Candida albicans Flu1-Mediated Efflux of Salivary Histatin 5 Reduces Its Cytosolic Concentration and Fungicidal Activity

Abstract

Histatin 5 (Hst 5) is a salivary human antimicrobial peptide that is toxic to the opportunistic yeast Candida albicans. Fungicidal activity of Hst 5 requires intracellular translocation and accumulation to a threshold concentration for it to disrupt cellular processes. Previously, we observed that total cytosolic levels of Hst 5 were gradually reduced from intact cells, suggesting that C. albicans possesses a transport mechanism for efflux of Hst 5. Since we identified C. albicans polyamine transporters responsible for Hst 5 uptake, we hypothesized that one or more polyamine efflux transporters may be involved in the efflux of Hst 5. C. albicans FLU1 and TPO2 were found to be the closest homologs of Saccharomyces cerevisiae TPO1, which encodes a major spermidine efflux transporter, indicating that the products of these two genes may be involved in efflux of Hst 5. We found that flu1Δ/Δ cells, but not tpo2Δ/Δ cells, had significant reductions in their rates of Hst 5 efflux and had significantly higher cytoplasmic Hst 5 and Hst 5 susceptibilities than did the wild type. We also found that flu1Δ/Δ cells had reduced biofilm formation compared to wild-type cells in the presence of Hst 5. Transcriptional levels of FLU1 were not altered over the course of treatment with Hst 5; therefore, Hst 5 is not likely to induce FLU1 gene overexpression as a potential mechanism of resistance. Thus, Flu1, but not Tpo2, mediates efflux of Hst 5 and is responsible for reduction of its toxicity in C. albicans.

INTRODUCTION

Candida albicans is a human fungal pathogen that causes serious infections in immunocompromised populations (1–3). In hospitalized patients, this organism can disseminate hematogenously and infect virtually all organs (4–6). The incidence and mortality rates associated with this infectious disease have remained unchanged for more than a decade despite major advances in the field of antifungal therapy (6, 7).

Azoles are one of the most widely used groups of antifungal drugs used to treat candidiasis patients and represent a class of five-membered, nitrogen-containing, heterocyclic compounds (8, 9). The azole drugs enter C. albicans cells by facilitated diffusion (10) and inhibit the biosynthesis of ergosterol, which is a major component of C. albicans membranes (11–13). However, the widespread use of azoles coupled with the fungistatic nature of these drugs has led to the emergence of resistance in clinical isolates (8, 14, 15). Mechanisms of azole resistance include alteration (either by mutation or by overexpression) of the drug target enzyme lanosterol demethylase (encoded by ERG11) and/or constitutive overexpression of multidrug transporters (16).

There are two main classes of multidrug transporters that are found to be upregulated in drug-resistant clinical isolates and experimentally evolved populations of C. albicans; these include transporters containing ATP-binding cassettes (ABC) and transporters of the major facilitator superfamily (MFS) (17–20). The ABC transporter superfamily, including Cdr1 and Cdr2, consists of membrane proteins that have two membrane-spanning domains and two nucleotide-binding domains that utilize ATP to drive substrates across the membrane (21, 22). PDR5 encodes an ABC protein, which is involved in multidrug resistance in Saccharomyces cerevisiae. Pdr5 has a nucleotide-binding domain followed by a transmembrane domain, which is repeated in the second half of the protein. CDR1 and CDR2 were both identified based on their functional complementation of an S. cerevisiae Pdr5 mutant (19, 23). Other studies showed that a C. albicans CDR1 homozygous deletion mutant was hypersensitive to azoles, whereas a CDR2 homozygous deletion mutant did not have altered sensitivity (24). However, the combined deletion of both CDR1 and CDR2 resulted in an increased hypersensitivity compared to that resulting from the deletion of CDR1 alone, suggesting that the Cdr2 protein also plays a role in azole resistance (24–26). MFS drug pumps, including Mdr1 and Flu1, have no nucleotide-binding domain but instead use the proton motive force of the membrane as an energy source (21, 22). The Mdr1 plasma membrane protein functions as a drug/H⁺ antiporter in C. albicans, which exchanges H⁺ with antifungal compounds (27–29). Both MDR1 and FLU1 were overexpressed specifically in fluconazole-resistant C. albicans isolates (23, 30–32). Like the CDR genes, FLU1 was also identified during genomic library screening for complementation of fluconazole hypersusceptibility in the S. cerevisiae Pdr5 mutant strain (33). Flu1 can mediate fluconazole resistance when expressed in S. cerevisiae; however, disruption of C. albicans FLU1 had only a small effect on its fluconazole susceptibility (33). Disruption of FLU1 in a background of C. albicans mutants with deletions in several multidrug efflux transporter genes, including CDR1, CDR2, and MDR1, resulted in enhanced susceptibility to several azole derivatives, indicating that FLU1 also contributes to azole resistance (30). Since a BLAST search of available databases with C. albicans FLU1 revealed high similarity with C. albicans MDR1 (33, 34), there has been a focus on its function as a fluconazole transporter. However, we found that the FLU1 gene product also has high homology with S. cerevisiae Tpo1, a major plasma membrane polyamine efflux transporter (35), suggesting other functions for this protein.

