Eur. J . Biochem. 56, 375-383 (1975)

Glucagon and Insulin Control of Gluconeogenesis in the Perfused Isolated Rat Liver. Effects on Cellular Metabolite Distribution Roberto PARRILLA, Isabel JIMENEZ, and Matilde S. AYUSO-PARRILLA Department of Metabolism, Instituto G. Maraiion, Consejo Superior de Investigaciones Cientificas, Madrid (Received March 17/April 24; 1975)

The metabolic effects of glucagon and glucagon plus insulin on the isolated rat livers perfused with 10 mM sodium L-lactate as substrate were studied. Glucagon stimulated gluconeogenesis, ketogenesis and ureogenesis at the concentration used of 2.1 nM. The addition of insulin to give a glucagon-to-insulin ratio of 0.2 reversed all the glucagon effects. The glucagon enhancement of gluconeogenesis was accompanied by a rise in the cytosolic and mitochondrial state of reduction of the NAD system and a fall in the [ATP]/[ADP] ratio. The analysis of the intermediary metabolite concentrations suggested, as possible sites of glucagon action, the steps between pyruvate and phosphoenolpyruvate as well as the reactions catalyzed by phosphofructokinase and/or fructose bisphosphatase. All the changes in metabolite contents were abolished when insulin was present. Glucagon increased the intramitochondrial concentration of all the metabolites, whose intracellular distribution was calculated. The finding of a significant rise in the calculated intramitochondrial concentration of oxaloacetate points to pyruvate carboxylation as an important site of glucagon interaction with the gluconeogenic pathway. A primary event in the glucagon action redistributing intracellular metabolites seems to be the mitochondrial entry of malate. The possibility is discussed that the changes in metabolite cellular distribution were brought about by the increased cellular state of reduction caused by the hormone.

Since an early study, in which the role of glucagon on hepatic gluconeogenesis was postulated [l - 41, a considerable amount of information has been accumulated regarding its mechanism of action. However, the efforts do not seem to have been succesful, since as yet no satisfactory explanation has been found to elucidate the mechanism leading to the enhancement of the hepatic gluconeogenic flux under the different situations in which the hormone is known to be active. Recent work has pointed to the activation of some metabolic event between pyruvate and phosphoenolpyruvate as the cause for the observed effect of glucagon stimulation of the gluconeogenic flux from subEnzymes. Aconitate hydratase (EC 4.2.1.3); alanine dehydrogenase (EC 1.4.1.1); fructose-1,6-bisphosphatase(EC 3.1.3.11); glutamate dehydrogenase (EC 1.4.1.2); glutamic-oxaloacetic transaminase (EC 2.6.1. l ) ; 3-hydroxybutyrate dehydrogenase (EC 1.1.1.30); isocitrate dehydrogenase (EC 1.1.1.42); lactate dehydrogenase (EC 1.1.1.27); malate dehydrogenase (EC 1.1.1.37); ‘malic’ enzyme (EC 1.1.1.40); phosphoenolpyruvate carboxykinase (EC 4.1.1.32); phosphofructokinase (EC 2.7.1.11);pyruvate carboxylase (EC 6.4.1.2).

Eur. J. Biochem. 56 (1975)

strates like lactate, pyruvate or alanine. For some authors pyruvate carboxylase was the non-equilibrium step susceptible to glucagon activation [ 5 ] , while others pointed to phosphoenolpyruvate carboxykinase as the site of the hormone action [6]. The conversion of pyruvate to phosphoenolpyruvate, due to the almost exclusive mitochondrial location of pyruvate carboxylase in rat liver [7,8], involves the intramitochondrial formation of oxaloacetate and its transport to the cytosol as aspartate or malate [9,10]. Theoretically the transport of these metabolites across the mitochondrial membrane may be subject to control. In this work the effect of glucagon on hepatic metabolite distribution has been studied in order to find out whether the hormone could accelerate the rate of carbon transport to the cytosol by altering the distribution of some key metabolites on both sides of the mitochondrial membrane. The study of the metabolite distribution also seems to be relevant from the point of view that many enzymes are subject to control by metabolites in the range of concentration found

376

under physiological conditions. The concentration of intermediates in the whole tissue may consequently obscure changes in metabolite concentrations in a cellular compartment, which may be of importance for the rate control of a given enzymic reaction. From a physiological point of view the glucagon effect probably should not be considered separately from that of insulin. Antagonistic effects of both hormones have been reported by several authors using different experimental models and under variable experimental conditions [ l l - 151. The hepatic tissue is greatly sensitive to variations in the concentration of both hormones, and recent work pointed to their molar ratio as very important in the control of the metabolic activity of the hepatocyte [16]. This work has also studied whether in the situation of perfused isolated liver insulin, administered so as to obtain glucagon : insulin molar ratios close to those normally found in the portal vein in vivo, inhibited glucagon metabolic effects and the molecular events leading to such an inhibition.

