*2.2. Nuclear Gene Mutations of ATP Synthase and Its Assembly Factors*

Mutations in the nDNA genes encoding for ATP synthase subunits are very rare. Only a few different cases have been discovered over the years. The first mutation was reported in the *ATP5F1E* gene, which encodes ε subunit in the central stalk of the enzyme [49]. More recently, different mutations were found in the *ATP5F1A (*α subunit) [50,51], in the *ATP5F1D* (δ subunit) [63] or in the *ATP5MK* (DAPIT subunit) [64] genes. These nDNA mutations cause a similar and marked decrease in the content of fully assembled ATP synthase complexes, with a consequent decrease in their activity [49–52] and are detailed below.

The first case, the mutation causing Tyr12Cys amino acid residue substitution in the ε subunit, was described in a 22-year old woman presenting a neonatal-onset lactic acidosis, 3-methylglutaconic aciduria, mild mental retardation and developed peripheral neuropathy [49]. An analysis on patient fibroblasts with the homozygous missense mutation c.35A>G showed a decrease in the mitochondrial ATP synthase activity (both in ATP synthesis and hydrolysis) and assembly, caused by a reduction in the enzyme subunit level. Unlike the expression of the other subunits, the c subunit was found to accumulate and aggregate in a detergent-insoluble form, an accumulation that was also described in other disorders, such as Batten disease and fragile X syndrome, which will be discussed below. Overall, these findings suggest that the ε subunit is important for proper biosynthesis and assembly of the ATP synthase and for the proper incorporation of the c subunit into the rotor structure [49]. In line with these results, the downregulation of the ε subunit in HEK293 cells [71] or in yeast [72], caused a decrease in mitochondrial content and the activity of ATP synthase, with an effect on c subunit accumulation in HEK293 cells depleted of the ε subunit [71].

One of the other mutations (c.985C>T) of the nuclear *ATP5F1A* gene encoding the α subunit mentioned above was found in two different siblings who died in the first weeks of life [50]. The severity of this phenotype depends on the fact that the wild-type allele of the mother was not expressed in the siblings. Patient fibroblasts showed a reduction in the oxygen consumption rate, possibly caused by impairment of ATP synthase assembly and function. These cells displayed, in line with a decreased enzyme assembly, a decreased content of the subunits α, β, OSCP or d, which are important for the enzyme catalysis [50]. The authors showed that the expression of the wild-type gene encoding the α subunit in patient fibroblasts rescued the ATP synthase complex content and activity [50]. The possible explanation of the functional effects of the described mutation proposed by the authors implies that the substitution of arginine 329 for cysteine abolishes the three α-β interactions in the catalytic core of the enzyme, with the consequent loss of stability of the entire ATP synthase complex [50]. Another case is the homozygous c.962A>G mutation in the *ATP5F1A* gene that was described in two sisters born from consanguineous first-cousin parents. They both died early after birth with microcephaly, pulmonary hypertension and heart failure [51]. Patient muscle tissue showed OXPHOS deficiency and mtDNA depletion. Additionally, in this case, the p.Tyr321Cys mutation involved a highly conserved residue. The expression of the analogous yeast variant (*ATP1*: p.Tyr315Cys) in an *ATP1* knockout strain reflected the same severe phenotype with mtDNA loss, decrease of mitochondrial membrane potential and petite phenotype [51]

Other examples of nuclear mutations known to cause a mitochondrial dysfunction involved the δ subunit and its *ATP5F1D* gene. Two different homozygous mutations, c.245C>T (p.Pro82Leu) and c.317T>G (p.Val106Gly), were found in two unrelated individuals with a metabolic disorder [52]. Cultured skin fibroblasts from these individuals showed an impaired ATP synthase assembly, as revealed through BN-PAGE, and decreased enzyme activity. Moreover, in both subjects, the amount of the δ subunit was unchanged but not that of other subunits like α, β or OSCP, which were decreased in abundance. Through in silico modeling, the authors found that each of the amino acid substitutions induces changes in the predicted structure of the protein. According to these data, they hypothesized that these changes can alter the ability of the δ subunit to bind and interact with the F<sup>1</sup> subunits and thus affect the proper assembly of the enzyme. The pathogenicity of the two *ATP5F1D* variants was corroborated by studies performed in Drosophila. Indeed, both the mutated proteins were unable to complement the phenotypic defects caused by the δ subunit depletion in Drosophila, whereas the human wild-type subunit did [52]. Interestingly, and in line with the effect on the enzyme assembly, fibroblasts from the patient with a c.245C>T mutation showed significant decrease in mitochondrial cristae content [63], a fact consistent with the role of ATP synthase dimers in maintaining normal mitochondrial cristae morphology [73,74].