Histatin 5 (Hst 5) is a histidine-rich, antifungal cationic protein (24 amino acids) secreted by the major salivary glands only in humans and higher primates (36, 37). The distinct mode of action displayed by Hst 5 compared with that of azoles or polyenes may be beneficial when resistance to conventional antifungals has occurred. Unlike other cationic peptides, the fungicidal mechanism of Hst 5 is not a result of cytolysis or membrane disruption. Instead, Hst 5 induces selective leakage of intracellular ions and ATP from yeast cells, resulting in gradual cell death similar to osmotically induced cell death (38, 39). These cytotoxic effects are initiated once Hst 5 translocates into the intracellular compartment and accumulates to a critical concentration (40). However, we also observed that total cytosolic levels of Hst 5 in C. albicans cells were gradually reduced while cells were still intact (as shown by the inability of propidium iodide to enter cells) and metabolically active (40), suggesting that C. albicans possesses a transport mechanism for efflux of Hst 5. Since we found that Hst 5 utilizes the C. albicans polyamine influx transporters Dur3 and Dur31 (both plasma membrane permeases) for its uptake (41), we hypothesize that one or more polyamine efflux transporters may be involved in efflux of Hst 5.

Although little is known about efflux of polyamines in C. albicans, polyamine transport in S. cerevisiae has been well studied and is carried out by the proteins Tpo1 to Tpo4 (42, 43). S. cerevisiae TPO2 and TPO3 encode transporters specific for spermine, whereas TPO1- and TPO4-encoded pumps utilize putrescine, spermidine, and spermine as the substrates (43). The S. cerevisiae TPO1 gene product is the major plasma membrane-bound exporter involved in the detoxification of excess levels of intracellular spermidine (44). However, Tpo1 can excrete various substances in addition to polyamines such as the immunosuppressive drug mycophenolic acid (45) and mediates growth resistance to cycloheximide and quinidine (46). The S. cerevisiae Tpo1 polyamine transporter is pH dependent, so that polyamine uptake occurs at pH 8 while excretion occurs at acidic pH (pH 5). Both processes are catalyzed by activity of a polyamine-H⁺ antiporter (47). We reasoned that C. albicans Tpo1 homologs are likely to be functionally similar with respect to pH-dependent excretion of polyamines and other small cationic substances, including Hst 5. We found that both C. albicans FLU1 and TPO2 have high homology with S. cerevisiae TPO1, suggesting that either may be involved in efflux of polyamines and Hst 5. However, we report here that Flu1, but not Tpo2, is involved in efflux of Hst 5 and is responsible for reduction of its toxicity in C. albicans.

MATERIALS AND METHODS

Yeast strains and peptides.

C. albicans strains used in this study are described in Table 1. C. albicans cdr1Δ/Δ, mdr1Δ/Δ, and cdr1/cdr2Δ/Δ strains were generously provided by D. Sanglard (University of Lausanne and University Hospital Center, Lausanne, Switzerland). Knockout mutants of C. albicans Flu1 (flu1Δ/Δ) and C. albicans Tpo2 (tpo2Δ/Δ) were constructed in a CAF4-2 parental background using the URA blaster strategy as detailed previously (40) with URA3 replacement at the RPS10 locus to generate Ura⁺ constructs. For restoration strain construction (flu1Δ/FLU1 or tpo2Δ/TPO2), a copy of wild-type (WT) FLU1 or TPO2 was introduced into the RPS10 locus of the respective flu1Δ/Δ or tpo2Δ/Δ strain using the vector CIp10 containing URA3 (41). Yeast cells were transformed with purified PCR products using the Frozen-EZ Yeast Transformation II kit (Zymo Research, CA). Cells were maintained in yeast extract-peptone-dextrose (YPD; Difco) medium with the addition of uridine when required and stored at −80°C. Spermidine was obtained from Sigma. Hst 5 and N-terminally biotin-labeled Hst 5 (BHst 5) were synthesized by Genemed Synthesis Inc. (San Antonio, TX). BHst 5 was verified to have biological activity similar to that of unlabeled Hst 5 by candidacidal assays (38, 48).

Table 1.

