Mechanisms of Oxidant-mediated Cell Injury THE GLYCOLYTIC AND MITOCHONDRIAL PATHWAYS OF ADP PHOSPHORYLATION ARE MAJOR INTRACELLULAR TARGETS INACTIVATED BY HYDROGEN

Inhibition of ADP phosphorylation by both glycolysis and mitochondria in P388D1 cells exposed to HzOz is described. Net glucose uptake and lactate production were inhibited by oxidant exposure (ED60 PM). Glycolysis was specifically inactivated at the glyc-eraldehyde-3-phosphate dehydrogenase step by three independent mechanisms: (a) direct inactivation of the intracellular enzyme (EDbo = 100 WM); (b) reduction of the intracellular concentration and redox potential of its nicotinamide cofactors; and (c) a cytosolic pH shift further from the enzyme optima. Consistent with inhibition of glycolysis at the glyceraldehyde-3-phos-phate dehydrogenase step, a rise in the intracellular concentration of glyceraldehyde 3-phosphate, dihydroxyacetone phosphate, and fructose 1,6-bisphos-phate was observed. The calculated combined inhibition of glyceraldehyde-3-phosphate dehydrogenase Res. Commun. 148, 120-125) demonstrated that glyceraldehyde-3-phos- phate dehydroyenase of human lung carcinoma cells was inhibited by exposure.

the intracellular concentration and redox potential of its nicotinamide cofactors; and ( c ) a cytosolic pH shift further from the enzyme optima. Consistent with inhibition of glycolysis at the glyceraldehyde-3-phosphate dehydrogenase step, a rise in the intracellular concentration of glyceraldehyde 3-phosphate, dihydroxyacetone phosphate, and fructose 1,6-bisphosphate was observed. The calculated combined inhibition of glyceraldehyde-3-phosphate dehydrogenase activity could be reasonably correlated with the depression in glycolytic flux rate with the appropriate modeling.
The steady-state contribution by mitochondria to the total intracellular ATP pool was indirectly determined by the use of various metabolic inhibitors and was found to rapidly decline following exposure to 300-800 MM HzOZ. The inhibition of ADP phosphorylation appeared to be related more to the direct inhibition of the ATPase-synthase complex rather than to the diminished capacity of the respiratory chain for coupled electron transport.
Both the estimated rates of ADP phosphorylation by glycolysis and mitochondria and the estimated rate of ATP hydrolysis by ongoing metabolism were utilized to model the approximate decline in intracellular ATP expected at 15-min exposure to various HzOz concentrations. Theoretical calculations and the measured intracellular ATP status were in good agreement. Oxidant exposure for 15 min resulted in dose-dependent killing of the cells (EDSo = 500 NM), indicating a close correlation between HzOz-mediated loss of intracellular ATP and cell viability. The possible contri-HL23584, Office of Naval Research Contract 105-837, and Grant * This work was supported by National Institutes of Health Grant 764-H from the Council for Tobacco Research. This is Publication 4608-1" from the Research Institute of Scripps Clinic. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ Present address: Dept. of Surgery, Surgical Service (112), University of Michigan, 2215 Fuller Rd., Ann Arbor, MI 48105. bution of impaired energy homeostasis during oxidantmediated injury to the process of cell dysfunction and death is discussed.
Role of Oxidants in Inflammatory Disease-Stimulated leukocytes generate superoxide anion (0;) as part of the host defense mechanism against foreign organisms in higher animals. In inflammatory disease states, oxidants and reactive free radicals are generated from the relatively unreactive 0; and contribute to the destruction of tissues by directly causing cell dysfunction and loss of viability (1-6). However, in the environment of stimulated neutrophils, reactive free radicals can only be detected in the presence of added iron salts (7), and even under these conditions stimulated neutrophil products actively antagonize their formation (8). HzOz (produced by the dismutation of O;), on the other hand, has been detected in inflammatory disease (9). Addition of H202 to cells causes loss of viability (2,3,10-17), and perhaps more important in the short term, causes alterations in cell morphology (15,16). Both of these events are likely to contribute, for example, to the loss of vascular endothelium integrity observed in HzOp-induced high permeability lung edema (17).

