Phosphorylated Extracellular Signal-regulated Protein Kinases 1 and 2 Phosphorylate Sp1 on Serine 59 and Regulate Cellular Senescence via Transcription of p21Sdi1/Cip1/Waf1*

Expression of p21Sdi1 downstream of p53 is essential for induction of cellular senescence, although cancer cell senescence can also occur in the p53 null condition. We report herein that senescence-associated phosphorylated extracellular signal-regulated protein kinases 1 and 2 (SA-pErk1/2) enhanced p21Sdi1 transcription by phosphorylating Sp1 on Ser59 downstream of protein kinase C (PKC) α. Reactive oxygen species (ROS), which was increased in cellular senescence, significantly activated both PKCα and PKCβI. However, PKCα, but not PKCβI, regulated ROS generation and cell proliferation in senescent cells along with activation of cdk2, proven by siRNAs. PKCα-siRNA also reduced SA-pErk1/2 expression in old human diploid fibroblast cells, accompanied with changes of senescence phenotypes to young cell-like. Regulation of SA-pErk1/2 was also confirmed by using catalytically active PKCα and its DN-mutant construct. These findings strongly suggest a new pathway to regulate senescence phenotypes by ROS via Sp1 phosphorylation between PKCα and SA-pErk1/2: employing GST-Sp1 mutants and MEK inhibitor analyses, we found that SA-pErk1/2 regulated Sp1 phosphorylation on the Ser59 residue in vivo, but not threonine, in cellular senescence, which regulated transcription of p21Sdi1 expression. In summary, PKCα, which was activated in senescent cells by ROS strongly activated Erk1/2, and the SA-pErk1/2 in turn phosphorylated Sp1 on Ser59. Sp1-enhanced transcription of p21Sdi1 resulted in regulation of cellular senescence in primary human diploid fibroblast cells.

Expression of p21 Sdi1 downstream of p53 is essential for induction of cellular senescence, although cancer cell senescence can also occur in the p53 null condition. We report herein that senescence-associated phosphorylated extracellular signal-regulated protein kinases 1 and 2 (SA-pErk1/2) enhanced p21 Sdi1 transcription by phosphorylating Sp1 on Ser 59 downstream of protein kinase C (PKC) ␣. Reactive oxygen species (ROS), which was increased in cellular senescence, significantly activated both PKC␣ and PKC␤I. However, PKC␣, but not PKC␤I, regulated ROS generation and cell proliferation in senescent cells along with activation of cdk2, proven by siRNAs. PKC␣-siRNA also reduced SA-pErk1/2 expression in old human diploid fibroblast cells, accompanied with changes of senescence phenotypes to young cell-like. Regulation of SA-pErk1/2 was also confirmed by using catalytically active PKC␣ and its DN-mutant construct. These findings strongly suggest a new pathway to regulate senescence phenotypes by ROS via Sp1 phosphorylation between PKC␣ and SA-pErk1/2: employing GST-Sp1 mutants and MEK inhibitor analyses, we found that SA-pErk1/2 regulated Sp1 phosphorylation on the Ser 59 residue in vivo, but not threonine, in cellular senescence, which regulated transcription of p21 Sdi1 expression. In summary, PKC␣, which was activated in senescent cells by ROS strongly activated Erk1/2, and the SA-pErk1/2 in turn phosphorylated Sp1 on Ser 59 . Sp1-enhanced transcription of p21 Sdi1 resulted in regulation of cellular senescence in primary human diploid fibroblast cells.
PKC 3 comprises a family of serine/threonine kinase that modulates a variety of signal transduction pathways, leading to gene expression, cell proliferation, and differentiation. PKC isoforms are classified into three subgroups. The conventional PKC isozymes, comprising, ␣, ␤I, ␤II, and ␥, are activated by Ca 2ϩ , phosphatidylserine, diacylglycerol, or phorbol esters; the novel PKC isozymes, consisting of PKC␦, -⑀, -, -, and -, are activated by phosphatidylserine, diacylglycerol, or phorbol esters, but insensitive to Ca 2ϩ ; and the atypical isoforms ( and /) are dependent on phosphatidylserine for activation, but not affected by Ca 2ϩ , diacylglycerol, or phorbol esters (1)(2)(3). Activated PKC translocates from cytosol to membranous organelles and/or to the nucleus (4,5). In addition, oxidative stress has been reported to induce prolonged activation of PKC within the cells (6 -9). The growth regulatory consequences of PKC activation suggest a link between PKC signaling and control of the cell cycle machinery. Activation of PKC has been shown to result in alterations of cell cycle progression in either stimulatory or inhibitory directions in several systems (10 -15). Among the isoforms, PKC␣ has been implicated in the control of G 1 /S transition (14, 16 -18) and recently in the regulation of cancer cell senescence (19), whereas PKC␤II has been shown to play a role in progression from G 2 into M phase (15), and PKC␦ has been associated with the control of M phase (20).
