Interaction of the antibiotic minocycline with liver mitochondria – role of membrane permeabilization in the impairment of respiration

Authors

  • Peter Schönfeld,

    Corresponding author
    1. Institute of Biochemistry and Cell Biology, Otto-von-Guericke-University, Magdeburg, Germany
    • Correspondence

      P. Schönfeld, Institute of Biochemistry and Cell Biology, Otto-von-Guericke-University, Leipziger Str. 44, 39120 Magdeburg, Germany

      Fax: +49 0391 67 15898

      Tel: +49 0391 67 15362

      E-mail: peter.schoenfeld@med.ovgu.de

      L. Wojtczak, Nencki Institute of Experimental Biology, Pasteura 3, 02-093 Warsaw, Poland

      Fax: +48 22822 5342

      Tel: +48 225892 315

      E-mail: l.wojtczak@nencki.gov.pl

    Search for more papers by this author
  • Detlef Siemen,

    1. Department of Neurology, Otto-von-Guericke-University, Magdeburg, Germany
    Search for more papers by this author
  • Peter Kreutzmann,

    1. Institute of Biochemistry and Cell Biology, Otto-von-Guericke-University, Magdeburg, Germany
    Search for more papers by this author
  • Claudia Franz,

    1. Institute of Biochemistry and Cell Biology, Otto-von-Guericke-University, Magdeburg, Germany
    Search for more papers by this author
  • Lech Wojtczak

    Corresponding author
    1. Nencki Institute of Experimental Biology, Warsaw, Poland
    • Correspondence

      P. Schönfeld, Institute of Biochemistry and Cell Biology, Otto-von-Guericke-University, Leipziger Str. 44, 39120 Magdeburg, Germany

      Fax: +49 0391 67 15898

      Tel: +49 0391 67 15362

      E-mail: peter.schoenfeld@med.ovgu.de

      L. Wojtczak, Nencki Institute of Experimental Biology, Pasteura 3, 02-093 Warsaw, Poland

      Fax: +48 22822 5342

      Tel: +48 225892 315

      E-mail: l.wojtczak@nencki.gov.pl

    Search for more papers by this author

Abstract

Several studies have proposed that the antibiotic minocycline (MC) has cytoprotective activities. Nevertheless, when cells have been exposed to MC at micromolar concentrations, detrimental effects have been also observed. Despite the known inhibitory activity of MC on ATP synthesis and the Ca2+ retention capacity of isolated rat liver and brain mitochondria, the underlying mechanism is still debated. Here, we present further arguments supporting our concept that MC acting on rat liver mitochondria suspended in KCl medium permeabilizes the inner membrane. Supplementation of the medium with cytochrome c and NAD+ strongly enhanced the respiration of MC-treated mitochondria, thus partly preventing or reversing the inhibitory effect of MC on state 3 or uncoupled respiration. These results indicate that MC produced depletion of mitochondrial cytochrome c and NAD+, thus impairing mitochondrial respiration. In addition, NADH oxidation by alamethicin-permeabilized mitochondria supplemented with cytochrome c was insensitive to 200 μm MC, arguing against direct impairment of respiratory chain complexes by MC. Finally, a surprising feature of MC was its accumulation or binding by intact rat liver mitochondria, but not by mitochondria permeabilized with alamethicin or disrupted by freezing and thawing.

Abbreviations
Ala

alamethicin

FCCP

carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone

MC

minocycline

SD

standard deviation

TEA

tetraethylammonium

Δψ

mitochondrial transmembrane potential

Introduction

Minocycline (MC) is a long-known semisynthetic second-generation tetracycline antibiotic, consisting of a four-ring carbocyclic skeleton (Fig. 1). The phenolic β-ketone–enol system connected to rings D, C and B enables MC to chelate divalent cations, including Mg2+ and Ca2+, a property that is shared with other members of the tetracycline family [1]. Tetracyclines exist in solution as a mixture of tautomers, and therefore can adopt different chemical structures, depending on the environment [2]. MC is more lipophilic than other tetracyclines [3], and, like other tetracyclines, has a zwitterionic character and can adopt, depending on the medium pH, a net positive or negative charge.

