Calcium Binding Protein Ncs1 Is Calcineurin Regulated in Cryptococcus neoformans and Essential for Cell Division and Virulence

Cryptococcus neoformans is the major cause of fungal meningitis in HIV-infected patients. Several studies have highlighted the important contributions of Ca2+ signaling and homeostasis to the virulence of C. neoformans. Here, we identify the cryptococcal ortholog of neuronal calcium sensor 1 (Ncs1) and demonstrate its role in Ca2+ homeostasis, bud emergence, cell cycle progression, and virulence. We also show that Ncs1 function is regulated by the calcineurin/Crz1 signaling cascade. Our work provides evidence of a link between Ca2+ homeostasis and cell cycle progression in C. neoformans.

neuronal calcium sensor 1 (Ncs1) homolog in C. neoformans, given that this protein is important for Ca 2ϩ regulated processes in a variety of eukaryotic cells (43).
For this purpose, we performed an in silico analysis at FungiDB to identify the NCS1 coding sequence in the C. neoformans H99 genome (CNAG_03370). Ncs1 is well conserved in eukaryotes, with orthologs sharing common regions, such as EF-hand domains and a myristoylation motif. Our analysis revealed that C. neoformans Ncs1 contains four predicted EF-hand domains that span the full length of the protein (Fig. 1). Moreover, the presence of an N-terminal myristoylation motif was identified using the NMT-themyr predictor database. Myristoylation, a lipid modification conserved among eukaryotic Ncs1 proteins (44), is important for cell signaling, proteinprotein interaction, and protein targeting to endomembrane systems and the plasma membrane (45). Comparative analysis of the C. neoformans Ncs1 protein sequence with the products encoded by Aspergillus fumigatus NCSA (Afu6g14240), Schizosaccharomyces pombe NCS1 (SPAC18B11.04), and S. cerevisiae FRQ1 (YDR373W), which are already functionally characterized (34,40,41), revealed high degrees of amino acid sequence similarity (86, 87, and 81%, respectively) ( Fig. 1).
Disruption of the NCS1 gene affects C. neoformans traits associated with calcium homeostasis. Calcium sensor proteins measure fluctuations in free cytosolic Ca 2ϩ and transduce the signal to downstream effectors (41,42,46). To determine whether Ncs1 plays a similar role in C. neoformans, we obtained a NCS1 gene knockout strain (ncs1⌬) from Madhani's mutant collection (47) and generated an NCS1 reconstituted (ncs1⌬::NCS1) strain (see Fig. S1 in the supplemental material) using the ncs1⌬ background. We then evaluated the ability of these mutant strains to grow under different stresses. We initially chose high Ca 2ϩ concentration (to alter Ca 2ϩ homeostasis) and high temperatures (37°C and 39°C), as the calcineurin (cna1⌬) and calmodulin (cam1⌬) mutants were shown to be sensitive under these growth conditions (24,25,42). We observed impaired ncs1⌬ strain growth in high Ca 2ϩ levels and at 39°C but not at 37°C; these growth defects were restored to wild-type (WT) levels in the ncs1⌬::NCS1 strain (Fig. 2). Lower Ca 2ϩ concentrations (ranging from 1 to 20 mM CaCl 2 ) did not influence ncs1⌬ strain growth (data not shown). Other traits associated with the

FIG 2
Disruption of NCS1 leads to stress sensitivity in C. neoformans. Spot plate assays of the WT, ncs1⌬ mutant, and ncs1⌬::NCS1 complemented cells were performed on YPD agar. The plates were incubated at 30°C (control for normal growth), 37°C, or 39°C and under stress induced by Ca 2ϩ (200 mM or 300 mM CaCl 2 at 30°C). The calcineurin (cna1Δ) and calmodulin (cam1⌬) mutants were included as controls, separated by a thin white line that indicates noncontiguous portions of the same image. All assays were conducted for 48 h. Ca 2ϩ -calcineurin pathway, such as growth in the presence of cell wall-perturbing agents (calcofluor white and Congo red) and osmotic stress (1 M NaCl) were evaluated in the ncs1Δ null