MSK1 and JNKs Mediate Phosphorylation of STAT3 in UVA-irradiated Mouse Epidermal JB6 Cells*

Phosphorylation of Tyr 705 and Ser 727 of signal transducer and activator of transcription 3 (STAT3) are known to be required for maximal activation by diverse stimuli. Tyr 705 phosphorylation is generally accepted to be mediated by the Janus kinase family. But the mech-anism for STAT3 (Ser 727 ) phosphorylation is not well understood. Here, we provide evidence that UVA-in-duced phosphorylation of STAT3 at Ser 727 is inhibited by pretreatment of JB6 cells with PD98059 or SB202190. Phosphorylation of STAT3 (Ser 727 ) is also markedly prevented by a dominant negative mutant of ERK2, c-Jun N-terminal kinase 1 (JNK1), or p38 kinase and in knockout Jnk1 (cid:1) / (cid:1) or Jnk2 (cid:1) / (cid:1) cells. Furthermore, STAT3 (Ser 727 ) phosphorylation is suppressed by C- or N-termi-nal “kinase-dead” mutants of mitogen- and stress-acti-vated protein kinase 1 (MSK1), a downstream kinase of ERKs and p38 kinase, and H89, a potential MSK1 inhibitor. In vitro experiments showed that active MSK1 and JNKs, but not ERKs or p38 kinase, phosphorylate STAT3 (Ser 727 ). Additionally, the role of MAPKs in mediating UVA-stimulated DNA binding activity of STAT3 was investigated. Overall, these results suggest

Signal transducer and activator of transcription 3 (STAT3) 1 was identified as a latent transcription factor that transduces signals to the nucleus and activates expression of many genes in response to cytokines and growth factors (1)(2)(3)(4). STAT3 activation is generally accepted to occur by cytokine-stimulated Tyr 705 phosphorylation (3,5,6). Tyr 705 phosphorylation of STAT3 by Janus kinases, receptors with an intrinsic tyrosine kinase activity (e.g. epidermal growth factor receptor (EGFR)), or nonreceptor tyrosine kinases (e.g. Src or Abl) (5)(6)(7)(8) appeared to be all that was required for its activation. However, other earlier studies showed that cytokine-stimulated STAT3 signaling activation required a secondary phosphorylation modification at the serine/threonine residues, possibly by H7-sensitive kinase (9) or the extracellular signal-regulated kinases (ERKs) (10). Later, a serine phosphorylation of STAT3, as well as STAT1, 4 and 5, was found to be induced in a stimulus-related manner (4). Further, phosphorylation of STAT3 at Ser 727 , like Tyr 705 , was shown to be essential for its full activation (11)(12)(13)(14). Additionally, STAT3 signaling activation by environmental stresses, such as heat and osmotic shock, short wave UV light, free radicals, or hypoxia, was also shown to involve induction of multiple signaling pathways (15)(16)(17). These findings, therefore, indicate that STAT3 may be a convergent point for integrating signals from multiple pathways (18). More interestingly, only Ser 727 phosphorylation of STAT3 was observed to be stimulated by insulin, anisomycin, tumor necrosis factor-␣, or arsenite and, to a weaker extent, by NaCl, okadaic acid, or lipopolysaccharide (16,19,20). In contrast, Tyr 705 phosphorylation was not detected in cells with these treatments (16,19,20). These observations indicate that Ser 727 phosphorylation may play an important role in mediating the activation of STAT3 independently of Tyr 705 phosphorylation. Moreover, Ser 727 lies within a potential mitogen-activated protein kinase (MAPK) consensus motif of the C-terminal domain (11,12,21). Ser 727 phosphorylation is, thereby, postulated to occur through activation of MAPKs, including ERKs, c-Jun N-terminal kinases (JNKs), and p38 kinase (4,22). However, identification of the kinases responsible for Ser 727 phosphorylation has not been fully resolved (4,23). Therefore, identifying the STAT3 (Ser 727 ) kinases will provide a clearer understanding of the mechanisms of STAT3 activation. Activation of STAT3 signaling was shown to be involved in regulation of diverse cell processes, including growth, differentiation, proliferation, transformation, as well as apoptosis (24 -26). Furthermore, STAT3 was confirmed to be an oncogene (27), and its mediated signaling pathways were initiated by carcinogens (28 -30). Solar UV irradiation is believed to be one of the most important skin carcinogens (31,32). The UV components of sunlight reaching the surface of the earth are UVB (290 -320 nm) and UVA (320 -400 nm). UVC (200 -290 nm) is completely absorbed by the ozone layer of the atmosphere of the earth (31) and, therefore, is unlikely to have major pathophysiological effects. UVB is also partially