Transcriptional regulation of endothelial nitric-oxide synthase by an interaction between casein kinase 2 and protein phosphatase 2A.

We previously demonstrated that lysophosphatidylcholine up-regulated endothelial nitric-oxide synthase promoter activity by increasing Sp1 binding via the action of protein serine/threonine phosphatase 2A (Cieslik, K., Zembowicz, A., Tang, J.-L., and Wu, K.K. (1998) J. Biol. Chem. 273, 14885-14890). To characterize the regulation of basal endothelial nitric-oxide synthase promoter activity and the signaling pathway through which lysophosphatidylcholine augments endothelial nitric-oxide synthase transcription, we used a casein kinase 2 inhibitor coupled with immunoprecipitation to demonstrate that basal Sp1 binding and endothelial nitric-oxide synthase promoter activity were controlled by casein kinase 2 complexed with protein serine/threonine phosphatase 2A. Casein kinase 2 catalyzed protein serine/threonine phosphatase 2A phosphorylation thereby inhibiting its activity. Lysophosphatidylcholine selectively activated p42/p44 mitogen-activated protein kinase. Purified extracellular regulated kinase 2 blocked casein kinase 2 activity and increased protein serine/threonine phosphatase 2A activity, resulting in an increased Sp1 binding and endothelial nitric-oxide synthase promoter activity. These results indicate that Sp1 binding to its cognate site on the endothelial nitric-oxide synthase promoter and its transactivation of endothelial nitric-oxide synthase is regulated by post-translational Sp1 phosphorylation and dephosphorylation through a dynamic interaction between casein kinase 2 and protein serine/threonine phosphatase 2A.

Endothelial nitric-oxide synthase (eNOS) 1 is a member of the NOS family, which catalyzes the oxidation of L-arginine to generate nitric oxide (NO) and L-citrulline (1)(2)(3)(4). Nitric oxide induces smooth muscle cell relaxation, inhibits platelet and leukocyte activation, and plays a key role in maintaining vascular integrity and tone (5). Altered endothelial NO production has been implicated in several important cardiovascular diseases: hypertension, coronary heart disease, diabetes, and ischemic stroke. eNOS is constitutively expressed in endothelial cells and has features of a housekeeping gene (4). However, it is inducible by various vasoactive agents including fluid shear stress (6), physical exercise (7), hypoxia (8), estrogen (9), low levels of oxidized low density lipoprotein (10), and lysophosphatidylcholine (lysoPC or LPC) (11). eNOS induction is considered to be important in fortifying the vasoprotective role of NO (12).
The mechanism by which eNOS is induced by diverse physical and chemical factors remains unclear. We have used ly-soPC as a model system to elucidate the mechanism by which eNOS gene transcription is regulated. Basal eNOS transcription requires binding of Sp1 to its cognate site (Ϫ104 to Ϫ90) at the 5Ј-flanking region of human eNOS gene (13,14). Recent work from our laboratory indicated that lysoPC increases Sp1 binding activity, thereby augmenting eNOS promoter activity (15). Our work demonstrated that lysoPC-induced increase in Sp1 binding activity was blocked by okadaic acid and correlated with an elevated PP2A activity (15), suggesting that eNOS transcription is regulated by post-translational modification of Sp1 via the actions of PP2A and a nuclear kinase, which was recently identified as casein kinase 2 (CK2) (16). The purpose of this study was to determine how Sp1 binding and eNOS promoter activities are controlled by CK2 and PP2A and to elucidate the signal pathway through which lysoPC alters this control mechanism. Our results indicate that CK2 complexes with PP2A and suppresses PP2A activity at the basal state, resulting in a low level of Sp1 binding and eNOS promoter activity. LysoPC activates selectively the extracellular signalregulated kinase (ERK-1 and ERK-2), which suppresses the CK2 activity and unleashes its inhibition of PP2A, leading to a higher Sp1 binding and eNOS promoter activity.