1666
Mechanisms of Oxidant Injury decline in intracellular ATP. First, the rate of decline in intracellular ATP is as rapid as when ADP phosphorylation is completely arrested with inhibitors (23). Second, glycolysis is inhibited in tumor cells (26)(27)(28)(29)(30) by HzO, exposure. The intracellular NAD' content is reduced (12, 25) (this cofactor is required for glycolysis), and, furthermore, the addition of exogenous nicotinamide to tumor cells partially restored the glycolytic activity (27,28). Glyceraldehyde-3-phosphate dehydrogenase has also been shown to be inactivated by H202 in vitro (31,61), raising the possibility that the enzyme could be inactivated by oxidants in vivo. P388D1 Model of Oxidant Targets-P388D1 cells, as other transformed cells, derive a greater part of their metabolic energy anaerobically by both glycolysis and glutaminolysis (32). In the absence of glutamine, glycolysis and oxidative phosphorylation compensate adequately, as in other tumor cells (32). This cell line, utilized as a target with glucose as the sole metabolizable carbon source, is, therefore, a useful model for a target maintaining ATP homeostasis by both glycolysis and oxidative phosphorylation. The impact of HzO, on both these pathways and the resulting effect on cellular ATP can be studied independently within the same cell. The metabolic properties of the P388D1 cell have many similarities to bovine endothelial cells (33): for example, a very high glycolytic activity under aerobic conditions, lactate rather than CO, accounting for the bulk of metabolized extracellular glucose; a highly activatable HMP shunt pathway; and finally the cells have a low mitochondrial respiration rate.

RESULTS
Glucose Uptake Studies-Overall clearance of glucose (5.5 mM) from the media by P388D1 cells is suppressed over long time courses following incubation with 250 pM H202, from 875 f 56 pmol/106 cells/min to approximately 400 pmol/106 cells/min averaged over a 30-min exposure period. At 5.5 mM extracellular glucose utilizing [3H]2-deoxyglucose tracer, the apparent velocity of net glucose uptake measured by this technique was similar to the estimated clearance of glucose from the media (Table 1).
Cytochalasin B inhibition studies of 2-deoxyglucose tracer uptake indicate that the membrane transport step exerts low rate control over the overall process of transport and phosphorylation of the tracer (Table 1) at physiological (5.5 mM) extracellular glucose.
At 100 ~L M extracellular glucose, the transport rate is sufficiently reduced such that the control strength (Cj) for the transport/phosphorylation process is considerably increased (Fig. 1, Table 1). Following HzOz injury, net uptake of tracer is inhibited in a dose-dependent fashion (Table 2). Furthermore, at this concentration of extracellular glucose, Cj is reduced following exposure to 0.1 mM HzOz (Fig. 1, Table 1) indicating that phosphorylation must, at least partially, contribute to reduced tracer uptake.
However, when the transport rate is reduced even further by incubating the cells with 10 p~ glucose, both the net uptake of tracer and Cj are unperturbed by exposure to 0.1 mM H202 (Table 1). Thus, transporter activity appears not to be inhibited by HZ02, and the dose-dependent decrease in net 2-deoxyglucose uptake reported in Table 1 is the result of increased efflux of the incompletely phosphorylated hexose.
Effects of H2OZ on Glucose Conversion to Lactate and C0,-Total lactate production was inhibited by Hz02 in a dosedependent manner (Table 1). Specific activities of the lactic acid produced by metabolism from ['4C]glucose in the presence and absence of 0.1 mM Hz02 were determined in separate experiments to validate these conclusions. Glucose (specific activity, 580 mCi/mol) was converted into lactate over a 10min period by the P388D1 cells (specific activity, 520 k 91 mCi/mol; n = 5) at a rate of formation of 1.7 f 0.3 nmol/106 cells/min; n = 5. In the presence of 0.1 mM HzOz, lactate was formed at a rate of 0.85 2 0.08 nmol/1O6 cells/min, n = 5 (specific activity, 590 f 59 mCi/mol, n = 5).
14C02 production from extracellularly labeled ['4C]glucose metabolized through the HMP shunt and the tricarboxylic acid cycle were both activated by HzO2 (Table 3) with little dose dependence up to 1 mM Hz02. Control rates of lactate carbon production were, within experimental error, identical to those of glucose carbon clearance from the media and establish lactate production as the major route of glucose utilization in these cells.
Effect of HzOz on Enzymes of the Glycolytic Pathway-No detectable effects of 5 mM Hz02 were observed on the kinetic parameters of any of the glycolytic enzymes, except for a (significantly) small effect on the maximal activity of hexokinase and a relatively large inhibition of glyceraldehyde-3phosphate dehydrogenase ( Table 4). The ICso for this effect was approximately 100 PM H202 (Fig. 2, open circles). Maximal inhibition was achieved in about 10 min of exposure to oxidant (Fig. 3C). H202 also inhibited the activity of purified rabbit muscle glyceraldehyde-3-phosphate dehydrogenase, ICso = 10 pM HzOz (Fig. 2, open circles).
