the journal of biowcical chemistry vol. no. issue of ... · the journal of biowcical ......

THE JOURNAL OF BIOWCICAL CHEMISTRY Vol. 269, No. 42, Issue of October 21, pp. 26100-26106, 1994 Printed in W.S.A.

Glucose-stimulated Synthesis of Fructose 2,G-Bisphosphate in Rat Liver DEPHOSPHORYLATION OF FRUCTOSE 6-PHOSPHATE,Z-KINASE:FRUCTOSE 2,6-BISPHOSPHATASE AND ACTIVATION BY A SUGAR PHOSPHATE*

(Received for publication, May 18, 1994, and in revised form, August 8, 1994)

Motonobu Nishimura, Sergei FedorovS, and Kosaku UyedaO From the Department of Veterans m a i m Medical Center and the Department of Biochemistry, The University of lbxas Southwestern Medical Center, Dallas, Texas 75216

The effect of glucose on hepatic fructose (Fru) 2,6-P2 in starved rats was investigated. When livers were perfused with high glucose (40 m ~ ) , hexose-P in the liver increased immediately reaching the maximum within in 2 min, but Fru 2,6-P2 after a lag period of 4 min increased linearly. The activation of Fru 6-P,2-kinase and inactivation of Fru 2,6-Pase also showed a similar lag period. Determination of the phosphate contents of the bifunctional enzyme after 10 min of glucose perfusion revealed that 90% of the enzyme was in the dephospho form while only lWo of the control liver enzyme was dephosphorylated.

Comparison of crude extracts of liver perfused with either high glucose or normal glucose (5.6 m ~ ) showed that high glucose livers contained 50%0 higher protein phosphatase activity, which dephosphorylated the bifunctional enzyme. Subcellular fractionation of the extract showed that activation of the protein phosphatase occurred in the cytosol. Desalting of the cytosolic fraction resulted in a 50% loss of the protein phosphatase activity. The low molecular weight activator in the cytosol was isolated, and by various chemical and enzymatic methods it was identified as xylulose 5-P. The activation of protein phosphatase by xylulose 5-P showed a highly sigmoidal saturation curve. The rate of formation of xylulose 5-P in the perfused liver showed a lag period of approximately 2 min, and after 4 min its concentration reached 10 p ~ , the minimum concentration necessary for the activation of the protein phosphatase.

We conclude that the mechanism of glucose-induced Fru 2,6-P, synthesis was not due to increased Fru 6-P as generally thought but occurred as a result of dephosphorylation of Fru I-P,2-kinase:Fru 2,6-Pase. Moreover, the dephosphorylation was enhanced by increased xylulose 5-P, which activated a specific protein phosphatase. The results suggest a mechanism for coordinated regulation of glycolysis and the pentose shunt pathway that is mediated by xylulose 5-P.

Glucose is known to stimulate both glycolysis and glycogen synthesis in liver (reviewed in Refs. 1,2). When a high concen-

* This work was supported by grants from the Department of Veter- ans Affairs and the National Institute of Diabetes and Digestive and Kidney Diseases. 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.

$ Present address: Dept. of Pharmacology, The University of Texas Southwestern Medical Center at Dallas, Dallas, TX.

Q To whom correspondence should be addressed: Dept. of Veterans Affairs Medical Center, 4500 S. Lancaster Rd., Dallas, TX 75216. Tel.: 214-372-7028; Fax: 214-372-7948.

tration of glucose is administered to fasted animals, the liver takes up glucose. Some glucose is converted to glycogen, and the rest is converted to lactate and fatty acids. Insulin and other hormones have a significant effect on this transition in glucose metabolism from catabolism to anabolism, but glucose by itself is able to turn on both glycogen synthesis and glycolysis. The mechanism of the stimulation of glycogen synthesis by glucose is relatively well understood (reviewed in Refs. 1-4). The binding of glucose to phosphorylase a stimulates phosphorylase phosphatase, and phosphorylase a is converted to a less active form, phosphorylase 6. This decrease in phosphorylase a concentration results in the activation of glycogen synthase phosphatase which in turn activates glycogen synthase.

The mechanism of the stimulation of glycolysis by glucose, however, is unclear. The administration of glucose to isolated rat hepatocytes or refeeding high carbohydrate diet to starved rats (5) increases fructose (Fru)l 2,6-P, level (5-7) and lactate production (9, lo), indicating that the increased glycolysis is the result of increased Fru 2,6-P,, which is the most potent activator of phosphofructokinase. Hexose-P also increases in a dose-dependent manner in cells incubated with glucose (5,7,9). Glucose utilization in hepatocytes is also reported to be stimulated by low concentration (0.25 mM) of fructose (11). This stimulation is also attributed to the increased Fru 2,6-P, as a result of increased Glc 6-P but not to any changes in the activity of Fru 6-P,2-kinase. Since Fru 6-P is the substrate for Fru 6-P,B-kinase and an inhibitor of Fru 2,6-Pase, it is assumed that the effect of glucose is the increased availability of the hexose-P (10). Liver Fru 6-P,2-kinase:Fru 2,6-Pase can be regu- lated in a reciprocal manner, i.e. inhibition of the kinase and activation of the phosphatase by phosphorylation catalyzed by CAMP-dependent protein kinase (12-14). Dephosphorylation of the enzyme produces the opposite effect, i.e. activation of kinase and inhibition of phosphatase, which results in increased Fru 2,6-P2 (12). However, no attempt was made to determine directly the phosphorylation state of the bifunctional enzyme in these glucose-stimulated hepatocytes. The regulation of an al- losteric enzyme, such as Fru 6-P,2-kinase:Fru 2,6-Pase, by its substrate (Fru 6-P) is one possible mechanism since the enzyme exhibits sigmoidal kinetics for the substrate (15). How- ever, regulation by phosphorylation and dephosphorylation mechanism is the result of hormonal and nutritional interven- tion and often is a more direct and efficient means of regulating enzyme activity than changes in substrate. In this paper we

6-phosphate,2-kinase:fructose-2,6-bisphosphatase; Glc 6-P, glucose The abbreviations used are: Fru 6-P,2-kinase:Fru 2,6-Pase, fructose

6-phosphate; Fru 6-P, fructose 6-phosphate; MOPS, 4-morpholinepro- panesulfonic acid; RT2KS30, mutant rat testis Fru 6-P,S-kinase:Fru 2,6-Pase containing SeP; Xu 5-P, xylulose 5-P; HPLC, high perform- ance liquid chromatography.

