fasted mice (9).

Insulin also suppresses hepatic gluconeogenesis by phosphorylating cAMP response element (CRE)-binding protein (CREB)-binding protein (CBP), CRTC2, and FoxO1, leading to the disassembly of the CREB-CBP complex (10–12). In addition, CRTC2 and FoxO1 are exported from the nucleus after their phosphorylation and subjected to cytoplasmic degradation. These effects of insulin action lead to suppression (∼60%) of glucose release and storage of glucose as glycogen (1, 13).

Reports from decades ago have suggested that the gluconeogenic pathway accounted for 50%–70% of newly synthesized glycogen (14). In the perfused rat liver or in primary rat hepatocytes, physiologic concentration of glucose had minimal effect on the glycogen synthesis when glucose was the sole substrate; however, efficient glycogen synthesis occurred when gluconeogenic precursors were added (15, 16). Studies from rat, mouse, and dog using radiotracer-labeling techniques have firmly established that the gluconeogenic pathway contributes substantially to hepatic glycogen formation during postprandial state (17–22). Human studies reached the same conclusion (23–26). Data from these studies indicate that a significant amount of gluconeogenesis still occurs even in the presence of elevated blood glucose levels and that physiologic hyperinsulinemia does not completely inhibit net gluconeogenic flux (17–27). Therefore, hepatic gluconeogenesis during the postprandial state has important implications for converting gluconeogenic precursors, such as lactate, fructose, and amino acids delivered from the gastrointestinal tract and other tissues, into glucose that can be stored as glycogen or released into blood as glucose. In fact, 15% of the glucose uptake by muscle is released as lactate, and lactate is then used in the liver to synthesize glycogen through Cori cycle (28).

However, the underlying mechanism of glycogen synthesis through the gluconeogenic pathway in the postprandial state has not been elucidated. Recently, we reported that coactivator p300 maintains basal gluconeogenesis in the fed and postprandial states (29). Therefore, we investigated the possible role of p300 in mediating glycogen synthesis. In the current study, we found that p300 is critical and unique in its ability to maintain hepatic glycogen synthesis in both the fed and postprandial states.

Materials and Methods

Adenoviruses

The BLOCK-iT adenoviral RNA interference expression system (Invitrogen, Carlsbad, California) was used to construct adenoviral short hairpin RNA (shRNA) for CBP, p300, CRTC2, and scrambled shRNA as previously described (12, 29).

Glucose production, glycogen synthesis, and glucose uptake assays

Mouse primary hepatocytes from fed mice were cultured in William's medium E supplemented with insulin-transferrin-selenium (BD Biosciences, Palo Alto, California) and dexamethasone. Glucose production assays were performed as previously described (12, 29). Primary hepatocytes were infected 16–24 hours after planting with adenoviral shRNAs. After incubation for 72 hours, cells were washed twice with PBS. For the glycogen synthesis assay, the medium was replaced with 2 mL of glycogen synthesis buffer consisting of 20 mM glucose and/or 20 mM sodium lactate and 2 mM sodium pyruvate, and supplemented with 20 nM insulin. After a 5-hour incubation, cells were washed 5 times with PBS and collected, and then subjected to 3 freeze-thaw cycles. The glucose and glycogen readings were normalized to the total protein concentration. To measure the rate of glucose uptake, primary hepatocytes from both wild-type (WT) and p300G422S mice were subjected to serum starvation in Williams E medium (11.1 mM glucose) without serum for 4 hours, after which cells were incubated in fresh William E medium supplemented with 1 μCi/mL [2-³H]deoxyglucose (PerkinElmer Life Sciences, Wellesley, Massachucetts) for 10 minutes. Reactions were terminated by 3 washes with ice-cold PBS. Cells were collected and lysed in Lysis Buffer (Cell Signaling Technology, Danvers, Massachusetts). After centrifugation at 3000 rpm for 10 minutes, a portion of the lysate was used for the measurement of protein concentration; the remainder was used for scintillation counting, with results expressed as counts per minute in the cell lysate/mg protein.

