ERK3 (extracellular-signal-regulated kinase 3) is an atypical MAPK (mitogen-activated protein kinase) that is suggested to play a role in cell-cycle progression and cellular differentiation. However, it is not known whether the function of ERK3 is regulated during the cell cycle. In the present paper, we report that ERK3 is stoichiometrically hyperphosphorylated during entry into mitosis and is dephosphorylated at the M→G1 transition. The phosphorylation of ERK3 is associated with the accumulation of the protein in mitosis. In vitro phosphorylation of a series of ERK3-deletion mutants by mitotic cell extracts revealed that phosphorylation is confined to the unique C-terminal extension of the protein. Using MS analysis, we identified four novel phosphorylation sites, Ser684, Ser688, Thr698 and Ser705, located at the extreme C-terminus of ERK3. All four sites are followed by a proline residue. We have shown that purified cyclin B-Cdk1 (cyclindependent kinase 1) phosphorylates these sites in vitro and demonstrate that Cdk1 acts as a major Thr698 kinase in vivo. Reciprocally, we found that the phosphatases Cdc14A and Cdc14B (Cdc is cell-division cycle) bind to ERK3 and reverse its C-terminal phosphorylation in mitosis. Importantly, alanine substitution of the four C-terminal phosphorylation sites markedly decreased the half-life of ERK3 in mitosis, thereby linking phosphorylation to the stabilization of the kinase. The results of the present study identify a novel regulatory mechanism of ERK3 that operates in a cell-cycle-dependent manner.
- cell cycle
- extracellular-signal-regulated kinase 3 (ERK3)
- mitogen-activated protein kinase (MAPK)
- protein phosphorylation
ERK3 (extracellular-signal-regulated kinase 3) is an atypical member of the MAPK (mitogen-activated protein kinase) family of serine/threonine protein kinases. Together with the paralogous kinase ERK4, ERK3 defines a distinct subfamily of MAPKs that is found exclusively in vertebrates . Structurally, the ERK3 protein is composed of a kinase domain at the N-terminus followed by a long C-terminal extension rich in serine and threonine residues [2–4]. The kinase domain of ERK3 displays ~45% amino acid identity with that of the MAPKs ERK1/ERK2 and is catalytically active. One notable feature that distinguishes these two subfamilies of MAPKs is the presence of a single phospho-acceptor site in the activation loop of ERK3, instead of the canonical Thr-X-Tyr motif. The unique C-terminal extension of ERK3 is conserved throughout vertebrate evolution, suggesting an important function. Previous findings indicate that this region is involved in mediating protein–protein interactions [5,6].
Although the exact cellular functions of ERK3 remain to be defined, several lines of evidence link ERK3 signalling to the regulation of cell-cycle progression and cellular differentiation. The levels of ERK3 mRNA and protein are up-regulated during cell-cycle exit and terminal differentiation of P19 embryonal carcinoma cells  and C2C12 myoblasts  in vitro. ERK3 expression is also induced following treatment of Raji lymphoma cells with gangliosides  or after plating of squamous cell carcinoma lines on type IV collagen , two conditions associated with proliferation arrest. In further support of this idea, overexpression of ERK3 was shown to inhibit G1- to S-phase progression in fibroblasts [7,10]. Interestingly, two studies have documented that ERK3 interacts with the cell-cycle regulators cyclin D3  and Cdc (cell-division cycle)14A  through its unique C-terminal domain, suggesting a possible implication of this region in cell-cycle control.
The activity of ERK3 is regulated in large part by post-translational mechanisms. ERK3 is a highly unstable protein (half-life ~30–60 min) that is constitutively degraded by the ubiquitin–proteasome pathway in exponentially proliferating cells . Analysis of a series of chimaeras made between the stable ERK1 protein and ERK3 established two degrons in the N-terminal lobe of the ERK3 kinase domain that are both necessary and sufficient to target ERK3 for degradation . Analogous to other MAPKs, ERK3 is phosphorylated on the activation loop residue Ser189 in intact cells [7,11,12]. Activation loop phosphorylation of ERK3 stimulates its intrinsic catalytic activity and is required for the formation of a stable complex with the substrate MK5 (MAPK-activated protein kinase 5) . Notably, phosphopeptide mapping analysis revealed that ERK3 is phosphorylated on additional residues outside of the activation loop . We have identified one of these sites as Ser386, which lies within a putative MK5 consensus phosphorylation sequence . It is not known whether the phosphorylation and activity of ERK3 varies during the cell cycle.
