Mice with a Deletion in the Gene for CCAAT/Enhancer-binding Protein β Have an Attenuated Response to cAMP and Impaired Carbohydrate Metabolism*

Fifty percent of the mice homozygous for a deletion in the gene for CCAAT/enhancer-binding protein β (C/EBPβ−/− mice; B phenotype) die within 1 to 2 h after birth of hypoglycemia. They do not mobilize their hepatic glycogen or induce the cytosolic form of phosphoenolpyruvate carboxykinase (PEPCK). Administration of cAMP resulted in mobilization of glycogen, induction of PEPCK mRNA, and a normal blood glucose; these mice survived beyond 2 h postpartum. Adult C/EBPβ−/− mice (A phenotype) also had difficulty in maintaining blood glucose levels during starvation. Fasting these mice for 16 or 30 h resulted in lower levels of hepatic PEPCK mRNA, blood glucose, β-hydroxybutyrate, blood urea nitrogen, and gluconeogenesis when compared with control mice. The concentration of hepatic cAMP in these mice was 50% of controls, but injection of theophylline, together with glucagon, resulted in a normal cAMP levels. Agonists (glucagon, epinephrine, and isoproterenol) and other effectors of activation of adenylyl cyclase were the same in liver membranes isolated from C/EBPβ−/− mice and littermates. The hepatic activity of cAMP-dependent protein kinase was 80% of wild type mice. There was a 79% increase in the concentration of RIα and 27% increase in RIIα in the particulate fraction of the livers of C/EBPβ−/− mice relative to wild type mice, with no change in the catalytic subunit (Cα). Thus, a 45% increase in hepatic cAMP (relative to the wild type) would be required in C/EBPβ−/− mice to activate protein kinase A by 50%. In addition, the total activity of phosphodiesterase in the livers of C/EBPβ−/− mice, as well as the concentration of mRNA for phosphodiesterase 3A (PDE3A) and PDE3B was approximately 25% higher than in control animals, suggesting accelerated degradation of cAMP. C/EBPβ influences the regulation of carbohydrate metabolism by altering the level of hepatic cAMP and the activity of protein kinase A.

purchased from Sigma and for protein kinase A were from Upstate Biotechnology (Lake Placid, NY). The SuperSignal® chemiluminescent substrate kit was from Pierce. Antibodies against RI␣ were from Biomol (Plymouth Meeting, PA), against RII␣ were from Santa Cruz Biotechnology (Santa Cruz, CA), and against C␣ were from Upstate Biotechnology (Lake Placid, NY). Imobolin-P® polyvinylidene difluoride membranes were purchased from Millipore Corp (Bedford, MA).
Experimental Animals-C/EBP␤Ϫ/Ϫ mice were obtained for this study by breeding female heterozygous animals with a targeted deletion in the gene for C/EBP␤ with heterozygous male mice. The generation of the C/EBP␤Ϫ/Ϫ mice and their genetic background have been described by Screpanti et al. (9). Briefly, ES cell clones from the CCE cell line (derived from the 129/Sv/Ev strain) carrying the mutation were injected into C57Bl6 blastocysts and were transplanted into the uteri of F1 (CBAϫC57Bl6) foster mothers. Male chimeras were mated to MF1 females, and offspring heterozygous for the mutant allele were intercrossed to obtain homozygous mice. Adult male and female mice were 8 -12 weeks old at the time of their use. Screening for C/EBP␤Ϫ/Ϫ mice was carried out by Southern analysis as described previously (9). The animals were given free access to water and standard chow (Tekland F6 8664 containing 24% protein, 6% fat, and 4.5% crude fiber and the remainder carbohydrate). The composition of the high carbohydrate diet used in this study was described previously (11). The mice were killed between 9 and 11 a.m., and where indicated, they were injected with Bt 2 cAMP (35 mg/kg of body weight) and theophylline (30 mg/kg of body weight).
Perinatal Studies-Mice were delivered at day 19 of gestation by cesarean section, and where indicated, they were injected with 125 mg of Bt 2 cAMP/kg of body weight (12). All pups were maintained in a humidicrib at 37°C from birth to the completion of the experiment (up to 4 h). When the pups of phenotype B became lethargic due to a low blood glucose concentration, they were killed together with their littermates. Blood was collected by decapitation, blood glucose was measured using a Glucometer (Ames Products, Indianapolis, IN), and insulin was measured by radioimmunoassay. The liver was freeze-clamped, and total RNA was isolated as described below, glycogen was measured (13), and cAMP was determined using an enzyme immunoassay procedure. Glycogen was extracted from frozen livers by homogenization in 6% KOH, precipitated in ethanol, and hydrolyzed by boiling in 1 N HCl, and glucose was measured using a Beckman glucose analyzer. Statistical comparison between groups were performed using Student's t test.
Metabolic Measurements-Adult C/EBP␤Ϫ/Ϫmice (A phenotype) were given 5 mg of glucose/kg of body weight orally by gavage and then fasted for 16 and 30 h. Blood was taken from the tails of mice at 30 min, and after 16 and 30 h of fasting, and plasma was separated and frozen. Blood glucose was measured using glucose oxidase method (Glucose Trender Kit, Sigma), whereas the concentration of ␤-hydroxybutyrate was determined using an enzymatic kit (Sigma). The blood urea nitrogen (BUN) levels were measured on a Beckman automated analyzer at MetroHealth, Schwartz Nutrition Center, Cleveland, Ohio. Ammonia levels were measured on blood plasma after an overnight fast by an enzymatic kit (Sigma) based on reductive amination using Lglutamate dehydrogenase (14). Insulin was determined using a radioimmunoassay.
Systemic Glucose Production-Mice were fasted overnight before injecting into the tail vein 100 l of 5 Ci of D-[3-3 H]glucose in 0.9% NaCl. Blood samples (25 l) were obtained via the tail vein at 5, 15, and 30 min, and serum was used for the determination of glucose (glucose oxidase method). For the determination of radioactivity, 10 l of blood was deproteinized with 200 l of 20% trichloroacetic acid. Samples were centrifuged, and the supernatants were evaporated to dryness overnight at 65°C. The residues were reconstituted in 200 l of water, 5 ml of scintillation fluid was added, and the samples were counted in a ␤-scintillation spectrometer. The rate of systemic glucose production was calculated using steady-state equations. (15) Statistical comparison between groups was made using Student's t test.