Construction of eNOS 5Ј-Promoter in Luciferase Expression
Vectors-A 5Ј-flanking fragment at nucleotide position from Ϫ1322 to ϩ22 was obtained by polymerase chain reaction, using genomic DNA as a template and synthetic oligomers as primers: EN1322G (5Ј-AAA-GATCTTCCATCTCCCTCCTCCTG-3Ј) and EN3H (5Ј-GGGAAGCTT-GTTACTGTGCGTCCACTCTG-3Ј). The polymerase chain reaction product purified from agarose gel was digested with BglII/HindIII and cloned into the promoterless luciferase reporter vector pGL3.
Cell Culture and Transient Expression-ECV-304 (spontaneously transformed human umbilical vein endothelial cell line) was cultured in complete Medium 199 (Life Technologies, Inc.) containing 10% fetal bovine serum. Transient transfection by Lipofectin (Life Technologies, Inc.) was performed as described (13). ECV-304 was incubated in serum-free medium containing a mixture of 10 l of Lipofectin and 2 g of promoter construct at 37°C for 5 h. Medium was removed; cells were washed and incubated with fresh complete medium for 24 h. Cells were then washed and incubated in Medium 199 containing 0.5% fetal bovine serum for 16 h. The medium was replaced with fresh Medium 199 containing 5% fetal bovine serum in the presence or absence of PD 98059 (Calbiochem) or 5,6-dichloro-1-␤-D-ribofuranosylbenzimidazole (DRB) (Calbiochem). After 1 h of incubation, 100 M lysoPC (Avanti Polar Lipids) was added to the medium and cells were incubated for an additional 6 h at 37°C. The cells were harvested and luciferase activity was determined by luciferase assay in a luminometer (Analytical Luminescence Laboratories, Monolight model 2010) as described (17).
Protein Phosphatase 2A Assay-PP2A assay kits were obtained from Promega. The assay is based on determining the amount of free phosphate generated in the reaction by measuring the absorbance of a molybdate malachite green-phosphate complex. 10 g of NE was incubated on a 96-well plate together with a peptide substrate RRA(pT)VA and buffer (50 mM imidazole, pH 7.2, 0.2 mM EGTA, 0.02% ␤-mercaptoethanol, 0.1 mg/ml BSA) for 30 min at 30°C. After incubation, the molybdate complex dye was added and incubated for an additional 30 min at room temperature for color development. The level of molybdate malachite green-phosphate complex formed was monitored at 630 nm.
Immunoprecipitation-0.5 mg of NE with 4 l of anti-human CK2␣subunit rabbit polyclonal IgG1 (Upstate Biotechnology, Inc.), or 16 l of anti-PP2A catalytic subunit rabbit polyclonal IgG (Promega) was incubated overnight at 4°C in 1 ml of TBS buffer (10 mM Tris-HCl, pH 8.0, 150 mM NaCl) containing 5 mM EDTA, 1 g/ml leupeptin, 0.5 mM PMSF. 20 l of Protein G Plus-Agarose (Santa Cruz Biotechnology) was added to each sample and incubated for an additional 2 h. Immunoprecipitates (IP) were spun down at 2500 rpm for 5 min, washed four times with TBS, and resuspended in 100 l of TBS or CK2 assay buffer.
Western Blot Analysis-100 g of NE were electrophoresed on SDS-PAGE gel and transferred to a nitrocellulose membrane in a transfer buffer (25 mM Tris base, 0.2 M glycine, 20% methanol). Membranes were blocked overnight with 3% milk solution in TBS and incubated with anti-PP2Ac at 1:500 or anti-CK2␣ at 1:100. The secondary anti-rabbit horseradish peroxidase-conjugated antibody was added and detected by enhanced chemiluminescence (ECL) system (Pierce).
Casein Kinase 2 Assay-IP were incubated in 100 l of CK2 assay buffer containing 100 M [␥-32 P]ATP (activity 100 cpm/pmol), 20 mM MgCl 2 , 40 mM Tris-HCl, pH 7.9, 150 mM NaCl, 0.1 mg/ml BSA, and 10 l of 5% casein solution. Reaction was performed at 37°C for 20 min and stopped by adding an equal volume of 2ϫ Laemmli buffer. Samples were boiled for 5 min and subjected to SDS-PAGE and autoradiography.