absorbed by the ozone layer, and efficient protection can be provided by using sunscreen (31,32). UVA thus constitutes more than 90% of solar UV and is a major contributor to carcinogenesis (31)(32)(33). To date, an array of signaling pathways are known to be activated in UV-induced carcinogenesis (34,35). Recently, STAT1 was reported to be activated through tyrosine phosphorylation in response to UVA (36). However, whether STAT3 signaling in oncogenesis is stimulated by UVA is not clear. In the UVB/UVC response, activation of STAT3 signaling was triggered by phosphorylation at Ser 727 but not at Tyr 705 (16,17,23). Moreover, Ser 727 phosphorylation was also shown to be critical for constitutive or aberrant activation of STAT3 signaling in tumorigen-esis (4,28,30,37). To facilitate an understanding of the role for the STAT3 signaling pathway in UVA-induced skin carcinogenesis, the serine/threonine kinase pathways through which Ser 727 in STAT3 is phosphorylated in UVA-irradiated epidermal JB6 cells were investigated. We provide evidence that UVA-induced Ser 727 phosphorylation of STAT3 may occur through JNKs and mitogen-and stress-activated protein kinase 1 (MSK1), a downstream kinase of both ERKs and p38 kinase.
Analysis of Phosphorylated Proteins by Western Blotting-After starvation, experimental cells were or were not pretreated for 1 h with AG1478, PD135035 (Calbiochem, Inc., San Diego, CA), PD98059, SB202190, rapamycin (Sigma), or H89 (Alexis Biochemicals, Inc., San Diego, CA) at the doses indicated in the figures and then irradiated with UVA, UVB, or UVC as described previously (42,43,46). Here, the nonirradiated cells in the UV box were used as negative controls. The treated cell lysates were harvested in SDS sample buffer and resolved by 8% SDS-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes followed by Western blotting analysis with specific antibodies against phospho-STAT3 (Ser 727 ), phospho-STAT3 (Tyr 705 ), phospho-MSK1 (Ser 376 ), STAT3 (Cell Signaling, Inc., Beverly, MA), or MSK1 (Upstate Biotechnology, Inc., Lake Placid, NY). For a detailed description of the assay see Zhang et al. (42,43). Additionally, the intensity of the bands in some Western blots was calculated using the Image-Quant Microsoft system or the recommended system (www.totallab.com).
In Vitro Assay for STAT3 Phosphorylation by Protein Kinases-After starvation for 48 h in 0.1% FBS with minimum essential medium, JB6 Cl 41 cells were disrupted in 300 l of lysis buffer A containing 20 mM Tris, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% (v/v) Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM ␤-glycerol phosphate, 1 mM phenylmethylsulfonyl fluoride, 1 mM Na 3 VO 4 , 1 g/ml leupeptin, and 10 g/ml aprotinin. The cell lysates were clarified by centrifuging at 17,000 ϫ g at 4°C for 5 min, and then equal amounts of protein were subjected to immunoprecipitation with a STAT3 antibody (42). The immune complex containing STAT3 proteins was incubated at 30°C for 60 min with active ERK1, ERK2, JNK1, JNK2, p38 kinase, or MSK1 (Upstate Biotechnology) in kinase buffer (50 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 1 mM EGTA, 1 mM dithiothreitol, 5 mM ATP, and 0.01% Brij 35) (Cell Signaling) after washing twice with lysis buffer A and twice with kinase buffer. Subsequently, Ser 727 phosphorylation of STAT3 was detected by Western blot analysis as described previously (42). Nonphosphorylated STAT3 was used as an internal control to verify equal protein loading. To further evaluate STAT3 phosphorylation catalyzed by MSK1, the immunoprecipitated STAT3 from nonirradiated JB6 cell lysates was incubated with active MSK1 or JNK1 at the indicated doses in the above kinase buffer containing 1 Ci of [␥-32 P]ATP. After quantifying by scintillation counting, radioactive phosphate incorporated into immunoprecipitated STAT3 was calculated (www.upstatebiotech.com). In addition, to assess the specificity of the phospho-STAT3 (Ser 727 ) antibody, UVA-irradiated cell lysates were subjected to immunoprecipitation with a phospho-STAT3 (Ser 727 ) antibody preincubated with a 5-fold concentration of bovine serum albumin (BSA) or STAT3 blocking peptides containing phospho-specific Ser 727 (sc-8001P) or Tyr 705 (sc-7993P), a C terminus with no Ser 727 or Tyr 705 sites (sc-482P), or an internal domain (sc-483P) of STAT3 p92 of mouse origin (Santa Cruz) for 1 h at room temperature. Then phosphorylated STAT3 p92 was visualized using Western blots with the phospho-STAT3 (Ser 727 ) antibody.
