Stoichiometry of Mitochondrial H+ Translocation Coupled to Succinate Oxidation at Level Flow* of

The mechanistic stoichiometry vectorial H+ trans- location coupled to succinate oxidation by rat liver mitochondria in the presence of a permeant cation has been determined under level flow conditions with a membraneless fast responding 0, electrode kinetically matched with a glass pH electrode. The reactions were initiated by rapid injection of O2 into the anaerobically preincubated test system under conditions in which interfering H+ backflow was minimized. The rates of O2 uptake and H+ ejection, obtained from computer-fitted regression lines, were monotonic and first order over 75% of the course of O2 consumption. Extrapola- tion of the observed rates t.o zero time, at which zero A& and thus level flow prevails, yielded vectorial H+/ 0 flow ratios above 7 and closely approaching 8. The mitochondria undergo no irreversible change and give identical H+/O ratios on repeated tests. In a further refinement, the lower and upper limits of the mechanistic H+/O ratio were determined to be 7.55 and 8.56, respectively, from plots of the rates of 0, uptake uersus H+ ejection at increasing malonate and increasing val- inomycin concentrations, respectively. It is therefore concluded that the mechanistic H+/O ratio for energy- conserving sites 2 + 3 is 8, in confirmation of earlier measurements. KC1 concentration is critical for maxi- mal observed H+/O ratios. Optimum conditions and possible errors in determination of mechanistic H+/O translocation ratios are discussed.

The mechanistic stoichiometry of vectorial H+ translocation coupled to succinate oxidation by rat liver mitochondria in the presence of a permeant cation has been determined under level flow conditions with a membraneless fast responding 0, electrode kinetically matched with a glass pH electrode. The reactions were initiated by rapid injection of O2 into the anaerobically preincubated test system under conditions in which interfering H+ backflow was minimized. The rates of O2 uptake and H+ ejection, obtained from computerfitted regression lines, were monotonic and first order over 75% of the course of O2 consumption. Extrapolation of the observed rates t.o zero time, at which zero A& and thus level flow prevails, yielded vectorial H+/ 0 flow ratios above 7 and closely approaching 8. The mitochondria undergo no irreversible change and give identical H+/O ratios on repeated tests. In a further refinement, the lower and upper limits of the mechanistic H+/O ratio were determined to be 7.55 and 8.56, respectively, from plots of the rates of 0, uptake uersus H+ ejection at increasing malonate and increasing valinomycin concentrations, respectively. It is therefore concluded that the mechanistic H+/O ratio for energyconserving sites 2 + 3 is 8 Previous reports from this laboratory have concluded that the mechanistic stoichiometry of H' translocation coupled to oxidation of succinate by mitochondria from various animal tissues is 8 H'/atom of oxygen reduced (1)(2)(3)(4)(5)(6)(7). Independent kinetic and thermodynamic studies also support an H'/O ratio of 8 (8)(9)(10)(11)(12)(13)(14). However, some laboratories have supported values close to 6 (15-20), and others (21)(22)(23) continue to adhere to the value 4 observed with the original oxygen pulse method (24). Because of the fundamental implications of the H+/O stoichiometry with respect to the mechanism and site of the H+-translocating reactions in the respiratory chain, this disagreement requires resolution.
This paper describes our "second generation" measure-* This work was supported in part by Grant GM05919 from the National Institute of General Medical Sciences. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisenent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Fellow of the Consejo Nacional de Investigaciones Cientificas y Tecnicas de la Republica Argentina.
Supported by Grant CA25360 from the National Cancer Institute.
To whom reprint requests should be sent.
