Glycerophosphocholine provision rescues Candida albicans growth and signaling phenotypes associated with phosphate limitation

ABSTRACT The fungal pathogen Candida albicans must acquire phosphate to colonize, infect, and proliferate in the human host. C. albicans has four inorganic phosphate (Pi) transporters, Pho84 being the major high-affinity transporter; its cells can also use glycerophosphocholine (GPC) as their sole phosphate source. GPC is a lipid metabolite derived from deacylation of the lipid phosphatidylcholine. GPC is found in multiple human tissues, including the renal medulla, where it acts as an osmolyte. C. albicans imports GPC into the cell via the Git3 and Git4 transporters. Internalized GPC can be hydrolyzed to release Pi. To determine if GPC import and subsequent metabolism affect phosphate homeostasis upon Pi limitation, we monitored growth and phenotypic outputs in cells provided with either Pi or GPC. In pho84∆/∆ mutant cells that exhibit phenotypes associated with Pi limitation, GPC provision rescued sensitivity to osmotic and cell wall stresses. The glycerophosphodiesterase Gde1 was required for phenotypic rescue of osmotic stress by GPC provision. GPC provision, like Pi provision, resulted in repression of the PHO regulon and activation of TORC1 signaling. Pi uptake was similar to GPC uptake when phosphate availability was low (200 µM). While available at lower concentrations than Pi in the human host, GPC is an advantageous Pi source for the fungus because it simultaneously serves as a choline source. In summary, we find GPC is capable of substituting for Pi in C. albicans by many though not all criteria and may contribute to phosphate availability for the fungus in the human host. IMPORTANCE Candida albicans is the most commonly isolated species from patients suffering from invasive fungal disease. C. albicans is most commonly a commensal organism colonizing a variety of niches in the human host. The fungus must compete for resources with the host flora to acquire essential nutrients such as phosphate. Phosphate acquisition and homeostasis have been shown to play a key role in C. albicans virulence, with several genes involved in these processes being required for normal virulence and several being upregulated during infection. In addition to inorganic phosphate (Pi), C. albicans can utilize the lipid-derived metabolite glycerophosphocholine (GPC) as a phosphate source. As GPC is available within the human host, we examined the role of GPC in phosphate homeostasis in C. albicans. We find that GPC can substitute for Pi by many though not all criteria and is likely a relevant physiological phosphate source for C. albicans.

