Dopamine-derived Dopaminochrome Promotes H2O2 Release at Mitochondrial Complex I

Inhibitors of Complex I of the mitochondrial respiratory chain, such as rotenone, promote Parkinson disease-like symptoms and signs of oxidative stress. Dopamine (DA) oxidation products may be implicated in such a process. We show here that the o-quinone dopaminochrome (DACHR), a relatively stable DA oxidation product, promotes concentration (0.1–0.2 μm)- and respiration-dependent generation of H2O2 at Complex I in brain mitochondria, with further stimulation by low concentrations of rotenone (5–30 nm). The rotenone effect required that contaminating Ca2+ (8–10 μm) was not removed. DACHR apparently extracts an electron from the constitutively autoxidizable site in Complex I, producing a semiquinone, which then transfers an electron to O2, generating \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \(\mathrm{O}_{2}^{{\bar{{\cdot}}}}\) \end{document} and then H2O2. Mitochondrial removal of H2O2 monoamine, formed by either oxidase activity or DACHR, was performed largely by glutathione peroxidase and glutathione reductase, which were negatively regulated by low intramitochondrial Ca2+ levels. Thus, the H2O2 formed accumulated in the medium if contaminating Ca2+ was present; in the absence of Ca2+, H2O2 was completely removed if it originated from monoamine oxidase, but was less completely removed if it originated from DACHR. We propose that the primary action of rotenone is to promote extracellular \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \(\mathrm{O}_{2}^{{\bar{{\cdot}}}}\) \end{document} release via activation of NADPH oxidase in the microglia. In turn, \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \(\mathrm{O}_{2}^{{\bar{{\cdot}}}}\) \end{document} oxidizes DA to DACHR extracellularly. (The reaction is favored by the lack of GSH, which would otherwise preferably produce GSH adducts of dopaminoquinone.) Once formed, DACHR (which is resistant to GSH) enters neurons to activate the rotenone-stimulated redox cycle described.

Inhibitors of Complex I of the mitochondrial respiratory chain, such as rotenone, promote Parkinson disease-like symptoms and signs of oxidative stress. Dopamine (DA) oxidation products may be implicated in such a process. We show here that the o-quinone dopaminochrome (DACHR), a relatively stable DA oxidation product, promotes concentration (0.1-0.2 M)and respiration-dependent generation of H 2 O 2 at Complex I in brain mitochondria, with further stimulation by low concentrations of rotenone (5-30 nM). The rotenone effect required that contaminating Ca 2؉ (8 -10 M) was not removed. DACHR apparently extracts an electron from the constitutively autoxidizable site in Complex I, producing a semiquinone, which then transfers an electron to O 2  oxidizes DA to DACHR extracellularly. (The reaction is favored by the lack of GSH, which would otherwise preferably produce GSH adducts of dopaminoquinone.) Once formed, DACHR (which is resistant to GSH) enters neurons to activate the rotenone-stimulated redox cycle described.
Parkinson disease (PD) 1 is one of the major human neurodegenerative disorders and is clinically characterized by resting tremors, rigidity, slowness of voluntary movement, and postural instability. The neuropathological hallmark of PD is the progressive loss of the nigrostriatal dopamine (DA)-containing neurons, the cell bodies of which are in the substantia nigra pars compacta and nerve terminals in the striatum (1,2). Lewy bodies are characteristic aggregates that form in affected cells, and increasing evidence suggests that ␣-synuclein, ubiquitin, and iron constitute a major fraction of Lewy body aggregates (3). Defects in mitochondrial Complex I activity and DA deficiency have been detected in patients suffering from the disease (4 -6). Mitochondrial Complex I inhibitors such as 1-methyl-4-phenylpyridinium (MPP ϩ ), a metabolic product of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), and rotenone have been shown to induce symptoms similar to those of PD in experimental animals and humans (7,8). Several in vitro studies using MPP ϩ and rotenone have shown that these compounds can induce oxidative stress, apoptosis, and other biochemical changes similar to those observed in patients afflicted with idiopathic PD (9,10). The conversion of MPTP to MPP ϩ is a complex process that is required for toxicity and that is catalyzed by monoamine oxidase (MAO) type B, localized in non-DA glial cells, followed by spontaneous oxidation to MPP ϩ (11,12). The latter is taken up by DA neurons via the DA transporter and is concentrated by an active process in mitochondria (13). Rotenone needs no specific transport mechanism. It is the classical Complex I inhibitor that indiscriminately affects respiration in every cell type. However, its low specificity is accompanied by the striking specificity of the pathology following its slow infusion, which includes degeneration of the substantia nigra pars compacta neurons accompanied by the formation of inclusions similar to authentic Lewy bodies (14). It therefore appears that some link must exist between the presence of the neurotransmitter DA and mitochondrial Complex I inhibitors for the development of toxicity.
