Membrane potential and delta pH dependency of reverse electron transport-associated hydrogen peroxide production in brain and heart mitochondria

Succinate-driven reverse electron transport (RET) is one of the main sources of mitochondrial reactive oxygen species (mtROS) in ischemia-reperfusion injury. RET is dependent on mitochondrial membrane potential (Δψm) and transmembrane pH difference (ΔpH), components of the proton motive force (pmf); a decrease in Δψm and/or ΔpH inhibits RET. In this study we aimed to determine which component of the pmf displays the more dominant effect on RET-provoked ROS generation in isolated guinea pig brain and heart mitochondria respiring on succinate or α-glycerophosphate (α-GP). Δψm was detected via safranin fluorescence and a TPP+ electrode, the rate of H2O2 formation was measured by Amplex UltraRed, the intramitochondrial pH (pHin) was assessed via BCECF fluorescence. Ionophores were used to dissect the effects of the two components of pmf. The K+/H+ exchanger, nigericin lowered pHin and ΔpH, followed by a compensatory increase in Δψm that led to an augmented H2O2 production. Valinomycin, a K+ ionophore, at low [K+] increased ΔpH and pHin, decreased Δψm, which resulted in a decline in H2O2 formation. It was concluded that Δψm is dominant over ∆pH in modulating the succinate- and α-GP-evoked RET. The elevation of extramitochondrial pH was accompanied by an enhanced H2O2 release and a decreased ∆pH. This phenomenon reveals that from the pH component not ∆pH, but rather absolute value of pH has higher impact on the rate of mtROS formation. Minor decrease of Δψm might be applied as a therapeutic strategy to attenuate RET-driven ROS generation in ischemia-reperfusion injury.


