The Golgi localization of phosphatidylinositol transfer protein beta requires the protein kinase C-dependent phosphorylation of serine 262 and is essential for maintaining plasma membrane sphingomyelin levels.

Recombinant mouse phosphatidylinositol transfer protein (PI-TP)beta is a substrate for protein kinase C (PKC)-dependent phosphorylation in vitro. Based on site-directed mutagenesis and two-dimensional tryptic peptide mapping, Ser(262) was identified as the major site of phosphorylation and Ser(165) as a minor phosphorylation site. The phospholipid transfer activities of wild-type PI-TP beta and PI-TP beta(S262A) were identical, whereas PI-TP beta(S165A) was completely inactive. PKC-dependent phosphorylation of Ser(262) also had no effect on the transfer activity of PI-TP beta. To investigate the role of Ser(262) in the functioning of PI-TP beta, wtPI-TP beta and PI-TP beta(S262A) were overexpressed in NIH3T3 fibroblast cells. Two-dimensional PAGE analysis of cell lysates was used to separate PI-TP beta from its phosphorylated form. After Western blotting, wtPI-TP beta was found to be 85% phosphorylated, whereas PI-TP beta(S262A) was not phosphorylated. In the presence of the PKC inhibitor GF 109203X, the phosphorylated form of wtPI-TP beta was strongly reduced. Immunolocalization showed that wtPI-TP beta was predominantly associated with the Golgi membranes. In the presence of the PKC inhibitor, wtPI-TP beta was distributed throughout the cell similar to what was observed for PI-TP beta(S262A). In contrast to wtPI-TP beta overexpressors, cells overexpressing PI-TP beta(S262A) were unable to rapidly replenish sphingomyelin in the plasma membrane upon degradation by sphingomyelinase. This implies that PKC-dependent association with the Golgi complex is a prerequisite for PI-TP beta to express its effect on sphingomyelin metabolism.

to enzyme (1). In mammalian tissues at least two isoforms, PI-TP␣ and PI-TP␤, are found. PI-TP␣ is able to transfer phosphatidylinositol (PI) and, to a lesser extent, phosphatidylcholine (PC) (2)(3)(4)(5)(6) and is mainly localized in the cytosol and in the nucleus (7). Similar to PI-TP␣, PI-TP␤ is able to transfer PI and PC but is also able to transfer sphingomyelin (SM) (8). PI-TP␤ is mainly associated with the Golgi apparatus (7). The primary sequences of PI-TP␣ and PI-TP␤ are very similar, with an identity of 77% and a similarity of 94% (9).
To date, little is known about the exact cellular function of PI-TP␣ and PI-TP␤. In a cell-free system containing trans-Golgi membranes, both PI-TP␣ and PI-TP␤ stimulated the formation of constitutive secretory vesicles and immature granules (10). In permeabilized, cytosol-depleted HL60 cells, both isoforms restored GTP␥S-stimulated protein secretion and phospholipase C-mediated inositol lipid signaling (11,12). In these assays, PI-TP␣ and PI-TP␤ were found to function equally well despite their different biochemical properties and cellular localizations. On the other hand, NIH3T3 cells with increased expression of either PI-TP␣ or PI-TP␤ demonstrated remarkable differences in lipid metabolic pathways. Cells overexpressing PI-TP␣ (SPI␣ cells) showed an enhanced PLA 2mediated PI degradation resulting in increased levels of lyso-PI, glycerophosphoinositol, Ins(1)P, and Ins(2)P (13). This was not observed in cells overexpressing PI-TP␤ (SPI␤ cells). However, in SPI␤ cells (but not in SPI␣ cells) it was shown that under conditions in which plasma membrane SM was hydrolyzed to ceramide by exogenous sphingomyelinase, PI-TP␤ was involved in maintaining the steady-state levels of SM (14). It was recently postulated that PI-TP␤ plays a key role in SM metabolism, making it a potential regulator of pathways for diacylglycerol production and consumption in the mammalian trans-Golgi network (15). Disruption of the PI-TP␤ gene in mice leads to early failure in embryonic development (16).