MSK1 Activity Assay-The immune complex MSK1 activity assay with the Akt/SGK peptide as a substrate was employed according to the recommended procedure of Upstate Biotechnology, Inc. (www.upstatebiotech.com) (44). In brief, treated or untreated cell lysates were subjected to immunoprecipitation with a MSK1 antibody as reported (42). Then the immune complex containing MSK1 proteins was incubated with agitation for 15 min at 30°C in a mixture of the following: 20 l of assay dilution buffer (20 mM MOPS, pH 7.2, 25 mM ␤-glycerol phosphate, 5 mM EGTA, 1 mM Na 3 VO 4 , and 1 mM dithiothreitol), 10 l of the Akt/SGK substrate peptide in assay dilution buffer, and 10 l of [␥-32 P] ATP solution (1 Ci/l, diluted in 75 mM MgCl 2 and 500 M unlabeled ATP). To stop the reaction, the samples were spotted onto a numbered P81 paper square and washed three times (5 min each) with 0.75% phosphoric acid and once (3 min) with acetone. Each sample paper was transferred into a scintillation vial containing 5 ml of scintillation fluid and counted in a ␤-scintillation counter.
STAT3 DNA Binding Mobility Shift Assay-Electrophoretic mobility shift assays were used to detect STAT3 DNA binding activity after exposure of the cells to UVA irradiation (16,36). Briefly, 1 h after UVA irradiation, the cells were disrupted in 500 l of lysis buffer B (50 mM KCl, 0.5% Nonidet P-40, 25 mM HEPES, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 20 g/ml aprotinin, and 100 M dithiothreitol). After centrifugation at 17,000 ϫ g for 1 min, the nuclear pellets were washed with the same buffer without Nonidet P-40 and then resuspended in 300 l of extraction buffer (500 mM KCl and 10% glycerol with the same concentrations of HEPES, phenylmethylsulfonyl fluoride, leupeptin, aprotinin, and dithiothreitol as the buffer B). After an additional centrifugation for 5 min, the supernatant fractions were harvested as nuclear protein extracts and stored at Ϫ70°C. The STAT3-binding oligonucleotide (5Ј-GATCCTTCTGGGAATTCCTA-GATC-3Ј) probe from Santa Cruz Biotechnology (Santa Cruz, CA) (66) was labeled with [␥-32 P]ATP using the Klenow fragment (Life Science Co., Gaithersburg, MD). The nuclear extract (3 g) was added into the DNA binding buffer containing 5 ϫ 10 4 cpm 32 P-labeled probe, 1.5 g of poly(dI⅐dC), and 3 g of BSA. The reaction mixture was incubated on ice for 10 min followed by an additional incubation at room temperature for 20 min. Then the DNA-protein complexes were resolved on a 6% nondenaturing polyacrylamide gel. For specific competition, the nuclear extracts were preincubated for 20 min with 2 g of unlabeled cold probe or for 4 h with 0.2 g of STAT3 antibody before addition of the labeled STAT3-binding probe. After autoradiography, the radioactivity was quantified using the Image Quant software (Molecular Dynamics, Sunnyvale, CA).

RESULTS
Induction of Ser 727 Phosphorylation of STAT3 by UVA Irradiation-Ser 727 phosphorylation of STAT3 was shown to be required for its maximal activation by cytokine stimulation (11,12,14,47). A recent report indicated that STAT1, which is highly homologous to STAT3, was activated by UVA irradiation (36). But whether UVA stimulates STAT3 signaling through Ser 727 phosphorylation is as yet unknown. This question was investigated here by using Western blot analysis with a specific antibody to detect phosphorylation of STAT3 at Ser 727 . After epidermal JB6 Cl 41 cells were exposed to UVA irradiation, Ser 727 phosphorylation was induced in a dose-dependent ( Fig.  1A) and time-dependent (Fig. 1B) manner. Induction of Ser 727 phosphorylation occurred 15 min after irradiation with UVA at 160 kJ/m 2 (Fig. 1B), increased to maximal induction at 30 min ( Fig. 1), and then gradually decreased to basal level by 120 min, and a second peak of induction recurred at 360 min following irradiation (Fig. 1C). In contrast, a weaker Tyr 705 phosphorylation induced by UVA was also observed in the same experiments ( Fig. 1, A and B). Tyr 705 phosphorylation was induced at 5 min following UVA irradiation and then decreased to basal level at 30 min after irradiation (Fig. 1B). In addition, the phosphorylation at Ser 727 (Fig. 1A), but not at Tyr 705 (Fig. 1A), was detected in JB6 cells irradiated with UVB or UVC, consistent with previous reports (16,17,23,48). These findings, therefore, suggest that changes in Ser 727 phosphorylation of STAT3 may indirectly reflect regulation of STAT3 signaling.
