Effect of stimulation of glucokinase (GK) export from the nucleus by small amounts of sorbitol on hepatic glucose flux in response to elevated plasma glucose was examined in 6-h fasted Zucker diabetic fatty rats at 10 wk of age. Under basal conditions, plasma glucose, insulin, and glucagon were ∼8 mM, 2,000 pmol/l, and 60 ng/l, respectively. Endogenous glucose production (EGP) was 44 ± 4 μmol·kg−1·min−1. When plasma glucose was raised to ∼17 mM, GK was still predominantly localized with its inhibitory protein in the nucleus. EGP was not suppressed. When sorbitol was infused at 5.6 and 16.7 μmol·kg−1·min−1, along with the increase in plasma glucose, GK was exported to the cytoplasm. EGP (23 ± 19 and 12 ± 5 μmol·kg−1·min−1) was suppressed without a decrease in glucose 6-phosphatase flux (145 ± 23 and 126 ± 16 vs. 122 ± 10 μmol·kg−1·min−1 without sorbitol) but increased in glucose phosphorylation as indicated by increases in glucose recycling (122 ± 17 and 114 ± 19 vs. 71 ± 11 μmol·kg−1·min−1), glucose-6-phosphate content (254 ± 32 and 260 ± 35 vs. 188 ± 20 nmol/g liver), fractional contribution of plasma glucose to uridine 5′-diphosphate-glucose flux (43 ± 8 and 42 ± 8 vs. 27 ± 6%), and glycogen synthesis from plasma glucose (20 ± 4 and 22 ± 5 vs. 9 ± 4 μmol glucose/g liver). The decreased glucose effectiveness to suppress EGP and stimulate hepatic glucose uptake may result from failure of the sugar to activate GK by stimulating the translocation of the enzyme.
- obese-type 2 diabetes
- glucokinase regulatory protein
excessive postprandial hyperglycemia is a prominent and early defect in subjects with type 2 diabetes, which is the result, at least in part, from an inadequate suppression of net hepatic glucose production (NHGP) (15, 34, 36) and an insufficient increase in hepatic glucose uptake (HGU) (6, 7, 21). Glucose-induced suppression of NHGP was associated with increased hepatic glucose phosphorylation (47) and was abolished by inhibiting hepatic glucokinase (GK) activity in normal rats (5). This suggests that increased hepatic GK flux mediates glucose's effect not only to increase HGU, but also to decrease NHGP by the opposition of glucose 6-phosphatase flux. Basu et al. (6, 7) showed that in type 2 diabetic patients, lower splanchnic glucose uptake in the face of hyperglycemic hyperinsulinemia was associated with a blunted increase in glucose phosphorylation. This implies impaired activity of GK in the liver of these patients.
Studies using mouse models showed a large impact of altered hepatic GK expression in regulating postabsorptive blood glucose levels and hyperglycemia-induced HGU (35). However, decreased hepatic GK activity is not always present in patients (8, 12) and animals (30, 31, 45, 51) with obesity and obese-type 2 diabetes. Functional GK activity in the liver is also allosterically regulated by its interaction with a 68-kDa GK regulatory protein (GKRP), which binds to GK and inhibits the enzyme by a decrease in the apparent enzyme affinity for glucose (58). Studies using both cultured hepatocytes (2, 10, 32) and performed in vivo (13, 22, 23) demonstrated that GK is sequestered in the nucleus by GKRP under conditions of low glucose and insulin concentrations and that in response to a rise in glucose and/or insulin concentration, GK is dissociated from GKRP and is rapidly transported to the cytoplasm. It is likely, therefore, that stimulation of GK translocation by glucose plays an important physiological role in glucose effectiveness to stimulate its own uptake and inhibit its own production by the liver in normal subjects. Recently, we reported that at an early diabetic stage in Zucker diabetic fatty (ZDF) rats, a model of obese-type 2 diabetes mellitus, inadequate suppression of endogenous glucose production (EGP) and a defect in HGU in response to hyperglycemia and hyperinsulinemia were accompanied by a lack of GK translocation from the nucleus to the cytoplasm but not altered expression of the enzyme in hepatocytes (23). In view of the reported impact of GK translocation in the control of hepatic glucose phosphorylation (2), these findings imply a pathogenic role for the lack of GK translocation in the blunted response of HGU to hyperglycemia. Even in normal animals, however, it remains unknown whether GK translocation from the nucleus to the cytoplasm is a required process to increase glucose phosphorylation. GK in its free form in the nucleus could phosphorylate glucose, because the molecular masses of glucose and glucose 6-phosphate (G-6-P) are below the exclusion limit of the nuclear pore (44) and thus should allow the passive diffusion through the nuclear membrane. The immunohistological technique used in our previous study (23) did not allow us to determine whether GK was still binding to GKRP in the nucleus in the presence of hyperglycemia. Furthermore, it has been shown that the catalytic activity of GK is allosterically inhibited by some metabolic intermediates, a long-chain acyl-CoA (54, 59) and N-acetyl-glucosamine (41, 59), and by phosphorylation of the enzyme (19, 40). Therefore, it is possible that the unresponsiveness of hepatic glucose flux to hyperglycemia and hyperinsulinemia at the early stage of diabetes in ZDF rats results from an inhibition of the catalytic activity of GK by factor(s) unrelated to GKRP.
