Cdk3-Mediated c-Jun Phosphorylation at Ser63 and Ser73 Enhances Cell Transformation

Abstract

c-Jun is a component of the activator protein-1 (AP-1) complex, which plays a crucial role in the regulation of gene expression, cell proliferation and cell transformation as well as cancer development. Herein, we found that Cdk3, but not Cdk2 or JNK, is a novel kinase of c-Jun induced by stimulation with growth factors such as epidermal growth factor (EGF). Cdk3 was shown to phosphorylate c-Jun at Ser63 and Ser73 in vitro and ex vivo. EGF-induced Cdk3 activation caused c-Jun phosphorylation at Ser63 and Ser73, resulting in increased AP-1 transactivation. Ectopic expression of Cdk3 resulted in anchorage-independent cell transformation of JB6 Cl41 cells induced by EGF and foci formation stimulated by constitutively active Ras (Ras^G12V), which was mediated by AP-1 in NIH3T3 cells. These results demonstrated that the Cdk3/c-Jun signaling axis plays an important role in EGF-stimulated cell proliferation and cell transformation.

Introduction

c-Jun was originally identified as the normal cellular counterpart of the viral Jun oncoprotein (v-Jun) encoded by an avian sarcoma virus (ASV17) (1). Exposure of cells to phorbol esters, growth factors and transforming oncogenes, such as Ras, led to an increase in c-Jun transcriptional activity (2). Subsequent studies identified serines 63 and 73 as major sites of c-Jun phosphorylation, and modification of these residues stimulated c-Jun transactivation (3–5). Initially, the prototypic mitogen-activated protein (MAP) kinase extracellular signal-regulated kinases (ERKs) were believed to modify these residues (4). However, the cloning of a new subfamily of MAP kinases, the c-Jun N-terminal kinases (JNKs) (6), or stress-activated protein kinases (SAPKs) (7), led to a reevaluation of this scenario (8). Cellular JNKs are activated by cytokines (e.g., TNF, IL-1) and by exposure to environmental stresses (e.g., osmotic stress, redox stress, radiation) (9). We found that when mouse skin was exposed to 12-O-tetradecanoylphorbol-13-acetate (TPA), JNK phosphorylation at Thr183/Tyr185 was not changed (10). TPA or EGF is known to induce ERKs phosphorylation and increase AP-1 activity along with malignant cell transformation (11). Although ERKs were reported to phosphorylate c-Jun at Ser63/73 (12), others found that purified JNKs, not ERKs, phosphorylated c-Jun at Ser63/73 (13). This suggested that another c-Jun kinase might exist, which is induced by growth factors.

Cyclin-dependent kinase3 (Cdk3) was originally classified as a cyclin-dependent kinase because of its high-sequence identity (76%) with both Cdc2 and Cdk2 (14). However, Cdk3 functions in the regulation of the G1/S cell cycle transition in a manner distinct from Cdk2 (15). Cdk3 participates in the G1/S progression, at least partially, by binding with E2F-1, E2F-2, or E2F-3 through DP-1 and enhancing their transcriptional activities, whereas Cdk2 phosphorylates protein substrates to promote S-phase entry (16). Transfection of a dominant-negative mutant of Cdk3 (Cdk3-DN) caused an accumulation of cells in G1 (15, 17). Cdk3 activity appears early in the G1-phase (18), peaks at mid G1 (17), and is required for entry into S-phase (15), resulting in increased proliferation as well as cell transformation (17). In addition, Cdk3 mediates phosphorylation of the retinoblastoma (Rb) protein (Ser 807/811) during the G0 phase (19), but phosphorylation of Rb by Cdk3 alone is insufficient to inactivate Rb to promote G1 entry, which requires cyclin C (Cyc C) participation (20). Cdk3-DN inhibited Rb phosphorylation (Ser 807/811), preventing G0 exit (19). Although Cdk3 is a key regulator in the control of the G1/S cell cycle transition, the role of Cdk3 in cell cycle regulation as well as proliferation and cell transformation is only partially elucidated. Furthermore, the downstream target of Cdk3 involved in proliferation and cell transformation has not been clearly identified.

Here, we show that c-Jun is a novel substrate of Cdk3 induced by stimulation with growth factors such as EGF. Cdk3 phosphorylates c-Jun at Ser63/73, and increases AP-1 activity resulting in enhanced anchorage-independent cell transformation. This report is the first to show that c-Jun is a novel substrate of Cdk3 induced by growth factors, which occurs independently of the JNKs signaling pathway.

Materials and Methods

Chemical reagents, including Tris, HCl, and SDS for molecular biology and buffer preparation, were purchased from Sigma-Aldrich (St. Louis, MO). The checkmate mammalian two-hybrid system was from Promega Co. (Madison, WI). Cell culture media and supplements were obtained from Life Technologies (Rockville, MD). Cdk3, Cdk2, and JNK1 active kinases were from Upstate Biotechnology, Inc. (Lake Placid, NY). Antibodies against His, Cdk3, phospho-c-Jun (Ser63/73), Cyc C, and JNKs were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Antibodies against phospho-c-Jun (Ser63), phospho-c-Jun (Ser73), c-Jun (mouse), β-actin, GST and phospho-JNKs were from Cell Signal Technology, Inc. (Beverly, MA). The QuikChange II Site-Directed Mutagenesis Kit was obtained from Stratagene, Inc. (La Jolla, CA) and Taq DNA polymerase from Qiagen, Inc. (Valencia, CA). JetPEI was purchased from Qbiogen, Inc. (Montreal, Quebec, Canada). G418 was from Biomol International, L.P. (Plymouth Meeting, PA) and EGF was purchased from BD Bioscience (Bedford, MA).