Figure 1.

Chemical structure of MC.

Several studies have reported nonantibiotic, cytoprotective activities of MC (for reviews, see [4, 5]). Nevertheless, when cells were exposed to MC concentrations in the range of 50–100 μm, toxic effects were observed [6]. The fluorescence properties of tetracyclines revealed five decades ago that they specifically bind to mitochondria of living cells [7]. This observation fits well with accumulating evidence that mitochondria are the primary targets for MC. Thus, mitochondria isolated from rat liver or brain respond to micromolar concentrations of MC with impaired Ca2+ retention capacity [8-15]. Efforts to understand the detrimental activities of MC have been confused by the use of nonenergized or energized mitochondria and the application of nonionic (sucrose/mannitol) or ionic (KCl) incubation media. It has been proposed that the effects of MC on mitochondrial respiration may be attributable to activation of ion-conducting pathways [11] or the formation of Ca2+-dependent ion channels in the inner mitochondrial membrane [14]. Alternatively, the decrease in phosphorylating respiration caused by MC has been attributed to direct interactions of MC with the electron transport chain [13] and/or with the adenine nucleotide translocase [15].

The present investigation was an attempt to look more closely at the mechanisms of MC interaction with isolated mitochondria, to elucidate the ways in which this antibiotic impairs mitochondrial functions. Here, we show that the inhibitory effect of MC on mitochondrial respiration with NAD-linked substrates can be, to a great part, prevented or reversed by supplementation with cytochrome c and NAD+. The inhibition of succinate oxidation could be partly restored by supplementation with cytochrome c alone. We have also confirmed previous observations that MC induces rapid mitochondrial swelling in media containing KCl [11], presumably because of increased permeability to K+ and Cl. On that basis, we conclude that impairment by MC of mitochondrial respiration is mainly caused by depletion of essential water-soluble cofactors of the respiratory chain. Independently of this mechanism, we have also demonstrated accumulation of MC within mitochondria that is partly energy-dependent.

Results

Mitochondrial respiration

The present results fully confirmed previous observations by ourselves [11] and other authors [9, 13, 15] that MC at micromolar concentrations decreased mitochondrial respiration. However, here we found that this inhibition of oxygen uptake could be reversed to a great extent by subsequent addition of NAD+ and cyttochrome c (Fig. 2A). Moreover, if the incubation medium was supplemented with NAD+ plus cytochrome c before addition of the antibiotic, 50 μm MC exerted a much smaller effect (Fig. 2B). It was subsequently found that the optimum protection against MC inhibition of mitochondrial respiration with glutamate and malate as substrates was obtained with 0.5 mm NAD+ and 5 μm cytochrome c. When the respiratory substrate was succinate (plus rotenone), supplementation with cytochrome c alone was sufficient (not shown). Similar protection was also observed with 100 μm MC (not shown).

Figure 2.

Effect of MC on the respiration of phosphorylating rat liver mitochondria. Mitochondria were suspended in 2 mL of KCl medium (0.5 mg protein·mL−1) containing 5 mm glutamate and 5 mm malate. Additions were: 50 μm MC (corresponding to 100 nmol·mg−1 protein), 0.5 mm NAD+, and 5 μm cytochrome c. State 3 respiration was adjusted by addition of 0.67 mm ADP. (A) State 3 respiration was first induced by ADP, and this was followed by addition of MC and, finally, by addition of NAD+ plus cytochrome c. (B) The medium was supplemented with NAD+ plus cytochrome c. ADP was added where indicated, followed by MC. Solid lines: oxygen concentration traces. Dashed lines: rates of oxygen uptake. Rates of respiration (nmol O2·min−1 per mg protein) are indicated by numbers next to the oxygen traces. Cyt c, cytochrome c; RLM, rat liver mitochondria.