C. albicans strains used in this study

Strain	Parent	Genotype	Phenotype	Reference
CAF4-2	CAF2-1	Δura3::imm434/ura3::imm434	Ura−	59
flu1Δ/Δ mutant	CAF4-2	Δura3::imm434/ura3::imm434 Δflu1::FRT/RP10::URA	Ura+	This study
tpo2Δ/Δ mutant	CAF4-2	Δura3::imm434/ura3::imm434 Δtpo2::FRT/RP10::URA	Ura+	This study
flu1Δ/FLU1 restoration	flu1Δ/Δ mutant	Δura3::imm434/ura3::imm434 flu1/FLU1::FRT/RP10::FLU1 ORFa	Ura+	This study
tpo2Δ/TPO2 restoration	tpo2Δ/Δ mutant	Δura3::imm434/ura3::imm434 tpo2/TPO2::FRT/RP10::TPO2 ORF	Ura+	This study
DSY448	CAF2-1	Δcdr1::hisG-URA3-hisG/Δcdr1::hisG	Ura+	23
DSY465	CAF2-1	Δmdr1::hisG-URA3-hisG/Δmdr1::hisG	Ura+	23
DSY654	CAF2-1	Δcdr1::hisG/Δcdr1::hisG Δcdr2::hisG-URA3-hisG/Δcdr2::hisG	Ura+	23

^aORF, open reading frame.

Phylogenetic analysis.

Candidate polyamine excretion proteins in C. albicans were identified by homology with TPO1 to TPO4 in S. cerevisiae using CLUSTAL 2.0.12 and the Candida Genome Database. CLUSTAL 2.0.12 was used to calculate a phylogenetic tree with use of the neighbor-joining algorithm.

Spermidine efflux assays.

Spermidine was labeled (BODIPY-Spd) as described previously using boron dipyrromethene difluoride (BODIPY) 630/650 succinimidyl ester (Invitrogen) (41). C. albicans cells (WT, flu1Δ/Δ, tpo2Δ/Δ, flu1Δ/FLU1, and tpo2Δ/TPO2) were grown in YPD broth overnight at 30°C with shaking and then subcultured and regrown to mid-log phase. Harvested cells were washed twice in 100 mM sodium phosphate buffer (NaPB) adjusted to pH 5.0 or pH 7.4, and then cells (5 × 10⁶) were incubated with 100 μM BODIPY-Spd in 100 mM NaPB for 15 min, allowing BODIPY-Spd to enter the cells (41). After 15 min of incubation, cells were washed with 100 mM NaPB (pH 5.0/pH 7.4), resuspended in 250 μl 100 mM NaPB (pH 5.0/pH 7.4), and loaded in a 96-well plate to a concentration of 2 × 10⁶ cells/well. A standard curve for BODIPY-Spd was determined to be linear (R² = 0.9899) within the range of 3.5 nmol to 10 nmol spermidine. Released BODIPY-Spd was detected using a Bio-Tek multifunction plate reader and Gen5 software and was quantified by reference to the standard curve. Statistical analysis and determination of kinetic parameters were performed using Prism version 5.0 (GraphPad Software, San Diego, CA). Each experiment was performed in triplicate in at least two biological replicates.

Detection of extracellular Hst 5 and cytoplasmic transported Hst 5.

For assays measuring Hst 5 efflux from C. albicans, cells (1 × 10⁷) were first loaded with Hst 5 by incubation with BHst 5 (31 μM) for 20 min in 10 mM NaPB (pH 7) and then washed and resuspended in NaPB (pH 5) for 0 min, 2 min, 10 min, 20 min, and 30 min. Supernatants were collected at each time point and assayed by a slot blot assay. For the time course assay of cytosolic levels of Hst 5, cells (1 × 10⁷) were suspended in 1 ml of NaPB, and BHst 5 was added to a final concentration of 31 μM. The cell mixtures were incubated with constant shaking for 1 min, 30 min, and 60 min. The reaction was stopped at each time point, and the cell pellet was subjected to cytoplasmic extraction as described before (49). Equal amounts of protein from each extract were subjected to SDS-PAGE using 12.5% acrylamide gels. Extracellular and cytoplasmic BHst 5 (∼3 kDa) was detected using streptavidin conjugated with horseradish peroxidase (Pierce). Quantitative analysis of BHst 5 proteins was performed using a Bio-Rad GS-700 imaging densitometer (Arcus II; Agfa) and Quantity One software (version 4.2). Concentrations of cytoplasmic BHst 5 from at least three independent experiments were averaged, and the means were curve fitted using Prism 5.0 software. Differences between experimental groups were evaluated for significance by unpaired t test, using Prism 5.0 software.

Total RNA isolation, cDNA synthesis, and real-time reverse transcription-PCR (RT-PCR).

