OUP user menu

Clinical manifestation and a new ISCU mutation in iron–sulphur cluster deficiency myopathy

Gittan Kollberg, Már Tulinius, Atle Melberg, Niklas Darin, Oluf Andersen, Daniel Holmgren, Anders Oldfors, Elisabeth Holme
DOI: http://dx.doi.org/10.1093/brain/awp152 2170-2179 First published online: 30 June 2009

Summary

Myopathy with deficiency of succinate dehydrogenase and aconitase is a recessively inherited disorder characterized by childhood-onset early fatigue, dyspnoea and palpitations on trivial exercise. The disease is non-progressive, but life-threatening episodes of widespread weakness, severe metabolic acidosis and rhabdomyolysis may occur. The disease has so far only been identified in northern Sweden. The clinical, histochemical and biochemical phenotype is very homogenous and the patients are homozygous for a deep intronic IVS5 + 382G>C splicing affecting mutation in ISCU, which encodes the differently spliced cytosolic and mitochondrial iron–sulphur cluster assembly protein IscU. Iron–sulphur cluster containing proteins are essential for iron homeostasis and respiratory chain function, with IscU being among the most conserved proteins in evolution. We identified a shared homozygous segment of only 405 000 base pair with the deep intronic mutation in eight patients with a phenotype consistent with the original description of the disease. Two other patients, two brothers, had an identical biochemical and histochemical phenotype which is probably pathognomonic for muscle iron–sulphur cluster deficiency, but they presented with a disease where the clinical phenotype was characterized by early onset of a slowly progressive severe muscle weakness, severe exercise intolerance and cardiomyopathy. The brothers were compound heterozygous for the deep intronic mutation and had a c.149 G>A missense mutation in exon 3 changing a completely conserved glycine residue to a glutamate. The missense mutation was inherited from their mother who was of Finnish descent. The intronic mutation affects mRNA splicing and results in inclusion of pseudoexons in most transcripts in muscle. The pseudoexon inclusion results in a change in the reading frame and appearance of a premature stop codon. In western blot analysis of protein extracts from fibroblasts, there was no pronounced reduction of IscU in any of the patients, but the analysis revealed that the species corresponding to mitochondrial IscU migrates slower than a species present only in whole cells. In protein extracted from isolated skeletal muscle mitochondria the western blot analysis revealed a severe deficiency of IscU in the homozygous patients and appearance of a faint new fraction that could represent a truncated protein. There was only a slight reduction of mitochondrial IscU in the compound heterozygotes, despite their severe phenotype, indicating that the IscU expressed in these patients is non-functional.

  • mitochondrial myopathy
  • Fe–S cluster deficiency
  • ISCU
  • aconitase deficiency
  • pseudoexon

Introduction

Myopathy with deficiency of succinate dehydrogenase and aconitase (MIM 255125) is caused by a deep intronic splicing affecting mutation in ISCU, encoding the iron–sulphur (Fe–S) cluster assembly protein IscU (Mochel et al., 2008; Olsson et al., 2008). The disease is recessively inherited and characterized by childhood-onset early fatigue, dyspnoea, hyperkinetic circulation with tachycardia and palpitations on trivial exercise. It is a non-progressive disorder, but life-threatening episodes with muscle weakness, tachycardia, metabolic acidosis and rhabdomyolysis may occur. The heart and visceral smooth muscles are not clinically affected. Muscle aconitase deficiency as well as deficiency of several Fe–S cluster containing proteins in the respiratory chain has been found (Hall et al., 1993; Haller et al., 1991). There has been a clustering of cases in northern Sweden where the disease was first recognized (Larsson et al., 1964; Linderholm et al., 1969), and it was recently given an alternative name; ‘Myopathy with exercise intolerance, Swedish type’.

Iron–sulphur cluster-containing enzymes catalyse some of the most basic redox transformations in nature and are essential constituents for generation of energy for metabolic purposes. The Fe–S cluster assembly process is fundamental and the IscU protein is among the most conserved proteins in evolution with homologues in all superkingdoms (Hwang et al., 1996). Most insights to the complex process of Fe–S clustering formation and transfer have come from studies on Escherichia coli and yeast (Johnson et al., 2005). The main components of the machinery are IscS, a cysteine desulphurase and sulphur donor, IscA, a Fe-chaperon, IscU, the scaffold protein on which the Fe–S cluster is built, and the chaperon proteins HscA and HscB, which interact with IscU to accomplish transfer of Fe–S clusters to apoproteins (Yang et al., 2006). Frataxin, the protein deficient in Friedreich ataxia, interacts together with another protein component (ISD11) with both the chaperons (HscA, HscB) and IscU, and has recently been proposed to be involved in Fe–S cluster insertion into apoproteins (Shan et al., 2007). In model systems, deletion of genes involved in Fe–S cluster assembly are lethal, thus showing their extreme importance for a functional biological system, which has recently been reviewed (Rouault and Tong, 2008). Ferrochelatase, a mitochondrial membrane-associated protein, is a Fe–S cluster protein in mammals and the final enzyme in the heme biosynthetic pathway. Thus, heme containing enzymes as well as Fe–S containing enzymes can be affected in patients with ISCU mutations due to defects in the mechanism of Fe–S cluster assembly.