In the present study, we report on the identification, by MS, of four residues in the C-terminal extension of ERK3 (Ser684, Ser688, Thr698 and Ser705) that are specifically phosphorylated in mitosis. We provide evidence that phosphorylation of Thr698 is catalysed by the mitotic kinase cyclin B-Cdk1 (cyclin-dependent kinase 1), and that mitotic phosphorylation is reversed by the action of the phosphatases Cdc14A and Cdc14B. Importantly, we show that phosphorylation of the four C-terminal sites stabilizes the protein during mitosis. Our results identify a novel mechanism of regulation of ERK3 and a further connection of the kinase with the regulation of cell-cycle progression.
MATERIALS AND METHODS
Reagents and antibodies
Thymidine, nocodazole, taxol, demecolcine and cycloheximide were obtained from Sigma–Aldrich. Roscovitine and SU9516 were from Calbiochem. Active cyclin B-Cdc2 and λ protein phosphatase were from New England Biolabs. Histone H1 was obtained from Roche. The ERK3 phospho-Thr698-specific antibody was generated by GeneScript by immunization of rabbits with the synthetic peptide CTLpTPSAMKS. Polyclonal anti-ERK3 antibody E3-CTE4 has been described previously . Commercial antibodies were from the following suppliers: rabbit monoclonal anti-ERK3 from Epitomics; monoclonal anti-actin (AC-40) and polyclonal anti-FLAG from Sigma–Aldrich; monoclonal anti-cyclin B1 (GNS1) and anti-cyclin A (E23) from Neomarkers; monoclonal anti-phospho-histone H3(Ser10) (6G3) from Cell Signaling Technology; rabbit polyclonal anti-GAPDH (glyceraldehyde-3-phosphate dehydrogenase) and anti-Myc from Santa Cruz Biotechnology; and HRP (horseradish peroxidase)-conjugated goat anti-mouse and anti-rabbit IgG from Bio-Rad. An anti-FLAG M2–agarose antibody was obtained from Sigma–Aldrich.
Plasmid constructs and mutagenesis
pcDNA3-EGFP-ERK3 has been described previously . pcDNA3-EGFP-ERK3-Flag was constructed by inserting the FLAG sequence into EcoRI/XbaI sites of pcDNA3-GFP (green fluorescent protein) followed by cloning of ERK3 into the EcoRI site. pcDNA3-Myc6-ERK3 and pcDNA3-Flag-ERK3 wild-type and mutants were obtained by cloning ERK3 into the EcoRI site of pcDNA3-Myc6 and pcDNA3-Flag. pHGST.1-ERK3 has been described previously . The hCdc14A cDNA was kindly provided by Professor H. Charbonneau (Department of Biochemistry, Purdue University, West Lafayette, IN, U.S.A.). EcoRI and XhoI sites were added by PCR and the hCdc14A sequence was subcloned into pcDNA3-Flag. The catalytically inactive mutant of hCdc14A (C238G) was generated by PCR. Wild-type and catalytically inactive pcDNA3-Flag-Cdc14B have been described previously . pcDNA3-Myc6-cyclin A2, pcDNA3-Myc6-cyclin B1, pcDNA3-Myc6-cyclin D1 and pcDNA3-Myc6-cyclin E2 have been described elsewhere [12,15]. Mutations were introduced into the human ERK3 cDNA by PCR and were confirmed by DNA sequencing.
Purification of recombinant proteins and in vitro translation
Purification of His6–ERK3–GST (glutathione transferase) protein was carried out as described previously . The yield and purity of the ERK3 preparation was evaluated by SDS/PAGE and Coomassie Blue staining using BSA as standard.
In vitro transcription/translation reactions were performed using the TNT® system from Promega according to the manufacturer's instructions.
Cell culture, transfections and cell synchronization
HeLa and HEK (human embryonic kidney)-293 cells were grown in DMEM (Dulbecco's modified Eagle's medium) and MEM (minimal essential medium) respectively, supplemented with 10% FBS (fetal bovine serum) and antibiotics. HeLa cells were transfected with polyethylenimine. HEK-293 cells were transfected using the calcium phosphate co-precipitation method.
