Computational design of highly selective antimicrobial peptides

Area of Science:

• Biochemistry
• Computational Biology
• Pharmacology

Background:

• Antimicrobial peptides (AMPs) are crucial for innate immunity.
• Developing selective AMPs with high therapeutic indices (TI) is a significant challenge.
• Existing AMP databases lack comprehensive structure-selectivity relationships.

Purpose of the Study:

• To create a structure-selectivity database (AMPad) for frog-derived AMPs.
• To develop a predictive model for AMP selectivity using sequence moments.
• To design and synthesize novel, highly selective AMPs.

Main Methods:

• Construction of the AMPad database integrating peptide sequence, structure, and therapeutic index (TI).
• Application of sequence moments and the D-descriptor to analyze lengthwise asymmetry of physicochemical properties.
• Development of the AMP-Designer algorithm incorporating D-descriptor, amino acid selectivity, and motif regularity.

Main Results:

• The D-descriptor effectively identifies highly selective antimicrobial peptides.
• The AMP-Designer algorithm successfully predicted TI for known and de novo designed AMPs.
• A novel peptide, adepantin 1, exhibited potent antibacterial activity (vs. E. coli) and a high TI (>200), surpassing existing AMPs in selectivity.

Conclusions:

• Sequence moments and the D-descriptor offer a powerful approach for designing selective AMPs.
• The AMP-Designer algorithm represents a significant advancement in rational AMP design.
• Adepantin 1 demonstrates the potential of algorithm-driven design for novel therapeutics with improved safety profiles.