RNA Extraction and Northern Blot Analysis-Total RNA was extracted from the liver and kidney of mice using a Quick Prep total RNA kit (Amersham Pharmacia Biotech) by a modified acid-phenol guanidine thiocyanate procedure that has been described in detail previously (16). Northern blot analysis was carried out as described previously (17) using 20 g of total RNA. After electrophoresis of the RNA, RNA in the gels was transferred to Gene Screen Plus® membrane and hybridized with a probe. The probe for PEPCK was a 1.1-kilobase PstI fragment from the 3Ј-end of the PEPCK cDNA that was isolated as described previously (16). The probe for glucose-6-phosphatase (Glc-6-P) mRNA was a fragment (XbaI-PstI) from Glc-6-P cDNA (18). The C/EBP␤ cDNA probe was a 0.7-kilobase BamHI fragment of the mouse cDNA (9). The concentration of 18 S rRNA was determined by Northern blotting using a 752-nucleotide SacI fragment of a cDNA made from mouse 18 S rRNA. The signal from this hybridization was used to standardize the concentration of RNA on the Northern blots. All probes were labeled by using [␣ 32 P]dATP and the Strip-EZ RNA & DNA probe synthesis and removal kit (Ambion Inc., Austin, TX).
DNA Analysis-DNA was isolated from the tail of mice by lysis overnight at 55°C in a buffer containing 50 mM KCl, 10 mM Tris-HCl, pH 8.3, 2.5 mM MgCl 2 , 0.1% gelatin, 0.45% Nonidet P-40, 0.45% Tween 20, and 24 g of proteinase K. The DNA was digested with EcoRI, and the resulting fragments were separated by electrophoresis on 1% agarose gel, transferred to Gene Screen Plus®, and hybridized to a cDNA probe for C/EBP␤.
Glucagon-stimulated cAMP Production in Liver in Vivo-Fed C/EBP ␤Ϫ/Ϫ mice and control littermates (WT) (8 -12 weeks of age) were anesthetized with avertin, the liver was clamped, and a biopsy was taken for the measurement of basal cAMP. The mice were then injected with glucagon (50 g/kg of body weight) via the portal vein, and the liver was biopsied 1 min later for the assay of glucagon-induced cAMP. C/EBP␤Ϫ/Ϫ and wild type mice were given an intraperitoneal injection of either theophylline (30 mg/kg of body weight) or RO 20-1724 (15 mg/kg of body weight). Thirty min later, the mice were anesthetized with avertin, the liver was clamped to prevent bleeding, and a liver biopsy was obtained. This was used for the basal concentration of cAMP. Glucagon (50 g/kg of body weight) was injected into the portal vein, and one min later, another piece of liver was removed for the quantitation of cAMP. The liver samples were quickly frozen and assayed for cAMP using an enzyme immunoassay (Amersham Pharmacia Biotech).
Quantitative Reverse Transcription-PCR-Quantitative competitive reverse transcription-PCR was used to measure the relative concentrations of phosphodiesterase mRNA in the livers of C/EBP␤Ϫ/Ϫ and control littermates. This procedure involves four steps: first, total RNA was isolated from the liver; second, reverse transcription was performed to create cDNAs; third, competitive PCR was preformed; and fourth, the DNA bands were quantitated. The RNA was isolated from the livers of mice using the Amersham Pharmacia Biotech QuickPrep® total RNA extraction kit. Trace amount of genomic DNA was removed from the sample by treatment with 10 units of RNase-free DNase I, and the RNA was further purified by using a Qiagen RNeasy Mini Kit®. Reverse transcription was performed at 42°C for 2 h using an Ambion RETROscript kit®. Briefly, 2 g of total RNA was reverse-transcribed in a reaction mixture containing 0.5 mM dNTP, 5 M random primer, 1ϫ buffer (50 mM Tris-HCl, pH 8.3, 75 mM KCl, 30 mM MgCl 2 , and 50 mM dithiothreitol), 10 units of RNase inhibitor, and 100 units of reverse transcriptase. The competitive PCR reaction was carried out in a volume of 25 l that contained 2 l of the solution from the reversetranscribed reaction with varied concentrations of an internal DNA control fragment as well as 0.2 mM dNTP mix, 1ϫ PCR buffer (20 mM Tris-HCl, pH 8.4, and 50 mM KCl), 3 mM MgCl 2, 0.8 M mixed primer, and 0.5 units of SuperTaq DNA polymerase. The competitive PCR reaction was performed at 94°C for 1 min, at 55°C for 1 min, and at 72°C for 1 min for 30 cycles. For each RNA sample, one reverse transcriptase reaction and eight PCR reactions were performed. The internal DNA control fragments were constructed as follows. The internal control for PDE3A was obtained from a 482-bp segment of the mouse PDE3A by deleting a 178-bp StyI fragment. The internal control for PDE3B was generated by deleting a 124-bp SacII segment of the 682-bp XhoI-StuI cDNA fragment, and for the PDE4B internal control, the fragment was obtained by deleting a 152-bp NsiI fragment from the 1968-bp EcoRI cDNA.
After the PCR reaction, 8 l of the reaction solution was subjected to electrophoresis in a 1.6% agarose gel. The DNA bands were recorded using an IS-500 digital imaging system (Alpha Innotech Corp. San Leandro, CA), and the bands were quantitated with ImageQuant® software (Molecular Dynamics, Sunnyvale, CA). The concentration of mRNA was determined as the point where the internal control band is equal in intensity to that of the test sample. To find that point, the ratios of the intensities of test cDNA and internal control bands were calculated and fitted to a regression curve using GraphPad Prism® software (GraphPad Software, San Diego, CA).
