We have identified a mutation in the myotilin gene in a large North American family of German descent expressing an autosomal dominant form of limb girdle muscular dystrophy (LGMD1A). We have previously mapped this gene to 5q31. Symptoms of this adult onset disease are progressive weakness of the hip and shoulder girdles, as well as a distinctive dysarthric pattern of speech. Muscle of affected individuals shows degeneration of myofibers, variations in fiber size, fiber splitting, centrally located myonuclei and a large number of autophagic vesicles. Affected muscle also exhibits disorganization and streaming of the Z-line similar to that seen in nemaline myopathy. We have identified a C450T missense mutation in the myotilin gene that is predicted to result in the conversion of residue 57 from threonine to isoleucine. This mutation has not been found in 396 control chromosomes. The mutant allele is transcribed and normal levels of correctly localized myotilin protein are seen in LGMD1A muscle. Myotilin is a sarcomeric protein that binds to α-actinin and is localized in the Z-line. The observed missense mutation does not disrupt binding to α-actinin.
Received 1 May 2000; Revised and Accepted 26 June 2000.
Limb girdle muscular dystrophy (LGMD) encompasses a clinically diverse group of disorders often characterized by proximal muscle weakness, elevated serum creatine kinase values and absent or reduced deep tendon reflexes. Both dominant (LGMD1) and recessive (LGMD2) forms have been reported. A number of the recessive limb girdle dystrophies are caused by mutations in the dystrophin-associated glycoproteins that make up the large multiprotein complex on the sarcolemmal membrane (1). Among the dominant forms, five have been mapped: LGMD1A to 5q31 (2); LGMD1B to 1q11–21 (3); LGMD1D to 6q23 (4); LGMD1E to 7q (5); and vocal cord and pharyngeal weakness with autosomal dominant distal myopathy (VCPDM) to 5q (6). Mutations in caveolin 3 have been described in LGMD1C (7,8) and mutations in the lamin A/C gene have been identified in autosomal dominant Emory–Dreyfus muscular dystrophy (9). We have previously reported the only known pedigree in which autosomal dominant limb girdle muscular dystrophy 1A (LGMD1A) is segregating (10). This pedigree was linked to 5q (11), within a 2 Mb interval bounded by D5S479 and D5S594 (2,12,13).
Individuals affected with LGMD1A exhibit proximal leg and arm weakness with a mean age at onset of 27 years, later progressing to include distal weakness. Approximately half of the affected individuals exhibit a distinctive nasal, dysarthric pattern of speech. Tightened heel chords and reduced knee and elbow deep tendon reflexes are frequently seen, although nerve conduction studies are normal (10). Creatine kinase levels are elevated, ranging from 1.6- to 9-fold higher than the normal limit of 120 IU/l for males and 80 IU/l for females. EMG changes indicative of primary myopathy are observed. We describe here the identification in this family of a mutation in the myotilin gene, which encodes a 498 amino acid polypeptide with a molecular weight of 57 kDa (14). Myotilin is a sarcomeric protein that binds to α-actinin and is associated with the Z-line. Its C-terminus contains two Ig domains homologous to the Z-disk-associated Ig-like domains 7 and 8 of the giant muscle protein titin. In contrast, the N-terminus is unique and contains no known structural domains. Other features are a 23 amino acid hydrophobic stretch (residues 57–79) and a region of high serine content between residues 28 and 124 (27 of 96 residues).
Positional identification of the myotilin mutation
Analysis of the previously identified 2 Mb candidate interval for LGMD1A revealed 28 apparently unique expressed sequence tags (ESTs). The EST AA086281 was investigated as a candidate gene because it is contained in UNIGENE cluster 84665, which is composed almost exclusively of skeletal muscle-derived ESTs. A P1-derived artificial chromosome (PAC) clone 38I10 (GenBank accession no. AC004820) was identified that contained ESTs AA086281 and AA180059 from this UNIGENE cluster. Comparison of the sequence of these ESTs to the known sequence of PAC 38I10 allowed putative intron–exon boundaries to be identified. Intronic primers were designed and exons were amplified from the genomic DNA of two affected and two control individuals. Sequence analysis identified a C→T missense mutation in affected individuals. Using the predicted exonic sequence around this mutation as probe, 19 cDNA clones were isolated from a human skeletal muscle cDNA library, sequenced and assembled into a full-length 2244 bp cDNA. At this time it became apparent that the positionally identified gene is identical to a gene termed myotilin (GenBank accession no. AF144477) independently isolated through a yeast two-hybrid screen to identify novel cytoskeletal components (14). This gene has also been investigated as a possible tumor suppressor (15).
