Phosphatidic Acid Mediates the Nem1-Spo7/Pah1 Phosphatase Cascade in Yeast Lipid Synthesis

In the yeast Saccharomyces cerevisiae, the PAH1-encoded Mg2+-dependent phosphatidate (PA) phosphatase Pah1 regulates the bifurcation of PA to diacylglycerol (DAG) for triacylglycerol (TAG) synthesis and to CDP-DAG for phospholipid synthesis. Pah1 function is mainly regulated via control of its cellular location by phosphorylation and dephosphorylation. Pah1 phosphorylated by multiple protein kinases is sequestered in the cytosol apart from its substrate PA in the membrane. The phosphorylated Pah1 is then recruited and dephosphorylated by the protein phosphatase complex Nem1 (catalytic subunit)-Spo7 (regulatory subunit) in the endoplasmic reticulum. The dephosphorylated Pah1 hops onto and scoots along the membrane to recognize PA for its dephosphorylation to DAG. Here, we developed a proteoliposome model system that mimics the Nem1-Spo7/Pah1 phosphatase cascade to provide a tool for studying Pah1 regulation. Purified Nem1-Spo7 was reconstituted into phospholipid vesicles prepared in accordance with the phospholipid composition of the nuclear/endoplasmic reticulum membrane. The Nem1-Spo7 phosphatase reconstituted in the proteoliposomes, which were measured 60 nm in an average diameter, was catalytically active on Pah1 phosphorylated by Pho85-Pho80, and its active site was located at the external side of the phospholipid bilayer. Moreover, we determined that PA stimulated the Nem1-Spo7 activity, and the regulatory effect was governed by the nature of the phosphate headgroup but not by the fatty acyl moiety of PA. The reconstitution system for the Nem1-Spo7/Pah1 phosphatase cascade, which starts with the phosphorylation of Pah1 by Pho85-Pho80 and ends with the production of DAG, is a significant advance to understand a regulatory cascade in yeast lipid synthesis.