MATERIALS AND METHODS Reagen ts

Most of the reagents were obtained from Sigma (Sigma Chemical Co., St. Louis, Missouri, U.S.A.). Enzymes were obtained from Boehringer (Mannheim, Germany). Animals

Male albino rats of the Wistar strain of 200 f 10 g in body weight were used. These rats were purchased when approximately 120 g in weight and kept in our Center under controlled conditions of food intake, light and temperature. Prior to their experimental use all the animals were starved for 24- 36 h. Liver Perfusiori

Livers were perfused with 100 ml of blood-free recirculating medium whose composition was KrebsHenseleit bicarbonate buffer 1171 containing 4 % fraction V bovine albumin. The pH of the perfusion fluid, after equilibration with a 95% 0, : 5 % C 0 2 gaseous phase, was 7.4. The day before to its experimental use the albumin was dissolved in a small volume of bicarbonate buffer and dialyzed overnight at 4 ° C against two or three changes of the same buffer. The fluid was made sterile before using by filtering it through Millipore membranes having 0.45-pm pore size.

Hormonal Control of Hepatic Gluconeogenesis

The experimental set-up for liver perfusion was provided with a discs oxygenator similar to the one previously used for dog [ 181 or rat [ 191 liver perfusion. The medium was made to circulate through the liver at a rate of 28 ml/min which, according to previous experiments [20], was found to maintain an adequate oxygen supply to the organ. In all the experiments the preparation was allowed to equilibrate by perfusing it for 30 min without any substrate. At this time a perfusate sample was taken and then substrate and hormones were added. A sufficient amount of 1 M sodium L-lactate was added to give a final perfusate concentration of 9- 11 mM. Stock solutions of insulin (1 mg/ml) made up in doubledistilled water, pH 3.0, and of glucagon (1 mg/ml) in glycine buffer, pH 10, were stored deep-frozen in aliquots. Prior to their experimental use they were diluted with albumin buffer and added to the perfusate to give a final concentration of 10.5 nM insulin and 2.1 nm glucagon. Oxygen tension in the perfusion fluid was determined polarographically using a platinum electrode (Yellow Spring, Ohio, U.S.A.). Analytical Techniques

1-ml perfusate samples were taken every 10 min and rapidly poured into test-tubes containing 2 ml of cold 6 % perchloric acid. The precipitate was removed and the clear supernatant brought to pH 6 with potassium carbonate. At the end of the perfusion the livers were frozen in situ using aluminium clamps [21] cooled in liquid nitrogen. Representative portions of the frozen tissue were freeze-dried and extracted with 35 volumes of 8 % perchloric acid (w/v). Acid extracts were brought to pH 6.0 with potassium carbonate and immediately used for metabolite assays. Tricarboxylic-acid-cycle intermediates were measured fluorimetrically 1221. 3-Hydroxybutyrate and acetoacetate were measured according to Williamson et al. [23].Alanine was measured with alanine dehydrogenase and NAD' in Tris- hydrazine buffer. Other intermediates were measured according to previously described methods 124,251. CoA and acyl-CoA derivatives were extracted and assayed basically as described by Williamson et al. [22]. Calculation of the Metabolite Distribution

The distribution of malate and oxaloacetate has been calculated [26] assuming that the cytosolic malate and lactate dehydrogenase are in equilibrium with the same pool of cytosolic NAD' . It is assumed also that the malate and 3-hydroxybutyrate dehydroEur. J Biochem. 56 (1975)