ments of the stoichiometry of H' ejection coupled to the oxidation of succinate by rat liver mitochondria. They represent substantial refinements in both measurement techniques and experimental design. Determination of the mechanistic H+/O stoichiometry of electron transport from rates is subject to two types of uncertainty: 1) technical errors in measurement of H' and oxygen changes, and 2) inadequate experimental design. It has been claimed, for example, that H+/O ratios approaching 8 for succinate oxidation, as this laboratory and others have reported, are overestimated because of underestimation of the rate of 0 2 uptake with Clark-type O2 electrodes (20,22,23). However, most of the earlier measurements from this laboratory employed 0, electrodes fitted with stretched thin membranes, which greatly shorten the response time; moreover, the H+/O ratios were determined at rates of O2 uptake within the capabilities of the electrodes used (1)(2)(3). The mechanistic H+/O ratio may also be underestimated because of the failure to account for or to minimize energy leaks. Such leaks, which can take the form of H' backflow, cycling of other cations (Ca2+, K+), and pump slippage (ll), are responsible for state 4 respiration. Such leaks cannot be totally inhibited, cannot be readily measured, and are not constant in rate. Energy leaks constitute a major problem in establishing the true mechanistic stoichiometry of oxidative phosphorylation. Indeed, the long accepted P/O ratio of 3 for electron transport from NADH to oxygen (cf. Ref. 25) has recently been questioned, and values of 2.0 (26), 2.5 (181, and 2.67 (15) have been proposed. Energy leakage in the form of H+ backflow is especially critical in direct measurement of the vectorial H+/O ratio from the rates of H' ejection and 0 2 consumption in the presence of a permeant cation, since the rate of H+ backflow increases as H' is ejected and ApH, the driving force for H+ backflow across the membrane, increases.
However, H+ backflow can be minimized or allowed for by appropriate recognition of the kinetic and thermodynamic factors influencing the stoichiometry of processes coupled to electron transport (11,(27)(28)(29)(30)(31)(32)(33)(34)(35). One such approach, taken in this study, is afforded by initial rate measurements extrapolated to the state of level flow, i.e. the condition in which electron transport is opposed by zero resistance in the form of ApH; under these conditions, the coupling coefficient is near-maximal and H+ backflow is at a minimum (32, 33).
The H+/O stoichiometry of succinate oxidation reported in this paper is the first to be obtained with: ( a ) response matched electrodes, validated against a reaction of known stoichiometry; and ( b ) reaction conditions arranged to obtain the H'/O ratio at level flow. In addition, a refinement was introduced which permitted determination of the upper and lower limits of the mechanistic H'/O ratio for succinate oxidation. From the data reported here, it is concluded that the mechanistic H+/O translocation ratio for succinate oxi-dation by rat liver mitochondria is 8. Preliminary reports of some of the data have been made (7,37,38). A. -5

EXPERIMENTAL PROCEDURES
The rate measurements were carried out in a closed, temperaturecontrolled chamber (1.35 or 1.66 ml) resembling that in Ref. 39. When filled with the reaction system, it has no gas phase. The reaction chamber was fitted with a fast responding 0 2 electrode and a combination pH electrode (Beckman Altex 531167). The cell stopper also contained a narrow Teflon port for making additions. The reaction system was stirred with a Teflon-coated magnetic bar driven at -1500 rpm. The stirring rate and cell geometry required many careful empirical adjustments to minimize mixing time, electrode response time, and noise of mechanical origin. Within the narrow range of 0 2 concentration changes used (0-50 nmol of O/ml at 25 "C), there was no significant "cross-talk" interference between the responses of the H* and O2 electrodes arising from use of a common reference cell (provided by the combination pH electrode); in any case, this type of interference could be controlled by an external bucking circuit for precise compensation of the current generated by the 0 2 electrode. In some experiments, separate reference electrodes were used and showed responses to H' and O2 changes identical with those observed with a common reference cell and compensating circuit. The electrode signals were suitably amplified and fed into a Soltec model 330 multichannel recorder (full scale response time, -250 ms) usually run at a chart speed of 120 cm/min. Data accessing and processing were carried out with a Hewlett-Packard 9872C digitizer and plotter, 694UB multiprogrammer, and 9845B computer. Line fitting was carried out by regression analysis; in the case of semilogarithmic plots, the lines were fitted by a pseudo-linear approximation method (40).
The 90% response time of the fast responding 0 2 electrode depends upon the number of layers of sintered glass applied to the platinum tip (41, 42). Although it can be made as short as 1 ms, for the experiments described here, the number of layers was adjusted so that the response time matched closely that of the H+ electrodes used.