Requirement of EGFR for UVA-induced Ser 727 Phosphorylation of STAT3-EGFR-mediated activation of STAT3, but not STAT1, has been shown to be an early event in carcinogenesis (30, 49 -51). Furthermore, UVA, like UVB or UVC, was reported to initiate several potential EGFR signaling pathways including phosphatidylinositol 3-kinase and Ras/MAPKs (30,52,53). However, although Janus kinase-independent Tyr 705 phosphorylation is known to be mediated by EGFR, which has an intrinsic tyrosine kinase activity (7), whether Ser 727 phosphorylation of STAT3 by serine/threonine kinases is EGFR-dependent is unclear. To determine this, we used EGFR-deficient (Egfr Ϫ/Ϫ ) cells that were characterized and identified in our laboratory 3 and two EGFR-specific tyrosine kinase inhibitors, AG1478 and PD153035 (52,54,55). Our data demonstrated that a dose-dependent Ser 727 phosphorylation stimulated by UVA irradiation was markedly prevented in Egfr Ϫ/Ϫ cells compared with wild-type Egfr ϩ/ϩ control cells ( Fig. 2A). Furthermore, UVA-induced Ser 727 phosphorylation was inhibited by pretreatment of JB6 cells with either AG1478 or PD153035 (Fig. 2B), but total levels of STAT3 were unchanged in Egfr Ϫ/Ϫ cells or after pretreatment with EGFR inhibitors (Fig. 2). On the other hand, the phosphorylation of STAT3 at Ser 727 was not totally blocked by deficiency of EGFR or inhibition of EGFR kinase (Fig. 2). Therefore, these results indicate that EGFRdependent and -independent signaling may be involved in Ser 727 phosphorylation of STAT3 in response to UVA. In addition, AG1478 and PD153035 had no marked effect on Tyr 705 phosphorylation levels (Fig. 1B), based on the observation that the UVA-stimulated increase of Tyr 705 phosphorylation did not occur 30 min after irradiation of JB6 cells (Fig. 1, A and B). However, a weaker Tyr 705 phosphorylation was induced in UVA-irradiated Egfr ϩ/ϩ cells, but no induction was observed in corresponding control Egfr Ϫ/Ϫ cells ( Fig. 2A), suggesting that EGFR may be required for UVA-induced Tyr 705 phosphorylation.
Involvement of MAPKs in UVA-induced Ser 727 Phosphorylation of STAT3-In the UVA response, activation of Ras/ERKs cascades is triggered by EGFR-dependent signaling, but JNKs and p38 kinase activation is not. 4 EGFR signaling to STAT3 and subsequent phosphorylation of STAT3 at Ser 727 is proposed to be mediated by MAPKs, based on the fact that Ser 727 is located in a conserved Pro-Met-Ser-Pro motif of the transcriptional activation domain (11,12). However, a role of MAPKs in mediating STAT3 is not well elucidated. To assess the role, we first treated JB6 cells with PD98059, an inhibitor of MEK1, which is a kinase that leads to ERKs activation (56), or with SB202190, a specific inhibitor of p38 kinase (57,58). The results from these experiments showed that UVA-induced Ser 727 phosphorylation of STAT3 was markedly suppressed by pretreatment with either PD98059 (Fig. 3A) or SB202190 (Fig.  3B), compared with treatment with UVA only. In addition, induction of STAT3 (Ser 727 ) phosphorylation after exposure of JB6 cells to UVA irradiation was also partially impaired by pretreatment with rapamycin, an inhibitor of mTOR (Fig. 3C), 4 Y. Zhang and Z. Dong, unpublished data. which was used as an internal control. These observations suggest that besides mTOR (59), Ser 727 phosphorylation and maximal activation of STAT3 may be mediated in vivo by ERKs or p38 kinase. However, LY294002 pretreatment had no effect on Ser 727 phosphorylation (Fig. 3D), indicating that Ser 727 phosphorylation of STAT3 occurred through phosphatidylinositol 3-kinase-independent pathways.
