The UMD-DYSF mutations database
The protein

Dysferlin (O75923) is a 237 kDa protein of 2080 amino acids.

Domain organization

Dysferlin is a type II transmembrane protein that belongs to the ferlin family, a protein family composed of dysferlin (FER1L1), otoferlin (FER1L2), myoferlin (FER1L3), FER1L4, FER1L5 and FER1L6.
Structure of the dysferlin sequence consists in seven C2 domains (C2A to C2G, Therrien et al. 2006), a central region of unknown function composed of ferlin family domains and two nested DysF domains, and a carboxy-terminal transmembrane segment that anchors the protein to the membrane.

Coordinates of dysferlin structural domains as annotated in the UMD-DYSF (Show/Hide)

Role and structure of C2 domains (Show/Hide)

C2 domains are found in proteins that participate in signal transduction via second messengers lipids or in membrane trafficking. They are independently folded protein domains that are commonly described to bind phospholipids and proteins in a Ca(2+)-dependent manner, although some domains can bind to phospholipids independent of calcium and others have not shown calcium binding properties.
Two topologies of C2 domains have been described (type I and type II, Nalefski et al. 1996). Type I C2 domains contain usually all of the five conserved Glu/Asp residues while type II C2 domains lack one or more of the five conserved Ca2+ coordinating residues. Moreover, the loops implicated in membrane recognition and binding are positioned between different beta-strands in type I and II topologies (Therrien et al. 2006, Nalefski et al. 1996). Based on residues conservation, Therrien et al. predicts a type I topology for dysferlin C2C, C2D, C2F and C2G domains and a type II topology for C2A, C2B and C2E.
In dysferlin, the C2A domain has been shown to bind phospholipids in vitro in a Ca(2+)-dependent manner and mutations in this domain can lead to muscular dystrophy (Davis et al. 2002). Considering their sequence structure, C2D, C2F and C2G domains could in theory bind calcium ions and phospholipids whereas C2B and C2E are less likely to bind calcium ions (Therrien et al. 2006).
The 3D structure of a C2 domain has first been reported for the mouse synaptotagmin III (type I topology, Sutton et al. 1995) and more recently for the human myoferlin (type II topology, Protein Data Bank 2DMH,) and the rat otoferlin (type II topology, Helfmann et al. 2011).
Synaptotagmins are single-pass membrane proteins that contain two cytosolic C2A and B domains and a N-terminal transmembrane domain. They are thought to play a role in the regulation of signal transmission at the synapse by mediating synaptic vesicles fusion with the plasma membrane. The crystal structure of the synaptotagmin III C2A domain exhibits five aspartic acid residues which interact with calcium ions and mediate phospholipid binding.
Otoferlin participates in membrane fusion events during exocytosis in auditory hair cells. The crystal structure of its N-terminal C2A domain has revealed important structural and functional differences with synaptotagmin C2A domains: in contrast with synaptotagmin, the otoferlin N-terminal C2A domain does not bind calcium ions and/or phospholipids (Helfmann et al. 2011).

Role and structure of DysF domains (Show/Hide)

Tissue distribution and subcellular location

Dysferlin is a ubiquitously expressed protein, with particular high expression in skeletal muscles, cardiomyocytes, placenta and monocytes. It is a major component of the sarcolemma (Anderson et al. 1999), colocalizes with the dihydropyrimidine receptor (DHPR) in T-tubules (Ampong et al. 2005, Klinge et al. 2010) and is also found in cytoplasmic vesicles (Piccolo et al. 2000). It is not an integral component of the dystrophin-glycoprotein complex.

Involvement in disease

The dysferlin gene was first identified as the genetic causes of Miyoshi myopathy and limb girdle muscular dystrophy type 2B (LGMD2B) in 1998 (Bashir et al. 1998, Liu et al. 1998). Since then, other phenotypes caused by mutations in DYSF have been described (see THE CLINICS part). Patients with mutations in the dysferlin gene show severe reduction or complete absence of dysferlin protein in skeletal muscles. In addition, dysferlin-null mice present an accumulation of vesicles at the sarcolemma, near membrane damage sites (Bansal et al. 2003).
Secondary reduction of dysferlin level is also observed in other neuromuscular disorders, including limb girdle muscular dystrophies such as some sarcoglycanopathies (Piccolo et al. 2000), calpainopathies (Chrobakova et al. 2004) or caveolinopathies (Sugie et al. 2004).
In different forms of muscular dystrophy (dystrophinopathies, sarcoglycanopathies, calpainopathies, fukutin-related protein, and lamin A/C) an altered and predominantly cytoplasmic localization of dysferlin has been described in isolated myofibers (Klinge et al. 2010).

Functional role

Dysferlin deficiency leads to membrane resealing defaults and to severe disruption of the sarcolemma structure (Matsuda et al. 1999). Although its precise biochemical role remains unknown, dysferlin has been shown to play a key role in membrane repair (Bansal et al. 2003) by mediating the fusion of intracellular vesicles to the sarcolemma during two-photons-induced membrane repair in skeletal muscle fibers. Recently, it has been proposed that dysferlin may also be involved in T-tubule formation and maintenance (Klinge et al. 2010).

