ABSTRACT
Mitochondrial dysfunction in pathogenic fungi or model yeast causes altered susceptibilities to antifungal drugs. Here we have characterized the role of mitochondrial complex I (CI) of Candida albicans in antifungal susceptibility. Inhibitors of CI to CV, except for CII, increased the susceptibility of both patient and lab isolates, even those with a resistance phenotype. In addition, in a C. albicans library of 12 CI null mutants, 10 displayed hypersusceptibility to fluconazole and were severely growth inhibited on glycerol, implying a role for each gene in cell respiration. We chose two other hypersusceptible null mutants of C. albicans, the goa1Δ and ndh51Δ mutants, for transcriptional profiling by RNA-Seq. Goa1p is required for CI activity, while Ndh51p is a CI subunit. RNA-Seq revealed that both the ndh51Δ mutant and especially the goa1Δ mutant had significant downregulation of transporter genes, including CDR1 and CDR2, which encode efflux proteins. In the goa1Δ mutant, we noted the downregulation of genes required for the biogenesis and replication of peroxisomes, as well as metabolic pathways assigned to peroxisomes such as β-oxidation of fatty acids, glyoxylate bypass, and acetyl coenzyme A (acetyl-CoA) transferases that are known to shuttle acetyl-CoA between peroxisomes and mitochondria. The transcriptome profile of the ndh51Δ mutant did not include downregulation of peroxisome genes but had, instead, extensive downregulation of the ergosterol synthesis gene family. Our data establish that cell energy is required for azole susceptibility and that downregulation of efflux genes may be an outcome of that dysfunction. However, there are mutant-specific changes that may also increase the susceptibility of both of these C. albicans mutants to azoles.
INTRODUCTION
Global infectious diseases caused by fungi represent a major health problem (1). The development of new therapies, whether to potentiate immunity or kill pathogens with antifungal drugs, is slow. The ideal antifungal agent should have broad fungicidal activity, not select for resistant strains, and have minimal adverse effects. None of the current drugs fulfill all or even some of the criteria for an ideal agent. The azole antifungals, including the newer triazoles, target the ergosterol synthesis pathway by inhibiting the Erg11p 14-α-demethylase protein. Problems with their use include the selection of resistant strains because of their fungistatic property. Consequently, inherently resistant Candida spp. have become more common among clinical isolates during treatment (1). Azole resistance in C. glabrata has resulted in a modification of breakpoints for the standardized CLSI susceptibility assay to reflect its reduced susceptibility (1).
Resistance to fluconazole among Candida spp. occurs through several mechanisms or combinations of mechanisms. The Erg11p target may be overexpressed, or key mutations in Erg11p may result in reduced drug binding. The latter problem also appears to be similar to the “hot-spot” mutations in the Fks1p (β-1,3-glucan synthase) observed in echinocandin-resistant strains of C. glabrata (2). Also, an inability of drugs to penetrate biofilms contributes to resistance (3). Triazole resistance also occurs as a result of the upregulation of multidrug efflux transporters of two gene families, the ABC (ATP-binding cassette) transporters and the major facilitator gene superfamily (MFS). C. albicans has two CDR genes (CDR1 and CDR2) and a single MDR1 (Candida drug resistance and multiple-drug resistance, respectively) gene that is not related to MFS genes of other organisms. The major facilitator efflux pump superfamily proteins transport small solutes in response to chemiosmotic ion gradients (4). Regulation of efflux protein expression can also occur by gain-of-function mutations in the transcription factors Tac1p and Mrr1p that are associated with greater Cdr1p/Cdr2p or Mdr1p expression, respectively, in resistant cells (5, 6).
An association of decreased fluconazole susceptibility with mitochondrial dysfunction in C. glabrata has been reported by several laboratories (7–12). Ferrari et al. (9) demonstrated that a mitochondrial deficiency in C. glabrata strain BPY41 caused resistance to azoles associated with the upregulation of ABC transporter genes CgCDR1, CgCDR2, and CgSNQ2. Strains BYP40 (azole sensitive) and BYP41 (azole resistant), related and both isolated from the same patient, were compared to determine if mitochondrial dysfunction confers a selective advantage in virulence. While displaying a growth defect in vitro, strain BYP41 (a petite mutant) was more virulent in mouse infections. The gain of fitness in strain BYP41 was an apparent result of several gene changes, including increased oxidoreductive metabolism, carbohydrate metabolism, and stress responses, while genes associated with mitochondrial functions were downregulated in microarray analyses. Kaur et al. (10) screened a transposon mutant library of C. glabrata for strains altered in azole susceptibility. They identified several such mutants with altered susceptibility that were functionally associated with retrograde signaling from the mitochondria to the nucleus (CgRTG2, increased susceptibility) or mitochondrial biogenesis (CgSUV3, CgMRP14, and CgSHE9, decreased susceptibility). In S. cerevisiae, dysfunctional mitochondria activate the pleiotropic drug resistance pathway for adaptation (13).
