Antimicrobial Peptides for Therapeutic Applications: A Review

Antimicrobial peptides (AMPs) have been considered as potential therapeutic sources of future antibiotics because of their broad-spectrum activities and different mechanisms of action compared to conventional antibiotics. Although AMPs possess considerable benefits as new generation antibiotics, their clinical and commercial development still have some limitations, such as potential toxicity, susceptibility to proteases, and high cost of peptide production. In order to overcome those obstacles, extensive efforts have been carried out. For instance, unusual amino acids or peptido-mimetics are introduced to avoid the proteolytic degradation and the design of short peptides retaining antimicrobial activities is proposed as a solution for the cost issue. In this review, we focus on small peptides, especially those with less than twelve amino acids, and provide an overview of the relationships between their three-dimensional structures and antimicrobial activities. The efforts to develop highly active AMPs with shorter sequences are also described.

AMPs can be classified into four groups based on their structures: α-helical peptides, β-sheet peptides, extended peptides, and loop peptides [ 1 , 2 , 3 , 6 ]. The α-helical AMPs, including magainin, cecropin, and pexiganan, constitute a representative class of AMPs that are the most well established in structure-activity relationships. This group of peptides is usually unstructured in aqueous solution and forms amphipathic helices in membranes or membrane-mimicking environments. Most α-helical AMPs disrupt bacterial membranes, and several mechanisms of action employed by various AMPs have been proposed. The α-helical amphipathic peptides form barrel-like bundles in the bacterial membranes, and these transmembrane clusters line amphipathic pores (barrel-stave model). Many α-helical AMPs, including some magainins and cecropins, can dirupt bacterial membranes by forming carpet-like clusters of peptides. The peptides adsorb and align in parallel to the surface of bacterial membranes, then the membranes are collapsed into micelle-like structures by high concentrations of peptides (carpet model). The α-helical AMPs, such as some magainins and protegrins, form toroidal pores to disrupt the bacterial membranes (toroidal pore model) [ 1 , 2 , 3 , 4 , 5 , 6 , 7 , 9 , 11 ]. The β-sheet AMPs, such as α-, β-defensins, and protegrin, are stabilized by disulfide bridges, and form relatively rigid structures. Many of β-sheet AMPs exert their antimicrobial activities by disrupting bacterial membranes. They are perpendicularly inserted or tilted into the lipid bilayer to form toroidal pores, and hydrophilic regions of the peptides are associated with the polar head groups of the membranes [ 3 ]. The extended AMPs, which are predominantly rich in specific amino acids such as proline, tryptophan, arginine, and histidine, have no regular secondary structure elements. Indolicidin is a tryptophan/proline-rich extended peptide and Bac5 and Bac7 are proline/arginine-rich peptides [ 17 , 18 ]. Many extended AMPs are not active against the membranes of pathogens, but they can achieve their antimicrobial activities by penetrating across the membranes and interacting with bacterial proteins inside [ 3 ]. On the other hand, some extended peptides, such as indolicidin, are membrane active and induce membrane leakage. Indolicidin is a 13-amino acid AMP containing five tryptophan and three proline residues. The peptide adopts a poly-L-II helical structure in the presence of liposomes, and the high content of tryptophan residues is responsible for the interaction with lipid membranes [ 17 ]. The loop AMPs, including bactenecin, adopt a loop formation with one disulfide bridge. Understanding the structure-activity relationships (SAR) of AMPs is essential for the design and development of novel antimicrobial agents with improved properties. In particular, the atomic level structures of AMPs can provide versatile information for all stages of drug development, including the peptide design and modification for pharmaceutical application. Pexiganan (also known as MSI-78), a synthetic variant of magainin 2, is one of the best-investigated AMPs in terms of drug development [ 19 , 20 ]. Pexiganan has reached clinical trials as a novel topical broad-spectrum antibiotic for the treatment of mild-to-moderate diabetic foot ulcer infections [ 4 , 20 , 21 ]. The three-dimensional structure of pexiganan, determined by Nuclear Magnetic Resonance (NMR) spectroscopy, revealed that the peptide forms a dimeric antiparallel α-helical structure in the presence of membrane mimetics [ 22 , 23 ]. The atomic resolution structure also demonstrated that the side chains of three phenylalanine residues are important for the self-dimerization [ 23 ]. In addition to the peptide structure, its orientation in membrane is critical to understand the exact mode of membrane interaction. A solid-state NMR study of pexiganan suggested that the peptide adopts a helical conformation interacting with the membrane surface and then the dimeric peptide is inserted into the membrane [ 22 ]. Knowledge on the functional structures of AMPs also enables a rational design of synthetic model peptides. In this respect, minimalistic de novo approaches to design model amphipathic helical peptides are noteworthy. For example, certain 14- or 15-mer peptides (namely LK peptides) composed of two kinds of amino acids (leucines and lysines), have been characterized as possessing strong antimicrobial activity [ 24 , 25 ]. Then, by introducing a single tryptophan residue at a specific position, 9- to 11-mer LKW peptides could exhibit antimicrobial activity. Now, the de novo designed model peptides remain to be further improved by optimizing the amino acid sequence to enhance stability and to reduce potential toxicity. In addition, their activities need to be checked using clinically isolated bacteria. In the LKW peptide design, the tryptophan residue was specifically incorporated in expectation of str

