Mitochondrial Physiology and Vegetal Molecules: Therapeutic Potential of Natural Compounds on Mitochondrial Health

Mitochondrial Physiology and Vegetal Molecules: Therapeutic potential of natural compounds on mitochondrial health provides a comprehensive overview of mitochondrial physiology throughout the human life span, as well as the effect of molecules of vegetal origin on mitochondrial health. The editor has lined up a team of worldwide experts to cover the most exciting and high-impact advancements of research in this area.

This book is structured into two parts that provide a balance of both foundational and applied content. Part I provides an overview of mitochondrial physiology including its structure, dynamics, biogenesis, membranes, DNA transcription, and translation in the mitochondria. Part I also covers other themes such as apoptosis. Part II then covers the effect of specific vegetable-derived molecules on mitochondrial health, including anthocyanins, caffeine, cannabinoids, carnosic and rosmarinic acids, citrus flavonoids, polyphenols, pterostilbene, resveratrol, and sulforaphane, among others.

Mitochondrial Physiology and Vegetal Molecules: Therapeutic potential of natural compounds on mitochondrial health is a complete resource for researchers in this exciting field. Its comprehensive coverage makes it particularly interesting to bioscience researchers willing to understand the foundations of mitochondrial physiology throughout the human life span. Clinician researchers, MDs, nutritionists, pharmacologists, and sports scientists may be attracted to the detailed information on the health effects of vegetal origin molecules on the organelle.

• Contains detailed information on plant products and their effect on mitochondria
• Proposes therapies and reviews mechanisms of absorption at the cellular level
• Discusses the limited bioavailability of plant molecules/compounds in the human organism
• Includes coverage of specific conditions such as Sports and affective disorders, among others
• Presents the protective effects of plant products in mitochondrial health through all stages of life

• Language: English
• Publisher: Academic Press
• Release date: Jul 14, 2021
• ISBN: 9780128215838

Book preview

Part I

Overview of mitochondrial physiology

Chapter 1: Mitochondria: Ultrastructure, dynamics, biogenesis, and main functions

M. Rigoulet; C.L. Bouchez; T. Molinié; S. Cuvellier; J.P. Mazat; S. Ransac; S. Duvezin-Caubet; P. Paumard; Anne Devin    Univ. Bordeaux, CNRS, IBGC, UMR 5095, Bordeaux, France

Abstract

Mitochondria are double-membrane-enclosed organelles that ensure a number of central functions in the vast majority of eukaryotic cells. These include, among others, key steps in lipid and amino acid metabolism, biosynthesis of iron–sulfur clusters, heme, and further prosthetic groups, regulation of programmed cell death, and most prominently the final transformation of proteins, fats, and sugars into the cellular energy currency adenosine-5′-triphosphate (ATP) by oxidative phosphorylation (OXPHOS). Severe mitochondrial dysfunctions are associated with so-called mitochondrial diseases, and are largely associated with specific or broad OXPHOS defects. In addition, alterations of mitochondrial properties have been associated with other diseases associated with metabolic and/or physiologic alterations. This chapter aims to give a general introduction on mitochondrial properties and functions that will be further developed in subsequent chapters.

Keywords

Mitochondria; Oxidative phosphorylation; Mitochondrial dynamics; Mitochondria ultrastructure; Respiratory chain

Acknowledgments

The authors received continuous support from the CNRS (Centre National de la Recherche Scientifique), the Comité de Dordogne & Gironde de la Ligue Nationale Contre le Cancer, The Fondation ARC pour la recherche sur le Cancer, the Plan cancer 2014-2019 No BIO 2014 06, the ANR, and Bordeaux University. The authors thank Claudine David, Manuel Rojo, and Arnaud Mourier for their help in refining this chapter.

1: Introduction

Mitochondria are double-membrane-enclosed organelles that ensure a number of central functions in the vast majority of eukaryotic cells. These include, among others, key steps in lipid and amino acid metabolism, biosynthesis of iron–sulfur clusters, heme, and further prosthetic groups, regulation of programmed cell death, and most prominently the final transformation of proteins, fats, and sugars into the cellular energy currency adenosine-5′-triphosphate (ATP) by oxidative phosphorylation (OXPHOS). Severe mitochondrial dysfunctions are associated with so-called mitochondrial diseases, and are largely associated with specific or broad OXPHOS defects. In addition, alterations of mitochondrial properties have been associated with other diseases associated with metabolic and/or physiologic alterations. This chapter aims to give a general introduction on mitochondrial properties and functions that will be further developed in subsequent chapters.

2: Mitochondrial ultrastructure

Mitochondria display a unique and highly characteristic ultrastructure that allow their unambiguous identification in electron microscopy sections. Mitochondria are enveloped by an outer membrane and an inner membrane enclosing a narrow intermembrane space. The outer membrane represents the border to the cytosol and the inner membrane separates the intermembrane space from the matrix. The inner membrane is not limited to the mitochondrial boundary, but forms protrusions into the matrix—known as cristae—that provide a tremendous increase in surface (Frey and Mannella, 2000). The direct contact between outer and inner membranes as well as the transition between inner boundary and inner cristae membranes occurs at specific structures known as contact sites and cristae junctions, respectively (Fig. 1A). These structures are made and regulated by specific protein complexes known as MICOS (mitochondrial contact site and cristae organizing system) that are essential for several mitochondrial functions, including the maintenance of mitochondrial ultrastructure as well as the transport and exchange of proteins and lipids (reviewed in Van Der Laan et al., 2016).

Fig. 1Fig. 1

Fig. 1 Mitochondrial ultrastructure. (A) Mitochondria are enveloped by an outer and an inner membrane enclosing a narrow intermembrane space. The inner membrane encloses the mitochondrial matrix and forms invaginations called cristae. The direct contact between outer and inner membranes as well as the transition between inner boundary and inner cristae membranes occurs at specific structures known as contact sites and cristae junctions, respectively. Outer mitochondrial membranes and the endoplasmic reticulum are in close or direct contact at sites termed MAMs (mitochondria-associated membranes). (B) Mitochondria form filaments containing numerous mtDNA nucleoids in wild-type (WT) cells and display fragmented mitochondria in fusion-defective cells lacking fusion factor Mfn1 (MFN1 KO). Morphology and mitochondrial DNA distribution in cultured cells was visualized with antibodies against outer membrane Tom20 (red) and DNA (green), respectively. The inset represents an enlargement of the indicated area. Bar = 20 μm.

These ultrastructural characteristics are common to the mitochondria of all eukaryotes. However, the topology, size, and relative abundance of the different membranes and subcompartments vary significantly between species and tissues and as a function of the metabolic or physiological state (Mannella, 2008). This is in accordance with the fact that each mitochondrial membrane and compartment hosts specific functions and activities.

