turnover in Saccharomyces cerevisiae

Proc. Natl. Acad. Sci. USA Vol. 93, pp. 6399-6404, June 1996 Biochemistry

31p NMR magnetization transfer study of the control of ATP turnover in Saccharomyces cerevisiae (yeast/mitochondria/P:O ratio)

JONATHAN G. SHELDON, SIMON-PETER WILLIAMS, ALEXANDRA M. FULTON, AND KEVIN M. BRINDLE* Department of Biochemistry, University of Cambridge, Tennis Court Road, Cambridge CB2 lQW, United Kingdom

Communicated by Robert G. Shulman, Yale University, New Haven, CT, January 22, 1996 (received for review June 8, 1995)

31P NMR magnetization transfer measureABSTRACT ments have been used to measure the steady state flux between P1 and ATP in yeast cells genetically modified to overexpress an adenine nucleotide translocase isoform. An increase in Pi -> ATP flux and apparent ratio of moles of ATP synthesized/ atoms of oxygen consumed (P:O ratio), when these cells were incubated with glucose, demonstrated that the reactions catalyzed by the translocase and F1Fo ATP synthase were readily reversible in vivo. However, when the same cells were incubated with ethanol alone, translocase overexpression had no effect on the measured Pi -> ATP flux or apparent P:O ratio, suggesting that the synthase was now operating irreversibly. This change was accompanied by an increase in the intracellular ADP concentration. These observations are consistent with a model proposed for the kinetic control of mitochondrial ATP synthesis, which was based on isotope exchange measurements with isolated mammalian mitochondria [LaNoue, K. F., Jeffries, F. M. H. & Radda, G. K. (1986) Biochemistry 25, 7667-7675].

tochondria and rat hepatocytes have shown that the P:O ratio also depends on the rate of respiration, being close to zero at low rates of respiration and approaching its maximal value at high rates of respiration (16-18). These latter studies undermine the value of the P:O ratio as an indicator of the irreversibility of the ATP synthase reaction in NMR magnetization transfer measurements. Both studies suggest that an ATP < Pi exchange reaction, catalyzed by the mitochondria, could make a significant contribution to the P1 -> ATP flux measured using magnetization transfer techniques at relatively low rates of respiration. We have investigated this possibility by measuring the Pi -> ATP flux in yeast cells overexpressing a mitochondrial adenine nucleotide translocase isoform. The kinetic properties of the translocase have been suggested to limit the rate of ATP reentry into the mitochondria and, thus, the rate of an ATP < P1 exchange reaction (15). If this is the case and there is an exchange contribution to the observed Pi -* ATP flux, then the measured flux should increase in cells overexpressing the translocase in the absence of any change in oxygen consumption. A preliminary account of related work has been published (8).

31P NMR magnetization transfer techniques have been used to measure directly the rate of ATP turnover in a variety of systems, including Escherichia coli (1, 2), yeast (3-8), rat brain (9), and the perfused rat heart (10-12). A primary objective of these studies, particularly those on the heart, has been to determine the extent to which the F1Fo ATP synthase is displaced from equilibrium and, thus, to assess its potential for exerting kinetic control over flux in mitochondrial oxidative phosphorylation (13). These measurements have also been used to determine the degree of mitochondrial coupling, as reflected by the apparent ratio of moles of ATP synthesized/ atoms of oxygen consumed (P:O ratio) (14). Previous magnetization transfer measurements of Pi -- ATP flux in vivo, in which the contribution of the glycolytic enzymes glyceraldehyde-3-phosphate dehydrogenase (GAPDH; EC 1.2.1.12) and phosphoglycerate kinase (PGK; EC 2.7.2.3) to this measured flux has been removed (see Fig. 1), have produced values for the apparent P:O ratio of between 2 and 3 (6, 12). The similarity of these values to those measured with isolated mitochondria catalyzing net ATP synthesis in vitro was taken to indicate that the reaction catalyzed by the mitochondrial ATP synthase is effectively irreversible in vivo and that the magnetization transfer experiment is measuring net ATP turnover. However, isotope exchange measurements on isolated rat heart and liver mitochondria have shown that the unidirectionality of ATP synthesis depends on the rate of respiration (15). At high rates of respiration, the reaction was unidirectional, whereas at low rates, it was near-toequilibrium. This study suggested that an exchange reaction between P1 and ATP should be observable in NMR magnetization transfer measurements in the perfused rat heart at low workloads (15). Determination of P:O ratios in isolated mi-

