The Vacuolar Ca2+ ATPase Pump Pmc1p Is Required for Candida albicans Pathogenesis

Maintenance of Ca2+ homeostasis is important for fungal cells to respond to a multitude of stresses, as well as antifungal treatment, and for virulence in animal models. Here, we demonstrate that a P-type ATPase, Pmc1p, is required for Candida albicans to respond to a variety of stresses, affects azole susceptibility, and is required to sustain tissue invasive hyphal growth and to cause disease in a mouse model of disseminated infection. Defining the mechanisms responsible for maintaining proper Ca2+ homeostasis in this important human pathogen can ultimately provide opportunities to devise new chemotherapeutic interventions that dysregulate intracellular signaling and induce Ca2+ toxicity.

reticulum and Golgi apparatus (4). In yeast, more than 90% of the intracellular Ca 2ϩ is sequestered in the vacuole (4). Upon appropriate stimulation, Ca 2ϩ ions are rapidly released into the cytoplasm from extracellular sources as well as intracellular stores to activate Ca 2ϩ -dependent proteins such as calmodulin (5,6). Resting state equilibrium is subsequently restored through calcium efflux, resulting in transient cytoplasmic Ca 2ϩ fluxes that relay the extracellular signal. Failure to remove excess intracellular Ca 2ϩ from the cytoplasm or to restore presignal equilibrium dysregulates intracellular signaling and can lead to cell death (7,8). As such, the mechanisms responsible for Ca 2ϩ detoxification are of critical importance to eukaryotic cell viability.
In fungi, two systems are responsible for sequestering cytoplasmic Ca 2ϩ ions into the vacuole, namely, the H ϩ /Ca 2ϩ exchanger Vcx1p (9)(10)(11), which uses the proton gradient across the vacuolar membrane (generated by the V-ATPase) to drive Ca 2ϩ transport, and the P-type ATPase Pmc1p (11,12). Both Pmc1p and Vcx1p are required for Cryptococcus neoformans to colonize lung or brain tissue in a mouse model of infection (13)(14)(15). While Aspergillus fumigatus, a major cause of pulmonary as well as disseminated infections of humans, has three Pmc1p homologues (PMCA, PMCB, and PMCC), with PMCC seemingly essential for viability and PMCA required for virulence in a neutropenic mouse model of invasive pulmonary aspergillosis (16). Ca 2ϩ -dependent signaling, particularly through the calcineurin signaling pathway, is also required for Candida albicans to tolerate the azole antifungals (17,18). However, deletion of the PMC1 gene has been reported to result in fluconazole resistance in C. albicans (18). The purpose of this study was to determine if Pmc1p or Vcx1p is required for C. albicans pathogenicity and how these pumps impact antifungal resistance.

RESULTS
Pmc1p is required for Candida albicans stress tolerance. To determine if Pmc1p or Vcx1p is required for C. albicans pathogenesis, we constructed pmc1Δ/Δ and vcx1Δ/Δ mutants using a PCR-based approach (19). Complemented strains were made by reintroducing a wild-type PMC1 or VCX1 allele into the pmc1Δ/Δ or vcx1Δ/Δ mutant, respectively, using an integrating vector that fully restores the IRO1-URA3 locus. We initially examined phenotypes that have been associated with loss of Pmc1p or Vcx1p function in fungi (10-12, 15, 16, 18). Each of the described phenotypes was verified using at least two independently derived clones for each genotype. While the pmc1Δ/Δ and vcx1Δ/Δ mutants grew to a similar extent as the wild-type control strain on yeast extract-peptone-dextrose (YPD) agar plates, the pmc1Δ/Δ mutant was severely impaired by high concentrations of CaCl 2 ( Fig. 1A; see also Fig. S1 in the supplemental material). The pmc1Δ/Δ mutant was more resistant than the wild type to LiCl and CdSO 4 but hypersensitive to the membrane stressor SDS (Fig. 1A). Reintroduction of PMC1 into the pmc1Δ/Δ mutant only partially restored wild-type phenotypes with respect to LiCl, CaCl 2 , and CdSO 4 sensitivity, suggesting that PMC1 may be haploinsufficient. The PMC1 reconstituted strain also remained completely sensitive to SDS at the concentration tested (0.05%), suggesting that the capacity of C. albicans to survive plasma membrane stress is especially sensitive to Pmc1p deficiencies. The growth of the vcx1Δ/Δ mutant, on the other hand, was unaffected by CaCl 2 , LiCl, or CdSO 4 supplements (Fig. 1A). The vcx1Δ/Δ mutant was, however, sensitive to SDS, albeit to a lesser extent than the pmc1Δ/Δ mutant. These results confirm the importance of C. albicans Pmc1p for Ca 2ϩ homeostasis and resistance to ionic as well as membrane stress.