26100

Glucose-induced Activation of Fru 6-82-kinase:Fru 2,6-Pase 26101

have investigated the mechanism of glucose-induced synthesis of Fru 2,6-P, using perfused liver.

EXPERIMENTAL PROCEDURES Material~-[y-~'P]ATP (3000 Ci/mmol) was purchased from Amer-

sham Corp. The catalytic subunit of CAMP-dependent protein kinase was purchased from Promega (Madison, WI). Antiserum against rat liver Fru 6-P,2-kinase:Fru 2,6-Pase was raised in a goat as described (16). Rabbit anti-goat IgG was purchased from Miles-Yeda, Ltd. (Elkhart, IN). Rabbit muscle phosphofructokinase (17) and rat liver Fru 6-P,2-kinase:Fru 2,6-Pase (15) were prepared as before. A mutant enzyme of rat testis Fru 6-P,2-kinase:Fru 2,6-Pase (RT2KS30) which con- tains a phosphorylation site a t S e P and is similar to the liver isozyme with respect to the kinetic properties (18) was used as a substrate for the protein kinase and protein phosphatases. This enzyme is over- expressed in Escherichia coli and was purified as described (18,19). All other enzymes were obtained from either Boehringer Mannheim or Sigma All other chemicals and resins were analytical reagent grade and purchased from commercial sources. A CarboPac PA1 column was purchased from Dionex Corp. (Sunnyvale, CAI.

Rats-Male Sprague-Dawley rats weighing 200-250 g purchased from Sasco Co. (Omaha, NE) were used. Rats were fed ad libitum with the standard NIH diet or fasted for 24 h.

Liver Perfusions-Rats were anesthetized by intraperitoneal injec- tion of sodium pentobarbital (60 mg/kg), and the livers were perfused in situ with Krebs-Henseleit bicarbonate buffer equilibrated with 95% 0, 5% CO,. The medium also contained either 5.6 or 40 nm glucose. Flow rate was set to approximately 4 ml/min/g liver.

Metabolite Measurements-At the indicated time a piece of perfused liver was cut out and immediately freeze-clamped with a tong precooled in liquid nitrogen. The frozen tissue samples were ground in a porcelain mortar prechilled with liquid nitrogen. Fru 2,6-P, was extracted from the tissue powder with 0.1 M glycine/NaOH buffer, pH 10.0 (1:3 weight/ volume) and assayed by the method of Thomas and Uyeda (20). Xu 5-P is assayed by the method of Casazza and Veech (21) except the formation of Fru 6-P coupled to NADPH formation was measured fluorometrically by adding phosphoglucose isomerase (1 unit) and Glc 6-P dehydrogenase (0.4 unit) with excitatiou and emission wave lengths a t 354 and 452 nm, respectively. All other metabolites were assayed spectro- photometrically on perchlorate (3.6%) extracts oflivers using enzymatic methods as described (22).

Assay Method for Fru 6-P,S-Kinase-The reaction mixture contained in a final volume of 0.1 ml: 100 mM Tris-HC1, pH 7.5,O.l mM EDTA, 0.1 nm EGTA, 5 nm ATP, 5 mM phosphate, 10 nm MgCl,, and indicated amounts of Fru 6-P. The mixture was incubated at 30 "C, and at timed intervals 10-pl aliquots were transferred into 90 pl of 0.1 N NaOH, and the solution was heated for 1 min at 80 "C to stop the reaction. Suitable aliquots of the heated reaction mixture were then assayed for Fru 2,6-P, as described. One unit of activity is defined as the amount of enzyme that catalyzes the formation of 1 pmol of Fru 2,6-P,Jmin under these conditions.

Assay Method for Fru 2,6-Pase-This coupled assay measures con- tinuously the formation of Fru 6-P coupled to NADPH formation using phosphoglucose isomerase and Glc 6-P dehydrogenase as described pre- viously (19,231. The reaction mixture contained in a final volume of 1.0 ml: 100 mM Tris-HC1, pH 7.5, 0.2 mM EDTA, 100 w NADP, 17 1.1~ Fru 2,6-P,, 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 at 354 and 452 nm, respectively.

Partial Purification of Fru 6-P,S-kinase:Fru 2,6-Pase from Perfused Rat Liuers-Apiece of perfused liver (approximately 4 g) was excised at an indicated time and quickly homogenized using a Polytron homoge- nizer in 10 ml of ice-cold 50 mM Tris-phosphate, pH 8.0, 100 mM KF, 5 mM EDTA, 5 mM EGTA, 1 nm dithiothreitol, 1 mM benzamidine, and 0.2 mM phenylmethanesulfonyl fluoride (buffer A). The homogenate was centrifuged at 33,000 x g for 30 min, and the supernatant solution (4 ml) was applied on a Blue Sepharose column (0.7 x 5 cm) which had been equilibrated with buffer A. The column was washed with 40 ml of buffer A, and the enzyme was eluted with buffer A containing 2 M NaCl(5 ml). Fractions (1 ml) were collected and assayed for Fru 2,6-Pase activity. Usually, fraction 3 contained most of the enzyme, and the recovery of the enzyme was approximately 84 5% of that in the crude extract.

Determination of P'PIPhosphate Incorporation in Fru 6-P,S-kinase: Fru 2,6-Pase of Perfused Rat Liuers-The reaction mixture contained in final volume of 0.3 ml: 0.1 milliunit of Fru 6-P,B-kinase:Fru 2,6-Pase, 50

mM potassium phosphate, pH 7.4,0.2 mM [y-32P1ATP (1000 countsfmid pmol), 5 mM MgCl,, 0.2 mM EDTA, 0.2 mM EGTA, and 2 mM dithiothreitol. The reaction was initiated by addition of catalytic subunit of CAMP-dependent protein kinase (100 units), and the reaction mixture was incubated at 30 "C for 30 min. At the end of incubation the mixture was diluted with 2 ml of ice-cold potassium phosphate, pH 7.4, and 0.15 M NaCl (PBS) and concentrated to minimal volume by Centricon-30 filter (Amicon-Beverly, MA). This dilution and concentration was re- peated three times. Pure rat liver Fru 6-P,2-kinase:Fru 2,6-Pase (0.1 milliunit) as a carrier and 20 pl of antiserum against Fru B-P,a-kinase: Fru 2,6-Pase were added to the concentrated 32P-labeled enzyme solution, and the mixture was incubated for 14 h at 4 "C. Rabbit anti-goat IgG (0.3 mg) was added to the reaction mixture and incubated at 37 "C for 2 h. The precipitate was removed by centrifugation at 5,000 x g for 10 min, washed three times with PBS, dissolved in 65 mM Tris-HC1, pH 6.8, containing 5% P-mercaptoethanol, 2% sodium dodecyl sulfate. A suitable aliquot of the dissolved protein solution was removed and mixed with scintillation mixture (Optifluor-Packard, CT), and radioac- tivity was determined with a scintillation counter (TM Analytic).