Normal human diploid fibroblasts (HDFs) have a defined proliferative capacity in tissue culture. Hayflick and Moorhead (21) described in detail about the limited proliferative capacity of normal cells and postulated that it was an in vitro manifestation of human aging. Cellular senescence plays a key role in complex biological processes, including development, aging, and tumorigenesis. Hallmarks of cellular senescence are metabolically active but not responsive to mitogens, therefore, decreased cell growth, cell cycle arrest, and tumor suppressor effects are accompanied with flat and large cell shapes. Characteristic features of cellular senescence are cytoplasmic sequestration of phospho-extracellular signal-regulated protein kinase 1/2 (SA-pErk1/2), and increased levels of ROS (22). The phenomena are well corroborated by the results that activation of H-ras induces cellular senescence of mouse embryo fibroblasts (23) and primary culture of HDF cells (24), and that infection of HDF with virus carrying H-ras double mutants, G12V/ T35S, G12V/E37G, and G12V/Y40C, accumulates SA-pErk1/2 with increased MEK activity. On the other hand, treatment of old HDF cells with 12-O-tetradecanoylphorbol-13-acetate (TPA) reverses cellular senescence phenotypes (25): increased DNA synthesis, phosphorylation of pRB, reduced expressions of p21 Sdi1 and SA-pErk1/2, as well as reduced expression of ␤-galactosidase along with morphological changes of the senescent cells. It has been known that TPA initially activates PKC, and then down-regulates its isoforms (26,27). Therefore, PKC isozymes have strongly been suggested to be an important in vivo regulator of senescence phenotypes. However, there has been no report on the crosstalk between PKC isozymes and SA-pErk1/2 and the role of SA-pErk1/2 in the process of cellular senescence.
In addition to the above mentioned characteristics of cellular senescence, p21 Sdi1 has been well known as a key molecule to induce normal cell senescence (28,29). However, there is no report on the induction mechanism of p21 Sdi1 expression during cellular senescence, except transcriptional activation of the p21 Waf1/Cip1 promoter by Sp1 after phosphorylation on the two threonine residues (Thr 453 and Thr 739 ) by mitogen-activated protein kinase (30), but not serine residues (31). In this study, we investigated the role of SA-pErk1/2 in cellular senescence downstream of the PKC␣; PKC␣ was responsible for ROS generation and activation of Erk1/2 in senescent cells. SA-pErk1/2 enhanced the expression of p21 Sdi1 via phosphorylation of Sp1 on Ser 59 in vivo downstream of PKC␣, resulting in induction of senescence phenotypes.
Cell Culture-Primary culture of HDF was prepared and maintained in our laboratory (22,24) in Dulbecco's modified Eagle's medium (DMEM-high glucose, Invitrogen) supplemented with 10% fetal bovine serum (Invitrogen). The number of population doublings of HDF was calculated based on the equation, PDs ϭ log[A/BC]/log2; A, B, and C indicate the numbers of collected cells, plated cells, and the attachment efficiency, respectively. Doubling time of HDF cells was measured with PDs, and young cells used in this study represent cells with doubling time of around 24 h. Doubling time of mid-old and old cells were 7-10 and 14 days, respectively. All cells used in this study were maintained in a 5% CO 2 incubator at 37°C.
siRNA Transfection-Young and mid-old cells were transfected with siRNA against PKC␣, PKC␤I, and Sp1 (Santa Cruz) using Oligofectamine (Invitrogen) following the protocol provided by the manufacturer.
Subcellular Fractionation-Cytoplasmic and nuclear extracts were prepared as described previously (32) with minor modifications. Briefly, the cells were washed twice with ice-cold phosphate-buffered saline, and transferred to 1.5-ml Eppendorf tubes. The cells were then centrifuged at 300 ϫ g for 4 min at 4°C. After centrifugation, the pellet was resuspended in 400 l of cold buffer A (10 mM HEPES-KOH (pH 7.5), 10 mM KCl, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride (PMSF), and 1 g/ml leupeptin). After incubation on ice for 15 min, 12.5 l of 10% Nonidet P-40 was added, and the mixture vortexed briefly and incubated on ice for 10 min. The nuclei were pelleted by centrifugation at 1,500 ϫ g for 5 min at 4°C, whereas the supernatant (cytoplasmic extracts) was recovered by centrifugation at 13,000 ϫ g for 15 min. Nuclei were washed twice with 1 ml of ice-cold buffer A, and then resuspended in 50 l of ice-cold buffer B (20 mM HEPES-KOH (pH 7.5), 0.4 M NaCl, 1 mM dithiothreitol, 1 mM PMSF, and 1 g/ml leupeptin), followed by incubation on ice for 30 min. The mixture was then centrifuged at 18,000 ϫ g for 5 min, and the supernatant was collected as a nuclear extract. To determine subcellular distribution of PKC isozymes, soluble and particulate fractionation was performed as described previously with some modifications (33). For soluble and particulate fractions, cells were washed twice with ice-cold phosphate-buffered saline and scraped into a homogenization buffer containing 25 mM Tris/ HCl (pH 7.4), 2 mM EDTA, 0.25 M sucrose, 1 g/ml leupeptin, and 1 mM PMSF and then lysed twice by sonication (Sonic Dismembrator 550, Fisher) at level 2 for 10 s. The lysates were centrifuged at 500 ϫ g for 5 min, and the low-speed supernatant was centrifuged at 100,000 ϫ g for 1 h. The high-speed supernatant constituted the soluble fraction. The high-speed pellet was washed three times and extracted in ice-cold homogenization buffer containing 1% Triton X-100 for 30 min. The Tritonsoluble component (particulate fraction) was separated from the Triton-insoluble material by centrifugation at 100,000 ϫ g for 1 h.
Immunoprecipitation-Immunoprecipitation (IP) was performed with young and mid-old HDF cell lysates (500 g or 1 mg) in the IP buffer containing 50 mM Tris/HCl (pH 7.4), 150 mM NaCl, 0.5% deoxycholic acid, 1% Nonidet P-40, 1 mM PMSF, and 1 g/ml leupeptin by the standard method. Whole cell lysates were pre-cleared with protein G-agarose beads (Invitrogen) for 1 h at 4°C before precipitation overnight with anti-PKC␣, anti-PKC␤I, anti-cdk2, or anti-Sp1 antibody at 4°C. The immunoprecipitates were washed 3 times with IP buffer, and then subjected to kinase assay or immunoblot analysis with anti-PKC␣, anti-PKC␤I, anti-cdk2, or anti-Sp1 antibody.
