This study was performed to isolate and characterize novel antifungal lipopeptide from Bacillus cereus.
This study was performed to isolate and characterize novel antifungal lipopeptide from Bacillus cereus.
Elucidation of its chemical structure was carried out by electrospray ionization mass spectra (ESI-MS) and Fourier transform infrared spectroscopy (FT-IR). The compound is a cyclic heptapeptide and composed of amino acids, Leu–Asp–Val–Leu–Leu–Leu–Leu. The in vitro activity of Kannurin against various pathogenic yeasts was assessed by CLSI M27-A and moulds by M38-A. It demonstrated broad-spectrum, fungicidal activity against clinically relevant yeasts and moulds. Kannurin exhibited low haemolytic activity and remained active over a wide pH and temperature range. In addition, Kannurin did not bind with melanin particles and was as active in inhibiting biofilms.
An antifungal surfactin-like lipopeptide produced by Bacillus cereus strain AK1 was purified and chemically characterized. We propose to name this lipopeptide compound ‘Kannurin’. To our knowledge, this is the first report of Bacillus cereus producing surfactin-like lipopeptide antibiotic with stronger antifungal activity.
Our results provide a valuable contribution towards a better understanding of the lipopeptide of Bacillus cereus. Moreover, it raises the possibility of using as an alternative antibiotic in clinical medicine.
During the past decades, there has been a concomitant increase in the incidence of mycosis and also the emergence of multidrug-resistant pathogens that has caused serious problems worldwide. Invasive fungal infections, primarily those caused by Candida species, Cryptococcus neoformans and Aspergillus species, constitute a major cause of morbidity and mortality in severely immunocompromised host (Silveira and Husain 2007; Antachopoulos and Walsh 2012). In this connection, much interest has been focused on novel therapeutic approaches using peptide antibiotics. Peptide antibiotics are quite diverse, amphipathic and either ribosomally or non-ribosomally synthesized (Hancock and Chapple 1999). Bacteria and fungi use non-ribosomal peptide synthetases (NRPSs) to produce broad structural and biologically active peptides. This often contains unnatural amino acids (D-amino acids or hydroxy amino acids) and other molecules, not found in ribosomally produced peptides (Ajesh and Sreejith 2009). The assets of these peptides are their diverse potential applications as single antimicrobials or in combination with other antibiotics (Marr et al. 2006).
The genus Bacillus produces a large number of peptide antibiotics representing different basic chemical structures, finding wide applications (Abriouel et al. 2011). Lipopeptides are produced in bacteria and fungi during cultivation on various carbon sources (Makovitzki et al. 2006). A prominent group of bioactive lipopeptides produced by Bacillus species is constituted by iturins, surfactins and lichenysins (Bonmatin et al. 2003). Iturins produced by Bacillus subtilis are a prominent group of cyclic peptidolipids with seven alpha-amino acids and one beta-amino fatty acid (Ajesh and Sreejith 2009). They exhibit strong antifungal activities against various varieties of pathogenic yeasts and fungi, and their activity is related to their interaction with the cytoplasmic membrane of target cells, leading to an increase in K+ permeability (Maget- Dana and Peypoux 1994). Surfactins are cyclic lipopeptides containing seven residues of D- and L-amino acids and one residue of a β-hydroxy fatty acid (Kluge et al. 1998). Fengycin, another lipopeptide complex produced by the B. subtilis strain F-29-3, exhibited strong inhibitory activity against filamentous fungi, but not against yeast (Vanittanakom et al. 1986).
The present study describes the identification of an antifungal lipopeptide-producing strain Bacillus cereus AK1 isolated from soil and the purification, structural elucidation and antifungal spectrum of the lipopeptide Kannurin.
The antifungal lipopeptide-producing strain was isolated from a soil sample collected at the suburb of Kannur city (India) and cultured in (LB) medium at 37°C. To identify the antifungal peptide-producing strain, a series of biochemical tests were performed (Logan and Turnbull 1995). The bacterial genomic DNA was isolated as described by Sambrook and Russel (2001) and amplified by PCR (Applied Biosystems) using the eubacterial specific primers 16SF (AGAGTTTGATCCTGGCTCAG) and 16SR (ACGGCTACCTTGTTACGACTT). The sequence of the PCR product was compared with the acquired sequence of GenBank using the BLAST programme. Also cry gene detection based on PCR was carried out to distinguish the isolate from Bacillus thuringiensis using the forward (5′-GGATTGGAATGGGAAACA-3′) and reverse (5′-AAATAGCCGCATTGACAC-3′) primers (Merck Millipore, India; Tohidi et al. 2013).