MATERIALS AND METHODS Materials. Yeast growth media were obtained from Difco. Low-gelling-temperature agarose, Mes, N-(2-hydroxyethyl)piperazine-N'-(2-ethanesulfonic acid) (Hepes), carboxyatractyloside, NADH, D-sorbitol, D-mannitol, and a glucose assay kit using glucose oxidase and o-dianisidine were obtained from Sigma. Triethanolamine hydrochloride, 3-phosphoglycerate, GAPDH, ADP, and ATP were obtained from Boehringer Mannheim. Protein concentrations were determined using the method of Bradford (19) with a kit obtained from Bio-Rad. Bovine serum albumin was used as the standard. Glusulase was obtained from DuPont. Preparation of Cells for NMR Experiments. Cells of the Saccharomyces cerevisiae strain BC3 (his 3-15, leu 2-3, leu 2-112, trp 1-1, ura 3-52pgk::TRPJ) (20), transformed with the single copy plasmid pYC775, were kindly provided by Peter Piper (University College, London). The plasmid pYC775 contains the yeast PGK gene with a partially deleted promoter (21). These cells were transformed by the method of Hinnen and coworkers (22) with one of two multicopy plasmids carrying the LEU 2 gene. The adenine nucleotide translocase expression vector, pBF3, was constructed by inserting the 2.6-kb BamHI fragment from YEp 6-19-28 (23), containing the gene encoding the translocase isoform AAC1 and its promoter, into the BamHI restriction site of the vector pMA 3a (24). The vector pMA 3a, lacking the translocase gene and its promoter, served as a control plasmid. Abbreviations: GAPDH, glyceraldehyde-3-phosphate dehydrogenase; P:O ratio, ratio of moles of ATP synthesized/atoms of oxygen consumed; PGK, phosphoglycerate kinase; NDP, nucleoside diphosphate; NTP, nucleoside triphosphate. *To whom reprint requests should be addressed.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

6399

6400

Proc. Natl. Acad. Sci. USA 93

Biochemistry: Sheldon et al.

Cells for NMR experiments were grown aerobically at 30°C on a synthetic medium containing 6.7 g/liter of yeast nitrogen base, 2% (wt/vol) glucose, and a mixture of 0.002% (wt/vol) histidine and 0.002% (wt/vol) uracil. When the cells had reached late-log phase growth, approximately 2.5 x 107 cells/ ml, they were transferred to a medium containing 2% (wt/vol) Bacto-Peptone, 1% (wt/vol) yeast extract, and 2% (wt/vol) glucose at a cell density of 4 x 105 cells/ml. The cells were harvested 24 h later and washed twice at 4°C in a buffer containing 2 mM MgSO4, 1.7 mM NaCl, 2 mM KCI, and 50 mM Mes at pH 6. Cells (4 g) were mixed with 8 ml of a solution of 1.8% wt/vol low-gelling-temperature agarose, and the mixture extruded to form fine agarose gel threads that entrapped the cells, as described (4, 25). The immobilized cells were perifused in the NMR spectrometer at '50 ml/min with 2 liters of the above mentioned buffer, containing glucose and/or ethanol where stated. This buffer was maintained at a temperature of 30°C and sparged with oxygen. Oxygen consumption was estimated by measuring the dissolved oxygen concentration in the influent and effluent buffer. Comparison of the oxygen consumption rates displayed by the cells when immobilized and when in dilute suspension showed that they were not oxygen-limited in the agarose gel threads at this cell density (data not shown). After perifusion, cell extracts were prepared by perchloric acid extraction as described (25, 26). NMR Measurements of Metabolite Concentrations in Cell Extracts. Lyophilized samples were reconstituted in 50 mM triethanolamine buffer, pH 8.0, containing 5 mM EDTA and 20% (vol/vol) D20, for a field-frequency lock. Spectra were acquired at a 31P NMR frequency of 161 MHz and were the sum of 10,000 scans, collected into 8000 data points with a 300 pulse, a 4.6-s interpulse delay, and a sweep width of 10 kHz. WALTZ proton decoupling was applied during the acquisition period and the sample temperature was maintained at 30°C. The resonances were assigned by comparison with previously published spectra (27). Metabolite concentrations were determined as described (25, 26). NMR Measurements on Cells. Magnetization transfer experiments were performed and analyzed as described (7). Spectra in which the ATP y-phosphate resonance was irradiated were alternated with those in which a control irradiation was applied. A total of 128 scans (in blocks of 16) were collected at the two irradiation frequencies. The inversion recovery T1 measurements on the cytosolic Pi resonance involved measuring the Pi z magnetization at 7 different delay values after its inversion. The delays ranged from 0.0625 s to 4 s. Each spectrum was the sum of 96 scans (collected in blocks of 16 scans) with a delay of 5 s between the end of the acquisition period and the next 1800 pulse. The resonances were assigned by comparison with previously published spectra (27). Metabolite concentrations were determined as described above except that resonance intensities were determined by weighing peaks cut from plotted spectra. The intracellular pH was estimated from the frequency difference between the methylenediphosphonate resonance and the intracellular Pi resonance (4, 7). The ratio of Mz, the steady state Piz magnetization measured in the presence of selective saturation of the y-phosphate resonance of ATP, to the equilibrium Pi z magnetization, Mo, measured in the presence of the control irradiation, is given by Mz/Mo = 1/(1 + k Tip), where k is a first-order rate constant describing loss of z magnetization from Pi due to chemical exchange and T1p is the spin lattice relaxation time of the phosphorus nucleus in Pi. The relaxation rate of the Pi z magnetization in the inversion recovery experiment, in the presence of selective saturation of the y-phosphate resonance of ATP, is given by 1/Tip + k. A combination of these two experiments yields a value for k and T1p. Multiplication of k by the Pi concentration gives the Pi -> ATP flux (see Table 2).