Mitochondrial ultrastructure was suspected to be regulated by a number of factors since 1962. Yotsuyanagi first described these modifications in aerobically growing yeast (Yotsuyanagi, 1962). He showed that when glucose was used as a carbon source, mitochondrial ultrastructure switched from poorly defined cristae during the initial logarithmic growth phase to a classical and easily recognizable structure during the final aerobic growth and then to an unconventional shape and ultrastructures when entering the stationary phase. These results suggested a close relationship between energy metabolism and mitochondrial morphology that was later investigated on mitochondria isolated from rat livers (Hackenbrock, 1966). In this model, mitochondrial reorganization (orthodox vs condensed cristae) was reported at the minute time scale as various substrates of the OXPHOS system were added. One of the mechanisms proposed to explain these rapid mitochondrial modifications was water efflux. It was demonstrated that the mitochondrial metabolic water produced during oxidative phosphorylation is responsible for these intramitochondrial volume fluctuations (Casteilla et al., 2011). The F1F0 ATP synthase was also suspected to influence the shape of the cristae (Allen et al., 1989) and it was indeed demonstrated that the dimerization/oligomerization of this enzyme was involved in this process (Paumard et al., 2002; Habersetzer et al., 2013) as the enzyme determines the shape of cristae edges (Strauss et al., 2008; Davies et al., 2012). Recent studies using a combination of electron microscopy, super resolution imaging microscopy (dSTORM), and classical biochemistry methods tend to demonstrate that the curvature of cristae edges is regulated in hepatocellular carcinoma HepG2 cells. Indeed, upon moderate hypoxia (5% O2), the decrease in the amount of some constituents of the MICOS complex (MIC 10 and MIC60) coupled with a reduction in the relative amount of ATP synthase dimers/oligomers was correlated with cristae remodeling exhibiting wider edges (Plecita-Hlavata et al., 2016). Conversely, sharper cristae edges where reported in cells cultivated under normoxia or hypoxia after the addition of a membrane-permeable Krebs cycle substrate precursor dimethyl-2-oxoglutarate. The data suggest the essential role of the elongation and strengthening of ATP synthase oligomers for the bending of the membrane at the cristae edges accompanied by MIC60 relocation to accommodate overall cristae reshaping (Dlasková et al., 2019). Altogether, these ultrastructural modifications may be involved in the optimization of the OXPHOS system as proposed by the modelization of the effect of the membrane curvature on the electric field generated by the respiratory chain proton pump activity (Strauss et al., 2008).

Electron microscopy analysis often depicts close proximity or even direct contacts between outer mitochondrial membranes and the endoplasmic reticulum (ER) (Fig. 1A) that are termed MAMs (mitochondrial associated membranes) or ERMES (ER-mitochondria encounter structure). The molecular composition of yeast ERMES complex is well established (Kornmann and Walter, 2010), but that of mammalian MAMs remains highly debated (Herrera-Cruz and Simmen, 2017). These contacts mediate the interorganellar exchange of calcium and lipids, are involved in mitochondrial fission (see below), and play a role in insulin signaling, autophagy, and apoptosis. Consequently, alterations of MAM properties and of MAM-dependent functions have been associated with human diseases (Vance, 2014; Rieusset, 2017).

3: Mitochondrial morphology, distribution, and dynamics

Mitochondria are mobile structures that do not distribute homogeneously throughout cells, but localize to intracellular sites (e.g., apical membranes of secretary cells, synapses of neurons) where their function (e.g., ATP synthesis by OXPHOS) is required. Mitochondrial distribution and mobility are ensured by adaptors and motors that mediate interactions with the microtubule and actin cytoskeletons (Frederick and Shaw, 2007; López-Doménech et al., 2018).

Electron microscopy of ultrathin sections has often depicted mitochondria as isolated punctate structures that resemble bacteria (their phylogenetic ancestors), in form and in size. This appearance and the resemblance to that of their prokaryotic ancestors led to the long-standing concept that the mitochondrial compartment is composed of a collection of numerous small and independent organelles. However, work in the last decades has revealed that mitochondria continuously fuse and divide and that the equilibrium between antagonizing fusion and fission determines their morphology: mitochondria appear as a network of interconnected filaments at high fusion/fission ratios and as a collection of small punctate structures when the fusion/fission ratio shifts toward fission (Fig. 1B). Fusion enables molecular exchanges between mitochondria (proteins, RNA, DNA) and leads to the existence of a single mitochondrial compartment within cells. Division enables severing of mitochondrial filaments and segregation of functional and defective mitochondria, a requisite for the selective degradation of individual mitochondria by autophagy or mitophagy.

Fusion and fission are energy-consuming processes that depend on the hydrolysis of nucleotides by specific GTPases of the dynamin superfamily (see below). In addition, the fusion of inner membranes (but not of outer membranes) also depends on the inner membrane electrical potential difference (Malka et al., 2005). It is now well established that mitochondrial dynamics and bioenergetics are linked: the fusion/fission steady state is modulated by the bioenergetic status of cells and mitochondria and reciprocally, fusion/fission dynamics can modulate mitochondrial bioenergetics (Sauvanet et al., 2010; Wai and Langer, 2016). Accordingly, conditions leading to mitochondrial dysfunction and/or depolarization, like the permeabilization of the outer membrane at the onset of apoptosis, are accompanied by mitochondrial fragmentation (Arnoult, 2006).

Several components of the mitochondrial fusion and fission machineries have been identified and many of them, notably some GTP-binding proteins of the dynamin superfamily, are conserved between animals and fungi (Westermann, 2010). In mammals, outer membrane fusion is governed by mitofusins MFN1 and MFN2, two GTPases anchored to the outer mitochondrial membrane. Further reports reveal that mitofusin MFN2 can modulate contacts between ER and mitochondria, but the precise role of MFN2 in MAM biogenesis/modulation remains controversial (Martins De Brito and Scorrano, 2008; Filadi et al., 2015). Inner membrane fusion depends on OPA1, a GTPase that localizes to the intermembrane space. The size, membrane association, and activity of OPA1 are regulated by mitochondrial metalloproteases (see below) that generate isoforms of different lengths. Of note, proteolytic processing of OPA1 is modulated by bioenergetics (Guillery et al., 2008) and altered by mitochondrial dysfunctions (Duvezin-Caubet et al., 2006). Mitochondrial fission is modulated by a machinery that relies on two cytosolic GTPases (DRP1/DNM1L and dynamin 2) and their complex interactions with outer membrane receptors, ER, MAMs, and the actin cytoskeleton (Sesaki et al., 2014; Manor et al., 2015; Lee et al., 2016). To date, no specific machinery for inner membrane fission has been identified.

Mitochondrial fusion and fission are essential processes that are required for embryonic development of mice (Bertholet et al., 2016). Further work has revealed that alterations of mitochondrial fusion/fission dynamics provoke defects in cellular bioenergetics and/or physiology (Mourier et al., 2015) and that mutations of genes encoding fusion/fission proteins provoke hereditary neuropathies with highly variable symptoms (Bertholet et al., 2016). While mutations of a fusion factor of the outer membrane (MFN2) are associated with a peripheral sensory and motor neuropathy (Charcot–Marie–Tooth disease of type 2A, CMT2A), that of an inner membrane fusion factor (OPA1) or of a fission factor (DNM1L) have been linked to optic nerve atrophy (dominant optic atrophy) (Gerber et al., 2017).

