Mitochondrial Structure and Bioenergetics in Normal and Disease Conditions
Mitochondria are ubiquitous intracellular organelles found in almost all eukaryotes and involved in various aspects of cellular life, with a primary role in energy production. The interest in this organelle has grown stronger with the discovery of their link to various pathologies, including cancer, aging and neurodegenerative diseases. Indeed, dysfunctional mitochondria cannot provide the required energy to tissues with a high-energy demand, such as heart, brain and muscles, leading to a large spectrum of clinical phenotypes. Mitochondrial defects are at the origin of a group of clinically heterogeneous pathologies, called mitochondrial diseases, with an incidence of 1 in 5000 live births. Primary mitochondrial diseases are associated with genetic mutations both in nuclear and mitochondrial DNA (mtDNA), affecting genes involved in every aspect of the organelle function. As a consequence, it is difficult to find a common cause for mitochondrial diseases and, subsequently, to offer a precise clinical definition of the pathology. Moreover, the complexity of this condition makes it challenging to identify possible therapies or drug targets.
Keywords: ATP production, biogenesis of the respiratory chain, mitochondrial disease, mi-tochondrial electrochemical gradient, mitochondrial potential, mitochondrial proton pumping, mitochondrial respiratory chain, oxidative phosphorylation, respiratory complex, respiratory supercomplex
According to current theories, mitochondria evolved from free-living bacteria and participated in the origin of eukaryotic cells through a process known as endosymbiosis [1]. The endosymbiotic hypothesis proposes that original anaerobic eukaryotic (modern views indicate endosymbiosis with a non-eukaryotic archaeon organism) cells engulfed the primitive mitochondria and established a favorable interaction (although this occurred after massive negative selection due to gene shuffling) for both the organisms. Indeed, mitochondria were able to drastically improve the cell energy production, generating adenosine triphosphate (ATP) through the respiratory chain, while the host cell offered a safe environment for bacterial proliferation [2].
This theory was supported by the discovery in the 1960s of mitochondrial DNA (mtDNA) and an independent mitochondrial translation system. Indeed, mitochondria contain their own genetic material, mtDNA, which maintains the typical features of bacterial DNA: it is a circular 16,569-base pairs (bp) double-stranded molecule, does not contain introns and is polycistronic [3]. In fact, apart from one non-coding region, called the displacement loop or D-loop, each gene is contiguous to the next one, albeit some are partly overlapped. Moreover, contrary to nuclear DNA, mtDNA is present in many copies in the cell, between 100 and 10,000 copies, proportionally to the energy demand of the specific tissue [3]. Finally, the mtDNA genetic code differs slightly from nuclear DNA, presenting different codons encoding for tryptophan and methionine and only two stop codons. This is true in different species, the genetic code being different, for instance, between vertebrates and other metazoans, but it is a universal code in plants; this implies that the change in the genetic code of mtDNA occurred several times during evolution and it cannot be the primary cause for the maintenance of mtDNA within the organelle.
During evolution, most of the mitochondrial genes were lost or transferred to nuclear DNA, and today, mtDNA only contains 37 genes: 11 messenger ribonucleic acids mRNAs, translated to 13 proteins, 2 ribosomal RNAs (rRNAs, 12S and 16S) and 22 tRNAs [4]. The structure of mtDNA is represented in Figure 1.
As for nuclear DNA mutations, alterations in mtDNA can have important pathological consequences. However, because of the differences between mitochondrial and nuclear DNA, mtDNA inheritance does not follow the canonical mendelian genetics. Firstly, in sexuate organisms, mtDNA is maternally inherited [5]; therefore, only the mother can transmit mutant mtDNA to offspring. Secondly, whilst nuclear genes are present in only two copies per cell, every cell contains multiple copies of mtDNA. These copies can all be identical in sequence, giving a condition known as homoplasmy. However, inheritance of mutated copies, replication errors, oxidative stress or inefficient DNA repair can lead to mtDNA mutations in a percentage of copies, causing heteroplasmy [3]. The proportion of mutant DNA versus the wild-type variant has a strong impact on the development and the severity of the pathological phenotypes.