(1996)

Enzyme and Metabolite Assays. Glucose consumption and ethanol production were determined by sampling the cell perifusate. Ethanol concentrations were assayed using enzymatic methods as described (28). PGK was assayed spectrophotometrically in cell extracts as described (7, 29). The adenine nucleotide translocase was assayed by titration of its activity using the quasi-irreversible inhibitor carboxyatractyloside. Yeast mitochondria were isolated as described (30) and incubated at a protein concentration of 0.2 mg/ml in the following basal medium: 10 mM Tris maleate/0.65 M mannitol, pH 6.7, comprising 0.5 mM ADP and 1 mM NADH. The reaction was started by the addition of Pi to a final concentration of 7.7 mM. Mitochondrial oxygen consumption, which was measured polarographically at 27°C using a Clark-type oxygen electrode (Hansatech Instruments, Ltd., Kings Lynn, U.K.), was titrated with a carboxyatractyloside solution (31). Oxygen Consumption Measurements on Cell Suspensions. Oxygen consumption rates were measured at 30°C using a Clark-type oxygen electrode. Cells were resuspended at a concentration of 1 mg of dry mass/ml in the same buffer that was used for the NMR experiments.

RESULTS The coupled reactions catalyzed by the glycolytic enzymes GAPDH and PGK can make a significant contribution to the

Pi -> ATP flux measured using magnetization transfer techniques (Fig. 1) (5-8, 12, 29). As these enzymes catalyze reactions that are near-to-equilibrium in the glycolytic pathway, the Pi -> ATP flux due to these enzymes may significantly exceed the net glycolytic flux (7, 12). This exchange was removed here by inserting a PGK gene, with a partially deleted promoter, into a single copy vector and transforming this into cells in which the chromosomal gene had been disrupted. These cells, which had -2% of wild-type PGK activity, had a glycolytic flux rate that was slightly lower than that observed at wild-type levels of PGK but showed no Pi ATP exchange due to the glycolytic enzymes (Table 1). Under anaerobic

P~~~~~~~~P > ~~~ADPAD I ~~~~Translocase nt ATP

F F ATP Synthase

ATP

Mitochondrion GAPDH

GAP +

NAD+

+

P.i

1,3-DPG + NADH + H PGK

1,3-DPG + ADP X 3PGA + ATP FIG. 1. Pathways for exchange between Pi and ATP that have been shown to be measurable in 31P NMR magnetization transfer experiments in vivo. GAP, glyceraldehyde 3-phosphate; 1,3-DPG, 1,3diphosphoglycerate; 3PGA. 3-phosphoglycerate.

Biochemistry: Sheldon et aL

Proc. Natl. Acad. Sci. USA 93 (1996)

conditions, the Pi -> ATP flux in these cells, determined using magnetization transfer techniques, was very similar to the net glycolytic flux through the triose phosphates, as estimated from measurements of glucose consumption. The rate of mitochondrial ATP synthesis under aerobic conditions can then be calculated from the total P1

->

ATP flux measured in

magnetization transfer experiment simply by deducting the contribution of net glycolytic flux. The results of these calculations are shown in Table 2. Overexpression of the mitochondrial adenine nucleotide translocase was achieved by transforming the cells with a multicopy vector (pBF3) that contained the gene for the AAC1 isoform of the carrier (23, 33). Measurements of oxygen consumption in mitochondria isolated from these cells, after titration of carrier activity with the quasi-irreversible inhibitor carboxyatractyloside, demonstrated that there was an approximately 2-fold overexpression of functional carrier compared with controls (Fig. 2). The low initial slopes of the titration curves are consistent with previous titrations of carrier activity, which demonstrated that it had a low flux control coefficient for respiration in isolated yeast mitochondria (31). Increased expression of the translocase resulted in signifia

cant increases in the

Pi

-> ATP flux in cells incubated with

glucose or glucose plus ethanol but not in cells incubated with ethanol alone (Table 2). Translocase overexpression had no effect, however, on oxygen consumption under all three substrate conditions or on the rate of glycolysis in those cells incubated with glucose. Therefore, overexpression of the translocase increased the apparent P:O ratio in cells incubated with glucose or glucose plus ethanol but not in cells incubated with ethanol alone. The increased oxygen consumption and decreased glucose consumption rates after addition of ethanol to cells incubated with glucose has been observed (34), where the stimulation of oxygen consumption was attributed to increased provision of NADH to the mitochondria. Incubation of cells with ethanol alone resulted in a relatively large increase in the ADP concentration, a decrease in the ATP concentration and a small, but significant, decrease in intracellular pH compared with cells incubated with glucose or glucose plus ethanol (Table 3). The apparent P:O ratios of 1.0 in control cells incubated with glucose or glucose plus ethanol (Table 1) are at the lower end of the range of values measured with isolated yeast mitochondria (35, 36) and suggest that the mitochondria are partially uncoupled. Further evidence for this is shown in Table 4. Inhibition of the translocase or mitochondrial ATP synthase in intact cells, with bongkrekic acid or triethyltin, respectively, resulted in only a 40-50% inhibition of the basal oxygen consumption rate. The remaining 50-60% of oxygen consumption is presumably due to proton leak across the mitochondrial membrane (see Discussion). Results similar to these have been obtained by Beauvoit and coworkers (34). The

6401

titration of oxygen consumption in control cells with bongkrekic acid demonstrates that the translocase has a low flux control coefficient for respiration in the intact cell as well as in isolated mitochondria (Fig. 2). This titration also showed that, at an inhibitor concentration of 25 ,uM, the inhibition of respiration was maximal. Similar titrations with triethyltin showed that 100 ,tM gave maximal inhibition (data not shown). The triethyltin titrations of oxygen consumption in control cells and cells overexpressing the translocase were identical (data not shown).

