Partial tandem duplication of mtDNA–tRNA^Phe impairs mtDNA translation in late-onset mitochondrial myopathy

Abstract

An 80-year-old woman (PI) has been suffering of late onset progressive weakness and wasting of lower-limb muscles, accompanied by high creatine kinase levels in blood. A muscle biopsy, performed at 63 years, showed myopathic features with partial deficiency of cytochrome c oxidase. A second biopsy taken 7 years later confirmed the presence of a mitochondrial myopathy but also of vacuolar degeneration and other morphological features resembling inclusion body myopathy. Her 46-year-old daughter (PII) and 50-year-old son (PIII) are clinically normal, but the creatine kinase levels were moderately elevated and the EMG was consistently myopathic in both. Analysis of mitochondrial DNA sequence revealed in all three patients a novel, homoplasmic 15 bp tandem duplication adjacent to the 5′ end of mitochondrial tRNA^Phe gene, encompassing the first 11 nucleotides of this gene and the four terminal nucleotides of the adjacent D-loop region. Both mutant fibroblasts and cybrids showed low oxygen consumption rate, reduced mitochondrial protein synthesis, and decreased mitochondrial tRNA^Phe amount. These findings are consistent with an unconventional pathogenic mechanism causing the tandem duplication to interfere with the maturation of the mitochondrial tRNA^Phe transcript.

1. Introduction

Human mitochondrial DNA (mtDNA) is a circular double-stranded molecule of 16.6 kb, which contains 27 genes encoding 13 subunits of the respiratory chain (RC) complexes, 22 tRNA and 2 rRNA transcripts, and a ≈ 1 kb non-coding control region, the D-Loop, between the mt-tRNA^Pro and mt-tRNA^Phe genes. Whilst pathogenic single-base missense mutations are frequent in mtDNA, small rearrangements such as micro-deletions or micro-duplications are rare. In particular, duplications of a few base pairs in the D-Loop regulatory region have been reported as somatic changes related to aging, but none has been associated with a specific clinical syndrome. Likewise, no rearrangements of tRNA^Phe gene are known, whereas single-base missense changes have been associated with a wide spectrum of clinical manifestations, including tubulo-interstitial nephritis and stroke, MELAS, epileptic syndromes in childhood, myoclonic epilepsy and migraine, ataxia and deafness, progressive neurodegeneration with dementia and parkinsonism, myopathy with acute rhabdomyolysis, exercise intolerance associated with retinal dystrophy or progressive external ophthalmoplegia (PEO) (see http://www.MitoMap.org).

Here we report the identification of a 15 bp homoplasmic tandem duplication, between the mt-tRNA^Phe gene and the D-loop, determining reduced amount of tRNA^Phe gene transcript and reduced mtDNA translation products, associated with late-onset, isolated myopathy.

2. Case report

From 60 years of age the index patient, a woman now 80 years old, has been developing slowly progressive muscle weakness and hypotrophy of the lower limb muscles, including the ileopsoas, quadriceps, glutei and gastrocnemius. She complained neither of exercise intolerance nor myalgia, myoglobinuria or dysphagia. The family history is negative for neuromuscular disorders. The neurological examination consistently showed progressive, severe weakness of the lower limb muscles with bilateral hypotrophy of the quadriceps, and absence of the tendon reflexes, without fasciculations. She has neither PEO, nor pyramidal signs, nor ataxia. The CK levels were consistently elevated, about four-fold the upper normal limit. Electromyography (EMG) showed myopathic features, and a muscle tomography showed connective and adipose substitutions in the proximal muscles of the lower limb. A brain NMR was normal. The disease progressed up to the present condition, characterized by anserine gait, inability to stand up unaided from a chair, need of a cane to walk, and bilateral foot drop.

Although her 46-year-old daughter (PII) consistently showed moderate increase of CK levels and myopathic features at the EMG, she does not complain of fatigue or muscle weakness and her neurological examination is normal. A myopathic EMG was also found in her 50-year-old asymptomatic brother (PIII).

3. Materials and methods

3.1. Morphological and biochemical analysis

Quadriceps muscle and skin biopsies were obtained after written informed consent from PI and PII. Cryostat sections were analyzed using standard techniques [1,2].