Assays for Mitogen-activated Protein Kinases (MAPK)-ECV-304 or passage 1 near-confluent human umbilical vein endothelial cells were incubated with lysoPC or PD 98059 in Medium 199 containing 5% fetal bovine serum. At indicated time points, cells were lysed in 500 l of lysis buffer containing a protease inhibitor mixture (Roche Molecular Biochemicals), 10 mM sodium pyrophosphate, 2 mM sodium orthovanadate, and 100 mM sodium fluoride. Cell residues were spun down and removed. Supernatants were collected and subjected to SDS-PAGE. The proteins were transferred to nitrocellulose membrane and blotted with an antibody specific for phosphorylated p44/p42 MAPK (ERK1/2), P38 MAPK, or JNK and also with an antibody specific for unphosphorylated p44/p42, P38 or JNK MAPK (all from New England Biolabs). The protein band recognized by the specific antibody was detected by ECL system. To assay ERK1/2 activity, 200 l of cell lysates containing approximately 200 g of proteins were immunoprecipitated with an antibody specific for phosphorylated p44/p42 at 4°C overnight. Sepharose beads (Amersham Pharmacia Biotech) were added, and the precipitates were washed twice with 500 l of lysis buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton, 2.5 mM sodium pyrophosphate, 1 mM ␤-glycerol phosphate, 1 mM sodium orthovanadate, 1 g/ml leupeptin) and twice with 500 l of kinase buffer (25 mM Tris, pH 7.5, 5 mM ␤-glycerol phosphate, 2 mM DTT, 0.1 mM sodium orthovanadate, 10 mM MgCl 2 ). The pellet was suspended in 50 l of kinase buffer containing 200 M ATP and 2 g of Elk1 fusion protein (New England Biolabs) and incubated for 30 min at 30°C. The reaction was terminated with 25 l of SDS buffer. The samples were subjected to SDS-PAGE, transferred to nitrocellulose membrane, and blotted with an antibody specific for phosphorylated Elk-1 (New England Biolabs). Phosphorylated Elk-1 band was detected by the ECL system.

Control of Sp1
Binding and eNOS Promoter Activity by CK2-In our previous report (15), we showed by using luciferase reporter constructs that Sp1 sites at Ϫ104 and Ϫ90 of eNOS promoter region were essential for eNOS promoter activity. Mutation of these sites by site-directed mutagenesis abrogated the promoter function. When nuclear extracts prepared from unstimulated or lysoPC stimulated human umbilical vein endothelial cells were incubated with labeled native promoter fragments or consensus Sp1 sequences, two Sp1-DNA complexes were detected that were competed out by a 50-fold molar excess of unlabeled probes. We estimated by Sp1 oligonucleotide competition that lysoPC increased Sp1 binding activity by about 5-fold over the basal activity (15). In the present study, nuclear extracts from ECV-304 also formed two complexes with consensus Sp1 sequence (Fig. 1A), which were competed out by a 100-fold molar excess of unlabeled probes (data not shown). The density of these two bands was enhanced by lysoPC treatment (Fig. 1A, lane 2 versus lane 3). Transfection of a promoter fragment (Ϫ1322 to ϩ22) luciferase reporter construct in ECV-304 expressed luciferase activity which was augmented by lysoPC (Fig. 1B). To determine whether CK2 is involved in regulating the Sp1 binding activity, we pretreated ECV-304 cells with DRB, a selective CK2 inhibitor. DRB at 6 M markedly increased the basal Sp1 binding activity (Fig. 1A, lane 4) and caused a lesser increase in lysoPC-induced Sp1 binding. DRB treatment also increased the basal promoter activity (Fig. 1B). These results suggest that the basal Sp1 binding to eNOS promoter region and consequently basal eNOS promoter activity was controlled by CK2. They are consistent with a previous report, which showed that CK2 was involved in Sp1 phosphorylation and lowered the Sp1 binding activity in liver differentiation (16).
Inhibition of PP2A Activity by CK2-We then determined whether PP2A activity was influenced by DRB treatment. Pretreatment of cells with DRB increased the basal PP2A activity (Fig. 2). LysoPC-induced PP2A activity was also elevated by DRB. These results are in agreement with those of Sp1 binding and eNOS promoter activity. CK2 and PP2A protein levels in NE were not affected by lysoPC treatment; nor were they influenced by DRB treatment (Fig. 3).