Bacillus cereus strain AK1 was inoculated 1% by volume from a 16-h culture into 100 ml in a medium containing 3% peptone, 0·5% yeast extract and 0·5% NaCl (pH 7·0) for 72 h at 30°C at 150 rpm (Wakayama et al. 1984). The culture was centrifuged at 21 000 g for 15 min, and the supernatant was sterilized by membrane filtration (Advantec, Toyo Roshi Kaisha, Ltd, Tokyo, Japan). The sterile culture supernatant fluid was subjected to precipitation by the addition of CaCl2 (final concentration, 1%), and the mixture was kept at 4°C for overnight. The precipitated proteins were pelleted by centrifugation at 10 000 g for 20 min and dissolved in 100 mmol 1−1 EDTA–0·05 mol l−1 Tris–hydrochloride buffer (pH 8·0) and dialysed against 0·05 mol l−1 sodium phosphate buffer, pH 7·0, and dialysed using a benzoylated membrane (Sigma, St. Louis, MO, USA), overnight, at 4°C. Ethanol (final concentration, 80%) was added to the dialysate, and the mixture was allowed to stand for 6 h at 4°C. Precipitates were pelleted by centrifugation at 15 000 g for 20 min. Then, the supernatant was dried in vacuum, dissolved in minimal volume of water and acidified with HCl to pH 3·0. Precipitates were collected by centrifugation (15 000 g, 4°C), resuspended in 0·05 mol l−1 ammonium bicarbonate and dialysed against deionized water. The dialysate as obtained above was submitted to filtration column of Sephadex G-25 (120 × 1·5 cm, Bio-Rad) using 0·05 mol l−1 ammonium bicarbonate buffer with the flow of 30 ml h−1, and fractions of 2·0 ml per tube were taken, and the absorbance was monitored at 220 nm. The fractions were tested for antifungal activity and freeze-dried. The preparation thus obtained was then stored at −40°C for further analysis. For final purification, concentrated active fractions from the Sephadex G-25 gel filtration were applied to a C18 reverse-phase high-performance liquid chromatograph (RP-HPLC) column (Shimadzu, Japan). Mobile phases A (water/TFA (99·95 : 0·05, v/v) and B (acetonitrile/water/TFA (80 : 19·95 : 0·05)) were prepared, and gradient elution starting with 100% A, 0% B changing to 0% A, 100% B over 80 min at a flow rate of 1·0 ml min−1 was performed.
ESI-MS were acquired on an Agilent ion trap mass spectrometer (6340 Series) coupled to an Agilent 1200 series HPLC. The samples were infused to the mass spectrometer through a reversed-phase column (Zorbax SB-C18, 2·1 × 35 cm) with solvent A (0·1% formic acid in water) and solvent B (0·1% formic acid in acetonitrile). The flow rate was maintained at 0·2 ml min−1, and the UV absorbance was monitored at 210 nm. MS data were acquired over an m/z range, 100–1500. MS/MS data were collected using collision-induced dissociation (CID) with the mass spectrometer operated in a data-dependent mode. All data were acquired in positive ionization mode and processed with Bruker Data Analysis software, version 4.0 (Bruker Daltonik GmbH, Bremen, Germany). The FT-IR spectrum of the sample was taken on Perkin-Elmer Spectrum 400 equipped with Pike GladiATR having diamond crystal. The spectrum was recorded from 4000 to 400 cm−1. The spectrum was taken as an average of 30 scans. The resolution of the spectrum was 4 cm−1.
