Evaluation of the Relationship between the Intra- and Extramitochondrial [ATP]/[ADP] Ratios Using Phosphoenolpyruvate Carboxykinase”

The ratio of free ATP to free ADP in the mitochon- drial matrix ([ATPt]/[ADP,]) has been rfieasured in suspensions of isolated mitochondria under conditions of active phosphorylation of extramitochondrial ADP. These measurements utilized phosphoenolpyruvate carboxykinase which is present in the matrix of mitochondria from the livers of guinea pigs, chickens, and pigeons. Mitochondria isolated from these sources also contain nucleoside diphosphate kinase, malate dehy- drogenase, and glutamate dehydrogenase or 3-OH-bu-tyrate dehydrogenase. Together these enzymes catalyze the synthesis of phosphenolpyruvate and CO, from oxaloacetate with oxidative phosphorylation as an energy source. These reactions have been shown to be fully reversible in suspensions of mitochondria isolated from the above sources. With oxidative phosphorylation as the source of ATP, phosphoenolpyruvate was synthesized from malate and conversely addition of phosphoenolpyruvate, ADP, and COz led to synthesis of malate and ATP. The forward and reverse reactions were allowed to continue until the rate of change of metabolite concentrations was minimal and then the latter were measured. The intramitochondrial [Mg-ATPf]/[MgADPf] was calculated from the equilibrium constants for the reactions and

These measurements utilized phosphoenolpyruvate carboxykinase which is present in the matrix of mitochondria from the livers of guinea pigs, chickens, and pigeons. Mitochondria isolated from these sources also contain nucleoside diphosphate kinase, malate dehydrogenase, and glutamate dehydrogenase or 3-OH-butyrate dehydrogenase. Together these enzymes catalyze the synthesis of phosphenolpyruvate and CO, from oxaloacetate with oxidative phosphorylation as an energy source. These reactions have been shown to be fully reversible in suspensions of mitochondria isolated from the above sources. With oxidative phosphorylation as the source of ATP, phosphoenolpyruvate was synthesized from malate and conversely addition of phosphoenolpyruvate, ADP, and COz led to synthesis of malate and ATP. The forward and reverse reactions were allowed to continue until the rate of change of metabolite concentrations was minimal and then the latter were measured. The intramitochondrial [Mg-ATPf]/[MgADPf] was calculated from the equilibrium constants for the reactions and the measured steady state concentrations of the metabolites in both the intra-and extramitochondrial spaces. The value of the intramitochondrial [MgATPt]/[MgADPt] was found to exceed the extramitochondrial value (adjusted to the same free Mg2' concentration) by a factor (2S.E.) of 0.83 f 0.22 (n = 17) for the forward reaction and 1.81 f 0.54 ( n = 11) for the reverse reaction. It is concluded that the adenine nucleotide translocase catalyzes electroneutral exchange of ATP for ADP and that this reaction does not contribute significantly to the energetics of mitochondrial oxidative phosphorylation.
In eukaryotic cells synthesis of ATP from ADP and inorganic phosphate (Pi) occurs in the mitochondrial matrix while most of the ATP utilization takes place in the cytosol. Thus the process of providing ATP for cellular function may be divided into two steps, ATP synthesis per se and translocation *This work was supported by grants GM 12202 and AM 25551 from the National Institutes of Health. 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. of ATP, ADP, and orthophosphate (Pi) across the inner mitochondrial membrane. The translocation of ATP and ADP is generally considered to be accomplished by a 1:l exchange catalyzed by a transmembrane protein, the adenine nucleotide translocase. This reaction has been studied extensively and it has been proposed that the translocator catalyzes an exchange of ATP4-for ADP3-, the electrical imbalance resulting in a coupling of the exchange to the transmembrane electrical potential (for review see Refs. [1][2][3]. Experimental support for this hypothesis is largely circumstantial and relies on three types of experimental observations. 1 Kinetic measurements indicate that ATP is preferentially exported while ADP is preferentially imported (11,12). 3. The distribution of adenine nucleotides across the membrane seems to be dependent on the K' gradient in the presence of valinomycin (13).