Complex of CK2 with PP2A in Nuclear Extracts-It has been shown in several signaling systems that kinase forms a complex with its phosphatase partner in achieving a dynamic and specific on and off signaling (18 -21). We suspected that CK2 might complex with PP2A in regulating Sp1-activated eNOS transcription. NE obtained from basal or lysoPC-treated cells were immunoprecipitated with anti-PP2Ac and the CK2 activity in IP was determined. The results from one of five experi-ments are shown in Fig. 4A.  4. Association of PP2A with CK2. A, Autoradiographs of a representative experiment. NE (C denotes basal and L denotes lysoPCtreated) or purified CK2 (0.4 g, 10 units/g) and PP2A (0.4 g, 1 unit/g) were immunoprecipitated with anti-PP2Ac, anti-CK2␣, or preimmune IgG (pre-IgG, only the basal results shown), and after washing, the CK2 activity of the immunocomplex was measured using casein as a phosphorylation substrate. Lane 1 shows addition of purified CK2 in the incubation buffer without subjecting to immunoprecipitation to serve as an internal positive control. B, densitometric analysis of casein phosphorylation in five experiments. Numbers shown on the horizontal axis correspond to those shown in A. Purified CK2 activity (lane 1) was arbitrarily set as 100%. The data denote mean Ϯ standard deviation. Differences between lanes 4 and 5 (p Ͻ 0.01) and between lanes 8 and 9 (p Ͻ 0.01) are statistically significant. Interaction of Purified CK2 with PP2A-To confirm CK2 and PP2A complex formation, purified CK2 holoenzyme was incubated with PP2A holoenzyme or catalytic subunit PP2Ac and immunoprecipitated with anti-PP2Ac. CK2 activity was detected in IP from CK2 ϩ PP2Ac (Fig. 5A, lane 3) as well as CK2 ϩ PP2A (Fig. 5A, lane 5). Preimmune IgG did not bring down a complex with CK2 activity (Fig. 5A, lanes 2 and 4) while purified CK2 treated with anti-PP2Ac did not yield any precipitate with CK2 activity (Fig. 5A, lane 7). Thus, CK2 activity detected in PP2A immunoprecipitate was specific for CK2-PP2A complex. The stoichiometry of CK2 and PP2A activities in the complex appeared to be 1 to 1 (Fig. 5, B-D). As inhibition of CK2 by DRB increased PP2A activity (Fig. 2), we wondered whether CK2 catalyzes the phosphorylation of PP2A. Purified PP2Ac was incubated in CK2 assay buffer with purified CK2 and [␥-32 P]ATP. PP2A phosphorylation was analyzed by SDS-PAGE and autoradiography. Fig. 5E shows phosphorylation of the 36-kDa catalytic subunit of PP2A.
Activation of ERK1/2 by Lyso-PC-To discern the signaling pathway through which lysoPC alters CK2-PP2A activities thereby enhancing Sp1 binding and eNOS promoter activity, we evaluated the effect of lysoPC on the MAPK pathways. LysoPC selectively induced ERK1 and ERK2 phosphorylation in a time-and concentration-dependent manner without alter-ing their protein levels (Fig. 6, A and B). In accordance with ERK1/2 phosphorylation, ERK activity as evaluated by using Elk-1 as a substrate was increased by lysoPC in a concentration-related manner (Fig. 6C). LysoPC did not induce phosphorylation of JNK/SAPK or p38 MAPK (data not shown). Both ERK1/2 phosphorylation and ERK catalytic activity induced by lysoPC were blocked by PD 98059 (Fig. 6, B and C), a selective inhibitor of MEK-1, which is the upstream kinase that catalyzes ERK phosphorylation (18).