The reliability of the matched electrodes and the recording system was then validated against the rates of H+ and O2 changes in two scalar reactions in which there is a fixed stoichiometric relationship between H' and O2 changes NADH + H+ + $4202 NAD' + HZ0 (1) and 2 ferrocytochrome c + 2 H' + 1/z02 2 ferricytochrome c + Hz0 (2) Fig . 1A shows the traces of a representative electrode validation test using scalar Reaction 2 as described in Ref. 36. A known amount of 0 2 was injected into an anaerobic suspension of rat liver mitoplasts, with ferrocytochrome c as electron donor, in the presence of antimycin A, rotenone, and FCCP,' the latter to cancel vectorial H+ translocation. Scalar uptake of 0 2 and H' proceeded at rapid but declining rates, owing to formation of ferricytochrome c, which inhibits oxidation of ferrocytochrome c (43). Both rates as recorded were very close to simple exponentials up to -90% of the reaction course, as shown in the linear plots ( Fig. 1B) of log rate uersu.s time, constructed after converting points from the unsmoothed traces into digital form at 0.5-s intervals. From the means of 20 points on the plots, the calculated average H+/O rate ratio was 2.03. Also shown (Fig. 1C) is a plot of the accumulative H+/O uptake ratio with time, which gave a value of 1.98, also close to the theoretical value 2.00. Frequent checks of the electrode response times were made by addition of appropriate small amounts of 0.1 N HCl containing dissolved 0 2 to the standard test medium (mitochondria not included) and with the recorder run at the maximum chart speed (120 cm/min). Rat liver mitochondria were isolated from homogenates in unbuffered 0.25 M sucrose or in H medium (44). Oxygen solubility in the test media was determined from rate measurements of H+ uptake during the course of Reaction 1 in noneoupled NADH oxidase preparations, with NADH present in excess over the dissolved oxygen in   (46). For the airequilibrated standard medium used here (50 mM KCl, 200 mM sucrose, and 2.0 mM Hepes a t pH 7.05), the 0 2 solubility was 480 nmol of O/ ml a t 25 "C and 760 mm; at other temperatures, the O2 solubility was determined from a standard plot.
Over the range of O2 concentrations used (0-25 WM), the fast tion. The O2 electrode was calibrated by addition to the antimycin A-responding O2 electrodes showed a linear response to O2 concentrainhibited anaerobic system of known amounts of dissolved oxygen in aliquots of the standard KCl/sucrose/Hepes medium (above) equilibrated against air for extended periods at the temperature of the H' / 0 measurements. Calibration of the pH electrode was made at the end of each H+/O ratio measurement, after AH+ backflow was largely complete, by recording the immediate electrode response given by addition of internal standards of 50-100 nmol of HCl, within the range of the AH+ changes observed in the experiment. The glass pH electrodes used undergo a significant increase in their 90% response times after several repeated H+/O ratio measurements, presumably because of binding of protein material; they can be "rejuvenated by placing them in 0.1 PI NaOH for -30 s at room temperature, followed by 1-2 min in 0.1 N HC1 and exhaustive washing with water. However, with long use and many rejuvenation treatments, the response time of these pH electrodes increases irreversibly.

Rationale and
Optimization of Conditions-The basic plan of most of the H+/O ratio measurements was to initiate electron flow from succinate to oxygen by rapid injection of a known small amount of oxygen into an anaerobic suspension of de-energized rat liver mitochondria in the presence of succinate, K' , and valinomycin; the ensuing 0, consumption 4804 H"/O Ratio of Succinate Oxidation The test system, optimized from earlier experiments (1-7, 361, consisted of a basic supporting medium of 50 mM KC1 + 200 mM sucrose + 2.0 mM K'Hepes, pH 7.05, supplemented with succinate and rotenone to prevent electron flow from malate and endogenous site 1 substrates, N-ethylmaleimide to prevent fast reuptake of H' with H2POi on the H+/H2PO; symporter (1,47), and valinomycin to allow K' to enter in exchange for the ejected H+. As will be shown, Ca2+ can replace K+ + valinomycin as charge-compensating cation (cf. Ref. 1). Although oligomycin was added in most experiments to inhibit synthesis or breakdown of ATP, its addition generally had little perceptible effect on the H+/O ratios under the conditions described here. In any case, excess oligomycin was avoided because of its capacity to induce chloride-dependent uncoupling (48). It is essential that the test medium contains no added phosphate or other permeant H+-carrying anions, such as acetate, since their entry in response to the alkaline-inside ApH generated by electron flow (49,50) causes underestimation of H' ejection. Although Mg2+ is often added in H+/O determinations because it inhibits endogenous K'/ H' exchange (51), it also inhibits valinomycin-induced K' uptake by rat liver mitochondria (52) and excesses must therefore be avoided. In any case, little if any beneficial effect of Mg2+ up to 2.0 mM was noted in the experiments reported here, and it was not always added. That Aj&p is essentially zero at the time of 0, injection in the experiments described here is ensured by the presence of valinomycin, which collapses A$, and by the relatively long anaerobic preincubation period of several minutes. As will be shown, the pH gradient across the membrane in the presence of valinomycin decays rapidly, with a tl,z of about 11 s at 25 "c.