Effects of HzOz on Intracellular Metabolites-Fructose 1,6bisphosphate and total GAP plus dihydroxyacetone phosphate was increased by H20z exposure (Table 4), consistent with glyceraldehyde-3-phosphate dehydrogenase inhibition. However, glucose-6-P and fructose-6-P concentrations were reduced by H202 exposure, indicating that the hexokinase step must also be (indirectly) inhibited. The only known allosteric regulator of hexokinase (apart from ATP at >5 mM) is glucose 1,6-dlphosphate (50). Table 4 indicates that the intracellular level of glucose 1,6-diphosphate is suppressed following HzO2 exposure, and since ATP levels do not increase, allosteric inhibition of the enzyme cannot account for the data. Intracellular I;-Y Measurements-Exposure to H202 (>lo0 p~) produced a modest fall of about 0.2 pH units after 15 min of exposure (mean (0.1-5 mM) f S.D. = 6.56 f 0.03), while 50 PM produced an intermediate value ( Table 4). The fall in pH was monitored as a function of time (not shown), exposing the cells to 1 mM Hz02. The pH declined to its minimal level within the first 10 min of exposure and remained constant for a further 20 min. In separate experiments, the sensitivity of glyceraldehyde-3-phosphate dehydrogenase to pH was measured. Between pH 8.0 and 7.4, the enzyme activity was relatively constant. Between pH 7.4 and 6.0, a near linear dependence of the activity (A) was observed (dA/dpH = -0.64). The expected relative intracellular activity of the enzyme in control cells at the prevailing intracellular pH will, therefore, be 56% of its activity at pH 7.4. Exposure to 50 p M would be expected to depress the activity of the intracellular enzyme 5% below control rate, while >lo0 p~ depresses this activity by 12%. Influence of H202 Exposure on Intracellular NAD' and NADH-H2O2 up to 1 mM was without detectable effect on the NAD' content of the mitochondrial compartment. (Con-trol mean f S.D. = 8.6 f 1.9 ( n = 12) pmol/106 cells; Hz02 (500-1000 fiM) for 30 min = 8.2 2 3.7 ( n = 12) pmOl/1o6 Cells.) Time courses of intracellular NAD' and NADH content were monitored for 30 min following exposure to H202. It was found that in triplicate experiments (not shown) 50,100, 250, and 500 NM H20, had similar (within experimental error) effects on depressing the intracellular concentration of nicotinamide nucleotides. These data were, therefore, combined and are presented in Fig. 3, panels A and B. The threshold detectable H202 concentration modulating NAD'/NADH was 37.5 PM (not shown). 1 mM H202 resulted in a less severe decline in NAD' but a greater decline in NADH. The mean values for two separate experiments after 15 min of exposure were 82 and 27 pmol/106 cells, respectively.
NAD' and NADH Binding to Cytosol in Vitro-The total percent bound NAD' (closed squares) and NADH (open squares) to cytosolic proteins is presented in Fig. 4. The nonsaturable component, estimated at 10 mM NAD' or NADH, was 35% in both cases (not shown). From these binding curves, the estimated free intracellular [NAD'] and [NADH] are calculated, as described in the previous section, over the time course of the experiment (Table 6). Using these values of [NAD(H)] the theoretical velocities of GAP oxidation can be calculated as a percentage of control and are shown in Table 5. The impact of the change in the NAD(H) concentration due to H202 exposure on the activity of glyceraldehyde-3-phosphate dehydrogenase is assessed at the time point of interest in these studies, i.e. after 15 min of exposure, by converting the fractional V,,, activity to a percentage of the control value. This value is the useful parameter because it determines the effect of a change in the nicotinamide redox state and concentration on glyceraldehyde-3-phosphate dehydrogenase activity (with respect to GAP oxidation) under any given set of conditions (Le. constant GAP, phosphate, 1,3-diphosphoglycerate, and pH).
Measurement of the Effects of H202 on Mitochondrial Respiration-Atractyloside and oligomycin-sensitive respiration (measured as the difference between rates in the presence and absence of inhibition, Fig. 5, left panel) in control cells were 0.073 k 0.012 and 0.56 f 0.11 nmol 02/min/106 cells, respectively. H202 exposure inhibited these processes such that atractyloside and oligomycin sensitivity was unmeasurable at 800 p~ and 5 mM H202, respectively. Atractyloside-sensitive oxygen consumption was significantly increased above control levels between 100 and 300 PM H202 and presumably represents mitochondrial activation of ADP phosphorylation in response to glycotic inhibition by H202 at these concentrations. The time course for oligomycin-sensitive respiration inhibition (Fig. 3, panel A ) demonstrates that the greatest inhibition occurs within the first 10 min of oxidant exposure. H202 also depressed basal and uncoupled respiration in a dose-dependent manner (Fig. 3, panel B).