Cell Culture and Transfections

HEK293 (10% FBS-DMEM), SaoS-2 (15% FBS-McCoy’s 5A), and JB6 C141 mouse skin epidermal cells (5% FBS-MEM) were cultured with growth medium supplemented with antibiotics at 37°C and 5% CO₂. NIH3T3 cells were cultured with DMEM with 10% calf bovine serum (CBS) and antibiotics at 37°C and 5% CO₂. Cells were maintained by splitting at 80–90% confluence and media changed every 3 days. For transfection experiments, cells were split and the expression vector induced when cells were 50 to 60% confluent using jetPEI (Qbiogen, Inc.), following the manufacturer’s suggested protocol. For stable transfection, JB6 cells (5.0×10⁵) in 5% FBS-MEM were seeded in 100-mm culture dishes. After culturing at 37°C in 5% CO₂ for 16 h, the cells were transfected with 2 μg of pRcCMV-HA-Cdk3, pCMV-HA-Cdk3DN, or pcDNA3.1 (“mock”) using jetPEI. Cdk3, Cdk3-DN, and mock stably-transfected cells were obtained by selection for G418 resistance (400 μg/ml) and further confirmed by assessing Cdk3 activity and expression.

Construction of Expression Vectors

pCMV-HA-Cdk2 (pHA-Cdk2, Cdk2 in pCMV-Bam-neo), pCMV-HA-Cdk2-DN (pHA-Cdk2-DN, dominant-negative Cdk2 in pCMV-Bam-neo), pRcCMV-HA-Cdk3 (pHA-Cdk3, Cdk3 in pRcCMV), and pcDNA3-Cyc C were gifts from Dr. Barrett J. Rollins (Department of Medical Oncology, Harvard Medical School, Boston, Massachusetts) (19). The pHis6-tagged c-Jun was kindly provided by Dr. Dirk Bohmann (European Molecular Biology Laboratory, Heidelberg, Germany) (21). The c-Jun fragment was generated by PCR and subcloned into the pGEX-5X-1 vector (Amersham Biosciences Corp., Piscataway, NJ) at the BamHI/XhoI site to generate a glutathione S-transferase (GST)-c-Jun plasmid (pGST-c-Jun) (22, 23). The mutant GST-c-Jun plasmid was generated by the QuikChange II Site-Directed Mutagenesis Kit and c-Jun mutant primers (for Ser63 mutation, sense: 5′-CGACCTTCTATGACGATGCC-3′; antisense: 5′-GGGCATCGTCATAGAAGGTCG-3′; and for Ser73, sense: 5′-AGCGGACCTTATGGCTACAGT-3′; antisense: 5′-ACTGTAGCCATAAGGTCCGCT-3′). Mutant plasmids were confirmed by DNA sequencing.

Reporter Gene Assays

The AP-1 luciferase reporter plasmid construct contains the −73 to +63 collagenase promoter sequence (23). AP-1 transactivation activity was analyzed by transfection of the AP-1-luc reporter plasmid with various expression vector combinations of Cdk3, Cdk2, c-Jun, c-Jun M63/73, Cdk3-DN, and/or Cyc C. Cells were disrupted with lysis buffer (Dual Luciferase Reporter Assay System, Promega) at room temperature for 30 min by gentle shaking and analyzed for firefly luciferase activity. The AP-1-luc luciferase activity was normalized against Renilla luciferase activity (phRL-SV40).

MTS Assay

To estimate cell proliferation, SaoS-2 cells harboring pU6pro-si-mock (“mock”) or pU6pro-si-Cdk3 (“si-Cdk3”) were seeded (4 × 10³) into 96-well plates in 100 μl of 15% FBS/McCoy’s 5A medium and incubated at 37°C, 5% CO₂. After culturing for 24 h, 20 μl of the CellTiter 96^® Aqueous One Solution (Promega, Madison, WI, USA) were added to each well and cells were then incubated for 1 h in a 37°C, 5% CO₂ incubator. To stop the reaction, 25 μl of 10% SDS were added and absorbance was measured at 492 and 690 nm.

Construction of si-RNA vectors

The pU6pro vector (provided by David L. Turner, University of Michigan, Ann Arbor, MI) was used to construct pU6pro-si-mock (“si-mock”) and Pu6pro-si-Cdk3 (“si-Cdk3) following the recommended protocol at (http://sitemaker.umich.edu/dlturner.vectors). For the si-mock and si-Cdk3, we synthesized primers for the si-mock (General scramble: sense, 5′-TTTGACTACCGTTGTTATAGGTGTTCAAGAGACACCTATAACAACGGTAGTTTTTT-3′ and antisense, 5′-CTAGAAAAAACTACCGTTGTTATAGGTGTCTCTTGAACACCTATAACAACGGTAGT-3′) and for si-Cdk3 (sense, 5′-TTTGTGAGTTGGGTGCCATCAAGTTCAAGAGACTTGATGGCACCCAACTCATTTTT-3′ and antisense, 5′-CTAGAAAAATGAGTTGGGTGCCATCAAGTCTCTTGAACTTGATGGCACCCAACTCA-3′). All constructs were confirmed by restriction enzyme mapping and DNA sequencing.

Anchorage-Independent Cell Transformation Assay

To determine Cdk3 function in cell transformation induced by growth factors, JB6 cells were stably transfected with a mock vector or Cdk3. Cells (8 × 10³/ml) were exposed to EGF (20 ng/ml) in 1 ml of 0.3% basal medium Eagle’s agar/10% FBS. Cultures were maintained in a 37°C, 5% CO₂ incubator for 10 days, and colonies were scored using a microscope and the Image-Pro PLUS computer software program (v.4, Media Cybermetics, Silver Spring, MD) as described (24).

Foci Formation Assay

Transformation of NIH3T3 cells was performed following standard protocols (25). Cells were transiently transfected with combinations of pcDNA3-H-Ras^G12V (100 ng), Cdk3 (2.5 μg), c-Jun (2.5 μg), or c-Jun M63/73 (2.5 μg) and pcDNA3-mock (as compensation to achieve equal amount of DNA) DNA and then cultured in 5% calf serum-DMEM for 2 weeks. The media were changed every 3 days. Foci were fixed, stained with 0.5% crystal violet, counted with a microscope and the Image-Pro PLUS (v.4) software program.