In search of mechanisms by which PI-TP activity is regulated, PI-TP␣ was shown to be phosphorylated by protein kinase C in vitro as well as in vivo (17,18). PI-TP␣ was exclusively phosphorylated on Ser 166 , with the PC-carrying form of PI-TP␣ more readily phosphorylated than the PI-carrying form (18). Furthermore, in NIH3T3 cells, PI-TP␣ was translocated from the cytosol to the Golgi membranes upon phosphorylation after stimulation with phorbol ester. This relocalization of PI-TP␣ coincided with an increased level of intracellular lyso-PI, indicating the activation of a PI-specific PLA 2 (17,18). Based on these findings, a model was proposed for the regulation of PI-TP␣ by PKC-dependent phosphorylation. In contrast to PI-TP␣, PI-TP␤ purified from bovine or rat brain could not be phosphorylated despite the fact that it contains the same serine residue (Ser 165 ) and an additional putative PKC phosphorylation site (Ser 262 ) not present in PI-TP␣ (8,17,19).
In this study, we report that murine PI-TP␤ can be phosphorylated by PKC in vitro as well as in situ. The major site of phosphorylation was Ser 262 , whereas Ser 165 was a minor site. By site-directed mutagenesis we have shown that Ser 165 is essential for the lipid transfer activity of the protein, whereas phosphorylation of Ser 262 is required for the association of PI-TP␤ with the Golgi membranes. This latter residue is also essential for the ability of PI-TP␤ to maintain steady-state levels of SM in the plasma membrane.
Site-directed Mutagenesis of PI-TP␤-The PI-TP␤ cDNA cloned into the pBluescript SK Ϫ vector (14) was used for site-directed mutagenesis using the QuikChange site-directed mutagenesis method according to the manufacturer's instruction. Ser 165 was replaced by Ala using the following mutagenic oligonucleotides: sense 165 primer, 5Ј-CCTGCAT-TATTCCACGCAGTGAAGACCAAGAGA-3Ј and antisense 165 primer, 5Ј-TCTCTTGGTCTTCACTGCGTGGAATAATGCAGG-3Ј. The bold nucleotides encode the mutated amino acid (Ser 165 to Ala 165 ) The underlined nucleotides are mutations that do not result in a change in amino acid composition, but they introduce a DraIII restriction site. The mutated construct is denoted as pBluePI-TP␤(S165A).
Ser 262 was replaced by Ala using the following mutagenic oligonucleotides: sense 262 primer, 5Ј-ATGCGTAAGAAGGGTGCGGTC-CGAGGCACGTCG-3Ј and antisense 262 primer, 5Ј-CGACGTGCCTCG-GACCGCACCCTTCTTACGCAT-3Ј. The bold nucleotides encode the mutated amino acid (Ser 262 to Ala 262 ). This mutation also results in the introduction of an RsrII restriction site. The resulting construct is denoted as pBluePI-TP␤(S262A). A mutant in which both Ser 165 and Ser 262 were replaced by Ala was generated using the same primers as for the mutation of Ser 262 to Ala 262 with pBluePI-TP␤(S165A) as target DNA in the mutagenesis reaction. The resulting construct is denoted as pBluePI-TP␤(S165A/S262A). Incorporation of the mutagenic oligonucleotides into the construct was checked by restriction enzyme analysis and by DNA sequencing. The three mutated and the wtPI-TP␤ cDNAs were cloned into the pET15b expression vector. Expression of these pET15b-PI-TP␤ constructs yielded wtPI-TP␤ or mutant PI-TP␤ fused to an N-terminal peptide containing six histidine residues (His 6 tag).
Purification of Wild-type and Mutant His 6 -tagged PI-TP␤-The E. coli strain BL21(DE3) was transformed by the pET15b-PI-TP␤ constructs. A 50-ml culture, grown overnight in Luria-Bertani (LB) medium containing 50 g/ml ampicillin was used to inoculate 2 liters of LB medium (also containing 50 g/ml ampicillin). The culture was grown at 18°C for 24 h, and the His 6 -tagged PI-TP␤s (wt and mutant proteins) were purified from these by Ni 2ϩ -Hybond affinity chromatography. After chromatography, the fractions were assayed for PI transfer activity and immunoreactivity (enzyme-linked immunosorbent assay). After the final purification step, the fractions containing His 6 -tagged PI-TP␤ were combined and concentrated to 10 ml. The purified His 6 -tagged PI-TP␤s were stored in 52% (v/v) glycerol at Ϫ20°C.
