Improved Cell Selectivity of Pseudin-2 via Substitution in the Leucine-Zipper Motif: In Vitro and In Vivo Antifungal Activity

Several antimicrobial peptides (AMPs) have been discovered, developed, and purified from natural sources and peptide engineering; however, the clinical applications of these AMPs are limited owing to their lack of abundance and side effects related to cytotoxicity, immunogenicity, and hemolytic activity. Accordingly, to improve cell selectivity for pseudin-2, an AMP from Pseudis paradoxa skin, in mammalian cells and pathogenic fungi, the sequence of pseudin-2 was modified by alanine or lysine at each position of two amino acids within the leucine-zipper motif. Alanine-substituted variants were highly selective toward fungi over HaCaT and erythrocytes and maintained their antifungal activities and mode of action (membranolysis). However, the antifungal activities of lysine-substituted peptides were reduced, and the compound could penetrate into fungal cells, followed by induction of mitochondrial reactive oxygen species and cell death. In vivo antifungal assays of analogous peptide showed excellent antifungal efficiency in a Candida tropicalis skin infection mouse model. Our results demonstrated the usefulness of selective amino acid substitution in the repeated sequence of the leucine-zipper motif for the design of AMPs with potent antimicrobial activities and low toxicity.


Introduction
The increasing emergence of pathogens showing resistance to conventional drugs and the increasing frequency of microbial infections in immunosuppressed hosts have necessitated the identification of new types of antibiotics showing different mechanisms for the prevention of infection [1,2]. One promising therapeutic strategy for the treatment of multidrug-resistant strains is antimicrobial peptides (AMPs) [3][4][5][6][7]. AMPs are involved in the innate host defense system as a primary barrier against infection in most organisms. Most AMPs share several common characteristics, including membranolytic or intracellular-damaging action, amphipathicity, 10-50 amino acids, and net positive charge; however, some anionic AMPs have been reported [4,5]. Studies of AMPs have investigated their abundance in nature, their mechanisms, and their roles in innate or adaptive immune systems, as well as the effects of amino acid modification and their functions in drug delivery systems [8][9][10][11][12][13][14][15].  Antibiotics 2020, 9, 921 3 of 15 In this study, we aimed to develop AMP analogous to pseudin-2 to improve cell selectivity, increase antifungal activity, and reduce cytotoxicity. To this end, two residues containing leucine, isoleucine, and/or phenylalanine at the "a" and "d" positions in the LZ sequence of pseudin-2 were replaced with alanine or lysine (P2-LZ1-LZ5). The lytic activities of the peptides were investigated using pathogenic fungi, erythrocytes, and mammalian cells. Their secondary structures and membrane-permeable effects in fungal and mammalian membrane-mimicking liposomes were then examined to investigate the relationships between structure and lytic activity, and one of the variants was applied for the analysis of in vivo antifungal activity.