Adenylyl Cyclase Activity-Partially purified liver plasma membranes were prepared from the livers of C/EBP␤ and control mice as described previously (19), with slight modifications. Briefly, the livers were homogenized (Dounce homogenizer) in 5 ml of Buffer A (10 mM Tris-HCl, pH 7.5, 1 mM EDTA, 100 M phenylmethylsulfonyl fluoride, 3 mM benzamidine, and 1 M leupeptin) containing 0.25 M sucrose. The homogenates were centrifuged at 17,000 ϫ g for 10 min, and the pellet was suspended in 4 ml of the homogenization buffer. This suspension was layered onto 42.3% (w/w) sucrose in Buffer A and centrifuged at 30,000 rpm for 90 min in a SW40 rotor. The interfacial material was collected and washed twice with Buffer A by centrifuging at 86,000 ϫ g for 45 min. The final pellet was re-suspended in Buffer A and stored at Ϫ80°C until used. The yield of membrane protein, as determined by the method of Lowry (20), was about 15 mg of protein/g of liver; there was no difference in the yield of liver membranes of C/EBP␤ and control mice. Adenylyl cyclase activity in the membrane preparations was assayed for 10 min at 30°C using 300 M [␣-32 P]ATP as a substrate, as described previously (19).
Protein Isolation for Western Blotting-Proteins were isolated from the livers of C/EBP␤Ϫ/Ϫ and wild type mice that were fasted overnight. A piece of frozen liver was homogenized in 15 ml/g of tissue of ice-cold homogenizing buffer consisting of 20 mM Tris, pH 7.6, 0.1 mM EDTA, 0.5 mM EGTA, 0.1% Triton-X, 250 mM sucrose, and 50 l/5 ml protease inhibitor mixture. Homogenates were centrifuged at 19,500 ϫ g for 30 min at 4°C to separate cellular debris, mitochondria, nuclei, and plasma membranes. This pellet was re-suspended in 4 ml of homogenizing buffer (N fraction). The supernatant was then centrifuged at 100,000 ϫ g for 30 min at 4°C. The resulting supernatant (the cytosolic fraction) was removed, and the remaining pellet (particulate fraction) was re-suspended in 4 ml of homogenizing buffer. The particulate fraction contained membranes of the endoplasmic reticulum and the Golgi complex. The concentration of protein was measured with the Bio-Rad protein assay using bovine serum albumin as a standard.
Electrophoresis and Western Blotting-The N fraction as well as the cytosolic and particulate fractions were sonicated for 20 s, and 20 g of protein was diluted in 2ϫ loading buffer containing 100 mmol/liter Tris-HCl, pH 6.8, 20% ␤-mercaptomethanol, 4% SDS, 0.2% bromphenol blue, 20% glycerol and separated by 10% SDS-PAGE. Proteins were electrophoretically transferred to Immobilon-P® polyvinylidene difluoride membranes and stained with Coomassie stain (45% methanol, 10% acetic acid, 2.5% Coomassie Blue R250) to ensure even loading. For detection of the subunits of protein kinase A, RI␣, RII␣, and C␣, polyvinylidene difluoride membranes were incubated in blocking buffer containing 5% nonfat dried milk in 10 mmol/liter Tris, pH 7.4, 150 mmol/liter NaCl, and Tween 20 (TBS-T) for 1 h at room temperature. Membranes were then incubated with primary antibody diluted in blocking buffer for 1 h as follows: RI␣ (1:250), RII␣ (1:1000), C␣ (1:250), washed 3 times for 5 min each in TBS-T. The membranes were then incubated with secondary antibody diluted in blocking buffer for 1 h as follows: rabbit anti-chicken IgG peroxidase (1:1000) for RI␣ and antirabbit IgG peroxidase (1:1000) for RII␣ and C␣. Membranes were washed 3 times for 5 min each in TBS-T. All incubations were at room temperature. Immunoreactive proteins were detected using the Super-Signal Chemiluminescent Substrate® kit, and the density of the immunoreactive bands was measured by scanning densitometry. Western blots were first reacted with RI␣, washed as described by Pierce, and reacted again with RII␣ and then with C␣.
Cyclic AMP-dependent Protein Kinase Assay-The activity of protein kinase A (PKA) was measured using the Kemptide assay (21) using cytosol and particulate fractions from the livers of fasted C/EBP␤Ϫ/Ϫ and wild type mice. Livers were homogenized in ice-cold Kemptide homogenizing buffer (1 g of liver/9 ml of buffer). The buffer contained 20 mM Tris, pH 7.6, 0.1 mM EDTA, 0.5 mM EGTA, 0.1% Triton-X, and 250 mM sucrose and 50 l/5 ml protease inhibitor mixture. The homogenate was centrifuged at 19,500 ϫ g for 30 min at 4°C. The cytosolic fraction was used for the measurement of PKA activity. Samples were diluted 1:1 with the Kemptide homogenization buffer, and the phosphorylation of the Kemptide substrate was measured using the PKA assay kit with [␥-32 P]-ATP. Duplicate samples were assayed at 30°C at 30 s intervals for 5 min for three C/EBP␤Ϫ/Ϫ and three wild type mice. The amount of labeled Kemptide was determined using a Beckman LS scintillation spectrometer. Activity in the presence of protein kinase A inhibitor peptide was subtracted from total activity to account for nonspecific activity.
cAMP Degradation in Vitro-Phosphodiesterase activity in homogenates of livers from C/EBP␤Ϫ/Ϫ and control mice was determined by the procedure of Shahid and Nicholson (22). About 0.2 g of liver was homogenized in 2 ml of Buffer A containing 0.25 M sucrose using a glass homogenizer fitted with a motor-driven Teflon pestle. The homogenates were filtered though two layers of cheese cloth and used immediately for the assay of phosphodiesterase activity. Approximately 25 g of protein equivalent of the homogenate was incubated with 1 M [ 3 H]cAMP for 5 min at 30°C. The reaction was terminated by placing the samples into boiling water; the unhydrolyzed [ 3 H]cAMP was separated by using alumina columns. In other assays in which the rate of cAMP degradation was also measured in the homogenate, 50 g of protein was incubated with 1 M [ 3 H]cAMP at 30°C, and the labeled cyclic nucleotide that was not hydrolyzed was measured at various time intervals up to 6 min by separation on alumina columns as described above.

