A mechanism for fatty acid inhibition of glucose utilization in liver. Role of xylulose 5-P.

The glucose-stimulated rise in Fru-2,6-P2 in liver results from xylulose 5-P activation of a specific protein phosphatase 2A which dephosphorylates Fru-6-P,2-kinase:Fru-2,6-bisphosphatase (Nishimura, M., and Uyeda, K. (1994) J. Biol. Chem. 269, 26100-26106). In order to determine the role of xylulose 5-P in regulating Fru-2,6-P2 in liver, the effect of fatty acids, various hexoses, and hormones was examined in perfused rat liver and in intact rats. When 24-h starved rat livers were perfused with acetate, butyrate, or propionate, Fru-2,6-P2 and xylulose 5-P decreased to the same extent and at similar rates. The activity ratios of the kinase and the phosphatase changed in a reciprocal manner, indicating that the phosphorylated form of the enzyme was increased by the fatty acids perfusion. The fatty acids caused the similar changes in the metabolites and the phosphorylation state of the bifunctional enzyme in livers of fed animals. Fructose, galactose, or mannose perfusion in starved rat liver increased both Fru-2,6-P2 and xylulose 5-P and converted the bifunctional enzyme to the dephospho form. Both the Fru-2,6-P2 and xylulose 5-P levels in rats fed a high fat diet decreased over 50% compared to that in control rats. These results indicated a close correlation between Fru-2,6-P2 and xylulose 5-P levels and the phosphorylation state of fructose 6-P,2-kinase:fructose 2,6-bisphosphatase. Fatty acid inhibition of glucose metabolism can be explained by a decrease in xylulose 5-P, which lowers xylulose 5-P-activated protein phosphatase 2A activity, resulting in more phosphorylated form of the bifunctional enzyme and consequently lower Fru-2,6-P2.

The administration of a high concentration of glucose to isolated hepatocytes, the perfusion of high glucose to liver, or the feeding of a high carbohydrate diet to starved (5) rats increases Fru-2,6-P 2 (6 -8). This rise was attributed to the increased concentration of Fru-6-P (9), which is the substrate for Fru-6-P,2-kinase and an inhibitor of Fru-2,6-Pase. More recently, however, Nishimura et al. (10) demonstrated that the glucose-dependent increase in Fru-2,6-P 2 is a result of dephosphorylation of the bifunctional enzyme resulting in activation of the kinase and inhibition of the phosphatase, which is the reversal of the effect of cAMP. The dephosphorylation of the enzyme is catalyzed by a specific protein phosphatase, which is activated specifically by xylulose 5-P (Xu-5-P) (Scheme I) (10). The protein phosphatase has been purified to homogeneity and characterized as a heterotrimeric protein phosphatase 2A (PP2A) (11). Although the PP2A dephosphorylates other substrates, such as pyruvate kinase and phosphorylase a, only the reaction with Fru-6-P,2-kinase:Fru-2,6-Pase as a substrate is stimulated by Xu-5-P (11). Based on these observations, we proposed the mechanism shown in Scheme I (11). According to this scheme, the Fru-2,6-P 2 concentration in liver is determined by the phosphorylation state of Fru-6-P,2-kinase:Fru-2,6-Pase, and the phosphorylation state of the enzyme is governed by the relative activities of protein kinase A and Xu-5-P-activated PP2A. The activities of protein kinase A and Xu-5-P-activated PP2A are controlled by cAMP and Xu-5-P, respectively. We proposed that Xu-5-P may serve as a second messenger, sensing the level of glucose in liver and antagonizing the effect of hormone mediated by cAMP, to stimulate glycolysis by raising Fru-2,6-P 2 level by dephosphorylation of the bifunctional enzyme (11).