Mammalian Two-Hybrid Assay

HEK293 cells (2.0×10⁴) were seeded into 48-well plates and incubated for 18 h before transfection. The DNAs, pACT-c-Jun, pBIND-Cdk3, and pG5-Luciferase (pG5-luc), were combined in the same molar ratio and the total amount of DNA was not more than 100 ng per well. The transfection was performed using jetPEI following the manufacturer’s recommended protocol. The cells were disrupted by addition of lysis buffer [25 mM Tris-HCl (pH 7.5), 5 mM β-glycerophosphate, 2 mM DTT, 0.1 mM Na₃VO₄, 10 mM MgCl₂, 1 mM aprotinin and 1 mM PMSF] directly into each well of the 48-well plate and aliquots of 20 μl were added to each well of a 96-well luminescence plate. The luminescence activity was measured automatically by computer program (MTX Lab, Inc., Vienna, VA). Equal transfection efficiency was normalized with Renilla luciferase activity and the relative firefly luciferase activity was calculated and normalized based on the pG5-luciferase basal control.

IP/Kinase Assay of Cdk3

To study the effect of EGF on the induction of Cdk3 activity, Cdk3 or Cdk3-DN stably-transfected cells (1.0×10⁶) were cultured for 12–24 h in 100-mm dishes. At 70–80% confluence, cells were stimulated with EGF at various doses (0, 10, 20, 40 ng/ml) or 20 ng EGF for different times (0.25, 0.5, 1, 3, 6 h), washed once with ice-cold PBS, harvested and disrupted in 250 μl of lysis buffer [25 mM Tris-HCl (pH 7.5), 5mM β-glycerophosphate, 0.1 mM Na₃ VO₄, 10 mM MgCl₂, 1 mM aprotinin, and 1 mM PMSF]. The clarified supernatant fractions containing equal amounts of protein were subjected to immunoprecipitation using a Cdk3 antibody. The Cdk3 kinase assay was carried out as described by Upstate (Upstate Biotechnology, Inc., Lake Placid, NY). Briefly, the immune complex was added to 2.5 μl of 10x kinase buffer [250 mM Tris-HCl (pH 7.5), 50 mM β-glycerophosphate, 20 mM DTT, 1 mM Na₃ VO₄, 100 mM MgCl₂], 2.5 μl (2.5 μg) of a GST-c-Jun fusion protein, 10 μl diluted ATP/cocktail (Upstate Biotechnology, Inc.), 10 μCi of [γ-³²P] ATP and H₂O to a final volume of 25 μl. The reaction was incubated at 30°C for 30 min and then resolved by 12% SDS-PAGE. Phosphorylated-GST-c-Jun was visualized by autoradiography.

Immunofluorescence Assay

Cdk3 and mock stably-transfected cells (1.0 × 10³) were seeded in eight-chamber slides and incubated 24 h at 37°C, 5% CO₂. The cells were washed at each time point, fixed in 4% formalin and permeabilized with 0.5% Triton X-100/1X PBS for 10 min. The cells were hybridized with a c-Jun mouse monoclonal antibody (1:200) and Cdk3 rabbit antibody (1:100) together at room temperature for 4 h. The cells were washed and hybridized at room temperature for 1 h with an anti-mouse goat antibody conjugated with Texas Red for detection of c-Jun and an anti-rabbit goat antibody conjugated with FITC for detection of Cdk3. Cells were washed again and observed under a fluorescence microscope (X200). To analyze co-localization of endogenous Cdk3 and phosphorylated c-Jun, SaoS-2 cells (3 × 10⁴) were seeded into 2-chamber slides and cultured overnight. The cells were fixed with 4% formalin, permeabilized with 0.5% Triton X-100 and then hybridized with a Cdk3 rabbit antibody and a phospho-c-Jun (Ser63/73) mouse antibody at 37 °C for 2 h. The slides were washed and incubated with anti-mouse goat antibody conjugated with Texas Red and anti-rabbit goat antibody conjugated with FITC at 37 °C for 1 h. The slides were observed under a confocal microscope.

Results

Cdk3 is a novel kinase to phosphorylate c-Jun at Ser63/73

Potential protein binding partners of Cdk3 were screened using the mammalian two-hybrid system. To identify partners, Cdk3, containing the full-length open reading frame, was cloned into the pBIND vector and used as bait. Various transcription factors were cloned into the pACT vector. Protein interactions between Cdk3 and individual transcription factors were measured using the pG5-luciferase reporter gene assay. Results indicated that c-Jun had a 15-fold higher interaction affinity with Cdk3 compared with the control group transfected with pG5-Luc and Cdk3 (Fig. 1A, lane 4). The pBIND-JNK1 interaction with c-Jun was used as positive control (Fig. 1A, lane 5). Based on this result, the interaction of c-Jun and Cdk3 was chosen for further analysis. To examine whether Cdk3 phosphorylated c-Jun, we generated a GST-c-Jun fusion protein (22) and conducted an in vitro kinase assay using [γ-³²P] ATP and commercially available active Cdk3 or JNK1 with the same enzyme activity units (10 mU). The results indicated that either Cdk3 or JNK1 phosphorylated c-Jun (Fig. 1B, lanes 2 and 4, respectively) and therefore Cdk3 is a newly discovered kinase that directly phosphorylates c-Jun in vitro.

Fig. 1.

Cdk3 phosphorylates c-Jun at Ser63 and Ser73 in vitro. A, Cdk3 binds with c-Jun. Novel binding partners of Cdk3 were screened by the mammalian two-hybrid assay as described in “Materials and Methods”. HEK293 cells were transfected with various combinations of plasmids as indicated and the binding affinity was measured using the pG5-luciferase reporter plasmid assay. Luciferase activity, which was normalized against Renilla luciferase activity, is expressed as the fold change in relative luminescence units normalized to the value (1.0) for cells transfected with pG5-Luc only. Data are expressed as means ± S.D. Significant differences were evaluated using the Student’s t test (* p < 0.05). B, Cdk3 phosphorylates c-Jun in vitro. The GST-c-Jun fusion protein served as substrate for active Cdk3 in the presence of [γ-³²P]ATP. The reaction mixture was resolved by 12% SDS-PAGE and ³²P-labeled GST-c-Jun was visualized by autoradiography (upper panel). The reaction of GST-c-Jun and JNK1 served as a positive control and Coommassie blue staining was used to verify equal loading (lower panel). C, Cdk3 phosphorylates c-Jun at Ser63/73. To confirm that Ser63 and Ser73 are phosphorylation target amino acids for Cdk3, an in vitro kinase assay was conducted with active Cdk3 and His-c-Jun. Phosphorylation of c-Jun at Ser63 and Ser73 was analyzed using phospho-specific antibodies and Western blotting. D, Cdk3 specifically phosphorylates c-Jun at Ser63 and Ser73. To further confirm that Cdk3 phosphorylates c-Jun at Ser63 and Ser73, a c-Jun protein harboring mutations at Ser63 and Ser73 was expressed in the BL21 bacterial strain and an in vitro kinase assay was conducted using Cdk3 and c-Jun. The phosphorylation of c-Jun was visualized by Western blot with a phospho-specific c-Jun antibody, which recognizes phosphorylation of both Ser63 and Ser73. JNK1 was used as a positive control. IB: immunoblot.

