Phosphorylation of Phosphatidate Phosphatase Regulates Its Membrane Association and Physiological Functions in Saccharomyces cerevisiae

The Saccharomyces cerevisiae PAH1-encoded phosphatidate phosphatase (PAP) catalyzes the penultimate step in the synthesis of triacylglycerol and plays a role in the transcriptional regulation of phospholipid synthesis genes. PAP is phosphorylated at multiple Ser and Thr residues and is dephosphorylated for in vivo function by the Nem1p-Spo7p protein phosphatase complex localized in the nuclear/endoplasmic reticulum membrane. In this work, we characterized seven previously identified phosphorylation sites of PAP that are within the Ser/Thr-Pro motif. When expressed on a low copy plasmid, wild type PAP could not complement the pah1Δ mutant in the absence of the Nem1p-Spo7p complex. However, phosphorylation-deficient PAP (PAP-7A) containing alanine substitutions for the seven phosphorylation sites bypassed the requirement of the phosphatase complex and complemented the pah1Δ nem1Δ mutant phenotypes, such as temperature sensitivity, nuclear/endoplasmic reticulum membrane expansion, decreased triacylglycerol synthesis, and derepression of INO1 expression. Subcellular fractionation coupled with immunoblot analysis showed that PAP-7A was highly enriched in the membrane fraction. In fluorescence spectroscopy analysis, the PAP-7A showed tighter association with phospholipid vesicles than wild type PAP. Using site-directed mutagenesis of PAP, we identified Ser602, Thr723, and Ser744, which belong to the seven phosphorylation sites, as the sites phosphorylated by the CDC28 (CDK1)-encoded cyclin-dependent kinase. Compared with the dephosphorylation mimic of the seven phosphorylation sites, alanine substitution for Ser602, Thr723, and/or Ser744 had a partial effect on circumventing the requirement for the Nem1p-Spo7p complex.

In the yeast Saccharomyces cerevisiae, the PAH1-encoded phosphatidate phosphatase (PAP) 2,3 catalyzes the dephosphor-ylation of PA, yielding DAG and P i (1,2). This reaction is dependent on Mg 2ϩ ions and is based on a DXDX(T/V) catalytic motif within a haloacid dehalogenase-like domain in the enzyme (2)(3)(4). PAP is associated with the cytosolic and membrane fractions of the cell, and the association with the membrane is peripheral in nature (2). Chromatin immunoprecipitation analysis indicates that PAP is also localized in the nucleus (5). The DAG generated in the PAP reaction is used for the synthesis of TAG (2) and for the synthesis of phosphatidylethanolamine and phosphatidylcholine via the CDPethanolamine and CDP-choline branches, respectively, of the Kennedy pathway (4,6). The enzyme also plays a major role in controlling the cellular concentration of its substrate PA (2), the precursor of phospholipids that are synthesized via the CDP-DAG pathway (6 -8). In addition, the substrate PA plays a signaling role in the transcriptional regulation of phospholipid synthesis genes (9). In fact, mutants defective in PAH1-encoded PAP activity exhibit a Ͼ90% reduction in TAG content, a derepression of phospholipid synthesis genes, and an expansion of the nuclear/ER membrane (3,5). Thus, the regulation of PAP activity governs the synthesis of TAG, the pathways by which phospholipids are synthesized, PA signaling, and the growth of the nuclear/ER membrane (6).
The importance of PAP in lipid metabolism and cell physiology is further emphasized by the fact that the overexpression of Lpin1-encoded PAP (also known as lipin 1) in mice leads to obesity and insulin sensitivity, whereas loss of lipin 1 prevents normal adipose tissue development, resulting in lipodystrophy and insulin resistance (10,11). Moreover, mice lacking PAP activity exhibit peripheral neuropathy (12)(13)(14) caused by degradation of myelin through the MEK/ERK signaling pathway that is activated by elevated levels of PA (14).
In humans, mutations in LPIN1-encoded PAP are associated with metabolic syndrome, type 2 diabetes, and recurrent acute myoglobinuria in children, whereas mutations in LPIN2-encoded PAP (also known as lipin 2) are the basis for the anemia and inflammatory disorders associated with the Majeed syndrome (15)(16)(17)(18).
PAP is subject to the covalent modification of phosphorylation (19,20). The yeast enzyme has been identified in proteome-wide in vitro phosphorylation analyses to be a target for multiple protein kinases, including those encoded by CDC28 (CDK1) (21), PHO85 (22,23), and DBF2 (24). Mass spectrometry analysis of purified PAP, in combination with immunoblot analysis using anti-MPM2 antibodies that recognize phosphorylated serine and threonine residues in a Ser/ Thr-Pro motif, has identified 16 sites of phosphorylation, seven of which (Ser 110 , Ser 114 , Ser 168 , Ser 602 , Thr 723 , Ser 744 , and Ser 748 ) are contained within the minimal Ser/Thr-Pro motif that is a target for protein kinases regulated during the cell cycle (25). Moreover, data indicate that PAP is phosphorylated by CDC28 (CDK1)-encoded CDK in a cell cycle-dependent manner (5). The nine remaining sites are putative targets for protein kinases, such as protein kinases A and C, casein kinases I and II, and MAPK, indicating that the regulation of PAP by phosphorylation is complex.
