Combined computational and rational design of α-helical antimicrobial peptides active against opportunistic pathogens

Given the current global problem of multidrug resistance in human opportunistic pathogens, antimicrobial peptides (AMPs) show significant potential as novel anti-infective agents1. Notably, they show broad spectrum antibacterial activity and their most common mode of action, namely the disruption of bacterial membrane and/or the formation of pores or lesions, makes it hard for bacteria to gender permanent resistance. The toxicity towards host cells is however a key limitation factor, which imposes the necessity to design artificial AMPs de novo, or redesigning/mutating known natural AMPs to have an increased selectivity index (SI) and reduce side effects. We report four novel Gly and Lys rich peptides which were computationally designed based on a set of training natural peptides with low reported MIC values against E.coli and taking into account a defined set of parameters ; net charge, hydrophobicity, number of Lysines, etc. One additional peptide was designed by creating the specular image of a previously designed sequence. All peptides were named kiadins, from the first three residues in the ‘parent’ peptide sequence1. Susceptibility testing was performed on clinically relevant, drug- resistant bacteria chosen among the ESKAPE pathogens ; carbapenem-resistant Acinetobacter baumannii and Pseudomonas aeruginosa, 3rd generation cephalosporin-resistant E. coli and Klebsiella pneumoniae, as well as meticillin- resistant Staphylococcus aureus (MRSA). The results show different range of antibacterial activity with respect to Gly substitutions, where the peptides with higher proportion of Gly showed significantly weaker bactericidal effects. The best antibacterial activity was observed for Kiadin-2, which exhibited MIC values from 0.25 to 8 µM, a promising SI values and low cytogenotoxicity on HPBLs, making it a suitable candidate for further development.