The protective effect of NAD+ and cytochrome c against MC inhibition for both state 3 and the uncoupled state is summarized in Fig. 3. As shown, 50 μm MC decreased mitochondrial respiration by 43% and 42% in state 3 (Fig. 3A) and under uncoupled conditions (Fig. 3B), respectively, whereas this inhibition amounted to as little as 17–23% (in both conditions) after supplementation with NAD+ plus cytochrome c. It has to be noted that this supplementation had negligible effects on state 3 and uncoupled respiration in the absence of MC. These results prompted us to hypothesize that the inhibitory effect of MC on mitochondrial respiration may be mainly attributable to depletion of essential water-soluble respiratory cofactors. This might result from the mitochondrial swelling produced by this antibiotic [11].

Figure 3.

Effect of supplementation with cytochrome c and NAD+ on the inhibition of mitochondrial respiration by MC. Mitochondria were suspended in KCl medium as in Fig. 2. The medium for the column at the right side of each panel also contained 0.5 mm NAD+ and 5 μm cytochrome c. MC (50 μm), ADP (0.67 mm) and FCCP (0.5 μm) were added where indicated. The results are means ± SDs for four to seven mitochondrial preparations. *Means are significantly different (P ≤ 0.05): *ADP versus ADP + MC, and **ADP + MC versus ADP + MC (NAD+, cytochrome c); *FCCP versus FCCP + MC, and **FCCP + MC versus FCCP + MC (NAD+, cytochrome c). Cyt c, cytochrome c.

To verify this assumption, we compared the effect of MC with that of a known swelling-inducing agent, alamethicin (Ala). This compound is known to form large-conductance water-filled pores in biological membranes [16], including the inner mitochondrial membrane [17], but exerts no documented inhibitory effect on complexes of the respiratory chain. Ala-treated mitochondria are fully uncoupled, as they are unable to maintain the transmembrane electrical potential and ion gradients. This was confirmed by the pilot experiment illustrated in Fig. 4. As shown, the uncoupled respiration with glutamate and malate as substrates was rapidly and almost completely inhibited by Ala, and could be restored to ~ 70% by subsequent addition of NAD+ and cytochrome c. Figure 4B,C shows that both NAD+ and cytochrome c are needed to maximally restore the oxygen uptake. Figure 4B,C also shows that Ala-treated mitochondria are fully uncoupled, as the presence of the chemical uncoupler carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone (FCCP) is not required for the maximum rate of respiration to be achieved. This is obviously attributable to complete permeability of Ala-treated mitochondria to ions. An analogy between Ala and MC in their action on mitochondria is further illustrated by Fig. 5. Neither compound, when added to mitochondria respiring with glutamate and malate in the resting state, had an apparent effect on the rate of oxygen uptake (Fig. 5A,C), unless the medium was supplemented with NAD+ and cytochrome c (Fig. 5B,D), because depletion of these cofactors prevented the maximum rate of uncoupled respiration being achieved.

Figure 4.

Effect of Ala on mitochondrial respiration. Experimental conditions were as in Figs 2 and 3. Ala (16 μg·mg−1 protein) was added where indicated. Rates of respiration (nmol O2·min−1 per mg protein) are indicated by numbers next to the oxygen traces. Cyt c, cytochrome c; RLM, rat liver mitochondria.

Figure 5.

Comparison of the effects of MC and Ala on basal mitochondrial respiration. Conditions were as in Fig. 2. MC (50 μm) and Ala (16 μg·mg−1 protein) were added where indicated. The incubation media in (B) and (D) contained 0.5 mm NAD+ and 5 μm cytochrome c. Rates of respiration (nmol O2·min−1 per mg protein) are indicated by numbers next to the oxygen traces. Cyt c, cytochrome c; RLM, rat liver mitochondria.

