Overexpression of the P46 (T1) translocase component of the glucose-6-phosphatase complex in hepatocytes impairs glycogen accumulation via hydrolysis of glucose 1-phosphate.

The final step of gluconeogenesis and glycogenolysis is catalyzed by the glucose-6-phosphatase (Glc-6-Pase) enzyme complex, located in the endoplasmic reticulum. The complex consists of a 36-kDa catalytic subunit (P36), a 46-kDa glucose 6-phosphate translocase (P46), and putative glucose and inorganic phosphate transporters. Mutations in the genes encoding P36 or P46 have been linked to glycogen storage diseases type Ia and type Ib, respectively. However, the relative roles of these two proteins in control of the rate of glucose 6-phosphate hydrolysis have not been defined. To gain insight into this area, we have constructed a recombinant adenovirus containing the cDNA encoding human P46 (AdCMV-P46) and treated rat hepatocytes with this virus, or a virus encoding P36 (AdCMV-P36), or the combination of both viruses, resulting in large and equivalent increases in expression of the transgenes within 8-24 h of viral treatment. The overexpressed P46 protein was appropriately targeted to hepatocyte microsomes and caused a 58% increase in glucose 6-phosphate hydrolysis in nondetergent-treated (intact) microsomal preparations relative to controls, whereas overexpression of P36 caused a 3.6-fold increase. Overexpression of P46 caused a 50% inhibition of glycogen accumulation in hepatocytes from fasted rats incubated at 25 mm glucose relative to cells treated with a control virus (AdCMV-betaGAL). Furthermore, in hepatocytes from fed rats cultured at 25 mm glucose and then exposed to 15 mm glucose, AdCMV-P46 treatment activated glycogenolysis, as indicated by a 50% reduction in glycogen content relative to AdCMV-betaGAL-treated controls. In contrast, overexpression of P46 had only small effects on glycolysis, whereas overexpression of P36 had large effects on both glycogen metabolism and glycolysis, even in the presence of co-overexpressed glucokinase. Finally, P46 overexpression enhanced glucose 1-phosphate but not fructose 6-phosphate hydrolysis in intact microsomes, providing a mechanism by which P46 overexpression may exert its preferential effects on glycogen metabolism.

The glucose-6-phosphatase enzyme complex catalyzes the final step of gluconeogenesis. The complex is composed of a catalytic subunit of 36 kDa (P36) 1 sequestered within the endoplasmic reticulum (ER), a 46-kDa glucose-6-phosphate translocase known as P46 or T1 that delivers glucose 6-phosphate to the catalytic subunit, and putative ER glucose and inorganic phosphate (P i ) transporters (T2 and T3) that move the reaction products back into the cytosol (1)(2)(3). However, only P36 and P46 have been clearly identified and cloned (4 -7). It remains uncertain whether P i transport is embodied in the function of P46 or encoded by a separate protein (8). Furthermore, an earlier report describing an ER-localized glucose transporter, , has recently been retracted (10).
Mutations in P36 and P46 have both been linked to glycogen storage diseases in human subjects (5,11). Patients with type Ia glycogen storage disease have mutations in the P36 gene (5) and a complete deficiency of glucose-6-phosphatase enzymatic activity, regardless of whether the assay is performed in intact or detergent-disrupted microsomal preparations (12). Patients with type Ib glycogen storage disease have mutations in the P46 gene (11) and have absent or reduced glucose-6-phosphatase enzymatic activity in intact microsomes but normal or increased activity in detergent-disrupted preparations (13).
Whereas the human genetic studies clearly demonstrate that both the P36 and P46 gene products are required for normal glucose 6-phosphate (Glc-6-P) hydrolysis, the relative contributions made by these proteins to the hydrolytic rate, and to overall regulation of carbohydrate metabolism, are incompletely understood. One way of gaining insight into this issue is to overexpress selectively single components of the glucose-6phosphatase complex. Thus, our laboratory has previously used recombinant adenovirus to overexpress P36 in INS-1 insulinoma cells (14), primary hepatocytes (15), or liver of normal rats (16). Expression of P36 caused clear increases in glucose 6-phosphate hydrolysis in both cultured cell models (14,15), and in intact rats, resulted in glucose intolerance, mild hyperinsulinism, and a 50% decrease in hepatic glycogen stores (16). However, the extent of overexpression of P36 in the INS-1 study (10-fold) was larger than the increment in glucose production measured by a 3 H 2 O incorporation assay (4-fold), raising the possibility that other components of the complex such as P46 might contribute to the overall rate at which Glc-6-P is hydrolyzed.
The purpose of the current study was to investigate the potential regulatory role of P46 in the glucose-6-phosphatase complex by adenovirus-mediated overexpression of the protein in rat hepatocytes. We find that overexpression of P46 increases Glc-6-P hydrolysis in intact microsomes, although not to the same extent as overexpression of the catalytic subunit.