Human ISCU encodes two isoforms of the Fe–S cluster assembly enzyme, IscU1 (NP_55116.1) and IscU2 (NP_998760.1), due to alternative splicing of the ISCU pre-mRNA. ISCU1 consists of six exons (1–6) and ISCU2 of five exons (1 and 3–6). Exons 3–6 are identical, and thus, the two proteins only differ in the N-terminus and localize to different cellular compartments. IscU1 localizes to the cytosol and the nucleus and IscU2 to mitochondria (Tong and Rouault, 2000).

We have investigated eight patients from six families descending from the northern part of Sweden with a phenotype consistent with the original description of myopathy with deficiency of succinate dehydrogenase and aconitase, and two brothers with the same biochemical and histochemical findings but with a considerably more severe clinical phenotype. We identified homozygosity for the deep intronic mutation in all individuals with the classical phenotype and further narrowed the shared homozygous segment to 405 kb. The two brothers were found to be heterozygous for the deep intronic mutation and a novel c.149 G>A missense mutation in exon 3.

Materials and Methods

Patients

The patients and their family members gave their written informed consent to participate in the study. The study was approved by the regional ethical committees in Gothenburg and Uppsala, Sweden.

Pedigrees of all families are shown in Fig. 1. Patients in Family A–F had childhood-onset early fatigue, dyspnoea and palpitation on trivial exercise. Patients A–E had no signs of cardiac involvement with normal findings on ECG and echocardiographic examinations. Patient F1 is hypertensive and has during her seventies developed a chronic aortic insufficiency with hypertrophy and dilatation of the left ventricle. Clinical features and muscle mitochondrial functions of patients in Family A–F are described in Table 1. A detailed description of the two patients in family G is given below and summarized in Table 2.

Figure 1

Pedigrees of the investigated families. Filled black symbols in each family represent persons affected by Fe–S cluster deficiency myopathy. Patients in families A–F suffer from the originally described form of ‘Myopathy with deficiency of succinate dehydrogenase and aconitase’, whereas Patients G1 and G2 have a different and more severe phenotype. Shaded symbols represent individuals that have not been genetically investigated. Circles represent female family members and squares represent males. Crossed symbols show deceased individuals.

View this table:
Table 1

Clinical features and muscle mitochondrial findings of patients with the classical form of myopathy with deficiency of succinate dehydrogenase and aconitase

PatientsNormal
ABCDE1E2F1aF2asubjects range
Clinical features
    Preschool age at onset++++++++
    Exercise intolerance++++++++
    Muscle weakness+++
    Recurrent myoglobinuria++
Polarography (nmol/min/mg)
    Pyruvate + malate1222251113NDNDND94–178
    Glutamate + malate4884816050NDNDND92–203
    Palmitoyl-carnitine + malate3640282029NDNDND28–80
    Succinate + rotenone3450354133NDNDND67–174
    Ascorbate + TMPD306361269215378NDNDND276–443
Respiratory chain enzymes
    NADH–ferricyanide reductase (nmol/min/mg)21602600290028302100NDNDND4970–9450
    Succinate–cytochrome c reductase (nmol/min/mg)3057506347NDNDND179–466
    Cytochrome oxidase (1/min/mg)4.16.28.57.36.6NDNDND7.1–20
Aconitase (µmol/min/mg)0.0880.0800.0600.0080.051NDNDND0.34–0.58
Citrate synthase (μmol/min/mg)5.63.33.82.73.3NDNDND2.3–3.2
Succinate dehydrogenase (nmol/min/mg)4825ND4116NDNDND97–324
Mitochondrial myopathy with succinate dehydrogenase deficiency (histochemistry)+++++++ND
Age, at muscle biopsy (years)1311525192267ND
  • ND = Not determined.

  • a The patients belong to family D in the original report on the disorder (Larsson et al., 1964; Linderholm et al., 1969).

View this table:
Table 2

Clinical features and muscle mitochondrial findings in Patients G1 and G2

PatientsNormal subjects
G1G2range
Clinical features
    Preschool age at onset++
    Exercise intolerance++
    Muscle weakness++
    Recurrent myoglobinuria
Polarography (nmol/min/mg)
    Pyruvate + malate231894–178
    Glutamate + malate582692–203
    Palmitoyl-carnitine + malate342828–80
    Succinate + rotenone452767–174
    Ascorbate + TMPD223149276–443
Respiratory chain enzymes
    NADH–ferricyanide reductase  (nmol/min/mg)233319484970–9450
    Succinate–cytochrome c reductase  (nmol/min/mg)4830179–466
    Cytochrome oxidase (1/min/mg)3.82.67.1–20
Aconitase (µmol/min/mg)0.067ND0.34–0.58
Citrate synthase (μmol/min/mg)4.42.3–3.2
Succinate dehydrogenase (nmol/min/mg)38ND97–324
Mitochondrial myopathy with succinate dehydrogenase deficiency (histochemistry)++
Age, at muscle biopsy (years)2113

Patient G1, a boy born in 1972, was the first child of healthy unrelated parents. Patient G2, born in 1976, is his younger brother. Four younger siblings born in 1978, 1981 (twins) and 1985 are healthy. There are no other cases of neuromuscular or cardiac disease in the family.

