Evidence for phosphorylation of CTP:phosphocholine cytidylyltransferase by multiple proline-directed protein kinases.

Reversible phosphorylation of CTP:phosphocholine cytidylyltransferase, the rate-limiting enzyme of phosphatidylcholine biosynthesis, is thought to play a role in regulating its activity. In the present study, the hypothesis that proline-directed kinases play a major role in phosphorylating cytidylyltransferase is substantiated using a c-Ha-ras-transfected clone of the human keratinocyte cell line HaCaT. Cellular extracts from epidermal growth factor-stimulated HaCaT cells and from ras-transfected HaCaT cells phosphorylated cytidylyltransferase much stronger as compared with extracts from quiescent HaCaT cells. The tryptic phosphopeptide pattern of cytidylyltransferase phosphorylated by cell-free extracts from ras-transfected HaCaT cells was similar compared with the patterns of cytidylyltransferase phosphorylated by p44mpkmitogen-activated protein kinase and p34cdc2 kinase in vitro, whereas in the case of casein kinase II the pattern was different. Furthermore, in c-Ha-ras-transfected HaCaT cells the in vivo phosphorylation state of cytidylyltransferase was 2-fold higher as compared with untransfected HaCaT cells. This higher phosphorylation of cytidylyltransferase in the ras-transfected clone was reduced to a level below the phosphorylation of cytidylyltransferase in untransfected cells, using olomoucine, a specific inhibitor of proline-directed kinases. The reduced phosphorylation of cytidylyltransferase in olomoucine-treated cells correlated with an enhanced stimulation of enzyme activity by oleic acid.

Reversible phosphorylation of CTP:phosphocholine cytidylyltransferase, the rate-limiting enzyme of phosphatidylcholine biosynthesis, is thought to play a role in regulating its activity. In the present study, the hypothesis that proline-directed kinases play a major role in phosphorylating cytidylyltransferase is substantiated using a c-Ha-ras-transfected clone of the human keratinocyte cell line HaCaT. Cellular extracts from epidermal growth factor-stimulated HaCaT cells and from ras-transfected HaCaT cells phosphorylated cytidylyltransferase much stronger as compared with extracts from quiescent HaCaT cells. The tryptic phosphopeptide pattern of cytidylyltransferase phosphorylated by cell-free extracts from ras-transfected HaCaT cells was similar compared with the patterns of cytidylyltransferase phosphorylated by p44 mpk mitogen-activated protein kinase and p34 cdc2 kinase in vitro, whereas in the case of casein kinase II the pattern was different. Furthermore, in c-Ha-ras-transfected HaCaT cells the in vivo phosphorylation state of cytidylyltransferase was 2-fold higher as compared with untransfected HaCaT cells. This higher phosphorylation of cytidylyltransferase in the ras-transfected clone was reduced to a level below the phosphorylation of cytidylyltransferase in untransfected cells, using olomoucine, a specific inhibitor of proline-directed kinases. The reduced phosphorylation of cytidylyltransferase in olomoucine-treated cells correlated with an enhanced stimulation of enzyme activity by oleic acid.
In mammalian cells, the main pathway for the biosynthesis of phosphatidylcholine (PC) 1 is via CDP-choline and CTP:phosphocholine cytidylyltransferase (EC 2.7.7.15) (CT) is the ratelimiting enzyme of this pathway (1). In addition to the structural function as a component of cellular membranes, PC has been identified to be involved into signal transduction via the PC cycle (2). In a recent study, the coordination of PC metab-olism with the cell cycle was investigated, and evidence was provided that the net biosynthesis of PC is restricted to the S-phase of the cell cycle (3).
CT exists as a soluble, inactive form that can be activated by translocation to membranes. Many mechanisms that modulate this translocation process have been firmly established (4 -7), and reversible phosphorylation of CT has been shown to influence translocation of CT between cytosol and membranes (7)(8)(9)(10). In many systems cytosolic CT is highly phosphorylated, whereas the membrane-bound, active form of CT is dephosphorylated. However, in a recent study it was shown that dephosphorylation of CT is not required for membrane binding (11).