In order to further check a possible inhibitory effect of MC on the respiratory chain, mitochondria permeabilized with Ala and supplemented with 5 μm cytochrome c were titrated with increasing concentrations of MC. It was found that the oxidation of (externally added) NADH or succinate by such a preparation was insensitive to concentrations of MC up to 200 μm. A significant decrease in the oxidation rate could only be observed at MC concentrations as high as 0.5 mm or 1 mm (Fig. 6). It has to be noted that the oxidation of NADH was rotenone-sensitive, so it occurred through the mitochondrial respiratory chain.

Figure 6.

Effect of MC on the respiration of Ala-permeabilized rat liver mitochondria. Mitochondria permeabilized with Ala (16 μg·mg−1 protein) were suspended in KCl medium supplemented with 5 μm cytochrome c [in (A), the medium also contained 5 mm glutamate and 5 mm malate], and titrated with increasing concentrations of MC. (A) Oxidation of external NADH. Mitochondria (0.17 mg protein·mL−1), 0.3 mm NADH and 2 μm rotenone were added where indicated. The results are mean values ± SDs for three mitochondrial preparations. *Significantly different from the control incubation (ANOVA, Tukey's post hoc test, P ≤ 0.001). (B) Oxidation of succinate. Mitochondria (0.33 mg protein·mL−1), 5 mm succinate and 2 μm rotenone were added where indicated. The results are mean values ± SDs for four mitochondrial preparations. *Significantly different from the control incubation (ANOVA, Tukey's post hoc test, P ≤ 0.001). Rot, rotenone; Succ, succinate.

MC-induced mitochondrial swelling and electron microscopy

MC-induced swelling of mitochondria in KCl medium, but not in sucrose medium, has already been documented [11]. Here, we found that micromolar concentrations of MC also induced swelling in isotonic KNO3 medium but not in media containing nonpermeant cations such as tetraethylammonium (TEA) or choline (Fig. 7A). In addition, we found that inhibition of the respiratory chain by rotenone (Fig. 7A, trace b) or other blockers, such as antimycin A and KCN (not shown), as well as de-energization by FCCP (not shown), prevented swelling in KCl medium. Remarkably, MC-induced swelling was followed by a slow partial recovery (Fig. 7A, traces a and c). Parallel monitoring of the mitochondrial transmembrane potential (Δψ) revealed that swelling in KCl or KNO3 medium was accompanied by a transient depolarization, whereas such a decrease in Δψ was not observed in choline chloride or TEA chloride medium (Fig. 7B). It has to be noted that MC did not produce mitochondrial depolarization in sucrose medium [11]. As expected, addition of the chemical protonophore FCCP resulted in complete depolarization in all media. These results, taken together, suggest that MC specifically permeabilizes the inner mitochondrial membrane to K+ and, possibly, to Cl.

Figure 7.

MC-induced swelling of rat liver mitochondria and the effect of MC on Δψ. Mitochondria (1 mg protein·mL−1) were suspended in either KCl medium or in media where KCl was replaced by 125 mm KNO3, 125 mm choline chloride (CholCl), or 125 mm TEA chloride (TEACl). All media contained 5 mm glutamate plus 5 mm malate. Rotenone (2 μm) was present where indicated. The tonicity of all media was approximately the same. In (A), swelling was initiated by the addition of 100 μm MC (corresponding to 100 nmol·mg−1 protein). In (B), the media also contained 10 μm safranin O. A decrease in safranin fluorescence corresponds to an increase in Δψ. Both panels show typical traces. ΔE; RLM rat liver mitochondria; Rot, rotenone.

A more detailed insight into the volume and structural changes of mitochondria induced by MC was enabled by electron microscopy. Rat liver mitochondria freshly suspended in KCl incubation medium supplemented with glutamate and malate adopted the so-called orthodox or slightly condensed conformation (Fig. 8A), whereas mitochondria treated with 100 μm MC were either swollen or strongly contracted (Fig. 8B). For comparison, mitochondria treated with Ala were also examined. They consisted of a uniform population of strongly swollen particles (Fig. 8C).