Animal experiments

All animal protocols were approved by the Institutional Animal Care and Use Committee of the Johns Hopkins University. Adenoviral shRNA knockdown experiments were conducted 64–72 hours after mice were injected with the adenovirus (12, 29). Because mice have a very long period of nocturnal feeding (6–8 h), we started the fasting period at 5–6 am, which represents the postprandial state. In the ¹⁴CO₂ incorporation assay, 4 month-old WT with adenoviral shRNAs injection (48 h) or 5 month-old p300G422S and littermate control mice were injected ip with a dose of 2 μCi/g of NaH¹⁴CO₃ at 3 am. Blood was collected by cardiac puncture after anesthetization at 5 am, and livers were removed, snap-frozen in liquid nitrogen, and stored at −80°C until use. In the [2-³H]glucose incorporation assay, 3 month-old male p300G422S and littermate control mice were injected ip with 1 μCi/g body weight of tracer at 3 am. After 2 hours, mice were humanely destroyed, and livers were removed and snap-frozen in liquid nitrogen. To measure the glucose 6-phosphate in the liver, 100 mg of hepatic tissue was rapidly homogenized in ice-cold PBS. Following deproteinization, we used a glucose 6-phosphate assay kit to determine glucose 6-phosphate concentrations (Abcam, Inc, Cambridge, Massachusetts). Serum lactate concentration was measured using a colorimetric lactate assay kit (Abcam). All the mice were humanely destroyed at 5–6 am (postprandial state), unless otherwise specified.

Glycogen measurement

To determine ¹⁴C and [2-³H]glucose incorporation into glycogen and the glycogen content in the liver, hepatic tissues were treated as described below. Hepatic tissue was heated at 95°C in 30% KOH for 2 hours. Following cooling, glycogen was precipitated on ice for 20 minutes with 95% ethanol. After centrifugation, glycogen was dissolved in 300 μL of H₂O and precipitated with ethanol; this process was then repeated. The glycogen pellet was washed with 66% ethanol, recentrifuged, and dissolved in 300 μL of H₂O, from which 30 μL was mixed with 3 mL of liquid scintillation mixture. ¹⁴C and ³H were qualified using a liquid scintillation counter. To determine the glycogen concentration, the above purified glycogen and glycogen standards were diluted in 1 M HCl, heated at 95°C for 1 hour, and then centrifuged. The supernatants were diluted 25 times with H₂O, and 20 μL of each sample was used to measure glycogen content with the EnzyChrom Glucose Assay Kit (BioAssay System, Hayward, California). To correct for glucose present in the samples, we also determined the glucose concentration in unheated samples. However, these values turned out to be zero, indicating the purity of the extracted glycogen. Glycogen concentrations in the primary hepatocytes were measured by using the Glycogen Assay Kit (Abcam) and by following the procedure provided by the manufacturer.