Polyacrylamide Gel Electrophoresis-Polyacrylamide slab gel electrophoresis in sodium dodecyl sulfate was carried out in 10% acrylam- ide, containing 0.1% sodium dodecyl sulfate according to the procedure of Laemmli (24). The gels were stained with Coomassie Brilliant Blue and destained. For autoradiography of the gels, Kodak X-Omat AR film was used.

Preparation of Liver Extract for Protein Phosphatase Assay-A piece of perfused liver was excised at the indicated time and homogenized immediately in a motor-driven Potter-Elvehjem tube in 1 volume of extraction buffer containing 0.25 M sucrose, 50 mM imidazole, pH 7.4, 1 nm dithiothreitol, 4 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, and 1 mM benzamidine, and the extract was centrifuged at 10,000 x g for 10 min. The resulting post-mitochondrial supernatant solution (extract) was fractionated by centrifugation (35 min at 220,000 x g ) into a cytosolic fraction and a glycogedmicrosomal fraction. The latter pellet was resuspended in the same extraction buffer to the volume of the original extract. The samples were desalted before assay by passing over a Sephadex G-50 spin column (0.5 x 5 cm) equilibrated in the extraction buffer without EDTA, if necessary. The protein concentration of the desalted sample was determined.

Preparation of Phosphorylated Substrates-Pure rat testis mutant bifunctional enzyme RT2KS30 (approximately 500 pg, 9.1 nmol) was phosphorylated in reaction mixtures containing 50 mM Tris-phosphate, pH 7.5,0.2 mM [y-32PlATP (1000 counts/midpmol), 10 mM MgCl,, 0.2 mM EDTA, 0.2 mM EGTA, 2 nm dithiothreitol, and 5.6 pg of the catalytic subunit of cAMP-dependent protein kinase. Pure rat liver bifunctional enzyme was phosphorylated with the same procedure. The reaction mixtures were incubated at 30 "C for 40 min and then passed twice over a Sephadex G-50 spin column (0 .5~5 cm) equilibrated with 20 mM MOPS, pH 7.0, and 1 mM dithiothreitol to remove free [32P]ATP. The specific activity of [32PlRT2KS30 was approximately lo6 countsfmid nmol.

Protein Phosphatase Assays-A reaction mixture contained in a final volume of 50 pl 20 mM MOPS, pH 7.0, 1 mM dithiothreitol, 25 pg of bovine serum albumin, and the [32Plphosphate-labeled substrates (4 w) for 10-20 min at 30 "C. Since large amounts of the mutant rat testis Fru 6-P,2-kinase:Fru 2,6-Pase enzyme (RT2KS30) are available in our lab- oratory, it was used as a substrate in most of the experiments described herein. An aliquot (15 pl) was removed from the reaction mixture at a given time and precipitated by adding 30 pl of cold trichloroacetic acid (10% final concentration) and 30 pl of 6% bovine serum albumin as a carrier. The precipitate was removed by centrifugation at 10,000 x g for 10 min, and the 32Pi in the supernatant solution was determined by mixing 60 p1 of solution with 5 ml of scintillation mixture (Optifluor- Packard, CT) and counting in a scintillation counter. One unit of protein phosphatase is defined as that amount which catalyses the release of 1.0 pmol of 32Pj/min.

Assay Method for the Protein Phosphatase Activator-The activator was assayed based on their ability to activate the protein phosphatase activity in the crude liver extract using 32P-labeled RT2KS30 as a substrate. The reaction mixture was the same as the protein phosphatase assay described above except a varying concentration of the activator was added. The cytosolic fraction of the extract of the liver perfused with 40 r m glucose for 10 min was used in the assay after desalting using a Sephadex G-50 spin column. Usually 50 p1 of the reaction mixture contained the amount of the extract equivalent to 5 mg of liver.

Other Methods-Protein concentration was determined with the Bradford method (25) using crystalline bovine serum albumin as a standard. Statistical analysis was performed by the unpaired t test.

26102 Glucose-induced Activation of Fru 6-P,B-kinase:Fru 2,6-Pase Tmm I

Effect of glucose on Fru 2,6-P, contents in perfused livers

Procedures.” Control perfused with 5.6 mM glucose; glucose perfused The experimental conditions were as described under “Experimental

with 40 mM. After 10 min of perfusion, the livers were freeze-clamped and assayed for Fru 2,6-P,. Values are mean t S.E. of six livers.

Rats Fru 2,6-P,

Control Glucose

nmollg liver Fed 8.1 f 0.7 Starved

11.4 -c 1.8 1.1 t 0.3 5.8 f 0.7

RESULTS

Increase in Hepatic Fru 2,6-P2 Content by High Concentra- tion ofGZucose-Livers from fed and fasted rats were perfused for 10 min with “normal” glucose (5.6 mM; control) or high glucose (40 mM; glucose) medium and the hepatic Fru 2,6-P, contents were determined. The Fru 2,6-P, contents were higher in the glucose-perfused livers than the control livers both in the fed and the fasted rats, but the difference was much greater in the fasted rats (5.3 x versus 1.4 x) (see Table I). If this rise in Fru 2,6-P2 were as a result of increased Fru 6-P in the glucose- perfused livers, one would expect the kinetics of these sugar-P formation to coincide. Fig. 1 shows the changes in the Glc 6-P, Fru 6-P, and Fru 2,6-P, levels of the fasted rat livers during the course of perfusion with high glucose medium. The Glc 6-P and Fru 6-P concentrations increased rapidly without any lag period, reaching the maximum levels within 2 min of the high glucose perfusion. However, the Fru 2,6-P, formation showed a lag period of 4 min and reached the maximum after 8-10 min. These results suggested that the increase in Fru 2,6-P, contents by the high concentration of glucose was not dependent on increased Fru 6-F! Furthermore, Glc 6-P or Fru 6-P was unable to activate Fru 2,6-P2 synthesis.