Figure 8.

Electron micrographs of rat liver mitochondria. Rat liver mitochondria (1 mg protein·mL−1) were suspended in KCl incubation medium containing 5 mm glutamate and 5 mm malate. They were incubated without any other additions (A), with 100 μm MC (B) or with Ala (10 μg·mg−1 protein) (C) for 5 min at 25 °C. Thereafter, the mitochondria were fixed, stained, embedded, and sliced, as described in Experimental procedures. c, condensed conformation; o, orthodox conformation; s, swollen conformation. Scale bars: 0.5 μm.

Energy-dependent uptake of MC by mitochondria

MC forms a coloured complex with Mg2+ in aqueous media. This property enabled us to follow changes in MC concentration after it had been added to mitochondrial suspensions. To elucidate this point, mitochondria in the KCl medium supplemented with succinate (plus rotenone) were exposed to MC for 2 min. Thereafter, the incubation mixture was centrifuged, MgCl2 was added to the supernatants, and the remaining MC concentration in the supernatants was measured photometrically as the Mg2+–MC complex (Fig. 9A). It was found that, in the presence of mitochondria, the MC concentration decreased by ~ 40%. The uptake of MC was smaller in mitochondria de-energized by FCCP, and was completely abolished by Ala or after a freezing–thawing treatment (Fig. 9B). Ala added 2 min after MC resulted in its complete release (not shown).

Figure 9.

Energy-dependent uptake of MC by rat liver mitochondria. KCl medium contained 100 μm MC. (A) Absorption spectra of the MC complex with Mg2+ (5 mm MgCl2) in KCl medium. Trace a: in the absence of mitochondria. Trace b: in the supernatant after preincubation with energized mitochondria (2 mg protein·mL−1). Trace c: in the supernatant after preincubation with mitochondria in the presence of 1 μm FCCP. For other conditions, see 'Experimental procedures'. (B) MC concentrations calculated for these experiments. The third and fourth column show data for mitochondria treated with Ala (10 μg·mg−1 protein) or exposed to three freezing–thawing cycles. Results are mean values ± SDs for four mitochondrial preparations. *Significantly different from the control incubation (ANOVA, Tukey's post hoc test, P ≤ 0.001). RLM, rat liver mitochondria.

Discussion

A wealth of reports on beneficial activities of MC on cell viability [4, 5] are contradicting other reports demonstrating that exposure of isolated mitochondria to moderate micromolar MC concentrations (25–200 μm) has detrimental effects on mitochondrial functions [8-15]. Mitochondria are likely to be exposed to such concentrations under in vivo conditions, when MC is injected in animal models at doses of 22–100 mg·kg−1 body weight several times daily [18]. In animal experiments, where high concentrations of MC or of its chlortetracycline analogue were applied [19, 20], and in studies on cell cultures, deleterious effects of MC have been found [21]. The results of the present study may help to explain the mechanism(s) of these deleterious activities. However, it has to be kept in mind that, with therapeutic doses in humans treated for infectious and inflammatory diseases in the range of 3 mg·kg−1 body weight/day [18], the MC concentrations in tissues and body fluids are much lower.

The present investigation sheds new light on the mechanism by which MC impairs oxygen uptake by isolated mitochondria. We have shown that the inhibition of mitochondrial respiration by NAD-linked substrates in KCl medium could, to a great extent, be reversed or prevented by supplementation of the medium with NAD+ and cytochrome c (Figs 2 and 3). This clearly indicates that the inhibitory effect of low MC concentrations is attributable, in the first instance, to depletion of these two water-soluble mediators of electron transport. We hypothesize that pyridine nucleotides, both oxidized and reduced, become released from the matrix and their binding sites within the inner membrane during MC-induced swelling. Elucidation of the underlying mechanism needs further investigations. In contrast, the release of cytochrome c is most likely attributable to large-amplitude swelling and the resulting disruption of the outer mitochondrial membrane (Fig. 8B). Impermeability of the intact outer mitochondrial membrane to cytochrome c is well established [22].