DISCUSSION NMR magnetization transfer measurements have been widely used to measure the kinetics of enzyme-catalyzed reactions in vitro and in vivo (37). The technique is analogous to isotope exchange measurements of enzyme kinetics except that the label, nuclear spin polarization, is very short-lived and its half-life depends on its chemical location (37). As with an isotope exchange measurement, there is no necessary relationship between the exchange measured between a pair of enzyme substrates and the overall flux catalyzed by that enzyme. In the case of creatine kinase, for example, the flux between phosphocreatine and ATP, measured in a 31P NMR magnetization transfer experiment, corresponds with the overall flux catalyzed by the enzyme only because the same step, the interconversion of the ternary enzyme complexes, limits both the exchange and the overall flux (37, 38). However, under some conditions, this result is not the case, and the exchange flux can deviate from the metabolically relevant overall flux catalyzed by the enzyme (39). Isolated yeast and mammalian mitochondria, when incubated with an NADH-linked substrate and an excess of ADP, catalyze net synthesis of ATP with a P:O ratio of -2.5 (17). Therefore, in a magnetization transfer experiment in vivo, in which the contribution of the glycolytic enzymes to the observed P1

-*

ATP flux has been accounted

for,

an

apparent P:O

ratio of -2.5 has been taken to indicate that the mitochondrial ATP synthase is operating irreversibly and that the technique is measuring net ATP synthesis in the cell (6, 12). There has been no evidence for an ATP Pi exchange catalyzed by the ATP synthase, which would increase the apparent P:O ratio and which has been observed in isotope exchange measurements on isolated mitochondria (15). In studies on the perfused rat heart, these 31P NMR magnetization transfer measurements of

Pi

->

ATP

flux, in combination with

measure-

ments of the effects of different substrates on the steady state levels of ADP, led to the proposal that the reactions catalyzed

by cytochrome c oxidase and the ATP synthase were displaced from equilibrium. These enzymes were considered, therefore, to have the potential to exert control over the rate of oxidative phosphorylation in vivo (40).

Table 1. Effect of reduced PGK expression levels on magnetization transfer measurements of Pi - ATP flux in cells perifused anaerobically with glucose Cells transformed with plasmid pMA777 (PGK 1800 pYC775 plus pBF3 (PGK 40 units/ml of cell units/ml of cell water) water) Glucose consumption (n = 4) 0.53 ± 0.03 (n = 2) 0.61 ± 0.02 (n = 4) Ethanol production (n = 4) 0.64 + 0.02 (n = 2) 0.80 + 0.05 (n = 4) Lower limit for net glycolytic flux through the triose phosphates 0.79 + 0.04 0.92 ± 0.03 0.7 ± 0.1 3.7 ± 0.5 Pi--* ATP flux (n = 15) All fluxes are quoted in units of ,umol s- '(ml of cell water)-1. The values are shown as the mean ± 1 SEM. Plasmid pMA777 directs near wild-type levels of PGK expression. The net glycolytic flux through the triose phosphates was calculated by correcting the glucose consumption rate for glucose utilization by nonglycolytic pathways, using the parameters as described (6, 32). These fluxes were used here to calculate a lower limit for net glycolytic flux through the trioses. The upper limit is twice the glucose consumption rate.

6402

Proc. Natl. Acad. Sci. USA 93

Biochemistry: Sheldon et al.

(1996)

Table 2. Effect of increased adenine nucleotide translocase expression on magnetization transfer measurements of Pi -- ATP flux 100 mM Ethanol + 110 mM 100 mM Ethanol glucose 110 mM Glucose Translocase Translocase Translocase Control overexpressor overexpressor Control Control overexpressor 0.39 ± 0.01 0.37 ± 0.03 0.46 + 0.02 0.45 ± 0.03 Glucose consumption (n =3)t (n =3) (n =3)t (n =3) 0.49 ± 0.01 0.49 ± 0.01 Ethanol production (n =3) (n =3) 0.83 ± 0.02 0.85 + 0.02 0.64 ± 0.02 0.75 + 0.01 0.64 ± 0.03 0.75 ± 0.02 Mi/Mo (n =12) (n= 10) (n =8) (n =8) (n =8) (n =8) 0.64 ± 0.01 0.81 ± 0.01 0.82 ± 0.02 0.69 ± 0.2 1/[1/(Tip + k)] s (n =2) (n =2) (n =2) (n =2) 0.57 ± 0.07 0.91 ± 0.06 1.38 ± 0.11 0.55 ± 0.10 0.92 ± 0.08 1.42 ± 0.21 Total Pi -- ATP flux (n =8) (n =8)* (n =8)* (n= 10) (n= 12) (n =8) 0.23 ± 0.01 0.44 ± 0.03 0.23 ± 0.01 0.42 ± 0.01 0.31 ± 0.01 0.34 ± 0.02 Oxygen consumption (n =2) (n =2) (n =3) (n =3)t (n 3)t (n =3) 0.55 ± 0.10 0.57 ± 0.07 1.06 ± 28* 0.59 +0.09 1.08 +0.13* 0.55 ± 0.15 Pi -- ATP flux catalyzed (0.90 ± 0.11*) (0.41 ± 0.07) (0.85 ± 0.22*) (0.33 ± 0.09) by the mitochondria 2.4 + 0.5 2.5 + 0.3 2.5 ± 0.3* 1.8 ± 0.5 3.1 ± 0.8* 1.4 ± 0.2 Apparent P:O ratio (1.0 ± 0.2) (2.1 ± 0.3*) (2.5 ± 0.7*) (1.0 ± 0.3) All fluxes are quoted in units of gmol s-1 (ml of cell water)-', with the exception of oxygen consumption, which is quoted as ,uatoms s-1 (ml of cell water)-1. The results are shown as the mean ± 1 SEM. The P1 -> ATP flux catalyzed by the mitochondria was calculated by subtracting the contribution of glycolysis from the total P1 -- ATP flux measured in the magnetization transfer experiment (see Table 1). The net glycolytic flux through the triose phosphates was calculated by correcting the glucose consumption rate for glucose utilization by nonglycolytic pathways (6, 32). These fluxes were used to estimate upper and lower limits for net glycolytic flux through the trioses and thus lower and upper limits, respectively, on the P1 -> ATP flux due to the mitochondria and the apparent P:O ratio. The values for the lower limits are shown in parentheses. The upper limit for net glycolytic flux was obtained by assuming that 17.5% of the glucose consumed was converted into polysaccharide, and the lower limit was obtained by assuming that 40.8% was converted into polysaccharide. Tip was 1.06 ± 0.04 s. This value, which is very similar to those measured previously (6, 7), was used in the calculations of P1 -* ATP flux in cells incubated with ethanol alone. *Significantly different from control (P < 0.05). tSignificantly different from cells incubated with glucose alone (P < 0.05).