Characteristics of Newborn C/EBP␤Ϫ/Ϫ Mice Exhibiting the A and B
Phenotypes-Mice heterozygous for a deletion in the gene for C/EBP␤ were mated, and the pups were delivered by cesarean section on days 19 to 20 of gestation. The newborn mice were stimulated by touch after birth and maintained at 37°C in a humidicrib. C/EBP␤Ϫ/Ϫ mice identified as being of the B phenotype were typically lethargic, had difficulty breathing, and remained cyanotic until they died within 2 h after delivery. The C/EBP␤Ϫ/Ϫ pups that were viable and breathing well were classified as the phenotype A. The plasma glucose levels of the B phenotype were about 50% that of control pups (wild type) (p Ͻ 0.01) (Fig. 1). The C/EBP␤Ϫ/Ϫ mice (B phenotype) also had difficulty in mobilizing hepatic glycogen compared with wild type littermates, as was evident from the higher level (p Ͻ 0.03) of hepatic glycogen in these animals (70 mg/g of liver glycogen in C/EBP␤Ϫ/Ϫ versus 40 mg/g of liver glycogen in the wild type mice). Because maintenance of normal blood glucose in the neonate depends on the capacity to mobilize hepatic glycogen as well as to initiate hepatic gluconeogenesis, we measured the ability of these animals to initiate gluconeogenesis by measuring the level of mRNA for PEPCK, the last of the gluconeogenic enzymes to develop in newborn mammals (23). As shown in Fig. 1, the level of PEPCK mRNA in C/EBP␤Ϫ/Ϫ mice (B phenotype) was only 30% that of either the A phenotype or the wild type control animals. PEPCK gene expression is down-regulated by insulin (24). We therefore measured the concentration of insulin in the plasma of C/EBP␤Ϫ/Ϫ mice and control mice; the insulin levels were found to be the same (data not shown).
The high levels of glycogen in the livers of animals of the B phenotype suggested a defect in the ability of these mice to mobilize their hepatic glycogen. C/EBP␤Ϫ/Ϫ mice at 19 to 20 days of fetal life were delivered by cesarean section and given an intraperitoneal injection of Bt 2 cAMP (125 mg/kg) immediately after delivery. These mice mobilized their hepatic glycogen to the same extent as control littermates (to a level of 35 mg/g of liver) (data not shown). Surprisingly, mice with the B phenotype responded immediately to the Bt 2 cAMP by breathing normally and becoming less lethargic. All of the C/EBP␤ Ϫ/Ϫ mice injected with Bt 2 cAMP survived for up to 4 h, the duration of the experiment (Fig. 2). Since the administration of Bt 2 cAMP rescued the C/EBP␤Ϫ/Ϫ mice (B phenotype) from death within the first 2 h after birth, we considered it important to determine whether endogenous levels of cAMP in the livers of B-phenotype mice were different from the normal littermates. For this experiment, the animals were delivered at 19 -20 days of fetal life. The newborn animals were treated as one group representing a mixture of both A and B phenotypes. The C/EBP ␤Ϫ/Ϫ mice had the same level of cAMP as control mice (data not shown), suggesting that there is no defect upstream of cAMP production. The biochemical basis for the decreased viability of C/EBP ␤Ϫ/Ϫ mice of the B phenotype remains enigmatic, but it seems likely that their inability to respond to the normal concentration of cAMP may be due to an alteration in downstream target(s) required to activate critical metabolic processes during the perinatal period.
Metabolic Characteristics of C/EBP␤Ϫ/Ϫ (phenotype A) and Wild Type Mice-Based on the blunted response of the C/EBP␤Ϫ/Ϫ mice (B phenotype) to the endogenous concentrations of hepatic cAMP and the resulting aberrations in carbohydrate metabolism, we extended this study to adult animals (the A phenotype) to determine the differences in metabolic response from adult wild type control mice. To simulate the fed state, adult mice were given glucose orally (5 g/kg), and blood was taken from the tail vein 30 min later. The mice were then fasted for 30 h, blood was collected from the tail vein at 16 and 30 h, and the concentrations of glucose, ␤-OH butyrate, and BUN were determined. The concentration of glucose in the blood of the C/EBP␤Ϫ/Ϫ mice was 25% lower than that of wild type control animals, whereas the level ␤-hydroxybutyrate was 50% lower than the wild type animals (Fig. 3). The lowered concentration of ␤-hydroxybutyrate probably reflects a decreased oxidation of fatty acids in the C/EBP␤Ϫ/Ϫ mice because these animals exhibit lower fasting free fatty acids levels than wild type counterparts (4).
The adult C/EBP␤Ϫ/Ϫ mice resemble "sparse fur mice" (25) that have a defect in the urea cycle enzyme, ornithine transcarbamylase, and exhibit lower than normal levels of BUN, elevated blood ammonia, and premature hair loss. For this reason, we also investigated alterations in ammonia metabolism in the C/EBP␤Ϫ/Ϫ mice. In 16-h fasted animals, the concentration of BUN was 40% lower in C/EBP␤Ϫ/Ϫ mice compared with control littermates, suggesting derangement in ammonia metabolism or urea production; however, this value returned to control levels after 30 h of fasting (Fig. 4). To confirm a lower rate of flux through the urea cycle, the concentration of ammonia was determined in the plasma of C/EBP␤Ϫ/Ϫ mice that had been fasted for 16 h (Fig. 4, inset). C/EBP␤Ϫ/Ϫ mice had blood ammonia levels of 425 g/dl as compared with 200 g/dl in wild type mice. This suggests that ammonia metabolism in the C/EBP␤Ϫ/Ϫ mice is compromised, an abnormality that may contribute to the premature death of these animals, as noted earlier by Screpanti et al. (9).