The physiological role of phosphorylation and the kinases involved in this process are still discussed. Whereas CT is a substrate for cAMP-dependent kinase in vitro (12), neither cAMP-dependent protein kinase (13)(14)(15) nor protein kinase C (16) seem to phosphorylate CT in vivo. In rat hepatocytes and HeLa cells only serine residues of CT are phosphorylated in vivo (14,17), and the phosphorylation sites of CT from rat liver were identified (18). This study revealed that phosphorylation of CT is confined to the carboxyl-terminal region of CT and that many serine residues reside in potential sites for proline-directed kinases. We have shown recently that growth factors can stimulate phosphorylation of CT in HeLa cells and that CT is a substrate for p44 mpk MAP kinase in vitro (19), suggesting that the ras/Raf/MAP kinase-signaling pathway is involved in this process.
In the present study, we investigated the phosphorylation of CT in vivo and in vitro. In ras-transfected HaCaT cells, the phosphorylation of CT was increased by 2-fold as compared with the phosphorylation of CT in untransfected cells. The enhanced phosphorylation of CT in ras-transfected cells was reduced in the presence of olomoucine, a specific inhibitor of proline-directed kinases, such as p34 cdc2 and p44 mpk MAP kinase (20). Using this experimental approach we could show that phosphorylation of CT interferes with the activation of the enzyme by oleic acid and protects the enzyme against proteolytic digestion. Furthermore, cell-free extracts from quiescent and EGF-stimulated HaCaT cells as well as ras-transfected HaCaT cells were used to phosphorylate CT in vitro. The tryptic phosphopeptide patterns of CT phosphorylated with cellfree extracts from HaCaT and ras-transfected HaCaT cells were compared with the patterns of CT phosphorylated with purified MAP kinase, cyclin-dependent kinase, and casein kinase II in vitro. The results presented here substantiate the hypothesis that p44 mpk MAP kinase is involved in the phosphorylation of CT, but other proline-directed kinases, such as p34 cdc2 kinase, represent probable candidates as well.  1 The abbreviations used are: PC, phosphatidylcholine; CT, CTP: phosphocholine cytidylyltransferase; MAP, mitogen-activated protein; EGF, epidermal growth factor; MBP, myelin basic protein; SA 2, polyclonal peptide-specific antibody against cytidylyltransferase residues 1-17; PAGE, polyacrylamide gel electrophoresis; MOPS, 4-morpholinepropanesulfonic acid. schweig, Germany). [␥-32 P]ATP (110 TBq/mmol) was from Hartmann Analytic (Braunschweig, Germany). 33 P as orthophosphoric acid (370 MBq/ml) was from ICN Biomedicals (Meckenheim, Germany). Silica Gel 60 high performance thin layer chromatography plates, cellulose plates, and all solvents and reagents (reagent grade) were purchased from Merck (Darmstadt, Germany). Dithiothreitol, phenylmethylsulfonyl fluoride, myelin basic protein, oleic acid, phosphatidylcholine, and staurosporine were from Sigma (Mü nchen, Germany). Olomoucine and iso-olomoucine were from LC Laboratories (Grü nberg, Germany), and stock solutions of 40 mM were prepared in Me 2 SO. Human EGF was from Biomol (Hamburg, Germany). Polyclonal MAP kinase antibody (ERK-1) was from Santa Cruz Biotechnology (Santa Cruz, CA). p44 mpk MAP kinase and p34 cdc2 kinase were from UBI (Lake Placid, NY), and casein kinase II was purchased from Boehringer (Mannheim, Germany).

Materials-The
Cell Culture-The spontaneously immortalized human keratinocyte cell line HaCaT (21) and the ras-transfected clone I-7 (22) were gifts from Dr. N. Fusenig (Deutsches Krebsforschungzentrum, Heidelberg, Germany). Cells were grown in RPMI medium supplemented with 10% heat-inactivated fetal calf serum, 0.35 g/liter glutamine, 100 000 IU/ liter penicillin, and 0.1 g/liter streptomycin in plastic culture dishes (Nunc, Denmark). Media and culture reagents were obtained from Life Technologies, Inc. (Karlsruhe, Germany). Penicillin and streptomycin were from Boehringer. For the experiments, normal HaCaT cells were used between passages 32 and 45, and ras-transfected HaCaT cells were used between passages 12 and 25.