4: Biogenesis, maintenance, and quality control

Present-day mitochondria are the descendants of prokaryotic endosymbionts that were internalized by the ancestor of today’s eukaryotic cells; they carry a relatively small mitochondrial genome that represents a relic of this evolutionary past (Gray et al., 1999). The mitochondrial genome encodes a few proteins (13 in humans) that are essential components of the OXPHOS machinery, as well as ribosomal and transfer RNAs (rRNAs, tRNAs) essential for their intramitochondrial translation. Of note, all human mtDNA-encoded genes give rise to hydrophobic, transmembrane subunits of the respiratory complexes and the ATP synthase. The respiratory complexes and the ATP synthase that catalyze the OXPHOS reactions in the inner mitochondrial membrane therefore have a double genetic origin: while the majority of their components are encoded by the nuclear genome, few essential subunits are encoded by a relatively small mitochondrial genome (16.5 kb in humans) that is present in thousands of copies per cell and localizes to hundreds of nucleoprotein complexes termed nucleoids (Taylor and Turnbull, 2005; Malka et al., 2006). The mitochondrial proteome comprises more than 1000 different proteins (Calvo et al., 2016). The vast majority of these proteins (>  99%) is thus encoded by the nuclear genome, synthesized on cytosolic ribosomes and imported into the organelle. The proper biogenesis of mitochondria requires a number of processes ensuring for instance the maintenance of the mitochondrial genome, the coordination of the two genomes’ expression, the import of nuclear-encoded components, as well as the maturation, folding, membrane insertion, and the final assembly in functional molecular complexes. The biogenesis of mitochondria is therefore a real challenge for the cell with the coordination of two genomes, the membrane compartmentalization, the elaboration of multisubunit complexes containing cofactors/prosthetic groups (see Tang et al., 2020 for a recent review). Furthermore, mitochondria are the main source of reactive oxygen species (ROS) and their components are thus the first targets of ROS-induced damage. An array of proteins with chaperone and protease functions are therefore required to ensure the protein quality control (PQC) and proteostasis of mitochondria (Voos, 2013; Voos et al., 2016). For instance, no less than 45 proteases are identified in mitochondria; 23 are resident and 22 are transient proteases translocated to mitochondria only in certain conditions that perform additional proteolytic activities essentially related to apoptosis or autophagy (Quirós et al., 2015; Levytskyy et al., 2017; Deshwal et al., 2020).

4.1: Mitochondrial protein import and sorting

The vast majority of mitochondrial proteins are translated on cytosolic ribosomes and subsequently targeted to and imported into mitochondria. Targeting to the matrix, the inner membrane and the intermembrane space mainly rely on the presence of cleavable targeting signals: 60% of all mitochondrial proteins have an N-terminal mitochondrial targeting signal (MTS) on their precursor that is cleaved upon/after mitochondrial import (see below). Other proteins of the inner membrane and intermembrane space have internal targeting signal that, in some cases, have not been identified or characterized yet. The targeting of outer membrane protein follows somewhat different paths. Highly hydrophobic channels/beta-barrels are imported/inserted by a dedicated machinery (see below), but the machineries allowing for the targeting of proteins with a single membrane anchor remain elusive (Costello et al., 2017). It is important to note that targeting, folding, processing, and assembly appear exquisitely interlinked and thus, that translocases, proteases, and chaperones fulfill complementary, linked, and partially overlapping functions that are difficult to dissociate/describe separately. In addition, several of these protein import machineries interact with each other and with other complexes such as MICOS or ERMES (Pfanner et al., 2019). Therefore, all these machineries must be considered as a dynamic network rather than individual complexes and interorganellar interactions have to be considered. Of note, most of the genes encoding these proteins are essential and/or linked to a variety of diseases, showing that these functions are not dispensable. Below, we present a succinct view of the main machineries mediating import and sorting of mitochondrial proteins (see Wiedemann and Pfanner, 2017; Ghifari et al., 2018; Kang et al., 2018; Kolli et al., 2018 and references therein for more extensive reviews) (Fig. 2).

Fig. 2

Fig. 2 Overview of the main machineries for protein import and quality control in mitochondria. Most mitochondrial proteins are nuclear-encoded and imported into mitochondria. There are several types of targeting signals that determine the routes and final destination of precursor proteins. The main protein import machineries are as follows: the translocase of outer membrane (TOM), the translocase of inner membrane 23 (TIM23) coupled to presequence-associated motor (PAM), the sorting and assembly machinery (SAM) complex, mitochondrial import and assembly machinery (MIA) composed by Mia40 and the sulfhydryl oxidase augmenter of liver regeneration (ALR), small TIMs, and the carrier translocase TIM22. A subset of precursor proteins is subjected to proteolytic maturation, for example by the mitochondrial processing peptidase (MPP) for matrix proteins. In order to acquire a proper folding to become functional, mitochondrial proteins can be assisted by molecular chaperones, such as HSP60/HSP10 and mtHSP70. Finally, damaged, unfolded, or unassembled proteins are degraded by quality control proteases such as the Lon protease homolog (Lon) and the ATP-dependent Clp protease (ClpXP) in the matrix, and the intermembrane-oriented ( i -AAA) or the matrix-oriented ( m -AAA) proteases of the AAA family in the inner membrane.

4.1.1: TOM (translocase of outer mitochondrial membrane) complex

The translocase of outer membrane represents the main entry gate for proteins bearing MTSs. Precursors of mitochondrial proteins are chaperoned by combinations of cytosolic factors that are determined by the sequence/nature of their targeting signal, hydrophobicity, and many other factors. These precursors are recognized by the receptor proteins of the TOM complex. When exiting the TOM complex, precursor proteins follow different routes depending on their targeting signals and final destination within mitochondria.

4.1.2: TIM23 presequence translocase

TIM23 is the so-called presequence pathway. It resides in the inner membrane (TIM: translocase of inner membrane) and recognizes precursor proteins with cleavable amphipathic N-terminal targeting sequences emerging from the TOM complex. The inner membrane electrical potential difference is required to insert the positively charged N-terminal presequence precursors into the TIM23 pore, but it is not sufficient for a complete translocation across the IM. The mitochondrial HSP70 chaperone protein (mtHSP70), together with its regulating factors, forms the presequence-associated motor (PAM) that interacts with the TIM23 core components and hydrolyzes ATP as a secondary driving force to entirely pull the proteins inside the matrix. In most cases the presequence is processed in the matrix by the mitochondrial processing peptidase (MPP) and/or additional matrix proteases. Alternatively, TIM23 ensures the lateral insertion of precursors into the inner membrane. In this process named stop-transfer, translocation is interrupted when a hydrophobic stretch reaches the membrane pore and the protein is then released in the inner membrane. Some of these precursors are further processed by the inner membrane protease (IMP) to release the mature protein in the intermembrane space. Consequently, TIM23 allows proteins to be directed either to the matrix, the inner membrane, or the intermembrane space.

