1 Introduction

In 1973 Boveris and Chance [1]found that an uncoupler of oxidative phosphorylation (CCCP) or ADP+P_i inhibit H₂O₂ formation by mitochondria oxidizing succinate and glutamate. Later Liu and Huang [2, 3]reported a very steep dependence of the O^−⋅ ₂ and H₂O₂ production by mitochondria upon the ΔΨ level on the mitochondrial membrane. In their experiment, ΔΨ was decreased by adding different concentrations of malonate which decreases the state 4 respiration rate. Therefore one might suggest that formation of reactive oxygen species (ROS) is a function of respiratory chain electron transport rather than of ΔΨ.

Our group proposed [4-6]that activation of ROS production in state 4, when protonic potential is high and respiration rate is limited by lack of ADP, is due to the fact that some transients of the respiratory chain electron transport, capable of reducing O₂ to O^−⋅ ₂, such as CoQH^⋅, become long-lived. It was assumed that mitochondria are equipped with a special mechanism (`mild uncoupling') preventing protonic potential from being too high.

In this paper, we re-investigate the problem of mitochondrial ROS-protonic potential relationships. It was found that any decrease in the state 4 ΔΨ resulted in a strong inhibition of H₂O₂ formation by mitochondria independently of whether activators (uncoupler or ADP+P_i) or inhibitors (malonate) of electron transport were added to cause a lowering of the ΔΨ. A threshold ΔΨ value required for the elevation of H₂O₂ generation was found.

2 Materials and methods

Mitochondria were isolated from rat heart muscle. Cooled muscles, purified from fat and tendons, were minced and passed through a stainless steel press with holes about 1 mm in diameter. The tissue was then homogenized for 3 min with a Teflon pestle in a glass (Pyrex) homogenizer, the tissue:isolation mixture ratio being 1:8. After the first centrifugation (10 min, 700×g), the supernatant was decanted and filtered through gauze, then centrifuged (10 min, 16 000×g). The sediment was suspended in 1 ml isolation mixture (250 mM sucrose, 10 mM MOPS, 1 mM EDTA, pH 7.4) supplemented with BSA (3 mg/ml). Then the mixture was centrifuged once more and washed with the isolation medium without BSA (10 min, 16 000×g). The final mitochondrial sediment was suspended in the isolation mixture (70–90 mg protein/ml) and stored on ice.

In all the experiments shown in the figures, the above procedure was supplemented with the treatment of mitochondria by 2 mM H₂O₂ and 7 mM aminotriazole which were added to the supernatant before the second centrifugation.

Oxygen consumption was recorded with a Clark-type oxygen electrode and a Rank Brothers polarograph.

The ΔΨ level was estimated using safranin O. The ratio of dye (nmol) to mitochondrial protein (mg) was 20:1. It should be noted that the safranin O response has been shown to linearly depend on ΔΨ in the region 50–170 mV [7-9].

H₂O₂ generation was measured by the scopoletin/horseradish peroxidase method [10]. Scopoletin fluorescence response was calibrated with glucose/glucose oxidase [11]and by adding known amounts of H₂O₂. In 1, 2, hydrogen peroxide production rates are expressed as percent of the rate which was obtained in state 4 mitochondria. This rate was 0.14±0.04 nmol H₂O₂/min/mg protein.

Oligomycin, antimycin A, SF6847, and rotenone were dissolved in twice distilled ethanol. All reagents were first tested to avoid interference with the scopoletin/peroxidase assay.

EDTA, EGTA, rotenone, and scopoletin were from Serva; ADP, oligomycin, MOPS, fatty acid-free BSA, catalase, antimycin A, and aminotriazole were from Sigma; 30% H₂O₂ was from Merck.

3 Results

In the literature there is no consensus about the ability of mitochondria to produce measurable amounts of ROS in state 4. Some authors describe such a production [1, 10-12], others fail to do this [13]or succeeded under special conditions (e.g. only in the case when high succinate concentration was used [14]).

In our hands, the rate of H₂O₂ production in state 4 was negligible but increased during storage of isolated mitochondria. It was suggested that such an increase is due to exhaustion of endogenous mitochondrial antioxidants (e.g. reduced glutathione). To stimulate this process, we modified the procedure of isolation of mitochondria to include a short H₂O₂ treatment at the stage of washing of mitochondria with the BSA solution in the cold. The washing mixture was supplemented with 2 mM H₂O₂ and 7 mM aminotriazole, the catalase inhibitor. This modification gave mitochondria which always produce H₂O₂ at a measurable rate when oxidizing 6 mM succinate in state 4. The H₂O₂ treated mitochondria show a 3.5–4-fold stimulation of state 4 respiration by uncouplers.

Fig. 1 shows the effects of uncouplers and respiration inhibitors on the H₂O₂ production by the rat heart mitochondria pretreated with H₂O₂ and aminotriazole. It is seen (Fig. 1, upper curve) that a very low concentration of uncoupler (1×10⁻⁹ M SF6847) strongly inhibits state 4 H₂O₂ production. This effect seems to be due to the SF6847 uncoupling activity rather than to other possible effects of this compound since 6-ketocholestanol, a recoupler specific for SF6847 and carbonylcyanide phenylhydrazone derivatives [15, 16], caused a transient activation of the H₂O₂ generation inhibited by SF6847. Subsequent antimycin A addition strongly stimulated the H₂O₂ production, this fact confirming the original observation of Boveris and Chance [1].