Prior to EGF stimulation, confluent cells were starved for 24 h in keratinocyte basal medium (Clonetics, San Diego, CA) containing no supplements. Under these conditions, proliferation of HaCaT cells and ras-transfected HaCaT cells was totally abolished. 33 P Labeling of Cells-For this, 2 ϫ 10 6 confluent cells were washed two times with 5 ml of phosphate-free Dulbecco's minimal essential medium (Life Technologies, Inc.) and then incubated in 2 ml of phosphate-free Dulbecco's minimal essential medium containing 300 Ci of carrier-free [ 33 P]PO 4 and different concentrations of olomoucine or isoolomoucine. After 2 h the cells were washed twice with ice-cold phosphate-buffered saline and lysed by sonication in 1 ml of 50 mM Tris-HCl, pH 7.2, 150 mM NaCl, 2 mM EGTA, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% sodium dodecylsulfate, 1 mM phenylmethylsulfonyl fluoride, 5 g/ml leupeptin, 20 mM NaF, and 1 mM sodium vanadate. Cytidylyltransferase was immunoprecipitated from the samples as described below using CT antibody SA 2.
MAP Kinase Assay-Stimulation of MAP kinase activity was measured by a modified method of Dubois and Bensaude (23). After treatment with keratinocyte basal medium containing 100 ng/ml EGF, cells (2 ϫ 10 6 /dish) were washed with ice-cold buffer A (12.5 mM MOPS (pH 7.2), 20 mM sodium glycerophosphate, 7.5 mM magnesium cloride, 1 mM ethylene glycol tetraacetic acid, 1 mM sodium vanadate, 0.2 mM phenylmethylsulfonyl fluoride, 5 mg/ml leupeptin, 1 mM dithiothreitol, 100 M sodium fluoride) and lysed on ice with 500 l of buffer A containing 1% Triton X-100. Cytosolic extracts were cleared by centrifugation for 2 min at 13,000 ϫ g. Phosphorylation of myelin basic protein (MBP) was performed by incubating 2 l of cytosolic extracts at 30°C with 20 l of buffer A containing [␥-32 P]ATP (50 Ci/ml, 50 M), staurosporine (1 M) and MBP (0.4 mg/ml) for 20 min. The reaction was stopped by the addition of 5 ϫ Laemmli buffer (24). After electrophoresis in 12.5% polyacrylamide gels, the gels were dried and autoradiographed. Bands corresponding to MBP were excised from the gel, and incorporated radioactivity was determined by scintillation counting.
Western Blot Analysis-Immunoblotting was performed as described (25). Briefly, 2 ϫ 10 6 cells that had been previously treated with 100 ng/ml EGF in keratinocyte basal medium were homogenized in 300 l of buffer containing 50 mM Tris-HCl (pH 7.2), 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% sodium dodecylsulfate, 1 mM sodium vanadate, 1 mM NaF, 1 mM phenylmethylsulfonyl fluoride, and 5 g/ml leupeptin. Proteins were separated by 10% SDS-PAGE and transferred to nitrocellulose membrane. Both MAP kinase isoforms, p42 mpk and p44 mpk , were detected with a polyclonal rabbit antibody (1:200) from Santa Cruz Biotechnology. The same procedure was applied to detect CT in cellular extracts from HaCaT cells using CT antibody SA 2.
