Loss of Ypk1, the Yeast Homolog to the Human Serum- and Glucocorticoid-induced Protein Kinase, Accelerates Phospholipase B1-mediated Phosphatidylcholine Deacylation*

Background: Ypk1 is a protein kinase known to regulate sphingolipid homeostasis. Plb1 deacylates phosphatidylcholine to produce glycerophosphocholine. Results: Loss of Ypk1 or disruption of sphingolipid synthesis by other means elevates Plb1-mediated phosphatidylcholine turnover. Conclusion: Accelerated turnover of phosphatidylcholine compensates for aberrant sphingolipid synthesis. Significance: Sphingolipid synthesis is coordinated with phosphatidylcholine metabolism to maintain membrane lipid homeostasis. Ypk1, the yeast homolog of the human serum- and glucocorticoid-induced kinase (Sgk1), affects diverse cellular activities, including sphingolipid homeostasis. We now report that Ypk1 also impacts the turnover of the major phospholipid, phosphatidylcholine (PC). Pulse-chase radiolabeling reveals that a ypk1Δ mutant exhibits increased PC deacylation and glycerophosphocholine production compared with wild type yeast. Deletion of PLB1, a gene encoding a B-type phospholipase that hydrolyzes PC, in a ypk1Δ mutant curtails the increased PC deacylation. In contrast to previous data, we find that Plb1 resides in the ER and in the medium. Consistent with a link between Ypk1 and Plb1, the levels of both Plb1 protein and PLB1 message are elevated in a ypk1Δ strain compared with wild type yeast. Furthermore, deletion of PLB1 in a ypk1Δ mutant exacerbates phenotypes associated with loss of YPK1, including slowed growth and sensitivity to cell wall perturbation, suggesting that increased Plb1 activity buffers against the loss of Ypk1. Because Plb1 lacks a consensus phosphorylation site for Ypk1, we probed other processes under the control of Ypk1 that might be linked to PC turnover. Inhibition of sphingolipid biosynthesis by the drug myriocin or through utilization of a lcb1-100 mutant results in increased PLB1 expression. Furthermore, we discovered that the increase in PLB1 expression observed upon inhibition of sphingolipid synthesis or loss of Ypk1 is under the control of the Crz1 transcription factor. Taken together, these results suggest a functional interaction between Ypk1 and Plb1 in which altered sphingolipid metabolism up-regulates PLB1 expression via Crz1.

Ypk1, the yeast homolog of the human serum-and glucocorticoid-induced kinase (Sgk1), affects diverse cellular activities, including sphingolipid homeostasis. We now report that Ypk1 also impacts the turnover of the major phospholipid, phosphatidylcholine (PC). Pulse-chase radiolabeling reveals that a ypk1⌬ mutant exhibits increased PC deacylation and glycerophosphocholine production compared with wild type yeast. Deletion of PLB1, a gene encoding a B-type phospholipase that hydrolyzes PC, in a ypk1⌬ mutant curtails the increased PC deacylation. In contrast to previous data, we find that Plb1 resides in the ER and in the medium. Consistent with a link between Ypk1 and Plb1, the levels of both Plb1 protein and PLB1 message are elevated in a ypk1⌬ strain compared with wild type yeast. Furthermore, deletion of PLB1 in a ypk1⌬ mutant exacerbates phenotypes associated with loss of YPK1, including slowed growth and sensitivity to cell wall perturbation, suggesting that increased Plb1 activity buffers against the loss of Ypk1. Because Plb1 lacks a consensus phosphorylation site for Ypk1, we probed other processes under the control of Ypk1 that might be linked to PC turnover. Inhibition of sphingolipid biosynthesis by the drug myriocin or through utilization of a lcb1-100 mutant results in increased PLB1 expression. Furthermore, we discovered that the increase in PLB1 expression observed upon inhibition of sphingolipid synthesis or loss of Ypk1 is under the control of the Crz1 transcription factor. Taken together, these results suggest a functional interaction between Ypk1 and Plb1 in which altered sphingolipid metabolism up-regulates PLB1 expression via Crz1.
Ypk1, the yeast homolog of the human serum-and glucocorticoid-induced kinase (Sgk1), is a serine/threonine protein kinase known to affect multiple downstream processes, including lipid homeostasis (1)(2)(3)(4)(5), actin dynamics (6,7), cell wall integrity (6,7), endocytosis (3,8), and lipid flippase activity (1). Ypk1 has a functionally redundant homolog, Ypk2 (Ykr2). Either Ypk1 or Ypk2 is required for cell viability, but Ypk1 plays a more prominent role in several processes (6,9). For example, deletion of YPK1 results in a growth defect and increased sensitivity to several drugs, phenotypes that are not observed in the ypk2⌬ mutant (6). Ypk1, a cytosolic protein, is recruited to the plasma membrane through its interaction with Slm1/2, proteins that contain a pleckstrin homology domain that mediates binding to phosphatidylinositol 4,5-bisphosphate (10,11). At the plasma membrane, Ypk1 is first activated by TORC2 (target of rapamycin complex 2) (10,11) and then by Pkh1/2 phosphorylation (6). TORC2 has been shown to be activated by sphingolipid depletion (12) and elevated reactive oxygen species (13). In turn, Pkh1 and Pkh2, homologs of the PDK1 (mammalian phosphoinositide-dependent protein kinase 1) family, preferentially phosphorylate and activate Ypk1 and Ypk2, respectively, although some crossover is observed (6). Pkh1 and Ypk1 are localized exclusively to the cytosol, whereas Pkh2 and Ypk2 also enter the nucleus. Exogenous addition of the long chain base phytosphingosine activates Pkh1/Pkh2 in vitro as measured by an increase in PKC phosphorylation (14), whereas a more recent paper suggests that the ability of Pkh1/2 to activate Ypk1/2 is unaffected by long chain bases but instead requires the complex sphingolipid, mannosylinositol phosphorylceramide (1).