Overexpression of P46 also causes pronounced inhibition of glycogen synthesis and activation of glycogenolysis, but has only small effects on glycolysis, whereas overexpression of P36 has potent effects on both pathways. The preferential effect of P46 overexpression on glycogen metabolism may be related to its capacity to enhance the hydrolysis of a hexose phosphate intermediate of glycogen metabolism, glucose 1-phosphate.
Hepatocyte Isolation, Culture, and Viral Treatment-Hepatocytes were isolated from ad libitum fed or overnight (18 h) fasted male Wistar rats (180 -225 g) using collagenase perfusion (21). All reagents were purchased from Sigma unless stated otherwise. Cells were suspended in Dulbecco's modified Eagle's medium supplemented with 25 mM glucose, 10% fetal bovine serum, 100 nM insulin (Life Technologies, Inc.), 100 nM dexamethasone, and 2% penicillin/streptomycin (attachment medium) and plated at a density of 7 ϫ 10 4 cells/cm 2 in either 12-well or 6-cm tissue culture plates pre-coated with 0.1% collagen solution at 37°C for 2 h. For studies of transgene expression and P46 subcellular localization or studies of glycogen synthesis, lactate production, or glucose usage, hepatocytes from overnight fasted rats were treated with various adenoviruses immediately after attachment or were left untreated and then incubated with culture medium (consisting of Dulbecco's modified Eagle's medium supplemented with 0.2% bovine fraction V albumin, 1 nM insulin, 10 nM dexamethasone, 2% penicillin/streptomycin) containing either 5 or 25 mM glucose for 8 or 24 h. For studies on glycogen degradation, hepatocytes from ad libitum fed rats were incubated with culture medium containing 25 mM glucose for 24 h and were then incubated with adenoviruses added singly or in combinations or were left untreated for 2 h at 37°C. Virus-containing medium was then removed, and cells were cultured for an additional 24 h in culture medium containing 15 mM glucose. For all experiments, viral particle numbers were estimated by measurement of A 260 of viral stocks, and for each virus, 500 particles/cell were added.
RNA Blot Hybridization Analysis-Total RNA was isolated from hepatocytes by extraction with the TRIZOL reagent (Life Technologies, Inc.). 10 g of total RNA was analyzed using a procedure described previously (22). A 1459-base pair polymerase chain reaction product derived from the P46 cDNA (7) was randomly labeled using the Random Primers DNA labeling system (Life Technologies, Inc.) and was used to detect P46 mRNA by blot hybridization, whereas a 989-base pair polymerase chain reaction product derived from the P36 cDNA was similarly labeled and used to detect expression of P36 mRNA.
Immunoblot Analysis-Hepatocytes were homogenized using a Brinkman homogenizer in homogenization (H) buffer containing 10 mM Tris/HCl (pH 7.0), 150 mM KF, 15 mM EDTA, 600 mM sucrose, 15 mM 2-mercaptoethanol, 10 g/ml leupeptin, 1 mM benzamidine, and 1 mM phenylmethylsulfonyl fluoride. Immunoblot analysis was performed on the total cell homogenate or on various subcellular fractions prepared as described previously (15). Briefly, total cell homogenate was centrifuged at 10,000 ϫ g for 15 min at 4°C, and the pellet was resuspended in H buffer. The supernatant from this spin was recovered and recentrifuged at 105,000 ϫ g for 1 h at 4°C. The pellet was resuspended in H buffer as the microsomal fraction, and the supernatant was recovered as the cytosolic fraction. 25-75 g of protein from each fraction was then subjected to polyacrylamide electrophoresis. Blots were incubated with an anti-P36 antibody (23; a gift of Dr. Rebecca Taub, University of Pennsylvania), or an anti-peptide antibody specific for human P46 (24). After incubation with the primary antibodies, blots were treated with anti-rabbit Ig horseradish peroxidase-linked secondary antibody from sheep (Amersham Pharmacia Biotech).

Measurement of Hexose Phosphate Hydrolysis in Hepatocyte
Microsomes-Hepatic microsomes were collected as described above for immunoblot analysis, except that a glass homogenizer was used for preparation of the initial hepatocyte homogenate. The resuspended microsomal fraction was incubated in the presence or absence of 0.5% cholic acid at 4°C for 20 min. Reaction buffer containing 60 mM sodium cacodylate and 10 mM glucose 6-phosphate, 10 mM mannose 6-phosphate, 10 mM glucose 1-phosphate, or 10 mM fructose 6-phosphate was added to each microsomal sample. Hydrolysis reactions were carried out at 37°C for 9 min, based on time course studies showing linearity of the reaction at this time and substrate concentration (data not shown). Reactions were terminated by addition of trichloroacetic acid (final concentration, 4%). After centrifugation at 3000 rpm for 5 min, the inorganic phosphate concentration was assayed in the supernatant using a kit from Sigma. Results were normalized to the total protein content of the microsomal sample.