To determine the sites of c-Jun that are phosphorylated by Cdk3, we compared the amino acid similarity among various Cdk3 substrates. The retinoblastoma (Rb) (19) and ik3-1/Cables (26) proteins exhibit a conserved SP motif (Supplementary Fig. 1A) and further comparison revealed that the c-Jun protein exhibited the same SP motif at serines 63 and 73 (Ser63/73; Supplementary Fig. 1A). We conducted in vitro kinase reactions and analyzed phosphorylation of c-Jun at Ser63/73 by Western blot (Fig. 1C). The results confirmed that Cdk3 phosphorylated c-Jun at either Ser63 or Ser73.

Although Cdk3, but not Cdk2, was shown to rescue dominant negative (DN)-Cdk3-induced G1 block, Cdk2 could still phosphorylate c-Jun because of its high sequence identity with Cdk3 and its ability to complement Cdc28 mutation in yeast (14). To examine this idea, we compared the ability of active Cdk3, active Cdk2, and active JNK1 to phosphorylate a GST-c-Jun fusion protein. In vitro kinase assay and Western blot results demonstrated that Cdk3 or JNK1 could phosphorylate c-Jun at Ser63/73 (Supplementary Fig. 1B, lanes 2 and 3), whereas Cdk2 could not (Supplementary Fig. 1B, lane 4). Furthermore, neither Cdk3 nor JNK1 was able to phosphorylate a mutant (GST-c-Jun M63/73) c-Jun (Fig. 1D, lanes 2, 5 vs. lanes 7, 8, respectively). These data indicated that Cdk3, but not Cdk2, specifically phosphorylates c-Jun at Ser63/73 in vitro.

Cdk3-mediated c-Jun phosphorylation at Ser63 and Ser73 modulates c-Jun/AP-1 activity ex vivo

We next examined whether Cdk3-mediated c-Jun phosphorylation at Ser63/73 could occur ex vivo. pHis-c-Jun and pHA-Cdk3 or pHA-Cdk2 were transfected into JB6 Cl41 cells and immunoprecipitated and visualized with a His or HA antibody, respectively (Fig. 2A). The results indicated that His-c-Jun was co-immunoprecipitated with Cdk3, but not Cdk2 (Fig. 2A). We then transfected JB6 cells with various combinations of pHA-Cdk3, pHA-Cdk3-DN, pHis-c-Jun, or pHis-c-Jun M63/73 to verify that ectopic expression of Cdk3 could induce c-Jun phosphorylation at Ser63/73 (Fig. 2B). Results indicated that c-Jun phosphorylation at Ser63 and Ser73 was induced by co-transfection of pHA-Cdk3 (Fig. 2B, lane1), but not by Cdk3-DN (Fig. 2B, lane 2); and the mutant His-c-Jun M63/73 could not be phosphorylated by pHA-Cdk3 (Fig. 2B, lane 4). Moreover, the phosphorylation of c-Jun at Ser63/73 was specifically induced by Cdk3, but not by Cdk2 (Fig. 2C, lanes 4 vs. 1). These results showed that only Cdk3, and not Cdk2, induced c-Jun phosphorylation at Ser63/73 ex vivo. To determine the biological significance of c-Jun phosphorylation by Cdk3, we examined AP-1 activity using an AP-1-luciferase reporter gene assay (pAP-1-Luc). Various combination of pHA-Cdk3, pHA-Cdk2, pHA-Cdk3-DN, and the pAP-1-Luc plasmid were transfected into JB6 cells; and AP-1 transactivation activity increased by about 11-fold by Cdk3 transfection compared with pAP-1 luciferase alone (Fig. 2D, lane 2, left graph). However, AP-1 transactivation activity was not changed by transfection with either Cdk2 or Cdk3-DN compared with pAP-1-Luc alone (Fig. 2D, lane 1 vs. lanes 3, 4 left graph). Furthermore, combined transfection of Cdk3 or Cdk3-DN with p5xGal4-luciferase reporter and pGal4-c-Jun plasmids indicated that AP-1 transactivation corresponded with an increase of c-Jun transactivation activity (Fig. 2D, right graph, lane 2). Taken together, these results demonstrated that Cdk3 phosphorylation of c-Jun at Ser63/73 induced formation of the c-Jun/AP-1 complex, resulting in increased AP-1 transactivation activity.

Fig. 2.

Cdk3 phosphorylates c-Jun at Ser63 and Ser73 ex vivo. A, Cdk3 interacts with c-Jun ex vivo. pHA-Cdk3 and pHis-c-Jun or pHA-Cdk2 and pHis-c-Jun plasmids were transiently co-transfected into HEK293 cells and cultured 48 h at 37°C in a 5% CO₂ incubator. The cells were then disrupted and His-c-Jun was immunoprecipitated with a specific HisG antibody. The samples were resolved by SDS-PAGE and co-precipitation of Cdk3 and Cdk2 was analyzed using a specific HA antibody and Western blotting. B, Cdk3 phosphorylates c-Jun ex vivo. Various combinations of plasmids were transiently transfected into HEK293 cells and the c-Jun phosphorylated (Ser63 or Ser73) protein level was analyzed using phospho-specific antibodies and Western blotting. β-actin was used to verify equal protein loading. C, Cdk3, but not Cdk2, specifically phosphorylates c-Jun at both Ser63 and Ser73. To confirm that Cdk3 specifically phosphorylates c-Jun ex vivo, various combinations of plasmids containing sequences for Cdk3, Cdk2, Cdk2-DN or c-Jun were transfected into HEK293 cells and the c-Jun phosphorylated (Ser63 or Ser73) protein level was analyzed using phospho-specific antibodies and Western blotting. β-actin was used to verify equal protein loading. D, Cdk3 mediates AP-1 activity and c-Jun transactivation. (Left panel) Cdk3, Cdk2, or Cdk3-DN was transfected with AP-1-Luc into HEK293 cells. The cells were cultured for 36 h and firefly luciferase activity was measured as described in “Materials and Methods”. (Right panel) A pGal4-c-Jun fusion plasmid was co-transfected with Cdk3 or Cdk3-DN together with the p5xGal4-luciferase reporter plasmid. The cells were cultured for 36 h and firefly luciferase activity was measured as described in “Materials and Methods”. For both panels, the firefly luciferase activity was normalized against Renilla luciferase activity (phRL-SV40). Data are expressed as means ± S.D. and significant differences were determined using the Student’s t test (*, p < 0.05).