mucosa and the gastrointestinal and genitourinary tracts of many healthy individuals (2,3).Within these host niches, C. albicans competes for resources with the host flora, necessitating a variety of strategies to utilize host nutrients (4,5).Phosphate is a required nutrient and plays a key role in C. albicans survival and growth and in its ability to invade the host.Several genes involved in phosphate acquisition and homeostasis are upregulated during C. albicans infection (6)(7)(8).
The phosphate homeostatic system in C. albicans, known as the PHO regulon, is conserved among human fungal pathogens.The major transcriptional regulator of the PHO regulon is Pho4 (9,10).Cells lacking Pho4 are unable to upregulate many genes involved in phosphate acquisition and are more sensitive to a variety of stressors, including osmotic and cell wall stresses (6,10).Pho84, a major high-affinity inorganic H+/phosphate (P i ) symporter, a member of the major facilitator superfamily, is among the targets of Pho4.Pho84 is one of four predicted phosphate importers in C. albicans (Fig. 1).Deletion of pho84 causes increased sensitivity to external stressors, increases ROS levels, and diminishes levels of nucleotide sugars required for cell wall synthesis (11,12).Loss of Pho84 also causes in vitro hyphal growth defects and decreases the virulence of C. albicans in a Drosophila model and in murine models of oropharyngeal and systemic candidiasis (11).
In addition to P i transporters, Pho4 also regulates the glycerophosphodiester transporters, Git3 and Git4 (13) (Fig. 1).Glycerophosphodiesters are common lipid metabolites produced by cytosolic and secretory phospholipases of the A and B type that deacylate glycerophospholipids such as phosphatidylcholine (PC) (14)(15)(16).Impor tantly, phospholipase activities from both C. albicans and the human host contribute to GPC production through hydrolysis of available PC derived from their own or each other's cellular membranes (14)(15)(16)(17)(18). Metabolomic studies have identified GPC in serum, in the gastrointestinal and urinary tract, and in various human tissues (19)(20)(21).
The Git transporters are homologs of the S. cerevisiae Git1 transporter, which is specific for the glycerophosphodiester, glycerophosphoinositol (GPI) (15), not to be confused with glycosylphosphatidylinositol, the lipid anchor linking proteins to the membrane.Whereas S. cerevisiae has a single GIT1 gene, C. albicans has 4 Git1 homologs, Git1-4 (22).While Git2 has yet to be characterized, Git1 is specific for GPI, and Git3 and Git4 transport GPC (Fig. 1).GPC import occurs under phosphate replete conditions and is upregulated by phosphate limitation.Of note, the rate of GPC transport in C. albicans under phosphate-limiting conditions is roughly 50× greater than that observed for S. cerevisiae, suggesting differential importance for GPC acquisition between the two organisms (13).Indeed, inhibiting GPC import by deleting git2-4 results in a decrease in the virulence of C. albicans in a mouse model of bloodstream infection (13).Once internalized, GPC can be hydrolyzed to choline and glycerol-3-phosphate by glycero phosphodiesterases.A single glycerophosphodiesterase, Gde1, has been characterized, and its expression is also upregulated by phosphate limitation (13).In order to release P i from glycerol 3-phosphate, a phosphomonoesterase is required.Phosphomonoesterases such as Rhr2 and Dog1 have been identified in C. albicans; however, there may be other phosphomonoesterases that can hydrolyze glycerol 3-phosphate (23,24).
Based on limited quantitative studies, GPC appears to be available at concentrations of ≥1 order of magnitude lower than those of P i in serum (19,(25)(26)(27).Nonetheless, our previous findings led us to hypothesize that GPC import and subsequent metabolism affect phosphate homeostasis.To test this hypothesis, we examined the ability of GPC provision to affect established aspects of phosphate limitation and signaling in C. albicans.We report that provision of GPC rescued several growth defects of a pho84∆/∆ mutant.In addition, GPC provision, like P i provision, resulted in repression of the PHO regulon though to a lesser degree.Strikingly, GPC activated TORC1 in cells lacking the major P i transporter Pho84.Through radiolabel uptake analysis, we found that P i uptake was roughly 2× as great as GPC uptake under low total phosphate conditions, but similar to GPC when ambient phosphate concentrations were moderate.To illustrate the nutritional utility of GPC transport and subsequent metabolism for the cell, we employed a choline auxotrophic strain to show that provision of GPC can simultaneously act as both sole phosphate and sole choline source.Overall, our studies indicate that GPC can substitute for P i in several though not all measures of phosphate homeostasis and is likely a physiologically relevant phosphate source in vivo.

Strains and media
C. albicans strains used in this study can be found in Table 1.Strains were grown aerobically at 30°C unless otherwise stated.Turbidity was monitored by measurement of absorbance at 600 nm (A 600 ) on a BioMate 150 Thermo Scientific spectrophotome ter.The medium used for this study was synthetic complete (SC) (yeast nitrogen base [YNB]) containing 2% glucose and amino acids, as described previously (28).Medium phosphate concentrations were controlled by omitting KH 2 PO 4 (1 g/L) from the synthetic mix and replacing it with KCL (1 g/L).KH 2 PO 4 or GPC was added back into media at high (10 mM), medium (1 mM), or low (200 µM) concentrations unless otherwise stated.Strains were maintained on YPD agar (yeast extract 10 g, peptone 20 g, and dextrose).

Growth assays
Overnight cultures were used to inoculate 200 µL of media at A 600 = 0.1 in a 96-well plate.Plates were incubated at 30°C, with intermittent shaking prior to each reading using a Molecular Devices SpectraMax i3 instrument.A 600 readings were taken at 30-minute intervals, and time zero values were subtracted from each timepoint to reflect overall growth.Data points represent the mean and standard deviation of a minimum of three independent replicates.

