Investigation of Cryptococcus neoformans magnesium transporters reveals important role of vacuolar magnesium transporter in regulating fungal virulence factors

Abstract Cryptococcus neoformans is an important opportunistic fungal pathogen in humans. Recent studies have demonstrated that metals are critical factors for the regulation of fungal virulence in hosts. In this study, we systemically investigated the function of C. neoformans magnesium transporters in controlling the intracellular Mg balance and virulence‐associated factors. We identified three Mg transporters in C. neoformans: Mgt1, Mgt2, and Mgt3. While we could not detect a Mg2+‐related growth phenotype in mgt1 and mgt3 knockout strains, a GAL7p‐Mgt2 strain showed significant Mg‐dependent growth defects in the presence of glucose. Further analysis demonstrated that MGT2 is a homolog of MNR2 in Saccharomyces cerevisiae, which is localized to the vacuolar membrane and participates in intracellular Mg transport. Interestingly, a transcriptome analysis showed that Mgt2 influenced the expression of 19 genes, which were independent of Mg2+. We showed that melanin synthesis in C. neoformans required Mg2+ and Mgt2, and that capsule production was negatively regulated by Mg2+ and Mgt2. Repressing the expression of MGT2‐induced capsule, which resulted in an increased fungal burden in the lungs. Cumulatively, this study sets the stage for further evaluation of the important role of Mg homeostasis in the regulation of melanin and capsule in C. neoformans.


| INTRODUCTION
Cryptococcus neoformans is a major fungal pathogen in humans. It primarily afflicts immunodeficient and immunocompetent individuals, causing about 600,000 deaths per year globally Kronstad et al., 2011;May, Stone, Wiesner, Bicanic, & Nielsen, 2016). Cryptococcus infections begin in the lung tissue. By unknown mechanisms, fungal cells disseminate into the central nervous system, resulting in lethal meningitis Kronstad et al., 2011;May et al., 2016). Recent epidemiological studies have revealed that infections of Cryptococcus species are a significant health issue in the Pacific Northwest in Northern America, in addition to Europe, Africa, and China (Byrnes et al., 2010(Byrnes et al., , 2011Liu et al., 2017;May et al., 2016). Two major fungal virulence factors have been extensively studied. These are the generation of a cell-surface polysaccharide capsule structure and melanization using L-DOPA as a substrate Kronstad et al., 2011). The capsular structure is critical for shedding the fungal cells from the host immune cells and manipulating the responses of immune cells to C. neoformans invasion (May et al., 2016;Zaragoza et al., 2009). Melanin formation in C. neoformans protects fungal cells from many environmental stresses, such as UV radiation and free-radical toxicity (Wang & Casadevall, 1994;Williamson, 1997).
Of all-important metals in biology, magnesium (Mg) is one of the most abundant metals in the cell and it plays critical roles as an enzymatic cofactor and structural component of biological molecules.
Although Mg 2+ homeostasis plays vital functions in fungal biological processes, the role of Mg 2+ in regulating virulence factors in human fungal pathogens remains largely unknown. In this study, we examined the function of three C. neoformans Mg 2+ transporters with respect to the regulation of fungal virulence-associated factors. We demonstrated that a vacuolar Mg 2+ transporter (Mgt2) in C. neoformans positively regulated Mg 2+ delivery and melanin formation, but negatively controlled capsule production. Using transcriptome analysis, we showed that knockdown of MGT2 differentially regulated the expression of 118 genes. Of these, 19 genes were influenced by MGT2 knockdown independently of exogenous Mg. In animal infection models, the mgt2 mutant displayed an elevated fungal burden in lung tissues as a result of an enlarged capsular structure.
The transformant selection was performed using YPD agar supplemented with 200 μg/ml of G418. A final concentration of 50 mmol/L MgSO 4 was used for GAL7p-MGT2 cell growth.