It has to be noted that both NAD+ and cytochrome c were needed to obtain the maximum effect. In this respect, the effect of MC resembled that of the antibiotic Ala, which is known to form water-filled, large-conductance pores in biological membranes, resulting in large-amplitude mitochondrial swelling [16, 17]. These pores are known to enable the passage of low molecular mass hydrophilic substances, such as pyridine nucleotides and ATP, across the inner mitochondrial membrane. In fact, an analogy between the effects of MC and Ala on mitochondrial respiration was demonstrated (Figs 4 and 5). However, it has to be remembered that the pore-forming ability of Ala is independent of the medium composition and energy supply [16, 17], whereas that of MC is dependent on the presence of K+ as the cation and Cl or NO3 as the anion. On the other hand, we also found [11] that neither swelling, nor the release of accumulated Ca2+ nor dissipation of Δψ was sensitive to cyclosporin A. These observations further substantiate the notion that MC-associated permeabilization of the inner mitochondrial membrane is different from the cyclosporin A-sensitive permeability transition. Taking these findings together, we can conclude that large-amplitude swelling may be responsible for depletion of mitochondria of their NAD+ and cytochrome c, thus impairing their respiratory function.

Here, we also found that MC induced mitochondrial swelling in KNO3 but not in choline chloride or TEA chloride (Fig. 7A). In addition, we have demonstrated previously [11] that N,N′-dicyclohexylcarbodiimide (which is known to inhibit the K+-uniporter [23, 24]) reduces MC-induced mitochondrial swelling. This indicates that MC specifically permeabilizes the inner mitochondrial membrane to K+, but not to choline or TEA cations. Moreover, as Cl, in contrast to NO3, is considered to be unable to cross the inner mitochondrial membrane, it can be concluded that MC also permeabilizes the mitochondrial membrane to Cl. This view is supported by the observation [11] that MC-induced mitochondrial swelling was inhibited by tributyltin, a known blocker of the mitochondrial inner membrane anion channel [25].

To explain this mechanism, it has to be recalled that tetracyclines show strong chelating properties towards divalent cations, among them Mg2+ [1, 26], and there is reason to expect that they can thus modulate the endogenous content of mitochondrial divalent alkali metal cations. Indeed, we have reported for the first time that MC depletes rat liver mitochondria of endogenous Mg2+ [11]. We propose that the observed activation of the K+-uniporter and the anion channel is most likely attributable to chelation by MC of mitochondrial Mg2+, as membrane-bound Mg2+ has long been known to be the main factor responsible for the low permeability of the inner mitochondrial membrane to monovalent cations [27-30]. In analogy, we have reported previously that nonesterified long-chain fatty acids induce passive swelling of liver mitochondria suspended in KCl medium, and that this swelling is associated with depletion of endogenous Mg2+ [31-33]. In those studies, fatty acids failed to trigger mitochondrial swelling in sucrose or choline chloride media, as did MC in the present investigation.