STAT3 protein was immunoprecipitated from 24-h starved JB6 cells and used as a substrate of kinases for in vitro experiments. The kinase reactions showed that JNK1 and JNK2 (Fig. 5B), but not ERK1, ERK2, or p38 kinase (Fig. 5A), phosphorylated Ser 727 of the STAT3 protein. Additionally, the pu-rified JNK1, JNK2, ERK1, ERK2, and p38 kinase were confirmed to be active as previously reported (Ref. 43 and data not shown). Overall, these in vivo and in vitro results indicate that after exposure of JB6 cells to UVA irradiation, JNKs may directly mediate Ser 727 phosphorylation of STAT3, whereas ERKs and p38 kinase appear to regulate the process indirectly through a serine/threonine kinase. Additionally, a phospho-STAT3 (Ser 727 ) antibody was preincubated with BSA or STAT3 blocking peptides and subjected to immunoprecipitation to examine specificity of the antibody. The results in Fig. 5C showed that Ser 727 phosphorylation of STAT3 immunoprecipitated from UVA-irradiated JB6 cell lysates was suppressed by preincubation with a Ser 727 phospho-specific blocking peptide (sc-8001P) but not by a Tyr 705 phospho-specific peptide (sc-7993P), a C terminus with no Ser 727 or Tyr 705 residues (sc-482P), an internal domain (SC-483P) of STAT3 p92 of mouse origin or by BSA. These data suggest that the phospho-STAT3 (Ser 727 ) antibody specifically recognizes phosphorylated STAT3 (Ser 727 ).

MSK1 Phosphorylation of STAT3 at Ser 727 Both in Vivo and in Vitro-
To determine an ERK-or p38 kinase-dependent serine/threonine kinase for Ser 727 phosphorylation of STAT3, we employed additional in vitro kinase experiments using active MSK1 (44). The results demonstrated that immunoprecipitated STAT3 proteins were phosphorylated at Ser 727 by active MSK1 in vitro (Fig. 5B). To further evaluate STAT3 phosphorylation by MSK1 in vitro, we determined catalytic kinetics of MSK1 and JNK1 in kinase assays by using [␥-32 P]ATP and immunoprecipitated STAT3 protein. The results in Fig. 5D showed that MSK1 was more effective than JNK1 for catalysis of STAT3 phosphorylation. To examine the role of MSK1 in UVA-induced Ser 727 phosphorylation, we prepared JB6 cell lines stably expressing an N-terminal or C-terminal "kinasedead" mutant (Nd or Cd) of MSK1, as well as wild-type MSK1 (MSK1-WT) (44). As expected, the UVA-stimulated increase in MSK1 activity was significantly inhibited by overexpression of MSK1-Cd or -Nd (p Ͻ 0.01, Fig. 6A) compared with MSK1-WT or empty vector CMVS. Furthermore, MSK1-Cd or -Nd markedly prevented UVA-induced Ser 727 phosphorylation of STAT3 (Fig. 6B). Additionally, a dose-dependent inhibition of UVAstimulated Ser 727 phosphorylation of STAT3 (Fig. 6C) was observed following pretreatment of JB6 cells with H89, a potential MSK1 inhibitor (60,61). 2 These data suggest that the Ser 727 phosphorylation of STAT3 may be mediated by MSK1 in UVA-irradiated JB6 cells.
