Antifungal Activity of Microbial Secondary Metabolites

Secondary metabolites are well known for their ability to impede other microorganisms. Reanalysis of a screen of natural products using the Caenorhabditis elegans-Candida albicans infection model identified twelve microbial secondary metabolites capable of conferring an increase in survival to infected nematodes. In this screen, the two compound treatments conferring the highest survival rates were members of the epipolythiodioxopiperazine (ETP) family of fungal secondary metabolites, acetylgliotoxin and a derivative of hyalodendrin. The abundance of fungal secondary metabolites indentified in this screen prompted further studies investigating the interaction between opportunistic pathogenic fungi and Aspergillus fumigatus, because of the ability of the fungus to produce a plethora of secondary metabolites, including the well studied ETP gliotoxin. We found that cell-free supernatant of A. fumigatus was able to inhibit the growth of Candida albicans through the production of a secreted product. Comparative studies between a wild-type and an A. fumigatus ΔgliP strain unable to synthesize gliotoxin demonstrate that this secondary metabolite is the major factor responsible for the inhibition. Although toxic to organisms, gliotoxin conferred an increase in survival to C. albicans-infected C. elegans in a dose dependent manner. As A. fumigatus produces gliotoxin in vivo, we propose that in addition to being a virulence factor, gliotoxin may also provide an advantage to A. fumigatus when infecting a host that harbors other opportunistic fungi.


Introduction
Microbial secondary metabolites have provided numerous pharmaceutical agents ranging from antibiotics to immunosuppressive compounds. Synthesis of these low molecular weight compounds is not required for normal growth of the microbe, however these compounds may provide several benefits to the organism. Fungi have the ability to produce a plethora of secondary metabolites, typically dependent on the stage of development of the fungus and environmental factors ranging from nutrient concentrations to light and temperature [1,2]. Fungi belonging to the genus Aspergillus are especially capable of producing a diverse array of these compounds [3,4]. The filamentous fungus Aspergillus fumigatus secretes more than 226 secondary metabolites including commonly studied polyketides, such as cyclic peptides, alkaloids, and sesquiterpenoids [4]. Members of another class of secondary metabolites produced by A. fumigatus, termed the epipolythiodioxopiperazines (ETPs), are characterized by an internal disulphide bridge across a diketopiperazine ring, where the first and best characterized member being gliotoxin [5].
A. fumigatus spores are ubiquitous in the environment and are commonly inhaled. Invasive aspergillosis usually only effects immune-compromised patients (those with leukemia, transplantation) or patients with other medical conditions such as cystic fibrosis, chronic obstructive pulmonary disease, or severe asthma, as the primary route to an established infection is through the lungs [6]. Among different Aspergillus species, only those associated with aspergillosis, such as A. fumigatus, A. terreus, A. flavus, and A. niger, produce gliotoxin [7,8]. Conversely, A. nidulans, a saprobe not normally associated with invasive aspergillosis, does not have the secondary metabolite gene cluster necessary to produce gliotoxin or any other ETP [9]. The role of gliotoxin in mammalian virulence is not fully known as conflicting results exist (recently reviewed in [10]). In A. fumigatus, the gliotoxin secondary metabolite gene cluster is composed of 12 genes approximately 28 kb in length (the gli cluster) [11]. Dioxopiperazine synthase (GliP) is required in the first step for the biosynthesis of gliotoxin generating the characteristic diketopiperizine ring [5,12].
Factors which enable A. fumigatus to colonize and remain established within the host by competing for limited available nutritional resources are currently unknown; however gliotoxin has potent antifungal activity against Candida albicans, Cryptococcus neoformans, and other fungi [13]. This is interesting because pathogenic fungi, such as C. albicans and C. neoformans, primarily infect or colonize hospitalized patients and particularly the same patient population as A. fumigatus, providing an environment conducive of pathogen-pathogen interactions between these fungi, in particular within the pulmonary system. For example, concurrent co-infection/colonization of Aspergillus spp. and Candida spp. can occur in patients [14]. Moreover, Candida spp. can colonize the respiratory tract of hospitalized patients, and the ability of a fungus such as A. fumigatus to compete against a previously established Candida spp. colonization may be necessary for the second pathogen to develop an infection.
Here we reanalyzed the results from a recently published in vivo Candida-infected nematode survival assay to identify secondary metabolites capable of prolonging nematode survival. We found that two members of the ETP class of secondary metabolites we able to significantly increase nematode survival after infection with C. albicans. As pathogenic fungi are capable of producing these compounds within a host, the inhibitory action of gliotoxin against C. albicans was further studied. This research investigates the potential antagonistic activities mediated by secondary metabolites that may be occurring among fungi within a host.