4.1.3: The carrier pathway into the inner membrane

The TIM22 complex (translocase of the inner mitochondrial membrane 22) is required for the import of a subset of multispanning hydrophobic proteins of the inner membrane that do not exhibit any cleavable presequence but rather several internal targeting elements. It is also called the carrier translocase as it mediates the import of a large family of carriers that contain six transmembrane domains and catalyzes the exchange of metabolites across the inner membrane (Taylor, 2017). TIM22 substrates are first taken up at the TOM exit by soluble chaperones of the small TIM chaperone family in the intermembrane space. The precursor proteins are then delivered to the TIM22 complex from the intermembrane space side and inserted into the inner membrane.

4.1.4: The intermembrane space, mitochondrial import, and assembly machinery (MIA)

The intermembrane space contains a number of proteins with characteristic cysteine motifs. The MIA complex couples oxidative protein folding to the biogenesis of intermembrane space proteins. It mediates the import and oxidative folding of substrates with conserved cysteine motifs (CX3C, CX9C). Key components are the oxidoreductase Mia40 (CHCHD4) and the sulfhydryl oxidase ALR (augmenter of liver regeneration). Mia40 engages in an intermolecular disulfide bond with the incoming substrate. Following oxidation of the substrate, MIA40 is reoxidized by ALR. In yeast it has been shown that reduced FAD (FADH2) in Erv1 (yeast ALR) is reoxidized either by transferring the electrons directly to molecular oxygen or to cytochrome c (Allen et al., 2005; Bihlmaier et al., 2007; Dabir et al., 2007; Banci et al., 2012).

4.1.5: Import of mitochondrial outer membrane proteins

Integral outer mitochondrial membrane proteins are of two types: (1) Beta-barrel proteins that are anchored in the outer membrane by multiple transmembrane Beta strands and that are evolutionarily conserved from gram-negative bacteria (mitochondria’s ancestor); and (2) proteins that are anchored in the outer membrane by one or more alpha-helical segments. A large variety of import pathways has been reported over the recent years for mitochondrial outer membrane proteins and the best described system is the sorting and assembly machinery (SAM) complex for Beta-barrel proteins. There are several Beta-barrel proteins in the outer mitochondrial membrane such as VDAC (voltage-dependent anion channel), TOM40 or SAM50 (channel-forming proteins). The members of this family contain an array of multiple Beta strands arranged as transmembrane segments in the outer membrane. The MTS in these proteins consists of the last Beta strands and the loop in between arranged as a hairpin. This signal is recognized by the TOM complex which translocates the precursors. The interaction between the TOM complex and the SAM complex promotes the release of Beta-barrel proteins into the membrane.

4.2: Mitochondrial protein homeostasis: biogenesis, regulation, and quality control

The following paragraphs will give the reader a glimpse of the many chaperones and proteases and of the diversity of their roles in mitochondria. We refer to extensive reviews for a more exhaustive view.

4.2.1: Protein trafficking, processing, and activation: Processing peptidases

Most mitochondrial proteins are imported from the cytosol thanks to targeting signals. The majority of these proteins require further processing of this targeting signal once they reach their final location in order to be active. Processing peptidases in the different mitochondrial compartments perform this maturation. In the matrix, MPP is in charge of removing the N-terminal targeting peptides from precursor proteins. Likewise, IMMPL1/2 (mitochondrial inner membrane protease) removes hydrophobic sorting signals from proteins sorted to the intermembrane space. Several mitochondrial proteins undergo additional proteolytic cleavages in order to regulate their activities (ribosome assembly and consequently synthesis of mitochondrially encoded subunits of the respiratory chain (AFG3L2)), regulation of import (YME1L1), control of mitochondrial membrane composition (YME1L1), regulation of apoptotic resistance (YME1L1), mtDNA maintenance and gene expression (LONP1, HTRA2), mitochondrial dynamics (YME1L1 and OMA1) (Quirós et al., 2015; Deshwal et al., 2020).

4.2.2: Chaperones involved in the correct folding of proteins within mitochondria

During import precursor proteins must be maintained soluble and later must be folded and assembled (reviewed in Voos, 2013; Voos et al., 2016). As stated previously, MIA40 and small TIM in the intermembrane space and mtHSP70 (HSPA9) together with HSP60 (HSPD1) and HSP10 (HSPE1) in the mitochondrial matrix fulfill chaperone functions. MIA40 and small TIM function in an ATP-independent manner whereas in the mitochondrial matrix, chaperones are ATP-dependent. mtHSP70 has an essential role in protein import by acting in the import motor to translocate proteins in the matrix. It also has a concomitant role in maintaining the preproteins unfolded during import to stabilize them and to avoid irregular associations. mtHSP70 has a further role in the correct folding of polypeptides in the matrix together with HSP60, not only for incoming polypeptides but also for the biosynthesis and assembly of mtDNA-encoded subunits of the OXPHOS system. HSP60 is a member of the chaperonins/GroEL family also located in the matrix. It forms large complexes that interact with HSP10. ATP hydrolysis-driven conformational changes allow the folding of substrate proteins. Both mtHSP70 and HSP60 are essential proteins highlighting their unique and important roles in cellular metabolism.

4.2.3: Additional factors required for the biosynthesis of OXPHOS complexes

In addition to classical chaperones with general functions in mitochondria, numerous substrate-specific factors are required to obtain functional complexes of the respiratory chain or functional ATP synthase (Tang et al., 2020). Indeed, the biogenesis of large hetero-oligomeric complexes such as OXPHOS complexes is a real challenge for the cells due to the large number of subunits added to their dual genetic origin and the presence of metal cofactors. Therefore, these processes are usually assisted by dedicated chaperones and multiple quality control checkpoints. A striking example is the biogenesis of the cytochrome c oxidase, the copper-heme A terminal oxidase of the respiratory chain (complex IV). This transmembrane complex consists of 14 subunits of which three are mtDNA-encoded. To date, 30 COX assembly factors have been described, the roles of which range from mitochondrial mRNA stability and translation activation to metalation and assembly of hemes and of the different subunits in functional complexes and supercomplexes (Timón-Gómez et al., 2018; Grevel et al., 2019).

4.2.4: Protein quality control

PQC stands for the mechanisms and factors required for the maintenance of protein homeostasis in mitochondria in normal or under stress conditions. These factors include chaperones to take care of unfolded or misfolded proteins but also proteases to degrade proteins that are terminally damaged (oxidized, aggregated) and/or do not reach their final destination (not assembled properly) (Voos et al., 2016; Jadiya and Tomar, 2020). mtHSP70 and HSP60 indeed have important roles in prevention of aggregation, refolding of misfolded proteins to an active conformation, or stabilization in a soluble state awaiting for degradation by proteases. They are strongly induced in stress conditions where damaged/unfolded proteins accumulate (elevated temperatures, exposure to toxic compounds, elevated ROS, mitochondrial unfolded protein response) and were accordingly initially recognized as heat shock proteins (HSPs) (Voos, 2013) (Table 1).

