Evaluation of Antifungal Efficacy of Three New Cyclic Lipopeptides of the Class Bacillomycin from Bacillus subtilis RLID 12.1

ABSTRACT

New lipopeptide homologues (AF₃, AF₄, and AF₅) with antifungal activities against Candida and Cryptococcus spp. were purified from a cell-free supernatant of Bacillus subtilis RLID 12.1. The lipopeptides AF₃, AF₄, and AF₅ were identified with the same peptide sequence Asn-Pro-Tyr-Asn-Gln-Thr-Ser with variations in the fatty acid branching type and chain length (anteiso-C₁₇, iso-C₁₇, and iso-C₁₈, respectively). Upon comparing the three homologues for MICs against 81 Candida (n = 64) and Cryptococcus (n = 17) clinical isolates and their cytotoxicities, we found that AF₄ was the most promising antifungal lipopeptide, since it demonstrated 100% inhibition at geometric mean MICs of 3.31, 3.41, 3.48, and 2.83 μg/ml against Candida albicans, Candida tropicalis, Candida auris, and Cryptococcus neoformans, respectively, with low hemolysis values (<6%) and 50% inhibitory concentrations (13.31 μg/ml). The additive effects among the homologues AF₃, AF₄, and AF₅ were evaluated against three Candida species, along with the cytotoxicity studies. Five combinations exhibited good additive interaction effects: AF₃/AF₄ (at corresponding concentrations of 4 and 4 μg/ml [4/4 μg/ml]), AF₃/AF₅ (4/4 μg/ml), AF₃/AF₅ (2/4 μg/ml), AF₄/AF₅ (4/4 μg/ml), and AF₄/AF₅ (2/4 μg/ml) in planktonic cell inhibition and AF₃/AF₄ (4/4 μg/ml), AF₃/AF₅ (4/4 μg/ml), and AF₃/AF₅ (2/4 μg/ml) in the inhibition of biofilm formation. However, combinations AF₃/AF₄ and AF₃/AF₅, which showed >70% cell survival with low hemolysis (<5%), were found to be comparatively effective. We describe here the additive effects of lipopeptide homologues showing reduced cytotoxicity against mammalian cells; these combinations might serve as a potent antibiofilm-forming substitute.

INTRODUCTION

Over the last three decades, the incidence of invasive fungal infections (IFIs) pertaining to Candida spp. and Cryptococcus spp. has been on the rise across the globe (1, 2). The management of IFIs is difficult due to the change in epidemiology, less-than-optimum diagnostic capability, and the emergence of antifungal resistance (3). In a recent study in India, more than 27 intensive care units (ICUs) reported 1,400 candidemia cases. Two major observations were made in that study: that multidrug-resistant Candida auris had spread across India and that 11.8% Candida isolates are resistant to azoles. Only 55% of isolates were clearly susceptible, while the others are in a susceptible dose-dependent range (1). Other than azoles, the rise of echinocandin-resistant Candida glabrata in many countries has become a matter of concern in managing invasive candidiasis (4).

The antifungal agents active against IFIs are broadly classified into polyenes, azoles, echinocandins, allylamines, and flucytosine. Majority of these agents are fungistatic and toxic and create problems during management due to drug interactions. The newest generation of native peptide molecules, termed “natural antibiotics,” are unusually active against a large spectrum of microorganisms, including bacteria and filamentous fungi. Bacillus spp. are well known for the production of such antifungal lipopeptides such as bacillomycin, iturin A, mojavensin, mycosubtilin, bacillopeptin, and fengycin homologues. These lipopeptides target the cytoplasmic membrane by forming ion-conducting pores in the lipid membrane (5). However, their strong hemolytic activity at their corresponding MIC values and their limited antifungal activity have restricted their use in clinical applications (5–7). In the present investigation, antifungal lipopeptides produced by Bacillus subtilis RLID 12.1 were extracted, purified to homogeneity, and studied for their therapeutic efficiency. Further, interactive studies of the three purified lipopeptide homologues/isomers (AF₃, AF₄, and AF₅) were also examined. Antifungal compound production by B. subtilis RLID 12.1 was screened and reported in our previous paper (8).

RESULTS

The production of antifungal compounds by B. subtilis RLID 12.1 started at 36 h (early log phase) and attained a maximum at 60 h of growth (late log phase) (see Fig. S1 in the supplemental material). The pH values increased gradually from the initial pH of 7.4 to 9.5 until the death phase, and it was maintained at pH ∼8.5 during the maximum production of antifungal compounds (see Fig. S1 in the supplemental material). The antifungal compound was partially purified by adsorption chromatography at the different solvent ratio of chloroform and methanol and was further purified using reversed-phase high-pressure liquid chromatography (RP-HPLC) (see Fig. S2 in the supplemental material). All of the peaks were collected and tested for antifungal activity against C. albicans SC5314; of these, five fractions (AF₁, AF₂, AF₃, AF₄, and AF₅) were found to be bioactive. Thin-layer chromatography (TLC) of all the five different HPLC fractions were carried out, and the R_f values of bioactive fractions were observed in a bioautography assay. Fraction AF₁ showed two active spots at an R_f of 0.51 and 0.43, AF₂ at an R_f of 0.51 and 0.39, AF₃ at an R_f of 0.53, AF₄ at an R_f of 0.53, and AF₅ at an R_f of 0.55. Fractions AF₁, AF₂, AF₃, AF₄, and AF₅ showed yellow spots when exposed to iodine vapor and blue spots after Serva Blue W staining, suggesting their lipopeptide nature.