To study the heat stability of the peptide, 10 μl of aliquots and 20 μg of lipopeptide dissolved in sodium phosphate buffer were heated for 15 min at various temperatures (25, 30, 40, 50, 60 and 70°C). Next, the heat treated samples were pipetted into 3-mm wells on Sabouraud dextrose agar (SDA) plate and seeded with C. neoformans. Following incubation, the lawn culture was inspected for zones of inhibition. The pH of the purified peptide was adjusted to the following levels: 5·0, 6·0·7·0, 8·0 and 9·0. After storage for 6 h at 30°C, the samples were assayed for antifungal activity.
In vitro antifungal assay was carried out by CLSI M27-A2 (2002a) and moulds by M38-A2 (2002b). RPMI 1640 (Sigma-Aldrich, St. Louis, MO, USA) was used as the assay medium for all the yeasts and mould strains except Cryptococcus neoformans, for which Dulbecco's modified Eagle's medium (DMEM; Sigma-Aldrich) was used. Cells from overnight-grown liquid cultures (for yeast species) or from the conidial suspension (for mould species) by centrifugation were harvested and washed twice in sterile PBS. Then, the cell concentration was adjusted to 1·0 × 106 cells ml−1 for Candida species and 1·0 × 105 for mould species, in RPMI 1640, or 1·0 × 107 cells ml−1 in DMEM for Cryptococcus and used as the inoculum. Peptide solution was added to the wells of a 96-well plate (100 μl per well) and serially diluted twofold. The final concentrations of peptide mixtures ranged from 0·25 to 512 μg ml−1. After inoculation (100 μl per well, 5 × 103 cells ml−1), the 96-well plate was incubated at 30°C for 48 h, and the absorbance was measured at 620 nm using a microtitre plate reader (iMark Microplate Absorbance Reader; Bio-Rad, Japan). For determining the minimal fungicidal concentration (MFC), 100 μl of cell suspension was taken from each well, centrifuged and washed three times with fresh Sabouraud broth. Then, each cell suspension was vortexed vigorously for 10 s, plated on a Sabouraud dextrose agar plate and incubated at 30°C for 48 h. The MFCs were assessed visually as the lowest concentration of the peptide at which there was no growth.
Human red blood cells were washed and diluted in PBS to a concentration of 2%. 25 μl of the erythrocyte suspension was mixed with 50 μl of lipopeptide, amphotericin B and fluconazole solutions diluted in twofold steps (5–100 μg ml−1) and incubated at 37°C for 1 h. The samples were then centrifuged for 5 min at 250 g, and the optical density was measured at 655 nm. 100% haemolysis was achieved with 0·2% Triton X-100 in physiological solution.
Two millilitres each of stock solutions of amphotericin B (0·25 mg ml−1) in dimethyl sulfoxide and lipopeptide, Kannurin (5 mg ml−1) was incubated with cryptococcal and synthetic melanin (20 mg each) for 2 h at 30°C. Cryptococcal melanin was obtained from C. neoformans cells as previously described (Rosas et al. 2000).The supernatants of the amphotericin B and Kannurin lipopeptide were used in the killing assay with C. neoformans strain as described previously (van Duin et al. 2002).
The MICs for biofilms of Candida albicans (clinical isolate) and Cryptococcus neoformans MTCC 4406 were determined using a microtitre-based XTT (2, 3-bis (2-methoxy-4-nitro-5-sulfophenyl)-5-[(phenylamino) carbonyl]-2 H-tetrazolium hydroxide) assay (Ramage et al. 2001). Kannurin and amphotericin B were tested at a concentration of 0·125–64 μg ml−1.
Conversion of blastoconidia to hyphal forms was carried out by transferring the C. albicans cells to human serum followed by cultivation at 37°C for 36 h. The dimorphic transition was investigated from cultures containing various concentrations of lipopeptide (0, 4 and 8 μg ml−1) and detected by phase-contrast microscopy.
The experimental data were subjected to statistical analysis. Statistical analyses of the differences between mean values obtained for experimental groups were performed using OriginPro 8 computer software (Origin Lab, Northampton, Massachusetts). P values were calculated by t-test. Statistical significance value was set at P < 0·05.
Biochemical reactions and 16S rDNA analysis indicate that it was Bacillus cereus. In addition, the cry gene analysis study successfully discriminated the isolate from B. thuringiensis. The phylogenetic tree (Fig. 1) was inferred by neighbour-joining method. The 16S rDNA sequence obtained in this study has been registered at GenBank database, and the accession number was JX512716.1.
