Polyphemusin-1

The cyclic beta-sheet structure possessed by the 10-residue antibiotic peptide gramicidin S was taken as the structural framework for the de novo design of biologically active peptides with membrane-active properties. We have shown from previous studies that gramicidin S is a broad-spectrum antibiotic effective against Gram-positive bacteria, Gram-negative bacteria, and fungi, but is toxic to human red blood cells. We tested the effect of ring size on antimicrobial activity and hemolytic activity on peptides varying from 4 to 16 residues. Interestingly, we were able to dissociate hemolytic activity and antimicrobial activity by increasing the ring size of the peptide to 14 residues (peptide GS14). Furthermore, we increased specificity for microbial membranes while decreasing toxicity to red blood cells by substituting enantiomers (D-amino acids for L-amino acids and vice versa) into the GS14 sequence. The enantiomeric substitutions all disrupted beta-sheet structure in benign medium and decreased peptide amphipathicity. The least amphipathic peptide, produced by substituting a D-Lys at position 4 of GS14 (peptide GS14K4), also had the highest therapeutic index, i.e., highest degree of specificity for microbial cells over human cells. Solution structures of GS14 analogs solved by NMR spectroscopy showed that the D-amino acid side chain was located on the nonpolar face of GS14K4. Another analog, a beta-sheet peptide with reduced amphipathicity (peptide GS14 K3L4), also had a lysine (lysine 3) on the nonpolar face as determined by the NMR structure. Both GS14K4 and GS14 K3L4 had reduced amphipathicity relative to GS14 and much higher therapeutic indices. Finally, the alteration of the nonpolar face hydrophobicity of GS14K4 analogs provided a range of activities and specificities, where the peptides with the intermediate hydrophobicities among the series had the highest therapeutic indices. The optimal peptide hydrophobicities varied depending on the microorganism being tested, with higher hydrophobicity requirements against Gram-positive bacteria and yeast compared with Gram-negative microorganisms. The net result of these studies suggests that it is possible to rationally design a cyclic membrane-active antimicrobial peptide with high specificity towards prokaryotic (bacterial and fungal) membranes and minimal toxicity to eukaryotic (e.g., mammalian) membranes.

Antimicrobial peptides (AMPs) with amphipathic β-hairpin structures have been demonstrated to possess potent antimicrobial activities and great cell selectivities. However, our understanding of β-hairpin antimicrobial peptides lags behind that of α-helices, mainly because it is difficult for short peptides to form robust β-hairpin structures. Tryptophan zipper (trpzip) peptides are among the most stable β-hairpin peptides known to fold spontaneously without requiring covalent disulfide constraint or metal binding. To develop model β-hairpin AMPs with small size and remarkable stability, a series of amphiphilic linear peptides were designed based on the trpzip motif. The sequence of designed peptides is (WK) n (D) PG(KW) n -NH2 (n = 1, 2, 3, 4, 5), and the antimicrobial activity and membrane interaction mechanism of the peptides were evaluated. The results showed that these peptides readily fold into β-hairpin structures in aqueous and membrane-mimicking environments and exhibit broad-spectrum antimicrobial activities against both gram-positive and gram-negative bacteria. The antibacterial potency of the peptides initially increased and then decreased with increasing chain length. WK3, a 14-residue peptide, displayed excellent antimicrobial activity with minimal hemolytic activity and cytotoxicity, suggesting that it possesses great cell selectivity. Scanning electron microscopy (SEM), transmission electron microscopy (TEM), fluorescence spectroscopy, and flow cytometry indicated that representative peptides WK3 and WK4 exert their activities by permeabilizing the microbial membrane and damaging cell membrane integrity. This study reveals the application potential of the designed peptides as promising antimicrobial agents for the control of infectious diseases, and it also provides new insights into the design and optimization of highly stable β-hairpin AMPs with great antimicrobial activities and cell selectivities.

Tachyplesin-1 (TP1; 1) is a cationic β-hairpin antimicrobial peptide with a membranolytic mechanism of action. While it possesses broad-spectrum, potent antimicrobial activity, 1 is highly hemolytic against mammalian erythrocytes, which precludes it from further development. In this study, we report a template-based approach to investigate the structure-function and structure-toxicity relationships of each amino acid of 1. We modulated charge and hydrophobicity by residue modification and truncation of the peptide. Antimicrobial activity was then assessed against six key bacterial pathogens and two fungi, with toxicity profiled against mammalian cells. The internal disulfide bridge Cys7-Cys12 of 1 was shown to play an important role in broad-spectrum antimicrobial activity against all pathogenic strains tested. Novel peptides based on the progenitor were then designed, including 5 (TP1[F4A]), 12 (TP1[I11A]), and 19 (TP1[C3A,C16A]). These had 26- to 64-fold improved activity/toxicity indices and show promise for further development. Structural studies of 5 (TP1[F4A]) and 12 (TP1[I11A]) identified a conserved β-hairpin secondary structure motif correlating with their very high stablility in mouse and human plasma. Membrane binding affinity determined by surface plasmon resonance confirmed their selectivity toward bacterial membranes, but the degree of membrane binding did not correlate with the degree of hemolysis, suggesting that other factors may drive toxicity.