Abstract In the yeast Saccharomyces cerevisiae, the PAH1encoded Mg 2þ -dependent phosphatidate (PA) phosphatase Pah1 regulates the bifurcation of PA to diacylglycerol (DAG) for triacylglycerol (TAG) synthesis and to CDP-DAG for phospholipid synthesis. Pah1 function is mainly regulated via control of its cellular location by phosphorylation and dephosphorylation. Pah1 phosphorylated by multiple protein kinases is sequestered in the cytosol apart from its substrate PA in the membrane. The phosphorylated Pah1 is then recruited and dephosphorylated by the protein phosphatase complex Nem1 (catalytic subunit)-Spo7 (regulatory subunit) in the endoplasmic reticulum. The dephosphorylated Pah1 hops onto and scoots along the membrane to recognize PA for its dephosphorylation to DAG. Here, we developed a proteoliposome model system that mimics the Nem1-Spo7/Pah1 phosphatase cascade to provide a tool for studying Pah1 regulation. Purified Nem1-Spo7 was reconstituted into phospholipid vesicles prepared in accordance with the phospholipid composition of the nuclear/endoplasmic reticulum membrane. The Nem1-Spo7 phosphatase reconstituted in the proteoliposomes, which were measured 60 nm in an average diameter, was catalytically active on Pah1 phosphorylated by Pho85-Pho80, and its active site was located at the external side of the phospholipid bilayer. Moreover, we determined that PA stimulated the Nem1-Spo7 activity, and the regulatory effect was governed by the nature of the phosphate headgroup but not by the fatty acyl moiety of PA.
The reconstitution system for the Nem1-Spo7/Pah1 phosphatase cascade, which starts with the phosphorylation of Pah1 by Pho85-Pho80 and ends with the production of DAG, is a significant advance to understand a regulatory cascade in yeast lipid synthesis.
The most important factor responsible for Pah1 function is its cellular location, which is controlled by its phosphorylation and dephosphorylation (54) (Fig. 1). After its expression, Pah1 in the cytosol is Model for the phosphorylation/dephosphorylation-mediated regulation of Pah1 PAP location and mode of action of the enzyme at the nuclear/ER membrane via the hopping and scooting modes. Pah1 in the cytosol is phosphorylated by several protein kinases. The phosphorylated enzyme (gray circles) translocates to the nuclear/ER membrane via its recruitment and dephosphorylation by the Nem1-Spo7 protein phosphatase complex. Dephosphorylated Pah1 hops onto the membrane, scoots along the surface to bind its substrate, and catalyzes the dephosphorylated PA to produce DAG, which can be utilized in the formation of TAG stored in lipid droplets (LDs). Following the PAP reaction, Pah1 scoots along the membrane until it recognizes a PA molecule for another round of reaction. Pah1 catalyzing the PAP reaction is designated with blue color. PA is also converted to CDP-DAG for the synthesis of the membrane phospholipids PI, PS, phosphatidylethanolamine (PE), and PC. The phospholipids PE and PC may also be derived from DAG via the CDP-ethanolamine and CDP-choline branches of the Kennedy pathway when cells are supplemented with ethanolamine and/or choline. Further details of the lipid synthesis pathways may be found elsewhere (3)(4)(5). PA, phosphatidate; PAP, phosphatidate phosphatase; PC, phosphatidylcholine; PI, phosphatidylinositol; PS, phosphatidylserine. phosphorylated by multiple protein kinases (54). The phosphorylation stabilizes the enzyme (72) but makes it nonfunctional because it is sequestered in the cytosol apart from its substrate PA that resides in the nuclear/ ER membrane. Phosphorylated Pah1 translocates to the nuclear/ER membrane through its recruitment and dephosphorylation by the Nem1-Spo7 protein phosphatase complex (61-65, 71, 73-76). Dephosphorylated Pah1 hops onto the membrane via its N-terminal amphipathic helix (73), scoots along the surface to recognize its substrate PA, and catalyzes the dephosphorylated PA to produce DAG (77). Following the PAP reaction, Pah1 scoots along the membrane for another round of reaction (77).
The aim of this work is to develop a proteoliposome model system that mimics the Nem1-Spo7/Pah1 phosphatase cascade that occurs at the nuclear/ER membrane. In a previous study, we showed that a liposome composed of PC, PE, PI, PS, and PA, which mimics the phospholipid composition of the nuclear/ER membrane (1,78,79), provides a suitable model for the action of unphosphorylated Pah1 to hop and scoot along the membrane surface and catalyze the PAP reaction (77). Herein, we significantly advanced this in vitro model for studying the dephosphorylation of Pah1 at the membrane surface by the Nem1-Spo7 complex reconstituted in the PC/PE/PI/PS/PA liposomes. We discovered that the PAP substrate PA stimulates the Nem1-Spo7 complex-mediated dephosphorylation of Pah1 and demonstrated that the Nem1-Spo7/Pah1 phosphatase cascade, starting with the phosphorylation of Pah1 and ending with the generation of DAG from PA, was reconstituted in the Nem1-Spo7 proteoliposome system.

Materials
Avanti Polar Lipids and Analtech were the sources of lipids and silica gel GHL TLC plates, respectively. Bio-Rad was the source of molecular mass protein standards and reagents for protein assay, SDS-PAGE, and Western blotting. InstantBlue protein stain was from Expedeon. GE Healthcare was the source of the enhanced chemifluorescence Western blotting detection kit, Sephadex G-50 superfine, IgG-Sepharose, Q-Sepharose, and polyvinylidene difluoride paper. Milli-poreSigma was the source of ammonium molybdate, bovine serum albumin, Triton X-100, and rabbit anti-protein A antibody (product P3775, lot 025K4777). PerkinElmer Life Sciences and National Diagnostics, respectively, were the sources of radiochemicals and scintillation-counting supplies. Thermo Fisher Scientific was the source of malachite green, alkaline phosphatase-conjugated goat anti-rabbit IgG antibody (product no. 31340, lot number: NJ178812), and Pierce strong anion exchange spin columns and protein concentrators 30K. Qiagen was the source of nickel-nitrilotriacetic acid-agarose resin. Anti-Spo7 antibody was previously produced in New Zealand White rabbits (80). Wako Chemicals supplied Phos-tag™ Acrylamide AAL-107. All other chemicals were reagent grade.

