Correspondence: David A. Phoenix, University of Central Lancashire, Preston PR1 2HE, UK. Tel.: +44 (0) 1772 892504; fax: +44 (0) 1772 892936; e-mail: email@example.com
Fungal infections with multiple resistance to conventional antifungals are increasingly becoming a medical problem, and there is an urgent need for new antifungal compounds with novel mechanisms of action. Here, we show that aurein 2.5, a naturally occurring peptide antibiotic, displays activity against the fungal strains: Rhodotorula rubra and Schizosaccharomyces pombe (MICs < 130 μM). The peptide adopted high levels of membrane-interactive α-helical structure (> 65%) in the presence of lipid membranes derived from these organisms and showed strong propensities to penetrate (π ≥ 13 mN m−1) and lyse them (> 70%). Based on these data, we suggest that aurein 2.5 kills yeasts via membranolytic mechanisms and may act as a template for the development of therapeutically useful antifungal agents.
It is well established that the overuse of antibiotics has led to an increase in bacteria with multiple drug resistance (MDR), which has severely limited therapeutic options in treating infections related to bacterial pathogens (Hogberg et al., 2010). However, over the past few decades, MDR has also become an increasing problem in the treatment of infections due to yeasts, which is exacerbated by the fact that only a limited number of safe and effective antifungal drugs are currently available on the pharmaceutical market (Morschhaeuser, 2010). This problem has aroused growing concern by the appearance of yeasts with MDR as causative agents in food contamination (Stratford, 2006; Prasad & Goffeau, 2012) and as opportunistic pathogens in human infections (Pfaller & Diekema, 2004; Morschhaeuser, 2010). For example, Rhodotorula rubra is amongst the most significant and commonly reported food spoilage yeast species (Stratford, 2006), and recently, the organism has emerged as an opportunistic pathogen (Wirth & Goldani, 2012) with the ability to colonize and infect susceptible patients, particularly immunocompromised individuals (Tuon & Costa, 2008; Miceli et al., 2011). Fluconazole and itraconazole were commonly used to treat these infections, but there is increasing evidence that the Rhodotorula is developing resistance to these drugs (Gomez-Lopez et al., 2005; Krzysciak & Macura, 2010; Miceli et al., 2011). Moreover, this problem has been amplified by the propensity of Rhodotorula spp. and other fungi to form biofilms, which are generally highly resistant to conventional antifungal agents (Gomez-Lopez et al., 2005; Martinez & Fries, 2010; Nunes et al., 2013; Walraven & Lee, 2013). In combination, these observations have led to a growing view that food may be an underestimated source of environmental yeast pathogens and raised the disturbing possibility that the consumption of yeast-infected food could play a direct role in causing opportunistic infections (Wirth & Goldani, 2012).
There is an urgent need for new antifungal compounds with novel mechanisms of action (Miceli et al., 2011), and major candidates to serve in this capacity, include photodynamic agents (Calzavara-Pinton et al., 2012; Harris & Pierpoint, 2012) and antimicrobial peptides (AMPs) (Matejuk et al., 2010; Desbois et al., 2011; Wilmes et al., 2011; Mehra et al., 2012). These latter peptides are established as potent broad-range antimicrobial agents (Harris et al., 2009; Phoenix et al., 2013a, b) and of the c. 2500 listed in major databases, approximately 20% have been reported to possess antifungal activity (Wang et al., 2009; Thomas et al., 2010). Mining these databases shows that a particularly rich source of antifungal AMPs with chemotherapeutic potential is the skin gland secretions of anurans (Matejuk et al., 2010; Desbois et al., 2011; Wilmes et al., 2011; Mehra et al., 2012) with major examples including citropins and aureins from Litoria spp, which are Australasian tree frogs (Kamysz et al., 2006; Pukala et al., 2006). In an effort to expand the repertoire of these latter AMPs, this study has investigated the activity of aurein 2.5 when directed against R. rubra and Schizosaccharomyces pombe, a model yeast organism (Yanagida, 2002) that is closely related to Pneumocystis jiroveci, a yeast-like fungus, which commonly causes pneumonia in immunocompromised patients (Wazir & Ansari, 2004; Tyler et al., 2013).
