N-Acetylglucosamine Metabolism Promotes Survival of Candida albicans in the Phagosome

Candida albicans is the most important medically relevant fungal pathogen, with disseminated candidiasis being the fourth most common hospital-associated bloodstream infection. Macrophages and neutrophils are innate immune cells that play a key role in host defense by phagocytosing and destroying C. albicans cells. To survive this attack by macrophages, C. albicans generates energy by utilizing alternative carbon sources that are available in the phagosome. Interestingly, metabolism of amino acids and carboxylic acids by C. albicans raises the pH of the phagosome and thereby blocks the acidification of the phagosome, which is needed to initiate antimicrobial attack. In this work, we demonstrate that metabolism of a third type of carbon source, the amino sugar GlcNAc, also induces pH neutralization and survival of C. albicans upon phagocytosis. This mechanism is genetically and physiologically distinct from the previously described mechanisms of pH neutralization, indicating that the robust metabolic plasticity of C. albicans ensures survival upon macrophage phagocytosis.

carboxylic acid utilization in the phagosome. The results indicate that GlcNAc utilization represents a third pathway for blocking acidification of the phagosome.

RESULTS
Growth with N-acetylglucosamine induces rapid pH neutralization. C. albicans can neutralize acidic environments in the presence of nonpreferred carbon sources, including amino acids and carboxylic acids, and we have demonstrated that these two phenomena are genetically and physiologically distinct (15,28). Growth with N-acetylglucosamine (GlcNAc) as the sole carbon source can also promote neutralization of the ambient pH (29). We sought to further understand the physiology of this process as well as the genetic requirements to determine if this represented a third independent mechanism by which C. albicans neutralizes its environment. To do so, we grew C. albicans in minimal media with either GlcNAc or Casamino Acids (CAA) as the carbon source. These media also contained 0.5% glycerol, a carbon source that neither promotes nor inhibits pH changes (28), to minimize differences in growth rates when using mutant strains that cannot utilize GlcNAc (see below). Wild-type control cells alkalinized the medium and also excreted ammonia, as expected (28) (Fig. 1). Interestingly, cells grown in GlcNAc neutralized the acidic medium faster than those grown on CAA for the first 6 h, although the endpoint pH was roughly similar after 8 to 12 h (Fig. 1A). GlcNAc also more quickly induced large, flocculated, hyphal mats than did amino acids (not shown). Significant levels of ammonia were produced and excreted by cells grown on both media, but a smaller amount was observed in GlcNAc-grown cells, despite the higher rate of pH change (Fig. 1B). The difference in ammonia release may be related to the differences in metabolism of GlcNAc versus amino acids or buffering capacity of the media (see Discussion).
We also compared pH neutralization in GlcNAc to the phenomenon observed in the presence of carboxylic acids such as ␣-ketoglutarate (␣KG). While the magnitude and FIG 1 Growth with N-acetylglucosamine rapidly raises the environmental pH and releases detectable ammonia. (A) Wild-type C. albicans strain SC5314 was assessed for pH neutralization in YNBAG plus 20 mM N-acetylglucosamine or 1% Casamino Acids over 24 h. pH was adjusted to 4 with HCl prior to inoculation. (B) Ammonia released by C. albicans cells during pH neutralization on solid medium containing YNBAG plus 20 mM N-acetylglucosamine or 1% Casamino Acids. Ammonia was detected after 24, 48, and 72 h via Nessler's reagent. *, P Ͻ 0.05. Data are expressed as mean values Ϯ standard deviation (SD) from triplicate experiments. YNBAG medium consists of minimal YNB supplemented with 0.5% glycerol and 0.5% allantoin. kinetics of alkalinization of the media were similar between the two conditions, we did not detect ammonia excretion from the ␣KG-grown cells, nor did these cells germinate, as previously reported (15) (data not shown). Thus, the physiologies of carboxylic acidand GlcNAc-induced pH changes are substantially different.
