Disclosing the functional changes of two genetic alterations in a patient with Chronic Progressive External Ophthalmoplegia: Report of the novel mtDNA m.7486G>A variant

Highlights • Novel mt-tRNASer(UCN) (m.7486G>A) variant found in CPEO patient with 4,977 bp deletion.• The variant located in the anticodon loop meets the pathogenicity criteria.• Both genetic defects segregate with the biochemical phenotype in muscle.• Assembly impairment of MRC complexes was detected.• Mitochondrial translation defect and bioenergetic dysfunction were revealed.


Introduction
One of the most common presentations of mitochondrial DNA (mtDNA)-associated disease in adulthood is Chronic Progressive External Ophthalmoplegia (CPEO), which is characterized by progressive paralysis of the extraocular muscles (EOMs) leading to ptosis and impaired eye movement (ophthalmoplegia) [1].
Muscle biopsies of CPEO patients typically comprise subsarcolemmal accumulation of abnormal mitochondria known as ragged-red fibres (RRF) and a mosaic pattern of COX-deficient fibres showing abnormal COX activity. Mitochondrial dysfunction occurs in a tissue where the number of mutated mtDNA accumulates and exceeds a biochemical threshold (threshold effect), which has been shown to vary for different types of mutation, namely 50-80% for single, largescale mtDNA deletions and 70-95% for tRNA point mutations [24,25]. Currently, more than half of disease-related mtDNA point mutations have been reported within mt-tRNA genes that cause defective translation and, consequently, combined respiratory chain deficiency. Any deletion that eliminates an mt-tRNA gene causes the same impairment of overall mitochondrial-encoded proteins.
The aim of the present work was to: (i) clarify the molecular genetic defect and (ii) elucidate which cellular mechanisms are affected in a patient presenting with CPEO that harbours a novel mt-tRNA Ser(UCN) variant (m.7486G>A, ClinVar accession number: SCV000492500) in addition to the "common deletion".

Case report
The patient is a 62-year-old Portuguese Caucasian female followed at the Neurology Department of the Coimbra Hospital University Centre. The first clinical signs started at the age of 12 years old with slowly progressive ptosis of the right eyelid and later involvement of the other eyelid at the age of 45 years old. At this point, she was observed in the context of a corneal ulcer and complaints that she had "difficulties keeping her eyes open". The disease maintained a slow progression and she was diagnosed with CPEO at the age of 55 years old, with the neurological examination revealing severe bilateral eyelid ptosis, ophthalmoplegia and dysphonia. There was no reported family history of CPEO or any other mitochondrial disorder.
Biological samples (peripheral blood, skin and muscle biopsies) were collected from the patient investigated in the present study during diagnostic investigation. There were no samples available from maternal relatives.
The DNA samples of 200 adult healthy subjects of the same ethnic background were used as controls.
Informed consent was obtained from the participants, as recommended by the local Ethics Committee (CE-032/2014), following the Tenets of the Helsinki Declaration.
Quadruple immunofluorescence was performed on cryosectioned patient skeletal muscle (n = 1022) using antibodies detecting subunits of OXPHOS complexes: anti-NDUFB8 for Complex I (CI) and anti-COXI for Complex IV (CIV), as described previously [33]. Mitochondrial mass was quantified using an antibody to Porin and the myofibre boundaries were labelled with the antibody to Laminin.

Skin derived cultured fibroblasts
Fibroblasts from skin biopsies of the patient and three healthy Portuguese individuals (control group), without clinical evidence of mitochondrial disease, were grown in complete medium supplemented with 20% FBS (Gibco, Life Technologies) and antibiotics.

Genetic investigation in different tissues
Total DNA was extracted from several tissues including blood, dermal fibroblasts and muscle homogenate according to standard protocols [34,35].
Individual COX-positive and COX-deficient muscle fibres were isolated by laser microcapture, using a PALM Laser Capture Microdissection system, and lysed to obtain total cellular DNA, as previously described [36]. The stained sections were then used for single-cell molecular analyses.

