Collagen-derived cryptides : machine-learning prediction and molecular dynamic interaction against Klebsiella pneumoniae biofilm synthesis precursor

Abstract

Collagen-derived cryptic peptides (cryptides) are biologically active peptides derived from the proteolytic digestion of collagen protein. These cryptides possess a multitude of activities, including antihypertensive, antiproliferative, and antibacterial. The latter, however, has not been extensively studied. The cryptides are mainly obtained from the protein hydrolysate, followed by characterizations to elucidate the function, limiting the number of cryptides investigated within a short period. The recent threat of antimicrobial resistance microorganisms (AMR) to global health requires the rapid development of new therapeutic drugs. The current study aims to predict antimicrobial peptides (AMP) from collagen-derived cryptides, followed by elucidating their potential to inhibit biofilm-related precursors in Klebsiella pneumoniae using in silico approach. Therefore, cryptides derived from collagen amino acid sequences of various types and species were subjected to online machine-learning platforms (i.e., CAMPr3, DBAASP, dPABBs, Hemopred, and ToxinPred). The peptide-protein interaction was elucidated using molecular docking, molecular dynamics, and MM-PBSA analysis against MrkH, a K. pneumoniae’s transcriptional regulator of type 3 fimbriae that promote biofilm formation. As a result, six potential antibiofilm inhibitory cryptides were screened and docked against MrkH. All six peptides bind stronger than the MrkH ligand (c-di-GMP; C2E).