| Plasmid construction and mutant construction
The generation of C. neoformans strains in which Mg transporter genes were disrupted was performed as follows. All primers used for the generation of mutants are listed in Table S1. For the targeted disruption of the MGT1 or MGT3 open reading frame, the upstream sequence of MGT1 or MGT3 was amplified using primer pair CDp280/ CDp281 or CDp288/CDp289, respectively. The PCR products were cloned between the ApaI and XmaI restriction sites in a NEO R plasmid.
The downstream sequence of MGT1 or MGT3 was amplified using primer pair CDp282/CDp283 or CDp290/CDp291, respectively. The resulting PCR products were cloned between the BamHI and XbaI restriction sites of the NEO R plasmid. The resulting plasmids were sequenced, and the inserts were amplified using primer pair CDp23/ CDp24 to generate the knockout integration cassette. The PCR products were purified, concentrated, and delivered using a PDS-1000/He System (Bio-rad). Transformants were picked and genomic DNA was isolated for PCR confirmation using primer pair CDp428/CDp131 for MGT1 and CDp962/CDp131 for MGT3.
The C. neoformans mgt2 mutant was constructed by replacing the MGT2 endogenous promoter with the GAL7 promoter. The upstream sequence from MGT2 was amplified using primers CDp284 and CDp285, and then this was cloned between the ApaI and XmaI restriction sites in a NEO R plasmid. The C. neoformans GAL7 promoter or MGT2 ORF (first ~1 kb of genomic DNA sequence) were amplified using primer pair CDp611/CDp612 or CDp613/CDp614, respectively. The resulting PCR fragments were purified and mixed, and overlapping PCRs were performed to join the GAL7 and MGT2 fragments using primers CDp611 and CDp614. The overlapping PCR product was purified and cloned between the SpeI and SacII restriction sites in the NEO R plasmid. The integration cassette was amplified using primers CDp23 and CDp24. The PCR fragment was transformed into C. neoformans by biolistic transformation. After 4 hr of recovery incubation, cells were transferred onto YPGal agar supplemented with G418. Transformants were confirmed by PCR using primers CDp632 and CDp131 and by Southern blotting.
For the construction of an overexpression strain of Mgt2, the C. neoformans TEF1 promoter was used to constitutively drive gene expression at the CMT2 locus. The upstream or downstream sequence of CMT2 was amplified from genomic DNA using primer pair CDp732/CDp733 or CDp738/CDp739, respectively. The upstream fragment was then cloned between the HindIII and EcoRI restriction sites in a NAT R plasmid, and the downstream fragment was cloned between the XbaI and SacII restriction sites in the same plasmid. The C. neoformans TEF1 promoter was amplified using primers CDp2204 and CDp2205, and the MGT2 ORF was amplified using primers CDp2206 and CDp2207. The fragments of the TEF1 promoter and the MGT2 ORF were joined by overlapping PCR using primers CDp2204 and CDp287. The resulting fragment was cloned at the XbaI site in the plasmid. The integration cassette was transformed using biolistic transformation. Successful integration at the CMT2 locus was confirmed using primers CDp861 and CDp131.
For the expression of S. cerevisiae ALR1 or MNR2 in the GAL7-MGT2 strain, a C. neoformans GPD1 expression plasmid (pXL-HYG) was used as described previously. The genomic DNA sequence from S. cerevisiae ALR1 or MNR2 was amplified using primer pair CDp1699/ CDp1700 or CDp1954/CDp1955, respectively. The resulting PCR products were cloned into the PacI site in the pXL-HYG plasmid. The plasmids with the correct gene orientation were transformed in the C. neoformans GAL7-MGT2 strain.

| Genomic DNA isolation and Southern blotting
Genomic DNA isolation and Southern blotting were performed as previously described (Ding & Butler, 2007). Briefly, 5 ml of cell culture (YPD medium or YPGal medium) was washed twice with water and then resuspended in lysis buffer (100 mM Tris-HCl pH 8.0, 50 mmol/L EDTA, 1% SDS, 100 mM NaCl, 2% Triton X-100). Cells were disrupted using a bead beater (Biospec). The supernatant was mixed with 7 mol/L ammonia acetate (pH 7.0). Genomic DNA was extracted using chloroform and precipitated with isopropanol. DNA pellets were then washed, dried, and resuspended in water. Next, 1 μg of DNA was used for PCR. Genomic DNA (20 μg) was treated with RNase A and digested with StuI and SacII overnight for Southern hybridization analysis. Southern blotting was performed using a DIG High Prime DNA Labeling and Detection Starter Kit II (Roche), and the experimental procedure was carried out according to the manufacturer's manual.
The DIG-labeled DNA probe was synthesized by PCR using primers CDp1045 and CDp612 (Table S1).