Materials and methods
Aurein 2.5 (GLFDIVKKVVGAFGSL-NH2), which was synthesized by Pepceuticals (Leicestershire, UK) to purity > 95%, was used without further modification. All buffers were prepared using ultrapure water (resistivity 18 MΩ cm). Chloroform and methanol (HPLC grade) were obtained from Fisher (UK), and all other regents were purchased from Sigma-Aldrich (UK).
Antimicrobial activity of aurein 2.5
Cultures of R. rubra and S. pombe, which had been freeze-dried in 20% (v/v) glycerol and stored at −80 °C, were used to inoculate malt extract agar plates, which were incubated at 30 °C for 48 h. Inocula of the test organism were prepared by selecting and suspending single colonies in 10 mL of malt extract broth, which were incubated overnight at 30 °C. The cell suspensions were then centrifuged at 15 000 g for 15 min at 25 °C using a bench-top centrifuge, and the resulting pellet was washed twice in 1/4 strength Ringer's solution before being resuspended in 1.0 mL of the Ringers. These suspensions served as the cell inocula. Twofold dilutions of aurein 2.5 in phosphate-buffered saline (PBS) ranging from 2 mM down to 0.03125 mM were prepared for each test organism, dispensed in 0.5 mL aliquots in Eppendorf tubes. 0.1 mL of cell suspension was added to each peptide concentration, and the samples were then incubated at 30 °C for 48 h. The suspensions were then spotted onto malt agar plates and incubated at 37 °C for 48 h. The minimum inhibitory concentration (MIC) was taken as the lowest concentration of antimicrobial peptide, which induced the complete inhibition of visible growth after 48 h of incubation.
Total lipid extracts from cells of S. pombe and R. rubra
A protocol for fungal lipid extraction, which was based on the techniques of Bligh & Dyer (1959) and Somashekar et al. (2001), was used to obtain lipid extracts from membranes of S. pombe and R. rubra. Essentially, cultures of these organisms grown as described above, and the cell biomass homogenized using a pestle and mortar. The lipid was then extracted with 1.5 mL of chloroform:methanol (2 : 1) and then vortexed. Water (0.5 mL) was added and the whole vortexed for 5 min before being centrifuged at low speed (660 g, 5 min) to produce two phases. The lower organic layer was concentrated by removing the solvent with N2 (g).
Secondary structure determination of aurein 2.5
The secondary structure of aurein 2.5 in the presence of either S. pombe or R. rubra was determined using a J-815 spectropolarimeter (JASCO, UK). Samples were prepared by dissolving 0.1 mg mL−1 of aurein 2.5 in PBS or suspensions of small unilamellar vesicle (SUVs) to maintain a peptide to lipid ratio of 1 : 100. To prepare SUVs, lipid extract from S. pombe or R. rubra in chloroform was dried under N2 (g) and placed under vacuum for 4 h. The lipid film was rehydrated using PBS (pH 7.5) before being vortexed for 5 min. The lipid solution was sonicated for 30 min, and then underwent three cycles of freeze-thawing. Liposomes were then extruded 11 times through a 0.1-μm polycarbonate filter using an Avanti polar lipids mini-extruder apparatus. The samples were analysed at 20 °C using a 10 mm path-length cell over a wavelength ranging from 260 to 180 nm at 100 nm min−1, with a 1-nm bandwidth, and a data pitch of 0.5 nm. For all spectra acquired, four scans were added and averaged, followed by subtraction of the baseline acquired in the absence of peptide. The percentage α-helical content was then estimated using CDSSTR algorithm (protein reference set 3) on the DICHROWEB server (Whitmore & Wallace, 2004, 2008; Whitmore et al., 2010). These experiments were repeated four times and the percentage helicity was averaged.
Interaction of aurein 2.5 with monolayers formed yeast lipid extracts
Surface pressure measurements were performed in a 601M Langmuir Teflon trough (KSV NIMA, Coventry, UK) at 21 ± 1 °C equipped with a Wilhelmy plate. Langmuir monolayers were prepared by spreading lipid extract of S. pombe and R. rubra (Dennison et al., 2009b) in chloroform on the surface of an aqueous subphase 10 mM Tris buffer (pH 7.5). After spreading, 20 min was allowed for solvent evaporation to take place and for the monolayer to reach equilibrium. The lipid monolayer was compressed at a velocity of 10 cm2 min−1, to a surface pressure of 30 mN m−1, which is equivalent to that of the outer leaflet of a cell membrane (Seeling, 1987). After the pressure stabilization of the film was achieved, the barrier position was kept constant. Aurein 2.5 (1 mM) in 10 mM Tris (pH 7.5) was injected into the subphase to give a final peptide concentration of 4 μM underneath preassembled lipid monolayers, and its adsorption into the monolayer was monitored. The surface pressure (π) was determined using the Wilhelmy plate method, and surface pressure increases were recorded as a function of time.
