The Bethlem myopathy is a rare autosomal dominant proximal myopathy characterized by early childhood onset and joint contractures. Evidence for linkage and genetic heterogeneity has been established, with the majority of families linked to 21q22.3 and one large family linked to 2q37, implicating the three type VI collagen subunit genes, COL6A1 (chromosome 21), COL6A2 (chromosome 21) and COL6A3 (chromosome 2) as candidate genes. Mutations of the invariant glycine residues in the triple-helical domain-coding region of COL6A1 and COL6A2 have been reported previously in the chromosome 21-linked families. We report here the identification of a G→A mutation in the N-terminal globular domain-coding region of COL6A3 in a large American pedigree (19 affected, 12 unaffected), leading to the substitution of glycine by glutamic acid in the N2 motif, which is homologous to the type A domains of the von Willebrand factor. This mutation segregated to all affected family members, to no unaffected family members, and was not identified in 338 unrelated Caucasian control chromosomes. Thus mutations in either the triple-helical domain or the globular domain of type VI collagen appear to cause Bethlem myopathy.
Bethlem myopathy is a rare childhood onset, proximal muscular dystrophy with joint contractures most frequently affecting the elbows and ankles. It is inherited in an autosomal dominant fashion with virtually complete penetrance by age 5. A series of families has previously been linked to chromosome 21q22.3 (1). Evidence for genetic heterogeneity and linkage of a large American pedigree of French-Canadian descent to 2q37 subsequently was established (2,3). The families linked to chromosomes 21 and 2 are phenotypically identical (1,2). These results implicated the three type VI collagen subunit genes, COL6A1 (chromosome 21), COL6A2 (chromosome 21) and COL6A3 (chromosome 2) as candidate genes (4). Each type VI collagen chain comprises a central triple-helical domain of 335–336 amino acids with repeating Gly-X-Y sequences, flanked by N-and C-terminal globular domains primarily composed of 200 amino acid motifs similar to type A domains of the von Willebrand factor (vWF) (for a review, see ref. 5). The α3(VI) chain encoded by COL6A3 is much larger than the other two chains because it has multiple, alternatively spliced vWF A motifs in the N-terminal globular domain and three additional protein motifs in the C-terminal globular domain. Jôbsis et al. identified mutations in three chromosome 21-linked families, two in COL6A2 and one in COL6A1 (6). These mutations disrupted the Gly-X-Y motif of the triple-helical domain in each of the proteins encoded by these genes. Here we report the analysis of COL6A3 in the chromosome 2-linked pedigree. A missense mutation was found in the N-terminal globular domain rather than the triple-helical domain of the COL6A3-encoded protein. The mutation leads to a substitution of glycine by glutamic acid in the N2 subdomain, one of the vWF A-like motifs in the N-terminal globular domain.
Northern blot analysis showed that dermal fibroblasts from the two affected and two unaffected members of the chromosome 2-linked family [family 1489 (2,3)] produced mRNAs for the three collagen VI chains of the expected sizes (Fig. 1A). The ratios of the α3(VI) to α1(VI) mRNA in fibroblasts from affected members were not significantly different from the ratios in the unaffected individuals (Table 1). Metabolically labeled fibro-blasts were immunoprecipitated with an antiserum specific for the N-terminal globular domain of the α3(VI) chain, which recognized both triple-helical type VI collagen monomers and unassociated α3(VI) chains (8). The ratios of the α3(VI) to α1(VI) plus α2(VI) chains synthesized by the affected and unaffected fibroblasts were similar (Fig. 1B and Table 1). To test further if COL6A3 is indeed the gene at fault, family 1489 was genotyped using a highly informative dinucleotide polymorphism within COL6A3 (9). No recombination of the marker and the disease locus was found, suggesting the possibility of a point mutation or a small deletion/insertion in COL6A3.
Since mutations in the chromosome 21-linked families occur in the triple-helical domain, we first sequenced the COL6A3 triple-helical-coding regions of the two affected individuals. The sequences were identical to the cDNA sequence that we reported previously (7,10), except for a nucleotide variant that did not result in an amino acid substitution at position 7110 (Table 2). We then analyzed for changes in the sequences of the N-and C-terminal globular domains by conformation-sensitive gel electrophoresis (CSGE) (11,12). Several fragments coding for the C-terminal globular domain showed abnormal patterns of migration, and direct sequencing of the PCR products confirmed the sequence variations (Table 2). These sequence variations did not segregate with the disease phenotype and hence probably represent polymorphisms in the coding region.