Oxidative stress appears to contribute to the neurodegeneration observed in PD (8,(15)(16)(17). Brains of PD patients have decreased levels of GSH and exhibit oxidative damage to DNA, lipids, and protein (18,19). Reactive oxygen species responsible for damage are supposed to be produced during DA metabolism or oxidative phosphorylation (20). Normal O 2 -consuming metabolic processes (e.g. the mitochondrial electron transport chain and the oxidative deamination of DA by MAO) and nonenzymatic autoxidation of DA are believed to result in intraneuronal formation of O 2 . and H 2 O 2 (21 (25). When a thiol like GSH is present at the site of DAQ formation, a nucleophilic addition of the thiol at C-5 of DAQ takes place (see Scheme 1). This reaction prevails over the intramolecular cyclization (26). Neuromelanin, the dark pigment that accumulates with age in the substantia nigra, is the product of DA oxidation processes induced primarily by DAQ. Analysis of its composition indicates that it is formed in part from compounds originating from the interaction of DAQ with cellular GSH, leading to 5-S-thiol derivatives (mainly 5-S-cysteine, which is formed by degradation of the GSH adduct) (21). This is what is expected from the facile interaction of DAQ with GSH. However, ϳ50% of neuromelanin derives from cyclization products (i.e. from DACHR-like species) (27,28). As noted, such compounds have to originate from a cellular location where GSH is not present at sufficiently high concentrations. This observation is a strong indication that DACHR is formed in vivo.
The mechanism by which DA metabolites may be involved in reactive oxygen species generation is, however, not fully elucidated, and the molecular basis for facilitation of reactive oxygen species production in DA-containing neurons is not established. In this study, we report that the interaction between DACHR and respiring brain mitochondria generates H 2 O 2 . We show that DACHR dose-dependently increases the production of H 2 O 2 constitutively observed at Complex I of the mitochondrial respiratory chain and that the presence of very low, marginally inhibitory concentrations of the Complex I inhibitor rotenone further increase peroxide production. We also report on Ca 2ϩ control of H 2 O 2 generation and removal. The relevance of these results in the pathogenesis of PD is discussed.
Preparation of Rat Brain Mitochondria-Brain mitochondria were prepared as described (29). Specifically, the cerebral cortices of two 6 -7-week-old rats were rapidly removed into 20 ml of ice-cold isolation medium (320 mM sucrose, 5 mM MOPS, 0.5 mM EDTA, and 0.05 mM EGTA, pH 7.3) and homogenized. The homogenate was centrifuged at 900 ϫ g for 5 min at 4°C. The supernatant was centrifuged at 8500 ϫ g for 10 min, and the resulting pellet was resuspended in 1 ml of isolation medium. This was layered on a discontinuous gradient consisting of 4 ml of 6% Ficoll, 1.5 ml of 9% Ficoll, and 4 ml of 12% Ficoll (all prepared in isolation medium) and centrifuged at 75,000 ϫ g for 30 min. The myelin, synaptosomal, and free mitochondrial fractions formed above the 6% layer, as a doublet within the 9% layer, and as a pellet, respectively. The pellet was resuspended in 250 mM sucrose and 10 mM K-MOPS, pH 7.2, and centrifuged at 8500 ϫ g for 15 min before being resuspended in this last medium to 10 -20 mg of protein/ml by the Gornall protein assay. The mitochondria were well coupled as judged by the increase in the oxygen consumption rate upon addition of ADP (respiratory control ratio), which was between 3.5 and 6 with glutamate and malate as substrates.
