Regulation of the Glucose-6-phosphatase Gene by Glucose Occurs by Transcriptional and Post-transcriptional Mechanisms

To understand how glucose regulates the expression of the glucose-6-phosphatase gene, the effect of glucose was studied in primary cultures of rat hepatocytes. Glu-cose-6-phosphatase mRNA levels increased about 10-fold when hepatocytes were incubated with 20 m M glu- cose. The rate of transcription of the glucose-6-phosphatase gene increased about 3-fold in hepatocytes incubated with glucose. The half-life of glucose-6-phos-phatase mRNA was estimated to be 90 min in the absence of glucose an d 3 h in itspresence. Inhibition of the oxidative and the nonoxidative branches of the pentose phosphate pathway blocked the stimulation of glucose-6-phosphatase expression by glucose but not by xylitol or carbohydrates that enter the glycolytic/gluconeo-genic pathways at the level of the triose phosphates. These results indicate that (i) the glucose induction of the mRNA for the catalytic unit of glucose-6-phospha-tase occurs by transcriptional and post-transcriptional mechanisms and that (ii) xylitol and glucose increase the expression of this gene through different signaling pathways.


Duna Massillon ‡
From the Department of Nutrition, Case Western Reserve University School of Medicine, Cleveland,Ohio 44106 To understand how glucose regulates the expression of the glucose-6-phosphatase gene, the effect of glucose was studied in primary cultures of rat hepatocytes. Glucose-6-phosphatase mRNA levels increased about 10fold when hepatocytes were incubated with 20 mM glucose. The rate of transcription of the glucose-6phosphatase gene increased about 3-fold in hepatocytes incubated with glucose. The half-life of glucose-6-phosphatase mRNA was estimated to be 90 min in the absence of glucose and 3 h in its presence. Inhibition of the oxidative and the nonoxidative branches of the pentose phosphate pathway blocked the stimulation of glucose-6-phosphatase expression by glucose but not by xylitol or carbohydrates that enter the glycolytic/gluconeogenic pathways at the level of the triose phosphates. These results indicate that (i) the glucose induction of the mRNA for the catalytic unit of glucose-6-phosphatase occurs by transcriptional and post-transcriptional mechanisms and that (ii) xylitol and glucose increase the expression of this gene through different signaling pathways.
Glc-6-Pase 1 (EC 3.1.3.9) is a multicomponent protein complex comprising catalytic and transporting entities (1)(2)(3)(4)(5). The complex is tightly associated with the endoplasmic reticulum, and the enzymatic component catalyzes the hydrolysis of glucose 6-phosphate to glucose, a final common step to both the pathways of glycogenolysis and gluconeogenesis. Hepatic Glc-6-Pase activity is effectively regulated by hormonal and nutritional status. For example, fasting and hormones that increase cAMP concentration stimulate its gene expression while refeeding and insulin decrease it (1, 2, 6 -12).
The expression of the genes for several other proteins is regulated by glucose; these include genes for L-type pyruvate kinase (13,14), fatty-acid synthase (15,16), PEPCK (17,18), and the type 2 glucose transporter (GLUT-2) (19,20). The mechanism by which glucose regulates the expression of these genes remains largely unknown. Recently, Kahn and colleagues (21,22) have proposed a signaling pathway model to explain the molecular mechanism by which glucose regulates the expression of the L-type pyruvate kinase. This model consists of the following details: (i) the presence of glucose is sensed in the cell; (ii) this information is transduced by intracellular messengers; (iii) a second messenger, presumably xylulose 5-phosphate, rises and then modulates the activity of protein kinase and protein phosphatase involved in a cascade of phosphorylation/dephosphorylation. This cascade then leads to a modification of the phosphorylation state of the glucoseresponsive complex, followed by an increase in the transcriptional rate of the target gene. In this model, the presence of an active glucokinase that phosphorylates glucose to glucose 6-phosphate (Glc-6-P) is paramount.
