Structure and topology of the linkers in the conserved lepidosaur β-keratin chain with four 34-residue repeats support an interfilament role for the central linker

The beta-keratin chain with four 34-residue repeats that is conserved across the lepidosaurs (lizards, snakes and tuatara) contains three linker regions as well as a short, conserved N-terminal domain and a longer, more variable C-terminal domain. Earlier modelling had shown that only six classes of structure involving the four 34-residue repeats were possible. In three of these the 34-residue repeats were confined to a single filament (Classes 1, 2 and 3) whereas in the remaining three classes the repeats lay in two, three or four filaments, with some of the linkers forming interfilament connections (Classes 4, 5 and 6). In this work the members of each class of structure (a total of 20 arrangements) have been described and a comparison has been made of the topologies of each of the linker regions. This provides new constraints on the structure of the chain as a whole. Also, analysis of the sequences of the three linker regions has revealed that the central linker (and only the central linker) contains four short regions displaying a distinctive dipeptide repeat of the form (S-X)(2,3) separated by short regions containing proline and cysteine residues. By analogy with silk fibroin proteins this has the capability of forming a beta-sheet-like conformation. Using the topology and sequence data the evidence suggests that the four 34-residue repeat chain adopts a Class 4a structure with a beta-sandwich in filament 1 connected through the central linker to a beta-sandwich in filament 2.