For synchronization experiments, HeLa cells at 50–60% confluence were incubated with 2 mM thymidine for 18 h. The cells were then washed with PBS, and cultured in fresh medium for 8 h. To obtain mitotic cells, nocodazole (100 ng/ml), taxol (1 μM) or demecolcine (1 μM) were added for 18 h and round cells were collected by mitotic shake-off. Synchronization of cells at the G1/S transition was obtained by a double-thymidine block .
In vitro phosphorylation analysis
For identification of ERK3 phosphorylation sites, recombinant His6–ERK3–GST protein was phosphorylated in vitro by synchronized HeLa cell extracts. Exponentially proliferating or mitotic HeLa cells were lysed in buffer C [50 mM Tris/HCl (pH 7.4), 1.5 mM KCl, 1 mM DTT (dithiothreitol), 5 mM MgCl2, 0.2 mM sodium orthovanadate, 1 mM pepstatin A, 1 mM leupeptin and 0.1 mM PMSF] using a dounce homogenizer. Protein extract (100 μg) was incubated with 1 μg of His6–ERK3–GST on Ni-NTA (Ni2+-nitrilotriacetate) beads in kinase assay buffer [50 mM Tris/HCl (pH 7.4), 1.5 mM KCl, 1 mM DTT and 5 mM MgCl2] in the presence of 200 μM ATP. At the end of the reaction, the beads were washed four times with 800 μl of buffer D [50 mM Tris/HCl (pH 7.5), 150 mM NaCl, 10 mM NaF, 20 mM 2-glycerophosphate, 1% Triton X-100, 1 mM pepstatin A, 1 mM leupeptin, 0.1 mM PMSF and 0.1 mM sodium orthovanadate] and boiled in Laemmli buffer. The proteins were resolved by SDS/PAGE and detected by silver nitrate staining.
Phosphorylated His6–ERK3–GST was separated by SDS/PAGE and visualized by silver staining. The band was excised from the gel and cut into small pieces. The gel slices were then reduced in 10 mM DTT for 1 h at 56 °C and alkylated in 55 mM iodoacetamide for 1 h at room temperature (22 °C). After washing in 50 mM ammonium carbonate, the gel pieces were shrunk in 100% acetonitrile. Digestion was performed with trypsin in 50 mM ammonium carbonate for 4 h at 37 °C. The peptides were finally extracted in 90% acetonitrile/0.5 M urea and dried in a Speed Vac. Samples were resolubilized in 5% acetonitrile/0.2% formic acid and analysed on a Eksigent nano-LC system coupled to a Thermo LTQ Orbitrap mass spectrometer with a home-made C18 pre-column and an analytical column (10 cm×150 μm, Jupiter 3 μm C18). A volume of 10 μl of sample was injected for analysis. Tryptic peptide digests were first loaded on to the pre-column at a flow rate of 10 μl/min and subsequently eluted on to the analytical column using a gradient from 10–60% acetonitrile in 0.2% formic acid over 56 min at 600 nl/min. Database searches were run using Mascot version 2.1 (Matrix Science).
Immunoblot analysis, immunoprecipitation and kinase assays
Cell lysis and immunoblot analysis were performed as described previously . For co-immunoprecipitation experiments, cells were lysed in buffer E [50 mM Tris/HCl (pH 7.5), 125 mM NaCl, 5 mM EDTA, 0.2% Nonidet P40, 10 mM NaF, 1 mM DTT, 0.2 mM sodium orthovanadate, 1 mM pepstatin A, 1 mM leupeptin and 0.1 mM PMSF). Cell lysates were incubated with anti-FLAG M2–agarose beads for 2 h at 4 °C, and the precipitated complexes were analysed by immunoblotting.
Kinase assays with recombinant active cyclin B-Cdc2 were performed according to the manufacturer's instructions (New England Biolabs). Recombinant His6–ERK3–GST (2 μg of protein) was incubated with 2 units of enzyme, 200 μM ATP and 10 μCi of [γ32P]ATP for 30 min at 30 °C. The reaction products were analysed by SDS/PAGE and autoradiography.