Glycogen and Lactate Assays-At the conclusion of each experiment, culture medium was collected for measurement of lactate production, using a kit and protocol from Sigma. Hepatocytes were washed with Dulbecco's-phosphate-buffered saline once and scraped into 30% KOH solution, and the extracts were incubated in boiling water for 15 min. Glycogen content was measured as described (25).
Glucose Usage Assay-Glucose usage was measured as the conversion of [3-3 H]glucose to 3 H 2 O as described previously (15,26). Briefly, fasted hepatocytes were treated with adenoviruses immediately after attachment and then incubated in culture medium containing 3 mM glucose overnight. Cells were washed once with phosphate-buffered saline and then incubated with a Hanks' balanced salt solution containing 5 or 25 mM glucose containing tracer [3-3 H]glucose (Amersham Pharmacia Biotech, 10 Ci/nmol) at 37°C for 4 h. At the conclusion of this incubation, 3 H 2 O was collected, and the radioactivity was analyzed using a Beckman Counter. To monitor the effect of this treatment on glycogen content, another set of experiments was conducted under identical conditions, except that the cells were collected for measurement of glycogen content.
Statistical Analyses-At least three independent experiments were performed for each assay. Data were analyzed using the two-tailed Student's t test.

Expression of P46 with or without Co-overexpression of P36
in Isolated Hepatocytes-A recombinant adenovirus containing the cDNA encoding human P46 (AdCMV-P46) was constructed and used in conjunction with a previously prepared virus containing the cDNA encoding P36 from rat (AdCMV-P36, previously known as AdCMV-Glc-6-Pase; Ref. 14). As shown in Fig.  1, treatment of isolated hepatocytes with either or both of these viruses resulted in large increases in mRNA levels for the respective transgenes. Adenovirus-mediated transgene expression was clearly evident at 8 h after viral treatment and increased further between 8 and 24 h. Furthermore, co-treatment of hepatocytes with the AdCMV-P46 and AdCMV-P36 viruses resulted in robust coexpression of the two transgenes. To control for untoward effects of addition of two viruses, some cell batches were co-incubated with AdCMV-P46 and AdCMV-␤GAL, and such cells exhibited identical levels of P46 expression as in cells treated with AdCMV-P46 alone.
The glucose-6-phosphatase enzyme complex resides in the ER (1-3). To determine whether the increase in P46 mRNA shown in Fig. 1 corresponds to equivalent increases in P46 protein targeting to the ER, we prepared microsomal fractions of AdCMV-P46-treated and control cells and measured P46 protein by immunoblot analysis (Fig. 2A). Treatment of hepatocytes with AdCMV-P46 resulted in 6.1-9.9-fold increases in P46 protein levels in the microsomal preparations. We also performed subcellular fractionation of cells treated with Ad-CMV-P46 or AdCMV-␤GAL. As shown in Fig. 2B, AdCMV-P46 treatment caused a clear increase in P46 protein expression in total cell extracts (T) and in the microsomal fraction (M), with little or no increase in the pellet of a 10,000 ϫ g centrifugation (P) or in the cytosolic fraction (supernatant of 100,000 ϫ g spin; C). These findings suggest that most of the transgene-encoded P46 protein was correctly targeted to the ER.
A tight functional linkage between P46 and P36 has previously been demonstrated, such that knock-out of P36 in mice results in ablation of microsomal glucose 6-phosphate transport (27). These findings suggest that P46 and P36 could be physically associated in the ER, raising the possibility that overexpression of P46 might affect the stability, and thereby the expression level, of P36. To test this, we performed immunoblot analysis on microsomal preparations from cells treated with AdCMV-P46, AdCMV-P36, or both viruses. As shown in Fig. 2, cells treated with AdCMV-P46 or the combination of AdCMV-P46 ϩ AdCMV-␤GAL had the same levels of P36 protein as untreated cells or cells treated with AdCMV-␤GAL alone. Furthermore, the increase in P36 protein was equivalent in cells treated with AdCMV-P36 alone, AdCMV-P36 ϩ Ad-CMV-␤GAL, or AdCMV-P36 ϩ AdCMV-P46. Thus, overexpression of P46 had no effect on endogenous or overexpressed P36 protein levels.