Glycogen phosphorylase and GS activity assays

Hepatic tissues were homogenized in 500 μL TES buffer (20 mM Tris, pH7.4; 1 mM EDTA; 225 mM sucrose; 2.5 mM dithiothreitol; 0.1 mM phenylmethylsulfonylfluoride and protease inhibitor cocktail) with a glass Dounce homogenizer as described previously (30, 31). Samples were sonicated and centrifuged at 13 500 rpm for 10 minutes. Protein (100 μg) was used to measure glycogen phosphorylase activity in assay buffer containing: 50 mM potassium phosphate (pH 7.5); 10 mM MgCl₂; 100 μM EDTA (pH 8.0); 0.5 mM caffeine; 4 μM glucose 1,6-biphosphate; 0.5 mM NADP⁺; 1.5 U/mL glucose 6-phosphate dehydrogenase; 1 U/mL phosphoglucomutase; 0.5 mg/mL glycogen. Blank control contained all the reagents and 100 μg of protein except glucose 6-phosphate dehydrogenase and phosphoglucomutase. Sample absorbance at 340 nm was measured in a spectrophotometer after the mixture was incubated at 37°C for 30 minutes. The amount of reduced nicotinamide adenine dinucleotide phosphate in the samples was calculated by using a standard curve of known reduced nicotinamide adenine dinucleotide phosphate concentrations. GS activity was determined using a modified method of Thomas et al. (32). Hepatic tissue was homogenized in 500 μL of ice-cold GS assay buffer (50 mM Tris-HCl, pH 7.8, 10 mM EDTA, 100 mM KF plus proteinase inhibitor cocktail) with a glass Dounce homogenizer prior to centrifugation at 10 000 × g for 20 minutes. To measure the GS activity, 100 μg of total protein (5 μL) was added to a reaction (100 μL) containing: 50 mM Tris-HCl, pH 7.8, 10 mM EDTA, 100 mM KF, 10 mM UDP-[¹⁴C]glucose (0.1 μCi/μM), and 15 mg/mL glycogen, in the presence or absence of 10 mM glucose 6-phosphate, which allosterically activates GS. Assay conducted in the absence of glucose 6-phosphate was used to measure active GS, whereas assay in the presence of glucose 6-phosphate was used to measure total GS activity. Samples were incubated at 37°C, then spotted on Whatman filter paper (Millipore Corp, Bedford, Massachusetts). Filter paper was, dropped immediately in 70% ethanol, stirred with a stirring bar for 40 minutes, and then washed twice. Filters were air dried, and radioactivity was counted in 3 mL liquid scintillation fluid.

Immunoblot

Immunoblot was conducted as we previously described (12, 29)

Statistical analyses

Statistical significance was calculated with Student's t test and ANOVA test. Significance was accepted at the level of P < .05.

Results

The distinct role of CREB coactivators in mediating hepatic glucose production and maintaining blood glucose levels

Previous studies have shown that the CREB coactivators are important for regulating hepatic glucose production and maintaining euglycemia (12, 33). To assess further the role of these coactivators in mediating hepatic glucose production (HGP), we used adenoviral shRNAs delivered through tail vein injection to deplete these coactivators in the liver (Figure 1, A–D). Depletion of p300 significantly reduced blood glucose levels in both postprandial and fasting states (Figure 1B), whereas depletion of CBP (>8 h fast) and CRTC2 (>24 h fast) reduced blood glucose levels only after prolonged fasting (Figure 1, C and D). Depletion of p300 resulted in the greatest drop in blood glucose levels between 12 and 24 hours of fasting (ad-shSCR, −1.4; ad-shCBP, −1.5; ad-shCRTC2, −1.8; and ad-shp300, −2.2 mg/dL/h). These data indicate that p300 functions differently than CBP and CRTC2 in maintaining blood glucose levels.

Figure 1.

Important Role of CREB Coactivators in Regulating Hepatic Gluconeogenesis. A–D, Adenoviral shRNAs-mediated depletion of CREB coactivators significantly lowered blood glucose levels in postprandial and/or fasted states (24 h) (n = 4–5). The y-axis has been broken and begins at 30 mg/dL. Each insert shows the knockdown effect of target protein by adenoviral shRNA.

Having seen that depletion of p300 in the liver led to lower blood glucose levels and depletion of CBP and CRTC2 had minimal effect on blood glucose levels in the postprandial state (Figure 1, B–D), we assessed hepatic mRNA levels of gluconeogenic genes in mice after depletion of CREB coactivators. Depletion of p300 significantly decreased mRNA levels of gluconeogenic genes; in comparison, depletion of either CBP or CRTC2 had minimal effect on gene expression (Figure 2, A and B). In addition, adenoviral-mediated depletion of p300 significantly decreased Pck1 protein levels without affecting serum insulin levels (Figure 2, C and D).

Figure 2.

The p300 Coactivator Regulates Hepatic Gluconeogenic Gene Expression in the Postprandial State. A, Depletion of p300 significantly suppressed gluconeogenic gene expression in the liver of mice in postprandial state (left); however, depletion of CBP (right) had minimal effects on gluconeogenic gene expression (n = 4). B, Adenoviral shRNA-mediated depletion of CRTC2 had no significant effect on the mRNA levels of gluconeogenic genes in the postprandial state (n = 4). C and D, Hepatic protein levels of Pck1 and Fbp1(C) and serum insulin levels (D) in mice after CBP or p300 depletion as in (A) (n = 3–4). Each lane represents an individual mouse sample. Quantitative RT-PCR was used to measure gene expression (normalized to 36B4 expression levels) (A and B).