Construction of Recombinant Baculovirus and Expression of Cytidylyltransferase-CT was expressed in Sf 21 insect cells using the XPRESS System® from Invitrogen (San Diego, CA) as described (26) with minor modifications. Briefly, rat CT cDNA clone, which was a gift of Dr. R. Cornell (Burnaby, BC, Canada), was digested with BamHI. The fragment was ligated into BamHI-digested and phosphatasetreated pBlue Bac His vector. The pCT Blue Bac His plasmid DNA was mixed with wild type linear AcNPV DNA (3:1 ratio) and cotransfected into Sf 21 cells using cationic liposomes, according to the manufacturer's instructions. The culture supernatant was collected, titered, and used to infect a lawn of Sf 21 cells that were overlaid with agarose. Blue plaques were picked and subjected to three cycles of plaque purification until cells with inclusion bodies were not detected. For expression of CT, 9 ϫ 10 6 Sf 21 cells in a 15-cm diameter dish were infected with purified recombinant baculovirus in TC-100 medium (Life Technologies, Inc.) supplemented with 10% fetal calf serum. Expression of CT was assessed by Western blot analysis and measurement of CT activity. 64 h after infection, Sf 21 cells were collected in medium, pelleted, and stored at Ϫ80°C.
In Vitro Phosphorylation of Cytidylyltransferase-Rat liver CT expressed in Sf21 insect cells was used for in vitro phosphorylations. 20 nmol/min of CT activity was immunoprecipitated using the antibody SA 2 (25) bound to protein A-Sepharose. To immunoprecipitate CT, 50 l of antiserum SA 2 were incubated with 5 mg of protein A-Sepharose in 2 ml of buffer 1 (50 mM Tris (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% SDS, 0.5% sodium deoxycholate) for 3 h at 4°C after which the antibody coupled protein A-Sepharose was sedimented by centrifugation and incubated overnight at 4°C with homogenates of infected Sf 21 cells. The protein A-Sepharose was centrifuged at 13,000 ϫ g for 2 min, and the supernatant was discarded. The immune complex was washed three times with buffer 1 and one time with phosphatebuffered saline. CT was dephosphorylated with 5 units of alkaline phosphatase in 50 mM Tris-HCl (pH 7.7) for 15 or 30 min at 30°C. The immune complex was washed again three times with buffer 1 and one time with phosphorylation buffer containing 15 mM MOPS (pH 7.2), 2 mM dithiothreitol, 12.5 mM ␤-glycerophosphate, 7.5 mM MgCl 2 , 2 mM EGTA, 5 mM NaF, and 1 mM sodium vanadate. Then phosphorylation was performed in the presence of different kinases or cell-free extracts and 100 M [␥-32 P]ATP (20 Ci/assay) at 37°C in a total volume of 200 l. After 30 min the reaction was stopped by adding SDS sample buffer, and the mixtures were loaded on SDS-PAGE (10% acrylamide) and electrophoresed. Proteins were transferred to nitrocellulose membrane, and the membrane was exposed to Kodak XAR 5 film to detect cytidylyltransferase. Bands corresponding to CT were excised, and incorporated radioactivity was determined by scintillation counting.
Two-dimensional Peptide Mapping of Phosphorylated Cytidylyltransferase-Digestion of CT and peptide separation was performed as described by Boyle et al. (27). Briefly, bands corresponding to CT were excised from the nitrocellulose membrane, and the strips were incubated with 0.5% polyvinylpyrrolidone (PVP-360) in 100 mM acetic acid to prevent adsorption of trypsin to the nitrocellulose membrane. The membrane pieces were washed three times with water and subsequently incubated in 200 l of 50 mM ammonium bicarbonate (pH 8.2) containing 10 g of L-1-tosylamido-2-phenylethyl chloromethyl ketonetreated trypsin. Additional fresh trypsin (10 g) was added after 2 and 4 h of incubation. The total incubation time was 6 h. The membrane pieces were washed, and the combined supernatants were dried in a Speed-Vac® and washed by three cycles of resuspension and vacuum drying using 200 l of distilled water for each cycle. Samples were resuspended in 5 l of electrophoresis buffer (water/acetic acid/formic acid, 897:78:25, v/v/v) and spotted on cellulose thin layer chromatography sheets (0.1-mm thickness). Separation of the phosphopeptides was accomplished by electrophoresis at 1000 V for 45 min followed by ascending chromatography in n-butanol/water/pyridine/acetic acid (15: 12:10: Other Procedures-CT activity was measured by a modified method of Sohal and Cornell (28) using liposomes containing 400 M PC and different concentrations of oleic acid as described previously (15). Protein was determined by the BCA assay (29) with BSA as standard. Statistical comparisons were made in these studies with Student's t test.

