ARCHIVES

OF BIOCHEMISTRY

Vol. 283, No. 1, November

AND

BIOPHYSICS

15, pp. 51-59,199O

The Role of Inhibition of Pyruvate Kinase in the Stimulation of Gluconeogenesis by Glucagon: A Reevaluation’ Robert

C. Haynes,

Department

Jr.2 and Ralph

of Pharmacology,

University

A. Picking of Virginia,

Charlottesville,

Virginia

22908

Received May 9,199O

1. We tested and refined the 14C-tracer technique that has previously yielded the opposite conclusion, that is, that inhibition of pyruvate kinase is a relatively unimportant mechanism. The tracer procedure, as used by us, was found to be insensitive to the size of the pyruvate pool, and experiments using modifications of the technique to obviate a number of other potential errors support the earlier conclusion that control of pyruvate kinase is not the predominant mechanism. 2. Any stimulation of formation of glucose that results from inhibition of pyruvate kinase is the consequence of elevation of the steady-state concentrations of phosphoenolpyruvate and all subsequent intermediates in the gluconeogenic pathway. During ongoing stimulation of glucose synthesis by glucagon in isolated hepatocytes, the concentrations of all measured intermediate compounds between phosphoenolpyruvate and glucose were elevated except triose phosphates and fructose 1,6-bisphosphate. The failure of these compounds to rise above control levels indicates that not all gluconeogenic reactions beyond pyruvate kinase were accelerated thermodynamically as would occur with predominant control at pyruvate kinase. We conclude, therefore, that although glucagon inhibits flux through the pyruvate kinase reaction, this does not account for most of the stimulation of gluconeogenesis. Major control sites are also within the pyruvate-phosphoenolpyruvate segment and the fructose 1,6-bisphosphate 0 1990 Academic Press, Inc. cycle.

The elegant studies of Groen et al. (1) using perifused hepatocytes indicated that glucagon stimulates gluconeogenesis from lactate-pyruvate almost entirely by inhibition of pyruvate kinase. We were led to the same conclusion from similar experiments adapted to conventional incubation techniques (2). In the model of hormonal control of gluconeogenesis secondary to inhibition of pyruvate kinase, there is no change in flux at the phosphoenolpyruvate carboxykinase reaction (JPEPCK).3 Decreased utilization of phosphoenolpyruvate by the pyruvate kinase reaction accounts for the well-established rise in phosphoenolpyruvate concentration during stimulation by glucagon. The increased phosphoenolpyruvate available to enolase then accelerates formation of 2-phosphoglyceric acid followed by similar thermodynamic stimulations of all subsequent enzymatic reactions in the pathway leading to glucose. In contrast to this model, the study of Rognstad and Katz (3), in which a tracer technique was used to assay flux through pyruvate kinase (JPK), indicated JpK was surprisingly small in hepatocytes utilizing lactate for gluconeogenesis, and the decrement in JPK produced by glucagon was insufficient to account for the additional glucose formed. These findings supported a model in which JpEpCKis increased by the action of glucagon, and changes in JPK are relatively unimportant. We have also found in preliminary experiments using the technique of Rognstad and Katz that CAMP inhibited JPK, but not sufficiently to account for the increment in glucose formation (4). There is, therefore, a conflict in the literature (including publications from our own laboratory) about the relative importance of the mechanisms by which glucagon

i This work was supported by Grant DK14347 from The National Institutes of Health, USPHS. ’ To whom correspondence should be addressed.

3 Abbreviations used: J rx, rate of flux of the pyruvate kinase reacrate of flux of the phosphoenolpyruvate carboxykinase JPEPCK, reaction; Jglucoss , rate of synthesis of glucose; Mops, 4-morpholinepropanesulfonic acid; L/P, lactate/pyruvate ratio.

We have reexamined the concept that glucagon controls gluconeogenesis from lactate-pyruvate in isolated rat hepatocytes almost entirely by inhibition of flux through pyruvate kinase, thereby making gluconeogenesis more efficient.

0003-9361/90 $3.00 Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.

tion;

51

HAYNES

52

AND PICKING

stimulates gluconeogenesis. In this report we probe the opposing experimental systems and models for flaws that could account for this discrepancy. EXPERIMENTAL