It can be supposed that the primary event following Mg2+ complexation by MC is opening of the potassium channel, resulting in a rapid influx of K+ driven by Δψ. This is subsequently followed by influx of Cl or NO3, resulting in an increase in mitochondrial volume. This order of events is supported by a rapid decrease in Δψ in KCl or KNO3 medium but not in TEA chloride or choline chloride medium (Fig. 7B). It is noteworthy that rapid swelling is followed by slow contraction (Fig. 7A) and much more rapid restoration of Δψ (Fig. 7B). We have no clear explanation for the latter observation. It can only be speculated that added MC rapidly removes membrane-bound Mg2+ but is unable to completely complex Mg2+ located inside the matrix compartment, which then becomes slowly recruited and reseals the inner membrane. Mammalian mitochondria contain ~ 30–40 nmol Mg2+·mg−1 protein and 10 nmol Ca2+·mg−1 protein [34], which is half of the amount of MC added in swelling experiments (100 nmol·mg−1 protein; see legend to Fig. 7). However, we do not know the stoichiometry of MC–Mg2+ complex formation, or its dependence on pH, lipophilicity of the local microenvironment, etc. It is therefore likely that the natural low permeability to K+ and Cl can slowly be restored, followed by operation of the K+/H+-antiporter, ejecting K+ against H+. This is confirmed by electron microscopy, which reveals a mixture of swollen and condensed (‘contracted’) organelles after MC treatment (Fig. 8).

We thus hypothesize (see also [11]) that MC permeabilizes the inner mitochondrial membrane by activating transporters or channels for K+ and Cl, and that this occurs through chelation or depletion of mitochondrial Mg2+, thereby resulting in mitochondrial swelling. Alternatively, other authors [14, 15] have proposed that MC induces membrane permeability through a hypothetical channel-forming property of MC itself. Our results, however, are better interpreted in terms of specific activation by MC of the mitochondrial potassium channel and the inner membrane anion channel. First, Antonenko et al. [14] and Cuenca-Lopez et al. [15] drew their conclusions on channel formation by MC from experiments with black lipid membranes and liposomes, where the MC/membrane lipid ratio was several hundred times higher than in our experiments with mitochondria, the medium contained a high (5 mm) concentration of Ca2+, and the pH was 8.5 [14]. Second, we found specific permeabilization of the inner membrane to K+ and Cl, contrasting with the rather unspecific permeabilization of hypothetical MC channels to small ions [14]. In addition, the incubation medium used by Cuenca-Lopez et al. [15] usually contained 3 mm MgCl2, which was 15–60 times the concentration of MC (50–200 μm). Thus, they mostly studied the effects of the MC–Mg2+ complex rather than of free MC.

The discrepancy between our results and those of Cuenca-Lopez et al. [15] concerning depletion of mitochondrial cytochrome c can be explained by the fact that those authors measured the oxidation of external reduced cytochrome c in mitochondria de-energized by antimycin A (Fig. 4 of [15]), i.e. under conditions when no MC-induced swelling could occur (Fig. 7) and therefore no disruption of the outer membrane could be expected.

The present investigation also shows, to our knowledge for the first time, that MC can accumulate inside mitochondria in an energy-dependent way (Fig. 9). Uptake of MC by mitochondria may be explained by its moderate lipophilicity [3] and a positive net charge of the MC molecule in the incubation medium. Using a very rough calculation from the data presented in Fig. 9B, assuming a matrix volume of 1 μL·mg−1 total mitochondrial protein, we estimated 170-fold enrichment of MC inside energized mitochondria. On the other hand, some of the MC taken up by mitochondria may be tightly bound to the inner mitochondrial membrane by hydrophobic and/or electrostatic forces, which explains the observation that some accumulation could also occur in FCCP-uncoupled mitochondria (Fig. 9). The concept of energy-driven accumulation of MC by mitochondria fits well with a report [35] demonstrating that Escherichia coli cells take up MC from the incubation medium, resulting in a 100-fold enrichment. In addition, this uptake was completely suppressed by the protonophoric uncoupler 2,4-dinitrophenol. Binding of MC by mitochondria at 0 °C has already been reported by Antonenko et al. [14]. This binding, together with the lipophilic interaction between MC and lipid constituents of the inner membrane, may be responsible for the impairment of mitochondrial respiration by concentrations of MC of the order of 0.5 mm and higher.