4.1. hLF1-11 (Human Lactoferrin 1-11) Lactoferrin (LF) is an iron-binding glycoprotein which is responsible for part of the innate defense system. The LF can not only bind and trap Fe 3+ ion, but also interact with the bacterial membrane directly, which enables the LF to possess antibacterial activity [ 40 , 41 , 42 , 43 ]. The synthetic hLF1-11 peptide (GRRRRSVQWCA) is a lactoferrin derivative corresponding to the N-terminal eleven residues of human lactoferrin. The hLF1-11 peptide shows antimicrobial activity against both Gram-positive and -negative bacteria and various fungi. The synthetic peptide is also effective against methicillin-resistant Staphylococcus aureus (MRSA) and multidrug-resistant Acinetobacter baumannii strains [ 43 , 44 , 45 ]. Moreover, hLF1-11 is active against fluconazole-resistant Candida albicans , and can be used with classic antibiotics for a synergistic effect. Preincubation of fluconazole-resistant C. albicans with hLF1-11 significantly enhances the candidacidal effect of fluconazole [ 46 ]. The three-dimensional structures of several lactoferrin-derived peptides in membrane mimetic conditions have been solved by NMR spectroscopy [ 47 , 48 , 49 , 50 ]. The solution structure of LF11 (FQWQRNIRKVR), another N-terminal fragment based on human lactoferrin, showed conformational differences between in an anionic detergent (sodium dodecylsulfate; SDS) and in a zwitterionic detergent (dodecylphosphocholine; DPC) micelles. The structure of LF11 in SDS micelles was well defined and the tryptophan residue was significantly protected in the micelles. However, in DPC micelles, the structure was less defined and the tryptophan residue was less protected [ 48 ]. The conformation of hLF1-11 (GRRRRSVQWCA) also varied depending on the environments [ 43 ]. Molecular Dynamics (MD) simulation results of hLF1-11 in various solvents suggested that the peptide should be categorized as a loop peptide and adopts a more favorable conformation for the membrane interaction in membrane-mimicking environments. In particular, cationic residues (‑RRRR‑) of the hLF1-11 were rather flexible to be suitable for the interaction with the anionic bacterial membrane. Furthermore, the hydrophobic region, which was positioned approximately perpendicular to the cationic residues, enabled the peptide to bind to the membrane interior. Taken together, all these studies elucidated the structural selectivity of AMPs between bacterial and eukaryotic membranes. The safety and tolerability of hLF1-11 in healthy volunteers and haematopoietic stem cell transplantation (HSCT) recipients have been tested [ 43 , 51 ]. Intravenous administrations of hLF1-11 were safe and well tolerated in both healthy volunteers and the HSCT recipients. Although some adverse events (AE) were reported, all of them were mild in intensity and reversible. Pharmacodynamic evaluations including cytokine measurements also showed no significant changes in both populations.

Histatin 5, a parent molecule of P-113, is a small cationic peptide secreted into the saliva. Among the histatin family peptides, the histatin 5 exerts the most potent antimicrobial activity against bacteria and fungi [ 64 , 65 ]. The P-113 is an optimized fragment of the histatin 5 comprising twelve residues (residues 4-15 of histatin 5; AKRHHGYKRKFH), and retains antibacterial and anticandidal activity similar to that of the parent molecule, histatin 5 [ 66 ]. P-113 has potent candidacidal activity against various strains of C. albicans , C. glabrate , C. parapsilosis , and C. tropocalis that cause oral candidiasis. In addition to the broad spectrum of anticandidal activity, P-113 is also active against fluconazole resistant C. albicans and C. glabrata [ 66 ], which suggests that P-113 can be a promising antifungal agent for the treatment of oral candidiasis. Structural investigations of hsitatin 5 and P-113 in various solvents have been performed mainly by NMR spectroscopy [ 67 , 68 , 69 , 70 ]. The three-dimensional solution structures of histatin 5 revealed that the peptide prefers α-helical conformation in DMSO (dimethyl sulfoxide) or TFE (trifluoroethanol)/water, while it remains unstructured in water [ 68 , 69 ]. However, the helical conformation of histatin 5 in TFE/water system consists of two α-helices, one from Ser2 to Gly9 and the other from Lys11 to His19, whereas a single-alpha helix is formed in DMSO. MD simulation results of histatin 5 also supported that the α-helical regions of the peptide remain structured in the TFE system, but gradually unfold in water [ 70 ]. All the data suggest that the adaptation and maintenance of helical structure is essential for the antimicrobial activity of histatin 5. Structural properties of P-113, which are related to the antimicrobial activity, are also similar to those of histatin 5. Structural conversion of the P-113 from a disordered conformation in aqueous solution to an α-helical conformation in membrane mimetic environments was evidenced by a circular dichroism study [ 66 , 67 ]. The relationships between α-helical propensity and antimicrobial activity of P-113 can be demonstrated by comparing the structures and activities between the free- and metal bound-P-113 peptides. The solution structure of Zn(II)-bound P-113 (Zn(II)-P-113) contains a well defined N -terminal region (from Ala1 to His5) and a less defined C-terminal region (from Gly6 to His12). Imidazole rings of all three His residues (His4, His5, and His12) are coordinating one Zn(II) ion. The involvement of His12 in metal coordination makes the peptide to have lower propensity to adopt an alpha-helical conformation in C-terminal region, therefore resulting in the decrease of antimicrobial activity. The antimicrobial activity of Zn(II)-P-113 against a variety of microorganisms, including S. aureus , E. faecalis , and C. albicans , is much lower than that of metal free P-113 [ 71 ].