Statistical modeling of in vitro pepsin specificity.

The specificity of pepsin, the major protease of gastric digestion, has been previously investigated, but only regarding the primary sequence of the protein substrates. The present study aimed to consider in addition physicochemical and structural characteristics, at the molecular and sub-molecular scales. For six different proteins submitted to in vitro gastric digestion, the peptide bonds cleaved were determined from the peptides released and identified by LC-MS/MS. An original statistical approach, based on propensity scores calculated for each amino acid residue on both sides of the peptide bonds, concluded that preferential cleavage occurred after Leu and Phe, and before Ile. Moreover, reliable statistical models developed for predicting peptide bond cleavage, highlighted the predominant role of the amino acid residues at the N-terminal side of the peptide bonds, up to the seventh position (P7 and P7'). The significant influence of hydrophobicity, charge and structural constraints around the peptide bonds was also evidenced.