It is well known that glucose metabolism is inhibited by administration of fatty acid in liver and heart, so called "glucose sparing" effect (12)(13)(14)(15). Administration of palmitate and hexanoate has been shown to inhibit glycolysis in isolated hepatocytes, and the inhibition is attributed to decreased Fru-2,6-P 2 concentration (16,17). Short chain fatty acids including butyrate, propionate, and acetate also have been shown to have a similar effect (18). Hue et al. (16) suggested that the mechanism for this fall in Fru-2,6-P 2 could result from inhibition of the bifunctional enzyme by citrate. These results were obtained before the Xu-5-P-activated PP2A was discovered (17). Thus, the objectives of this current work were 2-fold: (a) examine the * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ To whom correspondence should be addressed: Dept. of Veterans Affairs, Medical Center, 4500 S. Lancaster Rd., Dallas, TX 75216. Tel.: 214-372-7028; Fax: 214-372-9534. 1 The abbreviations used are: Fru-2,6-P 2 , fructose 2,6-P 2 ; PP2A, protein phosphatase 2A; Xu-5-P, xylulose 5-P; Glu-6-P, glucose 6-P. validity of the proposed physiological roles of the Xu-5-P and the Xu-5-P-activated PP2A in regulation of Fru-2,6-P 2 in liver; and (b) determine whether the fatty acid sparing effect is related to citrate or Xu-5-P. If this hypothesis were true, one predicts that fatty acid feeding or perfusion with fatty acid would lower Xu-5-P concentration in liver, resulting in inhibition of the Xu-5-P-activated PP2A and thus increase the phospho form of Fru-6-P,2-kinase:Fru-2,6-Pase. The net result would be decreased Fru-2,6-P 2 level and inhibition of phosphofructokinase and glycolysis to lower glucose utilization. The results presented herein support the Xu-5-P-mediated mechanism for the fatty acid inhibition of glucose metabolism.

EXPERIMENTAL PROCEDURES
Materials-cAMP enzyme immunoassay system was purchased from Amersham Corp. Glucose, sucrose, galactose, deoxyglucose, sodium acetate, n-butyric acid, citrate lyase, and glucagon were purchased from Sigma. D-Mannose was purchased from California Co. for Biochemical Research (Los Angeles, CA), and fructose was obtained from Eastman Kodak (Rochester, NY). Insulin was purchased from Eli Lilly & Co (Indianapolis, IN). Propionic acid was purchased from Aldrich Chemical Co. Propionic acid and butyric acid were neutralized with NaOH before use. Casein, a mineral mixture, and a vitamin mixture were purchased from Harlan Co. (Madison, WI). Metamucil, oils, and lard were obtained from a local grocery store. All other chemicals and resins were analytical reagent grade and purchased from commercial sources.
The livers of rats starved for 24 h or fed ad libitum rats were perfused with various solutions from 9 a. m. Before perfusion, rats were anesthetized by intraperitoneal injection with sodium pentobarbital (60 mg/kg). The livers were perfused in situ with Krebs-Henseleit bicarbonate solution (120 mM NaCl, 4.5 mM KCl, 1.2 mM MgSO 4 , 1.2 mM KH 2 PO 4 , 25 mM NaHCO 3 , 2.5 mM CaCl 2 ) equilibrated with 95% O 2 and 5% CO 2 by continuous bubbling. The medium also contained 40 mM glucose, fatty acids, sugars, or hormones as indicated. Bovine serum albumin (0.1%) was present in the solutions containing hormones. The flow rate of the perfusate was set to approximately 4 ml/min/g liver. Middle lobes of the livers were cut out for the assay of Fru-6-P,2-kinase: Fru-2,6-bisphosphatase activities and left lobes were immediately freeze-clamped with a tong pre-cooled in liquid nitrogen. As zero time controls, the livers were cut out immediately after opening the abdomens.