Table 1

An array of mitochondrial proteins (molecular chaperones and proteases) are involved in the maintenance of protein homeostasis in mitochondria. The main components have been listed in this table together with their known implications in disease and in metabolism related to diabetes. (IMS: intermembrane space; AAA+: ATPases associated with diverse cellular activities).

Proteases involved in PQC are mostly chambered proteases of the AAA family (ATPases associated with diverse cellular activities) that hold a chaperone-like activity allowing for the necessary unfolding of substrate for their ATP-driven translocation inside the protease cavity and proteolytic cleavage: the Lon protease homolog (LONP1) and the ATP-dependent ClpP protease in the matrix, intermembrane-oriented i-AAA (YME1L1) and matrix-oriented m-AAA both anchored in the inner mitochondrial membrane (Voos, 2013; Quirós et al., 2015; Levytskyy et al., 2017; Deshwal et al., 2020). They degrade membrane spanning and membrane-associated subunits of the electron transport chain that are damaged, oxidized, or nonassembled, and have regulatory functions (mitochondrial protein synthesis, dynamics, calcium homeostasis). LONP1 substrates include, for instance, aconitase, succinate dehydrogenase subunit 5, and transcription factor A (TFAM), and several other OXPHOS complex subunits and heme-related enzymes.

Mitoproteases are also essential components of the mitochondrial stress response (for instance heat stress, elevated ROS levels, low inner membrane potential difference). UPRmt (mitochondrial unfolded protein response) is a good example of a recently described mitochondrial stress response (Tran and Van Aken, 2020). It is triggered by mitochondrial proteotoxic stress (accumulation of unfolded/unassembled proteins) that induces the expression of nuclear genes (retrograde signaling), notably nuclear-encoded mitochondrial quality control components (protease and chaperones) in order to restore mitochondrial protein homeostasis. There is evidence for the implication of CLPP and LONP1 in this process probably together with YME1L1. Another example is the inactivation of mitochondrial fusion through the proteolytic processing of OPA1 by the inner membrane metalloendopeptidase OMA1 in stress conditions (Head et al., 2009). Intricate links between mitochondrial quality control and energetic metabolism have been reported in human metabolic diseases or mouse models (Civitarese et al., 2010; Almontashiri et al., 2014; Cavalcanti et al., 2014).

When the intramitochondrial PQC machineries are overwhelmed, additional cellular responses can be triggered to remove the dysfunctional organelles for instance by selective autophagy (mitophagy) or in extreme cases programmed cell death (Voos et al., 2016; Jadiya and Tomar, 2020; Ma et al., 2020).

5: Mitochondrial functions

5.1: Mitochondrial oxidative phosphorylations and mitochondrial carriers

Mitochondria outer membrane has a high permeability, thanks to porine (VDAC) which, on isolated mitochondria, allows the free diffusion of molecules up to 5000 Da. However, it has been shown that in situ, permeability of this channel could be regulated (Rostovtseva et al., 2008) leading to subtle regulations of mitochondrial oxidative phosphorylations and regulation of metabolic channeling processes (Rigoulet et al., 2004). In eukaryotic cells, oxidative phosphorylations take place in the inner membrane of mitochondria. According to the chemiosmotic theory initially formulated by Peter Mitchell (Mitchell and Moyle, 1967b), this process converts a redox energy—coming from food conversion processes—into a proton osmotic gradient (electrochemical proton potential difference which has an electric component ∆  Ψ and a chemical component ∆  pH) that is itself converted into another kind of chemical energy usable for various cell functions, that is, the phosphate potential (ATP/ADP.Pi). Ever since Mitchell’s proposal of the chemiosmotic theory, numerous studies have validated this hypothesis which is now well accepted. Oxidative phosphorylations require the mitochondrial respiratory chain, the ATPsynthase, ATP/ADP antiporter, and the Pi−/H+ symporter. In contrast to the outer membrane, the mitochondrial inner membrane is impermeable particularly for molecules that harbor a positive or negative charge. Metabolite exchange between the intermembrane space and mitochondrial matrix is due to a family of proteins called mitochondrial carriers. The main carriers are listed in Table 2 (Monné and Palmieri, 2014; De Stefani et al., 2016). These carriers can be classified in regard to their dependence on the components of the inner membrane potential difference (∆  Ψ and ∆  pH) and the transport mechanism. Of interest for the purpose of this chapter is the ATP/ADP carrier (ANT) that catalyzes the electrogenic exchange (dependent on ∆  Ψ) of ATP⁴  − with ADP³  −; another electrogenic exchanger is the glutamate/aspartate exchanger that catalyzes the electrogenic exchange of aspartate with glutamate  +  H+. This mechanism is responsible for the vectorial aspect of the exchange (i.e., glutamate import and aspartate export), which is mandatory for the transfer of reduced equivalents in the mitochondrial matrix by the malate aspartate shuttle. The Pi−/H+ and pyruvate/H  + symporters are electroneutral but dependent on the ∆  pH which favors Pi and pyruvate import. Di- and tricarboxylate carriers are electroneutral. Calcium import/export to/from the mitochondrial matrix involves three distinct carriers. Calcium import goes through mitochondrial calcium uniport which is electrogenic and calcium release from the matrix of energized mitochondria requires an electrogenic exchanger (Ca²  +/3-4Na+). A Ca²  +/2 K+ and a Ca²  +/2H+ exchanger have also been proposed. However, their relevance is still controversial (De Stefani et al., 2016).

Table 2

Mitochondrial carriers allow an exchange of varied compounds between the intermembrane space and the mitochondrial matrix. These carriers can be classified into two categories: the electrogenic carriers, such as adenine nucleotide translocator (ANT), aspartate/glutamate carrier (AGC) and mitochondrial Ca²  + uniporter (MCU), and the electroneutral carriers, such as phosphate inorganic carrier (PiC), oxoglutarate carrier (OGC), citrate carrier (CIC), dicarboxylate carrier (DIC), tricarboxylate carrier and mitochondrial pyruvate carrier (MPC).