The molecular mechanism by which glucose regulates the expression of the Glc-6-Pase gene is currently unknown. Here we show that glucose regulation of the expression of this gene involves metabolism of glucose through the glycolytic pathway, transcriptional activation of the Glc-6-Pase gene promoter, and a decrease in the degradation of Glc-6-Pase mRNA.

EXPERIMENTAL PROCEDURES
Cell Culture-Hepatocytes were isolated from male Harlan Sprague-Dawley rats (200 -250 g) according to the two-step procedure of Berry and Friend (23) as modified by Leffert et al. (24). Cell viability was estimated by trypan blue dye exclusion, and only preparations with a viability of 85% or higher were used. The cells were cultured in 100-mm Petri dishes in RPMI 1640 (Life Technologies, Inc.) medium, supplemented with penicillin (100 units/ml), streptomycin (100 g/ml), fetal calf serum (10% v/v), calf serum (5% v/v), 10 g of insulin/ml (Sigma), and 100 nM dexamethasone (Sigma). After cell attachment (4 h), the medium was replaced with fresh medium supplemented with fetal calf serum (5% v/v), penicillin (100 units/ml), streptomycin (100 g/ml), without insulin or dexamethasone. The cells were then cultured for additional 16 h, at which time the cells were washed and fed the basal culture medium containing 5 mM glucose or other additions, as indicated in the legends to the figures. At the end of the test period, the cells were harvested and washed twice with 10 ml of cold phosphate-buffered saline.
bridization with the 32 P-labeled probe was carried out for 16 h in the same buffer. The membranes were washed twice for 10 min in 2ϫ SSC, 0.1% SDS at room temperature and once in 0.1ϫ SSC, 0.1% SDS for 15 min at 55°C. The membranes were then exposed to Fuji or Kodak x-ray films for 12-48 h at Ϫ80°C, using intensifying screens. 18 S ribosomal RNA was used to correct for loading irregularities. Quantification was done by scanning densitometry, using the one-dimensional analysis software from Kodak.
Nuclear Run-on Assay-The nuclear run-on assay was used to measure gene transcription rates. The procedure was as described by Aulak et al. 2 In brief, plates were washed twice in phosphate-buffered saline and scraped in the same buffer. Cells were pelleted by centrifugation at 600 ϫ g at 4°C for 5 min and then lysed in buffer A (10 mM HEPES (pH 7.4), 10 mM NaCl, 3 mM MgCl 2 , 0.5% Nonidet P-40) and incubated for 7 min at 4°C. Nuclei were isolated by centrifugation at 600 ϫ g at 4°C for 5 min and resuspended in 20 mM HEPES buffer (pH 8.3), containing 5 mM MgCl 2 , 0.1 mM EDTA, and 40% glycerol, and stored at Ϫ80°C until used. Nuclear run-on assays were performed as follows. Frozen nuclei (2 ϫ 10 7 nuclei, 100 l) were added to 100 l of a buffer containing 20 mM HEPES-KOH (pH 8.0), 25% glycerol, 10 mM MgCl 2 , 0.2 mM KCl, 1.2 mM ATP, 0.6 mM CTP, 0.6 mM GTP. After the addition of 40 units/ml RNase inhibitor and 50 Ci of [␣-32 P]UTP, the mixture was incubated at room temperature for 45 min. The labeled RNA was passed through a QIAshredder column (Qiagen Inc., Valencia, CA) and purified with the Qiagen RNeasy kit, according to the manufacturer's protocol. The RNA was resuspended in 100 l of diethyl pyrocarbonate-treated water. The labeled RNA from each sample was denatured and hybridized to dot-blots containing 2 g of purified cDNA fragments or rat genomic DNA immobilized onto nylon filters. Blots were hybridized for 72 h at 55°C in Church buffer (7% SDS, 0.5 M phosphate buffer (pH 7.5), 1% bovine serum albumin). The blots were then washed twice for 10 min in 2ϫ SSC, 0.1% SDS at room temperature and once in 0.1ϫ SSC, 0.1% SDS for 15 min at 55°C. The membranes were then exposed to Fuji or Kodak x-ray films for 12-48 h at Ϫ80°C, using intensifying screens. Quantification was done by scanning densitometry, using the one-dimensional analysis software from Kodak. Slots containing genomic DNA were used to normalize the efficiency of the nuclear run-on assays.