Results and Discussion
Secondary metabolites promoting Candida-infected Caenorhabditis elegans survival The C. albicans-C. elegans antifungal discovery assay allows simultaneous assessment of the ability of a compound to promote survival of infected nematodes and indicate if there is any associated potential toxicity [15]. Previously, a high-throughput screen of 2,560 natural products from the Analyiticon Discovery compound collection (www.ac-discovery.com) was conducted that identified several plant produced saponins that confer an increase in Candida-infected nematode survival [16]. In addition to these saponins, reanalysis of this screen also identified twelve microbial secondary metabolites that were able to prolong nematode survival and may have antifungal activity ( Figure 1; Table 1). These compounds were produced by bacteria and fungi, and several natural products that were closely related to known antifungal compounds were identified in this screen. Compounds conferring C. albicans-infected nematode survival rates greater then 40% were chosen for further discussion.
Interestingly, two members (A3 and A22) of the ETP family of secondary metabolites provided the highest worm survival of the natural products screened ( Figure 1; Table 1) and A22 (80%) and A3 (73%) promoted greater than 65% nematode survival ( Figure 1). Of note is that the dose response for nematode survival for both of these ETP compounds was comparable to amphotericin B (Figure 2). Toxicity for acetylgliotoxin was observed in the C. elegans assay at higher concentrations (.16 mg/ml; Figure 2), no toxicity was seen at the highest concentration tested for A22 (31 mg/ml; Figure 2). A number of ETP secondary metabolites are synthesized by fungi [5]. The most common form of this class of compounds contains a disulphide bridge, but sometimes versions containing one, three, or four sulfur atoms are also produced [5]. The disulphide bridge containing form of A22 has been previously isolated from Hyalodendron sp. and identified as hyalodendrin [17], one of the few ETP compounds produced by a basidiomycete [5].
Among the other secondary metabolites, compound A15 (tunicamycin V) is a member of the tunicamycin family of antimicrobial agents with a modified fatty acid side-chain ( Figure 1) [18]. Tunicamycin is a mixture of homologous nucleoside compounds and inhibits N-linked glycosylation [19]. Tunicamycin is highly toxic if swallowed, and targets the liver and nerves. However, the (E)-13-methyltetradec-2-enoic acid substituted analog tunicamycin V gave up to 60% protection to the worms and showed no signs of toxicity in the screen. Several tunicamycin homologues have previously been reported to inhibit C. albicans, and their degree of toxicity varied depending on the fatty-acid side chain [20].
Importantly, distinct structure activity relationships were discernible between several closely related analogs in the antifungal screen. For example, two members of the enniatin family of natural products, enniatin A (A5) and enniatin B1 (A12), produced by Fusarium spp. were active in the assay [21]. This class is characterized by the alternating arrangement of ester and amide linkages comprising an 18 membered macrocycle. Enniatin A and enniatin B1 differ by only two methyl substitutions and demonstrated protective activity with the isovaleroyl substituted enniatin A being twice as effective, 47% vs 20% respectively ( Figure 1). The protective activity of the enniatins is reflected in their MIC, as enniatin A inhibits C. albicans growth at half the concentration as enniatin B1 [22]. These compounds did not display toxicity to the worms in subsequent dose-response experiments, however the enniatins provided a low degree of protection to the worms.
Another notable secondary metabolite, desferrioxamine X belongs to the well-studied class of hydroxamate siderophores [23,24]. This class of cyclic hexadentate siderpohores are specific chelators of iron(III) and are produced by a variety of bacteria and fungi under iron deficient conditions. These hydroxamate based macrocycles sequester and solubilize iron(III) which is then actively transported into the organism. While siderophores have been utilized in humans for iron and aluminum overload therapy and some antibiotic applications, use of desferrioxamine in C. albicans treatment has not been extensively studied. However, desferrioxamine increases severity of mycoses caused by some fungi, in particular pathogenic members of the order Mucorales, as the fungi are able to uptake and utilize the iron chelated by the siderophore, although there appears to be no significant difference with C. albicans [25,26]. While the overall protective effect was moderate, 40%, desferrioxamine X showed no toxicity to the worms at higher concentrations.
Finally, the secondary metabolite ascochlorin (A1), also referred to as ilicicolin D, is an prenyl-phenol compound that was originally identified in extracts of the fungus Ascochyta viciae [27]. Ascochlorin inhibits mitochondrial electron transport via binding to the Q i and Q o sites of the cytochrome bc 1 complex [28]. Interestingly, another secondary metabolite identified in the screen, ilicicolin H (A4) produced by the fungus Cylindrocladium iliciola, also acts by inhibiting the cytochrome bc 1 complex, however this molecule binds at the Q n site [29]. Both ascochlorin and ilicicolin H conferred a similar C. albicans-infected nematode survival rate (40% and 33%, respectively; Figure 1 and Table 1).

