Influence of Iron Regulation on the Metabolome of Cryptococcus neoformans

Iron is an essential nutrient for virtually all organisms and acts as a cofactor for many key enzymes of major metabolic pathways. Furthermore, iron plays a critical role in pathogen-host interactions. In this study, we analyzed metabolomic changes associated with iron availability and the iron regulatory protein Cir1 in a human fungal pathogen Cryptococcus neoformans. Our metabolite analysis revealed that Cir1 influences the glycolytic pathway, ergosterol biosynthesis and inositol metabolism, which require numerous iron-dependent enzymes and play important roles in pathogenesis and antifungal sensitivity of the fungus. Moreover, we demonstrated that increased cellular iron content and altered gene expression in the cir1 mutant contributed to metabolite changes. Our study provides a new insight into iron regulation and the role of Cir1 in metabolome of C. neoformans.


Introduction
Iron is an essential nutrient for virtually all organisms and plays as a cofactor for many key enzymes of major metabolic pathways, such as the tricarboxylic acid cycle, respiration, amino acid biosynthesis, and the synthesis of lipids and sterols [1]. Acquisition of iron from the environment is problematic, mainly because of its insoluble nature under aerobic conditions. To overcome extremely low bioavailability of iron, organisms have developed various efficient iron transport systems. Examples include reduction of ferric iron to ferrous iron to increase its permeability and secretion of iron chelating molecules siderophores [2]. Although its essential roles, iron causes detrimental effects if exceed. Free iron catalyzes the formation of oxygen radicals via the Fenton reaction, which causes protein denaturation, DNA breakage, etc. Therefore, iron acquisition and homeostasis are tightly regulated in a cell.
Iron also plays a critical role in pathogen-host interactions. Typically, both the pathogen and the host require iron and compete over it during the infection process. While the hosts have developed various means to limit the availability of intra-or intercellular iron contents upon infection, the pathogens have evolved distinct systems for survival within the host environment. Therefore the mechanisms of iron transport and homeostasis in pathogens have been extensively studied [3].
Cryptococcus neoformans is a fungal pathogen and a leading cause of pulmonary and central nervous systemic mycosis in immunocompromised individuals such as HIV-infected patients. C. neoformans has been considered an excellent model fungal pathogen to study iron transport and homeostasis because of its intriguing connection with virulence. For example, iron deprivation influences the formation of the polysaccharide capsule and synthesis of melanin, which are considered major virulence factors [4]. Moreover, recent studies suggested that the high-affinity reductive iron transport pathway is required for full virulence in a mouse model of cryptococcosis and for proliferation in brain tissue [5,6]. A study aiming to identify the major iron regulatory protein in C. neoformans revealed that a GATA-type zinc finger transcription factor Cir1 controls expression of genes involved in iron transport and homeostasis [4]. The same study also showed that expression of genes required for major virulence factors such as melanin formation and capsule synthesis were regulated by Cir1. Furthermore, a number of genes required for signaling and metabolic pathways were shown to be influenced by deletion of CIR1 suggesting importance of the iron regulatory protein in physiology of C. neoformans. Finally, the strains lacking CIR1 became avirulent suggesting significance of the protein in pathogenesis of C. neoformans [4]. Although the previous study suggested that Cir1 plays a major regulatory role in transcriptional networks related to iron metabolism and virulence, the metabolic consequences of deletion of CIR1 are largely unknown. Therefore, we applied metabolite analysis to further understand the functions of Cir1 in C. neoformans.
Metabolomics profiles cellular metabolites and is a rapidly expanding area strategy in the post-genomics era [7]. In metabolite profiling, it is preferable to use a wide-spectrum metabolite analysis technique, which is rapid, reproducible and stable during sample analysis [8]. Recent technological advances in mass spectrometry have realized reliable and highly sensitive measurements of metabolites [9]. Among them, gas chromatography mass spectrometry (GC-MS) has been successfully applied to analyze and interpret multiparametric metabolic responses in living systems to pathopysiological and environmental perturbations [10]. In the last few decades, a number of chemometric tools have been applied to the interpretation and quality assessment of MS-based metabolomic data. Recently, an analysis of metabolite profiles was performed to understand how iron deficiency affected the metabolomes of a common laboratory strain Saccharomyces cerevisiae, which revealed that iron deficiency led to changes in glucose metabolism, amino acid biosynthesis, and lipid biosynthesis [11]. In the current study, we aimed to identify specific metabolic pathway that is mostly influenced by iron or Cir1 and to understand the role of Cir1 in the metabolome of C. neoformans. Our data showed that abundance of some of metabolites, such as those involved in the glycolytic pathway, ergosterol biosynthesis or inositol metabolism, was changed in the cir1 mutant.