Another finding linking Ypk1 to lipid metabolism is that overexpression of the B-type phospholipase, PLB1, rescues the lethality of the temperature-sensitive ypk1 ts ypk2⌬ mutant (1). The B-type phospholipases (PLBs) catalyze the deacylation of glycerophospholipids to produce the glycerophosphodiesters and free fatty acids (Fig. 1A). In Saccharomyces cerevisiae, four genes encode proteins with PLB activity, PLB1, PLB2, PLB3, and NTE1, which function in membrane remodeling in the cell. Based on in vivo studies, Plb1 primarily deacylates phosphatidylcholine (PC) to produce external glycerophosphocholine (GroPCho) (19). Biochemical studies have suggested that Plb1 is plasma membrane-associated and secreted from the cell (19 -21), whereas Nte1 is localized to the ER and is responsible for the production of internal GroPCho (22). The primary substrate for Plb2 appears to be exogenous phospholipids, and Plb3 acts on phosphatidylinositol (PI) to produce extracellular glycerophosphoinositol (20,21).
PC synthesis and catabolism are tightly regulated in yeast (23) and other eukaryotic cells (24), and disruptions to PC metabolism can result in aberrant physiological responses. In mammalian cells, increased choline uptake and incorporation into PC is a marker for proliferative growth (25), and conversely, inhibition of choline synthesis results in apoptosis (26). Aberrant PC levels and/or alterations in GroPCho have also been observed in patients with various cancers, such as colorectal (27) and prostate (28) cancer, and in Alzheimer disease (29,30). What has remained unknown, however, is how protein kinases and members of the B-type phospholipases, which deacylate phospholipids to produce GroPCho, are coordinately regulated to ensure lipid homeostasis. Here, we describe a novel role for Ypk1, the yeast homolog of Sgk1, in regulating Plb1mediated PC turnover.

EXPERIMENTAL PROCEDURES
Strains, Plasmids, Media, and Growth Conditions-The S. cerevisiae strains used in this study are shown in Table 1. Strains were grown aerobically with shaking at 30 or 37°C, as noted. Turbidity was monitored by measuring the optical density at 600 nm (A 600 ) on a BioMate 3 Thermo Spectronic spectrophotometer. Empty vectors, YEp351 and pRS426, and vectors overexpressing each of the PLB genes, YEp351-PLB1, YEp351-PLB2, YEp351-PLB3, and pRS426-NTE1, were gifts from Susan Henry (YEp351 series) (21) and Chris McMaster (pRS426-NTE1) (31). Media used for this study included rich yeast extract peptone dextrose (YPD) medium purchased from Fisher or synthetic complete yeast nitrogen base (YNB) medium containing amino acids, 2% glucose, and 75 M inositol, as described previously (32).
Construction of Strains-The S. cerevisiae WT strain BY4742 was purchased from Open Biosystems (Thermo Scientific, Huntsville, AL). The deletion strains were constructed using standard homologous recombination techniques (33). Drugresistant markers, hphMX, KanMX, natMX, and ble, were amplified from plasmids pAG32, pUG6, pAG25, and pUG66, respectively, using the primers listed in Table 2, and the resulting DNA fragments were inserted in place of the genes targeted for deletion. The plasmids were received from Euroscarf ( 35). PLB1 was disrupted by insertion of hphMX in the YPH499 background strains provided by J. Thorner (6). YPK1 was disrupted by insertion of hphMX in the ORM mutants and ORM phosphomimetic strains provided by D. Drubin (36). The N-terminal 3xFLAG-tagged ORM2 allele was previously shown to be fully functional (36). For deletion of YPK1 in the crz1⌬ strain, the URA3 cassette was amplified from plasmid pPMY-3xHA (37), and the resulting DNA fragment was used to disrupt YPK1.
[ 14 C]Choline Labeling and Preparation of External, Internal, and Membrane Fractions-Cells were grown to uniform labeling in synthetic medium containing 20 M choline and 1 Ci/ml [ 14 C]choline chloride of negligible concentration (NEC141VU250UC (PerkinElmer Life Sciences) or ARC 0208 (American Radiolabeled Chemicals)). Cells were harvested in log phase, washed free of excess label, and reinoculated at A 600 ϭ 0.2 into synthetic medium containing 10 mM nonradioactive choline. At various time points, 1-ml aliquots of the chase cultures were removed and separated into external, internal, and membrane fractions (32,38). In brief, cells were pelleted by centrifugation, and the supernatant was removed and retained as the extracellular fraction. The cell pellet was resuspended in 0.5 ml of 5% trichloroacetic acid (TCA), followed by a 20-min incubation on ice. Cells were pelleted again by centrifugation, and the supernatant was removed as the intracellular (soluble) fraction. The pellet was then resuspended in 0.5 ml of 1 M Tris (pH 8) and centrifuged, and the supernatant was added to the intracellular fraction. The final cell pellet was resuspended in 0.5 ml of 1 M Tris (pH 8) and saved as the membrane fraction. Total counts (external ϩ internal ϩ membrane) recovered at each time point as well as the percentages of counts in each fraction were tracked by liquid scintillation counting.
Analysis of Short Term Incorporation of [ 14 C]Choline into PC-Strains were grown to log phase (A 600 ϭ 0.5) in synthetic 75 M inositol medium, harvested by centrifugation, and concentrated to equivalent densities in fresh synthetic medium containing 75 M inositol and 100 M choline. To start the assay, 50 l of 0.005 mCi/ml [ 14 C]choline chloride was added, and cells were allowed to incubate for 30 min at 37°C. The cells were pelleted by centrifugation, washed with water, and suspended in 0.5 ml of 5% TCA for 10 min. After pelleting, the supernatant was discarded, and the membrane pellet was washed with water. Lipids were extracted from the remaining pellet as described above. Incorporation of [ 14 C]choline chloride into PC was quantified by liquid scintillation counting and verified by TLC.
