Rationalisation of Antifungal Properties of α-Helical Pore-Forming Peptide, Mastoparan B

The high mortality associated with invasive fungal infections, narrow spectrum of available antifungals, and increasing evolution of antifungal resistance necessitate the development of alternative therapies. Host defense peptides are regarded as the first line of defense against microbial invasion in both vertebrates and invertebrates. In this work, we investigated the effectiveness of four naturally occurring pore-forming antimicrobial peptides (melittin, magainin 2, cecropin A, and mastoparan B) against a panel of clinically relevant pathogens, including Candida albicans, Candida parapsilosis, Candida tropicalis, and Candida glabrata. We present data on the antifungal activities of the four pore-forming peptides, assessed with descriptive statistics, and their cytocompatibility with cultured human cells. Among the four peptides, mastoparan B (MB) displayed potent antifungal activity, whereas cecropin A was the least potent. We show that MB susceptibility of phylogenetically distant non-candida albicans can vary and be described by different intrinsic physicochemical parameters of pore-forming α-helical peptides. These findings have potential therapeutic implications for the design and development of safe antifungal peptide-based drugs.


Introduction
Invasive candidiasis (IC) is an infection caused by various Candida species, which results from candidemia and deep-seated tissue candidiasis. IC is observed among healthcareassociated infections in developed countries and is associated with high mortality rates (~40%) despite the institution of antifungal therapy [1]. Annually, the infection affects >250,000 people globally and is responsible for >50,000 deaths [2]. Candidemia, defined by the presence of Candida in the bloodstream, may cause infection of distant organs such as the liver or spleen, which may lead to inflammation and subsequent organ failure [2]. The incident rates of candidemia are higher among ageing populations and often comorbid with malignancy.
Candida albicans is the major etiological agent in candidemia in the past. The global shift in favor of non-albicans such as C. parapsilosis and C. tropicalis in Asia, Southern

Circular Dichroism (CD) Spectropolarimetry
Far UV-CD spectra of the peptides in phosphate-buffered saline (PBS) or trifluoroethanol (TFE) were recorded on a Chirascan CD Spectrometer (Applied Photophysics Ltd, Leatherhead, UK) using a 0.1 cm path length quartz cuvette at 20 • C. Spectra were recorded from 190 nm to 260 nm at a 1.0 nm interval with a scan rate of 50 nm/min, and the final spectrum was taken to be the average of three scans. The helical content was estimated, following the protocol reported before [20].

Zeta-Potential (ζ) Measurements
Measurements of the ζ-potential of the peptides were performed using Litesizer TM 500 (Anton Paar, Graz, Austria). Peptide samples were freshly prepared at 100 µg/mL in PBS, which was pre-filtered using a 0.22 µm membrane. Samples were transferred to an omega cuvette for measurements at 25 • . The ζ-potential in mV was automatically calculated by the Litesizer TM 500 (Anton Paar GmbH, Graz, Austria) using the Smoluchowski equation: where µ is the electrophoretic mobility measured by the Litesizer TM 500, η is the viscosity of PBS, and ε is the dielectric constant of PBS.
The experiments were repeated three times, and the results were expressed as mean ± standard deviation.

Determination of the Minimum Inhibitory Concentrations (MICs)
MICs of the test peptides were determined using the broth microdilution method in 96-well microtitre plates. Briefly, 100 µL of two-fold serial dilutions of each peptide (adjusted to a final concentration of 1-620 µM) were transferred to the microtitre plates in duplicate. The test strains were cultured in Sabouraud Dextrose Agar (SDA) plates incubated for 24 h at 37 • C. Inoculums of the fungal strains were prepared by selecting isolated fungal colonies from subcultures and suspending the colonies in Sabouraud Dextrose Broth (SDB). The suspension concentration was adjusted to the desired final concentration of 10 4 -10 5 CFU/mL. Final suspension (100 µL) was added to each well of the 96-well plates and incubated at 35 • C for 24 h. The optical density (OD) at 600 nm (OD 600 ) was determined using a TECAN Infinite 200 microplate reader (Mannedorf, Switzerland). The MIC was taken to be the lowest peptide concentration which showed no visible microbial growth as well as by measuring the absorbance OD 600 .