Initiation of the reaction by addition of succinate to an otherwise complete aerobic system was less satisfactory, since equilibration of succinate across the membrane via the dicarboxylate carrier, presumably in exchange for matrix P? , was relatively slow; furthermore, entry of succinate under these conditions was accompanied by a transient alkalinization owing to the appearance of Pfin the medium, pH 7.05. When electron flow is initiated by addition of valinomycin, the mitochondria are then already respiring in state 4, i.e. static head, in which the leak rate is maximal; in this case, the H' / 0 ratio was accordingly found to be significantly depressed.
Thus, in order to achieve the most reproducible early reaction rates, with least leakage and interference from extraneous H' movements, the reactions were started under anaerobic conditions by injection and rapid mixing of 02. In other experiments (data not shown), electron flow could be initiated by addition of excess EGTA to an aerobic system previously incubated with succinate and a low concentration of Zn", which almost completely blocks electron flow through site 2 (7,53). Addition of EGTA chelates the Zn2+ with high affinity and almost instantly releases the inhibition of succinate oxidation.
Finally, it must be pointed out that when electron flow is initiated in an anaerobic system with succinate as electron donor, not all the electrons already residing in the reduced respiratory chain will pass through all the H+-translocating sites between succinate and oxygen. Thus, the electrons already present in the redox centers of reduced cytochrome oxidase, which will emerge first when oxygen is introduced, will have passed through only the H+-translocating sites of cytochrome oxidase, but not through those of site 2. Therefore, the observed H+/O ratio with succinate at the very beginning of O2 reduction can be expected to be lower than the H+/O ratio prevailing when all the electrons reducing oxygen will pass through all the H+-translocating points in had been consumed. When the O2 became exhausted, there was a sudden increase in the rate of net H' reuptake, presumably because H+ reuptake was then no longer opposed by H' ejection. H' backflow then proceeded at a monotonic first order rate until the trace attained a stable end value exactly at the original H' base-line, with a half-time of 11 s. Thus, the H+ ejected during the consumption of 0 2 was completely reabsorbed again, via membrane leaks, presumably in exchange for matrix K' leaving on valinomycin. The precise return to the original H+ base-line indicates that exactly equivalent amounts of H+ ejection and K+ influx took place during succinate oxidation, followed by reversal of flows in the subsequent anaerobic period. The net amount of H' ejected at the peak was 32.8 nmol of H+/mg of protein, well within the matrix buffering power of about 50 nmol of OH-/mg of protein for the interval pH 7.0 + 8.0 (55). Fig. 3 shows the effect of omitting specific components, as well as the effect of cyanide and FCCP, on O2 consumption and H' ejection. In the absence of succinate, both O2 uptake and H' ejection were zero, showing that endogenous substrates made no contribution whatsoever to the electron flow observed with succinate, contrary to the claim that the presence of N-ethylmaleimide evokes oxidation of endogenous NADPH (56). Cyanide (74 p M ) gave about 96% inhibition of 0, uptake and H' ejection. FCCP stimulated O2 uptake slightly, but completely cancelled H+ ejection. When K' and valinomycin were omitted (Li+ was substituted for K' ), the rates of both 0 2 uptake and H+ ejection were only 11% those in the complete system, thus giving a respiratory control ratio Time-With the validated electrodes and the test system exactly as described in Figs. 2 and 3, many measurements of the rates of H+ ejection and 0 2 uptake during succinate oxidation in the presence of K+ + valinomycin were carried out, with the reaction initiated by the injection of a known amount of O2 as in Figs. 2 and 3. Fig. 4 shows the traces of two typical experiments. In Fig. 44, injection of 0 2 (48.2 ng atoms) into the anaerobic system was followed by 0 2 consumption and H' ejection. Net H+ ejection again became maximal before completion of 0 2 uptake, presumably because of the increasing rate of H' backflow as the reaction proceeded. After the initial dead time of about 0.8 s following injection of the 02, during which no 0 2 rate data could be obtained, points from the unsmoothed 0 2 trace at exact 0.2-s intervals were converted into digital form, and the rate law obeyed was analyzed by computer. The decrease in 0 2 concentration in the system was found to follow an apparent first order course very