Computational and In-Vitro Evaluation of Mucin-Degrading Enzymes for Mucosal Clearance in Cystic Fibrosis

Abstract

Cystic fibrosis airway disease is driven by accumulation of viscous mucus in which the gel-forming mucins MUC5AC and MUC5B determine viscoelasticity. Current mucolytics do not degrade the mucin scaffold, leaving an unmet need for agents acting on the glycoprotein itself. We screened twenty-four candidate mucin-degrading enzymes against four mucin target regions using AlphaFold 3, and evaluated representative enzymes experimentally. Sequences were retrieved programmatically from UniProt and submitted through AlphaFold Server; complexes were ranked by interface predicted TM-score (ipTM). The mucin-selective metalloprotease StcE was included as a positive control. Mucolytic activity was measured against porcine gastric mucin by inclined-slide gravity flow, in triplicate, with N-acetylcysteine as reference standard. None of 101 predicted complexes reached the ipTM threshold of 0.60, and StcE scored 0.32, indicating that the method failed to recognise a known interaction. Predicted confidence segregated by glycan dependence: enzymes cleaving the peptide backbone averaged 0.43 whereas glycan-recognising enzymes averaged 0.22–0.27. Experimentally, all agents reduced flow time dose-dependently. N-acetylcysteine achieved 86.2% reduction, papain 26.1% and α-fucosidase 17.8% (all p<0.001). The protease exceeded the glycosidase significantly (p<0.0001), mirroring the computational ordering. We conclude that protein-only structure prediction cannot model enzyme–mucin association because recognition is glycan-mediated, while confirming that proteolytic degradation reduces mucin viscosity more effectively than glycosidic degradation under these conditions.