Oxygen Consumption-Oxygen consumption was monitored with a Clark-type oxygen electrode in a 1.6-ml closed chamber thermostatted at 30°C with continuous stirring (30).
DACHR Generation-DACHR was generated in the test cuvette by incubating variable amounts of DA with tyrosinase (44 units/ml) for 5 min prior to the addition of mitochondria (31). The reaction proceeded without the liberation of reactive oxygen species, leading to the quantitative generation of DACHR as determined by the consumption of 2.0 nanoatoms of oxygen/nmol of DA and the appearance of typical absorbance at 470 nm (molar extinction coefficient of 3.24 ϫ 10 3 ) (Scheme 1) (28,32). It is shown that DA is oxidized by tyrosinase to DAQ, which readily undergoes irreversible spontaneous 1,4-intramolecular cyclization to give 5,6-dihydroxyindoline (leuco-DACHR), which in turn is oxidized to DACHR by one molecule of DAQ. Alternatively, the interaction of DA with tyrosinase was SCHEME 1 induced in the presence of 50 M GSH. Under these conditions, the high reactivity of DAQ with the thiol led to the quantitative formation of the GSH adduct at C-5 (5-S-glutathionyldopamine) as evidenced by the consumption of only 1 atom of oxygen/molecule of DA and the absence of an absorbance increase at 470 nm (Scheme 1). In our experiments, the formation of DACHR was conducted in the presence of bovine serum albumin, which decreased the yield of DACHR by 20 -25%, probably due to interaction of the thiols in bovine serum albumin with DAQ. The concentration of DACHR under each experimental condition was determined by the absorbance at 470 nm.
Hydrogen Peroxide Measurements-5 M Amplex Red and 15 g/ml HRP (3.75 units) were included in the incubations. H 2 O 2 was detected by the formation of the fluorescent Amplex Red oxidation product resorufin using excitation and emission wavelengths of 563 and 587 nm, respectively, on a Shimadzu RL-5000 spectrofluorometer in a stirred cuvette thermostatted at 30°C (33). The H 2 O 2 calibration scale is linear in the 0 -6 M range, and at the end of each assay, traces were calibrated by the addition of H 2 O 2 (500 pmol).

RESULTS
As shown in Fig. 1, H 2 O 2 was generated by brain mitochondria respiring on the NAD-linked substrates glutamate and malate. The release rate was low and could be reliably detected only with the high sensitivity reagent Amplex Red. H 2 O 2 production is to be ascribed to autoxidation of some redox component within  In the experiments described above, the Ca 2ϩ contamination of the incubation mixture (8 -10 M) was not removed. When such contaminating Ca 2ϩ was removed by EGTA, the constitutive H 2 O 2 production and the stimulated release induced by DACHR were essentially unchanged. However, rotenone stimulation was less evident (Fig. 2B). The decreased rotenone effect in the presence of EGTA was not due to a variation of the inhibitory potency of rotenone on respiration (Fig. 2C) and appeared to depend solely on the removal of free Ca 2ϩ , which does not mimic the heavy metal chelator TPEN (15-30 M) or the iron chelator deferoxamine (15 M) (data not shown). Adding small amounts of Ca 2ϩ to the incubation medium (with EGTA omitted) did not increase the rotenone effect further (data not shown).