Changes in Fru 6-P,2-kinase:Fru 2,6-Pase Activity during Glucose Perfusion-In order to determine if this glucose-dependent increase in Fru 2,6-P, contents was a result of activation of Fru 6-P,2-kinase and inhibition of Fru 2,6-Pase, the activity ratio of Fru 6-P,2-kinase at two different concentrations of Fru 6-P was determined in the perfused livers. The kinase was assayed at 0.05 mM (v) and 1.0 mM (V,,,) of Fru 6-P, where UN,,,,, = 1 for a fully dephosphorylated form of the bifunctional enzyme and UN,,, = 0.4 for a fully phosphorylated form. For this assay the enzyme in the crude extract of a fresh liver was partially purified by Blue Sepharose chromatography. Recovery of the enzyme was at least 85% of the activity in the liver extract. As shown in Table 11, the activity ratio of the enzyme in the fed rat liver increased from 0.81 to 0.95, while that in the fasted rat liver increased from 0.67 to 0.90. Thus, the increase was significantly greater in the enzyme from the fasted than from the fed liver. These results suggested that the hepatic Fru 6-P,2-kinase:Fru 2,6-Pase in the fasted rats was mostly in the phospho form before perfusion and that perfusion with a high concentration of glucose dephosphorylated the enzyme.

The kinetic changes in both Fru 6-P,2-kinase activity ratio and Fru 2,6-Pase activity with time after the initiation of glucose perfusion were also determined. The results demonstrated that the kinase activity ratio remains constant at below 0.5 for 4 min and thereafter increased rapidly to 0.8 after 10 min (Fig. 2). The phosphatase activity with time showed exactly opposite results. These results coincided with the changes in Fru 2,6-P, concentration with the perfusion time in the liver (Fig. 1) and appeared to be caused by the activation of the kinase and inhibition of the phosphatase. Such reciprocal changes in these enzyme activities indicated that dephosphorylation of the phosphorylated bifunctional enzyme was occurring after 4 min of exposure to high glucose concentration.

J

50 1 5

0 2 4 6 8 10

Perfusion Time (rnin)

FIG. 1. Changes in the Glc 6-P, Fru 6-P, and Fru 2,6-P2 levels with time during high glucose perfusion (40 m ~ ) in fasted rat livers. Glc 6-P, 0; Fru 6-P, 0 , Fru 2,6-P,, 0; levels in freeze-clamped livers were determined as described under “Experimental Procedures.” Values are mean t S.E. of six livers for each time point.

TABLE I1 Change in the Fru 6-P,B-kinase activity ratios and the phosphate

contents of the enzyme in the glucose-perfused livers The livers of the fed and fasted rats were perfused with the indicated

concentrations of glucose for 10 min. The bifunctional enzymes in the extracts were prepared as described under “Experimental Procedures,” and Fru B-P,B-kinase was assayed at 0.05 mM (v) and 1 mM CV,,,,) Fru

lyzed by protein kinase A was determined as described under ”Experi- 6-P. The phosphate incorporation into partially purified enzymes cata-

mental Procedures.” Values are mean S.D. of five livers of each sample.

Livers Glucose ulV,., Phosphate incorporation

mM mol/mllsubunit Fed 5.6 0.81 2 0.04 0.37 2 0.07

40 0.95 2 0.13 0.63 t 0.10 Fasted 5.6 0.67 f 0.12 0.10 f 0.03

40 0.90 0.10 0.91 * 0.12

Phosphorylation and Dephosphorylation of a Bifunctional Enzyme-To demonstrate directly whether the dephosphorylation of the enzyme had occurred during the glucose perfusion, the phosphate contents of the bifunctional enzymes in the livers were determined. The partially purified enzymes were phosphorylated by CAMP-dependent protein kinase in vitro in the presence of [32P]ATP, precipitated with specific antibodies, and analyzed by polyacrylamide gel electrophoresis. The amount of 32P incorporation in vitro was determined by the filter assay method of Corbin and Reimann (25) and indicated the amount of the dephospho form of the enzyme (Table 11). The Coomassie Blue-stained gel indicated the protein band corre- sponding to the Fru 6-P,2-kinase:Fru 2,6-Pase, and in addition, to heavy IgG bands were visible in all the samples (Fig. 3A). The autoradiogram of the same gel showed significant 32P incorporation into the enzymes from the liver perfused with high glucose medium (lanes 4 and 5, and 8 and 9, Fig. 3B), indicating that these enzymes were mostly in the dephospho form in vivo. This was in contrast to the enzymes from the liver perfused with control medium in which little 32P was incorporated (lanes 2 and 3, and 6 and 7, Fig. 3B), indicating that the enzymes existed in the phospho forms in these livers. The determination of stoichiometry of [32Plphosphate incorporation demonstrated that the enzyme in the fasted rat liver perfused with normal glucose concentration contained 0.1 mol of phos-

Glucose-induced Activation of Fru 6-Ee-kinase:Fru 2,6-Pase 26103

I I 0

A

B

1, l-1

0 2 4 6 8 10 Perfusion Time (min)

FIG. 2. Kinetic changes in both Fru B-P,a-kinase activity ratio and Fru 2,6-Pase activity during high glucose perfusion (40 m d of fasted rat livers. The bifunctional enzyme in the crude extract of a fresh liver was partially purified by Blue Sepharose before assay as described in “Experimental Procedures.” The Fru 6-P,B-kinase was assayed at 0.05 mM ( u ) and 1.0 mM (V,,,) of Fru 6-P and the ratio of the activities (urnmax) which represents the phosphorylation state of the enzyme was calculated (see text for details). Values are mean 2 S.E. of five livers for each time point.

phate incorporatedmol of enzyme subunit, indicating that 10% of the enzyme was in dephospho form (Table 11). However, when perfused with the high glucose concentration, 90% of the enzyme was in the dephospho form. Similarly, in the fed rats, 37% of the enzyme from the liver perfused in normal glucose concentration was in dephospho form, and the percentage of the dephosphorylated enzyme increased to 63% by high glucose perfusion (Table 11).