EGF-induced endogenous c-Jun phosphorylation at Ser63/73 is mediated by Cdk3 and not by JNK

To analyze the effect of Cdk3 on endogenous c-Jun/AP-1 activation ex vivo, we chose SaoS-2, HaCaT, and A431 cells because these cells are very responsive to EGF stimulation (27–29). SaoS-2, an osteosarcoma cell line, and A431, an epidermoid carcinoma cell line, showed a higher level of Cdk3 protein expression compared with HaCaT cells, which are an immortalized, non-malignant human keratinocyte cell line (Supplementary Fig. 2). We chose SaoS-2 cells for further study because this cell line has been used previously to study Cdk3 function (19). To examine the biological significance of endogenous c-Jun phosphorylation at Ser63/73 by Cdk3, we first analyzed c-Jun phosphorylation by inhibiting EGF stimulation with the EGFR inhibitor, AG1478. Results indicated that c-Jun phosphorylation induced by EGF was suppressed with EGFR inhibitor treatment (Fig. 3A, upper panels). Additional results indicated that blocking EGFR activation also inhibits c-Jun/AP-1 luciferase activity in a dose dependent manner (Fig. 3A, lower panel).

Fig. 3.

Cdk3-mediated c-Jun phosphorylation induces AP-1 activity ex vivo. A, (Western blot, upper panels) AG 1478, an EGFR inhibitor, suppresses c-Jun phosphorylation and AP-1 activity. SaoS-2 cells (2×10⁶) were seeded into 10-cm dishes and cultured overnight in a 37 °C, 5% CO₂ incubator. The cells were treated for 30 min with different doses of AG1478 as indicated, then stimulated with 20 ng/ml of EGF for 30 min and harvested. The cells were disrupted, 30 μg of total protein were resolved by 12% SDS-PAGE, and protein levels were visualized with specific antibodies and Western blotting. β-actin was used to verify equal protein loading. (Graph, lower panel) AG1478 inhibits AP-1 transcriptional activity. SaoS-2 cells (4×10⁴) were seeded into 12-well dishes and cultured overnight in a 37 °C, 5% CO₂ incubator. An AP-1-luciferase reporter plasmid (200 ng) was transfected and then SaoS-2 cells were treated with different doses of AG1478 for 12 h. The cells were harvested, disrupted with lysis buffer and luciferase activity was analyzed as described in “Materials and Methods”. The AP-1-luciferase activity was normalized against Renilla luciferase activity (phRL-SV40). Data are expressed as means ± S.D. and significant differences were determined using the Student’s t test (* p < 0.05). B, EGF stimulates phosphorylation of endogenous c-Jun. SaoS-2 cells (2×10⁶) were seeded into 10-cm dishes and cultured overnight in a 37 °C, 5% CO₂ incubator. The cells were or were not stimulated with 20 ng/ml of EGF and harvested at the indicated time point. The cells were disrupted and 30 μg of total protein were resolved by 12% SDS-PAGE and protein levels were visualized using specific antibodies and Western blotting. β-actin was used to verify equal protein loading. C, SaoS-2 cells (2×10⁶) were seeded into 10-cm dishes and cultured overnight in a 37 °C, 5% CO₂ incubator. The cells were treated with 10 μM PD98059 or SP600125 for 30 min and then stimulated with 20 ng/ml of EGF. The cells were harvested at the indicated time point and proteins were visualized by Western blotting with specific antibodies as described in “Materials and Methods”. β-actin was used to verify equal protein loading. D, SaoS-2 cells (1×10⁶) were seeded into 10-cm dishes and cultured overnight. The cells were transfected with 8 μg of pHA-mock, pHA-Cdk3 or pHA-Cdk3-DN and then cultured for 24 h. The cells were trypsinized and cells (4×10³) seeded into 96-well plates. Cell proliferation was measured at the indicated time point using the MTS CellTiter 96^® Aqueous One Solution assay as described in “Materials and Methods”. Data are expressed as means ± S.D. and significant differences were determined using the Student’s t test (* p < 0.05; ** p < 0.001).

Our previous study indicated that EGF stimulation induces cell transformation mediated through AP-1 activity (30). Furthermore, we found that EGF stimulation could not induce JNK phosphorylation at Thr183/Tyr185, a marker of JNK activation (31). To examine whether EGF-induced c-Jun phosphorylation is mediated by JNK or Cdk3, we analyzed phosphorylation of JNK and ERK after stimulation with EGF (20 ng/ml). Results indicated that although EGF stimulated phosphorylation of ERK at 5 min, continued over 30 min, and decreased at 60 min (Fig. 3B, topmost panel), JNK phosphorylation was not detected over the same period of time (Fig. 3B, upper middle panel), indicating that EGF does not activate JNK. On the other hand, c-Jun phosphorylation at Ser63/73 was increased at 5 min, continued to at least 60 min and then slightly decreased at 120 min after EGF stimulation (Fig. 3B, lower top panel). To avoid the possibility of the involvement of either ERK or JNK as c-Jun kinases, we used chemical compounds to inhibit MEK and JNK activity (Fig. 3C). SaoS-2 cells were pretreated with PD98059 (MEK1/2 inhibitor) or SP600125 (JNKs inhibitor) for 30 min and then stimulated with 20 ng/ml of EGF. The Western blot profiles indicated that MEK inhibition only suppressed ERK phosphorylation and not c-Jun phosphorylation at Ser63/73 (Fig. 3C, upper panel series). Furthermore, the JNKs inhibitor had no effect on EGF-induced c-Jun phosphorylation at either Ser63 or Ser73 (Fig. 3C, lower panel series). We found that transfection of HA-Cdk3 enhanced SaoS2 proliferation (Fig. 3D), whereas HA-Cdk3-DN slightly suppressed SaoS-2 proliferation (Fig. 3D). These results supported our hypothesis that EGF-mediated c-Jun phosphorylation does not occur not through ERKs or JNKs signaling.

