Differential Regulation of Mitogen- and Stress-activated Protein Kinase-1 and -2 (MSK1 and MSK2) by CK2 following UV Radiation

Cloning and Plasmid Constructions

Human MSK1 cDNA (gift from Dr. Peter Cheung) was PCR-amplified and cloned into the pcDNA3.1/V5-His mammalian expression vector by directional TOPO-cloning (Invitrogen). To obtain the full-length human MSK2 cDNA, the following nucleotide sequences were PCR-amplified from human cDNA expressed sequence tags (Open Biosystems). PCR-amplified nucleotides 1–1670 (IMAGE clone ID: 5216639) and nucleotides 1671–2218 (IMAGE clone ID: 2405246) were each cloned into TOPO-TA vectors (Invitrogen). PCR-amplified nucleotides 2219–2316 (IMAGE clone ID: 5763859) were cloned into the pcDNA3.1/V5-His mammalian expression vector by directional TOPO-cloning (2219–2316-pcDNA3.1/ V5-His). Enzymatic restriction fragments containing the MSK2 sequences 1–1670 and 1671–2218 were then ligated and subcloned into 2219–2316-pcDNA3.1/V5-His to generate the full-length MSK2-containing mammalian expression vector (MSK2-pcDNA3.1/V5-His) and into pGEX-4T3, yielding pGEX-4T3-MSK2. MSK1 and MSK2 point mutations were generated by QuikChange site-directed mutagenesis (Stratagene). The human CK2β cDNA (IMAGE clone ID: 8327487; Open Biosystems) was subcloned by enzymatic restriction digestion into HA_2-pcDNA3.1 (gift from Dr. Lea Harrington) and into pGEX-4T3 (GE Healthcare), yielding HA₂-pcDNA3.1-CK2β and pGEX-4T3-CK2β, respectively. For p65 cloning, RNA was harvested from cycling MDA-MB-231 cells using the RNeasy minikit from Qiagen and reverse transcribed to make cDNA using qScript cDNA Supermix (Quanta) from which p65 was PCR-amplified (primers: forward, 5′-ATGGACGAACTGTTCCCCCTCATC-3′; reverse, 5′-GGA GCT GAT CTG ACT CAG CAG GGC-3′) and subcloned into the pM vector (Clontech) in frame with the Gal4-DNA binding domain sequence (pGal4-p65). QuikChange site-directed mutagenesis (Stratagene) was performed to generate the p65-S276A mutant (pGal4-p65-S276A). All of the plasmid constructs were verified by sequence analysis.

Cell Culture and Transfections

All cell lines were grown at 37 °C in a humidified atmosphere containing 5% CO₂. HEK293T cells were cultured in Dulbecco's modified Eagle's medium and MDA-MB-231 cells in RPMI 1640 medium, both supplemented with 10% fetal bovine serum and 50 units of penicillin and 50 μg of streptomycin antibiotics per ml. Transient transfection of plasmids into 293T cells was performed using the Effectene transfection kit (Qiagen) according to the manufacturer's instructions. Transient transfection of plasmids into MDA-MB-231 cells was performed using the Lipofectamine^TM LTX Plus transfection reagent (Invitrogen) according to the manufacturer's instructions. All siRNA (non-targeting (NT) siRNA pool (D-001206-13), human CK2β siRNA pool (L-007679-00), human MSK1 siRNA pool (M-004665-02), and human MSK2 siRNA (J-004664-06) were purchased from Dharmacon) and esiRNA (non-targeting esiRNA and esiRNA against the 3′-untranslated region of MSK2 was generated according to standard protocols (17)) transfections were performed using Dharmafect1 reagent (Dharmacon) as per the manufacturer's instructions.

Antibodies

Commercial antibodies used in this study were purchased from R&D Systems (anti-MSK1 goat polyclonal and anti-pS196-MSK2 rabbit polyclonal), Santa Cruz Biotechnology, Inc. (Santa Cruz, CA) (anti-CK2α goat polyclonal, anti-CK2β rabbit polyclonal and mouse monoclonal, anti-α-tubulin mouse monoclonal, and anti-p65 mouse monoclonal), Abcam (anti-MSK2 rabbit polyclonal and anti-CK2α and anti-CK2α′ rabbit polyclonal), Oncogene (anti-α-tubulin mouse monoclonal), Millipore (anti-HA mouse monoclonal), Cell Signaling Technologies (anti-phospho-p65-Ser²⁷⁶ rabbit polyclonal), and Invitrogen (anti-V5 mouse monoclonal). Anti-V5-agarose affinity gel was purchased from Sigma.