Acid releasable inorganic phosphate assay
Acid-labile phosphate was measured by use of a colorimetric molybdate assay as previously described (30,31).Briefly, overnight cultures were grown overnight in low phosphate YNB containing 200 µM P i .Cells were washed twice, reinoculated at A 600 = 0.1 in YNB containing 200 µM P i or 200 µM GPC, and grown to a mid-log phase.Cells were then harvested and resuspended in 500 µL 0.1% Triton X-100 and lysed using zirconia/silica bead homogenization.Lysate protein concentrations were determined using a BCA Protein Assay Kit (Pierce).One hundred micrograms of whole-cell lysate was then boiled for 30 minutes in 1 M HCl before phosphate quantification using the colorimetric molybdate assay in biological triplicate.

P-orthophosphate and 14 C-choline-glycerophosphocholine uptake assays
Uptake assays were altered from references 32, 33.Cells were pregrown in low phos phate conditions (200 µM KH 2 PO 4 ).Cultures were then reinoculated in the indicated media conditions and grown into log phase at 30°C.Once in a log phase, cultures were reinoculated at A 600 = 0.1 and provided with either 32 P-orthophosphate or 14 C-cholineglycerophosphocholine at the indicated concentrations in separate cultures.After 1 hour, 1-mL aliquots were removed from each culture, centrifuged briefly, and separated into extracellular and cellular fractions.Radioactivity in each fraction was measured using liquid scintillation counting.

PHO84 promotor induction analysis
Cells of genotype PHO84/pPHO84-GFP-NAT1-PHO84 were grown in YPD liquid medium with additional 10 mM P i for 16 hours in order to maximally repress the PHO84 promoter and washed three times with 0.9% NaCl.Cultures were adjusted to A 600 = 0.01 in synthetic complete with increasing concentrations of KH 2 PO4 or GPC (0.05, 0.1, 0.2, 0.4, and 0.5 mM) and 50 µL/well of eight technical replicates for each condition was inoculated into a black 384-well plate with a transparent bottom.During incubation at 30°C, A 600 and GFP signal (Ex 485/20 nm; Em 528/20 nm) were recorded every 30 minutes at gain 50 in a Synergy 2 BioTek Plate Reader for 18 hours.Readings were graphed in GraphPad Prism at the 16-hour timepoint.

P-S6 western blotting in Pho84 strains with GPC
Cell lysis and western blotting were performed as described in (34).Rabbit anti-P-S6 (Cell Signaling Technology, #9611L) was used as the primary antibody.The loading control was Rabbit anti-Cdc28 (PSTAIRE, Santa Cruz #sc-53).Anti-rabbit IgG (Cell Signaling Technology, #7074S) was used as the secondary antibody.For densitometry, ImageJ (imagej.net/welcome)software (opensource) was used to quantitate signals obtained from Azure biosystems c600.

Provision of GPC rescued hypersensitivity of pho84∆/∆ cells to osmotic and cell wall, but not peroxide, stresses
Pho84 is the major high-affinity P i transporter in C. albicans (30), and its function can only partially be substituted by other phosphate transporters, such as Pho89 and Pho87 (Table 2).In the absence of Pho84, or of its major transcriptional regulator, Pho4, C. albicans becomes more susceptible to a variety of stresses including osmotic, cell wall, and oxidative stresses (8)(9)(10)(11)(12).We performed growth experiments to see if the provision of GPC can alleviate growth defects and stress hypersensitivities associated with phosphate limitation caused by loss of PHO84.As shown in Fig. 2A (top left), a pho84∆/∆ mutant displayed a growth defect in low phosphate (200 µM P i ) conditions in comparison to WT (Pho84 is primarily responsible for transport at this concentration, Table 2).Provision of 200 µM GPC instead of P i in the media rescued this phenotype.
A pho84∆/∆ mutant also showed increased sensitivity to osmotic stress caused by 1M NaCl addition when grown in media containing 200 µM P i in comparison to WT (Fig. 2B, top right).When provided 200 µM GPC instead of P i , this phenotype was rescued restoring growth to WT levels.Similarly, during cell wall stress induced by exposure to 20 µg/mL of the chitin-binding compound calcofluor white (CFW), pho84∆/∆ mutant cells grew poorly (Fig. 2C, bottom left), as we showed previously (12).This phenotype was rescued by provision of 200 µM GPC as phosphate source, which restored growth similar to the WT in these conditions (Fig. 2C, bottom left).Hypersensitivity of pho84∆/∆ cells to peroxide stress (11) (Fig. 2D, bottom right) induced by 1.5 mM H 2 O 2 was not clearly rescued by provision of GPC, although some growth began to occur in pho84∆/ ∆ cells after 30 hours.Of note, both WT and reintegrant strains displayed long lag times (note x-axis) upon provision of either P i or GPC in this experiment, indicating that 1.5 mM H 2 O 2 presented the cells with a severe stress.One interpretation is that the two-step requirement to release free P i from GPC precludes efficient rescue at this H 2 O 2 concentration.Further experimentation will be required to test whether pho84∆/∆ cells' elevated reactive oxygen species' content even in the absence of exogenous oxidative stress, and their hypersensitivity to superoxide and peroxide stresses (11), can be rescued by provision of GPC.Overall, these results suggest that GPC is transported into the cell and enters cellular metabolism efficiently enough to counter major phenotypes associated with severe P i limitation as experienced by cells lacking PHO84 in low ambient phosphate.