Changes in a Protein Phosphatase Activity ofFru 6-82-kinase: Fru 2,6-Pase in High Concentrations of Glucose-Livers of the 24-h fasted rats were perfused with either 5.6 mM glucose or 40 mM glucose for 10 min, and the activities of protein phosphatases in subcellular liver fractions were determined using the phosphorylated form of mutant rat testis Fru B-P,2-kinase:Fru 2,6-Pase (RT2KS30) as a substrate. The results, as summa- rized in Table 111, showed that the protein phosphatase activity in the post-mitochondrial supernatant solution of the liver perfused with high glucose medium was 50% higher than that of the liver perfused with normal glucose medium. The protein phosphatase activities were determined also in the cytosolic and the microsomaVglycogen fractions of the liver. The high glucose perfusion did not affect the protein phosphatase activities in the glycogedmicrosomal fractions of the liver. In contrast, phosphatase activities associated with the cytosolic fractions showed a 60% increase in the high glucose livers, and the increased activity observed in these liver extracts (postmito- chondrial supernatant) was recovered completely in the cytosolic fraction. These data suggested that the dephosphorylation of Fru 6-P,2-kinase:Fru 2,6-Pase induced by glucose was due to the activation of one or more cytosolic protein phosphatases. In addition, a similar difference in protein phosphatase activity in the cytosolic fraction was observed when the purified rat liver Fru 6-P,2-kinase:Fru 2,6-Pase was used as a substrate, demon- strating that the mutant testis enzyme is similar to the liver isozyme, justifying the use of the enzyme as a substrate for the protein phosphatase.

An Activator of a Cytosolic Protein Phosphatase-In order to characterize the specific cytosolic protein phosphatase cata- lyzing the dephosphorylation of FN 6-P,2-kinase:Fru 2,6-Pase which was activated by high glucose, the cytosolic fraction of the liver extract was desalted by Sephadex G-50 chromatography and assayed. The protein phosphatase activities in the

1 2 3 4 5 6 7 8 9 1 0 FIG. 3. Sodium dodecyl sulfate-polyacrylamide gel electro-

phoresis of phosphorylated Fru B-P,B-kinase:Fru 2,6-Pase from rat livers perfused with different concentrations of glucose. The enzyme was partially purified and phosphorylated as described under “Experimental Procedures.” The phosphorylated enzymes (0.1 milli- units) were subjected to gel electrophoresis, and the gels were stained with Coomassie Brilliant Blue and destained as described under “Ex- perimental Procedures.” Panel A, the gel was stained with Coomassie Brilliant Blue; lanes 1 and 10, pure enzyme as standard; lanes 2 and 3, livers from two fed rats perfused with 5.6 mM glucose; lanes 4 and 5, livers from fed rats perfused with 40 mM glucose; lanes 6 and 7, livers from 24-h fasted rats perfused with 5.6 mM glucose; lanes 8 and 9, livers

of the same gel showing the degree of phosphorylation of Fru 6-P,2- from 24-h fasted rats perfused with 40 mM glucose. B, an autoradiogram

kinase:Fru 2,6-Pase.

TABLE 111 Protein phosphatase activity for phosphorylated Fru 6”P,B-kinase:Fru

2,6-Pase in the rat liver extract The livers were perfused with the indicated glucose concentration for

10 min and fractionated by centrifugation as described under “Experi- mental Procedures.” Protein phosphatase in these fractions were assayed using the same concentration of either the phosphorylated RT2KS30 or the rat liver bifunctional enzyme as described under “Ex- perimental Procedures.” The numbers in parentheses were obtained with the rat liver bifunctional enzyme as the substrate instead of RT2KS30. These values are mean 2 S.E. of four to six rat livers.

Glucose cone. in Postmitachondrial perfusion buffer supernatant

Cytosol

m.u milliunitolg liver 5.6 0.90 2 0.03 0.40 = 0.06 0.50 0.09

(0.48 2 0.04) 40 1.31 2 0.09 0.43 2 0.09 0.79 2 0.03

(0.71 2 0.05)

desalted extracts of both control and glucose-treated livers de- creased by 17 and 26%, respectively, compared with those prior to the desalting (Fig. 4). The difference in the protein phosphatase activity between these two extracts also disappeared (Fig. 4). Similar loss of protein phosphatase activity was observed when these extracts were filtered through Centricon-3 (cut-off M, < 3,000) (Amicon-Beverly, MA). When the filtrate of the Centricon treatment was added back to the filtered liver extract, the activities of the protein phosphatase were completely recovered, and the difference between the control and the glucose extracts was also restored to their original values (Fig. 4).

It is known that protein phosphatase type 2C requires Mg2‘ while protein phosphatase type 2Ais activated by cations such as Mn2+ (reviewed in Ref. 27). As shown in Fig. 4, addition of Mn2+ and M e increased the protein phosphatase activity to the control level but failed to activate the protein phosphatase to the activity level found in the extracts of high glucose livers. These results ruled out type 2C phosphatase being involved in the activation. These results also showed that the Centricon filtrate contained a hitherto unknown activator (low molecular weight) of a cytosolic protein phosphatase which was synthe- sized by perfusion with high glucose. To see whether the dif-

Glucose-induced Activation of Fru 6-R.2-kinase:Fru 2,6-Pase

T T

filtrate kc)

filtrate [g)

MnClz MgCl,

+ c + + + + + + t

+ + + +

+ +

tein phosphatases activated by glucose. The protein phosphatase FIG. 4. Effects of small filtrate molecules on the cytosolic pro-

activities were assayed using the phosphorylated rat testis bifunctional

tracts obtained after 10 min of perfusion with either normal glucose (5.6 enzyme as a substrate in cytosolic fraction of the fasted rat liver ex-

m ~ ; control) or high glucose (40 mM; glucose). Extract, the cytosolic fraction of the liver extract; desalted extract, the cytosolic fraction of the liver extract after filtration with a Sephadex G-50 spin column; Filtrate (c), the Centricon-3 filtrate of the extract of the liver perfused with

of the liver perfused with high glucose; MnCl, and MgCl,, assays were normal glucose; Filtrate (g), the Centricon-3 filtrate of the liver extract

carried out in the presence of 1 mM MnCl, and MgCl,, respectively.

ference in the protein phosphatase activities between the control and the glucose extracts was caused only by this cytosolic activator, the filtrate of the high glucose liver extract was added to the desalted control liver extract (crude protein phosphatases). As shown in Fig. 4, the protein phosphatase activity in the control extract was activated more than the value before desalting but still lower than that in the high glucose liver extract before desalting. These results indicated that the livers perfused with high concentration of glucose appeared to con- tain two factors, a small M, activator and also an altered protein phosphatase which itself was activated by high glucose.

