**4. Mode of action of cationic antimicrobial peptides**

#### **4.1. Membrane effects**

CAPs can cause lysis of biological membranes or cross membranes by spontaneous lipid-as‐ sisted translocation. The initial interaction is electrostatic between the cationic residues of the peptide and the negatively charged constituents of the target cell, whether they are in an outer bacterial envelope, a viral envelope or a eukaryotic cell membrane. Because of the high proportion of uncharged zwitterionic lipids and sterols in normal eukaryotic membranes, however, these membranes are not as susceptible as negatively charged bacterial mem‐ branes or cancer cell membranes, which contain elevated levels of negatively charged sialy‐ lated glycoproteins (van Beek et al., 1973), *O*-glycosylated mucins (Yoon et al., 1996) and phosphatidylserine (Riedl et al., 2011).

Upon binding, the peptide assumes an amphipathic secondary structure that facilitates membrane disruption. A number of mechanisms have been proposed including micelle for‐ mation by the carpet model (Shai, 2002) and pore formation by the barrel-stave and toroidal pore models (Brogden, 2005; Hadley and Hancock, 2010). Additional variations on these ini‐ tial models for bacterial membrane interactions have been proposed including localized membrane disorganization by the aggregate model (Hancock and Rozek, 2002), the forma‐ tion of disordered toroidal pores, charged lipid clustering, membrane thinning, electropora‐ tion, non-lytic membrane depolarization, anion carriers and oxidized lipid targeting (Nguyen et al., 2011).

Pleurocidin has been proposed to act via the toroidal pore model, based on its interactions with model membranes (Saint et al., 2002). This finding is supported by NMR studies that show that in DOPC pleurocidin exists as a very large species, probably composed of 20-25 aggregating molecules (Syvitski et al., 2005), indicating that at high concentrations pleuroci‐ din forms pores within the membrane environment or disrupts the membrane by aggrega‐ tion. A more recent study proposed that CAPS similar to pleurocidin such as magainin 2 adopt a surface alignment in mixed zwitterionic/anionic membranes and forms disordered toroidal pores (Leontiadou et al., 2006). The interaction of pleurocidin with the anionic lipid phosphatidylglycerol (PG) rather than the zwitterionic lipid phosphatidylethanolamine (PE) in mixed membranes, resulting in disruption of membrane integrity, was confirmed using NMR (Mason et al., 2006). Interestingly, the three His residues, which become protonated at acidic pH, play no role in membrane disruption, most likely due to their position along the spine of the helix where they would be unlikely to interact with lipid head groups.

The Gly13 residue in the middle of pleurocidin confers flexibility between a longer α-helix at the amino terminus and a shorter α-helix at the carboxy terminus allowing it to interact with the negatively charged phospholipid membranes (Yang et al., 2006). Replacing Gly13 with Ala removes this hinge region, resulting in a much higher α-helical content and loss of cell selectivity. This analog is hemolytic and interacts with both negatively charged (mimick‐ ing bacterial and cancer cell membranes) and zwitterionic (mimicking erythrocyte and nor‐ mal mammalian cell membranes) phospholipids. Similarly, substitution of both Gly13 and Gly17 by Ala results in a large increase in α-helicity and a correspondingly dramatic in‐ crease in hemolytic activity, indicating that the hinge region facilitates flexibility and confers bacterial cell selectivity (Lim et al., 2004a). It is interesting that Gly13 is conserved in 15 of the 26 identified pleurocidin variants whereas Gly17 is conserved in only 6, indicating that Gly13 plays the more important role in peptide function.

Sterols are commonly found in eukaryotic but not bacterial cell membranes, and their pres‐ ence in anionic mixed membranes reduces the ability of pleurocidin to insert even though the peptide maintains its α-helical configuration (Mason et al., 2007). Cholesterol, found in membranes of higher eukaryotes, was shown to be more effective than ergosterol, common‐ ly found in those of fungi and certain protozoa, in reducing insertion of CAPs. These differ‐ ences in membrane sterol content explain the differential sensitivity of bacteria, lower eukaryotes and higher eukaryotes to CAPs such as pleurocidin. Using a fluorescent mem‐ brane probe, pleurocidin has been shown to structurally perturb the plasma membrane of the fungal pathogen *Candida albicans* (Jung et al., 2007), confirming earlier killing assay re‐ sults (Patrzykat et al., 2003).

