Complete loss of the DNAJB6 G/F domain and novel missense mutations cause distal-onset DNAJB6 myopathy

Abstract

Introduction: Protein aggregation is a common cause of neuropathology. The protein aggregation myopathy Limb-Girdle Muscular Dystrophy 1D (LGMD1D) is caused by mutations of amino acids Phe89 or Phe93 of DNAJB6, a co-chaperone of the HSP70 anti-aggregation protein. Another DNAJB6 mutation, Pro96Arg, was found to cause a distal-onset myopathy in one family.

Results: We detail the mutational, neuropathological, neurophysiological, neurological and radiological features of five new DNAJB6-myopathy families. One has the known Phe93Leu mutation and classic late-onset slowly progressive LGMD1D. Two have different mutations of Phe91 causing a variant childhood-onset severe limb-girdle myopathy. One has a Phe100Val mutation and distal-onset myopathy, unique early bulbar involvement, and a gender-modified wide age-of-onset range. The last has childhood-onset severe distal-onset myopathy and the first non-missense DNAJB6 mutation, c.346 + 5G > A, causing a splicing defect that entirely eliminates DNAJB6's G/F domain (ΔG/F), the domain that harbours all other mutations. Clinical and imaging examinations reveal that muscles considered uninvolved in DNAJB6-myopathy, e.g. lateral gastrocnemii, are affected in our patients with new mutations. Mutational modelling based on the known structure of the bacterial DNAJ2 protein indicates that all past and present mutated residues cluster within 15 Å in the G/F domain and all disturb the interface of this domain with the protein's J domain that confers the interaction with HSP70.

Conclusions: Our patients expand the phenotypic spectrum of DNAJB6-myopathy and allow tentative genotype-phenotype specifications. Combining with previous studies, the clinical severity spectrum is as follows: ΔG/F and Phe91 mutations, most severe; Phe100, Pro96, Phe89 mutations, intermediate; and Phe93, least severe. As it stands presently, proximal G/F domain mutations (Phe89, Phe91, Phe93) cause proximal limb-girdle myopathy, while distal G/F mutations (Pro96, Phe100) cause distal-onset myopathy. While all mutations affect the G/F-J interaction, each likely does so in different unknown extents or ways. One mutation, ΔG/F, causes its associated severe distal-onset myopathy phenotype in a clear way, through generation of a G/F domain-lacking DNAJB6 protein.

Fig. 1

Pedigrees of family 1 and sporadic patients

Fig. 2

Histopathological characterization of muscle biopsies in patients III:5, and 1 to 4s, showing, by H&E and Gomori trichrome stains, rimmed vacuoles, internalized myonuclei, and variability of fiber diameters in all patient muscles; by acid phosphatase staining, increased activity surrounding vacuoles in patient 3s

Fig. 3

Immunolocalization of various proteins in patients 1 to 4s and III:3, showing that DNAJB6 staining is present on the rim of nuclei and in the cytoplasm of the fibers, where it co-localizes at the z discs with desmin, and on the surface and cytoplasmic inclusion of vacuolated fibers; desmin is normally expressed; myotilin is present in few inclusions in rare fibers; LC3 is present in the lumen and on the rim of vacuoles and in subsarcolemmal sarcoplasmic inclusions; p62 is present in sarcoplasmic inclusions and vacuoles with variable intracellular distribution and co-localizes with TDP43 and with BAG3

Fig. 4

Western blot of DNAJB6 protein in muscle homogenates from sporadic patients compared to control showing both DNAJB6 isoforms

Fig. 5

a–e) Electropherograms showing the mutations in genomic DNA from patients II:6 and patients 1 to 4s. f Agarose gel showing the alternative transcript in patient 4s’ cDNA compared to control, and (g–i) sequences confirming skipping of exon 5, which encodes the G/F domain, in the smaller cDNA band and loss of amino acid 79_115 coded by exon 5

Fig. 6

Putative structure of DNAJB6b J and G/F domains obtained through homology modelling using the structure of TtDNAJ2 (pdbcode: 4 J80) as template. a Alignment of DNAJB6b and TtDNAJ2 highlighting the J and G/F domains. Identical residues are shown in bold, residues found mutated in patients are highlighted in red. b Ribbon representation of the homology model of DNAJB6. c A detail of the interface between J and G/F domains. The J domain is represented in surface and colored according to its electrostatic properties (blue for positive and red for negative charge, respectively) while the G/F domain is shown in cyan ribbon. Phenylalanine residues found at the interface are shown in sticks. d Detailed view of the interactions played by Phe100. Its corresponding valine residue found in patients is shown in magenta. The J domain is represented in blue while the G/F domain is represented in cyan. Residues close to Phe100 are also shown in sticks; Phe100 is predicted to establish interactions with both J domain residues (Val55, Lys61 and Ile64) and G/F domain residues (His82 and Val99). e Detailed view of the Phe93Leu mutation; Phe93 is predicted to establish a cation-π interation with J domain residue Lys47 and hydrophobic interactions with G/F domain Pro96. Pro96 faces the J domain helix III and its mutation to arginine is linked to distal-onset phenotype. f Detailed view of the Phe91Leu mutation; interacting residues are also shown. g Detailed view of the Phe89Ile mutation. Phe89 does not directly contact the J domain; rather it protrudes inside the G/F domain spiral where it is predicted to interact with Phe104. Color codes in (e, f and g) panels are the same as in (g)