Pmc1p influences Candida albicans antifungal susceptibility. A previous study reported that a C. albicans pmc1Δ/Δ mutant was resistant to the antifungal drug fluconazole (18). Similarly, we found our pmc1⌬/⌬ mutant to be more resistant than the wild type to fluconazole when grown at 30°C on YPD agar supplemented with the drug (Fig. 1A) or using Etest strips (see Fig. S2). However, when tested in liquid RPMI medium at 35°C according to the standards of the CLSI protocol (20), or on RPMI agar with Etest strips (Fig. S2), the pmc1⌬/⌬ mutant was as susceptible as the wild type. This indicated that the fluconazole resistance phenotype of the pmc1⌬/⌬ mutant was dependent on the medium or some other growth condition. To determine how temperature influ-ences the pmc1Δ/Δ mutants' susceptibility to fluconazole, we compared the sensitivity of our C. albicans strains on YPD agar using spot dilution assays, following incubation at either 30, 35, 37, or 42°C. This revealed that while the pmc1Δ/Δ mutant is more resistant to fluconazole than the wild type at 30°C, it was not significantly different at 35 or 37°C and, paradoxically, was more sensitive at 42°C (see Fig. S3). Thus, the effect of Pmc1p on C. albicans sensitivity to fluconazole was temperature dependent. Notably, the vcx1Δ/Δ mutants' sensitivity to fluconazole was indistinguishable from that of the wild type under all conditions tested.
Interestingly, the pmc1Δ/Δ mutant was slightly more sensitive than the wild type to the morpholine antifungal amorolfine (see Fig. S4). Amorolfine inhibits both C-8 sterol isomerase (Erg2p) and C-14 sterol reductase (Erg24p), both of which act downstream of Erg11p in the ergosterol biosynthetic pathway (21). Again, the susceptibility of the vcx1Δ/Δ mutant to amorolfine was indistinguishable from that of the wild-type control (data not shown).
Loss of Pmc1p impairs Candida albicans hyphal formation. The ability to form hyphae is important for C. albicans pathogenicity (22,23). We therefore examined the pmc1Δ/Δ and vcx1Δ/Δ mutants' capacity to form hyphae. The pmc1Δ/Δ mutant's ability to form hyphae was severely impaired on M199 or 10% fetal bovine serum (FBS) agar ( Fig. 2A), and it remained as yeast cells in liquid FBS (Fig. 3). While a significant fraction of pmc1Δ/Δ cells produced short filaments in liquid M199 (Fig. 3), these were substan- tially shorter than for the wild type, and many cells remained in the yeast form under these conditions. In contrast, the vcx1Δ/Δ mutant exhibited no detectable defects in hyphal growth under any of these conditions, indicating that Vcx1p is not required for hyphal growth in C. albicans.
Since the pmc1Δ/Δ mutant was hypersensitive to high levels of Ca 2ϩ and unable to form normal hyphae in FBS, we next evaluated if the defects in hyphal formation were due to the high levels of Ca 2ϩ found in serum (3.5 to 4 mM) (24). This was tested using the Ca 2ϩ chelator EGTA to sequester free Ca 2ϩ in the FBS. The addition of 5 mM EGTA to the FBS largely restored the ability of the pmc1Δ/Δ mutant to form filaments (Fig. 4). Interestingly, the ability of the wild type and the complemented control strain to form hyphae was reduced by the addition of EGTA to the FBS. These results establish that the calcium concentrations within host tissues and fluids, as well as the calcium detoxification functions performed by Pmc1p, are an essential determinant of fungal morphogenesis and therefore C. albicans pathogenicity.
Finally, since hyphal growth is intimately linked to the capacity of C. albicans to form biofilms (25), we compared the pmc1Δ/Δ, vcx1Δ/Δ, and wild-type control strains abilities to form biofilms using a simple in vitro assay. Again, the pmc1Δ/Δ mutant was significantly impaired in its capacity to form biofilms, while the vcx1Δ/Δ mutant was able to form biofilms to the same extent as the wild-type control and complemented strains (Fig. 2B).