Introduction
There is a large body of experimental evidence demonstrating pathologically enhanced mitochondrial reactive oxygen species (mtROS) production in several diseases such as diabetes, neurodegenerative conditions including Alzheimer's and Parkinson's diseases, diabetes, and ischemia-reperfusion injury; for review see (Beal 1996;Giacco and Brownlee 2010;Chouchani et al. 2014). Respiratory Complex I (CI) is a primary source of mtROS and its dysfunction is thought to be pathologically relevant (Cadenas et al. 1977, Grivennikova and Vinogradov 2006, Treberg et al. 2011. In isolated mitochondria CI-mediated mtROS generation can be initiated under the following conditions: 1) with NADH-linked substrates (such as glutamate and malate), which generate mtROS at a relatively low rate; 2) with NADH-linked substrates in the presence of a CI inhibitor, like rotenone producing high rate of ROS; 3) with FADH 2 -linked substrates like succinate or Zoccarato et al. 2004;Treberg et al. 2011;Orr et al. 2012) alpha-glycerophosphate (α-GP) (Tretter et al. 2007b, c). In hyperpolarized nonphosphorylating mitochondria, FADH 2 -linked substrates generate mtROS at a higher rate by supporting a reverse electron transport (RET) which occurs from Complex II (CII) or alphaglycerophosphate dehydrogenase (α-GPDH) to CI via the Qjunction (Treberg et al. 2011). According to prior reports, succinate-driven mtROS production appears to have the highest rate in isolated murine mitochondria in the absence of ADP compared to NADH-linked substrates initiated mtROS (Korshunov et al. 1997;Kwong and Sohal 1998;Votyakova and Reynolds 2001;Liu et al. 2002;Zoccarato et al. 2011); this is attributed primarily to RET towards CI and partially to the forward electron transport (FET) towards CIII (Grivennikova and Vinogradov 2006;Treberg et al. 2011;Zoccarato et al. 2011;Quinlan et al. 2013). Upon succinate oxidation, in the absence of ATP synthesis, FET secures the energy demands of RET. Rate of RET-associated ROS production is thought to be dependent on pmf which comprises mitochondrial transmembrane potential (Δψ m ) and mitochondrial transmembrane pH gradient (ΔpH) (Liu 1997;Votyakova and Reynolds 2001;Lambert and Brand 2004). It is a well-known phenomenon that high pmf, such as the one measured in the absence of ADP, is required for maintenance of RET. It has been shown that succinate-or α-GP-fuelled RET is very sensitive to minor changes in Δψ m in isolated mammalian (Tretter and Adam-Vizi 2007) and Drosophila (Miwa and Brand 2003) mitochondria.
More specifically, a 10% decrease in Δψ m (caused by an uncoupler agent) gave rise to a 90% decrease in succinatedriven ROS production in rat heart mitochondria (Korshunov et al. 1997). The other component of pmf, ΔpH, also appears to have a regulating effect on mtROS formation. Upon acidification of the matrix, mtROS generation is decelerated, which can be explained by the stabilisation of the semiquinone radicals (SQ .-) (Selivanov et al. 2008). The question arises as to which component of pmf plays the key role in the control of mtROS production. According to Lambert and Brand (Lambert and Brand 2004), succinate-driven ROS production is more dependent on ΔpH than on Δψ m , as detected in mitochondria isolated from rat skeletal muscle. On the contrary, Selivanov and co-workers (Selivanov et al. 2008) revealed that mtROS generation is significantly affected by the actual value of pH itself (extramitochondrial pH; pH extra and intramitochondrial pH; pH in ), and not much influenced by ΔpH or Δψ m , as measured in rat brain mitochondria.
The aim of the present study was to clarify which of the two components of pmf has a predominant role in the control of mtROS formation and to assess whether absolute pH value modulates RET-dependent mtROS production. We also aimed to test whether the effect of Δψ m , ΔpH, and absolute pH values on ROS formation is different in brain compared to heart muscle mitochondria. Δψ m and ΔpH usually change in the same direction; for example, uncoupling depolarisation (decrease of Δψ m ) is generally followed by a decrease in ΔpH as well. With ionophores, like valinomycin and nigericin, it is possible to dissect the two components of pmf: Δψ m and ΔpH can be varied in a different direction. Nigericin decreases pH in (Rottenberg and Lee 1975) and hyperpolarises Δψ m (Selivanov et al. 2008), whilst valinomycin elevates pH in (Selivanov et al. 2008) and depolarizes Δψ m (Selivanov et al. 2008) under specific conditions. In the present study, Δψ m , pH in , and H 2 O 2 production were measured systematically and ΔpH was calculated. To scrutinize Selivanov's theory, pH dependence of the above-mentioned parameters was examined. In contrast to Lambert and co-workers , we concluded that the succinate-driven RET-evoked ROS production is more dependent on Δψ m and less influenced by ΔpH in both guinea pig brain and heart mitochondria. Furthermore, we showed, in agreement with Selivanov and colleagues, that absolute pH rather than ΔpH itself modulates succinate-and α-GPdriven RET. Our results suggest that lowering Δψ m might be an effective solution to reduce the RET-provoked mtROS load in conditions like ischemia-reperfusion where oxidative stress and high Δψ m prevail.

Preparation of mitochondria
Mitochondria were prepared from albino guinea pig brain cortex using a Percoll gradient (Rosenthal et al. 1987,Tretter andAdam-Vizi 2007) and from whole heart using differential centrifugation (Mela andSeitz 1979,Korshunov et al. 1997), as previously described. Animal experiments were performed in accordance with the Guidelines for Animal Experiments of Semmelweis University. A modified biuret method was used to determine mitochondrial protein concentration (Bradford 1976).

Brain mitochondria
The brain was rapidly homogenized in Buffer A (in mM: 225 mannitol, 75 sucrose, 5 HEPES, 1 EGTA, pH 7.4) and centrifuged for 3 min at 1300 g. The supernatant was centrifuged for 10 min at 20,000 g, and the resulting pellet was resuspended in 15% Percoll and layered on a discontinuous gradient consisting of 40 and 23% Percoll. This was centrifuged for 8 min at 30,700 g using no brake. After resuspension of the lower fraction in Buffer A, centrifugation was applied at 16,600 g for 10 min. Pellet was resuspended in Buffer A and centrifuged at 6300 g for 10 min. Subsequently, supernatant was discharged, and the pellet was resuspended in Buffer B (in mM: 225 mannitol, 75 sucrose, 5 HEPES, pH 7.4). All operations above were performed either on ice or at 4°C (Komary et al. 2008).