DACHR-and MAO-dependent Production of Peroxide and Its Removal by Mitochondria: Effect of Ca 2ϩ -The results presented above were obtained by monitoring H 2 O 2 as soon as it formed with the high affinity trap Amplex Red/HRP. Such a situation does not exist in vivo, where mitochondrially produced H 2 O 2 has access to cellular peroxidases, notably the mitochondrial glutathione peroxidase. Regeneration of GSH is via glutathione reductase, which utilizes NADPH produced by the energy-dependent transhydrogenase. We recently reported that this system is under the control of intramitochondrial Ca 2ϩ , which acts as an inhibitor of both glutathione reductase (maximal inhibition of 70%, EC 50 ϭ 0.9 M Ca 2ϩ ) and glutathione peroxidase (maximum inhibition of 40%, EC 50 ϭ 0.9 M Ca 2ϩ ) (29). Such Ca 2ϩ -mediated control of H 2 O 2 removal was evident when we analyzed the ability of mitochondria to dispose of H 2 O 2 generated by MAO (which is located at the outer mitochondrial membrane) using DA as the substrate (Table I) To address the relationship between production and removal of H 2 O 2 when the latter originates from DACHR redox cycling, we performed a series of experiments with DACHR in which we monitored the intrinsic rate of H 2 O 2 production using the Amplex Red/HRP trap and compared it with the accumulated H 2 O 2 in the suspension medium as determined by adding the Amplex Red/HRP detection system at a fixed time (5 min) of the reaction (Table II). As described above (Fig. 2), trap measurements with glutamate/malate showed a strong stimulation of peroxide production by DACHR, which was further increased by rotenone in the absence of EGTA. Residual H 2 O 2 accumulation in the suspension medium was relatively high in the absence of EGTA without and especially with rotenone, suggesting that the peroxidase activity was not sufficient to dispose of all the H 2 O 2 produced. In the presence of EGTA, accumulated H 2 (29). Rotenone at a high concentration acts as an inhibitor, an indication that the autoxidizable carrier is located uphill of the rotenone inhibition site (29,35,36). Furthermore, H 2 O 2 release is inhibited by modestly increasing intramitochondrial Ca 2ϩ (i.e. in the presence of contaminating Ca 2ϩ ) and is partially inhibited by the contemporary oxidation of NAD-linked substrates (29). In vivo, succinate is never oxidized alone by the mitochondria, and its concentration is dictated by the concentration of other tricarboxylic acid cycle substrates, which are NAD-dependent. Some of the experiments in Table II   Generation, but DACHR-dependent Peroxide Production Is Not Prevented by GSH-When the tyrosinase-catalyzed oxidation of DA was performed in the presence of GSH, no DACHR formation took place, and 5-S-glutathionyldopamine was formed instead (see "Experimental Procedures" and Scheme 1). This is explained by the high reactivity of tyrosinase-produced DAQ with thiols, taking precedence over cyclization of DAQ to leuco-DACHR. When tested during respiration-dependent H 2 O 2 production, the GSH adduct exhibited no activity both without and with rotenone (Fig. 3). However, if DACHR was first produced (i.e. in the absence of GSH), and GSH was supplied later together with mitochondria, the respiration-promoted H 2 O 2 release and stimulation by rotenone were unmodified. Accordingly, the DACHR absorbance at 470 nm was not altered by GSH (data not shown). Thus, once formed, DACHR is relatively stable and is capable of performing redox cycling also in the presence of GSH. DISCUSSION Reactive oxygen species originating from DA oxidation products have long been suspected to participate in the pathogenesis of PD. DA oxidation to DAQ depends on its interaction with O 2 . , peroxynitrite, and other reactive species such as ⅐ OH originating via Fenton chemistry from H 2 O 2 and metals such as iron, which is elevated in PD substantia nigra. DAQ is unstable and readily undergoes additional reactions with thiols such as GSH or cyclization to leuco-DACHR, followed by oxidation to DACHR. Furthermore, DA can be transformed into DAQ by  cyclooxygenase, being used as an electron donor in the peroxidase component of the cyclooxygenase reaction, where prostaglandin G 2 (but also H 2 O 2 ) is the electron acceptor (25). Although cyclooxygenase-1 is constitutively present in the microglia, cyclooxygenase-2 has recently been shown to be present in DAergic neurons of the substantia nigra of PD patients (and to be induced upon MPTP treatment), and this may represent a strong contribution to DA oxidation (38). The involvement of Complex I of the respiratory chain in the pathogenesis of PD has long been suspected because Complex I inhibitors such as MPTP-derived MPP ϩ and, more recently, rotenone have been shown to promote PD-like symptoms (7,8) and because the activity of Complex I has been reported to be decreased in affected individuals (4 -6). Particularly impressive is the finding that a slow infusion of low concentrations of rotenone, a very specific, highly hydrophobic inhibitor, induces typical PDlike lesions, including the formation of Lewy bodies (14).