Role of p300 in glycogen synthesis in hepatocytes

Previous studies demonstrated that gluconeogenesis is essential for postprandial glycogen formation (14, 17–19). Efficient glycogen synthesis requires both glucose and gluconeogenic precursors, and this finding has been termed the “glucose paradox” (14). We therefore conducted a glycogen synthesis assay in primary hepatocytes, and in good agreement with previous studies, glycogen synthesis was markedly activated only in the presence of both high concentrations of glucose and of the gluconeogenic precursors lactate and pyruvate (Figure 3A). Thus, p300 maintains basal gluconeogenesis (29), and depletion of p300 led to lower blood glucose levels in the postprandial state (Figure 1B). To determine whether p300 and other CREB coactivators have an effect on glycogen synthesis, we employed adenoviral shRNAs to deplete the coactivators in primary hepatocytes. As shown in Figure 3B, depletion of p300 significantly decreased glycogen formation in the presence of either glucose alone or glucose plus lactate and pyruvate as substrates, whereas CBP depletion had a lesser effect and CRTC2 depletion had no effect on glycogen formation. Decreased glycogen content in hepatocytes with p300 depletion is likely due to compensation by CBP, the function of which can be inhibited by insulin-mediated phosphorylation. Furthermore, depletion of p300 significantly decreased hepatic glycogen content in mice during the postprandial state (Figure 3C); in contrast, depletion of either CBP or CRTC2 did not lead to a decrease in hepatic glycogen content (Figure 3, C and D). An early study had shown that glucose is broken down to lactate in the liver and released into circulation after feeding in dogs (34). One potential explanation for glucose-stimulated glycogen synthesis is that glucose was metabolized to lactate, and lactate was then used to synthesize glycogen through gluconeogenesis.

Figure 3.

The p300 Coactivator Regulates Hepatic Glycogen Synthesis. A, Glycogen synthesis in primary hepatocytes is dramatically increased by 20 mM glucose and by the addition of gluconeogenic precursors (lactate and pyruvate). B, Depletion of p300 greatly decreased glycogen synthesis wherease depletion of CBP had a mild effect on glycogen synthesis. C, Only the depletion of p300 significantly decreased hepatic glycogen storage in mice humanely destroyed in the postprandial state (n = 7). D, Depletion of CRTC2 increased glycogen content in the liver of mice humanely destroyed in the postprandial state (n = 4). E, Depletion of p300 significantly decreased ¹⁴C-incorporated glycogen in the ¹⁴CO₂ incorporation assay in the postprandial state (n = 4). Mice were administered NaH¹⁴CO₃ (2 μCi/g) through ip injection 48 hours after adenoviral shRNA injection. F, The mRNA levels of genes related to glycogen metabolism in the liver of mice with adenoviral-mediated depletion of either CBP or p300 (n = 3–4). The mRNA measurements were normalized to 36B4 levels. Means ± SEM are shown. Asterisk (*) signifies that groups with same treatment are significantly different (P < .05).

To unequivocally prove that p300 regulates glycogen synthesis through gluconeogenesis, we used adenoviral shRNA to deplete p300 and administered trace quantities of NaH¹⁴CO₃ to mice via ip injection in the postprandial state. In this experiment, ¹⁴CO₂ is incorporated into oxaloacetate through pyruvate carboxylase and subsequent intermediates in the gluconeogenic pathway. Depletion of p300 resulted in a 6-fold decrease in ¹⁴C-incorporated glycogen (Figure 3E). The above data suggest that p300 plays a critical role in maintaining glyconeogenesis through gluconeogenesis in the postprandial state. However, mRNA levels of genes related to glycogen synthesis were not significantly changed after p300 depletion (Figure 3F).

