*11.2. Accumulation of the Reduced Form of Pyridine Nucleotides and CoQ*

Direct consequences of defective CIII are energy failure and metabolic derangements, as indicated by a huge number of case reports describing lactic acidosis and hypoglycaemia as recurrent clinical phenotypes of patients bearing mutations in different CIII-related genes, i.e., *MTCYB* [120,141], *UQCRC2* [127,142], *UQCC3* [86] and others. Metabolomics analyses in liver of the mouse model of CIII dysfunction (*Bcs1lc.232A>G* mutant) revealed a decrease in carbohydrate intermediates, demonstrating an increase in glycolysis to compensate for the reduced mitochondrial ATP production [143]. Subsequently, targeted metabolomics detected increases in glucogenic and ketogenic amino acids in circulation, supporting a starvation-like condition [144]. Of note is that this mouse model, despite the severe CIII dysfunction, does not present significant perturbations in the SCs organization [115].

On the other hand, when the assembly of CIII-containing SCs is compromised, in addition to CIII, the amount of CI collapses as well, leading to elevation of the cellular ratio of reduced and oxidized pyridine nucleotides (NADH/NAD<sup>+</sup> ). The inability to oxidize NADH in the mitochondrial matrix affects not only the efficiency of OXPHOS but also the flux of metabolites through the Krebs cycle. As a consequence, the cells become heavily dependent on aerobic glycolysis for survival. The glycolytic flux relies on the activity of glyceraldehyde-3-phosphate dehydrogenase which requires NAD<sup>+</sup> , generated from NADH oxidation by the cytosolic lactate dehydrogenase enzyme. In agreement with this notion, we found that the amount of lactate released into the growth medium by the homoplasmic cells bearing the 18-bp *MTCYB* deletion was significantly greater than WT cells. Noticeably, cells bearing the p.278Y>C *MTCYB* mutation impairing CIII activity without affecting SCs organization failed to increase lactate release [121], in accord with the clinical phenotype of patients, presenting lactic acidosis in the patient bearing the 18-bp *MTCYB* deletion [120], but not in that with the p.278Y>C mutation [145]. The molecular mechanism underlining this metabolic switch is unknown, although the possible role for UQCC3 may be worth investigating [88].

Interestingly, previous studies described benefits in lifespan and energetic function of defective CI by interventions targeting NADH elevation, such as supplementation with NAD-precursor [146] inhibition of mTOR [147] and of mitochondrial serine catabolism [148], as well as hypoxia treatment [149]. To circumvent the CI deficiency, some studies took advantage of the xenotopic expression of the single-subunit yeast enzyme NADH dehydrogenase (Ndi1) [150–152]. In yeast, Ndi1 catalyses the oxidation of NADH in the matrix like CI, but is unable to restore the proton pumping. Ndi1 protein

expression in human cultured cells lacking CI restored the NADH-dependent respiration as well as the growth in glucose-free medium containing galactose [153,154]. Recently, McEllroy et al. (2020) generated a mouse that conditionally expresses Ndi1, confirming that its expression dramatically prolong lifespan, but was unable to significantly improve motor function in a mouse model of Leigh syndrome due to loss of the NDUFS4 CI subunit [155]. In the absence of structural data showing Ndi1 association with SCs, it is reasonable to speculate that the ability of Ndi1 to ameliorate the cell viability does not depend on association with other respiratory complexes, rather it depends on the restoration of NADH oxidation allowing for a compensatory increase in glycolysis and sufficient metabolite flux in the Krebs cycle.

At cellular level, the primary consequence of the specific drop/lack of CIII is the blockade of CoQH<sup>2</sup> oxidation, preventing the NADH and FADH<sup>2</sup> oxidation by CI and CII. As mentioned above, elevation of CoQH2/CoQ ratio causes reverse electron transport through CI, with local generation of superoxide, triggering CI subunits degradation and tuning the amount of this complex [94]. This is in agreement with previous data showing that CIII can be released from CI-containing SCs under metabolic conditions (e.g., starvation) when electron flux from FAD overwhelms the oxidation of CoQ, supporting the plasticity model of SCs organization [81]. The alternative oxidase (AOX) is a single-protein electron transport system present in bacteria, lower eukaryotes and plants that can perform CoQH<sup>2</sup> oxidation instead of CIII and CIV, by transferring electrons directly from quinols to oxygen without proton translocation [156]. The xenotopic expression of tunicated AOX in mouse was recently investigated, failing to show any association of AOX with SCs. This finding supports the notion that xenotopically expressed AOX acts as a freely diffusible redox partner [157]. Of note is that the expression of AOX from *Emericella nidulans* in *MTCYB* KO cells induced CoQH<sup>2</sup> oxidation, thus reducing the oxidative stress and inhibiting CI degradation. Despite the increased CI amount, the SCs were not restored due to the lack of CIII [94], further corroborating the central role of CIII as a scaffold for incorporation of CI and CIV [118]. Furthermore, AOX was reported to provide a full functional rescue of the cardiomyopathy of the *Bcs1l*c.232A>G mutant mice, by restoring respiration to wild-type level. Noticeably, the CIII and CI+CIII2+CIV assembly was partially rescued in cardiac mitochondria, likely secondary to the general improvement in mitochondrial structure and function [144].