In the absence of respiratory chain inhibitors, Complex I is the main site where O 2 . and H 2 O 2 are generated (29,36).
is formed from both NAD-linked substrates and succinate (in this case, via reverse electron flow to Complex I). In the first case, the production is very slow and is increased by fully inhibitory concentrations of the Complex I inhibitor rotenone, which acts distal to the autoxidizable component. Succinatedependent H 2 O 2 production is much faster. It is inhibited by rotenone (which interrupts the backwards electron transfer), by decreasing ⌬⌿ (such as during ADP-stimulated respiration), and by small amounts of Ca 2ϩ (such as present as a contaminant in saline medium) (29).
We have shown in this study that H 2 O 2 is formed following the interaction of DACHR with respiring mitochondria, apparently exclusively at the level of Complex I, with properties that are complex and very similar to those of constitutive H 2 O 2 generation, but at higher rates. Stimulation of H 2 O 2 release above basal levels was detected at DACHR concentrations as low as 0.1-0.2 M, and the rate increased strongly and essentially linearly with increasing DACHR concentrations. In essence, it appears that the site within Complex I responsible for the monoelectronic reduction of O 2 to O 2 . is also capable of transferring an electron to the o-quinone DACHR, with the likely generation of a highly reactive semiquinone radical, which in turn transfers an electron to O 2 , forming O 2 . and regenerating DACHR, which operates redox cycling. Electron transfer from Complex I to DACHR takes precedence over transfer to O 2 , given that its rate is much faster at much lower concentrations. A similar mode of interaction has previously been proposed for adrenochrome interaction with Complex I (39). Furthermore, a low concentration of rotenone, which marginally inhibits respiration and induces a negligible increase in the constitutive H 2 O 2 output, induced a strong potentiation of peroxide release by DACHR. The rotenone-induced extra electron leakage to DACHR was largely prevented by removing contaminating Ca 2ϩ . When some succinate was present together with glutamate/ malate, which may mimic a more physiological situation, H 2 O 2 release was generally higher than with glutamate/malate alone also when DACHR was omitted. In particular, the succinate potentiation of constitutive H 2 O 2 release was highest with EGTA because the succinate component of the overall H 2 O 2 production was Ca 2ϩ -inhibited. In this situation of higher constitutive H 2 O 2 production, the inclusion of DACHR promoted an additional large increase, as may be expected if the sites of O 2 and DACHR reduction are the same (and if the reaction with DACHR takes precedence over that with O 2 ). The succinate-dependent increased constitutive H 2 O 2 production was largely eliminated when ⌬⌿ was decreased by ADP. Accordingly, also the high rate of H 2 O 2 output induced by DACHR with the three substrates decreased to values similar to the no-succinate situation in the presence of ADP. In summary, the high peroxide output obtained with succinate plus glutamate/malate in the presence of EGTA with (and, to a lesser extent, also without) DACHR was readily eliminated as soon as ⌬⌿ was decreased. It is likely that, in vivo, ⌬⌿ is not at its maximum and that these high levels of H 2 O 2 released are not normally observed. When EGTA was omitted, the succinate-dependent component of H 2 O 2 release was largely eliminated, and peroxide production was closer to that observed without succinate both without and with DACHR. The site of O 2 . /H 2 O 2 production with DACHR is located on the inner face of the inner mitochondrial membrane. An important point that has been addressed in this study is how the mitochondrial peroxidase(s) deals with mitochondrially produced H 2 O 2 . Allowing H 2 O 2 formed at a known rate to accumulate in the suspension medium for some minutes before introducing the detection system provided information about how H 2 O 2 is handled. We found recently that the glutathione reductase/glutathione