Phospholipid Transfer Activity Assay-PI, PC, and SM transfer activities were determined in a continuous fluorescence assay using donor vesicles consisting of either Pyr-PI, Pyr-PC or Pyr-SM and PA, egg-PC, and TNP-PE (10:10:70:10, mol%) and acceptor vesicles consisting of PC and PA (95:5, mol%) (3,20). Measurements were performed using a fluorimeter (Photon Technology International) equipped with a thermostated cuvette holder and a stirring device.
Purification of Protein Kinase C-PKC was purified from rat brain by a modified procedure previously described by Huang et al. (21). Rat brains (20 -40 g of tissue) were homogenized, and the cytosolic fraction was subsequently purified on DEAE-Sepharose, Sephacryl 200, and phenyl-Sepharose columns. The purified enzyme has a specific activity of 200 nmol of phosphate/min/mg protein when assayed with histone III as substrate. The purified enzyme is stable for several months when kept at Ϫ80°C in 50% glycerol and 0.01% Triton X-100.
Phosphorylation of PI-TP␤ in Vitro by Protein Kinase C-His 6 -tagged PI-TP␤ (1 and 2 g) was phosphorylated in a reaction volume of 60 l containing 20 mM Tris-HCl, pH 7.5, 7.5 mM magnesium acetate, 10 g/ml leupeptin, 10 M ATP, and 1-2 Ci of [␥-32 P]ATP. The Ca 2ϩ / phospholipid-independent phosphorylation was determined in the presence of 1 mM EGTA, and the Ca 2ϩ /phospholipid-dependent phosphorylation was determined in the presence of 1 mM Ca 2ϩ , 96 g/ml phosphatidylserine, and 3.2 g/ml diacylglycerol. The mixture was incubated for 10 min at 37°C, and the reaction was terminated by the addition of 600 l of cold acetone. Bovine serum albumin (1 g) was added, and after 30 min on ice the precipitated protein was spun down, dissolved in sample buffer (125 mM Tris-HCl, pH 6.8, 5% (w/v) SDS, 12.5% (v/v) 2-mercaptoethanol, and 10% (v/v) glycerol), and analyzed by SDS-PAGE (15% gel) followed by autoradiography.
In some experiments, phosphorylation of PI-TP␤ (2 g/assay) was optimized by increasing the amount of ATP to 1 mM and by extending the time of incubation to 30 min. To estimate the extent of phosphorylation, the samples were submitted to SDS-PAGE. The band containing 32 P-labeled PI-TP␤ was cut out from the dried gel, and the labeled protein was eluted by incubation with Soluene-350 (Packard Bioscience) for 2 h at 50°C. Radioactivity was determined by liquid scintillation counting. The stoichiometry of phosphorylation was calculated from the PI-TP␤ 32 P radioactivity, the amount of protein applied to the gel, and the specific activity of ATP. Phosphopeptide and Phosphoamino Acid Analysis-After identification by autoradiography, the 32 P-labeled bands were excised from the dried gel and eluted as described by Boyle et al. (22). Briefly, the gel slices were homogenized in 50 mM ammonium bicarbonate, pH 7.3-7.6. SDS (final concentration, 0.1%) and 2-mercaptoethanol (final concentration, 1%) were added, and the sample was boiled for 5 min. After incubation of the mixture at 37°C for 2 h the gel was spun down, and the supernatant containing the 32 P-labeled proteins was collected. A second elution with 0.1% SDS and 1% 2-mercaptoethanol was carried out on the gel pellet. Carrier protein (boiled RNase, 10 g) and trichloroacetic acid (final concentration, 10%) were added to the combined supernatant fractions, and the samples were incubated on ice for 1 h. The trichloroacetic acid precipitate was washed with cold ethanol and dried. For phosphoamino acid analysis, the pellet was dissolved in 6 M HCl and hydrolyzed for 1 h at 110°C. The HCl was removed by lyophilization, and the pellet was dissolved in pH 1.9 buffer: glacial acetic acid/formic acid/H 2 O (88%) (78:25:897, v/v/v). A mixture of phosphoserine, phosphothreonine, and phosphotyrosine (1 g of each) was added. The 32 P-labeled phosphoamino acids were separated by twodimensional electrophoresis on 20 ϫ 20-cm cellulose TLC plates. The first dimension was in buffer, pH 1.9, and the second dimension was in glacial acetic acid/pyridine/H 2 O (50:5:945, v/v/v), pH 3.5. After electrophoresis the plates were dried, the phosphoamino acids were visualized by staining with 0.2% (w/v) ninhydrin in acetone, and the 32 P-labeled amino acids were identified by autoradiography.
