Mitochondrial DNA mutations at MT-ATP6 gene are relatively common in individuals suffering from striatal necrosis syndromes. These patients usually do not show apparent histochemical and/or biochemical signs of oxidative phosphorylation dysfunction. Because of this, MT-ATP6 is not typically analyzed in many other mitochondrial disorders that have not been previously associated to mutations in this gene. To correct this bias, we have performed a screening of the MT-ATP6 gene in a large collection of patients suspected of suffering different mitochondrial DNA (mtDNA) disorders. In three cases, biochemical, molecular-genetics and other analyses in patient tissues and cybrids were also carried out. We found three new pathologic mutations. Two of them in patients showing phenotypes that have not been commonly associated to mutations in the MT-ATP6 gene. These results remark the importance of sequencing the MT-ATP6 gene in patients with striatal necrosis syndromes, but also within other mitochondrial pathologies. This gene should be sequenced at least in all those patients suspected of suffering an mtDNA disorder disclosing normal results for histochemical and biochemical analyses of respiratory chain.
Approximately 10–20% of individuals with Leigh syndrome (LS) and more than 50% of individuals with neurogenic muscle weakness, ataxia and retinitis pigmentosa (NARP) have a mutation at mitochondrial DNA (mtDNA) nucleotide position 8993 of the MT-ATP6 gene (1). These two clinical conditions have been termed ‘striatal necrosis syndromes’ and have been also associated to mutations at nucleotide position 9176 of the MT-ATP6 gene (2). Thus, these mutations have been frequently analyzed in persons suffering these pathologies. In these patients, histologic examination of muscle biopsies shows only minimal, if any, changes and the specific activities of respiratory complexes are usually normal (1), because the p.MT-ATP6 polypeptide is not a component of the respiratory chain but of the oxidative phosphorylation system (OXPHOS) complex V (CV) or ATP synthase. Moreover, for ADP phosphorylation to take place, mitochondria need to be maintained intact and in a coupled state. Then, ATP synthesis can only be measured on freshly isolated mitochondria (3). Therefore, ATP synthase-specific activity is not regularly measured in patients' tissues. Because of the general lack of positive histochemical and biochemical evidences on the implication of p.MT-ATP6 in a particular pathology based in MT-ATP6 mutations, its screening in other mitochondrial diseases has been infrequent.
To avoid these biases, we have performed a screening of the MT-ATP6 gene in a big collection of patients suspected of suffering mtDNA disorders, as NARP, maternally inherited LS (MILS) and other with very different clinical phenotypes.
Three out of the 550 patients studied harbored interesting mutations in their MT-ATP6 genes. These mutations consisted of transitions at m.9025G>A, m.9029A>G and m.9032T>C nucleotide positions for P1, P2 and P3, respectively (Fig. 1). To rule out the presence of other potentially pathological mutations, we sequenced the complete mtDNAs of these three patients, but no other variant was a better candidate for being the pathogenic mutation (Supplementary Material, Note S1). These analyses also provided us their mtDNA haplogroups (U5b2b3, W1* and U5b1f). Next, we analyzed the presence of these mutations in 21 580 human p.MT-ATP6 sequences obtained from GenBank plus 4352 from different publications (Supplementary Material, References S1). The m.9025G>A, m.9029A>G and m.9032T>C transitions were found 13, 1 and 0 times, respectively, in apparently healthy individuals and always in different mtDNA genetic backgrounds and distinct from our patients (Supplementary Material, Note S2). To rule out these mutations as more geographically restricted polymorphisms, we also studied 13 and 18 Spaniard population samples from mtDNA haplogroups W and U5b, respectively. The mutations were not found in any of them. In addition of the 550 patients from the initial screening, we also analyzed by real-time polymerase chain reaction (PCR) the m.9025G>A/m.9032T>C and m.9029A>G transitions in all remaining samples of our laboratory from individuals suspected of suffering from MILS/NARP and optic neuropathy, respectively. We found the m.9025G>A mutation in a MILS patient (P4, Supplementary Material, Note S1). Finally, the m.9025G>A mutation had been found in one patient with colon cancer from haplogroup H7d3a and another with motor neuropathy and the m.9029A>G transition had also been found in one patient from haplogroup H with suspected mitochondrial disorder (Supplementary Material, References S2).