Preparation of Nem1-Spo7 proteoliposomes
Nem1-Spo7 proteoliposomes were prepared by the sizeexclusion chromatography as described by Mimms et al. (87) with minor modifications (88)(89)(90). Unless otherwise indicated, the dioleoyl derivatives of PC, PE, PS, PA, and soybean PI were used in this work. Under most conditions used in this study,  Briefly, chloroform was evaporated from phospholipid mixtures under a stream of nitrogen to form a thin film, and residual solvent was removed in vacuo. Phospholipids were resuspended in 280 mM octyl glucoside to a final phospholipid concentration of 20 mM. The octyl glucoside/ phospholipid-mixed micelles were then mixed with 9 μg of Nem1-Spo7 complex solubilized in 1.7 mM Triton X-100. The final molar ratios of octyl glucoside to Triton X-100 and octyl glucoside to phospholipids were 164:1 and 14:1, respectively. The reconstitution mixture (137 μl) was applied to and eluted from a Sephadex G-50 superfine column (1 × 4 cm) equilibrated at 4 • C with a chromatography buffer consisting of 25 mM Tris-HCl (pH 8.0), 250 mM NaCl, and 10 mM 2-mercaptoethanol at 4 • C. The presence of phospholipids and the Nem1-Spo7 complex, respectively, in the column fractions was analyzed by primuline staining of lipid spots on a TLC plate and Western blotting with anti-protein A and anti-Spo7 antibodies. Relative amounts were determined by fluorimaging; the image intensities were quantified with ImageJ software. The size of the proteoliposomes in the peak fraction, which we used in these studies, was determined by light scattering using a Brookhaven Instruments Particle Size Analyzer. Proteoliposomes were stored for no longer than one week at 4 • C.

Lipid analysis of Nem1-Spo7 proteoliposomes
Lipids were extracted from proteoliposomes by the method of Bligh and Dyer (91). Phospholipids and DAG, respectively, were resolved by one-dimensional TLC on silica gel plates using the solvent systems chloroform/ethanol/water/triethylamine (30:35:7:35, v/v) (92) and hexane/diethyl ether/ glacial acetic acid (40:10:1, v/v) (93). The resolved lipids were stained with 0.5% primuline. The amounts of PA and DAG in the proteoliposomes were determined from standard curves of each lipid on TLC plates. Fluorimaging using a Storm 860 Molecular Imager (GE Healthcare) was used to acquire fluorescence signals from the plates; the intensities of the images were analyzed by ImageJ software. The identity of the lipids was confirmed by comparison with the migration of authentic standards.

Nem1-Spo7 protein phosphatase assay
Nem1-Spo7 phosphatase activity was measured at 30 • C by following an increase in electrophoretic mobility of Pah1 in SDS-PAGE (70,94) or by following the release of 32 P i from [ 32 P]Pah1 (200-1,400 cpm/nmol) (70). The reaction mixture contained 100 mM sodium acetate (pH 5.0), 10 mM MgCl 2 , 1 mM DTT, 0.2 μM phosphorylated Pah1, and Nem1-Spo7 proteoliposomes in a total volume of 20 μl. In the nonradioactive assay, the reaction was terminated with Laemmli sample buffer (95) followed by SDS-PAGE in the absence and presence of the Phos-tag reagent (70,94). In the radioactive assay, the reaction was terminated by the addition of 20% trichloroacetic acid and 0.4 mg/ml bovine serum albumin. The mixture was cooled on ice for 15 min and then centrifuged for 20 min at 15,000 g to precipitate the radioactive substrate. The supernatant containing 32 P i was measured for radioactivity by scintillation counting. One unit of Nem1-Spo7 phosphatase activity was defined as the amount of enzyme that catalyzed the formation of 1 nmol phosphate per min. The Nem1-Spo7 phosphatase reactions, which were conducted in triplicate at 30 • C, were linear with time and protein concentration. To prepare [ 32 P]Pah1, E. coli-expressed unphosphorylated Pah1 was phosphorylated by the Pho85-Pho80 protein kinase complex using 100 μM [γ-32 P]ATP (5,000-10,000 cpm/pmol) as described previously (64). [ 32 P]Pah1 was purified by Q-Sepharose chromatography to remove the Pho85-Pho80 protein kinase (70).