The affinity of GKRP for binding GK is modified by the competitive binding of phosphoesters, fructose-6-phosphate (F-6-P), and fructose-1-phosphate (F-1-P) (62). The binding of F-6-P to GKRP is required for the rat isoform of the protein to bind GK, and displacement of F-6-P by F-1-P at the common binding site on GKRP dissociates GK from GKRP by decreasing the affinity of the protein for GK (9, 58). Small amounts of sorbitol or fructose, precursors of F-1-P, cause GK translocation from the nucleus to the cytoplasm and an increase in glucose phosphorylation and glycogen synthesis from glucose in cultured hepatocytes (2) during a euglycemic-hyperinsulinemic clamp in healthy humans (46) and in the presence of hyperglycemia with and without hyperinsulinemia in conscious normal dogs (49, 50). To test our hypothesis that the failure of the elevated glucose to stimulate the dissociation of GK-GKRP complex and subsequent redistribution of GK to the cytoplasm in hepatocytes is responsible for both the impaired suppression of NHGP and a defect in HGU in response to a rise in plasma glucose in prediabetic ZDF rats, we examined the effect of small amounts of sorbitol on GK localization, impaired glucose-induced suppression of NHGP and stimulation of HGU in ZDF rats in the early stage of diabetes.
RESEARCH DESIGN AND METHODS
Animals and surgical procedures.
Six-week-old male ZDF (ZDF/Gmi-fa/fa) rats were purchased from Charles River Laboratory (Wilmington, MA). The animals were fed with Formulab Diet (No. 5008; Purina Mills, St. Louis, MO) and were given water ad libitum in an environmentally controlled room with a 12:12-h light-dark cycle. Surgery was performed at the age of 8 wk as previously described (13, 23). Briefly, rats were anesthetized with pentobarbital sodium (42.5 mg/kg body wt ip). The catheters were placed into the ileal vein, left carotid artery, and right jugular vein and externalized at the back of the neck and filled with heparinized saline (200 units heparin/ml saline). All experiments were conducted in accordance with the Guide for the Care and Use of Laboratory Animals provided by the National Institutes of Health, and all protocols were approved by the Vanderbilt University Institutional Animal Care and Use Committee.
At the age of 10 wk, the animals were fasted for 6 h prior to each study that consisted of a 2-h tracer equilibrium period (−180 to −60 min), a 1-h basal period (−60 to 0 min), and a 1-h test period (0 to 60 min). At −180 min, both [2-3H]- and [3-3H]glucose were given at 60 μCi in a bolus through the jugular vein catheter followed by continuous infusion of 0.6 μCi/min. During the test period, somatostatin was infused through the jugular vein catheter at 3 μg·kg−1·min−1 to inhibit endogenous insulin and glucagon secretion. Insulin and glucagon were infused into the hepatic portal system through the ileal vein catheter at 5.0 mU·kg−1·min−1 and 2.6 ng·kg−1·min−1, respectively, to maintain basal plasma insulin (∼1,500 pmol/l) and glucagon (∼55 ng/l) levels. Plasma glucose levels were kept constant at 15 mmol/l by infusion of a 50% glucose solution into the systemic circulation through the jugular vein catheter at a variable rate. In the first group (no sorbitol; S-0 group), vehicle (saline) was infused at 20 μl·kg−1·min−1. In the second (low-sorbitol; S-1) and third (high-sorbitol; S-2) groups, sorbitol was infused at 5.6 or 16.7 μmol·kg−1·min−1, respectively, into the portal vein through the ileal vein catheter. Blood samples were taken directly from the arterial catheter, and blood glucose levels were monitored using the Accu-Chek glucometer (Roche Diagnostics, Indianapolis, IN). At each sampling time, the blood cells obtained were washed with saline and then given back to the animals, to maintain their normal hematocrit. At the conclusion of the experiment, the animal was anesthetized with an infusion of pentobarbital sodium (40 mg/kg iv), and a laparotomy was performed immediately. The median lobe of the liver was excised and dropped into ice-cold phosphate-buffered saline for immunohistochemical analysis. The left lobe of the liver and the skeletal muscle (vastus lateralis) were frozen in situ using Wollenberg tongs precooled in liquid nitrogen. The tissue sampling procedures took less than 20 s to perform.