Mitochondria are surrounded by two phospholipidic membranes, the outer mitochondrial membrane (OMM) and the inner mitochondrial membrane (IMM), which divide the organelle into two spaces, the matrix and the intermembrane space (IMS) [7]. The two membranes present significant differences in lipid composition, characteristics and roles of the transmembrane proteins, permeability and shape and are the result of the endosymbiotic origin of the organelle. Indeed, the OMM is more similar in lipid composition to eukaryotic cell membranes, while the IMM resembles the cardiolipin-containing bacterial membranes [2]. The IMM is characterized by a higher protein/lipid ratio and forms highly packed invaginations in the matrix, called cristae [8]. Embedded in the cristae resides, together with many other proteins, the oxidative phosphorylation (OXPHOS) machinery and one proposed reason for the IMM folding is to increase the available surface for energy production. The part of the IMM that does not protrude in the matrix but, instead, runs parallel to the OMM is called the inner boundary membrane (IBM). Cristae and the IBM are connected via narrow tubular or slit-like structures, the cristae junctions (CJs) [9] (Figure 2).
Moreover, the OMM and IMM differ largely in their permeability. While the OMM allows the passage of ions and small molecules through voltage-dependent anion channels (VDACs) [10], only water, oxygen (O2) and carbon dioxide (CO2) can pass freely through the IMM. This selectivity allows the formation of an electrochemical gradient across the membrane, which forms the basis for ATP production, and the tight regulation of other ions concentrations, such as calcium, largely used in cell signaling [11,12].
Mitochondria are internally organized into cristae, invaginations of the IMM, which can be dynamically reorganized according to various stimuli, such as changes in energy requirements or apoptotic signals, becoming more or less compact [13]. The formation and maintenance of these structures is a complex process that requires the participation of many proteins, including the mitochondrial contact site and cristae organizing system (MICOS) and the fusion protein Optic atrophy type 1 (OPA1). The MICOS and OPA1 have been widely associated with the regulation of the cristae architecture, specifically at the cristae junction [14].
The MICOS complex was originally characterized and largely studied in yeast, where it is composed of six subunits: Mic10/MINOS1, Mic12/Aim5, Mic19/CHCHD3, Mic26/ApoO, Mic27/ApoOL and Mic60/Mitofilin [15]. In mammals, this system is more complicated and two additional subunits have been described: Mic25/CHCHD6 [16] and Mic13/QIL1 [17]. All of these subunits are transmembrane proteins, except for Mic19 and Mic25.
Mic60 and Mic10 are the core subunits of the two subcomplexes forming the MICOS architecture (Figure 3). From one side, Mic60 contacts directly with Mic19 and Mic25, while the remaining subunits assemble with Mic10 [18]. Mic10 interacts with Mic26 and Mic27 and is stabilized by QIL1 [17]. Mic10 has been shown to be able to induce membrane curvature even in the absence of the other MICOS subunits [19]. Both subcomplexes are necessary for the formation and stabilization of the cristae at the CJ.
The Mic60/Mic19 module is also associated with the OMM. Indeed, Mic19 has been shown to form a bridge between Mic60 and Sorting Assembly Machinery 50 kDa subunit (SAM50), the outer membrane protein that regulates the import and assembly of β-barrel proteins [20], creating tight OMM and IMM contact sites. Other proteins have been specifically associated with this OMM–IMM junction such as Metaxin-1 and 2 (Mtx1/Mtx2) and DnaJ homolog subfamily C member 11 (DNAJC11) [21]. These OMM–IMM interactions create a specific environment of closed apposition of the two membranes favorable for protein import and lipid and phospholipid transport [22].
Immunoprecipitation experiments indicate that both Mic19 and Mic60 interact with OPA1. OPA1 is an IMM protein with a double role in mitochondrial dynamics and architecture: it is involved in mitochondrial fusion of the IMM and it participates in cristae remodeling [23,24]. Indeed, OPA1 has been shown to be upstream and epistatic to MIC60 and to be the sole regulator of cristae width and junction diameter and number [25].
Finally, the formation of the cristae rims has been proposed to be also linked to the dimerization of the ATP synthase. Indeed, Complex V (CV) dimers reside at the apex of the cristae and, when dimerization is abolished, the IMM loses the typical tubular organization of the cristae and forms onion-like structures instead, a typical feature associated with MICOS deregulation [26].