The decline in ADP phosphorylation appears to be related more to inactivation of the ATPase-synthase rather than to the decline in the rate of the electron transport, since over the region of inhibition (0-1 mM H202) the ratio of respiration rates attainable by uncoupling respiration from the proton electrochemical gradient remained substantially greater than basal rates of respiration. However, the capacity of the respiratory chain for electron transport is compromised by the oxidant, and at 2.5 mM H202, may contribute to limiting ADP phosphorylation. The observation that the decline in basal respiration by H202 exposure was much less inhibited than the oligomycin-sensitive portion indicates that H202 may induce an increase in electron transport coupled to ion translocation and/or molecular slip. These data indicate oxidants induce alterations in the function of a number of respiratory chain components.
Rate of Fall in Intracellular ATP following Exposure to Inhibitors of Glycolysis and Oxidative Phosphorylation-The initial rate of fall (nmol/min/106 cells) in intracellular ATP was as follows (Fig. 6): deoxyglucose plus oligomycin = 1.23, deoxyglucose alone = 0.75, atractyloside alone = 0.2. These data indicate that resting mitochondria are contributing 0.2 nmol/min/106 cells to the intracellular ATP pool, rising to 0.48 nmol/min/106 cells under conditions where they are compensating for ATP loss due to inhibition of ADP phosphorylation by glycolysis.
In order to test whether atractyloside rather than oligomycin-sensitive rates of respiration indeed represents the component of mitochondrial respiration linked to net ATP export from the mitochondria, it is necessary to convert the rates of inhibitor-sensitive oxygen consumption (previous section) to rates of expected ADP phosphorylation. Using a P/ O2 ratio of 6 for NADH-linked electron transport, atractyloside (0.44 nmol/min/106 cells) rather than oligomycin-sensitive respiration (3.36 nmol/ATP/min/lO' cells) clearly is closer to the expected contribution to the ATP pool from mitochondria (0.2 nmol/min/106 cells).
Effects of Hz02 on ATP Hydrolysis-In order to assess whether the presence of H202 accelerated or depressed the overall utilization of ATP, the rate of fall in ATP was monitored every 10 s over a 3-min time course in the presence of 2-deoxyglucose and KCN at different H202 concentrations. Rates were determined from the apparently linear (within experimental error) portion of the curves (Fig. 6). ATP utilization was constant as a function of time within experimental error from 20 to 120 s (2.3 nmol/106 cells/min) and was unperturbed by H202 over the dose range 50-1000 ~L M (2.20 f 0.06 nmol/106 cells/min).
Growth of P388D1 Cells following Exposure-Exposure of the cells to Hz02 resulted in growth curves that could all be fitted to a single exponential within experimental error, since the correlation coefficients (Table 6, fourth column) of the log plots were all very close to unity. The lag time of growth was approximately 19 h at all doses tested. The analysis of the data by this method enables a reasonable estimate to be made of the numbers of cells that are potential survivors at the 15th min of exposure to oxidant in terms of viability and retaining replicative capacity (column 2).

DISCUSSION
In this study, we identify intracellular targets of H20z injury that interfere with ADP phosphorylation. As the metabolic perturbations of interest occur largely within the first 10 min of exposure to H202 (Fig. 3), we focus on measuring the extent of the injury at the 15th min of exposure to oxidant. By integrating our experimental observations at this time point, we will attempt to evaluate whether the expected inhibition of net ADP phosphorylation by the range of oxidant concentrations correlates with the actual observed ATP status.
Inhibition of Glycolysis by H202-The premise that the glycolytic pathway is inhibited by H202 in these studies, as has been shown for other cell types (26)(27)(28)(29)(30), is drawn from the observation that lactate production is inhibited in a dosedependent fashion. Lactate production is the major pathway for extracellular glucose utilization in control cells (Table 1) since the flux of glucose 6-phosphate and pyruvate from extracellular glucose necessary for the HMP shunt and tricarboxylic acid cycle, respectively, is only 2% of the total substrate flux through glycolysis. 50 FM H202 maximally activates the flux of glucose carbon through the HMP shunt pathway, and under these conditions, still only represents diversion of 12% of the total metabolized glucose from glycolysis. During injury, the amount of glucose carbon partitioning via acetyl-coA increases but represents only a small fraction of the estimated total acetyl-CoA flux required by the mitochondria to theoretically support the observed rates of oxygen consumption. Thus, in these cells under the conditions studied, glycolysis, HMP shunt, and tricarboxylic acid cycle operate relatively independently in terms of metabolite availability, which is useful when assessing the impact of a perturbant on intracellular ATP metabolism.