Mammalian Cell Viability Assays
Cell viability was determined using CellTiter 96 ® Aqueous One solution cell proliferation assay kit (Promega, Madison, USA), according to the supplier's instructions. HaCaT and HDF cells were cultured in Dulbecco's Modified Eagle Medium (DMEM) supplemented with fetal bovine serum 10% (v/v) and antibiotics (50 U/mL penicillin and 50 µg/mL streptomycin). Cells were seeded in 96-well tissue culture plates at a cell density of 4.5 × 10 3 cells/well and incubated at 37 • C for 24 h. The cells were then treated with a range of peptide concentrations (2-600 µM) obtained from serial dilution in the cell culture medium and the plates containing treated cells were incubated again at 37 • C for another 24 h. MTS solution (20 µL, provided in the assay kit) was next added to each well followed by incubation at 37 • C for 2 h. Absorbance was measured at 490 nm using a TECAN microplate reader and normalized with positive and negative control values to obtain % cell viability. In another set of experiment, the peptide treated cells were fixed with formaldehyde (3.7%, v/v) at room temperature for 15 min. After fixation, the cells were first washed with PBS, then Triton-X-100 (0.3%, v/v) was added into each well to permeabilize the cells. The cells were washed again with PBS and the adherent cells were stained with fluorescein isothiocyanate labelled anti-α-tubulin antibody (green), rhodamine phalloidin (red) and Hoechst 33342 (blue) to visualize cytoskeleton and nuclei morphologies. Cell images were acquired by an automated microscope IN Cell Analyzer 2200 (Model 2200, GE Healthcare UK Limited, Buckinghamshire, UK) equipped with the IN Cell Investigator software.

Cell Migration Assay
NTLG1 cells were grown to confluency in keratinocyte serum-free medium in a 96-well plate. Identical scratches were created using AccuWound 96 Scratch Tool (Agilent Technologies Inc., Santa Clara, CA, USA) and then washed in PBS before being replenished with fresh medium containing peptides. Wound closure was monitored by EVOS M5000 (Thermofisher Scientific, Singapore) and subsequently quantified by ImageJ to determine the effect of peptide relative to untreated control.

Sytox™ Green Uptake Assay
Overnight cultured fungal colonies were suspended in 16% SDB in PBS and adjusted to obtain an OD of 0.4 at 600 nm. A 5 mM Sytox™ green stock solution in dimethylformamide was added to the suspension to a final concentration of 1 µM and incubated for 15 min at room temperature in dark. The suspension (600 µL) was then added to a quartz cuvette and placed inside the Quanta Master spectrofluorometer. Once a constant plateau was reached, peptide solution (0.5×-4× MIC) was added to the cuvette. The increase in emission intensity at 515 nm was monitored after the addition of various concentrations of the peptides (λ ex = 485 nm).

Time-Kill Kinetics Studies (TKS)
TKS was performed on C. parapsilosis 1002 (1.2 × 10 4 CFU/mL) and C. glabrata 1003 (3.4 × 10 4 CFU/mL) inoculum and exposed to MB peptide at 0.5×-4× MIC. At various time intervals, 100 µL of aliquots serial dilutions were plated on SDA plates and incubated at 35 • C for 48 h, and then the CFUs were counted and expressed as log 10 CFU/mL. PBSexposed strains were included as an untreated control. To determine the fungicidal activity of MB, C. parapsilosis 1002 strains (~1.5-2.0 × 10 5 CFU/mL) in SDB were exposed to the peptide (9.92 µM) for 6 h. Viable cells were enumerated as above after 48 h. To determine the effect of monovalent or divalent ions, the broth was adjusted to appropriate concentrations of the ions.