377

R. Parrilla, 1. Jimenez, and M. S. Ayuso-Parrilla

genases are in equilibrium with the same NAD' mitochondrial pool. Glutamate and aspartate distributions have been calculated according to Greenbaum et al. [27]. The assumptions made on these calculations are as follows. (a) Glutamic-oxaloacetic transaminase is in equilibrium in both cellular compartments; (b) ammonia is evenly distributed; (c) 3-hydroxybutyrate and glutamate dehydrogenases are in equilibrium with the same NAD' mitochondrial pool. 2-Oxoglutarate distribution was calculated from the glutamic-oxaloacetic transaminase equilibrium reaction once the glutamate and aspartate distributions were known. For calculating the isocitrate distribution [26] the cytosolic isocitrate was first obtained by combining the equilibrium reactions of isocitrate dehydrogenase and malic enzyme. Cytosolic or mitochondrial citrate were calculated from the aconitate hydratase reaction [26]. K values used in these calculations were those reported by Williamson et al. [28] and Krebs and Veech [29]. For the calculation of metabolite concentrations, the cytosolic water content was taken to be 2.323 ml/g dry wt liver and mitochondrial water content 0.237 ml/g dry wt liver. These values were calculated taking the intracellular water as 65% of the total tissue water content, the mitochondrial water 0.8 pl/ mg protein [30] and a mitochondrial protein content of 60mg/g fresh wt tissue [31]. The ratio fresh/dry liver was found to be 4.94 k 0.06, mean of 80 determinations. RESULTS Hormonal Effects on Glucose Synthesis The addition of glucagon to the perfusion medium produced a 34% stimulatory effect on the rate of gluconeogenesis from L-lactate (Table 1). A similar effect of glucagon on glucose output using lactate as substrate has been previously described [I]. The addition of insulin to the perfusion fluid so as to give a glucagon : insulin ratio of 0.2 virtually abolished the glucagon effect on glucose production. Glucagon enhanced lactate uptake in such a way that all the glucose produced could be accounted for by the substrate uptake. The ratio lactate removed to glucose formed was about 2.3 under all the situations studied (Table 1). The effects of glucagon and glucagon plus insulin on the metabolic balances across the livers also appear summarized in Table 1. Glucose production could always be accounted for by the lactate uptake. Glucagon enhanced ketogenesis and ureogenesis by almost 50 % and both effects were prevented Eur. J. Biochem. 56 (1975)

Table 1. Effects of glucagon and glucagon plus insulin on the rute of metabolite changes in the perfusion medium Livers from fasted rats were pre-perfused for 30 min and then sodium L-lactate (10 mM initial concentration) and hormones were added and the livers perfused for 60 min. Metabolic rates were calculated over the last 30-min interval of perfusion. The results are means of at least eight experiments k S.E. Metabolite change Rate of metabolite change in presence of -~-~ __ no glucagon glucagon addition + insulin pmol 100 g body wt-' Glucose formed 123 k 1 0 Lactate used 287 k 1 6 Lactate removed/ 2.3 k 0.2 glucose formed Lactate accounted for as glucose 246 k 2 0 -7.9 0.8 Pyruvate formed Ketone bodies formed 7 f 0.5 Lactate un41.9 accounted for Urea formed 9 k 0.7

*

h-'

165 i 8 375 k 2 4 2.2

* 0.05

120 * l o 264 k 2 0 2.3 k 0.19

330 k 16 -6.5 k 0.7

240 *20 1.1 -9.4*

k 0.9

6 & 0.6

10

41.5 16 i 1.3

27.4 13

* 0.9

by the addition of insulin to the perfusion fluid. The amounts of lactate unaccounted for and thus potential substrate of respiration were very similar in control and glucagon-treated livers, about 40 pmol . 100 g body wt.-' h-', and lower when insulin was present. Effects of Glucagon or Glucagon plus Insulin on the Hepatic Content of Intermediary Metabolites The changes in the hepatic content of gluconeogenic intermediates 60 min after the additions of glucagon or glucagon plus insulin are shown in Table 2. The additions of glucagon caused a decreased in lactate, pyruvate and fructose bisphosphate while all other intermediates increased their concentrations above the control. The lactate-to-pyruvate ratios rose from 15 to 18 under glucagon stimulation indicating a shift to a more reduced state of the cytosolic NAD couple. When insulin was added together with glucagon the hepatic content of virtually all the intermediates returned to values very close to those found in the control livers. These values from Table 2 appear represented in the form of a crossover plot in Fig. 1. The idea is not to evaluate the data according to the classical crossover theorem [32] but to facilitate the evaluation of the data. According to the variations in metabolite content two sites of interaction emerged as probable points of glucagon action. One between pyruvate and phosphoenolpyruvate and the second one between fructose bisphosphate and hexose phosphate. +

378

Hormonal Control of Hepatic Gluconeogenesis

Table 2. ,5flects qf'glucagon and glucagon plus insulin on the hepatic steadj,-.staie c,onceiitraiion of gluconeogenic intermediates Livers from fasted rats were preperfused for 30 min and then sodium L-lactate (10 mM initial concentration) and hormones were added. Livers were freeze-clamped in situ 60 min later. Representative portions of the liver were stored in liquid nitrogen and processed as described in Methods. The results are means of at least eight observations & S.E:. Metdbohte