Isolation of the Cytosolic Activator of Fru 6-82"kinase:Fru 2,6-Pase Dephosphorylating Enzyme-The low molecular weight activator was acid-stable (in 1 N HC1 for 15 min at room temperature) and was not adsorbed to charcoal. However, it was adsorbed on Dowex 1-HCO, and could be eluted with 0.15 M NH,HCO,, where most sugar monophosphates were eluted. We have developed the following purification procedure for the isolation of the activator. Livers (1 g) perfused with high glucose medium were homogenized in 1 volume of cold 0.3 N perchloric acid containing 1 m~ EDTA. The homogenate was quickly centrifuged at 9000 x g, and perchloric acid in the supernatant solution was removed with addition of the same volume of tri-n-octylamine/CH,CI (1:3.6 v/v) mixture according to the method of Khym (28). An equal volume of acid-washed charcoal (40 mg/ml) was added, and the charcoal was removed by centrifugation followed by filtration. The charcoal-treated extract was adsorbed on a CarboPac PA1 column and eluted with 50 mM NH,HCO,, pH 8.3, at 1 mumin a t 25 "C using a Dionex HPLC (Dionex, CAI. The activator was usually eluted after 9.3 min and was lyophilized. The lyophilized activator was dissolved in H,O, adsorbed on the CarboPac PA1 column, and eluted with 50 m~ CH,COONH,, pH 5.5, at 1 mumin at 25 "C. The activator was eluted after 7.1 min and lyophilized.

Characterization of the Activator-The results obtained SO far suggested that the activator was not a nucleotide and was likely a sugar monophosphate. When the purified activator was treated with alkaline phosphatase, its ability to activate the

\ Activator (0)

Xu5P (0)

I 1 0 20 30 40 Incubation Time (min)

FIG. 5. Comparison of acid hydrolysis rate of authentic Xu 5-P and the protein phosphatase activator. The reaction mixture contained in 0.2 ml of 1 N HCl, authentic Xu 5-P (20 nmol) or the activator (20 nmol) which was eluate of a Carbopac column as described under "Results," and the mixture was incubated at 100 "C. At indicated intervals an aliquot (30 111) was removed, neutralized to pH 7.0, and the

phatase activator as described under "Experimental Procedures." Data neutralized sample (10 111) was assayed for the residual protein phos-

are representative of results from three experiments.

protein phosphatase was completely lost. When the alkaline phosphatase-treated sample was chromatographed on the Car- boPac PA1 column and eluted with 50 mM NaOH at l mumin at 25 "C using HPLC, a major peak of sugar appeared at 12.8 min which corresponded to xylulose. All other sugars examined, including glucose, galactose, mannose, fructose, ribose, and ribulose were eluted before 10 min. More definitive identification of the activator was made by HPLC chromatography on a Car- boPac PA1 column in 50 m~ CH,COONH,, pH 5.5. The activator was eluted at 7.1 min from this column which was similar to the elution time of authentic xylulose 5-phosphate. None of the other sugar monophosphate examined, including sedohep- tulose 7-P, erythrose 4-P, Glc 6-P, Fru 6-P, Rib 5-P, ribulose 5-P, was eluted from the column at the same time.

The acid hydrolysis rates in 1 N HCl at 100 "C of authentic xylulose 5-phosphate and the isolated activator were also identical (Fig. 5) . The half-lives of the isolated activator and xylulose 5-phosphate under these conditions were approximately 12 min. Finally, incubation of the activator with transketolase in the presence of excess erythrose 4-P resulted in complete loss of its ability to activate the protein phosphatase (data not shown). All these results indicated that the activator in the liver perfused with high glucose medium was xylulose 5-phosphate.

Activation of a Protein Phosphatase for Fru 6-EZ-kinase:Fru 2,6-Pase by Xylulose 5-Phosphate-The activation of the protein phosphatase in the desalted extract of high glucose livers by increasing xylulose 5-phosphate was determined, and the result showed a highly sigmoidal saturation curve for xylulose 5-P. The marked activation of the protein phosphatase occurred between 10 and 20 p~ of xylulose 5-phosphate (Fig. 6).

Formation of Xylulose 5-P in Perfused Liver-In order to ascertain further that xylulose 5-P is the activator in vivo, its level in the liver perfused for varying lengths of time with 40 mM glucose was determined enzymatically. As shown in Fig. 7, the formation of xylulose 5-P began after 2 min of perfusion and proceeded linearly thereafter for 10 min. It should be noted that the rate of xylulose 5-P formation was much slower than

Glucose-induced Activation of Fru 6-RB-kinase:Fru 2,6-Pase 26105

XU 5-P (pM)

FIG. 6. Activation of the cytosolic protein phosphatase activity by increasing Xu 6-P concentration. The cytosolic protein phosphatase activity of the desalted extract was determined in the presence of 0-100 1.1~ Xu 5-P. The liver extract used in this experiment was the cytosolic fraction of liver perfused with high glucose (40 mM) for 10 min. Xu 5-P-dependent protein phosphatase activity was determined after subtracting the activity in the absence of Xu 5-P. The control activity was 0.38 L 0.05 milliunitdg liver, which is comparable to the value in Fig. 4. Values are mean S.E.

r

0 2 4 6 8 10 Perfusion Time (min)

FIG. 7. Changes of Xu 5-P level with time during high glucose perfusion (40 111~) in fasted rat livers. The perfused livers were removed and freeze-clamped immediately. Xu 5-P in perchlorate extracts of the livers was determined enzymatically as described under “Experimental Procedures.” Values are mean L S.E. of three livers for each time point.

the rates of Glc 6-P and Fru 6-P (Fig. 1) but slightly ahead of Fru 2,6-P, formation. The xylulose 5-P concentration did not rise to 10 nmoVg liver, the minimum concentration required for the activation of the protein phosphatase, until 4 min, the time at which the dephosphorylation and the activation of Fru 6-P,2- kinase:Fru 2,6-Pase (Fig. 2) and Fru 2,6-P, concentration began to increase (Fig. l).