We then proceeded to sequence the region of the N-terminal globular domain that does not undergo alternative splicing. A 2.7 kb fragment coding for this region was PCR amplified, subcloned and sequenced. A G→A missense mutation at nucleotide 5291 was found in approximately half of the clones from each affected individual. The mutation led to a substitution of glycine by glutamic acid at amino acid position 1679 in the N2 subdomain and introduced an additional AluI site (Fig. 2). Genomic DNA from all members of family 1489 was PCR amplified and digested with AluI. The mutation was found to segregate to all 19 affected individuals in the pedigree and to no unaffected family members (Fig. 3). A series of 169 unrelated control individuals (n = 338 chromosomes) was analyzed for evidence of the mutation by AluI digestion, and it was not identified in any control. The family member classified as of uncertain diagnostic status failed to inherit the mutation. The diagnostic classification of uncertain for this 48-year-old male was based on a slightly elevated creatine kinase value of 294 (normal range 61–200) in the presence of an entirely normal neurological evaluation. There are numerous causes of elevated creatine kinase values in the absence of neurological disorder, including strenuous exercise, cardiac abnormalities and laboratory error. Thus, it is likely that this mildly elevated creatine kinase is not predictive of disease status in this patient, which is expected based on the virtually complete penetrance of the disease gene.
We report a G→A transition mutation in the N-terminal globular domain of COL6A3 in the chromosome 2-linked Bethlem myopathy family. The study provides evidence that a missense mutation in the globular domain of the COL6A3-encoded protein may also cause Bethlem myopathy.
Initial scanning of mutations by CSGE has identified several sequence variants in the C-terminal globular domain-coding region (Table 2). Of interest is an insertion of three nucleotides (GCT) in a region containing three repeats of GCT trinucleotides. This leads to an additional alanine in a stretch of four consecutive alanines in the C3 domain of COL6A3-encoded protein, which contains many repetitive amino acids (7). The reason why the G→A mutation at position 5291 was not detected by the CSGE is not clear. A possible explanation is that the size of the PCR product (693 bp) was larger than the optimal fragment size of 500 bp for the CSGE.
Northern and biosynthetic studies indicate similar levels oftype VI collagen mRNA and protein production in affected and unaffected fibroblasts. The results suggest that mutant α3(VI) chains are probably produced in fibroblasts from affected individuals, although we could not directly assess the amount of the mutant chains because they differ from the normal chains by a single amino acid. The absence of detectable type VI collagen abnormality in fibroblasts may be due to the fact that the disease phenotype is manifested only in the skeletal muscle.
The mutation in family 1489 occurs in one of the 12 vWF A domains of the α3(VI) chain (N2 motif) at a position occupied predominantly by glycine or alanine when sequences of 75 vWF A domains in 25 proteins were aligned (ref. 13 and Fig. 4). The vWF A domain is present in one or more copies in complement components, integrin receptors and extracellular matrix proteins, all of which are involved in cell-cell, cell-matrix and matrix-matrix interactions (13,14). It has been shown that the A domains in vWF and in integrins are directly involved in ligand binding (15,16). The crystal structures of the A domains from two leukocyte integrin receptors have been reported (17,18). They consist of alternating α helices and β strands resembling the classic dinucleotide-binding fold. Specifically, a central sheet comprising one short antiparallel and five parallel β strands is surrounded by seven α helices. The ligand-binding site is thought to be in a pocket where the C-termini of the parallel β strands are located. The Gly1679Glu mutation in the COL6A3-encoded protein would correspond to the center of the second β strand (β2). The substitution of a charged amino acid is likely to alter the secondary structure of the β sheet, which may impair protein folding or disrupt ligand binding. It is of interest to note that mutations of type 2b von Willebrand disease are clustered in an area within or adjacent to the β2 strand of the A domain of vWF (ref. 18 and Fig. 4).