closely, as indicated by the linearity of the semilog plot of 0, disappearance uerssus time, fitted by regression analysis (Fig. 5, plot A). The  ions of H+/s for the system or 1640 ng ions of H+ ejected/ min.mg of protein (Fig. 5, plot B ) . The Fig. 2), but also by the finding that a second H+/O ratio measurement could be carried out on the same mitochondrial system after the first addition of 0 2 had been completely consumed and the resulting H+ ejection had undergone complete back-leakage. The second set of traces (Fig. 4B) shows somewhat lower rates of 0, uptake and H+ ejection, probably due to more complete  Fig. 44, the dead time of about 0.9 s, through which extrapolation of the zero time rate of H+ ejection was carried out, represents 16% of the reaction period of about 5 s that was useful for determination of the H+/O ratio. As was shown above, extrapolation of the O2 uptake rate to zero time is a valid procedure because of the close agreement between the extrapolated zero time O2 and the amount of 0, actually added. In such experiments, the dead time varies from 0.6 to 0.9 s, but cannot be shortened significantly. However, conditions can be arranged to allow the reaction to proceed over a longer period than shown in Fig. 4, so that the dead time constitutes a smaller fraction of the total reaction time. This has been accomplished by slowing down succinate oxidation with appropriate concentrations of cyanide or malonate. The traces of such an experiment with cyanide are shown in Fig. 6. Data from the traces were converted into digital form and processed exactly as in Fig. 4. The linear semilog plots in Fig. 7 again demonstrate that both 0 2 uptake and H+ ejection follow simple exponential kinetics.

The H+/O Flow Ratio in a System with Minimal Dead Time-As can be seen from the traces in
The extrapolated zero time rates were 475 ng ions of H+/min. mg and 59.3 ng atoms of 02/min. mg, to give a zero time H+/ 0 flow ratio of 8.01. Since the initial dead time was decreased from about 16 to less than 7% of the useful reaction course, this experiment significantly reduced the extrapolation uncertainty.  Fig. 4, the rate of utilization of electron transport energy to drive net K+ uptake and H+ ejection will increase. Thus, both J H and Jo will increase as valinomycin concentration is increased. If JL remained constant as valinomycin is increased, the slope of a plot of JH versus Jo would be equal to the mechanistic stoichiometry n. However, the leak rate J L is not likely to remain constant, since under these conditions it will depend on the magnitude of, among other factors, A;H+ , the major driving force for H' backflow in the presence of valinomycin. Since A;, + will become smaller as it is increasingly utilized to drive K+ uptake, JL is therefore likely to decrease as valinomycin concentration is increased. In this case, the slope of a plot of JH versus JO would be expected to be greater than the mechanistic stoichiometry n and would thus set its upper limit. However, if the rate of 0, uptake and H+ ejection during succinate oxidation is experimentally varied by adding in-creasing concentrations of malonate at a fixed concentration of valinomycin, the rate of energy production would be varied rather than its consumption. In this case, both electron flow and H+ ejection will decrease with increasing malonate concentration; AjiH+ would again be expected to decrease, with a consequent decrease in the leak rate. The slope of a plot of JH versus Jo would then be less than the mechanistic stoichiometry n and would set its lower limit. The opposite effects of these agents on the rates of Jo (and JH) thus will yield plots of JH versus Jo whose slopes are, respectively, above and below n, the mechanistic H+/O ratio. Such experiments are described below. .. The points designated 0 were obtained from a series of nine similar experiments in which the KC1 concentration was 2.0 mM rather than 50 mM, in order to obtain more points at relatively low rates of 0% uptake. In this series, 48 mM LiCl was also present. The slope of the plot of JO versus J H was n = 8.56, the upper limit of the H'/O ratio at level flow. FIG. 9 (right). Titration of the rates of 0 2 uptake and H+ ejection with malonate; the lower limit of the H+/O ratio.  Hf/O Ratio of Succinate Oxidation ejection and O2 consumption were computed as in Fig. 5. The paired initial rates from a total of 15 experiments were then plotted against each other (Fig. 8). As can be seen, the data yielded a plot of J H versus Jo, fitted by linear regression, having a slope of 8.56, which, by the rationale described above, is taken as the upper limit of n, the mechanistic stoichiometry of H+ ejection coupled to succinate oxidation.