Expression and Purification of Fusion Proteins

GST-CK2β and GST-MSK2 recombinant proteins were produced in Escherichia coli BL21(DE3)/pLysS (Novagen). The expression, extraction, and purification of GST fusion proteins were performed as previously described (18).

Preparation of Cell Extracts, Immunoprecipitation, Pull-downs, and Immunoblotting

For whole cell extract (WCE) preparation, cells were rinsed once in ice-cold phosphate-buffered saline and lysed in 50 mm Tris-HCl, pH 7.5, 100 mm NaCl, 1% Triton X-100, 0.5 mm dithiothreitol, 0.5 mm EDTA, 1× Complete mini protease inhibitor mixture (Roche Applied Science), 1 mm sodium orthovanadate, 40 mm β-glycerophosphate, and 50 mm sodium fluoride. For inhibition studies, cells were pretreated with DMSO, DMAT (10 μm), SB203580 (10 μm), or H89 (10 μm) (Sigma) for 2 h at 37 °C. For UV-C treatment, cultured cells were washed once with phosphate-buffered saline and then UV-irradiated (40 or 200 J/m², as indicated) in the presence of phosphate-buffered saline. Following UV irradiation, cells were placed again in culture medium and incubated at 37 °C for specified periods until cell harvest. Pretreatment with 10 μm anisomycin was performed for 20 min at 37 °C. λ-Protein phosphatase treatments were performed as previously described (19). For immunoprecipitations, clarified WCEs were incubated with 1 μg of the relevant antibody on ice for 60 min with occasional gentle agitation, followed by the addition of 30 μl of protein A-coupled Sepharose beads (ThermoFisher) and incubated for an additional 1 h at 4 °C, or WCEs were incubated with V5-agarose (Sigma) at 4 °C for 2 h. Immunoprecipitates were washed three times with lysis buffer, resuspended in SDS-PAGE sample buffer, and resolved by SDS-PAGE. Proteins were transferred to polyvinylidene difluoride membrane (Millipore) and immunoblotted as indicated. Detection was performed using SuperSignal enhanced chemiluminescence (Pierce). For pull-down assays, 1.5 mg of WCE was mixed at 4 °C for 2 h with glutathione-Sepharose beads bound with 1 μg of GST fusion protein. The complexes were then washed and analyzed for associated proteins by immunoblotting as described above.

In Vitro Protein Kinase Assays

MDA-MB-231 or HEK293T cells or those expressing ectopically expressing MSK2-V5 or MSK1-V5 were mock- or UV-irradiated with 200 J/m² UV-C. Thirty minutes post-treatment, cells were harvested as described above, WCEs were incubated with antibodies as indicated, and immunoprecipitates were washed three times with buffer containing 400 mm NaCl and then once with 50 mm Tris-HCl, pH 7.5, 0.1 mm EGTA, 10 mm β-mercaptoethanol. Immunoprecipitation kinase assays were performed with 10 mm MOPS, pH 7.2, 12.5 mm β-glycerol phosphate, 2.5 mm EGTA, 0.5 mm sodium orthovanadate, 0.5 mm dithiothreitol, 6.25 mm magnesium acetate, 62.5 μm cold ATP, 10 μCi of γ-[³²P]ATP, and 20 μm crosstide (GRPRTSSFAEG-KK) (Millipore) harboring two lysine residues (underlined) at the C terminus to facilitate binding to P81 paper. The reaction mixtures were incubated for exactly 10 min at 30 °C with constant shaking and then terminated by centrifugation and spotting of the supernatants onto P81 ion exchange paper (Millipore). P81 papers were washed three times for 5 min each with 0.75% phosphoric acid then once with acetone for 5 min. Incorporation of radiolabel was measured by scintillation counting. For in vitro CK2 kinase assays, 1 μg of purified recombinant GST or GST-MSK2 was incubated with 10 ng of CK2 (Upstate) in a reaction buffer containing 5 μCi of γ-[³²P]ATP for 20 min at 30 °C according to the manufacturer's instructions. The reaction was stopped by the addition of Laemmli sample buffer, and phosphorylated proteins were visualized by resolving the samples on SDS-PAGE followed by autoradiography.

Reporter Assays

Where indicated, MDA-MB-231 cells were transfected with scrambled esiRNA or MSK2 esiRNA, and 24 h later, cells were transfected with a Gal4-responsive firefly luciferase expression plasmid, pGL2–5xGal4 (gift from Dr. Peter Cheung), the pRL-TK Renilla luciferase normalizing control (Promega), and, as indicated, with pGal4, pGal4-p65, pGal4-p65-S276A, MSK2-V5, MSK2-S324A-V5, or “kinase-dead” MSK2-S196A-V5. 48 h later, cells were mock- or UV-C-irradiated (200 J/m²) in phosphate-buffered saline, medium was replaced, and cells were incubated for 8 h at 37 °C. Firefly and Renilla luciferase activities were measured using the Dual-Luciferase Assay System (Promega) as per the manufacturer's instructions. In all experiments, transfections were performed in triplicate, and the luciferase values were normalized to Renilla luciferase activity.

