Mitochondrial Structure and Bioenergetics in Normal and Disease Conditions

The proton gradient produced during respiration, similar to an accumulator, supplies energy to operate the ATP synthase (complex V) which, through dissipating it, provides the energy to condense ADP and Pi into ATP. Thus, respiration is distinct but coupled to phosphorylation, in the oxidative phosphorylation pathway. In a normal adult human being, this process leads to the daily production of approximately 70 kg of ATP, which provides the energy necessary for all exergonic reactions of the organism.

Mitochondria also play a role in the regulation of programmed cell death, called apoptosis, which is required for embryonic development and numerous physiological functions. Apoptosis leads to a controlled and programmed cell death, which can occur as a response to various damages or stressors, such as DNA damage, oxidative stress, immune reactions and absence of certain growth factors, hormones and cytokines, or as a natural part of development and aging [58]. Different apoptotic pathways exist, characterized by different triggers but with a common final execution pathway. Indeed, these different pathways lead to activation of initiator caspases (as caspase 8 and 9), which then activate executioner caspases (as caspase 3 or 7), to finally induce the degradation of cellular components. The extrinsic or death receptor pathway, which does not directly involve the mitochondria, is activated by extracellular ligands binding to death receptors on the plasma membrane and leads to the formation of the death-inducing signaling complex (DISC), which subsequently activates the initiator caspase 8 and then the executioner caspase 3 [59]. The best characterized is the mitochondrial or intrinsic pathway of apoptosis. Mitochondrial apoptosis is initiated by internal signals of stress or damage that usually lead to a bioenergetic failure and decrease ∆P and it consists in the mitochondrial outer membrane permeabilization (MOMP), regulated by the Bcl-2 family of proteins. Upon stressors, the pro-apoptotic members BAX and BAK oligomerize at the OMM [60,61], where they induce the release of pro-apoptotic proteins from the IMS into the cytosol, including cytochrome c [62]. Once in the cytosol, cytochrome c binds and activates apoptotic protease activating factor-1 (Apaf-1) as well as procaspase-9, forming a complex known as the “apoptosome”.

Active caspase 9 is then able to cleave and activate caspase 3, starting the communal execution pathway [63,64]. The execution pathway leads to DNA fragmentation, degradation of cytoskeletal and nuclear proteins, cross-linking of proteins and formation of apoptotic bodies. Thus, bioenergetic failure often leads to mitochondrial apoptosis, although it may also cause autophagy of individual spent mitochondria, or even to ablation of a portion of dysfunctional portions of mitochondria that are left surviving. Therefore, apoptosis and related phenomena constitute one of the possible deleterious outcomes of mitochondrial disease. Importantly, in addition to the release of cytochrome c and formation of the apoptosome, other mitochondrial proteins can determine apoptosis through alternative mechanisms, such as the apoptosis inducing factor 1 (AIF1) [65], a redox mitochondrial membrane-bound protein that in stress conditions can be cleaved by cathepsin or other proteases, released outside mitochondria and there activate caspases independently from the formation of apoptosome, eventually leading to apoptosis. Recessive mutations of AIF1 have been reported in severe infantile syndromes associated with multiple defects of the mitochondrial respiratory chain activities [66].

Calcium is largely used in cells as a signaling molecule; therefore, its regulation is critical. Cellular organelles such as the ER and mitochondria are able to sequester and release calcium, regulating the cellular concentration of the ions. Vice versa, calcium signaling has a role in mitochondrial functionality, even if not all the molecular mechanisms involved are clear yet. What is largely accepted is that calcium in the mitochondrial matrix regulates various enzymes, such as pyruvate, isocitrate and 2-oxoglutarate dehydrogenases, modulating, as a consequence, mitochondrial respiration, and the induction of the mitochondrial permeability transition pore [67].

Calcium passes the OMM barrier through a VDAC, which is characterized by high-conductance and weak anion selectivity [68]. The passage through the IMM, instead, is more controlled and involves the mitochondrial calcium uniporter (MCU), which transports Ca2+ inside the matrix, and a Na+/Ca2+ exchanger, mostly expressed in excitable cells, such as muscles and brain, or an H+/Ca2+ exchanger, in other cell types, which release calcium from the matrix to the IMS [69].