Identification of antifungal substances.

The major peaks of AF₁, AF₂, AF₃, AF₄, and AF₅ were obtained at m/z [M+H]⁺ 1,043.59, 1,057.69, 1,071.57, 1,071.68, and 1,085.61, respectively, using matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS). There was an exact 14-Da difference among the peaks of AF₁, AF₂, AF₃, AF₄, and AF₅ corresponding to the mass of CH₂, indicating the homologous nature of the compounds. Since AF₃, AF₄, and AF₅ exhibited lower MIC values (see Table S1 in the supplemental material) with the anti-Candida potential, further focus was given to molecular characterization of these compounds. Mass spectra obtained from MALDI-TOF tandem mass spectrometry (MS/MS) and electrospray ionization Fourier transform ion cyclotron resonance (ESI-FT-ICR) MS/MS analyses (the spectral data are not shown here) revealed that all three lipopeptides—AF₃, AF₄, and AF₅—share the similar peptide sequence Asn-Pro-Tyr-Asn-Gln-Thr-Ser-Xaa, where Xaa is a β-amino fatty acid variant. Using gas chromatography-mass spectrometry (GC-MS), the β-amino fatty acids of AF₃, AF₄, and AF₅ were identified as anteiso-C₁₇, iso-C₁₇, and iso-C₁₈, respectively.

Determination of MIC and MFC.

The MIC and minimum fungicidal concentration (MFC) values of the purified compounds were evaluated against 81 yeast strains, which included standard and clinical isolates of Candida spp. (n = 64) and Cryptococcus spp. (n = 17) (Table 1 and see Table S1 in the supplemental material). Amphotericin B and fluconazole were also tested simultaneously. The MIC₉₀ value of Amphotericin B ranged from 0.06 to 1.0 μg/ml. For fluconazole, MIC₅₀ value ranged from 0.25 to 4 μg/ml; two clinical isolates of C. albicans, four isolates of C. krusei, and three isolates of C. parapsilosis demonstrated a higher MIC₅₀ range of about 8 to 64 μg/ml. The MIC ranges of HPLC-purified fractions for the test compounds AF₁, AF₂, AF₃, AF₄, and AF₅ against Candida spp. were 8 to 20 μg/ml, 8 to 20 μg/ml, 4 to 16 μg/ml, 2 to 4 μg/ml (except for C. parapsilosis [16 μg/ml]), and 2 to 16 µg/ml (except for C. parapsilosis [32 μg/ml]), respectively, whereas for the same fractions MIC range recorded against Cryptococcus spp. were 8 to 16 μg/ml, 8 to 10 μg/ml, 2 to 8 μg/ml, 1 to 4 μg/ml (except one C. neoformans var. grubii [8 μg/ml]), and 2 to 4 μg/ml (except one strain of C. neoformans var. grubii and C. neoformans var. gattii [8 μg/ml]), respectively (see Table S1 in the supplemental material). The MIC and MFC values for all the compounds were found to be same for most of the organisms tested (Table 1 and see Table S1 in the supplemental material). Comparisons of MICs for five different compounds were statistically significant (P < 0.05). No trailing endpoints were encountered for any of the five compounds. Overall, 95% of the MIC and MFC endpoint values for AF₄ were the same, except for the MFCs of two clinical isolates of fluconazole-resistant C. parapsilosis and C. albicans, where it was determined to be 1- or 2-fold higher than the respective MIC value.

TABLE 1.

MICs, MFCs, and MIC_gs of Candida and Cryptococcus spp.

Yeasts (no. of strains)	MIC, MFC, or MICg (μg/ml)
	AF3		AF4		AF5
	MIC	MFC	MIC	MFC	MIC	MFC
Candida spp. (64)
C. albicans (11)*	7.51	7.51	3.31	3.76	4.26	4.26
C. tropicalis (13)*	7.19	7.19	3.41	3.41	4.95	4.95
C. krusei (4)a	8–16	8–16	4	4	4–8	4–16
C. glabrata (11)*	7.05	7.51	3.31	3.53	5.84	5.84
C. haemulonii (6)a	8	8	4	4	4–16	4–32
C. parapsilosis (8)a	8–16	8–16	4–16	4–16	4–32	8–32
C. auris (10)*	8.57	10.55	3.48	3.48	6.498	6.498
C. viswanathii (1)a	4	4	2	2	2	2
Cryptococcus spp. (17)
C. neoformans var. grubii (11)*	6.17	6.17	2.83	2.83	2.59	2.83
C. neoformans var. gattii (6)a	4–8	4–8	2–4	2–8	2–8	2–8

^aSome values are expressed as MIC ranges. For strains indicated by an asterisk (*) in column 1, the geometric mean MIC (MIC_g) is indicated rather than the MIC.