The simultaneous mutation of the seven sites within the Ser/Thr-Pro motif to a nonphosphorylatable alanine residue (also known as 7A mutations) results in a 1.8-fold increase in PAP activity (25). In addition, the overexpression of the PAP-7A mutant enzyme causes inositol auxotrophy by alleviating the PA-mediated inhibition of Opi1p repressor activity on INO1, the gene that encodes inositol-3-phosphate synthase (9,25). Moreover, cells that lack the Nem1p-Spo7p protein phosphatase complex, which is responsible for the dephosphorylation of PAP, show phenotypes characteristic of cells lacking PAP activity (5,26). These observations support the conclusion that phosphorylation of the seven sites negatively regulates PAP function in vivo (25). Due to the importance of these phosphorylations in controlling PAP function, the major aims of this work were 1) to further characterize the physiological consequences of the seven sites of phosphorylation and 2) to establish that CDC28 (CDK1)encoded CDK is in fact the relevant protein kinase responsible for these phosphorylations. We showed that lack of phosphorylation at the seven sites caused a great increase in the amount of PAP associated with membranes and that this correlated with a significant increase in TAG synthesis in cells lacking the Nem1p-Spo7p protein phosphatase complex. We also showed that among the seven sites of phosphorylation, only Ser 602 , Thr 723 , and Ser 744 were targets of CDC28 (CDK1)-encoded CDK. Moreover, mutations of these sites, individually and in combination, partially mimicked the physiological consequences of the 7A mutations.

EXPERIMENTAL PROCEDURES
Materials-All chemicals were reagent grade or better. Growth medium supplies were obtained from Difco. New England Biolabs was the source of modifying enzymes, recombinant Vent DNA polymerase, restriction endonucleases, and recombinant human CDK1-cyclin B complex. The DNA gel extraction kit, plasmid DNA purification kit, and nickelnitrilotriacetic acid-agarose resin were purchased from Qiagen. Sigma-Aldrich was the source of aprotinin, benzamidine, bovine serum albumin, leupeptin, pepstatin, phenylmethylsulfonyl fluoride, phosphoamino acids, L-1-tosylamido-2-phenylethyl chloromethyl ketone-trypsin, the CDK peptide substrate, and Triton X-100. PCR primers were prepared commercially by Genosys Biotechnologies. Tobacco etch virus protease was purchased from Invitrogen. The QuikChange site-directed mutagenesis kit was purchased from Stratagene. Carrier DNA for yeast transformation was from Clontech. GE Healthcare supplied IgG-Sepharose, polyvinylidene difluoride paper, and the enhanced chemifluorescence Western blotting detection kit. DNA size ladders, electrophoresis reagents, immunochemical reagents, molecular mass protein standards, and protein assay reagents were from Bio-Rad. Lipids were from Avanti Polar Lipids, and thin layer chromatography plates (cellulose and silica gel 60) were from EM Science. Scintillation counting supplies and acrylamide solutions were from National Diagnostics, and radiochemicals were PerkinElmer Life Sciences. Alkaline phosphatase-conjugated goat anti-rabbit IgG antibodies were from Thermo Scientific. Mouse anti-phosphoglycerate kinase antibodies and alkaline phosphatase-conjugated goat anti-mouse IgG antibodies were from Invitrogen and Pierce, respectively.
Strains and Growth Conditions-The strains used in this work are listed in Table 1. The pah1⌬ nem1⌬ mutant was constructed by transforming a pah1⌬::TRP1 fragment (5) into a nem1⌬::HIS3 mutant (26). The nem1⌬::HIS3 spo7⌬::HIS3 HEH2-CHERRY::TRP1 mutant was constructed by integration of plasmid YIplac204-HEH2-CHERRY at TRP1 in the nem1⌬::HIS3 spo7⌬::HIS3 mutant (26). The homologous recombinations were verified by PCR analysis. Plasmids were propagated in Escherichia coli strain DH5␣. E. coli cells were grown at 37°C in 1% tryptone, 0.5% yeast extract, 1% NaCl, pH 7 (LB medium). Ampicillin (100 g/ml) was added to select for cells carrying plasmids. PAP was expressed in E. coli BL21(DE3)pLysS cells bearing wild type and mutant PAH1 derivatives of plasmid pET-15b as described by Han et al. (27). Yeast cultures were generally grown in standard synthetic complete medium. Appropriate amino acids were omitted from the synthetic growth medium to select for cells carrying specific plasmids. Cells were also cultured in inositollacking synthetic medium (28) containing 2% glucose or 2% galactose. Inositol (75 M) and choline (1 mM) were added to this growth medium where indicated. Cell numbers in liquid cultures were determined spectrophotometrically at an absorbance of 600 nm. The growth medium was supplemented with agar (2% for yeast or 1.5% for E. coli) for growth on plates. For the heterologous expression of wild type and mutant forms of PAH1-encoded PAP enzymes, E. coli BL21(DE3)pLysS cells bearing pET-15b-based plasmids were grown to A 600 ϭ 0.5 at room temperature in 500 ml of LB medium containing ampicillin (100 g/ml) and chloramphenicol (34 g/ml). Cultures were incubated for 1 h with 1 mM isopropyl-␤-D-thiogalactoside to induce the expression of His 6 -tagged PAP enzymes. DNA Manipulations, PCR Amplification of DNA, Site-directed Mutagenesis, and DNA Sequencing-Standard protocols were used to isolate genomic and plasmid DNA, to digest and ligate DNA, and to amplify DNA by PCR (29,30). DNA sequencing was performed by GENEWIZ, Inc. (South Plainfield, NJ) using Applied Biosystems BigDye version 3.1. The reactions were then run on an Applied Biosystems 3730xl DNA Analyzer. Site-specific mutations were generated with the QuikChange site-directed mutagenesis kit using appropriate primers and plasmid templates, and mutant constructs were confirmed by DNA sequencing. Transformation of E. coli (29) and yeast (31,32) with plasmids was performed as described previously.