Patient G1 was born after normal pregnancy and delivery with a birth weight of 3700 g. In the newborn period he had sucking difficulties. He developed tracheomalacia with stridorous episodes, which recurred until 5 years of age. From an early age he failed to gain weight adequately. Mental and fine motor development were normal. He sat at age 8 months and walked at the age of 14 months. By 2 years, he had a waddling gait. He never learned to run. He tired easily and complained of fatigue when walking. At 5 years of age he had difficulties rising from the floor (positive Gower's sign) and climbing stairs. During childhood muscle wasting was noticed but otherwise the condition remained the same.

At the age of 17 years, he was admitted for investigation with special respect to mitochondrial disease. He could walk shorter distances (100 m) without support, but otherwise he used a wheel-chair. He had no exercise-induced muscle pains. His height was 176 cm (−0.5 SD) and weight was 36.7 kg (−4 SD). Muscles innervated by the cranial nerves were normal. Myopia was noticed but otherwise there were normal findings on ophthalmologic examination. The spine was kyphotic and there was general muscle weakness and wasting, more prominent in proximal than distal muscles. Electromyography showed myopathic potentials in proximal muscle groups of the upper extremities. Nerve conduction velocities were normal.

An Ergometer exercise test was performed with a maximum workload of 20 Watt (0.5 Watt/kg). He exercised for only 3 min and attained a maximal cardiac frequency of 190 beats per minute. The lactate concentration was pathological at rest, 5.6 mmol/l (reference interval <1.7 mmol/l) and increased to a maximal concentration of 13.6 mmol/l 10 min after he stopped exercise. There were no ECG changes during the exercise test. Cardiomyopathy was diagnosed at the age of 7.8 years. At the time of diagnosis a soft systolic murmur was heard. The patient showed no clinical signs of heart failure. ECG presented a left-oriented QRS-axis, left ventricular hypertrophy and shallow T-waves in leads V5–V6. Echocardiography revealed a normal left ventricular systolic function and moderate left ventricular hypertrophy without outflow tract obstruction (Holmgren et al., 2003).

Through the years the ECG and echocardiographic examinations have shown a stable pattern with a tendency towards a reduction of the left ventricular hypertrophy. At the latest check-up, at the age of 23 years, the patient had no clinical signs of heart failure. His weight was 45 kg. ECG showed sinus tachycardia (120 beats per minute). A mild hypertrophy of the left ventricle without obstruction was documented on echocardiographic, intra ventricular septum in diastole was 9.0 mm (reference value 7.0–10.1 mm) and the left ventricle posterior wall in diastole was 10.0 mm (7.0–9.6 mm). The left ventricular size was within the normal range, 43.0 mm (37.4–47.6 mm) with good systolic function. At the age of 23 years he had neck flexor weakness, and generalized muscle wastage, which was more pronounced proximally. Muscle strength was grade 4 out of 5 on the graded Medical Research Council (MRC) scale in the limb girdle muscles and limb muscles.

Patient G2: This boy, born in 1976 after a normal pregnancy and delivery, was the second child of six siblings. The birth weight was 3670 g. The newborn period was normal. Mental and fine motor development were normal. He sat at age 8 months and walked at 13 months. From age two years the parents noticed a waddling gait and that he fatigued easily. He could only walk short distances and could never run.

At the age of 13 years, he was admitted for investigation with special respect to mitochondrial disease. His height was 144 cm (−2 SD) and weight was 27.5 kg (−3 SD). He mostly used a wheel-chair, but was able to walk shorter distances (30 m) without support. A severe muscle weakness with wasting was noticed which was more prominent in the proximal than the distal muscles. Electromyography showed myopathic potentials in proximal muscle groups of the upper and lower extremities. Nerve conduction velocities were normal.