Metabolites in blood and tissue.
The glycogen and G-6-P levels in liver and skeletal muscle, plasma insulin and glucagon, and blood lactate and alanine were determined as previously described (23). The recombinant Escherichia coli F-1-P kinase, with COOH-terminal histidine tag, was expressed and used to analyze the F-1-P in liver extracts. This process was performed according to Veiga-da-Cunha's method (61) with a minor modification. The open reading frame of the gene encoding F-1-P kinase was amplified by a PCR technique, which used a 100-μl reaction mixture that contained 0.5 μg genomic DNA from E. coli K-12, 2.5 units AmpliTaq (Roche) polymerase, and primers (5′-GGCGCCATATGAGCAGACGTGTTGCTACT-3′ and 5′-CTGGATCCCGGTTAAAAGGTTGTAAGTC-3′). This introduced an NdeI restriction site at the initial ATG codon, and a BamHI restriction site after the last codon, respectively. The PCR product was digested with NdeI and BamHI, and cloned into the corresponding sites of pET-21b(+) (Novagen, La Jolla, CA), and its DNA sequence was verified by the dideoxy termination method. Recombinant enzyme was overexpressed by isopropyl-1-thio-β-d-galactopyranoside induction and purified to apparent homogeneity using a Ni-NTA column according to the manufacturer's protocol (Qiagen, Valencia, CA). The resulting purified enzyme had a specific activity (SA) of 115 U/mg.
The SA (dpm/μmol) of plasma [2-3H]- and [3-3H]glucose were determined as described previously (24). Plasma samples were deproteinized using Ba(OH)2 and ZnSO4. [2-3H] and [3-3H] radioactivity in plasma glucose and glycogen were determined by selective enzymatic detritiation of [2-3H]glucose. External standards of [2-3H]- and [3-3H]glucose, suspended in control rat plasma, were processed in parallel with each assay performed to calculate the degree of detritiation of each isotope during each sample plasma assay. Overall, completion of detritiation of [2-3H]glucose was 97.2 ± 0.4%, while 99.8 ± 0.3 of [3-3H]glucose remained intact. The liver content of uridine 5′-diphosphate (UDP)-glucose and UDP-galactose was obtained through two sequential chromatographic separations, and the amount of [3H] radioactivity in each fraction was measured, as described previously (23, 24).
Immunohistochemical and Western blot analyses of GK and GKRP.
In the liver, the immunostaining for GK and GKRP expression, along with the Western blot analyses of these proteins, was performed using sheep anti-rat GST-GK serum and rabbit anti-rat GST-GKRP serum, as reported previously (13, 23, 24). Quantitative image analysis of GK and GKRP immunofluorescence in the nucleus and cytoplasm of hepatocytes was performed using a Zeiss LSM510 confocal laser scanning microscope (13, 23, 24). The internal He/Ne laser and external argon-krypton laser at 543, 647, and 488 nm were used to optimally excite Cy3, Cy5, and YoPro-1 fluorescence, respectively. After the transfer of image files to a Power Macintosh Imaging workstation, the image was converted to TIFF and the individual cells were quantified using nuclear/cytoplasmic pixel density rationing with National Institutes of Health Image (version 1.56). In the section incubated with preimmune serum, three microscopic areas were randomly selected and GKRP immunofluorescence (Cy5) in the nucleus was measured in all of the cells in which the immunoreactive area with YoPro-1 (a cross section of the nucleus) was above a round 18 × 18 pixel area. GK and GKRP are not expressed homogenously in all the parenchymal cells in the liver. GK is expressed from periportal to perivenous areas in an increasing gradient in normal (18) and ZDF rats (data not shown). Even in perivenous areas, the intensities of immunofluorescence of GK and GKRP vary among hepatocytes (23). Furthermore, it is possible that the extent of GK translocation in response to increased plasma glucose differs throughout the liver lobule. To avoid intentional selection and to obtain results reflecting changes in the whole liver, five fields were randomly selected from each section. To detect a clear immunofluorescence signal in the nucleus, we selected cells that exhibited the immunoreactive area with YoPro-1 (a cross sectional area of the nucleus) above the 18 × 18 pixel area. To identify Cy3 (GK)-positive cells among these selected cells, we selected cells with a higher immunofluorescence intensity of Cy5 (GKRP) in the nucleus compared with controls stained with preimmune serum. We did not detect any cells in which GK was present in the nucleus in the absence of GKRP (data not shown). Ten to twenty cells were Cy5 positive in each microscopic field, and thus the total number of chosen cells was 50–100 cells for each section. For each cell, an ∼18 × 18 pixel area (181 square pixels) was analyzed in the nucleus and the cytoplasm by measuring mean pixel density (range = 0–255 grayscale levels). Nuclear-to-cytoplasm pixel density ratios of GK and GKRP were determined by digital image analysis by using Scion Image. The ratios of nuclear-to-cytoplasmic fluorescence of GK and GKRP were averaged for each animal, and the average value was normalized to that of a standard liver sample, stained on the same day. The results for each group were expressed as the means ± SE of the normalized value in five animals.
