**5. SCs Association Protects from ROS Damage**

The supramolecular organization of the respiratory chain has revealed a hitherto unknown link between mitochondrial structure and generation of the so-called ROS. In this section we will provide evidence that there is a biunivocal relation between ROS and SC association, i.e., SC association protects mitochondrial structure from ROS damage, but at the same time the SC association limits ROS generation from the respiratory chain.

ROS, in the past considered exclusively as major factors for cellular toxicity, are recognized today to be important physiological factors involved in cell signaling, by affecting the redox state of signaling proteins [221].

ROS is a collective term including oxygen derivatives, either radical or non-radical, that are oxidizing agents and/or are easily converted into radicals. Transition metals such as iron and copper, when in a free state, have a strong capacity to reduce O2. Stepwise addition of an electron generates in sequence the superoxide radical anion O<sup>2</sup> <sup>−</sup>, the peroxide ion (which is a weaker acid and is protonated to hydrogen peroxide H2O2), and the hydroxyl radical OH. The hydroxyl radical is extremely reactive with a half-life of less than 1 ns; thus, it reacts close to its site of formation.

#### *5.1. Mitochondrial Sources of ROS*

In the cell environment, ROS can stem from exogenous and endogenous sources [222]. Exogenous sources of ROS include UV and visible light, ionizing radiation, drugs and environmental toxins. On the other hand, endogenous sources of ROS embrace enzyme

activities such as: xanthine oxidase, cytochrome P-450 enzymes in the endoplasmic reticulum, peroxisomal flavin oxidases and plasma membrane NADPH oxidases. However, the membrane-ETS in the mtIM is probably the major source of ROS. Other mitochondrial enzyme systems can significantly contribute to ROS generation [223], as dihydrolipoamide dehydrogenase (a subunit of the α-ketoglutarate and pyruvate dehydrogenase complexes), monoamine oxidase, mitochondrial nitric oxide synthase and the adaptor protein p66Shc . These systems are not relevant to this review and will not be considered here. α

−

ROS from the Respiratory Chain

In mammalian mitochondria, and particularly in the respiratory chain and related enzymes, at least ten different sites of superoxide/H2O<sup>2</sup> production have been identified [224]. The FMN and CoQ-binding sites of CI and the Q<sup>O</sup> site (at the outer or positive side) of CIII are considered as the most important sites for superoxide production. However, other sites have also been defined (Figure 6).

α **Figure 6.** Major sites of superoxide production from the respiratory chain. The arrows represent the sources of superoxide at different sites in relation to the inner mitochondrial membrane. Blue, CI, NADH-ubiquinone oxidoreductase; yellow, CII, succinate-ubiquinone oxidoreductase; red, CIII, ubiquinol-cytochrome *c* oxidoreductase; green, CIV, cytochrome oxidase; NDi and NDe, internal and external alternative NAD(P)H dehydrogenases; AOX, alternative oxidase; αGP, glycerol-3-phosphate; ETF, electron transfer flavoprotein; Q, Coenzyme Q (ubiquinone); Cyt *C*, cytochrome *c*; (see text for details).

Most of superoxide is generated at the matrix side of the mtIM, as superoxide is detected in SMP, which are inside-out vesicles, namely have an opposite orientation with respect to mitochondria. However, a significant amount of superoxide is formed at the outer face of the mtIM [225]. It is likely that CI releases ROS in the matrix, while CIII releases ROS mostly in the IMS. The superoxide released at the IMS may be directly exported to the cytoplasm through an anion channel related to the Voltage-Dependent Anion Channel, VDAC.

CI is a major source of superoxide production in different mitochondrial types. Several prosthetic groups in CI were proposed to directly reduce oxygen, including FMN [226–228], ubisemiquinone [229,230], and iron sulfur (FeS) cluster N2 [231–233].

Grivennikova and Vinogradov [234] observed both superoxide and hydrogen peroxide generation by CI and attributed superoxide formation to FeS cluster N2, an hydrogen peroxide formation from 2-electron oxidation of fully reduced FMN. Rotenone, a specific inhibitor of CI, enhances ROS formation during forward electron transfer in the respiratory chain [231].

The highest rate of ROS generation in isolated mitochondria occurs during oxidation of succinate [235]. Since this ROS production is inhibited by rotenone [229,236,237], it is commonly attributed to CI during the energy-dependent reverse electron transfer from suc-

cinate to NAD<sup>+</sup> through CoQ [238–240]. See also Section 3 for the role of the reduced CoQ pool in reverse electron transfer and ROS-dependent CI destruction (cf. Guáras et al. [120]).