180

A

160

Metdbolile content in presence of -

tio

-

'idditioii

glucagon

glucagon inaulin

+

iimol g dry wt ~

LdCldte Pyruvdtc Phosphomoipyruvate %Phosphoglyceratc Triow phosphate Glycerol 3 pho\phcitc Fructose bi\phosphdte Fructox 6phosph'itc Glucose 1 phosphd t c Glucose 6phosphdtc [Lactate] [pyruvdtel -

-

* By t-test P

14595 t 9 1 0 995 f 119 466

351 36

& 66

+ +

13834f948 808) 76" 638f

62

517f

4

41

1034 5 38

*

42"

13721 k 9 4 8 944 f 101 492

& 62

77"

360

I

60

4

39

f

6

1 1 . 5 2 i 102"

933 & 66

49

y

5

46+

2

48

*

4

28

I

3

41 f

5"

30 f

5

2

21 f

2"

15

2

16"

203

f 18

16 5

180

15

1 5 3 ~1 2

232f 18+

14"

136f

18

~ -

0 05

i

The hepatic content of tricarboxylic-acid-cycle intermediates. amino acids and CoA derivatives are shown in Table 3. Glucagon increased the content of all the tricarboxylic-acid-cycle intermediates but oxaloacetate which was decreased. Aspartate, alanine and acetyl-CoA did also show a mild decrease in their content. The decrease in oxaloacetate is particularly striking and seems to be in conflict with the observed increase in the gluconeogenic flux. Seemingly the slight decrease in the acetyl-CoA content does not support the idea that glucagon might control gluconeogenesis at the pyruvate carboxylase step [33,34]. Glucagon did also decrease the content of CoA and total soluble CoA and increased the acetyl-CoA : CoA and 3-hydroxybutyrate : acetoacetate ratios. Insulin again in most instances shifted the hepatic content of metabolites to the values found in control livers. The only exceptions were alanine and CoA, which moved farther in the same direction in the presence of insulin. The effect of glucagon increasing the 3-hydroxybutyrate : acetoacetate ratio is the result of a more reduced state of the mitochondria1 N A D ' couple

60' Lac

'

Fyr

Oaa

~

Ma1

f-hr 1

Tri-f

f-Gri

Fru-P Glc-I-P Glc 'Fru-P Glc-6-P

Gro-F

Fig. 1. Crossover plots uf the steady-stuie ~oi11~1~i1t~utioii q/ gluwneogenic intermediates in ( A ) glucagon or B J glucagon-l)lus-in.sulintreated livers. (-~---) Connects glycerol 3-phosphate which is not in the direct sequence. The graph has been made using data from Tables 1 and 3. Each point is the mean value of cight experiments and the vertical bars the standard errors of the means. Abbreviations used: Lac, lactate; Pyr, pyruvate; Oaa, oxaloacetatc; Mal, malate: P-Pyr. phosphoenolpyruvate ; P-Gri, 3-phosphoglycerate: Tri-P, triose phosphate; Gro-P, glycerol 3-phosphate: Fru-P,. fructose bisphosphate; Fru-P, fructose 6-phosphate: Clc-1-P, glucose 1phosphate; Glc-6-P, glucose &phosphate

and parallels similar changes (Table 2) described above in the cytosolic compartment. The fact that insulin lowered those ratios suggests that this shift in the state of reduction is a necessary event during glucagon stimulation of gluconeogenesis. Changes in the Hepatic Content of Adenine Nucleotides

The hormonal treatment did not alter the total adenine nucleotide content of the liver nor the adenylate kinase mass action ratio, which at all times remained close to the value described for the experimentally determined equilibrium constant (Table 4) [35]. Glucagon treatment lowered the [ATP]/[ADP] ratio from 3.3 to 2.6 and the presence of insulin brought the ratio back to 3.1. The lower ratio was attained by a decrease in ATP and a rise in ADP content. The addition of 10 mM L-lactate produced an increase of 276 patoms. 100 g body wt-' . h-' in the hepatic oxygen uptake (Table 4). This increase accounts for all the energy demands for the observed rate of glucose production (Table 1). Glucagon produced a further increase in the oxygen uptake from 276 to 339 patoms while insulin lowered the uptake to 300 patoms. Fur. J . Biochein. 56 (197.5)