Log-phase cultures (1 × 10⁷ cells/ml) of C. albicans (wild-type and flu1Δ/Δ and tpo2Δ/Δ mutant strains) were exposed to 31 μM Hst 5 at 30°C with shaking for 1 min, 30 min, 60 min, and 90 min; total RNA isolation was performed using the RNeasy minikit (Qiagen). The absence of genomic DNA contamination was confirmed by PCR amplification of the 18S rRNA housekeeping gene. One microgram of total RNA was used per reaction for the first-strand synthesis (cDNA) using the iScript cDNA synthesis kit (Bio-Rad). The sequences of primers were as follows: 5′-GACCAATGACTGCTGCTGAA-3′ and 5′-GCCAGTTTTGACGTTTGGAT-3′ for DUR3, 5′-GATCATCTGTGCTGCTGGAA-3′ and 5′-AGCAGCTGAAGCCAATGT-3′ for DUR31, 5′-CAACGATATTGCTCCTGAAG-3′ and 5′-TGGCTCTTCTCGATAATTCA-3′ for FLU1, and 5′-CGATGGAAGTTTGAGGCAATA-3′ and 5′-CTCTCGGCCAAGGCTTATACT-3′ for 18S RNA. Amplifications (amplicons) were 150 to 200 bp in length, and the annealing temperatures were 54°C. Amplification and detection were carried out in 96-well plates on an iCycler iQ real-time detection system (Bio-Rad). All samples contained 10 μl iQ SYBR Green supermix (2× concentration), 1 μl forward primer, 1 μl reverse primer, 1 μl template (cDNA), and 17 μl nuclease-free water. Fluorescent data were collected and analyzed with iCycler iQ software. The threshold cycle (ΔC_T) value was obtained by calculating the difference between C_T values of the target gene and the normalizer (18S). The ΔC_T at 1 min was used as a reference (baseline). The ΔΔC_T values were then calculated as the difference between the ΔC_T of each sample and the baseline and were transformed to absolute values (2^−ΔΔCT) to calculate comparative expression levels. Mean mRNA levels for each gene were calculated from at least three independent biological replicates. Differences between experimental groups were analyzed with Expert qPCR Analysis software.

Candidacidal assays of Hst 5.

The susceptibility of C. albicans cells to Hst 5 was measured using microdilution plate assays as we have previously described (50). Briefly, 10 ml of YPD medium with uridine (50 μg/ml) was inoculated with single colonies of each strain. Cells were grown overnight at room temperature. Overnight cultures were diluted to an A₆₀₀ of 0.3 to 0.4 and then were incubated at 30°C with shaking (250 rpm) until an A₆₀₀ of ∼1.0 was attained. Cells were washed twice with 10 mM NaPB, pH 7.4, then cells (1 × 10⁶) were mixed with different concentrations of Hst 5 at 30°C for 30 min and diluted in 10 mM NaPB, and aliquots of 500 cells were spread onto YPD agar plates and incubated for 36 to 48 h until colonies could be visualized. Cell survival was expressed as a percentage compared with untreated control cells, and the percent killing was calculated as [1 − (number of colonies from peptide-treated cells/number of colonies from control cells)] × 100%. Assays were performed in triplicate for each strain.

Biofilm formation upon treatment with Hst 5.

Overnight-grown cells were diluted to an A₆₀₀ of 0.3 to 0.4 and were incubated at 30°C with shaking at 250 rpm for 4 h. Cells were harvested, washed, and resuspended to an A₆₀₀ of ∼1 in phosphate-buffered saline (PBS; pH 7.4). One milliliter of cells was added to each well of a 12-well tissue culture plate (Becton, Dickinson, Franklin Lakes, NJ). The plate was incubated at 37°C for 3 h, then nonadherent cells were removed by gently washing the wells with PBS, and fresh medium (1 ml yeast nitrogen base [YNB]) was added and incubated for 24 h at 37°C. After 24 h, all samples were washed with 1 ml PBS twice. For control groups, 1 ml YNB was added to each well after washing. For test groups, medium was removed and replaced with 500 μl Hst 5 (60 μM) for 1 h, after which 500 μl YNB medium was added. All samples were incubated at 37°C for another 36 h. Biofilm formed was then removed mechanically and collected into preweighed microcentrifuge tubes, and dry weights of cells per well were calculated. Experiments were conducted in triplicate. The data represent mean dry weight of cells with standard error. The percentage of biofilm reduction was calculated as [1 − (dry weight of Hst 5-treated cells/dry weight of control cells)] × 100%. Differences between experimental groups were evaluated for significance by an unpaired t test, using Prism 5.0 software.

RESULTS

Identification of C. albicans TPO transporter family members.

To identify likely C. albicans polyamine efflux transporters, we screened the C. albicans genome database (CGD) for orthologs of the major S. cerevisiae transporters responsible for polyamine efflux, which include TPO1, TPO2, TPO3, and TPO4. Homology searching using S. cerevisiae TPO genes showed five related members of the C. albicans TPO family, including FLU1 (orf19.6577), TPO2 (orf19.7148), TPO3 (orf19.4737), orf19.341, and TPO4 (orf19.473) (Fig. 1). The orf19.6577 gene was identified previously as FLU1 (fluconazole resistance), as this gene encodes a multidrug efflux pump at the plasma membrane and plays a major role in fluconazole resistance (33). However, our analysis showed that C. albicans FLU1 is a TPO family member also annotated in CGD as TPO1. C. albicans FLU1 and TPO2 were found to be the closest homologs to S. cerevisiae TPO1, which encodes a major spermidine efflux transporter. C. albicans TPO3 (orf19.4737) and orf19.341 were found to be most closely related to S. cerevisiae TPO2 and TPO3, while C. albicans TPO4 (orf19.473) was homologous to S. cerevisiae TPO4.