FIG 2
Candida albicans pmc1Δ/Δ mutant is defective in hyphal growth and biofilm formation. The abilities of the wild-type (GP1), pmc1Δ/Δ and vcx1Δ/Δ mutants, and isogenic control strains to form hyphae and biofilms were compared. (A) Each strain was resuspended at 1 ϫ 10 7 cells/ml in sterile deionized water, and 2.5 l was spotted onto either M199 or 10% FBS agar plates. The resulting colonies were imaged after 96 h of incubation at 37°C. (B) Each strain was suspended at 1 ϫ 10 6 cells/ml in RPMI medium (pH 7.0), and 200 l was dispensed into the wells of a flat-bottomed 96-well plate. After incubating at 37°C for 24 h, the plate was rinsed with PBS, and biofilm formation was detected using a crystal violet staining procedure. Crystal violet was subsequently eluted with 95% ethanol, and resolubilized dye was quantified by measuring the OD 570 in a microplate reader. The means and standard errors of the means from three biological replicates are shown. The mean of each group was compared using a two-way analysis of variance (ANOVA), and P values of relevant comparisons are indicated.

Pmc1p is required for Candida albicans virulence in a mouse model of disseminated infection.
To determine if either Pmc1p or Vcx1p is required for C. albicans to cause disease within its mammalian host, we compared the virulence of the pmc1Δ/Δ and vcx1Δ/Δ mutants with that of the wild type in a mouse model of disseminated infection (Fig. 5). All mice infected with wild-type C. albicans succumbed within 7 days of infection. However, all 7 of the mice infected with the pmc1Δ/Δ mutant survived the duration of the experiment (14 days), with 3 having undetectable levels of fungal colonization within their kidneys and the remainder having relatively low levels (ranging from 3.32 ϫ 10 3 to 2.34 ϫ 10 4 CFU/g of kidney). In contrast, the vcx1Δ/Δ mutant was as virulent as the wild-type and revertant control strains, as determined by the comparable survival times of mice infected with each strain. These data indicate that Pmc1p, but not Vcx1p, is required for C. albicans pathogenicity following dissemination through the bloodstream (Fig. 5).

DISCUSSION
In this study, we explored the contributions of the Pmc1p Ca 2ϩ pump and the Vcx1p Ca 2ϩ exchanger to stress response and pathogenesis of C. albicans. It is evident from these results that the inability of the pmc1Δ/Δ mutant to properly remove Ca 2ϩ from the cytoplasm into the vacuole renders the excess Ca 2ϩ toxic to the mutant in a Ca 2ϩ -rich medium.
Another interesting finding was the impairment of hyphal formation in the pmc1Δ/Δ mutant in M199 medium and FBS. This suggests that the regulation of Ca 2ϩ fluxes by Pmc1p into the vacuole is important for the yeast-to-hyphae transition and follows the findings of previous work indicating that Ca 2ϩ release from the fungal vacuole into the cytoplasm via the Ca 2ϩ channel Yvc1p is important for hyphal formation (26). This highlights the importance of Ca 2ϩ uptake into and release from the fungal vacuole in morphogenesis. Interestingly, removal of free Ca 2ϩ from the FBS allowed the pmc1Δ/Δ mutant to form filaments. Thus, while the concentration of calcium in FBS (ϳ3.5 to 4 mM) (24) was substantially lower than that used to test the calcium toxicity herein, it was nonetheless sufficient to impair the mutant's capacity to form hyphae, a characteristic that is intimately associated with the ability of C. albicans to cause disease (22,23). Our results further suggest that the pmc1Δ/Δ mutant does not have a mechanical defect in its ability to form hyphae, but rather its inability to remove excess Ca 2ϩ from the cytoplasm interferes with the requisite signaling events. It is likely that the dysregulation of Ca 2ϩ -based signaling also underlies the wide range of physiological and stress-related phenotypes of the pmc1Δ/Δ mutant.