PROCEDURES

Materials. L-[1-i4C]Alanine was obtained from Amersham Corp. Glucagon, 8bromoadenosine 3’:5’-cyclic monophosphate, and phenylephrine were purchased from Sigma. Bromododecane was from Aldrich Chemical Co., Inc. Hepatocyte isolation and incubation. Hepatocytes were isolated essentially as described by Seglen (5). Male Wistar strain rats (250-300 g) were starved overnight and anesthetized with intraperitioneal pentobarbital. Viability of cells was assessed by exclusion of 0.2% trypan blue and ranged between 89 and 97%. Incubations were at 30°C in Erlenmeyer flasks in a rotary shaker bath set at 140 rpm. Rates of glucose synthesis from pyruvate-lactate were linear for at least 2 h. Assays. In some experiments l-ml aliquots of the cell suspension were layered onto 0.7 ml bromododecane (d 1.038) lying over 0.2 ml 12% perchloric acid in a microfuge tube. Cells were centrifuged through the oil into the acid by a force of 12000g for 15 s. Perchloric acid extracts of the cells or of cell suspensions were neutralized with K&O3 in the presence of 1 mM EDTA prior to analyses of metabolites by fluorometric, enzymatic assays (6, 7). Extracts used for assays of triose phosphates and fructose 1,6-bisphosphate were neutralized to pH 6.5-6.9. The concentration of cytosolic oxalacetate was calculated as described previously from the cellular malate content using the lactate dehydrogenase and malate dehydrogenase equilibrium constants, assuming equilibrium for both reactions (2). For estimation of flux through the pyruvate kinase reaction (Jrx) we have used a modification of the procedure that was adapted by Rognstad and Katz (3) from the original technique of Rognstad (8). Hepatocytes were incubated at 0.5 mg protein/ml in 6 ml of KrebsRingers-bicarbonate medium; the temperature was 30°C with a gas phase of 95% 02/5% CO*. Substrates were initially 10 mM lactate/l mM pyruvate. At 45 min, aliquots of the cell suspension were removed for determination of glucose; flasks were sealed, [i4C]NaHC03 (10 to 20 &i) was injected, and control vehicle or agonists were added. Incubations were terminated after an additional 45 min by injection of 0.53 ml of 20% perchloric acid. Carbon dioxide was then trapped in 250 ~1 4 N NaOH contained in a cup suspended within the flask. To insure complete collection, flasks were shaken for 2 h then allowed to sit overnight. Control experiments demonstrated that labeled pyruvate was not lost from the medium under these conditions. The acidified medium was diluted to 20 ml and centrifuged to remove protein. The supernatant was brought to 5 mM Mops and neutralized with 4.8 M KOH. Two units of urease/ml, 0.8 mM NADH, and 10 units of lactic dehydrogenase/ml were added, and the mixture was incubated 30 min at 37°C to reduce pyruvate to lactate and destroy any urea present. The solution was then acidified to 0.08 M perchloric acid and bubbled with air for 15 min to remove any i4C02 released from urea. Following this, 15 ml was taken for ion exchange chromatography on a 1 X 8-cm Dowex50 (H+) column placed above a 1 X 8-cm Dowex-1 (acetate) column. The nonionic fraction was eluted with 40 ml H20, and the first 35 ml, which contained glucose, was collected for measurement of radioactivity. The columns were disconnected; then anions bound to the Dowex1 column were eluted with 30 ml of 2 M acetic acid (lactate fraction) and 25 ml of 3 M HCl (dicarboxylic acids and organic phosphates). Amino acids were eluted from the Dowex-50 with 25 ml of 4 M NHIOH. Aliquots of these fractions were then taken for scintillation counting. Quenching was insignificant in these samples. Uncorrected Jrs was calculated as the product of 2 X rate of glucose synthesis X the quotient of the radioactivity in the lactate fraction and the glucose (nonionic) fractions. Final estimation of Jrx required the application of the partition factor obtained from parallel incubations with L-[1-i4C]alanine

(8) modified as noted trapped as alanine and [l-i4C]pyruvate. Flux (JPEPCK) was calculated thesis.

RESULTS

AND

below by including corrections for pyruvate extracellular conversion of L-[1-“Clalanine to through phosphoenolpyruvate carboxykinase as the sum of Jrx and 2 X rate of glucose syn-

DISCUSSION

Critique of Rognstad-Katz Experiments with Modifications to Obviate Potential Errors The paper of Rognstad and Katz (3) reported the results of three experiments using hepatocytes from fasted rats in which decreases in JPK were found to be too small to account for the stimulation of gluconeogenesis produced by glucagon. The technique for estimating JPK as developed by Rognstad (8) employs the addition of [14C]NaHC03 to hepatocyte incubations in order to label phosphoenolpyruvate via pyruvate carboxylase and the subsequent randomization of the labeled oxalacetate generated by pyruvate carboxylase. The pyruvate kinase reaction then leads to formation of [l-‘4C]pyruvate, which is largely trapped in the substrate pool of unlabeled pyruvate and serves as an index of the rate of pyruvate kinase activity. The specific activity of labeled glucose synthesized is assumed to be twice that of phosphoenolpyruvate, so that knowledge of the rate of glucose synthesis (J,,,,,,,) and the specific activity of glucose can be used to determine JPK . A correction factor for metabolic loss of labeled pyruvate is obtained by labeling the pyruvate pool with [l14C]alanine in parallel incubations and determining the partition of radioactivity among the pyruvate pool, CO,, and glucose. A detailed description of this technique, as we have applied it, is under Assays. In using this method to assess the importance of control of JPK by hormones, the most serious potential errors are those that lead to underestimation of JPK, inasmuch as the primary question asked is whether hormone-produced decrements in JPK are sufficient to account for the increments in gluconeogenesis. Rognstad and Katz acknowledged the probability of underestimation of JPK in their experiments as a result of simplifications in the Rognstad technique (8) that were made to expedite analysis. Because their paper presents a conclusion completely contrary to that resulting from more recent studies (1,2), it seemed to us of great importance that the experiments of Rognstad and Katz be repeated after correcting, as far as possible, technical procedures used by them that may have led to errors of underestimation of JPK. As originally described (8)) the method for estimation of JPK utilized a substrate pool of 20 mM pyruvate to trap the [ 1-14C]pyruvate generated by the pyruvate kinase reaction, whereas Rognstad and Katz used 20 mM lactate added initially in incubations that continued for 30 min. Thus, there was only a small and variable pool of pyruvate for trapping labeled pyruvate, probably reaching a

PYRUVATE

KINASE

AND

CONTROL

OF GLUCONEOGENESIS

TABLE

IN RAT

53

I

The Independence of JPK Values and Substrate Pool Size. The Inhibition of JPK by a High Lactate/Pyruvate (L/P) Ratio in the Presence of an Enlarged Substrate Pool L/P = 10 Pyruvate 0.3

Pyruvate

(mM)