These mutations were firstly discovered by conformation sensitive gel electrophoresis (CSGE) and/or sequencing (Supplementary Material, Fig. S1A). To rule out sequencing artifacts and confirm the presence of these mutations by another different method, we performed PCR-restriction fragment length polymorphism (RFLP) analyses. These mutations were found in heteroplasmy in at least one individual from each pedigree (Supplementary Material, Fig. S1B). However, the mtDNA region around these mutations is highly represented in nuclear mitochondrial DNA sequences (NUMTs) (4) and short amplicons for m.9025G>A and m.9029A>G, but not for 9032T>C, analyses were observed from rho0 cells (Supplementary Material, Fig. S1C). Therefore, to ratify heteroplasmies, determine their true percentages and their association with the phenotype, we designed a larger amplicon (Supplementary Material, Fig. S1D) and performed a nesting radioactive PCR-RFLP in probands and relatives. The m.9025G>A mutation was found in homoplasmy in proband's muscle, liver and blood and in mother's and sister's blood (Fig. 2A). The m.9029A>G mutation was found in homoplasmy in the patient's blood but in heteroplasmy in his mother (95%) and siblings (brother 85%, sister 95%) (Fig. 2B). This heteroplasmy is authentic because only two NUMTs in chromosomes 5 and 17 have enough length to be amplified with our PCR conditions and, despite that their targets for the genotyping restriction enzyme are identical to that of the normal mtDNA sequence, there are many mispairings in the primers-matched sequences (Supplementary Material, Note S3) and this segment was not amplified in rho0 cells (Supplementary Material, Fig. S1D) or the patient blood (Fig. 2B). The m.9032T>C transition was found in heteroplasmy in proband's urine (96%), hair (93%) and blood (70%) and also in mother's hair (73%), urine (57%) and blood (42%) but it was not found in hair, urine, buccal mucosa or blood of a healthy maternal aunt (Fig. 2C).
As the mutation m.9025G>A was found in homoplasmy in the asymptomatic mother and sister and m.9029A>G was found at a high percentage in the asymptomatic mother and siblings, we looked for potential compensatory factors, such as the mtDNA levels, in symptomatic and asymptomatic family members (5). There were no differences in mtDNA levels in members of the m.9025G>A pedigree, but they were significantly lower in the m.9029A>G patient than in his relatives (Fig. 3).
The m.9025G>A, m.9029A>G and m.9032T>C transitions provoke glycine to serine, histidine to arginine and leucine to proline amino acid substitutions, respectively. The MutPred score mean ± standard deviation for all of the 24 206 possible amino acid variations defined by a single point mutation away from the revised Cambridge reference sequence (rCRS) is 0.642 ± 0.154 (6). These values for all of the 111 glycine to serine, 97 histidine to arginine and 553 leucine to proline possible substitutions were 0.749 ± 0.120, 0.728 ± 0.166 and 0.737 ± 0.105, respectively, and they were significantly higher than those for all of the possible amino acid substitutions (P < 0.0001). The particular values for these changes were 0.860 (G167S), 0.827 (H168R) and 0.837 (L169P) (Table 1). Because a higher pathogenicity score correspond to a greater likelihood that the amino acid variation is pathogenic, these results suggest that these particular amino acid substitutions will have phenotypic consequences.