Evidence has been accumulating, however, that the data on which this hypothesis is based must be re-evaluated. First, the overall regulation (for review see Ref. 14) and energetics (see for example Refs. [15][16][17][18] of mitochondrial oxidative phosphorylation make it unlikely that there is energy loss in the translocation step although the current estimates of the membrane potential and the ratio of [ATP],/[ADP], to [ATPIi/[ADPIi would require such an energy loss. Second, the methods used for measuring the intramitochondrial adenine nucleotides may give rise to erroneous values for the intramitochondrial [ATP]J[ADP], (10). Third, our recent measurements indicate that the intramitochondrial ADP is preferentially bound to the mitochondrial contents and the intramitochondrial ratio of free nucleotides, [ATP,]/[ADPf], is much higher than that measured for the total nucleotides (10).
In the present paper, we report measurements of intramitochondrial [ATPf]/[ADPf] using the combination of two enzymatic reactions, P-enolpyruvate carboxykinase (EC 4.1.1.32) and nucleoside diphosphate kinase (EC 2.7.4.6). We show that the intramitochondrial [ATPf]/[ADPf] is much higher than that reported for the total nucleotides and is equal to or slightly greater than the comparable extramitochondrial [ATP]/[ADP] under conditions of active phosphorylation of extramitochondrial ADP.
to use an enzyme present in the mitochondrial matrix which catalyzes a reversible reaction utilizing ATP and ADP. Analysis is greatly facilitated if the reaction employs extramitochondrial reactants and generates extramitochondrial products since calculations of [ATP,]/[ADP,] can then be made from both the intramitochondrial and extramitochondrial reactants. Since the extramitochondrial metabolites are in dilute aqueous solution while the intramitochondrial metabolites may be bound to intramitochondrial enzymes, calculations using the metabolite concentrations in the extramitochondrial compartment are less likely to be subject to errors caused by binding phenomena. In addition, the intramitochondrial metabolites may change during quenching of the mitochondrial suspension since they are in the presence of very high enzyme concentrations while the extramitochondrial metabolites are much less susceptible to very rapid enzymatic changes. If the calculations based on intra-and extramitochondrial metabolites agree, then it is very unlikely that significant error arises from either of these sources.
A survey of the intramitochondrial enzymes identifies only one, P-enolpyruvate carboxykinase, which is suitable for this purpose at the present time. This enzyme constitutes an important part of gluconeogenesis (19,20) and may be localized either in the cytosol or mitochondria, depending on the animal species (21). In addition, the reaction has been found to be reversible in intact rat liver cells (22) and thus in vitro reversal would not require unphysiological or extreme experimental conditions. The reaction can also be carried out under conditions such that the mitochondria catalyze conversion of extramitochondrial reactants to extramitochondrial products.
The There are two aspects of this reaction which must be overcome by making use of other enzymes; it is specific for GTP and GDP instead of ATP and ADP, and oxaloacetate cannot be measured reliably by direct analysis because of its very low concentration in the mitochondrial matrix. The P-enolpyruvate carboxykinase reaction can be coupled to the adenine nucleotides through nucleoside diphosphate kinase which catalyzes the transformation: Guanine nucleotides are good substrates for the enzyme but other nucleotides are also active. Nucleoside diphosphate kinase has been reported to be localized in both the matrix and intermembrane spaces in mitochondria from rat liver (24). The function of this enzyme in the matrix space has been demonstrated in mitochondria from rabbit liver (25), guinea pig liver (Refs. 26 and 27 and this work) as well as pigeon and chicken liver (this work) where oxidative phosphorylation can support high rates of P-enolpyruvate and ATP synthesis. If experimental conditions are established in This magnesium-GTP and magnesium-GDP complexes in solution exist primarily as the MgGTP2-and MgGDP-complexes. Since small fractions are present as MgGTP-, MgGTP3-, MgGDP2-, and MgGDPO, MgGTP and MgGDP will be used to designate the metal chelates of GTP and GDP. Likewise, the magnesium chelates of ATP and ADP will be designated as MgATP and MgADP. For similar reasons, carboxylic and phosphoryl-containing compounds are expressed without their charges. which nucleoside diphosphate kinase is the only significant source of GTP, i.e. GTP is generated indirectly by oxidative phosphorylation and not directly by substrate level phosphorylation, we can make use of the fact that the equilibrium constant for Equation 3 is 1.0 (10, 23) and write [

K7
X - (9) where K5, K,, and K7 represent the equilibrium constants for  . Thus neither calculation of the oxaloacetate concentration nor the P-enolpyruvate carboxykinase equilibrium is sensitive to a difference in pH between the intra-and extramitochondrial spaces. The intramitochondrial pH is important to the calculations primarily in its effect on glutamate dehydrogenase through the distribution of NH4+. The ammonium ion will be excluded to the extent that the mitochondrial matrix is more alkaline than the suspending medium ( [NH4']i = [NH,+]o/antilog(pHi -pH,)).