An Ergometer exercise test was performed with a maximum workload of 10 Watt. The patient exercised for 5 min and attained a maximal cardiac frequency of 190 beats per minute. The lactate concentration was pathological at rest, 6.1 mmol/l (reference interval <1.7 mmol/l) and increased to a maximum concentration of 14.5 mmol/l after he stopped exercise. Cardiomyopathy was diagnosed at the age of 3.5 years. At the time of diagnosis a discrete systolic murmur was heard but the patient had no clinical signs of heart failure. ECG showed a left oriented QRS-axis, left ventricular hypertrophy and negative T-waves in leads V5–V6. Echo revealed a normal left ventricular systolic function and moderate left ventricular hypertrophy without obstruction; maximal flow velocity in the outflow tract <1.5 m/s (Holmgren et al., 2003). From 8 years of age he was treated with propranolol for several years. Through the years, the ECG and echocardiographic examinations have shown a stable pattern. At the latest check-up, at the age of 31 years, the patient had no clinical signs of heart failure. ECG showed sinus tachycardia (110 beats per minute). A mild hypertrophy of the left ventricle without obstruction was documented on echocardiography, intra-ventricular septum in diastole was 11.0 mm (7.0–10.0 mm), and left ventricle posterior wall in diastole was 12.0 mm (7.0–9.6 mm). Left ventricular dimension was normal and the left ventricle inner diameter in diastole was 45.0 mm (37.4–47.6 mm), with a good systolic function and fractional shortening of 42% (>26%). He developed a severe progressive scoliosis and has been wheelchair bound since his late teens. At age 31 years he is not able to stand upright, has neck flexor weakness, and muscle strength MRC grades 2 and 3 in the shoulder and pelvic girdle, respectively, and grade 4 in more distal limb muscles. He has respiratory insufficiency and requires nocturnal assisted ventilation. Patient G2 is more severely affected than Patient G1. The disease has been slowly progressive in both.

Morphologic and biochemical analysis

The patients underwent open muscle biopsies from the vastus lateralis or the deltoid muscle. Mitochondria were isolated from fresh muscle tissue by homogenization and differential centrifugation as described (Bookelman et al., 1978). Polarographic and spectrophotometric analysis of the respiratory chain, morphologic and histochemical analysis of fresh-frozen muscle tissue, and electron microscopy were performed as described previously (Tulinius et al., 1991). The polarography was performed immediately after preparation of mitochondria. Enzymatic and western blot analyses were performed on fresh-frozen and thawed aliquots of the mitochondrial preparation. Histochemical staining of iron resulting in Prussian Blue colour was performed as described (Stevens, 1990). Aconitase activity was measured essentially as described (Fansler and Lowenstein, 1969). Succinate dehydrogenase activity was measured with dichloro-phenol-indo-phenol as electron acceptor as described (Hoppel and Cooper, 1968).

Cell cultures

Skin biopsies were obtained from the patients and immediately transferred to sterile culture medium. Skin fibroblasts were grown in Eagle's minimum essential medium (Invitrogen) supplemented with 9% foetal calf serum. In the downstream applications, 2 × 106 cells were used for RNA extraction, 4 × 106 cells for total protein extraction and 8 × 106 to 10 × 106 cells for isolation of mitochondria by digitonin-permeabilization and differential centrifugation.

Genotyping, DNA and RNA isolation, PCR, RT–PCR and sequencing analysis

Single nucleotide polymorphism genotyping using Affymetrix GeneChip® human mapping 250 K array was performed by AROS Applied Biotechnology. The search was performed in three families consisting of patients (A, B and C), parents and one non-affected sibling, and in three single unrelated patients (D, E1 and F1).

Total DNA was extracted from muscle tissue or blood using the QIAamp DNA Mini Kit (Qiagen). Total RNA was extracted from 10 mg of fresh-frozen skeletal muscle tissue or cultured skin fibroblasts using the AllPrep DNA/RNA Mini kit (Qiagen). RT-PCR was performed with the One-step RT-PCR kit (Qiagen) with primers amplifying exon 4–6 of ISCU cDNA. For genomic DNA, primers amplifying exons and exon–intron boundaries of ISCU were used, as well as primers designed to amplify intron 5. PCR products were separated on agarose gels stained with GelStar® and visualized on a Dark Reader Blue light transilluminator. Sequencing analysis was performed using an ABI PRISM® 3100 Genetic Analyzer and the BigDye Terminator v.1.1 Cycle Sequencing Kit (Applied Biosystems). For primer sequences and PCR conditions see Supplementary Table.

Cloning

The cloning procedure was performed using the Original TA Cloning® Kit (Invitrogen). Gelpurified cDNA fragments of aberrant size were ligated into a pCR®2.1 vector and transformed into E. coli. Bacterial colonies were grown on selective Ampicillin plates for 18 h. White colonies were resuspended in aliquots of ReddyMix™ PCR Master Mix with primers specific for the inserted fragment. Purified PCR products representing the inserts were sequenced and analysed.