Heme is an iron-containing porphyrin, essential in numerous biological processes, such as oxygen transport and storage, drug and steroid metabolism, signal transduction and microRNA processing [70]. Moreover, heme is incorporated in some subunits of the electron transport chain and it is necessary for cellular respiration [71].

The synthesis of this compound occurs both partially in the mitochondria and in the cytosol [72]. Heme is generated by the insertion of ferrous iron into the tetrapyrrole macrocycle of protoporphyrin IX [73] catalyzed by a mitochondrial matrix enzyme, called ferrochelatase. Protoporphyrin IX is produced starting from glycine and succinyl-CoA.

Most of heme production takes place in erythroid progenitors, followed by the liver for the formation of heme-containing enzymes [73]. The synthesis pathway is conserved in these two cellular types, while its regulation differs. The heme synthesis machinery in the liver has a rapid turnover in order to respond quickly to changes in metabolic requirements, while the synthesis in developing red cells is tied to the availability of iron.

Iron-sulfur (Fe/S) clusters are prosthetic groups with a variety of biological functions. Indeed, several enzymes, such as glycosylases, helicases, primases and respiratory chain enzymes, require the incorporation of Fe/S centers for their activity [74]. Both the proteins containing these centers and the proteins that are part of the biosynthesis machinery are highly conserved in prokaryotes and eukaryotes, suggesting an important role in the origin of life [75]. These cofactors originated probably in an environment characterized by low oxygen and co-evolved when the oxygen levels started increasing in the atmosphere, leading to the adaptation of anaerobic electron transport chains for an aerobic habitat. Most Fe/S proteins contain a rhomboid [Fe2/S2], a cuboidal [Fe3/S4] or a cubane [Fe4/S4] cluster [76]. The most common protein ligand is cysteine, but also histidine, serine and arginine can form a bond [74].

In yeast, Fe/S clusters synthesis takes place exclusively in mitochondria, but an Fe-S cluster can also be exported from mitochondria by a specific ABC transporter in the inner mitochondrial membrane [77]. The central actor of the synthesis pathway found in eukaryotic mitochondria is known as iron-sulfur cluster assembly enzyme (ISCU), which acts as a scaffold for the initial synthesis of a [2Fe–2S] cluster. This cluster will be the basis for the formation of both mitochondrial and cytosolic Fe/S groups. The sulfide ions used in this process are obtained from cysteine side chains, thanks to the activity of enzymes called cysteine desulfurases [78], while it is not clear how iron is delivered to the ISCU, although frataxin may be involved [79].

The enzymatic machinery performing cellular respiration, the electron transport chain (ETC), is composed of four protein complexes embedded in the IMM and two mobile electron carriers (ubiquinone, or coenzyme Q, and cytochrome c) (Figure 5). Electrons are transported from electron carriers reduced during glycolysis and the Krebs cycle (NADH and FADH2) to coenzyme Q and cytochrome c and eventually transferred to O2, forming H2O. The energy liberated by this chain of redox reactions leads to the generation of an electrochemical proton gradient across the IMM, which is used by complex V or F1Fo ATP synthase to generate ATP [80,81].

The formation of the electrochemical gradient is made possible by the nature of the phospholipidic bilayer that forms the IMM. Indeed, the membrane is impermeable to the passage of protons, which require protein transporters to cross it. These transporters are part of complexes I, III and IV and the energy necessary for the proton pumping and the generation of the electrochemical gradient derives from the transport of electrons.