Selection of Suitable Species for Preparation of Liver Mitochondria-Although gluconeogenesis and P-enolpyruvate carboxykinase are common to the livers of many animals the subcellular distribution of the enzyme is quite distinctive. In the livers of some animals such as the rat, mouse, and hamster, it is entirely cytosolic while in guinea pig liver it is both in the mitochondria and the cytosol. In the liver of rabbits and a t least some birds (pigeons, chickens), P-enolpyruvate carboxykinase is predominantly, 90% or greater, mitochondrial with very little activity in the cytosol (29)(30)(31). In the present study it was necessary to use mitochondria with high Penolpyruvate carboxykinase activity; hence chicken and pigeon livers were utilized. In addition to P-enolpyruvate carboxykinase, however, three other enzymes are needed for the complete assay system, malate dehydrogenase, nucleoside diphosphate kinase, and an enzyme for measurement of the intramitochondrial [NAD+]/[NADH] (3-OH-butyrate dehydrogenase or glutamate dehydrogenase). By also using guinea pig liver mitochondria it was possible to measure [NAD']/ [NADH] from equilibrium in two different enzymes, 3-OHbutyrate dehydrogenase (guinea pig liver) and glutamate dehydrogenase (chicken and pigeon liver). [MgADP,I.

Methods
Determination of K4 for P-enolpyruvate Carboxykime-P-enolpyruvate carboxykinase (EC 4.1.1.32) was purified from rat liver cytosol (32). When MgGTP and MgGDP are used as substrates, the enzyme gives increased catalytic rates in the presence of micromolar Mn2+, which was, therefore, included in the assay mixtures. Since the equilibrium constant is a thermodynamic constant, it is independent of enzyme source. The equilibrium constant for the P-enolpyruvate carboxykinase reaction was determined by a modification of the initial rate method described previously (33). Reaction mixtures contained 50 mM Hepes2-K+ of the desired pH, 120 mM KCl, 2 mM free MgClz, 10 p M MnC12, 50 p~ GTP, 1 mM NAD' , 5 mM malate, and 6 units/ ml of malate dehydrogenase as components of the forward reaction (P-enolpyruvate formation). Components of the reverse reaction were added as a stock solution (pH 8) containing approximately 100 mM NaHC03, 10 mM P-enolpyruvate, and 10 mM GDP. Additional MgClz was added to the reaction mixtures to give total MgCl, = 2 mM + GDP. Free Mg2f of 2 mM maintained essentially all of the GTP and GDP as the M$+ chelates. The final volume of the reaction mixtures was 1.0 ml. The pH of reaction mixtures was tested directly after addition of all reaction components. There was no change in pH during the time required to complete the rate measurements. Stock solutions of P-enolpyruvate, GDP, and GTP were standardized by enzymatic assays. Oxaloacetate was generated in the reaction mixtures from malate, NAD+, and malate dehydrogenase and its concentration calculated from the absorbance change at 340 nm. The reaction was initiated by addition of malate dehydrogenase to a reaction mixture which had been warmed to the desired temperature. When the change in absorbance at 340 nm became stable P-enolpyruvate carboxykinase was added and the further change in absorbance monitored at the same wavelength. The addition of P-enolpyruvate carboxykinase was approximately 4 min after the addition of the HC03stock solution. Since initial rates of oxaloacetate formation or removal were measured, there was no significant alteration in the COz equilibrium as a consequence of the measurement. In every K4 determination, the ratio of substrates and products was varied over a range which allowed the initial reaction to both produce and consume oxaloacetate. The equilibrium constant was calculated from Equation 2. The COz concentration was calculated from the Henderson-Hasselbalch equation using the measured pH, the concentration of added bicarbonate, and a pK. of 6.1. To ensure equilibration of COZ and HC03-prior to addition of P-enolpyruvate carboxykinase, reaction mixtures in some experiments contained 20 or 100 units/ml of carbonic anhydrase (EC 4.2.1.1).