Activities of glycogen synthase and phosphorylase and GK in freeze-clamped liver samples were measured by using the method described previously (23, 24). Glucose phosphorylation by GK was determined as the absorbance change in the presence of 8 or 100 mM glucose minus the absorbance change in the presence of 0.5 mM glucose under conditions in which the absorbance was increasing linearly between 10 and 30 min.
Rates of [2-3H]- and [3-3H]glucose-determined glucose turnover were calculated as the ratio of the rate of infusion of [2-3H]- and [3-3H]glucose (dpm/min) and the [2-3H]- and [3-3H]-labeled SA in plasma glucose, respectively, as described previously (23, 24): [2-3H]glucose turnover rate = infusion rate of [2-3H] glucose/plasma SA of [2-3H]glucose and [3-3H]glucose turnover rate = infusion rate of [3-3H]glucose/plasma SA of [3-3H]glucose, respectively. EGP was determined as the difference between [3-3H]glucose turnover rate and the glucose infusion rate (GIR). To estimate the amount of [2-3H]- and [3-3H]glucose incorporated into hepatic and skeletal muscle glycogen via the direct pathway (GLY-[2-3H]glucose and GLY-[3-3H]glucose, respectively), [2-3H] and [3-3H] radioactivities incorporated into glycogen were divided by the [2-3H] and [3-3H]-labeled SA of plasma glucose, respectively: GLY-[2-3H]glucose = [2-3H] radioactivity in glycogen/plasma SA of [2-3H]glucose and GLY-[3-3H]glucose = [3-3H] radioactivity in glycogen/plasma SA of [3-3H]glucose, respectively. The fractional detritiation of [2-3H]G-6-P (D[2–3H]) by exchange reaction of [3H]- of [2-3H]G-6-P with [H+] of H2O mediated by hexose isomerase was calculated as the ratio of GLY-[2-3H]glucose to GLY-[3-3H]glucose: D[2–3H] = 1-GLY-[2-3H]glucose/GLY-[3-3H]glucose. This calculation is based on the assumption that the ratio of [2-3H] to [3-3H]glucose incorporated into glycogen approximates that of the G-6-P pool. The ratio of [2-3H]glucose-to-[3-3H]glucose incorporated into glycogen was not different among the three groups (0.28 ± 0.02 in S-0, 0.26 ± 0.03 in S-1, and 0.30 ± 0.04 in S-2). Glucose recycling (GC) is defined as input of extracellular glucose into the G-6-P pool followed by exit of plasma-derived G-6-P back into the extracellular pool: GC = ([2-3H]glucose turnover rate − [3-3H]glucose turnover rate)/D[2–3H]. The glucose-6-phosphatase (G-6-Pase)-flux was calculated as the sum of EGP plus GC. The fractional contribution of plasma glucose via the direct pathway (UDP-glucose-glucose) was calculated as the ratio of [3-3H]SA in hepatic UDP-glucose to [3-3H]-labeled SA in plasma glucose: UDP-glucose-glucose = ([3H]SA of hepatic UDP-glucose/[3-3H]-labeled SA in plasma glucose) × ([3-3H] radioactivity in hepatic glycogen/total [3H] radioactivity in hepatic glycogen).
The data collected are expressed as means ± SE. The significance of the differences between the groups of the time course data was analyzed by two methods: a two-way, repeated-measures ANOVA and a one-way ANOVA or Student's t-test. Differences were considered significant when P < 0.05.
During the basal period, levels of plasma glucose, insulin, and glucagon were ∼8.4 mM, ∼1,900 pmol/l, and ∼68 ng/l, respectively (Fig. 1). The [2-3H]- (Fig. 2A) and [3-3H]glucose determined glucose turnover rates (Fig. 2B), and the difference between the [2-3H]- and [3-3H]-glucose turnover rates (Fig. 2C) (as average of the value at −60, −30, and 0 min) were similar in the vehicle (S-0) group (70 ± 10, 43 ± 5, and 27 ± 7 μmol·kg−1·min−1, respectively), the S-1 group (74 ± 5, 44 ± 3, and 28 ± 6 μmol·kg−1·min−1, respectively), and the S-2 group (71 ± 10, 45 ± 4, and 28 ± 8 μmol·kg−1·min−1, respectively). Blood lactate and alanine levels were not different among the groups (Fig. 3).