Physicochemical Properties of Pore-Forming Peptides
We compared the hydrophobicity, overall net charge, and isoelectric points of the four pore-forming peptides-MB, MG, MC, and CA (Table 1). These parameters are the key determinants of solubility and aggregation status (pI), interaction with acidic lipids (net charge), and membrane discrimination. The results suggest that MB is the most hydrophobic and MG is the least cationic among the four peptides (Table 1). Zeta-potential measurements further confirmed that MG had the lowest ζ-potential value, whereas ML had the highest value among these peptides. To determine the secondary structures of the peptides, we recorded their CD spectra in PBS and in a membrane-mimetic solvent (30% TFE). All the peptides displayed an unordered secondary structure in PBS, whereas a clear transition to α-helical structure was observed in TFE ( Figure S5). Quantitative estimation of the α-helical content suggested that CA had the lowest % α-helicity whereas ML displayed the highest helical content in TFE. The helical content was comparable for MB and MG peptides (Table 1).

Antifungal Activity of Pore-Forming Peptides
We determined the MIC values of the peptides against 35 strains of "Candida spp.", covering both albicans and NACs strains ( Table 2). The different groups of Candida strains displayed varying degrees of susceptibility to the antimicrobial peptides. All the peptides were potent against the tested C. albicans strains. MB, ML, and CA were effective with a median MIC of 1.26, 1.41, and 2.00 µM, respectively, whereas MG required higher concentration (median MIC 13.24 µM) for complete inhibition. Among the non-albicans strains, C. tropicalis strains were susceptible to all the four pore-forming peptides with median MIC ranging from 0.62 µM (MB) to 4.22 µM (ML). C. parapsilosis strains displayed resistance to CA as no inhibitory effects were observed even at the highest concentration tested but were susceptible to other pore-forming peptides. The median MICs were 4.96 µM, 5.62 µM, and 13.28 µM for MB, ML, and MG, respectively. Among the fungal species, C. glabrata displayed heightened resistance to ML as MIC values could be determined against three clinical isolates only. At least 7 out of 10 C. glabrata strains were susceptible to MB and MG with a median MIC of 4.96 µM and 26.48 µM, respectively. However, none of the C. glabrata strains were susceptible to CA.

Host Cell Cytocompatibility of Pore-Forming Peptides
Next, we determined the cytotoxicity of peptides for HaCaT and HDF cells. The dosedependent changes in the metabolic activity and cell morphology were investigated by MTS-based assays and automated imaging, respectively. Among the four pore-forming peptides, ML displayed heightened toxicity for both the mammalian cell-types as a complete loss of viability was observed at one to two times the median MIC value of C. albicans (Figure 1c). Compared to ML, cytotoxic effects of MB were instead observed at elevated concentrations (60× median MIC value against C. albicans) (Figure 1a). MG and CA did not show any major cytotoxic effect for both cell-types at the range of concentrations tested (Figure 1b,d).
To obtain better insight into the status of the cell health, we performed the highcontent analysis of the peptide treated cells (Figure 2). No significant change in the cellular morphology was observed for HaCaT cells exposed to 38.8 µM MB when compared to untreated cells. However, substantial disruption of the microcolonies and reduction in actin staining was observed at a MB peptide concentration of 77.5 µM. HaCaT cells exposed to 5.6 µM of ML appeared fragmented with a lower number of microcolonies compared to untreated cells, thus confirming the cytotoxic effects of the peptide. Cells treated with MG or CA retained the microcolonies with well-developed adhesion systems that appeared to be similar to the untreated control, indicating a lack of significant cytotoxic effects of the two pore-forming peptides. Similar concentration-dependent changes in the morphology were observed for HDFs treated with MB or ML, whereas MG and CA did not show any major signs of cytotoxicity for HDFs, thus corroborating the MTS results ( Figure 2). indicated that MB was less cytotoxic than ML, while the other two pore-forming (MG and CA) peptides remained the least toxic for the skin cells, corroborating the MTS and high content analysis.  One of the hallmarks of wound repair is the migration of cells to the injured sites. To infer whether the peptides alter the cell migration, we determined the effect of peptides on the migration of immortalized human keratinocytes by scratch wound assay (Figure 3a). The results suggested a significant decrease in the wound closure after treatment with MB (4.8 µM) or ML (2.8 µM) when compared to untreated control (Figure 3b). However, no difference was observed when treated with MG or CA peptides. Overall, these results indicated that MB was less cytotoxic than ML, while the other two pore-forming (MG and CA) peptides remained the least toxic for the skin cells, corroborating the MTS and high content analysis.