Increased gluconeogenesis leads to elevated glycogen synthesis

We previously reported that CBP was phosphorylated by insulin and metformin at S436 via atypical protein kinase C ι/λ (12). This pathway was subsequently confirmed by other investigators (35). We determined hepatic phosphorylation levels of atypical protein kinase C ι/λ, CBP and Akt from fed and 16-hour fasted mice and found that they all decreased with fasting (Figure 4A). Our recent study showed that p300 maintains basal gluconeogenesis in the liver due to the lack of a corresponding phosphorylation site found in CBP at S436, and p300 constitutively binds to the cAMP response elements (CREs) of genes related to gluconeogenesis (29). In comparison, mutant CBP did not dissociate from the CRE site in the liver of CBPS436A mutant mice in the fed state, and insulin and metformin treatment had no effect on its binding (12). These data suggest that mutant CBP functions like p300 and binds constitutively to the hepatic CRE sites. Because the depletion of p300 decreased glycogen synthesis in hepatocytes (Figure 3, B and C), it can be predicted that up-regulation of gluconeogenesis in CBPS436A mutant mice would result in increased glycogen formation. We therefore conducted glucose production and glycogen synthesis assays in primary hepatocytes from CBPS436A knock-in and littermate control mice. Consistent with our previous publication (36), primary hepatocytes from CBPS436A mice produced significantly more glucose than hepatocytes from control mice in both basal and cAMP treatment groups (Figure 4B). We now find that primary hepatocytes from CBPS436A mice also synthesized significantly more glycogen than primary hepatocytes from control mice in the presence of either glucose or in the presence of glucose, lactate, and pyruvate (Figure 4C).

Figure 4.

Up-Regulation of Gluconeogenesis Increases Glycogen Synthesis. A, The phosphorylation status of CBP, CREB, PKCι/λ, and Akt in the liver from fed and 16-hour fasted mice (n = 5). B, Primary hepatocytes from CBPS436A mice produced significantly more glucose than hepatocytes from control mice in the presence or absence of cAMP treatment. Glucose production measurements were normalized to protein levels and the untreated control group. The data shown here are relative to 1 (normalized data). C, Primary hepatocytes from a CBPS436A mouse synthesized significantly more glycogen than primary hepatocytes from a control mouse in each treatment.

Decreased gluconeogenesis in p300G422S knock-in mice results in the lower hepatic glycogen formation in the postprandial state

To prove that distinct functions of p300 and CBP in hepatic glucose production are due to the absence or presence of this phosphorylation event, we used the p300G422S knock-in mouse model bearing the identical phosphorylation site found in CBP at S436. We first confirmed that G422S in p300 can be phosphorylated by insulin and metformin (29). Compared with WT littermates, mutant p300 was absent from the CRE site of genes important for gluconeogenesis in the fed and postprandial states, and p300G422S knock-in mice also exhibited lower blood glucose levels in the postprandial state (Figure 5A) (29). In addition, p300G422S knock-in mice displayed hypersensitivity to insulin (Figure 5B) and enhanced glucose tolerance (Figure 5C). Lower blood glucose levels in p300G422S knock-in mice were also associated with significantly lower protein levels of G6pc and Pck1 (Figure 5D) and lower hepatic glycogen content in the postprandial state (Figure 5E). We observed similar protein levels of Gck and Glut2 in the liver of WT and p300G422S mice (Figure 5D). Furthermore, the mRNA levels of genes related to glycolysis and glycogen metabolism did not differ significantly between WT and p300G422S mice (Figure 5F). Of note, p300G422S knock-in mice had lower mRNA levels of Scd1 and Fasn, which would decrease lipogenesis in the liver and increase glucose tolerance (Figure 5, B, C, and F). Finally, compared with WT mice, p300G422S displayed significantly lower hepatic glycogen levels, which were quickly depleted during fasting (Figure 5, G and H).

Figure 5.