Heart mitochondria
Mitochondria from heart were isolated following the modified protocol of Korshunov and co-workers (Korshunov et al. 1997). The heart was repeatedly washed in homogenisation buffer (in mM: 200 mannitol, 50 sucrose, 5 NaCl, 5 MOPS, 1 EGTA, 0.1% BSA, pH 7.15) to remove residual blood. Afterwards, it was cut into small pieces with scissors under 2.5 ml homogenisation buffer supplemented with 10 U protease (Protease from Bacillus licheniformis, Type VIII). After adding 17 ml of homogenisation buffer, the preparation was properly homogenised and centrifuged for 10 min at 10,500 g. The supernatant was discharged, the pellet was resuspended in 25 ml homogenisation buffer, and then centrifuged for 10 min at 3000 g. The supernatant was centrifuged for 10 min at 10,500 g and the formed pellet was resuspended in the homogenisation buffer. All operations above were performed either at 4°C or on ice.

Buffers
Depending on the requirement of K + of the applied ionophore (nigericin or valinomycin), one of the following media was applied in the relevant experiments: Standard medium A (high K + content for nigericin; in mM): 125 KCl, 20 HEPES, 2 KH 2 PO 4 , 0.1 EGTA, 1 MgCl 2 , and 0.025% BSA.
Standard medium B (low K + content for valinomycin to avoid mitochondrial swelling; in mM): 240 sacharose, 10 Tris, 2 KH 2 PO 4 , 4 KCl, 0.1 EGTA, 1 MgCl 2 , and 0.025% BSA. pH of the respiratory media was adjusted prior to the measurements, in the absence of mitochondria, with HCl or NaOH to 6.4, 6.8, 7.0, 7.2, 7.4, 7.6 or 8.0. Addition of mitochondria suspended in buffered solution and addition of high concentrations of respiratory substrates (succinate or α-GP) could shift the pH of the incubation medium slightly. In order to calculate an accurate ΔpH, pH extra measured in the presence of mitochondria and respiratory substrate was applied in this study.

Measurement of mitochondrial H 2 O 2 production
The assay is based on detection of H 2 O 2 in the medium using the Amplex UltraRed fluorescent dye. In the presence of horseradish peroxidase (HRP), Amplex UltraRed reacts with H 2 O 2 in a 1:1 stoichiometry producing fluorescent Amplex UltroxRed. HRP (5 U/2 ml) and Amplex UltraRed (3 μM) were added to standard medium A or B. Subsequently, mitochondria (0.05 mg/ml) and succinate (5 mM) or α-GP (20 mM) were added. Resorufin fluorescence was detected using a Photon Technology International (PTI; Lawrenceville, NJ, USA) Deltascan fluorescence spectrophotometer. The excitation wavelength was 550 nm, while the emission was detected at 585 nm. At the end of each experiment, the fluorescence signal was calibrated with 100 pmol H 2 O 2 . All the measurements were performed at 37°C.

Measurement with safranine-O
Δψ m was assessed using safranine-O, a lipophilic cationic fluorescent dye, which accumulates in the mitochondrial membrane upon hyperpolarisation resulting in fluorescence quenching (Akerman and Wikstrom 1976). Safranine (2 μM) fluorescence (495 nm for excitation, 585 nm for emission) was detected using a Hitachi F-4500 spectrofluorimeter (Hitachi High Technologies, Maidenhead, UK). All measurements were carried out at 37°C in standard medium A or B, as previously described.

Measurement with TPP + electrode
Δ ψ m was estimated via the distribution of t he tetraphenylphosphonium ion (TPP + ). TPP + was detected using a custom-made TPP + -selective electrode (Kamo et al. 1979), as described previously (Tretter et al. 2007a). Δψ m was calculated using the Nernst equation and the reported binding correction factor for brain mitochondria, as previously described (Rottenberg 1984;Rolfe et al. 1994). The calculation was performed according to Rottenberg and co-workers (Rottenberg 1984) assuming that the matrix volume of the mitochondria is 1 μl/mg protein (D.G. Nicholls, personal communication). The sensitivity of the TPP + electrode was found to be decreased at low Δψ m (less than~120 mV) (Starkov and Fiskum 2003).