During the test period, plasma glucose levels were increased twofold, while arterial plasma insulin and glucagon levels were kept at basal levels. There were no differences in these parameters among the three groups (Fig. 1). In the S-0 group, the [2-3H]- and [3-3H]glucose-determined glucose turnover rates and the difference in glucose turnover rate between [2-3H] and [3-3H]glucose were increased by 40% (108 ± 14), 30% (57 ± 5), and 90% (51 ± 10 μmol·kg−1·min−1), respectively (Fig. 2). Blood lactate and alanine levels were not changed (Fig. 3). EGP was not changed from that in the basal period (Fig. 4A). In the S-1 and S-2 groups, compared with the S-0 group, [3-3H]glucose-determined glucose turnover rates (70 ± 5 and 76 ± 2 μmol·kg−1·min−1, respectively) were slightly higher, but without significant differences. On the other hand, [2-3H]glucose-determined glucose turnover (160 ± 14 and 165 ± 27 μmol·kg−1·min−1, respectively), and the difference between [2-3H]- and [3-3H]glucose turnover rates (75 ± 14 and 91 ± 28 μmol·kg−1·min−1, respectively) were significantly higher (Fig. 2C). GIRs required to maintain hyperglycemia were fivefold higher in S-1 and S-2 (Fig. 2D). Blood lactate levels were increased twofold in the S-1 and S-2 groups (2.7 ± 0.4 mM at 60 min and 2.8 ± 0.3 mM at 60 min, respectively) (Fig. 3A), while blood alanine levels were not increased (Fig. 3B). EGP tended to decrease in S-1 (Fig. 4A; 23 ± 19) and was significantly suppressed in S-2 (Fig. 4A; 12 ± 5 vs. 51 ± 2 μmol·kg−1·min−1 in S-0, P < 0.05), while G-6-Pase flux in the S-1 and S-2 groups was not different from that in the S-0 group (Fig. 4B; 145 ± 23 in S-1 and 126 ± 16 in S-2 vs. 122 ± 10 μmol·kg−1·min−1 in S-0). On the other hand, GC rates (Fig. 4C; 122 ± 17 in S-1 and 114 ± 19 in S-2 vs. 71 ± 11 μmol·kg−1·min−1 in S-0, P < 0.05), G-6-P content (Fig. 4D; 254 ± 32 in S-1 and 260 ± 35 in S-2 vs. 188 ± 20 μmol·kg−1·min−1 in S-0, P < 0.05), and fractional contribution of plasma glucose to the flux through UDP-glucose pool in the liver (Fig. 4F; 43 ± 8 in S-1 and 42 ± 8 in S-3 vs. 27 ± 6% in S-0, P < 0.05) were significantly higher than that in the S-0 group. Furthermore, the incorporation of plasma glucose into hepatic glycogen via the direct pathway was significantly increased (Fig. 4H; 19.8 ± 3.7 and 22.4 ± 4.5 vs. 8.6 ± 3.6, P < 0.05), while glycogen synthesis via the indirect pathway was not. Glycogen content in the S-1 and S-2 groups tended to be higher but was not significantly different, compared with that in the S-0 group (Fig. 4G).
Hepatic F-1-P and GK translocation.
In response to sorbitol infusion, hepatic F-1-P content was increased dose dependently (49.4 ± 12.5 in S-1 and 71.4 ± 15.2 in S-2 vs. 14.5 ± 1.6 nmol/g liver in S-0, P < 0.05), while hepatic F-6-P content remained unchanged (42 ± 8 in S-1 and 46 ± 8 in S-2 vs. 38 ± 3 nmol/g liver in S-0). As a result, the ratio of F-1-P to F-6-P was markedly increased by sorbitol infusion (1.2 ± 0.2 in S-1 and 1.6 ± 0.2 in S-2 vs. 0.38 ± 0.5 in S-0, P < 0.05) (Fig. 5).
Immunofluorescence of both GK and GKRP were predominantly detected within the nucleus in the S-0 group (Fig. 6, D–F). In the S-1 (Fig. 6, G and H) and S-2 (Fig. 6, J and K) groups, the intensity of immunofluorescence of GK was markedly reduced in the nucleus. As a result, the ratio of nucleic to cytoplasmic immunofluorescence was significantly decreased (Fig. 6M). The immunofluorescence of GKRP was still predominantly detected in the nucleus in all of the S-0 (Fig. 6, D and F), S-1 (Fig. 6, G and I), and S-2 groups (Fig. 6, J and L), and thus the ratio of nucleic-to-cytoplasmic immunofluorescence of GKRP was not significantly different among the groups (Fig. 6N). Total amounts of GK and GKRP protein as determined by Western blot analyses were similar, regardless of experimental conditions (the relative band density values of GK were 1.00 ± 0.20, 1.20 ± 0.53, and 1.18 ± 0.15, and those of GKRP were 1.00 ± 0.09, 0.83 ± 0.12, and 0.83 ± 0.02 for the S-0, S-1, and S-2 groups, respectively).