The primer pairs in Table 1 were used to amplify each of the 10 myotilin exons from genomic DNA of two LGMD1A patients and two controls. Sequence analysis of these genomic PCR products demonstrated a C450→T missense mutation in exon 2 of affected individuals (Fig. 1a), which is predicted to convert residue 57 from threonine to isoleucine (T57I). This portion of exon 2 was amplified from 69 affected individuals and 194 unaffected family members and subjected to single strand conformation polymorphism (SSCP) analysis (Fig. 1b). The mutation segregates perfectly with the disease-associated haplotype. Eight clinically asymptomatic (normal examination, normal creatine kinase) individuals in the pedigree carry the disease-associated haplotype and the mutation. All of these individuals were last examined at ages well below the average age at onset for disease in the pedigree (27.1 ± 8.5 years); therefore, it is probable that these individuals will become symptomatic on subsequent evaluation. This mutation has not been detected in 396 control chromosomes of Caucasian North American origin. It was also not detected in an additional 200 chromosomes of Finnish origin and 60 chromosomes of Japanese origin. A single nucleotide polymorphism (1875G/A; NCBI SNP ID rs4288) in the 3′-untranslated region is present in both normal and affected individuals and appears to be phenotypically silent.
Individuals affected with LGMD1A do not display any defects in splicing of the only observed skeletal muscle-specific myotilin transcript. The splice donor and acceptor regions bounding each exon were sequenced and no mutations were found. Further, a series of overlapping reverse transcription–polymerase chain reaction (RT–PCR) products crossing all splice junctions was amplified from LGMD1A and control muscle RNA and no disease-specific splicing defects were detected (data not shown).
Patient histological findings
A biopsy was performed on the deltoid muscle of a 36-year-old, moderately affected individual. Hematoxylin and eosin stained sections showed degeneration of myofibers, variations in fiber size, fiber splitting and large numbers of centrally located myonulei (Fig. 2a). Affected muscle also exhibits interstitial fibrosis and extensive fatty infiltration late in the disease course, although there is no evidence of inflammatory cell infiltrate. ATPase staining showed some pseudogrouping due to fiber splitting, but there was no evidence for selective myopathic involvement of specific fiber types or for any alteration in the relative numbers of fiber types. The dystrophin-associated glycoprotein complex is intact in LGMD1A muscle (data not shown). There are large numbers of rimmed vacuoles (Fig. 2b), which on electron microscopic analysis appear to be autophagic vesicles, with no evidence of any amyloid material (Fig. 2c). The presence of a large number of vacuoles is sufficiently characteristic to warrant a molecular analysis of the myotilin gene in patients with unlinked myopathies. There are no abnormal accumulations of glycogen and no abnormalities of mitochondria or the tubular systems are observed. LGMD1A muscle also exhibits patches of striking Z-line irregularity (Fig. 2d). Although such Z-line streaming is a relatively non-specific feature present in many myopathic disorders and can sometimes be seen even in presumably normal, asymptomatic muscle, the extent and severity of streaming seen in LGMD1A muscle is very unusual. These Z-line abnormalities are entirely consistent with myotilin’s interaction with α-actinin and its localization near the Z-line.