Interaction of aurein 2.5 with lipid vesicles
SUVs formed from lipid extracts of S. pombe or R. rubra were prepared as described above except the lipid film was rehydrated with 1 mL of 5.0 mM HEPES (pH 7.5) containing 70 mM calcein. These vesicles were passed down a Sephadex G75 column, which was rehydrated overnight in 20 mM HEPES, 150 mM NaCl and 1.0 mM EDTA in order to separate the calcein-entrapped vesicles from free calcein by gel filtration eluting with 5 mM HEPES (pH 7.5). The calcein release assay was then performed by combining 2 mL 20 mM HEPES, 150 mM NaCl and 1.0 mM EDTA (pH 7.4), 20 μL calcein vesicles. The fluorescence intensities of calcein was measured using a FP-6500 spectrofluorometer (JASCO, Tokyo, Japan), with an excitation wavelength of 490 nm and an emission wavelength of 520 nm. The experiment was repeated containing aurein 2.5 in PBS (1–1000 μM) which was added to 20 μL calcein vesicles and incubated for 30 min before combining with 2 mL 20 mM HEPES, 150 mM NaCl and 1.0 mM EDTA (pH 7.4). To measure maximum fluorescence, 20 μL of Triton ×100 was used to dissolve the vesicles. The percentage of dye leakage was then calculated.
The antimicrobial activity of aurein 2.5 was tested against S. pombe and R. rubra. After 48-h incubation, aurein 2.5 was found to inhibit S. pombe at an MIC of 62 μM and R. rubra at an MIC of 125 μM, clearly indicating strong antifungal activity. It has been previously suggested that an essential requirement for the biological activity of aurein 2.5 is the adoption of α-helical secondary structure at the membrane interface (Dennison et al., 2012a, b). Accordingly, CD conformational analyses of aurein 2.5 were performed in the presence of lipid vesicle extracts formed from either S. pombe or R. rubra, respectively (Fig. 1). The results of these secondary structure analyses were found to be in alignment with prediction methods such as DSSP/stride (Heinig & Frishman, 2004). It can be seen from Fig. 1 that in each case, the CD spectrum of aurein 2.5 includes two minima at 208 and 222 nm, which is characteristic of the presence of an α-helical conformation. Analysis of these spectra indicated that in the presence of vesicles derived from S. pombe, the peptide possessed 75% α-helicity, whilst in the presence of those from R. rubra, it exhibited 66% α-helicity (Fig. 1). It is known that α-helical aurein 2.5 possesses levels of amphiphilicity that are associated with membrane interaction (Dennison et al., 2012a, b). Accordingly, lipid monolayers, which provide a sensitive model for mimicking biological membranes, were used to assess the interactions of aurein 2.5 with lipid mimics of S. pombe and R. rubra membranes (Fig. 2). It can be seen from Fig. 2 that the peptide rapidly partitioned into these yeast membrane mimics, inducing surface pressure changes of 16.6 mN m−1 in the case of S. pombe and 13.8 mN m−1 in the case of R. rubra. These high levels of interaction are consistent with penetration of the fatty acyl core region of the monolayer and are often associated with AMPs that induce microbial cell death via membranolysis (Phoenix et al., 2013a, b). Aurein 2.5 was tested for membranolytic potential using a calcein release assay, which evaluated the ability of the peptide to induce leakage from lipid vesicles derived from either S. pombe or R. rubra (Fig. 3). It can be seen from Fig. 3 that in the case of both organisms, the lytic ability of the peptide was concentration dependent and in the case of both organisms exceeded 50% at aurein 2.5 concentrations of c. 500 μM.