Determination of hemolytic concentration (HC).

To determine the effects of any potential cytotoxicity, the effects of various concentrations of all the five purified antifungal compounds on 5% human erythrocytes were studied (Fig. 1). HC₁₀ and HC₅₀ indicate the concentrations of antifungal compound exhibiting 10% and 50% hemolysis, respectively. AF₃ showed HC₁₀ and HC₅₀ at 8 and 16.4 μg/ml respectively; however, AF₄ displayed HC_<10 (∼5.6%) at 8 μg/ml and HC₅₀ at 17.7 μg/ml, whereas AF₅ showed HC₅₀ at 8 μg/ml (Fig. 1). In contrast, AF₁ and AF₂ exhibited 100% hemolysis at concentrations less than their MICs.

FIG 1.

Hemolytic activities of AF₃, AF₄, and AF₅ fractions against human erythrocytes. The dashed lines represent HC₁₀ and HC₅₀. Each data point represents the mean result ± the standard deviation (error bars) of experiments performed in triplicate.

Biofilm formation inhibition.

Biofilm formation was quantified using an XTT dye reduction assay for the selected seven Candida species after 48 h (see Fig. S3 in the supplemental material). AF₃, AF₄, and AF₅ were tested for their ability to inhibit biofilm formation (to about 50%) by the preadhered cells. AF₃ reduced biofilm formation by the adhered cells at concentrations equal to the MIC for C. albicans SC5314 and C. glabrata ATCC 2001 and at 2× MIC for the other tested organisms (Table 2). In the case of AF_4, the inhibition of the biofilm formation by the adhered cells was observed at 4× MIC for C. krusei NCCPF 440022 and C. parapsilosis NCCPF 450033, whereas for the remaining organisms 2× MIC yielded 50% inhibition (Table 2). AF₅ showed 50% inhibition at concentrations of 1× and 4× for C. parapsilosis NCCPF 450033 and C. auris NCCPF 470067, respectively, and at 2× MIC for the remaining tested organisms (Table 2).

TABLE 2.

Determination of SMIC₅₀s

Organism	Strain	MIC or SMIC50 (μg/ml)a
		MIC (planktonic cells)			SMIC50 (biofilm)
		AF3	AF4	AF5	AF3	AF4	AF5
C. albicans	SC5314	8	2	4	1×	2×	2×
	ATCC 24433	8	4	4	2×	2×	2×
C. auris	NCCPF 470067	8	2	4	2×	2×	4×
C. haemulonii	ATCC 22991	8	4	4	2×	2×	2×
C. krusei	NCCPF 440022	16	4	8	2×	4×	2×
C. glabrata	ATCC 2001	8	4	4	1×	2×	2×
C. parapsilosis	NCCPF 450033	16	4	8	2×	4×	1×

^aSMIC₅₀ (sessile MIC) indicates the antifungal concentration which produces 50% inhibition of biofilm formation.

Time-kill assay.

Time-kill kinetics studies were performed with 2× MIC of AF₃, AF₄, and AF₅ on C. albicans ATCC 24433 in RPMI 1640 medium. AF₃ and AF₅ displayed a 2-log₁₀ decrease (99%) in CFU/ml by 8 h of treatment, whereas AF₄ decreased the number of viable cells to 99% (2 log₁₀) at between 8 and 12 h of treatment (Fig. 2).

FIG 2.

Time-kill kinetic studies of AF₃, AF₄, and AF₅ against C. albicans ATCC 24433. Each data point represents the mean result ± the standard deviation (error bars) of experiments performed in triplicate.

Cytotoxicity assay.

The cytotoxic activity of all the antifungal compounds against human embryonic kidney 293 (HEK293), human immortalized keratinocyte (HaCaT), human cervical cancer (HeLa), and human type II alveolar epithelial (A549) cell lines were assessed using an MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide] assay (Fig. 3). The 50% inhibitory concentrations (IC₅₀s) varied depending on the cell line used (Table 3). All three compounds showed dose-dependent cytotoxicity effects against the HEK293, HaCaT, and HeLa cell lines, with >70% cell survival at 8 μg/ml when the HaCaT and A549 cell lines were used (Fig. 3). AF₄ showed an IC₅₀ of 13.31 μg/ml for the untransformed cell lines but IC₅₀s of 23.18 and 34 μg/ml for the HeLa and A549 cell lines, respectively, whereas AF₃ exhibited an IC₅₀ of 13 to 15.5 μg/ml for normal and HeLa cell lines. The IC₅₀s of AF₅ were in the range of 9 to 13 μg/ml.

FIG 3.

Cytotoxicity profiles of AF₃, AF₄, and AF₅ lipopeptides at various concentrations against the mammalian cell lines. (a) HaCaT; (b) HEK293; (c) A549; (d) HeLa. Each data point represents the mean result ± the standard deviation (error bars) from experiments performed in triplicate.

TABLE 3.