The pathways responsible for the observed exchange between Pi and ATP in a magnetization transfer experiment are shown in

Fig.

1. To

measure

the

Pi ->

ATP flux

catalyzed by

the mitochondria, it is first necessary to remove, or to be able for, the Pi enzymes GAPDH and enzymes was removed

to account

->

ATP flux

catalyzed by

the

glycolytic

PGK. The exchange catalyzed by these here, while maintaining close to wildtype levels of glycolytic flux, by reducing the cellular concentration of PGK. The data in Table 1 show that in anaerobic cells with

%2%

of

wild-type

PGK

activity,

the

Pi

->

ATP flux

measured in a magnetization transfer experiment was similar to the net rate of glycolytic ATP synthesis, as calculated from measurements of

glucose consumption (i.e.,

the

Pi

++ ATP

exchange catalyzed by GAPDH and PGK was abolished). These data also show that there are no other quantitatively significant Pi

ATP flux could be estimated

by subtracting

the

net

Pi

->

ATP flux in

glycolysis,

calculated from

glucose

ATP flux consumption measurements, from the total Pi measured in a magnetization transfer experiment. When this was done, a P:O ratio of 1.0 was found in cells incubated with glucose or glucose plus ethanol (Table 2). When the cells were incubated with ethanol alone, the apparent P:O ratio was increased to -2.5 (Table 2). This is at the higher end of the range of values measured in vitro (35, 36) and is comparable with the ratios measured in a previous 31P NMR magnetization transfer study of mitochondrial ATP synthesis in yeast (6). This increase in apparent P:O ratio could indicate an increase in mitochondrial coupling efficiency in the absence of glucose. However, we cannot rule out the possibility that there is still some residual glycolytic exchange contribution under these conditions. Increased expression of the adenine nucleotide translocase by a factor of -2 (Fig. 2) resulted in an approximately proportional -*

increase in the

Pi

--

ATP flux and apparent P:O ratio in

cells

Table 3. Effect of substrate on intracellular pH and the concentrations of P1, NDP, and NTP 100 mM Ethanol + 110 mM 100 mM Ethanol glucose 110 mM Glucose Translocase Translocase Translocase Control Control Control overexpressor overexpressor overexpressor 7.18 + 0.02 7.19 ± 0.03 7.27 ± 0.01 7.30 ± 0.01 7.29 ± 0.01 7.25 ± 0.02 pH (n =10)* (n =8) (n =12)* (n =8) (n =8) (n= 8) 2.67 ± 0.11 2.43 ± 0.09 2.95 ± 0.08 2.70 ± 0.09 mM Pi 2.72 ± 0.10 2.48 ± 0.23 (n= 10) (n =8) (n =12) (n =8) (n =8) (n =8) 0.75 ± 0.07* 1.48 ± 0.05 1.34 ± 0.05 0.75 ± 0.07* 1.22 ± 0.06 1.18 ± 0.16 mM NTP (n + 2) 0.33 ± 0.01* 0.33 ± 0.03* 0.12 ± 0.01 0.12 + 0.01 0.11 ± 0.01 0.10 ± 0.01 mM NDP (n = 2) 2.3 ± 0.1* 2.3 ± 0.3* 12.3 ± 0.4 11.2 ± 0.4 11.1 ± 0.5 11.8 ± 1.7 NTP/NDP ratio pH and the intracellular Pi concentrations were determined from the cell spectra. The concentrations of NTP and NDP were determined from extract spectra. Between 50% and 70% of the NTP pool has been shown in several different yeast strains to be ATP (26). The numbers in parentheses represent the number of determinations. The results are quoted as the mean ± 1 SEM. NDP, nucleoside diphosphate; NTP,-nucleoside triphosphate. *Significantly different from cells incubated with glucose or glucose plus ethanol (P < 0.001).