Systemic Glucose Production-Because C/EBP␤Ϫ/Ϫ mice have difficulty maintaining normal levels of blood glucose during fasting, systemic glucose production was determined in vivo. In these experiments [ 3 H]glucose was injected into the tail vein of conscious C/EBP␤Ϫ/Ϫ mice that had been fasted overnight, and the rate of dilution of the [ 3 H]glucose was then measured. As shown in Fig. 5, glucose production in the C/EBP␤Ϫ/Ϫ mice was half that of control littermates, suggesting a defect in gluconeogenesis. The levels of insulin were the same in both the wild type and C/EBP␤Ϫ/Ϫ mice (data not shown). The levels of mRNA for two gluconeogenic enzymes, PEPCK and Glc-6-P, were then measured in the livers of C/EBP␤Ϫ/Ϫ mice and control littermates fed a high carbohydrate diet for 1 week and then injected with Bt 2 cAMP (Fig. 6). In livers of normal animals, both hepatic PEPCK and Glc-6-P mRNA levels were repressed by a high carbohydrate diet were induced by the administration of Bt 2 cAMP. The inhibitory response of the PEPCK gene to a high carbohydrate diet and normal induction by Bt 2 cAMP was blunted in the livers of C/EBP␤Ϫ/Ϫ mice in comparison to control animals. In addition, a high carbohydrate diet, rather than repressing Glc-6-P mRNA as in control animals, induced its levels in the livers of the C/EBP␤Ϫ/Ϫ mice. One possible explanation for this finding is that C/EBP␤ is involved in the repression of Glc-6-P gene transcription by carbohydrate but not in the repression of PEPCK gene transcription. Although PEPCK and Glc-6-P share a common set of regulatory signals, they respond in a different manner to high concentrations of glucose (18). Our findings suggest that C/EBP␤ is involved in controlling the response of the gene for Glc-6-P to dietary carbohydrate.
cAMP Metabolism in the C/EBP␤Ϫ/Ϫ A-phenotype and Wild Type Mice-We previously noted that the basal level of cAMP in the livers of C/EBP␤Ϫ/Ϫ mice was about half that of littermates (4). The concentration of hepatic cAMP was 296.68 Ϯ 32.98 pmol/g as compared with 581.63 Ϯ 92.98 pmol/g of liver in control littermates (Fig. 7A). After the administration of glucagon into the portal vein, the concentration of hepatic cAMP increased to 495.32 Ϯ 84.17 pmol/g of liver in C/EBP␤Ϫ/Ϫ mice and to 1162.96 Ϯ 171.11 pmol/g of liver in the wild type mice (Fig. 7A). Although the fold change in the concentration of cAMP was about the same, the absolute level of the cyclic nucleotide in the livers of C/EBP␤Ϫ/Ϫ mice was markedly different from that noted in wild type animals. The relatively lower concentration (50%) of hepatic cAMP in the C/EBP␤Ϫ/Ϫ mice might be due to the inability of the liver to synthesize cAMP at the appropriate rate or might have resulted from an increase in cAMP degradation. To test these possibilities, basal and activated adenylyl cyclase activities were measured in liver plasma membranes of C/EBP␤Ϫ/Ϫ and wild type mice (Table I). In response to glucagon, cholera toxin, or forskolin, and isoproterenol, membranes from the C/EBP␤Ϫ/Ϫ mice synthesized cAMP at the same rate as those from control littermates. This indicates that the capacity to produce cAMP was intact in the livers of these mice and was not altered by a deletion of the gene for C/EBP␤. However, there appeared to be an accelerated rate of degradation of cAMP in levels of these mice.
The degradation of cAMP was tested by first administering the phosphodiesterase (PDE) inhibitors theophylline (a nonselective inhibitor of PDE) or RO 20-1724 (a specific inhibitor of PDE 4) to fed C/EBP␤Ϫ/Ϫ and control mice. The concentration of cAMP was then measured in liver biopsies taken before and after glucagon injection into the portal vein (Fig. 7, panels B and C). After the administration of theophylline or RO 20-1724, the basal levels of cAMP in the livers of C/EBP␤Ϫ/Ϫ mice were the same as control animals, and there was no significant difference in the response of cAMP to glucagon injection.
Because these data suggest an accelerated rate of degradation of cAMP in the livers of C/EBP␤Ϫ/Ϫ mice, we determined the total activity of PDE in liver homogenates of fed mice using 1 M cAMP as substrate. The specific activity of PDE was 56.7 pmol/min/mg of protein in C/EBP␤Ϫ/Ϫ mice as compared with 44.1 in the livers of control animals (Fig. 8). The level of mRNA for PDE 3A, 3B, and 4B was also measured using a liver biopsy taken 1 min after glucagon injection. The results show that the concentrations of PDE 3A and PDE 3B mRNA were 25% higher in C/EBP␤Ϫ/Ϫ mice (p Ͻ 0.01 for PDE 3B); no difference in the levels of PDE 4 was detectable. It is important to note that PDE 3B is the major phosphodiesterase isoform in the liver, and its mRNA was 100-fold higher that of PDE3A. In agreement with these findings, the rate of cAMP degradation in vitro was determined from parallel experiments in which diluted liver homogenates were incubated with 1 M [ 3 H]cAMP. As shown in Fig. 9, the disappearance (degradation) of [ 3 H]cAMP by liver extracts from C/EBP ␤Ϫ/Ϫ mice was more rapid than noted for controls.
Since cAMP levels can affect the activity of PKA, we next determined the levels of both the regulatory (RI␣ and RII␣) and the catalytic (C␣) subunits of PKA in the nuclear (N), cytosolic (C) and particulate (P) fractions of livers from adult C/EBP␤Ϫ/Ϫ and wild type mice by Western blotting (Fig. 10). We found significant changes in the regulatory subunits of PKA in the livers of C/EBP␤Ϫ/Ϫ mice. The concentration of RI␣ was 79% higher in the particulate fraction and 17% higher in the cytosolic fraction as compared with wild type mice, whereas the concentration of RII␣ in the livers of C/EBP␤Ϫ/Ϫ mice was increased by 27% in particulate fraction and 5% in the cytosolic fraction as compared with wild type littermates. In contrast, the concentration of the catalytic subunit of PKA in C/EBP␤Ϫ/Ϫ mice was the same as wild type (Fig. 10). RII␣ antibody reacted with two protein bands, one at 56 kDa and another at 52 kDa; the 56-kDa band is characteristic of a phosphorylated form of RII␣ (28). The concentration of the 52-kDa band (non-phosphorylated RII␣) was higher in the P and C fractions as compared with wild type, whereas the phosphorylated band was similar to wild type mice in the P and C fractions. The potential physiological significance of this result is not clear.
