EBAMP: An efficient de novo broad-spectrum antimicrobial peptide discovery framework

Area of Science:

• Computational chemistry
• Microbiology
• Peptide design

Background:

• Designing novel antimicrobial peptides (AMPs) is complex due to the vast sequence space and poorly understood mechanisms of action.
• The rise of multidrug-resistant pathogens necessitates innovative strategies for antimicrobial drug discovery.

Purpose of the Study:

• To introduce EBAMP, a generative-discriminative framework for the de novo design of broad-spectrum antimicrobial peptides targeting both bacteria and fungi.
• To validate the efficacy and safety of EBAMP-designed AMPs through experimental testing and in vivo studies.

Main Methods:

• Utilized a Transformer-based generative model for exploring peptide sequence space.
• Implemented advanced feature-based screening for selecting multiobjective AMP candidates.
• Conducted experimental validation including in vitro antimicrobial assays, cytotoxicity, hemolysis tests, and in vivo wound infection models.
• Performed alanine substitution analysis and molecular dynamics simulations to identify critical functional residues.

Main Results:

• Out of 256 designed sequences, 96 (37.5%) demonstrated bactericidal activity.
• The top 10 designed AMPs exhibited potent activity (2 μg/mL) against multidrug-resistant bacteria and fungi with low cytotoxicity and hemolysis.
• In vivo studies showed significant inhibition of Acinetobacter baumannii and Candida auris infections in a mouse wound model.
• Comparative analysis indicated therapeutic efficiency comparable to conventional antibiotics but with a lower propensity for resistance development.

Conclusions:

• EBAMP is an effective framework for the de novo discovery of broad-spectrum antimicrobial peptides.
• The designed AMPs show promise as a new class of therapeutics to address the challenge of antibiotic resistance.
• Functional analysis provides insights into the mechanisms of action for future AMP optimization.