LC-MS Analysis of Extracellular Choline and GroPCho-Cells were grown in synthetic medium containing 75 M inositol, 20 M choline, and 2 mM KH 2 PO 4 instead of 7 mM KH 2 PO 4 . The cells were pelleted by centrifugation at 1,600 ϫ g for 3 min, and the supernatant containing the extracellular metabolites was filtered through 0.2-m cellulose acetate filters. The method of analysis for these extracellular metabolites by mass spectrometry was described previously (45,46). A 250-l aliquot of supernatant was added to an internal standard (cholined 9 ), and water-soluble lipid metabolites were extracted by the addition of 2.25 ml of chloroform/methanol (2:1, v/v). The suspension was agitated with a Vortex mixer followed by centrifugation for 3 min at 1,600 ϫ g. The upper aqueous phase was dried under N 2 and resuspended in 125 l of acetonitrile/methanol (75:25, v/v) with 10 mM ammonium acetate, pH 4.5. After centrifugation at 16,000 ϫ g for 10 min, an aliquot of the supernatant was transferred to an LC vial and diluted 20-fold in acetonitrile/methanol (75:25, v/v). A 10-l volume was injected

sequences of primers used for gene deletions
The deleted gene was replaced with the gene name in parentheses. Boldface sequences are homologous to the template plasmids.

Gene name
Primer Reverse ATGTATTCACTAAGTCTATCTAATGCTTCTACCTTGCACCATTGAGCTACCTAGCTGTTCGCATAGGCCACTAGTGGATCTG into the Agilent 1200 series Rapid Resolution LC system coupled to an Agilent 6440 triple quadrupole mass spectrometer. Separations were performed on a Waters Xbridge HILIC (150 ϫ 4.6 mm, 5 m) column with an isocratic elution at a flow rate of 0.5 ml/min at room temperature. The mobile phase was acetonitrile/water ( Each sample was run in duplicate. Data were acquired and analyzed using MassHunter Work Station software. The peak areas of the target metabolites were normalized to the internal standard and to the density of cells (A 600 ) from which metabolites were extracted.

RNA Extraction and Real-time Quantitative RT-PCR (qRT-PCR) Gene Expression Analysis-Cultures
were grown in synthetic medium containing 75 M inositol under the conditions specified. RNA was extracted using a hot phenol-chloroform extraction (47). A 7-g sample of RNA was DNase-treated with 2 units of DNase at 37°C for 30 min using the TURBO DNAfree TM kit (Applied Biosystems) per the manufacturer's specifications. Message abundance was quantified by qRT-PCR using the Verso TM SYBR Green 1-Step QRT-PCR ROX kit (Thermo Scientific) on the Applied Biosystems StepOne-Plus TM real-time PCR system. The primers used are shown in Table 3. Standard curves for each primer pair were optimized to determine acceptable primer concentrations (70 -100 nM) and template concentrations (15-25 ng/25-l reaction). Reverse transcription was carried out at 50°C for 15 min, followed by 95°C for 15 min for RT inactivation and polymerase activation. Amplification parameters were as follows: 40 cycles at 95°C for 15 s, 54.5°C for 30 s, and 72°C for 40 s. Primer specificity was verified by melt curve analysis and visualization of amplicons by gel electrophoresis. Primer sets did not produce amplicons when RNA from the respective deletion strains was used as template. Control reactions without template and without reverse transcriptase were performed to verify the absence of genomic DNA contamination in the RNA samples and/or the reagents. RNA from at least three independent cultures of all strains was analyzed in triplicate reactions per primer set. Data were normalized to the endogenous control gene, SNR17, and analyzed using the comparative ⌬⌬C T approach (48,49).
Construction of the Chromosomal PLB1-I-3xHA Allele-The plasmid pPMY-3xHA (37), kindly provided by Nancy Hollingsworth (Stony Brook University), was used as the template to amplify a 3xHA-URA3 module for insertion into the genome between amino acids 30 and 31 of Plb1. Because Plb1 is predicted to be GPI-anchored to the plasma membrane, insertion at this point would avoid cleavage of the HA tag from either end because both the N-terminal localization signal and C-terminal GPI anchor attachment site are cleaved during processing and GPI anchor attachment. Insertion did not interfere with the predicted catalytic site for Plb1. The primers used to amplify the 3xHA-URA3 module are listed in Table 2. The PCR product was transformed into WT and ypk1⌬ strains, and transformants were selected on synthetic plates lacking uracil. To verify integration at the correct location, DNA was extracted and used as template with two sets of primers. Verified transformants were grown in YPD liquid medium to allow for recombination and loss of the URA3 marker. Cells that had lost the URA3 marker were selected on synthetic plates containing 5-fluoroorotic acid. DNA was extracted from the 5-fluoroorotic acidresistant colonies and used as a template for PCR. Strains producing the expected amplicon size, indicating insertion of the 3xHA and loss of the URA3, were named JPV767 (WT ϩ PLB1-I-3xHA) and JPV 745 (ypk1⌬ ϩ PLB1-I-3xHA). These strains therefore contain an internal 3xHA tag between nucleotides 90 and 91 of PLB1 in their genomic DNA. The function of Plb1 containing the 3xHA tag was verified by observing the expected Plb1 activity in PC turnover/GroPCho production experiments.
Protein Extraction and Sucrose Gradient Separation-For total protein extraction from cell lysates, cultures were grown in synthetic ϩ 75 M inositol medium to log phase (A 600 Ϸ 0.7). Cells were collected by centrifugation of a 125-ml culture, and the supernatant was removed for evaluation of secreted protein. The supernatant was concentrated 100-fold using an Amicon Ultra 30K centrifugal filter (Millipore). The cell pellet was resuspended in 200 l of SUME protein lysis buffer (1% SDS, 8 M urea, 10 mM MOPS, 10 mM EDTA, 1 mM 4-benzenesulfonyl fluoride hydrochloride, 1ϫ Yeast/Fungal Protease Arrest TM mixture (G-Biosciences catalog no. 786-333)) and lysed with glass beads by eight 1-min alternate cycles of agitation on a Vortex mixer and incubation on ice. Centrifugation at 2,000 ϫ g for 5 min removed unbroken cells and debris. Protein concentration in the concentrated supernatant and the cell lysate was determined by a Bradford assay kit (Thermo Fisher).