IC₅₀s of AF₃, AF₄, and AF₅ against four mammalian cell lines

Cell line	IC50 (μg/ml)
	AF3	AF4	AF5
HEK293	14.53	13.12	9.5
HaCaT	15.69	13.33	11.45
A549	>32	34.4	>32
HeLa	13.04	23.18	12.57

Interaction of three lipopeptides.

Interactions among AF₃, AF₄, and AF₅ were studied using checkerboard analysis against three organisms, which were selected based on their susceptibility to the three homologues or isomers. Of the three organisms, the two clinical isolates C. parapsilosis NCCPF 450033 and C. krusei NCCPF 440022 demonstrated high MICs compared to C. albicans ATCC 24433. Different combinations showing 80% (MIC₈₀) and 50% (MIC₅₀) inhibition with different fractional inhibitory concentration (FIC) index values are listed in Table 4 for both planktonic cells and biofilm formation inhibition. Interactive effect of AF₃, AF₄, and AF₅ were studied in planktonic cells at a final concentration of about 10⁵ cells/ml (see Table S2 in the supplemental material). MIC values of the individual compounds for the three organisms are summarized in Table 2 and see Table S2 in the supplemental material. Of eight combinations, 80% inhibition of C. albicans ATCC 24433 was shown by combinations AF₃/AF₅ (at corresponding concentrations of 2 and 4 μg/ml [2/4 μg/ml]), AF₄/AF₅ (2/4 μg/ml), and AF₃/AF₅ (4/4 μg/ml), with additive interactions. The combination AF₄/AF₅ (2/4 μg/ml) showed additive interaction with 50% inhibition of C. parapsilosis NCCPF 450033 and C. krusei NCCPF 440022 (Table 4). The combination of AF₃ and AF₅ inhibited the biofilm formation to about 50%, with considerable additive effect against C. albicans ATCC 24433. The AF₃/AF₄ (4/4 μg/ml) combination showed additive interaction in inhibiting both planktonic cells and biofilm formation in case of all the three selected strains (Table 4). On the other hand, the AF₄/AF₅ combination did not inhibit the biofilm formation effectively. The MIC needed for 50% biofilm formation inhibition (SMIC₅₀) was obtained using the AF₃/AF₅ combination at 4/4 μg/ml and at 2/4 μg/ml against C. krusei NCCPF 440022. Henceforth, these five combinations were evaluated for cytotoxicity. All of these combinations showed negligible hemolysis (<5%) and cell viability (>70%) for both the untransformed and cancer cell lines tested (Table 5).

TABLE 4.

Percent inhibitions and FIC indices for the interaction effects of homologues AF₃, AF₄, and AF₅ determined using C. albicans ATCC 24433, C. parapsilosis NCCPF 450033, and C. krusei NCCPF 440022 as target organisms^a

Combination	Concn (μg/ml)b	C. albicans ATCC 24433				C. parapsilosis NCCPF 450033				C. krusei NCCPF 440022
		Planktonic cells		Biofilm formation		Planktonic cells		Biofilm formation		Planktonic cells		Biofilm formation
		%I	FIC	%I	FIC	%I	FIC	%I	FIC	%I	FIC	%I	FIC
AF3/AF4	4/4	80	0.75	50	0.75	50	0.75	50	0.375	80	0.625	50	0.375
	2/4	80	0.625	50	0.625	<50		<50		<50		<50
AF3/AF5	4/4	80	0.5	50	0.375	50	0.531	<50		50	0.75	50	0.375
	0.5/8	80	0.53	50	0.5	<50		<50		80	0.515	<50
	2/4	80	0.375	50	0.3125	<50		<50		<50		50	0.375
AF4/AF5	4/4	80	0.75	50	1	50	0.75	<50		80	0.75	<50
	4/2	80	0.625	50	0.75	<50		<50		<50		<50
	2/4	80	0.5	50	0.75	50	0.5	<50		80	0.5	<50

^aThe percent inhibition (%I) and fractional inhibitory concentration (FIC) index values for the interaction effects of homologues AF₃, AF₄, and AF₅ determined using C. albicans ATCC 24433, C. parapsilosis NCCPF 450033, and C. krusei NCCPF 440022 as the target organisms were determined.

^bConcentrations are expressed as follows for the combinations listed in column 1: concentration of first component/conentration of second component.

TABLE 5.

Hemolytic and cytotoxic effects of the AF₃, AF₄, and AF₅ combinations in interaction study^a

Combination	Concn (μg/ml)b	Mean hemolysis (%) ± SD	Mean cell viability (%) ± SD
			HaCaT	HEK293	A549	HeLa
AF3/AF4	4/4	3.87 ± 3.7	68.05 ± 2.3	68.81 ± 0.9	76.61 ± 5.4	90.04 ± 3.8
AF3/AF5	4/4	3.43 ± 3.1	78.27 ± 1.7	70.42 ± 5.8	84.07 ± 3.5	93.16 ± 4.6
	2/4	4.11 ± 5.9	75.56 ± 2.5	72.58 ± 1.8	63.17 ± 5.2	70.16 ± 6.4
AF4/AF5	4/4	3.12 ± 3.4	73.08 ± 3.9	60.90 ± 2.6	60.93 ± 2.8	69.01 ± 0.5
	2/4	2.96 ± 4.1	70.31 ± 6.4	71.98 ± 4.2	72.94 ± 3.3	77.26 ± 6.9

^aEach data point represents the standard deviation from three experiments.