Immunoprecipitation and in Vitro Kinase Assays-For the IP kinase assay, the immunoprecipitates of PKC␣, PKC␤I, and cdk2 were washed twice with IP buffer, and then twice with kinase buffer (50 mM HEPES, pH 7.5, 10 mM MgCl 2 , 1 mM dithiothreitol, 2.5 mM EGTA, 1 mM sodium fluoride, 0.1 mM Na 3 VO 4 , 10 mM ␤-glycerophosphate) before resuspended in 20 l of kinase buffer. Kinase assay was initiated by adding 40 l of kinase buffer containing 5 g of MBP or histone H1 substrate and 5 Ci of [␥-32 P]ATP. For in vitro kinase assay, recombinant GST-hSp1 proteins (wild type and mutant) were used as substrates. The kinase assay was initiated by adding PKC␣ (Calbiochem) or Erk1 (Upstate) to 30 l of kinase buffer containing 3 g of GST-hSp1 and 5 Ci of [␥-32 P]ATP. The reactions were carried out at 30°C for 30 min, and terminated by adding SDS sample buffer followed by boiling for 5 min. The reaction products were analyzed by SDS-PAGE and autoradiography.

Measurement of ROS by FACS Analysis-The level of H 2 O 2 in the senescent cells was measured using FACScan (BD Biosciences) by measuring the fluorescence of DCF arising from oxidation of H 2 -DCFDA.
Electroporation-An MP-100 Microporator (Digital Bio Technology) was used to electroporate HDF cells. Cells (2 ϫ 10 5 ) were resuspended in 10 l of microporation buffer (Digital Bio Technology) containing 1 g of PKC␣-DN or PKC␣-CA DNA. The cells were pulsed by using a 1 ϫ 30-ms protocol with 1400 V and then plated on 6-well plates. At 48 h after electroporation, cells were harvested for Western blot analysis. The pcDNA3-HA was used as a control.
Plasmid Transfection and Promoter Analysis-NIH3T3 cells in 6-well plates (5 ϫ 10 4 cells/well) were transiently transfected with 1 g of a p21 promoter-driven luciferase reporter plasmid, pGL2-0.3 (35), together with 1 g of pcDNA3-HA(vector), PKC␣-DN, or PKC␣-CA using Lipofectamine (Invitrogen). A Renilla luciferase expression vector (100 ng, pRL-TK, Promega) was co-transfected for normalization of transfection efficiency. Forty-eight hours later, cell extracts were subjected to luciferase determination using the Dual Luciferase Reporter Assay System (Promega). Results were expressed as the ratio between firefly and Renilla luciferase. Wild type (WT) and S59A point mutant (S59A) of full-length Sp1 were cloned into the EcoRI and SalI sites of p3xFLAG-CMV (Sigma). Huh7 cells in 6-well plates (5 ϫ 10 4 cells/well) were transiently transfected with 500 ng of p21 promoter-driven luciferase reporter plasmid (pGL2-0.3) and WT-Sp1 or S59A-Sp1 (500 ng) using Lipofectamine. A Renilla luciferase expression vector (50 ng, pRL-TK) was cotransfected for normalization of transfection efficiency. Twenty-four hours later, serum was deprived for 1 day before stimulation with EGF for 6 h. And then, the cell extracts were subjected to promoter analysis by measuring the luminescence with TD-20/20 luminometer (Turner Designs).
Statistical Analysis-Data were presented as mean Ϯ S.D. and analyzed by Student's t test. p value less than 0.05 was considered as statistically significant.

PKC Isozymes are Activated in the Replicatively Senescent
Cells-Our previous study showed that treatment of senescent cells with TPA reverses senescence phenotypes to young celllike (25). In the present study, we attempted to determine the PKC isozymes possibly involved in the process. Therefore, primary cultures of young and old HDF cells were treated with either TPA (50 ng/ml) or DMSO (0.01%) for 15 min, and the cell lysates were then separated into soluble (S) and particulate (P) fractions by centrifugation. Thirty g each of the fractions were resolved on 8% SDS-PAGE and then hybridized with anti-PKC antibodies. As shown in Fig. 1A, PKC␣, PKC␤I, and PKC in both young and old cells were activated by TPA treatment. Fig.  1B shows that there was no change in the expressions of PKC␣ and PKC␤I proteins, whereas the kinase activities were significantly higher in the older cells than the younger cells. To investigate whether activation of PKC regulated cellular senescence or not, GF109203X was applied to senescent cells for 3 days, and the result showed that inhibition of PKC isozymes clearly changed the large and flat cells to small and slim fibroblasts (Fig.  1C). Furthermore, immunoblot analyses revealed that the inhibition of PKC activities with GF109203X significantly reduced expressions of molecular markers of senescence, such as p21 Sdi1 , p53, and SA-pErk1/2 (Fig. 1D). These features are well accordant with a recent report that PKC␣ works as a mediator of G 2 /M arrest and senescence of non-small cell lung cancer cells, and that it induces p21 Sdi1 , an obligatory event for conferring the senescence phenotype (19).

