Oncogenic Ha-Ras transformation modulates the transcription of the CTP:phosphocholine cytidylyltransferase alpha gene via p42/44MAPK and transcription factor Sp3.

We have shown previously that expression of the murine CTP:phosphocholine cytidylyltransferase (CT) alpha gene is regulated during cell proliferation (Golfman, L. S., Bakovic, M., and Vance, D. E. (2001) J. Biol. Chem. 276, 43688-43692). We have now characterized the role of Ha-Ras in the transcriptional regulation of the CTalpha gene. The expression of CTalpha and CTbeta2 proteins and mRNAs was stimulated in C3H10T1/2 murine fibroblasts expressing oncogenic Ha-Ras. Incubation of cells with the specific inhibitor (PD98059) of p42/44(MAPK) decreased the expression of both CT isoforms. Transfection of fibroblasts with CTalpha promoter-luciferase constructs resulted in an approximately 2-fold enhanced luciferase expression in Ha-Ras-transformed, compared with nontransformed, fibroblasts. Electromobility shift assays indicated enhanced binding of the Sp3 transcription factor to the CTalpha promoter in Ha-Ras-transformed cells. Expression of several forms of Sp3 was increased in nuclear extracts of Ha-Ras-transformed fibroblasts compared with nontransformed cells. Tyrosine phosphorylation of one Sp3 form was decreased, whereas phosphorylation of two other forms of Sp3 was increased in nuclear extracts of Ha-Ras-transformed cells. When control fibroblasts were transfected with a Sp3-expressing plasmid, an enhanced expression of CTalpha and CTbeta was observed. However, the expression of CTalpha or CTbeta was not increased in Ha-Ras-transformed cells transfected with a Sp3 plasmid presumably because expression was already maximally enhanced. The results suggest that Sp3 is a downstream effector of a Ras/p42/44(MAPK) signaling pathway which increases CTalpha gene transcription.

Two genes that encode CT activity have been identified and characterized. CT␣ is expressed in many cells and tissues (7)(8)(9)(10)(11)(12)(13) and has a predicted structure that contains catalytic, phosphorylation, and lipid binding domains as well as a nuclear localization sequence (8, 14 -21). Recently, CT␤ has been identified in human tissues and appears to exist as two splice variants, CT␤1 and CT␤2, differing at their C termini (22,23). Like CT␣, CT␤1/2 contains catalytic and lipid binding domains. However, CT␤1 lacks the phosphorylation domain, and both CT␤1 and CT␤2 lack the nuclear localization sequence (23).
In the last several years, studies have shown that CT can be regulated at both the transcriptional level and post-transcriptionally. CT mRNA is increased after partial hepatectomy in rats (13), after stimulation with colony-stimulating factor 1 in macrophages (24), and during development and growth (25)(26)(27)(28). It remains to be determined whether these increases in the message levels are the result of CT␣ and/or CT␤1/2 mRNA stability, an increase in gene transcription, or a combination of both.
The murine CT␣ gene (Ctpct) was cloned and characterized by Tang and co-workers (29). The Ctpct promoter contains several putative elements for binding transcription factors, including Ap1, an overlapping site for nuclear factor-B, E2F, and Elk1, one sterol response element, as well as three elements for Sp-related factors (30). We have shown that Sp1, Sp2, and Sp3 bind competitively to three GC-rich elements and that relative promoter activity depends upon the abundance of these factors (31). We further established that transcription enhancer factor-4 can bind to an upstream regulatory element (Ϫ103/Ϫ82) and enhance the CT␣ gene expression through its interactions with the basal transcriptional machinery (32). In agreement with the finding that lipoprotein deficiency induces the expression of CT␣ mRNA and protein in alveolar type II epithelial cells (33), and our observation that the CT␣ promoter contains a putative sterol response element (30), studies by Kast et al. (34) indicate a role for cholesterol/sterol response element-binding protein and the functionality of the sterol response element in the regulation of CT␣ gene expression in Chinese hamster ovary cells and THP-1 cells. On the other hand, Lagace et al. (35) have shown that cholesterol/sterol response element-binding protein can stimulate PC biosynthe-sis without altering transcription of the CT␣ gene. Recently we discovered that increased transcription of the CT␣ gene occurs during the S phase of the cell cycle in murine fibroblasts (36). However, this type of regulation was not evident in type II lung epithelial cells (37), further suggesting a cell type-specific regulation of the CT␣ gene.
The importance of PC metabolism in cell proliferation has been under intense investigation, and it has become clear that the GTP exchange protein, Ras, might play an important role in linking these two processes (38 -44). Both PC synthesis and degradation are stimulated in Ras-transformed C3H10T1/2 murine embryonic fibroblasts (45), NIH-3T3 fibroblasts (46), and in human keratinocyte cell line, HaCaT (47). The oncogenic transformation in mouse cells leads to increased activity of choline kinase and decreased activity of CT (45,46), whereas in human keratinocytes CT activity and choline uptake were increased, but choline kinase activity did not change (47). Based upon those data and our finding that CT␣ promoter activity and CT␣ mRNA increased after growth stimulation by serum (36), we investigated the role of Ha-Ras in the regulation of the CT␣ gene. We demonstrate that the Ras/p42/44 MAPK signaling pathway plays a role in the regulation of expression of both CT␣ and CT␤ which is at least partially mediated by the transcription factor Sp3.