Effect of P46 Overexpression on Glucose 6-Phosphate and Mannose 6-Phosphate Hydrolysis-Since overexpressed P46 is clearly targeted to hepatocyte microsomes (see Fig. 2), we next evaluated the effect of overexpression of this protein on glucose 6-phosphate (Glc-6-P) and mannose 6-phosphate (Man-6-P) hydrolysis via measurement of P i production in intact and detergent-disrupted microsomes. Comparison of hexose phosphate hydrolysis in intact versus detergent-treated microsomes allows us to differentiate between the activity of the intact system, which is a function of the combined actions of P36 and P46, compared with detergent-treated samples, which measure total phosphohydrolase (P36) activity. As shown in Fig. 3A, microsomes isolated from control hepatocytes (untreated or AdCMV-␤GAL-treated) exhibited appropriate latency for Man-6-P hydrolysis, in that P i production in intact microsomes was only 33-38% that in detergent-disrupted microsomes. Treatment of cells with either AdCMV-P46 or AdCMV-P46 ϩ Ad-CMV-␤GAL caused a 60% increase in Man-6-P hydrolysis compared with untreated control cells or AdCMV-␤GAL-treated controls in intact microsomes (nondetergent-treated) but had no effect on total Man-6-P phosphohydrolase activity in disrupted microsomes (detergent-treated). In contrast, treatment of hepatocytes with AdCMV-P36 resulted in a 3.6-fold increase in Man-6-P hydrolysis in intact microsomes relative to either control group, but also a 4.1-4.8-fold increase in total Man-6-P phosphohydrolase activity in disrupted microsomes, such that the percentage of Man-6-P hydrolyzed by inact microsomes was still only 30% that hydrolyzed by disrupted microsomes. Finally, co-overexpression of P46 and P36 did not activate Man-6-P hydrolysis further in either intact or disrupted microsomes relative to cells with overexpression of P36 alone.
A different pattern was noted when Glc-6-P was used as substrate instead of Man-6-P (Fig. 3B). In the untreated and AdCMV-␤GAL-treated control groups, Glc-6-P hydrolysis in intact microsomes approached that of disrupted microsomes (75%), due to the specificity of P46 for Glc-6-P. In cells treated with either AdCMV-P46 or AdCMV-P46 ϩ AdCMV-␤GAL, Glc-6-P hydrolysis in intact microsomes was increased by 33-58% relative to either control group, with no change in total Glc-6-P phosphohydrolase activity as measured in disrupted microsomes. Interestingly, these changes mean that Glc-6-P hydrolysis in intact microsomes from AdCMV-P46-treated cells tended to be higher than in disrupted microsomes. Treatment of cells with AdCMV-P36 increased Glc-6-P hydrolysis by 3.6and 4.3-fold in intact and disrupted microsomes, respectively, relative to preparations from untreated or AdCMV-␤GALtreated control cells. Finally, co-treatment of hepatocytes with AdCMV-P46 ϩ AdCMV-P36 caused an additional 23% increase in Glc-6-P hydrolysis in intact microsomes relative to cells treated with AdCMV-P36 alone, with no effect on total phosphohydrolase activity. Taken together, these experiments demonstrate that P46 overexpression in hepatocytes causes a clear increase in Man-6-P and and Glc-6-P hydrolysis in intact but not disrupted microsomal preparations.
Effect of P46 Overexpression on Glycogen Synthesis-To begin to assess the metabolic impact of overexpression of P46, hepatocytes isolated from fasted rats were treated with the various combinations of adenoviruses described in Fig. 1 and   FIG. 1. Overexpression of T1 translocase and Glc-6-Pase in isolated hepatocytes. Hepatocytes isolated from overnight fasted rats were treated with an adenovirus containing the ␤-galactosidase gene, AdCMV-␤Gal (BG), a virus containing the cDNA encoding P46, Ad-CMV-P46 (P46), or an adenovirus containing the cDNA encoding the catalytic subunit of glucose-6-phosphatase, AdCMV-P36 (P36), or with combinations of these viruses, such as AdCMV-P46/AdCMV-␤Gal (P46/BG) and AdCMV-P46/AdCMV-P36 (P46/P36). Other groups of control cells were not treated with virus (null). After the cells were incubated in culture medium containing 5 or 25 mM glucose for 8 or 24 h, total RNA was isolated for Northern blot analysis. A radiolabeled probe specific for the P46 transcript was used in A and C, and a probe specific for the P36 catalytic subunit mRNA was used in B and D.

FIG. 2. Immunoblot analysis of P46 and P36 expression in hepatocytes.