1.0

L/P 2.0

(nmol/min/mg J gl”COSe

= 1 mM

5

20

protein)

JPEPCK J PK

2.34 zk 0.40* 6.70 f l.lO* 2.02 * 0.41

3.31 r!r 0.08 9.47 f 0.46 2.85 + 0.59

3.11 + 0.28, 8.56 -t 0.41 2.35 f 0.35

2.43 310.18* 7.05 f 0.67* 2.20 f 0.35

3.20 + 0.16 8.10 + 0.66 1.69 + 0.33*

JPK /JPEPCK

0.30 f 0.03

0.30 + 0.05

0.28 f 0.05

0.31 f 0.02

0.21 + 0.02’

Note. Experiments to estimate J PK were performed as dekribed under Experimental Procedures except the incubation was in a modified Krebs-Ringers-bicarbonate medium containing 20 mM Hepes, pH 7.4, and 6 mM NaHC03 with NaCl adjusted accordingly. The gas phase was air. The amounts of pyruvate and lactate added initially resulted in the concentrations indicated during the second 45min period (2). Values are given as means f standard errors for three experiments. The significance of difference of mean values from those obtained using 1 mM pyruvate and a L/P of 10 was estimated by Student’s t test, one tailed, pair comparison. Asterisks indicate significance at P = 0.05 or less.

final concentration of 1 to 2 mM. If isotopic equilibration incomplete trapping of labeled pyruvate. Formation of between pyruvate and lactate in the medium in these ex[ l-‘4C]pyruvate was possibly undeiestimated in the exhowever, for several periments was much more rapid than metabolism of the periments of these investigators, other reasons. First, only radioactive lactate was isolated labeled pyruvate, the effective pool size was 20 mM, but this cannot be assumed to be so. and counted, leading to an approximately 10% underestimation if the lactate/pyruvate ratio was 10 and isotoAlthough the use of parallel incubations with L-[lpic equilibration between lactate and pyruvate was com14C]alanine can, in theory, correct for losses of labeled pyruvate, we judged it essential to test the effect of the plete. Neither of these conditions can be assumed, so the error may have been greater or less. As indicated under pyruvate pool size directly. For these experiments, the of both medium hepatocyte concentration was kept quite small, 0.5 mg Assays, we measured the radioactivity cell protein per milliliter, so that substrate pools could pyruvate and lactate by reducing pyruvate to lactate bebe maintained essentially constant at desired levels (2). fore isolating the lactate for counting. In addition, Rognstad and Katz did not take into acThe results of Table I, left side, show that when the laccount the partial conversion of [1-14C]pyruvate to [ltate/pyruvate ratio (L/P) was maintained at 10, changes in pool size of nearly sevenfold had no effect on the ap- 14C]alanine in calculating JPK. Smith and Freedland (9) parent JPKwhen expressed relative to flux through phosdemonstrated that 56% of the radioactivity of the amino phoenolpyruvate carboxykinase (JPEPCK).The correcacid fraction from incubations of hepatocytes with [‘“Cltion factors derived from parallel experiments were 1.32 NaHC03 is in alanine. Since this radioactive alanine is + .06,1.2 + 0.06, and 1.15 f .05 for the experiments using presumed to arise entirely through transamination of 0.3, 1.0, and 2 mM pyruvate, respectively, with the L/P the [l-‘4C]pyruvate that is formed by the pyruvate kiequal to 10. The constant JPK/JPEPCKratio observed as nase reaction, it is reasonable to include this fraction of the lactate-pyruvate substrate pool was varied provides the radioactivity of the amino acid fraction with the pyassurance that trapping of labeled pyruvate was not sig- ruvate pool in calculating JPK. We have found inclusion nificantly dependent on pool size. The modest decline in of the radioactive alanine increases the radioactive “pycorrection factor when the lactate pool was largest does ruvate” pool between 15 and 140% thus affecting the esindicate a small effect, of pool size on trapping efficiency; timate of JpK significantly in some cases. nevertheless, the data demonstrate this was fully comRognstad and Katz elected not. to use correction facpensated for by the correction factor. It is to be noted in tors derived from parallel incubations with L-[1-14C]alathese experiments that decreasing pyruvate to 0.3 mM, a nine (8) and accepted the consequent underestimation concentration less than maximally effective as sub- of JpK. For the most accurate measurements this adjuststrate, led to decreases in mean values of JpEpCK,Jglucose ment should be included. We found, in addition, that the and JPK , but the rate of JPK relative to the flow of 3 car- incubation medium from which hepatocytes had been bon units into glucose remained approximately 3 to 7. removed was capable of converting L-[l-14C]alanine to These data indicate that the experiments of Rognstad labeled pyruvate, presumably through isotopic exchange and Katz probably did not underestimate JPK because of catalyzed on pyruvate aminotransferase free in the me-