mutant, with no effect observed (Fig. S2). We also evaluated whether the level of free intracellular Ca 2ϩ in C. neoformans is affected in the absence of Ncs1. Relative to the WT strain, the ncs1⌬ mutant had a higher basal level of free cytosolic Ca 2ϩ , which was reduced to WT levels in the ncs1⌬::NCS1 strain. This high-Ca 2ϩ -level phenotype was shared with that observed for the cna1⌬ and cam1⌬ mutant strains (Fig. 3A). In S. pombe, Ncs1 physically interacts with the Mid1 ortholog, Yam8, which is a stretch-activated Ca 2ϩ channel. S. pombe YAM8 gene disruption in the ncs1⌬ background restored the Ca 2ϩ -sensitive phenotype (35). In this study, the authors proposed that Ncs1 negatively regulates the Yam8 calcium channel. We therefore investigated whether the high-affinity Mid1-Cch1 calcium channel complex (48,49) is responsible for the increased intracellular Ca 2ϩ observed in the ncs1⌬ mutant. Specifically, we used real-time reverse transcriptionquantitative PCR (RT-qPCR) to compare the expressions of MID1 and CCH1 in WT and ncs1⌬ grown in yeast extract-peptone-dextrose (YPD) with or without 100 mM CaCl 2 for 24 h. The transcript levels of both genes increased by approximately 3-fold in the ncs1⌬ mutant strain, but only following growth in the presence of 100 mM CaCl 2 (Fig. 3B). No differences in CCH1 and MID1 expression was observed in the WT and the ncs1⌬ strain grown in YPD without CaCl 2 (Fig. 3B), suggesting that the Mid1-Cch1 complex, which imports Ca 2ϩ into the cytosol (29,48), is not the source of extra Ca 2ϩ in the ncs1⌬ mutant. In further support of this, we generated a mid1⌬ ncs1⌬ double mutant in C. neoformans and found that increased intracellular Ca 2ϩ accumulation and Ca 2ϩ sensitivity persisted in this mutant (Fig. S3). Despite these findings, we cannot rule out the involvement of other low-affinity calcium channels in contributing to the increased intracellular Ca 2ϩ observed in the ncs1⌬ mutant.
We also C-terminally tagged Ncs1 with green fluorescent protein (GFP) (NCS1::GFP) to assess Ncs1 subcellular localization. Faint, predominantly cytosolic, Ncs1 fluorescence was observed when the strain was cultured in the absence of Ca 2ϩ . However, the fluorescence was higher than that observed for the nonfluorescent WT control strain (Fig. 3C). Ncs1 fluorescence became more intense following culture in the presence of Ca 2ϩ (100 mM CaCl 2 ), with Ncs1 adopting a more punctate staining pattern: 24.5% Ϯ 1.1% and 36.3% Ϯ 3.0% of the cell population displayed puncta in the absence (MM) and presence (MM ϩ CaCl 2 ) of Ca 2ϩ , respectively (P ϭ 0.013, Welch's test with Ն200 cells per sample) (Fig. 3C). Increased Ncs1 fluorescence in the presence of Ca 2ϩ correlated with higher expression of NCS1 by the NCS1::GFP strain under the same condition (Fig. 3D). Taken together, these results suggest that Ncs1 responds to increase in intracellular Ca 2ϩ levels and participates in the regulation of calcium homeostasis in C. neoformans.
NCS1 is a calcineurin-Crz1 responsive gene. Given that NCS1 is a Ca 2ϩ -responsive gene in C. neoformans (Fig. 3D), we investigated whether NCS1 expression is regulated by the calcineurin signaling pathway via the transcription factor Crz1. NCS1 expression was analyzed in the presence and absence of the calcineurin inhibitor FK506 (Fig. 4A) and in the WT and crz1Δ mutant (Fig. 4B). The results demonstrated that FK506 treatment reduced NCS1 transcription in the WT (Fig. 4A) and that NCS1 expression was downregulated in the crz1Δ mutant at 25°C and 37°C (Fig. 4B). In further support of NCS1 being a Crz1 target, we identified two Crz1-binding consensus motifs (50) in the putative NCS1 regulatory region encompassing the 1,000-nucleotide sequence upstream of the transcription start site (Fig. 4C). These findings provide evidence that Ncs1 and calcineurin work together to regulate Ca 2ϩ homeostasis.