319

R. Parrilla, I. Jimenez, and M. S. Ayuso-Parrilla

Table 3. Effect of glucagon and glucagon plus insulin on the hepatic content of intermediary metabolites Experimental conditions were as described in Table 2. The results are means of at least eight experiments f S.E. Statistical significance estimated by t-test Metabolite

Content of metabolites in presence of no addition

glucagon

glucagon

+ insulin

nmolig dry wt Malate Oxaloacetate Acetyl-CoA Citrate Isocitrate 2-Oxoglutarate Glutamate Aspartate Alanine Ammonia CoA Total acid-soluble CoA Acid-insoluble acyl-CoA Acetyl-CoAiCoA 3-Hydroxybutyrate Acetoacetate [3-Hydroxybutyrate]/[acetoacetate] [Malate]/[oxaloacetate] a

1186 f 1 6 1 18.7 f 2.4 347 f 30 2118 f266 91- f 6 1763 k 125 9481 f 3 7 2 1902 f 155 3730 f 3 5 0 2470 f 2 6 5 412 f 29 1400 k 2 0 7 114 f 24 0.86f 0.06 2729 f 297 934 f 162 2.9 f 0.4 70 10

1334 f 101” 11.2 3.2‘ 336 f 17 2518 f376‘ 132 8‘ 2009 f 2 4 2 b 10297 f 470 1515 f 84” 3120 k 70 2500 f 300 315 14” 1269 f 95 159 f 14” 1.06f 0.06” 2639 f 198 709 123’ 4.6 f 0.7’ 134 k 20”

*

+

1249 f 112 16.3 f 3 335 f 60 2202 f 244 8 83 f 1766 168 8981 f 1 1 0 7 1766 f 179 2610 k 230’ 2550 f 325 332 19 1342 f 232 104 k 22 0.2 0.94 f 2668 i- 2.19 839 +_ 120 3.7 f 0.5 95 f 13

P < 0.1.

” P < 0.05.

P < 0.001.

Table 4. Eflects of glucagon and glucagon plus insulin on the hepatic content of adenine nucleotides Experimental conditions were as described in Table 2. The results are means of at least eight experiments i S.E. Nucleotide

Adenine nucleotide content in presence of no addition

glucagon

glucagon

+ insulin

8287 2622 1804 12172

f562 f110 f156 f764

nmol/g dry wt ATP ADP AMP Total adenine nucleotides [ATPI/[ADPl [ATP] [AMP]/[ADP]*

7740 2333 1595 11668

f 628 ,153 f 72 f698

3.36 f 2.10f

0.29 0.15

7422 2899 1790 12123

f 390 f218” f 82 f494

2.6 f 1.87f

0.21” 0.24

3.16f 2.01 f

-

0.26 0.18 -

A Oxygen uptake (watoms . 100 g body wt-’ h-’) a

276

339

300

By t-test P < 0.05.

Hepatic Metabolite Distribution The cytosolic concentrations of several intermediary metabolites are shown in Table 5. Glucagon decreased the cytosolic concentration of malate, oxaloacetate, glutamate and aspartate. 2-Oxoglutarate was increased while citrate and isocitrate remained unchanged. Insulin reversed the glucagon effects Eur. J. Biochem. 56 (1975)

except €or glutamate, citrate and isocitrate, which displayed a further decrease. The decrease in the cytosolic oxaloacetate in glucagon-treated livers may reflect an increased phosphoenolpyruvate formation; however, it is a rather conflicting finding considering that phosphoenolpyruvate carboxykinase responds readily to variations in the oxaloacetate concentration in the range found in vivo [36]

380

Hormonal Control of Hepatic Gluconeogenesis

Table 5. bYffrct5 of glucagon and glucagon plus insulin on the cytosolic concentrations of intermediary metabolites The results have been calculated as described in Methods using the data from Tables 2 and 3 Metabolite

Concentration of intermediary metabolites in presence of no addition

glucagon

glucagon

+ insulin ~

Malate Oxaloacelatz Citratc Isocitrate 2-Oxoglutarate Glutdinate Aspartate

~~

ninollg dry wt

mM

nmol/g dry wt

mM

nmolig dry wt

mM

1098 18.7 1723 81 1760 9266 1868

0.47 0.008 0.74 0.034 0.75 3.98 0.80

785.5 10.93 1702 80 1997 8686 1251

0.33 0.0047 0.73 0.034 0.86 3.74 0.53

865.5 15.94 1425 67 1761 8511 1619

0 37 0 0068 0 61 0 028 0 75 36 0 69

'Table 6. ~ f j ' k r s .o j glucagon and glucagon plus insulin on the mitochondrial concentrations of' intermediurj metabolircs The results havc been calculated as described in Methods using the data from Tables 2 and 3 Metabolite