EGF stimulation induces the interaction of endogenous Cdk3 and c-Jun

To analyze the ex vivo interaction of endogenous Cdk3 and c-Jun by EGF stimulation, we conducted a co-immunoprecipitation (IP) experiment. In the co-IP results, we found that Cdk3 and c-Jun binding increased at 5 min after EGF stimulation, further increased by 60 min, and decreased at 120 min (Fig. 4A, top panel), which closely matches the c-Jun phosphorylation pattern (Fig. 4A, 4^th panel). Concurrently, the total Cdk3 protein level was increased at 60 min and slightly deceased at 120 min (Fig. 4A, 3^rd panel). In addition, EGF stimulation increased Cdk3-mediated c-Jun phosphorylation as shown by IP-kinase assay results using His-c-Jun and [γ-³²P]ATP (Fig. 4B). Taken together, these results indicated that EGF stimulation enhanced the interaction of Cdk3 and c-Jun, which resulted in the phosphorylation of c-Jun at Ser63/73 by Cdk3. Next, we hypothesized that if Cdk3 phosphorylates c-Jun at Ser63/73 ex vivo, then these proteins should co-localize. SaoS-2 cells were or were not stimulated with EGF and co-localization of Cdk3 (FITC) and phospho-c-Jun (Ser63/73) was examined by immunofluorescence (Fig. 4C). The phosphorylated c-Jun and Cdk3 were barely detectable before EGF stimulation (Fig. 4C). However, after EGF stimulation, the phosphorylated c-Jun and Cdk3 levels were increased and co-localized together (Fig. 4C). These results further demonstrated that EGF induces c-Jun phosphorylation at Ser63/73, which is directly mediated by Cdk3.

Fig. 4.

EGF-induced c-Jun phosphorylation is mediated by Cdk3. A, Endogenous Cdk3 is activated by EGF stimulation. SaoS-2 cells (2×10⁶) were seeded into 10-cm dishes and cultured overnight in a 37 °C, 5% CO₂ incubator. The cells were stimulated with 20 ng/ml of EGF and harvested at the indicated time point. The cells were disrupted with kinase lysis buffer, immunoprecipitated with a c-Jun antibody and Cdk3 was detected with a Cdk3 antibody. Protein levels were analyzed from the same cell lysates using specific antibodies and Western blotting and β-actin was used to verify equal protein loading. B, Using the same cell lysates from (A), Cdk3 was immunoprecipitated with a Cdk3 antibody and an in vitro kinase assay was conducted using [γ-³²P]ATP and GST-c-Jun as described in “Materials and Methods”. The ³²P-labeled phospho-bands were visualized by autoradiography and immunoprecipitated Cdk3 was visualized by Western blotting. C, Endogenous Cdk3 co-localizes with phosphorylated c-Jun (Ser63 and Ser73). SaoS-2 cells (2 × 10⁴) were seeded into 2-chamber slides and cultured overnight. The cells were or were not stimulated with 20 ng/ml of EGF for 15 min and then an immunofluorescence assay was conducted as described in “Materials and Methods”. The cells were observed using a confocal microscope (200X).

Knockdown of Cdk3 inhibits cell proliferation and c-Jun/AP-1 activity

To analyze the knockdown effect of endogenous Cdk3, we constructed pU6pro-si-general scramble (si-mock) and pU6pro-si-Cdk3 (si-Cdk3) vectors as described in “Materials and Methods”. The si-mock or si-Cdk3 was transfected into SaoS2 cells and knockdown efficiency was analyzed by Western blot. Tranfection of si-Cdk3 suppressed the endogenous Cdk3 protein level by about 80% (Fig. 5A). Importantly, we found that knockdown of Cdk3 by si-Cdk3 stable transfection inhibited c-Jun phosphorylation at Ser63/73 (Fig. 5B). Suppression of c-Jun phosphorylation by increasing amounts of si-Cdk3 was directly associated with dose-dependent reductions in AP-1 luciferase activity (Fig. 5C) and proliferation (Fig. 5D). These results demonstrated that Cdk3-mediated c-Jun activation increases c-Jun/AP-1 transcriptional activity and induces proliferation.

Fig. 5.

Knockdown of endogenous Cdk3 inhibits EGF-induced c-Jun phosphorylation and c-Jun/AP-1 activity. A, Knockdown of Cdk3. The si-Cdk3 and si-mock vectors were constructed as described in “Materials and Methods”, transfected into SaoS-2 cells, and cells were then cultured for 24 h. The cells were disrupted and the endogenous protein level of Cdk3 was visualized by Western blotting. β-actin was used to verify equal protein loading. B, Knockdown of Cdk3 suppresses EGF-induced c-Jun phosphorylation. SaoS-2 cells (2×10⁶) stably expressing si-mock or si-Cdk3 were seeded into 10-cm dishes and cultured overnight in a 37 °C, 5% CO₂ incubator. The cells were stimulated with 20 ng/ml of EGF and harvested at the indicated time point. The cells were then disrupted, 30 μg of total protein were resolved by 12% SDS-PAGE and protein levels were visualized with specific antibodies and Western blotting. β-actin was used to verify equal protein loading. C, Knockdown of endogenous Cdk3 suppresses AP-1 transcriptional activity. SaoS-2 cells (4×10⁴) stably expressing si-mock or si-Cdk3 were seeded into 12-well dishes and cultured overnight in a 37 °C, 5% CO₂ incubator. An AP-1-luciferase reporter plasmid was transfected at different doses and cells were cultured for 24 h. The cells were then harvested, disrupted with luciferase lysis buffer and the luciferase activity was analyzed as described in “Materials and Methods”. The AP-1-luciferase activity is normalized against Renilla luciferase activity (phRL-SV40). Data are expressed as means ± S.D. and significant differences were determined using the Student’s t test (* p < 0.05). D, Knockdown of endogenous Cdk3 inhibits cell proliferation. SaoS-2 cells (5×10³) stably expressing si-mock or si-Cdk3 were seeded into 96-well plates and cultured with complete growth medium. Cell proliferation was analyzed using the MTS CellTiter 96^® Aqueous One Solution assay at the indicated time point as described in “Materials and Methods”. Data are expressed as means ± S.D. and significant differences were determined using the Student’s t test (* p < 0.05; ** p < 0.005).