| RNA isolation, real-time PCR, and transcriptome analysis
RNA samples used for quantitative real-time PCR and transcriptome library construction were isolated using TRIzol reagent (Thermo Fisher Scientific) followed by Turbo DNase I treatment (Thermo Fisher Scientific) to eliminate DNA contamination. Next, 1 μg of total RNA was reverse-transcribed into cDNA using the Go Script Reverse Transcription System (Promega). Real-time PCR was performed using a CFX Connect Thermocycler (Bio-Rad). The data were analyzed using 2 −ΔΔCt method. ACT1 was used as a loading control. Primers are listed in Table S1. For RNA-seq analysis, 3 μg of total RNA were processed using the TruSeq RNA Sample Preparation Kit (Illumina, San Diego, CA, USA).
Next, mRNA was purified from total RNA using the polyT oligo-attached magnetic beads. An Illumina proprietary fragmentation buffer was used for mRNA fragmentation. First-strand cDNA was synthesized with random hexamer primers using SuperScript II followed by second-strand cDNA synthesis using RNase H and DNA polymerase I. Overhang sequences were then blunted and adenylated at the 3′ end of the cDNA sequence. Subsequently, cDNA sequences 200 bp in length were selected using the AMPure XP system (Beckman Coulter) and enriched using Illumina PCR Primer Cocktail in a 15-cycle PCR reaction. The PCR products were purified and their integrity was confirmed using the Agilent high sensitivity DNA assay on a Bioanalyzer 2100 (Agilent). The sequencing library was then sequenced on a Hiseq platform (Illumina) by Shanghai Personal Biotechnology Cp. Ltd. Sequenced reads were mapped using TopHat (http://ccb.jhu.edu/software/tophat/), and the transcript assembly was performed using Cufflinks (http://cufflinks. cbcb.umd.edu) (Trapnell et al., 2012). Genes with differential expression were analyzed using DESeq, as previously described (Chen et al., 2014). Genes of the mutant strains with fold changes greater or less than 1.5 in comparison to that of the wild-type strain were considered to be upregulated or downregulated, respectively.

| Microscopy
The capsule was observed using a Nikon Eclipse Ni-E microscope with the 100× objective (Nikon). Lung tissue mucicarmine and hematoxylin/eosin staining were visualized using a Leica DM4000B LED microscope (Leica) with the 20× objective. Protein localization of GFP-Mgt2 was performed by inducing cells in medium supplemented with galactose and 200 μM EDTA for 6 hr. The GFP signal was observed using a Leica TCS-SP8 confocal microscope under the 65× oil objective (Leica).

| Animal virulence assay, fungal burden, and histopathology
All animal experiments were reviewed and ethically approved by the Research Ethics Committees of the College of Life and Health Sciences of Northeastern University. The animal infection experiments were performed as previously described .
Briefly, 6-8-week-old female C57BL/6 mice were used for survival rate and fungal burden analysis, and they were purchased from Changsheng Biotech (China). For the survival rate assay, mice were anesthetized and intranasally administered 10 5 fungal cells in 50 μl of PBS. Animals were monitored twice a day for morbidity. For the fungal burden analysis, infected mice at 14 days post infection were killed by exposure to CO 2 . Lung and brain tissues were collected, weighed, and homogenized in 5 ml of PBS. The suspensions were diluted with PBS, and serial dilutions were plated onto YPGal agar at 30°C for 3 days. For the histopathological assay, lung tissues were collected at 14 days post infection and fixed with paraformaldehyde. Fixed lung tissues were frozen and processed using a Crypstat Microtome Leica CM1850 (Leica). Tissue sections (10 μm in thickness) were stained with mucicarmine or hematoxylin/eosin (Sigma).

| Melanin formation and capsule induction
Melanin was quantified using L-DOPA agar medium. Overnight cell cultures were diluted to an A 600 of 1.0 and spotted onto L-DOPA plates supplemented with various concentrations of MgSO 4 (0 g/L, 0.25 g/L, 0.5 g/L, or 1 g/L). The plates were incubated at 37°C until black pigment was observed. Photographs were then taken. The capsule induction was performed using two induction conditions. Cells were either incubated in DMEM supplemented with 10% FCS at 37°C and 5% CO 2 for 3 days, or in ACSF medium supplemented with 0.1% glucose at 37°C for 3 days (Rathore, Raman, & Ramakrishnan, 2016).
The capsule was stained with India ink.