EXPERIMENTAL PROCEDURES
Antibodies and Reagents-All reagents were molecular biology grade. The inhibitors of p38 MAPK (SB202190) and MEK1 (PD98059) were from Calbiochem. The transfection reagent DOTAP was from Avanti Polar Lipids (Birmingham, AL). Phospho-MAPK antibody sampler containing phospho-p42/44 MAPK , phospho-p38 MAPK , and phospho-SAPK/JNK-specific rabbit polyclonal antibodies was from New England BioLabs, Mississauga, Canada. A murine monoclonal phosphotyrosinespecific antibody, PY20, was from BD Biosciences, Canada. Cell culture media and reagents were from Invitrogen.
Cell Culture and MAP Kinase Inhibition-Murine embryo fibroblasts (C3H10T1/2) and the Ha-Ras-transformed clone ras11A were kindly provided by Dr. C. Kent (University of Michigan) (45). They were grown at 37°C in Dulbecco's minimal essential medium supplemented with 10% fetal calf serum, 100 g/ml streptomycin, and 100 units/ml penicillin under a humidified 5% CO 2 atmosphere at 37°C. The ras11A cell line is neomycin-resistant and was grown in the presence of 400 g/ml G418. During experiments, G418 was not added to the medium, as recommended (45). The MAP kinase inhibitors PD98059 and SB202190 were dissolved in dimethyl sulfoxide. An aliquot of each inhibitor solution was added to the medium, and the final concentration of the vehicle in the medium was adjusted to 0.1% (v/v). The control medium contained the same amount of the vehicle.
Transient Transfections-Fibroblasts were transfected using a DOTAP liposomal method (30). Control and Ha-Ras-transformed fibroblasts were plated at a density of 2 ϫ 10 6 cells/60-mm dish and transfected the next day with 2.5 g of specific CT␣-luciferase reporter plasmid with or without the indicated amounts of Sp1 and Sp3 expression plasmids (pPacSp1 or pPacSp3) or control vector (pPacO). Transfected cells were grown overnight in normal medium, then growth was arrested in a low serum medium (0.5% fetal calf serum) for 2 additional days. The arrested cells were stimulated to grow by the addition of 10% serum and 24 h later collected for further analysis. The pBK⌬Gal vector, encoding ␤-galactosidase, was cotransfected as an internal control to measure differences in transfection efficiency. Luciferase and ␤-galactosidase activities were measured using a luciferase and ␤-galactosidase assay system (Promega). The amount of cellular protein was measured by the Bio-Rad method.
CT Enzymatic Activity-CT activity in total cell homogenates, cy- Immunobloting Analysis of CT␣ and CT␤-Cell lysates from control and Ha-Ras-transformed cells (20 -50 g of protein) were prepared as described previously (36) and separated on 10% denaturing polyacrylamide gels. The proteins were transferred onto polyvinylidene difluoride membranes (Bio-Rad) at ϳ150 mA and 4°C for 1 h. After checking for protein loading with Ponceau S dye, the membranes were incubated overnight with 5% skim milk in TTBS (20 mM Tris-HCl, pH 7.5, 500 mM NaCl, 0.05% Tween 20). The membranes were washed in TTBS and incubated at 25°C for 2 h with polyclonal antibodies raised against CT (anti-M), CT␣, CT␤1/2, and/or CT␤2. The CT (anti-M) antibody was a rabbit polyclonal antibody directed against a peptide corresponding to amino acids 256 -288 (M domain) of rat liver CT␣ from Dr. R. Cornell (51). The anti-CT␣ rabbit polyclonal antibody, corresponding to the first 17 amino acids of human CT␣, the rabbit anti-human CT␤1/␤2 (B2 epitope) antibody, corresponding to amino acids 5-22 of CT␤1/2, and the rabbit anti-human CT␤2 antibody (B3 epitope), corresponding to amino acids 347-365 of CT␤2, were all gifts from Dr. S. Jackowski (22,23).
Activation States of p42/44 MAPK and p38 MAPK -Nuclear extracts were prepared at described (30,52). Equal aliquots of nuclear proteins or cell lysates (20 -50 g) from control and transformed cells were separated by electrophoresis on a 10% SDS-polyacrylamide gel, then transferred to polyvinylidene difluoride membranes. To investigate the activation state of different MAP kinases in response to Ha-Ras transformation, we employed specific antibodies directed against the phosphorylated forms of p42/44 MAPK and p38 MAPK . The immunoblotting procedure was as described above for CT.
