Native rates of superoxide production from multiple sites in isolated mitochondria measured using endogenous reporters

Abstract

Individual sites of superoxide production in the mitochondrial respiratory chain have previously been defined and partially characterized using specific inhibitors, but the native contribution of each site to total superoxide production in the absence of inhibitors is unknown. We estimated rates of superoxide production (measured as H(2)O(2)) at various sites in rat muscle mitochondria using specific endogenous reporters. The rate of superoxide production by the complex I flavin (site I(F)) was calibrated to the reduction state of endogenous NAD(P)H. Similarly, the rate of superoxide production by the complex III site of quinol oxidation (site III(Qo)) was calibrated to the reduction state of endogenous cytochrome b(566). We then measured the endogenous reporters in mitochondria oxidizing NADH-generating substrates, without added respiratory inhibitors, with and without ATP synthesis. We used the calibrated reporters to calculate the rates of superoxide production from sites I(F) and III(Qo). The calculated rates of superoxide production accounted for much of the measured overall rates. During ATP synthesis, site I(F) was the dominant superoxide producer. Under nonphosphorylating conditions, overall rates were higher, and sites I(F) and III(Qo) and unidentified sites (perhaps the complex I site of quinone reduction, site I(Q)) all made substantial contributions to measured H(2)O(2) production.

FIGURE 1. Comparison of rates of superoxide production (measured as H₂O₂ production) by site I_F in CDNB-pretreated and control mitochondria

Site I_F was titrated by adding malate from 0.01 mM to 5 mM in separate runs followed by addition of rotenone. Points were fitted to give the parameter values in Eq. 1 (inset). The dashed line indicates a 1:1 relationship. Values are means ± SEM (n = 5).

FIGURE 2. The theory and assumptions of the reporter-based assay of superoxide production rate

(a) Theory. When the donor species in a superoxide-producing site (ovals) is reduced (darker shading) it donates electrons to oxygen to generate superoxide. The redox state of the donor is reported by an adjacent redox centre (rectangles). Two primary assumptions are made. First, the rate of superoxide production is a unique function (f) of the reduction state of the donor (in the simplest case, the rate of superoxide production is the product of a pseudo first-order rate constant and the concentration or % reduction of the donor). Second, the reduction state of a relevant, nearby reporter is a unique function of the reduction state of the donor (in the simplest case the two centres are at equilibrium and their redox states are related by the Nernst equation). It follows that the rate of superoxide production will be a unique function of the reduction state of the reporter (in the simplest case, that function can be derived from the two simplest cases above). These assumptions allow the rate of superoxide production by a particular donor to be calibrated to the reduction state of the appropriate reporter. In the present study we calibrated the rate of superoxide production from site I_F to the reduction state of NAD(P)H, and the rate of superoxide production from site III_Qo to the reduction state of cytochrome b₅₆₆. (b) Reactions of donor (FMNH₂) and reporter (NADH₂) at site I_F. (c) Reactions of donor (QH^·) and reporter (reduced cytochrome b₅₆₆) at site III_Qo.

FIGURE 3. Maximum observed rates of H₂O₂ production from different sites compared to the native rates observed in this study

Black bars indicate the approximate maximum rates of H₂O₂ production from different sites in CDNB-treated rat skeletal muscle mitochondria, to establish a basis for comparison with the native rates. Data were obtained in standard KCl buffer (see MATERIALS AND METHODS), but lacking phosphate and magnesium (phosphate lowers the rate of H₂O₂ production from both site I_Q and site III_Qo). Site I_F was assayed in the presence of 5 mM malate, 4 μM rotenone, and 4 μM FCCP. Site I_Q was assayed in the presence of 5 mM succinate as the rotenone-sensitive portion of the signal (no correction was made for the changes in rate of H₂O₂ production by other sites on addition of rotenone). Site I_Q data were adapted from [21]. Site III_Qo was assayed with 5 mM succinate and 2.5 mM malonate, in the presence of 2 μM antimycin A, as the myxothiazol-sensitive portion of the signal (no correction was made for the changes in rate of H₂O₂ production by other sites on addition of myxothiazol). Site III_Qo data were adapted from [20]. The final four bars (grey) show the current assay conditions (data from Fig. 8): CDNB-treated mitochondria oxidizing 5 mM malate or 5 mM glutamate plus 5 mM malate in the absence of inhibitors during ATP synthesis (st. 3) or in non-phosphorylating conditions (st. 4); the sites generating H₂O₂ in each condition will be determined in this study (Fig. 8). Data are means ± SEM (n = 4).