PKC␣ and ROS are Concurrently Regulated during Cellular
Senescence-Because of our earlier study that the level of H 2 O 2 is much higher in the senescent HDF than the young cells (22) and the reports that ROS stimulates PKC activity (36 -38), we assessed whether the H 2 O 2 accumulated in the senescent cells was responsible for activation of PKC␣ and PKC␤I. Therefore, in vitro kinase assay was performed with immunoprecipitates of PKC␣ and PKC␤I obtained from young and old HDF cells after treatment with H 2 O 2 , TPA, or NAC. As shown in Fig. 2A, not only TPA (50 ng/ml) but H 2 O 2 (1 mM) also increased the activities of PKC␣ and PKC␤I in young cells, whereas treatment of old cells with NAC significantly reduced the PKC activities. These data are well supported by the study that ROS regulates PKC activity in human granulocytes according to age (39).
To investigate which PKC isozyme can regulate ROS levels in cellular senescence, senescent cells were treated with either PKC␣-siRNA or PKC␤I-siRNA, and the changes of ROS levels FIGURE 1. Activation of PKC␣ and PKC␤I during replicative senescence. A, primary culture of HDF, young and old, cells were treated with either TPA (50 ng/ml) or DMSO (0.01%) for 15 min, and the cell lysates were fractionated into soluble (S) and particulate (P) fractions by centrifugation. Thirty g of the fractions were resolved on 8% SDS-PAGE and then hybridized with anti-PKC antibodies. Note activations of PKC␣, PKC␤I, and PKC after treatment with TPA in both the young and old cells. B, to evaluate expression and activity of the PKC isozymes in the cells, immunoblot analyses and in vitro kinase assays were performed with cell lysates obtained from the young and old HDF cells using anti-PKC antibodies and myelin basic protein (MBP) as a substrate, respectively. There was no significant change in the expressions of PKC␣ and PKC␤I, however, the activities of PKC␣ and PKC␤I were significantly higher in the old cells than that of the young cells. For activity assay, cell lysates were immunoprecipitated with either anti-PKC␣ or anti-PKC␤I antibodies. The precipitates were washed twice with precipitation and kinase buffers, and then resuspended in 20 l of kinase buffer. Kinase assay was performed by adding 40 l of the buffer containing 5 g of MBP and 5 Ci of [␥-32 P]ATP at 30°C for 30 min, and then terminated by adding SDS sample buffer and boiling for 5 min. The reaction products were analyzed by SDS-PAGE and autoradiography. C, to investigate whether activation of PKC regulates cellular senescence or not, PKC inhibitor, GF109203X (4 M), was applied to mid-old cells for 3 days, and the treatment was found to clearly change senescent cells to young cell-like; cells were changed from large with flat cytoplasm to small and slim. D, molecular changes of cellular senescence were also evaluated by immunoblot analyses after inhibiting PKC activities. Note reduced expressions of senescence markers, such as p21 Sdi1 , p53, and SA-pErk1/2, after treatment of senescent cells with PKC inhibitor, GF109203X.
were measured by FACS using H 2 DCF-DA. As already known, ROS levels of mid-old cells was higher than that of young cells (Fig. 2B), however, PKC␣-siRNA, but not PKC␤I-siRNA, significantly reduced the level in the mid-old cells (Fig. 2C). These findings show that H 2 O 2 increased PKC␣ activity in the senescence process and PKC␣ regulated the H 2 O 2 level in senescent cells. The observation can be supported by the reports that PKC␣, -␤II, -␦, and -phosphorylate p47 phox sites and activate NADPH oxidase (40,41).

Down-regulation of PKC␣ Inhibits Growth Arrest and Senescence
Phenotypes-Because PKC␣-siRNA reduced ROS accumulation in senescent cells, regulation of senescence phenotypes by PKC␣ was investigated. When PKC␣-siRNA and PKC␤I-siRNA were introduced into HDF cells, immunoblot analyses revealed reduced expressions of PKC␣ (Fig. 3A) and PKC␤I (Fig.  3D), respectively. Moreover, transfection of PKC␣-siRNA released growth arrest of the senescent cells (Fig. 3C), however, there was no change in young cells (Fig. 3B).
Here, GFP-siRNAs were employed for a control experiment. On the other hand, down-regulation of PKC-␤I failed to induce proliferation of senescent cells (Fig. 3F), but rather decreased the proliferation in young cells (Fig. 3E). Therefore, to confirm the effect of PKC␣ on cell proliferation, we assessed cell cycle profiles using siRNAs against GFP or PKC␣. In 24 h of treatment, progression of G 1 to S phase in addition to G 2 /M was more than 2-fold higher in the PKC␣-siRNA-treated cells than GFP control and no treatment control (Fig. 3G). PKC␣-mediated progression of cell division cycle was further proved by an increase of cdk2 activity in the PKC␣-siRNA-treated cells, but not the control, evaluated by the IP kinase assay with histone H1 as substrate (Fig. 3H). Indeed, down-regulation of PKC␣ induced G 1 to S transition and multiplication of the replicatively senescent HDF cells.
To investigate whether PKC␣ also regulated molecular markers of senescence, old HDF cells were treated with PKC␣-siRNA for 2 days, and a few markers were measured by immunoblot analyses while employing treatment with NAC as positive control. As expected, PKC␣-siRNA significantly reduced the expressions of p21 Sdi1 , p53, and SA-pErk1/2, as compared with those of the GFP-siRNA (Fig. 3I). It should be noted particularly that down-regulation of PKC␣ significantly reduced the expression of SA-pErk1/2. To evaluate distribution of SA-pErk1/2, immunoblot analyses were performed with nuclear and cytoplasmic fractions of the young and old HDF cells before (Fig. 3J) and after (Fig. 3K) transfection with siRNAs. As shown in the figures, the expression of SA-pErk1/2 was significantly reduced in both fractions of the old cells treated with PKC␣-siRNA, as compared with GFP-siRNA. Here, lamin B was used as a loading control and a marker of nuclear proteins.