Immunoprecipitation and Immunoblot Analysis of Sp3-related Proteins-Nuclear proteins were prepared as described (30,52). Cells grown in 100-mm dishes were scraped into 1 ml of an immunoprecipitation buffer (phosphate-buffered saline (PBS) containing 50 mM Tris, pH 8.0, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1% Triton X-100, 1 mM dithiothreitol). The cell debris was removed by centrifugation at 10,000 ϫ g for 10 min at 4°C. Total proteins (500 g) or nuclear proteins (100 g) from control and Ha-Ras-transformed cells were incubated with 2 g of Sp3-specific polyclonal antibody (H-225, 200 g/ml; sc-13018, Santa Cruz Biotechnology) for 1 h at 4°C. Subsequently, 20 g of protein A-agarose was added, and the samples were mixed gently overnight at 4°C. The immunoprecipitates were collected by centrifugation at 1,000 ϫ g for 5 min at 4°C, and the pellet was washed four times with PBS. Finally, the pellet was resuspended in 40 l of electrophoresis sample buffer and boiled for 2-3 min. Immunoprecipitated proteins were resolved on a 12% SDS-polyacrylamide gel. For immunoblotting, proteins were transferred to a polyvinylidene difluoride membrane that was blocked with PBS containing 6% powdered milk, 0.5% Tween 20, then probed with anti-Sp3 antibody at a dilution of 1:500 for 1 h at room temperature. After three washes with PBS containing 0.5% Tween 20, blots were incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG (dilution 1:2,000) for 1 h, washed with PBS-Tween, and proteins were visualized using enhanced chemilumi-nescence. Phosphorylated Sp3 was detected using an anti-phosphotyrosine antibody (PY-20 at 1:1,000 dilution) by reprobing the same blot using a protocol similar to that described for CT. In some experiments, Sp3-related proteins and tyrosine phosphorylation were measured directly by immunoblotting (31).
Electromobility Shift Assays-Nuclear extracts from Ha-Ras and control fibroblasts were prepared 24 h after serum stimulation as described (30,52). The double stranded CT␣ promoter probes D1 (Ϫ98/ ϩ38) and D2 (Ϫ130/ϩ38) containing Sp binding sites were prepared from LUC.D1 and LUC.D2 reporter vectors by restriction digestion and purification (30). The promoter fragments were 3Ј-end labeled with [ 32 P]dCTP and Klenow polymerase and the binding of Sp1 and Sp3 analyzed after separation of DNA-protein complexes on 5% nondenaturing polyacrylamide gels followed by autoradiography (30,31).
RNA Preparation and Reverse Transcriptase-mediated PCR of CT␣-Total RNA from control and Ha-Ras-transformed cells grown in the presence or absence of MAP kinase inhibitors was extracted using TRIAZOL reagent (Invitrogen) (36). The RNA was reverse transcribed with a first strand cDNA synthesis kit (Superscript II, Invitrogen) according to the manufacturer's protocol. The reverse-transcribed mRNA (0.5-3 g) was amplified using PCR primed with forward (5Ј-ATGCACAGTGTTCAGCCAA-3Ј) and reverse (5Ј-GGGCTTACTA-AAGTCAACTTCAA-3Ј) primers corresponding to the CT␣ gene. The signal from the CT␣ mRNA transcript was normalized to the signal obtained from glycerol-3-phosphate dehydrogenase as a control using the primer pair 5Ј-TCCACCACCTGTTGCTGTA-3Ј (forward) and 5Ј-ACCACAGTCCATGCCATCAC-3Ј (reverse) or with the signal from cyclophilin using the primers 5Ј-TCTTCTTGCTGGTCTTGCCATTCC-3Ј (forward) and 5Ј-TCCAAAGACAGCAGAAAACTT-3Ј (reverse). 30 cycles of PCR amplification at 95°C for 1 min, 45°for 1 min, and 72°C for 2 min produced ϳ200-bp fragments for CT␣ and ϳ250-bp fragments for glyceraldehyde-3-phosphate dehydrogenase; 33 cycles at 94°C for 1 min, 60°C for 2 min, 72°C for 2 min produced ϳ300-bp fragments for cyclophilin.
Statistical Analysis of Data-Calculations of the average values, standard deviations, and the comparison of means by Student's t test were performed by Microsoft Excel (Microsoft Inc.). Densitometric analyses of the CT␣ and/or CT␤ protein mass and mRNA expression were performed by Scion Image acquisition and analysis software (Scion Inc.).

Expression of Ha-Ras Increases Active p44/42 MAPK -It is well
documented that deregulated cell proliferation is a consequence of activated Ras signaling through the MAP kinase cascade, which is a critical component of the proliferative response (39 -41). Activated Ras signals directly through Raf kinase with the subsequent activation of MEK1/2 kinases and results in the phosphorylation and activation of p44/42 MAPK . Typically, the Raf/MEK1/2/p42/44 MAPK pathway is strongly stimulated by growth factors and mitogenic stimuli, whereas in contrast, two other signaling pathways, mediated by p38 MAPK and p46/54 JNK , are activated primarily by cellular stresses that include heat, UV radiation, and hypoxia (42)(43)(44)53) and usually are antiproliferative and apoptotic (54 -56). We, therefore, determined whether p44/42 MAPK was activated to a greater extent in fibroblasts constitutively expressing oncogenic Ha-Ras. The phosphorylation of p44/42 MAPK was analyzed by antibodies that specifically detected phosphorylated, and therefore active, forms of p42/44 MAPK and p38 MAPK . Using the anti-phospho-p42/44 MAPK antibody we observed bands corresponding to p42/p44 MAPK in the nuclear proteins from both control and Ha-Ras-transformed cells (Fig. 1). The nuclear proteins from transformed cells contained significantly more phosphorylated p42/44 MAPK , whereas p38 MAPK was only weakly phosphorylated in Ha-Ras-transformed cells relative to control cells (Fig. 1).