Cell-cycle distribution was monitored by FACS analysis after propidium iodide staining using a LSRII cytometer (Beckton Dickinson) as described previously .
ERK3 is hyperphosphorylated and accumulates in mitosis
In an effort to determine whether the levels or activity of ERK3 oscillate during the cell cycle, we monitored the expression of ERK3 at different stages of the cell cycle using highly synchronized HeLa cell populations. We observed that ERK3 protein levels are significantly higher in mitotic cells as compared with asynchronous or late-G1 cells (Figure 1A). Notably, the accumulation of ERK3 in mitosis was accompanied by a stoichiometric upward electrophoretic mobility shift of the protein. Essentially similar results were obtained using three different microtubule-disrupting agents, indicating that the slower migration of ERK3 is not artefactually caused by the method of cell synchronization (Figure 1A). The most likely explanation for the mobility shift of ERK3 observed in mitosis is a post-translational modification by phosphorylation. Given the importance of phosphorylation in the regulation of mitotic events, we set out to investigate whether ERK3 is differentially phosphorylated in mitosis. To test this idea, we immunoprecipitated endogenous ERK3 from asynchronous or nocodazole-arrested prometaphase cells and incubated the precipitates with λ protein phosphatase. λ phosphatase treatment of ERK3 isolated from exponentially proliferating cells resulted in a slight acceleration of its electrophoretic migration, consistent with the fact that ERK3 is phosphorylated on at least two residues in proliferating cells (Figure 1B) . However, phosphatase treatment of ERK3 isolated from mitotic cells markedly increased the mobility of the kinase, which now migrated at the same rate as the protein in asynchronous cells. This change in mobility was completely prevented by pre-incubation with a phosphatase inhibitor cocktail. We concluded from these experiments that ERK3 is hyperphosphorylated in mitosis. This phosphorylation of ERK3 in mitosis was observed in several other cell lines, including T98G, U2OS and HT-29 (results not shown).
We next analysed the kinetics of phosphorylation of ERK3 during the cell cycle. HeLa cells were first synchronized in mitosis by nocodazole treatment followed by mitotic shake-off, and the rounded cells were replated in fresh medium to allow synchronous exit from mitosis and entry into the next G1-phase. ERK3 was stoichiometrically hyperphosphorylated in mitotic cells (zero time point) as indicated by its retarded mobility on the gel (Figure 1C). By 2 h of nocodazole release, the slow-migrating form of ERK3 had completely shifted to a faster-migrating species, similar to that observed in asynchronous cells (Figure 1C). This mobility shift was associated with a decrease in ERK3 levels. The kinetics of dephosphorylation of ERK3 paralleled the degradation of cyclin B1, which occurs in late mitosis and early G1. These results indicate that the phosphorylation of ERK3 in mitosis is reversible. We next synchronized the cells at the G1/S transition by a double-thymidine block and allowed them to progress into mitosis in the presence of nocodazole to create a mitotic trap. Entry into mitosis was monitored by the phosphorylation of histone H3 on Ser10. ERK3 mobility started to shift upward between 8 and 9 h after release from G1/S, when cyclin B1 and phospho-H3 levels reached a maximum and the majority of cells have a 4N DNA content (Figure 1D). Notably, at 6 and 7 h after G1/S, when cells are mainly in G2 (as indicated by the high percentage of cells with 4N DNA and the absence of the phospho-H3 signal), ERK3 migration was not modified. These results demonstrate that ERK3 is reversibly hyperphosphorylated in mitosis, concomitant with its transient accumulation.