Sucrose gradient separation was performed as described (50,51). A 100-ml culture was grown to log phase, and 2 ml of 0.5 M NaN 3 was added at the time of harvest. The cells were collected by centrifugation and broken by glass bead lysis in a 10 mM Tris-HCl, pH 7.5, 1 mM EDTA, 10% sucrose solution. Cell lysates were cleared of debris by low speed centrifugation, and the resulting lysate was layered onto an 11-ml gradient of 30 -70% sucrose in the lysis buffer. The gradient was centrifuged at 100,000 ϫ g in a Beckman SW41 rotor for 18 h at 4°C. 1-ml fractions were collected from the top of the gradient by pipetting. Proteins of interest were analyzed by SDS-PAGE and Western blot analysis.
Western Blot Analysis-Total protein extracts were analyzed by SDS-PAGE and Western blotting. An equal amount of protein was loaded onto NuPAGE BisTris gels (Invitrogen, IM-8042) and transferred to a nitrocellulose membrane using the XCell Blot Module (Invitrogen). Membranes were incu- Sucrose gradient fractions were analyzed by SDS-PAGE and Western blot analysis as described for total protein extracts, but equal volumes of each fraction were run on the SDS-polyacrylamide gel instead of equal amounts of protein. Membranes were probed with primary antibodies against HA-Plb1, Sec61 (1:1,000), and Pma1 (1:4,000). Anti-Pma1 was provided by Dr. Carolyn Slayman (Yale University), and anti-Sec61 was obtained as described (52). The blots were then incubated with the appropriate secondary antibody, HRP-conjugated anti-rabbit or HRP-conjugated anti-mouse. For quantification of Western blots, densitometry was carried out using Kodak 1D software.
Statistical Analysis-GraphPad Prism was used for statistical analysis by Student's t test or analysis of variance as indicated.

Plb1-mediated Extracellular GroPCho Production Increases
upon Loss of YPK1-Plb1 deacylates phosphatidylcholine to produce free fatty acids and extracellular GroPCho (19) (Fig.  1A). To better define the link between Ypk1 and Plb1 activity, turnover of PC and the concomitant production of the GroPCho were examined. Cells grown to uniform labeling with [ 14 C]choline chloride were washed free of label and reinoculated into fresh medium containing 10 mM choline at 37°C. The chase was performed at elevated temperature to maximize turnover, because the Ypk1 pathway is activated by heat stress (1,53) and because PLB1 expression is up-regulated by elevated temperature (see Fig. 4A). These experiments were performed in both the BY4742 and YPH499 backgrounds because of slight differences in phenotypes observed in ypk1⌬ plb1⌬ double mutants constructed in the two backgrounds, as described below. At the indicated times, cells were harvested, the percentages of the total radioactivity in the extracellular, intracellular, and membrane fractions were determined, and the metabolites produced were identified. Radioactivity in the extracellular fraction, which is used to monitor in vivo Plb1 activity, consisted of [ 14 C]GroPCho and free [ 14 C]choline. As shown in Fig.  1, B and C, deletion of YPK1 resulted in increased production of extracellular GroPCho that is ameliorated by deletion of PLB1. This held true for both strain backgrounds, although the magnitude of the change varied from nearly 3-fold for YPH499 to roughly 5-fold for BY4742. The rest of the counts at each time point were in either the membrane or the intracellular fraction ( Table 4).
Examination of the intracellular fractions revealed that there is a decrease in internal GroPCho production in both the ypk1⌬ and the ypk1⌬plb1⌬ strains (Table 4). Internal GroPCho production occurs primarily through Nte1-associated turnover (22), suggesting that there may be a trade-off between Nte1 and Plb1 activity upon loss of Ypk1. Loss of Ypk1 also appears to cause increased extracellular choline. The phospholipase D that hydrolyzes PC into free choline and phosphatidic acid (54), Spo14, could be responsible for the increase. Although quantitative differences in the levels of choline metabolites were observed as a function of strain background, the general trends were consistent. Also, deletion of YPK2, the paralog of YPK1, did not affect GroPCho production (Fig. 1B).
To confirm the results obtained through radiolabeling and to obtain a more complete picture of the extracellular choline metabolites produced, we performed tandem MS. Through radiolabeling, we were able to monitor the turnover of Kennedy pathway-derived PC synthesized during the labeling period. Through MS, we could monitor all GroPCho produced during the experiment through both the PE methylation pathway and through the Kennedy pathway (23). Importantly, the MS data confirmed that GroPCho production rose upon deletion of YPK1 and that Plb1 plays a major role in facilitating the increase (Fig. 1D). Using MS, we also detected a decrease in extracellular GroPCho production by plb1⌬ yeast as compared with wild type (WT) cells that was not apparent through pulse-chase radiolabeling. As indicated by both the MS data and the radiolabeling data, Plb1 cannot be the only phospholipase responsible for extracellular GroPCho production, because strains lacking Plb1 still produce GroPCho. In fact, a strain in which all of the known PLBs are deleted, plb1,2,3⌬nte1⌬, exhibits greatly reduced but not completely eliminated GroPCho production (Fig. 1D), suggesting that one or more deacylating phospholipases have yet to be identified.

Turnover of Another Glycerophospholipid, Phosphatidylinositol, Is Unaffected by Deletion of YPK1-In a WT strain,
glycerophosphoinositol produced through the Plb3-mediated deacylation of PI is the major phospholipid metabolite found in growth medium (21). To monitor this metabolite, WT and ypk1⌬ cells grown to uniform labeling with [ 3 H]inositol were chased with non-radiolabeled inositol, similar to the conditions used in the [ 14 C]choline labeling experiments above. However, we observed no significant difference between WT and ypk1⌬ strains in the rate of PI turnover or in the percentage of watersoluble inositol-containing counts found in extracellular (WT, 31% Ϯ 1; ypk1⌬, 25% Ϯ 2) or intracellular (WT, 22% Ϯ 1; ypk1⌬, 26% Ϯ 3) fractions following a 5-h chase.