PKC␣ Regulates Expressions of SA-pErk1/2 and p21 Sdi1 -To investigate whether the expression of SA-pErk1/2 was regulated by PKC␣ or not, pcDNA3-HA (vector), dominant negative PKC␣ (PKC␣-DN), or catalytically active PKC␣ (PKC␣-CA) was transfected to HDF young cells by electroporation, and whole cell lysates (30 g) were subjected to immunoblot analyses in 2 days. As shown in Fig. 4A, PKC␣-CA significantly induced expressions of SA-pErk1/2 and p21 Sdi1 , whereas PKC␣-DN down-regulated expression of SA-pErk1/2, as compared with that of pcDNA3-HA. Moreover, DiC 8 (20 g/ml), a PKC activator, also increased expression of SA-pErk1/2 and p21 Sdi1 in young HDF cells (Fig. 4B) with a concomitant decrease of cell growth (Fig. 4C). These findings strongly suggested SA-pErk1/2 as a downstream effector of PKC␣ in the process of cellular senescence. To confirm the regulation of the phenotypes by active PKC␣, p21 Sdi1 promoter-driven luciferase assay was performed in NIH3T3 cells (Fig. 4D); cells were co-transfected with p21 Sdi1 promoter-luciferase (pGL2-0.3) together with pcDNA3-HA, PKC␣-DN, or PKC␣-CA. In 48 h, the cell lysates were subjected to luciferase (Luc) assay. As expected, luciferase activity was markedly increased after transfection with PKC␣-CA, indicating that transcription of p21 Sdi1 was increased only by PKC␣-CA, but not PKC␣-DN or the vector alone. Taken together, PKC␣ regulated transcription of p21 Sdi1 in addition to the induction of SA-pErk1/2 expression.
Ser 59 and Thr 739 in Sp1 are Phosphorylated by Active Erk1-To unequivocally establish the sites in Sp1 phosphorylated by active Erk1, site-directed mutagenesis of the GST-⌬1 and GST-⌬3 proteins was performed by point mutation of Ser 59 to alanine (Fig. 6A) and Thr 739 to alanine (Fig. 6C). As shown in Fig.  6, B and D, active Erk1 failed to phosphorylate GST-⌬1 S59A and GST-⌬3 T739A mutant proteins, as opposed to phosphorylation of the wild types, thus revealing Ser 59 in GST-⌬1 and Thr 739 in GST-⌬3 as in vitro targets of active Erk1.
PKC␣ and SA-pErk1/2 Phosphorylate Sp1 on Serine, but Not Threonine, in Senescent Cells-To investigate whether PKC␣ and SA-pErk1/2 regulate Sp1 phosphorylation in vivo or not, Sp1 was isolated from cell lysates by immunoprecipitation after FIGURE 3. Regulations of G 1 arrest and other senescent phenotypes by PKC␣, but not PKC␤I. Young and old HDF cells (4 ϫ 10 3 ) were cultured in 12-well plates for 16 h, and then treated with GFP-siRNA, PKC␣-siRNA, or PKC␤I-siRNA for the indicated times. Down-regulation of PKC␣ (A) significantly increased proliferation of senescent cells from 2 days after treatment with PKC␣-siRNA (C), as compared with that of the GFP-siRNA treated. However, there was no effect of PKC␣-siRNA in the young cells (B). Data indicate mean Ϯ S.D., *, p Ͻ 0.05 (paired t test). On the other hand, PKC␤I-siRNA failed to regain cell proliferation of the mid-old cells (F). Instead, down-regulation of PKC-␤I (D) significantly reduced growth of the young cells from 2 days after treatment (E), **, p Ͻ 0.01 (paired t test). When mid-old cells were treated with PKC␣-siRNA, cells in G 1 phase were significantly reduced in 24 h, as compared with that of the GFP-siRNA-treated cells (G), indicating release of senescent cells from G 1 arrest by PKC␣-siRNA. H, down-regulation of PKC␣ significantly increased cdk2 activity in senescent cells. Mid-old cells were treated with siRNAs for 24 h, and the cell lysates (200 g) were immunoprecipitated with anti-cdk2 antibody to perform an in vitro kinase assay using histone H1 as a substrate. Treatment of senescent cells with PKC␣-siRNA significantly induced ckd2 activity, but not GFP-siRNA. I, to confirm the reversal of the senescence program by regulating PKC␣ activity, markers of cellular senescence in the old cells were evaluated. Note the significant expressions of p21 Sdi1 , p53, and SA-pErk1/2 in the old cells compared with those of the young cells (left panel). GFP-siRNA and PKC␣-siRNA were transfected into old HDF cells and the changes of senescent markers were measured by immunoblot analyses. Note significantly reduced expressions of p21 Sdi1 , p53, and SA-pErk1/2 by treating the cells with PKC␣-siRNA, as compared with those of the GFP-siRNA-treated cells (middle panel). NAC also reduced the cellular senescence markers (right panel). Expression of tubulin indicates loading control between the lanes. J, young and old HDF cells were fractionated into nuclear (Nu) and cytosolic (Cyt) fractions as described under "Experimental Procedures." Thirty g of proteins were resolved on 10% SDS-PAGE and then hybridized with anti-pErk1/2 and anti-lamin B, a marker for nuclear fraction, antibodies. It shows increased expression of pErk1/2 in the old compared with the young cells. K, note significantly reduced SA-pErk1/2 in both the cytoplasm and nuclear fractions of the old cells, when treated with PKC␣-siRNA, compared with those of the GFP-siRNA-treated cells. treatment of old HDF cells with GFP-siRNA, PKC␣-siRNA, or NAC. The precipitates were then analyzed by immunoblotting with anti-Ser(P) and anti-Thr(P) antibodies. Unexpectedly, Sp1 was phosphorylated only on serine, but not threonine, and was regulated by treatment with PKC␣-siRNA (Fig. 7A) and NAC (Fig. 7B). To confirm in vivo phosphorylation of Sp1 by SA-pErk1/2, mid-old HDF cells were treated with U0126 for 1 h, and the Sp1 protein was extracted by immunoprecipitation. As shown in Fig. 7C, U0126 significantly reduced the expressions of pErk1/2 (80%) and Ser(P) (50%) in Sp1 as compared with the untreated control. At the same time, threonine phosphorylation was not found in the Sp1 molecules. These data strongly indicate serine residue as the in vivo phosphorylation site of Sp1 by not only PKC-␣ but also SA-pErk1/2, but not threonine, in the process of cellular senescence.
