|Home | About | Journals | Submit | Contact Us | Français|
Plakortide F acid (PFA), a marine-derived polyketide endoperoxide, exhibits strong inhibitory activity against the opportunistic fungal pathogens Candida albicans, Cryptococcus neoformans, and Aspergillus fumigatus. In the present study, transcriptional profiling coupled with mutant and biochemical analyses were conducted using the model organism Saccharomyces cerevisiae to investigate the mechanism of action of this compound. PFA elicited a transcriptome response indicative of a Ca2+ imbalance, affecting the expression of genes known to be responsive to altered cellular calcium levels. Several additional lines of evidence obtained supported a role for Ca2+ in PFA's activity. First, mutants lacking calcineurin and various Ca2+ transporters, including pumps (Pmr1 and Pmc1) and channels (Cch1 and Mid1), showed increased sensitivity to PFA. In addition, the calcineurin inhibitors FK506 and cyclosporine strongly enhanced PFA activity in wild-type cells. Furthermore, PFA activated the transcription of a lacZ reporter gene driven by the calcineurin-dependent response element. Finally, elemental analysis indicated a significant increase in intracellular calcium levels in PFA-treated cells. Collectively, our results demonstrate that PFA mediates its antifungal activity by perturbing Ca2+ homeostasis, thus representing a potentially novel mechanism distinct from that of currently used antifungal agents.
Plakortide F acid (PFA) belongs to a common class of cyclic peroxide metabolites that are known to be produced by members of the Plakortis genus of marine sponges within the Plakinidae family (30). Sponges within this family are prominent inhabitants of the Caribbean and Indo-Pacific coral reef systems and have been recognized as rich sources of bioactive secondary metabolites, such as cyclic peroxides containing five- or six-membered 1,2-dioxygenated rings (reviewed in reference 17). These polyketide-derived metabolites have been reported to possess a diversity of pharmacological properties, including antitumor, antiparasitic, and antimicrobial activities. For example, cyclic peroxides such as plakortides, plakortolides, plakorins, and plakorstatins that have been isolated from Plakortis spp. exhibit potent activity against a variety of human cancer cell lines (reviewed in reference 12). Several plakortides, including plakortides F, I, K, and L, have been found to show strong antimalarial activity, while cycloperoxide acids isolated from Plakortis spp. have demonstrated potent antifungal activity (reviewed in reference 12). PFA, a hexacyclic endoperoxide (see the structure shown in Table 2), is of particular interest as a potential therapeutic agent, as it exhibits activity against murine leukemia cells (30) and also displays potent inhibitory activity against human fungal pathogens and protozoal parasites (25).
Little is known concerning the mechanism of action of the sponge-derived cyclic peroxide compounds with respect to their antitumor, antiparasitic, and antimicrobial properties. However, prior studies have shown that cyclic peroxide compounds isolated from Plakortis spp. function as activators of sarcoplasmic/endoplasmic reticulum Ca2+-ATPase (SERCA) transporters. For example, the compound plakorin, isolated from an unidentified species of Plakortis collected in Okinawa, Japan, was shown to activate a SERCA transporter in rabbit skeletal muscle (27). Similarly, in a screening for SERCA activators, four different metabolites (3-epiplakortin and plakortides F, G, and H) isolated from the Jamaican sponge Plakortis halichondrioides were found to stimulate cardiac Ca2+ uptake in sarcoplasmic reticulum vesicles from canine ventricle tissue (29). While these results are intriguing, it remains to be determined whether the activation of sarcoplasmic reticulum Ca2+ ATPase reflects the primary mechanism of action of marine cyclic peroxides.
Further characterization of the mechanism of action of cyclic peroxides, such as PFA, is important for facilitating their development as potential pharmaceuticals; moreover, such studies could shed significant light on their roles in marine ecosystems, where they may serve as host defense compounds. In the present study, we have employed transcriptional profiling experiments coupled with genetic and biochemical analyses to gain insight into the mechanism of action of PFA. Using Saccharomyces cerevisiae as a model organism, we show that PFA exposure elicits a transcriptome response suggestive of Ca2+ stress, causing changes in the expression of a significant number of genes known to be responsive to elevated cytosolic Ca2+ levels as well as genes involved in Ca2+ homeostasis. This transcriptional response is distinct from that observed for antifungal drugs currently in use, suggesting that PFA inhibits fungal cell growth through a novel mechanism. S. cerevisiae mutants lacking various Ca2+ transporters or deficient in calcineurin (a key player in Ca2+ signal transduction) showed increased sensitivity to PFA compared to that of the corresponding wild-type strains. Moreover, elemental analysis based on inductively coupled plasma-mass spectrometry (ICP-MS) revealed that intracellular calcium levels were dramatically increased in S. cerevisiae cells following PFA exposure. Collectively, these results suggest that disruption of Ca2+ homeostasis plays a major role in the antifungal activity of the polyketide endoperoxide PFA.