Biochemistry: Sheldon et aL

Proc. Natl. Acad. Sci. USA 93 (1996)

8

3t7

1

0

{6

\

4)

on

0

1 2 3 nmol carboxyatractyloside/mg protein xlO

4

clear that the translocase has a low flux control coefficient for respiration in yeast. This is illustrated by the titrations of oxygen consumption in intact cells and isolated mitochondria, with bongkrekic acid (Table 4) and carboxyatractyloside (Fig. 2), respectively, and the absence of an increase in oxygen consumption in cells overexpressing the translocase (Table 2). These data imply that translocase overexpression has no effect on mitochondrial membrane potential as this potential can influence strongly the rate of respiration (43). The translocase also appears to have no direct effect on proton leak rate as the residual oxygen consumption rate, following full inhibition of the phosphorylation system with triethyltin or bongkrekic acid, is the same in both control cells and cells overexpressing the translocase (Table 4). As there is no evidence that the translocase has a high flux control coefficient for flux in oxidative phosphorylation, the most

FIG. 2. Titration of oxygen consumption with carboxyatractyloside in mitochondria isolated from control cells (0) and cells overexpressing the adenine nucleotide translocase (-).

incubated with glucose but had no effect on cells incubated with ethanol alone (Table 2). The explanation for these observations depends on whether the translocase has a high or low control coefficient for flux in oxidative phosphorylation. Utilization of the proton-motive force generated by the respiratory chain can be viewed as a competition between the proton leak pathway and the phosphorylation system (18). If the translocase has a high control coefficient for flux in the phosphorylation system, then increasing its activity could divert additional proton flux through the ATP synthase, at the expense of the leak pathway, and the apparent P:O ratio would increase. This could explain the increased apparent P:O ratios observed after an increase in translocase content, in cells incubated with glucose, or after a rise in the ADP concentration, in cells incubated with ethanol (Tables 2 and 3). However, there is no evidence that the translocase has a high control coefficient for flux in the phosphorylation system. The NDP/ NTP ratio and phosphorylation potential, which are increased in yeast cells in which flux in oxidative phosphorylation has been stimulated by oxygenation (41, 42), were unaffected by an increase in translocase content in cells incubated with glucose or glucose plus ethanol (Table 3). Furthermore, an inhibitor titration study on isolated yeast mitochondria showed that the translocase had a low control coefficient for flux in the phosphorylation system, although in this study a relatively high ADP concentration had been used (0.5 mM) (31). It is also

6403

likely explanation

for the increases in

Pi

-k

ATP flux and

apparent P:O ratio, in cells overexpressing the translocase, is that the translocase is catalyzing a reaction that is near-to-

equilibrium. Therefore, what is being measured in the magnetization transfer experiment in cells incubated with glucose is, in part, an exchange between P1 and ATP catalyzed by the mitochondrion rather than unidirectional flux. This interpretation is consistent with isotope exchange measurements of flux between P1 and ATP in isolated rat heart and liver mitochondria, which showed that whereas ATP synthesis was unidirectional in state 3, there was considerable exchange between P1 and ATP in state 4. The unidirectionality of ATP synthesis in state 3 was thought either to be a kinetic property of the translocase, limiting the rate of reentry of ATP into the mitochondria (see Fig. 1), or of the synthase itself (15). Furthermore, these

data

suggested

that

a

mitochondrial ATP

*->

P1

exchange reaction should have been observable by 31P NMR magnetization transfer measurements on the isolated perfused rat heart at lowworkloads (15). This isotope exchange study could explain the observations made here. At low ADP concentrations in cells incubated with glucose, an increase in translocase content could increase the rate of ATP transport into the mitochondria and, thus, the

rate of an ATP

*

Pi exchange

reaction. In cells

incubated with ethanol alone, the high ADP concentration could compete with ATP for entry into the mitochondria and, thus, inhibit the exchange. Under these circumstances, the flux between Pi and ATP would become unidirectional and equivalent to the net rate of mitochondrial ATP synthesis. A corollary of this explanation is that, in the magnetization transfer experiment, the intramitochondrial ATP -phosphate resonance cannot be directly saturated when the extramitochondrial resonance is irra-

Table 4. Effect of inhibitors and uncoupler on the respiration rate of cells incubated with glucose or ethanol Respiratory rate, nmol 0 min-'-(mg dry weight)-' Inhibitor Cell line With glucose (110 mM) With ethanol (100 mM) Basal respiratory rate Control 80 ± 2 70 ± 3 Translocase overexpressor 80 ± 3 70 ± 6 Control 3±1 Antimycin A, 25 ,ug/ml 3±1 Translocase overexpressor 2±1 3±1 Bongkrekic acid 2.5 ,uM Control 78 75 3.75 ,uM Control -83 72 5.0 ,uM Control 83 72 25 ,uM Control 46 ± 2 39 ± 2 Control 45 ± 6 36 ± 5 Triethyltin, 100 ,uM Translocase overexpressor 40 ± 2 31 ± 1 Control 200 ± 22 Carbonyl cyanide m-chlorophenyl 152 ± 11 Translocase overexpressor hydrazone, 20 ,uM 149 ± 6 200 ± 7 Respiratory rates were measured as described in Materials and Methods. Where errors are quoted, these values are the mean of three determinations ± SD. When bongkrekic acid was used, the buffer was changed from 50 mM Mes, pH 6.0, to 50 mM phthalate, pH 4.5. Antimycin A and carbonyl cyanide m-chlorophenyl hydrazone were dissolved in 95% ethanol. Bongkrekic acid was dissolved in 1 M ammonium hydroxide.