Cybrids [3], skin fibroblasts and 143B cells were grown in D-MEM glucose medium. Biochemical assays of the individual MRC complexes and citrate synthase (CS) were carried out as described [4]. Oxygen consumption rate (OCR) measurements were taken every 4 min after a 4 min mix period in a XF 96-well apparatus (Seahorse Bioscience) on 15–20 × 10³/well cells [5].

Quantitative Real-time PCR was carried out on an ABI Prism 7000 apparatus (Applied Biosystems).

3.2. Molecular biology

Circularization and cDNA amplification of mtDNA regions were carried out as described elsewhere [6]. Briefly, RNA extracted from isolated mitochondria was circularized by incubation for 2 h at 37 °C with T4 RNA ligase. In order to obtain all the cDNA species containing the tRNA^Phe and 12S-rRNA genes, the circularized RNAs were then retrotranscribed and PCR-amplified using divergently oriented primers (available upon request). The PCR products were cloned in a TOPO-TA vector (Invitrogen) and sequenced.

For mt-tRNA Northern-blot analysis mitochondrial RNA was extracted in acidic conditions [7], electrophoresed through a 10% polyacrylamide/8 M urea gel, electroblotted onto a Zeta-Probe membrane (Bio-Rad) and hybridized with different [³²P]γ-ATP labeled probes [8].

For in organello aminoacylation assays [9], mitochondria were incubated for 15′ at 37 °C in 10 mM Glutamate, 2.5 mM Malate, 1 mg/ml BSA, 1 mM ADP, a mixture of all amino acids 10 μM each (except the labeled one) and 75 μCi of [³H]-labeled aminoacid.

For mitochondrial protein synthesis analysis [10], cybrids were labeled for 2 h with [³⁵S]-methionine-cysteine in the presence of 100 mg/ml emetine. Equal amounts of total cellular protein were loaded on a 20–15% exponential gradient SDS–polyacrylamide gel.

Autoradiography was performed with a PhosphorImager apparatus and densitometric analysis was carried out by using Quantity One software (BioRad).

4. Results

4.1. Muscle biopsy

A muscle biopsy of PI, performed at 63 years, demonstrated the presence of mild fiber size variability with no type grouping, scattered ragged-red fibers (RRF) and focal COX deficiency (Fig. 1a–c). A second biopsy, performed 7 years later, confirmed the previous findings, and, in addition, the presence of sparse rimmed vacuoles, type grouping, and fibro-adipose substitution, but neither acid-phosphatase positive fibers, nor inflammatory cells (Fig. 1d–f). A muscle biopsy performed in PII was morphologically normal (not shown).

Fig. 1.

Morphological analysis: PI muscle biopsies at 63 (A) and 70 (B) years of age. Panels a, d: H&E; note the fiber size variation and numerous centralized nuclei in a, with severe worsening in d. Panels b, e: Gomori trichrome; a pre-ragged red fiber is present in b, whereas several fibers contain rimmed vacuoles in e. Panel c, f: COX + SDH histochemistry; numerous SDH-positive, COX-depleted blue fibers are present in both samples. Magnification of panels A (a, b, c), referring to the first biopsy, is 20×. Magnification of panels B (d, e, f), referring to the second biopsy, is 40×, due to the scarcity of muscle tissue.

4.2. Mutation analysis

Southern-blot analysis of mtDNA showed neither depletion, nor multiple or single large-scale rearrangements. Sequencing analysis of the entire mtDNA form PI and PII muscle and fibroblasts, and PI-PII-PIII lymphocytes, revealed a homoplasmic tandem direct duplication in the D-loop encompassing the first 11 bp at the 5′ end of the mtDNA–tRNA^Phe gene and four nucleotides at the end of the D-loop (Fig. 2). This mutation has never been described previously and was absent in 100 consecutive control mtDNA samples. Analysis of the GNE gene, associated with inherited IBM, failed to detect mutations in PI and PII DNA samples.

Fig. 2.

MtDNA analysis: (A) Electropherogram and schematic representation of the 15 bp duplication. Arrows indicate the H1 and H2 promoters with the corresponding nucleotide positions. Underlined letters indicate the duplicated stretch. The same color code is used in the scheme and sequence letters to indicate the D-loop (ocre) and tRNA^Phe gene (lilac). (B) Scheme of the human mtDNA. Hooked arrows indicate the two promoters of the heavy strand (H₁, H₂) and that of the light strand (L₁). The mt-tRNAs are denominated as the corresponding aminoacid residues expressed in the single-letter code.