The Distinct Role of CBP and p300 in Regulating Hepatic Glucose Production due to the Presence of Insulin Phosphorylation at CBP S436, and Which Does Not Exist in p300. A, P300G422S knock-in mice exhibited lower blood glucose levels in the postprandial state compared to littermate control mice. B and C, Insulin tolerance test (B) and glucose tolerance test (C) in 2-hour fasted mice (n = 4–5). D, Indicated protein levels in the liver of p300G422S knock-in mice and littermate controls in the postprandial state (n = 4). Densitometric analysis of the protein levels (right panel). E, Hepatic glycogen content in p300G422S knock-in mice and littermate controls in the postprandial state (n = 3–4). F, The mRNA levels of genes related to glucose, glycogen, and lipid metabolism in the liver of mice killed in postprandial state (n = 5). The mRNA measurements are normalized to 36B4 levels. G and H, Hepatic glycogen contents in WT (G) and p300KI (H) mice at indicated times during the 12-hour or 24-hour fasting period (n = 3–4).

To support the notion that decreased gluconeogenesis resulted in the lower glycogen content in the liver of p300G422S knock-in mice in the postprandial state (Figure 5E), we conducted glucose production assays in primary hepatocytes. Hepatocytes from p300G422S knock-in mice produced significantly less glucose than hepatocytes from littermate control mice in both basal and cAMP treatment groups (Figure 6A). Insulin and metformin had a greater inhibition of cAMP-stimulated glucose production in hepatocytes from p300 knock-in mice than in WT control mice (Figure 6, A and B). If gluconeogenesis is important for the glycogen synthesis, then decreased gluconeogenesis must result in the lower glycogen content. We, therefore, administered trace quantities of NaH¹⁴CO₃ via ip injection to p300G422S knock-in and WT littermate control mice in the postprandial state. Compared with WT control mice, p300G422S knock-in mice exhibited approximately 25× less [¹⁴C]glycogen content in the liver in the ¹⁴CO₂ incorporation assay (Figure 6C). However, the protein levels of Glut2 and Gck did not differ between p300G422S and littermate control mice (Figure 5D), and hepatocytes from p300G422S mice exhibited similar glucose uptake rates as hepatocytes from WT littermate control mice (Figure 6D). In addition, p300G422S mice displayed increased [³H]glycogen content in the liver during a [2-³H]glucose incorporation assay, even though this result did not reach statistical significance (Figure 6E). These data eliminate the possibility that lower glycogen content in p300G422S mice is due to decreased glucose uptake or decreased glycogen synthesis from the direct pathway.

Figure 6.

P300G422S Knock-In Mice Synthesize Less Glycogen in the Liver. A and B, Primary hepatocytes from a p300G422S knock-in mouse produced significantly less glucose than hepatocytes from a control mouse in the presence or absence of cAMP treatment. After 4 hours of serum starvation, cells were washed with PBS twice, after which 20 nM insulin was added 30 minutes prior to the addition of cAMP (A) and 5 mM metformin was added 4 hours before the addition of cAMP (B). C, P300G422S knock-in mice synthesized significantly less glycogen in the ¹⁴CO₂ incorporation assay (n = 4). D, [2-³H] deoxyglucose uptake in primary hepatocytes from p300KI and littermate control mice. Hepatocytes were treated as described in “Materials and Methods”. E, P300G422S and WT control mice synthesized similar amounts of [³H]glycogen in the [2-³H]glucose incorporation assay (n = 5∼6). F, Crude extract of hepatic tissues (100 μg) was used to measure glycogen phosphorylase (PYG) activity. G, Immunoblot analysis of hepatic protein levels with indicated antibodies of p300G422S knock-in mice and littermate controls in the postprandial state. Serine phosphorylation levels were examined in immunoprecipitated glycogen phosphorylase (n = 5). H, Indicated phosphoprotein or total protein levels were determined in the liver of WT mice killed at postprandial or fasted (24 h) state. Each lane represents an individual mouse sample.