C. albicans mutants deficient in calcineurin signaling were previously reported as exquisitely sensitive to Ca 2ϩ levels in serum (18,24) and had severe defects in the colonization of kidney tissue in a mouse model of disseminated infection (18,27). This makes sense, since Pmc1p is a downstream effector of calcineurin signaling (18,28); therefore, a loss of Pmc1p would be expected to produce a similar phenotypic profile  to that resulting from calcineurin dysfunction. Collectively, the phenotypic deficiencies of the pmc1Δ/Δ mutant resulting from its reduced capacity to sequester excess Ca 2ϩ into the vacuole likely underlie its inability to colonize or invade mammalian tissue in the mouse model of disseminated infection. Interestingly, loss of Vcx1p function had little consequence on C. albicans physiology in vitro or pathogenicity in vivo. Given that Pmc1p is a low-capacity high-affinity Ca 2ϩ transporter, while Vcx1p is a high-capacity low-affinity transporter, our results suggest that restoring a low cytoplasmic concentration of Ca 2ϩ following a signaling event may be more important with respect to avoiding toxicity than the rate at which Ca 2ϩ is initially removed.
While most Pmc1p homologs are not essential for fungal viability in vitro, their function appears to be essential for the survival of C. albicans, C. neoformans and A. fumigatus in vivo, i.e., within mammalian tissue (13,15,16). Furthermore, even partially suppressing the expression of either of two Pmc1p orthologues that localize to intracellular acidocalcisomes is sufficient to cause gross morphological abnormalities and severely inhibit the growth of Trypanosoma brucei (29), the causative agent of African sleeping sickness. Collectively, these data underscore the critical importance of vacuolar Ca 2ϩ uptake by the high-affinity Pmc1p transporter in promoting the survival of infectious eukaryotes within the mammalian host. Given the severity of the pathogenesis defects that result from loss of Pmc1p activity in each of these human pathogens, and in particular, the consequences upon fungal colonization of mammalian tissue, Pmc1p could provide a potentially efficacious target for antifungal therapy. Alternative approaches that prevent the sequestration of Ca 2ϩ within the fungal vacuole, or else mobilize intravacuolar calcium, leading to toxicity, may also be viable therapeutic strategies. Whether Pmc1p itself is vulnerable to small-molecule inhibition and if inhibitors with sufficient fungal selective activity can be derived to provide the basis of a viable antimicrobial pharmacotherapy remain to be determined.
Pmc1p is most closely related to the PMCA-type Ca 2ϩ pumps of mammals but has several important distinctions (11,30). First, fungal Pmc1p proteins localize to the vacuolar membrane rather than the plasma membrane, and thus sequester excess calcium into an intracellular store rather than out of the cell (11,30). Second, while fungal Pmc1p has been classified as a type 2B P-type ATPase, which are typically characterized by a cytoplasmic autoregulatory domain, fungal Pmc1p proteins completely lack this domain. In mammals, the autoinhibitory domain of all four PMCA pumps is found within an extended cytoplasmic C-terminal domain, shown by crystallography to wrap around and block the catalytic core of the protein (31,32). Intracellular Ca 2ϩ fluxes activate PMCA following Ca 2ϩ -calmodulin binding to specific sequences in its C terminus that release autoinhibition. Curiously, fungal Pmc proteins completely lack extended N or C termini or the calmodulin binding sites that form the basis of the autoregulatory domains found in higher eukaryotes (11,(31)(32)(33)(34). Thus, key structural determinants, as well as the molecular mechanisms by which the activity of these pumps are regulated in infectious fungi, are fundamentally different from those found in their mammalian host.
Previous reports, which were confirmed in this study, found that C. albicans pmc1Δ/Δ mutants are resistant to fluconazole on agar plates (18), while their susceptibility is not detectably different from that of the wild type when the standard CLSI broth microdilution protocol is used (20). Here, we determined that the pmc1Δ/Δ mutant's reduced susceptibility to fluconazole was temperature dependent and therefore may relate to a form of azole tolerance known as "trailing growth" (35) rather than outright azole resistance. Trailing growth is observed with a significant proportion of C. albicans isolates and manifests as significant residual growth in the presence of the azole that becomes apparent after 48 h of incubation. In its extreme form, trailing growth can be mistaken for true azole resistance; however, patients and experimental animals infected with trailing isolates generally respond well to treatment with the azoles (36,37). The effect of temperature on azole tolerance was also observed previously in an endosomal trafficking mutant lacking the Rab GTPase Vps21p. A C. albicans vps21Δ/Δ mutant displayed enhanced growth in the presence of fluconazole at 35°C but not at 30°C (38). This enhanced growth in the presence of fluconazole also resembled the trailing growth phenomenon and appeared to be dependent on elevated Ca 2ϩ signaling (20,38). Together, these findings further emphasize the importance of temperature and exogenous Ca 2ϩ levels, as well as intracellular Ca 2ϩ homeostasis, as determinants of the capacity of C. albicans to survive following exposure to the azole antifungals.