FIGURE 4. Design of the reporter-based superoxide assay

The assay was designed to measure the rate of H₂O₂ production using Amplex UltraRed, and the steady-state reduction levels of NAD(P)H and cytochrome b₅₆₆ under closely similar conditions. The timing of all additions was synchronized between the three assays. The addition of substrate, in this case 5 mM glutamate plus 5 mM malate, led to an increased rate of change of Amplex UltraRed fluorescence (a), and increased steady-state reduction levels of both NAD(P)H (b) and cytochrome b₅₆₆ (c). The gray traces in each graph show the control in the absence of substrates or inhibitors. The 100% value for each reporter in (b) and (c) was established by addition of its relevant downstream inhibitor (4 μM rotenone or 2 μM antimycin A), as described in MATERIALS AND METHODS. The horizontal bars in (c) indicate the regions of data that were averaged to give the mean value used to calculate the % reduction of cytochrome b₅₆₆. All of the data for the calibration curves (Figs 5 and 7) and measurements of native rates (Fig. 8) were collected in essentially this way.

FIGURE 5. Relationship between the rate of superoxide production by site I_F and the reduction state of NAD(P)H

(a) Dependence of the rate of H₂O₂ production on malate concentration after addition of 4 μM rotenone. (b) Dependence of %NAD(P)H reduction on malate concentration after addition of rotenone (100% reduction was subsequently established by addition of 5 mM malate). (c) Final calibration of the relationship between the rate of superoxide production from site I_F and NAD(P)H reduction state, obtained by combining panels (a) and (b). Filled symbols: control mitochondria; open symbols: CDNB-treated mitochondria (underlying data for CDNB-treated mitochondria is not shown). Where not visible, error bars are contained within the points. Lines show exponential relationships (for simplicity), fitted by non-linear regression to give the parameter values in Eq. 2. See MATERIALS AND METHODS. Data are means ± SEM (n = 6).

FIGURE 6. Contribution of site I_F during calibration of site III_Qo.

For the calibration curve in Fig. 7c, the rate of superoxide production by site III_Qo was measured as myxothiazol-sensitive H₂O₂ production at different succinate:malonate ratios in the presence of rotenone. (a) Typical Amplex UltraRed fluorescence trace. 2 μM myxothiazol was added where indicated. (b) Corresponding NAD(P)H autofluorescence trace indicating that NAD(P)H becomes reduced under these conditions (5 mM glutamate plus 5 mM malate were added at the end to establish 100% reduction of the NAD(P)H pool. Therefore, correction for the changes in superoxide production from site I_F before and after addition of myxothiazol was required. (c) The mean ± SEM reduction state of NAD(P)H at each succinate:malonate ratio in the absence and presence of 2 μM myxothiazol used to generate the correction using the I_F calibration curve in Fig. 5c (n = 3).

FIGURE 7. Relationship between the rate of superoxide production by site III_Qo and the reduction state of cytochrome b₅₆₆

(a) Dependence of the rate of myxothiazol-sensitive H₂O₂ production on succinate concentration at fixed succinate+malonate concentration in the presence of rotenone. Data were corrected for the contribution of site I_F (Fig. 6c and Fig. 5c). (b) Dependence of cytochrome b₅₆₆ reduction on succinate concentration in parallel incubations (100% reduction was subsequently established by addition of 2 μM Antimycin A). (c) Final calibration of the relationship between the rate of superoxide production from site III_Qo and cytochrome b₅₆₆ reduction state, obtained by combining panels (a) and (b). Filled symbols: control mitochondria; open symbols: control values after correction to CDNB-treated mitochondria using Eq. 1, assuming that 50% of superoxide from site III_Qo was produced in the matrix (see [26]). Where not visible, error bars are contained within the points. Lines show exponential relationships (for simplicity), fitted by non-linear regression to give the parameter values in Eq. 3. See MATERIALS AND METHODS. Data are means ± SEM (n = 9).

FIGURE 8. Reported and measured native rates of superoxide and H₂O₂ production by mitochondria oxidizing NAD-linked substrates in the absence of electron transport chain inhibitors

Data from Table 2. (a) 5 mM malate as substrate under phosphorylating conditions. (b) 5 mM malate under non-phosphorylating conditions. (c) 5 mM glutamate + 5 mM malate under phosphorylating conditions. (d) 5 mM glutamate + 5 mM malate under non-phosphorylating conditions. The two left-hand bars in each panel show results for control mitochondria; the two right hand bars show results for CDNB-treated mitochondria. Reported rates from site I_F are in dark grey; those from site III_Qo are in light grey. Black bars represent measured rates. Data are means ± SEM (n = 6). SEM values for reported rates were determined by error propagation; significance was tested using Welch’s t-test, see MATERIALS AND METHODS.