Preparation of Cell Extracts and Subcellular Fractions and
Purification of Enzymes-All steps were performed at 4°C. Cell extracts were prepared by disruption of yeast cells with glass beads (0.5-mm diameter) using a BioSpec Products Mini-BeadBeater-16 (34). The cell disruption buffer contained 50 mM Tris-HCl (pH 7.5), 0.3 M sucrose, 10 mM 2-mercaptoethanol, 0.5 mM phenylmethanesulfonyl fluoride, 1 mM benzamidine, 5 g/ml aprotinin, 5 g/ml leupeptin, and 5 g/ml pepstatin. The cytosolic (supernatant) and total membrane (pellet) fractions were prepared by centrifugation at 100,000 ϫ g for 1 h (34). The membrane pellets were suspended in the disruption buffer to the same volume of the cytosolic fraction. Protein concentration was estimated by the method of Bradford (35) using bovine serum albumin as the standard. His 6 -tagged wild type and mutant PAP enzymes expressed in E. coli were purified by affinity chromatography using nickel-nitrilotriacetic acid-agarose as described by Han et al. (2). Protein A-tagged wild type and 7A mutant forms of PAP expressed in S. cerevisiae were purified by affinity chromatography using IgG-Sepharose as described previously (25). The CDK-cyclin B complex was purified from S. cerevisiae that expressed protein A-tagged CDK by affinity chromatography using IgG-Sepharose (26). Elution of the untagged PAP enzymes and CDK from IgG-Sepharose columns was achieved by treatment with tobacco etch virus protease (36). SDS-PAGE analyses showed that the purified preparations of the wild type and mutant PAP enzymes and CDK-cyclin B complex were nearly homogeneous.
Phosphorylation Reactions-Phosphorylation reactions were routinely performed in triplicate for 5-10 min at 30°C in a total volume of 20 l. The standard reaction mixture contained 25 mM Tris-HCl (pH 7.5), 10 mM MgCl 2 , 2 mM dithiothreitol, 20 M [␥-32 P]ATP (2,400 cpm/pmol), 50 g/ml PAP, and yeast (1 g) or recombinant human (20 ng) CDK. At the end of the phosphorylation reactions, samples were treated with 4ϫ Laemmli sample buffer (37), subjected to SDS-PAGE, and transferred to polyvinylidene difluoride membranes. Phosphorimaging was used to visualize phosphorylated enzyme, and the extent of phosphorylation was quantified with ImageQuant software. For some experiments, the CDK reactions were terminated by spotting the reaction mixtures onto P81 phosphocellulose paper. The papers were washed three times with 75 mM phosphoric acid and then subjected to scintillation counting. A unit of CDK activity was defined as the amount of enzyme that catalyzed the formation of 1 nmol of phosphorylated product/min. SDS-PAGE and Immunoblot Analysis-SDS-PAGE (37) and immunoblotting (38) with polyvinylidene difluoride membrane were performed by standard protocols. Anti-PAP antibodies were prepared in rabbits against the C-terminal portion (residues 778 -794) of the protein at BioSynthesis, Inc. Rabbit anti-PAP antibodies, rabbit anti-phosphatidylserine synthase antibodies (39), and mouse anti-phosphoglycerate kinase antibodies were used at a concentration of 2 g/ml. Alkaline phosphatase-conjugated goat anti-rabbit IgG antibodies and goat anti-mouse IgG antibodies were used at a dilution of 1:5,000. Immune complexes were detected using the enhanced chemifluorescence Western blotting detection kit. Fluorimaging was used to acquire images from immunoblots, and the relative densities of the images were analyzed using ImageQuant software. Signals were in the linear range of detectability.
Phosphoamino Acid and Phosphopeptide Mapping Analyses-For phosphoamino acid analysis, 32 P-labeled PAP on polyvinylidene difluoride membrane was subjected to acid hydrolysis with 6 N HCl, followed by two-dimensional electrophoresis on cellulose thin-layer chromatography plates (40,41). Phosphorimaging was used to visualize 32 P-labeled phosphoamino acids, whereas ninhydrin spraying was used to visualize standard phosphoamino acids (41). For phosphopeptide mapping analysis, 32 P-labeled PAP on polyvinylidene difluoride membrane was subjected to proteolytic digestion with L-1-tosylamido-2-phenylethyl chloromethyl ketone-trypsin, followed by electrophoresis and TLC using cellulose thinlayer chromatography plates (42). Radioactive phosphopeptides were visualized by phosphorimaging analysis.
Preparation of 32 P-labeled PA and Measurement of PAP Activity-[ 32 P]PA was synthesized enzymatically from DAG and [␥-32 P]ATP with E. coli DAG kinase, and the radioactive product was purified by thin-layer chromatography (34). PAP activity was measured by following the release of water-soluble 32 P i from chloroform-soluble [ 32 P]PA (10,000 cpm/nmol) (34). The reaction mixture contained 50 mM Tris-HCl buffer (pH 7.5), 1 mM MgCl 2 , 0.2 mM PA, 2 mM Triton X-100, and enzyme protein in a total volume of 0.1 ml. All enzyme assays were conducted in triplicate at 30°C. The average S.D. value of the assays was Ϯ5%. The reactions were linear with time and protein concentration. A unit of PAP activity was defined as the amount of enzyme that catalyzed the formation of 1 nmol of product/min.
Labeling and Analysis of Lipids-Steady-state labeling of lipids with [2-14 C]acetate was performed as described previously (43). Lipids were extracted from labeled cells by the method of Bligh and Dyer (44). Lipids were analyzed by one-dimensional thin-layer chromatography on silica gel plates using the solvent system hexane/ethyl ether/acetic acid (40: 10:1) (45). The identity of the labeled TAG and total phospholipids on thin-layer chromatography plates was confirmed by comparison with standards after exposure to iodine vapor. Radiolabeled lipids were visualized by phosphorimaging analysis, and their relative quantities were analyzed using Image-Quant software.