This gradient produces the proton motive force (PMF or ∆p), which can be described as a measure of the potential energy stored across the IMM. Since protons are electrically charged particles, the PMF has both chemical and electric components. The electric component corresponds to the voltage difference across the membrane and the free energy is calculated as ΔG = −F∆Ψ (F = Faraday constant; ∆Ψ = membrane potential). The chemical component, instead, has a free energy calculated as ∆G = RT ln([H+]i/[H+]o), where [H+]i and [H+]o refer to the proton concentrations inside and outside the IMM, respectively, R is the gas constant of 1.987 cal/(degree·mol), and T is the temperature (in degrees Kelvin). Combining these two components, the PMF is calculated as ∆p = ∆Ψ − (RT/F) * ln([H+]i/[H+]o). Under physiological conditions, the magnitude of the PMF is about −220 mV [71]. As a consequence of the difference in protons concentration, the matrix side of the inner mitochondrial membrane is negatively charged and slightly alkaline (pH = 8).

The first actors in the electron transport chain are NADH and FADH2 (Figure 6). NAD+ and FAD+ are reduced to NADH and FADH2, respectively, during glycolysis or beta oxidation of fatty acids and the citric acid (Krebs) cycle. A 1:1 mixture of NADH and NAD+ has a redox potential of −320 mV, while the midpoint redox potential of FADH2 is around −220 mV. This means that both these molecules have a strong tendency to donate electrons [54]. The ΔG°’ values for these strongly exergonic reactions are −52.6 kcal/mol (NADH) and −43.4 kcal/mol (FADH2) [71].

NADH binds to complex I (CI, NADH dehydrogenase) and is oxidized to NAD+, donating two electrons to a flavin mononucleotide (FMN) (Figure 6), inserted in CI subunit NDUFV1. Electrons are then passed to a chain of eight iron-sulfur (Fe/S) clusters, in order to be eventually transferred to the oxidized form of coenzyme Q or ubiquinone (Q), which uptakes two protons to form ubiquinol (QH2). As the electrons are transferred from one redox center to the other, four protons are pumped through CI out of the matrix. Despite the numerous biochemical and structural studies on CI, a definitive model of redox-coupled proton pumping does not exist yet. Many models have been proposed: the first hypotheses suggested conformational changes in antiporter-like subunits in the P-module, allowed by the energy produced during electron transport [82,83], or transient hydration changes able to generate water-gated pathways for proton transfer between conserved ionizable residues along the membrane domain [84]. The energy necessary for the proton translocation could be provided by two processes: a two-stroke mechanism where the pumping is coupled with N2 (the terminal cluster in the Fe/S chain) reduction/re-oxidation, which occurs twice for every NADH oxidized, assuming the transfer of one electron at a time, or a single-stroke mechanism, where all four protons are translocated together after the reduction of coenzyme Q [85,86,87]. More recent analyses of the X-ray structure of the Y. lipolytica enzyme, instead, led to the hypothesis that proton pumping is linked to the coordinated conformational rearrangement of three loops in subunits ND1, NDUFS2 and ND3, triggered by the binding of negatively charged ubiquinone [88].

FADH2 derives from the oxidation of succinate to fumarate by complex II (CII, succinate dehydrogenase) during the Krebs cycle (and the last steps of beta oxidation), a reaction that reduces FAD+ to FADH2, a cofactor bound to the flavoprotein subunit (SDHA). Then, two electrons are transferred to the Fe/S clusters contained in SDHB, which will eventually pass them to Q (Figure 6). This process results in an increased ubiquinol pool but does not directly influence the proton gradient because CII is not a proton pump.

Coenzyme Q is a mobile cofactor that can interact with CI and CII and transports the electrons received to complex III (CIII, Q-cytochrome c oxidoreductase). CIII oxidizes QH2 to Q and passes the electrons to another soluble carrier, cytochrome c, during a process known as the Q-cycle (Figure 7). The Q-cycle consists of two parallel reactions, which involve the three prosthetic groups of the enzyme: the heme groups contained in cytochrome c1 and cytochrome b and the 2Fe/2S cluster contained in the Rieske protein/UQCRFS1 [89]. The first reaction requires the passage of one electron from QH2 bound to the Qo binding site to the iron/sulfur group and then to cytochrome c1, leading to the reduction of cytochrome c. Each cytochrome c is able to bind only one electron and, when reduced, moves from CIII to complex IV (CIV, cytochrome c oxidase). The second electron from QH2 is passed to the two heme b groups (bL and bH) contained in cytochrome b and terminates on a second ubiquinone molecule bound to a different binding site of the enzyme (Qi site). This ubiquinone is partially reduced to a semiquinone (Q−•) during the first Q-cycle and completely reduced to QH2 following a second catalytic cycle [90,91,92,93]. One QH2 molecule is then recycled and two electrons are eventually passed to two cytochrome c molecules. For each QH2 molecule that is oxidized, there is the release of two protons to the intermembrane space. QH2 has a redox potential around 0 mV, while CIII centers have a positive potential, allowing the passage of electrons. In order to permit the two branches of the Q-cycle, different centers of CIII must have different redox potentials, ranging between 72.5 mV for cytochrome b, low enough to allow the recycling of electrons through semiquinone, and 242 mV for cytochrome c1, which passes the electron to cytochrome c (251 mV) [94,95].