While the relationship of mitochondrial functions to azole susceptibility has been extensively studied in C. glabrata, much less has been reported about C. albicans. Cheng et al. (14) described a petite C. albicans strain (J5) whose decrease in susceptibility to fluconazole and voriconazole but not itraconazole was noted. The petite strain was originally isolated from wild-type C. albicans SC5314 by serial passage through mouse spleens and then shown to have uncoupled oxidative phosphorylation. Other C. albicans petite strains have been described, but most are induced by harsh chemical treatments, and therefore, correlations of specific mitochondrial dysfunctions with azole resistance are more difficult to interpret (15–17). Recently, knockout strains of C. albicans have been constructed that lack either GOA1 or NDH51 (18–21). Goa1p is required for optimum electron transport chain (ETC) complex I (CI) activity. Strains with genes deleted accumulate ROS, undergo apoptosis, and are avirulent. NDH51 encodes the 51-kDa subunit of the NADH dehydrogenase CI of the ETC whose deletion results in morphogenesis defects.
Our hypothesis is that mitochondrial dysfunction, especially related to energy production, can potentiate azole susceptibility in C. albicans and other fungi. If verified, combination therapy with a fungus-specific antimitochondrial compound could be effective in extending the usefulness of the triazole antifungal drugs. Using specific inhibitors of the ETC complexes (CI to CV), CI mutant libraries, as well as the NDH51 (ndh51Δ) and GOA1 (goa1Δ) knockout strains described above, we verify this hypothesis. We also look globally at mutant strains by using RNA-Seq analyses to define transcriptional changes that may impact susceptibility to triazoles.
MATERIALS AND METHODS
Strains, strain maintenance, and plasmids.All of the strains used in the present study are listed in Table S1 in the supplemental material and were maintained as frozen stocks in 96-well plates and propagated on yeast extract-peptone-dextrose (YPD) agar when needed (1% yeast extract, 2% peptone, 2% glucose, 2% agar). The goa1Δ and ndh51Δ null mutant strains and the corresponding strains with the genes goa1/GOA1 and ndh51/NDH51 reconstituted were grown similarly.
Antifungal agents.Piericidin A (PdA) was purchased from Enzo Life Sciences. C12E8 was purchased from Affymetrix, Inc. Other ETC complex inhibitors and the azole antifungals were purchased from Sigma Chemical Co. These compounds were reconstituted according to the manufacturers' directions.
Antifungal susceptibility testing.Drug susceptibility testing was carried out with flat-bottom, 96-well microtiter plates (Greiner Bio One) by using the broth microdilution protocol according to the Clinical and Laboratory Standards Institute M-27A methods. All strains were evaluated for growth in the presence of fluconazole, and growth was reported as a percentage of that of untreated cells. Overnight cultures were prepared in YPD and washed, and ∼103 cells/10 μl were inoculated into microtiter wells. CI to CV ETC inhibitors (Sigma-Aldrich), including rotenone (10 μg/ml), C12E8 (4 μg/ml), PdA (4 μg/ml), TTFA (10 μg/ml), antimycin A (10 μg/ml), KCN (10 μg/ml), and oligomycin (10 μg/ml), were used alone or in combination with fluconazole (22, 23). Growth was evaluated by measuring cell density (optical density at 595 nm [OD595]) after 24 h of incubation at 30°C. Experiments were repeated at least three times. Data were averaged, and the statistical significance of differences between treatments was determined.
MIC determinations were also done with lab stock cultures of azole-susceptible and -resistant strains. For these experiments, two of the most active inhibitors (C12E8 and PdA) from the previous screening were used with a 0.2-, 4-, or 32-μg/ml concentration of fluconazole. Strains were grown as described above. Relative growth was calculated on the basis of OD595 data and visualized by using heat maps.
Screening of C. albicans CI mutants for fluconazole susceptibility.Putative mitochondrial ETC CI proteins of C. albicans were identified through the Candida Genome Database by using FASTA and BLASTP. All protein alignments were manually reviewed. From a morphogenesis mutant library provided by Noble and Johnson (24), 12 putative CI mutants were grown overnight in YPD broth and then plated on YPD agar medium containing fluconazole or on YPG agar plates (1% yeast extract, 2% peptone, 3% glycerol, 2% agar) to evaluate growth on a nonfermentable carbon source. Plates were incubated at 30°C for 48 h and photographed.
Drop plate assays with antifungal drugs.The susceptibility of all strains to antifungal drugs was tested by plating 5 μl of serial dilutions of 5 × 100 to 5 × 10−4 cells onto YPD agar plates containing antifungal drugs. Yeast cells were obtained from overnight cultures grown in YPD broth at 30°C, washed with saline, and standardized by hemacytometer counts. Growth of each strain was evaluated after 48 h of incubation at 30°C.