Concentration of intermediary metabolites in presence of no addition

Malate Oxaloaceta te Citrate lsocitratc 2-Oxoglutarate Glutamate Aspartatc

glucagon

glucagon

+ insulin

nniollg wt

mM

nmol/g dry wt

mM

nmolig dry wt

mM

87 0.02 394 10 2.39 214 34

0.37 0.00008 1.6 0.042

548 0.07 815 51 11.14 1610 263

2.3 0.00029 3.44 0.21 0.047 6.7 1.1

383 0.06 776 16 4.04 469 146

1.6 0.00025 3.27 0.067 0.017 1.9 0.61

0.01 0.90 0.14

The intramitochondrial variations in the metabolite concentrations are described in Table 6. Glucagon stimulation produced changes in the opposite direction to that in the cytosol, that is, an increase in the concentration of all the metabolites whose distribution was studied. The increase in the intramitochondrial oxaloacetate concentration, despite the marked decrease in the whole tissue content, agrees with the idea of a possible activation of pyruvate carboxylase, whose cellular location in rat liver is almost exclusive of this compartment [8]. Particularly interesting is the almost eight-fold increase in the aspartate concentration after glucagon stimulation, considering that this intermediary seems to be the form under which oxaloacetate is transported to the cytosol during gluconeogenesis from lactate [9,37]. The presence of insulin once again resulted in a reversal of the glucagon efrects, returning the metabolite concentrations to values approaching those of control livers. Table 7 shows, as expected from the foregoing results, how glucagon increased the mitochondrial : cytosolic concentration gradients of all the metabolites studied. It was particularly striking to observe the twelve-fold increase in the aspartate gradient.

Table 7. EJfect.7 of glucagon and ghcagon p1u.s insulin on the mitochondriul :cjlosolic concentration gradients of in rcririeo'iurj' nWtUhCJ1ite.s

Gradients have been calculated using the concentrations (mM) shown in Tables 5 and 6 Metabolite

Malate Oxaloacetate Citrate Isocitrate 2-Oxoglutarate Glutamate Aspartate

Mitochondria1 : cylosolic concn gradient in prcsence of no addition

glucagon

glucagon + insulin

0.78 0.01 2.16 1.23 0.013 0.22 0.17

6.96 0.061 4.71 6.17 0.054 1.79 2.07

4.32 0.036 5.36 2.39 0.022 0.53 0.8X

DISCUSSION Sites of Glucagon Action

An apparent site of glucagon interaction with the gluconeogenic pathway was found to be between Eur. J. Biochem. 56 (1975)

381

R. Parrilla, I. Jimenez, and M. S. Ayuso-Parrilla

pyruvate and phosphoenolpyruvate (Fig. 1 and Table 2). The decrease in total oxaloacetate (Table 3) is in conflict with the increase in flux at any of the two non-equilibrium enzymes of this metabolic path. However, the study of the metabolite distribution in glucagon-treated livers shows how in spite of the decrease in total oxaloacetate content its intramitochondrial concentration was actually raised (Table 3), suggesting an activation of the intramitochondrial enzymic step of pyruvate carboxylation. The apparent lack of changes in acetyl-CoA (Table 3), an important activator of this enzyme, points to variations in the concentration of inhibitors like acetoacetyl-CoA [38] as possible modulators of the acetyl-CoA response. The possibility might exist that glucagon produced a redistribution of cellular acetyl-CoA leading to a rise within the intramitochondrial compartment without appreciable variations in its total content. The steady-state concentration of cytosolic oxaloacetate (Table 5) is well below the K, of phosphoenolpyruvate carboxykinase for this intermediary [36], which is the reason why variations in its concentration should be very important for the control of flux at this step. On these grounds the observed effect of glucagon decreasing oxaloacetate concentration does not seem to agree with the postulate that the hormone activates this enzymic step [6,39]. However, the possibility can not be excluded, provided that the K, of the enzyme for oxaloacetate were somehow lowered under glucagon treatment. Cyclic AMP, whose hepatic concentration is known to increase after glucagon treatment [40], has been suggested to be an important candidate as effector of phosphoenolpyruvate carboxykinase [27]. The finding of a fall in fructose bisphosphate and a rise in hexose phosphates in the presence of glucagon is in agreement with the previous report of Schimassek and Mitzkat [l], who originally suggested that the glucagon site of action was at the level of fructose-1,6bisphosphatase and/or phosphofructokinase. This assumption is based on the finding of decreased fructose bisphosphate and hexose phosphate concentrations (Table 2). The fact that glucagon apparently was not able to promote gluconeogenesis from substrates like oxaloacetate or fructose [4,40] suggested that these changes could probably be related to the glycogenolytic action of glucagon. However, recent work seems to indicate that glucagon is able to promote gluconeogenesis from substrates like dihydroxyacetone, fructose or xylitol [41- 431, pointing also to the enzymes phosphofructokinase and fructose1,6-bisphosphatase as probable sites of the hormone action. However, the aforementioned changes in metabolite contents can not be explained on the basis of the known regulatory properties of these two Eur. J. Biochem. 56 (1975)