ERK3 is specifically phosphorylated in the C-terminal extension in mitosis
ERK3 is phosphorylated on the activation loop residue Ser189 and on Ser386 in exponentially proliferating cells [7,11,12]. To determine whether these two serine residues contribute to the mobility shift of ERK3 in mitosis, we analysed the phosphorylation of in vitro translated wild-type ERK3 or S189A and S386A mutants by mitotic extracts. The migration of ERK3 was clearly shifted upward after 1 h of incubation with an extract of mitotic HeLa cells (Figure 2A). The ERK3 mutants S189A and S386A shifted to the same level as the wild-type protein, indicating that these sites are not responsible for the hyperphosphorylation of ERK3 in mitosis. To define the region(s) of ERK3 phosphorylated in mitosis, we analysed the electrophoretic mobility shift of a series of ERK3-deletion mutants (Figure 2B) following incubation with mitotic HeLa cell extract. Deletion of the last 43 amino acids of ERK3 resulted in a slight reduction in the mobility shift on the gel (Figure 2C). However, deletion of the entire C-terminal extension (mutants 1–399 and 1–365) completely abrogated the shift. Conversely, incubation of only the C-terminal extension of ERK3 (mutant 365–721) with a mitotic extract resulted in a marked electrophoretic mobility retardation (Figure 2C, right-hand panel). These results identify the C-terminal domain of ERK3 as the main region phosphorylated during mitosis.
To identify individual ERK3 sites that are specifically phosphorylated in mitosis, extracts from asynchronously proliferating or mitotic HeLa cells were incubated with purified recombinant His6–ERK3–GST and 200 μM ATP (Figure 3A). The reaction mixture was separated by SDS/PAGE, and the band corresponding to ERK3 was excised and digested in-gel with trypsin. The resulting peptides were subjected to LC-MS/MS (liquid chromatography-tandem MS) analysis on a LTQ Orbitrap. The analysis resulted in a coverage of 68% of the protein sequence. We confirmed the identification of Ser189 and Ser386 as two phosphorylation sites of ERK3 [7,11,12], thereby validating our MS analysis, and showed that these two residues are phosphorylated both by asynchronous and mitotic extracts. In addition, we identified four new phosphorylation sites, Ser684, Ser688, Thr698 and Ser705, all located in the C-terminal extension of ERK3, that are phosphorylated by mitotic extracts (Figure 3B). Of these four sites, only Ser684 was found to be phosphorylated by an extract of asynchronous cells in one single experiment (Figure 3B).
To verify whether these four C-terminal residues contribute to the mobility shift of ERK3 observed in mitosis, we transfected HeLa cells with wild-type GFP–ERK3 or a mutant where all four sites were substituted by alanine residues, GFP–ERK3(S684A/S688A/T698A/S705A) (referred to as C4A). Both GFP–ERK3 and the GFP–ERK3C4A mutant migrated similarly in lysates of exponentially proliferating cells (Figure 3C). However, their migration pattern was different in mitotic cell lysates. The upward mobility shift of the ERK3 C4A mutant band was less pronounced as compared with the wild-type protein (Figure 3C). This indicates that one or more of the four C-terminal residues identified by MS contribute to the hyperphosphorylation of ERK3 in mitosis. It should be noted that the migration of the C4A mutant was still slightly retarded in mitosis, suggesting the existence of additional phosphorylation sites (Figure 3C).
To determine whether Ser684, Ser688, Thr698 and Ser705 are phosphorylated in vivo, GFP–ERK3–FLAG was transiently expressed in HeLa cells and the cells were synchronized in mitosis as above. The ectopically expressed ERK3 was immunoprecipitated with an anti-FLAG antibody, separated by SDS/PAGE and the silver-stained band corresponding to ERK3 was processed for LC-MS/MS analysis (Supplementary Figure S1 at http://www.BiochemJ.org/bj/428/bj4280103add.htm). These analyses confirmed that Thr698 and Ser705 of ERK3 are phosphorylated in mitosis in intact cells. To facilitate the detection of phosphorylated Thr698 and Ser705, and to study their regulation in vivo, we generated phospho-specific antibodies against these two residues. Unfortunately, the antibody raised against phospho-Ser705 of ERK3 failed to detect any specific immunoreactive band in cells. However, immunoblotting with the anti-phospho-ERK3 (Thr698) antibody recognized ectopic wild-type GFP–ERK3, but not the T698A mutant, in lysates of mitotic cells (Figure 3D). As expected, no immunoreactivity could be detected in lysates of asynchronously proliferating cells. To confirm that the endogenous ERK3 protein is phosphorylated on Thr698, we immunoprecipitated ERK3 from asynchronous or nocodazole-arrested HeLa cells and analysed its phosphorylation by immunoblotting with an anti-phospho-ERK3 (Thr698) antibody. The antibody recognized an endogenous band of ~110 kDa that co-migrated with the ERK3 protein and was detected specifically in mitotic cells (Figure 3E). The band was not detected with a pre-immune serum.
