RESUMO
The antimicrobial activity of the collective molecules comprising human milk reflects an evolutionarily successful paradigm for preventing and limiting microbial infection. Understanding the molecules that participate in this process and how they work can yield insight into potentially new antimicrobial therapies. Upon proteolytic processing, antimicrobial peptides can be derived from milk proteins, such as lactoferrin, casein, and lysozyme. Similarly, using the HIV-1 gp41 protein template, we have demonstrated that the 28-residue C-terminus, when produced as an independent peptide, exhibits selective toxicity for bacteria over eukaryotic cells. Upon optimizing this sequence for cationic charge and hydrophobic character presented as a alpha-helical structure, we show improved capability of the parent LLP1 sequence to selectively kill bacteria in the host environment and that this activity is increased by the inclusion of Trp residues on the hydrophobic face. We report that it is possible to (i) design de novo antimicrobial peptides that demonstrate optimal antimicrobial activity with minimal inflammatory activity and (ii) design antimicrobial peptides to function in a defined environment. In the end, we describe a de novo designed antimicrobial peptide, WLBU2, which is selectively toxic to microbial pathogens in complex environments and does not stimulate a significant immunomodulatory response. In spite of these properties, WLBU2 activity against Pseudomonas aeruginosa in human milk is inferior to the host peptide LL37 with regard to antimicrobial potency. These studies demonstrate that antimicrobial peptides can be engineered for greater potency in one medium but may not be optimal for working in a different medium such as human milk.