S. cerevisiae strains used in this study are listed in Table Table11 . Synthetic dextrose (SD) medium, containing a 0.67% yeast nitrogen base (without amino acids) and 2% dextrose, was used to grow the wild-type S. cerevisiae S288C strain. The medium was buffered with 0.165 M 3-[N-morpholino]propanesulfonic acid (MOPS), and the pH was adjusted to 7.0. Synthetic complete (SC) medium, consisting of SD medium supplemented with complete supplement mixture (Sunrise Science Products) was used for growing the deletion mutants. For plasmid-containing strains, SC medium lacking uracil (SC-URA medium; no MOPS, no pH adjustment) was used to maintain plasmid selection. FK506, cyclosporine (CsA), and amiodarone (AMD) were obtained from Sigma-Aldrich, and stock solutions were made in dimethyl sulfoxide (DMSO). PFA was isolated from a Jamaican Plakortis sponge as previously described (25).
S. cerevisiae strain S288C was used in the microarray experiments, and all procedures, including IC50 determinations, were performed as previously described (2). For IC50 determinations with small-scale cultures, broth microdilution assays were performed according to the Clinical and Laboratory Standards Institute (CLSI) protocols (5), with the modification that inoculum size was 2 × 106 CFU/ml. This cell density, which is ~200 times greater than that in the standard protocol, was used in order to mimic the microarray culture conditions.
For determination of IC50s in large-scale cultures, an overnight culture was used to inoculate 50 ml of SD medium to an optical density at 600 nm (OD600) of 0.1 (~2 × 106 CFU/ml). Multiple cultures were started in order to test four to five different drug concentrations. The cultures were grown in an environmental shaker (200 rpm, 30°C) until an OD600 of 0.2 was reached. PFA was then added at 2- to 4-fold serial dilutions into each culture. Two rounds of experiments were conducted, with a broad range of PFA concentrations tested in the first round and a narrow range tested in duplicate experiments in the second round. The cultures were grown to late exponential phase (17 h), and the final OD600 was measured using a spectrophotometer (Amersham Ultrospec 2000). The IC50 was determined to be 0.62 μg/ml (see Fig. S1 in the supplemental material).
An overnight culture of S. cerevisiae S288C was used to inoculate 50 ml of SD medium to an OD600 of 0.1. Six independent 50-ml cultures were grown for each experiment, three for treated and three for untreated samples; thus, each treatment consisted of three biological replicates. The cultures were allowed to recover from stationary phase until an OD600 of 0.2 was reached. PFA was added to each culture at a concentration equivalent to the IC50 (0.62 μg/ml). Control cultures were simultaneously treated with 0.25% DMSO. The cultures were allowed to grow until an OD600 of 0.5 was reached (~4 h). Cells were harvested by centrifugation, flash frozen in liquid nitrogen, and stored at −80°C.
RNA isolation, target preparation, hybridization, and data analysis were performed as described previously (2). The Affymetrix GeneChip Yeast Genome 2.0 array was used in all experiments. Image analysis, scaling, and probe set-level data analysis were performed using the default parameters in GeneChip operating software v.1.1 (Affymetrix). Differentially expressed genes were identified using the statistical tools available in BRB-Array Tools v.3.5.0 software developed by Richard Simon and the BRB-Array Tools Development Team (http://linus.nci.nih.gov/BRB-ArrayTools.html), and gene data with a P value of ≤0.001 were considered to be significant. Gene annotations were obtained from the Saccharomyces Genome Database (SGD). The Gene Ontology (GO) Term Mapper tool (http://go.princeton.edu/cgi-bin/GOTermMapper) was used to distribute the genes into GO-based biological process categories, and overrepresented GO terms were identified using binomial distribution probability. Hierarchical cluster analysis was performed using Gene Cluster 3.0 (11) with the Pearson correlation similarity metric (uncentered for both genes and arrays) and the average linkage clustering method, and the visual presentation of the data was done by using Java Tree View (31). For comparison, the AMD profile (38) was obtained from the authors, and the profiles for response to CaCl2 in wild-type, crz1Δ mutant, and FK506-treated cells (37) were obtained from the Yeast Functional Genomics Database (YFGdb; http://yfgdb.princeton.edu/).