A change in the ratio of regulatory to catalytic subunits of PKA has a profound effect on the total activity of the enzyme at a given concentration of cAMP. In Table II the levels of RI␣, RII␣, and C␣ were measured by scanning the Western blots in Fig. 10. The values for the particulate and cytosolic fractions are represented as percent of wild type. For RII␣, the phospho- rylated (56 kDa band) and the unphosphorylated (52-kDa band) forms of RII␣ were combined when scanned to give the total RII␣. Using the equations of Houge et al. (26), we calculated the increase in cAMP required to give 50% activation of PKA in the cytosolic and particulate fractions of the livers of C/EBP␤Ϫ/Ϫ mice as compare with wild type animals. For example, the observed 79% increase in the RI␣ subunit in the particulate fraction of the livers of C/EBP␤Ϫ/Ϫ mice requires a 33% increase in cAMP for a 50% activation of PKA. With the observed increase in the level of regulatory subunits of PKA, the concentration of cAMP in the livers of C/EBP␤Ϫ/Ϫ mice is critical in determining the total activity of the enzyme. The result could be a failure to fully stimulate many of the cAMPdependent processes vital to the metabolic function of the liver. DISCUSSION The deletion of the gene for C/EBP␤ markedly alters the normal initiation of glucose homeostasis in the immediate perinatal period. Only 50% of the C/EBP␤Ϫ/Ϫ mice (animals with the A phenotype) survive the first hours after birth. C/EBP␤Ϫ/Ϫ mice (B phenotype) display profound hypoglycemia despite the fact that they have higher than normal levels of hepatic glycogen (Fig. 1). Kawai and Arinze (46) demonstrate that in newborn guinea pigs the response of hepatic glycogenolysis to administered glucagon in the first 3-4 h after birth is blunted when compared with the response beyond 4 h. A similar age-dependent response was observed with epinephrine and isoproterenol. In contrast, cAMP itself induced glycogenolysis independent of age, suggesting that the retarded rate of hepatic glycogen mobilization might be due to a delayed responsiveness of the receptor-coupling system in the livers of newborn guinea pigs. The results of the present studies suggest that C/EBP␤Ϫ/Ϫ mice (B phenotype) do not respond appropriately to the normal stimuli that occur at birth and that these can be by-passed by an injection of Bt 2 cAMP immediately after birth.
One possible downstream target is PKA. We show that in C/EBP␤Ϫ/Ϫ mice (A phenotype) there is a 79% increase in the concentration of the RI␣ regulatory subunit and a 27% increase in the concentration of the RII␣ in the particulate fractions of the liver as compared with wild type. There was no difference in the levels of the C␣ subunit of PKA. This change in the R/C ratio could account for the 25% decrease in PKA activity noted in the livers of C/EBP␤Ϫ/Ϫ mice (A phenotype). In fact, a calculation of the required increase in the concentration of cAMP needed to cause a 50% activation of PKA activity increases by 45% for fractions RI␣ ϩ RII␣ combined in the par- FIG. 5. Systemic glucose production by C/EBP␤؊/؊ mice and wild type mice after an overnight fast. C/EBP␤Ϫ/Ϫ mice and control littermates (WT) (8 -12 weeks of age) were fasted overnight, and the rate of hepatic glucose production and concentration of plasma glucose in the plasma were determined as outlined under "Experimental Procedures." Details of the procedure for the determination of systemic glucose output were presented by Liu et al. (4). The values are expressed as the mean Ϯ S.E. for six mice in each group. *, p Ͻ 0.04; **, p Ͻ 0.05.

FIG. 6. Northern blot analysis of hepatic PEPCK and Glc-6-
Pase mRNA in wild type and C/EBP␤؊/؊ mice fed a high carbohydrate diet and treated with cAMP. C/EBP␤Ϫ/Ϫ mice and control littermates (8 -12 weeks of age) were fed a high carbohydrate (CHO) diet for 1 week before the determination of hepatic PEPCK and Glc-6-P mRNA. The mice were killed, the livers were freeze-clamped, and total RNA was extracted. In some of the experiments shown in this figure, Bt 2 cAMP (35 mg/kg of body weight) and theophylline (30 mg/kg of body weight) were administered by intraperitoneal injection, and the mice were killed 2 h later for the determination of PEPCK and Glc-6-Pase mRNA. The level of mRNA was determined by Northern blotting. The hepatic mRNA was standardized against 18 S rRNA in the same tissue using a PhosphoImager® and expressed as a ratio. The values are expressed as the mean Ϯ S.E. for five mice in each group. *, p Ͻ 0.02. ticulate fractions from livers of C/EBP␤Ϫ/Ϫ mice as compared with wild type controls (Table II). A similar pattern of change in the ratio of RI␣ and RII␣ relative to C␣ has been noted in the regenerating liver (27). In the first 36 h of regeneration after 70% hepatectomy, the levels of both RI␣ and RII␣ increase 30 to 50%, whereas the concentration of C␣ in the liver remains constant. This leads to a disproportion between the R and C subunits of PKA that diminishes the concentration of C␣ during the cAMP burst that occurs with liver regeneration (27). The increase in the R subunit during liver regeneration has been interpreted as a response to the increase in cAMP, since the elevation of the R subunit of PKA may be a method of down-regulating PKA activity (hysteresis).