The mechanisms of protein and ion transport through the OMM and the IMM differ in many aspects [27]. As described previously, the IMM is much less permeable than the OMM and is characterized by α-helical transport proteins such as protein translocases and other carriers for metabolites and ions. The OMM, instead, contains channel-forming proteins, such as β-barrel transmembrane hydrophilic pores, that allow the passage of precursor proteins, small hydrophilic metabolites and ions.
Human mtDNA only contains 37 genes, while it has been estimated that the mitochondrial proteome is composed of approximately 1200 proteins [28]. Consequently, most of the proteins required are encoded by nuclear genes, translated in the cytosol and transported to or into the organelle. Indeed, newly synthesized proteins carrying a mitochondrial target signal are transported to the OMM by chaperones and, according to the nature of their signal, they can be inserted in the OMM, imported in the IMS or transported to the IMM translocases [29]. At the OMM, proteins transport is ensured by the two main complexes: the Translocase of the Outer Membrane (TOM) complex and Sorting and Assembly Machinery (SAM) complex [30,31]. For insertion of proteins in the OMM, integral proteins containing one or more spanning helices are not imported, but rather inserted in the OMM via initial interaction with TOM70 (receptor subunit of TOM) and then, in yeast, with Mim1, while β-barrel proteins pass through TOM40 (pore subunit of TOM) and are then transported by chaperones to SAM. The main component of SAM is SAM50, a β-barrel protein, which interacts with the new protein, accommodates the folding and inserts it laterally into the membrane [32].
Proteins localized in the IMS can have different fates: they can undergo modifications to stabilize the protein and prevent retrograde transport, such as insertion of a heme group or oxidation of cysteine residues in order to form a disulfide bridge via the Mia40 pathway [33]. In addition, modified proteins with a specific target peptide can be directed to the IMM or the matrix via TIM23 [34].
Proteins that have to be inserted in the IMM are delivered by the IMS chaperones TIM9-10 to the IMM insertase/translocase complex TIM22, the main pathway for import of polytopic inner membrane proteins [35]. The import through TIM22 requires the mitochondrial membrane potential, responsible for an electrophoretic effect on the positively charged targeting sequences of these proteins, but it is not ATP-driven [36]. Finally, Oxa1, a conserved membrane protein, mediates the insertion of both nuclear and mitochondria-encoded precursors into the inner mitochondrial membrane [37].
For matrix-targeted proteins, the translocase involved in this process is TIM23 and three different forces drive the transport: the membrane potential, the increasing affinities of the precursor proteins to the components on the trans side of the translocase, compared with the affinities to the cis side, and the motor force generated upon ATP hydrolysis by the chaperone mtHsp70 and its associated subunits [38].
Mitochondria are dynamic organelles that form a complex network of tube-like structures [39]. They undergo opposing fusion and fission events to generate a specific mitochondrial morphology network according to the cellular energy needs, the metabolic state of the cell or to adapt to cellular cues. Mitochondrial fusion allows the organelles to share metabolites, proteins and mtDNA, and a hyperfused mitochondrial morphology is associated with a mechanism of defense to enhance cell survival (and impede mitochondrial clearance). In contrast, while mitochondrial fragmentation is often associated with mitochondrial dysfunction and cell death, this process is also required for mitochondrial motility or segregation of damaged portions of the reticulum for degradation through a process known as mitophagy [40].
Mitochondrial dynamics are controlled by Guanosine Triphosphatase (GTPases) proteins belonging to the Dynamin family of proteins, where GTP hydrolysis leads to structural change, subsequently driving membranes remodeling [41]. During mitochondrial division, the constriction of the tubule and membrane scission of one mitochondrion into two separate organelles is carried out by recruitment of the cytosolic GTPase Dynamin-related/-like protein 1 (Drp1) to mitochondria–endoplasmic reticulum (ER) contact sites, via interaction with fission protein 1 (Fis1) and mitochondrial fission factor (Mff) [39]. At these sites, Drp1 oligomerizes into a ring-like structure and upon GTP hydrolysis drives mitochondrial division. Mitochondrial fusion, instead, is a two-step mechanism with the OMM localized GTPases mitofusin 1 and 2 (Mfn1 and Mfn2) ensuring OMM fusion, and the IMM GTPase OPA1 responsible for IMM fusion. Again, the shape and dynamics of mitochondria are tightly linked to their bioenergetic status. Failing mitochondria usually appear fragmented, due to a prevalence of the fission machinery, whereas mitochondria hyperfuse as a mechanism of defense against autophagy in conditions of stress, including bioenergetic stress. In addition, specific disorders of some main actors of mitochondrial fission (e.g., mutations in the dynamin-related protein, DRP, the key factor of mitochondrial fission [42,43]) or fusion (e.g., mitofusin 2, Mfn2 [44], and optic atrophy protein 1, OPA1 [45,46,47]) can be associated with OXPHOS failure and, in particular for some OPA1 mutations, with the accumulation of multiple mtDNA deleted species and multiple defects of the respiratory chain complex activities [48]. In addition to its profusion role, OPA1 acts also as a sealer of the cristae junctions and regulates the ordered distribution of the ETC complexes along the mitochondrial cristae [49,50].