Hepatocytes isolated from overnight fasted rats were treated with different adenoviruses as described and abbreviated in the legend to Fig. 1, and cultured in medium containing 25 mM glucose for 24 h. A, microsomal fractions were subjected to SDS-polyacrylamide gel electrophoresis and immunoblotted with anti-P36 or anti-P46 antibodies, as described under "Materials and Methods." The blot shown is representative of three independent experiments. B, immunoblot analysis of P46 protein was performed on total cell extract (T), the pellet from an initial 10,000 ϫ g centrifugation (P), cytosol-enriched fraction (C), and microsomal fraction (M), prepared as described under "Materials and Methods." then incubated in the presence of 5 or 25 mM glucose for 8 or 24 h prior to measurement of glycogen content. Fasted rats with depleted hepatic glycogen stores were used in these experiments to allow measurement of glycogen synthesis during the period of cell culture. As shown in Fig. 4, overexpression of P36, P46, or both proteins had no significant effect on glycogen storage at 5 mM glucose at either 8 or 24 h. However, in cells incubated at 25 mM glucose for 8 h, each of the 4 experimental groups (AdCMV-P46 alone, AdCMV-P36 alone, AdCMV-P46 ϩ AdCMV-␤GAL, and AdCMV-P46 ϩ AdCMV-P36) had similar decreases (31-37%) in glycogen stores relative to the two control groups (untreated hepatocytes or hepatocytes treated with AdCMV-␤GAL alone), although these decreases only achieved statistical significance in the AdCMV-P46 and the AdCMV-P46 ϩ AdCMV-␤GAL-treated groups. Results were more dramatic after 24 h of culture at 25 mM glucose. Thus, AdCMV-P46 or AdCMV-P46 ϩ AdCMV-␤GAL-treated cells had statistically significant 48 and 49% reductions in glycogen content relative to the untreated or AdCMV-␤Gal-treated control groups, respectively, whereas treatment with AdCMV-P36 caused slightly larger (71 and 62%) decreases. Interestingly, combined treatment with AdCMV-P36 ϩ AdCMV-P46 appeared to have an additive effect, causing 84 and 80% decreases in glycogen content relative to the untreated and AdCMV-␤GAL controls, respectively.
Effect of P46 Overexpression on Glycogen Degradation-To determine whether overexpression of components of the glucose-6-phosphatase enzyme complex are capable of enhancing the rate of glycogen degradation, glycogen-replete hepatocytes were isolated from fed rats, treated with the various recombinant adenoviruses, and cultured for 24 h in the presence of 25 mM glucose. Cells were then cultured an additional 24 h in 15 FIG. 4. Effect of P46 overexpression on glycogen synthesis. Hepatocytes isolated from overnight fasted rats were left untreated (null), or were treated with AdCMV-␤Gal (BG), AdCMV-P46 (P46), or AdCMV-P36 (P36) alone, or with combinations of viruses such as Ad-CMV-P46/AdCMV-␤Gal (P46/BG) and AdCMV-P46/AdCMV-P36 (P46/P36). After incubation in culture medium containing 5 or 25 mM glucose for 8 or 24 h, cells were collected for measurement of glycogen content. Data represent the mean Ϯ S.E. for five independent experiments. The various experimental groups were compared with the Ad-CMV-␤Gal-treated control group and were found to be different with the following levels of significance: *, p Ͻ 0.05; #, p Ͻ 0.005; and @, p Ͻ 0.0005.

FIG. 3. Mannose 6-phosphate and glucose 6-phosphate hydrolysis in intact and detergent-disrupted hepatic microsomes.
Hepatocytes isolated from overnight fasted rats were left untreated (null) or were treated with AdCMV-␤Gal (BG), AdCMV-P46 (P46), or AdCMV-P36 (P36) alone, or with combinations of viruses such as AdCMV-P46/AdCMV-␤Gal (P46/BG) and AdCMV-P46/AdCMV-P36 (P46/P36). Microsomes were prepared from overnight fasted rats and incubated in the absence (intact) or presence (detergent-disrupted) of 0.5% cholic acid. These preparations were then incubated in the presence of 10 mM mannose 6-phosphate (A) or 10 mM glucose 6-phosphate (B) for 9 min at 37°C. P i production was measured as an index of hydrolysis of these hexose phosphates, normalized to total microsomal protein in each sample. Data represent the mean Ϯ S.E. for six independent experiments. The symbols * and # indicate those samples for which mannose 6-phosphate or glucose 6-phosphate hydrolysis was increased relative to the corresponding AdCMV-␤GAL-treated control groups, with p Ͻ 0.05. mM glucose, whereupon they were collected for measurement of glycogen content. Importantly, treatment of hepatocytes with the AdCMV-␤GAL virus and culture for 24 h at 15 mM glucose did not cause lowering of glycogen levels relative to untreated cells cultured for 24 h at 25 mM glucose, showing that treatment with a control virus does not activate glycogenolysis in response to lowering of glucose from 25 to 15 mM (Fig. 5). In contrast, cells treated with AdCMV-P46 or AdCMV-P46 ϩ Ad-CMV-␤GAL exhibited a 46% decrease in glycogen content relative to either control group, whereas cells treated with Ad-CMV-P36 or AdCMV-P36 ϩ AdCMV-␤GAL had decreases of 73%. Finally, cells treated with the combination of AdCMV-P36 and AdCMV-P46 had the largest (80%) decline in glycogen content. Thus, both P46 and P36 overexpression stimulate glycogen degradation in glycogen-replete cells, and some additivity of the two effects can be demonstrated.