During the Q-cycle, CoQ is present in three different forms, according to its redox state: ubiquinone (Q), semiquinone (Q−•) and ubiquinol (QH2). According to the phase of the cycle and the consequent state of the reaction, CoQ molecules can bind CIII in different binding sites: Qo, which faces the IMS and catalyzes the oxidation of ubiquinol to ubiquinone, and Qi, which faces the matrix and catalyzes the reduction of ubiquinone to semiquinone and ubiquinol [96].

The last steps of oxidative phosphorylation take place in CIV (the terminal oxidase), which allows the passage of electrons from cytochrome c to oxygen (redox potential = 820 mV), generating water. Since cytochrome c carries only one electron, four molecules are oxidized in order to generate two H2O molecules from one molecule of O2. In the meantime, four substrate protons are taken from the matrix to form H2O and the other four protons are pumped into the IMS [97,98]. CIV contains two heme groups (cytochromes a and a3) and two copper atoms (CuA and CuB) [99]. Electrons are transferred through the CuA center and heme a to the heme a3/CuB group. When both heme a3 and CuB are reduced, one O2 molecule is recruited to form a peroxide bridge between these two prosthetic groups. This bond is broken by the reaction with protons picked up by the mitochondrial matrix and two H2O molecules are formed [100].

NADH dehydrogenase (Complex I, CI) is the first step of the electron transport chain and is composed of 44 different subunits in mammals [101], organized into three structural domains: a membrane arm, or P-module, and two peripheral domains, the N and the Q modules, protruding in the mitochondrial matrix. The N module contains the FMN cofactor and is responsible for the binding and the oxidation of NADH, while the Q module contains the ubiquinone binding site. The passage of electrons between these two extremities occurs in Fe/S clusters in both the N and the Q modules. The peripheral arm is composed of nuclear-encoded proteins, 7 “core” subunits (NDUFV1, NDUFV2, NDUFS1, NDUFS2, NDUFS3, NDUFS7 and NDUFS8) and 30 accessory subunits necessary to stabilize the enzyme and to protect it from reactive oxygen species (ROS) damage [102]. The P-module, instead, is deputed to the proton pumping activity and contains seven mtDNA-encoded proteins: ND1, ND2, ND3, ND4, ND4L, ND5 and ND6. ND1 forms the reduction site for ubiquinone, while ND2, ND4 and ND5 have been found to share a similar structure to sodium and potassium antiporters and may be involved in proton pumping [103]. The mammalian CI structure is represented in Figure 8.

Due to the large number of subunits forming CI, the assembly pathway of this enzyme is particularly complex and requires the involvement of many assembly factors. The first stage is the synthesis of the various subunits, both inside mitochondria and in the cytoplasm, coupled with the import in the organelle of the nuclear-encoded components. Most CI subunits have N-terminal mitochondrial targeting sequences (MTS), while 11 are imported into the organelle thanks to uncharacterized internal signals within the mature protein [104]. Several core subunits need further maturation and the insertion of the prosthetic groups. However, it is difficult to identify assembly factors with this role using traditional proteomic analysis, probably because of the transient and labile interaction between them and the forming enzyme [105]. The only assembly factor known to be involved in the incorporation of 4Fe/4S clusters in the peripheral arm is NUBPL, a member of the Mrp/NBP35 ATP-binding protein family [106,107]. Moreover, it is not clear if the insertion of the iron/sulfur clusters happens before or after the incorporation of the single subunit into the subcomplex.