Rapid adaptation to fluconazole, defined as growth that occurs within 48 h of treatment, was also examined. Strains of C. albicans were grown overnight at 30°C in synthetic complete medium and washed with saline, and about 104 cells were spread onto Sabouraud dextrose (SD) agar medium with or without 128 μg/ml of fluconazole. Cultures were incubated at 30°C for 4 and 6 days, and colony counts were determined. For these experiments, cdr1Δ cdr2Δ and tac1Δ knockout strains and strains that overexpressed CDR1 and MDR1 were used along with goa1Δ, ndh51Δ, goa1/GOA1, and ndh51/NDH51 mutant strains. Goa1p is required for ETC CI activity, while Ndh51p is a subunit of ETC CI.
RNA preparation and next-generation sequencing.Total RNA was extracted and submitted to Otogenetics Corporation (Norcross, GA) for RNA-Seq assays. In brief, the integrity and purity of total RNA were assessed by using an Agilent Bioanalyzer (Agilent Technologies) and measuring the OD260/OD280 ratio. One to 2 μg of cDNA was generated from 100 ng of total RNA by using the Clontech SmartPCR cDNA kit (catalog number 634925; Clontech Laboratories Inc., Mountain View, CA). cDNA was fragmented by using Covaris (Covaris, Inc., Woburn, MA), profiled by using an Agilent Bioanalyzer, and subjected to Illumina library preparation using NEBNext reagents (catalog number E6040; New England BioLabs, Ipswich, MA). The quality, quantity, and size distribution of the Illumina libraries were determined by using Agilent Bioanalyzer 2100. The libraries were then submitted to Illumina HiSeq2000 according to the standard operation. Paired-end 90- or 100-nucleotide reads were generated and subjected to data analysis by using the platform provided by DNAnexus, Inc. (Mountain View, CA). We obtained 11,279,512, 10,906,588, and 10,630,834 total reads for the wild type (SC5314) and the goa1Δ and ndh51Δ mutants, respectively.
All samples were mapped to SC5314 with TopHat 1.3.3 and then analyzed with Cufflinks 1.2.1 for expression level of genes and isoforms. Comparison of expression levels was conducted with Cuffdiff. We compared C. albicans SC5314 with the goa1Δ and ndh51Δ mutants, as well as both mutants with each other. RNA-Seq data for selected genes were validated by real-time PCR.
Relative quantification of differentially expressed genes by real-time PCR.Each C. albicans strain was grown overnight in 10 ml of YPD medium at 30°C. For RNA extractions, cultures were diluted to 108 cells per ml in 20 ml of fresh YPD medium, grown at 30°C for 2 h, and collected for subsequent RNA isolation. Total RNA was extracted by using TRIzol following a phosphate-buffered saline (PBS) wash. The quality and concentration of RNAs were measured with a nanospectrophotometer, and approximately 1 μg of the total RNA was subjected to first-strand cDNA synthesis (QuantiTect Reverse Transcription; Qiagen). Real-time PCR assays were performed with 20-μl reaction volumes that contained 1× iQ SyBR green Supermix (Bio-Rad), including a 0.2 μM concentration of each primer (see Table S2 in the supplemental material) and 8 μl of a 1:8 dilution of each cDNA from each strain. Quantitative reverse transcription-PCR for each experiment was performed in triplicate by using Bio-Rad iQ, and the transcription level of each gene was normalized to 18S rRNA levels. Data are presented as means ± standard deviations. The 2−ΔΔCT (where CT is the threshold cycle) method of analysis was used to determine the n-fold change in gene transcription.
Efflux of rhodamine 6G.Approximately 1 × 106 yeast cells from each overnight culture were transferred to YPD medium and allowed to grow for 4 h. Cells were pelleted, washed twice with PBS (pH 7.0, without glucose), and suspended in glucose-free PBS to 108/ml. Cell suspensions were incubated at 30°C with shaking (200 rpm) for 120 min under glucose starvation conditions. The de-energized cells were pelleted, washed, and then suspended in glucose-free PBS to 108/ml, and then rhodamine 6G was added at a final concentration of 10 μM. Cells were incubated for 30 min at 30°C, washed twice, and then suspended in glucose-free PBS to 108/ml. At 10-min intervals, cells (1-ml volumes) were removed by centrifugation and the absorption at 527 nm of triplicate 100-μl volumes of supernatant was measured. Energy-dependent efflux was measured after the addition of 2% (final concentration) glucose. Glucose-free controls were included in all experiments. Efflux of rhodamine 6G was calculated from a rhodamine 6G concentration curve.