The Rate of Short Term Incorporation of [ 14 C]Choline into PC Is Unaffected by Loss of YPK1-Because an increased rate of PC deacylation and GroPCho production was observed upon loss of Ypk1, we next tested if the rate of incorporation of label into PC was affected. The logic underlying this experiment was that Cki1 (choline kinase) and Pct1 (choline phosphate cytidylyltransferase), both of which are involved in PC synthesis by the Kennedy pathway, contain potential consensus Ypk1 phosphorylation sites (RXRXX(S/T), where is a preference for hydrophobic amino acids (9)). To detect changes in the synthesis rate via the Kennedy pathway, we measured the incorporation of [C 14 ]choline into PC following a 30-min pulse, which is short enough to result in minimal lipid turnover. No difference was found between WT and ypk1⌬ yeast (Fig. 2), suggesting that the difference in external GroPCho production is not due to increased PC synthesis but rather to altered patterns of turnover. This result also suggests that choline uptake was not significantly affected by loss of Ypk1, a suggestion substantiated by the fact that similar levels of incorporated label occurred in all strains following the radioactive pulse (data not shown).
In contrast, the Plb1 sequence lacks the Ypk1 consensus phosphorylation site. Furthermore, Plb1 was shown to have a low likelihood value for phosphorylation by Ypk1 in a large scale peptide screen for kinase activity (55). Similarly, the Plb1 sequence does not contain the Fpk1 phosphoacceptor site, RXSLDX 15-21 RXSLXD (1). Thus, direct regulation of Plb1 Values for the internal and membrane metabolites are derived from the experiment described in Fig. 1, A and B Strain Time

Cell-associated counts (percentage of total counts) Total counts Internal Choline Internal GroPCho
Ypk1 Regulates Phosphatidylcholine Turnover through phosphorylation by Ypk1 or its downstream kinase, Fpk1, is unlikely. In addition, a strain bearing a deletion in FPK1 did not significantly alter GroPCho production. The percent-ages of total [ 14 C]choline label incorporated into external GroPCho in the fpk1⌬ and WT strains following an 8-h chase were 8 Ϯ 1 and 5 Ϯ 1%, respectively. Together, these data strongly suggest that Fpk1 is not involved in the regulation of Plb1 by Ypk1. The Secreted and Cell-associated Forms of Plb1 Increase in ypk1⌬ Strain-Plb1 activity has been reported in plasma membrane fractions, in the periplasm, and in the culture supernatant (19,21). However, localization data were absent in large scale analyses using fluorescent tags fused to the C termini of yeast ORFs expressed under control of their endogenous promoters (56 -60) and in a study utilizing C-terminal fusions to an epitope tag with the target genes expressed under control of the GAL promoter (61). In addition, a localization screen of proteins involved in lipid metabolism, in which Plb1 was C-terminally tagged and expressed under the control of the TEF promoter, demonstrated ER and vesicle localization, but plasma membrane localization was not apparent; the medium/secreted fraction was not analyzed in this study (62). Complicating the question of Plb1 localization is that it has a predicted GPI anchor, which should lead to plasma membrane residence (63,64). During processing, GPI-anchored proteins are cleaved just after the anchor attachment site, which removes the C terminus (22). Therefore, if Plb1 were GPI-anchored, then C-terminal  tagging, as used in several of the studies cited above, would be unable to detect the final processed protein.
To establish the localization of Plb1 and to have a reliable epitope-tagged Plb1 for Western analysis, we constructed a WT and ypk1⌬ strain in which HA was inserted into the genomic copy of PLB1 near the N-terminal end (i.e. after the signal sequence but before the catalytic site). This PLB1-I-3xHA construct was judged functional by confirming that ypk1⌬ containing PLB1-I-3xHA displayed elevated GroPCho production indistinguishable from the ypk1⌬ strain (see "Experimental Procedures"; data not shown). Consistent with increased GroPCho production, ypk1⌬ containing the PLB1-I-3xHA allele displayed greater Plb1 abundance in total cell lysates and in the medium as compared with the WT strain (Fig. 3, A and  B). To examine intracellular localization, we performed sucrose gradient fractionation using lysates from WT strains that contained the integrated, functional PLB1-I-3xHA gene. Interestingly, we found Plb1 associated with the ER (as indicated by co-migration with Sec61), as expected for a protein trafficking through the secretory pathway, but the protein was absent from plasma membrane fractions, as indicated by the migration of Pma1 (Fig. 3C). These results, which are consistent with fluorescence microscopy of C-terminally tagged Plb1 under the control of the TEF promoter (62), lead us to conclude that the protein is not plasma membrane-associated and that if it is GPIanchored, it is readily released from the plasma membranes under the conditions used here. We suspect that the cell surface-associated Plb1 activity that releases GroPCho into the medium is due to its periplasmic localization.