Ha-Ras Transformation Decreases Total CT Enzymatic Activity but Increases CT␣ and CT␤2 Protein Mass in a p42/ 44 MAPK -dependent Manner-It was of interest to know whether or not CT activity was affected by the Ras-signaling pathway. Table I shows total CT enzymatic activities in control and Ha-Ras-transformed cells stimulated by serum in the presence and absence of specific inhibitors of p42/p44 MAPK and p38 MAPK . The flavone compound PD98059 is a specific inhibitor of MEK1/2 kinase and has been used extensively for investigating the physiological function of p42/p44 MAPK (57). The pyridylimidazole compound SB202190 is a p38 MAPK inhibitor (58,59). Neither compound inhibits other known related kinases (57)(58)(59). The results shown in Table I reveal that, similar to published data (45), in Ha-Ras-transformed cells CT activity was 65% (p Ͻ 0.05) less than in the control cells. An equivalent decrease was also observed in the cytosol and microsomes of Ha-Ras-transformed cells compared with control cells (data not shown). Paradoxically, in neither control cells nor Ha-Rastransformed cells was the CT activity modified by kinase inhibitors, suggesting a complex regulation of CT activity by Ras signaling (Table I).
We next performed immunoblotting analyses of the different CT isoforms (Fig. 2A). Densitometric analysis of the immunoreactive bands showed that CT␣ protein levels increased 1.8 Ϯ 0.2-fold in Ha-Ras-expressing cells relative to control cells. With the CT␤2-specific antibody (Fig. 2B) Ha-Ras-transformed cells contained significantly higher amounts (1.5-Ϯ 0.1-fold) of immunoreactive CT␤2 protein than did the control cells (CT␤1 was not detectable by the CT␤1/2 antibody).
Because cells constitutively expressing Ha-Ras contained more active nuclear p42/44 MAPK activity compared with the control cells (Fig. 1), we investigated whether or not the  MEK1/2 inhibitor PD98059 altered the amount of CT␣ and CT␤2 proteins (Fig. 2). Because MEK1/2 phosphorylates and activates p42/p44 MAPK , the inhibitor would be expected to inhibit the phosphorylation of p42/p44 MAPK and their downstream targets. The addition of the 20 -150 M MEK1/2 inhibitor to the cells reduced the amount of CT␣ ( Fig. 2A) and CT␤2 (Fig. 2B) in a dose-dependent manner. The magnitude of inhibition by 150 M PD98059 in control cells was 37 and 50% for CT␣ and CT␤2, respectively, and in Ha-Ras-transformed cells was 61 and 87% for CT␣ and CT␤2, respectively.
CT␣ mRNA Is Increased in Ha-Ras-transformed Cells in a p42/44 MAPK -dependent Manner-Based upon the above results and our previous findings that serum growth factors stimulate the expression of the CT␣ gene at the transcriptional level (36), we next investigated whether CT␣ mRNA was increased by overexpression of Ha-Ras. Fig. 3A shows that CT␣ mRNA is increased in Ha-Ras-transformed cells in the absence of serum (2.5-2.8-fold) as well as after serum stimulation (4.0 -6.5-fold). Fig. 3B demonstrates that 150 M p42/44 MAPK inhibitor (PD98059) reduces CT␣ mRNA expression by 70% in both control and Ha-Ras cells. The p38 MAPK inhibitor SB202190 did not decrease CT␣ mRNA expression in control and Ha-Ras cells. Together, these results suggest that the increase in CT␣ protein mass shown in Fig. 2 is caused by an increased transcription of the CT␣ gene in Ha-Ras-transformed cells. This increase is p42/44 MAPK -dependent and is increased further by serum.
CT␣ Promoter Activity Is Increased in Cells Overexpressing Ha-Ras-To examine further the transcriptional regulation of the CT␣ gene, we investigated CT␣ promoter activity in transiently transfected cells using a series of truncated promoterluciferase reporters (Fig. 4A). We have shown previously that C3H10T1/2 fibroblasts support the activation of CT␣ promoterluciferase constructs (30,31). Fig. 4A shows that control and Ha-Ras-transformed cells both have the ability to direct luciferase gene expression, as expected (31). Interestingly, the promoter activity of each of the promoter-reporter constructs tested, including the basal promoter construct Ϫ52/ϩ38 (LUC.D3), is persistently 1.5-3-fold higher in the Ha-Ras cells than in control cells, suggesting that the basal promoter region is sufficient to drive the increased luciferase expression in the cells constitutively expressing Ha-Ras.