RESULTS
Fluconazole susceptibility of C. albicans is increased by ETC complex inhibitors.We initially compared the susceptibility of C. albicans SC5314 to fluconazole in the absence or presence of mitochondrial CI to CV inhibitors. Included were the inhibitors PdA (class I/A-type CI inhibitor), rotenone (class II/B-type CI inhibitor), and C12E8 (type C CI inhibitor), TTFA (CII inhibitor), antimycin (CIII inhibitor), KCN (CIV inhibitor), and oligomycin (CV inhibitor). We observed that all of the inhibitors increased fluconazole susceptibility, except for TTFA (CII inhibitor), which had a minor effect (1 dilution) (see Table S3 in the supplemental material). The low activity of the CII inhibitor could be due to the smaller amount of ATP produced via oxidoreduction of CII substrates. Of the CI inhibitors, C12E8 increased the fluconazole susceptibility to 0.063 μg/ml (MIC for 80% of the strains tested), which is significantly higher than the susceptibility seen when the the CII, CIII, CIV, and CV inhibitors were used. These data suggest that reduced ETC complex activity (but not CII) increases the susceptibility of C. albicans to fluconazole.
CI inhibitors increase fluconazole hypersusceptibility of C. albicans lab and patient isolates.The CI inhibitors C128E and PdA were assayed in combination with 0.2, 4, or 32 μg/ml of fluconazole. Controls consisted of the inhibitors or fluconazole used alone (Fig. 1). Increased susceptibility of 14 of 42 strains to fluconazole was noted when cells were incubated in the presence of PdA or C12E8 and 0.2 μg/ml of fluconazole compared to that of four strains incubated with only 0.2 μg/ml of fluconazole. The number of hypersusceptible strains increased as the concentration of fluconazole was increased to 4 and 32 μg/ml and in the presence of either CI inhibitor. At 32 μg/ml fluconazole, a higher number of CDR-overexpressing isolates was hypersusceptible in the presence of PdA than in the presence of C12E8. Both inhibitors were about equally effective against MDR-overexpressing isolates at the same concentration of fluconazole. The susceptibility of cultures treated with either inhibitor alone was much lower. These data demonstrate that the fluconazole susceptibility of a variety of C. albicans strains, including those that are resistant, can be increased in the presence of ETC CI inhibitors.
Susceptibility profiles of C. albicans strains in the presence of fluconazole or the CI inhibitor C12E8 or PdA alone or in combination with fluconazole (0, 2, 4, or 32 μg/ml). Relative growth is indicated by the scale at the right. Susceptibility was strain dependent, but the combination of 4 μg/ml of fluconazole and C12E8 or PdA increased the susceptibility of the largest number of strains, including those that were resistant to fluconazole. On the left are strain numbers, and on the right, strains are clustered according to their susceptibility, source, and/or resistance mechanisms. The combination of fluconazole with either C12E8 or PdA produced fluconazole susceptibility higher than that obtained with the inhibitor or fluconazole alone.
C. albicans mitochondrial CI null mutant strains, fluconazole susceptibility, and growth on glycerol.To further verify the enhanced fluconazole susceptibility in the presence of CI inhibitors, the susceptibility of C. albicans CI mutants to fluconazole was examined. From a library of C. albicans homozygous deletion mutants, we selected 12 lacking putative CI subunit orthologs (24). The 12 mutants and control strains SN250 and SC5314 were grown on YPD, on YPD plus 16 μg/ml of fluconazole, or on YP-glycerol (YPG) alone (Fig. 2). Control strains SN250 and SC5314 grew on YPD, on YPD containing 16 μg/ml of fluconazole, and on YPG. Of the 12 putative CI mutants, 10 displayed hypersusceptibility to 16 μg/ml of fluconazole in YPD agar compared to control strains. Nine of the 10 hypersusceptible CI mutants did not grow on YPG medium, demonstrating that a mitochondrial respiratory deficiency was associated with fluconazole hypersensitivity. Two CI mutants (orf19.3611 and orf19.3290 mutants) that had fluconazole susceptibilities like those of control strains grew on YPG agar. On the basis of growth or lack of growth on YPG, we conclude that loss of respiratory activity is associated with fluconazole hypersensitivity in most CI mutants. The orf19.4758 mutant had increased susceptibility to fluconazole but grew on YPG agar, indicating that it may not be directly involved in cell respiration (Fig. 2). Growth in the presence of glycerol may indicate that orf19.3611 and orf19.3290 are also associated with nonrespiratory mitochondrial activities. Importantly, of these complex proteins, orf19.287 and orf19.6607 are fungus specific, and both deleted mutants are attenuated or avirulent (24). The others are orthologs of mammalian mitochondrial CI proteins.