Preparation of Unilamellar Phospholipid Vesicles-Unilamellar phospholipid vesicles were prepared with dioleoylphosphatidylcholine and dioleoyl-PA at a molar ratio of 10:1 by the lipid extrusion method of MacDonald et al. (46). Chloroform was evaporated from phospholipids under nitrogen to form a thin film. The phospholipids were then resuspended in 20 mM Tris-HCl buffer (pH 7.5) containing 150 mM NaCl and 1 mM EDTA. After five cycles of freezing and thawing, the phospholipid suspension was extruded 11 times through a polycarbonate filter (100-nm diameter).
Light Microscopy-The expression of the HEH2-CHERRY fusion gene in wild type and the nem1⌬ spo7⌬ mutant was used to visualize nuclear morphology. Images were acquired on an epifluorescence microscope consisting of an inverted microscope (Zeiss Axiovert 200M), a camera (Hamamatsu Orca ER cooled CCD type), and a 100ϫ plan-apochromatic 1.4 numerical aperture objective lens (Carl Zeiss Ltd.). The microscope was controlled by the Improvision OpenLab software (version 5). mCherry was recorded with Carl Zeiss filter set 00 (488000-0000-000) (excitation BP 530 -585 nm, emission LP 615). The brightness and contrast of the images were adjusted using Adobe Photoshop software.
Fluorescence Measurements-Fluorescence measurements were carried out in a FluoroMax-3 fluorimeter (HORIBA Jobin Yvon Inc.) at room temperature in 200 l of 20 mM Tris-HCl buffer (pH 7.5) containing 150 mM NaCl, 105 nM PAP, and the indicated concentrations of phospholipid vesicles. The excitation wavelength was 280 nm, and the emission spectra were collected from 300 -450 nm after a 10-min incubation period.
Analyses of Data-Kinetic data were analyzed according to the Michaelis-Menten and Hill equations using the Enzyme Kinetics module of SigmaPlot software. Dissociation constants for the interaction of PAP with phospholipid vesicles were determined by the method of Lear and DeGrado (47). Statistical analyses were performed with SigmaPlot software.
The p values of Ͻ0.05 were taken as a significant difference.

Phosphorylation of PAP by CDK-Data indicate that CDC28
(CDK1)-encoded CDK might be the protein kinase responsible for the phosphorylation of the seven sites within the Ser/ Thr-Pro motif of PAP previously shown to be phosphorylated in vivo (5,25). To examine the phosphorylation of PAP by CDK in vitro, we utilized a preparation of PAP that was heterologously expressed in E. coli (2). In this manner, our phosphorylation studies were conducted with a pristine substrate that was free from the endogenous phosphorylations that occur when PAP is expressed in S. cerevisiae (25). Purified PAP was incubated with yeast CDK in the presence of [␥-32 P]ATP, and its phosphorylation was monitored by following the incorporation of the radioactive ␥-phosphate into the enzyme. Phosphorimaging analysis of reaction products resolved by SDS-PAGE showed that purified PAP was a substrate for yeast CDK (Fig. 1A). In addition, PAP was also phosphorylated by recombinant human CDK1 (CDC2)-encoded protein kinase (Fig. 1A), which is functionally homologous to yeast CDC28 (CDK1)-encoded protein kinase (48,49). Phosphopeptide mapping analysis showed similar patterns, indicating that the yeast and human CDK enzymes phosphorylated PAP on the same sites (Fig. 1B). Phosphoamino acid analysis of the 32 P-labeled protein derived from phosphorylation by human CDK showed that PAP was phosphorylated at both serine and threonine residues (Fig. 1C). Due to the fact that human CDK had a 10-fold higher specific activity (as determined with the CDK peptide substrate PKTPKKAKKL) and was more stable to storage when compared with the yeast CDK preparation, we utilized human CDK for the remainder of this work.
Using PAP as substrate, CDK activity was dependent on the time of the reaction ( Fig. 2A) and the concentrations of the kinase (Fig. 2B), ATP (Fig. 2C), and PAP (Fig. 2D). The dependences of CDK activity on ATP and PAP followed saturation kinetics and positive cooperative kinetics, respectively. Analyses of the data according to the Michaelis-Menten and Hill equations showed that the K m values for ATP and PAP were FIGURE 1. PAP is phosphorylated by yeast and human CDK enzymes. A, purified recombinant PAP (50 g/ml) was phosphorylated with yeast (1 g) or human (20 ng) CDK and 20 M [␥-32 P]ATP (2,400 cpm/pmol) for 10 min, followed by SDS-PAGE, transfer to polyvinylidene difluoride membrane, and phosphorimaging analysis. B, polyvinylidene difluoride membranes containing 32 P-labeled PAP that were phosphorylated with yeast or human CDK were treated with L-1-tosylamido-2-phenylethyl chloromethyl ketone-trypsin. The resulting peptides were separated on cellulose thin layer plates by electrophoresis (from left to right) in the first dimension and by chromatography (from bottom to top) in the second dimension. The major (labeled 1 and 2) and minor (labeled 3) phosphopeptides were present in the maps of PAP phosphorylated by both CDK enzymes. The radioactive spots labeled 3 in the map for the yeast CDK phosphorylation were observed when the image was overexposed (not shown). C, polyvinylidene difluoride membrane containing 32 P-labeled PAP phosphorylated with human CDK was hydrolyzed with 6 N HCl for 90 min at 110°C, and the hydrolysate was separated by two-dimensional electrophoresis. The positions of phosphorylated PAP and the standard phosphoamino acids phosphoserine (P-Ser), phosphothreonine (P-Thr), and phosphotyrosine (P-Tyr) (dotted lines) are indicated. The data shown are representative of three independent experiments. 5.8 and 0.21 M, respectively, and that the Hill number for PAP was 2. These data indicated that PAP was a bona fide substrate for CDK.