Quantitative Real-time PCR

Total RNA was extracted from MDA-MB-231 cells using the RNeasy minikit (Qiagen). 5 μg of total RNA was reverse transcribed to make cDNA using qScript cDNA SuperMix (Quanta Biosciences). Quantitative real time PCR was performed with PerfeCTa^TM SYBR® Green SuperMix (Quanta Biosciences) as per the manufacturer's instructions in a 7900HT fast real-time PCR system (Applied Biosystems). Reactions were performed in triplicate and normalized to TATA-box binding protein expression.

Cell Viability Assays

MDA-MB-231 cells expressing NT or MSK2 siRNAs were mock- or UV-C-irradiated (200 J/m²), and cell viability was assessed 24 h later using the colorimetric CellTiter 96® AQ_ueous assay (Promega) by measuring absorbance at 490 nm as per the manufacturer's instructions.

CK2 Physically Interacts with MSK2 in Human Cells

Both CK2 and MSK1/2 interact with and are dependent on p38 for their activation following UV-C radiation. Therefore, it is possible that CK2 also interacts with and regulates MSK1/2 activity during the UV response. To explore this question, we initially examined whether MSK1 or MSK2 associate with CK2β, because many CK2 substrates may be initially bound by the CK2β regulatory subunit (10). Purified recombinant GST-CK2β or GST immobilized on glutathione-Sepharose beads were mixed with whole cell extracts from human MDA-MB-231 cells in pull-down assays and immunoblotted for endogenous MSK1 and MSK2 (Fig. 1A). As demonstrated in Fig. 1A, GST-CK2β was found to specifically interact with endogenous MSK2, but no detectable interaction was observed with the closely related isoform MSK1. We also noted that UV radiation was associated with decreased MSK2-GST-CK2β interactions, and GST-CK2β appeared to preferentially interact with a faster migrating MSK2 species (Fig. 1A). Co-immunoprecipitation analyses also identified a basal association between MSK2-V5 and ectopically expressed HA₂-CK2β immunoprecipitated from HEK293T whole cell extracts but not between HA₂-CK2β and MSK1-V5 (Fig. 1B). The reverse co-immunoprecipitation also confirmed an association between endogenous MSK2 but not MSK1 and endogenous CK2β from MDA-MB-231 cell extracts (Fig. 1C). Furthermore, MSK2 co-immunoprecipitated with endogenous CK2α from MDA-MB-231 cell extracts under basal conditions, suggesting that MSK2 interacts with the CK2 holoenzyme. Interestingly, following UV-C radiation, we observed a diminished association between MSK2 and CK2, but the interaction with p38 was maintained (Fig. 1C). There was no detectable interaction between endogenous MSK1 and p38.

MSK2 Activity Is Regulated by CK2 following UV-C Radiation

Given that CK2 appeared to interact with MSK2, we next examined whether CK2 might regulate MSK2 activity following UV-C radiation. To do so, endogenous MSK2 or MSK1 as a negative control was immunoprecipitated from MDA-MB-231 whole cell extracts that had been pretreated with or without the CK2 or p38 inhibitors (DMAT and SB203580, respectively), followed by UV-C radiation, and examined in in vitro kinase assays with the peptide substrate, crosstide (Fig. 2A). Although chemical inhibition of p38 reduced both MSK2 and MSK1 UV-induced activation to base-line values as expected, only UV-induced MSK2 activation was affected by CK2 inhibition, with an approximately 3.5-fold reduction (Fig. 2A). When these experiments were repeated in HEK293T cells, similar results were obtained (supplemental Fig. S1). As another means to inhibit CK2 function, we depleted CK2β by siRNA (which impairs CK2 holoenzyme formation and activation (3, 9)) in HEK293T cells and found that this was associated with an approximately 2-fold reduction in MSK2-V5 UV-induced kinase activity but had no effect on MSK1-V5 activity (Fig. 2B). As an additional control, we examined MSK2-V5 activation following anisomycin treatment, which specifically activates the p38 signaling cascade and found that siRNA-mediated CK2β depletion also similarly diminished MSK2-V5 kinase activation (Fig. 2C).