In Vivo Phosphorylation of Sp1 on Ser 59 Regulates the Expression of p21 Sdi1 -To investigate whether phosphorylation of Sp1 on Ser 59 regulates p21 Sdi1 expression, mid-old HDF cells were treated with siRNAs for 2 days, and then expressions of Sp1 and p21 Sdi1 were determined by immunoblot analyses. As shown in Fig. 7D, Sp1 expression was almost completely down-regulated (0.08) with concurrent reduction of p21 Sdi1 (0.42) expression when treated with Sp1-siRNAs, as compared with control scrambled siRNA. To confirm whether Ser(P) 59 in Sp1 affects FIGURE 5. In vitro phosphorylation of GST-Sp1 by PKC␣ and Erk1. To investigate whether phosphorylation of Sp1 was regulated directly by PKC␣ or SA-pErk1/2, a human Sp1 fusion construct (GST-Sp1) and its deletion mutants (GST-⌬1, 1-110 residues of Sp1; GST-⌬2, 301-350 residues of Sp1; GST-⌬3, 504 -785 residues of Sp1) were prepared as described under "Experimental Procedures," focusing on the potential phosphorylation sites (A). The recombinant proteins were expressed in E. coli and confirmed by Coomassie Blue stain after SDS-PAGE (B). In vitro phosphorylations of Sp1 by PKC␣ (C) and active Erk1 (D) were evaluated by an in vitro kinase assay. The recombinant proteins were subjected to kinase assay in the presence of 5 Ci of [␥-32 P]ATP and 10 times excess amount of unlabeled ATP. The reaction mixture was separated by SDS-PAGE and then visualized by autoradiography. Note phosphorylated GST-⌬1 and GST-⌬3, but not GST-⌬2, in addition to the GST-Sp1 full sequence by active PKC␣ and active Erk1. To identify phosphorylated residues in the recombinant proteins, immunoblot (IB) analyses were performed with anti-Ser(P) (E) and anti-Thr(P) (F) antibodies. Note the phosphorylations of serine and threonine residues in the GST-⌬1 and GST-⌬3, respectively, by active Erk1.
transcriptional activity of p21 Sdi1 or not, Huh7 cells were transfected with p21 Sdi1 -luciferase together with either WT-Sp1 or the S59A-Sp1 construct for 24 h, and then serum-starved for 24 h before treatment with EGF for 6 h. As expected, luciferase activity was significantly increased by EGF treatment in the WT-Sp1-transfected cells, as compared with the untreated control. However, the effect of EGF was not found in the S59A-Sp1-transfected cells (Fig. 7E). The data clearly indicate the role of Ser(P) 59 Sp1 for regulation of p21 Sdi1 transcription in response to EGF treatment.
The cellular response to ROS was not confined to PKC activation, but also induced expression of SA-pErk1/2 (supplemental Fig. S1). The findings strongly suggest concurrent regulations of PKC␣, pErk1/2, and ROS in cellular senescence. Indeed, ROS and PKC␣ were concurrently regulated in HDF cells, and this was confirmed by treatment of senescent HDF cells with PKC␣-siRNA. Activation of PKC␣ increased the expression of pErk1/2 in senescent cells, and SA-pErk1/2 phosphorylated Sp1 only on Ser 59 in vivo, consequently enhancing transcription of p21 Sdi1 . Elevated p21 Sdi1 inhibited the cell cycle at the G 1 phase, leading to growth arrest of senescent cells (Fig.  8). On the other hand, down-regulation of PKC␣ by treating old cells with PKC␣-siRNA significantly reduced SA-pErk1/2 in both nuclear and cytoplasmic fractions of old cells, and released senescent cells from G 1 arrest of the cell division cycle with activation of ckd2. Taken together, p21 Sdi1 expression seemed to be regulated via Sp1 phosphorylation on Ser 59 by SA-pErk1/2 downstream of PKC␣ in the replicative senescence process.

DISCUSSION
Replicative senescence of human diploid fibroblast has frequently been used as an aging model in vitro (46). One of the significant phenotypes of replicative senescence is the SA-pErk1/2 due to persistent activation of H-ras (23,24). On the other hand, the role of SA-pErk1/2 in cellular senescence has not yet been reported, although the expression is constitutively higher in the cytoplasm of senescent cells (22). To the best of our knowledge, therefore, this is the first study to confirm that SA-pErk1/2 activates the transcription factor, Sp1, via Ser 59 phosphorylation downstream of PKC␣, leading to transcription of p21 Sdi1 and resulting in replicative senescence of HDF cells. We do not yet know why SA-pErk1/2 phosphorylated only Ser 59 in vivo, but not Thr 739 , despite in vitro phosphorylation of both of these residues by active Erk1 (Fig. 6). Nevertheless, the present study is well supported by the recently proposed mechanism of TPA-induced Erk1/2 regulation during early events of ML-1 cell differentiation to macrophage via induction of p21 Sdi1 (47) and also the observation that PKC␣ induces senescence of human lung cancer cells (19).