^bConcentrations are expressed as follows for the combinations listed in column 1: concentration of first component/conentration of second component.

DISCUSSION

Spectral analyses of the antifungal lipopeptides produced by B. subtilis RLID 12.1 indicated that the amino acid sequences of all the three homologues/isomers were similar to the bacillomycin D homologues with variations in the amino acid positions. Bacillomycin D homologues were reported as cyclic lipopeptides with the sequence Asn-Tyr-Asn-Pro-Gln-Ser-Thr-Xaa, where Xaa is the β-amino fatty acid varying from C₁₄ to C₁₇ with n-, iso-, or anteiso-terminal structures (9, 10). Our analyses reveal the predicted sequences of AF₃, AF₄, and AF₅ with same amino acids as reported in bacillomycin D homologues, but the positions of the amino acids were altered as Asn-Pro-Tyr-Asn-Gln-Thr-Ser-Xaa, where Xaa corresponds to the C₁₇ (antesio- and iso-) and C₁₈ (iso-) β-amino fatty acid chains, respectively. Variations observed in the structural elucidation of cyclic lipopeptides with the fatty acid chain lengths of C₁₇ and C₁₈ might have improved the antagonistic nature of AF₃, AF₄, and AF₅ against Candida and Cryptococcus spp. with reduced cytotoxicity compared to the other antifungal lipopeptides reported in Bacillus spp. (11, 12).

Due to the high MIC values (see Table S1 in the supplemental material) obtained with AF₁ and AF₂, no fatty acid, biofilm inhibition, time-kill, and additive interaction studies were conducted for these compounds. The MIC and MFC values of AF₄ (Table 1) against C. albicans, C. tropicalis, C. auris, and C. neoformans are very encouraging. The MIC and MFC values of AF₅ were observed at the concentration range of 2 to 16 μg/ml. In the context of the Indian subcontinent (13), the rising incidence of infections caused by C. auris and the associated drug resistance have raised concerns among clinicians (1). In our study, we observed the geometric mean MIC (MIC_g) of 3.48 μg/ml for AF₄ against the clinical isolates of C. auris (Table 1), a concentration at which negligible hemolysis was observed. In a recent study from Hong Kong, an alarming rise in resistance has been noted, with 40% of C. albicans, 10% of C. tropicalis, 11.1% of C. parapsilosis, and 100% of C. glabrata strains reported as fluconazole resistant (14). In the present study, fluconazole-resistant C. albicans (n = 1), C. parapsilosis (n = 1), and C. krusei (n = 4) strains displayed MICs of 4, 8, and 4 μg/ml, respectively, in the case of AF₄ (see Table S1 in the supplemental material). An MIC range of 2 to 4 μg/ml was observed for C. glabrata, which is encouraging since limited treatment options are presently available for C. glabrata infection due to its azole and echinocandin resistance (4). Reports on lipopeptide activity against Cryptococcus spp. are scanty compared to Candida spp. Echinocandin lipopeptides were reported to show poor antifungal activity against C. neoformans (15). The anticryptococcal prowess of AF₄ and AF₅ with low MIC range is either superior or comparable to standard antifungals tested in the present study (Table 1; Table S1). Similar to the present work, a previous study reported promising antifungal activity of a novel membrane-active peptide and its d-enantiomer, with an MIC in the range of 2 to 4 μg/ml against pathogenic clinical isolates of C. neoformans (16). YM-47522 compound produced by Bacillus spp. showed antifungal activity against C. neoformans at ∼6.25 μg/ml (17).

Iturin-like lipopeptides disrupt the membrane integrity of mammalian cells at different concentrations (dose dependent), depending on the nature of the compound and its critical micelle concentrations (18), leading to limited clinical usage. The hemolytic study was conducted using all five compounds wherein AF₄ exhibited a negligible 0 to 5.6% hemolysis at an MIC of 1 to 8 μg/ml, and the recorded cell survival at about 80% at 4 μg/ml (Fig. 3) suggests this compound's antimicrobial efficacy. In previous studies, the MIC values of iturin A, fengycin, and lipopeptide 6-2 antiCA produced by B. subtilis and B. amyloliquefaciens against C. albicans were found to be 5 to 10, 15.62, and 7.0 μg/ml, respectively (19–21). However, percentage of hemolysis recorded was high for all these compounds at the respective MICs. Iturin A, bacillomycin D, bacillomycin Lc, and mycosubtilin were reported to induce 25, 28, 75, and 100% hemolysis, respectively, at MICs of 10 μg/ml, indicating the limited therapeutic applications of these compounds (11). Similarly, three bacillomycin D-like lipopeptides a₁, a₂, and a₃ from B. subtilis B38 had been reported with different antimicrobial potency, where a₃ was found to be very active compared to 50% hemolysis at the reported MIC range (12). The IC₅₀ varied depending on the nature of the cell lines tested. IC₅₀s of AF₃, AF₄, and AF₅ on A549 cell lines was found to be high compared to other cell lines tested. It is worth noting that the IC₅₀ of AF₄ is ∼13.3 μg/ml and that the MIC_g values of Candida spp. and Cryptococcus spp. were between 3.31 to 3.48 and 2.83 μg/ml, respectively, with hemolysis values not exceeding 4% at an MIC of 4 μg/ml. Therefore, considering the low MIC values, hemolytic values, and cytotoxicity values, AF₄ was the most potent compound of the five extracted antifungal lipopeptides of B. subtilis RLID 12.1.