Antifungal Activity of MB in the Presence of Mono and Divalent Cations
Time-kill kinetics studies suggested that MB elicited rapid candidacidal activity against C. parapsilosis. There was a >2 log 10 reduction in fungal viability within 30 min at 2× and 4× MIC (Figure 4a). A similar effect was observed against C. glabrata at 4× MIC (Figure 4b). To infer the pore-forming properties of MB, we determined ATP release of fungal strains after exposure to 4× MIC of MB. ATP release studies (data not shown) indicated that there was a complete loss of intracellular ATP within 6 h of exposure of C. parpsilosis to MB. To further discern the membrane perturbation by MB, we recorded the fluorescence influx of the membrane impermeable Sytox green dye upon addition of the peptide to fungal inoculum. MB elicited rapid and concentration-dependent increase in the uptake of Sytox green, indicating substantial membrane perturbation at 2× and 4× MIC values against both C. parapsilosis (Figure 4c) and C. glabrata (Figure 4d) strains. Thus, these results demonstrate that MB elicits rapid candidacidal properties by perturbing the cytoplasmic membrane and removing vital intracellular components. The fungicidal property of the peptide was dependent on the concentrations of monovalent (NaCl) or divalent (CaCl 2 ) ions present in the broth ( Table 3). The presence of a high concentration of Na + or Ca 2+ attenuated the fungicidal properties of MB.

Antifungal Activity of MB in the Presence of Mono and Divalent Cations
Time-kill kinetics studies suggested that MB elicited rapid candidacidal activity against C. parapsilosis. There was a >2 log10 reduction in fungal viability within 30 min at 2× and 4× MIC (Figure 4a). A similar effect was observed against C. glabrata at 4× MIC (Figure 4b). To infer the pore-forming properties of MB, we determined ATP release of fungal strains after exposure to 4× MIC of MB. ATP release studies (data not shown) indicated that there was a complete loss of intracellular ATP within 6 h of exposure of C. parpsilosis to MB. To further discern the membrane perturbation by MB, we recorded the fluorescence influx of the membrane impermeable Sytox green dye upon addition of the peptide to fungal inoculum. MB elicited rapid and concentration-dependent increase in the uptake of Sytox green, indicating substantial membrane perturbation at 2× and 4× MIC values against both C. parapsilosis (Figure 4c) and C. glabrata (Figure 4d) strains. Thus, these results demonstrate that MB elicits rapid candidacidal properties by perturbing the cytoplasmic membrane and removing vital intracellular components. The fungicidal property of the peptide was dependent on the concentrations of monovalent (NaCl) or divalent (CaCl2) ions present in the broth ( Table 3). The presence of a high concentration of Na + or Ca 2+ attenuated the fungicidal properties of MB.  3.5. Correlation of Physicochemical Properties of MB with Anti-C. glabrata Activity α-helical HDPs are well-characterized peptides for drug optimization by quantitative structure activity correlation [22][23][24][25][26]. Yount et al. reported that α-helical HDPs or their signature in mammalian proteome could be distinguished by hydrophobic moment, overall net charge, hydrophobicity, and lysine-to-arginine ratio [27]. These authors further demonstrated that the membrane permeation of α-helical HDPs could be predicated by integration of hydrophobic moment and charge. Among the four pore-forming peptides, MB displayed broad antifungal activities against both C. albicans as well as NACs. However, C. parapsilosis and C. glabrata strains displayed lower susceptibility to ML, MG, and CA when compared to C. albicans or C. tropicalis strains. Therefore, we hypothesized that the amino acid sequence of MB is optimized to confer potent antifungal activities against NACs. To understand the distinct properties of MB, we searched the PUBMED and Google Scholar using the key words such as "Antifungal peptides" and "Peptides and C. glabrata/tropicalis" and identified research articles that reported the MIC values of natural or synthetic peptides against the two pathogens. To determine the physicochemical parameter that better describes the antifungal properties of MB, we searched the literature that reported the MIC values of short α-helical peptides (≤18 residues). A total of 41 and 42 peptides that showed antifungal activity against C. glabrata and C. tropicalis, respectively, were identified from the search (Tables S1 and S2). We then computed their physicochemical parameters such as hydrophobicity (H), hydrophobic moment (µH), and overall net charge (Q) for these peptides.
The correlation between log 10 MIC values and the various physicochemical parameters are shown in Figures 4 and 5. The scattered data points around the lines of best fit are attributed to inter-laboratory variations and experimental conditions employed for the determination of MIC values. There was a statistically significant association between log 10 MIC of the peptides that displayed potent activity against C. glabrata strains and Q (r = 0.53 and p = 3 × 10 −4 ) (Figure 5c). There were no statistically significant associations (p > 0.05) observed between log 10 MIC of C. glabrata peptides and H (Figure 5a) or µH (Figure 5b). However, the integration of the latter two parameters with Q resulted in statistically significant associations with the log 10 MIC values of C. glabrata (Figure 5d,e). There was a statistically significant association between log 10 MIC of the peptides that displayed potent activity against C. tropicalis strains and H (r = 0.38 and p = 1.3 × 10 −2 ) (Figure 6a), as well as µH (r = 0.41 and p = 7.0 × 10 −3 ) (Figure 6b). However, there was no statistically significant association between the log 10 MIC values of C. tropicalis and Q or integration of the parameters (i.e., H*Q or µH*Q, Figure 6c-e).