54

HAYNES

AND

dium. This conversion was extensive; approximately 40% of the total label in pyruvate could be accounted for by this extracellular conversion. Since this gives an erroneously large pool of radioactive pyruvate-lactate, which is the denominator of the correction factor, the factor is too small. In our experiments, we have used only 0.5 mg cell protein per milliliter of incubation medium in contrast to the Rognstad-Katz study in which approximately 20-30 mg protein per milliliter was apparently used. In calculating the correction factor, the 40% decrement in the radioactivity of the lactate-pyruvate pool that we applied might be even larger with more concentrated cell suspensions. This modification of the correction factor increases the estimated values of JPK and therefore provides an increased stringency in testing the significance of any decreases in JPK mediated by glucagon. Underestimation of JPK can also result from radioactive contamination of the glucose (nonionic) fraction. Contamination of this fraction with radioactive urea is not unlikely, given that hepatocytes incubated with [14C]NaHC03 synthesize variable amounts of [14C]urea and that urea is an extremely weak acid (pk, = 0.1). In preliminary studies we found that [14C]urea accounted for as much as 0.6 of the radioactivity in the nonionic fraction; therefore, as noted under Assays, we have routinely removed urea from extracts before fractionation. In the resulting nonionic fractions, 98-99.5 percent of the radioactivity was found to be in glucose as indicated by its conversion to an anionic form with hexokinase and ATP. When lactate alone is the initial substrate for incubations of hepatocytes, as was true in the Rognstad-Katz study, the ratio of lactate to pyruvate remains high for a considerable period of time (10). The resulting elevated NADH/NAD+ ratio of the hepatocyte cytosol can inhibit JPK by lowering the steady-state concentration of phosphoenolpyruvate, and any fractional decrease in JPK caused by glucagon will appear less significant in controlling J,l,,,, . The decline in phosphoenolpyruvate is a consequence of the shift in equilibrium of the glyceraldehyde-3-phosphate dehydrogenase reaction toward glyceraldehyde-3-phosphate that results from the increased NADH/NAD+ ratio. This leads to similar forward shifts of preceding near-equilibrium reactions resulting in lower steady-state concentrations of intermediates back to, and including, phosphoenolpyruvate, providing the rate of formation of phosphoenolpyruvate is not altered. A high lactate/pyruvate ratio therefore depresses pyruvate kinase activity and results obtained under such conditions may not apply to more physiologic states. This effect of a high lactate/pyruvate ratio (20 to 1) is illustrated in the right hand column of Table I where it can be noted that the JPK/JPEPCK ratio is diminished relative to that occurring in cells incubated with a lactate/pyruvate ratio of 10 to 1. We used low con-

PICKING

centrations of cells that permitted nearly constant L/P ratios of about 10 and that limited substrate utilization to 10% or less during the incubation. The use of very small amounts of tissue necessitated longer incubation periods than in the original procedure (8). The use of these extended incubations is justified by the fact that the rates of glucose synthesis and JPK were linear during the 45-min period of incubation used in our experiments. Similarly the partition of 14C from L-[1-14C]alanine into COZ and glucose was constant over the same length of time (data not shown). One highly improbable but undetectable error in the Rognstad procedure would result if there is a preferential channeling of pyruvate derived from the pyruvate kinase reaction in the cytosol of the hepatocyte to oxidation by pyruvic dehydrogenase within mitochondria. Should such a channeling occur, and if it did not take place also with pyruvate derived from alanine, it would not be corrected for by the use of labeled alanine; consequently JPK would be underestimated. It should be noted that the control rates of JPK we have observed here (0.1 to 0.5 of the flux of phosphoenolpyruvate to glucose) correspond to the range of values reported by other workers who used different experimental systems and analytic techniques to measure JPK (11-14). When considering the role of pyruvate kinase activity in gluconeogenesis, attention should also be directed to the poor correlation between JPK and rates of glucose synthesis in control cells (Table II). Using the modifications of the Rognstad technique discussed above so as to avoid underestimation of JPK, we carried out a number of hepatocyte incubations that confirmed the general conclusion of Rognstad and Katz, that is, that inhibition of JPK is not the major mechanism by which glucagon stimulates gluconeogenesis from pyruvate-lactate. One pair of experiments in which the effects of glucagon on gluconeogenesis and JPK were assessed using cells incubated in three different media are presented in Table II. It is evident, in spite of differences of control values for JPK and different responses of JPK to glucagon with the three media, that inhibition of JPK accounted for, at most, slightly over half of the effect of glucagon and, in two media, less than one fourth. These estimates were obtained using correction factors of 1.1 to 1.3 derived from parallel incubations with L-[1-14C]alanine. Critique of Studies Using Substrate- Velocity Curves to Identify and Quuntitate Hormonal Control Mechanisms The experimental technique developed by Groen et-al. (1) involved determination of the rate of glucose synthesis by hepatocytes exposed to varying concentrations of lactate-pyruvate. In addition, the “steady-state” cellular concentrations of intermediate compounds of gluco-

PYRUVATE

KINASE

AND

CONTROL TABLE

The Influence

of Incubation

Media

on the Role of Inhibition

OF GLUCONEOGENESIS

IN RAT

55

II of JPK in Stimulation

of Gluconeogenesis

by Glucagon

Hormone effect due to inhibition Medium KRB

KRBH

KRBHK+

Treatment Control Glucagon ( 10d7 M) Difference Control Glucagon ( 10e7 M) Difference Control Glucagon (10-r M) Difference

J Gl”COSe (nmol/min/mg protein)

of JPK

JPK

(nmol/min/mg

2.28 3.32 1.04 2.72 4.34 1.62 2.30 3.43 1.13

2.16 1.52 0.64 2.76 0.98 1.78 1.14 0.86 0.28

protein)

(percent)

31

55

12

&EPCK

(nmol/min/mg/protein) 6.72 8.16 +1.44 8.20 9.66 +1.46 5.74 7.72 +1.98

Note. Estimations of JPK were carried out as described under Experimental Procedures except the incubation media were: (i) Krebs-Ringersbicarbonate (KRB). (ii) a modified Krebs-Ringers-bicarbonate medium that contained 20 mM Hepes, pH 7.4, and 6 mM NaHCO, (KRBH), and (iii) a modified Krebs-Ringers-bicarbonate medium that contained 20 mM Hepes, pH 7.4, 6 mM KHCOa, 5 mM Na+, and 141 mM K+ (KRBHK+). Isotonicity was maintained in the modified media by appropriate adjustment of NaCl or KC1 content. For the incubations using Krebs-Ringers-bicarbonate medium, the gas phase was 95% Oz, 5% CO,; when the modified media were used, the gas phase was air. Acid extracts derived from cell incubations in the high K+ medium were hydrolyzed at 95°C before determination of glucose and ion exchange chromatography because some glycogen was formed by hepatocytes incubated in this medium. Data are means from two closely agreeing experiments.