Protein blotting

Total proteins were extracted from isolated skeletal muscle mitochondria, from cultured skin fibroblasts and from the non soluble fraction of digitonin-permeabilized cultured skin fibroblasts. The pellets of mitochondria, cell fractions or cells were resuspended in LDS-NuPage sample buffer (Invitrogen) supplemented with 5% dithiothreitol. Total lysis was performed by sonication for 8 min and heating for 10 min at 75°C. Cell-debris was removed by centrifugation at 13 000 rpm for 5 min. The proteins were loaded and separated on a 4–12% Bis–Tris Gel (Invitrogen) and transferred to a nitrocellulose membrane by semidry electro blotting using a Trans-Blot SD Semi-Dry Transfer Cell (Bio-Rad). The membrane was blocked with 5% blotting-grade milk (Bio-Rad) and incubated with rabbit-anti-human IscU1/2 (FL-142) antibodies (Santa Cruz Biotechnology Inc) at 4°C over night and was subsequently incubated with secondary goat-anti-rabbit IgG horseradish peroxidase-conjugated antibodies for one hour at room temperature. The blot was developed using SuperSignal® enhanced chemilumeniscent substrate (Pierce).

Accession numbers

Human ISCU mRNA transcript variant 1, NM_014301; human ISCU mRNA transcript variant 2, NM_213595; Human ISCU gene contig reference NC_000012 REGION: 107480512..107487273, Contig AC008119.6.1.173153; UniProt entry Q9H1K1.

URLs

NCBI: http://ncbi.nlm.nih.gov/; Uniprot: www.expasy.org; mFold: http://frontend.bioinfo.rpi.edu/applications/mfold/cgi-bin/rna-form1.cgi; Analyzer Splice Tool: http://ast.bioinfo.tau.ac.il/SpliceSiteFrame.htm.

Results

Morphologic and biochemical analysis

Morphologic and histochemical investigation showed mitochondrial myopathy with succinate dehydrogenase and cytochrome c oxidase (COX) deficiency in all patients (Fig. 2). A consistent finding was a patchy distribution of weak enzyme activity, which was best visualized by COX staining (Fig. 2C and D). This specific pattern of succinate dehydrogenase and COX deficiency is probably a pathognomonic finding of Fe–S cluster deficiency myopathy. The histochemical staining of iron resulting in Prussian Blue colour revealed accumulation of iron in the patients’ muscle tissue (Fig. 3A). Electron microscopic visualization revealed accumulated dense particles within the mitochondrial matrix (Fig. 3B). As opposed to the patients in families A–F, the two patients in family G, G1 and G2, showed more pronounced signs of myopathy. They had an increased number of fibres with internalized nuclei and a slight focal increase of the interstitial connective tissue. The biochemical analysis verified succinate dehydrogenase, aconitase and complex I deficiency in isolated muscle mitochondria in patients with the classical form of the disease (Table 1) and in the two affected brothers from family G (Table 2).

Figure 2

Consecutive sections of muscle biopsy of Patient G2. (A) Abnormal accumulation of mitochondria in the muscle fibres as demonstrated by staining of NADH–tetrazolium reductase. (B) Reduced histochemical activity of succinate dehydrogenase. Inset is a normal control. (C and D) Patchy reduction of histochemical COX activity. Some muscle fibres show reduced COX activity in parts of the fibres. Bars correspond to 100 µm.

Figure 3

(A) Muscle biopsy from Patient E1 demonstrating punctuate accumulation of iron in the muscle fibres by Prussian blue reaction. Bar corresponds to 40 µm. (B) Electron micrographs illustrating four examples of the frequent dense intra-mitochondrial inclusions apparently corresponding to iron accumulation shown in A. Bar corresponds to 0.2 µm.

Genotyping, PCR, RT–PCR and sequencing analysis

The single nucleotide polymorphism genotyping resulted in identification of a single homozygous region on chromosome 12 spanning ∼405 kb, where the patients were identical in all positions (Supplementary Fig. 1). This region encompasses six genes of which ISCU recently was shown to be the affected gene (Mochel et al., 2008; Olsson et al., 2008). Direct sequencing of exons and flanking intronic sequences of ISCU revealed four single nucleotide polymorphisms in exon 1 (rs10778647, rs10778648, rs11837563 and rs2287555) but no other variants compared to the reference sequence. RT–PCR of ISCU cDNA from muscle tissue and cultured skin fibroblasts, spanning exons 4–6, followed by gel electrophoresis, revealed a product approximately 100 base pair larger in addition to the expected transcript of 388 base pair in the patients but not in the control subjects (Fig. 4A). Sequencing analysis revealed a mixture of sequences immediately after exon 5 (Fig. 4B). Subtraction of the expected wild-type sequence and a BLAST search identified an intronic sequence starting 390 base pair downstream exon 5. Cloning followed by sequencing of vectors containing the odd-sized transcript revealed two species, 388 + 86 and 388 + 100 base pair in size. The 86 and 100 base pair inserts matched intronic sequences between exons 5 and 6 (Fig. 4C). The inserts i.e. pseudoexons, were flanked by a 3′ acceptor splice site, which they shared, and a 5′ donor splice site that differed between the two species (Fig. 4E). PCR and sequencing of intron 5 revealed a deep intronic G>C mutation at position 382 downstream exon 5 (IVS5 + 382 G>C, position –8 from the inserted sequence) (Fig. 4D). Homozygosity for the variant IVS5 + 382 G > C was found in all affected individuals in families A–F. Parents were heterozygotes whilst unaffected siblings were either heterozygotes or homozygous wild type. The calculated score for the acceptor splice site of the identified pseudoexons in our patients increased from 87.12% in the wild-type allele to 90.21% in the mutant allele (Shapiro and Senapathy, 1987), predicting that this was a pre-existing but normally silent acceptor splice site. The scores calculated for the activated cryptic donor splice sites were 61.84% for the 100 base pair insertion and 63.35% for the 86 base pair insertion. Screening by sequencing for the deep intronic mutation in 100 control subjects that originate from the same area of northern Sweden as the patients, revealed one carrier.