Discussion and Conclusions
Invasive fungal infections caused by candida species represent the fourth leading cause of hospital acquired infections with high mortality [28]. Although C. albicans is the major etiological agent in candidiasis, the increased prevalence of NACs has been reported worldwide. Among the NAC species, C. glabrata remains the most common while C. parapsilosis, C. tropicalis, and C. krusei are also frequently isolated [29]. The rise of antifungal resistance among clinically relevant Candida species and the differences in the pattern of antifungal resistance limit the choice of antifungal therapy. In this study, we investigated the antifungal properties of four prolific pore-forming peptides against C. albicans and NAC strains. Investigation of the physicochemical parameters and secondary structural analysis suggest that the peptides are amphipathic and cationic, as indicated by high pI and positive zeta-potential values. Among the peptides, ML had the maximum zeta-potential value when compared to other peptides, though CA had the highest overall net charge. Similarly, the helical content of the peptides in membrane-mimetic solvent was the highest for ML and the lowest for CA.
All the four peptides displayed potent activity against C. albicans and C. tropicalis strains, though a higher MIC was observed for MG against C. albicans. Two C. tropicalis strains displayed higher MIC values against ML, MG, and CA peptides, but were susceptible to MB. Thus, MG was the least potent and MB was the most potent against the two strains. Both C. parapsilosis and C. glabrata were not susceptible to CA and showed enhanced resistance to ML. Memariani and Memariani compiled the antifungal properties of ML against a wide range of fungal species and indicated a more than 50 times increase in the MIC of ML against C. glabrata when compared to C. albicans [30], thus corroborating our in vitro assays. A total of 7 of the 10 C. glabrata strains were susceptible to MB peptide, whereas the other three strains displayed heightened resistance to the peptide. These in vitro results shed further light on the remarkable resistance of C. glabrata against prolific pore-forming peptides. A number of surveillance studies documented the enhanced resistance of C. glabrata against echinocandins, cyclic lipopeptides, underlining the need for alternative therapies for treating such infections [31][32][33]. The reduced susceptibility of C. glabrata to salivary proteins histatin-5, human-β-defensins 2 and 3, lactoferrin, sheep myeloid antimicrobial peptide, and other natural antimicrobial peptides has been reported. Phylogenetically, C. parapsilosis and C. glabrata are a distant lineage from C. albicans, and C. glabrata is more closely related to non-pathogenic Saccharomyces cerevisiae. Schmalreck et al. reported a linkage between azole susceptibility and phylogenetic relationship among 20 pathogenic fungal species covering nearly 10,000 yeast strains [34]. The observation that ML required a high concentration to elicit fungicidal activity against S. cerevisiae and C. glabrata augment the lineage specific susceptibility to pore-forming peptides [35].
The antifungal properties of ML and CA against C. albicans have been attributed to the interaction of the peptides with cytoplasmic membrane with a concomitant increase in membrane potential loss [36,37]. It has been shown that ML displayed lethal candidacidal properties through membrane disruption, and the lethal action was energyindependent [36]. We determined that ML was potent against various Gram-negative and Gram-positive bacteria at sub-µM concentration, but a higher concentration was required for the inhibition of C. parapsilosis and C. glabrata strains. The lack of activity is substantial against C. glabrata strains as no detectable MIC was observed within the tested concentration range. These results suggest the distinctive feature of the cell wall components of C. glabrata that would interfere with the interaction of peptides with cytoplasmic membrane [38].