MSK1 phosphorylation was shown to reflect its activity indirectly (44,60). A recent report showed that Ser 376 located in the hydrophobic motif of the MSK1 linker region is a critical site for MSK1 activity (62). Here, Ser 376 phosphorylation of MSK1 was analyzed by Western blot analysis with a phosphospecific antibody to MSK1 (Ser 376 ). The data demonstrated FIG. 5. Induction of STAT3 phosphorylation at Ser 727 by protein kinases in vitro. The immunoprecipitated STAT3 proteins from 24-h starved JB6 Cl 41 cell lysates were incubated for 60 min at 30°C in kinase buffer with active purified ERK1, ERK2, or p38 kinase (A) or JNK1, JNK2, or MSK1 (B) at the described doses. Total and phosphorylated (Ser 727 ) STAT3 proteins were detected by Western blot analysis as described for Fig. 4. C, JB6 cells were or were not irradiated with UVA (160 kJ/m 2 ), and the cell lysates were subjected to immunoprecipitation with a phospho-STAT3 (Ser 727 ) antibody preincubated with a 5-fold concentration of BSA or STAT3 blocking peptides, and subsequent determination of STAT3 phosphorylation was performed as described under "Materials and Methods." D, immunoprecipitated (IP) STAT3 from nonirradiated JB6 cells was incubated with active MSK1 or JNK1 at the indicated doses in kinase buffer containing [␥-32 P]ATP. Then radioactive phosphate incorporated into immunoprecipitated STAT3 was calculated after counting in a ␤-scintillation counter. The results are representative of at least three independent experiments. p, phosphorylated; np, nonphosphorylated; IP, immunoprecipitated; mU, milliunits; Ctrl, control. . Nontransfected JB6 Cl 41 cells (C) were irradiated after pretreatment with H89 at the indicated doses. A, MSK1 activity in immunoprecipitated complexes was assayed as previously reported (44,60). After subtraction of background, UVA-induced MSK1 activity was normalized to nonirradiated negative control values and is represented as a fold change. Each column or bar indicates the mean and standard deviation from three independent experiments performed in duplicate. UVA-induced MSK1 activity was significantly inhibited (*, p Ͻ 0.01) in MSK1-Cd or -Nd cells when compared with CMVS or MSK1-WT control cells using Student's t test. B and C, total and phosphorylated STAT3 (Ser 727 ) proteins were detected by Western blot analysis as described for Fig. 4. The data are representative of at least three independent experiments. p, phosphorylated; np, nonphosphorylated.

STAT3 Ser 727 Phosphorylation by MSK1 and JNKs
that UVA-stimulated enhancement of Ser 376 phosphorylation of MSK1 was completely blocked by pretreament with PD98059 (Fig. 7A) and partially inhibited by SB202190 (Fig. 7B) compared with treatment with UVA only. Furthermore, similar inhibitory effects on Ser 376 phosphorylation of MSK1 were observed in DNM-ERK2 (Fig. 7C) and DNM-p38 (Fig. 7D) cells. These data indicate that Ser 376 phosphorylation of MSK1 occurs through ERKs and p38 kinase in the UVA response. Together, our results suggest that ERKs and p38 kinase may mediate Ser 727 phosphorylation of STAT3 indirectly through MSK1. In addition, DNM-JNK1 and Jnk1 Ϫ/Ϫ and Jnk2 Ϫ/Ϫ cells also had inhibitory effects on MSK1 (Ser 376 ) phosphorylation compared with corresponding control cells (data not shown). These data suggest that JNKs may be a potent upstream kinase of MSK1, but the suggestion remains to be investigated further.
Involvement of MAPK Cascades in Mediating UVA-stimulated DNA Binding Activity of STAT3-Induction of STAT1 DNA binding activity by UVA was reported by Maziere et al. (36). Here, to characterize UVA-induced STAT3 DNA binding activity, nuclear extracts from UVA-irradiated JB6 cells expressing MSK1-WT or CMVS were subjected to electrophoretic mobility shift assay with a STAT3-specific high affinity consensus-binding site oligonucleotide probe (66). A marked stimulation of STAT3 DNA binding activity was observed 1 h after UVA irradiation (Fig. 8A), compared with that in sham-irradiated controls. Competition analysis showed that STAT3 binding to the labeled oligonucleotide was completely blocked by preincubation with an excess of unlabeled probe (Fig. 8A). Furthermore, prior incubation of UVA-treated nuclear extracts with a STAT3 antibody induced a weaker supershift but mark-edly inhibited the binding complex (Fig. 8A) compared with internal controls using normal serum. These data indicate that DNA binding activity of STAT3 is stimulated after irradiation of JB6 cells with UVA.