The increased occurrence of infections due to yeasts and fungi with MDR has led to a search for AMPs with activity against these organisms to serve as alternatives to conventional antifungal drugs (Wilmes et al., 2011; Mehra et al., 2012). To further this search, the present study has investigated the antifungal activity of aurein 2.5, which is an AMP found in the skin gland secretions of the Australian Bell Frogs Litoria aurea and Litoria raniformis (Dennison et al., 2009a, b). Aurein 2.5 was found to be highly effective against S. pombe and R. rubra with MICs (62 and 125 μM, respectively), which are comparable to those reported for the antifungal activity of other aureins and closely related AMPs (Kamysz et al., 2006; Pukala et al., 2006). The levels of aurein 2.5 required here were also comparable to those required for the peptide's antibacterial activity (Dennison et al., 2009a, b, 2012a, b), which would support the broad-range antimicrobial role predicted for aurein 2.5 in the skin of Litoria spp. (Bowie et al., 2012). Indeed, there is evidence to suggest that the panoply of aureins and closely related AMPs produced by these frogs has helped them to survive chytridiomycosis (Woodhams et al., 2006, 2007; Rollins-Smith, 2009), an emerging infectious disease that is induced by the skin-invasive fungus, Batrachochytrium dendrobatidi, and is responsible for a global decline in amphibians (Rollins-Smith et al., 2011).
Far less research has been directed towards elucidating the mechanisms underpinning the antifungal action of these peptides than has been directed towards deciphering their antibacterial activity. As a consequence, these mechanisms generally lack detail although it is becoming clear that they vary widely in nature (Matejuk et al., 2010; Desbois et al., 2011; Wilmes et al., 2011; Mehra et al., 2012). For example, histatins, which possess extended β-turn/polyproline II architecture, appear to utilize nonlytic antifungal mechanisms, involving specific cell surface receptors that promote translocation of these peptides across yeast membranes, in order to induce host cell death via several potential mechanisms, including altered mitochondrial function and the initiation of apoptosis (Matejuk et al., 2010; Phoenix et al., 2013a, b). In contrast, some plant defensins, which possess cysteine-stabilized β-sheet structures, are believed to utilize specific yeast sphingolipids as membrane receptors to promote host cell death via lytic mechanisms involving membranolysis and the induction of apoptosis (Wilmes et al., 2011). In the case of aureins and related AMPs, several lines of indirect evidence have suggested that this action may involve the nonspecific disruption of yeast membranes, which is strongly supported by our data (Rollins-Smith, 2009). In the presence of lipid membranes derived from S. pombe and R. rubra, aurein 2.5 adopted high levels of lipid interactive α-helical structure (> 65%; Fig. 1). Similar conformational changes have been observed for the peptide in its antibacterial action where it undergoes a transition from unordered structure to predominantly α-helical structure in the amphiphilic environment of the bacterial membrane interface (Dennison et al., 2009a, b, 2012a, b). This conformational behaviour is typical of α-helical AMPs and in most cases constitutes an initial step in their antimicrobial mechanisms, which is generally nonreceptor-mediated and proceeds via relatively nonspecific membranolytic actions (Phoenix et al., 2013a, b). Consistent with the use of a similar membranolytic mechanism in its antifungal action, aurein 2.5 showed strong propensities to penetrate (π ≥ 13 mN m−1; Fig. 2) and lyse (> 70%; Fig. 3) lipid membranes derived from S. pombe and R. rubra that were comparable to those reported in corresponding studies on the antibacterial activity of the peptide (Dennison et al., 2009a, b, 2012a, b). Based on parallels between the antifungal and antibacterial activity of aurein 2.5, we speculate that the peptide may kill yeasts via a membranolytic mechanism with similarities to models generally accepted to represent the antimicrobial action of α-helical AMPs (Phoenix et al., 2013a, b). On this basis, it would seem unlikely that the peptide would use the barrel-stave pore model, given that it is sixteen residues in length and therefore too short to span the bilayer (Phoenix et al., 2013a, b). It therefore seems most likely that aurein 2.5 may kill yeasts by lysing the membranes of these organisms using models such as the toroidal pore, tilted peptide or carpet mechanisms (Phoenix et al., 2013a, b).
In conclusion, despite the clear chemotherapeutic potential of many AMP's as antifungal agents, research into this potential remains limited. Indeed, currently, < 100 of these peptides have been reported to fulfil the criteria for medical application, whilst only a very few have been approved for use in clinical practice (Matejuk et al., 2010; Desbois et al., 2011; Wilmes et al., 2011; Mehra et al., 2012). In response, this study has shown aurein 2.5 has strong membranolytic activity against yeasts and suggest that the peptide may act as a template for the development of therapeutically useful agents against the organisms examined here and possibly other fungal pathogens.