Gde1 was required for GPC utilization as a phosphate source
Uptake of GPC requires the Git3 and Git4 transporters (13).Once internalized, GPC is hydrolyzed by the glycerophosphodiesterase Gde1 into choline and glycerol-3 phos phate (13).To verify that GPC must be hydrolyzed prior to its utilization as a phosphate source under stressed and non-stressed conditions, we employed a gde1∆/∆ mutant subjected to NaCl stress.As shown in Fig. 3A, WT and gde1∆/∆ grew identically when P i was the phosphate source, both in the presence and in the absence of 1 M NaCl.However, when GPC was provided as the phosphate source, gde1∆/∆ showed a growth defect in the absence of stress, and that defect increased in the presence of 1 M NaCl.The growth defect of gde1∆/∆ is shown in Fig. 3B, where only the 16-hour timepoint is presented.Of note, loss of Gde1 does not completely inhibit growth on GPC, indicating that there are other, as yet uncharacterized, enzymes that can hydrolyze GPC (13).

GPC and P i provision supported similar levels of intracellular acid-labile phosphate
To directly determine how provision of GPC compares with provision of P i in terms of internal phosphate stores, we quantified acid-labile phosphate within cells.Boiling acid treatment releases phosphate from polyphosphates and has been used to determine estimates of free P i arising from polyphosphates in yeast (30,35).
As shown in Fig. 4, GPC and P i provision supported similar levels of total free and acid-labile internal phosphate in a wild-type strain.This implies that despite the two steps of catabolism needed to release free phosphate from GPC, internal phosphate stores are not negatively impacted.Previous studies have shown that in the absence of Pho84, there is a significant decrease in the amount of acid-labile phosphate, interpreted to be representative of polyphosphate storage (30,35).We have repeated those findings and report that provision with GPC was not able to rescue this phenotype (Fig. 4).Given the predicted lower P i content of pho84∆/∆ cells at the beginning of the experiment (30), the provided GPC may have only been sufficient to enable biomass addition but not replenishment of polyphosphate stores in these cells.

GPC repressed PHO84 promoter activation
The PHO regulon is the key homeostatic mechanism for phosphate import and intracellular phosphate distribution.Pho4 is the primary transcriptional activator of the PHO regulon, upregulating the expression of PHO84, the genes encoding the GPC importers Git3 and Git4, and the glycerophosphodiesterase Gde1 among many other genes (9,10).Pho4 activity increases as the P i available within the cell decreases.To assess how provision of GPC impacts PHO regulon signaling, we measured the fluorescence produced through activation of the PHO84 promoter driving GFP in a strain containing two functional PHO84 alleles (12) as a proxy for Pho4 transcriptional activity.
As expected and previously shown (12), fluorescence from pPHO84-GFP decreased as P i concentration in the medium was increased from 0.05 mM to 0.5 mM (Fig. 5).When GPC was provided as the phosphate source, a comparable decrease in pPHO84-GFP fluorescence occurred (Fig. 5).Thus, GPC repressed the PHO regulon in a qualitatively similar manner to P i .At 0.05 mM and 0.1 mM P i , we observed a statistically significant but small increase in repression (roughly 110%) of pPHO84-GFP by GPC as compared with P i , a result for which we have no obvious explanation.At 0.2 mM, repression by both phosphorus sources appeared equivalent.A larger quantitative difference was noted at the 0.4 mM and 0.5 mM levels, where P i repressed expression to a greater extent (roughly 150%) than GPC.Pho4 also regulates the expression of Gde1, the enzyme needed for the first step in hydrolysis to release P i by GPC hydrolysis.It may be that at high GPC