Although the MIC values of AF₄ for planktonic cells were found to be low, the SMIC₅₀ was found to be high (4 to 16 μg/ml), i.e., about 4× MIC (Table 2). AF₄ showed 50% cytotoxicity for HEK293, HaCaT, and HeLa cells, as well as being hemolytic at 4× MIC (Fig. 3), making it less applicable for biofilm inhibition. Despite this, it is worth mentioning that AF₄ at 1× and 2× MIC showed SMIC₅₀s against a C. glabrata strain (Table 2), which is favorable since biofilms of C. glabrata are unmanageable by antimicrobial therapy due to the natural properties of growth (22).

Synergistic action is a very well-known property of antifungal lipopeptides, which enables their broad antimicrobial spectrum applications (23). To overcome the cytotoxic effect of AF₄ at high SMIC₅₀ and low IC₅₀ of AF₅, as indicated in Tables 2 and 3, we attempted to study the interactive effects of three homologues of AF₃, AF₄, and AF₅, which has not been reported before. Eight active combinations were obtained with various FIC values (Table 4) based on the susceptibilities of three selected organisms to lipopeptides. Although some combinations exhibited FIC values of ≤0.5, the interactions were altogether found to have an additive rather than a synergistic effect, since all of the lipopeptides follow the same mechanism of action. Lipopeptide molecules penetrate the cytoplasmic membrane and form oligomer-like structures that produce ion-conducting pores in the target cells (5). All five combinations—AF₃/AF₄ (4/4 μg/ml), AF₃/AF₅ (4/4 μg/ml), AF₃/AF₅ (2/4 μg/ml), AF₄/AF₅ (4/4 μg/ml), and AF₄/AF₅ (2/4 μg/ml)—showed effective additive actions in terms of planktonic cell inhibition, whereas for biofilm formation inhibition the AF₃/AF₄ (4/4 μg/ml), AF₃/AF₅ (4/4 μg/ml), and AF₃/AF₅ (2,4 μg/ml) combinations were suitable (Table 4). This observation highlights the fact that interactions among the lipopeptide homologues eventuate during their coproduction. As a result of an additive study, two specific combinations, AF₃/AF₄ and AF₃/AF₅, were found to exhibit strong anti-biofilm-forming potential with ≥70% cell survival and <4% hemolysis. Working on the similar line, additive type interaction was reported with [Ile⁷]surfactin and bacillomycin D in suppressing the gray mold disease on the cucumber leaves by Botrytis cinerea (24). Maget-Dana et al. (25) explained the mechanism of the synergistic effect of surfactin on the biological properties of iturin A. Synergistic action of surfactin and fengycin, as well as iturin and fengycin, was also reported in the literature (5, 26). Antifungal lipopeptides bacillomycin D and fengycin of Bacillus amyloliquefaciens strain FZB42 showed a synergistic effect against Fusarium oxysporum strain (27). Liu et al. reported a synergistic study between two surfactin homologues (C₁₄ and C₁₅) with ketoconazole against C. albicans with MIC values of 12.5 and 6.25 μg/ml (28). Tabbene et al. (12), in a separate study, reported bacillomycin D showing a synergistic effect with amphotericin B against Candida strains, with FIC indices ranging from 0.28 to 0.5. Interestingly, when these two drugs were used together at these dosages (displaying a synergism in the anti-Candida activity), the cytotoxic effect was reduced (29).

Studies of the MIC and MIC_g values for AF₄ lipopeptide against 81 yeast strains showed that all of these values were ≤4 μg/ml except for five clinical non-albicans Candida isolates which exhibited MICs of >4 μg/ml (8 μg/ml for four isolates and 16 μg/ml for one isolate); only one clinical C. neoformans var. grubii isolate exhibited an MIC value of 8 μg/ml (see Table S1 in the supplemental material). This explains why only 6.2% of the tested strains in the present study exhibited an MIC of 8 μg/ml for AF₄, at which the observed hemolysis value was <10% (∼5.6%) (Fig. 1). Besides this, the additive interaction study (Table 3) clearly indicates that the two combinations, AF₃/AF₄ and AF₄/AF₅, can be used against yeast strains that are less sensitive to AF₄ since they exhibited negligible hemolysis (<5%) and approximately 70% cell viability (when tested against four mammalian cell lines). These observations indicate that our novel compound AF₄ and its combinations may not have significant toxicity. However, we require further experimentation regarding the toxicity of our novel compound in future studies.