Structural modelling predicted that only 12 of the 26 described pleurocidin variants formed amphipathic α-helices and killing assays showed that these were cytotoxic to HL-60 human leukemia cells, resulting in rapid and complete killing (Morash et al., 2011). Cell death was non-apoptotic and probably occurred by the formation of ion channels that dissipated the membrane potential and led to cell lysis. The active pleurocidin variants were the more highly charged (>+6.5) members of the family; interestingly only one of them was also hemolytic. The anti-cancer activities of two pleurocidin variants (NRC-03 and NRC-07) against breast can‐ cer cells involve binding to negatively charged cell-surface molecules (Hilchie et al., 2011).

#### **4.2. Intracellular effects**

with chronic periodontitis (Turkoglu et al., 2011) and mycoplasma infection induces catheli‐ cidin expression in neutrophils of infected mice (Tani et al., 2011). Interestingly, increased endogenous glucocorticoid levels induced by psychological stress reduce CAP expression in

Expression of fish CAPs is also regulated in response to stress and disease (Douglas et al., 2003a; Sun et al., 2007), and novel variants are often induced by such stressors. Monitoring levels of CAPs in aquaculture settings provides early warning of immunosuppression due to chronic stress, and conversely induction of CAPs provides protection against anticipated

CAPs can cause lysis of biological membranes or cross membranes by spontaneous lipid-as‐ sisted translocation. The initial interaction is electrostatic between the cationic residues of the peptide and the negatively charged constituents of the target cell, whether they are in an outer bacterial envelope, a viral envelope or a eukaryotic cell membrane. Because of the high proportion of uncharged zwitterionic lipids and sterols in normal eukaryotic membranes, however, these membranes are not as susceptible as negatively charged bacterial mem‐ branes or cancer cell membranes, which contain elevated levels of negatively charged sialy‐ lated glycoproteins (van Beek et al., 1973), *O*-glycosylated mucins (Yoon et al., 1996) and

Upon binding, the peptide assumes an amphipathic secondary structure that facilitates membrane disruption. A number of mechanisms have been proposed including micelle for‐ mation by the carpet model (Shai, 2002) and pore formation by the barrel-stave and toroidal pore models (Brogden, 2005; Hadley and Hancock, 2010). Additional variations on these ini‐ tial models for bacterial membrane interactions have been proposed including localized membrane disorganization by the aggregate model (Hancock and Rozek, 2002), the forma‐ tion of disordered toroidal pores, charged lipid clustering, membrane thinning, electropora‐ tion, non-lytic membrane depolarization, anion carriers and oxidized lipid targeting

Pleurocidin has been proposed to act via the toroidal pore model, based on its interactions with model membranes (Saint et al., 2002). This finding is supported by NMR studies that show that in DOPC pleurocidin exists as a very large species, probably composed of 20-25 aggregating molecules (Syvitski et al., 2005), indicating that at high concentrations pleuroci‐ din forms pores within the membrane environment or disrupts the membrane by aggrega‐ tion. A more recent study proposed that CAPS similar to pleurocidin such as magainin 2 adopt a surface alignment in mixed zwitterionic/anionic membranes and forms disordered toroidal pores (Leontiadou et al., 2006). The interaction of pleurocidin with the anionic lipid phosphatidylglycerol (PG) rather than the zwitterionic lipid phosphatidylethanolamine (PE)

mice, leading to increased susceptibility to infection (Aberg et al., 2007).

stressful situations such as handling (Noga et al., 2011).

124 Using Old Solutions to New Problems - Natural Drug Discovery in the 21st Century

**4.1. Membrane effects**

(Nguyen et al., 2011).

phosphatidylserine (Riedl et al., 2011).