Although Plb1 is predicted to be 72 kDa, the species found in the lysate migrates at ϳ145 kDa. Although the 145-kDa species is the most abundant form in the medium, a less abundant, diffuse band at ϳ260 kDa is also present. Plb1 is known to be N-glycosylated (65,66), and multiple glycosylated orms of secreted phospholipase B have been detected previously (19,67,FIGURE 4. PLB1 message levels are elevated by heat stress and upon disruption of sphingolipid homeostasis. A, strains in the BY4742 background were grown to early log phase in synthetic medium at 30°C. Cultures were split and incubated at 30 or 37°C, as indicated, for 3 h. B, cells were grown in synthetic medium at 30°C to an A 600 of ϳ0.7, myriocin (dissolved in methanol) was added to a final concentration of 1 g/ml, and cultures were incubated for 2 h at 30°C. C, cells were grown in synthetic medium at 30°C to an A 600 of ϳ0.4, aureobasidin (dissolved in methanol) was added to a final concentration of 1 g/ml, and the cells were incubated for 3 h at 30°C. A-C, after harvesting the cells, total RNA was extracted, and PLB1 transcript levels were analyzed by real-time qRT-PCR using the comparative ⌬⌬C T method. Data were normalized to the amount of the endogenous control mRNA, SNR17, and expressed relative to the value in WT control cells. Values represent the means Ϯ S.E. (error bars) of at least three independent cultures assayed in triplicate. D, simplified schematic of sphingolipid synthesis in S. cerevisiae including targets of inhibiting drugs. E, WT cells in the BY4742 background were pregrown at 30°C in synthetic medium containing 0.2 Ci/ml [ 14 C]choline chloride, followed by a chase at 30°C in synthetic medium containing 10 mM choline and 1 g/ml myriocin in methanol or an equivalent volume of methanol. Extracellular, intracellular, and membrane fractions were isolated, and [ 14 C]choline-containing metabolites from extracellular fractions after 3 h were separated by anion exchange chromatography. Values are the means Ϯ S.E. of at least duplicate cultures and are representative of multiple experiments. Similar trends were observed in the YPH499 background (data not shown). F, the lcb1-100 and WT strains were grown in synthetic medium at room temperature (25°C) to an A 600 of ϳ0.4 and then shifted to the restrictive temperature, 37°C, for 90 min. RNA was extracted and analyzed as in A-C. Student's t test was performed for statistical analysis (B, p ϭ 0.05; C, p ϭ 0.008; F, p ϭ 0.0006). 68). Upon treatment with Endo H f , the protein migrates as one species at ϳ72 kDa (data not shown).
Compromised Sphingolipid Synthesis Up-regulates PLB1 Expression and Increases PC Turnover-Consistent with the observed increase in PC deacylation and Plb1 abundance (Figs.  1 and 3), PLB1 transcript levels also increased upon loss of YPK1 at both 30°C and after a 37°C temperature shift (Fig. 4A). In both WT and ypk1⌬ yeast, a shift to 37°C increased PLB1 message levels. The ypk1⌬ strain in the YPH499 background displayed a similar trend (data not shown).
The major role for Ypk1 described in the literature is the regulation of sphingolipid homeostasis (2). If the mechanism by which loss of Ypk1 up-regulates Plb1 expression involves compromised sphingolipid synthesis, we would expect reduction of sphingolipid synthesis by other means to have a similar effect on Plb1 expression. Indeed, PLB1 expression was elevated in a WT strain treated with either myriocin or aureobasidin (Fig. 4, B and C), which inhibit SPT and inositol phosphoceramide synthase, respectively (69, 70) (Fig. 4D). We also monitored PC turnover in WT cells treated with myriocin and saw the expected increase in external GroPCho production (Fig. 4E). Similar trends were observed in a YPH499 strain background. Sphingolipid synthesis can also be diminished genetically through the use of the lcb1-100 strain, which harbors temperature-sensitive SPT activity (71). As anticipated, the lcb1-100 strain shifted to the restrictive temperature also exhibited increased PLB1 expression as compared with WT (Fig. 4F).
The Crz1 Transcription Factor Is Responsible for PLB1 Upregulation either When Sphingolipid Biosynthesis Is Compromised or When Ypk1 Is Absent-A recent microarray study showed that many of the up/down-regulated genes, including PLB1, identified in a ypk1⌬ ypk2⌬ strain bearing an inactivated analog-sensitive allele of YPK1 (YPK1 as ) contain binding sites for the Crz1 transcription factor (18,19). Thus, we examined the role of Crz1 in mediating PLB1 expression in response to altered sphingolipid biosynthesis or when YPK1 was deleted. A ypk1⌬ mutant exhibited increased PLB1 message levels as compared with WT yeast, whereas PLB1 message levels were unchanged in a ypk1⌬crz1⌬ double mutant (Fig. 5A). In addition, up-regulation of PLB1 expression in response to myriocin treatment (Fig. 4B) observed in a WT strain did not occur in the crz1⌬ strain (Fig. 5B). These data indicate that Crz1 is required for the up-regulation of PLB1 expression that occurs upon loss of Ypk1.
Ameliorating Sphingolipid Synthesis by Deletion of ORM1 and ORM2 in a ypk1⌬ Strain Returns PLB1 Expression and Extracellular GroPCho Production to WT Levels-A primary mechanism by which Ypk1 regulates sphingolipid synthesis is through phosphorylation of Orm1 and Orm2 (17,36). Orm1 and Orm2 are ER transmembrane proteins that regulate sphingolipid biosynthesis by binding to and inhibiting SPT (17,18). Upon phosphorylation by Ypk1, the interaction between Orm1/2 and SPT is lessened, and sphingolipid biosynthesis is up-regulated. Thus, if Ypk1 regulates Plb1 via an Orm1/2-dependent effect on sphingolipid synthesis, we would expect PLB1 expression to mirror WT levels in an orm1⌬orm2⌬ypk1⌬ strain, as was found (Fig. 6A). Consistent with this result, a ypk1⌬ strain displays greater extracellular GroPCho production, but the levels of this metabolite nearly return to WT levels in the orm1⌬orm2⌬ypk1⌬ strain (Fig. 6B).