Another important finding is that the MICs of amphotericin B and fluconazole increased with the increased incubation periods, whereas the MICs of AF₃, AF₄, and AF₅ remained the same even after 72 and 96 h of incubation in the presence of Candida and Cryptococcus spp., respectively. Taken together, the total picture that emerges from the present study is the preliminary structural elucidation of the highly efficacious antifungal compound and also the finding of additive combinations showing antibiofilm-forming potential with less cytotoxicity. AF₄ and the five additive combinations may be considered an alternative to conventional antifungal agents, although more trials are warranted. Based on these findings, the combinations assessed here can be further evaluated for their synergistic action with existing antifungal drugs as well, which may ameliorate the antifungal efficacy reducing the SMIC₅₀, as well as the toxicity, which in turn help to combat rising antifungal resistance.

MATERIALS AND METHODS

Growth kinetics.

Modified Trypticase soy broth (with 0.5% yeast extract) was inoculated with freshly grown B. subtilis RLID 12.1 and incubated at 37°C with constant stirring at 110 rpm. After every 12 h, a 2-ml sample was drawn and centrifuged at 10,000 rpm for 15 min. The supernatant was membrane filtered through 0.45-μm-pore size disposable filter, concentrated using 10-kDa centrifugal filters (Millipore; Merck), and assayed for antifungal activity against C. albicans SC5314 (grown in Sabouraud dextrose agar). The growth of RLID 12.1 at an optical density at 600 nm and pH values at the respective time interval were recorded simultaneously.

Extraction and purification of antifungal compound.

Cell-free supernatant collected after 60 h was adjusted to pH 2.0 by the addition of 12 N HCl and extracted with n-butanol (30). The solvent-extracted supernatant was further purified using silica gel (230 to 400 mesh) as the stationary phase and chloroform with methanol as the mobile phase at different ratios. The pooled active adsorption chromatography fractions were dissolved in 70% methanol and fractionated by RP-HPLC in a semipreparative scale using Eclipse XDB-C18 column (9.4 by 250 mm; particle size, 5 μm). The solvent system used was water with 0.1% trifluoroacetic acid (TFA) (solvent A) and acetonitrile containing 0.1% TFA (solvent B). The gradients of solvent B used for purification were as follows: 0 to 45% for 0 to 10 min at the flow rate of 1 ml/min, 45 to 54% from 10 to 20 min at 0.5 ml/min, and 54 to 60% from 20 to 48 min at 0.5 ml/min. All fractions eluted from the column were monitored at 214 nm in a diode array detector, and all peaks obtained during HPLC were collected. Fractions of multiple runs were pooled, concentrated by speed vacuum, and tested for antifungal activity against C. albicans ATCC SC5314. The fractions that showed antifungal activity were rechromatographed using an analytical column (4.6 by 250 mm; particle size, 5 μm) with a gradient of 0 to 50% for 25 min at 0.7 ml/min. The peptide concentration was determined by the bicinchoninic acid method (Pierce BCA protein assay kit).

TLC analysis.

RP-HPLC-purified fractions were subjected to TLC using n-butanol–methanol–water (5:1:1 [vol/vol/vol]) as the mobile phase. The bioassays were performed using C. albicans SC5314. The developed TLC plates were sprayed with water, Serva Blue W stain, and iodine for the detection of the hydrophilic nature, the peptide nature, and the lipid moiety of the compounds, respectively (31).

Identification of the antifungal compounds.

Peptide sequence portions of lipopeptides were analyzed and confirmed by Ultraflex MALDI-TOF mass spectrometer (Bruker) using the matrix α-cyano-4-hydroxycinnamic acid (CHCA) (21) and an ESI-FT-ICR mass spectrometer (SolariX; Bruker).

Lipid moieties of lipopeptides were analyzed using GC-MS. Fatty acid methyl esters (FAMEs) were prepared by hydrolyzing the lipopeptides with 6 M HCl at 110°C for 16 h, followed by ether extraction, sulfuric acid treatment, and final extraction with n-hexane. FAMEs were analyzed using DB-5 GC column (0.25 mm, 30 m, 0.25 μm) in a GC-MS-QP 2010 quadruple mass spectrometer (Shimadzu) (32).

MICs and MFCs.

MICs were determined by the broth microdilution protocol according to the recommendations of the Clinical and Laboratory Standards Institute (CLSI; M27-A3) using RPMI 1640 medium with l-glutamine in morpholinepropanesulfonic acid (MOPS) buffered to pH 7.0 for Candida spp. and additional 2% glucose for Cryptococcus spp. (33). The clinical isolates were obtained from the National Culture Collection of Pathogenic Fungi (NCCPF), Post Graduate Institute of Education and Medical Research (PGIMER), Chandigarh, India. Each suspension was diluted to obtain desired final inoculum size (0.5 × 10³ to 2.5 × 10³ CFU/ml) and tested for its sensitivity to increasing concentrations of the purified antifungal compounds at final concentrations ranging from 0.5 to 32 μg/ml for AF₃, AF₄, and AF₅. In the case of AF₁ and AF₂, concentrations ranging from 0.5 to 32 μg/ml and 2.5 to 40 μg/ml were checked. Amphotericin B (range, 0.03 to 16 μg/ml) and fluconazole (range, 0.125 to 64 μg/ml) were used as positive controls to assess the antifungal susceptibility of all of the strains tested. MIC₅₀ and MIC₉₀ values were calculated as the concentrations of fluconazole and amphotericin B that could inhibit 50 and 90% of the isolates, respectively. For lipopeptides, the MIC endpoint was considered the lowest concentration that could inhibit ≥90% of the isolate (no visible growth). Geometric mean MIC (MIC_g) values were used to facilitate comparisons of the activities of the five HPLC-purified active compounds. The results were processed statistically using one-way analysis of variance with Tukey's multiple-comparison test (GraphPad Prism).