To confirm the microarray results, quantitative real-time reverse transcription (RT)-PCR was performed using the same RNA preparations that were used in the microarray experiments. RNA samples were treated with DNase, gene-specific primers were designed, and quantitative real-time PCRs were performed as described previously (1). For the primer sequences of each gene selected for the assays, see Table S1 in the supplemental material. Data were normalized to an internal control (18S rRNA), and the comparative threshold cycle (ΔΔCT) method was used to obtain the relative expression level for each gene.
Broth microdilution assays were performed according to CLSI guidelines (5). An overnight culture started from a single colony was diluted in the appropriate medium after comparison to the 0.5 McFarland standard to afford a final inoculum of ~1 × 104 CFU/ml. After dilution, 190 μl of the inoculum was added to a microplate containing 10 μl of PFA at various concentrations. The microplates were read at 600 nm prior to and after incubation (48 h, 30°C) using a BioTek Powerwave XS microplate reader.
To test whether FK506 or CsA could alter the sensitivity to PFA, broth microdilution assays were performed as described above except that 180 μl of the inoculum was added to a microplate containing 10 μl of PFA and 10 μl of either FK506 or CsA. The concentrations of FK506 and CsA used were 0.6 μg/ml and 3.0 μg/ml, respectively, which did not inhibit yeast cell growth (see Fig. S2 in the supplemental material).
The wild-type strain W303-1A was transformed with a plasmid (pAMS366) containing the lacZ reporter gene driven by the calcineurin-dependent response element (CDRE). Overnight cultures of individual transformants were grown in SC-URA broth and used to inoculate fresh medium at an OD600 of 0.1. After ~2.5 doublings (OD600, 0.6), DMSO (0.25%), PFA (IC50, 0.8 μg/ml), or AMD (IC50, 6.1 μg/ml) was added to the cultures. The cells were allowed to grow for 90 min (OD600, ~0.9) after treatment, and β-galactosidase activities were measured using a yeast β-galactosidase assay kit from Pierce Technologies. Three independent experiments were performed for individual transformants.
An overnight culture of strain S288C was grown in SD broth (MOPS buffered, pH 7.0) and used to inoculate fresh medium (200 ml) at an OD600 of 0.1. After one doubling, either PFA (0.62 μg/ml) or DMSO (0.25%) was added to the cultures. The cells were allowed to grow for 8 h (~2.5 doublings based on DMSO controls) after treatment and were harvested by centrifugation. The cells were washed three times with 1 μM EDTA and then washed three times with distilled water, and cell pellets were flash frozen in liquid nitrogen. Three independent experiments were performed with independently grown cultures.
The cell pellets were treated with 5 ml of concentrated nitric acid (Optima grade, 67 to 70%; Thermo Fisher Scientific) at 120°C for 20 min in a microwave digester (MARS5; CEM Corp., Matthews, NC) until all solid material was dissolved. The solution was cooled and filtered through a 0.45-μm Teflon filter (Phenomenex) and diluted 1:50 with ultrapure water. Dilutions of the multielement calibration standard 2A (Agilent Technologies) were prepared in the same concentration of nitric acid that was present in the final cell extracts. The elemental concentrations of the samples were measured using an Agilent 7500ce system for ICP-MS. The operating parameters used were the following: forward power, 1,500 W; reflected power, 2 W; analyzer pressure, 9.29 × 10−3 Pa; interface/backing (IF/BK) pressure, 3.80 × 102 Pa; and water temperature, 33°C. Autotuning of the instrument was carried out using Agilent ICP-MS tuning solution. The total analysis time was 5 min in multitune mode followed by a 2-min rinse in nitric acid at the same concentration as that in the samples. The instrument was operated in hydrogen mode for Fe and helium mode for other elements to remove spectral interferences. The integrated sample introduction system (ISIS) was used with a pump speed set at 0.1 revolutions per second (rps) during the analysis and washout to minimize the amount of matrix on the interface and optimize sample throughput.
The transcriptional profiling data described in this article are accessible through accession no. GSE26925 at NCBI's Gene Expression Omnibus (GEO; http://www.ncbi.nlm.nih.gov/geo/).