Localization and expression level of myotilin protein
To determine whether the T57I missense mutation alters the localization of myotilin protein, frozen sections of LGMD1A and control muscle were stained with antibody to myotilin (Fig. 3a). The normal pattern of sarcomeric staining was seen in both samples and there was no evidence of ectopic expression or of abnormal accumulation of myotilin protein in affected muscle. Further, western blot analysis of patient and control samples showed that the overall level of myotilin expression was normal in patient muscle (Fig. 3b). Patient muscle (Fig. 3b, lane 2) displayed a prominent immunoreactive band migrating at 75 kDa, but this band was also seen in control muscle (lane 1). The intensity of this band varied among different control muscle extracts. No alternatively spliced myotilin transcripts have been identified in skeletal muscle that could encode a protein of this size and it may correspond to myotilin protein that has been post-translationally modified.
The antibody used in these experiments does not distinguish between normal and mutant myotilin protein. Although there is insufficient patient tissue available to allow the direct detection of mutant protein, the presence of mutant myotilin transcript can be detected. Patient RNA was reverse transcribed and PCR amplified with primers flanking the mutation. This RT–PCR product was cloned and individual clones were analyzed by denaturing high performance liquid chromatography (DHPLC). The insert of each clone was PCR amplified and mixed with PCR product from wild-type myotilin. After denaturing and reannealing, heteroduplexes were detected by DHPLC. Of the 20 clones tested, 10 (50%) contained the C450→T missense mutation. It is likely that these mutant transcripts are translated into protein.
Binding of myotilin to α-actinin
Previous studies have demonstrated an interaction between myotilin and α-actinin and suggested that the binding site resides in the N-terminal 150 amino acids of myotilin (14). The yeast two-hybrid method was used to further map the interacting region and to test whether the T57I mutation affects this interaction. Figure 4 shows that wild-type full-length myotilin interacts with α-actinin and further mapping with deletion constructs indicates that the binding site resides between residues 79 and 150, whereas no binding activity was seen with residues 1–78. Thus, the α-actinin binding site in myotilin is C-terminal to the T57I amino acid substitution. Further, the wild-type and T57I mutant myotilins demonstrate comparable binding to α-actinin, indicating that the T57I missense mutation is not likely to cause muscle pathology by disrupting the binding of myotilin and α-actinin.
Thr57 is conserved in murine myotilin
Evolutionary conservation of Thr57 would support its importance in the structure and function of myotilin protein. The murine myotilin gene was cloned and sequenced. Two overlapping ESTs corresponding to murine myotilin were identified by searching public sequence databases. Oligonucleotide primers designed from these sequences were used to perform 5′ and 3′ rapid amplification of cDNA ends (RACE) and the products were cloned and sequenced. An alignment of the deduced amino acid sequence of human and murine myotilin is shown in Figure 5. These proteins share 89% identical and 93% similar residues and Thr57 is conserved between the two species.
We have identified a missense mutation in the myotilin gene of LGMD1A patients. This gene encodes a sarcomeric protein that is localized in the Z-disk. The actin-binding protein α-actinin, which cross-links thin filaments into antiparallel bundles in Z-lines, is also localized in the Z-disk and yeast two-hybrid analysis has shown that these two proteins bind to one another (14). Electron microscopic analysis of LGMD1A patient muscle biopsies reveals Z-line abnormalities, including extensive regions of Z-line streaming. This abnormal Z-line material often appears as rod-like bodies, closely resembling the pathognomic rods seen in nemaline myopathy. Mutations in two other sarcomeric proteins that localize to the Z-line have been shown to result in nemaline myopathy: nebulin (NEM2) (16) and α-tropomyosin (NEM1) (17). The latter presents striking parallels with LGMD1A because, not only do both the α-tropomyosin and myotilin proteins bind α-actinin, but also missense mutations in the corresponding genes give rise to autosomal dominant myopathy with disruptions of the Z-line. Mutations in the α-actin gene (ACTA1) also give rise to nemaline myopathy (18).