Glycogen synthase and phosphorylase activities.
As shown in Fig. 7, the activity of the active form of glycogen synthase, total activity of the synthase, and the ratio of the active form-to-total were significantly increased in the S-1 and S-2 groups at the end of the test period. However, the activity of the active form of glycogen phosphorylase, total activity of the enzyme, and the ratio of the active form to total activity in the S-1 and S-2 groups did not differ from the S-0 group.
The catalytic activities (nmol·min−1·mg protein−1) of GK measured with glucose concentrations at both 8 and 100 mM, which give near half-maximal and maximal velocities of the enzyme, respectively, in GK isolated from normal rat liver (53), were not significantly different among the control (18.5 ± 2.5 and 8.9 ± 1.2), S-1 (19.2 ± 2.6 and 9.4 ± 1.3), and S-2 (19.7 ± 2.5 and 9.6 ± 1.2, respectively) groups.
The present study using ZDF rats at the early stage of diabetes showed that GK translocation from the nucleus to the cytoplasm could be brought about by a small amount of sorbitol but not by elevated plasma glucose. Also overcoming impairment of GK translocation with a small amount of sorbitol improved the blunted response of EGP, hepatic glucose phosphorylation, and glycogen synthesis to hyperglycemia and hyperinsulinemia. The results from the present study indicate that decreased effectiveness of plasma glucose to suppress NHGP and to stimulate HGU at the early stage of diabetes of ZDF rats is, at least partly, due to the failure of glucose to activate GK via dissociation of GK from GKRP in the nucleus and/or obstruction of the binding of GK with GKRP in the cytoplasm in hepatocytes.
A defect in glucose-induced GK translocation in the ZDF rat is due to the failure of the sugar to dissociate the GK-GKRP complex and/or to obstruct the binding of GK with GKRP.
Our previous study showed that GK is still colocalized with GKRP in the nucleus in ZDF rats even in the presence of hyperglycemia that induces GK, but not GKRP, translocation from the nucleus to the cytoplasm in the normal rat (23). It was not clear, however, whether the defect in glucose-induced GK translocation is due to the failure of glucose to dissociate GK from GKRP or impairment of subsequent processes for the translocation and the retention of GK in the cytoplasm. Studies using HeLa cells transfected with GK and/or GKRP genes reported an essential role of the binding of GK by GKRP for importing GK into the nucleus (14, 48). Mice with an engineered mutant null for GKRP exhibited cytoplasmic localization of GK even at low blood glucose levels (20, 27), indicating that the binding with GKRP is essential to sequester GK in the nuclear compartment in vivo. On the other hand, it is possible that GK moves to the cytoplasm if GK dissociates from GKRP in the nucleus, depending on a functional nuclear export signal sequence in the GK molecule (48). If GK and GKRP are shuttling between the cytosolic and nuclear compartments, even under basal conditions, and GKRP serves as a chaperone to import GK into the nuclear compartment, the disruption of the binding of GK with GKRP in the cytoplasmic compartment may lead to GK accumulation in the cytoplasm. Indeed, studies using cultured hepatocytes isolated from normal rats showed that GK translocation is promoted by agents that favor the dissociation of the GK-GKRP complex either by binding to GKRP or to GK (3, 41). In the present study, intraportal administration of small amounts of sorbitol compartmentalized GK from GKRP in terms of the intracellular distribution, with predominant localization of GK in the cytoplasmic compartment and of GKRP in the nuclear compartment in ZDF rats (Fig. 6). The administration of sorbitol increased F-1-P content in the liver from basal (∼20 nmol/g liver) to 70 nmol/g liver without a change in F-6-P content (40–50 nmol/g liver) (Fig. 5). Studies with purified GK and GKRP proteins showed that the inhibition of GK catalytic activity by GKRP binding was completely relieved by F-1-P concentration above 0.05 mM in the presence of 0.05 mM F-6-P (16, 17). These findings suggest that GK could export to the cytoplasm once GK is dissociated from GKRP in ZDF rats. It is likely, therefore, that a defect of GK translocation in response to the rise in plasma glucose results from the failure of the sugar to dissociate the GK-GKRP complex in the nucleus and/or to obstruct the binding of GK with GKRP in the cytoplasm in ZDF rats.
A defect in GK translocation from the nucleus to the cytoplasm is responsible for insufficient suppression of NHGP and blunted HGU in response to elevated plasma glucose.