Kannurin was purified from supernatant fluid by calcium chloride precipitation, sequential Sephadex G-25 and C18 RP-HPLC. Total ion chromatogram and UV chromatogram revealed the peaks, numbered from 1 to 3 as shown in Fig. 2a. The mass spectra corresponding to peaks 1, 2 and 3, respectively, show protonated masses at m/z 1008·6, 1022·7 and 1036·7 (Fig. 2b). The mass difference of 14 Da observed between these successive molecules indicates that they are homologous in nature. The cyclic nature of the peptides was confirmed through saponification reaction, where an increase in mass of 18 Da indicates the addition of a water molecule followed by a ring opening (Sabareesh et al. 2007). Figure 3 shows LC-ESI-MS/MS of the saponified molecules corresponding to peaks 1, 2 and 3 and the sequences derived through de novo sequencing approach following Biemann's nomenclature (Biemann 1990). By similarity, it can be envisaged that the peptides belong to surfactin class of molecules that contain a β-hydroxy fatty acid with the backbone cyclization occurred through an ester bond between the hydroxyl group of the fatty acid and carboxylic group of a C-terminal amino acid (Grangemard et al. 1999).
FT-IR spectrum (Fig. 4) shows broad band between 3000 and 3600 cm−1, which is due the –OH and NH stretching. The peaks due to NH bond from the amide (3405 cm−1) and peak due to –OH (3282 cm−1) are clearly seen in the spectrum. The ester bond by the beta hydroxyl acid and the amino acid is indicated by the presence of the peak at 1720 cm−1. The numerous amide bonds are reflected by the intense peak at 1634, which is due to amide I formed by C=O stretching vibration from the amide bond. The value at 1532 arises from the amide II band, which results from the deformation mode of N–H bond combined with C–N stretching mode. The typical C–H stretching frequencies are observed at 2957, 2925 and 2855 cm−1. This can be due to the –CH stretching from the amino acids as well as the alkyl part of the surfactant. The amino acid sequence of Kannurin was assigned to be Leu–Asp–Val–Leu–Leu–Leu–Leu and is linked by hydroxyl fatty acid of the chain length 10–12 carbon atoms to form a cyclic lactone ring structure (Fig. 5).
The lipopeptide Kannurin was relatively heat stable and retained its activity when heated to temperatures as high as 70°C for 15 min (Fig. 6a). The lipopeptide was found active for wider pH range (Fig. 6b), but the maximum activity was retained at neutral pH.
As revealed in table 1, Kannurin inhibited the growth of Candida and Cryptococcal strains at concentrations ranging from 1 to 4 μg ml−1, while MFCs ranged from 2 to 4 μg ml−1. Comparison with amphotericin B showed that Kannurin was slightly less active than this antimycotic, while the lipopeptide had stronger fungicidal activity in comparison with fluconazole against yeast pathogens. The MIC of Kannurin was 1 μg ml−1 for C. albicans and C. parapsilosis. Kannurin was also active against C. tropicalis, C. krusei, Cryptococcus neoformans, C. laurentii and C. albidus; all species were inhibited by ≤2 μg ml−1, but did not inhibit Candida glabrata at up to 4 μg ml−1. Aspergillus species were susceptible to Kannurin at 4–16 μg ml−1 and Fusarium oxysporum, isolated at 4 μg ml−1. The MICs of the medically important Zygomycetes members (Mucor and Rhizopus sp.) were ranged from 2 to 4 μg ml−1. We pursued further investigations into common dematiaceous fungi (Alternaria sp., Curvularia sp. and Cladosporium sp.), and the MICs of Kannurin were in the range of 2 to 4 μg ml−1, while MFCs were slightly higher
|Species||MIC and MFC (μg ml−1)|
|Candida albicans ATCC 10231||1||2||1||1||2|
|Candida albicans (Clinical isolate)||2||2||0·5||1||2|
|Candida glabrata MTCC 3019||4||4||0·5||2||16|
|Candida tropicalis MTCC 1406||2||4||1||1||4|
|Candida parapsilosis MTCC 1965||1||2||0·5||1||4|
|Candida krusei (Clinical isolate)||2||2||0·25||2||32|