Secondary metabolites from A. fumigatus inhibit other opportunistic fungi
The secondary metabolites produced by A. fumigatus were chosen for additional investigation. This fungus was chosen for further studies with C. albicans and C. neoformans because of the numerous secondary metabolites known to be synthesized and have been characterized, including gliotoxin, the deacetylated version of A3. Although rare, concurrent fungal infection/ colonization between A. fumigatus and C. albicans or C. neoformans have been documented [14]. Isolation of Candida spp. from respiratory specimens is generally not indicative of colonization, but rather is the result of contamination of the bronchoscope from the gastrointestinal tract during the examination procedure. However, it is notable that in a study of postmortem examinations, six (4.8%) revealed concurrent Aspergillus spp. and Candida spp. infection and one with Aspergillus spp. and C. neoformans (,1%) [14].
When A. fumigatus was co-inoculated with C. albicans or C. neoformans on plates containing Spider medium at 37uC, yeast colonies proximal to A. fumigatus were unable to grow (data not shown), suggesting that A. fumigatus produces a secreted toxic agent to both C. albicans and C. neoformans at the stationary phase. In order to identify if any secondary metabolites of A. fumigatus were responsible for the growth inhibition to C. albicans and C.
neoformans, A. fumigatus supernatant (AFS) was collected. As little as 2 mg of AFS was able to form a zone of inhibition around C. albicans strains DAY185, 95-120, or 98-145 on plates containing Spider medium and grown at 37uC ( Figure 3); a zone of inhibition was also observed with C. neoformans KN99a on YPD plates grown at 30uC (Figure 3). AFS was also able to produce a slight zone of inhibition around the C. albicans strains at 30uC on YPD plates (data not shown).
To further assess the nature of the secreted product(s) in the supernatant that are responsible for the inhibition of C. albicans and C. neoformans, the supernatant was heated at 60uC for 2 hours to inactivate potential enzymatic activity. The inhibitory activity of the heat inactivated supernatant (HI-AFS) was assayed on C. albicans DAY185 and was able to produce a zone of inhibition on plates containing Spider medium and grown at 37uC ( Figure S1), similar to the results observed with AFS. As a negative control the fungal inhibitory potential of the C. neoformans supernatant was assessed on C. albicans growth. The C. neoformans supernatant was unable to produce a zone of inhibition with C. albicans DAY185 when grown on either YPD at 30uC or Spider medium at 37uC ( Figure S2). Taken together, these data suggest that AFS contains heat stable product(s), possibly secondary metabolites, which were secreted from the fungus capable of inhibiting the growth of C. albicans and C. neoformans.
Gliotoxin is the major secondary metabolite of A. fumigatus that is toxic to C. albicans and C. neoformans The A. fumigatus DgliP3 mutant strain that does not produce gliotoxin [12] was used to investigate if gliotoxin produced by A. fumigatus is responsible for the inhibitory activity of the supernatant to C. albicans and C. neoformans. Unlike the supernatant from the wild type AF293 strain, the supernatant of the DgliP isolate failed to produce a zone of inhibition for either C. albicans strains DAY185, 95-120, 98-145, or C. neoformans KN99a, demonstrating that gliotoxin was responsible for the zone of inhibition observed with all the C. albicans and C. neoformans strains. Further studies on the inhibitory activity of  Table 1. Compounds identified in a screen of natural products containing secondary metabolites.

