Predicting antimicrobial peptides from eukaryotic genomes: in silico strategies to develop antibiotics.

A remarkable and intriguing challenge for the modern medicine consists in the development of alternative therapies to avoid the problem of microbial resistance. The cationic antimicrobial peptides present a promise to be used to develop more efficient drugs applied to human health. The in silico analysis of genomic databases is a strategy utilized to predict peptides of therapeutic interest. Once the main antimicrobial peptides' physical-chemical properties are already known, the correlation of those features to search on these databases is a tool to shorten identifying new antibiotics. This study reports the identification of antimicrobial peptides by theoretical analyses by scanning the Paracoccidioides brasiliensis transcriptome and the human genome databases. The identified sequences were synthesized and investigated for hemocompatibility and also antimicrobial activity. Two peptides presented antifungal activity against Candida albicans. Furthermore, three peptides exhibited antibacterial effects against Staphylococcus aureus and Escherichia coli; finally one of them presented high potential to kill both pathogens with superior activity in comparison to chloramphenicol. None of them showed toxicity to mammalian cells. In silico structural analyses were performed in order to better understand function-structure relation, clearly demonstrating the necessity of cationic peptide surfaces and the exposition of hydrophobic amino acid residues. In summary, our results suggest that the use of computational programs in order to identify and evaluate antimicrobial peptides from genomic databases is a remarkable tool that could be used to abbreviate the search of peptides with biotechnological potential from natural resources.