Figure 4

Genetic analyses. (A) RT–PCR products spanning exons 4–6 showing a fragment of abnormal size in patients (lanes 1, 2 and 4) and its absence in control samples (lanes 3 and 5). RNA was extracted from muscle tissue of Patient D (lane 1), Patient G2 (lane 2), control muscle tissue (lane 3), cultured skin fibroblasts from Patient D (lane 4) and cultured skin fibroblasts from a normal control subject (lane 5). (B) Direct sequencing revealed a mixture of sequences immediately after exon 5. Cloning followed by direct sequencing of inserts revealed sequences of 388, 474 and 488 base pair in length representing fragments with the wild-type sequence, an 86 and 100-nt insertion, respectively. (C) The intronic sequences of the pseodoexons (in capital letters) between exons 5 and 6 (in lower case letters). The insertion of either of the pseudoexons results in premature translation termination due to the presence of a stop codon, included in both insertions (in bold). (D) Identification of a homozygous IVS5 + 382 G > C mutation in genomic DNA extracted from Patient C (upper panel), heterozygosity in one of the parents (middle panel) and homozygous wild type in a control subject (lower panel). The asterisk indicates the starting position of the pseudoexons in the patient's genomic DNA. (E) Schematic illustration of aberrant splicing, including the pseudoexons, compared to normal splicing of ISCU pre-mRNA. The arrow indicates the position for the mutation. (F) Identification of a heterozygous c. 149 G > A mutation in exon 3 in genomic DNA extracted from Patient G1. The mutation gives rise to a predicted G50E amino-acid change.

Patients G1 and G2 were compound heterozygous for the IVS5 + 382 G>C mutation and a c.149 G>A missense mutation in exon 3 changing a highly conserved glycine residue to a glutamate (Fig. 4F and Supplementary Fig. 2). Their father was not available for genetic testing but their mother, who was of Finnish descent, carried the missense mutation. Two healthy sisters were homozygous wild type in both alleles whereas one younger brother was heterozygous for the missense mutation but homozygous wild type for the intronic variant. Their grandmother of the maternal line was not a carrier of the missense mutation, which indicates either a de novo mutation in their mother, or that their maternal grandfather, who was not available for genetic testing, carried the mutation. The mutation was not found in 100 Finnish control individuals.

Protein blotting

In protein extracts from isolated skeletal muscle mitochondria, the western blot analysis revealed a severe deficiency in patients homozygous for the intronic mutation, whereas the reduction of IscU protein in the compound heterozygous patients was less severe (Fig. 5A). In addition to the severe reduction, the immunoblot analysis revealed a slightly faster migrating species in the muscle mitochondrial protein extracts from patients. In the lanes loaded with mitochondrial proteins from patients G1 and G2, the faster migrating species was barely detectable. In the protein fraction from cultured skin fibroblasts there was no significant difference between patients and controls, neither in the total cell fraction, nor in the non soluble fraction of digitonin-treated cells (Fig. 5B). A comparison of the migration pattern of muscle mitochondrial extracts and fibroblast extracts revealed that the minor fraction in fibroblast extracts migrated as the muscle mitochondrial fraction and the major fraction in fibroblast was only present in proteins from whole fibroblasts (Fig. 6).

Figure 5

Protein analyses. (A) Western blot analysis of mitochondrial protein extracts from skeletal muscle tissue with polyclonal rabbit anti human IscU-antibodies showing severely reduced levels of the IscU protein in patients B and C compared with a normal control sample, whereas in patients G1 and G2 the signal was only slightly reduced. Ten microgram of mitochondrial protein was loaded in each lane. In the lanes with mitochondrial protein extract from patients, there is a band beneath the predicted IscU protein that was absent in the control lane. (B) Western blot analysis of proteins extracted from cultured skin fibroblasts, whole cells (lanes 1, 3, 5 and 7) and the non soluble fraction of digitonin-treated cells (lanes 2, 4, 6 and 8) showing that there was no significant difference between patients and a normal control sample. Observe that the faster migrating species present in whole cells disappears in the protein fraction from digitonin-treated cells. A loading control is mounted below the blot. (C) Translation of the mRNAs with pseudoexonic inclusions is predicted to give a frame-shift and premature truncation of IscU.