For phosphopeptide mapping, the trichloroacetic acid pellet was dissolved in performic acid, and oxidation was performed for 1-2 h on ice. After lyophilization the sample was incubated with TPCK-trypsin in 50 mM ammonium bicarbonate (200 g/ml) at 37°C for 5 h. The incubation was repeated by the addition of fresh trypsin, and the sample was lyophilized. The phosphopeptides were separated on cellulose TLC plates. In the first dimension, electrophoresis was performed using the pH 1.9 buffer; in the second dimension, TLC was performed in n-butyl alcohol/pyridine/glacial acetic acid/H 2 O (75:50:15:60, v/v/v/v). Radioactive phosphopeptides were identified by autoradiography.
pBK-CMV-PI-TP␤ Constructs for Transfection of NIH3T3 Fibroblast Cells-PI-TP␤ (S262A) cDNA was isolated from the pBluescript SK Ϫ vector by digestion with BamHI and SacI and cloned into the corresponding sites of the pBK-CMV expression vector. PI-TP␤ expression was regulated by the cytomegalovirus immediate early promotor, and the SV40 poly(A) adenylation signal provided the signal for termination of eukaryotic transcription and polyadenylation.
Cell Culture and Transfection-NIH3T3 mouse fibroblasts were maintained in Dulbecco's modified Eagle's medium containing 10% newborn calf serum and buffered with NaHCO 3 (44 mM) at 5% CO 2 in a humidified atmosphere. NIH3T3 cells were transfected using the method of Chen and Okayama (23). Briefly, cells were seeded 5 h prior to transfection at 1 ϫ 10 4 cells/cm 2 and then transfected with 20 g of the pBK-CMV-PI-TP␤ constructs. After another 24 h, G418 (0.4 mg/ml) was added for selection of G418-resistant cells. Fresh medium containing G418 was added every 4 days, and resistant clones, denoted as SPI␤(S262A) cells, were identified after 3 weeks of growth.
Gel Electrophoresis and Western Blotting-The PI-TP␤ content of several G418-resistant clones was analyzed by immunoblotting with anti-PI-TP␤ antibodies. Cells were washed twice with PBS and lysed in 500 l of lysis buffer (20 mM Tris-HCl, 150 mM NaCl, 50 mM NaF, 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1% (w/v) Nonidet P-40). After centrifugation at 17,500 ϫ g for 10 min at 4°C, the protein concentration of the supernatant was determined using the Lowry assay (24). A 150-g portion of protein was subjected to SDS-PAGE on a 15% gel and analyzed by Western blotting using a PI-TP␤-specific antibody. The PI-TP␤ levels on the immunoblot were quantified using a Bio-Rad GS700 imaging densitometer equipped with an integrating program. Known amounts of PI-TP␤ were used as a standard.