| Atomic absorption spectroscopy
To quantify intracellular Mg concentrations, the wild-type and Gal7p-MGT2 cells were inoculated in YPGal medium overnight. The cultures were diluted to an A 600 of 0.2 in 10 ml of YPD medium and incubated on an orbital shaker at 30°C for 6 hr. Cells were then washed twice with 10 ml of 1 mmol/L EDTA and twice with deionized water to remove Mg ions from the medium and cell surface. The washed cells were collected in preweighed Eppendorf tubes, and the tubes were incubated at 80°C overnight to eliminate residual water. The dried cells were weighed. Next, 400 ml of 70% nitric acid was added to digest the cells at 100°C for 30 min. The resulting suspensions were diluted with 1% nitric acid. Mg concentrations were quantified using a JENA Atomic Absorption Spectroscopy ZEE nit 700P (JENA). The operation procedure was performed according to the manufacturer's manual. A standard curve was first constructed using an Mg atomic spectroscopy standard solution (Sigma). The Mg concentrations were then calculated using the standard curve.
These were re-named Mgt1, Mgt2, and Mgt3, respectively. The phylogenetic analysis showed that Mgt1 grouped with S. cerevisiae To determine the function of C. neoformans Mg 2+ transporters in maintaining Mg 2+ balance, Mg 2+ transporter knockout strains were generated using the targeted homologous integration method, to obtain mgt1Δ, mgt3Δ, and mgt1Δ mgt3Δ. In our hands, we could not disrupt MGT2 after multiple attempts, implying an essential function for MGT2 under normal yeast growth conditions. We therefore sub- Using the spotting assay, the GAL7p-MGT2 strain showed severe growth defects when glucose was added to suppress the activity of the galactose promoter in the medium. This observation was in agreement with the difficulties related to generating an mgt2Δ strain ( Figure 2a). The mgt1Δ, mgt3Δ, and mgt1Δ mgt3Δ strains had similar cell growth rates compared to the wild-type strain. The cell growth of the GAL7p-MGT2 strain was rescued when Mg was supplemented into the medium, but not with other metals (Figure 2a and b). Cumulatively, our data demonstrated that Mgt2 functions as a specific Mg transporter in C. neoformans. We speculated that the impairment in cell growth was due to an intracellular Mg deficiency. Therefore, atomic absorption spectroscopy was used to measure cellular Mg levels. Our data showed a 50% reduction in the Mg 2+ content in the GAL7p-MGT2 strain ( Figure 2c). These results strongly suggested that the Mgt2 protein is a critical C. neoformans Mg transporter involved in intracellular Mg accumulation.