Ncs1 activity is essential for C. neoformans virulence. As proven in other studies, the disruption of Ca 2ϩ homeostasis components is important for cryptococcal pathogenicity (29)(30)(31)(32)(33)51). To determine whether disruption of Ncs1-mediated calcium homeostasis also contributes to pathogenicity, we compared the virulence of the ncs1⌬ mutant strain to that observed for the WT and ncs1⌬::NCS1 strains in a mouse inhalation . Mean values were compared using one-way ANOVA and Dunnett's post hoc test. Statistical significance is represented as follows: ****, P Ͻ 0.0001, and *, P Ͻ 0.05. (B) The transcript levels of genes encoding the calcium transporters, CCH1 and MID1, were evaluated using RT-qPCR. The WT and ncs1⌬ strains (10 6 cells/ml) were incubated in YPD for 16 h with shaking, either at 37°C (control) or at 37°C supplemented with Ca 2ϩ (100 mM CaCl 2 ). RNA was extracted and cDNA synthesized. Each bar represents the mean Ϯ the standard deviation (n ϭ 3) for each gene in each strain normalized to actin. Statistical analysis was performed using Student's t test (*, P Ͻ 0.05). (C) Ncs1 was tagged with GFP (NCS1::GFP), and the effect of CaCl 2 supplementation on Ncs1 production and subcellular localization was assessed by fluorescence microscopy. YPD overnight cultures of the WT (autofluorescence background control) and the NCS1::GFP strain were washed twice with water and used to seed on minimal medium (MM) or MM supplemented with Ca 2ϩ (100 mM CaCl 2 ) at an OD 600 of 1. The cultures were further incubated for 4 h at 30°C prior to visualization. DIC and green fluorescent images are included. (D) The cultures prepared for panel C were also used to extract RNA and perform RT-qPCR to assess the effect of Ca 2ϩ on the transcript levels of NCS1, normalized to actin. Statistical analysis was performed using one-way ANOVA with Tukey post hoc test. Comparisons were conducted between WT cells grown in the absence or in the presence of Ca 2ϩ or between NCS1::GFP cells grown in the absence or in the presence of Ca 2ϩ . ****, P Ͻ 0.0001. model of cryptococcosis. In a Kaplan-Meier survival study, the ncs1⌬ null mutant strain was found to be hypovirulent (median lethal time [LT 50 ], 32.7 days) compared to the WT (LT 50 , 18.9; P Ͻ 0.0001) and the ncs1⌬::NCS1 strain (LT 50 , 17.4 days; P Ͻ 0.0001) (Fig. 5A). Although the disruption of NCS1 prolonged mouse survival, no difference in the fungal burdens in lung and brain were observed at time of death, when infected mice had lost 20% of their preinfection weight (Fig. 5B). Thus, the ncs1⌬ null mutant strain is capable of infecting the lung and brain tissue but potentially grows at a lower rate than the WT and the ncs1⌬::NCS1 strains.
Ncs1 is necessary for growth under host-mimicking conditions. We also analyzed the capability of ncs1⌬ mutant to synthesize the polysaccharide capsule, since this is the main cryptococcal virulence factor (1,12). We observed that when the ncs1⌬ strain was grown under capsule-inducing conditions (Dulbecco modified Eagle medium [DMEM] at 37°C and 5% CO 2 ), mutant cells produced smaller capsules than the WT and ncs1⌬::NCS1 strain (Fig. 6A). However, capsule size was not affected following growth in mouse serum (data not shown). Next, we compared growth of the ncs1⌬ mutant to that of the WT and ncs1⌬::NCS1 strains under the capsule induction condition utilized and found that the null mutant growth was drastically compromised (Fig. 6B). Similarly, growth of the ncs1⌬ strain was severely impaired in mouse serum over a 24-h period at 37°C with 5% CO 2 (Fig. 6C). Collectively, these results suggest that hypovirulence of the ncs1⌬ mutant is most likely associated with the observed growth defects, with reduced capsule size making only a minor contribution to this virulence phenotype.