The activity of PFA against several clinically important fungal pathogens was first assessed by liquid broth microdilution assays (Table (Table22 ). For comparison, the widely used antifungal drug amphotericin B was also assayed in parallel. PFA displayed potent antifungal activity, comparable to that of amphotericin B, against the major fungal pathogens Candida albicans, Cryptococcus neoformans, and Aspergillus fumigatus (Table (Table2).2). It exhibited the strongest activity against the most frequently encountered fungal pathogen, C. albicans, with a MIC of 0.08 μg/ml, which was approximately 4-fold lower than that obtained with amphotericin B. In addition, these results indicate that PFA possesses significantly greater antifungal activity against C. albicans than do previously reported cyclic peroxide acids isolated from marine sponges. For example, cyclic peroxide acids isolated from Plakortis angulospiculatus and Plakinastrella onkodes exhibited a MIC of 1.6 μg/ml against C. albicans, while a peroxide acid isolated from Plakortis halichondrioides displayed a MIC of 5 μg/ml against this pathogen (4, 19).
To understand the cellular effects of PFA in fungal cells, we made use of S. cerevisiae, an established model organism that has been utilized extensively for elucidating the molecular targets of antifungal and therapeutic compounds (reviewed in references 28 and 33). A transcriptional profiling study was conducted using S. cerevisiae cells exposed to 0.62 μg/ml PFA, a concentration sufficient to cause an approximately 50% reduction in growth. Genes exhibiting significant differential expression between PFA-treated and solvent-treated cells (P value, ≤0.001; ≥2-fold change) were identified using BRB Array Tools software (see Materials and Methods). In total, 457 PFA-responsive genes were identified from these studies, with 296 genes showing increased expression and 161 genes showing decreased expression (see Table S2 in the supplemental material).
A hierarchical cluster analysis (Fig. (Fig.1A)1A) revealed significant parallels between the global transcriptional response to PFA and the previously described transcriptional responses shown by S. cerevisiae cells when exposed to high levels of CaCl2 (37) and the calcium homeostasis-disrupting anti-arrhythmia drug AMD (20, 38). Among the 296 genes induced by PFA, 129 of these were also induced by 200 mM CaCl2 (at both 15-min and 30-min CaCl2 exposures) and 98 were also induced in response to 15 μM AMD (10-min AMD exposure). Similarly, of the 161 genes repressed by PFA, 103 were also repressed by 200 mM CaCl2 and 124 were repressed following exposure to 15 μM AMD.
Of further significance is the observation that 75 of the PFA-induced genes identified in the present work correspond to genes previously identified by Yoshimoto et al.; the transcriptional control of these genes requires calcineurin, a protein phosphatase that mediates cellular responses to Ca2+ signals via the transcription factor Crz1 (37). Calcineurin activity increases in response to elevated intracellular Ca2+ levels, leading to Crz1 dephosphorylation, which in turn induces the expression of target genes associated with diverse cellular processes, including signal transduction, cell wall organization, transport, and protein degradation (reviewed in reference 9). The study performed by Yoshimoto et al. demonstrated that when S. cerevisiae cells are exposed to FK506, a specific inhibitor of calcineurin, prior to CaCl2 treatment, the expression of these genes is repressed (37). A similar reduction in gene expression is observed when an S. cerevisiae strain lacking the CRZ1 gene (crz1Δ strain) is treated with CaCl2 (37). For comparison, we included the FK506:CaCl2 and crz1Δ:CaCl2 transcriptional profiles in our hierarchical cluster analysis. This analysis identified a cluster of 98 PFA-induced genes that are upregulated by CaCl2 and/or AMD but downregulated in response to CaCl2 in FK506-treated cells and/or in the crz1Δ strain (Fig. (Fig.1A).1A). Importantly, this cluster also includes 67 of the above-mentioned calcineurin-dependent genes present in the set described by Yoshimoto et al. (37). Thus, the observation that a large number of PFA-induced genes are also induced by CaCl2 and/or AMD, combined with the presence of a significant number of calcineurin-dependent genes within this data set, suggests that PFA exerts its effects in yeast cells via perturbations in cellular Ca2+ levels or Ca2+-related signaling events.
To further validate the transcriptional profiling data, the expression of 8 PFA-induced genes representing diverse functional categories was analyzed by quantitative real-time RT-PCR (Fig. (Fig.1B).1B). These genes were selected because they are either induced by both CaCl2 and AMD (37, 38) or are known to be associated with calcineurin (37) and/or Crz1 (http://www.yeastract.com) (34). For all eight genes, there was consistent correlation between transcriptional profiling and real-time RT-PCR data (Fig. (Fig.1B),1B), with similar fold change values observed for the two assays (compare values in Fig. Fig.1B1B with those in Table S2 in the supplemental material).