Recently, a novel homozygous splice-site mutation (c.87+1G>C) in the ATP synthase *ATP5MK* gene (encoding the DAPIT subunit) was described in three unrelated Ashkenazi Jewish families. The mutation negatively affects enzyme dimerization and ATP synthesis

rate. Rescue with wild-type *ATP5MK* cDNA in patient fibroblasts restored the DAPIT protein levels, and enhanced ATP synthase dimers and their activity [53].

The biosynthesis of the eukaryotic ATP synthase is a highly organized process that requires the action of specific assembly factors [56,75–79]. It was shown that mutations in some of these "chaperone" proteins, named ATPA12 and transmembrane protein 70 (TMEM70), can be responsible for secondary ATP synthase deficiencies [80], leading to altered assembly and compromised activity of the enzyme.

The ATP12 protein is known to interact with the unassembled α subunit and is essential for its incorporation into the ATP synthase complex [78]. In a genetic study, De Meirleir et al. discovered a homozygous T>A missense mutation in exon 3 of the *ATPAF2* gene in a girl [54]. This mutation caused the amino acid substitution of a conserved tryptophan to an arginine at position 94 (p.Trp94Arg), which decreased the solubility of the protein with a tendency to aggregate [55]. The consequence is a severe decrease in the ATP synthase complex assembly and activity [54], even if no alteration in the mitochondrial morphology was observed in the fibroblasts derived from the patient carrying this mutation [55].

The other important regulatory protein in ATP synthase assembly is TMEM70, localized in the inner mitochondrial membrane [77]. Different mutations have been found over the years for *TMEM70*, with a broad spectrum of phenotypes and severity. The most common features of the syndrome caused by *TMEM70* mutations are a severe neonatal lactic acidosis, 3-methlyglutaconinc aciduria, cardiomyopathy, facial dysmorphism and mental retardation [7,81]. Early evidence for the role of TMEM70 in the enzyme assembly came in 2008 [56] and was later confirmed when it was shown that TMEM70 promotes the ATP synthase assembly by interacting with subunit c. This interaction facilitates the incorporation of the c subunit into the rotor structure of the enzyme within the inner mitochondrial membrane [76,77]. The c.317–2A>G mutation, which was firstly reported at the end of the second intron of the *TMEM70* gene, resulted in aberrant splicing and the loss of the transcript. As a consequence, low ATP synthase activity and assembly were observed. Fibroblast carrying this mutation were complemented with the wild-type *TMEM70*, which rescued structural and functional changes of ATP synthase, suggesting, for the first time, the importance of TMEM70 in the enzyme assembly [56]. All patients affected by this common mutation that were later diagnosed and exhibited ATP synthase deficiencies similar to the aforementioned case [57–59]. Other, less-common mutations on the *TMEM70* gene were found in many patients from different ethnic groups, with various phenotypes ranging from the absence of TMEM70 protein due to the premature stop codon, to the synthesis of an incomplete truncated form of the factor, lacking functional or structural domains [7,58,81–84]. As expected, a mitochondrial defect characterized by a decrease in ATP synthase assembly and activity was described in these patients [58,83,84]. Mitochondrial ultrastructural analysis in some *TMEM70* mutant samples showed a fragmented mitochondrial network and impaired mitochondrial morphology, swollen mitochondria or altered and concentric cristae [57–59,84] in line with the role of properly assembled ATP synthase in cristae shaping [73,74]. However, it has been shown that the disrupted mitochondrial cristae architecture in some patients also impairs the activity and localization of other OXPHOS complexes, increasing the severity of the disease [57,58,84]. Importantly, the mitochondrial effects caused by *TMEM70* mutations could be completely restored by complementation with the wild-type gene [84].

#### **3. ATP Synthase Dysfunctions in other Human Diseases**

The ATP synthase dysfunctions involved in the pathogenic events, leading to cardiovascular, neurodegenerative and neurocognitive diseases are described below and shown in Figure 2.

eimer's, Amyotrophic Lateral Sclerosis (ALS), fragile X, Batten and Parkinson's diseases are shown with different colors in the external perimeter. **Figure 2.** Schematic representation of the ATP synthase modifications involved in the progression of human diseases. Cardiovascular, Alzheimer's, Amyotrophic Lateral Sclerosis (ALS), fragile X, Batten and Parkinson's diseases are shown with different colors in the external perimeter. Changes in ATP synthase catalytic activity, assembly, subunit levels or subunit post-translational modifications and their consequence on mitochondrial morphology are listed inside the circle in correspondence of each related disease.