Strains and Growth Conditions
The strains used in this study were routinely grown in YPD (1% yeast extract, 2.0% bacto-peptone, 2.0% glucose), YNB (yeast nitrogen base, Difco) with 2.0% glucose. Defined low-iron media containing 5 g of glucose, 5 g of asparagine, 400 mg of K 2 HPO 4 , 80 mg of MgSO 4 N7H 2 O, 250 mg of CaCl 2 N2H 2 O, 57 mg of boric acid, 5 mg of CuSO 4 N5H 2 O, 10 mg of MnCl 2 N4H 2 O, 2 mg of ZnSO 4 N7H 2 O, and 4.6 mg of sodium molybdate per liter. Lowiron water was prepared using Chelex-100 (Bio-Rad) resin. The strains used in this study were C. neoformans var. neoformans serotype D B3501A (the wild-type) and B3CIR572 (the cir1 mutant), which were constructed previously [4]. To extract metabolites, cells were pre-grown in 50 ml of low-iron YNB overnight at 30uC. After incubation, cells were collected by centrifugation at 4,000 rpm for 5 min, were washed twice with low-iron water, and were resuspended in 25 ml of defined low-iron media. Suspension of cells was diluted 1/10 in 50 ml of defined low-iron media, followed by incubating at the same temperature for 12 h with shaking. Cells were transferred to 150 ml of fresh defined low-iron media with or without 100 mM FeCl 3 , grown for additional 12 h, and were harvested. Cell pellets were used for metabolomic analysis. Total six independent cultures (biological replicates) for each strain were prepared for each growth condition and were subsequently analyzed throughout the study.

Preparation of Fungal Extracts
In order to separate broth and the cells, harvested culture broths were centrifuged at 1,500 rpm. Collected cells were subjected to glass bead lysis and metabolites were extracted with 75% boiling ethanol. Extracts were dried in a freezing dryer (12 h) which were derivatized in two steps to protect carbonyl functions. First, the dried samples were dissolved in 100 ml of 20 mg/ml solution of methoxyamine hydrochloride in pyridine (Sigma) and incubated at 75uC for 30 min. The volatility of polar compounds was increased by exchanging acidic protons against trimethysilyl group using 100 ml of N-methyl-N-trimethylsilyltrifluoroacetamide (MSTFA) (Sigma) at 70uC for 30 min.

GC-MS Based Non-targeted Metabolite Profiling
Each 1 ml of the derivatives was injected in a split mode (1:25) into a Varian CP-3800 gas chromatography system coupled to a Varian 4000 ion-trap EI/CI mass spectrometric detector system (Varian, CA). The samples were analyzed using a VF-1MS capillary column (30 m60.25 mm) coated with 0.25-mm low bleed polymer, a FactorFour TM column (Varian), equipped with an integrated 10 m guard column (Varian, CA, USA), and vaporized and partial least square-discriminate analysis (PLS-DA) score scatter plot (B) represent significant differences of metabolites between the wild-type (WT) and the cir1 mutant (cir1D). Data show difference of 24 samples (6 biological replicates of the wild-type or the cir1 mutant grown in low-or high-iron medium). Remarkable score plots were observed in clusters at the given strain, respectively. However, no statistically significant difference was observed when metabolites of the strains grown in low-iron medium were compared with that of the strains grown in high-iron medium. doi:10.1371/journal.pone.0041654.g001 to separate the derivative metabolites at 250uC. For analysis, the initial oven temperature was held at 70uC for 2 min; ramped to 300uC at a rate of 10uC /min and held for 5 min. Helium (purity .9.999%) was used as a carrier gas at a constant flow rate of 1 ml/min. The temperature of the EI ion source and injector were set to 200uC and 250uC, respectively. The electron impact ionization (70 eV) was utilized, and mass data were collected in a full-scan mode (m/z 50-1000). The metabolites were identified by comparison to the NIST 2005 database (version 2.0, FairCom Co., USA).