The transcriptional responses to PFA were further analyzed by organization into GO-based functional categories using the GO Term Mapper tool (http://go.princeton.edu/cgi-bin/GOTermMapper), as shown in Table Table3.3. Several overrepresented functional categories (P value, ≤0.05) for both PFA-upregulated and -downregulated genes are of particular interest with respect to the potential involvement of Ca2+ perturbations with this inhibitor. For example, the enrichment of PFA-induced genes within functional categories, such as “membrane organization,” “cell wall organization,” and “transport,” is potentially significant, given the importance of preserving membrane and cell wall integrity and the central role played by transporters and channels in response to ionic stress (e.g., see reference 37). As mentioned, many of the known calcineurin/Crz1-dependent genes as well as genes induced by CaCl2 and/or AMD (34, 37, 38) were upregulated by PFA, and these are also listed by functional category in Table Table33.
As expected, “response to stress” represented an additional major functional category for genes induced in response to PFA toxicity. Interestingly, many of the genes within this category are also induced by CaCl2 and/or AMD and are also associated with calcineurin/Crz1 (34, 37, 38) (Table (Table3).3). In fact, a comparison of the complete set of 457 PFA-responsive genes with a previously determined set of 868 S. cerevisiae signature stress-responsive genes (18) revealed 84 common members, 61 of which are also affected by CaCl2 treatment (15- and 30-min exposures) (37) (see Table S3 in the supplemental material). Thus, a majority of PFA-induced genes in the “response to stress” category are most likely involved in functions required for tolerating Ca2+ stress. PFA exposure also resulted in the downregulation of a significant number of cell cycle-associated genes, and many of these are also known to be transcriptionally repressed by CaCl2 (37) and/or AMD (38) (Table (Table3).3). The downregulation of cell cycle-related genes in response to PFA is noteworthy, given that Ca2+ stress results in the onset of calcineurin/Crz1-mediated cell cycle arrest (24), and AMD has also been shown to disrupt cell cycle progression in a calcineurin/Crz1-dependent manner (38).
To further evaluate the potential role of Ca2+ in the growth-inhibitory activity of PFA, a series of mutants harboring deletions in various calcium transport pathway components were first analyzed in broth microdilution assays. The major calcium transporters in S. cerevisiae include pumps (Pmr1 and Pmc1), channels (Cch1, Mid1, and Yvc1), and exchangers (Vcx1), which together coordinate the entry and exit of Ca2+ from the cytosol (reviewed in reference 35). Pmr1 is an ATP-driven pump required for sequestering Ca2+ and Mn2+ within the Golgi apparatus. The Ca2+ pump, Pmc1, and the H+/Ca2+ exchanger, Vcx1, both participate in Ca2+ detoxification by sequestering excess cellular Ca2+ into the vacuole, while the Ca2+ channel Yvc1 is involved in Ca2+ release from the vacuole. Finally, Ca2+ influx through the plasma membrane is mediated by the Cch1 and Mid1 Ca2+ channels. Yeast mutants lacking the above-mentioned calcium transporters were therefore analyzed for their sensitivity to PFA.
The pmr1/pmr1 mutant exhibited strong hypersensitivity to PFA compared to the corresponding BY4743 parental strain, decreasing the IC50 for PFA approximately 4-fold (Fig. (Fig.2A).2A). Hypersensitivity to PFA was also observed for the pmc1/pmc1, cch1/cch1, and mid1/mid1 mutants, which exhibited similar IC50 decreases with respect to BY4743, ranging from approximately 1.5- to 2-fold (Fig. (Fig.2A).2A). Altered sensitivity to PFA was not observed in tests performed with either the vcx1/vcx1 or yvc1/yvc1 mutant (data not shown). The PFA hypersensitivity of the Ca2+-accumulating pmr1/pmr1 and pmc1/pmc1 mutants is in agreement with the transcriptional response to PFA, which appears to be indicative of Ca2+ overload-related stress. The hypersensitivity of the cch1/cch1 and mid1/mid1 mutants is perhaps somewhat counterintuitive given that Cch1 and Mid1 are required for Ca2+ entry into the cells. However, it is possible that Ca2+ deprivation in these mutants might result in a compensatory induction in intracellular Ca2+ levels, causing PFA to be more toxic under these conditions (e.g., 26). Taken together, the data on PFA hypersensitivity shown by cells deficient in the Ca2+ transporter Pmr1, Pmc1, Cch1, or Mid1 are consistent with the hypothesis that PFA disrupts Ca2+ homeostasis in yeast cells.