The Cdk3-AP-1 signaling axis plays an important role in cell transformation

The phosphorylation of c-Jun at Ser63/73 is associated with cell transformation (3, 12). A persistent alteration in AP-1 activity results in enhanced oncogenic transformation (32) and AP-1 activation is required to promote skin carcinogenesis (33). The protein expression level of Cdk3 and c-Jun was observed to be higher in A431 malignant human skin cancer cells compared to immortalized HaCaT cells (Supplementary Fig. 2). These results suggested that the Cdk3-c-Jun/AP-1 signaling axis plays an important role in cell transformation. To examine this suggestion, we established JB6 Cl41 cells stably expressing a mock vector or Cdk3. Using these cells, we assessed EGF-induced anchorage-independent growth in soft agar (Fig. 6A) and found that Cdk3 stable cells induced about 4.6-fold more cell transformation compared with JB6/pcDNA4-mock cells (Fig. 6A). Furthermore, knockdown of c-Jun with siRNA-c-Jun in Cdk3 stable JB6 Cl41 cells suppressed colony formation in soft agar (Supplementary Fig. 3). Cdk3 overexpressing stable cells also showed elevated c-Jun phosphorylation levels compared with pcDNA4-mock cells (Fig. 6B, upper panels 1 and 2). However, phosphorylation of JNK was not detectable in either of the two cell types (Fig. 6B, upper panel 5). Furthermore, EGF-induced AP-1 activity was also enhanced by the introduction of pcDNA4-Cdk3 (Fig. 6B, lower panel), indicating that Cdk3-mediated cell transformation might be mediated by c-Jun phosphorylation. This notion was supported by results of a foci formation assay using NIH3T3 cells (Fig. 6C). Constitutively active Ras^G12V-induced foci formation was increased by co-transfection with Cdk3 (Fig. 6C, bottom row, 1^st plate). The increased foci number induced by the combination of Ras^G12V and Cdk3 was even more increased by triple-transfection of Ras^G12V, Cdk3, and c-Jun (Fig. 6C, bottom row, 2^nd plate), but not by including c-Jun M63/73 with Ras^G12V and Cdk3 (Fig. 6C, bottom row, 3^rd plate). Furthermore, AP-1 luciferase activity of NIH3T3 cells was also increased by co-transfection of Cdk3 and c-Jun, but not by co-transfection of cells with Cdk2/c-Jun (data not shown). In addition, co-transfection of Cdk3/c-Jun/Cyc C further enhanced AP-1 luciferase activity compared with Cdk3/c-Jun (Fig. 6D, lane 4 vs. lane 6). Therefore, based on overall results, we concluded that growth factor-mediated Cdk3 activation induces c-Jun phosphorylation at Ser63/73, resulting in enhanced c-Jun/AP-1 transcriptional activity and cell transformation.

Fig. 6.

The Cdk3/c-Jun signaling axis plays an important role in cell transformation by regulating c-Jun/AP-1 activity. A, EGF-induced cell transformation is enhanced by stable expression of Cdk3. JB6 Cl41 cells (8×10³) stably expressing mock or Cdk3 were used for an anchorage-independent cell transformation assay with EGF stimulation (20 ng/ml) as described in “Materials and Methods”. Data are expressed as means ± S.D. from 3 independent experiments. Significant differences were evaluated using the Student’s t test (* p < 0.05). B, (Western blot, upper panels) Cdk3 enhances c-Jun phosphorylation at Ser63 and Ser73 independently of JNK. JB6 Cl41 cells (1×10⁶) stably expressing mock or Cdk3 were seeded into 10-cm dishes and cultured for 24 h. The proteins were extracted and visualized by Western blotting with specific antibodies as indicated. β-actin was used to verify equal protein loading. (Graph, lower panel) EGF enhances Cdk3-mediated AP-1 transcriptional activity. JB6 Cl41 cells (2×10⁴) stably expressing mock or Cdk3 were seeded and cultured overnight in a 37 °C, 5% CO₂ incubator. An AP-1-luciferase reporter plasmid (200 ng) was transfected and then the cells were stimulated for 12 h with 20 ng/ml of EGF as indicated. The cells were disrupted with lysis buffer and the luciferase activity was analyzed as described in “Materials and Methods”. The AP-1-luciferase activity was normalized against Renilla luciferase activity (phRL-SV40). Data are expressed as means ± S.D. from triplicate experiments and significant differences were evaluated using the Student’s t test (* p < 0.05). C, Ras-Cdk3-c-Jun enhances foci formation in NIH3T3 cells. NIH3T3 cells were transfected with various combinations of Ras^G12V, Cdk3, c-Jun or mutant c-Jun as indicated and foci formation was determined as described in “Materials and Methods”. Data are expressed as means ± S.D. from 3 independent experiments and significant differences were evaluated using the Student’s t test (* p < 0.05). D, Cyclin C induces Cdk3-mediated AP-1 activity. NIH3T3 cells (2×10⁴) were seeded into 12-well dishes and cultured overnight. The cells were transfected with AP-1-Luc and various combinations of expression vectors as indicated. The cells were cultured 24 h, disrupted with lysis buffer and the luciferase activity was analyzed as described in “Materials and Methods”. The AP-1-luciferase activity was normalized against Renilla luciferase activity (phRL-SV40). Data are expressed as means ± S.D. from triplicate experiments. Significant differences were evaluated using the Student’s t test (* p < 0.05).