Hemolytic assay.

The hemolytic activity of the purified compounds on human erythrocytes was determined as reported previously (34). Briefly, a 5% (vol/vol) suspension of fresh human erythrocytes was incubated with each antifungal compound at final concentrations ranging from 0.5 to 32 μg/ml for 1 h at 37°C with gentle mixing. The tubes were centrifuged (3,000 rpm, 10 min), and the absorbance of the supernatants was measured at 450 nm. We used 1% Triton X-100 and phosphate-buffered saline (PBS) as positive and negative controls.

Inhibition of biofilm formation.

Seven species of Candida were selected for the biofilm inhibition studies. A 100-μl portion of 10⁶ yeast suspension was dispensed into each well of a microtiter plate, and the plates were incubated at 37°C for 8 h. The nonadhered cells were removed, and each well was washed twice with 150 μl of PBS. AF₃, AF₄, and AF₅ at 1× MIC, 2× MIC, and 4× MIC were prepared in RPMI 1640 medium, and 100 μl of sample was transferred to preadhered cells. The plates were incubated at 37°C for 48 h, the medium was aspirated, and each well was washed twice gently with 200 μl of PBS to remove the nonadherent cells. Biofilm without any lipopeptide was taken as the positive control. After treatment with peptides, the medium was removed, and each well was washed twice with 200 μl of PBS. The biofilm formation was quantified by adding 200 μl of using tetrazolium XTT [2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide] reduction assay as described below. Each experiment was repeated three times in duplicate. XTT with menadione (0.5 g/liter XTT, 1 μM menadione) was added, followed by incubation in the dark at 37°C for 2 h. Then, an aliquot of 100 μl from the colored supernatant was transferred to the fresh wells, and the absorbance of the supernatant was determined using a microtiter plate reader at 490 nm (35).

Time-kill assay.

The time-kill kinetics of the peptides on C. albicans ATCC 24433 were determined by treating the lipopeptides at 2× MIC with 10⁵ cells/ml in RPMI medium at 37°C. Cells without lipopeptide addition were used as a control. Aliquots (10 μl) were withdrawn at 4-h intervals, serially diluted, and spread on the Sabouraud dextrose agar plates until 12 h. Plates were incubated at 37°C, and the number of CFU/ml was determined after 48 h. Results were obtained from three independent experiments (36).

Cytotoxicity.

HEK293, HaCaT, HeLa, and A549 cell lines were cultured in complete medium (Dulbecco modified Eagle medium supplemented with 10% fetal bovine serum) in a humidified atmosphere of 5% CO₂ at 37°C. The cells were seeded in 96-well microplates at 0.1 ml/well at a density of 1.5 × 10⁴ cells/well. After 24 h of incubation, the media were removed, and fresh media containing serially diluted peptide (final concentrations, 0.5 to 32 μg/ml) were added and incubated for another 24 h at 37°C and 5% CO₂. Cell viability was determined by MTT assay after 24 h using MTT bromide (Hi-Media). MTT (final concentration, 0.5 mg/ml) solution was added to each well, followed again by incubation for 4 h. Then, 100 μl of dimethyl sulfoxide was added to dissolve the formed formazan. The plates were read on a microplate reader (Multiskan; Thermo Scientific) using test and reference wavelengths of 550 and 650 nm, respectively, to check the cell viability (37).

Interaction effect of three lipopeptide homologues.

Purified peptide-peptide interactions were assessed for AF₃, AF₄, and AF₅ at final concentrations of 0.5 to 32 μg/ml (AF₃), 0.5 to 16 μg/ml (AF₄), and 0.5 to 32 μg/ml (AF₅) for both planktonic cells and biofilm formation in sterile 96-well polystyrene microplates using a checkerboard method. Dilutions were prepared as described previously (38). In the case of planktonic cells, a final suspension of 10⁵ cell/ml (instead of 10³ cell/ml) was added to wells containing combinations. In case of anti-biofilm formation study, preadhered cell layers were prepared as described above, and the combinations were added accordingly to the wells, followed by incubation at 37°C for 48 h. Inhibition (50 and 80%) was quantified using XTT. The FIC is defined as the MIC of the drug used in combination divided by the MIC of drug tested alone. The effects of the peptide-peptide combinations were expressed in term of the FIC index. An FIC index of ≤0.5 was interpreted as a synergistic interaction; an FIC index from 0.5 to 4 indicated an additive interaction, and an FIC index of >4 was interpreted as antagonism. Also, hemolytic and cytotoxicity studies were carried out for the active combinations.