Fig 1.

Phylogenetic analysis of C. albicans TPO transporter family members and their relationship with Saccharomyces cerevisiae Tpo. Members of the C. albicans TPO transporter family (http://www.candidagenome.org) are shown with TPO-homologous protein sequences from the yeast S. cerevisiae. C. albicans FLU1 (orf19.6577)- and TPO2 (orf19.7148)-encoded proteins have the closest phylogenetic relationship to S. cerevisiae polyamine excretion protein Tpo1, a spermidine efflux transporter.

As we have shown that C. albicans cells grown in medium supplemented with spermidine had higher resistance to Hst 5 killing than did cells grown with putrescine and spermine (41), we hypothesized that the candidal homolog of S. cerevisiae Tpo1 would be a major efflux transporter for spermidine as well as Hst 5. To identify the nearest ortholog of ScTpo1, CLUSTALW analysis was performed by optimal alignment of the amino acid sequences using CLUSTAL2.0 software with both C. albicans Flu1 and Tpo2. However, both C. albicans Flu1 and Tpo2 proteins showed high homology to S. cerevisiae Tpo1p, having 60% and 59% similarities, respectively. Hence, we selected both C. albicans Tpo2 and Flu1 proteins as likely candidates to mediate efflux of Hst 5 in C. albicans.

Spermidine efflux in C. albicans is pH dependent and is reduced by FLU1 deletion.

Rates of spermidine efflux by S. cerevisiae Tpo1 are higher at pH 5 due to its polyamine-H⁺ antiport activity (35). Therefore, we examined whether this was also the case for C. albicans Tpo1 homologs. C. albicans WT, flu1Δ/Δ and tpo2Δ/Δ mutant, and flu1Δ/FLU1 and tpo2Δ/TPO2 restoration strains were incubated with BODIPY-Spd to load cells with labeled polyamine. No differences in total cellular concentrations of BODIPY-Spd after 15 min were detected between these strains. Cells were then washed and transferred to medium without polyamines, and efflux of BODIPY-Spd was measured for 5 min, during which efflux was linear, allowing calculation of efflux rates (Table 2). The total efflux of BODIPY-Spd from WT cells was increased by nearly 3-fold by placing cells in medium at pH 5 compared to placing them at pH 7.4. Thus, C. albicans Tpo1 homologs function similarly to S. cerevisiae Tpo1 with respect to pH dependence. C. albicans tpo2Δ/Δ and tpo2Δ/TPO2 cells had spermidine efflux rates identical to those of WT cells under both neutral and acidic conditions, showing that removal of TPO2 had no impact on spermidine efflux. However, C. albicans flu1Δ/Δ cells had a statistically significant reduction in their rate of spermidine efflux at pH 5, while no difference was detected under neutral pH conditions (Table 2). Restoration of FLU1 in flu1Δ/Δ (flu1Δ/FLU1) cells returned the efflux rate of BODIPY-Spd at acidic pH to that of WT cells (Table 2). These results suggest a role for Flu1 but not Tpo2 in efflux of spermidine in C. albicans, particularly under acidic conditions.

Table 2.

Spermidine efflux rates for C. albicans strains

pH	Efflux rate (nmol spermidine/2 × 106 cells/min) for straina:
	WT	flu1Δ/Δ mutant	tpo2Δ/Δ mutant	flu1Δ/FLU1 restoration	tpo2Δ/TPO2 restoration
7	0.23 ± 0.03	0.26 ± 0.03	0.26 ± 0.03	0.23 ± 0.04	0.25 ± 0.04
5	0.63 ± 0.03	0.53 ± 0.02b	0.62 ± 0.03	0.63 ± 0.03	0.62 ± 0.02

^aData presented are means ± standard deviations of duplicate assays.

^bP < 0.05.

FLU1 deletion reduces the efflux of Hst 5.