We found that stimulation of HDF cells with TPA significantly increased activities of PKC␣, PKC␤I, and PKC ( Fig. 1). At present, there are no reports on the role of PKC in senescence of HDF, whereas there are a few reports that may suggest a role of PKC for cellular senescence. Normal colon mucosa expresses mRNAs of the following isoforms of PKC, in decreasing order of abundance: PKC␦ Ͼ PKC Ͼ PKC␣ Ͼ PKC␤ Ͼ PKC⑀. PKC is normally expressed in the more differentiated cells of epithelial tissues. However, expression of PKC is reduced in colon cancer and returns to normal levels in the more differentiated epithelial cells (48). Inhibition of PKC expression significantly reduces keratinocyte growth, suggesting a potential role of PKC in keratinocytes (49). Induced differentiation of neoplastic keratinocytes accompanies an increase of PKC expression (50). PKC␣ was responsible for generation of ROS in senescent cells (Fig. 2), and PKC␣-siRNA, but not PKC␤I-siRNA, released senescent cells from growth arrest (Fig. 3) accompanied with reversal of the senescence phenotypes (Fig. 4).
Treatment of HDF cells with TPA initially activated PKC␣ and PKC␤I (Figs. 1A and 2A), and down-regulation of PKC␣ and PKC␤I occurred in 8 and 4 h, respectively (supplemental Fig. S2). Nevertheless, growth arrest was regulated by PKC␣ only, not PKC␤I (Fig. 3). The kinetics of PKC␣ reduction are well correlated with reversal of senescence morphology in 8 h of TPA treatment (25). Furthermore, down-regulation of PKC␣ was accompanied FIGURE 6. In vitro phosphorylation of Sp1 on Ser 59 and Thr 739 by active Erk1. Based on the above data, serine and threonine residues were suspected as the phosphorylation sites in the GST-⌬1 and GST-⌬3, respectively. Therefore, Ser 59 to Ala (A) and Thr 739 to Ala (C) mutant constructs were prepared by site-directed mutagenesis. Employing kinase analyses, we confirmed that active Erk1 phosphorylated GST-⌬1 and GST-⌬3 proteins as expected, however, it failed in GST-⌬1 and GST-⌬3 proteins with the Ser 59 to Ala mutant (B) and the Thr 739 to Ala mutant (D), indicating the Ser 59 and Thr 739 residues as in vitro phosphorylation sites of active Erk1.
with reduced expression of senescence markers, especially SA-pErk1/2 (Fig. 3, I and K). These findings strongly support SA-pErk1/2 as a downstream mediator of PKC␣, which is activated in response to ROS in cellular senescence.
In the present study, SA-pErk1/2 was found to be the kinase of Sp1 on Ser 59 , resulting in transcription of p21 Sdi1 , in good agreement with reports that activation of the H 2 O 2 -mediated ERK signaling pathway is required for p21 Sdi1 expression by TGF-␤1, and that elimination of ROS with either antioxidant or catalase is accompanied by the inhibition of ERK activation by TGF-␤1, resulting in attenuation of p21 Sdi1 expression without any alteration of nuclear translocation of Smads (31,51,52). Some studies have implicated pErk1/2 as a mediator of p53independent p21 Sdi1 induction (53,54). These observations are in good accordance with another report that PKC␣ and PKC⑀ stimulate p21 Sdi1 expression via Sp1 (42). DNA-PK, PKA, PKC, casein kinase II, Erk1/2, and cyclin-dependent kinase are responsible for Sp1 activation, whereas protein phosphatase 1 FIGURE 7. In vivo phosphorylation of Sp1 on Ser 59 by SA-pErk1/2 and regulation of p21 Sdi1 expression. To evaluate in vivo phosphorylation of Sp1 by PKC-␣, HDF old cells were treated with either GFP-siRNA or PKC␣-siRNA for 2 days (A), or treated with either vehicle or 10 mM NAC for 12 h (B). The cell lysates (500 g) were incubated with anti-Sp1 antibody and protein G-Sepharose beads for 8 h at 4°C with continuous agitation. Immunoprecipitated complexes were separated by SDS-PAGE, and subjected to immunoblot analyses with anti-Thr(P), anti-Ser(P), or anti-Sp1 specific antibodies. Note serine phosphorylation of Sp1 in the cells, but not with threonine. Moreover, the degree of Ser(P) in Sp1 was significantly reduced by treatment of the cells with PKC␣-siRNA and NAC. These findings indicate that Sp1 was in vivo phosphorylated only on the serine residue, but not on threonine. C, at the same time, to evaluate in vivo regulation of Sp1 phosphorylation by SA-pErk1/2, mid-old HDF cells were treated with U0126 for 1 h, and Sp1 protein was extracted by immunoprecipitation. U0126 significantly reduced expressions of pErk1/2 (0.2) and Ser(P) in Sp1 (0.5) than the untreated control. Moreover, threonine phosphorylation in the Sp1 molecule by pErk1/2 was not found. These experiments clearly indicate Sp1 phosphorylation on the serine residue by not only PKC-␣ but also SA-pErk1/2, but not threonine, in the process of cellular senescence. D, to investigate whether phosphorylation of Sp1 on Ser 59 regulates p21 Sdi1 expression, mid-old HDF cells were treated with scrambled-siRNA or Sp1-siRNA for 2 days, and then expressions of Sp1 and p21 Sdi1 were determined by immunoblot analyses. Expression of Sp1 was almost