|Mutation load||Phenotype||Pathologic pedigrees||Controls (>25 000)||Mt-CI (%) (3410)||B-CI (%) (1409)||MutPred||Cybrids|
|Previously reported mutations|
|Ho||CC, MN, LS-like||S||13||99.7||86.2||0.860||N|
|Mutation load||Phenotype||Pathologic pedigrees||Controls (>25 000)||Mt-CI (%) (3410)||B-CI (%) (1409)||MutPred||Cybrids|
|Previously reported mutations|
|Ho||CC, MN, LS-like||S||13||99.7||86.2||0.860||N|
To consider a previously published p.MT-ATP6 mutation as a pathologic one, the mutation should fulfill at least one of these two criteria: (i) the mutation should be present in more than one pathologic pedigree, its population frequency should be very low (below 0.05%) and it should not be located at internal branches of the mtDNA phylogenetic tree or (ii) phenotypic differences should be shown between cell models harboring different percentages of mutant and control mtDNAs. The m.8989G>C and m.9191T>C mutations do not fulfill these criteria. However, several other features suggest that they are pathologic mutations. AOSA, adult-onset spinocerebellar ataxia; B-CI, conservation index for bacterial subunit a; BSN, bilateral striatal necrosis; C, cybrids; CC, colon cancer; CMT, Charcot-Marie-Tooth; Ho, homoplasmy; He, heteroplasmy; IC, infantile cardiomyopathy; LHON, Leber hereditary optic neuropathy; LS, Leigh syndrome; MN, motor neuropathy; Mt-CI, conservation index for p.MT-ATP6; MutPred, a proxy index for the pathogenicity of an amino acid variation (6); N, No; NARP, neuropathy, ataxia and retinitis pigmentosa; O, one; S, several. A hyphen in the B-CI column signifies that we cannot trustworthily indicate the equivalent amino acid in the bacterial subunit a.
The G167, H168 and L169 amino acids are extremely preserved through evolution. Thus, they were conserved in 99.7, 98.1 and 87.9%, respectively, of 3410 published p.MT-ATP6 reference sequences from animals, plants, fungi and protists (Table 1). G167 was the second most conserved p.MT-ATP6 position after R159, a key residue for proton transport. As there is a bias in the type of organisms with their mtDNAs sequenced (most of the sequenced mtDNAs belong to animals), to confirm the clues about the functional importance of these amino acids, we also estimated the conservation indexes (CI) for these positions in 1409 bacteria that are organisms very distant from an evolutionary point of view. The CIs were 86.2, 36.0 and 33.4%, respectively (Table 1).
The G167, H168 and L169 amino acids are located in the transmembrane helix 4 (TMH4) of p.MT-ATP6 (Fig. 4A and B). TMH4 and TMH5 of p.MT-ATP6 and TMH2 of subunit c form the F1F0 ATP synthase proton channel (7). Four other positions affected by pathologic mutations (A155, L156, A162 and L170) are also located in this TMH4. Moreover, G167 and L169 are placed along with A155, L156, R159, A162 and L170 in the most conserved side of TMH4 (Fig. 4C). A structurally compensating replacement in any of these TMHs could explain the phenotypic differences between the homoplasmic m.9025G>A patient and her asymptomatic maternal relatives. However, there were not differences between the p.MT-ATP6 and c subunits (the three isoenzymes) from the patient and relatives.
The previous results suggested that these mutations might be functionally important and, maybe, they were responsible for patients' phenotypes. To confirm this possibility, we built seven transmitochondrial cell lines (cytoplasmic hybrids or cybrids) using the osteosarcoma 143B206 rho0 cell line as the common nuclear background (8) and platelets from each of these patients. Unfortunately, family of deceased P1 did not want to continue with the study and from P4, a patient studied 7 years ago, we only had muscle. Therefore, we could not produce cybrids with this mtDNA mutation. We used a haplogroup H cybrid harboring the homoplasmic m.8993T>G pathologic mutation as a positive control. As negative controls, we used the isogenic m.8993T cybrid and, as it has been previously shown that the mtDNA genetic background plays an important role in modulating the biochemical defects and clinical outcome in NARP/MILS (9), we also generated cybrids from healthy individuals from close haplogroups to those of the patients. Thus, H, N1b and Uk cybrids were negative controls (C5, C2 and C3) for mutant H, W1* and U5b cybrids (P5, P2 and P3), respectively.
First, we confirmed that the cybrids harbored the nuclear background of osteosarcoma 143B206 rho0 cells and the mtDNA of the donor platelets (Supplementary Material, Fig. S2).
