en.Wedoany.com Reported - A research team at the Perelman School of Medicine at the University of Pennsylvania has utilized artificial intelligence to discover a class of short peptides called "prionins" within prion proteins, traditionally associated with neurodegenerative diseases, which have the potential to kill bacteria. This finding expands the understanding of where antibiotics might be hidden and provides a new source of candidates for combating drug-resistant bacterial infections. The study has been published in Nature Microbiology.

For a long time, prions, these misfolded proteins, have been primarily linked to rare and fatal degenerative brain diseases. Previous sporadic studies have hinted that certain protein fragments associated with neurodegenerative diseases, such as beta-amyloid and cellular prion protein, may possess antimicrobial properties, but a large-scale systematic search was lacking. The University of Pennsylvania team used the deep learning platform APEX 1.1 to scan 19.3 million short peptide fragments from 2,897 prion and prion-like proteins, identifying 1,179 candidate antimicrobial peptides based on predicted antibiotic activity, which they named "prionins."

Based on the platform's predictions against 11 different bacterial pathogens (including drug-resistant strains), the research team selected 75 of the most promising peptides for experimental validation. The results showed that 59 of these peptides inhibited at least one bacterial pathogen, and 42 exhibited strong activity at low concentrations. Further experiments indicated that many active prionins exert their antibacterial effects by disrupting bacterial membranes, with low toxicity; 16 active peptides caused no measurable harm to human cells or red blood cells at the highest tested concentrations.

To verify practical efficacy, the researchers tested two prionins derived from fungi and roundworms in mouse models. In a standard skin infection model caused by the drug-resistant pathogen Acinetobacter baumannii, both peptides reduced bacterial levels, with effects comparable to polymyxin B, and no treatment-related weight loss was observed.

Senior author Dr. César de la Fuente (FRSB) noted that this work changes the perception of where antibiotics might be hidden, as artificial intelligence has uncovered the possibility of useful molecules encoded within prions, proteins long considered synonymous with disease. Co-first author Marcelo D. T. Torres stated that the AI search provided a shortlist of candidates, and the actual effectiveness of many molecules in laboratory and animal infection models makes this study a true discovery platform.

This research builds on the broader work of the de la Fuente lab in mining "encrypted peptides" from the biological world. The team has previously searched sources such as human proteins, extinct organisms, archaea, microbiomes, and venoms. This study extends the scope to prions and prion-like proteins, opening new possibilities at the intersection of neurodegenerative diseases and innate immunity. However, the researchers emphasize that this work does not prove that prionins are released during natural infections, nor does it alter the understanding of the pathogenic role of misfolded prions in neurodegenerative diseases. Its core significance lies in demonstrating that these proteins may be a rich and previously overlooked source of antibiotic candidates, offering a new direction for exploring the potential link between protein aggregation and host defense.

This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com