Experimental Modulation of H+ Ejection by Varying Valinomycin and
In the experiments for the determination of the lower limit of the H+/O ratio (Fig. 9), all components, concentrations, and conditions were fixed except the concentration of malonate, which was varied from 0 to 1.15 mM in order to vary the rate of electron transport from succinate to oxygen. Valinomycin was constant at 120 ng/mg of protein. The reactions were initiated by injection of a fixed amount of O2 into the anaerobic system as in the experiments described for Fig. 8.
The H+ ejection and O2 uptake rate data obtained from the traces were processed in the same way. The paired zero time rates of O2 uptake and H+ ejection at nine different concentrations of malonate were plotted, giving a best line (Fig. 9) having a slope of 7.55, which may be taken as the lower limit of the mechanistic H'/O ratio for succinate oxidation. The average of the nine individual H+/O ratio measurements was only slightly lower, 7.36. From the data in Figs. 8 and 9, it may be concluded that the true mechanistic H+/O ratio for electron transport from succinate to O2 lies between the limits 7.55 and 8.56.

Ca2+
As Charge-compensating Permeant Cation-K+ + Valinomycin can be replaced by Ca2+ as charge-compensating permeant cation, since the electrophoretic Ca2+ uniporter of rat liver mitochondria can function at an extremely high rate (61). In an H+/O ratio experiment arranged exactly as that in Fig. 2, with valinomycin replaced with 3.0 mM CaZ+, the extrapolated zero time H*/O ratio was 7.53, obtained from the kinetic plots shown in Fig. 10.
Influence of the Test Medium-Most of the experiments employing K+ + valinomycin were carried out with a medium of 50 mM KC1 + 200 mM sucrose, which was found to be optimal for H+/O measurements. Fig. 11 shows that in the absence of respiratory inhibitors, the medium must contain at least 20-30 mM K+ in order to yield maximal H+/O ratios; at K+ concentrations below 10 mM, the zero time H+/O ratio was depressed to about half of its maximal value, presumably because the rate of K+ entry via valinomycin is limited by K'  10. ea2+ as permeant cation. The test system was as in Fig. 2 with 3.0 mM Ca2+ instead of valinomycin. The processing of the data, carried out as in Figs. 5 and 7, is shown. The lower plot shows the decline of the H+/O ratio as the reaction proceeds. concentration. Maximal H+/O ratios are also observed when sucrose (200 mM) is replaced with 100 mM Licl or 100 mM NaC1. Although a medium of 150 mM KC1 gives maximal rates of H+ ejection and O2 uptake, the rate of H+ backflow then becomes significantly greater. Presumably, this effect is due to a combination of a maximal rate of K+ entry and the increasing permeability of the inner membrane to chloride as the matrix pH rises during H+ ejection. High KC1 media also cause dissociation of cytochrome c from its inner membrane binding sites into the intermembrane space (62), causing anomalies in respiratory control as well as changes in the properties of the cytochrome oxidase reaction, to be considered elsewhere?

DISCUSSION
The major improvements described here for measurement of the H+/O translocation ratio for succinate oxidation derive in part from the use of a fast responding O2 electrode prepared in such a way as to yield response characteristics nearly identical with those of the H+ electrode; the reliability of the electrodes and recording system was validated in tests utilizing the fixed H+/O stoichiometry of the scalar oxidation of ferrocytochrome c (36) by noncoupled mitochondrial preparations. Although the type of O2 electrode can be prepared to have much shorter response times (1 ms or less), the limiting factor is the relatively long relaxation time of commercial glass electrodes for measuring pH changes. Nevertheless, use of these electrodes was highly desirable since they reliably measure the true bulk phase thermodynamic activity of O2 and H' . Although it is possible to avoid the significant response time of pH electrodes by spectrophotometric measurement of pH changes in the presence of an acid-base indicator, data obtained with the latter approach are compromised to an unknown extent by the binding and/or transport of the indicator by the mitochondria, which can yield spurious pH change signals. Use of the electrodes thus provides a more reliable measure of the bulk phase 0 2 and H' changes.