**4. Mode of action of cationic antimicrobial peptides**

A number of CAPs, when administered at low doses, are able to penetrate the bacterial membrane and disrupt metabolic processes (see Marcos and Gandia, 2009; Nicolas, 2009). CAPs can assume different structures whilst exerting their inhibitory effects. For example, when bound to DNA, buforin II assumes an extended form whereas magainin and pleuroci‐ din assume an α-helical form and bind DNA less effectively (Lan et al., 2010). Combinations

of CAPs that exert their effects on different targets may exhibit synergy, and provide more potent killing activity.

Pleurocidin usually causes membrane disruption at high concentrations whereas at low con‐ centrations, it can translocate into the cytoplasm without causing cell lysis and exert its ef‐ fects intracellularly, inhibiting DNA, RNA and protein synthesis (Patrzykat et al., 2002). The intracellular effect of pleurocidin NRC-03 has since been probed using zebrafish embryos, and shown to target the mitochondria and generate superoxide (Morash et al., 2011). TU‐ NEL staining indicates that the DNA of some cells becomes degraded, whereas other cells undergo rapid lysis and cell death without DNA fragmentation. Pleurocidin variants NRC-03 and NRC-07 cause mitochondrial membrane damage and production of reactive oxygen species (ROS) in MDA-MB-231 breast cancer cells (Hilchie et al., 2011).

#### **4.3. Receptors and binding proteins**

In most cases, killing is by nonreceptor-mediated mechanisms since all D-amino acid enan‐ tiomers are generally as active as the natural L-amino acid peptides. However, stereospecific receptor-mediated translocation has been described (Nicolas, 2009) and some CAPs trans‐ duce their effects via signaling networks upon interaction with receptors. Quite a diversity of receptors have been described, possibly reflecting the diversity in peptide structures. Up‐ take of apidaecin into Gram negative bacteria is proposed to by an energy-dependent mech‐ anism involving a permease or transporter (Castle et al., 1999). The detection of peptide:receptor complexes is technically very difficult but some receptors have been identi‐ fied. For example, the outer membrane protein OprI from *P. aeruginosa* (Lin et al., 2010) and the outer membrane lipoprotein Lpp in Enterobacteriaceae (Chang et al., 2012) have been shown to serve as receptors for α-helical CAPs. Histatin 5 and some defensins bind Ssa1/2 proteins on the cell surface of *C. albicans* in order to exert their activity, and the potassium transporter TRK1 is also required for histatin5 fungicidal activity (Vylkova et al., 2007). In eukaryotic cells, formyl peptide receptor-like 1 (FPRL1) is used as a receptor for LL-37 to act as a chemoattractant (Yang et al., 2000) and induce angiogenesis (Koczulla et al., 2003). LL-37 has also been shown to mediate keratinocyte migration and cytokine release by trans‐ activation of the epidermal growth factor receptor (Tokumaru et al., 2005) and P2X7 recep‐ tor (Elssner et al., 2004), respectively. However, activation of receptors could be via CAPinduced changes in membrane fluidity, ion transport and/or receptor aggregation (Braff et al., 2005) rather than direct binding.

Activation of mast cells by pleurocidin is G protein-dependent and proposed to involve FPRL1 and G protein coupled receptor signaling pathway (Kulka, unpub). The observation that pleurocidin can bind and activate FPRL1 has some important implications for human dis‐ ease. The FPRL1 receptor subtype is also a receptor for the bacterially-derived peptide fMLP (N-formyl-L-methionyl-L-leucyl-L-phenylalanine) making it an important innate immune receptor (Selvatici et al., 2006). FPRL1 activates key components of the innate immune sys‐ tem and is responsible for chemotactic responses, superoxide anion production and degranu‐ lation by neutrophils, macrophages and mast cells. FPRL function has been shown to be important in chronic obstructive pulmonary disease (COPD) due to cigarette smoking (Car‐ dini et al., 2012). FPRL1 can also transactivate epidermal growth factor receptors (EGFR) making them a potentially important target in lung cancer (Cattaneo et al., 2011).