Mitochondrial Preparation-Mitochondria were prepared from either the livers of 48-h starved adult guinea pigs or the livers of chickens and pigeons essentially according to the method of Schneider (34). The tissue was homogenized in cold mannitol (0.22 M), sucrose (0.075 M) medium containing 0.2 mM EGTA at pH 7.2. The mitochondria were suspended at a protein concentration of 50-80 mg/ml in the isolation medium.
Incubation Conditions-Incubations were carried out by diluting the mitochondria to 8-12 mg of protein/ml in a medium consisting of the isolation medium diluted 20% with water and containing the various substrates and reactants as well as 25 mM NaHC03. This medium was equilibrated with a 95% 02:5% COZ gas mixture and after the addition of the mitochondria the gas phase was again thoroughly exchanged with 95% 02:5% 0 2 and the flasks sealed. Incubation of 2-ml samples in 25-ml Erlenmeyer flasks was carried out in a Dubnoff metabolic shaker water bath at the indicated temperature (30-35 "C), at approximately 80 cycles/min. At the end of the incubation period samples were quenched for total metabolite assay by pouring the contents into a 25-ml Erlenmeyer flask containing 0.66 ml of 24% perchloric acid (final 6% perchloric acid), precooled in a salt-ice bath to -3 to -6 "C and being continuously stirred by a magnetic stirrer. After 5 min at 0-4 "C the samples were centrifuged to remove the precipitated protein. The clear supernatant was then neutralized as given below. The intramitochondrial metabolites were measured from separate but parallel incubations by centrifuging the mitochondria from aliquots of the incubation mixture through a layer of silicone oil (specific gravity 1.05) into a 4% perchloric acid, 3% NaCl solution. After centrifugation for 70-90 s in a Beckman model B microfuge, the top and silicone layers were removed and the protein pellet on the bottom layer was thoroughly mixed with the perchloric acid solution using a glass stirring rod. The tubes were recentrifuged and aliquots of the clear supernatant collected. The perchloric acid extracts from both the total and mitochondrial samples were neutralized to pH 6.5-7.0 with a 3 M KzC03-0.5 M triethanolamine base mixture, the KClO, precipitate was removed by centrifugation, and the metabolite assays were carried out on the neutralized samples.
All assays were completed on the same day the experiment was run.
The total content of metabolites in the medium was used to calculate the extramitochondrial concentrations after separate experimental measurements of the supernatant medium showed its content to be experimentally indistinguishable from that of the total. The measured intramitochondrial water volume was always less than 2% of the total incubation volume. Mitochondrial Water Volume-At the end of each incubation and at the same time samples were being quenched for metabolite assays, 1-ml fractions of the incubation mixture were treated with 3H20 and "C sucrose for 40 s. Then 0.25-ml aliquots were added to 400-pl polypropylene microfuge tubes containing 20 pl of silicone oil (specific Materials The enzymes used for metabolite assays were obtained from Sigma except for pyruvate kinase and lactate dehydrogenase (Boehringer Mannheim) and carbonic anhydrase (Worthington). All other chemicals were of the highest quality available commercially.