Compared with their haplogroup controls, mutant cybrids harboring the homoplasmic m.8993T>G, m.9029A>G and heteroplasmic (25 and 80%) m.9032T>C mutations presented a decrease in the endogenous oxygen consumption (Fig. 5A), but there were no differences in leaking respiration (Fig. 5A). Uncoupled respiration was not different from control cybrids in m.8993T>G and m.9029A>G cybrids but was significantly lower in m.9032T>C cybrids (Fig. 5A).
The mitochondrial inner membrane potential was significantly lower in the m.8993T>G cybrid than in its haplogroup control cybrid. However, this parameter was significantly higher in m.9029A>G and m.9032T>C (25%) cybrids than in their controls (Fig. 5B). This potential was not different from that of the haplogroup control in the m.9032T>C (80%) cybrid (Fig. 5B).
ATP levels in mutant cybrids were lower than those of their control cybrids (Fig. 5C), except for the m.9032T>C (25%) cybrid.
Reactive oxygen species (ROS) and Mn superoxide dismutase (SOD2) mRNA levels in mutant cybrids were also significantly higher than those of their control cybrids, except for the m.9032T>C (25%) cybrid (Fig. 5D).
Therefore, when compared with normal mtDNAs, the relocation of these mutated mtDNAs to rho0 cell lines transfers altered biochemical features.
M.9025G>A associated with type IV 3-methylglutaconic aciduria
The m.9025G>A/p.MT-ATP6:G167S mutation fulfills some but fails in other of the accepted pathogenicity criteria. Thus, this mutation was found in homoplasmy in healthy relatives and, except for Leber hereditary optic neuropathy (LHON) and mitochondrial deafness, this is not a common observation in disorders due to mtDNA mutations. However, the homoplasmic m.9176T>C/p.MT-ATP6:L217P mutation has also been associated with a very mild disease course in some individuals from two pedigrees suffering hereditary spastic paraplegia-like disorder (10) and Charcot-Marie-Tooth hereditary neuropathy (11). Homoplasmic levels of the m.9185T>C/p.MT-ATP6:L220P mutation were found in several slightly affected individuals from another pathologic pedigree (12). The m.9025G>A mutation has been found in 13 out of more than 25 000 (0.05%) apparently healthy individuals, but all of them from different mtDNA genetic backgrounds, which means that this is not an ancient polymorphism. Moreover, other MT-ATP6 pathologic mutations, such as m.8851T>C/p.MT-ATP6:W109R, m.9011C>T/p.MT-ATP6:A162V and m.9176T>C/p.MT-ATP6:L217P, have been found in homoplasmy in individuals from different population studies (Supplementary Material, Note S2). In all these examples, other factors such as nuclear modifying genes, the mtDNA genetic background or environmental influences may be involved in determining the severity of the phenotypic expression and the large phenotypic variability of these MT-ATP6 mutations (10). On the other hand, G167 is the second most conserved amino acid position in the p.MT-ATP6 subunit and bioinformatics studies predict that the glycine to serine amino acid substitution at p.MT-ATP6 167 position would be pathologic (6).
The glycine of the equivalent position in bacteria is also extremely conserved (86.2% of 1409 bacteria species). Many bacteria lacking a glycine in this position have a glycine at the equivalent position to amino acid 203 in TMH5 that interacts with amino acid at position 167 in TMH4 (2). Thus, glycine is frequently a member of the interacting amino acid pair. This pair forms part of an aqueous pocket extending from D61 of subunit c to the periplasmic surface (13). Glycine is the smallest amino acid and the bigger hydrophilic S167 might form a hydrogen bond with the E203 (Fig. 4D) and alter the conformational movements of this helix (14). Also supporting an important role of this amino acid position, the ATP synthase of extreme alkaliphiles has special features that are required for non-fermentative growth and OXPHOS at high pH (≥10.5). At such pH values, maintenance of a cytoplasmic pH that is acid in, lowers the total chemiosmotic driving force, and yet a robust H+-coupled OXPHOS proceeds optimally. The K180 amino acid of these alkaliphilic bacteria (amino acid position 167 in humans) is implicated in the pH-dependent restriction of the proton flux through the ATP synthase to and from the bulk phase (15). Moreover, different replacements at the equivalent position to human 167 have been modeled in E. coli and the ATP synthase activity was greatly inhibited and the aerobic growth suppressed. Replacing this amino acid by a polar one caused proton leak (14,16–18).