Measurement of the intramitochondrial pH (pH in )
pH in of isolated mitochondria was measured with the acetoxymethyl ester form of 2,7-biscarboxyethyl-5(6)-carboxyfluorescein (BCECF/AM) (Jung et al. 1989), as described earlier (Sipos et al. 2005). Briefly; 100 μl mitochondria (35-40 mg/ml protein) were incubated with 50 μM BCECF/AM in Buffer C (in mM: 225 mannitol, 75 sucrose, 5 HEPES, 0.1 EGTA, pH 7.4) for 10 min at 25°C. Ice-cold Buffer C (325 μl) was supplemented with 0.1 mM ADP (in order to prevent permeability transition pore opening). Loaded mitochondria were centrifuged for 2 min at 13000 g, the supernatant was removed, the pellet was resuspended in 450 μl Buffer C, and this was centrifuged for 2 min at 13000 g. The new pellet was resuspended in 450 μl Buffer C minus ADP, left standing for hydrolysis (10 min), and then centrifuged for 2 min at 13000 g. All centrifugation steps were performed at 4°C. The supernatant was discharged. The pellet was supplemented with 13 μl Buffer C. BCECF-loaded mitochondria were used within 90 min. For fluorescence measurements, 3 μl aliquots of mitochondria were diluted in 2 ml of standard medium A or B. Fluorescence ratios were determined using the PTI Deltascan fluorescence spectrophotometer (440 or 505 nm for excitation, 540 nm for emission). Leaching of BCECF from mitochondria was determined by measuring the fluorescence of the supernatant of the centrifuged loaded mitochondria. Corrections were made by subtracting the fluorescence values of the supernatant from those of the experimental values. For calibration, the external and internal [H + ] were equilibrated at varying pH extra values by the addition of a mixture of 8 μM nigericin (K + /H + antiporter), 2.5 μM gramicidin (Na + /K + ionophore), and 8 μM monensin (Na + /H + antiporter), as previously described (Sipos et al. 2005).

Statistical analysis
The statistical differences in multiple comparisons were evaluated with ANOVA (SigmaPlot™, Version 11, Systat Software, Inc., San Jose, CA, USA). Values of p < 0.05 were considered to be statistically significant.