Inhibition of the glycolytic pathway at the glyceraldehyde-3-phosphate dehydrogenase step is supported by the observation that the intracellular concentrations of GAP and dihydroxyacetone phosphate are substantially elevated by oxidant exposure. Since no detectable effects were observed on aldolase activity, the increase in fructose 1,6-bisphosphate is presumably due to the enzyme equilibrium.
Experiments reported in this study demonstrate that, under conditions where net uptake of hexose tracer is reduced by HzOz exposure, no significant effect is observed on the activity of the hexose carrier. Inhibition of glucose phosphorylation is clearly not a direct consequence of reduced flux of substrate through glycolysis (since the intracellular concentration of glucose-6-P and fructose-6-P fall following oxidant injury); also allosteric inhibition by an increase in glucose 1,6-diphosphate has also been excluded. At least one explanation for decreased hexose phosphorylation is likely to be due to the loss of intracellular ATP inhibiting hexokinase activity. For example, reduction of intracellular ATP from control values (6 mM) to 1 mM would reduce the theoretical rate of glucose phosphorylation by 40% (results not shown). (Because of the lower K,,, for ATP, the activity of phosphofructokinase would be less sensitive to ATP status.) The significant decrease in hexokinase activity following 5 mM H202 exposure may also contribute to a lesser degree at lower concentrations. Other factors such as redistribution of the intracellular location of the enzyme (53, 54) or the modulation of an unidentified regulator of hexokinase may also contribute to the inhibition of glucose phosphorylation during oxidant exposure.
The fall in intracellular glucose-6-P and fructose-6-P following H202 exposure can be reasonably qualitatively explained by both a decrease in the rate of glucose phosphorylation and also activation of the HMP shunt following injury.
Evidence has been presented that H202 reduces the intracellular glyceraldehyde-3-phosphate dehydrogenase activity by essentially three independent mechanisms. First, the major effect appears to be direct inactivation of the enzyme by H202, both within the cell and the isolated enzyme. The differences in dose dependences in these two phenomena are probably related to the fact that the cell suspension is rapidly metabolizing the HzOz, so the integrated concentration over the 15min assay period is lower in the cell suspension (11). Also, intracellular protective antioxidant mechanisms may compete to some degree with some step(s) in the oxidation process. It appears that, once the enzyme is inactivated, intracellular reductants are poorly efficient at reactivating the e n~y m e .~ Second, the impact of the altered concentration and redox potential of the cytosolic nicotinamide cofactors is estimated to reduce the absolute activity of intracellular glyceraldehyde-3-phosphate dehydrogenase by a factor of two over the dose range of 50-500 FM HzOz after 15 min of exposure. The lack of dose dependence of the effect above 50 p~ HzOz may represent maximal activation of the nuclear enzyme, poly(ADP-ribose) polymerase, which quantitatively ADP-ribosylates nuclear proteins at the expense of NAD' during Hz02 stress (14). Last, the small pH shift occurring in response to HzO2 modulates the activity of the protein by about 12% and probably results from extensive ATP hydrolysis.
The effect of summing these purturbations on the expected velocity of GAP oxidation as a percentage of control activity by glyceraldehyde-3-phosphate dehydrogenase as a function of Hz02 concentration is presented in Fig. 8, panel A. The calculation is performed assuming no other constraint on the enzyme, at constant substrate (GAP and 1,3-diphosphoglycerate) concentrations. A small correction for the effects of an increase in measured intracellular GAP concentration as a function of Hz02 dose in glyceraldehyde-3-phosphate dehydrogenase activity is included (open circles).
The estimated contribution of the glycolytic pathway to ADP phosphorylation is readily obtained from the sum of lactate and pyruvate flux (estimated from 14C02 production via the tricarboxylic acid cycle). Errors introduced by metabolism of pyruvate via other pathways are minimized, since lactate is the major glycolytic product. The result of this determination (1 mol of lactate/pyruvate = 1 mol of ATP) is shown in Fig. 8,  Since the magnitude of inhibition of the enzyme compares with the observed decrease in flux rate through the pathway, it is clear that under normal conditions, glyceraldehyde-3phosphate dehydrogenase must be sharing some of the control over glycolysis with the major regulators, phosphofructokinase and hexokinase. There appears to be good evidence that this is the case under certain metabolic conditions in brain (55), ascites cells (56), and heart (57). In P388D1 cells the total activity of glyceraldehyde-3-phosphate dehydrogenase (measured as the maximal possible rate of GAP oxidation at saturating NAD' , HAs04, pH 8.0) is -&fold higher than the maximal phosphorylating activities of phosphofructokinase and hexokinase; however, with respect to GAP oxidation, glyceraldehyde-3-phosphate dehydrogenase is operating at a low efficiency in uninjured cells. First, the prevailing intracellular pH reduces its activity to 56% of its activity at optimum pH (>7.4). Second, the redox status of the nicotinamide co-factors reduces its optimal activity for oxidation of GAP by 51%, and third, the concentration of intracellular phosphate reduoos the maximal rate of GAP oxidation by 50%. The net effect of these factors is to reduce the theoretical intracellular activity of glyceraldehyde-3-phosphate dehydrogenase to 14% of its maximum activity with respect to GAP oxidation. When it is considered that in uninjured cells phosphofructokinase is presumably operating near maximal efficiency (since intracellular ATP and fructose-6-P are close to saturating), the maximal activities of phosphofructokinase (15 milliunits/106 cells) and the above estimated intracellular glycolytic activity of glyceraldehyde-3-phosphate dehydrogenase (11 milliunits/I06 cells) are very comparable.