RESULTS
Equilibrium Constant for P-enolpyruvate Carboxykinase-The variation of initial reaction rate for P-enolpyruvate carboxykinase as a function of substrate and product concentration is shown in Fig. 2. When the products COS, MgGDP, and P-enolpyruvate are absent, or present at low concentrations, * Equilibrium constants measured at 37 "C did not contain carbonic anhydrase since its presence was not required to attain equilibrium of bicarbonate and CO, at 30 "C. e Measurements were made in the absence or presence of carbonic anhydrase (20 units/ml). The addition of carbonic anhydrase had no significant effect on the equilibrium titration curves.
Measurements were made in the presence of 100,20, and 0 units/ ml of carbonic anhydrase. Carbonic anhydrase had no significant effect on the equilibrium titration curves. the initial reaction proceeds in the direction of oxaloacetate decarboxylation. At higher product concentrations, the reverse reaction becomes predominant and increases the oxaloacetate concentration. This demonstrates the reversibility of the enzyme with excess free Mg2' and MgGTP and MgGDP as the nucleotide substrates. The Key values for 30 "C were calculated from the results of Fig. 2 and additional experiments and the values at 37 "C from experiments carried out at t h e latter temperature (Table I). Table I indicates that the Key value varies within relatively narrow limits of 0.11 t o 0.29 M as pH is varied from 7.0 to 7.9 a t both 30 and 37 "C. Temperature has only a small effect on the equilibrium constant and in the present work a value of 0.20 was used for the calculations. The values summarized in this table can be compared with the K,, of 3.2 M previously reported (33) for rat liver P-enolpyruvate carboxykinase at 37 "C and pH 8.0 but using Mn2+ as the free metal ion, Mn nucleotides as substrates, and expressed in terms of HCOs-concentration. When expressed in terms of CO, concentration, equivalent to those of Table I, the Keq is 0.04 M, in reasonable agreement with the data reported here, considering the different stability constants for Mg2+ and MnZ+ nucleotide complexes.

The Synthesis
and Degradation of P-enolpyruvate by Chicken and Pigeon Liver Mitochondria-Chicken liver and pigeon liver mitochondria have very similar abilities to synthesize P-enolpyruvate from malate and to utilize P-enolpyruvate to make malate and ATP. Two typical experiments in which both the forward (synthesis of P-enolpyruvate) and reverse (synthesis of malate and ATP from P-enolpyruvate and CO,) reactions were measured are given in Table 11. Incubations were carried out for 15 min since preliminary experiments established that the reactions attained their steady state values within this interval at 30 "C and no further changes occurred when the incubations were extended to 30 min. In the forward reaction there was substantial synthesis and export of P-enolpyruvate as indicated by the appearance of 0.5-1.5 mM P-enolpyruvate in the extramitochondrial medium. The reverse reaction gave rise to even higher levels of malate synthesis with the extramitochondrial malate rising to between 1.5 and 2.6 mM.
It was noted in preliminary experiments that both chicken and pigeon liver mitochondria rapidly metabolize NH,' and this must be inhibited if more than a few micromolar ammonia is to remain in the suspension. The disappearance of NH,+ was inhibited by amino-oxyacetic acid, a transaminase inhibitor, and hence 1-3 mM aminooxyacetic acid was routinely added when glutamate dehydrogenase was used to establish the intramitochondrial [NAD+]/[NADH].