Compounds
Percent nematode survival Mode of action Reference(s) Binding to free protein thiols and generation of ROS [5] Acetylgliotoxin (A3) 73% Binding to free protein thiols and generation of ROS [5] Tunicamycin V (A15) 60% Inhibits the transfer of N-acetylglucosamine to a lipid intermediate [18,20] Napthoquinone (A23) 53% Complex naphthoquinine; diverse biological properties Enniatin A (A5) 47% Disrupt cellular physiological cation concentrations [22,47] Desferrioxamine X (A10) 40% Hydroxamate siderophore -specific chelators of iron(III) [23,24] Ascochlorin/Ilicicolin D (A1) 40% Inhibitor of the cytochrome bc1 complex in mitochondria [27,28] Ilicicolin H (A4) 33% Inhibitor of the cytochrome bc1 complex in mitochondria [29,48] (S)-Ochromycinone (A13) 33% Multiple biological properties, but exact MOA is unknown [49] Tricothecin (A14) 33% Inhibits translation by binding to ribosomes [50,51] Oligomycin A (A9) 27% Inhibitor of F 0 domain of the H + -ATP synthase [52] Enniatin B1 (A12) 20% Disrupt cellular physiological cation concentrations [22,47] doi:10.1371/journal.pone.0025321.t001 commercially available gliotoxin supports that this compound was responsible for the zone of inhibition observed with C. albicans and C. neoformans. The area of the zone of inhibition produced by pure gliotoxin was in a dose dependent manner as observed with AFS ( Figure 3). Pure gliotoxin produced a clear zone of inhibition of C. albicans on Spider medium grown at 37uC as observed previously with AFS, but the small zone of inhibition was not apparent with C. albicans on YPD media grown at 30uC suggesting gliotoxin may not be responsible for the slight observed inhibition in this condition (data not shown). These studies demonstrate that gliotoxin is the major component involved in the inhibitory activity of the A. fumigatus supernatant when grown in Spider medium at 37uC.

