Deciphering Features of Metalloprotease Cleavage Targets Using Protein Structure Prediction

IntroductionADAM10 (A Disintegrin and Metalloproteinase 10) cleaves specific substrates, influencing diverse physiological and pathological processes. However, its substrate specificity and cleavage sites remain insufficiently characterized. This study aimed to identify and classify substrate features and elucidate cleavage sites using computational approaches. MethodsProtein structure prediction was performed on 13 substrates with experimentally defined cleavage sites to analyze ADAM10-substrate interactions and assess their structural and spatial characteristics. ResultsBased on protein structure prediction of experimentally validated substrates, we identified four recurrent structural features associated with ADAM10-mediated cleavage. First, most substrates were predicted to interact with the proteolytically active form of ADAM10 (92.3%). Second, interaction sites were typically located in the extracellular region rather than intracellular or transmembrane domains (76.9%). Third, cleavage sites predominantly resided in unstructured or extended loop regions, corresponding to linear secondary structures ([~]70%). Lastly, cleavage sites were spatially enriched in octants (1, 4, 5, and 8) relative to the catalytic Zn{superscript 2}{square} ion in the active site (84.0%). These percentages were calculated per determinant. On the basis of these determinants, we developed a classification algorithm and organized 51 substrates accordingly. As a result, 82.4% of the substrates were assigned to either Group 1 (51.0%) or Group 2 (31.4%). DiscussionThis study presents a structure-informed classification approach for predicting ADAM10 cleavage substrates, enabling candidate identification without direct experimental validation. ConclusionsWe present a novel structure-based framework for classifying the substrates of ADAM10 and demonstrate its applicability to other Metalloproteases.