As noted under "Rationale," it was necessary to limit possible synthesis of intramitochondrial GTP by substrate level phosphorylation, which could occur by %ketoglutarate dehydrogenase. This enzyme is reported to be strongly inhibited by sodium arsenite, with 50% inhibition requiring less than 50 PM arsenite (47). In separate experiments it was established that the stimulation of respiration induced by addition of 2ketoglutarate to suspensions of substrate-depleted mitochondria (with ADP and Pi) was indeed blocked by arsenite with 0.25 mM giving essentially 100% inhibition. We have routinely included 0.5 mM sodium arsenite in the incubation medium and found that there was no significant difference between 0.25 and 0.5 mM, consistent with complete inhibition of 2ketoglutarate dehydrogenase by 0.25 mM arsenite. (See Table  111 for experiments with pigeon liver mitochondria in which the arsenite concentration was varied from 0.25 to 0.5 mM.) The Synthesis and Degradation of P-enolpyruvate by Guinea Pig Liver Mitochondria-Guinea pig liver mitochondria contain 3-OH-butyrate dehydrogenase activity (see Table VI1 TABLE I1  and "Discussion") and show respiratory control with 3-OHbutyrate as substrate. Addition of 3-OH-butyrate with malate gave very low rates of P-enolpyruvate synthesis and relatively high rates of oxidation of 3-OH-butyrate while addition of acetoacetate gave relatively high rates of P-enolpyruvate synthesis and acetoacetate was reduced to 3-OH-butyrate (data not shown). Substantial P-enolpyruvate synthesis with minimal changes in the [3-OH-butyrate]/[acetoacetate] was observed for ratios of 3 to 10. We, therefore, used these latter conditions for our experiments. When suspensions of guinea pig liver mitochondria were incubated under conditions for P-enolpyruvate synthesis, the increase in extramitochondrial P-enolpyruvate was observed over a relatively long time period (see for example Table IV and Refs. 26 and 27). The intramitochondrial [P-enolpyruvate] first rose rapidly and then slowly continued to increase, rising from 1.1 mM at 20 min to 1.8 mM at 35 min. The rate of P-enolpyruvate synthesis was, in our hands, quite variable from animal to animal and in general adult animals gave higher rates than younger ones. We did not attempt to explore further these differences but used only adult animals for the experiments reported here.

The synthesis and degradation of P-enolpyruvate by mitochondria prepared from chicken liver and pigeon liver
Addition of arsenite stimulated the synthesis of P-enolpyruvate by guinea pig liver mitochondria, presumably because it increased oxidation of 3-OH-butyrate and thereby increased the intramitochondrial  Table V).
Guinea pig liver mitochondria also catalyzed the reverse reaction, which, like the forward reaction, was slower than in mitochondria from the livers of chickens and pigeons. A comparison of the forward and reverse reactions can be made from the data in Table VI. The amount of malate synthesized and exported to the medium was relatively small, less than 1 mM, but the intramitochondrial levels were much higher, between 6 and 18 mM.

TABLE VI Synthesis and degradation of P-enolpyruvate by guinea pig liver mitochrondria
The mitochondria were incubated as described under "Methods." The media for the forward reactions contained 8 mM malate, 14 mM (A) or 6 mM (B) 3-OH-butyrate (3-OH-B), 1 mM ATP, 0.2 mM ADP, 2 mM orthophosphate, and 0.5 mM arsenite. The media for the reverse reaction were the same except they contained 7 mM 3-OH-butyrate, 1 mM acetoacetate (Acac), 0 mM malate, and 6 mM P-enolpyruvate. In experiment A the incubation temperature was 35 "C while in experiment B it was 30 "C. All   These activities were calculated from values given per g of tissue, wet weight, by assuming 60 mg of mitochondrial protein per g of tissue.
'This is a review which contains the data given. The reader is referred to the review for the original references. P-enolpyruvate + ox-

DISCUSSION
Understanding of the interrelationships of mitochondrial and cytosolic metabolism has proven extraordinarily controversial. This is partially attributable to the spatial separation of the two compartments which necessitates the operation of many vectorial translocation systems across the inner mitochondrial membrane. These transport systems generally exchange one compound for another of similar charge in order to maintain electrical and osmotic balance, but in each case there has been the possibility that the transport could be coupled to the transmembrane electrical and/or ion gradients to actively transport the compound. The variety of possible mechanisms, coupled with an inability to account quantitatively for all of the species crossing the membrane in a given experiment, has led to many different views of the intra-and extramitochondrial metabolic relationships. This is particularly true concerning the role of the adenine nucleotide translocase in cellular energy metabolism (for review see Refs. 1-3 and 14).