Plasmid construction. All oligonucleotides used in this study are listed in Table S1 in the supplemental material.
The PMC1 open reading frame (ORF) with 761 bp of 5=-and 434 nt of 3=-untranslated-region (UTR) sequences was amplified from SC5314 genomic DNA (gDNA) with primer pair PMC1AMPF-KpnI and PMC1AMPR-SacI and then cloned between the KpnI and SacI restriction sites of pLUX (43) to produce plasmid pLUXPMC1.
For construction of the plasmid pLUXVCX1, the VCX1 ORF with 747 bp of 5=-and 324 bp of 3=-UTR sequences was amplified using primer pair VCX1AMPF-KpnI and VCX1AMPR-SacI and cloned between the KpnI and SacI restriction sites of pLUX.
Candida albicans strains. All strains used in this study are listed in Table S2. Transformation of C. albicans with DNA constructs was performed using the lithium acetate method (40). Gene deletion strains were constructed by the PCR-based approach described by Wilson et al. (19), using the ura3Δ/ Δhis1Δ/Δarg4Δ/Δ strain BWP17, kindly provided by Aaron Mitchell (Carnegie Mellon University).
The pmc1Δ/Δ strain was constructed as follows. PMC1 deletion cassettes were amplified by PCR with primers PMC1DISF and PMC1DISR using pRS-ARG4ΔSpeI, pGEM-HIS1, or pDDB57 (containing a recyclable URA3-dpl200 marker) (20,21) as the template. Each PMC1 allele was then sequentially deleted from BWP17 using HIS1 and ARG4 markers to generate pmc1Δ/Δ ura3Δ/Δ gene deletion mutants. The correct integration of the deletion cassettes was confirmed at each step by PCR with the following primers sets: ARG4INTF2/PMC1AMPR-SacI and ARG4INTR2/PMC1AMPF-KpnI for ARG4 integration or HIS1INTR2/ PMC1AMPR-SacI and HIS1INTF2/PMC1AMPF-KpnI for HIS1 integration. The absence of an intact PMC1 allele was confirmed using primer pair PMC1DETF and PMC1DETR. Isogenic mutant and PMC1reconstituted strains were produced by transforming the pmc1Δ/Δ ura3Δ/Δ mutant with either pLUX (vector alone) or pLUXPMC1 after digestion with NheI. The correct integration of the pLUX vector fully restores URA3 and adjacent IRO1 loci, and this was confirmed by PCR using primer pair LUXINTDETF and LUXINTDETR.
VCX1 deletion cassettes were amplified by PCR with primers VCX1DISF and VCX1DISR using pRS-ARG4ΔSpeI, pGEM-HIS1, or pDDB57 as the templates. The vcx1Δ/Δ ura3Δ/Δ gene deletion mutants were produced by sequential deletion of each VCX1 allele using HIS1 and ARG4 markers. Correct integration of deletion cassettes was confirmed at each step by PCR with primers pairs ARG4INTF2/VCX1AMPR-SacI and ARG4INTR2/VCX1AMPF-KpnI (ARG4 integration), or HIS1INTR2/VCX1AMPR-SacI and HIS1INTF2/ VCX1AMPF-KpnI (HIS1 integration). Lack of an intact VCX1 allele was confirmed by using primer pair VCX1DETF/VCX1DETR. Isogenic vcx1Δ/Δ mutant and VCX1-reconstituted strains were produced by transforming the vcx1Δ/Δ ura3Δ/Δ mutant with either NheI-digested pLUX (vector alone) or pLUXVCX1. Correct integration of either plasmid was confirmed by PCR using primer pair LUXINTDETF/LUXINTDETR. Stress resistance and hyphal growth assays. Each C. albicans strain was grown overnight in YPD at 30°C. The cells were washed in sterile deionized water, the cell density was adjusted to 10 7 cells/ml, and 1:5 serial dilutions were performed in a 96-well plate. Each cell suspension was then applied to agar plates using a sterile multipronged applicator. Resistance to different stresses was determined on YPD agar containing 5 g/ml of fluconazole, 0.05% SDS, 1.5 M NaCl, 300 mM LiCl, 5 mM CdSO 4 , or 500 mM CaCl 2 , with incubation at 30°C for 48 to 96 h. To induce hyphal growth, for each strain, 2.5 l of a 10 7 cells/ml cell suspension was spotted on M199 or 10% fetal bovine serum (FBS) agar plates, and incubated for 96 h at 37°C.