To further confirm the specific role of Ca2+ in PFA activity, we made use of two previously described ion selectivity variants of Pmr1, D53A and Q783A mutants, which transport almost exclusively either Mn2+ or Ca2+, respectively (22, 36). As mentioned above, Pmr1 mediates the transport of both Ca2+ and Mn2+ into the Golgi apparatus; thus, a comparison of Mn2+- or Ca2+-specific Pmr1 transport variants provides a direct means of assessing whether the increased sensitivity of the pmr1/pmr1 mutant to PFA specifically occurred due to a loss of Ca2+ transport activity. A similar approach was previously employed to investigate the mechanism of inhibition of the Ca2+ homeostasis-disrupting drug AMD (20). Plasmids expressing the D53A or Q783A variant or the wild-type PMR1 gene were therefore introduced into haploid cells lacking the PMR1 gene (pmr1Δ cells), and the sensitivity of the resulting strains to PFA was examined (Fig. (Fig.2B).2B). PMR1 gene-lacking cells transformed with the YCplac33 parent vector were also included as controls. Importantly, cells expressing the D53A variant (Ca2+ transport deficient/Mn2+ transport competent), exhibited increased hypersensitivity to PFA relative to cells expressing the wild-type PMR1 gene, and they showed a sensitivity similar to that of the YCplac33-transformed pmr1Δ control (Fig. (Fig.2B).2B). In contrast, cells expressing the Q783A variant (Mn2+ transport deficient/Ca2+ transport competent) exhibited no hypersensitivity, and IC50s obtained were nearly identical to those observed for cells expressing the wild-type PMR1 gene (Fig. (Fig.2B).2B). Thus, these results strongly suggest that Ca2+ perturbations are involved in the inhibitory effects of PFA and corroborate the results obtained using Ca2+ transporter mutants (Fig. (Fig.2A2A).
Given the central role played by calcineurin/Crz1 in Ca2+-dependent signal transduction, we also investigated the effect of loss or inhibition of calcineurin and loss of Crz1 on PFA inhibitory activity. Calcineurin is a heterodimer comprised of a catalytic subunit, encoded by the functionally redundant CNA1 and CNA2 genes, and a regulatory subunit encoded by the CNB1 gene (reviewed in reference 9). Yeast strains deficient in the CNA1, CNA2, CNB1, and CRZ1 genes were therefore assayed for their sensitivity to PFA relative to the BY4743 parental strain. As was observed for the Ca2+ transporter mutants analyzed (Fig. (Fig.2A),2A), the cna1/cna1, cna2/cna2, cnb1/cnb1, and crz1/crz1 mutants all exhibited hypersensitivity to PFA, with the cnb1/cnb1 mutant showing the largest (almost 4-fold) IC50 reduction (Fig. (Fig.3A).3A). Given the redundant nature of the CNA1 and CNA2 genes and the general lack of major phenotypic effects associated with single cna1 or cna2 mutant strains (http://www.yeastgenome.org), we further verified our results by regenerating the yeast deletion strains deficient in these two genes (see the Supplemental Methods in the supplemental material) and analyzing them for PFA sensitivity in broth microdilution assays. The two newly derived deletion strains displayed the same level of hypersensitivity to PFA as the original strains (see Fig. S3 in the supplemental material). Thus, the increased sensitivity to PFA of the cna1/cna1 and cna2/cna2 mutants suggests that differences in the regulation, substrate specificity, or localization of the CNA1 and CNA2 genes could contribute to their response to PFA (10). Taken together, these results demonstrate that deletions in the genes encoding calcineurin subunits (CNA1, CNA2, CNB1) or a major calcineurin substrate (CRZ1) cause a significant alteration in PFA sensitivity.
We also analyzed the effects of the calcineurin-specific inhibitors FK506 and cyclosporine (CsA) on PFA activity, using concentrations that were not inhibitory to cell growth (see Fig. S2 in the supplemental material) for both drugs (FK506 at 0.6 μg/ml and CsA at 3 μg/ml). Remarkably, the inclusion of either FK506 or CsA increased the potency of PFA, decreasing the IC50 of PFA approximately 6-fold and 4-fold, respectively (Fig. (Fig.3B).3B). Unrelated inhibitors, such as ketoconazole (ergosterol biosynthesis inhibitor) or caspofungin (β-glucan synthase inhibitor), did not cause an enhancement in PFA activity (data not shown). Furthermore, the strong enhancement of PFA activity clearly occurred in a nonadditive manner, since both calcineurin inhibitors were supplied at concentrations that did not inhibit cell growth, thus indicating that PFA targets a cellular process closely related to that of FK506 and CsA.