The effect of Ser 727 phosphorylation on DNA binding activity of STAT3 is controversial (4). A study showed that blocking Ser 727 phosphorylation by mutation of STAT3 Ser 727 with Ala 727 abolished its transcriptional activity but had no effect on DNA binding activity of STAT3 (79). However, the possibility exists that regulation of the Ser 727 phosphorylation is involved in this process. For example, Chung et al. (69) showed that the Ser 727 phosphorylation negatively regulated DNA binding activity of STAT3. Conversely, the phosphorylation of Ser 727 , besides Tyr 705 , was confirmed to contribute to maximal activation of STAT3 (11,63). Additional studies also showed that Ser 727 phosphorylation corresponded to an increased DNA binding activity of STAT3 (12,64,65,78). To determine the role of MAPK cascades for Ser 727 phosphorylation in regulating STAT3 DNA binding activity, we analyzed the UVA-induced binding complex in JB6 cell lines expressing MSK1-Cd or -Nd, DNM-ERK2, DNM-JNK1, or DNM-p38 kinase. The results demonstrated that UVA-stimulated enhancement of STAT3 DNA binding activity was significantly abrogated by MSK1 kinase-dead mutants (Fig. 8B). Furthermore, the enhancement was also markedly suppressed in DNM-ERK2, DNM-JNK1, or DNM-p38 kinase cells (Fig. 8C) compared with that observed in control cells. Together with previous reports (12,64,65,78), our data suggest that UVA-induced Ser 727 phosphorylation of STAT3 may be mediated by ERK-and p38 kinase-dependent MSK1, as well as JNKs, and may be involved in functional regulation of the DNA binding activity of STAT3, but whether serine/threonine sites other than Ser 727 are involved in this process is not known. DISCUSSION STAT3 is confirmed to be an oncogene that plays a key signaling role in neoplastic transformation (27). Activation of STAT3 signaling has been increasingly associated with malignant progression (28). Furthermore, STAT3 activation requires phosphorylation of both Tyr 705 and Ser 727 in response to stimulation by cytokines and growth factors (4,11,63). But, in the UVC or UVB response, only Ser 727 phosphorylation of STAT3 was detected (16; 23). Induction of a strong Ser 727 phosphorylation during UVA irradiation was observed here, but Tyr 705 phosphorylation was weaker. These findings suggest that Ser 727 phosphorylation may result in STAT3 signaling activation independently of Tyr 705 phosphorylation in response to some environmental stimuli such as UV stress and thus contribute to UV-induced oncogenesis. This assertion is supported directly from the observation that a dominant negative STAT3 mutant with a change of Ser 727 to Ala 727 blocks phosphorylation and inhibits STAT3 signaling activation and Src transformation (27,28,67). Therefore, elucidating the signal transduction pathways leading to Ser 727 phosphorylation of STAT3 will help in understanding the molecular mechanisms involved in activation of STAT3 during UVA-induced carcinogenesis. In addition, this knowledge may lead to the development of novel preventive and therapeutic approaches to intervene in the process. However, the identity of the kinases that are responsible for catalyzing Ser 727 phosphorylation of STAT3 has been elusive (23,68). In contrast, Tyr 705 phosphorylation of STAT3 is a well characterized event known to be mediated by Janus kinase family kinases and receptor or nonreceptor tyrosine kinases (3,(5)(6)(7). In this report, evidence is provided showing that UVA-stimulated Ser 727 phosphorylation of STAT3 and its DNA binding activity may be mediated by ERK-and p38 kinasedependent MSK1, as well as JNKs. The UVA-induced Ser 727 phosphorylation of STAT3 was initiated by EGFR-dependent and -independent signaling pathways. Additionally, a model of the possible signal transduction pathways involved in Ser 727 phosphorylation of STAT3 is presented in Fig. 9.
Because the site of Ser 727 phosphorylation, -Pro-Met-Ser-Pro-, present in both STAT1 and STAT3 agrees with the MAPK action consensus sequence, -Pro-Xaa-(Ser/Thr)-Pro- (22), MAPK family members, including ERKs, JNKs, and p38 kinase, are hypothesized to mediate Ser 727 phosphorylation of STAT3 and regulate its activity in vivo. As expected, in 3T3 cell lysates stimulated with fetal calf serum or epidermal growth factor, a C-terminal peptide of STAT3 was reported to be phosphorylated at Ser 727 by immune complexes of ERK1 but not p38 kinase or JNK1 (69). Unexpectedly, the homologous Cterminal peptide of STAT1 was a relatively poor substrate for all MAPKs tested both in vitro and in vivo (69). However, our in vitro kinase experiments showed that the Ser 727 phosphorylation in intact STAT3 immunoprecipitated from quiescent JB6 cell lysates was induced by purified active JNKs but not by ERKs or p38 kinase. This finding agrees with the assumption of Lim and Cao (16) that was drawn from experiments using glutathione S-transferase-STAT3 fusion protein, containing an almost full-length STAT3. The discrepancies of these reactions may be related to their docking interactions contributing to regulation of the specificity and efficiency of the enzymatic reactions with substrates (70,71). In fact, the specificity of STAT signaling dependent on the SH2 and C-terminal domains was recently confirmed by the experimental evidence of Kovarik et al. (23). Additionally, the occurrences of these reactions may also vary with the molecular conformations in different cell contexts. For example, STAT3 has recently been shown to form stable homodimers independently of Tyr 705 phosphorylation (72), although STAT3 was thought to exist as monomers in the cytoplasm prior to its activation. Therefore, JNKs may play a direct role in regulation of Ser 727 phosphorylation of STAT3, whereas the regulation by ERKs and p38 kinase may be mediated indirectly through a serine/threonine kinase (e.g. MSK1).