Designation of Pseudin-2 Derivatives Based on the LZ Motif and Structural Orientation
Pseudin-2 self-assembles in aqueous solution [25]. We hypothesized that the high cytotoxicity of this peptide may be related to this aggregate state. As shown in Figure 1A,B, pseudin-2 possessed a heptad repeat sequence, known as the LZ motif. The LZ is an important structural feature mediating the interactions between DNA and eukaryotic transcription regulatory proteins. Amino acids forming a scissors shape are connected at seven sites, and the inner parts are interlocked at sites of leucine residues. In order to improve the cell selectivity of pseudin-2 for fungal and mammalian cells, analogs were designed on a heptad repeat with leucine, isoleucine, and phenylalanine residues at the "a" and "d" positions in P1 and P2 lines ( Figure 1A,B). The three-dimensional (3D) model structure adopted a typical α-helical structure, and we predicted this sequence to be oriented such that the dimers were facing each other at the P1 and P2 positions ( Figure 2). Four double mutated variants, namely, P2-LZ1, P2-LZ2, P2-LZ3, and P2-LZ4 were designed by replacing leucine, isoleucine, and/or phenylalanine residues with an alanine residue at positions 5 and 12, 2 and 9, 12 and 19, and 9 and 16 of pseudin-2, respectively. Alanine was chosen owing to its weakly hydrophobic characteristics. Another analog, P2-LZ5, was designed by substituting phenylalanine and isoleucine residues with lysine residues at positions 9 and 12. Lysine was used here to prevent the formation of a self-assemble structure by allowing electrical repulsion between the opposing residues of the peptides and to alter the secondary structure of the peptide.
Antibiotics 2020, 9, x FOR PEER REVIEW 3 of 16 phenylalanine residues with an alanine residue at positions 5 and 12, 2 and 9, 12 and 19, and 9 and 16 of pseudin-2, respectively. Alanine was chosen owing to its weakly hydrophobic characteristics. Another analog, P2-LZ5, was designed by substituting phenylalanine and isoleucine residues with lysine residues at positions 9 and 12. Lysine was used here to prevent the formation of a self-assemble structure by allowing electrical repulsion between the opposing residues of the peptides and to alter the secondary structure of the peptide.  the leucine-zipper orientation determined via 3D structure modeling using the PEP-FOLD server (https://bioserv.rpbs.univ-paris-diderot.fr/services/PEP-FOLD3/) [26,27]. Blue and red colors indicate amino acids that could be substituted at P1 and P2 positions, respectively. in the leucine-zipper orientation determined via 3D structure modeling using the PEP-FOLD server (https://bioserv.rpbs.univ-paris-diderot.fr/services/PEP-FOLD3/) [26,27]. Blue and red colors indicate amino acids that could be substituted at P1 and P2 positions, respectively. Table 1 shows the physicochemical characteristics of pseudin-2 and its derivatives, including the measured molecular mass, hydrophobicity, hydrophobic moment, and net charge. The hydrophobicity values of the analogs were markedly lower than those of pseudin-2. Peptides with high hydrophobicity are likely to be located in fatty acids of lipids inserted in the cell membrane or to be self-assembled, such as adjacent chains of a globular protein. The hydrophobic moment, another important factor mediating the roles of AMPs, is related to the orientation at the surface between polar and nonpolar phases [28,29]. The high hydrophobic moment of pseudin-2 indicates that it is oriented in an amphipathic perpendicular position relative to its axis, e.g., the LZ orientation. We hypothesized that the reduced hydrophobicity and hydrophobic moment of the analogs could be related to the decreased cytotoxicity against mammalian cells.

In Vitro Antifungal Activity
To elucidate the relationships between the amino acid substitutions and antifungal activity of the derivatives compared with those of pseudin-2, we evaluated the minimum inhibitory concentrations (MICs) in nine fungal strains, including molds and yeasts. As shown in Table 2, pseudin-2 and its derivatives exhibited more potent inhibition of fungal growth in yeast than in mold fungi, except for that in Fusarium oxysporum, a phytopathogenic mold. Furthermore, the derivatives had lower MICs at an acidic pH than at a neutral pH. The MICs of P2-LZ1 and P2-LZ3 with substituted amino acids at the P1 position were slightly increased in most fungal cells, whereas those of P2-LZ2 and P2-LZ4, which had substitutions at the P2 position, showed similar or lower MICs compared with pseudin-2. Notably, P2-LZ4 showed potent antifungal activity against pathogenic yeast cells, including Candida albicans, C. krusei, C. tropicalis, and Trichosporon beigelii, with MICs ranging from 1.5 to 6 µM at neutral pH; this activity was not observed against C. parapsilosis, a yeast fungus that causes serious sepsis or wound infections in immunocompromised patients. We suggest that the difference of MIC values against each fungus is determined by the compositions of the cell wall of each fungus and P2-LZ4 has a high affinity with them. The MICs of P2-LZ5, for which the net charge was increased by lysine substitution, were very different depending on the fungal species.

In Vitro Cytotoxic Effects
2,3-Bis-(2-Methoxy-4-Nitro-5-Sulfophenyl)-2H-Tetrazolium-5-Carboxanilide (XTT) assays were performed to evaluate the viability of human HaCaT cells after treatment with the compounds. Although the cytotoxicity of P2-LZ1 was lower than that of pseudin-2, both compounds exhibited significant, concentration-dependent cytotoxic effects against HaCaT cells. Other peptides were not toxic at 256 µM ( Figure 3A). Figure 3B shows the hemolytic effects of peptides at a concentration of 64 µM in erythrocytes. Hemolysis is an important response to intravenous application of peptides. Triton X-100, a detergent, and melittin, a cytotoxic peptide, cause severe hemolysis by inducing hemoglobin release from Antibiotics 2020, 9, 921 5 of 15 erythrocytes. However, erythrocytes treated with pseudin-2 or its analogs did not release hemoglobin based on visualization ( Figure 3B). The different cytotoxic patterns observed in HaCaT cells and mouse red blood cells (mRBCs) were related to differences in the lipid composition and cell membrane content of the two cell types.
Antifungal assays were performed in 10 mM sodium phosphate buffer (pH 7.2) and 10 mM MES buffer (pH 5.5) supplemented with culture medium. The values in parentheses are MICs under acidic conditions.