Hansford and coworkers [14]reported that ROS production by state 4 mitochondria in the presence of a high succinate concentration is arrested by rotenone. They assumed that NADH-CoQ reductase is responsible for all the ROS formation by the respiratory chain. We found (Fig. 1, lower curve) that rotenone is really inhibitory for production of H₂O₂ by state 4 succinate oxidation but a measurable rate (about 20%) of H₂O₂ formation is still observed after addition of a saturating rotenone concentration. After rotenone, additions of SF6847 and antimycin A were inhibitory and stimulatory, respectively. In line with some previous observations (for review, see [17]), myxothiazol added after antimycin A strongly inhibited the H₂O₂ production.

The state 4-state 3 transition also inhibited H₂O₂ generation by heart mitochondria. Fig. 2 shows that addition of P_i and ADP almost completely abolished H₂O₂ production. Inhibitors of ATP synthase (oligomycin) or of ADP/ATP translocase (carboxyatractylate) prevented the ADP effect.

The effects of various concentrations of SF6847 on mitochondrial respiration, the ΔΨ level, and the rate of H₂O₂ generation are compared in Fig. 3 . It is seen that H₂O₂ formation is the most sensitive parameter. For instance, 8×10⁻¹⁰ M SF6847 caused a small decrease in membrane potential (about 13%), a 2.2-fold increase in respiration rate and about 80% inhibition of H₂O₂ production.

A very steep dependence of H₂O₂ formation upon ΔΨ is illustrated by Fig. 4 . The ΔΨ level was varied by adding ADP and P_i or various concentrations of SF6847 or malonate.

4 Discussion

The above data show that rat heart mitochondria exhausted in their antioxidants by a short H₂O₂ treatment at low temperature can be used as a model to study ROS generation in state 4. This generation was found to strongly depend upon the ΔΨ level on the mitochondrial membrane. As is seen in Fig. 4, there is a threshold ΔΨ value above which even a small increase in ΔΨ gives rise to a large stimulation of the H₂O₂ production by mitochondria. This threshold is above the state 3 ΔΨ level. Importantly, such relationships could be shown by varying the ΔΨ values by agents causing both stimulation and inhibition of the respiratory chain electron transport. Stimulation was caused by uncouplers (CCCP [12]or SF6847, this paper) or ADP+P_i ([18]and this paper). As for inhibition, it was induced by malonate ([2, 3]and this paper). Thus, under the conditions used, ROS formation is a function of ΔΨ rather than of the electron transport rate. The described effect can be explained assuming that CoQH^⋅ and some other electron transport intermediates competent in the one electron O₂ reduction become long lived when ΔΨ is sufficiently high. In the Q-cycle this is a consequence of inhibition by ΔΨ of the heme b_l →heme b_h electron transport step. This results in complete reduction of b_l which now cannot oxidize CoQH^⋅ formed in the Q-cycle [5]. Apparently, similar relationships are also inherent in NADH-CoQ reductase producing a large portion of ROS generated by state 4 mitochondria [14].

It is not surprising that some special conditions such as H₂O₂ and aminotriazole pretreatment and high succinate concentration are required to obtain measurable H₂O₂ generation by mitochondria in state 4; otherwise the cell would be poisoned by ROS in any case when ADP was exhausted. Such exhaustion occurs quite often, e.g. in muscle under the activity-to-rest transition. This is why we postulated [4-6]the existence in mitochondria of a mechanism (called `mild uncoupling') which makes the maintenance of protonic potential impossible at a level essentially higher than the threshold value critical for fast ROS formation (Fig. 4). The mechanism in question was assumed to be under thyroid control being related to non-ohmic resistance of the mitochondrial membrane [5, 6].

Non-ohmicity of the inner mitochondrial membrane (strong increase in H⁺ conductance at high protonic potential levels) was suggested to be the first line of antioxygen defence of the aerobic eukaryotic cell. It fails to save the cell, e.g. when the respiratory chain is inhibited by antimycin A or (in the case of NAD-linked substrates) by rotenone. Such inhibition is caused by a rather wide range of hydrophobic xenobiotics.

On the other hand, the antimycin A-induced H₂O₂ generation is strongly inhibited by myxothiazol ([17]and this paper) which prevents the CoQH₂ oxidation to CoQH^⋅ in the Q-cycle. The same effect is caused by cyanide [19]and extraction of cytochrome c ([17]and Simonyan, Skulachev and Starkov, in preparation). Moreover, intermembrane cytochrome c can (i) organize a bypass of the superoxide-producing steps of the respiratory chain via NADH-cytochrome b ₅ reductase and cytochrome b ₅ in some tissues, and (ii) oxidize O^−⋅ ₂ to O₂, preventing H₂O₂ formation.

It was recently reported that release of mitochondrial cytochrome c is a physiological phenomenon involved in apoptosis [20-24]. On the other hand, we have suggested that ROS-induced apoptosis, a mechanism of purification of a tissue from ROS-producing cells, represents the last line of antioxygen defence of organisms [4-6]. Thus the following chain of events may be postulated to be involved in antioxygen defence: (i) mild uncoupling, (ii) release of cytochrome c, (iii) cytochrome c-mediated apoptosis.