4.3. Biochemistry

Individual RC complex activities measured by standard spectrophotometric assays, and normalized to the activity of citrate synthase, showed isolated, partial COX deficiency in the second muscle biopsy of PI (Fig. 3A), consistent with the histochemical findings (Fig. 1), whereas fibroblasts from PI and PII, as well as cybrid derivatives, showed no defect (not shown). However, the oxygen consumption rate (OCR) was ≈50% decreased in both PII homoplasmic mutant fibroblasts, compared to wt fibroblasts, and transmitochondrial mutant cybrids, compared to 143B parental cells or a wild type cybrid cell line, taken as suitable controls (Fig. 3B).

Fig. 3.

Biochemical analysis: (A) RC complex activities normalized to CS and expressed as percentage of the mean control value. (B) Oxygen consumption rate (OCR) in cells. Note that for fibroblasts, values were measured as O₂ fmol/min/cell, whereas for cybrids values were measured as O₂ fmol/min normalized to the amount of mtDNA. Ctrl: control. Vertical bars indicate the standard deviations. Unpaired, two-tail Student’s t test: *p < 0.05; **p < 0.0005.

The mtDNA translation analysis showed ≈50% reduced amount of mtDNA-specific proteins in mutant cybrids vs. control cybrid cells (Fig. 4A).

Fig. 4.

In organello translation and tRNA analysis: (A) Mitochondrial protein synthesis. [³⁵S]-labeled mtDNA translation products. ND: subunits of complex I; CO: subunits of complex IV (COX); cyb: cytochrome b (complex III); ATPase: subunits of complex V (ATP synthase). Note the reduced intensity of the bands from mutant cybrids compared to 143B cells. (B) Northern-blot analysis of total and aminoacylated mt-tRNA transcripts. Top panel: Northern-blot analysis indicating the total amount of each mt-tRNA transcript. Bottom panel: aminoacylation assay. Analysis of the aminoacylation capacity in mutant and control cells performed using ³H-labeled amino acids. (C) Quantitative analysis of mt-RNA transcripts. Quantitative real-time PCR of mitochondrial H-strand transcripts relative to ND6 mRNA in mutant cybrids and 143B control cells.

4.4. Analysis of tRNA^Phe transcript and aminoacylation

Total and aminoacylated mt-tRNA^Phe transcript from isolated mt-RNA of homoplasmic mutant cybrids was reduced by ≈50% relative to that of 143B control cells, whereas mt-tRNA^Arg and mt-tRNA^Val transcripts were present in comparable amount in both (Fig. 4B). Likewise, the amount of other H-strand dependent transcripts located downstream from the mt-tRNA^Phe, including 12S rRNA, 16S rRNA, COI and cyt b, was measured by RT-PCR and compared to the values obtained from the L-strand dependent ND6 gene transcript. The values were comparable in the two cell types, with the exception of the 12S rRNA/ND6 ratio that was higher in the mutant than in control cells (Fig. 4C). Taken together, these data indicate no impairment of H-strand polycistronic transcription in mutant mitochondria.

A total of 20/20 cDNA clones, obtained by retrotranscription and circularization (see Section 3), all corresponded to the normal mt-tRNA^Phe species, suggesting that aberrant maturation of tRNA^Phe is a rare event leading to rapid degradation of the anomalous transcript.

5. Discussion

Our index patient had a relatively mild clinical presentation characterized by late-onset myopathy restricted to the lower girdle and limbs. Interestingly, whilst a first muscle biopsy at the very onset of disease showed a mitochondrial myopathy with scattered COX-negative fibers, a second biopsy performed years later showed the additional presence of sparse rimmed vacuoles, resembling those found in inclusion body myopathy (IBM). Molecular abnormalities of mtDNA have been reported in IBM, including mtDNA multiple deletions, partial mtDNA depletion and point mutations, but have been considered secondary, unspecific alterations [11]. However, the following evidence points to a primary mitochondrial myopathy in our patient. First, the timing of the morphological alterations demonstrates the onset of a mitochondrial myopathy several years before the appearance of the IBM-like alterations. Second, we excluded mutations in the GNE gene, which are the most common cause of autosomal recessive IBM [12], whereas the absence of skeletal alterations and cognitive impairment exclude the possibility of a mutation in the VCP gene [13]. Third, rimmed vacuoles can be secondary to a series of myopathic conditions, including LGMD1A, LGMD2J, oculopharyngeal myopathy, myofibrillary myopathy and myophosphorylase deficiency. Fourth, we found a novel, homoplasmic, peculiar mtDNA mutation that is clearly associated with reduced amount of total and aminoacylated tRNA^Phe in primary fibroblasts and transmitochondrial cybrids, leading to reduced mtDNA translation and mitochondrial respiration. Importantly, the mutation preceded the onset and progression of symptoms in the index patient and in her two children, both of whom had laboratory signs of muscle involvement, but no overt clinical symptoms, in agreement with the very late onset of the disease in the proband.