–

#### *3.1. Cardiovascular Disease and Cardio-Protection*

– Myocardial cell death due to ischemia–reperfusion is a major cause of morbidity and mortality in western nations. Rouslin has first demonstrated downregulation of the mitochondrial ATP synthase activity in ischemic heart tissue from different animals [85,86]. Under ischemic condition, the ATP synthase works in reverse hydrolyzing ATP. Thus, inhibition of the ATP synthase hydrolytic activity under these conditions conserves cellular ATP levels. The membrane potential prevents uncontrolled influx of ATP into the mitochondrial matrix via the electrogenic ATP/ADP translocator, thus limiting ATP hydrolysis. Furthermore, it is stated that during ischemia, the mitochondrial ATPase inhibitor protein (IF1) binds to and inhibits the mitochondrial ATPase, thereby conserving ATP [85,87–89]. IF1 can also contribute to the myocardial ischemic preconditioning, reducing the mitochondria damage during early reperfusion [90]. Cyclophilin D (CyPD), the permeability transition pore modulator, which also inhibits the ATPase catalytic activity [91], may contribute in preventing ATP dissipation as an additional mechanism of ATP synthase modulation.

The reintroduction of oxygen during reperfusion allows the generation of ATP, but the damage to the electron transport chain results in increased mitochondrial generation of ROS. The catalytic activity of ATP synthase has been shown to be regulated in the presence of ROS in several cardiovascular studies, a fact which might be caused by oxidation of specific enzyme residues [92]. Mitochondrial Ca 2+ overload and increased ROS can result in opening of the mitochondrial permeability transition pore, which further compromises cellular energetics and induces cell death. Apart from permeability transition-directed therapies [93], the cardioprotective strategy of ischemic preconditioning (PC), first described in 1986, provided an indication of the magnitude of the possible cardioprotective effect [94]. The rate of ATP consumption during ischemia is slower in PC hearts [95,96]. The full mechanism by which PC reduces ATP break down during ischemia remains still

unknown. Interestingly, cardiac-specific overexpression of the antiapoptotic protein Bcl-2, overexpression of the cardioprotective PKC-ε and adenosine pretreatment have all been shown to slow the rate of ATP breakdown during ischemia [97–99]. Di Lisa et al. used the fluorescent membrane potential-sensitive dye JC1 to measure mitochondrial ∆ψ in anoxic rat cardiomyocytes and showed a biphasic decline in ∆ψ [100]. These authors showed that glycolytically generated ATP was used to maintain ∆ψ, since ∆ψ was shown to decline more rapidly during ischemia in the presence of oligomycin, an ATP synthase inhibitor. Leyssens et al. obtained similar results using JC1 to measure ∆ψ in rat cardiomyocytes metabolically inhibited with cyanide and 2-deoxyglucose [101]. These data support the conclusion that ATP synthase is a major consumer of ATP during ischemia and/or metabolic inhibition, and they further demonstrated that the consumption of glycolytic ATP is used to maintain ∆ψ [87].

It was proposed that PC promoted earlier binding of the IF1 to ATP synthase. However, studies by two different groups using submitochondrial particles found no evidence supporting inhibition of the ATPase in PC hearts [102,103]. Other groups, on the other hand, have reported that PC and diazoxide enhance the IF1 binding to ATP synthase [104–106]. A previous study has also reported that pharmacological PC with adenosine results in increased phosphorylation of the β subunit of the ATP synthase [107], although the functional effects of phosphorylation on the ATPase activity were not addressed. Subsequent studies aimed at generating different yeast mutants to better define the role of the β subunit phosphorylation, demonstrated its impact on enzyme assembly/stability and catalytic activity [108].

Additionally, changes in the amount of ATP synthase subunits have been shown in other cardiac patients. During an inflammatory cardiomyopathy occurring in patients affected by chronic Chagas disease, the most characteristic histopathological lesions are consistent with inflammation and a myocardial remodeling process such as T-cell/macrophagerich myocarditis, hypertrophy and fibrosis [109]. Protein analysis showed a 20% decrease in the ATP synthase subunits α and β in the myocardium of chronic Chagas patients compared to myocardial samples from individuals without cardiomyopathies [110]. Since the analysis of the mRNA levels did not show significant differences [110], it seems plausible that a modulation at the level of subunit turnover or enzyme assembly might occur in chronic Chagas patients.