To determine whether the Orm2 phosphorylation state regulates PLB1 expression or GroPCho production, two other mutant strains, orm1⌬orm2 Ala (dephosphorylated mimic) and orm1⌬orm2 Asp (phosphorylated mimic), harboring or lacking the YPK1 gene were employed. Specifically, these strains contain ORM2 that is mutated to alanine or aspartic acid at the Ypk1 phosphorylation sites, Ser-46, Ser-47, and Ser-48, so they mimic the constitutively dephosphorylated or phosphorylated states (2,36). These phosphorylation sites were shown to be immediately but transiently phosphorylated in response to heat stress in a Ypk1-dependent manner (36). Orm2 has also been shown to be the major Orm species based upon its greater expression than Orm1 and because only the orm2⌬ strain displays a growth defect (36). In our experiments, performed over the course of several h, the orm1⌬orm2⌬, orm1⌬orm2 Ala , and orm1⌬orm2 Asp strains all have similar PLB1 message levels and extracellular GroPCho production as compared with the WT strain (Fig. 6, A and B). In addition, the orm1⌬orm2 Asp ypk1⌬ strain behaves similarly to the ypk1⌬ strain. These results suggest that Orm2 phosphorylation at these specific amino acids is FIGURE 5. Crz1 mediates PLB1 up-regulation either when sphingolipid biosynthesis is compromised or when Ypk1 is absent. A, strains (BY4742) were grown in synthetic medium at 30°C to an A 600 of ϳ0.7 (log phase). An aliquot was diluted to an A 600 of ϳ0.2 and grown for 3 h at 37°C. B, WT and crz1⌬ strains were grown in synthetic medium at 30°C to an A 600 of ϳ0.7, myriocin (dissolved in methanol) was added to a final concentration of 1 g/ml, and cells were incubated for 3.5 h at 30°C. A and B, RNA was extracted, and PLB1 message levels were analyzed by qRT-PCR using the comparative ⌬⌬C T method. Data were normalized to the amount of the endogenous control mRNA, SNR17, and expressed relative to the value in WT cells grown at 30°C (A) and WT cells grown with vehicle, methanol (B). Values represent the means Ϯ S.E. (error bars) of three independent cultures assayed in triplicate. Student's t test was performed for statistical analysis; p Յ 0.0001. NOVEMBER 7, 2014 • VOLUME 289 • NUMBER 45 not crucial for PLB1 expression or PC turnover but that the Orm2 protein itself is required.

Ypk1 Regulates Phosphatidylcholine Turnover
Overexpression of only PLB1 rescues ypk1 ts ypk2⌬ lethality at the restrictive temperature. Plb1 was previously identified as a multicopy suppressor of lethality in the ypk1 ts ⌬ strain at the restrictive temperature of 37°C (6). To determine whether rescue was PLB1-specific or could be conferred by other PLB-encoding genes, growth was monitored in the ypk1 ts ypk2⌬ strain overexpressing PLB1, PLB2, PLB3, or NTE1. The genes were expressed from their endogenous promoters on multicopy plasmids (see "Experimental Procedures"). Although there was a 200 -300-fold increase in PLB1, PLB2, and PLB3 message levels and a 50-fold increase in NTE1 message levels, only overexpression of PLB1 rescued ypk1 ts ypk2⌬ lethality at the restrictive temperature (Fig. 7).
Deletion of Both PLB1 and YPK1 Leads to Synthetic Interactions-Based on the data presented in Fig. 7, PLB1 and YPK1 appear to genetically interact. To confirm this hypothesis, we examined specific phenotypes in strains lacking both genes. As shown in Fig. 8A, a negative genetic interaction was observed between YPK1 and PLB1 when growth was monitored,  Overexpression of only PLB1 rescues the lethality of the ypk1 ts ypk2⌬ strain at a restrictive temperature. WT and ypk1 ts ypk2⌬ strains (YPH499) harboring the indicated plasmids were pregrown in selective medium, harvested, and resuspended to equivalent densities. Cells from 5-fold serial dilutions of each culture were spotted onto selective medium and grown for 3 days at the permissive (30°C) and restrictive (35°C) temperatures. FIGURE 8. The slowed growth phenotype and sensitivity to the cell wall perturbing agent, calcofluor white, in a ypk1⌬ strain is exacerbated by deletion of PLB1. A, strains in the BY4742 and YPH499 backgrounds were pregrown in synthetic medium, harvested, and resuspended to equivalent densities. Cells from 5-fold serial dilutions were spotted onto synthetic plates and incubated for 48 (BY4742) or 32 h (YPH499). B, WT and deletion strains were pregrown in synthetic medium, harvested, and resuspended to equivalent densities. 5-fold (BY4742) or 10-fold (YPH499) dilutions were spotted onto YPD and YPD ϩ 8.5 g/ml calcofluor white plates and incubated at 30°C for 32 or 48 h, respectively. Data displayed are representative of multiple experiments.

Ypk1 Regulates Phosphatidylcholine Turnover
although the severity of the defect varied somewhat depending upon the strain background employed. On solid synthetic medium, ypk1⌬plb1⌬ isolates constructed in both the BY4742 or YPH499 backgrounds displayed a growth defect as compared with ypk1⌬ at 30 and 35°C (Fig. 8A). The ypk1⌬plb1⌬ strain constructed in the YPH499 background also exhibited a growth defect on YPD plates (Fig. 8B).
The Ypk1 pathway influences several downstream events, including cell wall integrity (6,7). Therefore, we examined the effect of the cell wall-perturbing agent, calcofluor white, on growth. A ypk1⌬plb1⌬ mutant exhibited increased sensitivity to calcofluor white as compared with a ypk1⌬ strain (Fig. 8B). This result was most striking in the YPH499 background, where no growth was observed in the ypk1⌬plb1⌬ strain in the presence of calcofluor white when present at a final concentration of 8.5 g/ml. These findings, together with the fact that PLB1 is a multicopy suppressor of the lethality of ypk1 ts ypk2⌬ yeast, indicate that increased PC turnover via Plb1 activity increases the fitness of cells when Ypk1 is absent. More generally, these findings suggest that sphingolipid synthesis is coordinated with PC turnover to maintain optimal lipid homeostasis.

DISCUSSION
PC is the most abundant glycerophospholipid in eukaryotic membranes, and its synthesis and catabolism are tightly regulated. In yeast, the known routes of PC catabolism occur through the action of the phospholipase D encoded by PLD1/ SPO14 (54), the ER-localized PLB encoded by NTE1 (22), and PLB1 (19). Here we show that Plb1 activity (Fig. 1) and expression (Figs. 4 and 5) were modulated by the serine/threonine kinase Ypk1. Furthermore, we show that regulation occurs via a mechanism in which the cell can detect compromised sphingo-lipid biosynthesis (Fig. 5). Blockage of SPT activity through use of an lcb1-100 mutant or by treatment with myriocin also results in increased PLB1 expression. Treatment of cells with aureobasidin A, which blocks downstream steps in the sphingolipid biosynthetic pathway, led to a similar effect, indicating that the formation of the complex sphingolipid, inositol phosphoceramide, at a minimum, is required to regulate PLB1 (Fig.  5D).