Serine 324 Is Required for UV-induced MSK2 Kinase Activation

The UV-induced MSK2 kinase activity is known to be p38-dependent, and our data suggest that it is also CK2-dependent (Fig. 2). Furthermore, we found that recombinant purified GST-MSK2 was directly phosphorylated by CK2 in vitro (Fig. 3A). We also noted that UV radiation induced an upward MSK2 electrophoretic mobility shift that was abolished by λ-protein phosphatase treatment, suggesting that the shift is largely due to MSK2 UV-induced phosphorylation (supplemental Fig. S6). CK2 inhibition partially reduced the UV-induced MSK2 mobility shift, whereas no such effect was noted with MSK1 (Fig. 2A). In contrast, p38 inhibition abolished the UV-induced mobility shifts of both MSK2 and MSK1 (Fig. 2A). CK2 and p38 inhibition had no effect on the migration of MSK2 under basal conditions (Fig. 2A). Collectively, these data suggested that CK2-dependent phosphorylation of MSK2 may modulate UV-induced MSK2 activity.

To determine which amino acid residue(s) of MSK2 might be phosphorylated by CK2, we identified several sites possessing the consensus CK2 phosphorylation motif ((S/T)XX(E/D/pS), where pS represents phosphoserine) (20). Of these possible sites of CK2 phosphorylation, only Ser⁸⁷, Ser¹¹⁹, and Ser³²⁴ were found to be unique to MSK2 and not conserved in MSK1. Furthermore, of these residues, only Ser³²⁴ was found to possess an additional acidic residue at the n + 1 position, which is found in 75% of CK2 sites of phosphorylation (Fig. 3B) (4). Therefore, we initially examined whether Ser³²⁴ might be involved in the CK2-dependent regulation of MSK2 kinase activity. To more reliably examine this effect in the context of mammalian cells, we performed site-directed mutagenesis to substitute Ser³²⁴ to alanine (S324A) in MSK2-V5 and compared the UV-induced in vitro kinase activity of MSK2-S324A-V5 with that of wild-type (WT) MSK2-V5. In Fig. 3C, in vitro kinase assays of anti-V5 (MSK2) immunoprecipitates from HEK293T cell extracts harvested under basal conditions and following UV-C radiation demonstrate that the MSK2 mutant protein harboring the S324A substitution was severely impaired for UV-induced kinase activation.

To test whether the phosphorylation of Ser³²⁴ might be an important determinant of MSK2 UV-induced kinase activation, Ser³²⁴ was also substituted to a phospho-mimicking amino acid (aspartic acid) (S324D) and compared with WT MSK2-V5 in anti-V5 (MSK2) in vitro kinase assays. As shown in Fig. 3C, the phospho-mimicking MSK2 mutant did indeed restore UV-induced MSK2 kinase activity to levels that were comparable with that of WT MSK2-V5.

MSK2 is thought to undergo sequential activation in vivo, whereby p38/ERK first phosphorylates Thr⁵⁶⁸ within the C-terminal catalytic domain and Ser³⁴³ within the linker region (21). Based on mutational analyses, it has been further proposed that the activated C-terminal kinase domain then autophosphorylates Ser³⁶⁰ (within the linker region) and Ser¹⁹⁶ (within the activation loop of the N-terminal kinase domain), which is essential for MSK2 activation (21). Therefore, to monitor the effect of Ser³²⁴, which lies within the linker region, on MSK2 activation following UV-C radiation, we monitored cell extracts from HEK293T cells expressing WT, S324A or S324D MSK2-V5 proteins subjected to mock or UV-C radiation with phospho-specific Ser(P)¹⁹⁶ immunoblotting (Fig. 3D, pS196). We found that MSK2-S324A-V5 did not undergo appreciable autophosphorylation at Ser¹⁹⁶ following UV-C radiation, in contrast to WT MSK2-V5 and MSK2-S324D-V5 (Fig. 3D). Therefore, these results suggest that the phosphorylation of Ser³²⁴ is important for subsequent autophosphorylation at Ser¹⁹⁶ and activation of MSK2 following UV-C radiation in vivo.

A Phospho-mimicking Substitution at Ser³²⁴ Is Resistant to CK2 Inhibition

MSK2 kinase activation is dependent on CK2, and the substitution of Ser³²⁴ to a phospho-mimicking amino acid residue relieves MSK2 inhibition following UV-C radiation. Therefore, if the phosphorylation of Ser³²⁴ is CK2-dependent, then one would expect that the phospho-mimicking mutant MSK2-S324D would be unaffected by CK2 inhibition. To investigate this possibility further, we compared WT MSK2-V5 and MSK2-S324D-V5 UV-induced activation with or without CK2β-depletion by siRNA in in vitro kinase assays. As shown in Fig. 3E, the ability of WT MSK2-V5 immunoprecipitated from UV-treated cell extracts to phosphorylate crosstide was inhibited by siRNA-mediated CK2β depletion, whereas the kinase activity of MSK2-S324D-V5 was not appreciably affected. These results are consistent with the notion that the phosphorylation of MSK2 at Ser³²⁴ is dependent on CK2 in response to UV-C radiation.