Sp3 Binding to the CT␣ Promoter Is Increased in Ha-Rastransformed Cells-The basal promoter region of the CT␣ gene contains complex, overlapping binding sites for nuclear factor-B, Elk1, E2F, and Sp1. Previously, we investigated the role of the Sp site located at position Ϫ39/Ϫ9 (30,31). The role of Sp1, Sp2, and Sp3 in binding to this site was established by mutation and overexpression analysis in insect and mammalian cells, including C3H10T1/2 murine fibroblasts (30,31). We have demonstrated previously that Sp1 and Sp3 specifically and competitively bind to this promoter region as well as to two other regions of the CT␣ promoter at Ϫ88/Ϫ50 and Ϫ148/Ϫ128. We therefore performed electromobility shift assays with nuclear proteins from control and Ha-Ras-transformed cells to determine whether the increase in CT␣ promoter activity in the Ha-Ras cells was caused by changes in protein binding to the promoter. Fig. 4B shows that the protein profile from control and Ha-Ras cells is the same, suggesting that stimulation by Ha-Ras transformation of CT␣ promoter activity is not the result of the binding of a new transcription factor. We demonstrated previously by supershift analysis in C3H10T1/2 cells using antibodies specific for the Sp1 and Sp3 proteins that these proteins bind the promoter probe Ϫ90/ϩ38 (30,31). From Fig. 4B, it can be seen that there is a stronger binding of the Sp3-related nuclear proteins in the transformed, compared with the control, cells. A similar binding profile was observed when a different promoter probe, Ϫ130/ϩ38, which contained Sp binding sites, was used (data not shown). From these results and from our previously published data (30,31), we conclude that the increase in CT␣ promoter activity in the Ha-Ras-transformed cells shown in Fig. 4A is, at least in part, a consequence of an increased binding of the transcription factor Sp3.

Ha-Ras Transformation Increases Sp3 Protein Expression and Modifies Sp3
Proteins Post-translationally-We next investigated whether the increase in Sp3 binding to the promoter (Fig. 4B) was caused by an increase in the levels of Sp3 protein and/or a consequence of differences in post-translational modifications of Sp3. The existence of multiple Sp3 proteins has been shown to be caused by the presence of an internal translation initiation start site, which results in a truncated Sp3 that has opposing activity to the full-length Sp3 protein (50,60). We therefore immunoprecipitated proteins from cell lysates and nuclear extracts from transformed and control cells using a Sp3-specific antibody. Fig. 5 (lanes 3 and 4) clearly demonstrates that in nuclear extracts from Ha-Ras-transformed cells the amounts of both the full-length Sp3 (a 1 , 116 kDa, a 2 , 97 kDa) and truncated Sp3 (b 1 , 70 kDa and b  kDa) are increased compared with control cells. The doublets typically observed for each isoform, a 1 /a 2 and b 1 /b 2 , likely result from differences in mobility of post-translationally modified Sp3 proteins, primarily from differences in phosphorylation. In the nucleus of Ha-Ras-transformed cells 2.1-fold more of the slowest migrating form of Sp3, a 1 , is present than in control nuclei. Moreover, the a 2 form is 1.4 -1.9-fold more abundant in Ha-Ras-transformed cells than in control cells. However, the a 3 form of Sp3 (ϳ90 kDa), which represents yet another modified full-length Sp3 protein, was more dominant in control cells than in Ha-Ras-transformed cells (2.2.-fold more in lysates and 2.1-fold more in the nuclear extract). No differences in the amounts of the truncated Sp3 b 1 and b 2 forms between control and Ha-Ras-transformed cells were observed in the whole cell lysates. However, in nuclear extracts the amount of the b 1 form was increased by 50%, and the b 2 form was increased by 30% in Ha-Ras cells relative to control cells.
Next, we inspected the murine Sp3 (GenBank XP_130306) for putative phosphorylation sites by using the NetPhos 2.0 algorithm (61). The computer analysis did not reveal any con-

FIG. 3. The amount of CT␣ mRNA is altered by serum, Ha-Ras transformation, and inhibition of p42/44 MAPK . A, effect of serum and
Ha-Ras transformation on CT␣ mRNA abundance in fibroblasts. Control and transformed cells were incubated without serum for 48 h. At this time, either total RNA was isolated, or the serum-starved cells were incubated for an additional 24 h with 10% serum. mRNA abundance was estimated by reverse transcriptase-based PCR. Total mRNA was reverse transcribed using primers specific for CT␣ and glyceraldehyde-3phosphate dehydrogenase (G3PDH) (Control), and PCR was performed within the linear range of concentrations of mRNA and number of PCR cycles. CT␣ mRNA expression was normalized to glyceraldehyde-3-phosphate dehydrogenase mRNA, and the abundance of CT␣ mRNA was calculated relative to that in control cells grown without serum. B, effect of the MEK1 inhibitor PD98059 and the p38 MAPK inhibitor SB202190 on CT␣ mRNA abundance in control and Ha-Ras-transformed fibroblasts. The experiment was performed as in A except that cyclophilin served as control for the reverse transcriptase-based PCR. mRNA abundance is given relative to that of control cells grown without inhibitors.

FIG. 4. Promoter activation of the CT␣ gene by Ha-Ras transformation and binding of transcription factor Sp3.
