N-MIX: an in silico framework for predicting ADAM10-mediated substrate cleavage sites through structural and spatial...

ADAM10 is a membrane-anchored metalloprotease that mediates ectodomain shedding of diverse substrates implicated in cancer, inflammation, and neurodegeneration. Despite its biological significance, ADAM10 substrate specificity remains poorly understood, as conventional sequence-based analyses have proven insufficient to identify conserved cleavage motifs. Through three-dimensional structural analysis of ADAM10-substrate complexes predicted by AlphaFold2, we identified four key determinants governing ADAM10-mediated proteolytic cleavage: (1) conserved hydrogen bond interactions at the enzyme-substrate interface mediated by the GLY329-SER340 region, (2) extracellular localization of cleavage sites, (3) flexible loop conformations with limited hydrogen-bond stabilization, and (4) defined spatial orientation relative to the catalytic zinc ion. These findings establish that ADAM10 cleavage specificity is governed by three-dimensional conformational context rather than primary sequence alone. Based on these determinants, we developed Novel-Metalloproteinase Interaction eXcision (N-MIX), a structure-based computational framework for prioritizing ADAM10-compatible cleavage regions. N-MIX classified 50 ADAM10 substrates into structurally defined groups and prioritized candidate cleavage regions, showing 62.9% Top-1 concordance and 92.6% Top-5 coverage against published experimental evidence. N-MIX provides a scalable, accessible platform for structure-based prioritization of ADAM10-compatible candidate cleavage regions, offering significant potential for AI-driven protease research and anticancer therapeutic target discovery.