We examined the stoichiometry of the phosphorylation of PAP by CDK. The enzyme was incubated with [␥-32 P]ATP (2,500 cpm/pmol) and recombinant human CDK for 60 min. Following the incubation, samples were subjected to SDS-PAGE followed by transfer to polyvinylidene difluoride membrane. The amount of phosphate incorporated into PAP was determined by ImageQuant analysis. At the point of maximum phosphate incorporation, CDK catalyzed the incorporation of 0.8 mol of phosphate/mol of PAP. This stoichiometry was low given the results of the peptide mapping experiments that indicated multiple sites of phosphorylation (Fig. 1B). An explanation for this result might be that the phosphorylation of one site inhibited the phosphorylation of another site (50) (see below).
To examine the effect of CDK on PAP activity, the enzyme was preincubated with a constant amount of CDK and unlabeled ATP for various times. Following the phosphorylations, the reaction mixtures were diluted 10-fold and used for the measurement of PAP activity. The timedependent increase in PAP phosphorylation had little effect on its enzyme activity (Fig. 3). Increasing the amounts of CDK in this experiment did not affect the extent of PAP phosphorylation or PAP activity. These results indicated that PAP phosphorylation by CDK had essentially no effect on its enzyme activity.
Identification of Ser 602 , Thr 723 , and Ser 744 as the CDK Phosphorylation Sites in PAP-The seven sites with the Ser/Thr-Pro motif previously shown to be phosphorylated in vivo (25) are putative CDK sites (51), which are located at the N-and C-terminal portions of the protein (Fig. 4A). Individual Ser/ Thr to alanine mutations of the seven sites were constructed by site-specific mutagenesis. Each of the phosphorylation site mutant enzymes was expressed and purified from E. coli and   and threonine (T) residues within a Ser/Thr-Pro motif that were previously identified as sites of phosphorylation (25). B, WT and the indicated phosphorylation site PAP mutant enzymes were expressed and purified from E. coli. The recombinant PAP enzymes were phosphorylated with human CDK (20 ng) and 20 M [␥-32 P]ATP (2,400 cpm/pmol) for 10 min. After phosphorylation, the samples were subjected to SDS-PAGE and transferred to polyvinylidene difluoride membrane. The 32 P-labeled proteins were digested with L-1-tosylamido-2-phenylethyl chloromethyl ketone-trypsin. The resulting peptides were separated on cellulose thin layer plates by electrophoresis (from left to right) in the first dimension and by chromatography (from bottom to top) in the second dimension. The radioactive spots labeled 3 were observed when the images were overexposed (not shown). The positions of the phosphopeptides (labeled 1, 2, and 3) that were absent in the S602A, T723A, S744A, and S602A/T723A mutants (indicated by dotted lines) but were present in the wild type enzyme are indicated. The data are representative of three independent experiments. used as a substrate for CDK. Of the seven PAP mutants, only S602A, T723A, and S744A showed defects in phosphorylation by CDK. As discussed above, the phosphopeptide map of wild type PAP contained two major phosphopeptides (labeled 1 and 2) and one minor phosphopeptide (labeled 3) (Figs. 1B and 4B). Phosphopeptides 1, 2, and 3 were missing in the S602A, T723A, and S744A mutant enzymes, respectively (Fig.  4B). These results indicated that Ser 602 , Thr 723 , and Ser 744 were contained in phosphopeptides 1, 2, and 3, respectively. The phosphopeptides that were attributed to Ser 744 were more evident in the phosphopeptide map of the S602A/ T723A double mutant, which lacked phosphopeptides 1 and 2 (Fig. 4B). The extents of phosphorylation at Ser 602 , Thr 723 , and Ser 744 differed, with Thr 723 being the most phosphorylated and Ser 744 being the least phosphorylated. The S602A and T723A mutations reduced the extent of phosphorylation by 50 and 70%, respectively, whereas the S744A mutation caused only a small decrease in the extent of phosphorylation (Fig. 5). In addition, the phosphopeptide map of the wild type enzyme showed that Thr 723 was more heavily labeled when compared with Ser 602 and that Ser 744 was hardly labeled when compared with the other two sites (Figs. 1B and 4B). These observations indicated that the phosphorylations at Ser 602 and Thr 723 might inhibit the phosphorylation at Ser 744 and may provide an explanation for a stoichi-ometry of less than the theoretical value of 3 (see above). The phosphopeptide maps of the S110A, S114A, S168A, and S748A mutant enzymes were indistinguishable from the map of wild type PAP (data not shown), indicating that Ser 110 , Ser 114 , Ser 168 , and Ser 748 were not sites of phosphorylation for CDK. Individually, these sites were not examined further in this work.
Expression of CDK Phosphorylation Site PAP Mutants in S. cerevisiae-Ser/Thr to alanine mutations were constructed for the CDK phosphorylation sites in PAP to examine the physiological effects of phosphorylation in pah1⌬ and pah1⌬ nem1⌬ mutants. The rationale for using the pah1⌬ nem1⌬ double mutant, which lacks the NEM1-encoded protein phosphatase catalytic subunit (5), was to assess the dependence of PAP function on the Nem1p-Spo7p complex. In addition, this afforded examination of the phosphorylation site mutations in a background that favored the phosphorylation of the other non-mutated phosphorylation sites in the PAP protein. Antibodies directed against the C-terminal portion of PAP recognized the wild type and mutant forms of the enzyme when expressed in pah1⌬ (Fig. 6A) and pah1⌬ nem1⌬ (Fig. 6B) cells. ImageQuant analysis of the immunoblots shown in Fig.  6 indicated that there were no major differences in the relative amounts of the S602A, T723A, and S744A mutant enzymes when compared with the wild type control. However,  the relative amount of the 7A (50%), S602A/T723A (37%), and S602A/T723A/S744A (30%) mutant enzymes was reduced when compared with the wild type control (Fig. 6). Why the phosphorylation state of PAP appeared to govern its relative abundance in the cell was unclear, but it may be related to an unknown mechanism that controls protein stability. As described previously (5,25), the phosphorylation state of PAP affected its electrophoretic mobility; the phosphorylated forms migrated more slowly upon SDS-PAGE (Fig. 6, A and B, compare wild type and 7A). The change in electrophoretic mobility can be attributed to the phosphorylation of Thr 723 (25) (Fig. 6, A and B).