peroxidase system, responsible for much of the peroxide removal in mitochondria, is inhibited by low intramitochondrial Ca 2ϩ levels (29). At 75 M DA, a relatively high concentration that may, however, be of physiological significance, MAO activity monitored by immediately measuring H 2 O 2 release was unaffected by mitochondrial respiration or by EGTA. In contrast, H 2 O 2 that accumulated in the medium was highest without respiration; it was greatly decreased with substrates and no EGTA, but was totally absent if EGTA was also present. Also the DACHR-dependent accumulation of H 2 O 2 was under the control of intramitochondrial Ca 2ϩ and was at its lowest when contaminating Ca 2ϩ was removed by EGTA. However, there are indications that the glutathione reductase/ glutathione peroxidase system is partially inhibited in the presence of DACHR. In fact, when the DACHR-induced peroxide production was high, i.e. in State 4 with succinate and EGTA (a rate that was very similar to the MAO activity at 75 M DA), accumulated peroxide was still relatively high, whereas it was undetectable when it originated from MAO (compare Tables I and II). Some H 2 O 2 accumulated under similar conditions even in the absence of succinate. A possible explanation for the decreased H 2 O 2 -removing activity with DACHR-derived peroxide is that, with DACHR, the primary product formed is the intramitochondrial DACHR semiquinone, followed by electron transfer to O 2 , forming O 2 . . It is likely that some intramitochondrial DACHR semiquinone accumulates, escaping reaction with O 2 , and that such a species or its derivatives interact with mitochondrial proteins, altering their activity. Decreased glutathione reductase/glutathione peroxidase activity in the presence of intramitochondrial DACHR semiquinone would likely secondarily affect also the removal of MAO-derived H 2 O 2 . It is important to emphasize that the duration of the experiments in this study was rather short. It follows that this aspect may likely become of great importance in longer time frames. The control of both MAO-and DACHR-induced H 2 O 2 production by intramitochondrial Ca 2ϩ (an increase in which prevents efficient peroxide removal, leading to a higher H 2 O 2 steady state) may be an important new aspect of cell physiology. The Ca 2ϩ increase required for the inhibition of H 2 O 2 removal is small (semimaximal effect at 0.9 M (29)) and readily achieved in vivo. It is the same order of increase required for the activation of pyruvate, isocitrate, and ␣-oxoglutarate dehydrogenases (40). Increased intracellular Ca 2ϩ levels in PD have been repeatedly suggested. Recently, it was shown that the inhibition of Ca 2ϩactivated proteases (calpains) attenuates MPTP toxicity, directly involving increased Ca 2ϩ levels in the pathogenesis of PD (41).
An increase in cytosolic Ca 2ϩ levels is readily followed by increased mitochondrial Ca 2ϩ levels (42).
Both in vitro and in vivo studies have established that the primary lesions in rotenone-induced toxicity are not in DAergic neurons, but in the microglia (43)(44)(45). Extremely low rotenone concentrations promote the extracellular release of NADPH oxidase-derived O 2 . from the microglia (43). The importance of NADPH oxidase in the pathogenesis of PD has been well documented (44,46). Such activation of glial derived O 2 . may indeed be the primary event in rotenone toxicity (and likely also in MPP ϩ toxicity because MPP ϩ is produced from MPTP in the microglia and is actively taken up first by the mitochondria in these cells). Although it seems unlikely that significant DA levels are present in the microglia, it is conceivable that microglial released O 2 . initiates DA oxidation by acting on extracellular DA in the synaptic cleft or its vicinity via monoelectronic DA oxidation to the DA semiquinone radical, followed by dismutation of the semiquinone to produce DAQ. The low concentration of GSH in the extracellular environment may favor the transformation into DACHR of the unstable DAQ, the interaction of which with thiols would otherwise be strongly favored.
