Evolutionary and structural analyses uncover a role for solvent interactions in the diversification of cocoonases in butterflies.

Multi-omic approaches promise to supply the power to detect genes underlying disease and fitness-related phenotypes. Optimal use of the resulting profusion of data requires detailed investigation of individual candidate genes, a challenging proposition. Here, we combine transcriptomic and genomic data with molecular modelling of candidate enzymes to characterize the evolutionary history and function of the serine protease cocoonase. butterflies possess the unique ability to feed on pollen; recent work has identified as a candidate gene in pollen digestion. was first described in moths, where it aids in eclosure from the cocoon and is present as a single copy gene. In heliconiine butterflies it is duplicated and highly expressed in the mouthparts of adults. At least six copies of are present in and copy number varies across sub-populations. Most genes are under purifying selection, however branch-site analyses suggest genes may have evolved under episodic diversifying selection. Molecular modelling of cocoonase proteins and examination of their predicted structures revealed that the active site region of each type has a similar structure to trypsin, with the same predicted substrate specificity across types. Variation among heliconiine cocoonases instead lies in the outward-facing residues involved in solvent interaction. Thus, the neofunctionalization of duplicates appears to have resulted from the need for these serine proteases to operate in diverse biochemical environments. We suggest that may have played a buffering role in feeding during the diversification of across the neotropics by enabling these butterflies to digest protein from a range of biochemical milieux.