Proc. Natl. Acad. Sci. USA 93

Biochemistry: Sheldon et al.

6404

diated, because if it were, the translocase step would be bypassed, and increasing translocase activity could have no effect on the measured

Pi

+-> ATP

exchange.

Failure to saturate the intrami-

tochondrial ATP y-phosphate resonance could occur if it had a significantly different chemical shift from the extramitochondrial resonance or if the intramitochondrial ATP pool was turning over very rapidly, making its resonances difficult to saturate directly, as is the case with enzyme-bound intermediates (38, 39). A significant exchange contribution to the measured Pi ATP flux would make the true P:O ratio for cells incubated with glucose even lower than the value of 1.0 quoted in Table 2. The experimental results shown in Table 4 suggest that this low P:O ratio could be due to proton leak across the mitochondrial membrane. Inhibition of the mitochondrial phosphorylation system, with triethyltin or bongkrekic acid, resulted in a reduction of the oxygen consumption rate of only 40-50%. Because inhibition of the mitochondrial respiration system, with antimycin A, showed that nonmitochondrial oxygen consumption was negligible, it seems reasonable to conclude that the remaining oxygen consumption is due to proton leak (16-18). This is likely to represent an upper limit on the oxygen consumption due to leak as the mitochondrial membrane potential may be increased in these phosphorylation-inhibited cells (16-18). Measurements of mitochondrial membrane potential, from measurements of labeled methyltriphenylphosphonium cation uptake, were unsatisfactory due to limited uptake of the ion (data not shown). This problem has been observed previously (44). However, if it is assumed that the leak is the same in inhibited and noninhibited cells incubated with glucose, then the data in Table 4 show that proton leak could account for most of the apparent uncoupling of oxidative phosphorylation observed in the magnetization transfer experiment. Subtraction of the putative oxygen consumption rate due to proton leak (Table 4), from the oxygen consumption rates measured in the magnetization transfer experiments on the cells incubated with glucose (Table 2), results in an increase in the apparent P:O ratio from -1.0 to -2.5. This direct demonstration of partial uncoupling of mitochondrial oxidative phosphorylation in vivo could explain the long-standing observation that experimentally observed growth yields in microbial cultures are 50% or less of theoretical yields (45, 46). In conclusion, we have demonstrated that yeast mitochondria can be partially uncoupled in cells incubated with glucose and that oxygen consumnption, which is not coupled to mitochondrial ATP synthesis, can be accounted for by proton leak. The apparent reversibility of mitochondrial ATP synthesis at low ADP concentrations is consistent with the model of kinetic control proposed by LaNoue and coworkers (15). In conjunction with isotope exchange measurements made in vitro (15) and studies demonstrating that the P:O ratio is a function of respiration rate in mammalian mitochondria (16-18), these -

data

suggest that

significant

an

ATP

P1

contribution to the P1

->

exchange

could

make

ATP flux measured

a

using

magnetization transfer techniques in mammalian tissues at low rates of respiration. We thank Dr. Michael Douglas for the translocase clone, Dr. Alan Kingsman for various plasmids, and Dr. Peter Piper for the cells underexpressing PGK. We are grateful to Dr. Andrew Halestrap for his generous gift of bongkrekic acid and to Dr. Martin Brand for his critical reading of this manuscript and for his help with the methyltriphenylphosphonium cation uptake experiments. We also acknowledge the contribution made by Corinne Spickett in the early stages of this project. This work was supported by the Wellcome Trust and the Biotechnology and Biological Sciences Research Council. 1. Brown, T. R., Ugurbil, K. & Shulman, R. G. (1977) Proc. Natl. Acad. Sci. USA 74, 5551-5553.

(1996)

