Evaluation of LLM-generated peptide as foundation template for discovery of effective encrypted AMPs against clinical...

Area of Science:

• Microbiology
• Biotechnology
• Computational Biology

Background:

• The rise of multidrug-resistant (MDR) pathogens, including carbapenem-resistant Acinetobacter baumannii (CRAB) and methicillin-resistant Staphylococcus aureus (MRSA), necessitates the urgent discovery of novel antimicrobial agents.
• Existing antimicrobial therapies face limitations due to widespread resistance, posing a significant global health threat.

Purpose of the Study:

• To develop and apply an innovative strategy combining generative large language models (LLMs) with sequence alignment for the rapid identification of novel antimicrobial peptides (AMPs).
• To evaluate the antimicrobial activity, therapeutic efficacy, and mechanism of action of newly discovered AMPs against clinically relevant MDR pathogens.

Main Methods:

• Utilized a generative LLM to design peptide templates, followed by sequence alignment to identify novel encrypted peptides.
• Screened identified peptides for antimicrobial activity against a panel of clinical MDR pathogens.
• Conducted in vivo studies to assess the therapeutic efficacy of lead candidates, such as PL-15, against CRAB infections.
• Performed mechanistic investigations to elucidate the bactericidal effects of PL-15, including membrane disruption, depolarization, and reactive oxygen species (ROS) generation.

Main Results:

• Successfully identified five novel encrypted peptides with significant antimicrobial activity against a broad spectrum of MDR pathogens.
• The peptide PL-15 demonstrated potent, broad-spectrum antimicrobial activity and comparable therapeutic efficacy to polymyxin B against CRAB infections in vivo.
• PL-15 was found to exert bactericidal effects by disrupting bacterial membranes, causing depolarization, increasing ROS levels, and inhibiting biofilm formation.
• PL-15 also exhibited significant antibiofilm activity, preventing new biofilm formation and eradicating established biofilms.

Conclusions:

• The combination of generative LLMs and sequence alignment is a powerful and accelerated approach for discovering novel antimicrobial peptides.
• PL-15 represents a promising candidate for further development as a therapeutic agent against challenging MDR bacterial infections.
• The findings underscore the potential of this integrated computational and experimental strategy to address the critical need for new antibiotics.