neogenesis were measured, and from the resulting data a number of curves relating these concentrations to Jglucose were generated. These substrate-velocity curves were then analyzed to localize and quantitate control mechanisms. These experiments leave unresolved two major questions. The first, acknowledged by the authors (1, 2), is the inability of this approach to differentiate between hormonal stimulation of the phosphoenolpyruvate carboxykinase reaction and an inhibition of pyruvate kinase. The second is whether there is hormonal control of gluconeogenesis at the fructose 1,6-bisphosphate cycle in cells from fasted rats. In the absence of convincing evidence of acute hormonal regulation of phosphoenolpyruvate carboxykinase activity and in the face of the firmly established control of pyruvate kinase by CAMP-dependent phosphorylation, Groen et al. (l), and Sistare and Haynes (2), attributed the evident control within the oxalacetatephosphoenolpyruvate segment to the latter mechanism. The results of Rognstad and Katz (3) confirmed and strengthened by the data of this paper, indicated that glucagon stimulates JpEpCK in hepatocytes synthesizing glucose from lactate-pyruvate. The same conclusion was reached by Pate1 and Olson (15) using a different approach. They showed that decarboxylation of [1-14C]lactate can be used as an index of JpEpCK in hepatocytes. Neither glucagon nor phenylephrine activated pyruvate dehydrogenase in their system while both agents stimulated decarboxylation of [1-14C]lactate. From this, and the fact that 3-mercaptopicolinate decreased both [14C]C02 production and glucose synthesis, they were

able to conclude that glucagon and phenylephrine stimulated JpEpCK. This conclusion obtained by two different techniques conflicts with the model ascribing near complete control to inhibition of pyruvate kinase, which involves no increase in JpEpCK. In view of the demonstration of increased JPEPCKwith hormone treatment, and the inability of the curve-subtraction technique to differentiate control at phosphoenolpyruvate carboxykinase from control at pyruvate kinase, it is appropriate to conclude that a significant portion of control at the pyruvate-phosphoenolpyruvate segment of gluconeogenesis is not due to inhibition of pyruvate kinase, but is caused by increased JpEpCK. A stimulation of JPEPCKby glucagon could be either thermodynamic, by an increase in cytosolic oxalacetate, or kinetic, by an activation of phosphoenolpyruvate carboxykinase or by both mechanisms. Experiments in which the effect of glucagon on oxalacetate concentration was examined have been contradictory. Some investigators have reported increased oxalacetate (2, 16), while others found no change (17) or a decrease (18) after glucagon treatment. The data of Table III show that under the conditions of our experiments there was no effect of glucagon on the steady-state levels of oxalacetate. There was neither an increase that would be anticipated if the reaction were stimulated thermodynamically, nor a decrease that could result from activation of the enzyme. The lack of change in oxalacetate with glucagon was not an analytical artifact as evidenced by the rise in oxalacetate in cells treated with phenylephrine in the same experiments. Phenylephrine, therefore, appears to drive phosphoenolpyruvate carboxykinase

56

HAYNES

AND TABLE

PICKING III

Effects of Glucagon and Phenylephrine and Their Interaction on Glucose Synthesis, Substrate Concentrations, and Concentrations of Intracellular Compounds Associated with Gluconeogenesis

Treatment Control

Glucagon, 0.1 pM Phenylephrine, 25 pM Phenylephrine, 25 pM ghcagOn, 0.1 pM

Triose

Malate

JGlumk

Fructose-l,& bisphosphate

Oxalacetate (MM)

phosphates

L/P

(llIld/ mg protein)

2.3 + 0.1 2.1 * 0.1 1.8 f 0.1’

6.9 + 0.3 7.4 * 0.3 8.6 + 0.6*

1.17 f 0.07 1.22 k 0.09 2.09 f 0.25*

15.6 k 1.0 15.2 Itr 1.3 22.1 + 2.0*

8754 87 + 5

101+ 4*

14 + 1 14+ 1 22 k 1+

1.4 * 0.1*

12.7 k 0.7*

2.70 + 0.24*

19.6 k 1.5;

93 f 4

13k 1

(nmol/min/ mg protein)

Lactate bM)

Pymvate bM)

2.65 f 0.08 3.76 + O.ll* 4.54 + 0.10*

15.5 f 0.3 15.7 + 1.0 15.1 + 0.7

3.61 lr 0.18*

16.5 + 0.7

(pmol/w)

(PmQw)

Note. Hepatocytes, 10 mg cell protein/ml, were incubated in Krebs-Ringers-bicarbonate with lactate initially 20 mM. Gas phase was 95% 02, 5% CO*. After 60 min, agonists or control vehicle was added, and the incubation was then continued for an additional 60 min. At 80 and 100 min, cells were separated from the medium and extracted as described under Experimental Procedures. The data presented in the table, except rates of glucose synthesis, are the means f SE of the 80- and lOO-min assays which were pooled in each experiment. There were five replications of this experiment. Asterisks indicate significant differences from controls of P = 0.05 or less, Dunnett’s t test, one tail.