PAP assay
PAP activity was measured at 30 • C for 15 min by following the release of water-soluble P i from chloroform-soluble PA; the P i produced in the reaction was measured with malachite green-molybdate reagent (96,97). The reaction mixture contained 50 mM Tris-HCl (pH 7.5), 1 mM MgCl 2 , enzyme protein, and the PA-containing liposomes (77) in a final volume of 10 μl. Enzyme assays were conducted in triplicate, and the average SD of the assays was ± 5%. All enzyme reactions were linear with time and protein concentration. One unit of enzymatic activity was defined as the amount of enzyme that catalyzed the formation of 1 nmol of product per minute.

SDS-PAGE and Western blotting
SDS-PAGE (95) and Western blotting (98,99) with polyvinylidene difluoride membrane were performed by standard procedures. Phos-tag™ AAL-107 (20 μM) and MnCl 2 (100 μM) were added to 5% polyacrylamide gels for analysis of the phosphorylation state of Pah1. The samples for Western blotting were normalized to total protein loading. The membrane was cut and the upper and lower portions, respectively, were probed with rabbit anti-protein A (2 μg/ml) and rabbit anti-Spo7 (1 μg/ml) antibodies. The goat anti-rabbit IgG antibody conjugated with alkaline phosphatase was used at a dilution of 1:4,000. Immune complexes were assayed with the enhanced chemifluorescence Western blotting substrate. Fluorescence signals from the blots were visualized by fluorimaging with a Storm 865 Molecular Imager (GE Healthcare) and image intensities were analyzed with ImageJ software.

Data analysis
Microsoft Excel software was used for the statistical analysis of the data; the P values < 0.05 were taken as a significant difference. SigmaPlot software was used to analyze kinetic data.