Phosphorylation-deficient PAP Complements the Temperature-sensitive Phenotype of the pah1⌬ nem1⌬ Mutant-Mutants defective in PAP activity exhibit changes in cell physiology that are reflected in growth inhibition at elevated temperature (2,3,5). We examined whether phosphorylationdeficient PAP enzymes expressed via low copy plasmids could complement the temperature-sensitive phenotype of pah1⌬ and pah1⌬ nem1⌬ mutant cells. Wild type PAH1 complemented the temperature sensitivity of the pah1⌬ mutant (Fig. 7A). However, it did not complement this phenotype in the pah1⌬ nem1⌬ mutant (Fig. 7B), indicating the requirement of the Nem1p-Spo7p protein phosphatase complex for PAP function in vivo. In contrast to wild type PAH1, the phosphorylation-deficient 7A allele and, to a lesser extent, the CDK phosphorylation site mutant alleles complemented (e.g. bypassed) the temperature sensitivity of the pah1⌬ nem1⌬ mutant (Fig. 7B).

Phosphorylation-deficient PAP Complements the Nuclear/ER Membrane Expansion Phenotype of the nem1⌬ spo7⌬
Mutant-Cells lacking a functional Nem1p-Spo7p protein phosphatase complex (e.g. nem1⌬, spo7⌬, and nem1⌬ spo7⌬ mutants) exhibit an abnormally expanded structure of the nuclear/ER membrane (26) (Fig. 8A). This phenotype is also exhibited by mutants defective in PAP activity (3,5). We questioned whether low copy expression of wild type and phosphorylation-deficient PAH1 alleles could complement the defect of nuclear morphology exhibited by nem1⌬ spo7⌬ mutant cells. Wild type PAH1 did not complement the phenotype (Fig. 8B). However, normal nuclear morphology was observed when the mutant was transformed with the 7A allele (Fig. 8B), indicating that the lack of phosphorylation of the seven sites bypassed the requirement of the Nem1p-Spo7p protein phosphatase complex. On the other hand, the individual S602A, T723A, and S744A CDK phosphorylation site mutations and the double and triple combinations of these mutations did not complement the nem1⌬ spo7⌬ mutant phenotype (Fig. 8B).
Overexpression of Phosphorylation-deficient PAP Inhibits Growth on Media Lacking Inositol-The INO1-encoded inositol-3-phosphate synthase catalyzes the committed step for the synthesis of inositol in S. cerevisiae (8,52). Cells that express INO1 are prototrophic for inositol, whereas cells with repressed INO1 expression are auxotrophic for inositol (8,52). The repression of INO1 is mediated by the Opi1p repressor, whose function is controlled by its interaction with PA at the nuclear/ER membrane (9,53). Conditions that cause a decrease in PA concentration lead to the dissociation of Opi1p from the membrane and its translocation into the nucleus (9,  53). In the nucleus, Opi1p represses the expression of INO1 by binding to the transcriptional activator Ino2p that is associated with the INO1 promoter (9,53). Previous studies have shown that PAP activity plays a role in controlling PA content and the transcriptional regulation of INO1 as well as other phospholipid synthesis genes (3,5,25). For example, the loss of PAH1 causes an increase in PA content and the derepression of INO1 and other phospholipid synthesis genes (3,5), whereas the massive overexpression (driven by the GAL1/10 promoter) of PAH1-7A causes a decrease in PA content and the repression of INO1, rendering cells auxotrophic for inositol (25) (Fig. 9). The decrease in PA would also decrease the synthesis of phospholipids via the CDP-DAG pathway. Inositol supplementation could only partially complement the inositol auxotrophy caused by the overexpression of the 7A mutant, but supplementation of inositol plus choline fully complemented this growth defect (Fig. 9). The enhanced effect of choline might be attributed to the consumption of the PAP-mediated production of DAG for phospholipid synthesis via the Kennedy pathway as well as to the alleviation of any toxicity that might be caused by the accumulation of DAG. The S602A/T723A/S744A allele partially mimicked the growth properties of the 7A mutant, whereas the individual CDK phosphorylation site mutations did not affect the growth on media lacking inositol (Fig. 9).
Effects of the Phosphorylation-deficient PAP on the Amounts of TAG and Phospholipids-Because PAP catalyzes the penultimate step in the synthesis of TAG, we examined the effects of the 7A and CDK phosphorylation site mutations on TAG content. Because the effects of PAH1-encoded PAP activity on TAG content are most pronounced in stationary phase cells (2,3), our experiments were performed at this phase of growth. The phosphorylation-deficient PAP mutant enzymes were expressed in pah1⌬ and pah1⌬ nem1⌬ cells. Cells were labeled to steady state with [2-14 C]acetate followed by the extraction and analysis of lipids. In pah1⌬ cells expressing wild type PAH1, TAG accounted for 22% of the total 14 C-labeled lipids (Fig. 10A). The TAG content was not significantly affected by the 7A mutations or the CDK phosphorylation site mutations (Fig. 10A). In contrast, when wild type PAH1 was expressed in pah1⌬ nem1⌬ cells, TAG accounted for only 4.2% of the total lipids (Fig. 10B), indicating that the dephos-phorylation of PAP by the Nem1p-Spo7p protein phosphatase complex is required for PAP function in the synthesis of TAG. The decrease in TAG content in these cells was accompanied by an increase in total phospholipids; pah1⌬ nem1⌬ cells expressing PAH1 had 65% phospholipids (Fig. 10B), whereas pah1⌬ cells expressing PAH1 had 36% phospholipids (Fig. 10A).