MSK2 and MSK1 are closely related isoforms but appear to be differentially regulated by CK2. To further investigate this interesting observation, we noted that Ser³²⁴ at the equivalent position in other mammalian MSK2 proteins was highly conserved, but it was an aspartic acid residue in MSK1 proteins (Fig. 3B). Therefore, we substituted the human MSK1 aspartic acid residue (Asp³⁴¹), located at the equivalent position to MSK2-Ser³²⁴, to alanine and compared this mutant with wild-type MSK1 in in vitro kinase assays. As shown in Fig. 4, substitution of aspartic acid to alanine at amino acid residue 341 had no effect on MSK1 kinase activity, further suggesting that MSK1 and MSK2 are differentially regulated at this position.

MSK2 Is Required for the UV-induced Phosphorylation of p65 at Ser²⁷⁶ in Vivo

The mitogen- and stress-activated protein kinases have been shown to phosphorylate the p65 subunit of NF-κB at Ser²⁷⁶ in response to TNFα stimulation and following interleukin-1β treatment (11, 22). Therefore, we wondered whether p65-Ser²⁷⁶ might be phosphorylated by MSK1/2 following UV-C radiation in vivo. To address this question, we first examined whether the UV-induced phosphorylation of p65 at Ser²⁷⁶ in MDA-MB-231 cells (a cell line that has been commonly used in the examination of NF-κB activation (23)) was affected by the inhibition of MSK1/2 (H89), CK2 (DMAT), or p38 (SB203580). Indeed, the immunoblotting of such treated MDA-MB-231 cell extracts with a phospho-specific Ser(P)²⁷⁶ antibody demonstrated that the UV-induced phosphorylation of p65 at Ser²⁷⁶ was undetectable following MSK1/2 or p38 inhibition (Fig. 5A). In addition, CK2 inhibition led to a substantial reduction in p65-Ser²⁷⁶ phosphorylation in vivo (Fig. 5A).

Because H89 can inhibit both MSK1 and MSK2, we then examined MDA-MB-231 cells depleted of MSK1, MSK2, or both kinases using RNA interference. We found that MSK2 depletion reduced p65-Ser²⁷⁶ phosphorylation following UV-C radiation, whereas the depletion of MSK1 had no detectable effect (Fig. 5B). In addition, when we depleted endogenous MSK2 by esiRNA directed against the 3′-untranslated region and then reconstituted with WT MSK2-V5, MSK2-S324A-V5, or empty vector, we found that only WT MSK2-V5 could rescue the UV-induced phosphorylation of p65 at Ser²⁷⁶ (Fig. 5C). Collectively, our results suggest that Ser³²⁴ is important for MSK2 activation and for the in vivo phosphorylation of p65 at Ser²⁷⁶ following UV-C radiation.

UV-induced p65 Transactivation Is Dependent on MSK2 and p65-Ser²⁷⁶

To determine the effect of UV-C radiation on NF-κB activity in MDA-MB-231 cells, genes known to be modulated by NF-κB in response to UV radiation (24) were initially examined using quantitative real time PCR (Fig. 6A). We found that UV-C radiation was associated with an increase in the endogenous expression of all such genes examined (Fig. 6A). We then performed luciferase reporter assays whereby p65 was fused to the Gal4 DNA binding element (Gal4-p65), and transactivation was compared with the mutant Gal4-p65-S276A. As demonstrated in Fig. 6B, UV-C radiation was associated with an increase in p65-dependent transactivation capacity, which was virtually abolished when p65-Ser²⁷⁶ was substituted to alanine. To test whether this effect was dependent on MSK2, we depleted MDA-MB-231 cells of MSK2 using RNA interference and similarly did not detect UV-induced p65-dependent transcriptional activation (Fig. 6C). Furthermore, expression of WT MSK2-V5 in MDA-MB-231 cells depleted of endogenous MSK2 (by esiRNA directed against the 3′-untranslated region) restored UV-induced p65 transactivation potential, whereas MSK2-S324A-V5 was much less effective (Fig. 6D). The RNA interference-mediated depletion of MSK2 from MDA-MB-231 cells was also found to reduce cellular viability following UV-C radiation (Fig. 6E).

Supplemental Data