Results
In order to dissect Δψ m and ΔpH, the two components of pmf, ionophores were introduced throughout the experiments. The standard media A contained 2 mM K 2 HPO 4 and 125 mM KCl, whilst standard medium B was supplemented with 2 mM K 2 HPO 4 and 4 mM K + . ADP was absent providing a high Δψ m to support RET in succinate-or α-GP-energised mitochondria. At the end of each experiment the uncoupler FCCP was given to eliminate any Δψ m and abolish the succinate-or α-GP-driven RET.
Effects of nigericin on pH in , ΔpH, Δψ m , and mtROS production in brain mitochondria at medium pH 7.0 Nigericin, a K + /H + antiporter, allows the electroneutral transport of these two ions in opposite directions across the mitochondrial inner membrane following the K + concentration gradient (Henderson et al. 1969;Rottenberg and Lee 1975). As displayed in Fig. 1, nigericin (20 nM) decreased pH in (Fig.  1a) at pH extra = 6.84 ± 0.01 (medium pH = 7.0) by 0.13 ± 0.04 pH unit and ΔpH from 0.23 ± 0.06 to 0.089 ± 0.02 (Fig. 1c). In addition, nigericin increased Δψ m by 7.78 ± 2.5 mV; Δψ m could not be increased any further by subsequent additions of nigericin. In contrast to Lambert and co-workers , we found that nigericin increased the rate of H 2 O 2 generation by 52 ± 11% (from 1894 ± 169 to 2871 ± 169 pmol/min/mg protein) in succinate-respiring brain mitochondria (Fig. 1e). We can conclude that in succinatesupported mitochondria, nigericin decreased ΔpH and induced mitochondrial hyperpolarization, simultaneously elevating H 2 O 2 production.
In order to gain a deeper insight into the effects of nigericin on RET, α-GP was also applied as a respiratory substrate. Unlike succinate, α-GP does not enter the mitochondria, it is oxidized by α-GPDH on the outer surface of the inner mitochondrial membrane and does not form NADH. Addition of rotenone diminished the H 2 O 2 production both in succinate and α-GP energised mitochondria, which points to a CIrelated ROS production, likely RET (Votyakova and Reynolds 2001). Both respiratory substrates upon their oxidation by succinate dehydrogenase (SDH) or α-GPDH reduce the coenzyme Q (Q; ubiquinone)-junction bypassing CI. Similarly to that observed with succinate, nigericin decreased pH in , increased Δψ m (data not shown), and stimulated H 2 O 2 production (Fig. 2c) in α-GP-energised mitochondria as well.
Effects of valinomycin on pH in , ΔpH, Δψ m , and mtROS production in brain mitochondria at pH 7.0 Valinomycin is a K + ionophore transporting K + along its electrochemical gradient across the mitochondrial inner membrane. In succinate-supported mitochondria, valinomycin (0.25 nM) increased pH in by 0.38 ± 0.04 pH unit (Fig. 1b) , ΔpH from 0.39 ± 0.001 to 0.75 ± 0.04, and depolarized Δψ m in a dose-dependent manner (Fig. 1d). We found that valinomycin decreased the rate of H 2 O 2 generation by 44.5 ± 4% when mitochondria were supported by succinate (Fig. 1f, trace b). Valinomycin displayed similar effects on α-GP-respiring brain mitochondria. At pH extra = 7.22 ± 0.01 (Fig. 2d), valinomycin alkalized the mitochondrial matrix by 0.26 ± 0.02 pH unit, while ΔpH was increased from 0.32 ± Fig. 1 Effect of nigericin (a, c, e) and valinomycin (b, d, f) on pH in and ΔpH (a, b), Δψ m (c, d) and the rate of H 2 O 2 production (e, f) in succinate-energised brain mitochondria. Mitochondria (0.05 or 0.1 mg/ml) were incubated in different standard media as described under Materials and Methods. Succinate (5 mM), FCCP (250 nM), valinomycin (0.25 nM), nigericin (20 nM) and cocktail (gramicidin, monensin, nigericin) were given as indicated. ΔpH (a, b) values were calculated from the difference between pH in and pH extra . In A and B each experiment was calibrated by KOH. In E and F results (slope) are expressed in pmol/min/mg protein and each experiment was calibrated by 100 pmol H 2 O 2 . For (a, b, c, d, e, f) traces are representative of at least three independent experiments 0.01 to 0.59 ± 0.02 ( Fig. 3d) with α-GP. Simultaneously, a decreased rate of the α-GP-evoked H 2 O 2 production (by 45 ± 14%) was measured, similarly to that observed in succinatesupported mitochondria.
Effects of pH extra on H 2 O 2 production, pH in , ΔpH, and Δψ m in succinate-and α-GP-respiring brain mitochondria To examine the influence of changes in pH extra on ΔpH and H 2 O 2 production, experiments were carried out in standard media A (nigericin) or B (valinomycin) varying pH from 6.4 to 8.0 (see Materials and Methods).
Δψ m Measuring Δψ m by a TPP + electrode, it was concluded that nigericin always increased the Δψ m approximately to the same level (~− 195 -200 mV), even at different pH extra values in brain mitochondria. At pH extra = 6.45 ± 0.004, Fig. 2 Effect of nigericin (a, c) and valinomycin (b, d) on the rate of succinate (a, b) and αglycerophosphate (c, d)-driven H 2 O 2 production as a function of pH extra in brain mitochondria. Mitochondria (0.05 mg/ml) were incubated in the standard media as described under Materials and Methods. Succinate (5 mM), αglycerophosphate (α-GP; 20 mM), valinomycin (0.25 nM) and nigericin (20 nM) were added. The results are expressed as the rate of H 2 O 2 production in pmol/min/mg protein mean ± SEM (n > 4) and pH extra given as mean ± SEM (n > 4) and written in the graphs; ***p < 0.001; **p < 0.01 nigericin hyperpolarized the membrane by 12.5 mV, at pH extra = 6.84 ± 0.013 by 19 mV, and at pH extra = 7.30 ± 0.047 by 8.5 mV. Taken together, these data show that Δψ m and the rate of H 2 O 2 production were the highest when nigericin was present and the medium was the most alkaline. pH in and ΔpH As shown in Fig. 3, upon elevation of pH extra , ΔpH was concomitantly decreased in both succinate-and α-GP-respiring mitochondria. The addition of nigericin was followed by acidification of the mitochondrial matrix, resulting in a drop of ΔpH (Fig. 3a, c). At the most alkaline pH extra (7.45 ± 0.02), in the presence of succinate, nigericin could neither decrease pH in nor ΔpH. However, in α-GP-respiring mitochondria, nigericin reduced both pH in and ΔpH at all measured pH extra values (Fig. 3c). Valinomycin treatment of both succinate-and α-GP-respiring brain mitochondria caused alkalinization of the mitochondrial matrix and a corresponding elevation of ΔpH (Fig. 3b, d).

