Conservation of three-dimensional structure in Lepidoptera and Trichoptera L-fibroins for 290 million years

Abstract

The divergence of sister orders, Trichoptera (caddisflies) and Lepidoptera (moths and butterflies), from a silk-spinning ancestor occurred around 290 million years ago. Trichoptera larvae are mainly aquatic and Lepidoptera larvae are almost entirely terrestrial, distinct habitats that required molecular adaptation of their silk for deployment in water versus air. The major protein components of their silks are Heavy chain and Light chain fibroins. In an effort to identify molecular changes in L-fibroins that may have contributed to the divergent use of silk in water versus air, we used the ColabFold implementation of AlphaFold2 to predict three-dimensional structures of L-fibroins from both orders. Comparison of the structures revealed that despite the ancient divergence, profoundly different habitats, and low sequence conservation, a 10-helix core structure was strongly conserved in L-fibroins from both orders. Previously known intra- and intermolecular disulfide linkages were accurately predicted. Structure variations observed outside of the core may represent molecular changes that contributed to the evolution of silks adapted to water or air. The distributions of electrostatic potential, for example, are not conserved and present distinct order-specific surfaces for potential interactions with or modulation by external factors. Also, the interactions of L-fibroins with the H-fibroin C-termini are different between the orders; lepidopteran L-fibroins have N-terminal insertions, not present in trichopteran L-fibroins, which form an unstructured ribbon in isolation but become part of an intermolecular {beta}-sheet when folded with their corresponding H-fibroin C-termini. The results may serve as another example of protein structure prediction from deep sequence data of understudied proteins made possible by AlphaFold2.