Effect of P46 Overexpression on Lactate Production in Hepatocytes-We next tested whether overexpression of specific elements of the glucose-6-phosphatase enzyme complex can inhibit lactate production in hepatocytes from fasted rats. At 8 h after viral treatment and in the presence of 25 mM glucose, the experimental and control groups had produced the same amount of lactate (Fig. 6). At 24 h, AdCMV-P46 or AdCMV-P46 ϩ AdCMV-␤GAL-treated cells appeared to accumulate slightly less lactate than controls, but statistical significance was not achieved. In contrast, treatment with AdCMV-P36 alone caused a statistically significant 33% decrease in lactate output relative to controls, whereas the combination of AdCMV-P36 ϩ AdCMV-P46 treatment resulted in a 44% decrease. Thus, the effects of P46 overexpression on glycogen synthesis are clearly more dramatic than its effects on glycolytic flux, as assessed by measurement of lactate production.
Metabolic Effects of P46 and P36 Overexpression in Hepatocytes with Overexpressed Glucokinase-Isolated hepatocytes are known to rapidly lose differentiated function, including expression of glucokinase (28). Decreased expression of glucokinase can in turn impact the expression of other glycolytic and gluconeogenic enzymes (29,30). We therefore performed a final set of studies in which glucokinase expression was maintained in hepatocytes isolated from fasted rats at a high constant level via adenovirus-mediated expression of the enzyme (AdCMV-GKL virus, Ref. 20). Cells prepared in this way were then treated with the combinations of viruses encoding components of the glucose-6-phosphatase complex and were used for analysis of glycogen content and glycolytic flux via measurement of [3-3 H]glucose usage. As shown in Fig. 7, treatment of hepatocytes with AdCMV-GKL or AdCMV-GKL ϩ AdCMV-␤GAL resulted in an approximate doubling (217% increase) in[3-3 H]glucose usage relative to cells treated with AdCMV-␤GAL alone, consistent with our previous findings (15). Combined treatment of hepatocytes with AdCMV-GKL and AdCMV-P46 did not result in a significant decrease in glucose usage relative to cells treated with AdCMV-GKL ϩ AdCMV-␤GAL, whereas cotreatment with AdCMV-GKL ϩ AdCMV-P36 resulted in a 39% decrease. Thus, overexpression of P36, but not P46, partially counteracts the stimulatory effect of glucokinase overexpression on glucose usage in fasted hepatocytes.
To investigate the effects of P46 and P36 expression on glycogen metabolism in cells with constant glucokinase expression, cells subjected to the same treatment as described for Fig.  7 were used for measurement of glycogen content. As shown in Fig. 8, treatment of hepatocytes with AdCMV-GKL alone or AdCMV-GKL ϩ AdCMV-␤GAL resulted in 15.6-and 11.3-fold increases in glycogen content relative to cells treated with AdCMV-␤GAL alone. Co-treatment of cells with AdCMV-GKL ϩ AdCMV-P46 resulted in a small (13.4%, p ϭ 0.09) decrease in glycogen storage relative to the AdCMV-GKL ϩ AdCMV-␤GAL-treated group. In contrast, co-treatment of cells with AdCMV-GKL ϩ AdCMV-P36 completely prevented the glucokinase-induced increase in glycogen storage. Thus, whereas P46 expression exerts a counteractive effect on glucokinase-mediated stimulation of glycogen synthesis, this effect is small compared with that achieved by overexpression of P36.
P46 Overexpression Activates Glucose 1-Phosphate Hydrolysis-Our data showing that P46 overexpression increases the rate of Man-6-P hydrolysis led us to investigate whether hydrolysis of other hexose phosphates is also increased. Hepatocytes isolated from overnight fasted rats were treated as described in the legend for Fig. 4. The culture media were collected and analyzed for lactate concentration. Data represent the mean Ϯ S.E. for five independent experiments. The various experimental groups were compared with the AdCMV-␤Gal-treated control group and were found to be different with the following levels of significance: *, p Ͻ 0.05, and **, p Ͻ 0.0005. cogen synthesis and glycogen degradation. Hydrolysis of Glc-1-P as a consequence of overexpression of P46 could explain both the reduced glycogen accumulation in hepatocytes exposed to high glucose and the increased rate of glycogen degradation in hepatocytes exposed to a decline in glucose concentration (the latter via alteration of the equilibrium of the glycogen phosphorylase reaction).