Hz02 Inhibition of Mitochondrial ATP Production-The apparent mitochondrial ATP output declines after exposure to Hz02, although 100-300 p~ H20z results in an increase in net mitochondrial ATP production, presumably as a direct consequence of glycolysis inhibition. However, over the same concentration range of H2O2, the ATPase/synthase activity is inhibited. This indicates that the observed compensatory response of the mitochondria up to 300 pM oxidant is itself augmented by injury to the mitochondria. The estimated contribution of the mitochondria to the total cellular ATP  . 8. Panel A: 0, estimated intracellular glyceraldehyde-3-phosphate dehydrogenase (GAPDH) activity (with respect to GAP oxidation rate) modeled from the combined inhibitory effects of HzOz as percent control values; 0, the same analysis taking into account activation of GAP oxidation due to increasing intracellular GAP concentration. Panel B: 0, estimated rate of ADP phosphorylation by the glycolytic pathway after 15 min of exposure to HZ02; 0, estimated rate of net ADP phosphorylation by the mitochondria contributing to the intracellular ATP pool. Panel C, combined estimated ADP phosphorylation rates by glycolysis and mitochondria from panel B.
pool as a function of H202 dose is shown in Fig. 8, panel B, open circles. It appears from the respiratory inhibitor studies that the bulk of the ADP phosphorylated by these mitochondria is hydrolyzed within the organelle matrix. Thus, the activity of the Fo-ATPase appears to be comparable to the Fo-ATP synthase activity of the Fo-ATPase-synthase complex. Indeed, the total intramitochondrial turnover of ATP (2.9 nmol/min/106 cells), estimated from the oligomycin minus atractyloside-inhibited oxygen consumption, is similar to ATP being turned over by the whole cell (2.3 nmol/min/106 cells). It is of interest that P388D1 mitochondria degrade considerably more chemical energy by the phosphorylation/ hydrolysis cycle than they conserve as net exported ATP. This observation may provide an explanation for the increased metabolic rate of tumor cells.
Effects of the Combined H20z-mediated Inhibition of ADP Phosphorylation by Both Glycolysis and Oxidative Phosphorylation-The calculated combined rates of ADP phosphorylation by both the glycolytic and mitochondrial pathways, as a function of H202 exposure, are shown in Fig. 8, panel C. In uninjured cells, this rate (2.1 nmol/min/106 cells) compares very favorably to the estimated rate of cellular ATP hydrolysis (2.3 nmoles/min/106 cells) estimated from Fig. 6.
In order to test whether the observed ATP status following oxidant injury can be reasonably correlated with inhibition of net ADP phosphorylation, it is necessary to perform the following calculations. Total ADP phosphorylation is inhibited by 50% after 15 min of exposure to -300 ~L M H,Oz (1.05 nmol/min/106 cells). Assuming for the moment that ATP hydrolysis (2.3 nmol/min/106 cells) proceeds at the same rate to complete loss of ATP, then the ATP pool (7 nmol/106 cells) will be depleted in 7/(2.3 -1.05) = -6 min, assuming no contribution to the ATP pool from phosphocreatine or from the adenylate kinase equilibrium.
The activity of phosphocreatine phosphokinase (135 microunits/106 cells) is too low in these cells to significantly contribute to ADP phosphorylation over the time course of interest and was found to be somewhat inhibited by H202 exposure. Adenylate kinase activity, on the other hand (-5 milliunits/106 cells), appears to be sufficient to prevent significant accumulation of ADP above control values following Hz02 exposure (23) and was not significantly inhibited by the oxidant. Thus, ATP depletion would be theoretically extended to -11 min by the adenylate kinase equilibrium.