Mice models for dysferlinopathies

Researchers working on the field of dysferlinopathies have now a plentiful hand of dysferlin deficient models that can be analysed and that can undoubtedly accelerate the pace of dysferlin function studies. Dysferlinopathy profits from two naturally occurring models, the SJL/J and the A/J mice (Bittner et al. 1999, Vafiadaki et al. 2001) and two KO animals (Ho et al. 2004, Bansal et al. 2003).

Detailed information on mice models. (Show/Hide)

The SJL/J mouse strain was demonstrated to have a splicing mutation in the dysferlin gene, making it a natural model for LGMD2B and Miyoshi myopathy ( Bittner et al. 1999, Vafiadaki et al. 2001).This strain has been widely used as an animal model for experimental autoimmune encephalitis, inflammatory muscle disease, lymphomas and as a background strain for a variety of diseases (motor neurone disease, multiple sclerosis and atherosclerosis). Genomic analysis identified the mutation as a 141 bp genomic deletion involving a tandem repeat and altering the 3’ splicing site of exon 45 therefore leading to the splicing of this 171 bp long exon. A low expression of a non active dysferlin mutant was detected on Western-blot. SJL/J mice have inflammatory and degenerative/regenerative signs. Those dystrophic features on skeletal muscle appear at 8 weeks of age and are associated with slowly progressive muscle weakness. Moreover, SJL/J mice have more CD4+ T cells than CD8+ T cells and less B cells than wild-type mice, whereas macrophages are dominant. It is probable that the inflammatory changes seen in SJL/J mice were the result of the defect of dysferlin in muscle and the immunological abnormality.

Another strain of mice carrying a deficient dysferlin was fortuitously discovered in the Jackson laboratory. The A/J strain presents mild progressive muscular dystrophy (Ho et al. 2004). This dystrophic pattern was the consequence of an increased membrane permeability that was observed in a large number of Evans blue dye positive fibres. This was the first indication of a potential role of dysferlin in maintaining membrane integrity. Genetic analysis of this strain showed that a ETn retrotransposon inserted itself in the intron 4 of the dysferlin gene. No dysferlin positive signal was observed on Western-blot, indicating that this insertion destabilizes the whole dysferlin mRNA. Interestingly, whereas the SJL/J strain was known as a very aggressive mouse the A/J mice were peaceful, strongly arguing against the fact that dysferlin deficiency could be at the origin of this nasty comportment.

However, the SJL/J and A/J mice were reported to be naturally occurring mouse models without a strain matched control. Therefore, analysis of dysferlin only on studies in those mice could be misleading. To tackle this issue, the same teams who identified the muscular phenotype in the A/J mice developed a dysferlin deficient mouse model by gene targeting (Ho et al. 2004). Briefly, they knocked-out the exon 45 of dysferlin (the same exon affected in the SJL/J mice) by inserting the neomycin gene cassette by homologous recombination. No dysferlin protein was produced. These dysferlin deficient mice were identified as Dysf^-/-. At 5 months of age, Dysf^-/- mice developed muscular dystrophy shown by centrally placed nuclei, fibres hypertrophy and splitting. This phenotype was primarily seen on proximal muscles but tend to spread on distal muscles that nevertheless remained less affected.

At the same time, Kevin Campbell’s team developed another dysferlin knock-out mouse denoted dysferlin-null (Bansal et al. 2003). In this model the three last exons were deleted. Again, those mice developed a progressive muscular dystrophy correlated with the absence of dysferlin positive signal on Western-blot. The time of onset (2 months) and the severity were almost identical to those observed for SJL/J and Dysf^-/- mice, suggesting that the variability and severity observed on those models were independent of the mutation. From afunctional point of view, the analysis of this last model was very fructuous as it confirmed by two-photon membrane lesion assays that dysferlin was involved in membrane repair.

Additional documentation can be found on the Jain Foundation web site.

Known protein-protein interactions
Dysferlin has been shown to interact with the following proteins:

Alpha-tubulin (TUBA1B, TUBA4A). Polymers of alpha- and beta-tubulins, microtubules are major structural components of eukaryotic cells and are involved in many cellular processes including vesicular transport and protein trafficking to the plasma membrane. Alpha-tubulin interacts with C2A and C2B domains of synaptotagmins I and IX (Honda et al. 2002, Haberman et al. 2003) and it has been suggested that synaptotagmin I interaction with microtubules could retain synaptic vesicles as a readily releasable pool, close to the exocytosis active zone.
Alpha-tubulin also interacts with C2A and C2B domains of dysferlin, in a Ca2+-independent manner (other dysferlin C2 domains were not found to bind alpha-tubulin) (Azakir et al. 2010). In myoblasts, microtubules nucleate at the centrosome and project towards the plasma membrane and co-localization of dysferlin and alpha-tubulin was found in the perinuclear region, as well as in vesicular structures (Azakir et al. 2010). In myotubes, co-localization of dysferlin with alpha-tubulin was found in thin longitudinal structures indicative of microtubules (Azakir et al. 2010). It is thus hypothesized that interaction of dysferlin with microtubules could provide a ready pool of dysferlin for membrane resealing.