The clinical features suggest that the functional effects of the homoplasmic mtDNA duplication are mild. Accordingly, the spectrophotometric assays of individual RC complex activities showed only partial COX deficiency in skeletal muscle of PI, concordantly with the presence of scattered COX-defective muscle fibers. This uneven distribution of the COX reaction again suggests a dynamic disease mechanism, characterized by slow progression of the biochemical impairment over time, so that at the time of the biopsy, some, but not all, fibers had eventually overreached the threshold for histochemical detection. Although no obvious defect was found by spectrophotometric assays of individual RC complexes in cultured cells, the overall mitochondrial respiration was significantly reduced in both mutant fibroblasts and transmitochondrial cybrid derivatives, which contain homoplasmic mutant mtDNA on a heterologous nuclear background. This result unequivocally attributes the biochemical defect to the mtDNA gene complement derived from the mutant fibroblast cell line.

Human mtDNA encodes two rRNAs, 22 tRNAs, and 13 polypeptides. These genes are non-randomly distributed along the molecule: with a few exceptions, the tRNA genes are interspersed between the protein-encoding, and the rRNA, genes, with no intervening non-coding sequences [14,15]. Transcription of these genes is initiated from each mtDNA strand at three promoters, two for the heavy strand, H₁ and H₂, and one for the light strand, L₁. H₁ and L₁ are contained in the D-loop region, the major, non-coding “regulatory” segment of the mtDNA molecule [16,17], whereas H₂ is contained in the adjacent tRNA^Phe gene. Transcription gives rise to large polycistronic RNA precursors, in which mRNA and rRNA sequences are punctuated by tRNAs. According to the current model for RNA processing in human mitochondria, the mt-tRNAs act as cleavage signals for their own excision, and the concomitant release of mature rRNA and mRNA species [18,15]. The cleavage is carried out by two endonucleases, RNase P and RNase Z, operating at the 5′ and 3′ ends of the tRNA, respectively, and is followed by the addition of a CCA triplet at the 3′ end of each mt-tRNA, mediated by the tRNA-nucleotidyl transferase [19]. The novel mutation found in our family is a 15 bp perfect tandem duplication immediately adjacent and identical to the first 11 nucleotides of the mt-tRNA^Phe 5′ end, plus a 4 bp sequence of the D-loop. One possibility is that the duplication, which is in close proximity to the H₁ and H₂ promoters, can interfere with transcription initiation of the corresponding polycistronic pre-mRNAs. However, with the exception of the mt-tRNA^Phe transcript, other mtDNA-heavy strand transcripts measured in mutant cybrids were present in amount comparable to that of 143B cells, suggesting that the overall synthesis of the polycistronic H-strand dependent RNAs is unaffected by the mutation. A second possibility is that the duplication creates an aberrant 5′-end cleavage site for mt-tRNA^Phe, next to the normal one. This mechanism could interfere, for instance by steric hindrance, with the cleavage of the mt-tRNA^Phe at the normal 5′-end site, but it could also generate a structurally aberrant, non-functioning and highly unstable RNA species, that would rapidly be degraded. This is a common mechanism for pathogenic mutations in mt-tRNA genes [20]. Both mechanisms would result in decreased steady-state levels of mt-tRNA^Phe transcript, which is what we observed experimentally. Decreased availability of mt-tRNA^Phe can in turn explain the observed decrease of mitochondrial protein synthesis and eventually the defect of respiration documented in mutant mitochondria.

Taken together, these data prove a causative link between an unusual mtDNA mutation and a partial defect in mitochondrial respiration, through an unconventional mechanism involving mtDNA translation, and affecting either the maturation or the stability of mt-tRNA^Phe (or both). The clinical and morphological features of the disease raise the interesting, but still unproven, possibility that the mitochondrial dysfunction associated with the mtDNA mutation can eventually evolve into an IBM-like condition.