The Activities of Nucleoside Diphosphate Kinase, P-enolpyruvate Carboxykinase, and 3-OH-butyrate Dehydrogenase in Relationship to the Measured Approach to Equilibrium-In order for an enzymatic system to approach equilibrium in a reasonable time the capacities of the individual enzymes of the system have to be substantially higher than the changes in metabolite concentrations which are required to bring about equilibrium. We have relied in our studies on the activities of 5 different enzymes: malate dehydrogenase, nucleoside diphosphate kinase, P-enolpyruvate carboxykinase, 3-OH-butyrate dehydrogenase (guinea pig liver), and glutamate dehydrogenase (pigeon and chicken liver). Of those, malate dehydrogenase clearly has high enough activity to fulfill the above conditions (Table VII). The same is true of glutamate dehydrogenase in avian mitochondria, whereas the other three enzymes deserve some discussion.
Under our conditions, guinea pig liver mitochondria synthesize 0.3-0.9 mM P-enolpyruvate in 30 min (30 "C and 10 mg of protein/ml). The activity of P-enolpyruvate carboxykinase has been reported to be about 30 nmol/mg of protein/ min (Table VII) which means that the enzyme can synthesize 9 mM P-enolpyruvate in 30 min in our conditions. This activity is more than 10-fold greater than the amount actually produced. The same applies to nucleoside diphosphate kinase activity which is 90 nmol/mg of protein/min at 37 "C or about 45 nmol/mg of protein/min a t 30 "C. In agreement with the suggestion that nucleoside diphosphate kinase activity is sufficiently high to allow efficient utilization of ATP for Penolpyruvate synthesis are the observations of previous authors (Refs. 21,27,and 48; for review see Ref. 49). For example Bryla et al. (48) showed that in the presence of fluorocitrate (which blocks the citric acid cycle activity and thereby synthesis of GTP by substrate level phosphorylation) synthesis of P-enolpyruvate from ATP via nucleoside diphosphate kinase was 10-15%, or 2-3 nmol/mg of protein/min, of the rate observed when GTP was directly synthesized by oxidation of 2-ketoglutarate. This rate is in good agreement with the value found in our experiments which were carried out with arsenite as the inhibitor of substrate level phosphorylation. The agreement between these measurements further means that the rate of the GTP synthesis through substrate level phosphorylation is, under such conditions, minimal.
The activities of P-enolpyruvate carboxykinase and nucleoside diphosphate kinase in pigeon liver mitochondria are 4fold and 10-fold higher, respectively, than in guinea pig liver mitochondria, and this is consistent with the more rapid equilibration which is observed in the former system.
The activity of 3-OH-butyrate dehydrogenase in guinea pig liver mitochondria is 17-40 nmol/mg of protein/min at 30 "C, again 10-fold higher than the rate of P-enolpyruvate synthesis observed in our experiments. The addition of 3-OH butyrate alone inhibits and that of acetoacetate stimulates synthesis of P-enolpyruvate as expected if the former reduces and the latter oxidizes the intramitochondrial NAD couple. Hence the use of 3-OH-butyrate dehydrogenase to indicate the intramitochondrial [NAD']/[NADH] and thereby the concentration of oxaloacetate seems to be justified. Finally, it should be noted that as the steady state is approached the calculated mass action ratio approaches the equilibrium constant irrespective of 1) whether the reaction is carried out in the direction of P-enolpyruvate synthesis or its degradation, 2) whether 3-OH-butyrate dehydrogenase or glutamate dehydrogenase is used to calculate the oxaloacetate concentration, and 3) the absolute activities of nucleoside diphosphate kinase and P-enolpyruvate carboxykinase since these are 4-to 10-fold higher in pigeon liver than in guinea pig liver mitochondria. This behavior strongly supports our conclusion that the reactions are nearing equilibrium.