c-Ha-ras Transfection Alters MAP Kinase Activation in Ha-
CaT Cells-Activation of the MAP kinase cascade was measured by incubating the cells with EGF, and MAP kinase activity was assessed by measuring the rate of phosphorylation of MBP, a known substrate of MAP kinases. Fig. 1A shows the phosphorylation of MBP after 2 and 10 min of EGF stimulation in untransfected HaCaT cells. After 10 min of EGF treatment, phosphorylation of MBP was 220% as compared with the control. In contrast, MBP was highly phosphorylated without EGF stimulation in ras-transfected HaCaT cells, and the addition of Phosphorylation of CTP:Phosphocholine Cytidylyltransferase EGF had no further stimulating effect on phosphorylation of MBP. After 10 min of EGF treatment, incorporation of phosphate into MBP was 95% of basal activity (determined by scintillation counting of excised bands).
MAP kinases are activated by phosphorylation (30). It is known that the phosphorylation of both isoforms, p42 mpk and p44 mpk , leads to a mobility shift in the SDS gel. As shown in the Western blot of Fig. 1B, p42 mpk and p44 mpk are expressed in HaCaT cells and become phosphorylated time-dependently after EGF stimulation. In ras-transfected HaCaT cells, both isoforms of MAP kinase appear as phosphorylated, high molecular weight forms that were not influenced by EGF stimulation.
Phosphorylation of Cytidylyltransferase in Vivo-To investigate the phosphorylation state of CT in cell culture experiments, HaCaT cells and the ras-transfected clone were metabolically labeled with [ 33 P]orthophosphoric acid. After 2 h of incubation time, phosphorylation of CT was elevated by about 2-fold in ras-transfected cells as compared with the phosphorylation of CT in untransfected HaCaT cells (Fig. 2). In additional experiments, the activity of CT was assayed in HaCaT cells and ras-transfected HaCaT cells. In the absence of exogenous liposomes CT activity was 0.31 Ϯ 0.05 nmol/min/mg of protein in HaCaT cells and 0.37 Ϯ 0.03 nmol/min/mg of protein in ras-transfected cells (n ϭ 3), indicating that the increased phosphorylation of CT in ras-transfected cells did not have a significant effect on its basal activity.
To address the question of which kinase(s) phosphorylate CT in the cells, the purine analogue olomoucine was used in cell culture experiments. This compound was shown to specifically inhibit the activity of cell cycle-regulating kinases and p44 mpk MAP kinase (20). Iso-olomoucine, on the contrary, is confirmed as a general kinase inhibitor with much lesser specificity for these kinases. Fig. 2 shows that the phosphorylation of CT observed in ras-transfected HaCaT cells is inhibited by olomoucine in a concentration-dependent manner. In the presence of 100 M olomoucine, the phosphorylation of CT in ras-transfected HaCaT cells is reduced by 72%, whereas the phosphorylation of CT is reduced by only 39% in the presence of the same concentration of iso-olomoucine (see Fig. 2).
Effect of Reduced Phosphorylation on CT Activity and Enzyme Stability-To investigate the role of phosphorylation of CT by proline-directed kinases, ras-transfected HaCaT cells were treated with 100 M olomoucine and 0.25% Me 2 SO as controls for 4 h. As already shown in the previous chapter, proline-directed phosphorylation of CT was significantly reduced by treatment with olomoucine. After extraction of CT, lipid-mediated activation of the enzyme was determined in the presence of different amounts of oleic acid. Whereas the basal activity was not significantly influenced by olomoucine pretreatment of the cells, activation in the presence of 100 and 400 M oleic acid was enhanced by about 2-fold (Fig. 3).