Ncs1 is important for the release of daughter cells. Microscopic analysis to evaluate the size of the polysaccharide capsule and the growth rate in mouse serum revealed that some ncs1⌬ cells displayed aberrant morphology and cell division (Fig. 7A), suggesting that Ncs1 could play a role in cell cycle progression. We therefore investigated the growth defect further by determining the time it took for buds to emerge using time-lapse microscopy ( Fig. 7B and Movies S1 and S2). Given that the mutant was severely attenuated in growth when cultured in DMEM or exposed to mouse serum, we chose YPD medium for this analysis, as it is a richer medium in which mutant growth is not as compromised. To avoid bias due to lack of synchronization, we only measured the time of bud emergence in cells after the bud of the first daughter cell had separated from the mother or, in the case of the mutant cells, where progeny did not detach from mother cell, after the second bud emergence. The results demonstrate that it took ϳ70 min for buds to emerge in the WT cells and more than 140 min for buds to emerge in isolated and clumped ncs1Δ mutant cells (Fig. 7C). Furthermore, buds were slow to be released in some ncs1Δ mutant cells, resulting in more extensive cell clumping.
As cell division is linked to the cell cycle, we evaluated whether cells lacking NCS1 displayed defects in cell cycle regulation by measuring the levels of two transcripts associated with different stages of the cell cycle: the G 1 cyclin encoded by CNL1 (52) and the S phase DNA replication licensing factor encoded by MCM2 (53,54). We also measured the transcript levels of the G protein-coupled receptor encoded by GPA2, which displays oscillatory expression during the cell cycle (53,54). All three genes were upregulated in the ncs1⌬ strain compared to the WT after 4 h of growth in YPD (Fig. 7D), reinforcing that cell cycle progression is altered in the ncs1⌬ mutant strain. were infected with 500,000 cells of the WT, ncs1⌬, or ncs1⌬::NCS1 strain. Mice were monitored daily and euthanized by CO 2 asphyxiation when they had lost 20% of their preinfection weight. (A) Median mouse survival differences were estimated using a Kaplan-Meier log-rank Mantel-Cox test. The increase in median survival of ncs1⌬-infected mice relative to the other two infection groups was statistically significant (P Ͻ 0.0001). (B) Lungs, brain, and spleen were removed posteuthanasia, weighed, homogenized, serially diluted, and plated onto Sabouraud dextrose agar plates to determine fungal burden by quantitative culture (CFU) following 3 days of growth at 30°C. CFU were adjusted to reflect CFU/gram of tissue and CFU/milliliter of blood (normalized CFU). Statistical significance was determined using one-way ANOVA. However, no differences in organ burden were found.

FIG 6
Ncs1 is necessary for growth under host-mimicking conditions. (A) Capsule sizes of WT, ncs1⌬, and ncs1⌬::NCS1 cells were determined following incubation in capsule-inducing medium (DMEM) for 72 h (37°C and 5% CO 2 ). Capsules were visualized by India ink staining and light microscopy, and measurements were performed using ImageJ software for at least 50 cells of each strain. Relative capsule size was defined as the distance between the cell wall and the capsule outer border by cell diameter. Statistical analysis was performed using one-way ANOVA, with Tukey post hoc test. **** P Ͻ 0.0001, and ***, P Ͻ 0.001, compared to the WT. (B) Growth of the WT, ncs1⌬, and ncs1⌬::NCS1 cells in DMEM (37°C and 5% CO 2 ) for 24 or 48 h was assessed by quantitative culture (CFU). The results represent the mean Ϯ standard deviation (three biological replicates) of each strain normalized to the CFU of the inoculum, described as fold change. Statistical analysis was performed using one-way ANOVA with Dunnett's post hoc test. Significant differences compared to WT are marked (****, P Ͻ 0.0001). (C) Growth of the WT, ncs1⌬, and ncs1⌬::NCS1 cells for 24 h at 37°C 5% CO 2 in heat-inactivated mouse serum was by quantitated (CFU). The results are expressed as a fold change relative to the initial inoculum (10 4 cells/ml) and represent the means Ϯ standard deviations (three biological replicates). Statistical analysis was performed using one-way ANOVA and Dunnett's post hoc test (*, P Ͻ 0.05, and ****, P Ͻ 0.0001, relative to the WT).