Effect of an MEK-1 Inhibitor on Sp1 Binding, eNOS Promoter Activity, and PP2A Activity-The results shown above are consistent with a selective activation of MEK-1 and consequently ERK1/2 by lysoPC. To ascertain that this pathway is involved in regulating Sp1 binding, we pretreated cells with PD 98059, stimulated them with and without lysoPC, and determined Sp1 binding activity in the nuclear extracts of these treated cells. PD 98059 at 10 M reduced the lysoPC-induced Sp1 binding activity to approximately the basal level without affecting the basal binding activity, while at 50 M, it suppressed both basal and lysoPC-induced Sp1 binding activities (Fig. 7A). PD 98059 at 10 M similarly suppressed lysoPCinduced eNOS promoter activity without a significant effect on the basal promoter activity (Fig. 7B). These results indicate that lysoPC increases Sp1 binding and eNOS promoter activities through the MEK-1 signaling pathway. This signaling pathway was involved in modulating PP2A activity. PD 98059 at 10 M suppressed PP2A activity to the basal level (Fig. 7C), without a significant effect on PP2A protein level (Fig. 3A). The CK2 protein levels were also unaffected by PD 98059 (Fig. 3B).
Effects of Purified Activated ERK2 on CK2 and PP2A Activities-Activated ERK1/2 are translocated to the nucleus where they target transcription factors including Elk-1 (18). However, our results implied that ERK1/2 might target CK2 and/or PP2A. To discern this, we incubated purified activated ERK2 (p42 MAPK, the major isoform detected as shown in Fig. 6, A and B) with CK2 in the presence and absence of PP2A. The CK2 activity was reduced when CK2 alone or mixed with PP2A was coincubated with ERK2 (Fig. 8, A and B, lanes 2 and 3). By contrast, the PP2A activity was not suppressed when purified PP2A was incubated with ERK2 (Fig. 8C, lane 1 versus lane 3). PP2A activity was suppressed by CK2 (Fig. 8C, lane 1 versus  lane 2), and this suppression was totally abrogated by omitting ATP from the assay mixture. In the absence of ATP, the PP2A activity of PP2AϩCK2 mixture was not statistically different from that of PP2A alone (n ϭ 3, p Ͼ 0.05). These results are consistent with the notion that CK2-catalyzed PP2A phosphorylation (Fig. 5E) is a key mechanism for PP2A inhibition by CK2. Addition of ERK2 to CK2/PP2A mixture increased the PP2A activity (Fig. 8C, lane 2 versus lane 4, p Ͻ 0.01). These results indicate that ERK 2 targets primarily CK2. It reduces CK2 activity through which it increases PP2A activity. DISCUSSION Our findings shed light on how basal and lysoPC-induced eNOS transcriptions are controlled and regulated. Previous studies have shown that basal eNOS promoter activity depends almost entirely on binding of Sp1 to a cognate site at the proximal region of eNOS promoter (13,14). Results from this study indicate that basal Sp1 binding activity is controlled by CK2, and lysoPC treatment of endothelial cells leads to suppression of the basal CK2 activity (Fig. 4) with an increase in PP2A activity (Fig. 2). These results implied inhibition of PP2A by CK2 at the basal state. This notion was supported by two FIG. 6. Signaling through the MEK-1 and ERK1/2 pathway. A and B, phosphorylated ERK1/2 (p44/p42 MAPK) and unphosphorylated ERK were determined by Western blot analysis using an antibody specific for phosphorylated ERK1/2 and an antibody against nonphosphorylated ERK, respectively. Concentrations shown in B for lysoPC and PD 98059 are micromolar. C, ERK activity was measured by detecting Elk-1 phosphorylation with an antibody specific for phosphorylated Elk-1. A single phosphorylated Elk-1 band was detected. pieces of evidence from our experimental data. 1) Inhibition of CK2 activity with DRB raised the PP2A activity (Fig. 2), and 2) coincubation of PP2A with CK2 suppressed PP2A activity (Fig.  8B). LysoPC enhances Sp1 binding and eNOS promoter activities by about 2-fold, while it reduces the CK2 activity in the CK2-PP2A complex to half of the basal level, consistent with the notion that reduction of CK2 activity accounts for increased eNOS promoter activity in lysoPC-treated cells. Our results further implied that lysoPC unleashes CK2's control of PP2A by an ERK1/2-mediated inhibition of CK2. This is supported by the following evidence. 1) Coincubation of ERK2 with purified CK2 or CK2 ϩ PP2A resulted in a reduction of CK2 activity (Fig. 8A) and an increase in PP2A activity (Fig. 8B), and 2) lysoPC selectively activates the ERK1/2 pathway (Fig. 6, A-C) through which PP2A activity is enhanced in cells (Fig. 2). From these experimental results, a model for basal and lysoPC-induced eNOS transcription regulation is proposed as depicted in Fig. 9. According to this model, Sp1 binding to its DNA motif which has been suggested to be governed by a balance between Sp1 phosphorylation and dephosphorylation (16,22) is controlled by CK2 interacting with PP2A. At the basal state, CK2 dominates over PP2A thereby maintaining a high level of Sp1 phosphorylation with a controlled basal Sp1 binding activity and eNOS promoter activity. Stimulation with lysoPC leads to a reduction of CK2 activity and increased PP2A activity via ERK activation, tilting the balance to favor Sp1 dephosphorylation and consequently unleashing the control accompanied by increased Sp1 binding and gene transcription.