Figure 6

Comparison of the migration pattern of species immunoreactive to IscU antibodies in the different investigated protein fractions. Lanes 1 and 3, proteins extracted from the non-soluble fraction of digitonin-permeabelised cultured skin fibroblasts. Lanes 2 and 4, proteins extracted from intact fibroblasts. Lane 5, proteins extracted from isolated skeletal muscle mitochondria. Lane 1, 2 and 5, a normal control subject. Lanes 3 and 4, patient B.

Discussion

The originally described form of myopathy with deficiency of succinate dehydrogenase and aconitase caused by the deep intronic splicing affecting mutation has so far only been identified in patients from northern Sweden (Mochel et al., 2008; Olsson et al., 2008). An extensive genealogical study, tracing ancestors of nine families with 19 affected individuals for 300 years, failed to identify a common link between families (Drugge et al., 1995). The shared homozygous segment surrounding ISCU in our patients was only 405 kb in length, which indicates that the mutation is ancient in the studied population. Patients homozygous for the deep intronic mutation suffer from a pure myopathy and the phenotype of the patients is very homogenous.

Studies of a certain disease associated with one specific mutation in a population, generally lead to the finding of both a wider clinical spectrum than originally described and occasional patients with novel mutations, most often in compound with the common mutation.

In this study, the divergent and more severe clinical phenotype of the two brothers was associated with a novel ISCU mutation in compound with the common mutation.

Both brothers are severely disabled. They have progressive muscle weakness and muscle wasting and they presented with signs of muscle weakness at two years of age. Although early fatigue was noted on exercise in childhood the complaint never was that of tachycardia, palpitations and breathlessness on trivial exercise, which is so typical for the classical form. This does not necessarily mean that they do not respond with a hyperkinetic circulation to exercise but may be because they are too weak to elicit the symptoms. They show more severe muscle weakness and wasting than the patients with the classical form, who have normal or near normal muscle strength at rest as we also observed in patients in this study. The two brothers had pathological ECG findings and mild cardiac hypertrophy without dilatation in contrast to the findings in the patients with the classical form, who had normal ECG and echocardiographic findings. This indicates that the heart was primarily affected although it cannot be excluded that the cardiac findings are secondary to hyperdynamic circulation.

Consequences of the deep intronic mutation are insertion of pseudoexons between the last two exons and changing of the reading frame. Pseudoexons are defined as intronic inclusions flanked by good-to-consensus acceptor and donor-site signals that despite this are not normally recognized by the splicing machinery (Sun and Chasin, 2000). The reason for their exclusion in the normal splicing is not fully understood, but the RNA secondary structure has been suggested as one of the key regulatory elements in the pre-mRNA splicing process (Buratti and Baralle, 2004; Buratti et al., 2007). The relatively high splice score for the pre-existing acceptor splice site in the wild-type sequence (87.12%) indicates that the RNA is folded in a way that normally will keep the splice site silent. A prediction of RNA folding of the pseudoexons with flanking sequences using the mFold software (Zuker, 2003) revealed a stable G–C base pairing at the site for the mutation in the wild-type sequence, which is disrupted in the mutant pre-mRNA. In addition to the predicted change in RNA folding, the mutation also raised the score for splicing to 90.21%. Recent works have described the presence of one 100 base pair insert as a consequence of the deep intronic mutation (Mochel et al., 2008; Olsson et al., 2008), but our findings show that there are actually two different inserts present, the 100 base pair insert and one slightly shorter of 86 base pair. These two inserts have the same acceptor splice site but different donor splice site. The presence of two different inserts gives us a clue of the actual order of the splicing process in the mutant pre-mRNA. Nevertheless, the outcome will not be changed since both mRNAs with inserts harbour the same premature termination codon. The missense mutation in exon 3 changes a glycine residue to a glutamate at amino-acid position 50 (Supplementary Fig. 2). This amino-acid residue is totally conserved among species from bacteria to mammals.

Since Fe–S clusters are essential enzyme cofactors in many proteins, defects in the biosynthesis of these clusters can affect many cellular processes. The lethality in several model systems after disruption of different genes involved in the Fe–S cluster assembly process indicates that some residual normal activity of these proteins is necessary to be compatible with life (Rouault and Tong, 2008). It is rather common that splicing affecting gene variants give rise to both normal and aberrant transcript, as in our patients (Fig. 4A). The distribution between normal and aberrant transcripts may vary between tissues and the disease will only be manifest in tissues where the level of normal transcripts is below a certain threshold level (Pagani and Baralle, 2004). The difference in the level of properly spliced mRNA between severely affected and not affected tissue might be rather small. One might speculate that the levels of transcripts of normal size in patients homozygous for the deep intronic mutation is sufficient to overcome the need for IscU in certain tissues like heart, visceral smooth muscles and brain. The compound heterozygous patients were more severely affected, indicating that the allele with the missense mutation gives rise to a non-functional product. Nevertheless, both brothers have normal intelligence and there are no signs of encephalopathy.