Growth of wild-type (SC5314 and SN250) and CI mutant strains of C. albicans on YPD lacking fluconazole (FLC; middle panel) and on YPG (right panel) and their susceptibilities to fluconazole (left panel). Of the 12 CI mutants, 10 have increased susceptibility to fluconazole and 9 are also defective for growth on YPG. Two strains (orf19.3611 and orf19.3290 mutants) have wild-type susceptibilities to fluconazole and are able to grow on YPG. Growth of the two resistant mutants on YPG indicates that their corresponding genes are probably not part of the respiratory pathway of mitochondria.
C. albicans CI knockout strains and azole susceptibility.C. albicans GOA1 and NDH51 have been linked to mitochondrial functions (18–21). Goa1p is required for optimum CI activity, while Ndh51p is a subunit of CI. The goa1Δ mutant has a severe deficiency in mitochondrial membrane potential and ATP formation, overproduces reactive oxidant species (ROS), has a shorter chronological aging pattern associated with apoptotic events, and has reduced CI activity (18, 19). Ndh51p is a subunit protein of CI and has impaired morphogenesis (20). The association of both genes with CI functions led us to study their antifungal susceptibility in vitro (Fig. 3).
(A) Antifungal susceptibility profiles of C. albicans wild-type, goa1Δ mutant, and goa1/GOA1-reconstituted strains in drop plate assays compared to those on YPD alone. The drug concentration is indicated below each panel. Susceptibility profiles were also determined for 5-flucytosine, micafungin, and amphotericin B (see Fig. S1 in the supplemental material). FLC, fluconazole. (B) Susceptibilities of the wild type, the goa1Δ mutant, and the goa1/GOA1-reconstituted strain to fluconazole. Data are presented as the percentage of growth compared with that of untreated cells (mean ± standard deviation of three independent experiments). (C) Susceptibility profiles of the wild type, the ndh51Δ mutant, and the ndh51/NDN51 mutant to fluconazole. The ndh51Δ mutant is hypersensitive to fluconazole compared to the control strains.
In comparison with the wild type and the goa1/GOA1 mutant, the goa1Δ mutant was hypersensitive on drop plates at 16 μg/ml of fluconazole, as well as itraconazole (0.4 μg/ml), ketoconazole (0.8 μg/ml), and miconazole (1.0 μg/ml [not shown]) (Fig. 3A) but equal in sensitivity to wild-type cells in YPD plus 5-flucytosine, amphotericin B, or micafungin (see Fig. S1 in the supplemental material). The goa1Δ null mutant was also growth inhibited in the presence of fluconazole compared to the control strains (Fig. 3B). The ndh51Δ mutant was also hypersensitive to fluconazole (16 μg/ml) compared to the strain with the reconstituted gene (ndh51/NDH51) and the wild-type strain (Fig. 3C). Both null mutant strains are unable to grow in YPG, indicating that the deleted genes have a role in mitochondrial respiration (data not shown). Thus, hypersusceptibility to azoles is associated with deletions of ETC CI genes or inhibition of all ETC complexes except CII.
Adaptation to fluconazole is reduced in null mutant strains.When large numbers of cells are plated on a high concentration of fluconazole, a rapid adaptation to decreased susceptibility has been observed in wild-type cells (25). At a fluconazole concentration of 128 μg/ml in YPD, adaptation occurred by 4 to 6 days in the wild-type (SC5314) strain, the MDR1- and CDR1/CDR2-overexpressing strains, and the strains with reconstituted genes (goa1/GOA1, nhd51/NDH51) (Fig. 4). However, the goa1Δ, ndh51Δ, tac1Δ, and cdr1/2Δ mutants were unable to adapt to high drug concentrations by 4 or 6 days of incubation (no increase in colony numbers) (Fig. 4).
Rapid acquisition of adaptation to fluconazole in strains of C. albicans. All strains were grown overnight in SD medium, standardized to cell number, and plated on SD medium containing 128 μg/ml of fluconazole. Cells from treated or untreated cultures were grown for 4 or 6 days on YPD agar, and the number of colonies of each strain was determined. The goa1Δ, ndh51Δ, Δcdr1/2, and Δtac1 mutants were unable to grow in the presence of fluconazole, while the wild-type (SC5314) and CDR1/CDR2- and MDR1-overexpressing strains adapted to fluconazole 4 and 6 days following incubation with the drug.
RNA-Seq.We used RNA-Seq to define the transcriptional profiles of the goa1Δ and ndh51Δ mutants. In the goa1Δ mutant, a total of 388 genes were downregulated 4-fold or more. Gene ontology (GO) analysis of the five highest categories of downregulated genes in this mutant showed that they are associated with transmembrane transporter activity (13%), amino acid metabolism (9%), carbohydrate metabolism, (7%), mitochondria (4%), and peroxisomal functions such as β-oxidation (3%) (Fig. 5). The statistically significant (P value of 3.93e−7) GO term enrichment of transporter genes was calculated for those with >4-fold-decreased expression in the goa1Δ mutant. The ABC family of proteins was among the products of the downregulated transporter genes in the goa1Δ mutant, suggesting that deletion of GOA1 perhaps caused a significant change in membrane fluidity or that reduced ATP levels in the null mutant may partially explain the downregulation of energy-requiring transporters (18, 19).