Spectral Data Processing
The raw GC-MS data were converted into CDF (NetCDF) files by using the Vx capture software (version 2.1, Laporte, MN) and subsequently processed by XCMS toolbox for automatic peak detection and alignment. R-program version 2.9.0 (www.r-project. org) and XCMS version 1.16.3 were used. The XCMS parameters for the R language were performed by simple commands as XCMS's default settings (http://masspec.scripps.edu/xcms/ documentation.php). For multivariate statistical analysis, the XCMS output was further processed using Microsoft Excel (Microsoft, Redmond, WA, USA). The data were arranged on a three-dimensional matrix consisting of arbitrary peak index (RTm/z pair), sample names (observations), and peak area (variables).

Multivariate Statistical Analysis
The resulting three-dimensional data table was entered into the SIMCA-P + (version 12.0, Umetrics, Umeå, Sweden) software package for multivariate statistical analysis. Unsupervised principal component analysis (PCA) was performed to observe general clustering and trends of metabolite differences among all samples. Supervised partial least squares discriminant analysis (PLS-DA) model was used to maximize metabolite variations and identify significantly altered metabolites responsible for such variations. [12] The quality of the fitting model can be explained by R 2 X and Q 2 Y values. R 2 X display the variance explained in the model and indicates the goodness of fit. Q 2 Y display the cumulative predicted variation in Y. The values of these parameters approaching 1.0 indicated a stable model with a predictive reliability. The VIP (variable importance in the projection) values were calculated and used to identify the most significantly different metabolites for clustering of specific class-separating and were further validated by Student's t-test using STATISTICA (version 7.0, StatSoft Inc., Tulsa, OK). Metabolites with VIP values greater than 1.0 and p values less than 0.05 (threshold) were selected as discriminating metabolites between the wild-type and the mutant strains. Box and whisker plots were generated in STATISTICA to show the significantly difference of chemistry measurement between the 2 classes of samples. Pair-wise metabolite-to-metabolite correlations were calculated using the Pearson's correlation coefficient (rMet) test using the R programs.

Determination of Cellular Iron Concentration
To determine iron concentration, the wild-type strain and the cir1 mutants were grown as describe above. Cell cultures were washed three times with low-iron water and mixed with lysis Figure 2. Box-and-whisker plots illustrating discriminative metabolites of the cir1 mutant (cir1D) and the wild-type (WT) grown in low-(L.iron) or high-iron (H.iron) medium. A total of 18 discriminative metabolites were identified in the cir1 mutant. Among them, 12 metabolites were increased in the cir1 mutant (A), and 6 were decreased compared to that of the wild-type (B

Phenotypic Analysis
To estimate antifungal drug sensitivity, the wild-type and the cir1 mutant strain were grown in YPD overnight at 30uC with shaking. After incubation, 10-fold dilutions of cells (starting at 10 4 cells) were spotted onto YPD plates containing various azole antifungal drugs. The plates were incubated at 30uC for 2 days and photographed.