G6PDH and Lactate Assays-G6PDH activity was measured spectrophotometrically by the rate of production of NADPH at 340 nm (25). The assay medium contained 50 mM Tris (pH 8.0), 10 mM MgCl 2 , 2 mM Glc-6-P, and 0.9 mM NADP. Since 6-PGDH, contained in the extract, also produces NADPH under this assay condition, the assay reflects the total activity of both dehydrogenases. Lactate was assayed by the rate of formation of NADH at 340 nm (25).
Statistics-Results are expressed as means Ϯ S.E. Statistical analysis was performed with Student's t test.

Glucose Induces Glc-6-Pase mRNA Production in Primary
Cultures of Rat Hepatocytes-We have previously shown in conscious normal nondiabetic and diabetic rats that glucose is a major stimulator of Glc-6-Pase gene expression, independent of insulin or other hormones (11). To determine the mechanism by which glucose stimulates Glc-6-Pase gene expression, we have studied this effect in primary cultures of rat hepatocytes because cell cultures provide discrete advantages over intact animals in dissecting out the contribution of nutrients without interference from hormonal secretion. Primary cultures provide an additional advantage in the sense that the cells still display the same regulation as seen in vivo, as opposed to transformed cell lines. First we established the optimal conditions to study glucose regulation of Glc-6-Pase in cultured rat hepatocytes. The addition of glucose to the culture medium caused a marked dose-dependent increase in the induction of the mRNA levels for Glc-6-Pase (Fig. 1A) with a maximal induction at 20 mM glucose. Under the culture conditions used, glucose decreases the mRNA levels of the phosphoenolpyruvate carboxykinase (PEPCK) gene (Fig. 1B) as would be expected (12). These results establish that in this primary culture sys- tem, the hepatocytes maintain the glucose regulation as seen in vivo. Fig. 2 shows the time course of glucose-induced accumulation of Glc-6-Pase mRNA. The increase in the mRNA levels was already present after 1 h; peak values were observed at 2-4 h, after which the mRNA level declined rapidly, returning to near basal values after 8 h. Subsequent experiments were therefore performed with cells incubated with 20 mM glucose for 4 h.
Role of Glucose Phosphorylation-In intact animals, metabolism of glucose beyond glucose 6-phosphate (Glc-6-P) is necessary for the glucose-induced stimulation of the expression of the Glc-6-Pase gene (12). This effect is also demonstrable in primary cultures of rat hepatocytes. When glucosamine (2 mM) was used to inhibit glucokinase activity, glucose failed to increase Glc-6-Pase mRNA levels (Fig. 3). Also, 3-ortho-methylglucose, an analog of glucose that is transported into the cell but not metabolized, was unable to increase Glc-6-Pase mRNA levels. Similarly, when hepatocytes were cultured in the presence of 2-deoxyglucose, another glucose analog that is phosphorylated but not metabolized further, Glc-6-Pase mRNA levels were not changed. These results mimic the results obtained with intact animals (12) and provide more evidence that glucose needs to be metabolized to stimulate the expression of Glc-6-Pase.
To determine whether substrates that enter the glycolytic pathway downstream of glucose 6-phosphate can stimulate Glc-6-Pase expression, lactate and pyruvate were tested. When a combination of lactate (10 mM) and pyruvate (1 mM) was provided as substrate, no increase in Glc-6-Pase mRNA levels was observed (Fig. 3). Two other sugars, fructose and mannose, that enter the glycolytic pathway at the level of triose phosphate were also tested. Both sugars increased Glc-6-Pase mRNA concentrations in a dose-dependent manner (Fig. 4). Glycerol (2 mM) was also able to elicit a small increase in Glc-6-Pase mRNA; however, higher concentrations (Ͼ2 mM) of glycerol had an inhibitory effect.