RNA-Seq transcription profile of the goa1Δ mutant. Functional categories are indicated as percentages of the total number of downregulated genes.
The transcriptional profile of the ndh51Δ mutant suggests similarities to, as well as differences from, the goa1Δ mutant (Fig. 6). As in the goa1Δ mutant, in the ndh51Δ mutant, a large number of transporter genes were also downregulated (see Fig. S2 in the supplemental material), although the relative abundance was less than that in the goa1Δ mutant (13% versus 8%). The transcription level of the peroxisomal gene cluster was about 4-fold lower in the ndh51Δ mutant than in the goa1Δ mutant. However, the transcription level of the mitochondrial cluster of genes was 2-fold greater in the ndh51Δ mutant, while that of the sterol synthesis cluster was 4-fold greater in the ndh51Δ mutant than in the goa1Δ mutant (discussed below).
RNA-Seq transcription profile of the ndh51Δ mutant. Functional categories are indicated as percentages of the total number of downregulated genes.
Efflux pumps and susceptibility.The ABC drug transporters were represented among the downregulated transporters of both mutants. Their association with azole resistance is well established (26–28). To verify that the hypersusceptibility of the goa1Δ and ndh51Δ mutants was associated with the downregulation of CDR1- and CDR2-encoded drug efflux transporters, we used real-time PCR to measure the transcription of the efflux pump genes CDR1, CDR2, and MDR1. The transcription of the ABC family members CDR1 and CDR2, but not MDR1, was significantly lower in the goa1Δ and ndh51Δ mutants than in the wild type (SC5314) and the strains with the goa1/GOA1 and ndh51/NDH51 genes reconstituted (Fig. 7).
Real-time PCR of CDR1, CDR2, and MDR1 is shown for the wild type and the goa1Δ and ndh51Δ mutants, the strains with reconstituted goa1/GOA1 and ndh51/NDH51 genes. The expression of both CDR1 and CDR2 is significantly reduced in both mutant strains but not in the wild type or the strains with reconstituted genes.
Next, rhodamine 6G, a substrate for CDR pumps, was used to measure efflux activity in the same strains, as well as a CDR-overexpressing strain. Without glucose, minimal efflux of rhodamine 6G was observed in all of the strains. Upon the addition of glucose to provide an energy source, increased export of rhodamine 6G resulted over the next 60 min, especially in the CDR1/CDR2-overexpressing strain (Fig. 8). In addition to the CDR1/CDR2-overexpressing strain, the strains with the reconstituted genes and the wild-type strain also displayed high efflux activity. In comparison, both null mutants (the goa1Δ and ndh51Δ mutants) were unable to transport rhodamine 6G, unlike the wild type and the strains with reconstituted genes. These data point to a deficiency in rhodamine 6G transport that apparently relates in part to the lack of transcription of CDR1/CDR2.
Efflux of fluorescent rhodamine 6G, a substrate of Cdr1p and Cdr2p pumps. All strains were grown overnight in YPD, starved for 2 h in PBS, incubated with rhodamine 6G, and then transferred to buffer. At 30 min, glucose was added to the cultures and the efflux of fluorescent rhodamine was subsequently measured for a total of 90 min. The CDR-overexpressing (OE) strain exhibited the greatest rhodamine efflux, followed by the strains with reconstituted genes, i.e., the strains with the genes goa1/GOA1 and ndh51/NDH51 reconstituted (intermediate efflux). The goa1Δ and ndh51Δ mutants did not show rhodamine efflux.
Other transcriptional clusters: peroxisomes and mitochondria are affected.Downregulation of a number of peroxisomal and mitochondrial genes was noted in the goa1Δ mutant (Tables 1 and 2). Peroxisomes are the cell sites of β-oxidation of fatty acids, gluconeogenesis, the glyoxylate bypass pathway, and the breakdown of peroxides (29). Many of these genes were downregulated (Table 1). Also, in the goa1Δ mutant, 15 genes associated with either peroxisome biogenesis or replication (PEX genes) were downregulated (Table 2). The peroxisomal genes that were strongly represented among the downregulated genes in the goa1Δ mutant include PEX4 (14.3-fold) and PEX13 (4.4-fold) (Table 2). However, unlike the goa1Δ mutant, neither a pex4 nor a pex13 null mutant was hypersensitive to fluconazole (data not shown).