Quantitative Real-time RT-PCR
Primers for real-time RT-PCR were designed using Primer Express software 3.0 (Applied Biosystems) and are listed in Table S1. Cell cultures prepared as described above. Total

Results and Discussion
Non-targeted Metabolite Profiling of C. neoformans by Using MS The current study focused on determining the influence of Cir1 on the metabolome in C. neoformans. A total of 6 biological replicates, cultured in the absence or presence of an inorganic iron source (FeCl 3 ) were prepared for the wild-type and the mutant strain lacking CIR1, and their metabolites were analyzed using MS based non-targeted metabolite profiling. In general, ion chromato-gram does not provide an evidence of difference and similarity of complex sample sets and, therefore, often requires further sophisticated data analysis. We chose XCMS software for GC-EI-MS/MS data processing, and applied a normalization method to the total spectral area to obtain aligned metabolome data of the wild-type and the cir1 mutants individually [13,14,15]. The processed metabolite profiles (972 peaks, extracted by XCMS from the GC-EI-MS/MS data sets) were used to compare metabolite profiles of the strains in each growth condition via multivariate statistical analysis (Table S2).
Because of the qualitative similarity and complexity of the metabolome data sets, visualization of preliminary metabolite changes is challenging; thus, chemometric multivariate statistical analysis was carried out using principal component analysis (PCA) [16]. Data in the score plot of mass spectra showed clear separation, with good cross-validation, between the wild-type and the cir1 mutant. Moreover, we evaluated the quality of our . The cir1 mutant displayed decreased sensitivity to azole antifungal drugs. The growth of the wild-type (WT) or the cir1 mutant (cir1D) in media containing an antifungal drug was monitored. Ten-fold serial dilutions of cells (starting at 10 4 cells) were spotted onto YPD plates with the drug indicated. Plates were incubated at 30uC for 2 days. doi:10.1371/journal.pone.0041654.g004 models by using R 2 X and Q 2 values, which indicate goodness of fit and predictability of models, respectively [17]. According to our PCA score plots, differences of metabolite composition between the wild-type and the cir1 mutant could be explained by 0.74 of the R 2 X and 0.63 of the Q 2 values ( Figure 1A). These results suggest that metabolite profiles are consistent within biological replicates and that deletion of CIR1 resulted in distinct metabolic characteristics in the mutant cells.

Discriminative Metabolites in the Strain Lacking CIR1
We subsequently searched for discriminative metabolites that contributed to the separation of the metabolite profiles of the cir1 mutant from the wild-type. In order to discover them, student's ttest was conducted for annotated peaks by interpreting the first weight vector in the PLS-DA, assuming that metabolite concentrations follow a normal distribution in the same genetic background ( Figure 1B). PLS-DA is commonly used to detect hidden variables that focuses on class separation [18]. Discrimi- native metabolites were selected using variable importance of the projection (VIP .1.0) and their p-value statistics (p,0.005).
Total 18 discriminative metabolites were identified in the cir1 mutant, and their levels were significantly different from those of the wild-type (Table 1). To illucidate inter-individual variations in the wild-type and the cir1 mutant according to iron availability, the yields of these metabolites were compared using box-and-whiskers plots (Figure 2). Most of these discriminative metabolites exhibited small changes when they were compared between low-and highiron medium indicating that iron availability triggers only marginal differences in C. neoformans at least under the current experimental conditions used. The previous transcriptome analysis revealed that, in the wild-type strain, 483 genes were downregulated and 250 genes were up-regulated more than 2-fold in low-iron versus high-iron medium based on Q-value statistics. The majority of differentially expressed genes in the wild-type strain upon iron availability included those related to iron transport and homeostasis, and DNA metabolism and repair. This and the results of our current metabolite profiling suggested that altered transcript levels in the wild-type strain did not produce a significant change in the metabolic flux in the wild-type strain. Our data also agreed with the recent study by Shakoury-Elizeh et al., who suggested that alterations in protein or transcript levels do not always correlate with the metabolic flux in S. cerevisiae because a metabolic pathway is largely determined by enzyme kinetics and substrate concentrations [11]. The influence of the gene CIR1 on metabolism was dramatic in C. neoformans. Glycolytic pathway intermediates and ergosterol related compounds were increased in the cir1 mutants while lactic acid, propylstearate, propylpalmitate, peiperidone and sucrose were decreased compared to the wildtype. Among the discriminative metabolites in the cir1 mutant, the highest magnitude changes occurred in glucose, inositol, arabitol, ergosterol and ergostene.