Metabolite Measurements-The freeze-clamped livers were ground in a porcelain mortar prechilled with liquid nitrogen, and the powder was stored at Ϫ70°C until analysis. HClO 4 extracts of frozen livers were prepared as described (21). For Fru-2,6-P 2 assay, the liver power was homogenized in EDTA/NaOH buffer (pH Ͼ 10) and assayed according to the method of Uyeda et al. (22). Xu-5-P was assayed by the method of Casazza and Veech (23) except the formation of Fru-6-P coupled to NADPH formation was measured fluorometrically by adding phosphoglucose isomerase (1 unit) and Glu-6-P dehydrogenase (0.4 unit) with excitation and emission wavelengths at 354 and 452 nm, respectively. Citrate was determined by the method of Dagley (24). Glyceraldehyde 3-phosphate was calculated from determined dihydroxyacetone phosphate by the method of Michal and Beutler (25), and all other metabolites were assayed spectrophotometrically by enzymatic methods described previously (26).
Assay Method for Fru-6-P,2-kinase-The reaction mixture contained in a final volume of 0.1 ml: 100 mM Tris-HCl (pH 7.5), 0.1 mM EDTA, 0.1 mM EGTA, 5 mM ATP, 5 mM phosphate, 10 mM MgCl 2 , and 0.05 mM (v) or 2 mM (V max ) Fru-6-P. The mixture was incubated at 30°C, and at timed intervals 10-l aliquots were transferred into 90 l of 0.1 N NaOH, and the solution was heated for 1 min at 80°C to stop the reaction. Suitable aliquots of this mixture were then assayed for Fru-2,6-P 2 as described. One unit of activity is defined as the amount of enzyme that catalyzes the formation of 1 nmol of Fru-2,6-P 2 /min under these conditions.
Assay Method for Fru-2,6-Pase-This assay measures continuously the formation of Fru-6-P coupled to NADPH formation using phosphoglucose isomerase and Glu-6-P dehydrogenase as described previously (27). The reaction mixture contained in a final volume of 1.0 ml: 100 mM Tris-HCl (pH 7.5), 0.2 mM EDTA, 100 mM NADP, 5 M (v) or 40 M (V max ) Fru-2,6-P 2 , 0.4 unit of Glu-6-P dehydrogenase, and 1 unit of phosphoglucose isomerase. The reaction was initiated with the addition of the Fru-6-P,2-kinase:Fru-2,6-Pase, and it was followed at room temperature fluorometrically at excitation and emission wavelengths of 354 and 452 nm, respectively.
Partial Purification of Fru-6-P,2-kinase:Fru-2,6-Pase from Rat Livers-A middle lobe of a perfused liver was excised and quickly homogenized using a Polytron homogenizer in 10 ml of ice-cold buffer consisting of 50 mM Tris phosphate (pH 8.0), 100 mM potassium fluoride, 5 mM EDTA, 5 mM EGTA, 1 mM dithiothreitol, 1 mM benzamidine, and 0.2 mM phenymethanesulfonyl fluoride) (buffer A). The homogenate was centrifuged at 3,000 ϫ g for 5 min, then the supernatant solution was recentrifuged at 33,000 ϫ g for 30 min, and the resulted supernatant solution was applied to a Blue Sepharose column (0.7 ϫ 5 cm) which had been equilibrated with buffer A (4). The column was washed with 40 ml of buffer A, and the enzyme was eluted with 5 ml of buffer A containing 2 M NaCl. The major fraction of the enzyme was eluted between 2 and 3 ml which were pooled for assay of Fru-6-P,2-kinase and Fru-2,6-Pase activities.
Other Methods-Protein concentration was determined with the Bradford method (28) using crystalline bovine serum albumin as a standard.