Sequence alignment revealed that the four C-terminal serine/threonine sites identified are conserved in mammals (Figure 3F). The amino acid sequences surrounding these residues are also perfectly conserved, suggesting an important function.
Identification of cyclin B-Cdk1 as a candidate for the C-terminal phosphorylation of ERK3 in mitosis
The four residues Ser684, Ser688, Thr698 and Ser705 are all followed by a proline residue and lie within a minimal Cdk consensus site . The activity of the kinase cyclin B-Cdk1 peaks in mitosis and is required for normal entry and progression into mitosis . To investigate whether ERK3 is a candidate substrate for cyclin B-Cdk1, we first tested the ability of purified active cyclin B-Cdk1 to phosphorylate recombinant His6–ERK3–GST in vitro. As previously reported , we found that active cyclin B-Cdk1 is able to catalyse the phosphorylation of ERK3 in vitro (Figure 4A). Mutation of the four serine/threonine-proline sites (ERK3C4A) resulted in a complete loss of the phosphorylation signal, confirming that these sites are the targets of the mitotic kinase. To address the role of Cdk1 in the in vivo phosphorylation of ERK3, HeLa cells were co-transfected with GFP–ERK3 and different cyclin regulatory subunits. The C-terminal phosphorylation of ERK3 was monitored by immunoblotting with the anti-phospho-ERK3 (Thr698) antibody. Expression of the mitotic cyclins A2 and B1, but not cyclins D1 and E1, induced the phosphorylation of ERK3 on Thr698 (Figure 4B). Reciprocally, treatment of mitotic HeLa cells with two distinct pharmacological inhibitors of Cdk1 and Cdk2, roscovitine and SU9516, resulted in a marked decrease in the extent of Thr698 phosphorylation (Figure 4C). Together, these data strongly suggest that cyclin B-Cdk1 is a bona fide physiological kinase of ERK3 in mitosis.
Cdc14A and Cdc14B dephosphorylate ERK3 at mitosis exit
The results presented above show that ERK3 is hyperphosphorylated on Cdk sites located in the C-terminal extension in mitosis. Upon mitosis exit and entry into G1, ERK3 migrates more rapidly on the gel, indicating that it is dephosphorylated as cells enter the next cell cycle (Figure 1C). Mitotic exit is controlled in part by the activity of protein phosphatases, among which Cdc14 plays a key role [20,21]. Cdc14 is a proline-directed phosphatase that opposes Cdk action to regulate various cell-cycle events. Mammalian cells express two Cdc14 homologues, termed Cdc14A and Cdc14B, whose specific substrates are poorly documented . In a recent study, Hansen et al.  reported that Cdc14A physically interacts with ERK3. However, the authors did not examine the functional consequence of this interaction on the phosphorylation of ERK3 in intact cells. We first wanted to determine whether the interaction of Cdc14 with ERK3 is specific to the Cdc14A isoform by co-immunoprecipitation experiments. We found that both Cdc14A and Cdc14B physically interact with ERK3 when co-expressed in HEK-293 cells (Figure 5A). Replacement of the four C-terminal serine/threonine-proline sites of ERK3 by alanine residues or phospho-mimetic aspartate or glutamate residues had no effect on the interaction with Cdc14 isoforms.
We next investigated the role of Cdc14 in the dephosphorylation of ERK3 serine/threonine-proline sites during mitosis exit. HeLa cells were co-transfected with GFP–ERK3 and wild-type or catalytically inactive forms of either FLAG–Cdc14A or FLAG–Cdc14B, and the cells were synchronized in mitosis with nocodazole. Immunoblot analysis revealed that overexpression of Cdc14A and Cdc14B, but not the phosphatase-dead mutants, leads to dephosphorylation of Thr698 and the appearance of a faster-migrating form of ERK3 in mitotic cells (Figure 5B). These results suggest that phosphorylation of ERK3 by cyclin B-Cdk1 during mitosis is antagonized by Cdc14 as the cell exits mitosis.