To estimate the extent of phosphorylation of PI-TP␤, cell lysates were subjected to two-dimensional PAGE analysis. In some experiments GF 109203X was added to the cells for 16 h prior to lysis. Cells were lysed in 400 l of 20 mM Tris-HCl, pH 8.0, containing 1% Triton X-100. After centrifugation at 17.500 ϫ g for 10 min at 4°C, 100 g of supernatant protein was precipitated using the Two-Dimensional Clean-Up kit (Amersham Biosciences) according to the manufacturer's instructions. The protein pellets were solubilized in 150 l of sample buffer (7.7 M urea, 2.2 M thiourea, 2% Triton X-100, 2% CHAPS, 50 mM dithiothreitol, 0.2% carrier ampholytes (pH 3-10), and 0.0002% bromphenol blue) and run on 7-cm immobilized pH gradient strips, pH 5-8, (Bio-Rad) for a total of 24,000 V-h. The strips were equilibrated for 15 min in 6 M urea, 50 mM Tris-HCl, pH 8.8, 2% SDS, 20% glycerol, and 2% (w/v) dithiothreitol and for an additional 15 min in the same solution except that dithiothreitol was replaced with 2.5% iodoacetamide. Finally, the strips were run in the second dimension on a 10% SDS-PAGE Mini Protean 3 gel (Bio-Rad) and further analyzed by Western blotting using the PI-TP␤-specific antibody.
Immunolocalization-The localization of PI-TP␤ was determined by indirect immunofluorescence using a polyclonal PI-TP␤-specific antibody. Cells were grown on glass coverslips. In some experiments the 14. Briefly, cells were seeded at 0.8 ϫ 10 5 cells/55 cm 2 (wtNIH3T3 cells) and 1.0 ϫ 10 5 cells/55 cm 2 (SPI␤ and SPI␤(S262A) cells) in order to obtain identical cell densities at the time of the experiment. After 2 days of growth, the medium was removed, and the cells were labeled for 60 h with 0.5 Ci/ml [methyl-3 H]choline chloride in Dulbecco's modified Eagle's medium containing 10% newborn calf serum. The cells were washed once with PBS and chased for 2 h with fresh medium. After washing with PBS again, the cells were treated with 200 milliunits/ml bSMase for 30 min. The cells were washed twice with PBS to remove the bSMase and subsequently incubated in fresh medium for 6 h. The cells were harvested by scraping in PBS and sedimented by centrifugation at 350 ϫ g for 5 min. Cell pellets were resuspended in distilled water. An aliquot of the lysate was used for protein determination, and the remainder was used for lipid extraction according to the method of Bligh and Dyer (25

Determination of the Phosphorylation Sites-Prediction
analysis of the PI-TP␤ amino acid sequence indicated the presence of eight putative sites for PKC-dependent phosphorylation. These consensus sites contained six threonine (Thr 58 , Thr 132 , Thr 168 , Thr 197 , Thr 250 , and Thr 256 ) and two serine (Ser 165 and Ser 262 ) residues. Phosphoamino acid analysis of in vitro phosphorylated His 6 -tagged wtPI-TP␤ demonstrated that this protein was exclusively phosphorylated on a serine residue (Fig. 1).
To establish which serine residues were phosphorylated by PKC, mutants of PI-TP␤ were generated in which either Ser 165 or Ser 262 was replaced with an alanine residue. A double mutant in which both Ser 165 and Ser 262 were replaced was also prepared. His 6 -tagged-wtPI-TP␤, -PI-TP␤(S165A), -PI-TP␤(S262A), and -PI-TP␤(S165A/S262A) purified by affinity chromatography were phosphorylated by PKC at two different protein concentrations ( Fig. 2A). As shown in Fig. 2B, wtPI-TP␤ was a substrate for PKC (lanes 1-4). Phosphorylation was strongly reduced in the absence of Ca 2ϩ and PS (lanes  1 and 3). Phosphorylation of PI-TP␤(S165A) was comparable with that of wtPI-TP␤, indicating that Ser 165 had little or no phosphorylation by PKC (lanes 5-8). PI-TP␤(S262A) was a bad substrate for PKC, indicating that Ser 262 was the major site of phosphorylation (lanes 9 -12). The double mutant PI-TP␤(S165A/S262A) was not phosphorylated (lanes 13-16). These data show that PI-TP␤ contains two phosphorylation sites of which Ser 262 is the major and Ser 165 the minor site.