A, CT␣ promoterluciferase reporter activity in control and Ha-Ras-transformed fibroblasts. A series of truncation mutants of the CT␣ promoter linked to a luciferase reporter, spanning Ϫ201 bp to ϩ38 bp of the promoter region (2.5 g), and the SV-40 promoter-luciferase control vector (SV-40; 2.5 g) were transiently cotransfected with the pSV␤-galactosidase expression vector (2.5 g). Luciferase activity was normalized to ␤-galactosidase activity and is relative to that of the SV-40 control. Cells were serum deprived for 3 days prior to the transfection, then stimulated with serum for 24 h, and luciferase assays were performed. Results are the means Ϯ S.D. of three independent experiments, each performed in triplicate. B, Ha-Ras transformation increases the binding of Sp3, but not Sp1, to the CT␣ promoter. Electrophoretic mobility shift assays were performed using control and Ha-Ras-transformed nuclear fractions and were incubated with 32 P-labeled oligonucleotide consisting of the region Ϫ90/ϩ38 bp of the CT␣ proximal promoter and basal Sp-cis-acting elements. The complexes were resolved by electrophoresis on a nondenaturing 5% polyacrylamide gel. The bands corresponding to the complexes for Sp3 (I and II) and Sp1 are indicated. This experiment was repeated twice with similar results. served tyrosine phosphorylation sites in the Sp3 protein. There are, however, two nonconsensus tyrosine kinase sites at positions 87 (LQGNYIQSP) and 347 (CGKVYGKTS) which prompted our further analysis, shown in Fig. 5 (lanes 1 and 2). The immunoprecipitated Sp3 proteins were reprobed with an anti-phosphotyrosine antibody and compared with the Sp3 immunoblots. Differences in phosphorylation of Sp3 proteins between control and Ha-Ras-transformed cells are evident in both whole cell lysates and nuclear extracts. In addition, an unknown protein of molecular mass Ͼ150 kDa was immunoprecipitated which was more highly phosphorylated in the Ha-Ras cells. Furthermore, Fig. 5 (lanes 1 and 2) shows less phosphorylation of Sp3 a 1 and more phosphorylation of Sp3 a 3 in nuclear extracts from the transformed cells than from the control cells. Phosphorylation of Sp3 a 2 and Sp3 b 1 /b 2 was not observed. The results in Fig. 5 indicate that the full-length Sp3 a 3 and an unknown protein that coimmunoprecipitates with Sp3, are more phosphorylated in the Ha-Ras-transformed cells, but other Sp3 forms are less phosphorylated compared with control cells. Furthermore, the amounts of full-length Sp3 (a 1 and a 2 ) and truncated Sp3 (b 1 and b 2 ) are augmented by Ha-Ras transformation resulting in stronger binding of these Sp3 species to the CT␣ promoter (Fig. 4B).
Based upon our previous results (30,31), which demonstrated that Sp1 also plays an important role in the regulation of expression of Ctpct, we next determined whether the amount of Sp1 was increased by increased expression of Ha-Ras. We found, however, that the amount of Sp1 protein was unchanged (data not shown), consistent with the data in Fig. 4B that show no difference in binding of Sp1 to the promoter probe between control and Ha-Ras-transformed cells.
Transient Expression of Sp3 Increases the Amount of Both CT␣ and CT␤2 Protein-Mutation and transfection analyses (31) revealed that Sp3 is functionally equivalent to Sp1 and can act as an independent transcriptional activator of the CT␣ promoter. We have also demonstrated that overexpression of Sp3 in mammalian cells (including C3H10T1/2 murine embryo fibroblasts) stimulates CT␣ promoter-luciferase reporter activ-ity (31). Thus, to determine whether Sp3 increases the amount of CT proteins we analyzed the expression of CT proteins after transfection of control and Ha-Ras fibroblasts with various amounts (0 -12 g) of the Sp3-expression plasmid, pPacSp3. The levels of both CT␣ (1.2-fold and 1.4-fold) and CT␤2 (1.15fold and 1.5-fold) proteins increased in experiments with 3 and 6 g of the pPacSp3 plasmid, respectively, for control cells only (Fig. 6). No increase in CT␣ and CT␤2 proteins was apparent in Ha-Ras cells after transfection with 3 and 6 g of pPacSp3 because CT expression was already increased. At higher plasmid concentration, 12 g of pPacSp3, the CT␣ protein decreased to the "basal" levels (transfections with empty plasmid pPacO), and in case of CT␤2 the detectable protein declined below the basal level (0.7-fold and 0.5-fold in the Ha Ras and control cells, respectively). These results show for the first time that expression of Sp3 increases the level of CT␣ and CT␤2 proteins (Fig. 6). DISCUSSION We have found that in C3H10T1/2 fibroblasts Ha-Ras activates the expression of Sp3 via the p42/44 MAP kinase pathway, thereby increasing the amounts of CT␣ and CT␤2 mRNA and protein. Our data indicate that when Sp3 binds to one or more sites on the CT␣ promoter, transcription of the CT␣ gene is stimulated. Because the CT␤ promoter has not yet been characterized, it is not possible to determine whether Sp3 directly or indirectly governs the increased amount of CT␤2 in the Ha-Ras-transformed cells.