The observed alterations on the relative location of the isoforms of the R subunit of PKA in the livers of C/EBP␤Ϫ/Ϫ mice may also be of significance. RI␣ is the predominant regulatory subunit of PKA in the cytosol of hepatocytes, whereas RII␣ predominates in the cytoskeleton, the Golgi apparatus, microtubules, and nucleus (28). We have noted an increase in the concentration of both of the R subunits of PKA in the particulate fraction of liver cell and a decrease in the presence of these proteins in the cytosolic fraction (Fig. 10). In contrast to our results, Ekanger et al. (27), in their studies of regenerating liver, note that the change in the increased concentrations of the RI␣ and RII␣ (relative to the C␣) was constant in the cytosol and particulate fractions of the liver. The reason for the selective increase that we observe in the particulate fraction of C/EBP␤Ϫ/Ϫ mice is not clear, but it could be related to the concentration of PKA-anchoring protein in the liver. Anchoring proteins bind specifically to RII and control the movement of PKA to the particulate fraction of the liver, thus partitioning its activity in the cell (29). The concentration of anchoring protein in the livers of C/EBP␤Ϫ/Ϫ mice has not been determined.
Alterations of the concentration of the regulatory subunits in various fractions of the liver have profound implications in the response of PKA to cAMP. For example, O'Brien et al. (30) demonstrate that dietary protein restriction or reduction of the caloric content of the diet resulted in a loss of RI␣ in the cytosol of rat livers and an increase in the amount of the RII␣ subunit. This was accompanied by a sharp reduction in the level of the catalytic subunit of PKA in the particulate fraction of the liver cell. In addition, the activation of glycogen phosphorylase and the phosphorylation of the cAMP regulatory element-binding protein (CREB) by glucagon was lower in hepatocytes isolated from rats fed a 0.5% protein diet as compared with control animals that had been fed a standard diet containing 15% protein (30). This dietary-induced shift in the ratio of the RI␣ and RII␣ subunits of PKA in the liver would explain the blunted response of these animals to glucagon, despite the fact that the concentration of hepatic cAMP is the same as wild type mice, since the two subunits of PKA have different affinities for cAMP (31).
Although the C/EBP␤Ϫ/Ϫ mice (A phenotype) survive to adulthood, they display critical metabolic abnormalities. They have pronounced hypoglycemia associated with fasting and an FIG. 7. Effect of glucagon, theophylline, and RO-20-1724 on the concentration of cAMP in the livers of adult C/EBP␤؊/؊ mice and wild type mice. Panel A, fed C/EBP ␤Ϫ/Ϫ mice and control littermates (WT) (8 -12 weeks of age) were anesthetized with avertin, the liver was clamped, and a biopsy was taken for the measurement of basal cAMP. Glucagon (50 g/kg of body weight) was then injected via the portal vein, and the liver was biopsied 1 min later for the assay of glucagon-induced cAMP. The data in this panel were redrawn from Liu et al. (4). *, p Ͻ 0.05. Panel B, fed C/EBP ␤Ϫ/Ϫ mice and control littermates (WT) (8 -12 weeks of age) were administered theophylline (30 mg/kg of body weight) by intraperitoneal injection. Thirty min later, the animals were anesthetized with avertin, the liver was clamped, and a biopsy was taken for the measurement of basal cAMP. Glucagon (50 g/kg of body weight) was then injected via the portal vein, and the liver was biopsied 1 min later for the assay of glucagon-induced cAMP. Panel C, the protocol was the same as in B except that the RO 20-1724 (15 mg/kg of body weight) was injected instead of theophylline. The values are expressed as the mean Ϯ the S.E. of the mean for 6 mice in each group. impaired hepatic glucose production (4). There was a blunted rate of hepatic glucose production caused by glucagon injection into 18-h fasted mice that were infused with somatostatin to clamp the insulin and glucagon output from the pancreas (4). The level of cAMP in the livers of C/EBP␤Ϫ/Ϫ mice (A phenotype) was about half that noted in control littermates, and the response to glucagon was also less robust. In addition, these mice have a lower rate of epinephrine-induced release of free fatty acids from epididymal adipose tissue in vitro (4). This may explain the lower concentration of blood ketone bodies noted in the blood of C/EBP␤Ϫ/Ϫ mice after fasting (Fig. 3). In addition to alterations in the rate of hepatic glucose output, insulin sensitivity in the C/EBP␤Ϫ/Ϫ mice was greater, resulting in a rate of whole body glucose disposal that was 77% higher than noted in control littermates (32). This is in part due to an increased response of muscle from the C/EBP␤Ϫ/Ϫ mice to insulin stimulation; the insulin-stimulated phosphorylation of the insulin receptor and phosphatidylinositol 3-kinase activities as well as insulin receptor kinase substrate-1 and Akt-Ser 473 were all about 2-fold greater in the skeletal muscle of the C/EBP ␤Ϫ/Ϫ mice as compared with littermates (32). This suggests that the marked drop in the concentration of blood glucose in the C/EBP ␤Ϫ/Ϫ mice (A phenotype) during fasting is due in part to an accelerated rate of removal of glucose by muscle as well as a diminished rate of gluconeogenesis. This may also contribute to the profound hypoglycemia noted in C/EBP␤Ϫ/Ϫ mice (B phenotype) in the immediate perinatal period.