The respiratory chain comprises four complexes: I–IV. Complexes I, III, and IV couple electron transfer to proton extrusion whereas complex II only transfers electrons to the quinone pool. Complex I (NADH–CoQ oxidoreductase) catalyzes the electron transfer from NADH to quinone. Complex II (succinate-CoQ oxidoreductase) couples succinate oxidation to quinone reduction. Complex III (quinol-cytochrome c oxidoreductase) couples quinol oxidation to cytochrome c reduction. Complex IV (cytochrome c oxidase) catalyzes the last step of electron transfer from reduced cytochrome c to oxygen. Several other enzymes exist in the inner mitochondrial membrane, which can feed the respiratory chain with electrons such as glycerol-3-phosphate (G3P) dehydrogenase, electron transfer flavoprotein dehydrogenase, dihydroorotate dehydrogenase, and choline dehydrogenase (see below). Complexes I, III, and IV use the redox span along the respiratory chain (Gibbs energy of the redox reaction) to pump out protons against the electrochemical proton gradient (see Fig. 3). The inner membrane electrical potential difference is used by some mitochondrial carriers and mainly by the ATPsynthase which is able to convert passive proton transfer into ATP synthesis from ADP and Pi (Fig. 3). The structure of the respiratory chain as a succession of complexes relies on both functional studies and biochemical isolation (Hatefi, 1985). Electron transfer in between the complexes depends both on the quinone pool and cytochrome c which are electron mobile carriers. Use of nonionic detergents allowed the isolation of more organized structures, called supercomplexes, in which some of the above-described complexes are associated with specific stoichiometries (Schägger and Pfeiffer, 2000). The composition of these supercomplexes has been shown to be variable and to depend on physiological conditions. It has been proposed that these supercomplexes would be an advantage to the functioning of the respiratory chain (such as electron channeling, kinetic efficiency, decrease in ROS production). However, these proposals are still a matter of debate (Trouillard et al., 2011; Moreno-Loshuertos and Antonio, 2016).

Fig. 3

Fig. 3 Scheme of the mammalian oxidative phosphorylation (OXPHOS) system. The OXPHOS system can be described as two functional entities: the respiratory chain (RC) and the phosphorylating system. The RC is composed of holoenzymes such as NADH dehydrogenase (complex I), succinate dehydrogenase (complex II), cytochrome c reductase (complex III), and cytochrome c oxidase (complex IV). There are also other dehydrogenases such as the glycerol-3-phosphate dehydrogenase (G3PDH) which can be connected to the RC. The RC is also composed of mobile electron carriers such as coenzyme Q and cytochrome c . The RC is functionally coupled to the phosphorylating system composed of the ATP synthase, adenine nucleotide transporter (ANT), and phosphate inorganic carrier (PiC).

5.2: Thermokinetic control and yield of oxidative phosphorylations

According to the chemiosmotic hypothesis (Mitchell, 1961, 1976b), the energy transduction between redox free energy and phosphate potential is allowed by inner mitochondrial membrane proton pumps structurally independent but functionally connected by an inner membrane electrical potential difference between two bulk phases: intermembranal space and mitochondrial matrix (∆  μH+). By analogy with the functioning of an electric battery, Mitchell introduced the term protonmotive force (∆  p = ∆  μH+/F), which is largely used in bioenergetics. Two main questions to be considered are (1) the yield of energy conversion (ATP/O) and (2) the relative fluxes involved in this process where both ATP synthesis and NADH oxidation are of importance. It is thus necessary to understand how the values of the different coupled fluxes are determined in an integrated system, such as oxidative phosphorylations. Obviously, the coupling of fluxes is mediated by forces, but at first sight the quantitative relationships between fluxes may depend on the properties of the whole system. In a complex metabolic network like mitochondrial oxidative phosphorylations, a very simple quantitative analysis is the determination of the yield of the overall reaction, as is currently performed by measuring ATP synthesis flux over oxygen consumption flux (ATP/O). With this approach, the yield may vary and several mechanisms possibly involved have been proposed. The first mechanism decreasing the coupling efficiency, the proton leak, is a direct consequence of the nature of the energetic intermediary, the protonmotive force. Indeed, biological membranes always present some proton conductance (LH), and the resulting proton flux is strictly dependent on protonmotive force (JH = LH.∆p). This membrane conductance is a specific property of the membrane itself, but is not entirely independent from the protonmotive force. Obviously, the size of this proton leak may modulate the yield of oxidative phosphorylations (ATP/O), and it has been clearly shown that protonophores uncouple oxidative phosphorylations by increasing proton membrane conductance (Bielawski et al., 1966; Hopfer et al., 1967; Mitchell and Moyle, 1967a, 1967c; Liberman and Topaly, 1968; LeBlanc Jr., 1971; McLaughlin, 1972). In this uncoupling process, the first event is a dissipation of protonmotive force which leads to an increase in respiratory rate and a decrease in the ATP synthesis rate, thus leading to a decrease in the ATP/O ratio. Even if much experimental work has established strong evidence that some protonophoric action can quantitatively account for the uncoupling of oxidative phosphorylations (Hanstein, 1976), no definitive proof that the uncoupling is exclusively and quantitatively due to an increase in cation membranal conductance has been obtained. In fact, from a growing number of data, it is evident that the question of the multiplicity of uncoupling mechanisms is still largely open. The first member of the uncoupling protein (UCP) family, brown adipose tissue uncoupling protein 1 (UCP1), was identified in 1976 (Ricquier and Bouillaud, 2000). Its uncoupling activity has largely been demonstrated as well as the regulators of this activity. It is activated by fatty acids and inhibited by GDP. Its activation ensures thermogenesis. Based on sequence homology, closely related proteins were described in mammals. To date, the actual function of these proteins (i.e., UCP2 and UCP3) is still a matter of debate (Negre-Salvayre et al., 1997; Bouillaud et al., 2016). The definition of the function of these proteins is crucial in terms of bioenergetics in the sense that a slight uncoupling would lead to a decrease in protonmotive force and in ATP synthesis as well as in mitochondrial ROS production.

Among the multiplicity of uncoupling mechanisms, two kinds of experimental evidence can be noted: it has been shown that (i) some uncoupling effects are not linked to a significant decrease in protonmotive force, which indicates that the decrease in oxidative phosphorylations efficiency is not, in this case, the consequence of an increase in membranal proton conductance (Rottenberg, 1983; Rottenberg and Hashimoto, 1986; Luvisetto et al., 1987; Pick et al., 1987; Luvisetto and Azzone, 1989; Rigoulet et al., 1989; Rottenberg and Koeppe, 1989); (ii) direct or indirect estimations of the coupled flow through different proton pumps indicate that their intrinsic stoichiometry, that is, the H+/2e− stoichiometry of the respiratory chain and the H+/ATP stoichiometry of the ATPsynthase, may vary as a function of many physical parameters or some drug addition (Luvisetto et al., 1987; Murphy and Brand, 1987; Rigoulet et al., 1989; Luvisetto and Azzone, 1989; Van Walraven et al., 1990; Bechmann and Weiss, 1991; Capitanio et al., 1991; Papa et al., 1991; Steverding and Kadenbach, 1991; Luvisetto et al., 1991; Ouhabi et al., 1991; Groth and Junge, 1993; Krenn et al., 1993; Fitton et al., 1994). This led Azzone et al. (Pietrobon et al., 1981; Pietrobon et al., 1983) to propose another mechanism causing a loss of oxidative phosphorylations yield. Such a new possibility called slip is a decrease in the efficiency of a proton pump due to partial and variable decoupling of chemical reaction and proton transport inside a respiratory complex. We have described a new energy wastage mechanism of interest. In isolated yeast mitochondria, the membrane proton conductance is shown to be strictly dependent on external dehydrogenases activity. An increase in their activity leads to an increase in membrane proton conductance. This proton permeability is independent of the respiratory chain and ATP synthase proton pumps. This mechanism is an active proton leak. In such a process, the high cellular redox constraints can be alleviated without affecting much the ATP synthesis flux and allows a decrease in mitochondrial ROS production (Mourier et al., 2010).