Mechanistic Studies on Activation of the CT␣ Gene-The focus of the current study was to elucidate the mechanism by which the expression of CT␣ mRNA and protein is increased in Ha-Ras-transformed fibroblasts. Previously, we established that growth stimulation by serum increased the expression of CT␣ mRNA during the cell cycle (36). The results presented here demonstrate that the p42/44 MAPK signaling pathway is responsible for the serum-induced increase in CT␣ mRNA and protein that is further magnified by constitutive activation of p42/44 MAPK in transformed cells. In accordance with those results, when signaling via p42/44 MAPK was inhibited, or when serum was eliminated from the growth medium, the expression of CT␣ mRNA and protein was decreased. To identify the regulatory cis-acting elements responsible for the Ras/p42/ 44 MAPK regulation of the CT␣ gene, we used luciferase-reporter mutants spanning the promoter region from Ϫ201 bp to ϩ38 bp. We demonstrated that the promoter activity for all constructs was persistently higher in transformed cells.
Previously we established that the CT␣ basal promoter is completely inactive if transcription is not supported by Sp1 or Sp3 nuclear factors, as in insect cells naturally lacking those factors (31). Transient expression of Sp1 or Sp3 initiates CT␣ basal transcription in insect cells, suggesting that the interactions of Sp1 and Sp3 with the basal promoter and general transcription factors are critical for the CT␣ transcriptional activation (31). The basal CT␣ promoter Sp element is located in the vicinity of the transcription initiation site, at position Ϫ22/Ϫ15 bp, and could bind Sp1 and Sp3 together with other nuclear proteins, possibly including the basal transcription factors (30,31). We have now established that Sp3 from transformed cells binds stronger to the CT␣ promoter than Sp3 from untransformed cells and that Sp1 binds equally in both cells (Fig. 4B), suggesting that Sp3, not Sp1, could be solely responsible for the up-regulation of CT␣ gene in transformed cells.
The stronger binding of Sp3 to the CT␣ promoter in transformed cells is probably a combination of increased mass and modified phosphorylation of several Sp3 protein species. Sp3 exists in two main isoforms (50) of which we showed that the full-length Sp3 a (stimulatory form) is primarily overexpressed and post-translationally modified by Ha-Ras/p42/44 MAPK . The  1 and 2) was detected by anti-phosphotyrosine antibodies, and Sp3 proteins (lanes 3 and 4) were identified by using anti-Sp3 antibodies. The full-length Sp3 proteins, a1, a2, and a3, and truncated Sp3 proteins, b1 and b2, are described under "Results." An unknown protein that coprecipitates with Sp3 proteins is also indicated (*). The ϳ50 kDa bands in lanes 3 and 4 represent IgG. Molecular size markers are indicated on the left side. truncated form Sp3 b (repressive form) was increased modestly by the Ha-Ras transformation and was not affected by tyrosine phosphorylation (Fig. 5). At lower Sp3 cDNA concentrations the protein expression of both CT␣ and CT␤2 increased in control cells, whereas in transformed cells, overexpression of Sp3 did not alter the CT␣ and CT␤2 levels because they were already elevated. However, at higher Sp3 cDNA concentrations, the levels of CT␣ and CT␤2 were decreased below or returned to the basal level even in transformed cells, suggesting that Sp3 also acted as a transcriptional repressor. A dual property of Sp3 has been manifested in previous studies for the regulation of other genes (62,63) including our own work with the Sp3 regulation of CT␣ promoter in insect cells (31). That Sp3 is a multiple regulator is also supported by the findings that at higher concentrations, Sp3 is normally inhibitory by a DNA binding-independent mechanism that involves competition for components of the basal transcriptional machinery (62) or titration of the promoter-specific transcription factors (64).
Given the overwhelming evidence for the role of Sp1 in growth regulation and tumor development (62,63), it is surprising that Sp3 is solely implicated in the regulation of CT␣ during oncogenic transformation and that neither the Sp1 binding nor abundance appears to be important for CT␣ gene expression. Sp1 is a well known target for Ras/p42/44 MAPK regulation by phosphorylation, which typically does not involve changes in the Sp1 protein mass (65)(66)(67)(68). Even as a target for p42/44 MAPK , we report that Sp1 is not a significant regulator of CT␣ gene expression after oncogenic transformation and that Sp3 is predominantly regulating CT␣ by means of the Ras/p42/ 44 MAPK signaling pathway. To our knowledge this is the first time that Sp3, independently from Sp1, appears to regulate gene expression by means of the Ras/p42/44 MAPK signaling pathway. Interestingly, ERK2 (p42 MAPK ) but not ERK1 (p44 MAPK ) could also be preferentially stimulated by Sp3, not Sp1 (69), which taken together with our findings suggest the existence of a positive feedback mechanism for the observed selective up-regulation of Sp3 protein in transformed cells. The most recent evidence from gene disruption of Sp3 (70) suggests that Sp3 has a distinct regulatory function from Sp1 and other members of the broader Sp1/Krü ppel-like family of transcription factors (62). Sp3 is essential for postnatal survival, and Sp3-deficient embryos are growth-retarded and die at birth of respiratory failure (70). The observed breathing defect remains obscure, and surfactant protein expression was not found to be different from that in the Sp3 wild type embryos (70). However, given that CT␣ plays a central regulatory role in the production of the lung phospholipid surfactant, PC, it may be that the deletion of the Sp3 gene would diminish CT␣ expression and, thus, PC production causing lung failure. It is highly likely that Sp3 deletion will also abolish the expression of CT␤; however, the precise mechanisms for how Sp3 regulates the expression of CT␤ will remain unknown until the promoter of the murine CT␤ gene is isolated and fully characterized.