Biofilm formation assay. Biofilm formation was assessed using a protocol based on that described by O'Toole (41). Each C. albicans strain was grown overnight in YPD broth at 30°C. Each culture was then washed two times in sterile phosphate-buffered saline (PBS), the cell density was adjusted to 1 ϫ 10 6 per ml in RPMI medium (pH 7.0), and 200 l was dispensed into the wells of a flat-bottomed 96-well plate. After incubation at 37°C for 24 h, each well was rinsed 3 times with sterile PBS and then stained with 0.01% crystal violet for 15 min. Each well was again rinsed 3 times with sterile water, and the dye was eluted with 95% ethanol (200 l/well); 150 l of resolubilized dye from each well was then transferred to a new flat-bottomed microtiter plate, and the optical density at 570 nm (OD 570 ) was measured using a microplate reader.
Antifungal susceptibility testing. Antifungal susceptibility testing of all the strains included in this study was performed using the broth microdilution method described in the CLSI document M27-A3 (42) in a 96-well plate format. All drugs for susceptibility testing used in this study were diluted in dimethyl sulfoxide (DMSO) to 200 times the final concentration. RPMI 1640 medium (Sigma-Aldrich) was prepared according to the CLSI document; the medium was buffered with morpholinepropanesulfonic acid (MOPS) and pH adjusted using NaOH and HCl. Plates were incubated at 25°C, 35°C, or 42°C without shaking for 24 or 48 h. The content of each well was carefully resuspended by pipetting up and down before the OD 600 was measured using a Biotek Cytation 5 plate reader.
Susceptibility testing using fluconazole Etest strips was performed on agar plates with MOPSbuffered RPMI medium at pH 7. Cell suspensions were streaked onto the RPMI plates using sterile cotton applicators. Etest strips were applied on the surface of the agar, and the plates were incubated at 35°C for 24 to 48 h.
Ethics statement. The animals used in this study were housed in American Association for Accreditation of Laboratory Animal Care (AAALAC)-approved facilities at the University of Tennessee Health Science Center (UTHSC). The Institutional Animal Care and Use Committee (IACUC) at UTHSC approved the use of all animals and procedures (IACUC protocol numbers 15-081 and 16-156). Mice were given standard rodent chow and water ad libitum. Mice were monitored daily for signs of distress, including noticeable weight loss and lethargy, and for the body condition score. The IACUC at UTHSC uses the Public Health Policy on Humane Care and Use of Laboratory Animals (PHS) and the Guide for the Care and Use of Laboratory Animals as a basis for establishing and maintaining an institutional program for activities involving animals. To ensure high standards for animal welfare, the IACUC at UTHSC remains compliant with all applicable provisions of the Animal Welfare Act (AWAR), guidance from the Office of Laboratory Animal Welfare (OLAW), and the American Veterinary Medical Association Guidelines on Euthanasia.
Mouse model of disseminated candidiasis. C. albicans strains were grown overnight in YPD broth at 30°C with shaking. Stationary-phase cultures of C. albicans strains were washed twice in sterile, endotoxin-free phosphate-buffered saline (PBS) and resuspended in PBS at 5 ϫ 10 6 cells/ml. Groups of 6 BALB/c mice per C. albicans strain were then inoculated via tail vein injections with 100 l of the desired cell suspension (ϳ5 ϫ 10 5 cells). Viable cell counts of each inoculum were confirmed by plating appropriate dilutions on YPD agar plates and counting the colonies formed after 48 h of incubation at 30°C. Mice were then monitored for 14 days postinfection, and those showing signs of distress were humanely euthanized. Animals surviving to the end of the experiment (day 14) were euthanized, and their kidneys were extracted, weighed, and homogenized in PBS. Serial dilutions of kidney homogenates were plated on YPD agar plates containing 50 g/ml of chloramphenicol. The CFU/g of kidney tissue was then determined from the number of colonies formed on the plates after 48 h of incubation at 30°C.

ACKNOWLEDGMENTS
This work was funded by the National Institute of Allergy And Infectious Diseases of the National Institutes of Health under award number R01AI099080.
The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.