All these forward and against evidences on the pathogenicity of this mutation prevent its assignation as a pathologic change and make necessary more proofs.
The clinical phenotype and the increased levels of 3-methylglutaconic acid in our patient suggested a type IV 3-methylglutaconic aciduria. This is a mixed bag diagnostic but, frequently, a defect on the OXPHOS system has been found for these patients (Supplementary Material, References S3). Our results allowed us to rule out a defect of respiratory complexes, a coenzyme Q (CoQ) deficiency or an mtDNA deletion or depletion. The 3-methylglutaconic aciduria has sometimes been found in LS, but we were able to rule out several causes of LS such as pyruvate dehydrogenase (PDH) complex deficiency. LS caused by MT-ATP6 mutations frequently show normal activities for the electron transport chain complexes such as those found in our patient. Moreover, 3-methylglutaconic aciduria is frequently found in ATP synthase deficiency and, particularly interesting, it was found in a patient with the pathologic mutation m.8993T>G in the MT-ATP6 gene.
M.9029A>G associated with atypical LHON
The m.9029A>G/p.MT-ATP6:H168R mutation fulfills most of the accepted pathogenicity criteria. However, this mutation was also reported in two other individuals out of more than 25 000 including patients and controls. This observation is not rare for mtDNA mutations associated to hereditary optic neuropathies. Thus, the m.11778G>A and m.14484T>C transitions are frequently found in homoplasmy in normal females of LHON pedigrees. Moreover, two individuals without evidence of visual abnormalities and not being part of LHON families harbored these homoplasmic LHON pathogenic mutations (19). More importantly, one of these two individuals with the m.9029A>G mutation was suspected of suffering a mitochondrial disorder (20). On the other hand, and supporting its pathogenic role, this mutation is also the best candidate mutation to explain a pathologic phenotype (optic neuropathy) frequently found associated to pathologic mtDNA mutations. It was in homoplasmy in the proband and in heteroplasmy in the unaffected relatives. Despite the percentage of heteroplasmy was very high in the healthy relatives, maybe their higher mtDNA content may contribute to a moderate amount of normal mtDNA genotypes that compensate the pathologic mutation (21). The mutation changes a highly conserved amino acid in an important transmembrane alpha-helix which is part of the ATP synthase proton channel. Moreover, 3 (W109R, L156R and L217R) out of 12 previously described MT-ATP6 pathologic mutations incorporate an arginine. Finally, the m.9029A>G cybrid show OXPHOS dysfunction when compared with the cybrid harboring a normal mtDNA genotype from a close mtDNA haplogroup.
By using the Ag+-chemical labeling of cysteine side chains, introduced by site-directed mutagenesis, to map the aqueous accessible regions on subunit a (bacterial ortologue of mitochondrial p.MT-ATP6), it was found that the bacterial equivalent position to amino acid 168 might be part of an aqueous pathway to the intermembrane space (13). Moreover, it has been proposed that the proton motive force causes protonation of a network of residues in subunit a near the intermembrane surface, including amino acid 168 (22,23). Therefore, the H168R substitution, with very different pK, should have an important phenotypic effect. Interestingly, different amino acid replacements have been modeled at this equivalent position in E. coli (corresponding to position 219), showing that the proton translocation and the ATP synthase activity was greatly inhibited, and that the aerobic growth was also affected (16,24–27).
The main interest of m.9029A>G is that it is associated with an atypical LHON phenotype. Many patients are not fully evaluated early in the course of the disease or are referred electively for a subspecialty opinion years later. Properly evaluating an individual with suspected optic neuropathy years after the visual loss occurred may be difficult (28). Sometimes, rare optic neuropathies in patients harboring common LHON mutations can be confused with ischemic optic neuropathy (29,30), as it was the case of this patient. Moreover, LHON progression varies considerably, ranging from sudden and complete vision loss to progressive decline up to 2 years (31). However, the progressive loss of vision in our patient over 4 years, although highly unusual, has already been reported in an individual with the m.11778G>A pathologic mutation (32).