Cir1 Influences the Glycolytic Pathway and Respiration
The increase in glucose in the cir1 mutant was of particular interest, as it implied that major carbon assimilation processes are affected by the deletion of CIR1 in C. neoformans. Glucose is the main carbon source for energy production, is metabolized to pyruvate via glycolysis, and is oxidized via the tricarboxylic acid (TCA) cycle and respiration. It is known that iron influences both the TCA cycle and respiration. Two TCA cycle enzymes, aconitase and succinate dehydrogenase, contain Fe-S clusters, and complexes I-IV in the electron transport chain contain numerous Fe-S clusters as well as heme. In S. cerevisiae, expression of several genes involved in the TCA cycle and the electron transport chain were significantly down-regulated in cultures grown under iron deficient conditions [19]. Furthermore, recent metabolite analysis by Shakoury-Elizeh et al. revealed that iron deficiency caused depletion of glucose and glycolytic intermediates and increased flux of glucose through the glycolytic pathway for energy production but decreased respiration in S. cerevisiae [11]. Our metabolite analysis indicated that levels of glucose and glycolytic intermediate glycerol were increased in the cir1 mutants. These results imply possible up-regulation of the glycolytic pathway and respiration in the cir1 mutant, presumably due to increased intracellular iron levels, caused by deletion of CIR1. Indeed, a previous study showed that Cir1 acts not only as a transcriptional activator but also as a transcriptional repressor of genes required for iron uptake, which is also supported by the reduced growth phenotype of the mutant cells in the presence of excess iron or phleomycin, which induces DNA breakage via ironcatalyzed oxidation [4]. To confirm this hypothesis, we determined total cellular iron contents of the cells. The results suggested that total iron levels were significantly increased in the cir1 mutant compared to the wild-type strain both in low-and high-iron conditions ( Table 2). Taken together, these results show that deletion of CIR1 in C. neoformans leads to increased intracellular iron levels and subsequent up-regulation of ironrequiring process like the glycolytic pathway and respiration. These results are opposite to what had been observed in irondeficient S. cerevisiae cells. Therefore, we could suggest that flux of glucose might be reduced in C. neoformans due to deletion of CIR1, which resulted in increased glucose level in the cir1 mutants.

Cir1 Influences Ergosterol Biosynthesis
Levels of ergosterol and its derivatives were also increased in the cir1 mutant ( Figure 2). Ergosterol is the major fungal cell membrane constituent and is a target of the azole antifungal drugs. The biosynthesis pathway for ergosterol is highly conserved among fungi and the enzymes involved are encoded by approximately 22 genes [20]. Iron, as form of heme, is known to influence the synthesis of ergosterol because two enzymes, Erg5 and Erg11 encoding C-22 desaturase and C14a-methyl sterol demethylase respectively, belong to the heme containing cytochrome P450 protein family [21]. Moreover, Erg3 and Erg25 encoding D 5,6 sterol desaturase and C-4 methyl sterol oxidase respectively contain non-heme iron.
In the current study, the effect of deletion of CIR1 on ergosterol biosynthesis was further investigated by measuring expression of genes required for the pathway by Q-RT-PCR. These analyses suggested that in a high-iron medium, expression of ERG11 (CNA00300) and ERG24 (CNA01070) was up-regulated (2.55-fold and 2.70-fold, respectively) in the cir1 mutant than in the wild-type strain ( Figure 3). Moreover, in a low-iron medium, the expression of ERG11 was even more significantly increased (4.25-fold) in the cir1 mutant compared to the wild-type, which suggested that Cir1 implicates transcriptional regulation of ERG11 and ERG24. The previous study showed that the cir1 mutants are more resistant to miconazole, of which phenotype was supported by increased expression of genes encoding enzymes for ergosterol synthesis in transcriptome of the mutant cells [4]. For further confirmation of up-regulation of ergosterol biosynthesis in the cir1 mutants, we evaluated the sensitivity of the mutant cells to various azole antifungal drugs including miconazole ( Figure 4). Our data confirmed that the cir1 mutants are more resistant to fluconazole, miconazole and ketoconazole than the wild-type.
Taken together, the results of our current metabolite analysis suggested that ergosterol synthesis was largely affected by deletion of CIR1, which caused alteration of sensitivity to azole antifungal drugs. Up-regulation of genes required for the pathway and increased cellular iron levels might contribute altered ergosterol metabolism in the cir1 mutant. In addition to ergosterol itself, our metabolite analysis indicated that some of its derivatives were increased in the cir1 mutants. Abundance of ergostene and methylergostadiene were highly increased in the cir1 mutants implying that composition of membrane sterol was changed significantly in the mutant cells. The same ergosterol derivatives were already shown to be associated with resistance to azole antifungal drug in Candida albicans [22]. Our data revealed that loss of Cir1 functions resulted in remodeling of ergosterol metabolism and membrane biosynthesis.