thermodynamically. It is possible that the concentration of oxalacetate was not decreased after treatment with glucagon even if phosphoenolpyruvate carboxykinase were activated because oxalacetate concentration may be closely regulated by glutamate-aspartate amino transferase and malic dehydrogenase activity. In particular, a steady NADH/NAD+ ratio and the control of malate concentration by multiple reactions may stabilize the concentration of oxalacetate faster than it is metabolized. To test for possible regulation by glucagon of late reactions in glucose synthesis, especially at the fructose 1,6bisphosphate cycle, Groen et al. (1) examined substratevelocity curves relating J,l,,,, to cellular content of phosphoenolpyruvate, 3-phosphoglycerate, dihydroxyacetone phosphate, and fructose 1,6-bisphosphate; similarly, Sistare and Haynes (2) determined the substratevelocity curve for 3-phosphoglycerate-J,I,,,, . In each of these instances, values from control and glucagon-stimulated cells fell along single curves. This information, while not presented as such by these workers, could have been expressed as finding significantly higher concentrations of the metabolites in glucagon-treated cells relative to control cells for each lactate-pyruvate concentration used. This alternative conceptualization is intuitively easier to relate to stimulation of gluconeogenesis secondary to inhibition of pyruvate kinase and subsequent thermodynamic acceleration of all succeeding reactions. These data were interpreted by Groen et al. (1) as indicating glucagon had no “effect on the kinetics of fructose 1,6-bisphosphate conversion to glucose,” and by Sistare and Haynes (2) as indicating glucagon did not act on the fructose 1,6-bisphosphate cycle or on the glyceraldehyde-3-phosphate dehydrogenase reaction. When these data are recognized as demonstrating increased concentrations of intermediate compounds of gluconeogenesis following glucagon treatment, they can be directly compared with the observations of others. Investi-

gators who studied the effects of glucagon in perfused rat livers (16, 18-20) found that glucagon elevated cellular phosphoenolpyruvate and 3-phosphoglycerate in agreement with Groen et al. and Sistare and Haynes (1, 2). In contrast, however, to the results of Groen et al. (l), experiments from all four laboratories demonstrated that stimulation of glucose synthesis from lactate-pyruvate by glucagon was accompanied by either unchanged (18, 19) or decreased (16, 20) concentrations of triose phosphates and fructose 1,6-bisphosphate. These observations, obviously, do not support the concept of thermodynamic stimulation of all reactions downstream from pyruvate kinase that would result from predominant control by inhibition of that enzyme. If the concentrations of triose phosphates and fructose 1,6-bisphosphate are not elevated by glucagon, the increased flux through the fructose 1,6-bisphosphate cycle must be the result of kinetic changes in one or both enzymes of that cycle. Since fructose g-phosphate is increased by glucagon, this stimulation does not result from decreased product inhibition. With unchanged or depressed steady-state levels of fructose 1,6-bisphosphate the degree of response of Jg~,,,,, to glucagon ultimately depends on the extent of activation of the fructose 1,6-bisphosphate cycle in the direction of glucose synthesis. The experiments reporting no change or a decrease in triose phosphates and fructose 1,6-bisphosphate with glucagon treatment were all performed using perfused rat livers; the experiments indicating a rise in these compounds after glucagon were carried out with isolated, perifused hepatocytes (1). The use of dissimilar preparations may account for the reported differences. For example, metabolites in nonhepatocyte cells that are present in perfused livers might have dominated the analytical results. Isolated hepatocytes, on the other hand, may suffer significant loss of essential, small molecules during in vitro studies, and this loss might distort the response to glucagon. In this regard, it is noted that

PYRUVATE

KINASE

AND

CONTROL

the positive effects on triose phosphate and fructose 1,6bisphosphate were recorded in hepatocytes that had incubated over 2 h (1). The oleate used by Groen et al. (1) to supplement lactate-pyruvate may have changed the response to glucagon. Other than (1) there has apparently been only one report of studies using hepatocytes from fasted rats in which the effect of glucagon on fructose 1,6-bisphosphate was determined. Claus et al. (21) measured fructose 1,6-bisphosphate in hepatocytes incubated for a brief time with 10 mM pyruvate or 10 mM alanine. Fructose 1,6-bisphosphate was unchanged by glucagon in the cells incubated with alanine and was decreased in the cells incubated with pyruvate. No gluconeogenic responses were reported, and since glucagon is known to depress Jglucosein cells incubated with pyruvate, these data have limited relevance to this discussion. We felt it was essential to test whether the observations of Groen et al. (1) were peculiar to isolated hepatocytes. Therefore we carried out analyses of the “steadystate” concentrations of a number of gluconeogenic intermediates in conventionally incubated hepatocytes while gluconeogenesis from lactate-pyruvate was stimulated by glucagon or 8-bromoadenosine 3’:5’-cyclic monophosphate. The results presented in Fig. 1 can be seen to agree with the data from experiments with perfused livers (16,18-20) rather than those of Groen et al. (1). A second series of experiments reported in Table III showed again that glucagon did not increase the steady state concentration of triose phosphates and fructose 1,6-bisphosphate; however, these compounds were elevated by phenylephrine. This demonstrated that our methodology could detect hormonal perturbations of these compounds, and the mechanisms of action of phenylephrine and glucagon differ in their effects on the late reactions of gluconeogenesis. In a number of experiments, we also found that the presence of oleate did not change the glucagon response to that observed by Groen et al. (1) (data not presented). A careful examination of the data of Groen et al. (1) in regard to triose phosphates and fructose 1,6-bisphosphate reveals some problematic aspects. The effect of glucagon on fructose 1,6-bisphosphate was not determined directly, because the authors were concerned that extensive binding of this compound would make total amounts present a poor indicator of the free concentrations. The concentration of free fructose 1,6-bisphosphate was therefore estimated from the content of dihydroxyacetone phosphate and the triose phosphate isomerase and aldolase equilibria. Thus the results for both compounds were dependent on the quality of the data for dihydroxyacetone phosphate. In addition, if the artist represented the data accurately (Fig. 5B, Ref. (l)), glucagon increased the concentrations of dihydroxyacetone phosphate significantly at only two of the seven levels of substrates, and one of