In fact, studies of STAT1 showed that activation of the p38 kinase pathway was essential for Ser 727 phosphorylation and subsequent activation of STAT1 and indicated an indirect role for p38 kinase in this process (17,23,68,73). Although most studies on STAT3 Ser 727 phosphorylation revealed that regulation occurred through activation of ERK-dependent and -independent pathways (65,69,74), evidence was not provided as to whether ERKs is a direct Ser 727 kinase of STAT3. Here, however, we provide evidence that prevention of JNKs activation by DNM-JNK1 and Jnk1 or Jnk2 deficiency resulted in inhibition of Ser 727 phosphorylation of STAT3 and its DNA binding activity and that Ser 727 phosphorylation was induced by active JNKs in vitro, consistent with previous results (16,75). Thus, JNKs may be a candidate for the direct STAT3 Ser 727 kinase.
More interestingly, evidence was provided in this report that blocking MSK1 activation by H89, a potential MSK1 inhibitor (60), 2 or a N-terminal or C-terminal kinase-dead mutant of MSK1 (44) 2 resulted in inhibition of UVA-stimulated STAT3 Ser 727 phosphorylation and its DNA binding activity, indicating that MSK1 may play a role in the process. Furthermore, activation of ERKs and p38 kinase pathways were required for induction of Ser 727 phosphorylation of STAT3 during stimula- tion with UVA. This was supported by the findings that inhibition of UVA-activated ERKs by PD98059 and DNM-ERK2 and of p38 kinase by SB202190 and DNM-p38 blocked MSK1 phosphorylation and subsequently suppressed STAT3 Ser 727 phosphorylation and its DNA binding activity. Moreover, purified active MSK1 can induce phosphorylation at Ser 727 in vitro. At the same time, however, the in vitro induction of the Ser 727 phosphorylation by active ERK1, ERK2, or p38 kinase was undetectable. Taken together, these findings suggest that the regulation by ERKs and p38 kinase of UVA-induced STAT3 Ser 727 phosphorylation and its DNA binding activity may be mediated through MSK1, a direct downstream serine/threonine kinase dependent on both ERKs and p38 kinase. Therefore, MSK1 is likely to be considered as a potential STAT3 Ser 727 kinase.
However, the possibility of phosphorylation at other serine/ threonine sites cannot be ruled out, because mutation of STAT3 Ser 727 with Ala 727 blocked phosphorylation and transcriptional activity of STAT3 but did not produce any effect on its DNA binding activity (4,11). In addition, a relative weaker induction of serine phosphorylation of STAT3 by p38 kinase was shown with radioactive labeled phosphate (75) but was not detected with specific phospho-STAT3 Ser 727 antibody in our experiments or in those of Lim and Cao (16). Therefore, the action of MSK1 on STAT3 phosphorylation was postulated to require a prior phosphorylation at other serine/threonine sites that induce a change in its conformation. In fact, a recent study suggests that phosphorylation at other serine/threonine sites may occur by activation of the p38 kinase pathway and play a role in p38 kinase regulation of STAT3 signaling activation (76), because the p38 kinase pathway is required for interferon ␣-dependent activation of STATs but not for their Ser 727 phosphorylation.
In summary, following integrating signals from multiple signaling pathways (including MAPKs and others) induced concurrently by UV stimulation, STAT3 was phosphorylated at Ser 727 and/or other serine/threonine sites, and subsequently STAT3-mediated signaling pathways were activated to regulate expression of target genes. In the UVA response, Ser 727 phosphorylation of STAT3 is mediated by JNKs and MSK1, which is dependent upon ERKs and p38 kinase both in vivo and in vitro. The phosphorylation of STAT3 at Ser 727 and possibly other serine/threonine sites by MAPKs cascades may play a role in regulating STAT3 DNA binding activity and transcriptional activity as a consequence of integrating multiple signals. In addition, the Ser 727 and non-Ser 727 phosphorylation events may also require H7-sensitive kinase (9, 65), protein kinase C␦d (77), mTOR kinase (59), or other kinase pathways (37,78), and thus a goal of our future studies will be to provide a mechanistic explanation for our findings.