The effects of gliotoxin in C. albicans and C. neoformans
Pure gliotoxin and AFS were used to find the minimum inhibitory concentration (MIC) in vitro of C. albicans strains DAY185, 95-120, 98-145, and C. neoformans strain KN99a. The MIC of gliotoxin was 2.0 mg/ml for C. albicans and 4.0 mg/ml for C. neoformans (Table 2), whereas the MIC of AFS was 3.2 mg/ml for all the strains tested (Table 2). In murine studies, gliotoxin was able to be accumulated in lung tissue to a mean concentration of ,4 mg/g and was also detected in the sera of the animals, although at a significantly reduced concentration, 36 ng/mL [30]. Gliotoxin was capable of being detected in several patient serum samples where the concentration of one sample was 785 ng/mL (range 166-785 ng/mL) [30], suggesting gliotoxin can accumulate in patients with invasive aspergillosis at a concentration capable of inhibiting other fungal pathogens.
Gliotoxin is produced by a number of fungi [5], although whether gliotoxin is produced by C. albicans is not known, as conflicting studies exist suggesting gliotoxin is produced by some strains of C. albicans [31], however subsequent studies have shown that the fungus does not produce the ETP [32]. In support of the lack of gliotoxin production in C. albicans, the genome of the fungus does not contain a secondary metabolite gene cluster predicted to synthesize an ETP [5], and this study demonstrates that C. albicans is highly susceptible to gliotoxin (Figure 3; Table 2).
The toxicity of gliotoxin is possibly due to several mechanisms. Gliotoxin has the potential to induce production of reactive oxygen species (ROS) by a intracellular redox cycle, where the reduced compound oxidizes to reform the disulfide bridge, producing hydrogen peroxide and superoxide in the process [5]. ETP compounds also have the potential to react with numerous cellular proteins which have exposed cystine residues, and therefore have no ''specific'' mode of action [5]. In addition, it has been demonstrated that in mammalian cells the reduced form of gliotoxin is unable to cross the plasma membrane, resulting in accumulation inside the cell, and therefore the intracellular concentration of gliotoxin is several orders of magnitude higher and predominantly in the reduced form [33]. The hydrogen peroxide produced by the redox cycle of gliotoxin has been implicated in causing single-and double-stranded DNA breaks [34], and therefore some of the antifungal activity of gliotoxin may potentially also be derived by the damage of fungal DNA.
The effects of AFS and gliotoxin on C. albicans DAY185 were further evaluated using the BacLight live-dead staining kit (Molecular probes). Using this system, live fungi with intact membranes fluoresce green, while dead fungi with damaged membranes fluoresce red. The live-dead staining reflects the previous observation where DMSO treated cells were alive (green), and AFS (3.2 mg/ml) and gliotoxin (2.0 mg/ml) treated cells were dead (red) in Spider medium at 37uC. At 30uC, C. albicans yeast cells treated with DMSO and AFS (3.2 mg/ml) were alive (green, Figure 3). In contrast, gliotoxin treated cells did not grow and the nuclear contents were yellowish with the surrounding cytosol green (Figure 4, gliotoxin treated cells in YPD at 30uC).
The inhibitory activity of gliotoxin (and AFS) treated C. albicans cells was consistently higher against cells grown in Spider medium compared to cells grown in YPD medium (Figures 3, 4 and S1). Although the reason(s) for the observed differences are unknown, we speculate that gliotoxin may have increased activity against cells with a filamentous morphology which is favored by growing C. albicans in Spider medium. Additionally, physiological properties such as permeability or increased ROS generation could potentially be contributing to the antifungal activity of gliotoxin when grown under this condition.
A. fumigatus is resistant to the effects of gliotoxin, as one of the genes in the gli cluster, gliT, encodes a reductase that confers a high level of self-protection to the compound [35]. Additionally, another member of the gene cluster, gliA, encodes a major facilitator transporter that has been shown to be involved in efflux of the ETP [36]. Therefore, at least two mechanisms exist in A. fumigatus that confer tolerance to gliotoxin that are absent in most other fungi, including C. albicans and C. neoformans. In general there are fewer efflux transporters in C. albicans and C. neoformans when compared to A. fumigatus [37], and therefore are less likely to have the capability to transport gliotoxin out of the cell. Of note is that two fluconazole resistant isolates of C. albicans were as susceptible to gliotoxin as the wild-type isolate suggesting the efflux pumps conferring resistance to fluconazole do not confer resistance to gliotoxin ( Figure 3; Table 2). The lack of mechanisms able to  confer resistance to gliotoxin may account for the high level of inhibitory activity of the compound.

C. elegans survival as a marker for the evaluation of the antifungal activity of gliotoxin
The C. elegans-C. albicans model system was used to evaluate the efficacy of AFS and gliotoxin using nematode survival as a method to gauge the antifungal activity of the compound. Although other invertebrate host models exist [38,39] C. elegans is an ideal heterologous host to evaluate the effects of gliotoxin, as the nematode lacks a NF-kB homolog while other immune response pathways remain intact [40,41], and therefore some of the immunosuppressive activity derived by inactivation of this transcription factor does not interfere with the evaluation of the antifungal activity of gliotoxin. AFS and gliotoxin were able to inhibit the growth of C. albicans, prolonging the survival of C. elegans ( Figure 5). The effective concentration which resulted in 50% survival of nematodes (EC 50 ) of pure gliotoxin and AFS were determined. The EC 50 of AFS and gliotoxin were 1.0 mg/ml and 2.0 mg/ml, respectively (Table 2 and Figure 5). The highest concentration of AFS used in the C. elegans assay was 12.8 mg/ml ( Table 2) a concentration that appeared to be non-toxic to the nematode. As a control, the supernatant of C. neoformans was unable to inhibit the growth of C. albicans and prolong the survival of Candida-infected C. elegans ( Figure 5).