DISCUSSION Since the discovery of Fru 2,6-P,, glucose has been known to

increase the Fru 2,6-P, level in hepatocytes and other cells which then activates phosphofructokinase and glycolysis. The mechanism for this effect of glucose, however, is not understood. It is suggested that increased Fru 6-P levels may explain the increased Fru 2,6-P, since Fru 6-P is the substrate for Fru

B-P,a-kinase and the inhibitor of Fru 2,6-Pase (8). We presented evidence here that the increased Fru 2,6-P2 level in 24-h starved rat liver perfused with glucose was not dependent on the increased Fru 6-P. The rates of the formation of Fru 2,6-P, lagged approximately 4 min behind the rise in the hexose-P (Fig. 1). If it were due merely to the increased availability of Fru 6-P as the substrate for Fru 6-P,B-kinase, one would expect that the rates of formation of all these sugar-Ps should be similar if not identical. Instead, our results demonstrated con- clusively that the effect of glucose on the increased synthesis of Fru 2,6-P, was by activation of Fru 6-P,2-kinase and inhibition of Fru 2,6-Pase as a result of dephosphorylation of the phosphorylated bifunctional enzyme that existed in the starved rat. In previous studies (29, 30) using hepatocytes neither the enzyme activities nor the phosphorylation state of the bifunctional enzymes was determined. In contrast to these results, Hue et al. (9) reported that bolus administration of glucose to conscious mice causes increased Fru 2,6-P, but neither the increase of Fru 6-P nor the changes in the activities of the bifunctional enzyme occurred. It is possible that these differences are due to differences in the experimental conditions, such as conscious animals versus perfused liver, and that different mechanisms may be operative for the regulation of Fru 2,6-P, levels, depending upon the conditions. However, it is well established that hexose-P increases with glucose administration in isolated rat hepatocytes, perfused rat liver, and anesthetized mice (5, 7,8,29-33). More pertinent to our results are those of Casazza and Veech (34) who showed that hexose-P, as well as xylulose 5-P, increases significantly in rats which are starved and refed with high carbohydrate diet. Thus, the results reported here with perfused livers and those with whole animals appear to be similar, but the conscious animals may be different, the reason for which is unclear.

The discovery of xylulose 5-P as an activator of a specific protein phosphate for Fru g-P,a-kinase:Fru 2,6-Pase provides an explanation for the lag period observed in the Fru 2,6-P, formation. Xylulose 5-P formation was delayed for 2-3 min compared to the Fru 6-P formation, but slightly ahead of Fru 2,6-P, synthesis. However, the activation of the protein phosphatase requires at least 10 p~ xylulose 5-P, and its activation curve was highly sigmoidal. Apparently there is a threshold concentration (10 p ~ ) of the pentose-P necessary for activation of the protein phosphatase, which took at least 4 min. Further- more, xululose 5-P was a specific activator of the protein phosphatase because none of the other sugar-Ps, including Glc 6-P or Fru 6-P, was effective.

Our results also showed that xylulose 5-P was not the only factor activating the dephosphorylation of the bifunctional enzyme, but it appears that the protein phosphatase itself was also activated by glucose. The evidence supporting this premise is that addition of the filtrate of high glucose liver to the desalted control liver extract did not activate the protein phosphatase to the same level as found in the high glucose liver extract itself (Fig. 4). Thus two factors, namely xylulose 5-P and the altered protein phosphatase, possibly involving a covalent modification, are necessary to account fully for the effect of glucose. Moreover, both factors appear to act in concert for activation of the dephosphorylation reaction since the rates of the formation of xylulose 5-P and Fru 2,6-P2 were similar. If these factors were acting independently, one would expect that these rates might be different. One possible explanation for this concerted activation is that xylulose 5-P directly activates the protein phosphatase, and at the same time the pentose-P binds to the bifunctional enzyme which is the better substrate than free enzyme for the protein phosphatase. The activation of protein phosphatases by pentose-P has never been reported, but

26106 Glucose-induced Activation of Fru 6-P,2-kinase:Fru 2,6-Pase

there is precedence for an improved substrate for protein phosphatase. As discussed in the Introduction, phosphorylase phosphatase prefers glucose-bound phosphorylase a over free phosphorylase a as the substrate (reviewed in Ref. l), thus explaining glucose-stimulated conversion of phosphorylase a to phosphorylase b. However, in contrast to phosphorylase phosphatase, glucose did not activate directly the xylulose 5-P- dependent PPase for the bifunctional enzyme.

Glucose-dependent activation of Fru 2,6-P2 synthesis has been observed also in yeast (351, islet cells (361, and macroph- ages (37). It is interesting that in yeast the effect of glucose is mediated via CAMP which activates protein kinase. The CAMP- dependent protein kinase catalyzes the phosphorylation and the activation of Fru 6-P,2-kinase resulting in increased Fru 2,6-P, in yeast (35), opposite to the effect in liver. In the islet cells glucose raises the Fru 2,6-P, level, but the activities of the bifunctional enzyme are not altered, suggesting that covalent modification is not involved in these cells (36). These results may further suggest that the isozyme of the bifunctional enzyme in the islet cells may not be the liver type. More recently, Ogawa et al. (38) reported that rat liver perfused with a low concentration of fructose also increases Fru 2,6-P2, and this increase is attributable to the activation of Fru 6-P,2-kinase, since the activity ratio measured at two different concentrations of Fru 6-P changed. Although the phosphorylation state of the enzyme was not determined, the dephosphorylation of the bifunctional enzyme probably occurred as a result of fructose administration, and this could also be a result of the activation of the xylulose 5-P-dependent protein phosphatase due to increased Fru 6-P.