RESULTS
Effect of Fatty Acids-Livers from 24-h starved rats or rats fed with the standard diet were perfused with 40 mM glucose for 10 min followed by acetate (5 mM), propionate (10 mM), or butyrate (5 mM) for 10 min. During the initial perfusion with a high concentration of glucose, the liver Fru-2,6-P 2 concentration in the starved rat increased from 2.6 to 8.3 nmol/g (Table  I) and from 8.7 to 12.6 nmol/g in the fed rat liver (Table II). The continued perfusion with 40 mM glucose for an additional 10 min did not further increase the Fru-2,6-P 2 concentration (data not shown). Glucose, Glu 6-P, Fru-6-P, and Xu-5-P also increased similarly in both groups of the rat livers. Following the initial glucose perfusion the Fru-6-P,2-kinase and Fru-2,6-Pase activity ratios (v/V max ) were 0.93 and 0.67 in both the starved and ad libitum fed rat livers, respectively, indicating that the bifunctional enzyme was mostly in dephospho form in the glucose-perfused livers. The activity ratio of Fru-2,6-Pase at two different concentrations of Fru-2,6-P 2 for the phosphorylated and dephosphorylated forms are 1 and 0.4, respectively (4, 10). The activity ratios of Fru-6-P,2-kinase at two different concen- trations of Fru-6-P are v/V max ϭ 1 for a fully dephosphorylated form and v/V max ϭ 0.4 for fully phosphorylated form (4,10).
During the subsequent perfusion of the starved rat livers with acetate, propionate, or butyrate, Fru-2,6-P 2 decreased 40 -50% and Xu 5-P also decreased 40 -60%. The activity ratio of Fru-6-P,2-kinase decreased, while the activity ratio of Fru-2,6-Pase increased, indicating that the enzyme became more phosphorylated. Glucose and hexose-Ps were decreased 70 -90% by the fatty acid perfusion, but citrate increased 36 -126% (Table I). The ratios of Fru-2,6-P 2 /Xu 5-P concentration was fairly constant at 0.38 -0.57 in the starved rat livers perfused with high glucose or short chain fatty acids. On the other hand, the ratios of Fru-2,6-P 2 /citrate varied from 0.008 to 0.029, indicating no close correlation between citrate and Fru-2,6-P 2 concentrations.
Similarly the fatty acids caused 40 -50% and 50 -55% decrease in the Fru-2,6-P 2 and Xu-5-P concentrations, respectively, in the fed rat livers. The ratios of Fru-2,6-P 2 /Xu 5-P in these fed rats varied between 0.068 and 0.091, remaining relatively constant in spite of the fact that Xu-5-P concentrations were an order of magnitude higher in those livers than in those from the starved rat livers. In contrast to the fasted livers, however, the drop in glucose and hexose-Ps was less than 75 and 25%, respectively. The citrate concentration increased 33-243%, and again there was little correlation between citrate and Fru-2,6-P 2 (compare acetate versus propionate; Table II).
The time courses of decreases in Fru-2,6-P 2 , Xu-5-P, and glucose concentrations with butyrate perfusion in starved and fed rat livers showed that both Fru-2,6-P 2 and Xu-5-P decreased slowly over the period of 10 min, but glucose decreased more rapidly for the first 2 min in both starved and fed livers (data not shown). This decrease in the Fru-2,6-P 2 concentration was the result of inactivation of Fru-6-P,2-kinase and activation of Fru-2,6-Pase in both fasted and fed livers (Fig. 1,  A and B).
The results showed that these short chain fatty acids decreased glucose utilization, as indicated by the fact that the glucose and hexose-P levels in the perfused livers dropped significantly. Fru-2,6-P 2 and Xu-5-P did not decrease as much as glucose or hexose-Ps, but more importantly, both decreased nearly in concert in both fed and starved rat livers. Furthermore, the fatty acids caused conversion of the dephospho form of Fru-6-P,2-kinase:Fru-2,6-Pase to the phospho-form, judging from the changes in the activity ratios.