Whole mitochondrial genome sequencing
The presence of sequence variants and rearrangements were detected by subjecting the patient's mitochondrial genome to next generation sequencing (NGS) in all available tissues, enriched by a single amplicon long-range PCR followed by massively parallel sequencing [37], using the HiSeq2000 platform (Illumina technology).
Haplogroup of patient was determined using the Haplogrep ® tool [38].

In silico analysis
The in silico analysis included the evolutionary conservation of the MT-TS1 gene (mt-tRNA Ser(UCN) ) from mtDNA of different species, according to the proposed consensus panel [39] using ClustalOmega|EBI ® [40]. The location of the m.7486G>A sequence variation in the cloverleaf structure of mt-tRNA Ser(UCN) was verified from Mamit-tRNA ® database [41].

Screening for mtDNA rearrangements
The presence of rearrangements in mtDNA was confirmed in DNA derived from muscle homogenate, using an established triplicate long range PCR approach for large-scale multiple and single mtDNA deletions, amplifying approximately 10, 13 and 16 kb of the mtDNA in three separate reactions. For~10 kb amplification PCR forward-F (m.6122-6139) and reverse-R (m.16133-16153) primers were used. For~13 kb amplification PCR F (m.13965-13984) and R (m.129-110) primers were used and for~16 kb amplification PCR F (m.1157-1167) and R (m.19-1) were employed. Cycling conditions were: 94°C, 2 min; 35 cycles of 94°C, 30 sec and 65°C, 16 min; 72°C 16 min. Amplifying PCR products were separated in 0.7% agarose gels.

Single fibre studies
The m.7486G>A mt-tRNA mutation load was assessed by pyrosequencing technology, in all available tissues and in individual COX-positive and COX-deficient muscle fibres. The PyroMark Assay Design Software v.2.0. (Qiagen) was used to design locus-specific PCR and sequencing primers for the m.7486G>A variant (biotinylated forward primer: m.7466-7485; reverse primer: m.7583-7600; sequence primer: m.7488-7502) and pyrosequencing was performed on the Pyromark Q24 platform, according to the manufacturer's protocol. Pyromark Q24 software was used to quantify the m.7486G>A heteroplasmy levels by directly comparing the peak heights of both wild-type and mutant nucleotides at this position [42].
The multiplex MTND1/MTND4 real-time PCR assay was performed using DNA from individual COX-deficient and COX-positive isolated muscle fibres and muscle homogenate and the mtDNA deletion level was calculated from the proportion of wild-type (MTND4) to total (MTND1) copy number by the established ΔΔCt method [43]. PCR amplification was completed in a 25 µl reaction in triplicate for each sample, with each plate containing a serial dilution of p7D1 plasmid for standard curve generation, as reported previously [44].

Mitochondrial respiratory chain (MRC) enzymatic activity and respiratory rate evaluation
Spectrophotometric determination of the catalytic activity of the MRC complexes and segments was performed as previously described [45].
Oxygen consumption rate (OCR) was measured in adherent fibroblasts with a XF24 Extracellular Flux Analyser (Seahorse Bioscience, Billerica, MA, USA), essentially as described by Zhang and co-workers [46].
Each of the three controls' and patient's fibroblasts were seeded in XF24 cell culture microplates (Seahorse Bioscience). OCR was measured under basal conditions, and after sequentially adding oligomycin, FCCP and rotenone plus antimycin A were also added. Results were expressed as pmol of O2 per minute per mg of protein, and allowed the evaluation of the following bioenergetics parameters: basal respiration (BR), maximum respiration (MR), spare respiratory capacity (SRC), ATP production capacity (APC) and proton leak (PL). All determinations were performed in 9-12 replicates for each sample.

Analysis of mitochondrial membrane potential
Alterations in mitochondrial membrane potential (ΔΨm) were determined using 1.5 µM of the cationic fluorescent probe rhodamine 123 (Molecular probes, Invitrogen) in Krebs medium, for 1 h, at 37°C. Basal fluorescence (λ = 540 nm for excitation and λ = 590 nm for emission) was measured using a Microplate Spectrofluorometer Gemini EM (Molecular Devices, USA) at 37°C, for 5 min, followed by the addition of 2 µg/ml oligomycin and 2 µM FCCP. Results were expressed as the difference between the basal fluorescence values and the increase in rhodamine 123 fluorescence levels following addition of oligomycin plus FCCP.

