engleski

Designing peptide antibiotics with potential medical applications

In order to distinguish peptide antibiotics from peptide toxins and hemolytics we tried to learn as much as possible from natural design of antimicrobial peptides isolated from different frog species. About 50 such peptides sequences were collected from the literature, all of them having known hemolytic and antibacterial activitiy. All had linear sequences reported to form helices in organic solvents. Eliminating all pairs with more than 70% sequence identity left us with 37 nonhomologous peptides that had mainly antibiotic activity. Few remaining peptides with high hemolytic activity can be easily recognized and eliminated by using the threshold level for membrane buried helix preferences, because hemolytics have much higher preference than antibiotics to enter deeper into the bilayer membrane. The therapeutic index of antibiotic peptides depended strongly on amino acid composition of their sequence, on hydrophobic moment profile for the formation of an amphipathic helix and on profile for the formation of membrane-buried helix. Averaged amino-acid attributes for the whole sequence were inferior to sequence-dependent profiles of amino-acid attributes in our attempts to estimate therapeutic index. The SPLIT 3.5 server (split.pmfst.hr/split) is now supplied with the AMP link for predicting the therapeutic index of antimicrobial peptides. That link can be used for rational design of novel selective peptide antibiotics.

peptide antibiotics; antibacterial peptides; hemolytic activity; antibiotics design; SPLIT server

The scientific program and abstracts of invited lectures for the conference: The 1st Split Meeting on Development and Applications of Novel Methods and Models in Computational Biophysics and Structural Bioinformatics are available on-line at the: http://www.pmfst.hr/se/z_fizika/konf/