It is considered that LHON is caused by mutations in mtDNA-encoded complex I subunits. Mutations in MT-ND genes have been associated to LHON but also to LS. Moreover, some LHON mutations have been sometimes associated to LS (33). Therefore, it would be possible that MT-ATP6 mutations, usually associated to LS, were etiologic factors for LHON. A lowered rate of oxidative ATP synthesis appears to be the common biochemical background of pathogenic importance in all LHON complex I mutations (34). This background is also accompanied by an increased ROS production (35). The p.MT-ATP6 mutations that cause LS also reduce the ATP production (36) and increase ROS production (37). Very interestingly, the few histopathological studies reporting ocular findings in LS patients indicate that these individuals show features indistinguishable from LHON (38,39). Thus, maybe LHON is not a defect of mtDNA-encoded complex I subunits but of mtDNA-encoded OXPHOS subunits. Our results on ROS and ATP levels from cybrids harboring the m.9029A>G transition show that, in addition to MT-ND mutations, other mtDNA mutations could be associated to hereditary optic neuropathies. In fact, and although nowadays evidences are not too strong, other MT-ATP6 mutations, such as m.9101T>C/p.MT-ATP6:I192T (40) and m.9011C>T/p.MT-ATP6:A162V (41) have been postulated as etiologic factors for LHON patients. The m.9101T>C transition induced a proton leak that reduced OXPHOS efficiency (40) and this phenotype was transferred to cybrids (42). Moreover, cybrids harboring the m.9011C>T transition showed decreased calcium levels (43).
M.9032T>C associated with NARP
The m.9032T>C/p.MT-ATP6:L169P mutation fulfills all the accepted pathogenicity criteria. It was the best mtDNA candidate mutation to explain a pathologic phenotype (NARP) frequently found associated to pathologic mutations in the MT-ATP6 gene. It has not been found in normal individuals. It is present in heteroplasmy in the affected individuals, but it was not present in a healthy relative. It changes a highly conserved amino acid in an important transmembrane alpha-helix which is part of the ATP synthase proton channel. This mutation incorporates the same amino acid than other pathologic mutations of the same subunit, such as A105P, A155P, L156P, L170P, L217P, L220P and L222P, also associated to NARP/MILS phenotypes. A proline inside a transmembrane alpha-helix provokes the loss of hydrogen bonds in the peptide bonds, destabilizes the secondary structure and introduces a kink in the alpha-helix (44,45). Finally, cybrids harboring mutant mtDNA genotypes show OXPHOS dysfunction when compared with the cybrid harboring a normal mtDNA genotype from the same mtDNA haplogroup.
Our results on P3 patient (m.9032T>C) remark the importance of sequencing the entire MT-ATP6 gene in NARP/MILS patients. These analyses are providing new pathologic mutations, such as m.8618-8619insT (46), m.8839G>C (47) or m.8989G>C (48), as etiologic factors for these syndromes. Recent results also suggest that MT-ATP6 sequencing in other mitochondrial pathologies, no previously associated to MT-ATP6 mutations, unmask already known pathologic mutations as responsible for these clinical conditions, such as m.9185T>C in the Charcot-Marie-Tooth disease (49) or in the episodic weakness (50). Finally, these studies are also revealing new MT-ATP6 mutations in other mitochondrial disorders, such as m.8528T>C in infantile cardiomyopathy (51) or m.9025G>A and m.9029A>G in 3-methylglutaconic aciduria and hereditary optic neuropathy, respectively (this study). To increase the success rate in the finding of MT-ATP6 pathologic mutations, this gene should be sequenced at least in those patients suspected of suffering a mtDNA disorder and negatives for histochemical and biochemical analyses of respiratory chain defects. However, a positive marker is still required to improve the diagnostic efficiency of p.MT-ATP6 defects. Maybe, 3-methylglutaconic acid can help to point out MT-ATP6 mutations.