In pah1⌬ nem1⌬ cells, the expression of the 7A mutant allele caused a 5-fold increase in the amount of TAG when compared with the wild type PAH1 allele (Fig. 10B), and the amount of TAG (22%) in pah1⌬ nem1⌬ cells expressing the 7A mutant allele reached to the level of TAG found in pah1⌬ cells expressing the wild type allele (Fig. 10A). In addition, the expression of the 7A mutant allele in pah1⌬ nem1⌬ cells caused a decrease (from 65 to 44%) in total phospholipids when compared with the wild type allele (Fig. 10B). When expressed in pah1⌬ nem1⌬ cells, the CDK phosphorylation site mutant alleles alone, and in combinations, also caused increases (1.6 -2.9-fold) in the amounts of TAG but not to the same extent as the 7A allele (Fig. 10B). These data indicated that the phosphorylations of PAP at the seven sites had a major effect on TAG synthesis and that the phosphorylations at the three CDK sites played a partial role in this regulation.

Effects of the Phosphorylation Site Mutations on the Localization of PAP-
The effects of the 7A, S602A/T723A, and S602A/T723A/S744A mutations on the localization of PAP were examined by immunoblot analysis of cell fractions derived from pah1⌬ and pah1⌬ nem1⌬ mutant cells (Fig. 11, A  and B). In both genetic backgrounds, most of the wild type PAP enzyme was associated with the cytosolic fraction (82-88%) (Fig. 11C). The 7A mutations had a dramatic effect on the association of PAP with membranes (Fig. 11, A and B). In contrast to the wild type enzyme, most of the 7A mutant enzyme was associated with the membranes of both pah1⌬ (73%) and pah1⌬ nem1⌬ (76%) mutant cells (Fig. 11). In other words, the 7A mutation caused a ϳ10-fold increase in the association of PAP with the membrane. The combinations of the CDK phosphorylation mutations also affected the association of PAP with membranes but not to the same extent as the 7A mutations. The S602A/T723A and S602A/T723A/ S744A mutations caused an increase in membrane association by 2-3.6-fold and by 3.4 -4-fold, respectively (Fig. 11C). The individual CDK phosphorylation site mutations did not have a significant effect on the membrane association of PAP (data not shown).
Effects of the Phosphorylation Site Mutations on the Interaction of PAP with Phospholipid Vesicles-To gain further insight into the effect of phosphorylation on PAP association with membranes, we used a phospholipid vesicle binding assay based on fluorescence emission. For this study, we used the wild type and 7A mutant enzymes that were expressed in S. cerevisiae and were phosphorylated in vivo (25). The interaction of PAP with PA-containing phospholipid vesicles results in an increase in fluorescence emission intensity of tryptophan residues as well as a shift from 350 to 343 nm in the wavelength of the maximum emission. 4 The increase in fluorescence is probably a result of a change from a hydrophilic to a more hydrophobic environment around the tryptophan residues (54). For wild type PAP, there was a dose-dependent increase in fluorescence by the addition of phospholipid vesicles (Fig. 12A). The 7A mutations caused a significant increase in the fluorescence of PAP (Fig. 12A), indicating an increase in phospholipid vesicle interaction. The dissociation constant (K d ) for the phospholipid vesicles of the PAP-7A enzyme was 4.4-fold lower when compared with the constant determined for the wild type PAP enzyme (Fig. 12B).

DISCUSSION
The PAH1-encoded PAP, the enzyme that catalyzes the penultimate step in the synthesis of TAG and that plays an important role in the transcriptional regulation of phospholipid synthesis genes, is subject to multiple site phosphorylations (2,3,5,25). In fact, a bioinformatics analysis indicates 4 Z. Xu and G. M. Carman, unpublished data.  that PAP may be one of the most heavily phosphorylated proteins in S. cerevisiae, with Ͼ90 putative target sites for several protein kinases (51). In this work, we extended studies on the PAP-7A mutant enzyme to shed light on the physiological consequences of phosphorylation of PAP at the seven sites previously shown to be phosphorylated in vivo (25) and established the target sites and contribution of CDC28 (CDK1)encoded CDK to this regulation.
That PAP is a target for CDC28 (CDK1)-encoded CDK phosphorylation in vivo is supported by the observations that its electrophoretic mobility increases in a temperature-sensitive cdc28-4 mutant defective in CDK activity and in a cyclin clb3⌬ clb4⌬ mutant but decreases in cells at the mitotic phase of the cell cycle (5). Moreover, the slower migrating PAP protein is recognized by the anti-MPM2 antibody that is specific for cell cycle-regulated phosphoepitopes having the minimal Ser/Thr-Pro motif (5,55), and the seven sites of phosphorylation previously identified in PAP have this motif (25). The studies performed with unphosphorylated enzyme preparations isolated from E. coli demonstrated that PAP was a bona fide substrate of CDK with targets on serine and threonine residues. The phosphorylation of PAP was dependent on time and the concentration of the kinase, and the K m values for PAP and ATP were in the low micromolar range, indicating that PAP was a good substrate for CDK. Biochemical and molecular approaches were used to identify Ser 602 , Thr 723 , and Ser 744 as the targets of CDK phosphorylation, and these residues were among the seven sites previously shown to be phosphorylated in vivo (25).