In Vitro Cytotoxic Effects
2,3-Bis-(2-Methoxy-4-Nitro-5-Sulfophenyl)-2H-Tetrazolium-5-Carboxanilide (XTT) assays were performed to evaluate the viability of human HaCaT cells after treatment with the compounds. Although the cytotoxicity of P2-LZ1 was lower than that of pseudin-2, both compounds exhibited significant, concentration-dependent cytotoxic effects against HaCaT cells. Other peptides were not toxic at 256 µM ( Figure 3A). Figure 3B shows the hemolytic effects of peptides at a concentration of 64 µM in erythrocytes. Hemolysis is an important response to intravenous application of peptides. Triton X-100, a detergent, and melittin, a cytotoxic peptide, cause severe hemolysis by inducing hemoglobin release from erythrocytes. However, erythrocytes treated with pseudin-2 or its analogs did not release hemoglobin based on visualization ( Figure 3B). The different cytotoxic patterns observed in HaCaT cells and mouse red blood cells (mRBCs) were related to differences in the lipid composition and cell membrane content of the two cell types.

Mode of the Antifungal Actions of the Designed Peptides
Although C. albicans is the major fungus causing candidiasis, we used C. tropicalis as a model fungus in the mechanism study because pseudin-2 and analogs showed the best activity in C. tropicalis among the candida species (Table 2).
To ascertain the cellular distributions of the designed peptides, the localization of rhodaminelabeled peptides in C. tropicalis cells was observed by confocal laser-scanning microscopy (CLSM). Melittin, which is derived from honeybee venom and forms pores in the plasma membrane of eukaryotic cells [30], and pseudin-2 [17] are membrane-active peptides. Both rhodamine-labeled peptides were more accumulated on the surface than in the cytosol of fungal cells ( Figure 5). As shown in our report [17], pseudin-2 enters into the cytoplasm through a hole made by itself and is binds to nucleic acids. We suggested that P2-LZ1, -LZ2, -LZ3, and -LZ4 would have a similar mechanism of action to pseudin-2. On the other hand, P2-LZ5 was detected inside the cells, but not in the nucleus. This result suggests that the antifungal mechanisms of P2-LZ5 may be different from other peptides.