Utilization of GlcNAc is a trait shared with other CUG clade species, each of which is capable of neutralizing acidic environments when grown on this carbon source (see Fig. S1 in the supplemental material). In contrast, GlcNAc induces filamentation only in C. albicans. The GlcNAc gene cluster of HXK1, NAG1, and DAC1 is syntenic in all CUG clade species, with the direction of transcription of each gene conserved (37), indicating that the acquisition of these genes, while predating the divergence of this clade, is a relatively recent event. The more distantly related Candida glabrata, a non-CUG species, cannot utilize GlcNAc and does not alter pH when grown on this medium. Not surprisingly, the C. glabrata genome does not encode homologs of the GlcNAc transporter or catabolic enzymes.
pH neutralization with GlcNAc is genetically distinct from pH neutralization with amino acids. We next sought to compare the two mechanisms of pH neutralization using a genetic approach. First, we examined a mutant strain (hxk1⌬ nag1⌬ dac1⌬) that lacks the three enzymes responsible for the conversion of GlcNAc to fructose-6-phosphate, an intermediate of glycolysis, referred to as the "h-d mutant" (29). When the h-d mutant was grown with GlcNAc as the sole carbon source, there were no appreciable changes in pH ( Fig. 2A). A second mutant, lacking the GlcNAc-specific transporter Ngt1, also failed to neutralize the media, although there was a slight increase in pH ( Fig. 2A), perhaps resulting from low-affinity transport through another membrane permease. Both strains exhibited robust extracellular pH neutralization when CAA was present as the carbon source, indicating that the genes responsible for transport and catabolism of GlcNAc do not affect growth or neutralization with amino acids.
To determine whether the genes required for amino-acid-driven pH neutralization affected GlcNAc-induced pH changes, we used a deletion of STP2, which encodes a transcription factor that regulates many genes necessary for amino acid catabolism and pH neutralization (24,28,38). This strain was unaltered in its ability to change the pH with GlcNAc as the carbon source, indicating that Stp2 is not required for this process (Fig. 2B). When assessed for ammonia release, the ngt1⌬ and h-d mutants were impaired when grown on GlcNAc, but not amino acids, as the carbon source ( Fig. 2C and D). The stp2⌬ mutant, in contrast, showed the opposite pattern. Both results were consistent with the effects on the ambient pH of the media, although it was somewhat unexpected that GlcNAc elicited about half the concentration of ammonia relative to CAA given the comparable change in pH ( Fig. 2C and D).
We have previously implicated several members of a multigene family termed ATO (for ammonia transport outward) in ammonia release during pH neutralization, in particular ATO1 and ATO5 (39). To ask whether these facilitated pH neutralization in the presence of GlcNAc, we assayed ammonia release with both amino acids and GlcNAc as the carbon source in both ato5⌬ and ATO1* mutants. (ATO1* is a dominant-negative mutation in Ato1.) We confirmed our earlier observations in the presence of CAA (Fig. 3A); in contrast, these strains phenocopied the parental strain when GlcNAc was used as the carbon source, indicating that the ATO genes are not required for this process (Fig. 3B).
Both transport and catabolism of GlcNAc are necessary for robust virulence during interactions with murine macrophages. The h-d mutant has a reduced virulence phenotype in the murine model of disseminated candidiasis (29,40). We sought to investigate if this may be due to defects in interactions with murine innate immune cells. For this reason, we examined C. albicans survival upon phagocytosis by RAW264.7 cells as well as the ability of C. albicans to elicit cell damage. When assessed for survival after phagocytosis, it was clear that the ngt1⌬ and h-d mutants had diminished survival compared to the parental strain (Fig. 4A). When the supernatant was examined for levels of lactate dehydrogenase (LDH), a proxy for macrophage membrane damage, it was also clear that the ngt1⌬ and h-d mutants induced less cytotoxicity than the parent strain (Fig. 4B). The magnitudes of the defects in survival and macrophage damage were similar to those observed in the stp2⌬ strain, suggesting that both GlcNAc and amino acids are relevant nutrients in the macrophage phagosome. This is consistent with the modest impairment of virulence observed with both the stp2Δ and h-d mutants in the mouse disseminated candidiasis model (24,29).
