Application of Rosetta and Foldit for Structural Modeling of DYSF

David Baker, PhD

University of Washington (Seattle, Washington)

David Baker, PhD is the Head of the Institute for Protein Design and a Professor of Biochemistry at the University of Washington (Seattle, Washington)

Past Projects

Autosomal recessive mutations in the human gene encoding the dysferlin protein (DYSF) are the cause of rare forms of muscular dystrophies known as dysferlinopathies, and more specifically, as limb girdle muscular dystrophy type 2B (LGMD2B) or Miyoshi myopathy (MM).  Dysferlin is a 237 kDa type II transmembrane skeletal muscle protein associated with the sarcolemma, where its function appears to be required for efficient muscle contraction and muscle membrane repair. Dysferlin is a member of the ferlin gene family of which all members consist of multiple tandem cytosolic C2 domains and a C-terminal transmembrane domain. Dysferlin contains seven calcium sensitive C2 domains that associate with lipid membranes in a calcium-dependent manner and seem to play a role in calcium-mediated membrane fusion events.  Several dysferlin protein binding partners have been identified including the protein caveolin-3, a skeletal muscle membrane protein important in the formation of caveolae.  In all diagnosed cases of LGMD2B / MM, where the amount of cellular dysferlin protein is <20% than that of normal people, the phenotype is pathogenic and is always caused by primary dysferlin gene mutations.  

The dysferlin gene spans 233,140 base pairs of genomic DNA and generates a 6.9 kb-wide transcript with 55 exons.  Numerous studies of LGMD2B / MM patients have identified a wide variety of mutations riddled throughout the gene including missense (~43%), nonsense (~14%), splice site (~8%), frameshift (~30.5%), and in rare instances non-stop mutations (~2%) or gene rearrangements (3%).  In order to gain a better understanding of the structure and function of DYSF, we propose to conduct Rosetta structural modeling of the full length and splice isoforms of DYSF (Aim 1a).  We will also model possible protein-protein interactions of DYSF domains modeled in Aim 1a with two (2) other proteins known to interact with DYSF and whose structures are known or can be reliably modeled (Aim 1b).  We will also map ~200 of the known LGMD2B and MM missense mutations onto the modeled DYSF structures of Aim 1b and predict their effects on the structure and stability of DYSF (Aim 2). Finally, we will engage the Foldit citizen scientist community in DYSF related puzzles to educate the community on dysferlinopathies and obtain human intuition insights on the globular or extended structure of DYSF to facilitate Aims 1-2 (Aim 3).