Heart mitochondria. Effects of nigericin and valinomycin on mitochondrial parameters
Detecting RET is also relevant in organs other than the brain, like heart, regarding their exposure to oxidative stress under pathological conditions, like ischemia-reperfusion (Chouchani et al. 2014). To deepen our understanding on RET in heart mitochondria, effects of ΔpH and Δψ m on Fig. 3 Effect of nigericin (a, c) and valinomycin (b, d) on ΔpH in succinate (a, b) and αglycerophosphate (c, d)-respiring brain mitochondria as a function of pH extra . Mitochondria were incubated in the standard media as described under Materials and Methods. Succinate (5 mM), αglycerophosphate (α-GP; 20 mM), valinomycin (0.25 nM) and nigericin (20 nM) were used. The results are expressed as pH value mean ± SEM (n > 4) and written in the graphs; ***p < 0.001; *p < 0.05 succinate-supported H 2 O 2 production were investigated applying the above-mentioned ionophores.
Similarly to brain, in heart mitochondria nigericin hyperpolarized the membrane at various pH extra values. In the absence of nigericin, Δψ m of succinate-supported, nonphosphorylating mitochondria was similar at all pH extra values analogously to brain. In contrast to that observed in brain mitochondria, in heart, the addition of nigericin led to an increase of the rate of succinate-evoked H 2 O 2 generation only between pH extra = 6.46 ± 0.005 and 7.03 ± 0.008 (Fig. 4a). At a more alkaline pH (pH extra = 7.54 ± 0.002), nigericin decreased the rate of H 2 O 2 formation by 22 ± 8% (Fig. 4a, white  circles). In the absence of nigericin, upon elevation of pH extra , the rate of the succinate-initiated H 2 O 2 generation was steeply increasing (Fig. 4a, black circles). In the absence of nigericin, ΔpH decreased with incrementing pH extra (pH extra from 6.46 ± 0.005 to 7.03 ± 0.008) until pH extra 7.03 ± 0.008; at pH extra above such value , ΔpH increased (Fig. 4b, black circles). Contrary to this, in the presence of nigericin, ΔpH was slightly ascending upon pH extra elevation (Fig. 4b, white circles) and at pH extra = 7.54 ± 0.002 there was no statistically significant difference between ΔpH in the presence of nigericin compared to ΔpH in its absence.

Discussion
There is a lack of consensus regarding the role of ΔpH and Δψ m on mtROS generation Selivanov et al. 2008), therefore, in our study, we aimed to clarify the dependence of succinate-and α-GP-driven H 2 O 2 production on components of pmf. The results presented above allow the conclusion that Δψ m displays a stronger influence on the succinate-or α-GP-supported, RET-initiated H 2 O 2 production than ΔpH. In this study, we did not only measure H 2 O 2 production and Δψ m , but we also detected matrix pH (pH in ) with the fluorescent dye BCECF and calculated ΔpH. Under most physiological conditions depolarization of the inner membrane (decrease of the absolute value of Δψ m ) is associated with a decrease of ΔpH and an elevation of matrix [H + ]. It is unfeasible to create conditions where one of the components of pmf is maintained constant whilst the other one is independently altered. With ionophores however, these two parameters can be changed in opposite directions. In order to increase Δψ m , nigericin was applied, which decreased ΔpH and increased H 2 O 2 production ( Fig.1) suggesting that mtROS production is directly proportional to Δψ m . If ΔpH was the dominant factor of the RET-initiated H 2 O 2 formation, then H 2 O 2 production should have been decreased. To increase ΔpH, valinomycin was added, which simultaneously depolarized the inner membrane and decreased the rate of H 2 O 2 generation which changed in accordance with Δψ m values. If ΔpH had been the major player in H 2 O 2 production, then H 2 O 2 production should have been higher in the presence of valinomycin than in its absence.
Our measurements were carried out not only in brain but also in heart mitochondria, both displaying similar effects. Based on these observations, we can exclude the tissue specific modification of RET-supported H 2 O 2 generation in these tissues.
In summary, our studies with the two ionophores showed that RET-evoked H 2 O 2 production always varied in accordance with changes of Δψ m , which leads to the conclusion that Δψ m has a greater influence on mitochondrial RETinitiated H 2 O 2 formation than ΔpH. In addition, we also showed that elevation of pH extra resulted in increased H 2 O 2 generation, a finding that suggests a clear correlation between absolute pH and H 2 O 2 production.