We also tested the possibility that phosphorylation of CT affects enzyme stability to endogenous proteases. For this, cellular extracts from olomoucine-and Me 2 SO-treated cells were incubated for different time periods at 37°C in the absence of protease inhibitors. After the incubation, the reaction was stopped by the addition of SDS-sample buffer and CT, and digestion products of CT were visualized by Western blot analysis using the CT-specific antibody SA 2. As shown in Fig. 4,

FIG. 1. Activation of MAP kinases in HaCaT cells and rastransfected HaCaT cells. A, in vitro phosphorylation of MBP by cell
lysates from quiescent cells without stimulation (0 min) and from cells that had been stimulated for 2 and 10 min with 100 ng/ml EGF. Equal amounts of protein (8 g) were loaded into each lane of the gel. Bands corresponding to MBP were excised from the gel, and incorporated radioactivity was determined by scintillation counting. B, the same cell lysates (25 g protein/lane) were analyzed by Western blot using a rabbit anti-MAP kinase antibody that recognizes both isoforms, p42 mapk and p44 mapk . The phosphorylated high molecular mass forms are indicated by asterisks. The experiments were repeated, and similar results were obtained.

FIG. 2. Phosphorylation of cytidylyltransferase in vivo.
2 ϫ 10 6 ras-transfected HaCaT cells were incubated in phosphate-free medium containing 150 Ci/ml [ 33 P]P i (carrier-free) and different concentrations of olomoucine or iso-olomoucine for 2 h. As an additional control, 2 ϫ 10 6 untransfected HaCaT cells were labeled with phosphate-free medium containing 150 Ci/ml [ 33 P]P i (carrier-free) in the absence of kinase inhibitors. Subsequently, the cells were washed twice with cold phosphate-buffered saline and homogenized by sonication. CT was immunoprecipitated with the antibody SA 2 and immunoprecipitated proteins were separated on SDS-PAGE (10%) as described under "Experimental Procedures." The dried gel was exposed to Kodak XAR-5 film for 24 h. Bands corresponding to CT were analyzed by video densitometry. The optical density of immunoprecipitated CT from rastransfected cells in the absence of kinase inhibitors was defined as 100%. The experiment was repeated, and similar results were obtained. degradation products of CT with an apparent molecular mass below 30 kDa were observed after 20 min in extracts from olomoucine-treated cells. On the other hand, no degradation products of this size were observed in extracts from control cells.
Phosphorylation of Cytidylyltransferase with Cell-free Extracts of HaCaT Cells in Vitro-To substantiate the hypothesis that proline-directed kinases are involved in the phosphorylation of CT, immunoprecipitated CT was incubated with cellfree extracts from HaCaT cells and ras-transfected HaCaT cells (corresponding to the conditions shown in Fig. 1). Phosphorylation of CT was very low when extracts from quiescent HaCaT cells were used in the assay, and the phosphorylation was significantly stimulated by pretreatment of quiescent HaCaT cells with 100 ng/ml EGF. On the other hand, extracts from quiescent ras-transfected HaCaT cells strongly phosphorylated CT, and pretreatment of ras-transfected cells with 100 ng/ml EGF did not cause higher phosphorylation of CT (Fig. 5). The weak phosphorylation of immunoprecipitated CT in the absence of cytosol was probably due to the activity of endogenous kinases (see also Fig. 6) and did not disturb the in vitro assay.
Additionally, the effect of olomoucine on the phosphorylation of CT by cellular extracts from ras-transfected cells in vitro was investigated. The strong phosphorylation of CT catalyzed by cellular extracts from ras-transfected HaCaT cells was reduced in the presence of olomoucine far more potently than in the presence of the unspecific kinase inhibitor iso-olomoucine. To verify that olomoucine inhibits p44 mpk MAP kinase and p34 cdc2 kinase and not casein kinase II in our system, we investigated the effect of olomoucine on the phosphorylation of CT by the purified kinases in vitro. Although the phosphorylation of CT is reduced in the case of p44 mpk MAP kinase and p34 cdc2 kinase, olomoucine has no inhibitory effect on casein kinase II activity (see Table I).