Relationship    All of the values designated with asterisks in the "Ratio" column were used in calculating the mean f S.E. values given in the text. In the forward and reverse reaction for pigeon and chicken liver mitochondria all data were used except those with very high NH4+ from Table IIA  arise from the same factor, i.e. preferential binding of M e to nucleotide triphosphates. Therefore, the mutual relations between the two ratios will remain the same at any magnesium concentration. Comparison of the two values and consequently the conclusions presented in this paper are not significantly dependent on the actual intramitochondrial The experiments are carried out using mitochondria suspended in a medium containing 0.2 mM EGTA and with no added Mg2+ to minimize ATPase activity. In order to compare directly the intra-and extramitochondrial adenine nucleotide ratios it is necessary that they be calculated for the same concentration of free M e by correcting for the 6.5-fold difference in M e binding by ADP and ATP under approximately physiological conditions (57) (4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14). This is an observation analogous to that for the cytoplasm of muscle cells where the [Mg*+I.
[MgATPf]/[MgATPf] values calculated from the creatine phosphokinase reaction greatly exceed the measured total [ATP]/[ADP] (see for example Ref. 59). In muscle this is due to preferential binding of ADP to myosin (60, 61). An analogous ADP binding system may be present in the mitochondria due to tight binding of ADP to F,-ATPase (62), ADP-ATP translocase (1)(2)(3), and other enzymes. Measurements of the [ATP]/[ADP] ratio in concentrated mitochondrial lysates are consistent with the existence of a specific high affinity binding site of ADP with a concentration of approximately 2-fold greater than that of cytochrome a (10). The presence of such a binding site would mean that for a measured total [ATP]/ [ADP] of 10, more than 90% of the ADP may be bound, giving [ATPf]/[ADPf] values in excess of 100. The exact value of the latter is dependent on both the binding affinity and the stoichiometry of the binding site. In mitochondria there is also evidence that in many cases the measured total intramitochondrial [ATP]/[ADP] ratios are less than the correct values due to hydrolysis of ATP which occurs during the process of quenching the reaction mixture (10). The combined effect of ATP hydrolysis duing sample quenching and the presence of specific high affinity ADP binding sites can account for the reports that the measured intramitochondrial [ATP] ,/[ADP] is much lower than the intramitochondrial If, as our measurements indicate, the intramitochondrial [ATPf]/[ADPf] is equal to or slightly greater than the extramitochondrial [ATP]/[ADP], this not only removes the thermodynamic arguments for coupling of adenine nucleotide transport to the membrane potential but also makes such a coupling inconsistent with the observed reversibility of the partial reactions of oxidative phosphorylation (excluding oxygen reduction). In this case electrogenic exchange of ATP4for ADP3-across a membrane with a transmembrane electrical potential of 120-150 mV (negative inside) would lead to a loss of free energy of 2.7-4.1 kcal/mol of ATP transported outward. This loss in energy would make the ATP-driven reversed electron transfer very difficult to observe, in marked contrast to the experimental results.
Many aspects of cellular metabolism can be more readily understood if the mitochondrial [MgATPf]/[MgADPf] is, as our measurements indicate, equal to or slightly greater than the cytoplasmic [MgATPf]/[MgADPf]. This would make it reasonable that in guinea pig liver synthesis of P-enolpyruvate for gluconeogenesis could equally well occur in the cytoplasmic compartment, in the mitochondrial compartment, or in both compartments at the same time (63). In addition, it is consistent with the observation, both in vivo and in vitro, that the respiratory rate is dependent on extramitochondrial [ATP]/[ADP] [Pi] and that the overall reactions of the first two sites of oxidation are near equilibrium (15)(16)(17)(18).
In Summary-Advantage has been taken of the intramitochondrial localization of the enzyme P-enolpyruvate carboxykinase in guinea pig liver, chicken liver, and pigeon liver. This enzyme reaction has been shown to be freely reversible in suspensions of isolated mitochondria from the above sources. When the forward and reverse reactions attained steady state values the data were used, in conjunction with the measured equilibrium constant, to calculate the intramitochondrial [