2. Mitsumori, F., Rees, D., Brindle, K. M., Radda, G. K. & Campbell, I. D. (1988) Biochim. Biophys. Acta 969, 185-193. 3. Alger, J. R., den Hollander, J. A. & Shulman, R. G. (1982) Biochemistry 21, 2957-2963. 4. Brindle, K. & Krikler, S. (1985) Biochim. Biophys. Acta 847, 285-292. 5. Campbell, S. L., Jones, K. A. & Shulman, R. G. (1985) FEBS Lett. 193, 189-192. 6. Campbell-Burk, S. L., Jones, K. A. & Shulman, R. G. (1987) Biochemistry 26, 7483-7492. 7. Brindle, K. M. (1988) Biochemistry 27, 6187-6196. 8. Brindle, KI, Fulton, S., Sheldon, J. & Williams, S. (1995) Biochem. Soc. Trans. 23, 374-379. 9. Shoubridge, E. A., Briggs, R. W. & Radda, G. K. (1982) FEBS Lett. 140, 288-292. 10. Matthews, P. M., Bland, J. L., Gadian, D. G. & Radda, G. K. (1981) Biochem. Biophys. Res. Commun. 103,1052-1059. 11. Kingsley-Hickman, P., Sako, E. Y., Andreone, P. A., St. Cyr, J. A., Michurski, S., Foker, J. E., From, A. H. L., Petein, M. & Ugurbil, K. (1986) FEBS Lett. 198, 159-163. 12. Kingsley-Hickman, P. B., Sako, E. Y., Mohanakrishnan, P., Robitaille, P. M. L., Foker, J. E., From, A. H. L. & Ugurbil, K. (1987) Biochemistry 26, 7501-7510. 13. Ugurbil, K. & From, A. H. L. (1992) in Cardiovascular Magnetic Resonance Spectroscopy, eds. Schafer, S. & Balaban, R. S. (Kluwer, New York), pp. 63-92. 14. Sako, E. Y., Kingsley-Hickman, P. B., From, A. H. L., Foker, J. E. & Ugurbil, K. (1988) J. Biol. Chem. 263, 10600-10607. 15. LaNoue, K. F., Jeffries, F. M. H. & Radda, G. K. (1986) Biochemistry 25, 7667-7675. 16. Brand, M. D., Chien, L.-F. & Rolfe, D. F. S. (1993) Biochem. Soc. Trans. 21, 757-762. 17. Brand, M. D., Harper, M.-E. & Taylor, H. C. (1993) Biochem. J. 291, 739-748. 18. Harper, M.-E. & Brand, M. D. (1993) J. Biol. Chem. 268, 14850-14860. 19. Bradford, M. M. (1976) AnaL Biochem. 72, 248-254. 20. Piper, P. W. & Curran, B. P. G. (1990) Curr. Genet. 17, 119-123. 21. Ogden, J. E., Stanway, C., Kim, S., Mellor, J., Kingsman, A. J. & Kingsman, S. M. (1986) Mol. Cell. Biol. 6, 4335-4343. 22. Hinnen, A., Hicks, J. B. & Fink, G. R. (1978) Proc. Natl. Acad. Sci. USA 75, 1929-1933. 23. O'Malley, K., Pratt, P., Robertson, J., Lilly, M. & Douglas, M. G. (1982) J. Biol. Chem. 257, 2097-2103. 24. Kingsman, S. M., Dobson, M. J., Tuite, M. F., Mellor, J., Roberts, N. A. & Kingsman, A. J. (1983) in Gene Expression in Yeast: Proceedings of theAlko Yeast Symposium, Helsinki, eds. Korhola, M. & Vaisanen, E. (Foundation for Biotechnical and Industrial Fermentation Research), Vol. 1, pp. 95-114. 25. Brindle, K. M., Fulton, A. M. & Williams, S.-P. (1995) in Bioenergetics: A Practical Approach, eds. Brown, G. & Cooper, C. (IRL, Oxford), pp. 159-182. 26. Brindle, K., Braddock, P. & Fulton, S. (1990) Biochemistry 29, 3295-3302. 27. Den Hollander, J. A., Ugurbil, K., Brown, T. R. & Shulman, R. G. (1981) Biochemistry 20, 5871-5880. 28. Bergmeyer, H. U. (1974) inMethods in EnzymaticAnalysis (Academic, New York). 29. Brindle, K. M. & Radda, G. K. (1987) Biochim. Biophys. Acta 928, 45-55. 30. Guerin, B., Labbe, P. & Somlo, M. (1979) Methods Enzymol. 55, 149-159. 31. Mazat, J.-P., Jean-Bart, E., Rigoulet, M. & Guerin, B. (1986) Biochim. Biophys. Acta 849, 7-15. 32. Den Hollander, J. A., Ugurbil, K., Brown, T. R., Bednar, M., Redfield, C. & Shulman, R. G. (1986) Biochemistry 25, 203-211. 33. Gawaz, M., Douglas, M. G. & Klingenberg, M. (1990) J. Biol. Chem. 265, 14202-14208. 34. Beauvoit, B., Rigoulet, M., Bunoust, O., Raffard, G., Canioni, P. & Guerin, B. (1993) Eur. J. Biochem. 214, 163-172. 35. Fitton, V., Rigoulet, M., Ouhabi, R. & Gu6rin, B. (1994) Biochemistry 33, 9692-9698. 36. Mackler, B. & Haynes, B. (1973) Biochim. Biophys. Acta 292, 88-91. 37. Brindle, K. M. (1988) Prog. NMR Spectrosc. 20, 257-293. 38. Brindle, K. M. & Radda, G. K. (1985) Biochim. Biophys. Acta 829,188-201. 39. Kupriyanov, V. V., Balaban, R. S., Lyulina, N. V., Steinschneider, A. Ya. & Saks, V. A. (1990) Biochim. Biophys. Acta 1020, 290-304. 40. From, A. H. L., Bache, R. J., Zhang, J. & Ugurbil, K. (1994) in NMR in Physiology and Biomedicine, ed. Gillies, R. J. (Academic, San Diego), pp. 413-437. 41. Den Hollander, J. A., Ugurbil, K. & Shulman, R. G. (1986) Biochemistry 25, 212-219. 42. Davies, S. E. C. & Brindle, K. M. (1992) Biochemistry 31, 4729-4735. 43. Brown, G. C. & Brand, M. D. (1986) Biochem. J. 234, 75-81. 44. Beauvoit, B., Rigoulet, M., Raffard, G., Canioni, P. & Guerin, B. (1991) Biochemistry 30, 11212-11220. 45. Tempest, D. W. & Neijssel, 0. M. (1984) Annu. Rev. Microbiol. 38, 459486. 46. Verduyn, C., Stouthamer, A. H., Scheffers, W. A. & van Dijken,J. P. (1991) Antonie van Leeuwenhoek 59, 49-63.