completely down-regulated with concurrently reduced p21 Sdi1 expression when treated with Sp1-siRNAs. E, to confirm whether Ser(P) 59 in Sp1 affects transcription of p21 Sdi1 or not, Huh7 cells were transfected with p21 Sdi1 -luciferase along with either WT-Sp1 or the S59A-Sp1 construct for 24 h, and then serum-starved for 24 h before treatment with EGF for 6 h. As expected, luciferase activity was significantly increased by EGF treatment in the WT-Sp1transfected cells, as compared with the control (*, p Ͻ 0.01). However, the effect of EGF was not found in the S59A-Sp1-transfected cells, indicating the role of Ser(P) 59 -Sp1 in regulation of p21 Sdi1 transcription after EGF treatment. and 2A dephosphorylate. Furthermore, PKC␣ was constitutively activated in senescent cells without any change of expression levels (Fig. 1B). One possible cause of PKC␣ activation might be ROS accumulated in senescent cells (22). ROS such as H 2 O 2 can activate PKC by receptor tyrosine kinase (36 -38). In accordant with this notion, NAC reduced PKC␣ activity in senescent cells ( Fig. 2A). Another cause of PKC activation might be higher levels of diacylglycerol, a PKC activator, in senescent cells than in young cells (55). Indeed, down-regulation or activation of PKC␣ using PKC␣-siRNA and the catalytic subunit of PKC␣, respectively, regulated senescence phenotypes oppositely (Figs. 3 and 4). Furthermore, it has been reported that Sp1 is phosphorylated on Ser 59 , Ser 101 , Ser 131 , Thr 278 , Thr 355 , Thr 453 , Ser 641 , Thr 668 , Ser 670 , Thr 681 , and Thr 739 (56 -65).
When phosphorylation of Sp1 by PKC␣ and active Erk1 was investigated in vitro with deletion mutants of human Sp1 proteins, PKC␣ and active Erk1 were found to phosphorylate Sp1 on both serine and threonine residues, however, Sp1-⌬1 was a weaker substrate for PKC␣ than Sp1-⌬3. In contrast, however, Sp1-⌬1 was a slightly better substrate for Erk1 than Sp1-⌬3 (Fig. 5, C versus D). As Sp1-⌬3 has more sites phosphorylated by PKC (64,65) than by Erk (59,63), it might be a better substrate for PKC␣. When phosphorylation of the Sp1-⌬1 S59A mutant by PKC␣ was examined, phosphoryla-tion was found to be significantly reduced with some residual phosphorylation (1.00 versus 0.17; supplemental Fig. S3), strongly suggesting that PKC␣ phosphorylates Sp1-⌬1 not only on Ser 59 but also with some additional target site. In contrast, Ser 59 may be the only target for Erk1 (Fig. 5D). Employing GST-Sp1 mutant proteins (S59A and T739A), we confirmed indeed that both Ser 59 and Thr 739 residues of Sp1 were in vitro phosphorylated by active Erk1 (Fig. 6), however, Sp1 immunoprecipitates isolated from senescent HDF cells contained only Ser(P) (Fig. 7, A-C). The result indicates that phosphorylation of Ser 59 by pErk1/2 occurs both in vitro (Fig. 6A) and in vivo (Fig. 7C) in senescent cells, thereby increasing transcription of p21 Sdi1 by phosphorylated Sp1 on the Ser 59 residue in response to growth factor signals (Fig.  7E). In the present study, it seems that Ser 59 of Sp1 is a major phosphorylation residue by Erk1/2 for inducing transcription of p21 Sdi , although PKC␣ also phosphorylates Sp1 on Ser 59 directly. However, we need to explain why in vivo phosphorylation of Sp1 occurred only on Ser 59 , but not on Thr 739 , in the senescent cells. It has been reported that Sp1 is phosphorylated on Thr 453 and Thr 739 residues by Erk1/2 in SL2 Drosophila cells, and Sp1 has been suggested as a key molecular link between activation of RAS and vascular endothelial growth factor transcription in tumor cells (59). Therefore, we suggest p21 Sdi1 as a strong candidate of the Sp1-regulated gene downstream of PKC␣ and SA-pErk1/2 in the cellular senescence process.
In conclusion, we presented a novel pathway of p21 Sdi1 that was induced in a senescence process by Sp1, which was phosphorylated on Ser 59 by SA-pErk1/2 in response to the ROS signal and the constitutive activation of RAS. Our data may also explain an interesting report that hepatocellular carcinoma patients with activation of RAS/MAPK pathway, assessed by positive staining for pErk1/2, have a longer time to cancer progression (178 versus 46 days) (66), which strongly suggests that patients with higher pErk1/2 may express more p21 Sdi1 , thus inducing growth arrest and senescence phenotypes in tumor tissue. Elevated p21 Sdi1 induces cell-cycle arrest of the actively growing cells at G 1 phase, leading to growth arrest of senescent cells. On the other hand, down-regulation of PKC␣ expression by treatment of old cells with PKC␣-siRNA significantly reduces SA-pErk1/2 in both nuclear and cytoplasmic fractions of the old cells, and inhibits Sp1 phosphorylation on Ser 59 , but not Thr 739 , and transcription of p21 Sdi1 , resulting in the release of senescent cells from G 1 arrest. In summary, in vivo regulation of p21 Sdi1 expression occurs via Sp1 phosphorylation on Ser 59 by pErk1/2 downstream of PKC␣ in the process of cellular senescence.