RNA-Seq data for genes downregulated in the goa1Δ mutant
Peroxisome (PEX) genes downregulated in the goa1Δ mutant
Peroxisomes are functionally connected to mitochondrial activities in that they have a number of biochemical pathways in common (30). We next explored genes associated with these processes. Cross talk among peroxisomes and mitochondria requires the transport of acetyl coenzyme A (acetyl-CoA), which is not able to cross membranes because of its amphiphilic nature and bulkiness (31–33). Acetyl-CoA is a central intermediate in carbon metabolism that completely depends on the shuttling molecule carnitine to enter organelles in C. albicans (31, 32). In the goa1Δ mutant, genes encoding carnitine transferases (CTN1, CAT2, CTN3, and CRC1) that shuttle acetyl-CoA between mitochondria and peroxisomes are downregulated (Table 1). These data perhaps indicate an inability of these two organelles to coordinate acetyl-CoA transport, pending further study. The cat2 gene of C. albicans is essential for optimum β-oxidation but not required for virulence (34). In addition, the cat2Δ mutant and wild-type strains of C. albicans were equally sensitive to fluconazole (data not shown). Although a peroxisomal carnitine transporter(s) has not been identified, the C. albicans mitochondrial carnitine carrier protein encoded by CRC1 was found to be associated with the mitochondrial inner membrane (35). CRC1 was downregulated 4.7-fold in the goa1Δ mutant.
Genes encoding the glyoxylate bypass enzymes are strongly downregulated in the goa1Δ mutant only. Those genes include the isocitrate lyase gene ICL1 (17.8-fold), the malate synthase gene MLS1 (14.6-fold), and ACO1 (2.6-fold) (Table 1). Furthermore, the expression of several mitochondrial genes functionally related to the glyoxylate cycle was also lower in the goa1Δ mutant, including the citrate synthase gene CIT1 (9.1-fold). Citrate is transported from mitochondria to peroxisomes and converted to isocitrate for the glyoxylate cycle. The genes for malate dehydrogenase (MDH1; 9.1-fold), an intermediate transporter (SFC1; 42.5-fold), and fructose-1,6-bisphosphatase (FBP1), a key enzyme of gluconeogenesis, which is one of the subsequent reactions of the glyoxylate cycle, were downregulated by 4.3-fold in the goa1Δ mutant (Table 1). Interestingly, a majority of the genes of the glyoxylate cycle and gluconeogenesis were reported to be highly induced during phagocytosis and required for adaptation during carbon starvation (36–38). Their downregulation in the goa1Δ mutant is consistent with our previous findings that the null mutant was more readily killed by neutrophils and is avirulent (18). In C. albicans, β-oxidation of fatty acids is confined to peroxisomes, with a greater number of isozymes than in S. cerevisiae (12, 31). A majority of the genes associated with fatty acid β-oxidation are downregulated in the goa1Δ mutant (Table 1), indicating an inability of this mutant to utilize fatty acids as a sole carbon source. This hypothesis was confirmed by spot assays of the goa1Δ mutant on a variety of carbon sources, including oleate and other nonfermentation carbon sources (see Fig. S3 in the supplemental material).
In addition, other mitochondrial carrier protein genes were downregulated, including YMC2 (8.0-fold), TIM22 (7.0-fold), orf19.7267 (6.5-fold), orf19.3518 (3.9-fold), and orf19.4966 (3.6-fold) (data not shown). Thus, deletion of GOA1 may affect the transport between subcellular compartments of peroxisomes and mitochondria. Localization to mitochondria of nucleus-encoded proteins requires functional mitochondrial membrane receptors and transporter systems of the outer and inner membranes. In this regard, the inner, mitochondrial cristae house the respiratory complexes, including CI to CV. Interestingly, orf19.6062.3, the ortholog of which has a role in the maintenance of mitochondrial cristae and inner membrane architecture (39), was downregulated more than 200-fold in the goa1Δ mutant (data not shown). Furthermore, orf19.3089, with similar functions in the inner membrane structure, was downregulated 4.0-fold.
RNA-Seq analyses of the goa1Δ mutant are summarized in Fig. 9 to indicate all of the possible interactions of proteins from peroxisomes and mitochondria that reflect the transcriptome of the goa1Δ mutant.
Downregulation (brown) of genes of peroxisomes (left) or mitochondria (right) in the goa1Δ mutant. See the text for descriptions of the genes shown. The relative level of downregulation is shown as circles that vary in diameter.
The transcription profile of the ndh51Δ mutant includes downregulation of the ergosterol synthesis pathway.Compared to the goa1Δ mutant, transcriptional changes in the peroxisomal genes were not among those significantly downregulated in the ndh51Δ mutant. However, genes involved in ergosterol synthesis that were downregulated in the null mutant were much less strongly represented in the goa1Δ mutant (Fig. 10).
Ergosterol synthesis pathway (left) and transcription levels that correspond to different pathway genes (right). Downregulation of genes (green) or minimal change (black) is shown. All genes are aligned on the right for both strains. Downregulation of ERG genes is more common in the ndh51Δ mutant than in the goa1Δ mutant. P, phosphate; PP, pyrophosphate.