Relative quantification of MCR complexes
Samples were processed according to the protocol described elsewhere [47]. Briefly, following supplementation of samples with BN-sample buffer, the molecular weight marker (NativeMARK Unstained Protein Standard, Life Technologies) and 30 µg of samples were loaded into polyacrylamide gels and run at 80V at 4°C. The gel containing the proteins of interest was eletrotransferred to a PVDF membrane (Hybond P 0.5 µm, Amersham) for 2 h at 0.2A, at 4°C. Afterwards, membranes were incubated with monoclonal primary antibodies [anti-NDUFA9 for CI; anti-SDHA for Complex II (CII); anti-UQCRC2 for Complex III (CIII); anti-COX IV for CIV; and anti-ATP5A for Complex V (CV)]. Membranes were incubated with the anti-mouse HRP-conjugated secondary antibody solution. Subsequently, detection was carried out using a chemiluminescence substrate (Clarity Western ECL Substrate, Bio-Rad), through the ChemiDocTM XRS+ System (Bio-Rad). Protein band intensities were calculated by Quantity One® 1-D software (Bio-Rad) from at least 3 independent experiments. Relative semi-quantitation of each complex assembled was performed in comparison to CII levels.

Transmission electron microscopy (TEM)
Muscle fibres from a fresh biopsy were fixed with 4% glutaraldehyde in 0.2M sodium cacodylate buffer (pH 7.2) for 4 h. After rinsing twice in the same buffer, fibres were postfixed in 1% osmium tetroxide for 2 h. Following rinsing in buffer, samples were then dehydrated in a graded ethanol series (75-100%) before being embedded in Epoxy resin (Fluka Analytical). Semi-thin sections (2 µm) were obtained and stained with toluidine blue for light microscopy in order to identify the area of interest.
Fibroblasts were collected and centrifuged at 775 xg for 5 min. The supernatant was discarded and pellet cells were fixed with 2.5% glutaraldehyde in 0.1M sodium cacodylate buffer (pH 7.2) supplemented with 1 mM calcium chloride for 2 h. Following rinsing in the same buffer, post-fixation was performed using 1% osmium tetroxide for 1 h. After rinsing twice in buffer, and distilled water and, 1% aqueous uranyl acetate was added to the cells, for contrast enhancement during 1 h in the dark. After rinsing in distilled water, samples were dehydrated in a graded acetone series (30-100%), impregnated and embedded in Epoxy resin (Fluka Analytical).
Finally, for both preparations, ultrathin sections (70 nm) were mounted on copper grids (300 mesh) and stained with lead citrate 0.2%, for 7 min. Observations were carried out on a FEI-Tecnai G2 Spirit Bio Twin at 100 kV and images were acquired using the software AnalySIS 3.2.

Statistical analysis
Data were analysed using GraphPad Prism version 5.00 software for Windows, San Diego, California, USA. Normality tests were applied in order to assure the Gaussian distribution of the results. Statistical analysis of the patient's vs. controls' results was assessed by a Student's t-test (or nonparametric Mann-Whitney test).

Histochemistry and quadruple immunofluorescence presented evidences for mitochondrial dysfunction in muscle
Histological and histochemical examination of cryosectioned patient skeletal muscle revealed the presence of approximately 10% RRF as a result of mitochondria subsarcolemmal accumulation (Fig. 1A,B) and a significant proportion (40%-50%) of COX-deficient fibres (Fig. 1C-E). Assessment by quadruple immunofluorescence ( Fig. 2A) showed an equal down regulation (43%) of both CI and CIV levels in individual fibres (Fig. 2B); the MRC profile ( Fig. 2B) was similar to the profiles previously reported for single, largescale deletion [33].