GPC activates TORC1 signaling independently of Pho84
The mechanisms by which the cell senses cellular phosphate is an active area of research.
However, it has been established that phosphate, in addition to nitrogen and carbon, is one of the nutrients sensed by the C. albicans and S. cerevisiae TOR (target of rapamycin) complex 1 (TORC1) signaling pathway (12,30).TORC1 signaling is highly conserved within eukaryotes and controls cellular growth and proliferation in depend ence on nutrient availability.During phosphate starvation or upon PHO84 deletion, TORC1 signaling is decreased.TORC1 signaling is activated by the upstream GTPase, Gtr1 (12,30).In a recent study, Pho84 is hypothesized to have transceptor activity, affecting TORC1 signaling in combination with P i import (36).To examine if the provision of phosphorus as GPC is able to activate TORC1 similarly to the provision of P i , we monitored the phosphorylation of ribosomal protein S6, a known downstream target of TORC1 (34).In a wild-type strain, provision of GPC activated TORC1 signaling similarly to P i provision.Consistent with previous studies, we found that TORC1 signaling was decreased in pho84∆/∆ compared with wild-type cells when they were provided with P i (Fig. 6).When GPC was provided to pho84∆/∆ cells at concentrations of 0.1 and 10 mM, TORC1 signaling was restored.Thus, GPC activated TORC1 independently of Pho84 to return signaling to wild-type levels (Fig. 6).

Comparison of GPC and P i transport
In the experiments represented in Fig. 2-6, either GPC or P i was provided as the sole phosphate source.In the human host, GPC and P i are both available albeit at different concentrations and compartment distributions (37).Human serum P i is measured routinely as part of standard electrolyte panels and ranges from 0.8 mM to 1.3 mM (26,27).GPC concentrations in serum have been measured much less frequently and range from 3 mM to 35 mM (19,25).However, the relative concentrations of GPC and P i in host We therefore examined the transport of GPC and P i when both metabolites were available.Equimolar amounts of P i and GPC were provided in experiments in which one of the two was radiolabeled ( 14 C-choline-GPC or 32 P-orthophosphate).We used low and moderate total phosphate concentrations, either 200 µM total phosphate (100 µM P i plus 100 µM GPC) or 500 µM total phosphate (250 µM P i plus 250 µM GPC).In this way, we compared two conditions in which the PHO regulon was induced, and therefore, the transporters of both GPC and P i were expected to be expressed, albeit at higher levels in lower phosphate concentrations.To control for a background level of GPC transport, we used a git2-4∆/∆ strain that showed negligible GPC transport under the conditions examined.We also observed negligible P i transport in pho84∆/∆ cells (Table 2).At a total ambient phosphate concentration of 200 µM, P i and GPC were imported at similar rates in the course of the 1-hour assay: roughly 10 nanomoles of P i as compared with 7 nanomoles of GPC per mL of cell suspension at an OD 600 of 1 (ODU) (Table 2).At 500 µM ambient phosphate, P i import was roughly 2× greater than that of GPC: 23 nanomoles of P i as compared with 11 nanomoles of GPC (Table 2).We concluded that when both sources of phosphate were available at the same concentration, GPC was imported at a substantial rate, especially when P i was limiting.GPC transport may be more quickly saturable since P i import increased more with increasing ambient P i availability than GPC import.