Caveolin-3 (CAV3/LGMD1C) is a scaffolding protein within caveolar membranes and T-tubules that belongs to the dystrophin-glycoprotein complex. It regulates the sarcolemmal membrane stability and is involved in vesicular trafficking and signal transduction. Defects in caveolin-3 can be at the origin of several skeletal muscle disease phenotypes: limb girdle muscular dystrophy 1C, rippling muscle disease, distal myopathy and isolated hyperCKaemia (Gazzerro et al. 2010). Caveolin-3 inhibits endocytocis of sarcolemmal dysferlin via a clathrin-independent pathway (Hernandez et al. 2008) and mutations in CAV3 are associated with a mislocalization of dysferlin in muscle cells (Matsuda et al. 2001) with accumulation in the cytoplasm or irregular membrane distribution of dysferlin.

MG53 (TRIM72): MG53 is specifically expressed in striated muscles where it interacts with caveolin-3 to regulate membrane budding and exocytosis by nucleating recruitment of intracellular vesicles to injury sites for membrane patch formation (Cai et al. 2009a, Cai et al. 2009b). MG53 is required for the transport of dysferlin to sites of cell injury during repair patch formation (Cai et al. 2009c).

Annexins A1 and A2 (ANXA1/2) are involved in membrane trafficking and exocytosis, transmembrane channel activity, phospholipase A2 inhibition and cell-matrix interactions (Raynal et al. 1994). Annexins are Ca(2+)-dependent phospholipid binding proteins which promote membrane fusion by aggregating intracellular vesicles and lipid rafts to the cytosolic surface of the plasma membrane. Following membrane injury, the disruption of annexin-dysferlin interaction mediates sarcolemma repair via a Ca(2+)-dependent aggregation and fusion of vesicles to the membrane (Lennon et al. 2003).

Beta-parvin/affixin (PARVB) is an actin and integrin-linked kinase–binding protein that plays a role in the regulation of cell adhesion and cytoskeleton organization. Dysferlin was shown to interact with beta-parvin at the sarcolemma (Matsuda et al. 2005).

AHNAK and AHNAK2 proteins are both localized in Z-band regions of cardiomyocytes and bind to sarcolemma and T-tubule membranes. AHNAK is thought to be involved in calcium homeostasis and in signal transduction at the cell membrane: in vitro, the AHNAK protein can produce diacylglycerol and inositol triphosphate by activating phospholipase C (Sekiya et al. 1999). In addition, predicted 2D-structure of AHNAK2 suggests a 3D-spatial organization similar to beta-propeller proteins such as RCC1, clathrin and G beta-subunit and indicates a role of AHNAK proteins as scaffolding proteins (Rosa and Barbacid 1997, Komuro et al. 2004).
AHNAK interacts with dysferlin C2A domain in a Ca(2+)-independent manner (Huang et al. 2007) and this interaction is regulated by the calpain-3 protease activity (Huang et al. 2008). Furthermore, in dysferlinopathies, dysferlin deficiency in muscles correlates with a secondary loss of AHNAK and both proteins show an increased localization in cytoplasm in rat regenerating muscles (Huang et al. 2007). Thus it has been hypothesized that dysferlin may recruit AHNAK/AHNAK2 to the plasma membrane.

Calpain-3 (CAPN3/LGMD2A). One of the most common LGMD2 forms is caused by loss of function mutations in the LGMD2A gene that affect its product calpain-3, a protein that belongs to the calcium-dependant cysteine protease family and that seems to play a central role for remodelling the sarcomeres (Duguez et al. 2006). Calpain-3 interacts with dysferlin (Huang et al. 2005) and its expression is reduced in dysferlin deficient muscle cells and vice versa. Moreover, AHNAK, another dysferlin-interacting protein is a substrate of calpain-3 and accumulates at the sarcolemma in CAPN3 deficient patients (Huang et al. 2008). It is thus proposed that calpain-3 may play an important role in cytoskeleton-membrane interactions. However, calpain-3 have not yet been proved to play a role in the early stage of membrane repair (Mellgren et al. 2009).

Schematic representation of structural and protein-protein interacting domains alongthe protein sequence (Show/Hide)

Domain Name	Short Name	Start	End
C2A domain	C2A	2	85
C2B domain	CBA	222	302
Ferlin family domain	Fer	303	374
C2C domain	C2C	381	479
Ferlin family domain	Fer	694	759
Ferlin family domain	Fer	786	861
Outer DysF domain, N-terminal region	DysFN	874	933
Inner DysF domain, N-terminal region	DysFN	946	1002
Inner DysF domain, C-terminal region	DysFC	1011	1049
Outer DysF domain, C-terminal region	DysFC	1068	1101
C2D domain	C2D	1154	1244
C2E domain	C2E	1335	1421
C2F domain	C2F	1580	1663
C2G domain	C2G	1813	1926
Transmembrane domain	TM	2045	2067

The UMD-DYSF mutations database The protein

The UMD-DYSF mutations database
The protein