To test this idea, we compared the hydrolysis of glucose 6-phosphate, glucose 1-phosphate (Glc-1-P), fructose 6-phosphate (Fru-6-P), and mannose 6-phosphate in intact and disrupted microsomes in hepatocytes with or without overexpressed P46. Interestingly, distinct results were obtained with each of these sugar phosphates (Fig. 9). Thus, Glc-6-P and Man-6-P were handled much as described in Fig. 3. Glc-6-P hydrolysis was nearly equal in intact and disrupted microsomes and was increased significantly in intact microsomes by overexpression of P46. Man-6-P hydrolysis was much lower in intact than disrupted microsomes, and activity in intact microsomes was significantly increased by P46 overexpression. Surprisingly, both Glc-1-P and Fru-6-P were hydrolyzed to equal extent by intact and disrupted microsomes prepared from control hepatocytes, similar to the pattern with Glc-6-P. However, Glc-1-P hydrolysis, but not Fru-6-P hydrolysis, was increased in intact microsomes by P46 overexpression. Thus, these data are consistent with a model in which P46 overexpression had preferential effects on glycogen deposition because of its specific ability to stimulate hydrolysis of a hexose phosphate intermediate of glycogen metabolism. DISCUSSION The cDNA encoding the T1 translocase/P46 component of the glucose-6-phosphatase complex was cloned originally on the basis of its homology to the bacterial gene products UhpT, GlpT, and UhpC, which are hexose phosphate transporter, glycerol phosphate transporter, and hexose phosphate receptor proteins, respectively (7,8). The transcript corresponding to the putative P46 clone was found to be expressed at highest levels in liver and kidney, consistent with a role in the gluconeogenic glucose-6-phosphatase complex. The strongest evidence that the newly cloned gene was a component of the mammalian glucose-6-phosphatase enzyme complex was the identification of mutations in the gene in two human subjects with type 1b glycogen storage disease (11). However, the precise biochemical function of P46 in mammalian cells has not been elucidated. Therefore, the purpose of the current study was to provide insight into the metabolic and biochemical properties of this protein through its adenovirus-mediated overexpression in isolated rat hepatocytes.
Type 1b glycogen storage disease is characterized by an absence of glucose-6-phosphatase enzymatic activity in intact liver microsomes but normal or elevated activity in detergentdisrupted preparations (12). The biochemical phenotype of type 1b glycogen storage disease patients lends support to the substrate-transport model for the glucose-6-phosphatase complex, in which the role of P46 is suggested to be as an ER-localized transporter for glucose 6-phosphate, allowing the luminally oriented Glc-6-Pase catalytic subunit to gain access to its substrate (2). The substrate-transport model in its original form held that the translocase function was rate-limiting for the system, based on higher rates of glucose 6-phosphate hydrolysis in detergent-disrupted compared with intact microsomes. However, subsequent rapid kinetic studies revealed that in the first 10 s, rates of glucose 6-phosphate hydrolysis are equal in intact and disrupted microsomes, whereas at longer time points the rate decreases in the intact preparations but maintains linearity in the disrupted ones (31). This was interpreted to indicate a close interaction of the translocase and phosphohydrolase components of the glucose-6-phosphatase complex, with regulation of activity through a conformational change in one or both interacting proteins (3,31). The tight association of the two proteins is supported by recent studies in which knockout of the Glc-6-Pase catalytic subunit causes loss of glucose 6-phosphate transport into microsomes (27).
The current study provides the first demonstration that overexpression of P46 is sufficient to enhance the activity of the glucose-6-phosphatase enzyme complex. Key data supporting this point include the following. 1) The P46 transgene product is targeted to a microsome-enriched fraction, suggesting normal delivery of the overexpressed protein to the endoplasmic reticulum in intact cells. 2) Overexpression of P46 enhances glucose 6-phosphate, mannose 6-phosphate, and glucose 1-phosphate hydrolysis in intact microsomes, without an effect on total phosphohydrolase activity in disrupted microsomes. However, the data also show that the impact of P46 overexpression on Glc-6-P hydrolysis in intact microsomes is much less than the effect of overexpression of P36, suggesting that most of the flux control is vested in the phosphohydrolase component of the complex. These data do not clearly support or refute any of the foregoing models of glucose-6-phosphatase complex function, since the increased Glc-6-P hydrolysis observed with P46 overexpression can be accommodated by the substrate transport model or a conformational model. 3) Overexpression of P46 clearly inhibits glycogen accumulation in hepatocytes from fasted rats and also causes activation of glycogenolysis in hepatocytes from fed rats. Interestingly, P46 overexpression has minimal effects on lactate production or [3-3 H]glucose usage, in contrast to the more pronounced effects of overexpressed P36. Overexpressed P46 is also much less effective than overexpressed P36 at countering the enhancement of glycolytic flux and glycogen synthesis induced by glucokinase overexpression.