Nemaline-like rods are a relatively non-specific feature and have been found associated with such diverse disorders as mitochondrial myopathy (19,20), rhabdomyoma (21), HIV infection (22), chronic alcoholism (23), central core disease (24) and paranoid schizophrenia (25). They have also been induced experimentally in rat muscle by tenotomy (26). However, we have presented evidence that the Z-line streaming observed in the muscles of LGMD1A patients is directly involved in the disease process. We have provided quantitative data showing that the transcript from the mutant myotilin allele is present in the muscle tissue of LGMD1A patients and thus is likely to be translated. We have shown by western analysis that the overall level of myotilin protein is normal in LGMD1A muscle and immunocytochemical staining shows this myotilin to be correctly localized to the Z-line, with no evidence of precipitates or abnormal deposits. Thus, myotilin is localized precisely to the site of the observed myopathic changes and interacts with other proteins localized there, consistent with the observed myotilin defect playing a causative role in the Z-line abnormalities seen in LGMD1A patients.
The nemaline myopathies are a heterogeneous group of disorders and include a severe neonatal form that is often fatal during the first year of life, a less severe congenital form and an adult onset form (27–29). The less severe congenital and adult onset forms are non-progressive or slowly progressive (30). Type I muscle fibers tend to predominate, with the disappearance of type IIB (fast glycolytic) fibers and an increase in undifferentiated type IIC fibers (31–33). LGMD1A most closely resembles adult onset nemaline myopathy. Both disorders display Z-line streaming, similar ages at onset and normal or mildly elevated serum creatine kinase levels. However, there are significant differences between the two disorders. Although there is substantial variation among individuals, in general LGMD1A displays a more rapid clinical progression of disease than does adult onset nemaline myopathy. LGMD1A shows no evidence of preferential involvement of specific fiber types and the autophagic vesicles so characteristic of LGMD1A are not found in nemaline myopathy.
The precise mechanism by which the T57I mutation in myotilin disrupts the Z-disk is not yet understood and will require better knowledge of the function of myotilin in normal muscle. Because both the expression level and localization of myotilin appear to be normal in LGMD1A muscle, we speculate that the T57I mutant myotilin may perturb the normal structure of the Z‐disk. This could be the result of a gain of function or, perhaps more likely, the disruption of an existing protein–protein interaction required for normal structure and function of the Z‐disk. The yeast two-hybrid experiments presented here would seem to preclude a direct effect on α-actinin binding and suggest that the T57I substitution may interfere with binding of another, as yet unidentified, sarcomeric protein.
The identification of additional pedigrees carrying myotilin mutations would provide further insight into possible structure–function relationships in the myotilin protein. To date only one other autosomal dominant myopathic pedigree has been linked to 5q: the minimal candidate region for vocal cord and pharyngeal weakness with autosomal dominant distal myopathy (VCPDM) spans 7 cM and contains the myotilin locus (6). Through a combination of RT–PCR and sequence analysis of material from several VCPDM patients, splicing and coding defects in the skeletal muscle isoform of myotilin have been excluded as the cause of this disease. We are currently screening a collection of 108 individuals with unlinked limb girdle muscular dystrophy for defects in myotilin and will screen additional samples as they become available. At the same time, we are developing animal models for LGMD1A and are using yeast two-hybrid analysis to define the regions responsible for the binding of myotilin to various muscle proteins, as well as the effects of the T57I mutation on those interactions. These investigations will shed additional light on this interesting and novel protein and its role in normal and diseased muscle.
MATERIALS AND METHODS
The Clontech (Palo Alto, CA) human skeletal muscle 5′-stretch plus cDNA library was plated at a density of 50 000 clones per 127 mm nylon membrane (Hybond; Amersham, Arlington Heights, IL). Probe was prepared using the primers 5′-CTCCAGATTGCAGCCTCCT-3′ (forward) and 5′-TACTGCTATTGTAATCAGGC-3′ (reverse) to amplify genomic DNA, and the resultant 279 bp product was labeled with random primer extension using the Readiprime kit (Amersham). Nineteen positively hybridizing clones were plaque-purified twice before sequencing.