Mode of the Antifungal Actions of the Designed Peptides
Although C. albicans is the major fungus causing candidiasis, we used C. tropicalis as a model fungus in the mechanism study because pseudin-2 and analogs showed the best activity in C. tropicalis among the candida species (Table 2).
To ascertain the cellular distributions of the designed peptides, the localization of rhodamine-labeled peptides in C. tropicalis cells was observed by confocal laser-scanning microscopy (CLSM). Melittin, which is derived from honeybee venom and forms pores in the plasma membrane of eukaryotic cells [30], and pseudin-2 [17] are membrane-active peptides. Both rhodamine-labeled peptides were more accumulated on the surface than in the cytosol of fungal cells ( Figure 5). As shown in our report [17], pseudin-2 enters into the cytoplasm through a hole made by itself and is binds to nucleic acids. We suggested that P2-LZ1, -LZ2, -LZ3, and -LZ4 would have a similar mechanism of action to pseudin-2. On the other hand, P2-LZ5 was detected inside the cells, but not in the nucleus. This result suggests that the antifungal mechanisms of P2-LZ5 may be different from other peptides.
In general, studies of the mode of action of AMPs and peptides begin by measuring the membrane permeability of microbes in the presence of the peptides. Therefore, we assessed changes in the fluorescence of SYTOX Green, an impermeable nuclear and chromosome counterstain in live cells, over time ( Figure 6). Melittin and pseudin-2 induced gradual, concentration-dependent increases in fluorescence intensity, although the degree of fluorescence intensity differed. Changes in the fluorescence of LZ-P peptides were similar to those of pseudin-2, except for P2-LZ5. Furthermore, although the time of maximum fluorescence emission differed depending on the treatment concentration, the fluorescence intensity was highest at 9-12 min in the presence of 64 µM peptides, indicating that the antifungal mechanism involved membranolysis. The fluorescence intensity in the presence of P2-LZ5 was much lower than that in the presence of the other peptides. Typically, cell-permeable AMPs show no significant increases in fluorescence intensity when the peptide concentration is increased [31,32], and the membrane permeability of the dye is low when the peptide is applied at concentrations above the MIC value. Therefore, we believe that these peptides could penetrate into the cytosol of fungal cells. In general, studies of the mode of action of AMPs and peptides begin by measuring the membrane permeability of microbes in the presence of the peptides. Therefore, we assessed changes in the fluorescence of SYTOX Green, an impermeable nuclear and chromosome counterstain in live cells, over time ( Figure 6). Melittin and pseudin-2 induced gradual, concentration-dependent increases in fluorescence intensity, although the degree of fluorescence intensity differed. Changes in the fluorescence of LZ-P peptides were similar to those of pseudin-2, except for P2-LZ5. Furthermore, although the time of maximum fluorescence emission differed depending on the treatment concentration, the fluorescence intensity was highest at 9-12 min in the presence of 64 µM peptides, indicating that the antifungal mechanism involved membranolysis. The fluorescence intensity in the presence of P2-LZ5 was much lower than that in the presence of the other peptides. Typically, cellpermeable AMPs show no significant increases in fluorescence intensity when the peptide concentration is increased [31,32], and the membrane permeability of the dye is low when the peptide is applied at concentrations above the MIC value. Therefore, we believe that these peptides could penetrate into the cytosol of fungal cells. Analysis of the degree of dye leakage, indicative of the ability of a peptide to disrupt the membrane, showed that melittin and pseudin-2 caused a significant release of calcein from PC/PE/PI/ergosterol vesicles as a fungus-mimic membrane ( Figure 7A). Melittin, pseudin-2, P2-LZ1, P2-LZ2, and P2-LZ4 peptides induced 84.1%, 80.1%, 81.8%, 77.0%, and 80.5% leakage of calcein, respectively at a molar ratio of 0.05, whereas the leakage percentages of P2-LZ3 and P2-LZ5 were  Analysis of the degree of dye leakage, indicative of the ability of a peptide to disrupt the membrane, showed that melittin and pseudin-2 caused a significant release of calcein from PC/PE/PI/ergosterol vesicles as a fungus-mimic membrane ( Figure 7A). Melittin, pseudin-2, P2-LZ1, P2-LZ2, and P2-LZ4 peptides induced 84.1%, 80.1%, 81.8%, 77.0%, and 80.5% leakage of calcein, respectively at a molar ratio of 0.05, whereas the leakage percentages of P2-LZ3 and P2-LZ5 were 48.3% and 33.8%, respectively, at the same molar ratio ( Figure 7A). In addition, melittin and pseudin-2 induced significant dye leakage from a mammalian-mimic membrane composed of PC/CH/sphingomyelin (SM) vesicles, whereas P2-LZ peptides caused less than 20% calcein efflux at a peptide/lipid molar ratio of 0.1, except for P2-LZ1 ( Figure 7B). These results were consistent with the observed toxic effects of the peptides in HaCaT cells. The antimicrobial and cytotoxic effects of pseudin-2 and its analogs may be related to the cell membrane composition, and the cell selectivity of these peptides may be associated with the orientation of the peptides in aqueous solution owing to changes in hydrophobicity and the hydrophobic moment. Analysis of the degree of dye leakage, indicative of the ability of a peptide to disrupt the membrane, showed that melittin and pseudin-2 caused a significant release of calcein from PC/PE/PI/ergosterol vesicles as a fungus-mimic membrane ( Figure 7A). Melittin, pseudin-2, P2-LZ1, P2-LZ2, and P2-LZ4 peptides induced 84.1%, 80.1%, 81.8%, 77.0%, and 80.5% leakage of calcein, respectively at a molar ratio of 0.05, whereas the leakage percentages of P2-LZ3 and P2-LZ5 were 48.3% and 33.8%, respectively, at the same molar ratio ( Figure 7A). In addition, melittin and pseudin-2 induced significant dye leakage from a mammalian-mimic membrane composed of PC/CH/sphingomyelin (SM) vesicles, whereas P2-LZ peptides caused less than 20% calcein efflux at a peptide/lipid molar ratio of 0.1, except for P2-LZ1 ( Figure 7B). These results were consistent with the observed toxic effects of the peptides in HaCaT cells. The antimicrobial and cytotoxic effects of pseudin-2 and its analogs may be related to the cell membrane composition, and the cell selectivity of these peptides may be associated with the orientation of the peptides in aqueous solution owing to changes in hydrophobicity and the hydrophobic moment. To investigate the modes of action of P2-LZ peptides in fungal cells, we assessed the morphology of C. tropicalis cells in the presence of peptides by scanning electron microscopy (SEM). As shown in Figure 8, control cells exhibited smooth surfaces without damage, whereas melittin formed large To investigate the modes of action of P2-LZ peptides in fungal cells, we assessed the morphology of C. tropicalis cells in the presence of peptides by scanning electron microscopy (SEM). As shown in Figure 8, control cells exhibited smooth surfaces without damage, whereas melittin formed large holes in the surface of fungal cells. In the presence of pseudin-2, the fungal cell surface showed many holes and torn parts, and the cell morphology was wrinkled; similar results were observed for P2-LZ1 and P2-LZ3. In contrast, P2-LZ2 and P2-LZ4 induced cell swelling as well as holes in the cell surface. These results suggested that the antifungal mechanisms of P2-LZ1 and P2-LZ3 were different from those of P2-LZ2 and P2-LZ4. In contrast, C. tropicalis cells were relatively large, wrinkled, and swollen following the addition of P2-LZ5. In particular, the surface pores in these cells appeared to be created by pushing into the cell from the outside. This shape was not observed for the other peptide treatments.