Type VI collagen microfibrils are assembled by end-to-end association of tetramers comprising four triple-helical monomers (5). It is believed that the microfibrils are stabilized by interactions of globular domains with each other and with the triple-helical domains. Evidence thus far indicates that one each of the three subunits is required for the assembly of a stable type VI collagen triple helix (19–21). The α1(VI) and α2(VI) chains, lacking domains N2-N10 and C3-C5, are much smaller than the α3(VI) chain (7,22). The precise function of the extended N-and C-terminal globular domains of the α3(VI) chain is not yet known. The identification of a mutation in the N2 motif implies a critical role for the N-terminal globular domain of the α3(VI) chain. It seems possible that this region may contain binding sites for interacting with other parts of the type VI collagen molecule and thus may play a role in microfibril assembly. Alternatively, it may be responsible for the direct binding of molecules in the muscle basement membrane. Type VI collagen microfibrils previously have been localized close to some basement membranes (23). An interaction between the N-terminal globular domain of the α1(VI) chain and the C-terminus of type IV collagen, a ubiquitous basement membrane component, recently has been reported (24). Whether the N-terminal globular domain of the α3(VI) chain interacts with type IV collagen remains an interesting question.
Materials and Methods
Family materials and diagnostic classification
Thirty six members of family 1489 are included in the present study (2,3). Methods for DNA extraction and diagnostic criteria for the Bethlem myopathy are as previously described (3). The present study includes the DNA of 19 affected individuals, 12 unaffected individuals, four unrelated spouses and one family member of uncertain diagnostic status. All pedigree, clinical, DNA sample and genotyping data are stored in the PEDIGENE database management system (25). This study was approved by the Institutional Review Board at Duke University Medical Center.
Northern blot and biosynthesis
Dermal fibroblasts were established from two affected (0118, 9033) and two unaffected individuals (0119, 9034), and were cultured in Dulbecco's modified Eagle medium with 10% fetal bovine serum (Gibco BRL). Total RNA was isolated form fibroblasts by the guanidinium thiocyanate-phenol-chloroform extraction method (26). Southern and northern hybridization analyses were performed by standard protocols (27), using overlapping COL6A3 cDNA clones as probes (7). Biosynthesis of type VI collagen was analyzed by metabolic labeling of cultured fibroblasts with [35S]methionine and immuno-precipitation with polyclonal antibodies against the α3(VI) chain (8) by the method described previously (28).
Total RNA was reverse transcribed with Superscript II reverse transcriptase (Gibco BRL) and amplified by PCR. A 1176 bp region containing the triple-helical domain-coding region was amplified with AmpliTaq PCR kit (Perkin Elmer) using HCOL6A3-1 and HCOL6A3-2 primers (Table 3). The PCR conditions were 94°C for 30 s, then 35 cycles of 94°C for 30 s, 60°C for 30 s and 72°C for 1 min, and a final extension at 72°C for 5 min. Direct sequencing of the PCR products was performed by cycle sequencing with fluorescence-labeled dideoxynucleo-tides on an automatic DNA sequencer (Applied Biosystems). Fourteen primer pairs were used to generate overlapping fragments, ranging in size from 524 to 939 bp, for CSGE analysis. The sequences of these primers will be provided upon request. CSGE analysis was carried out by the methods described by Ganguly et al. (11) and Williams et al. (12). The 2.7 kb PCR product covering part of the N-terminal globular domain was amplified using the HCOL6A3-F7 and HCOL6A3-B10 primers (Table 3), and subcloned into the pCR II TA cloning vector (Invitrogen). Eight independent clones from each affected individual were sequenced manually using [33P]dATP and the Sequenase kit (US Biochemical).
For the analysis of the dinucleotide repeat polymorphism, genomic DNA from 36 family members was PCR amplified using primers and conditions as described (9). Exons encoding the N2 domain were PCR amplified from genomic DNA of 36 family members and 169 unrelated controls using HCOL6A3-F9A and HCOL6A3-B9 primers (Table 3). PCR conditions were an initial denaturation of genomic DNA at 96°C for 2 min then 35 cycles of 94°C for 30 s, 60°C for 30 s and 72°C for 1 min, and a final extension at 72°C for 5 min. The resultant 492 bp PCR products were cut with AluI and analyzed on a 1% agarose gel.
conformation sensitive gel electrophoresis
poly-merase chain reaction
von Willebrand factor
The authors thank Pieter Bolhuis for his stimulating discussions and Rupert Timpl for the antibodies. This work is supported by NIH grants AR38912 and AR38923 (M.-L.C.), NS26630 (M.A.P.-V), and by grants from the Muscular Dystrophy Association (M.C.S. and M.A.P.-V).
Chu, M-L. (EMBL X52022)