Xylulose 5-P is one of the intermediates in the pentose shunt pathway. The roles of this pathway are to provide reducing equivalents in the form of NADPH for fatty acid synthesis and ribose 5-P for nucleotide synthesis. The final products of the pentose cycle are Fru 6-P and glyceraldehyde 3-P which are also intermediates of glycolysis. Casazza and Veech (34) determined the concentrations of all the intermediates of the pentose cycle and concluded that all the reactions in the nonoxidative part of the pathway are in equilibrium in ad libitum fed and starved rats, thus glycolysis and pentose pathway are interde- pendent. However, xylulose 5-P (and ribulose 5-P) in rats fed high carbohydrate diet after starvation increased from 3.9 nmoVg liver to 77 nmoVg liver, more than an order of magni- tude higher than the estimated value based on the equilibrium. The mechanism for the large increase in those pentose-Ps is not known. Under the same conditions Fru 6-P also rose from 16.4 to 55.7 nmoVg (34). Thus, it is clear that both xylulose 5-P and Fru 6-P levels are increased by glucose administration in intact animals as well as in the perfused livers. It is tempting to propose that the xylulose 5-P-mediated increased synthesis of Fru 2,6-P, may provide a mechanism for a coordinated regulation of glycolysis and the pentose shunt pathway in liver under high glucose intake. Thus, xylulose 5-P may serve as a "feed- forward activator of phosphofructokinase by stimulating Fru 2,6-P2 synthesis and increase glycolysis in order to metabolize Fru 6-P generated by both pathways.

In summary, we have shown that the glucose-induced increase in the Fru 2,6-P, level in the perfused livers of starved rats was as a result of dephosphorylation of Fru 6-P,2-kinase: Fru 2,6-Pase which activated the kinase and inhibited the phosphatase. This dephosphorylation reaction was stimulated in part by activation of a protein phosphatase by increased levels of xylulose 5-P. This activation of protein phosphatase activity required a threshold level (10 PM) of xylulose 5-P. Direct activation of the protein phosphatase seems to be also required to fully account for the maximum activation of the dephosphorylation of the bifunctional enzyme induced by high glucose.

Acknowledgments-We thank Drs. Sarah McIntire and Paul A. Srere for critical review of this manuscript.

1. 2. 3.

4.

5.

6. 7.

8.

10. 9.

11.

12.

13.

14.

15. 16.

17.

18. 19.

20. 21. 22.

23.

25. 24.

26. 27. 28. 29. 30. 31.

32, 33

34 35 36 37 38

REFERENCES Stalmans, W. (1976) Curr: Top. Cell Regul. 11, 51-97 Hers, H. G. (1976) Annu. Rev. Biochem. 46, 167-189 Roach, P. J. (1986) in The Enzymes (Boyer, P. D., and Krebs, E. G., eds) Vol. 17,

Stalmans, W., Bollen, M., and Mvumbi, L. (1987) Diabetes Metab. Reu. 3,

Richards, C. S., and Uyeda, K (1980) Biochem. Biophys. Res. Commun. 109,

van Schaftingen, E., Hue, L., and Hers, H. G. (1980)Biochem. J. 192,887-895 Hue, L., Blackmore, P. F., and Exton, J. H. (1981) J. Bid. Chem. 256, 8900-

Neely, P., El-Maghrabi, R., Pilkis, S. J., and Claus, T. H. (1981) Diabetes 30,

Hue, L., Sobrino, F., and Bosca, L. (1984) Biochem. J. 224, 779-786

Fillat, C., Gomez-Faix, A. M., and Guinovart, J. J. (1993) Arch. Biochem. Hue, L., and Rider, M. H. (1987) Biochem. J. 245,313-324

Furuya, E., Yokoyama, M., and Uyeda, K. (1982) Proc. Natl. Acad. Sci. U. S. A.

El-Maghrabi, M. R., Claus, T. H., Pilkis, J., and Pilkis, S. J. (1982) Proc. Natl.

van Schaftingen, E., and Hers, H. G. (1981) Biophys. Biochem. Res. Commun.

pp. 499-539, Academic Press, Orlando

127-161

394-401

8903

1062-1064

Biophys. 300,564-569

79,325-329

Acad. Sci. U. S. A. 79,315-319 .~

Sakakibara, R., Kitajima S., and Uyeda, K. (1984) J. Bid. Chem. 269,41-46 Thomas, H., Taniyama, M., Kitajima, S., and Uyeda, K. (1986) Fed. Proc. 45,

103,362-368

Uyeda, K., Miyatake, A,, Luby, L. J., and Richards, E. G. (1978) J. Biol. Chem.

Abe, Y., and Uyeda, K. (1994) Biochemistry, in press Tominaga, N., Minami, Y., Sakakibara, R., and Uyeda, K. (1993) J. Biol. Chem.

Thomas, H., and Uyeda, K. (1986)Anal. Biochem. 154,50-56 Casazza, J. P., and Veech, R. L. (1986) J. Biol. Chem. 261,690-698 Lowry, 0. H., and Passonneau, J. V. (1972) A Flexible System of Enzymatic

Kitajima, S., Sakakibara, R., and Uyeda, K. (1984) J. Biol. Chem. 269,6896-

Bradford, M. M. (1976) Anal. Biochem. 72,24%254 Laemmli, U. K. (1970) Nature 227,680-685

Corbin, J. D., and Reimann, E. M. (1974) Methods Enzymol. 38, 287-290 Ingebritsen, T. S., and Cohen, P. (1983) Science 221,331-338 Khym, J. X. (1975) Clin. Chem. 21,1245-1252 Bontemps, F., Hue, L., and Hers, H. G. (1978) Biochem. J. 174,603-611 Hue, L. (1981)Adu. Enzymol. 62,247-331 Buschiazzo, H., Exton, J. H., and Park, C. R. (1970) Proc. Natl. Acad. Sci.

Katz, J., Wals, P. A,, and Rognstad, R. (1978) J. Biol. Chem. 263,4530-4536 Nordlie, R. C., Sukalski, K. A., and Alvares, F. L. (1980) J. Bid. Chem. 266,

Casazza, J. P., and Veech, R. L. (1986) Biochem. J. 238,635-641 Francois, J., Eraso, F!, and Gancedo, C. (1987) E m J. Biochem. 164,369-373 Malaisse, W. J., Malaisse-Lagae, F., and Sener, A. (1982) Diabetes 31, 90-93 Bustos, R., and Sobrino, F. (1989) FEBS Lett. 261,143-146 Ogawa, A., Nishi, T., Furuya, E., Watanabe, F., Sakai, A,, Kido, Y., Tsujinaka,

T., and Mori, T. (1993) Biochem. Mol. Biol. Int. 30,83-90

1013

263,8319-8327

288, 15951-15957

Analysis, pp. 146-218, Academic Press, New York

6903

U. S. A. 66,383387

1834-1838

the journal of biowcical chemistry vol. no. issue of ... · the journal of biowcical ......

Documents