MATERIALS AND METHODS
Patients screening and case reports
The MT-ATP6 gene was analyzed by CSGE or by sequencing of DNA obtained from blood or muscle samples of 550 patients that were referred to our laboratory after a suspect of mitochondrial disease. All samples were collected with informed consent of the patients and family members and the study was approved by the Ethics Review Committees of the involved Hospitals and of the Government of Aragón (CEICA 11/2010). These patients included individuals with MILS or NARP, in which we had previously ruled out the presence of the m.8993 and m.9176 mutations by PCR-RFLP. Many others suffered from other pathologies, such as optic neuropathies. Three of these patients presented interesting mutations.
The first patient (P1), a girl, is the first child born from healthy non-consanguineous parents. Familial antecedents were uneventful. The pregnancy was normal and the serology negative. The delivery was at 40 weeks by caesarean section with meconium-stained amniotic fluid. Since birth and due to respiratory depression, bradycardia and hypotonia, mechanic ventilation was applied. Clinical examination showed microcephaly, severe axial hypotonia, low deep tendon reflexes and pin cataract on the left eye. Computerized tomography scan and magnetic resonance imaging (MRI) showed cerebellum and brainstem atrophy and diffuse increased signal on cerebellar white matter. Motor and sensitive neuropathy was detected with electromyography and nerve conduction velocity. The respiratory problems were aggravated in the following days with severe bradycardia that did not respond to reanimation procedures and lead her to death at 41 days of life. Necropsy studies confirmed severe atrophy of the cerebellum and brainstem. There was loss of granule and Purkinje cells and important astrocytosis of the cerebellar white matter. Clinical picture was of LS but ganglia were not affected on MRI or on the anatomopathologic study.
Lactate, pyruvate and alanine were elevated in blood. Urine organic acid analysis disclosed increased 3-methylglutaconic acid (34 mmol/mol creatinine; control range 5–15). In muscle, normalized respiratory complexes I+III (CI + III), II + III (CII + III) and IV (CIV) activities were not diminished. In the liver, CI + III was severely decreased, CII + III increased and CIV normal. PDH was normal as well (Supplementary Material, Table S1). Normal muscle CI + III and CII + III and elevated liver CII + III activities ruled out a CoQ deficit. Total and non-esterified carnitine levels in the muscle were 26 and 36%, respectively, of the lower limit for the control ranges (Supplementary Material, Table S1).
The second patient (P2) is a 38-year-old caucasian male who first came to our clinic complaining of a 4-year history of progressive visual loss. He was tentatively diagnosed of retinal ischemia 4 months prior to consultation. On examination, his best-corrected visual acuity was 20/80 in the right eye and 20/70 in the left eye. The anterior segments examination was within normal limits. The intraocular pressure was 12 mmHg in both eyes. Posterior segment evaluation showed the temporal half of the optic disc to be pale in both eyes. Visual evoked potentials showed a markedly increased latency in both eyes. Paracentral visual field defects were seen by the Goldmann visual field analysis. He was diagnosed as suffering an atypical LHON. His visual acuity and the anterior and posterior segment exam and visual fields were unchanged after four and a half years of oral multivitamin therapy on daily basis, especially B6 and B12. Family history was negative for visual loss or neurological disease.
The third patient (P3) is a male resulting from an uncomplicated pregnancy and delivery. He is the second child from non-consanguineous parents. Since birth he showed a mild motor delay. At 12 months, he was admitted to the hospital because of recurrent vomiting episodes and metabolic acidosis that persisted throughout his childhood. His height and weight remained steady at high percentile during development. He had learning difficulties at school. At the age of 7 years, an ophthalmologic evaluation revealed decreased visual acuity and mild bilateral macular pigmentary changes being diagnosed with pigmentary retinopathy. Currently, the patient is 16 years old and the most recent examination has revealed a mild ataxic gait, hypoactive reflexes and bilateral hypoacusia along with visual deficit. He has remained relatively stable for the last years. The laboratory tests, the screening for metabolic disorders including lactate and pyruvate levels in basal conditions, the thyroid and liver function, and the serum creatine kinase level have been repeatedly normal, and the karyotype was also normal. No lactate determination was recorded during any of the metabolic acidosis episodes in his pediatrics history. The electrophysiological study showed a subclinical sensory neuropathy and the brain MRI detects cerebellar vermis atrophy. His mother presented with progressive both unstable gait and bilateral sensorineural hearing loss since more than 15 years ago. She also presented bilateral concentric visual loss with a final diagnosis of retinitis pigmentosa. Examination disclosed a proximal weakness of lower limbs. The laboratory tests, electrophysiogical studies and brain MRI are normal. The grandmother presented visual disturbances without a definitive diagnosis. The clinical phenotype of both patients (mother and son) was highly suggested of NARP syndrome. New sequence accession numbers: GenBank KJ742709-KJ742715.