Identification of Potential Metabolic Sites Mediating the Effect of Glucose on Glc-6-Pase Gene Expression-Doiron et al. (26) have recently challenged the idea of Glc-6-P (27,28) as the signaling molecule that transduces the effect of glucose in regulating the expression of certain genes. Instead, they propose that xylulose 5-phosphate, a metabolite in the pentose phosphate pathway, may be the major messenger candidate that transduces the effect of glucose. Glycolytic intermediates downstream of triose phosphates, such as phosphoglycerate and phosphoenolpyruvate, have been proposed as well (29). The infusion of xylitol into intact animals mimics the effect of glucose on Glc-6-Pase mRNA (12). Because xylitol is metabolized to xylulose 5-phosphate, we tested whether xylitol can alter Glc-6-Pase mRNA accumulation. At 5 mM, xylitol caused a much higher increase in the abundance of Glc-6-Pase mRNA than that caused by 20 mM glucose (Fig. 5A). In a parallel experiment (Fig. 5B) xylitol inhibited PEPCK gene expression, a finding consistent with previous results with intact animals (12).
We then studied the effect of inhibiting the pentose phosphate pathway at the glucose-6-phosphate dehydrogenase (G6PDH) level, the first committed enzyme of the pathway, with 6-aminonicotinamide (6-AN). When hepatocytes were preincubated with 200 M 6-AN before the addition of glucose, the usual glucose-induced stimulation of Glc-6-Pase mRNA was completely prevented (Fig. 6A). To rule out any possible toxic effect of 6-AN, we measured the release of lactate dehydrogenase in the medium as an indicator of cell injury. We did not detect any lactate dehydrogenase activity during the duration of these experiments (results not shown).
Since the nonoxidative branch of the pentose phosphate pathway is reversible, intermediates from the glycolytic pathway (i.e. fructose 6-phosphate and glyceraldehyde 3-phosphate) could be re-entering this pathway through the reactions catalyzed by transaldolase and transketolase. To evaluate the role of the nonoxidative branch of the pentose phosphate pathway in transducing the observed glucose signaling, we inhibited this portion of the pathway with oxythiamine, a transketolase inhibitor (30,31), the rationale being that this compound would interfere with interconversions in this portion of the pathway. At concentrations that completely inhibit transketolases in vitro (30), oxythiamine did not prevent the accumulation of Glc-6-Pase mRNA by glucose (Fig. 6A), suggesting that metabolism through the nonoxidative portion of the pathway may not be required for the response to glucose. To evaluate the effect of simultaneously inhibiting the oxidative and nonoxidative branches of the pentose phosphate pathway, we incubated the hepatocytes with 6-AN and oxythiamine and then re-assessed the response to glucose and xylitol. The combination of the two inhibitors had no effect on the stimulation of Glc-6-Pase gene expression by xylitol (Fig. 6B), but it prevented glucose from inducing an accumulation of Glc-6-Pase mRNA. Although this effect was not surprising, since 6-AN alone could prevent the glucose effect, it established that a different mechanism underlies the xylitol and glucose induction of this mRNA. When the inhibitor combination was tested on the inductive effects of fructose or glycerol, the combination had no effect on the ability of these substrates to induce Glc-6-Pase mRNA levels (Fig. 6C).
Because it has been reported that 6-AN can act as an inhibitor of the glycolytic enzyme, phosphoglucoisomerase (32, 33), we assayed lactate production (as an indication of glycolytic activity) in cells incubated with 6-AN. Lactate levels were 4-fold higher in cells incubated with 20 mM glucose than in cells incubated with 5 mM glucose. Preincubation of the cells with 200 M 6-AN caused a more than 60% inhibition of glycolysis; this inhibition was evident at the earliest time point (1 h) measured. 6-AN also blocked glycolysis from mannose (Fig.  7B). Since the metabolism of mannose involves its isomerization to fructose 6-phosphate by phosphomannose isomerase, this result suggests that indeed phosphohexoisomerases were inhibited in the presence of 6-AN. Interestingly, mannose-induced accumulation of Glc-6-Pase mRNA was also inhibited (results not shown). Lactate production from xylitol was significantly less compared with glucose. These results could be explained by the well documented blocking effect of xylitol on glycolysis (34,35). Xylitol is known to exert an ATP-depleting action by trapping phosphate inside the cells since the bulk of phosphorylated intermediates following xylitol metabolism enters the gluconeogenic pathway instead of glycolysis (34, 35) so the ATP used is not being regenerated.