Mitochondrial Reactive Oxygen Species (ROS) Generation in Response to P2-LZ5
To investigate the antifungal mechanism of P2-LZ5, the production of mitochondrial superoxide (SOX) in fungal cells was measured in the presence of peptide. The mitochondrial membrane of eukaryotic cells is anionic, similar to the bacterial cell membrane. When cationic peptides enter the cells, they can target mitochondria and cause disruption of membrane potential, release of cytochrome c, or generation of ROS, leading to cell apoptosis [33]. As shown in Figure 9, cytometry peaks indicated that MitoSOX Red fluorescence was significantly increased in the presence of P2-LZ5, as demonstrated by the right shift in the fluorescence emission. While pseudin-2 showed a little right shift and Antibiotics 2020, 9, 921 9 of 15 others did not. As shown in the results of other mechanisms, they did not induce intracellular ROS because they can act rapidly on fungal cell membranes and kill fungal cells. Figure 10 indicates that P2-LZ5 had different antifungal mechanisms compared with others. We propose that the P2-LZ5 peptides entered into the fungal cells can easily bind to the mitochondrial membrane with a negative charge via electrostatic interaction, resulting in the induction of destabilizing membrane potential and mitochondrial ROS.
Antibiotics 2020, 9, x FOR PEER REVIEW 9 of 16 holes in the surface of fungal cells. In the presence of pseudin-2, the fungal cell surface showed many holes and torn parts, and the cell morphology was wrinkled; similar results were observed for P2-LZ1 and P2-LZ3. In contrast, P2-LZ2 and P2-LZ4 induced cell swelling as well as holes in the cell surface. These results suggested that the antifungal mechanisms of P2-LZ1 and P2-LZ3 were different from those of P2-LZ2 and P2-LZ4. In contrast, C. tropicalis cells were relatively large, wrinkled, and swollen following the addition of P2-LZ5. In particular, the surface pores in these cells appeared to be created by pushing into the cell from the outside. This shape was not observed for the other peptide treatments.