| Mgt2 is an Mnr2 homolog
Our phylogenetic analysis demonstrated that Mgt2 is a putative Mg 2+ transporter and shares significant sequence similarity to Mnr2 from S. cerevisiae (Figure 1b). To assess whether Mgt2 functionally resembles Mnr2, a GAL7p-GFP-MGT2 strain was generated. This The genome of baker's yeast encodes two Mg importers, Alr1 and Alr2 (Da Costa et al., 2007;Graschopf et al., 2001). Mnr2 is a vacuolar Mg transporter that pumps Mg into the cytosol (Pisat et al., 2009). A Mg 2+ /H + exchanger also participates in Mg transport into the cytosol. Mg ions are delivered into the mitochondria via the Lpe10 and Mrs2 transporters (Gregan et al., 2001;Sponder et al., 2010). (b) Phylogenetic analysis of Mg transporters in S. cerevisiae and C. neoformans. A phylogenetic tree was generated as previously described (Ding et al., 2011). Numbers indicate the bootstrap resampling. Mg transporters in S. cerevisiae are labeled in black and those from C. neoformans are labeled in red. (c) Realtime PCR quantifications of MGT1, MGT2, and MGT3. Overnight C. neoformans cell cultures were diluted to an A 600 of 0.2 in YPD medium, and the cultures were incubated on an orbital shaker at 30°C until an A 600 of 1.0. A final concentration of 50 mmol/L of MgSO 4 or 200 μM EDTA were added to the cultures, and the incubations continued for 6 hr before RNA samples were isolated. Next, real-time PCR was performed. Expression levels were normalized to ACT1. Four biological samples were used for the analysis. Standard deviations were determined, and the Student's t-test was performed. (d). Scheme showing the construction of the GAL7p-MGT2 strain. The construction of the integration cassette harboring a GAL7 promoter is described in the Material and Methods. Briefly, the promoter and ORF fragments from MGT2 were amplified using PCR and cloned into a NEO R -GAL7p plasmid. The resulting plasmid was introduced into the wild-type strain using biolistic transformation. (e) Southern blot analysis of GAL7p-MGT2 strains. The Southern blot analysis was performed as described in the Material and Methods. A probe for the GAL7p promoter was used. The restriction enzymes StuI and SacII were used for genomic DNA digestion. To verify the correct integration of the GAL7p-MGT2 cassette, two GAL7p DNA species should be detected: one from the integrated cassette (resulting in a 1800-bp band) and the other from the wild-type GAL7p (resulting in a 2400-bp band). For the wild-type strain, only one band was detected at 2400 bp. Eight isolates (including two biological integrations) were used for the analysis. All isolates demonstrated correct integrations except isolate 3, which contained a random integration of the cassette in the genome We also observed GFP signal in unidentified intracellular spaces.
Because S. cerevisiae Mg 2+ transporters have been shown to localize to the plasma membrane, vacuolar membrane, and mitochondria (Bui et al., 1999;Cui et al., 2015;Pisat et al., 2009), we employed a mitochondrial staining method using MitoTracker to ascertain whether Mgt2 was a mitochondrial-localized Mg 2+ transporter in C. neoformans. However, no colocalization of Mgt2 and mitochondria was observed, indicating that Mgt2 is not a mitochondrialassociated Mg 2+ transporter (Figure 3b). A vacuolar specific dye,  Overnight galactose liquid cell cultures of the wild-type H99, mgt1Δ, GAL7p-MGT2, mgt3Δ, and mgt1Δ mgt3Δ were diluted in PBS to an A 600 of 1.0. Ten-fold serial cell dilutions were spotted onto YPGal agar, YPD agar, or YPD agar supplemented with 1 mmol/L, 5 mmol/L, 10 mmol/L, or 50 mmol/L MgSO 4 . Plates were incubated at 30°C for 3 days before photographs were taken. (b) Cell growth assay of Mg transporter mutants following exposure to other metals. The spotting assay was performed as described in Figure 2a, except YPD agar supplemented with 50 mmol/L Ca, 200 μmol/L Fe, 1 mmol/L Cu, and 200 μmol/L Zn was used. (C) Atomic absorption spectroscopy analysis of Mg transporter mutants. Overnight galactose liquid cell cultures of the wild-type H99, GAL7p-MGT2, and mgt1Δ mgt3Δ strains were diluted in YPD medium to an A 600 of 1.0, and the cultures were incubated for 6 hr. Cells were collected, washed, and placed in preweighed Eppendorf tubes. Tubes were then dried in an oven overnight. Atomic absorption spectroscopy was performed as described in the Material and Methods. Relative Mg concentrations were calculated. Four biological samples were used for the analysis. Standard deviations were determined, and the Student's t-test was performed. Triple asterisks represent p < .001 and NS indicates no statistical significance