All the four peptides displayed varying degree of cytotoxicity for skin cells, as determined by MTS and high-throughput screening. ML displayed heightened toxicity to both keratinocytes and fibroblasts followed by MB, whereas MG and CA remained non-cytotoxic, even having elevated concentration. In support of these observations, MG and CA did not affect the migration of keratinocytes, whereas a significant delay was observed for both MB and ML peptides, the latter causing the maximum delay in wound closure.
Among the various α-helical peptides, MB displayed the most potent antifungal activity against both C. albicans and NACs. Secondary structure analysis and zeta potential measurements indicated that ML had the highest α-helical content and zeta potential values. There was no clear correlation between the antifungal properties of α-helical peptides with the experimentally determined physicochemical parameter. In addition, MB displayed salt-concentration-dependent antifungal activity, a common trait associated with cationic antimicrobial peptides. Sytox green uptake assay indicated rapid uptake of the DNA-binding probe upon the addition of peptide to C. glabrata or C. parapsilosis strains, indicating that the peptide perturbs the cytoplasmic membrane of the yeast strains. Time-kill kinetics studies confirmed the substantial lethality associated with peptide-induced membrane perturbation.
To determine the physicochemical properties that can describe the antifungal activity of MB, a comprehensive literature search on short α-helical peptides that reported the MIC values of the peptides against C. glabrata strains was carried out. For a comparison, we have included C. tropicalis strains in order to infer whether phylogenetically distant species share a common susceptibility pattern against α-helical peptides. The analysis indicated that the overall net charge of the peptides and H*Q or µH*Q could describe the anti-C. glabrata activity of these peptides. Similarly, H and µH of the peptides could describe the anti-C. tropicalis activity of peptides. The correlation between physicochemical parameters and antifungal susceptibility is striking considering that the above analysis was carried out among a diverse heterogenous group of peptides. Taken together, these observations suggest different physicochemical requirements for antifungal peptide susceptibility among phylogenetically different Candida spp. strains. The approach reported here enabled us to determine the key physicochemical parameters of MB.
In conclusion, we compared the antifungal properties of prolific pore-forming antimicrobial peptides and their cytotoxicity for skin cells. Among the four pore-forming peptides, MB possessed physicochemical properties that can be fine-tuned to be effective against various Candida spp. strains.

Supplementary Materials:
The following supporting information can be downloaded online: Figure  S1: RP-HPLC of MB; Figure S2: RP-HPLC of MG; Figure S3: RP-HPLC of ML; Figure S4: RP-HPLC of CA; Figure S5: CD spectra of pore-forming peptides in (a) PBS and in (b) 30% TFE. The peptide concentration was 200 µM; Table S1: Amino acid sequence, physicochemical parameters and MIC values of peptides against C. glabrata reported in the literature; Table S2: Amino acid sequence, physicochemical parameters and MIC values of peptides against C. tropicalis reported in the literature.