Cir1 Regulates Inositol Metabolism
Inositol (myo-inositol) is an essential structural basis for many secondary messengers including inositol phosphates, phosphatidylinositol (PI) and phosphatidylinositol phosphate (PIP) lipids in a cell. Inositol can be synthesized from glucose 6-phosphate by 2 sequential processes: the first, conversion of glucose 6-phosphate to inositol 3-phosphate, is mediated by inositol-3-phosphate synthase, and the second, dephosphorylation of inositol 3-phosphate to generate inositol, by inositol monophosphatase. Cells can also import inositol from the environment via inositol transporters [23]. Several studies have shown that inositol and its derivatives play an important role in signaling pathways, sexual reproduction and virulence of the fungus [24,25,26,27]. Involvement of inositol in pathogenesis of C. neoformans has been of particular interest since its elevated levels was found in the mammalian brain, which is the most preferred host niche of the fungus [26,28,29,30]. The genome of C. neoformans contains an unusually large gene family encoding at least 10 inositol transporters [26], and the recent study by Wang et al. suggested that, among them, ITR1A and ITR3C are the major inositol transporters and are required for full virulence of C. neoformans [31]. The contribution of inositol metabolism to virulence has also been described in Candida albicans, which can acquire inositol by de novo synthesis or by importing it; either mechanism is sufficient to support the wildtype growth in vitro and full virulence [32]. More importantly, enzymes associated with inositol metabolism, such as phospholipases and lipases, contribute significantly to virulence of a number of other fungal pathogens [33].
We found that the level of inositol was significantly increased in the cir1 mutants than in the wild-type strains, suggesting that inositol metabolism is up-regulated by deletion of CIR1. This observation led us to compare the levels of expression of genes in inositol phosphate metabolism in the cir1 mutant and to compare with that in the wild-type. Several genes in the pathway were selected based on the previous transcriptome data and their relative expression levels were investigated. Figure 5 shows the results of Q-RT-PCR for inositol oxygenases (CNG03010 and CNH03170), myoinositol oxygenase (CNF03950), and methylmalonate-semialdehyde dehydrogenase (CND02030) in the pathway. Overall expression levels of these genes, except myoinositol oxygenase, were significantly increased in the cir1 mutant. These results suggested that inositol metabolism is indeed up-regulated in the cir1 mutant, which could explain why the metabolite analysis resulted in increased inositol contents in the mutant cells. We should also note here that the overall trends of expression of the genes were similar to previous transcriptome analysis [4]. Taken together, results of the current study suggested Cir1 influences on inositol metabolisms via transcriptional regulation of key genes in the pathway. In summary, our metabolome analysis revealed that deletion of the gene encoding a major iron regulatory transcription factor Cir1 largely influences iron required metabolisms such as the glycolytic pathway and respiration as well as ergosterol biosynthesis. Moreover, inositol metabolism is regulated by Cir1 in C. neoformans.  Author Contributions