Phosphorylation of Cytidylyltransferase by Different Kinases in Vitro-In order to investigate the phosphorylation of CT in more detail, we used three different kinases in an in vitro phosphorylation assay. Fig. 6 illustrates that immunoprecipitated CT is phosphorylated by MAP kinase (p44 mpk ), cyclin-dependent protein kinase (p34 cdc2 ), and casein kinase II. In all assays equal kinase activities were used (2 ϫ 10 Ϫ11 mol/min). As also mentioned in the previous section, phosphorylation of CT in the absence of exogenous kinases was very low. The radioactivity incorporated into CT was determined by scintillation counting of excised bands, and the background phosphorylation in the control was defined as 100%. In relation to the control, the incorporation of phosphate into CT catalyzed by p44 mpk and p34 cdc2 was 333 and 420%, respectively, whereas incorporation of phosphate into CT was only 165% in the case of casein kinase II.
In another set of experiments, immunoprecipitated CT was phosphorylated by MAP kinase, cyclin-dependent kinase, casein kinase II, and cell-free extracts from quiescent HaCaT cells and ras-transfected HaCaT cells, followed by digestion with L-1-tosylamido-2-phenylethyl chloromethyl ketonetreated trypsin and two-dimensional separation of phosphorylated peptides. In order to obtain a sufficiently high 32 P label in the phosphorylated peptides, the purified kinases were used in this assay at the highest specific activity available. The peptide maps obtained in the presence of purified kinases were compared with the patterns of digested CT phosphorylated by cell-free extracts from quiescent HaCaT cells and ras-transfected HaCaT cells (Fig. 7). Confirming the results in the previous section, phosphorylation of the different peptides by extracts from ras-transfected cells was higher as compared with phosphorylation by extracts from normal HaCaT cells. However, with the exception of spots 10 and 12, the patterns of HaCaT cells and ras-transfected HaCaT cells were identical. Furthermore, the patterns obtained by phosphorylation with MAP kinase and cdc2 kinase (especially the spots numbered 2, 3, 4, 5, 6, 7, 10, and 12) resembled the pattern of ras-transfected HaCaT cells. In contrast, when casein kinase II was used in the assay two predominant phosphopeptides were obtained (spots 4 and 7), resulting in a different phosphopeptide pattern as compared with the other patterns.

DISCUSSION
Reversible phosphorylation is a universal mechanism regulating enzymatic processes in eukaryotic cells. It has been known for some time that CT is regulated by reversible phosphorylation (1,7,9,31). However, the kinases involved in this process remain to be identified. Many approaches have used different protein kinase activators (13,14,16) and kinase inhibitors (15) as well as phosphatase inhibitors (10) to investigate the phosphorylation of CT. In the present paper, a different approach is presented using the human keratinocyte cell line HaCaT and a c-Ha-ras-transfected clone of this cell line. In untransfected HaCaT cells, MAP kinases became activated rapidly after EGF stimulation of the cells. On the other hand, MAP kinases were already fully activated in ras-transfected HaCaT cells without EGF stimulation, confirming the well known effect that EGF treatment and ras transfection stimulate MAP kinase activity (30). Cell culture experiments revealed that ras-transfected HaCaT cells contained a highly phosphorylated form of endogenous CT, suggesting that the activation of MAP kinases might also be important for the phosphorylation of CT in vivo. The increased phosphorylation of CT in ras-transfected cells did not influence CT activity when determined in the absence of stimulating liposomes. However, we could clearly demonstrate that activation of CT by oleic acid was enhanced when phosphorylation of the enzyme was reduced by olomoucine treatment. This is in accordance to previous findings by Yang and Jackowski (32), who showed that a mutant of CT lacking the COOH-terminal phosphorylation domain is much more sensitive to stimulation by lipids when compared with the wild type enzyme. Furthermore, we tested the hypothesis that phosphorylation of CT influences the stability of the enzyme. Using extracts from olomoucinetreated cells that mainly contained dephosphorylated CT, we could demonstrate that CT was more susceptible to digestion by endogenous proteases in vitro. In this context it has been shown that down-regulation of CT by cholecystokinin treatment of pancreatic acinar cells correlates with a decrease in CT phosphate levels, indicating that phosphorylation of CT protects the enzyme from degradation (33).