The MT-ATP6 gene was screened by CSGE and/or sequencing. The ATP5G1, ATP5G2 and ATP5G3 genes for the three isoforms of subunit c from the ATP synthase were screened by sequencing. The primers and conditions for these analyses and for the whole mtDNA sequencing are available upon request. The percentage of mutation was studied by the PCR-RFLP analysis (Supplementary Material, Table S2). For the hot-PCR, 2 μCi of [α−32P] dCTP (3000 Ci/mmol) was added and one additional cycle was performed. The mtDNA content was measured by using protocols described elsewhere (52).
The possibility of amplifying a nuclear sequence similar to the 1822 bp amplicon, between mtDNA positions 8018–9839, was checked in the mitochondrial pseudogenes database from mitomap (4). The 21 580 human, 3410 eukaryote and 1409 prokaryote p.MT-ATP6 sequences were obtained from GenBank (18 October 2013). The MutPred scores for particular amino acid substitutions were previously provided (6).
Modeling studies of p.MT-ATP6
The structure of E. coli subunit a (PDB 1C17) (53) was obtained with the RasMol2.6 program. The intramolecular interactions for the bacterial equivalent positions to p.MT-ATP6:G167 and p.MT-ATP6:S167 were obtained with Swiss-PDB-Viewer v.4.0.1. For this analysis, the E. coli equivalent position (histidine 245) in TMH5 was in silico mutated to glutamic.
Production, maintenance and functional investigations of transmitochondrial cell lines
To homogenize nuclear and environmental factors, we built transmitochondrial cell lines or cybrids with the osteosarcoma 143B rho0 nuclear background using patient and control platelets (54). The genetic fingerprints of the rho0 cell line and cybrids were performed with AmpFLSTR® Identifiler® PCR Amplification Kit (Life Technologies). These cybrids were grown in Dulbecco's modified eagle medium with no antibiotics and containing glucose (4.5 g/l), pyruvate (0.11 g/l) and fetal bovine serum (5%).
The analyses of oxygen consumption, ROS, SOD2 mRNA and ATP were performed in triplicate in three independent experiments, according to the previously described protocols (55). Mitochondrial ATP levels were measured using the CellTiter-Glow Luminiscent Cell Viability Assay (Promega) according to the manufacturer's instructions. Briefly, 20 000 cells/well were seeded 10–12 h before measurement. Then, cells were washed twice with PBS and incubated for 2 h in record solution with 5 mm 2-deoxy-d-glucose plus 1 mm pyruvate (56). Cells were lysed, and lysates were incubated with the luciferin/luciferase reagents. Samples were measured using a NovoStar MBG Labtech microplate luminometer, and the results referred to the protein quantity.
The determination of the mitochondrial inner membrane potential was also done in triplicate in three independent experiments using Mito-ID Membrane Potential Detection Kit (Enzo Life Sciences).
The statistical package StatView 6.0 was used to perform all the statistics. Data for mean and standard deviation are presented. The unpaired two-tailed t-test was used to compare parameters. P-values < 0.05 were considered statistically significant.
This work was supported by grants from the Instituto de Salud Carlos III (FIS-PI10/00662, PI11/01301, PI11/02350); and Asociación de Enfermos de Patología Mitocondrial (AEPMI). L.L. has a fellowship from Instituto de Salud Carlos III (FI12/00217). R.A. is supported by the ‘intensificación de la actividad investigadora’ program (ISCIII). CIBERER is an initiative of the ISCIII.
We would like to thank Santiago Morales for his help with the figures.
Conflict of Interest statement. None declared.