Nigericin
Nigericin, as a K + /H + antiporter, is responsible for the electroneutral exchange of K + and H + (Henderson et al. 1969;Bernardi 1999). In our preliminary experiments, the dose-dependent effects of nigericin on Δψ m were studied, and the lowest possible concentration was used which created a maximal mitochondrial hyperpolarization measured by safranin fluorescence (data not shown). Contrary, Lambert and colleagues  as well as Selivanov's group (Selivanov et al. 2008) applied 100 nM nigericin, which in our hands did neither increase Δψ m further, nor dissipate ΔpH completely but established a new equilibrium with lower pH in . ΔpH, after administration of 100 nM nigericin, could be decreased further by addition of 250 nM FCCP and mixture of ionophores (see Materials and Methods). To eliminate confounding factors that could have influenced ROS production (e.g. succinate transport or further metabolism of succinate in the tricarboxylic acid cycle), not only succinate, but also α-GP was used to energize mitochondria and support RET-mediated ROS production. Results with α-GP were qualitatively equivalent to those obtained in succinate-supported mitochondria (Figs. 2 and 3). The stimulating effect of nigericin on H 2 O 2 generation was more pronounced at acidic pH extra . Interestingly, in heart mitochondria, at alkaline pH extra , nigericin decreased the rate of H 2 O 2 release (Fig. 4a). It appears that in heart mitochondria, the diminution in the rate of H 2 O 2 production at alkaline pH cannot be explained by depolarisation of the mitochondrial membrane.

Valinomycin
In the presence of valinomycin, the mitochondrial membrane is permeable to K + ; its effect is highly dependent on the K + concentration of the medium and the applied valinomycin concentration. High K + concentrations in the presence of 2 mM KH 2 PO 4 and valinomycin lead to high amplitude mitochondrial swelling (Ligeti and Fonyo 1977;Bernardi 1999), therefore, 4 mM KCl was used in valinomycin experiments. It is well known that in isolated mitochondria the highest ΔpH can be achieved at low K + concentration (Mitchell and Moyle 1968;Nicholls 1974;Nicholls 2005). It is noteworthy that ΔpH in low K + medium is about 0.6-0.8 pH unit, but at high K + medium it is only 0.3 pH unit. In our experiments valinomycin caused matrix alkalization and concomitant ΔpH elevation. This observation can be explained by the fact that the valinomycin-induced entry of K + into the mitochondrial matrix usually triggers H + extrusion and P i / OH − exchange (Garlid and Paucek 2003). The H + extrusion generally mediates a compensatory decrease in Δψ m and an elevation of respiration both in the succinate-or α-GPsupported mitochondria. Valinomycin-caused depolarisation led to inhibition of RET-supported H 2 O 2 production.