|Cryptococcus neoformans MTCC 4406||2||4||0·25||1||4|
|Cryptococcus laurentii MTCC 2898||2||4||0·5||0·5||16|
|Cryptococcus albidus MTCC 4746||2||2||0·25||0·5||8|
|Opportunistic moniliaceous moulds|
|Aspergillus fumigatus (Clinical isolate)||4||8||0·5||2||32|
|Aspergillus flavus ATCC 11496||8||16||2||4||>64|
|Aspergillus niger ATCC 16404||16||>32||0·5||1||>64|
|Fusarium oxysporum (Clinical isolate)||4||16||2||4||32|
|Zygomycetes of medical importance|
|Mucor sp. (Food isolate)||2||4||0·5||1||16|
|Rhizopus sp. (Clinical isolate)||4||8||1||2||32|
|Alternaria sp. (Soil isolate)||2||4||2||4||64|
|Curvularia sp. (Plant isolate)||2||4||1||4||32|
|Cladosporium sp. (Plant isolate)||4||8||2||4||32|
Kannurin showed very less haemolytic activity, when compared with amphotericin B, but almost similar to fluconazole (Fig. 7). These results suggest that Kannurin is much less harmful to the cytoplasmic membranes of higher animals.
Time-kill assays (Fig. 8) demonstrated that incubation of melanin particles to amphotericin B with synthetic and cryptococcal melanin significantly reduced their toxicities for C. neoformans at a concentration of 0·25 and 0·5 μg ml−1. Our findings are consistent with the notion that amphotericin B binds to melanin (van Duin et al. 2002). In contrast, incubation of Kannurin with melanin did not affect its ability to inhibit C. neoformans in vitro.
Amphotericin B and Kannurin exhibited only moderate activity against biofilms of C. neoformans and C. albicans as indicated by the SMIC50 and SMIC80 (Fig. 9).
As seen in the Fig. 10b,c, the lipopeptide destroyed the hyphal forms at 4 and 8 μg ml−1 concentrations.
This study reports a novel antifungal cyclic lipoheptapeptide designated as Kannurin, which contains a β-hydroxy fatty acid in its side chain, isolated from Bacillus cereus strain AK1. It can be concluded from the FT-IR and ESI-MS that Kannurin belongs to surfactin class of lipopeptide molecules that contain a β-hydroxy fatty acid. The lipopeptide composed of amino acids, Leu–Asp–Val–Leu–Leu–Leu–Leu, and is unique in not having Glu or Gln among the repertoire, which is usually found in variants of surfactin family.
Kannurin had broad range of activities against various pathogenic yeasts. It was also fungicidal against opportunistic moniliaceous moulds and those belonging to the members of zygomycetes. Also, Kannurin exhibited good activity against clinically important dematiaceous fungi. Our study indicates that lipopeptide Kannurin may be productive in the field of antifungal therapy when other antimycotics often fail to eradicate the pathogen. Kannurin also has pH stability, thermo stability and low haemolytic activity.
Melanin production is a major virulence factor in many pathogenic fungi and can bind substances, such as heavy metals, antibiotics as well as antimicrobial peptides. Here, we show that pre-incubation of Kannurin with melanin did not affect its fungicidal activity. More cases of mycoses are related to biofilm formation on inert or biological surfaces, and our results suggest that Kannurin is as effective as amphotericin B against in vitro biofilms of tested strains. Dimorphism plays a crucial role in pathogenesis of C. albicans (Lo et al. 1997). Interestingly, in our study, the filamentous form of Candida albicans was destroyed by Kannurin at the concentrations tested.
The last decades of the twentieth century witnessed the discovery of hundreds of antifungal peptides, but only few among them made their entry as useful drugs. In this point of view, Kannurin purified from Bacillus cereus promises to have potential pharmaceutical and biotechnological applications.
We thank School of Biotechnology, Amrita Vishwa Vidyapeetham and School of Chemical Sciences, Mahatma Gandhi University, Kerala, India, for providing ESI-MS and FT-IR facility, respectively.
No conflict of interest declared.