ERK3 stability is regulated by phosphorylation in mitosis
To determine whether the fluctuations in ERK3 levels observed during mitosis progression results from a change in protein stability, we measured the half-life of the kinase in asynchronous and mitotic HeLa cells by cycloheximide-chase experiments. As previously reported , ERK3 was found to be highly unstable in asynchronously proliferating cells with a half-life of ~22 min (Figures 6A and 6B). However, in prometaphase-arrested cells the half-life of ERK3 markedly increased to ~115 min. The behaviour of ERK3 resembles that of cyclin B1, which is stabilized during mitosis and is degraded as cells progress into G1-phase .
We next asked whether the C-terminal phosphorylation sites identified are involved in the stabilization of the ERK3 protein during mitosis. To address this question, we transfected HeLa cells with the wild-type or C4A mutant FLAG–ERK3 and synchronized the cells in mitosis by a double-thymidine nocodazole block. The ectopically expressed ERK3 protein was immunoprecipitated with anti-FLAG–agarose beads and revealed by immunoblotting with an anti-ERK3 antibody. The half-life of ectopic FLAG–ERK3 in prometaphase-arrested cells was estimated to be ~102 min, which is essentially similar to the endogenous protein. In these conditions, the ERK3 C4A mutant was degraded much more rapidly with a half-life of ~20 min. We conclude from these results that phosphorylation of the four C-terminal serine/threonine-proline sites stabilizes ERK3 protein, leading to its accumulation during mitosis.
Signalling by various subfamilies of MAPKs has been shown to play an important role in the regulation of cell-cycle progression. Sustained activation of the MAPKs ERK1/ERK2 is necessary for G1- to S-phase progression, and is associated with induction of positive regulators of the cell cycle and inactivation of antiproliferative genes [24,25]. Activation of p38 MAPK by DNA damage or various stresses triggers a G2 checkpoint that delays entry into mitosis [26,27]. JNK (c-Jun N-terminal kinase) 1 and JNK2 MAPKs positively regulate cell proliferation, at least in part, by augmenting c-Jun expression . More recently, ERK5 was shown to promote G2-M progression by regulating the transcription of G2/M-specific genes through activation of NF-κB (nuclear factor κB) . Mounting evidence also links ERK3 signalling to the regulation of cell-cycle progression. The results of the present study further substantiate this idea by demonstrating that ERK3 is phosphorylated and stabilized in a cell-cycle-dependent manner. We show that ERK3 is stoichiometrically hyperphosphorylated as cells enter mitosis and is dephosphorylated at the M→G1 transition.
By MS analysis, we identified four residues (Ser684, Ser688, Thr698 and Ser705) in the unique C-terminal extension of ERK3 that are specifically phosphorylated by mitotic cell extracts in vitro. All four residues are followed by a proline residue and therefore represent candidate Cdk substrates. In the present study, we show compelling evidence that cyclin B-Cdk1 is a major ERK3 C-terminal kinase in mitosis. This conclusion is supported by the following observations: (i) active cyclin B-Cdk1 phosphorylates recombinant ERK3 in vitro and alanine mutation of the four serine/threonine-proline sites completely abolishes phosphorylation; (ii) ERK3 is hyperphosphorylated when the activity of Cdk1 is highest in the cell cycle; (iii) immunoblotting with a phospho-specific antibody against Thr698 recognizes an immunoreactive band that co-migrates with ERK3 specifically in mitotic cells; (iv) ectopic expression of mitotic cyclins induces the phosphorylation of ERK3 on Thr698; and (v) pharmacological inhibition of Cdk1 during mitosis results in a marked decrease in the phosphorylation of ERK3 on Thr698. This identifies ERK3 as a novel substrate of Cdk1. However, we do not exclude the possibility that ERK3 is phosphorylated by another mitotic kinase. Indeed, migration of an ERK3 mutant with all four C-terminal serine/threonine-proline sites replaced by alanine still exhibited a slight mobility shift in mitosis, suggesting the existence of additional phosphorylation sites. This idea is consistent with the results of in vitro phosphorylation of ERK3-deletion mutants, showing that the electrophoretic mobility of the 1–678 and 1–542 truncation mutants is still shifted upon phosphorylation by mitotic extracts (Figure 2C). Future identification of these additional phosphorylation sites of ERK3 will be important to understand the function of the kinase in cell-cycle regulation.