Extracellularly formed DACHR would be taken up by neuronal DA transporters to initiate the redox cycling described above, further amplified by the rotenone effect on neuronal mitochondria, in particular in situations in which mitochondrial Ca 2ϩ levels are elevated. The hypothesis that DACHR formation is primarily an extracellular event is supported by the finding that ϳ50% of DA oxidation products in neuromelanin are derived from DAQ cyclization to leuco-DACHR (27,28), which requires low GSH concentrations. Once formed and transferred inside the DAergic neurons, DACHR is resistant to the interaction with GSH and can promote long lasting redox cycling. Intracellularly formed DAQ, via cyclooxygenase-2 expression in PD DAergic neurons (38) or by other means, may also originate DACHR, probably at a later time given the decreased GSH in PD; it may be easily transformed into thiol derivatives, which themselves are toxic (47). The elevation of neuronal intracellular Ca 2ϩ levels could itself depend on the formation of extracellular O 2 . -derived H 2 O 2 . In fact, low extracellular H 2 O 2 levels were recently reported to facilitate Ca 2ϩ influx by promoting the opening of voltage-gated Ca 2ϩ channels (48 -50). H 2 O 2 also promotes iron uptake via the transferrin receptor (51). DACHR-dependent oxidative stress, potentiated by increased intracellular Ca 2ϩ levels and further amplified by the increased iron availability, may secondarily promote the production and toxicity of ␣-synuclein by favoring its polymerization (52-54), cyclooxygenase-2 generation, etc. The suggested sequence of events may be important in the pathogenesis of PD, also independently of the toxicological situations represented by MPTP or rotenone intoxication. In fact, a glial inflammatory component involving NADPH oxidase activation in the generation of extracellular O 2 . is likely an obligatory step in the induction of PD (44,55). Data on the efficacy of anti-inflammatory agents in PD are accumulating (55,56). In summary, we have shown that the DA oxidation product DACHR promotes respiration-dependent H 2 O 2 production in mitochondria. H 2 O 2 was released at the same site in Complex I responsible for H 2 O 2 release in the absence of DACHR. DACHR-induced H 2 O 2 was visible at submicromolar concentrations and was strongly stimulated by extremely low rotenone concentrations. The rotenone effect required that contaminating Ca 2ϩ (8 -10 M) was not removed from the incubation medium. H 2 O 2 derived from DACHR or from MAO activity was subject to removal by mitochondrial glutathione reductase/glutathione peroxidase. In turn, the activity of these enzymes was under the inhibitory control of intramitochondrial Ca 2ϩ , which controlled peroxide accumulation in the suspension medium. In the presence of contaminating Ca 2ϩ , H 2 O 2 (from MAO as well as DACHR) accumulated to a relatively large extent in the medium, whereas in the absence of Ca 2ϩ , MAO-derived H 2 O 2 was completely removed. Also DACHR-derived peroxide was largely eliminated, but the removal was less efficient when the DACHR-dependent production rate was high. Based on the finding that very low rotenone concentrations promote the activation of NADPH oxidase in the microglia, we propose that, in rotenone-dependent parkinsonism, NADPH oxidase-derived superoxide triggers the extracellular oxidation of DA to DAQ and DACHR and that this species enters neurons to initiate the (rotenone-potentiated) peroxide production, supported by the increase in intramitochondrial Ca 2ϩ levels (also possibly dependent on NADPH oxidase-derived H 2 O 2 activation of voltage-dependent Ca 2ϩ channels).