In summary, the present investigation shows that impairment of mitochondrial respiration by micromolar concentrations of MC is primarily attributable to depletion of cytochrome c and nicotinamide nucleotides, and can be largely prevented by supplementation with these water-soluble respiratory mediators. There is no indication of inhibition of respiratory chain complexes by MC at concentrations up to 200 μm, as shown by the lack of inhibition of NADH oxidation by Ala-permeabilized mitochondria.

Experimental procedures

Biological material

Procedures for animal use were in accordance with the guidelines of the Animal Health and Care Committee of the State Sachsen-Anhalt, Germany. Liver mitochondria were prepared from adult Wistar rats (average weight of 220 g) by differential centrifugation, essentially according to our standard protocol [36]. Tissue homogenization was performed in 250 mm sucrose containing 10 mm Tris/HCl (pH 7.4), 1 mm EDTA, and 1% fatty acid-free BSA. The final pellet was resuspended in BSA-free homogenization medium. The protein content in the mitochondrial stock suspensions was determined by use of the biuret method, with BSA as standard.

Incubations

All incubations, unless indicated otherwise, were performed in a standard incubation medium composed of 125 mm KCl, 10 mm Tris/HCl, 10 μm EGTA, and 5 mm Pi (pH 7.2), thermostated at 37 °C. The substrates used were 5 mm glutamate plus 5 mm malate or 5 mm succinate (plus 2 μm rotenone).

Oxygen consumption

Oxygen uptake by the mitochondria was measured with an Oroboros oxygraph (Bioenergetics and Biomedical Instruments, Innsbruck, Austria). Changes in both the oxygen concentration in the incubation medium (μm) and the rates of oxygen uptake (nmol·min−1 per mg of mitochondrial protein) were recorded.

Swelling and membrane potential

Mitochondrial volume changes were followed photometrically at 540 nm in cuvettes of 1 cm light path; Δψ was measured fluorimetrically with safranin O [37], with excitation and emission wavelengths of 495 nm and 586 nm, respectively.

Electron microscopy

For electron microscopy, mitochondria (1 mg protein·mL−1) were suspended in KCl incubation medium containing 5 mm glutamate and 5 mm malate at 25 °C, and treated as specified in Results. Thereafter, the incubation mixture was mixed with the same volume of phosphate buffer (0.1 m) supplemented with 2% glutaraldehyde plus 2% formaldehyde. The mixture was shaken for 1 h at 4 °C. Thereafter, mitochondria were sedimented by centrifugation at 10 000 g for 10 min at 4 °C. After being washed three times in 0.1 m cacodylate buffer, mitochondria were fixed for 60 min in 1% OsO4 and 0.4% potassium ferrocyanide in 0.1 m cacodylate buffer, dehydrated in a graded ethanol series, including 1 h of block staining with 2% uranyl acetate in 70% ethanol, incubated overnight in Durcupan at room temperature, and polymerized at 65 °C for 48 h. Ultrathin sections (70 nm) were obtained with an Ultracut UC6 (Leica), and examined on a Zeiss EM 900. Pictures were taken with a 2k-CDD-camera (TRS).

Chemicals

All chemicals, including MC hydrochloride, were purchased from Sigma-Aldrich at the highest quality commercially available. Stock solutions of MC hydrochloride in water were prepared daily. No significant pH changes were observed when MC concentrations in the incubation medium did not exceed 1 mm.

Statistical analysis

Data are represented as means ± standard deviations (SDs) for four to six different incubations with separate mitochondrial preparations. Statistical significance was determined by one-way ANOVA combined with the Holm–Sidak method (Fig. 3), or by one-way ANOVA combined with the Tukey post hoc test (Figs 6 and 9). Statistical calculations were carried out with sigmaplot software (Systat Software, Erkrath, Germany).

Acknowledgements

This work was supported financially by the Kultusministerium of Sachsen-Anhalt. We thank H. Goldammer for her excellent technical assistance. We are also grateful to K. Richter for preparing electron micrographs of mitochondria and helpful discussions.

Ancillary