The reduced fitness of the stp2⌬ mutant strain during interactions with macrophages is associated with impaired germination of phagocytosed cells (24). To ask whether this was also the case for the GlcNAc catabolic machinery, we assayed the morphology of ngt1⌬ and h-d mutant cells upon phagocytosis by murine macrophages. We measured the length of the germ tube (if any) from the body of the mother yeast cell to the tip of the projection. The ngt1⌬ and h-d mutant strains had a higher proportion of cells with either shorter hyphal projections or that had not germinated relative to the wild-type control (Fig. 5A). These results indicate that utilization of GlcNAc contributes to hyphal growth of phagocytosed cells, although the effects on morphology are less pronounced than when amino acid catabolism is impaired via mutation of STP2.
ngt1⌬ and h-d strains occupy an acidic phagolysosome. We previously demonstrated that phagocytosed C. albicans cells inhabit a compartment (phagosome or phagolysosome) of neutral rather than the expected acidic pH and that the SPS amino acid sensor, Stp2, Ato1, and Ato5 are required to maintain neutrality (21,24,39). Furthermore, neutral pH induces hyphal germination (24). The macrophage-associated phenotypes observed with the GlcNAc mutants are consistent with occupancy in an acidic phagolysosome. To test this, we utilized the acidotropic dye LysoTracker red (LR), which accumulates and fluoresces in acidic organelles. As previously published for the stp2⌬ mutant, the ngt1⌬ and h-d mutant cells colocalize with LR to a greater extent than the parental strain, indicating these cells are present in an acidic compartment ( Fig. 5B and C). We measured LR fluorescence intensity immediately outside the fungal cells, in the lumen of the phagosome, as described previously (21,39), and found that while the ngt1⌬ and h-d mutant cells were not as impaired as the stp2Δ cells, there was a statistically significant increase in LR staining relative to phagosomes containing wild-type cells, indicating that these mutants are in a more acidic environment.

DISCUSSION
This study elaborates upon the finding that C. albicans can utilize N-acetylglucosamine (GlcNAc) to manipulate the environmental pH (29). This phenomenon bears similarities to the extracellular neutralization that results from catabolism of amino acids (28), in the overall magnitude of the pH change, and the excretion of ammonia as a driving albicans strains were cocultured with RAW264.7 macrophages in RPMI medium for 2 h. Samples were counterstained with calcofluor white to identify internalized cells, and hyphal length was measured using Slidebook 6 software. (B) RAW264.7 macrophages preloaded with the acidophilic dye LysoTracker red were cocultured with SC5314, the stp2⌬ mutant, the h-d mutant, or the ngt1⌬ mutant expressing a C-terminal Pma1-GFP fusion. Phagosomal pH was measured by quantifying LysoTracker red fluorescent intensity using Slidebook 6 software. **, P Ͻ 0.01; ***, P Ͻ 0.001; ****, P Ͻ 0.0001; ns, not significant. The boxes represent cells with fluorescence intensities from 25% to 75% and the whiskers are 5% to 95%. At least 50 cells were counted per strain. (C) Representative images of each strain. C. albicans fluorescence results with GFP, Lysotracker red, and mCherry are shown.
GlcNAc Promotes Fungal Survival in the Phagosome force. By genetically blocking the utilization of GlcNAc, either by disrupting the transport from the extracellular environment (ngt1⌬) or by abolishing the enzymes responsible for its catabolism (hxk1⌬, nag1⌬, dac1⌬, or h-d), we have shown that utilization of GlcNAc is required for robust neutralization, as opposed to a strictly signaling role as described for GlcNAc-induced morphogenesis (41). In contrast, none of the described mutations that abolish amino-acid-induced pH changes, including in mutants lacking the Stp2 transcription factor or the Ato1 or Ato5 putative ammonia/ acetate transporters (21,24,39), affect neutralization in the presence of GlcNAc. Like these mutants, however, the ngt1⌬ and h-d mutants are impaired in several measures of fitness within macrophages as a result of occupying a more acidic phagolysosome. Thus, despite some similarities, the amino-acid-induced neutralization and GlcNAcinduced neutralization occur by distinct processes.
In addition to amino acids, C. albicans can raise the environmental pH when utilizing carboxylic acids, such as ␣-ketoglutarate, pyruvate, or lactate, as a carbon source (15). This pH change is not accompanied by hyphal morphogenesis or detectable ammonia release and thus differs from either the amino-acid-or GlcNAc-induced mechanism. Like these processes, however, cells impaired in carboxylic acid utilization are less able to resist killing by macrophages (15). Thus, C. albicans has at least three independent pathways through which it can manipulate the pH of its surroundings, each based on nonpreferred nutrients that appear to be available in the macrophage phagosome. There is evidence that combining amino acid and carboxylic acid mutations results in additive defects in macrophage interactions (15). We speculate that disabling all three systems would reduce fungal survival after phagocytosis to a much greater degree than was observed in any of the single mutants.