Effect of Different Hexoses-To test specificity for glucose, starved and fed rat livers were perfused for 10 min with glucose, fructose, galactose, or mannose. The higher concentration of some of the sugars could not be perfused due to a significant drop in the ATP level in liver (29,30). The results (Table III) showed that all these hexoses were able to increase the Fru-2,6-P 2 and Xu-5-P levels 50 -90% and 67-104%, respectively, compared to those perfused with 5 mM glucose in the starved rat livers. The concentration ratios of Fru-2,6-P 2 /Xu-5-P were the same among all those sugars, in spite of the fact that the Fru-2,6-P 2 level doubled with some of the sugars (compare fructose in Table III and galactose in Table IV). The increase in Fru-6-P,2-kinase and the decrease in Fru-2,6-Pase activity ratios corresponded closely with the changes in Fru-2,6-P 2 as well as Xu-5-P. However, there was less close correspondence between Fru-2,6-P 2 , Xu-5-P, or the enzyme activity ratios with the changes in the glucose, hexose-P, and citrate levels. Thus, these results suggested that other hexoses besides glucose were able to increase the levels of both Xu-5-P and Fru-2,6-P 2 and also stimulate dephosphorylation of the bifunctional enzyme.
In the fed rat livers, similar results were obtained with all the hexoses. However, mannose and 40 mM glucose increased Xu-5-P by 4 -9-fold, and Fru-2,6-P 2 level increased 23-39% compared to the liver perfused with 5 mM glucose. Thus, there were large variations in Fru-2,6-P 2 /Xu-5-P ratios, especially in the fed rat liver perfused with 40 mM glucose because of an extremely high concentration of Xu-5-P.
Effect of Glucagon and Insulin-Glucagon lowered Fru-2,6-P 2 and Xu-5-P by 63 and 50%, respectively, while cAMP increased 67% in the livers of fed rats perfused with the hormone (Table V). Glucose and hexose-Ps also decreased by approximately 25%. This decrease in Fru-2,6-P 2 could be attributed to increased phosphorylation of the bifunctional enzyme since the activity ratio of Fru-6-P,2-kinase decreased. Insulin had little effect on those metabolites and on the bifunctional enzyme in the fed livers. The similar lack of effect of insulin has been reported with hepatocytes isolated from fed rat livers (31). Citrate increased slightly with glucagon administration while Fru-2,6-P 2 decreased 63%, indicating no corresponding changes between these metabolites. The metabolite contents in the livers of the starved rats perfused with glucagon or insulin were not different from those in the control livers (data not shown).
Effect of Feeding High Carbohydrate or High Fat Diet-Starved rats (48 h) were fed various diets for 3 h, killed, and the livers removed immediately for analysis. The starved rats had the lowest concentration of Fru-2,6-P 2 (2.6 nmol/g) (Table VI), as expected, while the ad libitum fed rats contained the highest level of the hexose-P 2 (8.7 nmol/g). Feeding the high sucrose diet after starvation also raised the Fru-2,6-P 2 to near the maximum level (8.3 nmol/g liver). A high fat diet (30% fat) containing 26% starch significantly raised the Fru-2,6-P 2 to 4.4 nmol/g and high fat (30% fat), but a no carbohydrate diet increased it only to 3.6 nmol/g.