The administration of small amounts of sorbitol increased hepatic glucose phosphorylation as indicated by increases in GC, G-6-P content, the fractional contribution of plasma glucose to UDPG flux, and glucose incorporation into hepatic glycogen via the direct pathway (Fig. 4), along with activated glycogen synthesis (Fig. 7). Sorbitol administration also suppressed EGP markedly without alteration of G-6-Pase flux (Fig. 4), indicating that the suppression depended on increased GK flux. Hepatic glucose flux during hyperinsulinemic hyperglycemia in the presence of sorbitol in ZDF rats is similar to hepatic glucose flux seen in normal rats under similar conditions (23). Therefore, GK was catalytically active in the presence of sorbitol in ZDF rats. There is no information about the direct effect of sorbitol and its metabolites, fructose and F-1-P, per se on the catalytic activity of GK. Therefore, it is likely that a defect in the increase in glucose phosphorylation in response to hyperglycemia is due to the failure of elevated glucose to dissociate GK from GKRP, not to the direct inhibition of the catalytic activity of GK by factor(s) other than GKRP. Still, it remains unknown whether GK export from the nucleus to the cytoplasm is necessary to increase glucose phosphorylation.
Glycogen synthesis depends on the activity of glycogen synthase. While an activation of hepatic glycogen synthase has been repeatedly observed with administration of small amounts of fructose, a precursor of F-1-P, in in vivo and in vitro studies (57), several studies (38, 42) have suggested that the effect was secondary to increased G-6-P, a potent activator of glycogen synthase (63). Since the G-6-P content in the liver was significantly increased by sorbitol administration (Fig. 4), it is possible that the sorbitol-induced activation of glycogen synthase (Fig. 7) is due to its ability to increase glucose phosphorylation. Therefore, the restoration of glycogen synthesis by the direct pathway in response to hyperglycemia by the administration of small amounts of sorbitol may be secondary to restored glucose phosphorylation.
Possible mechanism responsible for the failure of glucose to dissociate the GK-GKRP complex and/or to obstruct the binding of GK with GKRP.
Niculescu et al. (41) showed that high concentrations of glucose form F-1-P in cultured hepatocytes. However, the glucose effect on GK translocation could not be explained solely by increased F-1-P, since more translocation of GK was observed with the same concentration of F-1-P in the presence of an elevated concentration of glucose than in the presence of fructose or sorbitol, at a low concentration of glucose (41). The effect of inhibition of aldose reductase in the polyol pathway in which glucose is converted to F-1-P on GK translocation is controversial (1, 41). In our in vivo studies, furthermore, hepatic F-1-P contents (nmol/g liver) were not increased during a hyperglycemic clamp (17 ± 3) compared with that under the basal condition (15 ± 4) in normal rats and were not different between 10-wk-old ZDF rats and the lean littermates (unpublished data). Therefore, it is unlikely that the glucose effect is mediated by F-1-P.
Glucose analogs, which are not phosphorylated by GK and hexokinase, could induce GK translocation in cultured hepatocytes (2), indicating that increased intracellular glucose concentration per se serves as a signal. GLUT2, a low-affinity high-turnover transporter, is the major glucose transporter isoform expressed in hepatocytes (39), and the transporter capacity via GLUT2 is present in vast excess of the GK trapping reaction (25). Thereby, the intracellular concentration of glucose in hepatocytes is equal to or slightly higher than the plasma concentration (64) and is rapidly equilibrated with a change in plasma glucose concentration in normal rats (43). The amount of GLUT2 in the liver of ZDF rats has been reported to be increased by 30–40%, relative to lean controls (52). The molecular mass of glucose, below the exclusion limit of the nuclear pore (44), should allow passive diffusion through the nuclear membrane. Therefore, it is unlikely that the elevated plasma glucose concentration failed to increase the intranuclear glucose concentration in hepatocytes in ZDF rats.