Phosphopeptide Mapping-Two-dimensional analysis of 32 Plabeled peptides derived from a tryptic digest of wtPI-TP␤ showed three major and several minor spots (Fig. 3, left panel). The tryptic peptide map of 32 P-labeled PI-TP␤(S165A) yielded the same three major spots (Fig. 3, middle panel), strongly suggesting that all three spots contained phosphorylated Ser 262 as a result of partial cleavage of the bonds in the sequence Arg-Lys-Lys-Gly-Ser 262 -Val-Arg. This was confirmed by the tryptic peptide map of PI-TP␤(S262A), which lacked these three spots (Fig. 3, right panel). Instead, phosphorylation of PI-TP␤(S262A) yielded one spot representing the peptide containing 32 P-labeled Ser 165 . This labeled peptide (indicated by the arrowheads in Fig. 3) was absent from the phosphopeptide map of PI-TP␤(S165A) but was present in the phosphopeptide map wtPI-TP␤ as a very minor spot. This indicates that the phosphorylation of wtPI-TP␤ is almost exclusively restricted to Ser 262 .
Phospholipid Transfer Activity of wtPI-TP␤ and the PI-TP␤ Mutants-The PI transfer activity of the PI-TP␤s was determined in the continuous fluorescent phospholipid transfer assay. As shown in Fig. 4A, PI-TP␤(S262A) and wtPI-TP␤ expressed an equal activity toward PI (curves 1 and 2). However, when Ser 165 was replaced with an alanine residue, the ensuing mutants PI-TP␤(S165A) and PI-TP␤(S165A/S262A) were completely inactive in the PI transfer assay (curves 3-4). The PC and SM transfer activities of the PI-TP␤s were also determined, and the results were comparable with those observed for the PI transfer activity (data not shown).
To establish whether phosphorylation had an effect on transfer activity, wtPI-TP␤ was phosphorylated by PKC under optimized conditions (1 mM ATP, 30 min of incubation), yielding a stoichiometry of 0.5 mol of phosphate/mol PI-TP␤. Under these conditions the SM transfer activity was not affected (Fig. 4B). Phosphorylation of PI-TP␤ also had no effect on the PC and PI transfer activities (data not shown).
Overexpression of PI-TP␤(S262A) in NIH3T3 Cells-To establish the effect of the major site of phosphorylation on the cellular function of PI-TP␤, mouse NIH3T3 cells were transfected with the pBK-CMV-PI-TP␤(S262A) construct. Stable cell lines were selected using the antibiotic G418; from the G418resistant clones that appeared after 3 weeks of selection, 7 cell lines were analyzed by Western blotting. Two cell lines, designated as SPI␤(S262A)6 and SPI␤(S262A)7, were selected for further experiments. By scanning the immunoblot, it was estimated that SPI␤(S262A) 6  were subjected to two-dimensional PAGE followed by Western blotting using a PI-TP␤-specific antibody. In the first dimension an immobilized pH gradient strip (pH 5-8) was used. The Western blot of the entire two-dimensional gel is shown. Phosphorylated PI-TP␤ runs at pH 6.2 and its non-phosphorylated form at pH 6.5. For further details, see "Experimental Procedures." were analyzed by two-dimensional PAGE and Western blotting using an anti-PI-TP␤ antibody. In initial experiments, immobilized pH gradient strips of pH 3-10 were used showing two spots of PI-TP␤ (estimated pH 6.5) that were poorly separated. To improve resolution, further analysis was carried out using strips of pH 5-8. As shown in Fig. 5A, PI-TP␤ collected in a minor spot at pH 6.5 and a major spot at pH 6.2. Incubation of the SPI␤ cells with the PKC inhibitor GF 109203X (5 M) resulted in a shift of PI-TP␤ from the pH 6.2 to the pH 6.5 form (Fig. 5B). The shift to the pH 6.5 form was virtually complete when the cells were incubated with 10 M PKC inhibitor (Fig.  5C). From this we infer that the pH 6.2 spot represents the phosphorylated form. In support of this, a similar analysis of lysates from SPI␤(S262A) cells demonstrated that the mutated PI-TP␤ collected exclusively at pH 6.5 (Fig. 5D). These data indicate that in situ PI-TP␤ is phosphorylated at Ser 262 and that this phosphorylation is dependent on PKC. From densitometric analysis we estimate that in situ PI-TP␤ is at least 85% phosphorylated. Addition of the antibody to the cell lysate followed by incubation with protein A linked to beads failed to immunoprecipitate PI-TP␤. Hence, immunoprecipitation of PI-TP␤ from cells labeled with inorganic 32 P could not be carried out.
