Divergent Loop Architecture Shapes Pocket2 Variation in Shark Legumains

Abstract

Legumain (AEP) is a cysteine protease with a highly conserved catalytic core but variable surface features whose evolutionary and structural diversity remain incompletely understood. To investigate how these features differ across vertebrates, we compared human legumain with six shark orthologs using sequence alignment, AlphaFold based structural modeling, pocket detection, and qualitative docking. All shark sequences retained the canonical His-Cys dyad and {beta} sandwich fold, and structural superposition revealed strong global conservation (RMSD = 0.39 [A]). A single surface exposed loop adjacent to a shallow cavity--designated Pocket 2--displayed pronounced sequence divergence. Structural models reproduced this pattern, with Pocket 2 showing the greatest variation in geometry and residue composition across species, including Callorhinchus milii, which clustered with elasmobranchs in Pocket 2 features. Although loop conformations were predicted with lower confidence, Pocket 2 was consistently detected in all models and exhibited interspecific differences in volume, shape, and physicochemical environment. Docking of a structurally characterized reference ligand (5KN) into Pocket 2 revealed species specific differences in modeled binding orientations and interaction patterns that were consistent with these geometric variations, though not interpretable as quantitative affinity predictions. Together, these results identify Pocket 2 as a structurally conserved but locally variable region of legumain and highlight it as a candidate site for future experimental and computational investigation. This study is hypothesis generating and provides a framework for examining how localized structural variation may arise within an otherwise conserved protease family.