Our work has uncovered a surprising heterogeneity in the handling of various hexose phosphates by the intact glucose-6phosphatase system. Thus, we have shown that both Glc-1-P and Fru-6-P are hydrolyzed by intact microsomes from control hepatocytes, although at approximately one-third to one-half the rate of Glc-6-P when all of these hexose phosphates are present at a concentration of 10 mM. Moreover, overexpression of P46 increases Glc-1-P, but not Fru-6-P hydrolysis, suggesting that the former sugar may be a real physiological substrate for the glucose-6-phosphatase system. These findings also suggest a mechanism by which P46 overexpression preferentially impairs glycogen metabolism relative to its minimal effects on glycolysis in hepatocytes from fasted rats, while also explaining why P46 overexpression increases glycogenolysis in glycogenreplete cells. Glc-1-P is a hexose phosphate intermediate that is specific to glycogen metabolism. Enhanced hydrolysis of Glc-1-P during periods of glycogen synthesis would reduce the amount of substrate for the UDPG pyrophosphorylase reaction, whose product UDP-glucose is the immediate precursor of glycogen synthesis. In glycogen-replete hepatocytes with P46 overexpression subjected to a sudden drop in glucose concentration, depletion of Glc-1-P would alter the equilibrium of the glycogen phosphorylase reaction in favor of glycogen degradation and Glc-1-P formation.
Although the ability of the glucose-6-phosphatase complex to hydrolyze Glc-1-P seems to explain the specific effects of overexpressed P46 on glycogen metabolism, we have also considered the possibility that P46 facilitates specific interactions between the glucose-6-phosphatase complex with proteins or enzymes that regulate glycogen metabolism. There is growing evidence that glucose disposal in general, and glycogen metabolism in particular, are spatially organized pathways within liver cells. For example, in the fasted state, cytosolic glucokinase enzyme activity is low due to sequestration of the enzyme in the nucleus via its binding to an inhibitory glucokinase regulatory protein (32)(33)(34). In the postprandial state, glucose stimulates the translocation of glucokinase from the nucleus to FIG. 9. Hydrolysis of glucose 1-phosphate, but not fructose 6-phosphate, is increased by P46 overexpression. Hepatocytes isolated from 18-h fasted rats were treated with AdCMV-P46 (P46), AdCMV-␤GAL (BG), or were left untreated (null). Cells were incubated in culture medium containing 3 mM glucose for 18 h, prior to harvesting of microsomes. Microsomes were incubated in the absence (intact) or presence (detergent-disrupted) of 0.5% cholic acid. These preparations were then incubated in the presence of 10 mM glucose 6-phosphate (G6P), 10 mM mannose 6-phosphate (M6P), 10 mM glucose 1-phosphate (G1P), or 10 mM fructose 6-phosphate (F6P) for 9 min at 37°C. P i production was measured as an index of hydrolysis of these hexose phosphates, normalized to total microsomal protein in each sample. Data represent the mean Ϯ S.E. for five independent experiments. The symbol * indicates those samples for which hexose phosphate hydrolysis was increased relative to the corresponding AdCMV-␤GAL-treated control group, with p Ͻ 0.05. the cytosol. The enzymes of glycogen metabolism also exhibit spatial organization. Thus, in liver cells, glycogen synthase is translocated from an intracellular site to the cell membrane in response to glucose and insulin, resulting in synthesis of glycogen in a gradient from the membrane surface toward the interior of the cell (3,35,36). Controlled movement of glycogen synthase within cells is complemented by targeting of protein phosphatase-1 (PP1) to the glycogen particle. This targeting is facilitated by glycogen targeting subunits of protein phosphatase-1, which also appear to bind to the key enzymes of glycogen metabolism, thereby serving as a molecular scaffold for glycogen synthesis (reviewed in Ref. 37). The existence of this glycogen "metabolon" (38), coupled with the remarkable organization of the glucose-6-phosphatase complex, suggests the possibility that these two metabolic machines could interact. Further work will be required to determine whether such interaction actually occurs, and if so, the role of the P46 T1 translocase in this process.
The findings described herein may also have relevance to understanding of regulation of hepatic glucose metabolism and the control of blood glucose homeostasis. Hepatic glycogen stores are decreased in all forms of human diabetes (39 -41), but the mechanisms responsible for this defect are not well understood. One possibility is enhanced flux through the glucose-6-phosphatase complex. Consistent with this idea, it has recently been demonstrated that patients with type 2 diabetes have a reduced capacity to regulate endogenous glucose production by glucose per se, leading the authors to suggest that regulation of glucose-6-phosphatase is impaired (42). To date, investigation of the molecular basis of up-regulated glucose-6phosphatase activity in diabetes has focused on the expression of the P36 catalytic subunit, leading to the finding that P36 expression is increased in response to hyperglycemia and hyperlipidemia in rodents and cultured cells (30,43,44). Recently, it was found that P46 expression was also up-regulated in diabetes (45), mainly by hyperglycemia (46). Thus, the potential contribution of changes in expression and activity of the P46 subunit should be considered more carefully, particularly with regard to its effect on hepatic glycogen storage in type 2 diabetes.