The murine myotilin cDNA was isolated from the Clontech mouse skeletal muscle Marathon cDNA library. Two ESTs (AI553071 and AA422457) corresponding to murine myotilin were identified through a TBLASTN search with the deduced amino acid sequence of human myotilin protein. This sequence was used to design primers for 5′ RACE (5′-GGTGGAGGGATAGCTGAGATCTGGCACTCCAGC-3′) and 3′ RACE (5′-GCTGGAGAAGCCACCTTCACAGTGCAGCTGG-3′). The products were cloned, sequenced and the open reading frame translated. The nucleotide sequence has been submitted to GenBank (accession no. AF230979).
To detect the C450→T mutation, a 279 bp fragment was amplified from genomic DNA using primers 5′-CTCCAGATTGCAGCCTCCT-3′ (forward) and 5′-TACTGCTATTGTAATCAGGC-3′ (reverse) with touchdown cycling conditions of 94°C for 20 s, 65°C for 15 s and 72°C for 20 s, with the annealing temperature decreasing 1°C per cycle to a final value of 55°C. After 45 cycles, 2 vol of SSCP stop buffer were added (0.1% xylene cyanol, 0.1% bromophenol blue, 10 mM NaOH, 95% formamide), samples were denatured for 3 min at 94°C and chilled on ice for 5 min. Samples were electrophoresed at 12 W for 3.5 h at room temperature in a 0.5× MDE gel (FMC Bioproducts, Rockland, ME) with 0.6× Tris–borate/EDTA buffer. To detect the G1875→A polymorphism, a 188 bp PCR product was generated using primers 5′-ACAGAAAGATGCTGGGTGG-3′ (forward) and 5′GTAGGCTCACAAATCGGAG-3′ (reverse). PCR amplification and SSCP was performed as above, except that samples were electrophoresed at 8 W for 2.5 h.
Lymphocytes were isolated from whole blood and genomic DNA extracted using Puregene kits (Gentra Systems, Minneapolis, MN). Genomic PCR products and cDNA clones were sequenced using the ABI Prism BigDye terminator kit (Perkin Elmer, Foster City, CA) and a Perkin Elmer ABI 373 sequencer or the Amersham Pharmacia Thermo Sequenase labeled primer cycle sequencing kit and a LiCor 4200 Gene Reader sequencer (LiCor, Lincoln, NE). Overlapping sequences were assembled using Sequencher software (Ann Arbor, MI).
Muscle tissue was isolated during diagnostic biopsy and frozen in liquid nitrogen-cooled isopentane. Cryosections were stained with hemotoxylin and eosin. For electron microscopy, muscle tissue was fixed in 0.1 M cacodylate-buffered Karnovsky’s fixative solution (3% glutaraldehyde and 3% formaldehyde) for 4 h at 4° C, pH 7.4, washed overnight in cacodylate buffer, post-fixed in a buffered solution of 1% osmium tetroxide and dehydrated through a graded ethanol series. Each specimen was propylene oxide embedded in epoxy resin and polymerized for 3 days at 45°C and 1 day at 60°C. Ultra-thin sections were post-stained in 1% uranyl acetate and lead citrate.
Tissue collected during biopsy of the deltoid muscle was snap-frozen in liquid nitrogen-cooled isopentane and 7 µm sections were fixed on glass slides for 5 min with cold acetone. Phosphate-buffered saline (PBS: 0.15 g/l Na2HPO4, 0.2 g/l KH2PO4, 0.2 g/l KCl, 8.0 g/l NaCl, pH 6.85) was used for washing and all subsequent incubations. Sections were blocked for 30 min with 5% goat serum, washed and incubated for 60 min at room temperature with 32 µg/ml affinity-purified anti-myotilin antibody in 1% goat serum (14). Samples were washed three times for 10 min each in PBS, stained for 30 min at room temperature with 0.4 µg/ml Cy3-conjugated goat anti-rabbit IgG (H+L) secondary antibody from Amersham Pharmacia, washed three times for 10 min each and mounted with Vectashield (Vector, Burlingame, CA).