Since Hst 5 is taken up via C. albicans Dur3 and Dur31 spermidine uptake transporters, we expected that Hst 5 might be transported out of the cells by efflux through polyamine transporters (Tpo proteins). Therefore, we examined Hst 5 efflux from C. albicans cells loaded with sublethal concentrations of biotin-labeled Hst 5 (BHst 5). C. albicans wild-type, tpo2Δ/Δ and flu1Δ/Δ mutant, and flu1Δ/FLU1 and tpo2Δ/TPO2 restoration strains were incubated first with BHst 5 for 20 min, as we have shown that this exposure time results in cells that have a significant cytosolic load of Hst 5 but still retain cellular integrity as measured by lack of propidium iodide uptake (40). After washing to remove extracellular BHst 5, cells were resuspended in buffer adjusted to pH 5, cell supernatants were collected, and BHst 5 was quantified at 1 min, 2 min, 10 min, 20 min, and 30 min. We found that the amount of extracellular BHst 5 in the flu1Δ/Δ strain was reduced compared with wild-type cells, while there was no difference in concentrations of extracellular BHst 5 with the tpo2Δ/Δ strain or in the restoration strains (flu1Δ/FLU1 and tpo2Δ/TPO2) (Fig. 2A), thus showing a role for Flu1p in the efflux of Hst 5. We expected that decreased efflux of cytosolic Hst 5 as a result of deletion of FLU1 would be accompanied by an increase in total intracellular Hst 5. Therefore, total cytosolic content of BHst 5 was compared after 30 min and 60 min in wild-type, flu1Δ/Δ, and tpo2Δ/Δ strains and their respective restoration strains. We found that flu1Δ/Δ cells had significantly (P < 0.05) larger amounts of cytoplasmic BHst 5 than did wild-type, tpo2Δ/Δ, or tpo2Δ/TPO2 cells (Fig. 2B). Furthermore, levels of cytoplasmic BHst 5 were returned to wild-type levels in the flu1Δ/FLU1 restoration strain (Fig. 2B). To determine whether azole drug transporters might be involved in efflux of Hst 5, the sensitivity to Hst 5 was also examined in wild-type and azole drug efflux transporter-deficient strains (cdr1Δ/Δ, mdr1Δ/Δ, and cdr1/cdr2Δ/Δ strains). There was no significant difference of sensitivity to Hst 5 between wild-type and mutant strains (Fig. 3), thus showing that other azole drug transporters are not involved in reducing the toxicity of Hst 5.

Fig 2.

Efflux of BHst 5 is reduced, resulting in increased cytoplasmic accumulation in the C. albicans flu1Δ/Δ strain. (A) C. albicans (wild-type, flu1Δ/Δ and tpo2Δ/Δ mutant, and flu1Δ/FLU1 and tpo2Δ/TPO2 restoration strains) were loaded with biotin-labeled Hst 5 (BHst 5) (31 μM) for 20 min and then resuspended in NaPB (pH 5) for 0 min, 2 min, 10 min, 20 min, and 30 min. Supernatants were collected at each time point and immunoblotted to detect extracellular concentrations of BHst 5 by densitometry for each time point. Only the flu1Δ/Δ mutant had significantly (P < 0.01) reduced BHst 5 in cell supernatants. (B) Cytosolic concentrations of BHst 5 were quantified following incubation of cells with BHst 5 (31 μM) for 1 min, 30 min, and 60 min. Cytoplasmic fractions were isolated at each time point, resolved using 12.5% SDS-PAGE, immunoblotted, and quantified as described above. Only the flu1Δ/Δ mutant had significantly (P < 0.05) increased cytosolic BHst 5 at both 30 min and 60 min compared to wild-type cells. Data shown are means from at least three independent experiments.

Fig 3.

C. albicans drug transporter-deficient strains are not altered in sensitivity to Hst 5. C. albicans deletion mutants (cdr1Δ/Δ, mdr1Δ/Δ, and cdr1/cdr2Δ/Δ) were exposed to Hst 5 (7.5 μM to 31 μM) using microdilution plate assays, and percent killing was calculated. There was no significant difference of sensitivity to Hst 5 between wild-type and mutant strains.

To rule out the possibility that the observed differences in cytosolic Hst 5 levels in flu1Δ/Δ cells were due to increased uptake of Hst 5, we examined expression levels of the two major Hst 5 uptake transporter-encoding genes (DUR3 and DUR31) after exposure to Hst 5. Real-time RT-PCR was performed to measure the transcript levels of DUR3 and DUR31, as well as FLU1 genes in C. albicans strains following treatment with sublethal doses of Hst 5 for 1 min, 30 min, 60 min, and 90 min. The 18S rRNA housekeeping gene was used as an internal control. Transcriptional levels of DUR3 and DUR31 in wild-type (Fig. 4) and flu1Δ/Δ and tpo2Δ/Δ (data not shown) strains were not altered over 60 min of exposure to Hst 5, as analyzed by Expert qPCR Analysis software. This suggested that the observed increase in cytosolic levels of Hst 5 in flu1Δ/Δ cells is due to reduced efflux of Hst 5. Further, we did not detect differences in spermidine uptake in flu1Δ/Δ cells (data not shown), pointing toward normal functioning of Dur uptake proteins up to 60 min. However, transcriptional levels of DUR3 but not DUR31 were increased by 5-fold after 90 min of exposure to Hst 5 (Fig. 4), suggesting that compensation for toxic effects of Hst 5 begins after longer exposure times. Interestingly, transcriptional levels of FLU1 did not change over a 90-min course of treatment with Hst 5 (Fig. 4). Thus, Hst 5 is not likely to induce FLU1 gene expression as a potential mechanism of development of resistance, at least within the early exposure times that we examined.

Fig 4.

Expression of C. albicans FLU1 is not altered following treatment of cells with Hst 5. Hst 5 (31 μM) was added to cultures at 0 min, and aliquots were removed for RNA extraction at 1, 30, 60, and 90 min as indicated. Transcriptional profiles of C. albicans DUR3, DUR31, and FLU1 were characterized by quantitative real-time RT-PCR, and expression levels were normalized to 18S RNA. Values for each gene are expressed as a change (n-fold) relative to the 1-min time point. Only DUR3 transcript levels were increased at 90 min compared to wild-type cells.