Mitochondrial Reactive Oxygen Species (ROS) Generation in Response to P2-LZ5
To investigate the antifungal mechanism of P2-LZ5, the production of mitochondrial superoxide (SOX) in fungal cells was measured in the presence of peptide. The mitochondrial membrane of eukaryotic cells is anionic, similar to the bacterial cell membrane. When cationic peptides enter the cells, they can target mitochondria and cause disruption of membrane potential, release of cytochrome c, or generation of ROS, leading to cell apoptosis [33]. As shown in Figure 9, cytometry peaks indicated that MitoSOX Red fluorescence was significantly increased in the presence of P2-LZ5, as demonstrated by the right shift in the fluorescence emission. While pseudin-2 showed a little right shift and others did not. As shown in the results of other mechanisms, they did not induce intracellular ROS because they can act rapidly on fungal cell membranes and kill fungal cells. Figure  10 indicates that P2-LZ5 had different antifungal mechanisms compared with others. We propose that the P2-LZ5 peptides entered into the fungal cells can easily bind to the mitochondrial membrane with a negative charge via electrostatic interaction, resulting in the induction of destabilizing membrane potential and mitochondrial ROS.

In Vivo Antifungal Effects
C. tropicalis induces various infectious diseases, including oropharyngeal candidiasis, oral thrush, vulvovaginal candidiasis, angular cheilitis, pulmonary candidiasis, and gastrointestinal candidiasis, depending on the specific tissue or organ that it colonizes. Among the analog peptides, P2-LZ4 has the best antifungal activity and has no cytotoxicity, therefore it was selected in the animal experiment. P2-LZ5, showing a different mechanism, increased net charge due to the increase of lysine residues, resulted in a significant decrease of its antifungal activity under PBS condition (high salt). We evaluated the in vivo fungicidal activities of pseudin-2 and P2-LZ4 peptides in a mouse model of C. tropicalis skin infection. At 24 h after subcutaneous fungal infection, mice were treated with phosphate-buffered saline (PBS), pseudin-2, or P2-LZ4. As shown in Figure 10A, control mice treated with PBS exhibited obvious swelling and redness in the dorsal skin, whereas psuedin-2 treated mice showed reduced swelling and redness. Surprisingly, no obvious skin lesions appeared in the presence of P2-LZ4. Next, histological evaluation of skin tissues was performed using Hematoxylin and Eosin (H&E) staining ( Figure 10B). Notably, C. tropicalis-infected mice showed markedly increased inflammatory cell numbers, inflammatory cell infiltration, and tissue necrosis ( Figure 10B(a2,b2)) compared with PBS-treated mice ( Figure 10B(a1,b1)). Although pseudin-2 treated mice were able to recover from histological lesions ( Figure 10B(a3,b3)), skin tissues from mice treated with P2-LZ4 were nearly identical to those of the PBS-treated mice ( Figure 10B(a4,b4)).

Peptide Synthesis
Microwave-assisted peptide synthesis (Discover Bio, CEM Co., Matthews, NC, USA) was used to synthesize the peptides.
The amidated peptides were obtained by Rink amide 4-methylbenzhydrylamine resin. Fmoc amino acids were coupled using microwave heating in the presence of DIC and Oxyma in dimethylformamide (DMF) and Fmoc deprotection was processed using 20% piperidine in DMF. After the final coupling and deprotecting steps, the resin was washed with dichloromethane and air-dried. The peptides were cleaved from the resin, using trifluoroacetic acid (TFA)/triisopropylsilane/DiH 2 O (95:2.5:2.5, v/v/v) for 2 h at room temperature, followed by precipitation and washing with ice-cold diethyl, and dried under a vacuum. The crude peptides were purified using a ZORBAX PrepHT Eclipse C 18 preparative column (21.2 × 150 mm, 5-µm) on a Shimadzu high-performance liquid chromatography (HPLC) system. The purity of the purified peptides was more than 98% in analytic HPLC. Their molecular masses were measured using a matrix-assisted laser desorption ionization mass spectrometer (MALDI II; Kratos Analytical Ltd., Manchester, UK) [17].

Antifungal Assay
Antifungal activity of peptides was performed by microtiter plate assays. Fungal conidia from four-day-old molds were collected using 0.08% Triton X-100 (v/v), and pre-cultivated yeast cells were adjusted to 2 × 10 4 spores/mL in 10 mM sodium phosphate buffer (SP, pH 7.2) or 10 mM 2-(N-morpholino)ethanesulfonic acid (MES) buffer (pH 5.5) supplemented with 20% culture medium, followed by addition to peptide solutions (0.25, 0.5, 1, 1. 5,2,4,6,8,12,16,32,48, and 64 µM) in 96-well microtiter plates. After incubation for 36 h at 28 • C, the hyphal growth of conidia and the proliferation of yeast cells were observed under a light microscope. All assays were performed in triplicate. The MIC values were defined as the lowest concentration of samples that reduced fungal germination (for mold) or cell proliferation (for yeast) by more than 90% [34].