The activity ratio of Fru-6-P,2-kinase also showed the corresponding changes, i.e. the highest with ad libitum and high carbohydrate diets, intermediate in high fat with no carbohydrate, and the lowest in the starved rat livers. Opposite effects were seen in the activity ratios of Fru-2,6-Pase. Xu-5-P showed values ranging from 6.3 to 17.3 nmol/g with different diets similar to the values seen with Fru-2,6-P 2 , except for rats fed a high sucrose diet in which the Xu-5-P level rose to 52 nmol/g. The Xu-5-P contents in the starved, ad libitum fed, and high carbohydrate diets were very similar to those reported previously by Casazza and Veech (19). The ratios of Fru-2,6-P 2 /Xu-5-P were nearly the same in all these livers except those from rats fed with a high sucrose diet in which the Xu-5-P concentration was 3ϫ and 13ϫ higher than the 48-h starved and the ad libitum fed rats, respectively. Glucose and hexose-Ps concentrations in these livers did not show large differences (5.1-9.6 mol/g) while Fru-2,6-P 2 and Xu-5-P concentrations varied significantly more among different diets. There was no correlative changes between citrate and Fru-2,6-P 2 in these fed rats. DISCUSSION One of the objectives of this investigation is to obtain in vivo evidence in support of the proposed role for Xu-5-P in regulating Fru-2,6-P 2 which ultimately regulates glucose metabolism in liver. In the present study, we demonstrated that perfusion of livers with short chain fatty acids or feeding high fat diet to whole animals produced a decrease in Fru-2,6-P 2 in liver, which was correlated with increased phosphorylation of the bifunctional enzyme, resulting in inhibition of the kinase and activation of the phosphatase. This increased phosphorylation state of Fru-6-P,2-kinase:Fru-2,6-Pase appeared to be closely correlated to a decreased Xu-5-P level in both the perfused  livers and the livers of rats fed a high fat diet. Such correlation was not seen with hexose-Ps. A similar correlation between Xu-5-P and Fru-2,6-P 2 was seen with the livers perfused with other sugars. Glucagon, which raises cAMP and is known to lower Fru-2,6-P 2 (reviewed in Refs. 1-3), also decreased Xu-5-P concentration. Moreover, when the increase in the Fru-2,6-P 2 and Xu-5-P concentrations were compared in the livers of rats fed with different diets (shown in Fig. 2, closed circles), the Fru-2,6-P 2 level reached the maximum value at 20 nmol/g Xu-5-P and the half-maximum at approximately 10 nmol/g (except those fed a high sucrose diet). The in vivo results were similar to the in vitro results obtained using pure Xu-5-Pactivated PP2A (11). When the activation of pure PP2A by varying Xu-5-P was determined (Fig. 2, open circles), K a Xu5P was also 10 M, and the maximum activation was obtained above 40 M Xu-5-P. Thus, there was remarkable agreement in the correlation of Fru-2,6-P 2 and Xu-5-P levels between in vivo and in vitro data and adds credence to our claim that Xu-5-P-activated PP2A plays an important role in the regulation of Fru-2,6-P 2 level in liver. Xu-5-P in rats fed with a high carbohydrate diet increased to 77 nmol/g (10) or 52 nmol/g (Table V), but Fru-2,6-P 2 (8.3 nmol/g) did not increase above that in ad libitum fed liver. This can be explained by the fact that the PP2A was saturated with Xu-5-P at 20 nmol/g, consequently higher concentrations of Xu-5-P would not have activated the enzyme further. Thus, the correlation between Fru-2,6-P 2 and Xu-5-P ended when Xu-5-P concentration increased above 20 nmol/g. The results of the

of sugars on metabolites and Fru-6-P,2-kinase and Fru-2,6-Pase in perfused livers of ad libitum fed rats
The livers were perfused with various sugars for 10 min. The activity ratio of the bifunctional enzyme and the content of the metabolites were determined as described under "Experimental Procedures." Values are mean Ϯ S.D. (n ϭ 4).     (Table III), however, are difficult to explain because in these livers Xu-5-P decreased from 171 nmol/g (in high glucose liver) to about 80 nmol/g with the fatty acid perfusion and the Fru-2,6-P 2 level decreased from 12.6 to about 6 nmol/g. If the PP2A was saturated at 20 nmol/g Xu-5-P, one would not expect to see any decrease in Fru-2,6-P 2 by fatty acid administration. We cannot offer any reasonable explanation for this observation. All these in vitro and in vivo results are consistent with our proposal that Xu-5-P serves as the key messenger for glucose or carbohydrate for the increased level of Fru-2,6-P 2 in liver by stimulating the specific PP2A to dephosphorylate the bifunctional enzyme. Furthermore, these results indicated that activation of the PP2A is extremely sensitive to the lowest concentration ranges of Xu-5-P in liver (levels comparable to those of Fru-2,6-P 2 ), thus making it suited for the role of a messenger. Our observation that other sugars alter the Xu-5-P level suggests that any substance, including nucleotide, nucleosides, triose-Ps, etc. which generates this pentose-P, may raise Fru-2,6-P 2 by the same mechanism. Thus, our original scheme (11) which showed only glucose, has been modified to include the other sugars as shown in Scheme I.