Genetic investigation revealed the presence of the "common deletion" and a novel mt-tRNA Ser(UCN) variant at high levels in muscle
The entire mitochondrial genome analysis by NGS revealed previously reported polymorphisms and a novel heteroplasmic variant in the mt-tRNA Ser(UCN) (m. 7486G>A, ClinVar accession number: SCV000492500) (Supplementary material -Table S1) in the patient's muscle homogenate, fibroblasts and lymphocytes. Besides, a heteroplasmic single, large-scale deletion was detected in muscle homogenate and fibroblasts of patient, but not in blood, with the sequence breakpoints determined as 8482-13447, corresponding to the "common deletion".
The previous analysis determined that the patient belongs to the haplogroup H+16311.
The sequence variation was present at residue 32 in the anticodon loop of mt-tRNA Ser(UCN) affecting a 3-methylcytidine post-transcriptional position (Fig. 3A). The evolutionary conservation showed that cytosine is conserved throughout evolution (Fig. 3B). Moreover, the same alteration was not detected in the DNA samples of 200 controls screened by Sanger sequencing (data not shown).
Long-range PCR for~10 kb amplification presented a product size band with~10 kb (wild type) and also~5 kb, confirming the presence of the single, large-scale deletion (Fig. 3D), which is in agreement with the deletion size detected by NGS.

Assembly of MRC complexes was impaired
A statistically significant reduction in the amount of fullyassembled MRC complexes containing mtDNA-encoded proteins was observed in the patient's fibroblasts compared to controls (Fig. 4A)

Mitochondrial dysfunction detected in patient's cells: bioenergetics impairment, membrane depolarization and morphological changes
Assessment of MRC enzymatic activity (Supplementary material - Table S4) in muscle homogenate revealed a reduction in CIV activity (44.8% of controls). Also, CV activity was diminished in muscle and skin fibroblasts (69% and 66% of the controls' mean, respectively).
Moreover, the addition of the uncoupler FCCP in cells preexposed to oligomycin (to inhibit CV) caused a rise in rhodamine 123 fluorescence that was significantly lower in the patient's cells than in controls' cells (**p = 0.0084) (Fig. 5G), suggesting mitochondrial membrane depolarization in patient's fibroblasts.