The immunoblot analysis of skeletal muscle mitochondria isolated from patients homozygous for the deep intronic mutation revealed a severe reduction of the IscU protein, i.e. there is an obvious correlation with the amount of protein and the reduced function of Fe–S cluster containing enzymes. In the mitochondria isolated from the compound heterozygous brothers, there was no significant reduction of the protein expression and it is reasonable to assume that their mutant form with the amino-acid substitution is expressed but non-functional.

Expression of the aberrant mRNA would result in a ∼1.4 kDa smaller product than the wild-type mitochondrial IscU. In one of the recent reports about this disease (Mochel et al., 2008), the studies performed in skeletal muscle tissue did not reveal any product that could represent a truncated mitochondrial IscU protein in patients. The authors suggested that the premature truncation codon in the ISCU mRNA eventually activated nonsense-mediated mRNA decay, which is a common feature, observed in diseases caused by nonsense mutations (Maquat, 2005). In this study, the mutant RNA was readily amplified by RT–PCR performed on skeletal muscle tissue and in the western blots loaded with protein extracts from isolated skeletal muscle mitochondria we could clearly observe a band that could represent a truncated IscU protein as shown in Fig. 5A. In the western blot analysis of proteins extracted from digitonin-permeabilized fibroblasts this truncated protein-product was not detectable. In the cultured skin fibroblasts, there was no obvious reduction of the IscU protein in patients compared to control subjects in accordance with findings from the recent reports.

An unexpected observation was that the immunoreactive species within mitochondria migrated differently than previously described and seems to be larger than the predicted 14 kDa (Tong and Rouault, 2000, 2006). This is especially obvious in the comparative Western blot loaded with total proteins from cultured skin fibroblasts, proteins extracted from digitonin-treated fibroblasts from the same culture and proteins from isolated skeletal muscle mitochondria (Fig. 6). Digitonin-permeabilization and differential centrifugation of cells eliminates cytosolic proteins, and the fraction of proteins used for Western blots should therefore contain only mitochondrial IscU2. The immunoreactive species in these fractions were of the same size as the immonoreactive species in proteins from isolated skeletal muscle mitochondria. The faster migrating species was present only in lanes loaded with proteins extracted from intact cells and disappeared in the membrane organelle protein fractions. These results are opposed to previous findings and urge further investigation on IscU processing.

As previously reported in myopathy with deficiency of succinate dehydrogenase and aconitase (Haller et al., 1991) there was iron accumulation in the muscle tissue of our patients and electron microscopic visualization clearly localized the corresponding inclusions to within the mitochondrial matrix. These findings are in accordance with findings in yeast strains (Schilke et al., 1999) and mammalian in vitro systems (Tong and Rouault, 2006) with compromised Fe–S cluster formation, and the findings in heart tissue in Friedreich ataxia, which are characterized by mitochondrial iron overload and mitochondrial oxidative damage (Rouault and Tong, 2005; Rötig et al., 1997). Treatment strategies in patients with Friedreich ataxia aim at diminishing the assumed oxidative damage to lipids, DNA and proteins that occurs during the progression of the disease due to iron accumulation. Most treatments are based on administration of antioxidants (Babady et al., 2007), but recently, a trial with selective iron chelation has shown promising results (Boddaert et al., 2007). In our patients it is hard to know whether the mitochondrial myopathy is due to the deficient respiratory chain function because of deficiency in Fe–S cluster enzyme subunits or due to secondary effects of iron overload and potential oxidative damage. Anyway, in our patients, there is another treatment strategy based on the possibility to manipulate gene expression that might be an option. A therapeutic approach in diseases caused by splicing defects could be to induce a shift towards a higher level of normally spliced transcripts in order to overcome the threshold level. The most promising strategy to manipulate gene expression is by the use of antisense oligonucleotides (Aartsma-Rus and van Ommen, 2007) e.g. to induce antisense-mediated exon skipping. The possibility to use antisense oligonucleotides to induce skipping of the pseudoexons in our patients would be particularly beneficial since it would result in a wild type mRNA restoration of the allele with the intronic mutation.

Supplementary material

Supplementary material is available at Brain online.

Acknowledgements

The authors thank Dr Ulf Jansson, Laboratory Medicine Västernorrland, Sundsvall, Sweden for providing control samples from the county of Västernorrland and Dr Marju Orho-Melander, University Hospital Malmö, Malmö, Sweden for providing control samples from a Finnish population.

Footnotes

  • Abbreviations:
    Abbreviations
    ECG
    electrocardiography
    Fe–S
    iron–sulphur
    MRC
    Medical Research Council
    RT–PCR
    reverse transcriptase polymerase chain reaction

References

View Abstract