enzymes. Phosphofructokinase is inhibited by ATP and citrate [44] but glucagon produced a slight decrease in ATP ; and although the citrate content rose its actual concentration at the cytosolic compartment remained unchanged (Table 5). On the other hand, fructose bisphosphatase is known to be inhibited by AMP [45], while glucagon produced a rise in AMP content (Table 4). Thus, unless glucagon produced changes in the adenine nucleotide content at the enzyme sites, which are not reflected by their total tissue content, it is not plausible to explain the observed metabolite changes on the grounds of the enzyme behaviour in vitro.

Hepatic Cellular Metabolite Distribution and Redox State Glucagon seems to produce a net malate mitochondrial entry, which leads to an almost eight-fold rise in the mitochondria1 : cytosolic gradient of this intermediary (Tables 5 - 7). The simultaneous rise in the cytosolic 2-oxoglutarate (Table 5) seems to indicate that malate could have been exchanged through the malate - 2-oxoglutarate translocator [46, 471. The rise in the intramitochondrial level of oxaloacetate probably facilitates aspartate formation, as judged by the eight-fold increase in the intramitochondrial concentration of this amino acid (Table 6). This finding may be of importance for the understanding of the observed glucagon effect on the gluconeogenic flux, considering that aspartate serves as the carbon carrier to the cytosol during gluconeogenesis from lactate [37,48]. The observed decrease in the cytosolic concentration of aspartate is probably the result of the rise in the steady-state concentration of 2-oxoglutarate to which it is linked through the cytosolic aspartate aminotransferase equilibrium reaction. On these grounds it seems reasonable to conclude that a primary event associated with glucagon stimulation of glucose production is a net malate intramitochondrial movement. The question is whether the hormone acts by directly accelerating malate exchange or whether this is only the consequence of some other metabolic event. The apparent intramitochondrial movement of glutamate and malate (Tables 5 and 6) could be secondary to the increase in redox state induced by the hormone (Tables 2 and 3). Teleologically this redistribution of metabolites would be necessary to maintain the redox potential gradient between the two cellular compartments when the redox state was increased. Assuming that the shift of the nicotinamide nucleotide system to a more reduced state were the cause for the calculated cellular redistribution of metabolites, the question would be through which

382

mechanism does glucagon increase the redox state. Struck et al. [33] originally suggested that glucagon effects could probably be the result of an enhanced lipid mobilization. This hypothesis was further developed by Williamson et al. [5,34]. The recent finding that glycodiazine. an inhibitor of lipolysis [49], does not prevent the glucagon stimulation of gluconeogenesis [50]has cast some doubts on this hypothesis. Nevertheless, the fact that the ratio lactate used : glucose formed remained unchanged after glucagon stimulation (Table 1) seems to support the idea that an enhanced lipid oxidation may be the source of the extra energy needed for glucose synthesis. The possibility may exist that glucagon activates some lipases which are not sensitive to glycodiazine inhibition. .4ctually from the work of Frolich and Wieland [50] it can be appreciated that glucagon produced a mild ketogenic effect in spite of the presence of glycodiazine. Since the lactate available for oxidation does not account for the ATP requirements for glucose synthesis after glucagon stimulation (Table l), it seems reasonable to conclude that the hormone may primarily act to mobilize endogenous fuels for respiration, these most probably being lipids in view of the increased reduction of the mitochondria1 NAD' system and the accelerated rate of ketone bodies production (Tables 1 and 2).