Discussion
In the present report, we describe a patient with histological and histochemical changes in muscle that were indicative of OXPHOS dysfunction and mitochondrial disease. Both immunofluorescence and MRC evaluation revealed decreased levels and activity of multiple complexes in muscle sections and homogenate, respectively, likely due to defective mitochondrial translation. Deep genetic investigation allowed the identification of the "common deletion" in muscle and fibroblasts, and a novel mt-tRNA Ser(UCN) variation (m.7486G>A) in muscle, fibroblasts and blood. For single-fibre studies, we have used the same muscle fibres to quantify the levels of the mtDNA deletion and the tRNA variant. Both presented high levels within the isolated COX-deficient fibres and low levels within COX-positive fibres. Consequently, it was not possible to clarify the segregation of the tRNA variant alone, within the COX-deficient fibres, because the biochemical segregation was similar for both alterations. The percentage of deleted molecules is probably enough to cause a clinical phenotype, while the presence of a novel heteroplasmic variant in the anticodon loop of the mt-tRNA Ser(UCN) was also relevant.
There are numerous examples of disease-causing mutations in mt-tRNA genes and more than 20 variants, in 13 different mt-tRNA genes, have been reported in association with CPEO and/or myopathy [48]. Three of these have been reported in the MT-TS1 gene (encoding mt-tRNA Ser(UCN) ): m.7506G>A [49], m.7451A>T [50] and m.7458G>A [51] have all been associated with PEO. Nevertheless, the interpretation of pathogenicity for mtDNA variants is complex and challenging. The pathogenic variants in mt-tRNA genes may impair transcription termination and tRNA maturation, reduce the aminoacylation, abolish post-transcriptional modification of tRNA, decrease the binding to translation factors or the mitoribosome, alter the structure, perturb the stability, and disturb codon reading, ultimately leading to loss of function [52,53]. The novel variation m.7486G>A affects a highly conserved nucleotide and alters the first base of the anticodon loop adjacent to the anticodon stem (position 32 of mt-tRNA Ser(UCN) ), disturbing a 3-methylcytidine (m 3 C) post-transcriptional modification position [53][54][55][56][57]. Little is known about the function of this type of modification, although a role in accurate codon recognition and translation efficiency has been suggested [58,59]. Thus, the m.7486G>A variation could affect the conformation of the anticodon loop by creating a U32 pair with A38, that may disturb the interaction with mitoribosome, thus reducing the contact time for codon recognition. Similar defects in the conformation of the anticodon loop were suggested for the m.7480T>C variation, reported as pathogenic and causing mitochondrial myopathy, located in the anticodon loop of mt-tRNA Ser(UCN) at position 38 (opposite to 32) [60]. The novel variation herein reported was identified (about 90%) in COXdeficient fibres of the affected tissue (muscle), segregating with the biochemical phenotype but maintaining the mt-tRNA Ser(UCN) steady-state levels (Supplementary material - Figure S1). Accordingly, there is a possibility that the novel sequence variant might compromise mt-tRNA Ser(UCN) anticodon function by disrupting the correct binding of mitoribosome and/or the recognition of the amino acid sequence, leading to consequences in mitochondrial translational accuracy, without any effect on mt-tRNA Ser(UCN) stability.
According to the scoring criteria [61] for characterisation of the pathogenicity for the novel m.7486G>A variant, the score totals 11 points, out of 20, as follows: (1) the variant was heteroplasmic in different tissues (2 points); (2) the base is conserved throughout evolution (2 points); (3) there was a strong histochemical evidence of mitochondrial disease (2 points); (4) biochemical defects were detected in complexes I, III or IV (2 points); and (5) the variant segregated with the biochemical phenotype (3 points). Some of these criteria were certainly influenced by the presence of the "common deletion". However, the specific criterion for the segregation with the biochemical defect suggested that the sequence variation is "probably pathogenic".
The defective mitochondrial translation is predicted to be the primary consequence of mutations affecting mt-tRNA genes, leading to OXPHOS deficiency. The "common deletion" would be expected to have a similar detrimental effect on translation of all subunits encoded by mitochondrial genome. Therefore, the most likely genetic cause of CPEO in this patient is the "common deletion", even though the m.7486G>A sequence variation holds a huge pathogenic potential associated to the disease.
Concerning ultrastructural alterations, the most relevant finding in fibroblasts was the presence of MLB, which could be interpreted as an expression of cellular damage, suggesting alterations of autophagy, similarly to the hypothesis proposed by Signorini and colleagues in skin fibroblasts of patients affected with Rett Syndrome [31] and Thomas and co-workers in fibroblasts of patients presenting Spinocerebellar Ataxia, autosomal recessive 20 [62]. The accumulation of abnormal mitochondria in the patient's skeletal muscle, presenting paracrystalline inclusions and concentric cristae, reinforces the mitochondrial dysfunction observed in this tissue, as previously described for myopathy caused by single, large-scale mtDNA deletion or mt-tRNA mutations [63]. These results strongly suggest that defects in the proteolytic systems, namely mitophagy, responsible for the elimination of damaged or dysfunctional mitochondria, can be implicated in the disease.
In conclusion, we describe a patient presenting with CPEO who harbours two genetic alterations contributing to the clinical phenotype: a single, large-scale mtDNA deletion and a novel mt-tRNA sequence variant. Both alterations were detected by NGS. However, due to technical limitations, more specific analyses were performed in order to confirm and quantify the level of heteroplasmy of the mtDNA deletion and the tRNA variant. Since both alterations show high levels in muscle, appear to segregate with the biochemical defect and may promote inefficient mitochondrial translation, it is difficult to assign the isolated impact of the novel mt-tRNA sequence variation, thus demonstrating the inherent difficulties in proving the pathogenicity of novel or rare mtDNA variants. The "common deletion" is the most likely genetic cause. However, the potential pathogenic effect of the novel mt-tRNA Ser(UCN) sequence variation cannot be ignored, and it may be valuable in evaluating further cases. Similar patients (i.e. presenting with two mtDNA genetic defects) are reported in the literature; one patient presented with myopathy and autoimmune polyendocrinopathy due to a single mtDNA deletion co-existing with the pathogenic m.3243A>G variant [64]; another patient presented with limb-girdle myopathy associated with two, heteroplasmic mtDNA sequence variants [65]; together with the case we report here, these reinforce the idea that multiple mtDNA variants may act synergistically to influence clinical phenotype.