Phosphorylation of CT by extracts from quiescent HaCaT cells was very low, and EGF stimulation as well as ras transfection of HaCaT cells obviously activated kinases or inactivated phosphatases. As a consequence, the phosphorylation of CT by extracts from EGF-stimulated or ras-transfected HaCaT cells was much stronger. Additionally, olomoucine, a purine analogue that was shown to specifically inhibit cyclin-dependent kinases and p44 mpk MAP kinase (20) reduced phosphorylation of CT in vivo and in vitro, suggesting an involvement of proline-directed kinases in this process.
Phosphorylation of CT by cAMP-dependent protein kinase or PKC has finally been ruled out. Because the cDNA of rat liver CT contains many consensus phosphorylation sequences for proline-directed kinases (like p34 cdc2 kinase and MAP kinase) and one consensus sequence for casein kinase II (34), we tested these kinases in their ability to phosphorylate rat liver CT in vitro. All three kinases were able to phosphorylate CT, and p34 cdc2 kinase catalyzed the strongest incorporation of 32 P into CT. The advantage of the approach presented here, using rat liver CT as a substrate for phosphorylation by cellular extracts and purified kinases, is that the tryptic phosphopeptide patterns of CT can be compared with each other. Using this system, difficulties arising from differences in the amino acid sequence of CT from different species are ruled out. The number of different spots obtained was similar to patterns presented by other groups, using rat liver CT as well (18). Interestingly, the pattern obtained after phosphorylation of CT by cellular extracts from ras-transfected HaCaT cells was similar FIG. 5. Phosphorylation of cytidylyltransferase by cellular extracts from HaCaT cells and ras-transfected HaCaT cells. 2 ϫ 10 6 serum-starved HaCaT cells and ras-transfected HaCaT cells were incubated with EGF (100 ng/ml) or with medium without supplements for 5 min. Cells were homogenized in 200 l of buffer A as described under "Experimental Procedures." 20 l of the extracts were used to phosphorylate rat liver CT that was previously dephosphorylated with 5 units of alkaline phosphatase. CT was immunoprecipitated with protein A-Sepharose using antibody SA 2 and separated by SDS-PAGE. A representative autoradiogram from two independent experiments is shown.

FIG. 6. Phosphorylation of cytidylyltransferase by different kinases in vitro.
Rat liver CT expressed in Sf21 insect cells was immunoprecipitated using protein A-Sepharose and antibody SA 2. The washed immune complex was dephosphorylated with alkaline phosphatase and incubated with 2 ϫ 10 Ϫ11 mol/min p44 MAP kinase, p34 cdc2 kinase, and casein kinase II for 30 min at 30°C. Phosphorylated CT was separated by SDS-PAGE. A representative autoradiogram from two independent experiments is shown. and iso-olomoucine in vitro Rat liver CT was phosphorylated by cellular extracts from ras-transfected HaCaT cells (50 g of cellular protein) and different purified kinases in the presence of different concentrations of olomoucine and iso-olomoucine. Immunoprecipitated CT was separated on SDS-PAGE (10%), and the gel was exposed to Kodak XAR-5 film. Bands corresponding to CT were analysed by video densitometry. The values are given as percentages of radioactivity incorporated in the absence of olomoucine or iso-olomoucine and represent the mean of two independent experiments. to the patterns obtained after phosphorylation by purified p44 mpk MAP kinase and p34 cdc2 kinase. On the other hand, phosphorylation of CT by casein kinase II revealed a different tryptic phosphopeptide pattern, suggesting that this kinase plays a minor role in phosphorylating CT. In this context, it is interesting to note that a mutant of CT with an alanine residue in the potential phosphorylation site for casein kinase II showed the same properties as compared with the wild type enzyme (35). The findings presented here, together with results showing that growth factors stimulate phosphorylation of CT in HeLa cells (19), substantiate the hypothesis that proline-directed kinases, such as MAP kinases or cyclin-dependent kinases, are involved in the phosphorylation of CT.