GPC was able to act as both a phosphate and a choline source
We used a choline auxotrophic strain to test the possibility that GPC can simultane ously act as the sole phosphate and the sole choline source (Fig. 7A).The pem1∆/∆ pem2∆/∆ strain is a choline auxotroph because it lacks the phosphatidylethanolamine methylation pathway, leaving the Kennedy pathway as the sole biosynthetic pathway for phosphatidylcholine biosynthesis in C. albicans (38).When this strain was grown without exogenous P i or choline, provision of GPC was able to support growth similarly to media containing both P i and choline (Fig. 7B).We concluded that GPC hydrolysis sufficed to simultaneously provide choline and phosphate to C. albicans.
(https://hmdb.ca), GPC levels that have been reported thus far have ranged roughly from 1 µM in saliva to 3-35 µM in serum to 30-500 µM in breast milk (19,25,41,44,45), with the concentrations varying widely depending upon the study and the methodology employed.Although quantitative metabolomic data on GPC are limited, it is clear that serum P i (at roughly 1 mM [26,27]) is more abundant than serum GPC (19,25).However, the relative concentrations of GPC and P i in host microenvironments have not been established and are undoubtedly dynamic, as GPC is liberated through phospholipasemediated PC hydrolysis, during which it may reach higher concentrations than those found in serum at a steady state, while serum and cytosolic P i concentrations are highly regulated (17,46,47).Both phospholipase B1 and B5 are required for full virulence of C. albicans.One potential mechanism for these phospholipases in pathogenesis is the release of GPC from host cells' plasma membranes that provides phosphorus as well as choline for the fungus (14,(48)(49)(50)(51). GPC is one of four major renal osmolytes found in tubular cells of the renal medulla during mammalian dehydration states, along with glycine betaine, myo-inositol, and sorbitol (52).In renal cells, GPC has been quantified by mass instead of volume and has been reported to be roughly 18 nmol/mg protein (53,54).In some human candidiasis syndromes and in the murine intravenous infection model, the kidney is a major target for disseminated candidiasis (55), though typically, the renal cortex and medulla are equally infected.Further work, especially regarding the contribution of fungal phospholipases to GPC release, should shed more light on these questions.
C. albicans upregulates high-affinity phosphate transporters during invasive infection suggesting that the host environment mimics P i starvation conditions (56)(57)(58)(59)(60)(61)(62).At the human serum P i concentration of 0.8-1.3mM (26,27), C. albicans would be expected to experience P i sufficiency, but given the acidic optima of all but one of its P i importers (our unpublished data), their activity may be inefficient at the normal pH of human serum (pH 7. 35-7.45).An analogous phenomenon was observed in the opportunistic fungal pathogen Cryptococcus neoformans and called "alkaline pH-simulated nutrient deprivation" (63,64).We examined the possibility that GPC is among the sources of phosphorus used by C. albicans in the host.
Previous studies have established that C. albicans can utilize GPC efficiently as a phosphate source (13).Importantly, loss of the Git3 transporter results in decreased virulence in a mouse model of bloodstream infection (13).We demonstrate here that provision of equimolar GPC can rescue several growth phenotypes associated with phosphate limitation imposed by the lack of PHO84 (Fig. 2).Further, provision of either P i or GPC at equimolar concentrations results in similar levels of total intracel lular phosphate in a WT strain, despite the need for a two-step catabolic process to release P i from GPC. Provision of GPC did not rescue P i stores of pho84∆/∆ cells (Fig. 4).One interpretation could be that the amount of GPC provided was not enough to support growth of pho84∆/∆ cells which requires incorporation of phosphorus into many macromolecules and to simultaneously restore depleted internal phosphate stores over the time span of the assay.