The comparison of growth and pH in the presence of amino acids versus GlcNAc highlights two questions regarding the mechanisms of pH change. First, the rise in pH is more rapid with GlcNAc than with amino acids, despite producing less excreted ammonia. This might reflect a shorter lag time in the transition from glucose-grown precursors to GlcNAc, which is metabolized largely through glycolysis, than for amino acids, which entails a switch to gluconeogenesis. Alternatively, the buffering capacity of media with GlcNAc might be less than for amino acids, making the released ammonia more effective. Second, hyphal growth in phagocytosed cells of the ngt1⌬ and h-d mutants is only modestly reduced relative to the severe defect of a strain lacking STP2 (24). GlcNAc is a potent morphogenetic inducer, sensed via an intracellular mechanism that does not require its catabolism through Hxk1 (41): thus, the inability to break down this compound might actually increase its concentration relative to phagosomes containing wild-type strains, potentiating signaling.
Evidence is mounting that C. albicans uses nonglucose carbon sources as a signal of the host environment. The Brown lab has convincingly demonstrated that growth of cells on lactate, an abundant carboxylic acid found in mammalian niches, affects cell wall structure, adhesion, and drug resistance of C. albicans, and these combine to alter recognition by macrophages (25)(26)(27). GlcNAc is also abundant throughout the mammalian host and triggers a hyphal morphogenetic program in C. albicans and in the dimorphic fungi Histoplasma capsulatum and Blastomyces dermatitidis; catabolism is not required for this effect in any of these species (41,42). In C. albicans, the cyclic AMP (cAMP)-dependent protein kinase A pathway is required for GlcNAc-induced morphogenesis, but not for its catabolism, while a newly identified transcription factor, Ron1, regulates catabolism and morphogenesis (43,44). GlcNAc has emerged as a signal of the host environment in pathogenic bacteria as well, including Pseudomonas aeruginosa, where it regulates production of phenazines, and Escherichia coli, where it promotes adhesion and biofilm formation via expression of fimbriae and curli fibers (35,45,46). Interestingly, host cells also use GlcNAc as a signaling molecule. A recent study reported that GlcNAc released from breakdown of bacterial cell wall peptidoglycan in the phagosome can trigger NLRP3 inflammasome activation in macrophages (47). Thus, a variety of species have evolved mechanisms to sense GlcNAc to regulate virulence and colonization determinants.
The macrophage-Candida coculture system has been a valuable model for identifying factors that mediate systemic virulence. Indeed, both the stp2⌬ and GlcNAcdeficient mutants are attenuated in the mouse disseminated hematogenous model (24,29,40). Several strains with impaired ability to utilize carboxylic acids are also attenuated in animals or in cell culture models (9,15,48,49). This reinforces the idea that many niches in the mammalian host are poor in glucose but replete in other nutrients. Some of these niches, including parts of the gastrointestinal tract, oral cavity, and vagina, have regions of low ambient pH. The role of pH modulation via alternative carbon metabolism by C. albicans in colonization and/or infection at these sites remains to be seen.