Role of Transcriptional and Post-transcriptional Mecha-nisms in Mediating the Effects of Glucose on Glc-6-Pase mRNA-Because the stimulatory effect of glucose on Glc-6-Pase mRNA level can result from either an increase in the transcription of the gene or a stabilization of the mRNA, both of these two possibilities were tested. The effect of glucose on the transcription rate of the Glc-6-Pase gene was measured by using nuclear run-on assays. In the presence of 20 mM glucose, the transcription rate increased 3-fold above the basal value (Fig. 8), a value less than that for Glc-6-Pase mRNA accumulation. To test the possibility that glucose might also stabilize Glc-6-Pase mRNA, the half-life of the Glc-6-Pase mRNA was measured. For this experiment, the transcription inhibitor actinomycin D (10 g/ml) was added to the hepatocytes after 4 h in culture in the presence of 20 mM glucose, and the culture was continued in the presence or absence of glucose. Total RNA was then isolated at various time points, and Glc-6-Pase mRNA abundance was quantified by Northern blot analysis. As shown in Fig. 9, the level of Glc-6-Pase mRNA decayed at a slower rate in the presence (t1 ⁄2 ϭ 3 h) of glucose than in its absence (t1 ⁄2 ϭ 90 min). Thus, we conclude that glucose-stimulated transcription of the Glc-6-Pase gene involves both transcriptional and post-transcriptional mechanisms. To examine whether de novo protein synthesis is required for the glucose-induced increase in Glc-6-Pase mRNA, cycloheximide (10 g/ml) was used to block protein synthesis. We found that glucose-induced Glc-6-Pase mRNA expression was similar in the presence or in the absence of cycloheximide (data not shown), indicating that glucose-induced Glc-6-Pase mRNA accumulation does not require the synthesis of new protein. DISCUSSION The molecular mechanisms underlying glucose-induced regulation of the Glc-6-Pase gene is not clearly understood. Here we present evidence that stimulation of Glc-6-Pase gene expression by glucose involves both transcriptional and posttranscriptional components. The specific DNA sequence(s) responsible for the transcriptional response remains to be established. The half-life of Glc-6-Pase mRNA was found to be 90 min. The fact that glucose prolongs this half-life suggests glucose-induced stabilization of the Glc-6-Pase mRNA. New protein synthesis does not seem to be necessary for glucose to induce Glc-6-Pase mRNA. This would argue against a mechanism whereby glucose stimulated the synthesis of a protein factor(s) that bind to the mRNA and thereby prevent its degradation. The stabilization of the Glc-6-Pase mRNA by glucose does not depend on the ability of this mRNA to be translated, since the 2-fold increase in the level of Glc-6-Pase protein was much lower than the increase in its mRNA. It has been extensively documented that mRNAs whose 3Ј-UTRs contain an AU-rich element (AURE) and/or an oligonucleotide (U) tend to be unstable (36). Such AURE in the 3Ј-UTR had been found in transcripts such as c-fos and c-myc (37,38) that are known to have a short half-life. Their significance in mRNA turnover has been demonstrated in vivo and in vitro (38 -41). The 3Ј-UTR of the Glc-6-Pase gene contains a few AUREs. Whether these AUREs play any role in Glc-6-Pase mRNA stability is not known.