The experiments described here also recognize that the rate of electron flow from succinate to oxygen is not precisely B. Reynafaje and A. L. Lehninger, unpublished observations. linear with time over any significant fraction of the reaction course, since in the presence of valinomycin the ApH across the membrane steadily rises, owing largely to alkalinization of the matrix as the ejected H+ is replaced by K' of the medium. As ApH rises, it imposes increasing resistance to electron flow and H' ejection, with the consequence that the rate of O2 uptake declines. Under the conditions described here, the observed rates of both O2 uptake and H' ejection declined in a manner that closely approximated a monotonic first order relationship up to 80% or more of the reaction course and thus allowed accurate extrapolation of the rates to zero time, the instant of level flow. True initial rates cannot be obtained simply from the slopes of tangents hand-drawn to the early portion of nonlinear traces, as has been done in most studies based on rate measurements, since this procedure not only underestimates the initial rates (591, but also can introduce subjective errors. The second important feature of the H+/O ratio measurements reported here is use of the principles developed in kinetic and nonequilibrium thermodynamic analyses of electron transport and oxidative phosphorylation (27-35). In particular, these relationships show that the closest approach to the mechanistic coupling ratios can be obtained in two different respiratory states, either from the H+/O flow ratio in the condition of level flow, at which electron transport and energy coupling proceed against zero load or resistance and the leaks are close to zero, or from the measured flows and forces in the condition of static head, i.e. state 4 (32,33). In this paper, we have employed the first of these approaches. Since the H' ejection and 0 2 uptake rates at the transient instant of level flow are not directly measurable, they must be obtained by extrapolation of the measured early rates to zero time. The accuracy of this extrapolation requires that (a) the methods reliably measure reaction rates, ( b ) the measurements begin as early as possible after the reaction is initiated, (c) the reaction rates are essentially monotonic over the measured period, ( d ) zero time is accurately known, and ( e ) A i s + is essentially zero at zero time. Within the limits of the instrumentation and the methods employed, these criteria have been met.
A further refinement was introduced to narrow further the value of the mechanistic H+/O ratio for succinate oxidation, from measurements of its upper and lower limits. This approach was based on Equation 3 which recognizes, as have numerous past studies, that respiratory energy transduction is always accompanied by losses which may take the form of ionic leaks, slippage of pumps, or other dissipative processes. Equation 3 requires no assumptions as to the mechanism or magnitude of such losses, but simply states that the observed H+/O flow ratio is less than the mechanistic H' /O ratio by the rate of the H' leak. From the appropriate kinetic data plotted in Figs. 8 and 9, the upper limit of the H+/O ratio for succinate oxidation was found to be 8.56 and its lower limit 7.55, strongly indicating that the true ratio is the integral value 8. The broader usefulness of this rationale for determination of the upper and lower limits of the stoichiometry of other types of energy-dependent membrane transport events is being developed f~r t h e r .~ The data reported here substantiate H+/O ratios approaching 8 for succinate oxidation obtained with less accurate methods in earlier studies from this laboratory on mitochondria from rat liver (11, rat heart (2), and different lines of tumor cells (3). They are also in agreement with H+/O ratios from other laboratories employing a variety of methods. POZzan et al. (8) and Azzone et al. (9) have reported H+/O ratios approaching 8 for succinate oxidation from rates obtained from tangents to early portions of 0, and H' traces in experiments similar to those described earlier from this laboratory (1-3). H' /O ratios approaching 8.0 for succinate oxidation have also been extrapolated from nonlinear rate data by Ho and Wang (13) and Ting and Wang (14) under conditions similar to those reported here. In a third approach, Lemasters and Billica (12) determined the P/O ratio and indirectly the H+/O stoichiometry of electron transport from thermodynamic measurements of the force ratio (-AG~/AG,,, where -AGR is the electromotive force span of electron transport) at static head in inverted inner membrane vesicles, in which no "proton motive" work is required to translocate adenine nucleotides, and the entire flow of respiratory energy is thus available to generate ATP. They found the ADP/O ratio of NADH oxidation to approach 4.5 in this stringent test, leading to their conclusion that the average H+/site ratio is 4.0, thus giving an H+/O ratio of 8 for succinate oxidation. Recently, Pietrohon et al. (11) have determined the H+/O ratio for succinate oxidation from measurements of both flows and forces in intact mitochondria respiring in different static head conditions. Their comprehensive measurements not only confirmed the H+/O ratio of 8 for succinate oxidation, but also reinforced the intrinsic value of nonequilibrium thermodynamic principles in quantitative analysis of energy coupling during mitochondrial electron transport.