The observation that expression of a single enzyme, such as AOX, can bypass defective oxidative reactions carried out by dozens of proteins is intriguing. Besides representing a useful tool for detecting the contribution of ATP requirement from the NAD<sup>+</sup> and CoQ regeneration, the oxidase may be of potential use for respiratory chain deficiencies, although a gene therapy approach seems quite problematic at present. In fact, it has to be considered that correction of the mutated gene in affected tissue of monogenic diseases by CRISPR/CAS9 genome editing, already shown to be promising in animal models, requires sophisticated gene-specific tools and is still under development for mtDNA interventions. Conversely, the expression of one protein such AOX might be, in theory, beneficial for restoring most of the metabolic stress induced by OXPHOS impairment caused by a wide variety of mutations.

#### *11.3. Elevation of Succinate and Effects on Gene Expression Regulation*

The lack of CIII and the extremely limited availability of oxidized CoQ results in the inability of the Krebs cycle to progress from succinate to fumarate, as demonstrated by the markedly increased levels of succinate and reduction of fumarate and malate detected by us in cells with the 18-bp *MTCYB* deletion [121] and also in cells bearing the 4-bp *MTCYB* deletion [158]. In cells with the p.278Y>C *MTCYB* mutation, with normal SCs organization, we also detected a weak increase of succinate, but increased malate and normal fumarate, suggesting that some succinate can be oxidized to fumarate which in turn produces malate, likely as a consequence of CIII assembly in the respirasome, that can preserve significant

electron transport [121]. Elevation of succinate and fumarate was reported in the liver of *Bcs1lc.232A>G* mutant mouse, in line with the blockade of the Krebs cycle flux [159].

Succinate accumulation in the cytosol has been shown to have a strong impact in gene expression regulation, by inhibiting the 2-oxoglutarate-dependent dioxygenases, which catalyse hydroxylation reactions on various types of substrates. In particular, succinate, competing with 2-oxoglutarate, inhibits the activity of prolyl hydroxylases, leading to stabilization of Hypoxia Inducible Factor-1α (HIF-1α) under normoxia, defined as a pseudohypoxic condition [160]. HIF-1α can then translocate from cytoplasm to the nucleus where it associates with HIF-1β, to activate transcription of HIF-1α-target genes, among which are those encoding glycolytic enzymes [161]. It is likely that the huge increase of succinate levels determined in cells with the 18-bp *MTCYB* deletion may be at least in part responsible for the glycolytic switch revealed by increased extracellular lactate production [121].

In addition to succinate and fumarate, the levels of the oncometabolite 2-hydroxyglutarate (L-2-HG) were increased in RISP/ UQCRFS1 KO cells [162] and also in cells bearing the 4-bp *MTCYB* deletion [158]. All these Krebs cycle metabolites competitively inhibit the activity of 2 oxoglutarate-dependent dioxygenases, including also JmJC domain-containing histone lysine demethylases and ten-eleven translocation TETs family of 5-methlycytosine hydroxylases, involved in oxidizing 5-methylcytosine into 5-hydroxymethylcytosine [163]. Accordingly, the DNA and histone methylation were increased upon loss of RISP in fetal hematopoietic stem cells, impairing their differentiation and maintenance of stemness [162]. The increase in these metabolites, through inhibition of the histone and DNA demethylases, can therefore represent a very important factor affecting the epigenetic landscape of the cells and causing wide-ranging effects on cell physiology (for a review on Krebs metabolites and epigenetics, see [164]).

#### **12. Conclusions and Perspectives**

The development of cryo-EM technology has provided a powerful tool to analyse at atomic level the specific associations of CI, CIII and CIV into the respirasome, the CI+CIII<sup>2</sup> and the CIII2+CIV1-2 SCs. However, these techniques suffer from some limitations, mainly associated with the mitochondrial purification procedures and the type/amount of detergent. In this regard, the development of the "in situ" reconstruction of SCs in eukaryotic cells in vivo by using the proximity-dependent labeling followed by mass spectrometry will open the possibility to identify potentially interacting proteins and their subcellular spatial localization [165,166]. Dynamic rearrangements between individual complexes and SCs have been demonstrated to occur (Table 1). It is necessary to identify the functional consequences of these arrangements as well as their implications in the regulation of the respiratory function under different physiological conditions. More biochemical and biophysical experiments, in combination with advances in super-resolution light microscopy are needed for clarify the functional mechanism of the SCs. This information will be crucial to elucidate the pathogenic mechanisms underlying the mitochondrial disorders associated with both nuclear and mtDNA mutations, and, hopefully, to identify effective treatments.


**Table 1.**Disease genes encoding structural subunits and assembly factors associated with CIII deficiency.


**Table 1.** *Cont.*


**Table 1.** *Cont.*

\*, a stop in the protein synthesis; ↓, decrease; ↑, increase; =, no change; n.d., not determined.

**Author Contributions:** Conceptualization, A.M.G. and M.R.; writing, original draft and figures preparation, M.R.; writing and editing, A.M.G.; supervision, C.Z. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Acknowledgments:** We thank Erika Fernandez-Vizarra, University of Glasgow, Scotland, UK, for kindly sharing some SILAC and complexome profiling results. This article is dedicated to the memory of Francisca Diaz, University of Miami, USA, who significant contributed to the study of SCs alteration in mitochondrial dysfunctions.

**Conflicts of Interest:** The authors declare no conflict of interest.