Under experimental conditions where ADP phosphoryla-tion by mitochondria and glycolysis is depressed by 50% (300 ~L M Hz02) the ATP pool is depleted to about 16% of its initial value after 15 min (Fig. 7), and it is of interest to note that almost 50% of the cells do not survive after this time point (Table 7).
Although such calculations give a degree of confidence to the accuracy of the model, it is clearly not reasonable to assume that the rate of intracellular ATP hydrolysis is independent of the intracellular ATP concentration during Hz02 injury. For example, when ADP phosphorylation is blocked by inhibitors (Fig. 6), the proportionality of ATP decline, as a function of time, is lost when ATP falls to 20% of its initial value, presumably because the activity of enzymes hydrolyzing ATP is sufficiently inhibited by ATP availability. Thus, a reduced rate of ADP phosphorylation can theoretically maintain lower steady-state levels of ATP and probably explains why H202 dose-dependent inhibition of ADP phosphorylation is not accompanied by proportional decreases in the measured ATP status (Fig. 7), a t least at the 15th min of injury.
It has yet to be determined whether loss of intracellular ATP is the sole determinant compromising cell viability during oxidant stress. Hepatocytes appear to tolerate chemically reduced ATP levels at 15-20% control values (58). It is perhaps more likely that other perturbations to cellular biochemistry, occurring during oxidant exposure, act synergistically with ATP loss to cause cell death (11-15).
Since ATP levels continue to decline after the 15th-min time point, especially in the continuing presence of the oxidant (23), it is not unreasonable to speculate that the decline in ATP will be at least sufficient to compromise cellular homeostatic processes dependent on ATP. Especially important in this regard are those processes protecting the cell against oxidants and those involved in restoring perturbations induced by the injury (11). We have recently reported' that falling intracellular ATP directly contributes to destabilization of F-actin filaments associated with oxidant-mediated changes in cell morphology (15). It is also of interest to note that the glycolytic and mitochondrial ADP phosphorylating pathways are themselves dependent on ATP for the phosphorylation of glycolytic intermediates and p-keto acids, respectively. Thus, a critical intracellular ATP concentration may be reached where ADP phosphorylation becomes rate-limited by ATP availability, in which case return to ATP homeostasis D. B. Hinshaw, B. C. Armstrong, and P. A. Hyslop, Am. J. Pathol., submitted for publication. may be impossible. This event may mark the watershed for cell death.
The active site of glyceraldehyde-3-phosphate dehydrogenase is known to have an essential thiol (Cys-149), the modification of which results in loss of enzyme activity (59, 61). The Fo-ATPase from Trypanosoma cruzi has been shown to be inhibited by oxidants i n vitro (60).
We have previously reported that bovine aortic endothelial cells (as a model for the target cells during oxidant injury in adult respiratory distress syndrome) behave in a similar manner to P388D1 cells, in terms of the effects of H202 on intracellular ATP (23), consistent with their metabolic similarity (44). Not all cell types may have the same sensitivity to H,O, exposure. For example, cells with a lower dependence on the glycolytic pathway for ATP generation and also a larger pool of mitochondria capable of increasing their compensatory rate of ADP phosphorylation to a greater degree than P388D, cells would be expected to be more resistant to oxidant injury.  HATERIAIS AND METHODS preparation for experimentation is described elsewhere (11).
Cell Culture. Conditions for the culturing Of Pl88D1 cells and determined by withdraving 100 "1 aliquats of cell suspension (2.2~10 c e l l s / m L l . Cells were raoidlv Oelleted in a Beckman microfuoe. and 10 "1 Of Glucose uvtake measurements. Net removal Of glucose from the m e p a vas the sbpektant was added-to i.i mI of an assay medium containing 0.5 U/ml glucose oxidase and a cocktail far detecting the H 0 generated by the oxidation of glucose exactly as described I n ref. 7 3 3 ) . A standard curve was mol/liter. Onioen uptake studies. Cell suspensions (2x106 mL"I7wergl chamber. Oxygen content was monitored polargraphically utilizing a and a method for exoerimental verification. Addition Of strestyloside to Calculation or theoretical mitochondrial Synthetic rate9 of ATP in vivo mitochondria respiring within intact cells arrests ADP translocation across the inner membrane of the Organelle. Concomitantly. oxygen sonaumption coupled to ADP phosphorylation associated with net ATP export from the mitochondria is inhibited. These conditions mimick classic State 4 respiration, which