Tissue collected during biopsy of the deltoid muscle was snap-frozen in liquid nitrogen, ground in a mortar and pestle, resuspended in 45 µl of homogenization buffer (1% SDS, 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 5 mM EGTA, 0.5 µg/ml leupeptin, 0.2 U/ml trasylol) per 50 mg tissue, vortexed for 45 s, boiled for 2 min and centrifuged at 12 000 g for 3 min to clarify. Protein concentration was measured using the Coomassie Plus Protein Assay Reagent (Pierce, Rockford, IL) and 20 mg of protein extract was electrophoresed on an 8% polyacrylamide gel electrophoresis gel. The gel was electrotransferred to a Bio-Rad (Richmond, CA) Trans Blot Transfer Medium, Pure Nitrocellulose 0.45 µm membrane, blocked with 5% non-fat dry milk + 0.1% Tween 20 in PBS and stained with 1.6 µg/ml anti-myotilin antibody (14) for 60 min at room temperature, washed three times for 10 min each in PBS + 0.1% Tween 20, stained with 0.53 µg/ml Jackson horseradish peroxidase-conjugated goat anti-rabbit secondary antibody, washed three times for 10 min each in PBS + 0.1% Tween 20 and detected with an ECL kit as described by Amersham Pharmacia.
Denaturing HPLC analysis of myotilin transcripts
To determine the relative amounts of normal and mutant myotilin transcript present in muscle LGMD1A muscle, a 410 bp RT–PCR product flanking the observed mutation was amplified from patient muscle RNA using primers 5′-CTCAACAAGGAAGAGCAGAC-3′ and 5′-GTTGTAACCCTTTGGCCTGG-3′. This product was cloned into plasmid pTAdv (Clontech) and 20 individual clones were selected for analysis. The insert from each clone was reamplified with the above primers and mixed with the same PCR product amplified from genomic DNA of an unaffected individual in order to allow formation of heteroduplexes. Genomic PCR amplification was performed in a 50 µl volume containing 20 mM Tris–HCl pH 8.4, 50 mM KCl, 0.2 mM dNTPs, 1.5 mM MgCl2, 0.5 U of Platinum Taq polymerase (Gibco Life Technologies, Grand Island, NY) and 0.20 µM concentrations of each primer. The cycling program started with an initial denaturation step of 2 min at 92°C. All subsequent denaturation steps were 92°C for 5 s and all elongation steps were 72°C for 45 s. The annealing temperature was decreased by 2°C over 10 cycles from 65 to 57°C. Thirty cycles were done with an annealing temperature of 55°C for 30 s. Heteroduplexes were formed by heating the samples to 95°C for 3 min and slowly cooling to 25°C over a period of 40 min. Four microliters of sample was injected into a Transgenomic WAVE DHPLC for separation on a DNasep column at 62°C as described by the manufacturer (Transgenomic, San Jose, CA).
Yeast two-hybrid analysis
Myotilin cDNA constructs encoding amino acid residues 1–78, 79–150, 1–498 (full-length) or 1–498T57L (LGMD1A patient mutation) were cloned into the pLexA yeast two-hybrid bait vector (34). Human skeletal muscle α-actinin-2 (amino acids 1–894) was cloned in prey vector pGAD10 (Clontech). The bait and prey plasmids were co-transformed into the Saccharomyces cerevisiae L40 reporter strain using a modified lithium acetate protocol. The transformants were plated onto synthetic medium lacking histidine, leucine, tryptophan, uracil and lysine and plates were incubated for 3 days at 30°C. HIS+ colonies were grown on synthetic medium lacking leucine and tryptophan, incubated for 3 days at 30°C and assayed for β-galactosidase activity by a filter assay (35) or liquid assay following the Matchmaker protocol (Clontech).
We would like to thank Dr Ed Bossen for helpful discussions of muscle pathology. This work was supported by a grant from the Muscular Dystrophy Association (M.C.S. and C.A.W.) and NIH grant NS26630 (M.P.V.).
These authors contributed equally to this work
To whom correspondence should be addressed. Tel: +1 919 684 3508; Fax: +1 919 684 2275; Email: firstname.lastname@example.org
|Oligonucleotide||Sequence||Position relative to exon|
|Oligonucleotide||Sequence||Position relative to exon|