Ser 262 and the Golgi Localization of PI-TP␤-In a previous study it was shown that PI-TP␤ in Swiss mouse 3T3 fibroblasts was mainly associated with the perinuclear Golgi (7,8). The Golgi localization of PI-TP␤ was confirmed in SPI␤ cells by indirect immunofluorescence using a specific anti-PI-TP␤ antibody (Fig. 6, panel A). Because of the very low amount of endogenous PI-TP␤, this Golgi labeling was hardly visible in NIH3T3cells (data not shown). Incubation of SPI␤ cells with the PKC inhibitor GF 109203X resulted in a redistribution of PI-TP␤ throughout the cell (panel B), strongly suggesting that phosphorylation of Ser 262 was a prerequisite for Golgi localization. This was confirmed by showing that PI-TP␤(S262A) was also present throughout the cytoplasm (panel C).
Effect of PI-TP␤(S262A) Overexpression on SM Synthesis-Previously it was shown that plasma membrane levels of SM were maintained in SPI␤ cells under conditions in which SM was hydrolyzed to ceramide by exogenous sphingomyelinase (14). This was not observed in wtNIH3T3 cells, which suggested that a certain level of PI-TP␤ was required for maintaining the steady-state SM level. In order to investigate whether the Golgi localization of PI-TP␤ played a role in the rapid conversion of ceramide into SM, the above experiment was repeated with the SPI␤(S262A) cell lines. Cells were incubated with [ 3 H]choline to label cellular SM and subjected to treatment with bSMase. After 30 min of SM degradation, the bSMase was removed, and the cells were allowed to recover in fresh medium for 6 h. After the incubation with bSMase, the hydrolysis of SM amounted to ϳ35% in NIH3T3, SPI␤, and SPI␤(S262A) cells (Fig. 7). At the end of the 6-h recovery period, SM levels were not restored in NIH3T3 and the SPI␤(S262A) cells, whereas SM was restored to basal levels in SPI␤ cells. Given that the in vitro SM transfer activity of PI-TP␤(S262A) was normal, these results strongly suggest that the Golgi localization is required for PI-TP␤ to be able to stimulate SM resynthesis. DISCUSSION In this study we have shown that PI-TP␤ is a substrate for PKC. By mutation analysis we could establish that Ser 262 was the main phosphorylation site. However, Ser 165 , which is analogous to Ser 166 in PI-TP␣, was also phosphorylated although to a very limited extent. In a previous study we had shown that the PKC-dependent phosphorylation of PI-TP␣ was restricted to Ser 166 (18). The PI-TP␣ isoform, which is highly homologous to PI-TP␤ (similarity of 94%), lacks Ser 262 (9). Assuming that phosphorylation of the PI-TP isoforms is important for the regulation of their function, we were interested to establish the role of Ser 262 in the function of PI-TP␤.
The peptide maps of wtPI-TP␤ and PI-TP␤(S165A) show three major spots, which are absent from the map of PI-TP␤(S262A) (Fig. 3). This indicates that these three spots represent peptides containing phosphorylated Ser 262 . The formation of these peptides is probably due to the presence of multiple tryptic cleavage sites in the amino acid sequence of the peptide Met-Arg-Lys-Lys-Gly-Ser 262 -Val-Arg. Partial digestion would yield the peptides Lys-Lys-Gly-Ser-Val-Arg, Lys- Gly-Ser-Val-Arg, and Gly-Ser-Val-Arg. According to the method described in Ref. 22, we could assign Lys-Lys-Gly-Ser-Val-Arg to spot 1, Lys-Gly-Ser-Val-Arg to spot 2, and Gly-Ser-Val-Arg to spot 3. The tryptic map of PI-TP␤(S262A) showed one spot representing a peptide containing phosphorylated Ser 165 . This spot was barely visible in the peptide map of wtPI-TP␤, indicating that phosphorylation was almost exclusively restricted to Ser 262 .