contains components Of oxygen consumption coupled to ion transport, matrix ATP hydrolysis and molecular slip. Oligomycin, on the other hand, inhibits the ATPaiSe-Bynthase complex and therefore the oxygen consumptian inhibited by this agent additionally'measllrea the component of State 4 Coupled to ADP phoBphorylation associated with matrix hydrolysis of ATP (45). glyCOlytiC pathway alone (5.5 mM 2-a)c) and also inhibition of both ~~ ~ ~ were assessed by sampling cella every 30 secande following inhibitor glycolytic and oxidative phosphorylation pathways (5.5 rm( W G + oligomycin) phosphorylation, the rate Of fall reflects ATP hydrolysie by metabolism and addition. In the caee of inhibitior, of both pathways for ADP phosphorylation Of 2-WG. Subtracting the latter component (estimated from the measured net uotake rates of qlucose from the medium1 Yields a rate of ATP cataboliam =lover than the re& rate of hydrolysis 0 ; iTP-&der"~-~~ steady-state due to adenylate kinase activity. However the contribution of the mitrochondria to ADP phosphorylation (providing it I s not the dominant Pathway phosphorylating ADP as is the ce8e in P388D1 cells) can be approximately estimated froh the difference in initial rate when oligomycin conditions should be maximallv stimulated hv the rfaa f n ln~racellular ADP is omitted from tho incubation. The mitrochondrial ATP ovtput under these was measuIed, in Order to validate this methodology for assessing net g~y l y t i c conversion Of glucose to lactate. Cells were incubated with U C glYCOIe + ( 1 mM) for 10 min. At the end Of his equilibration period, cell; ;ere pelleted and taken up I" fresh "c glucose media i HaOs., The assay was termmated after a further 10 nin. by pelleting the

i. 1 s
The lactate in the media was assayed utilizing lactate dehydrogenase (Sigma1 and hydrazine buffer (pH 9.2).
Lactate was quantified by measuring NADH fluorescence produced after 24 hr. incubation using lactic acid Standards (Sigma). The specific activity of the lactate was detemined by incubation Of an allquot Of the supernatant with lactate mancoxygenase Estimation of the impact of the CytoSOliC NAD+ and NADH concentration a number bf experimental Observables and assumptions: measurement of the cytosolic content of NAD+ and NADH; estimation Of the bound/free ratio of these components in the CytOSoli and measurement of the cytosolic volune. These parameters yield the apparent flee cytosolic nicotinamide nucleotide concentrations within the Cytosol. The effects Of manipulating the nicotinamide nucleotide p w 1 an GAPDH activity were assessed at saturating GAP and Na HAS0 . The measuremept of the initial velocity of GAP oxidation 1" th8 presence of NAD will be competitively inhibited in the presence Of NADH (401, and therefore the estimated rate of GAP oxidation (V) with respect to V (saturating NAD' GAP Na HAS0 Zero NADH) can be Calculated froPathe Dixon expressibn f& cdpetitive inhibition: Pellet and superpatant suspensions were then divided into 2 aliquats for determination of NAD and NADH, the supernatants containing CYtoSO11C nicotinamide nucleotides, and the pellets containing the mitochondrial fraction. One of each of the samples was acid treated while the Other wais treated extracts were then assayed for niptinanide nucleotides by the enzyme alkaline treated. exactly ais described by Jorgensen and Rasmvssen (51). The recycling assay Of Kat0 et al. (52). NAD and NADH and mixtures Of both Vere taken through the entire procedure as standards. Recovery was greater than 9 0 1 with no cross-interference between the two nicotinamide nucleotides.
cells Were reruspgnded In 115 mM KC1. 5 mH EGTA, 2 mM MgCl 1 mM d T , 10 nN HEPES p~ 7.4 at 4 c, in a total volume of IO m~. The cell2 were nitrogen cavitated, and homogenized in a tissue grlnder, and centrifuged at 100,000 xg for 30 ai". The supernatant vas dialyzed far 4 8 hrs. against 1 changes Of 2 volume in the dialysis sac "89 reduced by pressure diaqysis. to 2 nL, litres Of buffer containrng 50 mg Charcoal in a second dialysis Sac. The corresponding to the Original Cytosolic volume Of 2x10 cells. The recovery Of lactate dehydrogenase was greater than 90%. The reconstituted cytosol was divided into 2 aliquots. p d i n g studies were performed in 10,000 M.W. cut-off centricon tubes.
H NAD was obtained from NEN., and binding Was performed by adding 10 uL 'H NAD' to the centr-fcon, Such that the NAD+ concentration was 2 uM radiola eled with 100 nCi H. The rpm. Under these conditions 50 u1 buffer was eluted from thg centricon. 10 centricon Was incubated for 20 rain at 4 ' C , and then centrifuged at 5,000 determine bound/free ratio. Following this. the renaininq eluate "a8 u1 of both eluate and reconstituted CytoSol were counted for H in order to Kearurement of bound/free ratio of NAD' avd NADH. 2x109 P388D