DISCUSSION
Mitochondria are the major sites of ATP formation via ETC CI to CV. Indispensable for cell growth and macromolecular synthesis, they also are required for adaptation to ROS that is generated via electron flow from oxidoreduction reactions of the ETC. Antifungal drugs such as azoles also induce a stress response, and it appears that mitochondria are part of the adaptation network in S. cerevisiae and C. glabrata that leads to fluconazole resistance (13, 27, 40). In a C. glabrata petite mutant and gain-of-function mutations in CgPDR1, CgPdr1 is singularly used to elevate the expression of the CgCDR1 efflux pump (ScPdr5) and resistance to fluconazole (27, 40). However, a respiratory mutant of C. glabrata exhibited hypersusceptibility to several azoles (12).
In C. albicans, differences from the fungi mentioned above exist. A C. albicans library was used to screen S. cerevisiae mutants lacking either PDR1 or PDR3, both regulators of multidrug transporter genes (41). C. albicans CTA4, ASG1, and CTF1 each activated the transcription of the S. cerevisiae PDR5-lacZ reporter and conferred resistance to azoles. However, C. albicans mutants null for each gene showed no changes in azole susceptibility and no activation of MDR1, CDR1, or CDR2 expression. Of the former three genes, only ASG1 was required for growth on media containing acetate, ethanol, or acetic acid. These data point to a rewiring of gene function in orthologs of C. albicans. ASG1 does not appear to be associated with a general defect in the assimilation of a carbon source, and in this regard, the asg1 null mutant could assimilate glycerol (41). However, its association with mitochondria and nonrespiratory functions has not been reported. CTA4 of C. albicans is most similar to the MRR1 transcription factor which regulates MDR1.
We have attempted to link portions of the transcriptional profiles of the goa1Δ mutant and the ndh51Δ mutant with their hypersusceptibility to fluconazole. Our conclusions are as follows. (i) Both mutants are downregulated in transporters, and of special importance, the CDR1/CDR2 efflux drug pumps. (ii) Cell energy output is definitely reduced in the goa1Δ mutant (13), but less is known about this phenotype in the ndh51Δ mutant. Efflux pump activity requires energy. (iii) In the goa1Δ mutant, a reduction of cell energy is associated with a CI dysfunction (13); (iv) The downregulation of genes involved in gluconeogenesis, β-oxidation, acetyl-CoA shuttling and cross talk with mitochondria, and the glyoxylate cycle may also create unattainable energy demands. (v) While membrane defects have not been defined in either mutant, the downregulation of a large number of transporters, as well as genes involved in ergosterol synthesis, suggests membrane perturbation.
Are there therapeutic implications of this study? GOA1 is found only in the CTG clade of the subphylum Saccharomycotina, which includes most Candida species but not C. glabrata. Thus, an inhibitor of Goa1p would have a narrow spectrum but perhaps still be useful therapeutically against most triazole-resistant Candida spp. except C. glabrata. Ndh51p is highly conserved among fungi but also mammalian cells. Thus, target specificity is less rigorous in regard to this protein. Still, there does not seem to be a precedent to pursue mitochondrial proteins as antifungal drug targets. The reasons that compel such a study, include the following. (i) At least two and probably more CI fungus-specific proteins exist, and other fungus-specific targets are very likely part of the pathogen genome. (ii) Mitochondria are energy conduits that are needed for many cell processes. (iii) Conceivably, specific antifungal drugs that target mitochondria could act in synergy with triazoles. (iv) There are new initiatives to identify compounds that attenuate or boost mitochondrial functions in cancer, neurodegenerative diseases, and type II diabetes (42–45). Assays for the identification of these compounds are relatively high throughput, using a two-tier system of growth in the presence of glucose and galactose, the latter to identify mitochondrial respiratory inhibitors in mammalian cells (42). In regard to fungi, readers are directed to a review of the use of mitochondria as drug targets (46).
ACKNOWLEDGMENTS
We thank the Fungal Genetics Stock Center and Susan Noble for providing the C. albicans mutant library. Thanks also to Theodore White, Joachim Morschauser, Dominique Sanglard, Patrice LePape, and David Perlin for providing azole-resistant mutants. We also thank Margaret Hostetter for providing the isogenic set of parental, reconstituted-gene-containing, and NDH51 null mutant strains.
L.Z. is an awardee of the National Distinguished Young Scholar Program in China. Nuo Sun received a Georgetown University graduate student grant to support this research. We also thank the Biomedical Graduate Research Organization of the Georgetown University Medical Center for research funds.
FOOTNOTES
- Received 25 July 2012.
- Returned for modification 27 September 2012.
- Accepted 5 November 2012.
- Accepted manuscript posted online 12 November 2012.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/AAC.01520-12.
- Copyright © 2013, American Society for Microbiology. All Rights Reserved.