| Mgt2 modulates gene expression independent of Mg
The mgt2 knockout strain growth phenotype was complemented when exogenous Mg was supplemented in the growth medium. However, cell growth in the presence of Mg was only partially rescued, even at a very high level of Mg 2+ (Figure 2a). This growth phenotype suggests that Mgt2 influences normal cell growth via a Mg 2+ -independent pathway. To explore this hypothesis, we performed a high-throughput transcriptome analysis using RNA sequencing (RNA-seq). We com-  (Table S2). Among all differentially expressed genes, expression of MGT2 was the most significantly downregulated in the GAL7p-MGT2 cells, demonstrating the accuracy of our RNA-seq data (Table 1, Table S2 and S3). We observed that genes encoding deacetylase, mitochondrial alternative oxidase, alternative cyclin Pcl12, Fe-S cluster assembly factor, and proteins involved in sugar metabolism and transport were significantly downregulated. Genes encoding ferric-chelate reductase, sulfite transporter, septin ring protein, and glucose-methanol-choline oxidoreductase were significantly upregulated (Table S2). On the other hand, supplementing Mg into the medium of GAL7p-MGT2 cells only caused 27 genes to be differentially expressed. Of these, expression of 7 genes and 20 genes was up or downregulated, respectively (Table S3). We detected that genes encoding alcohol dehydrogenase, alternative cyclin pcl12, carnitine Overnight YPGal cultures were diluted to an A 600 of 1.0 in YPD medium. The YPD cultures were incubated at 30°C for 3 hr to repress the activity of the GAL7 promoter in the GAL7p-MGT2 strain. For the GAL7p-MGT2 strain, the YPD culture was divided in two, and MgSO 4 was added into one portion at a final concentration of 50 mmol/L. All cultures were incubated for additional 4 hr before RNA was isolated. Three biological replicates were carried out. B Gene ontology analysis. Gene ontology was performed using TopGO (Huber et al., 2015). GO terms were downloaded at http://www.bioconductor.org/packages/release/bioc/html/topGO.html. Significantly altered GO terms (classical fisher test, p < .05) are shown on the plot. Red bars indicate upregulation in the GAL7p-MGT2 strain and green bars indicate downregulation in the GAL7p-MGT2 strain. Hash signs show the unchanged GO terms after Mg treatment. (c) Venn diagram for transcriptomic analysis. Repressing MGT2 causes 118 genes to be differentially expressed compared to the wild-type strain H99. Repressing MGT2 in the presence of exogenous Mg causes 27 genes to be differentially expressed compared to the wild-type strain H99. Nineteen genes demonstrated similar expression dynamics in both growth conditions. (d) Real-time PCR confirmation of transcriptomic data. Cells were grown as described in Figure 4a, and RNA was isolated, digested with DNase I, and reverse-transcribed. Random genes were selected from 19 overlapping genes. Primers are listed in Table S1. Actin was used as a normalization control. Three biological replicates were performed. The Student's t-test was performed to determine statistical significance (***p < .005, **p < .01, *p < .05). transport, single-organism localization, and organic substance transport ( Figure 4b). Two processes were differentially regulated upon the Mg 2+ treatment, and these GO terms were enriched in monocarboxylic acid catabolic process and small molecule catabolic process.
The expression of 19 genes remained constant in GAL7p-MGT2 cells in YPD in the presence or absence of exogenous Mg (Figure 4b and Table 1), and this included MGT2 itself. Real-time PCR was performed for five randomly selected putative target genes of the 19-gene cohort to evaluate the RNA-seq data, and expression regulation patterns resembled those obtained in the RNA-seq analysis (Figure 4c).

| Mgt2 influences C. neoformans survival in the lungs
Because Mg 2+ is an essential metal and the mgt2 mutant showed striking growth defects, we hypothesized that Mgt2 participates in regulating fungal virulence in mice. C57BL/J mice were infected with mgt1Δ mgt3Δ or GAL7p-MGT2 strains (precultured in YPD medium), and the survival rates were recorded (Figure 5a). Mice infected with mgt1Δ mgt3Δ had similar death rates as the wild-type. Surprisingly, mice infected with two independent GAL7p-MGT2 strains had earlier death rates, but these data displayed poor reproducibility (Figure 5b). In the colony-forming unit (CFU) analysis, we detected increased levels of fungal proliferation in two independent GAL7p-MGT2 strains in the lungs (Figure 5c). This was subsequently recapitulated by the lung tissue mucicarmine stain in which more GAL7p-MGT2 cells were observed in lung tissues compared to the lung tissue infected with the wild-type strain (Figure 5f), with no significant changes observed for lung mass (data not shown). Furthermore, using the murine macrophage cell line RAW264.7, we showed that the mgt2 mutant inhibited the rate of phagocytisis, compared to the wild-type cells (Figure 5e), suggesting that Mgt2 may participate in capsule formation (discussed below). A comparison of the CFUs from brain tissues between H99 and GAL7p-MGT2 cells showed no significant difference (Figure 5d).
To gain further insights into the mechanism by which Mgt2 influences fungal proliferation in the lung, we analyzed melanin and capsule formation in the GAL7p-MGT2 strains. As shown in Figure 6a, melanin formation in the wild-type cells required Mg 2+ . The GAL7p-MGT2 strains demonstrated severe melanin production defects. The addition of Mg 2+ only partially rescued the pigmentation defects. We also tested the mgt1Δ mgt3Δ strain, and it showed similar melanin production deficits when compared with the wild-type strain (Figure 6a).
Our data demonstrated that Mg 2+ is a critical element for melanin formation in C. neoformans, and Mgt2 is required.
Next, we examined capsule formation in Mg 2+ transporter mutants in DMEM (Dulbecco's Modified Eagle Medium, high glucose) supplemented with 10% fetal calf serum (FCS) (Figure 6b). We observed a modest increase in capsule thickness in the GAL7p-MGT2 strain T A B L E 1 RNA-seq analysis of differentially expressed genes in the GAL7-MGT2 strain supplemented with or without MgSO 4