Although we did not identify all of the protein kinases responsible for the phosphorylation of the seven sites, this work advanced the understanding of how the phosphorylation of these sites regulated PAP function. Insights into this regulation were obtained through the analysis of phosphorylationdeficient mutants expressed in pah1⌬ nem1⌬ cells. The lipid composition analysis as well as the complementation analyses with respect to phenotypes related to temperature sensitivity and the anomalous nuclear/ER membrane expansion substantiated the importance of the Nem1p-Spo7p protein phosphatase complex in PAP function. Indeed, the wild type PAH1 gene did not complement physiological defects (e.g. reduced TAG content) caused by the pah1⌬ mutation unless a functional Nem1p-Spo7p protein phosphatase complex was present. The requirement of dephosphorylation was shown further by the ability of PAP-7A to complement the pah1⌬ phenotypes in the absence of functional Nem1p-Spo7p complex. Taken together, these data indicated that dephosphorylation at the seven sites was sufficient to confer the physiological functions of PAP. That CDK phosphorylation at Ser 602 , Thr 723 , and Ser 744 contributed to this regulation was primarily supported by the enzyme localization and lipid composition data. Indeed, the S602A/T723A and S602A/T723A/ S744A mutants partially mimicked the physiological consequences of the 7A mutations.
Based on the specific activities of purified wild type and 7A mutant enzymes (25), the phosphorylation of all seven sites would cause only a 1.8-fold decrease in PAP activity. This is a relatively small effect on the in vitro catalytic activity when compared with the large effects that the 7A mutations caused in vivo. Moreover, the CDK phosphorylation of the E. coliexpressed PAP had little effect on in vitro enzyme activity. Thus, the effects of phosphorylation on PAP functions in vivo should be mediated by an additional mechanism. The association of PAP with the membrane where its substrate PA resides is essential to its function in vivo. In the absence of the Nem1p-Spo7p complex, wild type PAP was enriched in the cytosol, where it was physiologically inactive, whereas the phosphorylation-deficient PAP-7A mutant enzyme was enriched in the membrane and was physiologically active. The requirement of the Nem1p-Spo7p protein phosphatase complex, which is located in the nuclear/ER membrane (5,26), indicates that PAP is recruited to the membrane for its physiological function. Indeed, the Nem1p-Spo7p-dependent membrane localization of PAP has been shown in the presence of elevated levels of PA (33). Moreover, dephosphorylated PAP anchors onto the nuclear/ER membrane via a short N-terminal amphipathic helix, allowing for the production of DAG for TAG synthesis (33). In the absence of the Nem1p-Spo7p complex, PAP-7A associates with the membrane for in vivo function (33). The phospholipid vesicle binding assays of the wild type and 7A mutant enzymes that were purified from yeast supported this conclusion. Thus, phosphorylation is a mechanism to control PAP function in vivo by regulating its association with the membrane. 5 Under normal physiological conditions (i.e. presence of the Nem1p-Spo7p complex), the level of wild type PAP detected on the membrane was very low. In fact, microscopic analysis of live S. cerevisiae cells expressing PAP-green fluorescent protein shows a cytoplasmic localization without a detectable fluorescence signal associated with the nuclear/ER membrane unless PA levels are overexpressed (33). Yet we know that PAP is physiologically active with respect to lipid metabolism throughout cell growth (2). Purified PAP has a relatively high catalytic efficiency when compared with other enzymes of phospholipid metabolism (1). The lethal phenotype of cells that overexpress the Nem1p-Spo7p complex (5) indicates that an excess of PAP function is detrimental to cell physiology. We speculate that under normal physiological conditions, the amount of PAP associated with membranes must be small to control its physiological activity (e.g. function), and this regulation is mediated by the amount of the Nem1p-Spo7p complex on the membrane. In support of this hypothesis, a global analysis of protein expression indicates that the expression level of Nem1p is 10-fold lower when compared with PAP (56).
Like yeast PAP, the mammalian enzyme (e.g. lipin 1) is also localized to the cytosol, ER, and nucleus (16,57,58). Covalent modifications of phosphorylation (59 -61) and sumoylation (62) as well as interaction with 14-3-3 proteins (63) govern lipin 1 localization. Interestingly, lipin 1 has been identified as one of the most heavily phosphorylated proteins in rat adipo- 5 Measurement of phosphorylated PAP activity in vitro using the substrate PA in a Triton X-100 (detergent) micellar assay is not constrained by the absence or presence of the Nem1p-Spo7p protein phosphatase complex in a membrane phospholipid bilayer.
cytes following treatment with insulin (59). This phosphorylation is dependent on the mTOR signaling pathway (59) and occurs at 19 sites (60). Phosphorylation also regulates lipin 1 and 2 in human cells (61). As in S. cerevisiae (5), the human lipins are phosphorylated during the mitotic phase of cell growth at sites within a Ser/Thr-Pro motif, and their PAP activities are inhibited by the mitotic phosphorylation (61). Similarly, phosphorylated forms of mammalian lipins are enriched in the cytosolic fraction, whereas the dephosphorylated forms are enriched in the membrane fraction (60,61). However, unlike the yeast PAP, the protein kinases responsible for the phosphorylations of mammalian lipins have yet to be identified.
The only other enzyme of lipid metabolism in S. cerevisiae known to be phosphorylated by CDC28 (CDK1)-encoded CDK is the TGL4-encoded TAG lipase (64). This particular lipase is important for the resumption of logarithmic growth from the stationary phase because it supplies precursors (e.g. DAG and fatty acids) for lipid synthesis (65,66). The phosphorylation of this TAG lipase at Thr 675 and Ser 890 occurs in a cell cycle-dependent manner but at an earlier point (G 1 /S transition) in the cell cycle when compared with the phosphorylation of PAP (G 2 /M transition) (5,64). Thus, the phosphorylation of PAP (synthesis) and TAG lipase (degradation) controls the homeostasis of TAG during cell cycle progression (66). However, we now know that CDC28 (CDK1)-encoded CDK is not the only cyclin-dependent kinase that phosphorylates and regulates PAP function during the cell cycle. The identification of the protein kinase(s) responsible for phosphorylating the remaining four sites among the seven sites that contain the Ser/Thr-Pro motif will permit studies to delineate the sequence of phosphorylations that occur during growth for the PAP-mediated regulation of lipid metabolism.