Cytotoxic and Hemolytic Assay
Cell viability was assessed using sodium XTT assays. HaCaT (human immortalized keratinocytes) cells were cultured in Dulbecco's modified Eagle medium (DMEM; ThermoFisher Scientific, Gibco, Waltman, MA, USA) supplemented with antibiotic-antimycotic (ThermoFisher Scientific, Gibco) and 10% fetal bovine serum (FBS; ThermoFisher Scientific, Gibco) at 37 • C in a humidified chamber containing 5% CO 2 . The cells were seeded at 5 × 10 4 cells/mL in flat-bottom 96-well microtiter plates in triplicate. Twenty-four hours later, cells were treated with two-fold serial dilutions of peptides (ranging from 256 to 4 µM) in medium, and the cells were further incubated for 24 h. Next, activated-XTT solution was added to each well and the plates were additionally incubated for 4 h. Absorbance in each well was measured at wavelengths of 480 and 650 nm using a microtiter SpectraMax M5 reader (Molecular Devices, Sunnyvale, CA, USA). Triton X-100 (0.1% v/v) and DMEM were used as controls for survival [31].
Mouse blood was collected into sodium heparin-coated tubes (BD Vacutainer; BD Diagnostics, Oxford, UK). The mRBCs were obtained by centrifugation (800× g, 10 min) and washing with PBS. The washed mRBCs were added to a final concentration of 8% (v/v) in 64 µM peptide solution, followed by incubation for 2 h at 37 • C. After centrifugation, each tube was digitally recorded.

CD Analysis
A Jasco J-810 spectropolarimeter (Jasco, MD, USA) was used to analyze the secondary structure of each peptide in artificial liposomes. A peptide solution mixed with PC/PE/PI/ergosterol (5:4:1:2, w/w/w/w) liposomes at a 1:10 molar ratio (30 µM peptide/300 µM lipid) was injected into a 0.1-cm path-length quartz cell. At least five scans were acquired and averaged from 190 to 250 nm. Samples were allowed to equilibrate to 25 • C prior to CD measurement. The experiments were run at 50 nm/min with 1 nm data intervals. The base line was adjusted with only liposome. Mean residue ellipticities ((θ), deg·cm 2 dmol −1 ) were calculated by following equation [31].
where θ obs is the measured signal (ellipticity) in millidegrees, l is the optical path-length of the cell in cm, and c is the concentration of peptide in mol/L [mean residue molar concentration: c = number of residues in the constructed of peptide × the molar concentration of the peptide].

CLSM
To conjugate the fluorescent probe at the N-terminus of the peptide, NHS-rhodamine was incubated with peptide solutions at a 1:1 molar ratio in PBS (pH 7.2) at room temperature for 1 h. Rhodamine-labeled peptides were purified using a C 18 column on an HPLC system. Peptide solutions were used with the mixtures of rhodamine-labeled peptides and rhodamine-free peptides (1:9, w/w) to minimize rhodamine effects. Peptides were then added to 200 µL of C. tropicalis cell suspensions at the appropriate MIC. After incubation for 1 h, the cells were pelleted by centrifugation at 3000× g for 5 min, washed with ice-cold PBS, and fixed with 2% glutaraldehyde (v/v) in PBS. The cellular distribution of rhodamine-labeled peptides was then examined using an inverted LSM510 laser scanning microscope (Carl Zeiss, Gőttingen, Germany). The 543 nm light from a helium neon laser was directed at a UV/488/543/633 beam splitter. Images were recorded digitally in a 512 × 512 pixel format [17].

SYTOX Green Uptake
C. tropicalis cells pre-grown at 28 • C were washed and suspended (2 × 10 6 cells/mL) in SP buffer (pH 7.2). The cells were pre-incubated with SYTOX Green (final concentration of 0.5 µM) for 15 min in the dark, followed by adding two-fold serial dilutions of peptides at concentrations of 1 to 64 µM. The fluorescence intensity was monitored for 30 min at 485 (Ex) and 520 (Em) nm using a microtiter SpectraMax M5 reader (Molecular Devices, San Jose, CA, USA) [35].

Conflicts of Interest:
The authors declare no conflict of interest.