The experiments reported here should not be taken as indicating that H' is normally ejected into the external bulk phase during oxidative phosphorylation nor that the net H+/ 0 stoichiometry is always 8. As is shown in Fig. 3 and in many previous reports, the near-quantitative ejection of H' in stoichiometry experiments such as these occurs only when a permeant cation such as K' or Ca2+ can enter the mitochondria in exchange for the ejected H+, or a permeant anion can leave with the H' . Several recent reports have provided evidence that H' transfer from the electron transport system to FoF,-ATPase takes place largely within or on the inner membrane, rather than (or in addition to) via the outer bulk aqueous phase (see below). In intact cells, respiration takes place somewhere between states 3 and 4, matched to the rate of cellular ATP utilization via the prevailing AGp (33). Under these conditions, there are finite energy leaks and/or slippage, and the coupling coefficient is therefore significantly less than 1.0 and the net H' /O ratio less than 8 (33). Nevertheless, knowledge of the intrinsic or mechanistic stoichiometry of electron transport is essential for establishing 1) which of the electron carriers of the respiratory chain participates in the translocation of H' across, within, or on the membrane; and 2) by what mechanism these processes occur, whether by ligand conduction (ie. loops), by conformational "pumps," or some combination thereof (20, 63). Space does not permit full discussion of those reports (15-24) concluding that the mechanistic H+/O ratio for succinate oxidation is less than 8; they will be reviewed in detail elsewhere! However, oxygen pulse measurements on succinate succinate oxidation from studies of the relative rates of electron flow from site 1, site 2, and site 3 substrates required to maintain a steady state membrane potential in respiring rat liver mitochondria. As they pointed out, this approach cannot yield absolute H+/O ratios, but gives values proportional to the charge/O (q'/O) ratios. Their results indicated that the intrinsic q'/O ratios for sites 1 + 2 + 3, sites 2 + 3, and site 3 could be given by the proportions 4 3 2 or 8 6 4 or 1296. They concluded that the absolute q+/O ratios are probably 8, 6, and 4; thus, the H+/O ratio for succinate oxidation would be 6. However, it may be noted that their ratio 12:9:6 can also accommodate, within experimental error, the directly determined Q'/O ratios of 12, 8, and 6 previously reported from this laboratory (1)(2)(3)(4)(5)65), in which the q+/O ratio for succinate oxidation is 8.
Some authors (19,21,22) have criticized an H+/O ratio of 8 for succinate oxidation as being thermodynamically unlikely on the grounds that electron transport from the succinate/ fumarate couple (Ed = 30 mV) to the HzO/Oz couple (Ed = 820 mV) supplies insufficient or just barely enough energy (a,' = 790 mV, equivalent to 1580 mV/2e-) to eject 8 H' at an assumed ApH+ of 200 mV (required, 1600 mV/2e-). For two reasons, this criticism is subject to question. First, the reported values of ap~+ under state 4 conditions show rather wide variations (18,66) which probably reflect the indirect methods used for measurement of ApH, A+, and matrix volume. Such measurements also may require significantly large corrections (not often applied) for binding of probe ions (an example of such a correction is given in Ref. 67) as well as the assumption that the probe ions come into complete equilibrium with the transmembrane A+ or ApH without affecting the integrity of the inner membrane. A recently described assessment of ApH in state 4 mitochondria by a noninvasive isotopic measurement (68) has shown ApH to be much smaller than indicated by transmembrane distribution of probe ions. Moreover, A~H + measurements on alkalophilic bacteria, which generate a large reversed ApH gradient but a "normal" A+, strongly suggest that ApH plays little if any role in the generation of ATP (69). Further uncertainties are furnished by the numerous recent reports (for review, see Refs. 66, 70-72; see also Refs. [73][74][75] of large quantitative inconsistencies between AbH+ and the rates of electron transport and between ApH+ and AG, and between the rates of generation of & i~+ and ATP, indicating that the properties of the bulk phase AbH+ as usually measured are not consistent with the central role proposed for it by the chemiosmotic hypothesis, i.e. that it is the sole intermediate vehicle for energy transduction between electron transport and phosphorylation. Such anomalies have led several laboratories to conclude that a significant fraction of energy coupling during respiratory chain phosphorylation takes place by localized, more direct energy transfer within or on the membrane rather than through bulk phase chemiosmotic gradients. Because of the now substantial evidence that AjiH+ as usually determined may not be a quantitative measure of the intermediate high energy state, thermodynamic objections to an H+/O ratio of 8 for succinate oxidation are not warranted.