Reconstitution of the Nem1-Spo7 protein phosphatase complex into phospholipid vesicles
We sought to reconstitute the Nem1-Spo7 phosphatase complex into PC/PE/PI/PS/PA vesicles to mimic the in vivo environment of the complex in the nuclear/ ER membrane (1,78,79) to recruit and dephosphorylate Pah1 (61,62,70,71,73). It has been shown that Pah1 PAP activity on PA in this phospholipid composition is higher than the enzyme activity on simple PC/PA phospholipid vesicles or Triton X-100/PA-mixed micelles (77). For the preparation of the Nem1-Spo7 proteoliposomes, we utilized the chromatographic method of Mimms et al. (87) that had been applied to the reconstitution of the membrane-associated phospholipid biosynthetic enzymes glycerol-3-phosphate acyltransferase (88), PI synthase (89), and PS synthase (90) into unilamellar phospholipid vesicles. Triton X-100/ Nem1-Spo7 complex micelles (70,71) were mixed with octyl glucoside/phospholipid micelles and fractionated by size-exclusion chromatography with Sephadex G-50 superfine. As expected (87)(88)(89)(90), the Nem1-Spo7 proteoliposomes were eluted from the column near the void volume, and the peak fractions of phospholipids, Nem1, and Spo7 were coincident ( Fig. 2A). The TLC analysis of the proteoliposomes confirmed that the vesicles were composed of PC, PE, PI, PS, and PA ( Fig. 2A, right), and the relative amounts of the phospholipids were estimated from their input amounts. The peak fractions of Triton X-100 and octyl glucoside micelles, which are well separated from the proteoliposomes, were contained in the later elution fractions that are not shown in Fig. 2A (87-90). Light scattering analysis of the Nem1-Spo7 proteoliposomes in the peak fractions, which were used in these studies, indicated the average diameter of 60 nm (Fig. 2B), a value within the range of vesicle sizes (40-90 nm) of reconstituted glycerol-3-phosphate acyltransferase, PI synthase, and PS synthase (88)(89)(90). Disruption of the proteoliposomes with Triton X-100 did not result in a significant increase in Nem1-Spo7 phosphatase activity, indicating that the phosphatase complex is asymmetrically reconstituted with its active site located outside of the vesicle. Proteoliposomes stored at 4 • C maintained Nem1-Spo7 phosphatase activity for at least one week.
In the electrophoretic mobility shift assay, the reconstituted Nem1-Spo7 catalyzed the time-dependent dephosphorylation of Pah1 phosphorylated in vivo (Fig. 3A) and in vitro (Fig. 3B) as indicated by an increase in the mobility of the protein. To better distinguish the phosphorylated and dephosphorylated forms A B Fig. 2. Preparation and analysis of Nem1-Spo7 proteoliposomes. A: octyl glucoside/phospholipid micelles were mixed with Triton X-100/Nem1-Spo7 complex micelles, followed by size-exclusion chromatography with Sephadex G-50 superfine. The elution fractions (0.5 ml) were analyzed for Nem1 and Spo7 by Western blotting (upper) with anti-protein A and anti-Spo7 antibodies, respectively, and for total phospholipids (PLs) by TLC and primuline staining (right). The amounts of Nem1, Spo7, and phospholipids were quantified by ImageJ software; the amount of each component in the peak fraction was set at 100%. The phospholipid composition of the peak fraction was analyzed by TLC; the positions of the individual phospholipids are indicated (right). The arrow in the figure indicates the position of the void volume. B: the peak fraction of Nem1-Spo7 proteoliposomes was subjected to particle size measurement by light scattering.
of Pah1, we examined its mobility in the polyacrylamide gel containing the Phos-tag reagent, which traps phosphorylated proteins (115) and retards their electrophoretic mobility (70, 101) (Fig. 4A). Using the Phos-tag polyacrylamide gel, we could more readily discern changes in electrophoretic mobility and showed that the dephosphorylation of endogenously phosphorylated Pah1 (Fig. 4B) and Pah1 phosphorylated by Pho85-Pho80 in vitro (Fig. 4C) is dependent on the amount of the Nem1-Spo7 proteoliposomes. In addition, the dephosphorylation of Pah1 resulted in its degradation as reflected by the reduction of protein abundance (Fig. 4B, C).
In the radioactive assay for Nem1-Spo7 activity, which is more sensitive and quantitative, a 32 P-labeled substrate was prepared by phosphorylating the E. coliexpressed unphosphorylated Pah1 by Pho85-Pho80 with [γ-32 P]ATP. The amount of 32 P i released from [ 32 P]Pah1 by Nem1-Spo7 was measured by scintillation counting after the removal of both phosphorylated and dephosphorylated forms of the protein by trichloroacetic acid precipitation (70). The Nem1-Spo7 phosphatase activity was linear over a 10-min incubation period and dependent on the concentration of phosphorylated Pah1 (Fig. 5). The radioactive assay was not performed on endogenously phosphorylated Pah1 owing to the cumbersome nature of purifying radiolabeled Pah1 from yeast cells. Overall, Nem1-Spo7 phosphatase activity in the proteoliposome system was dependent on time, the amount of the phosphatase complex, and the concentration of phosphorylated Pah1. The dephosphorylation of Pah1 was not observed when the phosphatase reaction was performed with liposomes without the reconstituted Nem1-Spo7 complex.