Plb1 has been widely accepted as a secreted and plasma membrane-associated protein based on cellular fractionation studies and based on its role in the production of extracellular GroPCho (19). Our results, utilizing a Plb1 construct with an internal HA tag to account for a potential GPI anchor, indicate that Plb1 is secreted (Fig. 4A) and fractionates with the ER marker, Sec61, but not with the plasma membrane marker, Pma1 (Fig. 4C). As a secreted and glycosylated protein, it is not unexpected to find Plb1 at the ER. However, our data along with the data of others suggest the intriguing possibility that Plb1 may be active intracellularly as well as at the cell surface, because it has been shown to metabolize excess intracellular lysophospholipids (72) and to be involved in PC acyl chain remodeling (73). The secreted/periplasmic form of Plb1 is most likely responsible for the production of extracellular GroPCho. Notably, extracellular GroPCho is not a dead end metabolite but can be recycled into PC synthesis following its uptake via glycerophosphodiester transporters that have been identified in both S. cerevisiae (74) and Candida albicans (46,75). Ypk1 regulates sphingolipid synthesis by phosphorylating Orm1/2 (2). Phosphorylated Orm1/2 cannot bind to SPT, which results in up-regulated sphingolipid synthesis. Our results show that the increase in PLB1 expression observed in a  strain is mediated by Orm1/2, because PLB1 transcript levels are depressed to WT levels in the orm1⌬orm2⌬ypk1⌬ strain (Fig. 6). Whereas the Orm1/2 proteins trigger the increased PLB1 response upon loss of YPK1, we find that the phosphorylation state of Orm2 at residues Ser-46, Ser-47, and Ser-48 is not involved in controlling PLB1 expression or PC turnover. Interestingly, Orm1 and Orm2 have also been found to physically interact with the ceramide synthase subunit, Lac1, suggesting that Orm1/2 regulate the sphingolipid biosynthetic pathway at multiple steps (76). These data suggest a complicated relationship between Orm1/2 and its role in regulating sphingolipid and phospholipid metabolism.
Phosphorylation of the lipid flippase activator, Fpk1, by Ypk1 down-regulates the Dnf1 and Dnf2 flippases (1). In addition, Fpk1 has been linked to elevated levels of reactive oxygen species that are produced upon loss of Ypk1 (13). Our results do not indicate a connection between Plb1 activity and Fpk1, because an fpk1⌬ strain does not display increased PC turnover, and Plb1 lacks a Fpk1/2 phosphoacceptor site.
We also discovered that the calcineurin-dependent responsive element-associated transcription factor, Crz1, is a key regulator of the increased PLB1 expression observed upon disruption of sphingolipid biosynthesis. Interestingly, Crz1 is not responsible for increased PLB1 transcription observed upon heat stress (Fig. 6A), suggesting multiple levels of PLB1 regulation, depending on the applied stress. Consistent with our findings, a ypk1⌬ypk2⌬ strain bearing a YPK1 as allele displays increased expression of several genes, many that have Crz1 binding sites in their promoters, including PLB1 (10). Also, others have shown that disruption of sphingolipid synthesis leads to increased expression of PMC1, which is a calcineurin-dependent gene (77). Based on these and other data, a model was proposed (10) in which the TORC2 and calcineurin pathways act antagonistically to maintain the optimal cell growth and survival. When Ypk1 is recruited to the plasma membrane by Slm1/2, there is a shift toward down-regulation of the calcineurin pathway and up-regulation of TORC2 signaling. Therefore, in a ypk1⌬ strain, activation of the calcineurin pathway would predominate, activating the downstream transcription factor, Crz1, which in turn activates various calcineurin-dependent responsive element-containing genes, including PLB1 (Fig. 9). Interestingly, treatment of cells with myriocin results in TORC2-dependent phosphorylation of Ypk1 in a manner that requires the Ypk1-recruiting proteins, Slm1/2, to translocate from eisosomes to membrane compartments containing TOR (11,12). Because Slm1 relocalization was suggested to be a result of an altered sphingolipid environment and an increase in reactive oxygen species (13), it is tempting to speculate that alterations in the eisosome lipid environment due to Plb1-mediated PC turnover may also contribute to Slm1 relocalization.
Our phenotypic analysis of the ypk1⌬plb1⌬ double mutant suggests that up-regulated Plb1 activity upon loss of Ypk1 is a functionally important compensatory mechanism, because slowed growth and sensitivity to calcofluor white are exacerbated upon deletion of PLB1 in the ypk1⌬ strain (Fig. 8). Two lines of evidence indicate Plb1 activity, as opposed to activity of the other PLBs, mediates this phenomenon: 1) only overexpression of PLB1 rescues growth of the ypk1 ts ypk2⌬ strain at a restrictive temperature (Fig. 8), and 2) turnover of PI, the primary substrate of Plb3, is unaltered in the ypk1⌬ strain. The mechanism by which Plb1 activity rescues phenotypes associated with loss of the Ypk1/2 proteins has not been elucidated and will be the focus of future work. However, one possibility is that PC levels, which are in turn modulated through Plb1 activity, affect the activities of essential plasma membrane proteins. It is also possible that the metabolic products of Plb1, GroPCho and free fatty acids, are responsible for the rescued phenotypes. Alternatively, as others have suggested, overexpression of PLB1 could induce greater damage to cell wall integrity, resulting in activation of additional cell wall integrity recovery pathways to facilitate rescue (6).
In summary, our results indicate that Ypk1, in addition to regulating sphingolipid synthesis and plasma membrane flippase activity, also regulates the turnover of the major phospholipid, PC. Future studies will focus on the mechanism by which altered sphingolipid biosynthesis, as engendered by loss of Ypk1, activates the myriad genes under the control of Crz1, such as Plb1.