To test whether GPC is used in conditions where both P i and GPC are available at equimolar concentrations, we examined the import of each phosphorus source in the presence of the other.GPC import is similar to P i under low ambient phosphate conditions and measures roughly half as much as P i import in moderate ambient phosphate (Table 2).Both import and hydrolysis of GPC are regulated by the PHO regulon; therefore, P i is the most readily usable form of phosphorus (8,13).In moder ate phosphate conditions, P i appears to be preferentially selected over GPC and the molecular mechanism of this selectivity remains to be discovered.Future work will examine whether the GPC transporters, Git3/4, are saturable at lower transport rates than Pho84, whether their expression is downregulated earlier than that of Pho84 in rising ambient P i concentrations, and whether mechanisms beyond the PHO regulon determine their expression and activity.GPC transporters Git3/4 compensate for loss of PHO84 under conditions where the other P i transporters which are present in the pho84∆/∆ mutant cannot (Fig. 2 and 3), so that 200 µM GPC rescues phenotypes in a pho84∆/∆ mutant that cannot be rescued by 200 µM P i .
GPC has the potential to provide the cell with choline and glycerol in addition to phosphate.Strains lacking a functional PE methylation pathway are choline auxotrophs as they require choline to make PC via the CDP-choline pathway.It has been shown previously that GPC can act as a choline source (38).Here, we further demonstrate the robustness of GPC import and catabolism by showing that GPC can simultaneously act as both a choline and phosphate source in a pem1∆/∆ pem2∆/∆ mutant (Fig. 7).This variety of metabolic uses for GPC may be one of the reasons that the loss of the major GPC transporter, Git3, leads to decrease in virulence in a mouse model (13).
Induction of the PHO84 promoter is a readout of PHO regulon activity in C. albi cans (12) as in S. cerevisiae (65).We observed repression of the PHO84 promoter with increasing provision of both P i and GPC in the medium, though P i was the more potent repressor at equimolar concentrations.Why P i has a stronger effect on the PHO regulon than GPC, when equal amounts of phosphate are delivered intracellularly, e.g., whether Pho84 has transceptor activity toward the PHO regulon, remains to be determined.
The ability of GPC to activate TORC1 signaling in cells lacking PHO84 (Fig. 6) suggests that it is intracellular P i , and not simply a direct signal from Pho84, that provides a crucial stimulus to TORC1.While Pho84 may also have a transceptor activity in addition to its role providing P i to the cytoplasm (30,36), our current findings indicate that in the presence of sufficient intracellular P i provided by GPC, this activity is not required for TORC1 activation.How intracellular P i availability is signaled to TORC1 remains an area of active investigation.
An open area of inquiry is the complete identification of gene products involved in GPC catabolism.As shown in Fig. 3, growth on GPC as a phosphate source is delayed in a gde1∆/∆ mutant, but there are undoubtedly other glycerophophodiesterases involved as well.Secondly, the phosphomonoesterase(s) responsible for the release of free phosphate from glycerol-3 phosphate have yet to be identified.While phosphomonoes terases, like Rhr2, Dog1, and multiple others are known, their role in GPC catabolism has not been established, and there are likely other phosphomonoesterase-encoding genes in the C. albicans genome (7,8,23,24).
Several aspects of the phosphate deprivation response are conserved among pathogenic and nonpathogenic fungi, including the Pho4 transcriptional regulator (10,66,67).However, others have noted that pathogenic fungi have an expanded range of Pho4 targets that include lipid metabolism (63).Lipid metabolism, both synthesis and turnover, is an ongoing process in pathogenic fungi and the human host.Our results show that C. albicans has adapted to use GPC, the product of PC metabolism, as part of its phosphate deprivation response, and that the choline released in the process can feed into PC biosynthesis.C. albicans GPC importers Git3/4 hence have significant roles in both phosphate homeostasis and lipid biosynthesis.Ongoing studies are exploring the possibility that GPC can be converted to PC through direct acylation, as recently shown in S. cerevisiae, plants, and mitis group streptococci (28,68,69).Interconvertibility of membrane organic phosphates with free cellular P i via GPC may contribute to C. albicans adaptation to insufficient P i access in the host.