Discussion

Growth factors and tumor promoters induce AP-1 activity, indicating that AP-1 is important in growth control as well as cell transformation. Several mechanisms account for stimulation of AP-1 activity by growth factors, pro-inflammatory cytokines and ultraviolet (UV) radiation (34). The most important mediator of the growth factor response is probably the ERK MAP kinase cascade, which is activated through phosphorylation of ternary complex factors (35). Our results demonstrated that EGF stimulation induced phosphorylation of ERK, but not JNK (Fig. 3B, 3C, 6B). In contrast, a primary stimulus that induces JNK phosphorylation is UV (34, 36, 37). Although UV is a well-known activator of JNK and AP-1, previous work indicated that ERK or other signaling pathways might also be required for AP-1 activation (38, 39). Those results indicated that the JNK and ERK signaling pathways might differentially lead to phosphorylation of c-Jun depending on stimulation or cell context. Furthermore, we found that the proteinase inhibitors I and II from potatoes specifically inhibited AP-1 activity through a pathway that is independent of the ERKs, p38, or JNKs pathway (40). Furthermore, treatment with PD98059, a MEK inhibitor, or SP600125, a JNKs inhibitor, did not affect EGF-induced c-Jun phosphorylation (Fig. 6B) in SaoS-2 cells, indicating that other signaling molecule(s) must exist to phosphorylate c-Jun. Thus the observation in the present study that Cdk3 is a novel kinase for phosphorylating c-Jun at Ser63 and Ser73 is extremely important in understanding how growth factors can induce cell proliferation or cell transformation through the regulation of AP-1 transactivation activity.

The initial conclusion that ERK was a c-Jun kinase (4) was re-evaluated when JNK was discovered and cloned (7). Purified JNK phosphorylated c-Jun at serines 63 and 73, whereas ERK did not (13). Many potent JNK agonists are poor ERK agonists and vice versa (7), suggesting that the level of c-Jun phosphorylation is associated with the relative activation of ERK and JNK in response to different stimuli. Evidence also indicates that the intensity and duration of JNK activation and the degree of c-Jun Ser63/73 phosphorylation correspond very well; but little or no correlation exists between c-Jun phosphorylation and ERK activity (7, 13, 41). Constitutive activation of growth-promoting signaling pathways such as Ras and ERK leads to cell transformation. The Ras proteins play a key role in signal transduction and regulation of proliferation (42, 43). Ras signaling is mediated through MAP kinase kinase, which in turn activates ERK. However, Ras only partially activates JNKs, suggesting that other kinase(s) that phosphorylate c-Jun at Ser63/73 must exist. When cells were stimulated with EGF, which is a well-known tumor promoter that activates Ras, ERK phosphorylation was increased (Fig. 3B), whereas JNK phosphorylation was not affected by EGF stimulation (Fig. 3B), indicating that c-Jun phosphorylation at Ser63/73 induced by EGF does occur not through JNK signaling. The purpose of the present study was to identify novel kinase(s), which can phosphorylate c-Jun; and Cdk3 was found to be a c-Jun binding partner (Fig. 1A). Cdk3 could phosphorylate c-Jun at Ser63/73 (Fig. 1B-C) and co-localize with phosphorylated c-Jun following EGF stimulation (Fig. 4C). EGF stimulation increased phosphorylated c-Jun and Cdk3 protein levels over a similar time period (Fig. 4A). An EGFR inhibitor suppressed EGF-induced c-Jun phosphorylation and AP-1 activation (Fig. 3A). Notably, knockdown of Cdk3 suppressed AP-1 luciferase activity and c-Jun phosphorylation as well as proliferation (Fig. 5B–D), indicating that the Cdk3/c-Jun signaling axis has an important function in proliferation induced by growth factors such as EGF.

The DNA binding and transcriptional activities of c-Jun are affected by events such as induction of AP-1 activity by TPA and other tumor promoters, which reflects a post-translational activating modification and stimulates c-Jun protein engagement in an autoregulatory loop (44, 45). AP-1 regulates cell cycle progression by binding with the cyclin D1 gene promoter directly (46) and suppresses stabilization of the p53 protein, resulting in accumulation of the p21 protein (47). AP-1 transactivation was required for tumor promotion in transgenic mice (33) and EGF stimulation was reported to induce cell transformation through AP-1 activation (30). A dominant negative c-Jun mutant blocked H-Ras and c-Jun-induced cell transformation (48). Stable expression of Cdk3 in JB6 Cl41 P+ cells enhanced cell transformation induced by EGF (Fig. 6A). Furthermore, NIH3T3 cells harboring constitutively active H-Ras^G12V and transfected with Cdk3 and c-Jun exhibited more efficient foci formation compared with H-Ras^G12V alone or H-Ras^G12V combined with Cdk3 (Fig. 6C); and the process was mediated through AP-1 activation (Fig. 6B, 6D).

Cdk3 induces G0 exit into G1 in resting cells (19) and participates in the G1 exit to S entry (16). Transfection of dominant negative Cdk3 caused G1 accumulation (15) and therefore, Cdk3 is involved in cell cycle progression. Cdk3 kinase activity was not affected by p16 or E2F-1, but was downregulated with transient p27^KIP1 expression (17). Ectopic Cdk3, but not Cdk2, enhanced Myc-induced proliferation and anchorage-independent growth-associated Myc activation (17), indicating that Cdk3 could have oncogenic potential. Recently, Ren and Ronallins (19) found that cyclin C is a novel binding partner with Cdk3 and Cdk3/cyclin C promotes Rb-dependent G0 exit, the regulation of which is similar to regulation of the G1/S transition (49). Our results demonstrated that knockdown of Cdk3 by si-RNA suppressed cell proliferation (Fig. 5D) through inhibition of c-Jun phosphorylation (Fig. 5B) and AP-1 transactivation activity (Fig. 5C). In summary, Cdk3 is a well-known kinase with a newly discovered function of phosphorylating c-Jun resulting in AP-1 transactivation, which is critical for proliferation as well as cell transformation. Growth factor-induced c-Jun phosphorylation occurs independently of JNK and this finding suggests that Cdk3 may be a new target for anticancer drug development.