To further explore the relationship between PFA toxicity and calcineurin/Crz1-mediated processes, we made use of a 4×-calcineurin-dependent response element (4X-CDRE) reporter construct (32) to monitor the effect of PFA on calcineurin/Crz1-dependent transcriptional activation. Yeast cells transformed with the 4X-CDRE::lacZ construct were exposed to IC50 levels of PFA, as well as the Ca2+ homeostasis disruptor AMD used as a positive control, and β-galactosidase activities were measured (Fig. (Fig.3C).3C). The activity of β-galactosidase was elevated approximately 4-fold compared to that of DMSO-treated cells for both PFA and AMD, thus demonstrating the ability of PFA to activate Ca2+/calcineurin-dependent transcription.
The results described above support the hypothesis that disruption of Ca2+ homeostasis and/or Ca2+-related signaling represents a key mechanism underlying the biological activity of PFA. We were therefore also interested in determining whether PFA exposure leads to increased intracellular calcium levels in yeast cells. Using the same experimental conditions as employed for transcriptional profiling, S. cerevisiae cells were exposed to an IC50 concentration of PFA (0.62 μg/ml) or solvent (0.25% DMSO) and then cultures were incubated for 2.5 doubling times (based on solvent-treated cells) prior to elemental analysis by ICP-MS. In addition to that for calcium, the levels for seven additional elements (Zn, Mn, Fe, Al, Na, Mg, and K) were readily quantifiable under these conditions. Importantly, an increase in cellular calcium content (approximately 7-fold) was observed for PFA-treated cells relative to that for DMSO-treated controls (Fig. (Fig.4).4). No significant changes were found in the concentrations of the remaining seven elements monitored, except zinc, for which a relatively modest (approximately 2-fold) increase was observed with PFA-treated cells (Fig. (Fig.4).4). The latter observation is not entirely unexpected given the apparently tight coordination between calcium and zinc homeostasis in yeast cells (15). These results provide direct confirmation that exposure to PFA is associated with a disruption in Ca2+ homeostasis in yeast, as suggested by our initial observations (Fig. (Fig.11 to to3).3). A similar increase in intracellular calcium has been reported for yeast cells exposed to the Ca2+ homeostasis-disrupting drug AMD (20).
The transcriptional response to PFA was also compared with the response to representatives of each of the four major classes of clinically used antifungal drugs, including amphotericin B (a polyene), ketoconazole (an azole), caspofungin (an echinocandin), and 5-fluorocytosine (a pyrimidine). Prior studies have shown that these four drugs exhibit gene expression profiles that are indicative of their mechanisms of action, i.e., disruption of membrane function, ergosterol biosynthesis, cell wall integrity, and nucleic acid synthesis, respectively (1). A hierarchical cluster analysis indicated no substantial overlap between the transcriptional profiles of PFA and these four antifungal drugs (Fig. (Fig.5).5). However, as described above, the transcriptional response to PFA most closely resembles the response exhibited by the Ca2+ homeostasis-disrupting drug AMD (Fig. (Fig.5).5). Thus, the mechanism of action of PFA is highly likely distinct from that of antifungal drugs currently in clinical use and potentially involves Ca2+ stress-mediated growth inhibition similar to that exhibited by AMD.
Diverse pharmacological properties have been attributed to marine-derived cyclic peroxides, yet at present a paucity of information exists concerning their mechanism of action. Using a combination of genomic, genetic, and biochemical approaches, we have shown that the marine cyclic peroxide PFA mediates its antifungal activity by interfering with Ca2+ homeostasis. This conclusion is based on several lines of evidence obtained from the present work: (i) the transcriptome response of S. cerevisiae to PFA was indicative of Ca2+ stress, (ii) four different Ca2+ transporter mutants showed increased sensitivity to PFA, (iii) loss or inhibition of calcineurin function significantly enhanced the antifungal activity of PFA, and PFA was shown to activate the calcineurin-dependent transcription of a reporter gene, and (iv) intracellular calcium levels were significantly increased in PFA-treated cells relative to those in controls.