Gene ID
Log 2 (MGT2 -v-H99) a Log 2 (MGT2 + Mg -v-H99)  RNA-seq analysis was performed and differential gene expression was compared between the wild-type and GAL7-MGT2 strains. Differential gene expression is represented by binary log of the fold change. Three biological replicates were performed. The full gene list is available in Table S2. b RNA-seq analysis was performed and differential gene expression was compared between the wild-type and GAL7-MGT2 strains (supplemented with 50 mM MgSO 4 ). Differential gene expression is represented by binary log of the fold change. Three biological replicates were performed. The full gene list is available in Table S3. compared to the wild-type strain (Figure 6c). The addition of Mg 2+ reduced the size of the capsule structure. To extend these observations, we constructed an Mgt2 overexpression strain in which MGT2 expression is under the control of the constitutive promoter, TEF1 (Figure 6d).
To avoid random integration in the genome, the construct was designed to target the metallothionein 2 locus (CMT2) because previous studies demonstrated that the disruption of CMT2 had no effect on the expression of virulence factors (Ding et al., 2013). The expression of MGT2 was measured using real-time PCR, which demonstrated a 15-fold increase in gene expression compared to that of the wild-type strain (Figure 6e).  Samanovic, Ding, Thiele, & Darwin, 2012). Fe has been widely accepted as a key regulator of pathogenicity in major human fungal pathogens (Bairwa et al., 2017;Ding et al., 2014;Jung et al., 2006;Lian et al., 2005;Nyhus et al., 1997;Saikia, Oliveira, Hu, & Kronstad, 2014). In C. neoformans, mutations in genes associated with Fe acquisition often lead to dramatic attenuations in fungal virulence. A C. neoformans GATA-type transcription factor, Cir1, is required for the regulation of expression of genes that are essential for Fe and Cu homeostasis. Moreover, important virulence factor genes, such as genes involved in the capsule and melanin biosynthesis processes, are also controlled by Cir1 (Jung et al., 2006).
Using metabolomic analysis, it was demonstrated that Cir1 modulates the glycolytic pathway, ergosterol processes, and inositol metabolism.
Numerous studies suggest that metal acquisition machinery has an essential function in C. neoformans virulence. Indeed, disrupting the high-affinity Fe 2+ permease Cft1 in C. neoformans significantly attenuates fungal virulence and fungal burden in animal models as a result of the poor adaptation of mutant cells during Fe limitation in the host.
However, the disruption of Cft2 has no effect on fungal proliferation in the host (Han, Do, & Jung, 2012;Kim et al., 2012). Cu has been suggested to be an essential metal for C. neoformans because of its function in melanin formation, capsule production, respiration, and as an antioxidant agent cofactor (Chun & Madhani, 2010;Ding et al., 2013;Jiang et al., 2009;Mauch, Cunha, & Dias, 2013;Walton et al., 2005;Waterman et al., 2007Waterman et al., , 2012Zhang et al., 2016). Recent studies of Cu homeostasis focused on two Cu + transporters, CTR1 and CTR4. It has been shown using two versions of a gene knockout system that a ctr4 mutant demonstrates high fungal burden in the lungs Zhang et al., 2016). Meanwhile, the ctr1 and ctr4 double mutant has melanin defects, and this strain shows an increased fungal burden and brain dissemination .
Mg 2+ is one of the most abundant divalent metal elements required for life, but its potential function in regulating fungal pathogenicity remains uninvestigated. Previous in vitro studies demonstrated that Mg 2+ is an important element in the regulation of biofilm formation and capsule production in C. neoformans (Robertson, Wolf, & Casadevall, 2012 Perez- Martin & Castillo-Lluva, 2008). Therefore, it is most likely that the growth defects of the GAL7-MGT2 strain after Mg treatment is due to the inhibition of gene expression of cell proliferation regulators.
Despite a previous study using semiquantitative RT-PCR show-ing that Mg 2+ influences the expression of genes involved in capsule production, we do not detect capsule-related RNA expression differences in our RNA-seq analysis. This is probably because different strain isolates were used in each study, causing different responses to metal ions Waterman et al., 2012;Zhang et al., 2016). In fact, C. neoformans is a fast-evolving human fungal patho-

CONFLICT OF INTEREST
None declared.