PP2A occupies a pivotal position in eNOS transcription. Its activity is regulated by phosphorylation through its complex formation with CK2. This mode of PP2A regulation is in keeping with the PP2A properties. PP2A is a heterotrimer composed of a 36-kDa catalytic C subunit and a 65-kDa regulatory A subunit that form the core enzyme and a regulatory B subunit that binds to the core enzyme to form the holoenzyme (for a review, see Ref. 23). Phosphorylation of the catalytic or regulatory subunit has been reported to influence its catalytic activity (24 -27). Our results are consistent with a previous report that CK2 catalyzes the phosphorylation of the PP2A catalytic subunit (21). Regulation of PP2A activity is facilitated by its binding to diverse proteins including several kinases (20,21,28,29). A recent report indicates that it binds to CK2␣ but not CK2 holoenzyme and its activity is enhanced by complex with CK2␣ (21). This is contrary to our results, which show that it complexes with either CK2 holoenzyme or CK2␣ and its activity is suppressed by either form of CK2 in the complex. When ATP was omitted from the reaction mixture, the suppressing effect of CK2 on PP2A was abrogated. In our experimental system, CK2-catalyzed phosphorylation of PP2A is responsible for its inhibition of PP2A activity.
Our experimental data support the notion that Sp1 is a substrate for CK2. 2 Recent reports showed that Sp1 is phosphorylated by CK2 and phosphorylated Sp1 has reduced DNA binding activity during terminal differentiation of the liver (16,30). It was subsequently shown that Thr-579 located within the second zinc finger at the C-terminal region of Sp1 is actively phosphorylated by CK2 and mutation of this residue eliminated Sp1 phosphorylation and CK2-mediated inhibition of 2 K. Cieslik and K. K. Wu, unpublished data. Sp1 binding activity (16). Other serine residues in this region are also phosphorylated, but their identities and their functional roles remain unclear. Two classes of protein serine/threonine phosphatases have been implicated to catalyze the dephosphorylation of CK2-mediated Sp1 phosphorylation. In the liver differentiation study, protein phosphatase 1 (PP1) was implicated based on okadaic acid inhibition of Sp1 binding activity (16). However, okadaic acid is a nonselective inhibitor for PP1 and PP2A (31). Our results provide direct evidence for the involvement of PP2A in regulating Sp1 binding activity.
LysoPC, a lipid mediator, is generated from phosphatidylcholine via the action of phospholipase A 2 (32). It induces transcription of a series of endothelial genes (33)(34)(35)(36). The signaling pathway through which lysoPC activates gene transcription has not been clearly defined. Several studies have shown that it activates protein kinase C (37,38). Others have shown that it activates MAP kinases (39,40). Our results indicate that lysoPC selectively activates ERK1/2. Our results demonstrate for the first time that ERK2 exerts a direct inhibitory effect on CK2. LysoPC induces MEK-1 and ERK1/2 activation probably by at least two upstream pathways: 1) through PKC activation, which may lead to Ras or Raf activation with subsequent MEK-1 and ERK1/2 activation, and 2) through Ras activation that results in Raf activation followed by MEK-1 and ERK activation.