In silico and in vitro evaluation of dermaseptin for antibacterial activity in resistant strains of S. aureus

Abstract

The current crisis of antibiotic-resistant bacteria such as Methicillin-resistant Staphylococcus aureus (MRSA) causes the need for a faster and more efficient methodology for drug discovery. Deep learning approaches show promising results for Virtual Screening (VS). Herein we perform an in silico and in vitro evaluation of a new peptide from the dermaseptin family for anti-MRSA activity. We perform VS using two neural networks: AMP-Net, which predicts antimicrobial properties and PLA-Net, which predicts interactions with human cell receptors. Moreover, we perform an antibacterial microdilution assay with two multi-resistant and one reference strain. The in silico screening showed that the peptide has a strong antibacterial and anti-tumorigenic effect. Also, the antibacterial assay shows a minimum inhibitory concentration of > 12.5 uM in MRSA strains, which presents the same efficiency as cefalexin and a higher efficiency than common antibiotics such as ampicillin and vancomycin. We conclude that this research shows the potential of VS to speed up the drug discovery and repurposing process. Finally, we state that further research should aim to assess the multi-functionality of this peptide as a potential anti-carcinogenic agent.