While the present results provide significant insights concerning PFA's mode of action, the specific mechanism underlying the disruption in Ca2+ homeostasis caused by PFA is unclear at this time. One possibility is that PFA interferes with the activity of Ca2+ ATPases, given that yeast mutants lacking Ca2+ ATPases, such as Pmr1 (Golgi-related pump) and Pmc1 (vacuole-related pump) accumulate excess calcium in the cytosol (7). Interestingly, the antimalarial drug artemisinin, which, like PFA, contains an endoperoxide moiety, has been shown to inhibit the SERCA-type Ca2+ transporter of the malaria parasite Plasmodium falciparum (14). Although yeast cells lack homologs of SERCA-type pumps, the Golgi-related pump Pmr1 is considered to be the functional equivalent of SERCA Ca2+ transporters (e.g., see reference 13). Another Ca2+ ATPase transporter, Spf1, which plays a role in endoplasmic reticulum (ER) function and in maintaining calcium homeostasis in yeast cells, has been identified (6). Given the presence of the endoperoxide moiety in the structures of both PFA and artemisinin, it is possible that PFA similarly inhibits the activity of a Ca2+ ATPase transporter in fungal cells.
PFA could also interfere with Ca2+ homeostasis by other mechanisms, such as by causing an influx of Ca2+ through interaction with the plasma membrane. It has been recently shown that the anti-arrhythmia drug AMD, due to its amphipathic nature, associates with the plasma membrane of yeast cells, causing membrane hyperpolarization that leads to the influx of Ca2+ into the cytoplasm (23). It has been proposed that yeast cells may possess hyperpolarization-activated Ca2+ channels, similar to those described for plant root hairs and pollen tubes that are required for cell elongation and growth. Given the presence of the long hydrocarbon chain and the carboxylic group in the structure of PFA (Table (Table2),2), it is possible that PFA might cause Ca2+ influx in fungal cells by hyperpolarization of the plasma membrane.
The work presented here indicates that PFA targets molecular pathways that are distinct from the pathways targeted by clinically used antifungal drugs. There appears to be no substantial overlap between the transcription profile of PFA and those of the four antifungal drugs, amphotericin B, ketoconazole, caspofungin, and 5-fluorocytosine. On the other hand, the transcriptional response to PFA resembles the response exhibited by the Ca2+ homeostasis-inhibiting drug AMD. PFA and AMD appear to have similarities with each other not only in their transcription profiles but also in their ability to increase cellular calcium levels and in their increased activities against Ca2+ transporter mutants (20, 38). Recent studies have shown that the fungicidal activity of AMD is tightly coupled to its ability to induce Ca2+ influx in yeast cells (26); thus, a similar mechanism of Ca2+-mediated cell death could be involved in PFA's antifungal activity.
Given the central role Ca2+ plays in various physiological processes, such as cell wall synthesis, cell cycle progression, and vesicular transport (24, 37), it would be expected that a severe Ca2+ imbalance would have devastating effects on fungal cell viability and growth. For example, yeast mutants deficient in both Pmr1 and Pmc1 Ca2+ ATPases are nonviable, thus demonstrating the inability of yeast to tolerate the resultant overaccumulation of cytosolic Ca2+ (8). Moreover, the observation that mutants of C. albicans and Cryptococcus neoformans lacking either specific Ca2+ transporters or the phosphatase calcineurin are avirulent in animal models of fungal infection demonstrates the requirement for normal Ca2+ homeostasis during fungal pathogenesis (3, 16, 21). Thus, the disruption of Ca2+ homeostasis represents a promising new target pathway for the development of antifungal drug therapies.
Taken together, our findings strongly suggest that calcium homeostasis represents a key cellular target for the antifungal marine endoperoxide PFA. PFA could therefore serve as a useful tool for the further characterization of this cellular process as an antifungal drug target and as a novel pharmacological probe for dissecting the molecular mechanisms underlying calcium homeostasis in fungal and potentially other eukaryotic cell types.
This work was supported in part by grant R01 AI27094 from the Public Health Service, National Institute of Allergy and Infectious Diseases, and USDA Agricultural Research Service Specific Cooperative Agreement 58-6408-2-0009.
We are grateful to Joseph Heitman (Duke University) and Rajini Rao and Kyle Cunningham (Johns Hopkins University) for providing strains and for helpful discussions. We thank Rajini Rao for providing transcriptional profiling data on amiodarone.
Published ahead of print on 7 February 2011.
§Supplemental material for this article may be found at http://aac.asm.org/.