Redefining the architecture of ferlin proteins: insights into multi-domain protein structure and function

Abstract

Ferlins are complex, multidomain proteins whose functions involve membrane fusion, membrane repair, and exocytosis. Progress in studying the domain composition of ferlins has been problematic due to the large size of the ferlin proteins and the difficulty in computing accurate residue ranges for each respective ferlin protein domain. However, recent advances in in silico protein folding methods have significantly enhanced the understanding of the domain structure of these complex proteins. To compute unbiased domain boundaries, we used RoseTTAFold to assemble full-length models for each of the six human ferlin proteins (dysferlin, myoferlin, otoferlin, Fer1L4, Fer1L5, and Fer1L6). Despite the differences in amino acid sequence between the ferlin proteins, the domain ranges and the distinct subdomains in some of the ferlin domains were remarkably consistent. Further, the RoseTTAFold/AlphaFold2 in silico boundary prediction methods allowed us to describe and characterize a previously unknown C2 domain, ubiquitous in all human ferlins, which we refer to as C2-FerA. At present, the ferlin domain-domain interactions implied by the full-length in silico models are predicted to have a low accuracy; however, the use of RoseTTAFold and AlphaFold2 as a domain finder has proven to be a powerful research tool in understanding ferlin structure.