Another objective of this work was to determine whether the well known phenomenon of fatty acid inhibition of glucose metabolism can be explained by the Xu-5-P-mediated regulation of Fru-2,6-P 2 . Previously, this inhibition was attributed to allosteric inhibition of Fru-6-P,2-kinase by citrate (16). Indeed, the citrate level does increase upon fatty acid administration, as was well known before (16,32,33) and confirmed here (Table V). However, the correlation between Fru-2,6-P 2 and citrate in the livers of rats fed with various diets or in the perfused livers was poor compared to that of Fru-2,6-P 2 and Xu-5-P. Moreover, the observation that the decreased Fru-2,6-P 2 correlates with increased phosphorylation state of the bifunctional enzyme upon fatty acid administration rules out the citrate inhibition of the enzyme, because citrate would not affect the phosphorylation state of the enzyme. Finally, all the citrate levels reported in the literature represent total cell contents of citrate and not that in cytoplasm where the bifunctional enzyme exists. Thus, we suggest that so-called "fatty acid sparing" cannot be explained by the citrate inhibition of Fru-2,6-P 2 synthesis but can be explained by the decreased level of Xu-5-P as a result of fatty acid feeding which results in decreased activity of the PP2A. It has been demonstrated that glycolysis, as measured by lactate production, in isolated hepatocytes requires 5 nmol/g Fru-2,6-P 2 (34). If this were true in whole animals, glycolysis must be completely inhibited in the livers of those fed with fatty acids containing 4.4 or 3.6 nmol/g Fru-2,6-P 2 ( Table VI).
The question then is how Xu-5-P concentration is decreased by fatty acid administration. Xu-5-P is generated in the pentose-P pathway in two ways: (a) from 6-P gluconate and NADP catalyzed by 6-P gluconate dehydrogenase and ribulose 5-P epimerase; and (b) from Fru-6-P and glyceraldehyde 3-P by transketolase reaction. The increased fatty acid oxidation may inhibit both pathways by: (a) decreased NADP/NADPH ratio and (b) decreased Fru-6-P and glyceraldehyde 3-P by inhibition of glycolysis by decreased Fru-2,6-P 2 . Casazza and Veech (19) have shown that the transketolase reaction (as well as all the other enzyme reactions in the nonoxidative pathway) is in equilibrium in ad libitum fed and starved rats, thus any decrease in Fru-6-P and/or glyceraldehyde 3-P results in decreased Xu-5-P. This points to an interesting synergism between phosphofructokinase and the bifunctional enzyme. The Xu-5-P concentration is autoregulated by the synergistic regulation of these enzymes via changes in the Fru-2,6-P 2 concentration. The resulting changes in phosphofructokinase activity would affect both glyceraldehyde 3-P and Fru-6-P concentrations (Scheme II) because Fru-6-P is the substrate for phosphofructokinase and glyceraldehyde 3-P is the product of phosphofructokinase and aldolase (Fru-6-P 3 Fru-1,6-P 2 3 glyceraldehyde 3P ϩ dihydroxyacetone-P). Thus, although the Fru-6-P level increases immediately after meal feeding, for example, the Xu-5-P level does not rise simultaneously because glyceraldehyde 3-P remains low due to low Fru-2,6-P 2 concentration and low phosphofructokinase activity. The Xu-5-P level remains low until Fru-2,6-P 2 begins to form, which activates phosphofructokinase to generate glyceraldehyde 3-P. Thus, both Xu-5-P and Fru-2,6-P 2 concentrations are coordinately regulated and depend on each other. This explains the previous observation (10) that Fru-6-P increases immediately in liver perfused with high concentration of glucose, but there is a lag period (4 min) in the Xu-5-P formation. Similarly, it may also explain a lag period of 2-4 h observed in Fru-2,6-P 2 formation in the livers of starved rats fed with regular lab chow (35,36). Previously, this delay was suggested to be related to accumulation of a critical concentration of glycogen before an increase in Fru-2,6-P 2 (35). These observations also explain why Xu-5-P, among all the pentose-P in the shunt pathway, is selected as an activator for the PP2A. The answer appears to be that Xu-5-P is a product of the transketolase reaction from Fru-6-P and glyceraldehyde 3-P, both of which are intermediates of glycolysis and considered as the branching point from glycolysis and the initial step in the pentose shunt pathway.