Ha-Ras Activation of CT␣ Expression-Expression of constitutively active Ras induces cellular proliferation and usually is transforming. The mechanism by which Ras mediates these effects is believed to be through a direct interaction with downstream effectors resulting in a persistent activation of MAP kinase signaling pathways (39 -41). It is well established that one characteristic of transformed cells is perturbation of lipid synthetic and degradative pathways (71)(72)(73), suggesting an active role of lipids in oncogenic transformation. Expression of oncogenic Ras correlates with increased levels of phosphocholine and phosphoethanolamine (45,(73)(74)(75), diacylglycerols (76), inositol phosphates (77), and arachidonic acid (78). The stimulated degradation of phospholipids is accompanied by an increased biosynthesis (45,(73)(74)(75), demonstrating an accelerated turnover of lipids in transformed cells. Increased amounts of phosphocholine and phosphoethanolamine have been implicated in tumor development (78), during which choline/ethanolamine kinases are typically up-regulated, whereas enzymes involved in the subsequent metabolism of phosphocholine and phosphoethanolamine (i.e. CTP:phosphocholine-and phosphoethanolamine cytidylyltransferases) are either up-regulated (47) or down-regulated (45,46) depending on the cell type.
Our data show that in Ha-Ras-transformed fibroblasts, the levels of CT␣ and CT␤2 proteins and mRNAs are increased relative to control cells. However, we were surprised that the increased amounts of CT␣ and CT␤2 proteins in the Ha-Rastransformed cells did not correlate with an increase in CT activity. On the contrary, CT activity was significantly lower in transformed cells than in control cells even though more CT protein was present. Other studies have also reported that CT activity is lower in Ha-Ras-transformed cells than in control cells (45). Moreover, CT activity was not affected by the specific inhibitor of p42/44 MAPK (PD98059) in transformed cells (that had constitutively active p42/44 MAPK ) or in control cells in which p42/44 MAPK was activated by serum. The inhibitor PD98059, however, decreased the protein mass of both CT␣ and CT␤2 as would be expected if their expression were regulated by p42/44 MAPK . Thus, Ha-Ras/p42/44 MAPK signaling simultaneously increases the amount of CT protein yet decreases CT activity.
CT phosphorylation has been studied extensively. It has been clearly established that soluble, inactive forms of CT are more highly phosphorylated than are membrane-bound, active forms (79,80). CT activity also varies inversely with CT phosphorylation during the cell-cycle (81), and the translocation of CT between membranes and cytosol is influenced by phosphorylation/dephosphorylation (4,79,80). CT dephosphorylation is secondary to its association with membranes, suggesting that CT interactions with lipids could be a more significant factor for enzymatic activity than phosphorylation (82,83). Insulin and epidermal growth factor stimulate the phosphorylation of FIG. 6. The amount of CT␣ protein is increased by transfection of cDNA encoding Sp3. Sp3 proteins were expressed in control and Ha-Ras-transformed cells by transient transfection of 3, 6, and 12 g of pPacSp3 or 12 g of pPacO control vector. The total amount of DNA was kept constant at 12 g by the addition of the appropriate amounts of empty control vector. After 48 h, cells were harvested and 50 g of proteins from cell lysates was loaded onto 7% SDS-polyacrylamide gels. The expression of CT␣ and CT␤2 was assessed by immunoblotting. Two independent experiments were performed with similar results.
CT in HeLa cells in vivo, and p42/44 MAPK is capable of phosphorylating purified CT␣ (65). Thus, phosphorylation of CT might be enhanced in the Ha-Ras-transformed cells because of constitutive activation of p42/44 MAPK .
CT␤ has been discovered only recently (22,23), and its posttranslational regulation by phosphorylation has not been reported. However, CT␤2 phosphorylation domains are highly homologous to the phosphorylation domains of CT␣ and possibly could be targeted by similar mechanisms. Which of the numerous Ser/Thr phosphorylation sites of CT␣ and CT␤ are selective targets for Ras/p42/44 MAPK signaling is presently unknown.
Alteration in lipid composition can also post-translationally regulate CT␣ (and possibly CT␤) activity and might contribute to the modification of CT activity in Ha-Ras-transformed cells. Ample evidence suggests that the cellular lipid content is changed after oncogenic transformation (45-47, 71-73, 76). For example, recent data show that increased cell proliferation and activation of the p42/44 MAPK signaling cascade alter the expression and/or activity of several genes of phospholipid metabolism, including phospholipase C (76), phospholipase A 2 (71), the level and/or phosphorylation of lipid transport proteins, apoA-I (85), apoC-III (86), low density lipoprotein receptor (87), and the activity of lipid-related transcription factors sterol response element-binding protein (88) and peroxisomal proliferator activated receptor-␥ (84) which regulate many genes involved in lipid metabolism, perhaps including the CT␣ gene (34).
Whether the decreased CT activity in Ha-Ras-transformed cells is linked to direct post-translational modifications of both CT isoforms, differences in the type, content and/or location of regulatory lipids, or other mechanisms, would be difficult to resolve.