Deletion of FLU1 increases sensitivity of C. albicans planktonic and biofilm cells to Hst 5.

To detect whether efflux of Hst 5 has an effect on its toxicity, the sensitivity to Hst 5 was examined in wild-type, flu1Δ/Δ and tpo2Δ/Δ mutant, and flu1Δ/FLU1 and tpo2Δ/TPO2 restoration strains under neutral (pH 7.4) and acidic (pH 5) conditions. Interestingly, all strains, including the wild-type, deletion, and restoration strains, had higher sensitivity to Hst 5 at pH 5 than at pH 7 (Fig. 5A), indicating the importance of the H⁺ electrochemical gradient in the Hst 5 candidacidal function. Deletion of FLU1 resulted in an increase in Hst 5 sensitivity under both neutral and acidic conditions (Fig. 5A) that was reversed upon gene restoration (Fig. 5B). At pH 7, the sensitivity of the flu1Δ/Δ strain was increased by 1.5-fold at higher (31 μM) Hst 5 concentrations (filled circles), while increased sensitivity of the flu1Δ/Δ strain was more evident at lower (<15 μM) Hst 5 concentrations at pH 5 (open circles) due to the overall higher sensitivity. However, there was no difference in Hst 5 sensitivity between tpo2Δ/Δ and wild-type cells. Next, we examined the susceptibility of biofilms formed by wild-type and flu1Δ/Δ cells to Hst 5. Biofilms formed by flu1Δ/Δ cells were slightly higher in total dry weight than those formed by the wild-type strain, showing that loss of Flu1 does not inhibit biofilm formation (Fig. 6). Reduction of biofilm formation in the presence of Hst 5 was detected in both wild-type and flu1Δ/Δ strains. The flu1Δ/Δ cells, like flu1Δ/Δ planktonic cells, had a significantly higher percentage of biofilm reduction (21.6 ± 2.2) than did the wild type (14.6 ± 1.2) (Fig. 6). Thus, the presence of Flu1p also reduces Hst 5 toxicity in C. albicans biofilms.

Fig 5.

The C. albicans flu1Δ/Δ mutant has increased sensitivity to Hst 5. C. albicans deletion mutants (flu1Δ/Δ and tpo2Δ/Δ and restoration strains) were exposed to Hst 5 (7.5 μM to 31 μM) for 30 min at pH 5 (open circles) and pH 7 (filled circles) using microdilution plate assays, and percent killing was calculated. The C. albicans flu1Δ/Δ strain was found to be hypersensitive to Hst 5 under both neutral and acidic conditions, while the tpo2Δ/Δ strain did not differ from the parental WT strain in its sensitivity to Hst 5. (B) Restoration strains (flu1Δ/FLU1 and tpo2Δ/TPO2) had sensitivity to Hst 5 similar to that of the parental strains under both neutral and acidic conditions.

Fig 6.

C. albicans flu1Δ/Δ biofilms have increased sensitivity to Hst 5. Biofilms were formed for 24 h in tissue culture plates using wild-type and flu1Δ/Δ cells, and then Hst 5 (60 μM) was added for 1 h and quenched with addition of fresh yeast nitrogen base (YNB) medium. Plates were incubated at 37°C for another 36 h, and then biofilms were removed mechanically and dry weights of cells per well were calculated. Reduction of biofilm formation in the presence of Hst 5 was detected in both the wild type (P = 0.0165) and the flu1Δ/Δ strain (P = 0.0133). The flu1Δ/Δ cells had a significantly higher percentage of biofilm reduction (21.6 ± 2.2) than did wild-type cells (14.6 ± 1.2).

DISCUSSION

Hst 5 has become a focus of interest as a natural therapeutic agent for oral candidiasis because of its potent candidacidal and candidastatic activities while being nontoxic to humans (51). Previously, we observed that the concentration of cytosolic Hst 5 was gradually reduced from intact C. albicans cells (40), suggesting that C. albicans possesses a transport mechanism for efflux of Hst 5 in order to reduce its toxicity. Therefore, identification of the route of Hst 5 efflux is crucial for understanding the mechanism for possible Hst 5 resistance.

Polyamines are essential aliphatic polycations needed for normal growth and morphogenesis in fungi, so that their depletion results in growth cessation while high intracellular accumulation may be cytotoxic (52). Thus, fungi must strictly regulate their intracellular polyamine pools using membrane-localized polyamine transporter systems that facilitate internalization of exogenous polyamines and excretion of excess intracellular polyamines. The major polyamines in fungi are putrescine, spermidine, and spermine, and all three have a net positive charge (pK_a values of 9 to 10), similar to Hst 5. We identified Dur3 and Dur31 polyamine transporters as major carriers for Hst 5 entry into C. albicans cells and found that Hst 5 uptake is competitive with the polyamine spermidine (41). Since energy-dependent uptake of Hst 5 into C. albicans is based upon its recognition as a polyamine homolog, it is likely that Hst 5 is also subject to excretion by C. albicans cells by polyamine efflux transporters.