In the fatty acid (short chain) perfused livers, the cAMP concentration increased from 0.7 to 0.8 nmol/g to 1.1-1.5 nmol/g in both starved and fed animals (Tables I and II). Since the phosphorylation state of Fru-6-P,2-kinase:Fru-2,6-Pase is determined by the relative activities of protein kinase A and PP2A (Scheme I), the results may suggest that the phosphorylation of the bifunctional enzyme was increased further by the activation of protein kinase A by increased cAMP. For a number of reasons, however, it is difficult to assess the physiological significance of the increased cAMP level in the fatty acid perfused livers. The cAMP concentration in the freezeclamped livers represents the value in the total liver and not FIG. 2. The relationship between Fru-2,6-P 2 and Xu-5-P in rat livers and the activation of PP2A by Xu-5-P. The activation of pure PP2A (E) by Xu-5-P is from Nishimura and Uyeda (11). The Fru-2,6-P 2 (q) and Xu-5-P levels were from the data in Tables II, III, and VI. SCHEME II. The abbreviations used are the following: Glc, glucose; G6P, Glc-6-P; F6P, Fru-2,6-P 2 ; GAP, glyceraldehyde 3-P; DHAP, dihydroxyacetone-P; Lac, lactate; TK, transketolase; F6P2K:F26Pase, Fru-6-P,2-kinase:Fru-2,6-Pase; P-F6P2K:2P, phosphorylated.
that in the cytoplasm. A similar increase in cAMP was not observed in the fatty acid fed animals (Table VI). Nevertheless, the observation that the phosphorylation state of Fru-6-P,2kinase:Fru-2,6-Pase was increased by fatty acids in both perfused liver and in the whole animal indicates that either protein kinase A activity remained the same or increased, while PP2A activity was inhibited as a result of lower Xu-5-P concentration. Thus, Xu-5-P and the PP2A appear to play a major role in regulation of this complex pair of reciprocating reactions, protein kinase A and PP2A, and Fru-6-P,2-kinase and Fru-2,6-Pase under these dietary conditions.
It is generally thought that the role of the pentose shunt pathway is to provide the source of: (a) the reducing equivalent in the form of NADPH for fatty acid and steroid biosynthesis and (b) ribose 5-P for nucleic acid and nucleotide biosynthesis. We presented herein the third role of the pathway: production of Xu-5-P which plays a vital role in regulation of glycolysis as well as the overall carbohydrate metabolism in liver.
In summary, we have shown that Fru-2,6-P 2 in liver, which is primarily determined by the phosphorylation states of Fru-6-P,2-kinase:Fru-2,6-Pase as we stated before (37), was correlated with changes in Xu-5-P in the livers perfused with different hexoses, fatty acids, and glucagon. A similar correlation between Xu-5-P and Fru-2,6-P 2 was demonstrated in rats fed with a high carbohydrate or high fat diet and in starved rats. These in vitro and in vivo results provide evidence in support of the proposed mechanism for regulation of the bifunctional enzyme by phosphorylation mediated by cAMP and dephosphorylation mediated by Xu-5-P.