The formation of superoxide in CIII depends on its peculiar mechanism of electron transfer, the so-called Q-cycle [241]. If the lifetime of the semiquinone (Q) at the outer Q<sup>O</sup> site, is prolonged, as in the controlled state when the electric potential is high, it reacts with oxygen, thus forming superoxide [242,243]. This conclusion is also reached by the superoxide increase induced by the specific CIII inhibitor, antimycin [244].

More recent studies, however, suggest that the oxygen reductant is the semiquinone formed in the so-called semi-reverse reaction in which cytochrome *b*<sup>L</sup> reduces the fully oxidized quinone [245,246]. Osyczka [247,248] proposes that a metastable radical state, nonreactive with oxygen, safely holds electrons at a local energetic minimum during the oxidation of ubiquinol. This intermediate state is formed by interaction of a radical with a metal cofactor of a catalytic site, presumably the FeS cluster, under physiological conditions.

The other sites in the respiratory chain which may be a source of ROS (e.g., CII, glycerol phosphate dehydrogenase, dihydroorotate dehydrogenase, ETF and ETF dehydrogenase) are reviewed in [249].

#### *5.2. Control of ROS Production*

The steady-state concentration of ROS in a cell depends on many factors, including the rate and site of production, their lifetimes and diffusion constants, the interconversions between different ROS, and the removal rate by different antioxidant systems.

The ROS generation by isolated mitochondria accounts for 0.1–0.2% of oxygen consumed, but the rate and extent vary among different tissues and substrates [250,251].

The mitochondrial ROS level is physiologically regulated, as expected for their role as cellular signaling molecules (cf. Section 5.3).

Mitochondrial ROS production increases in State 4 (controlled non-phosphorylating state) when the membrane potential is high and the electron transfer rate decreases [252]. In fact, under this condition the respiratory carriers capable of donating electrons to oxygen are more reduced. For this reason, uncoupling may limit the production of ROS by releasing excessive membrane proton potential. In rat hepatocytes, a futile cycle of H<sup>+</sup> pumping and proton leak may account for 20–25% of respiration [253] and even more in perfused rat muscle. Uncoupling may be achieved by activating proton leak through uncoupling proteins (UCP) [244] (for a different role postulated for uncoupling proteins, cf. Wojtczak et al. [254]). In this way, although a tissue may dissipate a conspicuous part of the energy conserved by its mitochondria, the mitochondrial respiratory chain is maintained under more oxidized conditions, thus preventing the overproduction of damaging species.

UCP2 structure supports a fatty acid cycling mechanism for uncoupling [255]: UCP2 expels fatty acids in their anionic form from the negative matrix, while the protonated acids flow back from the positive exterior by passive diffusion through the bilayer. The resulting effect is H<sup>+</sup> back diffusion to the matrix, so that ∆*p* dissipation decreases superoxide formation. UCP-mediated antioxidant protection and its impairment are expected to play a major role in cell physiology and pathology [256]. Mitochondrial uncoupling could be exploited to treat human diseases, such as obesity, cardiovascular diseases, or neurological disorders [257] (cf. Section 6).

Besides high membrane potential, other reasons may induce a decrease in the rate of electron transfer in CI leading to overproduction of ROS: for example, subunit phosphorylation that inhibits CI activity, enhances its ROS generating capacity [258–260]. Hence, endocrine alterations may affect ROS formation by changing the phosphorylation state of the complex.

Kadenbach et al. [261] proposed a ROS-generating mechanism that depends on regulation of cytochrome *c* oxidase. The cAMP-dependent phosphorylation of subunit I of the oxidase, by increasing its sensitivity to allosteric ATP inhibition, decreases H<sup>+</sup> translocation by the enzyme, with resulting decrease in mitochondrial membrane potential and, consequently, in ROS generation by the respiratory chain. On the contrary, dephosphorylation of CIV under stress conditions would induce increase in membrane potential and a burst of ROS generation by mitochondria. The discovery of a mitochondrial cAMP signalosome [262], modulating the allosteric inhibition of CIV by ATP, underlies a universal mechanism for metabolic regulation in eukaryotes.

In addition, cyt. *c* is regulated by different mechanisms, including post-translational modifications [263,264]. Specific cyt *c* residues may be phosphorylated in a tissue-specific mode. These modifications adjust electron flux and membrane potential to minimize ROS production under normal conditions. Conversely, under pathologic and acute stress conditions, such as ischemia-reperfusion, the dephosphorylation leads to membrane hyperpolarization, excessive ROS generation, and cyt. *c* release.

Other protein modifications in the respiratory complexes can modulate ROS production [265].