OF GLUCONEOGENESIS

60-

** PEP

** 3PG

IN RAT

* TRP

57

*

**

FDP

F6P

**

**

06’

JGLUCOSE

FIG. 1. The effect of glucagon and 8-bromoadenosine 3’:5’-cyclic monophosphate on concentrations of cytosolic intermediate compounds of gluconeogenesis. Hepatocytes, 10 mgprotein/ml, were incubated in Krebs-Ringers-bicarbonate medium with 95% 02, 5% CO* as the gas phase. Lactate was initially 20 mM. At 60 min, glucagon (10v7 M), 0; 8-bromoadenosine 3’:5’cyclic monophosphate (15 PM), 0; or control vehicle was added and the incubation continued an additional 60 min. At 90 min an aliquot of the cell suspension was withdrawn and the cells were separated and extracted by centrifugation through bromododecane into perchloric acid. These extracts were assayed as described under Experimental Procedures. The data are mean concentrations + SE displayed as percentages of mean control values. Control mean values are indicated as 100% + SE. Mean concentrations of intermediate compounds in control incubations expressed as pmol/mg protein were: phosphoenolpyruvate.(PEP), 141; 3.phosphoglyceric acid (3PG), 760; triose phosphates (TRP), 106; fructose 1,6-diphosphate (FDP), 16; fructose 6-phosphate (F6P), 77; glucose 6-phosphate (G6P), 214. The control rate of glucose synthesis was 3.1 nmol/min/mg protein. Asterisks indicate a significant (Jg,ucoee) difference from control (P = 0.05 or less) determined by Student’s t test, one tail. Replications were 15, except for fructose 6-phosphate which were 10.

these two points appears to represent a single determination. On the other hand, glucagon stimulated Jg~u,,,,at the four highest and possibly all other substrate levels (Fig. 4, Ref. (1)). Thus, there is an unsatisfactory documentation of increases of this intermediate at most substrate levels in glucagon-treated cells. Sistare and Haynes (2) were unable to measure triose phosphates and fructose 1,6-bisphosphate because of the small quantity of tissue used. Their conclusion (2) that glucagon does not affect enzyme reactions involved in conversion of phosphoenolpyruvate to glucose was based on 3-phosphoglycerate-Jg1J,1,,,,, curves in which both control and glucagon data fell on a single curve, that is, the concentration of 3-phosphoglycerate was increased with glucagon treatment at all substrate concentrations, a finding consistent with the early experiments in perfused livers (16, 18-20) and with isolated hepatocytes (this paper). An enhanced Jg~,,,,, in the presence of unchanged or decreased concentrations of fructose 1,6-bisphosphate together with elevated fructose 6-phosphate and glucose

58

HAYNES

AND

6-phosphate can probably only be explained by an overall activation of the fructose 1,6-bisphosphate cycle in the direction of glucose synthesis. An important question is whether fructose 1,6-bisphosphate is almost entirely protein bound and only a very small, analytically unmeasurable, free fraction serves as substrate for fructose-1,6-bisphosphatase and this undetectable free fraction is increased by glucagon. There are two observations that argue against this. First, in some instances (16,20, this paper) glucagon led to a decrease in fructose 1,6-bisphosphate, and it is highly unlikely that there was a rise in an occult, free fraction when the total quantity decreased. Second, phenylephrine was found to cause a measurable increase in fructose 1,6-bisphosphate, indicating that, if there had been such an effect of glucagon, it could have been detected. It should be noted that glucagon, when added together with phenylephrine, prevented the increase in fructose 1,6-bisphosphate (Table III). This is also consistent with an activation of the fructose 1,6-bisphosphate cycle by glucagon. The nature of this putative activation of the fructose 1,6-bisphosphate cycle by glucagon is not clear. Hue and Bartrons (22) found that the content of fructose 2,6-bisphosphate was extremely small in hepatocytes from fasted rats when the cells were incubated with 4 mM or greater lactate. Furthermore glucagon did not affect this regulatory factor under these conditions. In addition there have been no reports of an effect of glucagon on the activity of fructose-1,6-bisphosphatase when assayed in extracts of hepatocytes that were prepared from fasted rats and incubated with lactate-pyruvate. Since glucagon increases phosphorylation of the enzyme in “fasted” hepatocytes (23) and phosphorylation of the purified enzyme increases the rate of the reaction in the absence or presence of fructose 2,6-bisphosphate (24), it may be that the proper conditions have not yet been defined for extraction and assay of this enzyme that will permit the demonstration of effects of glucagon on its kinetics. Chan (20) found glucagon inhibits 6-phosphofructol-kinase when assayed in homogenates of hepatocytes that were prepared from fasted rats and incubated with lactate. We have confirmed these observations (data not shown). It is not certain, however, if the rate of 6-phosphofructo-1-kinase is sufficiently active under these incubation conditions for its inhibition to have a significant effect on Jglucose. The model for regulation of gluconeogenesis by glucagon that is supported by the data and analyses presented in this paper includes three sites of control: (i) stimulation of JpEpCK, (ii) inhibition of pyruvate kinase, and (iii) activation of the fructose 1,6-bisphosphate cycle. These have long been considered probable sites, and the model is therefore a revival of previous ones and rejects the current model that assigns nearly complete control to the inhibition of pyruvate kinase. The experiments of Rognstad and Katz (3) and those of this paper, which con-