PA stimulates Nem1-Spo7 phosphatase activity
The Nem1-Spo7 proteoliposomes contained PA because it is a component of the nuclear/ER membrane (1,78,79) and the substrate for the PAP reaction (34). The PA concentration in the nuclear/ER membrane is largely controlled by Pah1 PAP activity (33,35), and thus, we considered whether the PA concentration affects the Nem1-Spo7 activity on Pah1. To address this question, Pah1 phosphorylated by Pho85-Pho80 was incubated with Nem1-Spo7 proteoliposomes prepared A B Fig. 3. Nem1-Spo7 proteoliposomes catalyze the timedependent dephosphorylation of Pah1. Nem1-Spo7 proteoliposomes were incubated for the indicated time periods with purified Pah1 endogenously phosphorylated in yeast (yPah1) (A) or Escherichia coli-expressed and purified Pah1 phosphorylated by Pho85-Pho80 (ePah1 + Pho85-Pho80) (B). Following the incubations, the samples were subjected to SDS-PAGE using a 5% polyacrylamide gel followed by staining with Coomassie blue. The positions of phosphorylated (p-Pah1) and dephosphorylated Pah1 and molecular mass standards are indicated. The data shown is representative of three separate experiments. with and without 10 mol % PA, and its dephosphorylation was assessed by electrophoretic mobility in Phos-tag SDS-PAGE. Compared with Nem1-Spo7 phosphatase activity in the presence of PA, the enzyme activity was lower in the absence of PA as reflected by a less increase in Pah1 mobility (Fig. 6A). The stimulatory effect of PA on Nem1-Spo7 phosphatase was further examined using the radioactive assay with Pah1 phosphorylated by Pho85-Pho80 with [γ-32 P]ATP. The phosphatase complex was reconstituted in a series of phospholipid vesicles containing varying amounts of PA. Compared with Nem1-Spo7 activity in the absence of PA, the enzyme activity was stimulated by the phospholipid in a cooperative dose-dependent manner (Fig. 6B). Analysis of the data yielded a Hill number of ∼1.5. The Nem1-Spo7 phosphatase activity in the presence of 10 mol % PA was 3.6-fold greater than the enzyme activity in the absence of PA (Fig. 6B, C). In contrast to PA, thioPA (1, 2-diacyl-sn-glycero-3phosphorothioate) containing a sulfur atom in place of an oxygen in the phosphate headgroup did not show a significant stimulatory effect on Nem1-Spo7 phosphatase activity (Fig. 6C). Similarly, the PAP product DAG substituted for PA in the proteoliposome did not stimulate the Nem1-Spo7 phosphatase activity (Fig. 6C,  control). The stimulatory effect of PA was not affected by the composition of its fatty acyl groups, which include two saturated (16:0 to 16:0), two monounsaturated (18:1 to 18:1), or one saturated and one unsaturated (18:0-18:1) fatty acyl chains (Fig. 6D). Overall, these results indicate that the stimulatory effect of PA on Nem1-Spo7 phosphatase is governed by the phosphate headgroup but not by the fatty acyl groups of the phospholipid.

The Nem1-Spo7/Pah1 phosphatase cascade is reconstituted with proteoliposomes
After confirming that Nem1-Spo7 proteoliposomes are active on phosphorylated Pah1, we sought to establish the phosphatase cascade in the reconstituted assay system. For this experiment, the Nem1-Spo7 complex was reconstituted in phospholipid vesicles containing 10 mol % PA. The proteoliposomes were incubated with Pah1 phosphorylated by Pho85-Pho80 with unlabeled ATP. The reaction mixture was incubated for 45 min at pH 5.0 to allow for the maximum Nem1-Spo7 activity on phosphorylated Pah1 and to compensate for the reduced PAP activity of dephosphorylated Pah1 at pH 5.0 when compared with its optimum at pH 7.0-7.5 (35,116). An intermediate pH between the optimums for Nem1-Spo7 and Pah1 was not used for the assay because the former activity drops precipitously at pH above 5.0 (70). The amounts of PA and DAG in the proteasomes were determined at the 0and 45-min time intervals to assess the course of the PAP reaction. After the 45-min incubation, the amount of DAG produced was 0.7 nmol (Fig. 7), corresponding to 17% of PA dephosphorylation from its total vesicle concentration of 4.2 nmol. Considering the accessible amount of PA in the vesicles, its conversion to DAG would be 34% based on the equal distribution of the phospholipid between the inner and outer leaflets. As expected, DAG was not produced from the reaction when PA was omitted from the Nem1-Spo7 proteoliposomes. Moreover, the catalytic site mutant Pah1-D398E (83) (Fig. 7, inset) did not produce DAG in the reconstituted assay system. Taken together, these results demonstrate the control of PAP activity in the reconstituted Nem1-Spo7/Pah1 phosphatase cascade.