Next, we determined the enzymatic activity and phosphorylation status of glycogen phosphorylase. The dramatically decreased [¹⁴C]glycogen content was not due to a change of glycogen phosphorylase activity because the enzymatic activity and serine phosphorylation levels of glycogen phosphorylase were not significantly different between p300G422S and WT littermate control mice (Figure 6, F–H). However, p300G422S mice exhibited a significantly lower active GS and GS activity ratio (Figure 7, A–C). Intriguingly, GS was dephosphorylated in both p300G422S and littermate control mice in the postprandial state compared with the fasting state (Figure 6G). Moreover, p300G422S mice had similar amounts of phospho-GS and total GS protein levels in the liver as WT mice (Figure 6G). The above data indicate that the lower GS activity in the liver of p300G422S mice was not due to changes in GS protein phosphorylation or in total GS protein levels.

Figure 7.

Decreased Gluconeogenic Flux Affects Hepatic Glycogen Synthesis in p300KI Mice. A–C, GS activity in the absence or presence of 10 mM glucose 6-phosphate; active GS and GS active ratio were significantly decreased in p300G422S mice (A and B), and along with similar total GS activity between WT and p300KI mice (C) (P = .98). D, Glucose 6-phosphate levels in the liver of p300KI and littermate control mice killed at either postprandial or fasted (24 h) states. E, Serum lactate concentrations of p300KI and littermate control mice killed at postprandial state (n = 4). NS, not significant. F, A proposed model for maintaining blood glucose levels by CREB coactivators through gluconeogenesis and glycogenolysis.

Glucose 6-phosphate is a critical allosteric activator of GS (6, 8), and p300G422S mice had significantly lower gluconeogenesis in the postprandial state. It is possible that lower gluconeogenesis in the liver of p300G422S mice would lead to lower glucose 6-phosphate levels and subsequently affect GS activity. Indeed, compared with WT littermate controls, p300G422S mice exhibited significantly lower hepatic glucose 6-phosphate levels in the postprandial state (Figure 7D). The above data suggest that lower ¹⁴C incorporation into glycogen in p300G422S knock-in mice must be due to decreased gluconeogenesis. Additionally, the reduction in gluconeogenesis is intrinsic to the pathway as we suggest because levels of serum lactate, a critical substrate for gluconeogenesis, were higher in p300G422S knock-in mice compared with WT littermate control mice (Figure 7E), which suggests a reduction in lactate utilization in the liver.

Discussion

Hepatic glycogen is important for maintaining euglycemia in early fasting but is subsequently depleted during a prolonged fast (Figure 5G). It is generally believed that glucose is not the primary substrate for either glyconeogenesis or lipogenesis in the liver. In contrast, gluconeogenic precursors such as lactate and pyruvate are superior substrates for these processes (14, 15, 37, 38). It has been shown that carbon flux through the gluconeogenic pathway is essential for efficient glycogen synthesis, and the inhibition of PCK1 enzymatic activity led to approximately 85% decrease of glycogen content in the liver (17, 18). These reports also indicate that gluconeogenesis is still active in the postprandial states because glyconeogenesis occurs during this period. It is estimated that as much as 50%–70% of newly synthesized glycogen is formed via the gluconeogenic pathway (14, 21, 25). Moreover, an unexplained observation in humans has been that gluconeogenesis is not completely suppressed even in the presence of excess insulin (39–42).