PICKING

firmed and strengthened their data, indicate that inhibition of JPK accounts for a variable fraction of the effect of the hormone, approximately 10 to 50%. The hormone enhances JpEpCK thereby producing an increased influx of carbon into the cytosolic segment of the gluconeogenie pathway at phosphoenolpyruvate, so that both increased input (increased JPEP& and increased efficiency (decreased JPK) occur. The ultimate stimulatory control is localized to the fructose 1,6-bisphosphate cyof Jg,ume cle, however, because it accepts fructose 1,6-bisphosphate at control or even lower concentrations and converts it at an accelerated rate to fructose 6-phosphate, which is rapidly metabolized to glucose. The final rate of glucose synthesis, therefore, is dependent on the kinetic properties of the cycle enzymes. Although this problem has been actively studied for many years, the mechanisms by which JpEPCK and the fructose 1,6-bisphosphate cycle are stimulated by glucagon have not been elucidated. The demonstration of increased JpEpCK in the absence of elevated oxalacetate strongly indicates an activation of phosphoenolpyruvate carboxykinase. Most of the work on the fructose 1,6-bisphosphate cycle has been carried out either with cells from fed rats or in “fasted” cells supplied with substrates providing high levels of fructose l-phosphate and subsequently high levels of fructose 2,6-bisphosphate. The role of fructose 2,6-bisphosphate in control of gluconeogenesis from lactate-pyruvate in hepatocytes from fasted rats therefore remains unsubstantiated. Pilkis et al. have suggested an unidentified factor or factors may regulate the fructose 1,6-bisphosphate cycle (25). ACKNOWLEDGMENT The technical knowledged.

assistance

of Kenneth

W. Smith

is gratefully

ac-

REFERENCES 1. Groen, A. K., Vervoorn, R. C., Van der Meer, R., and Tager, J. M. (1983) J. Biol. Chem. 268,14,346-14,353.

2. Sistare, F. D., and Haynes, R. C., Jr. (1985) J. Biol. Chem. 260, 12,761-12,768.

3. Rognstad,

R., and Katz,

J. (1977) J. Biol.

Chem. 252,

1831-

1833.

4. Haynes, R. C., Jr. (1985) Biochem. Sot. Trans. 13,655-656. 5. Seglen, P. 0. (1976) Methods Cell Biol. 13, 29-83. 6. Lowry, 0. H., and Passonneau, J. V. (1972) A Flexible System of Enzymatic

Analysis,

Academic Press, New York.

7. Williamson, J. R., and Corkey, B. E. (1969) in Methods in Enzymology (Lowenstein, J. M., Ed.), Vol. 13, pp. 476-481, Academic Press, San Diego.

8. Rognstad,

R. (1975) Biochem. Biophys.

Res. Commun.

63, 900-

905. 9. Smith, S. B., and Freedland, R. A. (1981) Amer. J. Physiol. 240, E279-E285. 10. Sistare, F. D., and Haynes, R. C., Jr. (1985) J. Bid. Chem. 260, 12,748-12,753.

PYRUVATE

KINASE

AND

CONTROL

11. Friedmann, B., Goodman, E. H., Jr., Saunders, H. L., Kostos, V., and Weinhouse, S. (1971) Arch. Biochem. Biophys. 143, 566-

578. 12. Cohen, S. M. (1987) Biochemistry 13. Heath,

26,573-580.

D. F., and Rose, J. G. (1985) Biochem.

J. 227,

851-

867. 14. Grunnet,

N., and Katz, J. (1978) Biochem. J. 172,595-603.

15. Patel, T. B., and Olson, M. S. (1986) Biochim. Biophys. Acta 888, 316-324. 16. Williamson, J. R., Browning, E. T., Thurman, R. (1969) J. Bid. C&m. 244,5055-5064. 17. Ochs, R. S., and Lardy, 9962.

R. G., and Scholz,

H. A. (1983) J. Biol. Chem. 258,

OF GLUCONEOGENESIS

IN RAT

59

18. Parrilla, R., Jimenez, I., and Ayuso-Parrilla, M. S. (1975) Eur. J. Biochem. 56,375-383. 19. Exton, J. H., and Park, C. R. (1969) J. Biol. Chem. 244, 14241433. 20. Chan, T. M. (1984) J. Biol. Chem. 259,7426-7432. 21. Claus, T. H., El-Maghrabi, M. R., and Pilkis, S. J. (1979) J. Biol. Chem. 254,7855-7864. 22. Hue, L., and Bartrons, R. (1984) Biochem. J. 218,165-170. 23. Garrison, J. C., and Wagner, J. D. (1982) J. Biol. Chem. 257, 13,135-13,142. 24. Ekdahl, K. N., and Ekman, P. (1984) FEBS L&t. 167, 203-

209. 9956-

25. Pilkis, S. J., El-Maghrabi, Rev. Biochem. 57,755-783.

M. M., and Claus, T. H. (1988) Annu.

The role of inhibition of pyruvate kinase in the stimulation of gluconeogenesis by glucagon: a reevaluation.

We have reexamined the concept that glucagon controls gluconeogenesis from lactate-pyruvate in isolated rat hepatocytes almost entirely by inhibition ...
1MB Sizes 0 Downloads 0 Views