FIG 2
FIG 2 GPC rescued phosphate starvation phenotypes.(A) Cells were grown overnight in YNB containing 200 µM P i .Cells were then reinoculated into phosphate-free YNB containing either 200 µM P i or 200 µM GPC at an A 600 of 0.1, monitored every 30 minutes.(B) Grown as in panel A with the addition of 1 M NaCl to induce osmotic stress.(C) Grown as in panel A with the addition of 20 µg/mL calcofluor white to induce cell wall stress.(D) Grown as in panel A with the addition of 1.5 mM H 2 O 2 to induce oxidative stress.Data are displayed as the mean ± standard deviation of at least three replicates.

FIG 3 A
FIG 3 A gde1∆/∆ mutant displayed defective growth on GPC.(A) Cells were grown overnight in YNB containing 200 µM P i .Cells were then reinoculated into phosphate-free YNB containing either 200 µM P i or 200 µM GPC at an A 600 of 0.1 with or without 1 M NaCl.A 600 was monitored every 30 minutes using a plate reader.Data are displayed as the mean and standard deviation of at least three replicates.(B) The 16-hour timepoint from the data in A. A one-way ANOVA was performed.**, P < 0.001; ***, P < 0.0005; ****, P < 0.0001.

FIG 4
FIG 4 Internal phosphate levels were similar when cells were provided with either GPC or P i .Cells were grown overnight in YNB containing 200 µM P i .Cells were then reinoculated into phosphate-free YNB containing either 200 µM P i or 200 µM GPC at an A 600 of 0.2 and grown to mid-log phase.Cells were then harvested and assayed for acid-labile phosphate.Data represent the mean and standard deviation of biological triplicates.A one-way ANOVA was performed.

FIG 5
FIG 5 GPC repressed the PHO84 promoter.Cells expressing GFP under the control of the PHO84 promoter (JKC1659) were pregrown in YPD with an additional 10 mM P i overnight.Cells were then inoculated at an OD 600 = 0.01 into media without P i with indicated P i or GPC concentrations.The fluorescent signal and A 600 were followed over 18 hours.The 16-hour timepoint is shown.Data represent the mean and standard deviation of eight biological replicates.A one-way ANOVA was performed.a.u., arbitrary unit ***, P < 0.0005; ****, P < 0.0001.

FIG 6
FIG6 GPC activated TORC signaling.Cells pregrown in YPD overnight were washed three times in 0.9% NaCl and then inoculated into synthetic complete medium with 0.1 mM or 10 mM inorganic phosphate or glycerophosphocholine, at an A 600 of 0.2.Cells were collected after 2 hours of incubation (200 rpm) at 30°C.Total protein extracts were probed with antibody to phosphorylated Rps6 (P-S6) and Cdc28 as loading control.Dens, signal intensity ratio of P-S6 to Cdc28.

FIG 7
FIG 7 GPC served simultaneously as a P i and choline source.(A) pem1∆/∆pem2∆/∆ strain, a choline auxotroph, was grown overnight in YNB supplemented with 200 µM choline.Cells were then starved of phosphate and choline for 8 hours before restarting in phosphate-free YNB supplemented with either 200 µM GPC, 200 µM P i , 200 µM choline (cho), or 200 µM P i plus 200 µM choline as indicated.Cells were started at an A 600 of 0.2 and allowed to grow for 24 hours.Data represent the mean and standard deviation of biological triplicates.(B) Schematic of GPC and PC metabolism.GPC production and catabolism is indicated by black arrows.The Kennedy pathway for PC biosynthesis is shown in blue, and the PE methylation pathway for PC biosynthesis is shown in yellow.Shown in black is GPC production through the deacylation of PC by phospholipase of the b type or through a single deacylation of PC by a phospholipase of the A type to produce Lyso-PC followed by another deacylation by a lysophospholipase.GPC, glycerophosphocholine; G-3-P, glycerol-3-phosphate; P i , inorganic phosphate; LPC, lyso-phosphatidylcholine; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PLB, phospholipase B; PLA, phospholipase A; LysoPl, lysophospholipase; PME, phosphomonoesterase .A one-way ANOVA was performed.****, P < 0.0001.Made in Biorender.

TABLE 2
Comparison of P i and GPC import a a Cells were grown to log phase in YNB containing equivalent amounts of P i and GPC at two concentrations: very low phosphate (100 µM P i + 100 µM GPC) or low phosphate (250 µM P i + 250 µM GPC) conditions.Cells were then reinoculated into the same nutrient conditions at an A 600 of 0.1, but in media in which only one of the compounds was radiolabeled ( 14 C-choline-GPC or 32 P-orthophosphate, indicated by an asterisk at the top of the data column).Samples were incubated for 1 hour at which point cells were harvested and cellular radioactivity was determined per 1 ODU.Data represent the mean and standard deviation of biological triplicates.ND, not done.