DISCUSSION
Our results indicate that NCS1 expression in C. neoformans is regulated by Ca 2ϩ and the calcineurin/Crz1 pathway and corroborate findings on the Ncs1 ortholog in fission yeast (35). In contrast to our conclusions and those made in studies using S. pombe, the S. cerevisiae Ncs1 ortholog, Frq1, was found to be essential for viability and the level of FRQ1 expression was not influenced by the calcineurin/Crz1 pathway, as revealed by microarray analysis (55). This suggests that distinct calcium sensing mechanisms exist in fungal species despite widespread functional conservation of Ncs1 and other regulators of Ca 2ϩ homeostasis.
An interesting feature of the ncs1⌬ mutant is its attenuated virulence in a murine model of cryptococcosis. In contrast, attenuated virulence was not observed for the null Ncs1 ortholog mutant (NCSA) in A. fumigatus (41), reaffirming that processes regulated by Ncs1 orthologs in pathogenic fungi differ or that other genes can compensate in the absence of NCSA. Moreover, the cryptococcal ncs1⌬ strain took longer to achieve the growth densities associated with debilitating infection in the tissues of WT-infected mice. This slower growth phenotype in vivo correlated with the reduced rate of WT and ncs1⌬ cells were grown in minimal medium for 72 h at 37°C and 5% CO 2 (upper panel) or in heat-inactivated mouse serum for 24 h at 37°C and 5% CO 2 (lower panel), stained with India ink, and visualized by light microscopy. (B and C) Fungal cells were incubated in YPD medium for 16 h inside a chamber coupled to a confocal microscope (37°C and 5% CO 2 ), and bud emergence time was recorded using time-lapse microscopy. Time measurements were initiated after the first round of bud emergence to avoid errors associated with the lack of synchronization. Images were acquired every 30 s. The graph in panel C represents the mean time for buds to emerge (minutes) Ϯ standard deviation of at least 15 cells per strain. Statistical analysis was performed using the nonparametric Mann-Whitney test (***, P Ͻ 0.0001). (D) Transcript levels of genes encoding cell cycle regulators were assessed in WT and ncs1⌬ cells by RT-qPCR. Cells were grown in YPD at 37°C for 4 h. Results represent the mean transcript levels Ϯ standard deviations (three biological triplicates) with each gene normalized to ACT1 transcript levels. Statistical analysis was performed using Student's t test (**, P Ͻ 0.01).
proliferation of the ncs1⌬ strain in mouse serum and impaired bud emergence and release. These results confirm that Ncs1 is important for fungal adaptation to the host environment and for the establishment of disease and reaffirm the importance of Ca 2ϩ homeostasis and Ca 2ϩ signaling in cryptococcal virulence. Our findings also extend the set of calcium-related genes involved in virulence to include Ncs1.
Given that expression of ϳ40 virulence-associated genes is linked to the cell cycle in C. neoformans, the control of this process is fundamental to disease progression (53). In S. cerevisiae, Ca 2ϩ homeostasis is linked to cell cycle regulation, as a decrease in intracellular Ca 2ϩ leads to transient arrest in the G 1 phase, followed by interruption in the G 2 /M phase (27,(56)(57)(58). Moreover, bud emergence and the cell cycle depend on calcineurin activity, which regulates the availability of proteins involved in cell cycle regulation. These proteins include Swe1, a negative regulator of Cdc28/Clb complex, Cln2, a protein kinase required for cell cycle progression, and a G 2 cyclin (27). In this context, we speculate a potential role for calcineurin signaling in cell cycle regulation in C. neoformans. Our new data demonstrate that Ncs1 interferes with the transcription profile of genes associated with cell cycle progression (CLN1, GPA2, and MCM2). Notably, overexpression of the S. cerevisiae G 1 /S cyclin, Cln1, led to a filamentation phenotype (59). In our study, we showed that the cryptococcal ncs1Δ mutant was impaired in bud emergence and release, as seen in time-lapse microscopy. Furthermore, the C. neoformans cln1⌬ mutant exhibited aberrant bud emergence and cell division and a consistent delay in budding (60), which are phenotypes also observed in cryptococcal ncs1⌬ mutant cells. Additional experiments to confirm the impact of Ca 2ϩ homeostasis on cryptococcal cell cycle regulation are necessary to support this hypothesis.