Cdc14 is an evolutionarily conserved serine/threonine phosphatase that regulates various aspects of the cell-division cycle [20,21,30]. Mammalian cells express two Cdc14 homologues, named Cdc14A and Cdc14B, whose specific substrates and functions remain to be fully characterized. These phosphatases have been implicated in the regulation of the centrosome duplication cycle, chromosome segregation, cytokinesis, G1 length and the G2 DNA-damage checkpoint [15,31–33]. Genetic, biochemical and structural studies have shown that Cdc14 dephosphorylates phospho-serine/threonine residues immediately followed by a proline residue and opposes Cdk action [32,34,35]. To date, Cdc14A has been reported to dephosphorylate Cdc25A, Sirt2 (sirtuin 2), RN-Tre, iron regulatory protein 2, p53 and INCENP (inner centromere protein), whereas Cdc14B was shown to dephosphorylate Skp2, Sirt2 and p53 [15,36–41]. A recent study has shown that ERK3 can be dephosphorylated by Cdc14A in vitro . In the present study, we provide evidence that ERK3 is a substrate of Cdc14A and Cdc14B in intact cells. We show that ERK3 physically interacts with both Cdc14A and Cdc14B homologues in co-immunoprecipitation experiments. Most importantly, we show that overexpression of Cdc14A or Cdc14B, but not their catalytically inactive forms, dephosphorylate ERK3 in mitotic cells. These results add ERK3 to the growing list of Cdc14 substrates.
Cdc14 phosphatase isoforms exhibit distinct subcellular localizations during interphase and mitosis. Cdc14A is localized to the centrosome during interphase, and is redistributed to the cytoplasm during mitosis . Cdc14B is mainly localized to nucleoli in interphase cells, and is dispersed throughout the cell in prophase and metaphase [32,42]. We previously reported that ERK3 is found in both the cytoplasmic and nuclear compartments of exponentially proliferating cells . Interestingly, we observed that ERK3 localizes at the centrosome when co-expressed with Cdc14A, and is relocalized in the nucleus when co-transfected with the Cdc14B isoform (results not shown). The physiological significance of these observations for the regulation of ERK3 activity requires further investigation.
Why is ERK3 specifically stabilized in mitosis? The answer to this question will await the elucidation of ERK3 cellular functions. The high conservation of the C-terminus of ERK3 suggests that this region is likely to play an important functional role. Interestingly, ERK3 was previously identified as a potential regulator of mitosis in HT-29 colon cancer cells in a lentiviral-based shRNA (small-hairpin RNA) genetic screen . In this high-content screen, ERK3-knockdown resulted in an increase in mitotic index. Additional work is needed to connect this observation with our finding of ERK3 hyperphosphorylation and accumulation during mitosis.
Pierre-Luc Tanguay designed the study, carried out the experiments, analysed the data and drafted the manuscript. Geneviève Rodier contributed to the design of the study and to data analysis. Sylvain Meloche contributed to study design and data analysis and prepared the final manuscript. All authors read and approved the manuscript.
This work was supported by the Canadian Institutes of Health Research (CIHR) [grant number MOP93729 (to S.M.)]. P.-L.T. is the recipient of a studentship from the CIHR and G.R. is the recipient of fellowhips from the National Cancer Institute of Canada and the American Association for Cancer Research. S.M. holds the Canada Research Chair in Cellular Signaling.
We thank E. Bonneil and P. Thibault from the IRIC (Institut de Recherche en Immunologie et Cancérologie) Proteomic Facility for MS analysis, Dr M. Ewen and Professor H. Charbonneau for plasmids, and C. Julien for technical assistance.
Abbreviations: Cdc, cell-division cycle; Cdk, cyclin-dependent kinase; DTT, dithiothreitol; ERK, extracellular-signal-regulated kinase; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GFP, green fluorescent protein; GST, glutathione transferase; HEK, human embryonic kidney; JNK, c-Jun N-terminal kinase; LC-MS/MS, liquid chromatography-tandem MS; MAPK, mitogen-activated protein kinase; MK5, MAPK-activated protein kinase 5; Sirt2, sirtuin 2
- © The Authors Journal compilation © 2010 Biochemical Society