Replacement of Ser 165 with Ala yielded PI-TP␤(S165A) and the double mutated PI-TP␤(S165A/S262A), both of which in vitro completely lacked PI, PC, and SM transfer activity. On the other hand, PI-TP␤(S262A) was fully active. Mutation of the corresponding serine (Ser 166 ) in PI-TP␣ also yielded an inactive protein (18). From the three-dimensional structure it can be inferred that Ser 166 is exposed at the surface as part of the regulatory loop of PI-TP␣ (26). Hence it is possible that the loss of transfer activity is due to the inability of PI-TP␣(S166A) to properly interact with the membrane interface. However, at this stage we cannot exclude the possibility that replacement of Ser 166 with Ala affects the proper folding of the protein during synthesis in E. coli. The same explanations may hold for the lack of transfer activity observed in PI-TP␤(S165A) and PI-TP␤(S165A/S262A). After PKC-dependent phosphorylation, the phospholipid transfer activity of PI-TP␤ was unchanged, indicating that phosphorylation of Ser 262 had no effect (Fig. 4).
The lysates from SPI␤ cells contained two forms of PI-TP␤ that could be separated by isoelectric focusing. From densitometric analysis it was estimated that 85% of the PI-TP␤ collected at pH 6.2 and 15% at pH 6.5. Treatment of the cells with the PKC inhibitor GF 109203X shifted PI-TP␤ to pH 6.5, strongly suggesting that the spot at pH 6.2 represented the phosphorylated form of PI-TP␤. Because the lysate from the SPI␤(S262A) cells contained predominantly PI-TP␤ at pH 6.5, we conclude that in situ PI-TP␤ is constitutively phosphorylated at Ser 262 . Given that GF 109203X inhibits conventional and novel type PKCs (27, 28), we do not know which PKC isoform is involved in the phosphorylation of PI-TP␤. Because it is unlikely that PKC is constitutively active in these cells, it appears that phosphorylated PI-TP␤ in association with the Golgi is a poor substrate for protein phosphatase. In a previous study PI-TP␤ isolated from bovine brain could not be phosphorylated by PKC (8). Because bovine brain protein is 99% identical to murine PI-TP␤, we consider it likely that in this case also PI-TP␤ is mainly present in its phosphorylated form.
It has previously been reported that in Swiss mouse 3T3 fibroblasts, PI-TP␤ was predominantly associated with the Golgi (8). In the present study we have confirmed that PI-TP␤ was associated with the Golgi complex in the SPI␤ cells (Fig. 6,  panel A). By incubating SPI␤ cells with GF 109203X, a relocation of PI-TP␤ from the Golgi to the cytoplasm was observed (panel B). A similar distribution throughout the cytoplasm was observed for PI-TP␤(S262A) expressed in NIH3T3 cells (panel C). These observations indicate that Ser 262 has to be phosphorylated for PI-TP␤ to be associated with the Golgi system. It is to be noted that the phosphorylation site Ser 262 is only present in PI-TP␤, whereas the phosphorylation site Ser 165/166 is conserved in all PI-TPs identified so far, with the exception of PI-TP from Caenorhabditis elegans (26).
In contrast to wtPI-TP␤, PI-TP␤(S262A) that is overexpressed in mouse fibroblasts is not able to stimulate the resynthesis of SM after the breakdown of this lipid by sphingomyelinase. Because the mutant protein expresses full lipid transfer activity in vitro, we infer that the association of PI-TP␤ with the Golgi is a prerequisite for PI-TP␤ to stimulate rapid SM replenishment. SM and cholesterol regulation in the Golgi has also been linked to the Golgi localization and phosphorylation of the oxysterol-binding protein (29). Similar to its yeast analog Sec14p, PI-TP␤ may play a role in the budding of SM-containing vesicles from the Golgi (10,30). It has been well established that the intracellular transport of SM is linked to the assembly and dynamics of lipid rafts (31). We are currently investigating whether PI-TP␤ is involved in this process.