The second step of CI assembly is the formation of six independent modules, N, Q, ND1/PP-a, ND2/PP-b, ND4/PD and ND5/PD-b, and the incorporation of each of them in a specific order [108]. All the known CI assembly factors are summarized in Table 1.

The ND2 module is the first detectable after inhibition of mitochondrial protein biosynthesis [109]. This subassembly binds to numerous assembly factors: ACAD9, ECSIT, TMEM126B, NDUFAF1, COA1 and the putative assembly factor TMEM186, which form the mitochondrial complex I intermediate assembly (MCIA) [110]. Moreover, TMEM186 was found to strongly interact with the newly synthesized MT-ND3, which is added to the intermediate together with MT-ND6 and MT-ND4L, forming a 385 kDa structure.

In parallel to this, an intermediate of the Q module starts forming, binding to NDUFAF3 and NDUFAF4 and generating a ~170 kDa structure. This submodule will then bind to the assembly factor TIMMDC1 and the subunits ND1, NDUFA3, NDUFA8 and NDUFA13, to yield a 283 kDa complex [109].

The ND4 module, instead, involves the subunits NDUFB1, NDUFB4, NDUFB5, NDUFB6, NDUFB10 and NDUFB11, together with the assembly factors FOXRED1, ATP5SL and TMEM70. This 230 kDa module binds initially to the N2 module and then to the ND1/Q module intermediate [109].

The ND5 module, which forms the distal extremity of the membrane arm, is the second to last intermediate inserted into the forming enzyme. It is composed of the subunits NDUFB2, NDUFB3, NDUFB7, NDUFB8, NDUFB9 and NDUFAB1, and it is known to bind one assembly factor: DMAC1/TMEM261 [111]. This late intermediate lacking the N module is stabilized by NDUFAF2/NDUFA12L/B17.2L.

Finally, the N module, composed of NDUFV1, NDUFV2, NDUFS1 and NDUFA2, forming a 160 kDa assembly, is incorporated [109]. This last passage completes the assembly of the enzyme, which loses the interaction with the assembly factors and stabilizes as a ~1000 kDa complex.

Mutations affecting CI stability or activity are responsible for a wide range of pathological phenotypes [133]. Missense mutations affecting the mitochondrial-encoded subunits (ND subunits) have been associated with Leber’s hereditary optic neuropathy (LHON), mitochondrial encephalomyopathy, lactic acidosis and stroke-like syndrome (MELAS) and Leigh syndrome. Many mutations in nuclear-encoded subunits have been identified in patients with CI deficiency, causing Leigh syndrome, leukoencephalopathy, leukodystrophy, encephalopathy, cardiomyopathy and other neurological defects. In addition, assembly factors and chaperones involved in CI assembly can also be at the origin of the pathogenesis of these diseases [134]. The main pathological mutations found in CI subunits or assembly factors are summarized in Table 2.

Succinate dehydrogenase (SDH, complex II, CII) is a ~120 kDa integral membrane complex, involved in both the TCA cycle and the ETC. Indeed, this enzyme catalyzes the oxidation of succinate to fumarate, a central step of the citric acid cycle, and reduces FAD to FADH2, which then reduces ubiquinone to ubiquinol [178]. CII is the only complex of the chain that does not pump protons across the membrane and that is entirely encoded by the nuclear DNA.

CII is composed of four subunits, named SDHA-D, forming two domains (Figure 9). The hydrophilic head of CII comprises SDHA and SDHB and is required for the oxidation of succinate. FAD+ binds to SDHA and the electrons are transferred to SDHB, containing three Fe/S clusters ([2Fe-2S], [4Fe-4S] and [3Fe-4S]) [179]. The hydrophobic membrane domain of the enzyme is composed of SDHC and SDHD and contains a heme b group and two ubiquinone binding sites [180].