The concentration of cAMP in tissues is regulated not only by adenylyl cyclase but also by the activity of the various isoforms of PDE isozymes (33). The PDE families of enzymes are comprised of multiple isoforms within each family generated from alternative splicing of their precursor RNA. For example, the PDE3 family consists of PDE3A and PDE3B (34). PDE3A has been identified in smooth muscle, platelets, and cardiac tissue, whereas PDE3B is most abundant in adipocytes and liver. PDE4 is the largest member of the PDE families and is derived from at least four different gene PDE4 products (35). However, there is little information available concerning the tissue specificity of the members in this family (36). Our data show a 25% increase in both mRNA levels for PDE3A and PDE3B and a 25% increase in PDE activity in the livers of fasted C/EBP␤Ϫ/Ϫ mice. It is known that interleukin-3 and -4 activate PDE3 in FDCP2 myeloid cells (37), and the concentration of interleukin-6 increased in fasted C/EBP␤Ϫ/Ϫ mice (9). It is intriguing to speculate that the increased PDE activity observed in the C/EBP␤ mice is accomplished through the increase of interleukin-6, which could in turn cause a cascade effect through the insulin receptor substrate 2 (IRS-2), phosphatidylinositol 3-kinase, protein kinase B pathway and ultimately affect PDE3B activity. Another possibility would be a direct effect of C/EBP␤ on the promoter for the PDE3A and PDE3B genes. Little is known about the transcriptional regulation of PDE genes. It is known that cAMP down-regulates the expression of the gene for PDE3 (38). It is thus possible that in the absence of C/EBP␤ there is decreased inhibition of gene transcription, leading to FIG. 8. The activity of phosphodiesterase and the level of PDE 3A, PDE3B, and PDE4B in the livers of C/EBP␤؊/؊ mice and wild type mice. Livers from fed C/EBP␤Ϫ/Ϫ and control littermates (WT) (8 to 12 weeks of age) were analyzed for total phosphodiesterase activity. Fed C/EBP␤Ϫ/Ϫ mice and control littermates were anesthetized with avertin, glucagon (50 g/kg of body weight) was then injected via the portal vein, and the liver was biopsied 1 min later for the measurement of mRNA for PDE3A, PDE3B, and PDE4B as described under "Experimental Procedures." The concentration of PDE mRNA was assessed using competitive reverse transcription-PCR, as described under "Experimental Procedures." Values for PDE mRNA are expressed as a fold-change from the level of PDE3A mRNA in the livers of wild type mice. To quantify the relative abundance of the levels of PDE mRNA, values for PDE3A (WT) were designated as equal to 1. This gave a ratio of 1:129:2.9 for PDE3A, PDE3B, and PDE4B, respectively. The values are expressed as the mean Ϯ S.E. for four animals for PDE3A, three animals for PDE3B, and four animals for PDE 4B. The activity of phosphodiesterase was determined as described under "Experimental Procedures." The activity of phosphodiesterase is expressed as the mean Ϯ S.E. for eight animals in each group (*, p Ͻ 0.02). an accumulation of PDE3 mRNA. This would require that C/EBP␤ be involved as a negative regulator of PDE gene transcription, for which there is no direct information to date. However, a cAMP regulatory binding protein (CREB)-binding site is present in the PDE3B gene promoter 2 ; this might serve as a binding site for C/EBP␤, as occurs with the PEPCK gene promoter (39). A third possibility is that C/EBP␤ may directly regulate one of the G proteins in the adenylyl cyclase pathway. This is unlikely since we found no changes in the relative levels of G s ␣ 1 , G s ␣ 2 , G i ␣ 2 , G q ␣, G␤ 2 , and G␤ 1 in membranes of livers of C/EBP␤Ϫ/Ϫ mice and control littermates, as determined by Western blotting (data not shown). This result agrees with the data on adenylyl cyclase activation in the same liver membranes (Table I).
Abnormalities in the C/EBP␤Ϫ/Ϫ mice are not limited to carbohydrate and lipid metabolism; they also extend to amino acid metabolism. The concentration of BUN in C/EBP␤Ϫ/Ϫ mice after 16 h of fasting is half that of control littermates, a value that is reflected in the 2-fold increase in blood ammonia in the C/EBP␤Ϫ/Ϫ mice. After 30 h of fasting, the concentration of BUN decreases to near normal levels, indicating a sparing of amino nitrogen characteristic of prolonged starvation. The levels of BUN have been shown to increase in humans during the first few days of fasting and to return to normal levels by the end of 2 weeks (40). Rats also spare body protein; this is reflected in a decreased concentration of BUN at 24 h of fasting (41)(42)(43). We have determined the concentration of 20 amino acids in the blood of C/EBP ␤Ϫ/Ϫ mice and control littermates that were fasted for 30 h. The most notable difference was a 3-fold increase in the concentration of taurine relative to control littermates (1035 mol/liter versus 333 mol/ liter), a 2-fold increase in ornithine (127 mol/liter versus 64 mol/liter), and a 30% increase in glutamine (647 mol/liter versus 452 mol/liter). The higher levels of ammonia, ornithine, and glutamine in the blood of the C/EBP␤Ϫ/Ϫ mice may reflect a decreased rate of urea cycle activity in the livers of these mice. C/EBP␤ is required for the glucocorticoid induction of transcription of the gene for arginase, a critical urea cycle enzyme (44). It has been proposed that C/EBP␣ is involved in the regulation of basal expression of the gene for arginase and that C/EBP␤ controls the high level of expression of the gene in the adult mouse in response to dietary protein (45). C/EBP␤Ϫ/Ϫ mice have the appearance of sparse fur, a phenotype associated with a defect in the gene for ornithine transcarbamylase caused in part by elevated levels of blood ammonia.
The results of the present paper clearly demonstrate the far-reaching metabolic consequences caused by the absence C/EBP␤. A series of relatively small changes in the rate of hepatic cAMP degradation and a shift in the pattern of expression of the genes for the isoforms of the regulatory and catalytic subunits of PKA result in a lower concentration of cAMP in the 2 V. Manganiello personal communication.   Fig. 10. They represent the various isoforms of PKA in the fractions (C, cytosol; P, particulate) from the liver of C/EBP␤Ϫ/Ϫ mice and are expressed as a percentage of wild type control mice. The required increase in the concentration of cAMP in the livers of C/EBP␤Ϫ/Ϫ mice required for 50% activation of PKA was calculated from the following equations from Houge et al. (26).

RCϩ2cAMP^R-2cAMPϩC
(1) Kϭ͓RC͔͓cAMP͔ 2 /͓R-2cAMP͔͓C͔ (2) ͓cAMP͔ 50% activity ϭ ͓͑R-2cAMP͔K 50% activity ͒ 0.5 liver. When combined with the increased requirement for cAMP to attain activation of PKA in the livers of C/EBP␤Ϫ/Ϫ mice, the result is a failure to respond to glucagon in a normal fashion. Thus, a 25% increase in the activity of PDE3 in the liver ensures a rapid enough removal of hepatic cAMP to cause the animals to have lowered rate of hepatic glucose output in response to fasting and to glucagon administration.