One unresolved question about the glucose-induced accumulation of Glc-6-Pase mRNA is whether this effect requires glucose metabolism through the pentose phosphate pathway. If this pathway is important for the effect, then suppressing it should be expected to prevent the accumulation of Glc-6-Pase mRNA. Glucose-induced accumulation of the mRNA for Glc-6-Pase was indeed prevented when hepatocytes were incubated with 0.2 mM 6-AN. When metabolized to 6-amino-NAD and 6-amino-NADP, this nucleotide behaves as a competitive inhibitor of NAD(P ϩ )-requiring dehydrogenases that would include G6PDH, 6-PGDH, and glutathione reductase (32,33,(42)(43)(44)(45)(46)48). 6-Amino-NADP is an extremely potent competitive inhibitor of 6-PGDH, and concentrations that do not affect G6PDH may completely block 6-PGDH (43). On the other hand, no inhibition of NAD-dependent enzymes results from the use of this compound (48). The inhibition of glycolysis from glucose by 6-AN makes it difficult to interpret the results using glucose as an energy source. Nevertheless, this compound is very valuable to study the effect of substrates such as fructose and glycerol that enter glycolysis beyond the formation of fructose 6-phosphate.
When both the pentose phosphate pathway and phosphoglucoisomerase were inhibited with 6-AN and oxythiamine, stimulation of Glc-6-Pase gene expression by fructose or glycerol was not affected. These two substrates enter the glycolytic pathway at the level of triose phosphate(s), and the fact that the nonoxidative branch of the pentose phosphate pathway was blocked suggests that the pentose phosphate pathway is not required for the glucose effect on the Glc-6-Pase gene. In contrast to glucose, xylitol induced Glc-6-Pase mRNA even in the presence of inhibitors of the pentose phosphate pathway. Taken together, these results support the idea that xylitol and glucose signal through different pathways.
The signaling pathway from glucose to the DNA sequences termed carbohydrate responsive element (ChoRE/GIRE) is not totally characterized. Phosphorylation/dephosphorylation of transcription factors have been implicated in the glucose responsiveness of many genes. Xylulose 5-phosphate, produced by the pentose phosphate pathway, has been suggested as a secondary messenger in sensing glucose concentration in the hepatocyte. This proposal stems from the fact that xylitol, a precursor of xylulose 5-phosphate, is able to stimulate the expression of a number of genes, both in vivo (12) and in cultured cells (26,28). Xylulose 5-phosphate has been shown to activate the phosphatase 2A-mediated dephosphorylation (49, 50) of fructose-6-phosphate,2-kinase:fructose-2,6-bisphosphatase and to decrease the activity of protein kinase A. The same phosphatase is also involved in the dephosphorylation of the transcription factor Sp1 (51,52). Xylulose 5-phosphateactivated phosphatase is also activated by glucose or a glucose metabolite (50). Furthermore, glucose-induced transcription of the acetyl-CoA carboxylase gene is mediated by the transcription factor Sp1 (52). This stimulation is prevented by okadaic acid, an inhibitor of protein phosphatase types 1 and 2A (52). On the other hand, okadaic acid only partially inhibits the action of glucose and xylitol to stimulate the fructose-6-phosphate,2-kinase:fructose-2,6-biphosphatase (53). These results suggest that either dephosphorylation of Sp1 is not universal in the glucose signaling pathway or that glucose signaling also involves other transcription factors such as the upstreamstimulatory factor proteins (54). These upstream stimulatory factor proteins were among the first transcription factors suggested to link glucose action to the ChoRE/GIRE (54,55). The involvement of protein phosphatase(s) in the modulation of transcription factors involved in glucose signaling is worth exploring further.
The metabolic consequences of increased Glc-6-Pase gene expression by glucose are not known. More than 25 years ago, Nordlie and colleagues (47) proposed the controversial idea that Glc-6-Pase, in the presence of high glucose levels, might act as a phosphotransferase that uses carbamoyl phosphate as the phosphate donor. Whether Glc-6-Pase plays any role in glucose phosphorylation is debatable. Nonetheless, the para-doxical induction of Glc-6-Pase gene by glucose points toward a potential role of this enzyme in the removal of glucose from the circulation. Glucose stimulation might also be needed to make sure that the enzyme does not disappear when the cell shifts from fed to starved conditions. One speculation is that glucose stimulation of Glc-6-Pase gene may take place in vivo when glucose cycling is needed following hyperglycemia. Glc-6-Pase is a very complex protein with different protein components that are not yet well characterized.