DISCUSSION
In the yeast S. cerevisiae, Pah1 is a key regulator in lipid synthesis; it controls the bifurcation of PA between DAG and CDP-DAG for the synthesis of the neutral storage lipid TAG and membrane phospholipids, respectively (2, 50-54) (Fig. 1). The physiological function of Pah1 is largely controlled by its cellular location, which is mediated by phosphorylation and dephosphorylation (54). Phosphorylated Pah1 is sequestered in the cytosol apart from its nuclear/ER membraneassociated substrate PA, whereas dephosphorylated Pah1 associates with the membrane to catalyze the PAP reaction to generate DAG (54). Thus, the master regulators of Pah1 function are the protein kinases that phosphorylate Pah1 in the cytosol and the nuclear/ER membrane-associated Nem1-Spo7 protein phosphatase complex that recruits and dephosphorylates the enzyme at the membrane (61-65, 70, 71, 73-76). In this work, we developed a model system to examine Pah1 phosphorylation by Pho85-Pho80, the protein kinase that has the greatest effect on Pah1 function (54,64), and its dephosphorylation by the membrane-associated Nem1-Spo7 protein phosphatase complex.
We successfully reconstituted functional Nem1-Spo7 complex into phospholipid vesicles composed of PC/PE/PI/PS/PA, which constitutes the nuclear/ER membrane bilayer (1,78,79). The phospholipid vesicles afford a surface for PAP activity that rivals that of Triton X-100/PA-mixed micelles (77). The average size (60 nm) of the Nem1-Spo7 proteoliposomes prepared by size-exclusion chromatography was similar to that of other proteoliposomes reconstituted with lipid metabolic enzymes prepared by the same method (88)(89)(90). Moreover, the lack of latent Nem1-Spo7 activity upon proteoliposome disruption with detergent indicated that the active site of Nem1 is located at the external side of the vesicle. This model system permitted analysis of the Nem1-Spo7/Pah1 phosphatase cascade starting with the phosphorylation of Pah1 by Pho85-Pho80 and ending with the production of DAG. Analyzing Nem1-Spo7 proteoliposomes with and without PA revealed that the protein phosphatase activity is stimulated by the PAP substrate. The stimulatory effect of PA was governed by the nature of the phosphate headgroup. For example, thioPA, the phosphorothioate analog of PA, did not stimulate Nem1-Spo7 activity. Parenthetically, thioPA is not a substrate for the PAP reaction (100). Additionally, DAG lacking the phosphate moiety of a phospholipid molecule did not stimulate the protein phosphatase activity. The fatty acyl moiety of PA did not have a significant effect on its stimulatory effect on Nem1-Spo7 activity. Likewise, the PA fatty acyl groups do not affect Pah1 PAP activity (77). The mechanism by which PA stimulates Nem1-Spo7 phosphatase activity is unclear at present. We consider that the stimulatory effect on the enzyme activity might be governed by direct interaction (117) of PA with Nem1 and/or Spo7 and/or by the effect of PA on the membrane bilayer environment (118)(119)(120)(121). Additional work is required to address these notions.
The stimulatory effect of PA on Nem1-Spo7 activity was an unexpected finding that has implications for the regulation of the phosphatase cascade affecting the PA/DAG balance and lipid synthesis. We envisage a scenario where the cell senses elevated PA levels that fuel the recruitment and dephosphorylation of Pah1 for its function on the membrane. The activation of Pah1 in turn reduces PA levels through TAG synthesis controlled by its PAP activity. Reduced PA levels would signal the opposite effects such as reduced Nem1-Spo7 activity, Pah1 hyperphosphorylation, and reduced PAP activity. This situation, however, may be an oversimplification of the complex regulation that occurs in vivo with multiple lipid biosynthetic enzymes (e.g., glycerol-3-phosphate and lysoPA acyltransferases, DAG kinase, and phospholipase D) that affect PA levels (2)(3)(4)122).