5.3: Mitochondrial ROS and their detoxification

5.3.1: Dioxygen: A poison for life

Dioxygen owning two unpaired electrons each located in a different antibonding orbital is a biradical. These two electrons have the same spin quantum number or, as is often written, they have a parallel spin. This constitutes the most stable state, that is, ground state of dioxygen. If a diatomic oxygen molecule attempts to oxidize another atom or molecule by accepting a pair of electrons, both of these electrons must have an antiparallel spin, so as to fit the vacant space in the orbitals. In most cases, electrons forming a pair in an orbital have opposite spins in accordance with Pauli’s principle. This imposes a restriction on electron transfer which leads dioxygen to accept electrons one at a time. Consequently, dioxygen reacts only very slowly with all nonradicals. Even if the redox potential of the O2/H2O couple is extremely positive, oxygen in the air cannot immediately burn living organisms due to the structure of their organic compounds. This spin restriction of dioxygen reactivity allowed the complex evolution of organisms and a metabolic adaptation, that is, the use of dioxygen as last electron acceptor during catabolism leading to an efficient energy transfer in cells (Halliwell, 2006).

When a unique electron is accepted by the ground state dioxygen, it enters one of the antibonding orbitals and forms the superoxide radical, O2 glyph_rad −. This compound is a very active electron donor, potentially able to generate a toxic cascade of electron transfer reactions. However, at physiological pH, the main reason for superoxide disappearance in aqueous solution is its dismutation catalyzed by the superoxide dismutase enzyme (2 O2 glyph_rad − + 2H+ − > O2 + H2O2). It is often said that spontaneous dismutation is very rapid but it is worth noting that (i) such a reaction becomes slower as the pH rises; (ii) yeast mutant devoid of superoxide dismutases (sod 1 and sod 2) cannot survive when cells switch from fermentation to growth fueled by respiration (Longo et al., 1996; Longo et al., 1999). If dioxygen accepts two electrons, it generates the peroxide ion, that is, O2²  − which, in biological media, is found as hydrogen peroxide (H2O2). Hydrogen peroxide is not a radical but is considered a ROS since in the presence of a transition metal it is able to form the hydroxyl radical, OH glyph_rad (i.e., the Fenton reaction). In biological systems, this product can induce a cascade of radical reactions potentially involving any cell compound.

5.3.2: ROS production by the respiratory chain

It is generally believed that the main superoxide producer in the cell is the respiratory chain (Fig. 4). Indeed, two of the respiratory chain complexes (I and III) have been, for a long time, recognized as involved in superoxide production (Chance and Hollunger, 1961). Under physiological conditions, predictions estimate that the superoxide production is around 0.1% of the respiratory rate (Hansford et al., 1997; Tahara et al., 2009). However, one should keep in mind that such a production is highly dependent on the respiratory state/rate of the mitochondria. Complex III has the highest capacity of ROS production due to the formation during the catalytic process—the Q-cycle (Mitchell, 1976a; Crofts et al., 1983)—of an instable semiquinone SQ glyph_rad in Qo, the site of quinol oxidation that faces outward of the inner membrane (Brand, 2016). However, it has been shown that the Q-cycle can produce superoxides both on the inner and on the outer surface of the inner mitochondrial membrane (Boveris et al., 1976; Cadenas et al., 1977). Since superoxides do not cross this membrane, it is of importance to know the size of the relative superoxide fluxes in the intermembranal space and into the matrix.

Fig. 4

Fig. 4 Main enzymatic systems of detoxification of ROS generated by the OXPHOS system. NADH dehydrogenase (I), glyceraldehyde-3-phosphate dehydrogenase (G3PDH), succinate dehydrogenase (II), cytochrome c reductase (III) generate superoxide; CoQ: coenzyme Q; Cyt c : cytochrome c ; GPX: glutathione peroxidase; Gred: glutathione reductase; GSH: reductive form of glutathione; GSSG: oxidative form of glutathione; TPX: thiol peroxidase; Tred: thiol reductase; Trx: thioredoxin; O 2 glyph_rad − : superoxide; SOD: superoxide dismutase; H 2 O 2 : hydrogen peroxide; TH: transhydrogenase; G3PDH: mitochondrial glycerol-3-phosphate dehydrogenase.

Another respiratory complex in the respiratory chain involved in superoxide formation is complex I. In this large multisubunit complex, there is now general agreement that the electron transfer is working at near equilibrium and thus, superoxide production may be linked to both forward electron transport (FET) and reverse electron transport (RET) (Chance and Hollunger, 1961; Hinkle et al., 1967; Korshunov et al., 1997; Kwong and Sohal, 1998; Votyakova and Reynolds, 2001; Kushnareva et al., 2002). It is likely that different sites of superoxide production exist in complex I and that the sites involved are different for FET or RET (Vinogradov and Grivennikova, 2001; Batandier et al., 2006). The sites of complex I-associated superoxide generation are still controversial. Three main sites have been proposed: the flavine mononucleotide (FMN) (Johnson Jr et al., 2003; Kudin et al., 2004; Kussmaul and Hirst, 2006), the Fe glyph_sbnd S clusters (more likely N1a or N2) (Herrero and Barja, 2000; Genova et al., 2001; Kushnareva et al., 2002), and a ubiquinone specifically linked to complex I (IQ site) (Lambert and Brand, 2004). Rotenone inhibits electron transfer right upstream of the quinone binding site. Consequently, superoxides that are produced by complex I in the presence of NADH are most likely due to the electron carriers (flavin or Fe/S clusters). However, ROS generation flux seems mainly linked to reverse electron flux and the IQ site could be a major player in this flux. It should be stressed here that unlike ROS production for FET, ROS production from RET can originate either at the actual RET site or at the dehydrogenase level (see below).