Effects of pH extra on H 2 O 2 production
In agreement with the observations of Selivanov (Selivanov et al. 2008), in non-phosphorylating mitochondria, the acidification of the mitochondrial matrix is followed by an elevation in ΔpH and a decrease in the succinate-and α-GP-driven H 2 O 2 production. There is an inverse proportionality between ΔpH and H 2 O 2 formation, which weakens the notion of Lambert and Brand that ΔpH would exhibit a stronger effect on RET than Δψ m . Our measurements of pH in with BCECF have shown that ΔpH is greater at lower pH and varies with pH extra . Banh and Treberg observed an analogous pattern in glutamate and malate-respiring, non-phosphorylating, rat skeletal muscle mitochondria, where the H 2 O 2 generation was enhanced upon alkalization (Banh and Treberg 2013 (Ohnishi et al. 2005;Grivennikova and Vinogradov 2006). Two mechanistic models exist for the explanation of mtROS production by the CI: (1) the one-site model states that the O 2 .production site, during both FET and RET, is ultimately the reduced flavin (Galkin and Brandt 2005;Pryde and Hirst 2011), whereas (2) the two-site model suggests that during FET, the flavin of CI is responsible for O 2 .formation, while, under RET, the SQ .species, synthetized at the ubiquinone-binding Q-site (Q-binding site) of CI, are liable for the elevated O 2 .release (Brand 2010;Treberg et al. 2011). Both theories agree that the greatest drop in redox potential in the CI occurs between the N2 subunit and the ubiquinone (Q), whose interaction initiates conformational changes that are coupled to the proton translocation (Treberg et al. 2011).
Δψ m : There are speculations that the above mentioned conformational changes of the CI might also depend on Δψ m (Brandt 2006;Dlaskova et al. 2008). When Δψ m is adequately high, it decelerates the proton pumping activity of the CI, which may favour SQ .formation and hence O 2 .generation. ΔpH and pH extra : Our results do not support the hypothesis that ΔpH would influence the RET-initiated ROS production to a higher degree than Δψ m . The theory that tries to explain the influence of absolute pH on the H 2 O 2 formation assigns a potential role to SQ .formation at the Q-site of the CI (Ohnishi et al. 2005;Treberg et al. 2011). At the Q-site, Q is reduced by a single electron to SQ .-. SQ .can react further in two possible ways (Selivanov et al. 2008): (1) with a single electron plus two H + to form ubiquinol (QH 2 ) (SQ .-+ e − + 2 H + ↔ QH 2 ), or (2) with O 2 to form the highly reactive O 2 .-(SQ − + O 2 ↔ Q + O 2

.-
). At acidic pH, the first reaction is shifted towards QH 2 formation according to the Le Chatelier's principle (Selivanov et al. 2008).

Potential significance of our results: Mild uncoupling
In succinate-respiring mammalian mitochondria, mild uncoupling lowers Δψ m and consequently also the rate of ROS generation (Skulachev 1996;Korshunov et al. 1997;Miwa and Brand 2003). Mild uncoupling is a special condition where oxidative phosphorylation occurs at a relatively higher conductance of the inner mitochondrial membrane, this results in lowered pmf and a minor stimulation of respiration (Skulachev 1996;Brand et al. 2004). Our results support the notion that a minor decrease in Δψ m leads to a diminution of the succinate-evoked, RET-initiated H 2 O 2 release. Uncoupling proteins (UCP; like UCP1-3) and the adenine nucleotide transporter are also involved in mild uncoupling processes (Andreyev et al. 1988;Jezek 2002). Interestingly, O 2 .can activate UCPs in the matrix with the contribution of fatty acids resulting in mild uncoupling (Echtay et al. 2002) and consequently a slower ROS production. Although it is likely that in vivo, under physiological conditions, ATP synthesis caused depolarisation of Δψ m is sufficient to decrease ROS generation (Votyakova and Reynolds 2001;Starkov and Fiskum 2003), effects on mtROS of mild uncoupling and of Δψ m are possibly relevant to patological states.
In fact, it has been hypothesized that initiation of mild uncoupling might be beneficial in oxidative stress-related diseases characterized by high Δψ m such as in ischemiareperfusion injury (Kadenbach et al. 2011). This hypothesis has been corroborated by a report showing that under ischemia, succinate can accumulate in mouse heart owing to the reversal of SDH (Chouchani et al. 2014). In reperfusion, SDH returns to oxidize the accumulated succinate and this has been claimed to result in an enhanced RET-mediated mtROS formation (Chouchani et al. 2014).
In summary, data from our laboratory provided evidence that the succinate-or α-GP-evoked, RET-initiated H 2 O 2 production is more dependent on Δψ m than on ΔpH. Our findings have helped elucidating mechanisms underpinning mtROS production and support consideration of the therapeutic applications of mild uncoupling, which can be initiated by e.g. mitochondria-targeted antioxidants.