Concluding remarks
Relatively few studies exist describing the interactions between fungi, and usually these studies pertain to use of a fungus to control a phytopathogenic fungus in agriculture. As the opportunistic fungal pathogens C. albicans and A. fumigatus may exist within the same infected individual or share a similar environmental niche, it is imperative to understand the fungal-fungal interactions that maybe taking place. Additionally, the inhibitory activity of gliotoxin against pathogenic bacteria suggests this compound may also have a role in the interaction between medicallyrelevant bacteria and fungi capable of synthesizing the compound, as these interactions may occur more frequently [42,43]. These interactions create a competition between these microbes in order to obtain limited resources, and secondary metabolites provide a competitive advantage to the microbe harboring them. While gliotoxin production is related to virulence within a susceptible host, this study indicates it may also facilitate A. fumigatus colonization and maintenance within an individual by inhibition of other fungal pathogens that may exist within the host.

Materials and Methods
Media and strains C. albicans, C. neoformans, and A. fumigatus strains used in the assays are listed in Table S1. Yeast strains were grown on yeast extract-peptone-dextrose (YPD) (Difco) plates or in YPD liquid media containing kanamycin (45 mg/ml), ampicillin (100 mg/ml), and streptomycin (100 mg/ml) at 30uC. The inoculated liquid media were grown over night on a rotary shaker at 225 rpm. The cells were centrifuged, washed three times with phosphate buffered saline (PBS) and re-suspended in PBS at the required concentrations for experiments. The zone of inhibition assays were conducted on either YPD plates at 30uC or on plates containing Spider medium [44] at 37uC. Preparation of culture supernatants A. fumigatus and C. neoformans culture supernatants were obtained by growing single isolated colonies in glycerolarginine-yeast extract media [45] for 6-7 days at 30uC. The liquid supernatant was then centrifuged, filter sterilized, and lyophilized to obtain the crude concentrated fungal supernatants. Dried fungal supernatants were then weighed and dissolved in an appropriate amount of DMSO for use in subsequent assays. Stock solutions of the supernatants were prepared at a concentration of 200 mg/ml.

Minimal inhibitory concentration assay
The determination of the lowest concentration of the compounds with antifungal activity was accomplished using two-fold serial dilutions of the test compounds in RPMI 1640 media at 35uC for 24 hours [46]. The wells were assessed by a 96-well plate reader (V max kinetic microplate reader, Molecular Devices, Sunnyvale, CA) to determine the concentration exhibiting in vitro inhibition of C. albicans growth.

Live-Dead staining
The viability of C. albicans in the presence of fungal supernatants and gliotoxin was assessed by using the BacLight LIVE/DEAD staining system according to the manufacturer's protocol (Molecular Probes, Carlsbad, CA). The cells which retained the green fluorescence color were live whereas the red fluorescent cells were considered dead.

EC 50 assays
To measure the EC 50 of gliotoxin and fungal supernatants on worms, the C. albicans-C. elegans co-inoculation assay was performed using the C. elegans glp-4;sek-1 double mutant in all assays [15]. Nematodes were maintained on nematode growth medium with Escherichia coli strain OP50 as the food source. The screen medium was 20% brain heart infusion medium (BHI, Difco) in M9 buffer containing antibiotics kanamycin (90 mg/ml), ampicillin (200 mg/ ml), and streptomycin (200 mg/ml). M9 buffer was used to wash the worms as needed and for the diluting the screen media.
At the end of the incubation period, the entire wells were imaged, visually analyzed for in vitro fungal growth, followed by visual scoring of live and dead worms based on worm shape, as live worms appear sinusoidal and dead worms are rod shaped. The worms were also tested by using a platinum pick to score live or dead. For the determination of EC 50 , the test compounds were serially diluted two-fold. Half the maximum effective concentration that conferred 50% survival of the worms was determined as the EC 50 . Figure S1 Inhibition of C. albicans by heat inactivated A.fumigatus supernatant. C. albicans strain DAY185 was grown on YPD media at 30uC and Spider medium at 37uC overnight in the presence of discs containing heat inactivated supernatent from isolate AF293 (HI-AFS) at the indicated concentrations. (TIF) Figure S2 Absence of inhibition of C. albicans growth by C. neoformans supernatent. C. albicans strain DAY185 was grown on YPD media at 30uC and Spider medium at 37uC overnight in the presence of discs containing the supernatant of C. neoformans strain KN99a (CNS) and the heat inactivated supernatant of C. neoformans (HI-CNS) at the indicated concentrations. (TIF)