MATERIALS AND METHODS
Culture procedures. C. albicans strains were grown in YPD medium (1% yeast extract, 2% peptone, 2% glucose, with or without 2% agar for solid or liquid medium) routinely before all experiments. Experiments that required a change in carbon source utilized a minimal yeast nitrogen base (YNB) medium (YNBAG) with the specified carbon source and 0.5% allantoin as the nitrogen source plus 0.5% glycerol. The use of glycerol in this medium is to allow for some growth support of genetic mutant strains; glycerol does not affect pH neutralization (21,28). The media were adjusted to a starting pH of 4 using HCl. The murine macrophage cell line RAW264.7 was propagated routinely in RPMI 1640 with glutamate, 10% fetal bovine serum, and 1% penicillin-streptomycin in a 5% CO 2 incubator. Macrophages were used in experiments between passages 7 and 16. Cell counting was performed by a Countess II cell counter (Thermo Fisher) when required. The C. albicans strains that were used in this study are described in Table 1. Derivatives of the ngt1⌬, h-d, and stp2⌬ mutants with a plasma membrane-localized green fluorescent protein (GFP) to facilitate the LysoTracker quantification experiments were generated by integrating a PMA1-GFP translational fusion, as described previously (21). pH neutralization assay. Wild-type and mutant strains of C. albicans were assessed for pH neutralization in YNBG plus 20 mM N-acetylglucosamine or 1% Casamino Acids over 24 h. All C. albicans cultures were grown in YPD, rolling overnight in a 30°C incubator. Cultures were washed 3 times with doubledistilled water (ddH 2 O) for all pH neutralization and ammonia release experiments to wash away remaining YPD nutrients. Washed cultures were inoculated to achieve a starting optical density at 600 nm (OD 600 ) of 0.2. The growth and pH of the cultures were assessed at the indicated times using a standard spectrophotometer (OD 600 ) or pH electrode, as described previously (28).
Ammonia release detection. Ammonia release from neutralizing colonies was detected as described previously (28). Briefly, 3 l of C. albicans cultures adjusted to OD 600 of 1 was inoculated onto solid YNBAG plus 20 mM N-acetylglucosamine or 1% Casamino Acids. Ammonia was detected after 24, 48, and 72 h via Nessler's reagent from a 10% citric acid trap in each petri dish. Absorbance was read at 405 nm in an automatic plate reader (BioTek Synergy H4).
Endpoint survival assay. To assess survival of C. albicans strains in coculture with macrophages, we used a modified endpoint dilution assay (24). To do so, RAW264.7 murine macrophages were inoculated in a 96-well plate at a concentration of 2.5 ϫ 10 4 cells/ml and incubated overnight to reach a concentration of 5 ϫ 10 4 cells/ml. C. albicans strains were grown overnight in YPD at 30°C, collected by centrifugation, washed three times with phosphate-buffered saline (PBS), and diluted in PBS. Cocultures were inoculated starting at a multiplicity of infection (MOI) of 1:1 and serially diluted 5 times before incubation for a total of 16 h postinoculation. Graphs represent a visual count of colony-forming units (CFU) in row 5 of the serial dilution compared to C. albicans incubated without macrophages present.
Macrophage cytotoxicity. Fungus-induced damage to the macrophages was assayed by detection of lactate dehydrogenase (LDH) as described previously (24) using the Cytotox 96 kit (Promega). RAW264.7 murine macrophages were inoculated in a 96-well plate at a concentration of 2.5 ϫ  10 4 cells/ml and incubated overnight to reach a concentration of 5 ϫ 10 4 cells/ml. C. albicans strains were prepared as described above, inoculated at an MOI of 1:1, and incubated in RPMI at 37°C in 5% CO 2 for 16 h. The percentage of release is relative to chemically lysed macrophage controls, accounting for spontaneous release during incubation time. Absorbance was read at 450 nm in an automatic plate reader (BioTek Synergy H4). LysoTracker red pH assessment and morphological surveillance. SC5314, h-d mutant, ngt1⌬/⌬, and stp2⌬/⌬ strains were modified to feature a GFP-tagged Pma1, as previously described (21). RAW264.7 murine macrophages were inoculated in 8-well slides at a concentration of 2.5 ϫ 10 4 cells/ml and incubated overnight to reach a concentration of 5 ϫ 10 4 cells/ml. For LysoTracker experiments, macrophages were pretreated with 0.1 mM LysoTracker red for 1 h prior to inoculation. C. albicans cultures were prepared as described above and inoculated at an MOI ratio of 3:1 prior to incubation for 45 min in a 5% CO 2 incubator at 37°C. For morphological surveillance of C. albicans after phagocytosis, cocultures were incubated for 1, 2, and 3 h. Cells were then stained with 1:300 calcofluor white very briefly, washed three times with PBS, and fixed with 2.7% paraformaldehyde at 37°C for 15 min before storage at 4°C in PBS. Cocultures were imaged under ϫ60 magnification by differential inference contrast (DIC) and corresponding fluorescence spectrums. Image analysis was performed using SlideBook 6.0. The ratio of LysoTracker signal to GFP was obtained as previously described (21).
Statistical analyses. All statistical analyses were performed with Prism 6.