Therefore, alterations in mitodynamics are, nowadays, considered part of the possible causes of OXPHOS deficiency and bioenergetic failure leading to mitochondrial disease.
Mitochondria need to move within the cell and their localization is crucial for different functions such as cell division [51]. Directed mitochondrial transport happens on microtubule filaments, typically through force-generating motor proteins, classified into three families: myosins, kinesins and dyneins [52]. Milton, syntabulin and the GTPase Miro have been identified as mitochondria-specific molecules involved in microtubule-based transport [51].
Mitochondrial trafficking is fundamental for a constant supply of healthy mitochondria generating ATP at the right time and place, especially in neuronal cells, where organelles have to move over long distances along the axon, from the cell body to the presynaptic terminal [53].
Mitochondria are commonly known as the “powerhouse of the cell”, due to their role in energy production. However, in the last 30 years, mitochondria have been characterized also as a signaling organelle involved in numerous physiological functions, including calcium homeostasis, apoptosis and heme and iron-sulfur clusters synthesis.
The main source of energy in cells derives from the de-phosphorylation of an ATP molecule to an adenosine diphosphate (ADP) molecule. In order to make this process sustainable, the cell needs to use nutrients to re-generate the ATP molecules used.
During a series of chemical reactions, indeed, a glucose molecule is gradually broken down into carbon dioxide, and its hydrogen atoms are stripped and used to combine with oxygen to form water. The first stage of this mechanism takes place in the cytosol, is anaerobic and is mitochondria-independent [54]. This pathway is known as glycolysis, and it produces only two molecules of ATP from one molecule of glucose metabolized, generating two molecules of pyruvate.
To optimize ATP production, glycolysis is coupled to a second pathway, known as the citric acid cycle (or tricarboxylic acid cycle or Krebs cycle, Figure 4), which takes place in mitochondria and is aerobic. This cycle is composed of nine different enzymatic reactions and each round generates three nicotinamide adenine dinucleotide (NADH) molecules, one flavine adenine dinucleotide (FADH2) and one guanosine triphosphate (GTP). One molecule of glucose is catabolized by glycolysis into two molecules of the ketoacid pyruvate. Pyruvate enters through the mitochondrial pyruvate carrier (MPC) into the mitochondrial matrix [55], where it is decarboxylated, oxidized and coupled with coenzyme A by pyruvate dehydrogenase, PDH, to form acetyl-CoA. Thus, the two glucose-derived acetyl-CoA molecules in the mitochondrial matrix lead to a complex catabolic cycle, the Krebs cycle, which gives rise to two molecules of CO2, whereas their electron equivalents (hydrogen atoms) are stripped off and used to produce, in total, six NADH, two FADH2 and two GTP.
NADH and FADH2 molecules can then be used in a process known as oxidative phosphorylation, where the majority of the energy is finally converted to ATP. This process consists of the passage of electrons from NADH and FADH2 to the final acceptor, oxygen, through the electron transport chain (ETC). This process, which involves molecular oxygen as a sink for binding electrons, is called cellular respiration. Respiration is coupled with ATP production by ATP synthase or complex V of the ETC. In total, the complete oxidation of a single molecule of glucose is used by the cell to produce 30 ATP molecules [56].
Importantly, NADH and FADH2 can also be produced by a process known as fatty acid β-oxidation. A fatty acid is converted to fatty acyl-CoA and then to acyl carnitine, in order for it to enter the mitochondria and eventually be reconverted in intramitochondrial acyl-CoA. In the organelle, a long-chain acyl-CoA is broken down to acetyl-CoA molecules, producing one NADH and one FADH2 for each couple of carbons hydrolyzed from the acyl chain [57].