CBP and p300 are closely related proteins, but p300 lacks the phosphorylation site found in CBP at S436 and cannot be phosphorylated at this site. We recently reported that p300 constitutively binds to hepatic CREs such as found in the Ppargc1 gene promoter and maintains basal gluconeogenesis in the fed and postprandial states (29). In the current study, we found that depletion of hepatic p300 resulted in decreased hepatic mRNA levels of gluconeogenic genes and lower blood glucose levels in the postprandial state. Most importantly, depletion of p300 led to lower hepatic glycogen content and decreased [¹⁴C]glycogen in tracer incorporation assay in the postprandial state and early fasting stages (Figure 1, Figure 2A, and Figure 3, C and E). These data suggest that p300 is an important coactivator in maintaining glyconeogenesis. Moreover, p300G422S knock-in mice exhibited lower glycogen content and lower protein levels of gluconeogenic enzyme genes in the postprandial state (Figure 5, D–F), and hepatocytes from p300G422S knock-in mice produced significantly less glucose (Figure 6, A and B). These data suggest that decreased gluconeogenesis via introduction of an artificial phosphorylation site leads to lower hepatic glycogen formation and causes lower blood glucose levels in p300G422S knock-in mice in the postprandial and early fasted states. This was unequivocally substantiated by the significantly decreased [¹⁴C]glycogen in the p300G422S knock-in mice in the ¹⁴CO₂ incorporation assay (Figure 6C). P300G422S knock-in mice also had significantly lower mRNA and protein levels of G6pc than WT control mice in the postprandial state (Figure 5, D and F), which would, if anything, tend to divert the reduced gluconeogenic flux toward glycogen synthesis. These data suggest that constitutive binding of p300 to gluconeogenic enzyme gene promoters is also necessary for maintaining hepatic glycogen stores. On the other hand, mice containing CBPS436A, which functions like p300, displayed increased gluconeogenesis and glycogen synthesis in hepatocytes (Figure 4, B and C) (12, 36). Thus, the activation of gluconeogenic pathway by constitutive binding of p300 to the CREB coactivator complex maintains glyconeogenesis and sustains euglycemia through the activation of basal gluconeogenesis in the postprandial and early fasting states (Figure 7F).

Given that diabetic patients have increased hepatic gluconeogenesis (42, 43), one might predict that hepatic glycogen storage would be increased. In fact, however, previous studies have shown that diabetic patients have lower hepatic glycogen content due to decreased GS and/or increased glycogen phosphorylase activity (42, 43). This discrepancy might be explained by an increase in glucose 6-phosphatase in diabetic patients (44–46), which would increase the conversion of glucose 6-phosphate to glucose rather than being used for glycogen synthesis. Our mouse models of hepatic p300 depletion (Figures 2 and 3) and p300G422S mutation (Figures 5–7) demonstrate that p300 is critical for basal gluconeogenesis and glycogen synthesis and that p300 is unique among the CREB coactivators in having this function. In comparison, neither depletion of CBP nor CRTC2 resulted in the reduction of postprandial hepatic glycogen levels (Figure 3, C and D). Because total protein and phosphorylation levels of glycogen phosphorylase and synthase were normal in p300G422S mutant mice, the enzymatic activity of glycogen phosphorylase did not differ between WT and mutant mice (Figure 6, F and G). In contrast, we demonstrate that the lower glycogen content in mutant mice was due to the decreased GS activity resulting from decreased glucose 6-phosphate levels (Figure 7). Glucose 6-phosphate can be formed from glucose after its uptake and phosphorylation by glucokinase or be derived from gluconeogenesis. Given that glucose 6-phosphatase mRNA and protein levels are reduced, glucose uptake and protein levels of Glut2 and Gck are normal, and glycogen synthesis from direct pathways is unaffected in hepatocytes from p300G422S mice; the lower glucose 6-phosphate levels in the liver of p300G422S mice in the postprandial state must be due to a marked decrease in hepatic gluconeogenesis. Furthermore, p300 maintains constitutive activation of gluconeogenesis even in the postprandial state, which is due to the fact that p300 activity is not regulated by insulin (29). The latter finding may explain rodent and human data, which suggest that gluconeogenesis is unable to be completely inhibited even when high serum insulin levels are achieved in insulin-sensitive normal animals and human subjects.

Acknowledgments

This work was supported by grants from the National Institute of Diabetes and Digestive and Kidney Diseases, R00DK085142 (to L.H.) and R01DK063349 (to F.E.W.); and by the Baltimore Diabetes Research and Training Center, P60DK079637 (to F.E.W.).

Disclosure Summary: The authors have nothing to disclose.

Abbreviations

 
• Akt

  protein kinase B

 
• CBP

  CREB-binding protein

 
• CRE

  cAMP response element

 
• CREB

  CRE-binding protein

 
• GS

  glycogen synthase

 
• shRNA

  short hairpin RNA

 
• WT

  wild type.