The Flavin site IIF of respiratory complex II (or succinate dehydrogenase) is proposed to be the site of ROS production in complex II and displays a similar maximal capacity of ROS production as site IQ (Brand, 2016) although Grivennikova et al. claim that their data are indicative of the [3Fe glyph_sbnd 4S] center, close to the ubiquinone reduction site, as the site of superoxide generation in this complex (Grivennikova et al., 2017). Wild-type complex II makes little contribution to ROS production in isolated mammalian mitochondria under normal conditions (St-Pierre et al., 2002). However, mutations in this complex can lead to abundant ROS production and cause pathologies (Ackrell, 2002). ROS production by the Flavin site of complex II, site IIF, in isolated mitochondria (Quinlan et al., 2012) is observed in the presence of inhibitors of complexes I and III (Brand, 2016). It shows in all cases a bell-shaped response of ROS production as a function of succinate concentration, with a maximum in the region of the KM of complex II for succinate (100–500 μM) (Quinlan et al., 2012; Siebels and Dröse, 2013; Grivennikova et al., 2017). There is currently no satisfactory explanation for this observation. In rat skeletal muscle mitochondria, the maximum capacity for ROS production in site IIF is very high, of the same order as site IQ (Fig. 2 in Brand, 2016).

As previously mentioned, the respiratory chain complexes are still the most studied elements regarding the contribution of mitochondria to ROS production including from a theoretical point of view (Mazat et al., 2020). However, the mitochondrial dehydrogenases have also been shown to be involved in this process (see Brand, 2016 for a review).

The α-ketoglutarate dehydrogenase complex (αKGDHC) catalyzes the oxidation of α-ketoglutarate in succinyl-coA with the production of NADH in the Krebs cycle. It is composed of multiple copies of three enzymes: α-ketoglutarate dehydrogenase (E1), dihydrolipoamide succinyl-transferase (E2), and dihydrolipoyl dehydrogenase (Dld or LADH, E3) (Massey, 1960; Wacenknechtt et al., 1983; Sheu and Blass, 1999; Duvezin-Caubet et al., 2006). Starkov (2004) showed that this complex is able to produce ROS during its catalytic process in mammalian brain mitochondria. This event seems to be linked to the flavin cofactor of the Dld, which can generate superoxide anion (Massey, 1994; Bunik and Sievers, 2002). These observations suggest the possibility of ROS generation by other mitochondrial flavoenzymes such as the pyruvate dehydrogenase complex (PDHC), which is also composed of a Dld (Starkov, 2004).

One way to reoxidize the NADH produced during glycolysis is the glycerol-3-phosphate shuttle. This shuttle is made up of two isoforms of the enzyme called glycerol-3-phosphate dehydrogenase (G3PDH) which differ by their localization and cofactors (see below). The mitochondrial isoform (Klingenberg, 1970) is located on the external side of the mitochondrial inner membrane and is a FAD-linked enzyme which donates electrons to the respiratory chain via the ubiquinone pool. This enzyme has been shown to produce ROS as well.

Years ago, it was shown that pigeon and rat mitochondria respiration with palmitoyl-carnitine as substrate could lead to H2O2 release (Boveris et al., 1972; Boveris and Chance, 1973). St-Pierre et al. (2002) made the same observations and showed that this H2O2 production is independent of the respiratory chain complexes and is on the matrix side. It has been shown that low and physiological concentrations of palmitoyl-carnitine as substrate for skeletal muscle mitochondria are sufficient for ROS production with a relatively low dependence on ∆  Ψ (Seifert et al., 2010). In mammals, fatty acid catabolism takes place in the mitochondrial matrix and is linked to the electron transport by acetyl-CoA provided to the Krebs cycle and via the flavoprotein ETF (electron transfer flavoprotein). ETF is reduced by acyl CoA dehydrogenase and passes its electrons to the ubiquinone pool via the ETF CoQ oxidoreductase enzyme (ETF-QO) present on the inner mitochondrial membrane. Studies on the catalytic mechanism of ETF-QO suggest that electrons can escape from the ETF glyph_rad − or from the semiquinone formed during ETF oxidation (Beckmann and Frerman, 1985a, 1985b; Ramsay et al., 1987). Consequently, ETF and/or ETF-QO are supposed to generate ROS with palmitoyl-carnitine as substrate.

5.3.3: Control of ROS production flux by mitochondria

The redox level of the respiratory chain electron carriers, including ubisemiquinone, is thermokinetically controlled. Thus, the forces directly associated with respiratory chain activity, that are the redox potential of the NADH/NAD+ couple and the protonmotive force, are powerful regulators of the steady-state concentration of the free ubisemiquinone radical. An increase in electron supply to the respiratory chain or in protonmotive force must lead to a rise in ubisemiquinone radical content. Moreover, kinetic constraints exerted downstream of the quinone pool (at the level of cytochrome oxidase for instance) could further increase the level of free radical ubisemiquinones. On the other hand, moderate uncoupling will lead to a decrease in force and kinetic constraints and thus effectively decrease superoxide production by the respiratory chain (Skulachev, 1998). The opposite effect can be observed when the respiratory chain is inhibited (Boveris and Chance, 1973; Cadenas and Boveris, 1980). Other mechanisms that induce a decrease in the redox proton pump efficiency (such as slipping) are expected to be less effective because they do not significantly affect the protonmotive force (Rottenberg, 1983; Zoratti et al., 1986; Ouhabi et al., 1991; Rigoulet et al., 1998). It is well known that the transition nonphosphorylating to phosphorylating or the uncoupling state decrease ROS generation in isolated mitochondria (Korshunov et al., 1997). Free fatty acids exert different effects on mitochondria: they inhibit electron flux at complex I and probably complex III levels inducing an increase in ROS production associated with FET in isolated rat heart or liver mitochondria (Schönfeld and Wojtczak, 2007). They also induce a slight uncoupling effect which seems responsible for a large decrease in ROS formation linked to reverse electron transfer. This illustrates the subtle ROS production response to changes in electron flux through the respiratory chain.

One of the physiological functions of the uncoupling protein family (UCP) could be the modulation of the protonmotive force in response to an increase in mitochondrial ROS production (Echtay et al., 2000; Casteilla et al., 2001; Echtay et al., 2001, 2003).

5.3.4: ROS enzymatic detoxification

ROS are involved in a number of physiological processes such as cell signaling (Carrière et al., 2003, 2020; Frezza, 2017), mitochondrial biogenesis (Chevtzoff et al., 2010; Yoboue et al., 2012), and numerous physiopathological processes. Thus ROS production and ROS level are tightly controlled within the cell through a number of processes. We will focus here on ROS enzymatic detoxification processes, even though a number of antioxidant compounds may participate in controlling ROS levels. At the level of the mitochondrial inner membrane, ROS can be produced both on the matrix side and in the intermembrane space. As stated above, electron transfer through both mitochondrial dehydrogenases and oxidative phosphorylations complexes can generate the superoxide ion. Once produced, this unstable molecule is dismuted thanks to matricial MnSOD and cytosolic/intermembrane space Zn/Cu SOD (Fig. 4). These enzymes will generate H2O2 and oxygen from the superoxide ion. H2O2 itself is not a radical species but is considered as a highly ROS since it is the substrate of the Fenton reaction catalyzed by transition metals. This reaction generates OH glyph_rad −, which is the most reactive oxygen radical. It should be stressed here that both superoxide and OH glyph_rad − are highly reactive and poorly diffusible whereas H2O2 is highly diffusible through biological membranes. This leads to compartmentation of