GK has a single glucose-binding site (9, 33) and undergoes global conformational changes by glucose binding (33). The enzyme was observed to exist in a fully closed form when glucose and GK activator were bound to it and in a wide-open form when not liganded. Mutation studies in Xenopus GK showed that two regions of GK may be included in the binding site of GKRP (60). In the closed form, these sites are widely separated and the no-effect region is positioned between the two sites. In contrast, the two regions are much closer together in the wide-open form of GK. This suggests that GKRP specifically binds to the wide-open form of GK. Therefore, impaired glucose binding to GK might be expected to favor formation of the GK and GKRP complex. However, this is not the case. The conformational change of GK by glucose binding is closely linked to the catalytic activity. GK is a monomeric enzyme with positive cooperativity for glucose phosphorylation, and a reversible two-step reaction has been proposed as the underlying mechanism of the positive cooperativity (11, 29). The wide-open form of GK is essentially inactive for catalysis and initially binds glucose with a slow rate and a weak affinity to form a transient intermediate, followed by a conformational change to the closed form (a catalytically competent GK-glucose complex). The concentration of glucose determines the distribution between the two states, resulting de facto in two binding affinities and the sigmoidal glucose dependency of the reaction. Therefore, impaired glucose binding on GK should shift the S0.5 for glucose phosphorylation velocity of GK to the right with decreased Vmax. However, the catalytic property of isolated free GK related to glucose concentration was not different between the ZDF rats and the lean littermates (23) and was not affected by sorbitol administration. In vivo, glucose phosphorylation was increased when GK was dissociated by sorbitol (Fig. 4). These results indicate that glucose binding to GK in free form and subsequent change in conformation of the enzyme could normally occur under intracellular conditions existing in the liver of ZDF rats.
Although an increase in affinity of GKRP for GK may blunt glucose's ability to dissociate GK from GKRP, hepatic F-1-P contents were also not different between ZDF and the lean rats under basal conditions and were not increased by the hyperglycemic clamp in either ZDF rats or the littermates (unpublished data). Intracellular F-6-P content in the liver might not be higher in ZDF rats compared with the lean littermates as indicated by lower G-6-P content in ZDF rats (23). There is a rapid equilibration between G-6-P and F-6-P by high-G-6-P isomerase activity in the liver. While the inhibition of GK activity by GKRP is enhanced by decreased Mg2+-ATP concentration and pH (59), there are no data indicating that these parameters are altered in the liver of ZDF rats. On the other hand, Baltrusch et al. (4) reported that the mutation of the helical structure, which is exposed in the super open conformation of GK, increased fivefold the binding affinity for GKRP over that with wild-type GK; the mutant GK was trapped in the nucleus of cultured hepatocytes even in the presence of 20 mM glucose. It is possible, as speculated by Baltrusch et al. (4), that mutations within the α10-helix or modifications of the structure of the helix exists in ZDF rats.
The catalytic property of human GK was similar to that of rat GK (11), and nuclear localization of GK immunoreactivity was shown in the parenchymal cells of human liver samples taken postmortem (56). While human GKRP shares 88% identity with rat GKRP (65), human GKRP causes a greater inhibition of GK in the absence of F-6-P and has a higher affinity for F-6-P than rat GKRP (9). In addition, F-1-P counteracts the inhibition of GK by human GKRP even in the absence of F-6-P (9), implying that human GK activity is more strongly regulated by GKRP and that the regulation of GK by GKRP is more sensitively regulated by phosphoesters, F-1-P, and F-6-P. The effectiveness of glucose per se to inhibit EGP is markedly blunted in subjects with type 2 diabetes (55). In individuals with type 2 diabetes, the addition of a catalytic amount of fructose, a precursor of F-1-P along with sorbitol, decreases the glucose and insulin response to an oral glucose tolerance test (37) and nearly normalizes the ability of hyperglycemia per se to suppress NHGP (28). Thus, the impairment of short-term activation of GK by glucose via dissociating this enzyme from GKRP is a possible explanation for the decreased effectiveness of elevated plasma glucose to suppress NHGP and to stimulate HGU in subjects with type 2 diabetes. Recently, we reported that when postabsorptive and postprandial hyperglycemia were chronically normalized by the treatment with a SGLT inhibitor, glucose-induced GK translocation was restored in 14-wk-old ZDF rats (24). At 10 wk of age, ZDF rats exhibited marked hyperglycemia under nonfasted conditions (302 ± 21 mg/dl plasma) and in the postprandial phase (454 ± 46 mg/dl plasma), while their plasma glucose levels decreased to near normal levels after 6 h fast (Fig. 1), indicating that the rats were also exposed to hyperglycemia most of time. Individuals with type 2 diabetes with optimal glycemic control retain normal glucose effectiveness to suppress EGP, whereas those in moderate-to-poor control completely lack this regulation (55). Therefore, it is possible that chronic hyperglycemia (glucotoxicity) plays a role as the primary defect for causing impaired glucose-induced dissociation of GK from GKRP in humans as well as ZDF rats.
This research was supported by National Institute of Health, Diabetes and Digestive and Kidney Diseases Grant DK-60667 (to M. Shiota). The radioimmunoassay core laboratory and the Cell Imaging Core Resource at Vanderbilt University Medical Center are supported by National Institutes of Health Grants CA-68485 and DK-20593, respectively.
The authors acknowledge Drs. Mary C. Moore and Richard L. Printz, Vanderbilt University Medical Center, for reading this manuscript carefully.
↵* J. Shin and T. P. Torres contributed equally to this work.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society