A class of highly potent antibacterial peptides derived from pardaxin, a pore-forming peptide isolated from Moses sole fish Pardachirus marmoratus

Pardaxin, a 33-amino-acid pore-forming polypetide toxin isolated from the Red Sea Moses sole Pardachirus marmoratus, has a helix-hinge-helix structure. This is a common structural motif found both in antibacterial peptides that can act selectively on bacterial membranes (e.g., cecropin), and in cytotoxic peptides that can lyse both mammalian and bacterial cells (e.g., melittin). Herein we show that pardaxin possesses a high antibacterial activity with a significantly reduced hemolytic activity towards human red blood cells (hRBC), compared with melittin. Its potency is comparable to that of other known native antibacterial peptides such as magainin, cecropins and dermaseptins. To determine the structural features responsible for the selective hemolytic and antibacterial activities, and the structural requirements for a high antibacterial activity, 8 truncated and modified pardaxin analogues were synthesized and structurally and functionally characterized. Each peptide was synthesized with a free carboxylate or amino group (i.e., aminated form) at its C-terminus. The aminated form of pardaxin has both high hemolytic and antibacterial activity. A truncated analogue, with 11 amino acids removed from the C-terminal domain, had dramatically reduced hemolytic activity. However, the aminated form of this analogue was significantly more potent than pardaxin against most bacteria tested, suggesting that the C-terminal tail of pardaxin is responsible for non-selective activity against erythrocytes and bacteria. Furthermore, a positive charge added to its N-terminus significantly increased its antibacterial activity and abolished its low hemolytic activity. The 22-amino-acid C-terminal domain and the short II-amino-acid N-terminal domain were, in their aminated forms, active only against gram-positive bacteria. Secondary-structure determination using circular dichroism spectroscopy revealed that all the aminated analogues had 25-80% more α-helical content in 40% CF₃CH₂OH/water than their non-aminated forms. Using model phospholipid membranes it was found that all the analogues that were less hemolytic but had retained antibacterial activity could permeate acidicly charged phospholipid vesicles better than zwitterionic phospholipid vesicles, a property characteristics of all the native antibacterial peptides tested so far (e.g., cecropins, magainins and dermaseptins). Pardaxin and its analogues therefore represent a new class of antibacterial peptides that can serve as a basis for the design of therapeutic agents. Furthermore, negative-staining electron microscopy revealed that total inhibition of bacterial growth was due to total lysis of the bacterial wall. Therefore, it might be more difficult for bacteria to develop resistance to such a destructive mechanism, compared with the more specific mechanisms of the currently used antibiotics.