In addition to being a phospholipid intermediate in lipid biosynthetic pathways (Fig. 1), PA itself functions as a signaling molecule in various cellular functions. In mammalian cells, PA activates cell growth and proliferation, vesicular trafficking, secretion, endocytosis, and even hair growth (123)(124)(125)(126)(127)(128)(129)(130). In plants, PA is implicated in seed germination and stress responses to low temperature, salinity, and drought (125,126). In bacteria, PA plays a role in signaling and biofilm formation (131). In yeast, PA is required for suppression of the growth and membrane trafficking defects in the Sec14 PI/PC binding protein mutant (132)(133)(134) and in Spo20mediated fusion of vesicles with the prespore membrane during sporogenesis (135,136). Most germane to the control of lipid synthesis in yeast is the role of PA in the expression of lipid synthesis genes via the Henry (Opi1/Ino2-Ino4) regulatory circuit (2)(3)(4)137). PA, along with Scs2, has the ability to sequester the Opi1 repressor at the nuclear/ER membrane, permitting the Ino2-Ino4 complex-mediated transcriptional activation of UAS INO -containing phospholipid synthesis genes (2)(3)(4)137). Reduced PA levels permit Opi1 to dissociate from the nuclear/ER membrane and translocate into the nucleus where it interacts with Ino2 to attenuate transcription of UAS INO -containing genes (2)(3)(4)137).
The Pah1-mediated control of PA plays an important role in the transcriptional regulation of lipid synthesis genes (e.g., INO1, INO2, CHO1, and OPI3) via the Henry regulatory circuit (61,62,138,139). We posit that the PAmediated regulation of the Nem1-Spo7 phosphatase complex must be involved in this regulation. Interestingly, the fatty acyl species of PA do not stimulate the Nem1-Spo7 and PAP activities but affect the interaction of Opi1 with the phospholipid; the PA species 34:1, the most abundant PA species in S. cerevisiae (33), mediates expression of the UAS INO -containing INO1 gene (139). In addition, the Opi1 repressor function is also affected by the fatty acyl chain length of PA; Opi1 binding to PA is favored with its C16-over C18-chain length (140).
The reconstitution of the Nem1-Spo7/Pah1 axis is a significant advance in modeling a regulatory cascade in Fig. 7. Reconstitution of the Nem1-Spo7/Pah1 phosphatase cascade in proteoliposomes. The Nem1-Spo7 complex was reconstituted into liposomes with 10 mol % PA. Escherichia coliexpressed and purified unphosphorylated WT or D398E forms of Pah1 were phosphorylated by Pho85-Pho80 with unlabeled ATP. The phosphorylated forms of Pah1 were incubated with the Nem1-Spo7 proteoliposomes for 45 min. Following the incubation, the proteoliposomes were collected, the lipids extracted, and the amounts of PA and DAG were analyzed by TLC. Inset, unphosphorylated WT and D398E forms of Pah1 were assayed for PAP activity by following the release of P i from PA using liposomes composed of PC/PE/PI/PS/PA. The surface concentration of PA within the liposomes was 10 mol %. The data shown in the figure is the average of three independent experiments ± S.D. (error bars). The error bars are hidden behind the symbols. The individual data points are shown. PA, phosphatidate; PAP, phosphatidate phosphatase. lipid synthesis. This proteoliposome system will permit well-defined studies to examine the structure-function relationships of Nem1, Spo7, and Pah1 as well as the phosphorylation-mediated interactions of Pah1 with the Nem1-Spo7 complex at the membrane surface. In fact, in a recent study, the system provided important information on Nem1-Spo7 activity on Pah1 phosphorylated by Pho85-Pho80 and glycogen synthase kinase homolog Rim11, which showed a preference for dephosphorylating target sites of the Pho85-Pho80 protein kinase (69).

Data Availability
All data are contained within the article or supporting information.

Supplemental Data
This article contains supplemental data (100).

Acknowledgments
Azam Hassaninasab is acknowledged for assistance in the purification of the Pho85-Pho80 complex. Qingrong Huang and Spencer Knapp are acknowledged for use of the particle size analyzer and polarimeter, respectively. We acknowledge Shoily Khondker, Ruta Jog, and Geordan Stukey for helpful discussions during the course of this work.

Conflict of Interest
The authors declare that they have no conflicts of interest with the contents of this article.