Glucagon - Insulin Interactions The addition of insulin to the perfusion fluid prevented the described metabolic effects of glucagon. Similar antagonistic effects of both hormones have already been well described. However, as pointed out in a recent publication [I61 it is remarkable how molar ratios of both hormones, in the range of those found in the portal vein in vivo, completely abolished the enhancement of metabolism induced by glucagon alone. The fact that insulin inhibition runs parallel with a return to the control values of the intermediary metabolites suggests that both hormones may act on some early common step(s). The exact nature of this antagonism is unknown and it can not be explained on the basis of the above-reported experimental findings. Cyclic AMP has been suggested to play a role in this hormonal interaction in view of the reported antagonistic effect of both hormones on the hepatic content of this nucleotide [51] as well as the known inhibitory effects of insulin on the cyclic AMP metabolic effect [13,52]. Authors wish to thank Mr T. Fontela for his devoted and skilful1 technical assistance. This work has been supported by grants from Lilly Indiana S. A . . Funrlacibn Rodriguez Pascual and Comisibn Asesoru para (,I Dcstrrrollo de la lnvestigacicin.

Hormonal Control of Hepatic Gluconeogenesis

REFERENCES 1. Schimassek, H. & Mitzkat, H. J. (1963) Biochem. Z . 337, 510-518. 2. Exton, J. H., Jefferson, L. S., Jr, Butcher. R. W., Jr & Park, C. R. (1966) Am. J . Med. 40, 709-715. 3. Garcia, A., Williamson, J . R. & Cahill. G . F. (1966) Diabetes. 15, 188- 193. 4. Ross, B. D., Hems, R. & Krebs, H. A. (1967) Biochen7. .J. 102. 942 - 951, 5. Williamson, J. R., Browning, E. T., Thurman, R. G. & Scholz. R. (1969) J . B i d . Chem. 244, 5055 - 5064. 6. Exton, J. H. & Park, C. R. (1969) J . Biol. Cl7em. 244. 14241433. 7. Scrutton, M. C. & Utter, M. F. (1968) Annu. Rcv. Biochmnz. 37,249 - 302. 8. Utter, M. F. & Scrutton, M . C. (1969) in Currenr Topics in Cellular Regularion (Hoerecker, B. L. & Stadtman, E. R.. eds) vol. 1, pp. 253-296, Academic Press, New York. 9. Lardy, H. A., Patkau, V. & Walter. P. (1965) Proc. Nut1 Acud. Sci. U . S . A .53, 1410- 1415. 10. Krebs, H. A,, Gascoyne. T. & Notton. B. M. (1967) Biochi>m. J . 102,275 - 282. 31. Miller, L. L. (1961) Recent Prog. Horm. Rrs. 17, 539-564. 12. Jefferson, L. S., Exton, J. H., Butcher. R. W., Sutherland, E. W. & Park, C. R. (1968) J . Biol. C h r z . 243, 1031-1038. 13. Glinsmdn, W. H. & Mortimore, G . E. (1968) Am. J . Physiol. 215, 553- 559. 14. Menahan, L. A. & Wieland, 0. (1969) Eur. J . Biochem. 9, 55 - 62. 1 5. Parrilla, R., Toews, C. J. & Goodman, M. N. (1972) Diabetes, 21, 52. 16. Parrilla, R., Goodman, M . N. & Toews. C. J. (1974) Diahrtes. 23,725-731. 17. Krebs, H. A. & Henseleit, K. (1932) Hoppe-Seyler's Z. Phjsiol. Chem. 210,33-66. 18. Kerstens, P. J. (1964) in La Perfusion du Foie Isolc,, pp. 26- 38, Editions Arscia, Bruxelles. 19. Scholz, R. (1968) in Stofjkchsel der isoliert prrfudierten L.eber (Staub, W. & Scholz, R., eds) pp. 25-31, Springer Verlag, Berlin. 20. Goodman, M. N., Parrilla, R. & Toews, C. J. (1973) Am. J . Physiol. 225, 1384- 1388. 21. Wollenberger, A., Ristau, 0. & Schoffa, G. (1960) Pfliigers Arch. Gesamte Physiol. Menschm, Tierr. 270. 399 -412. 22. Williamson, J. R. & Corkey, B. E. (1968) Methods Enzjmol. 13,434- 512. 23. Williamson, D. H., Mellanby, J. & Krebs. H. A. (1962) Biochem. J . 82, 90-96. 24. Bergmeyer, H. U., ed. (1965) A4ethod.r of' Enz.vmatic Analysis, Academic Press, New York. 25. Parrilla, R. & Goodman, M. N. (1974) Biochem. J . 138, 341348. 26. Williamson, J. R. (1969) in The Ener