We hypothesize that Ca 2ϩ excess, or even other types of stress, leads to activation of the calcineurin pathway, which ultimately drives the expression of Ncs1 in a Crz1-dependent fashion. Therefore, Ca 2ϩ -activated Ncs1 would participate in a diverse array of cellular processes to cope with Ca 2ϩ excess, including the regulation of cell division via its potential association with Pik1, a protein implicated in cell septation in fission yeast (61). Two lines of evidences support this hypothesis: (i) yeast Pik1 forms puncta consistent with its localization in the Golgi apparatus (62) and we observed that cryptococcal Ncs1 also forms puncta, particularly when Ca 2ϩ is present, and (ii) yeast Frq1 physically interacts with Pik1 (62). Structural studies performed on Ncs1 revealed that the myristoyl group flips out following Ca 2ϩ binding, allowing Ncs1 to anchor reversibly to membranes (20,39,40). Thus, it is possible that Ca 2ϩ binding to Ncs1 exposes the hydrophobic N-myristoylation domain, promoting Ncs1 association with Pik1 in Golgi membranes and, hence, proper cell septation and division. Further experiments are required to confirm that the puncta assumed by Ncs1 upon addition of calcium colocalize with the Golgi.
In summary, we have characterized the Ncs1 homolog in C. neoformans, demonstrating its importance in Ca 2ϩ homeostasis and virulence. We showed that in contrast to S. cerevisiae, NCS1 is a calcineurin-responsive gene in C. neoformans, with calcineurin and Ncs1 working together to regulate calcium homeostasis and, hence, promote fungal growth and virulence. To our knowledge, this is the first report of a role for Ncs1 in fungal virulence using a mammalian infection model and of a potential correlation between Ca 2ϩ signaling and cell cycle progression in C. neoformans.

MATERIALS AND METHODS
Fungal strains and media. C. neoformans serotype A strain Kn99 was chosen to conduct the study as the wild type (WT). The NCS1 gene (CNAG_03370) deletion mutant (ncs1Δ), cna1⌬ mutant, and cam1⌬ mutant were obtained from H. Madhani's library (47). The ncs1Δ reconstituted strain (ncs1⌬::NCS1), the mid1Δ ncs1Δ double mutant, and the NCS1::GFP strain were all constructed using overlapping PCR as previously described (63), and site-directed homologous recombination was performed. Transformation was carried out using biolistic transformation, as previously described (64). The primer list is presented at Table S1, and the confirmations of the cassette's insertions are demonstrated in Fig. S1. Fungal cells were maintained on solid YPD medium (1% yeast extract, 2% peptone, 2% dextrose, and 1.5% agar). YPD plates containing hygromycin (200 g/ml) or G418 (100 g/ml) were used to select C. neoformans transformants.
Grande do Sul (UFRGS). Statistical analysis was done by timing how long each mother cell took to originate a bud. Measurements were performed at the beginning of the second budding in order to avoid errors associated with the lack of tools to synchronize cells.
RT-qPCR analysis. For gene expression analysis, strains were subjected to different conditions, as described in the figure legends. RT-qPCR technique was performed for all experiments as follows. Cryptococcal cells were washed once with PBS, then frozen in liquid nitrogen, and lyophilized. Cell lysis was performed by vortexing the tubes with the dry pelleted cells using acid-washed glass beads (Sigma-Aldrich Co., St. Louis, MO). Three independent sets of RNA samples for each strain were prepared using TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's protocol. Next, RNA samples were treated with DNase (Promega, Madison, WI), and a total of 300 ng of treated-RNA was used for reverse transcription with ImProm-II reverse transcriptase (Promega). RT-qPCR was performed on a real-time PCR StepOne real-time PCR system (Applied Biosystems, Foster City, CA). PCR thermal cycling conditions had an initial step at 94°C for 5 min, followed by 40 cycles at 94°C for 30 s, 60°C for 30 s, and 72°C for 60 s. Platinum SYBR green qPCR Supermix (Invitrogen, Carlsbad, CA) was used as the reaction mix, with 1 l of the cDNA (16 ng) template, in a final volume of 20 l. Each cDNA sample was done in technical triplicates. Melting-curve analysis was performed at the end of the reaction to confirm a single PCR product. The data were normalized to the actin cDNA levels. Relative expression was determined by the threshold cycle (2 ϪΔCT ) method (69).