S. cerevisiae Tpo1 is the major polyamine excretion protein and plays an important role in polyamine detoxification. We found that C. albicans Tpo2 is similar to its close homolog, S. cerevisiae Tpo1, in terms of its ability to transport spermidine; however, unlike S. cerevisiae Tpo1 and C. albicans Flu1, its transport abilities were not pH dependent, since tpo2Δ/Δ cells had spermidine efflux rates similar to those of the wild type under both neutral and acidic conditions (Table 2). Deletion of C. albicans TPO2 did not affect cytosolic Hst 5 levels and Hst 5 excretion or toxicity, thus showing that Hst 5 is not likely to be a substrate for transport via C. albicans Tpo2, in contrast to C. albicans flu1Δ/Δ cells. These results highlight the substrate specificity of polyamine transporters as described previously for S. cerevisiae Tpo proteins (43). For example, Glu-201, Glu-324, and Glu-574 residues of S. cerevisiae Tpo1 are important for spermine transport, while Ser-342 is required for spermidine efflux (43, 47). Thus, a unique set of amino acids on C. albicans Flu1 may be responsible for the transport of both polyamine and Hst 5 as the substrates and may be lacking in Tpo2. Although efflux of Hst 5 was reduced in flu1Δ/Δ cells (Fig. 2A), it was not completely stopped, suggesting that additional mechanisms or yet-to-be-identified transporters may be involved in the efflux of Hst 5.

Similarly, C. albicans Flu1 also has high homology to CaMdr1, which is responsible for fluconazole resistance and functions as a drug H⁺ antiporter. Therefore, we expected that azole drug efflux transporters (e.g., Cdr1, Mdr1, and Cdr2) would also be involved in efflux of Hst 5 in C. albicans. However, we did not find any significant difference in sensitivity to Hst 5 between the wild type and the transporter-deficient cdr1Δ/Δ, mdr1Δ/Δ, and cdr1/cdr2Δ/Δ mutant strains. Thus, C. albicans Mdr1 and Cdr1/2 transporters appear to possess high substrate specificity, perhaps through their transmembrane segments, which are envisaged to confer substrate specificity (53), and do not recognize linear basic peptides such as Hst 5. Although ABC transporters provide the greatest level of resistance to antimicrobial peptides (AMPs) in bacterial infections (54), there is little evidence showing that azole drug transporters (e.g., Cdr1, Mdr1, and Cdr2) are involved in resistance to AMPs in yeast cells. The fact that Hst 5 has been shown to be active against some fluconazole-resistant clinical isolates of C. albicans supports this concept (55).

Our finding that transcriptional levels of FLU1 did not change over a 90-min course of treatment with Hst 5 has important implications for potential use of Hst 5 and other AMPs as therapeutic agents. Unlike azole therapy, which has been shown to upregulate genes like MDR1 and FLU1 that are involved in azole resistance (56), we found no evidence of Hst 5-induced FLU1 gene overexpression, suggesting that C. albicans is not likely to develop resistance to Hst 5. Fluconazole resistance can occur without overexpression of CDR1/MDR1 due to alteration of the drug target enzyme (lanosterol demethylase) (16). Similarly, it is possible that fungi might develop resistance to Hst 5 through mechanisms other than FLU1 overexpression, such as changes in transporter affinity for Hst 5 with long-term exposure. However, to date no stable Hst 5-resistant C. albicans isolate has been identified. Surprisingly, transcriptional levels of DUR3 were increased significantly after 90 min of exposure to Hst 5, showing that cells mount compensatory responses upon accumulation of critical levels of cytosolic Hst 5. Intracellular Hst 5 may be treated as a polyamine by being localized within fungal cellular compartments (57) with excess levels trafficked to the vacuole for detoxification. Indeed, we have noted considerable vacuolar localization of Hst 5 (40). However, Hst 5 may overload vacuolar capacity, so that levels of trafficked Hst 5 (like its polyamine counterpart [58]) exceed the capacity of vacuoles to protect the cytosol from toxic levels. At the same time, accumulation of excess cytosolic Hst 5 may create anion channel-like pathways in the cell membrane, leading to changes in the membrane potential. The altered membrane potential may energize secondary transporters such as cations and polyamine transporters, which in turn increase the expression of genes encoding transport proteins such as Dur3.

In conclusion, this is the first report demonstrating that C. albicans utilizes the fungal polyamine efflux transporter Flu1 to pump Hst 5 out of the cell and thus reduce its toxicity. Although efflux of Hst 5 by Flu1 reduces its toxicity, Hst 5 treatment does not induce FLU1 gene overexpression during short-term exposure, underscoring its therapeutic potential. Further in vitro and in vivo research is needed to develop specific Flu1 inhibitors to increase the candidacidal activity of Hst 5.