Our original goal was to carry out high throughput screens using siRNAs against the human kinome to identify novel protein kinases that regulate Cdc25A stability and, in particular, to identify the protein kinase that phosphorylates Cdc25A on serine 82. With this goal in mind, stable cell lines that inducibly express a fusion protein between human Cdc25A and firefly luciferase (Cdc25A-FLuc) were generated to enable direct real-time monitoring of Cdc25A protein levels in cells. The promoter driving Cdc25A-FLuc expression contains a tetracycline response element regulated by rtTA (reverse tetracycline-controlled transactivator) and doxycycline (Dox), a tetracycline derivative. Expression of the Cdc25A-FLuc fusion protein was induced by addition of Dox to the culture media, and the level of Cdc25A-FLuc protein correlated with its luciferase activity (Fig. 1A, B). Furthermore, Cdc25A-FLuc protein levels were stabilized in cells incubated with either a proteosome inhibitor (MG132) or Chk1 inhibitors (UCN-01, Gö6976 and AZD7762) (Fig. 1C and data not shown). This data suggested that the Cdc25A-FLuc fusion protein was a valid substrate with which to identify novel Cdc25A regulatory kinases. As further controls, cells were incubated with siRNAs specific for protein kinases known to negatively regulate Cdc25A stability including Chk1 and GSK-3β (Fig. 2A). As expected, stabilization of the Cdc25A-FLuc reporter protein was observed under these conditions. As a negative control, cells were incubated with siRNAs specific for CK1α as CK1 inhibition was reported to have no effect on Cdc25A stability (Jin et al., 2003). Unexpectedly, enhanced bioluminescence indicative of Cdc25A-FLuc stabilization was observed in cells treated with CK1α-specific siRNAs (Fig. 2A). CK1 had been ruled out as a potential Cdc25A-regulatory kinase based on experiments performed with the casein kinase I inhibitors CKI-7 and IC261 (Jin et al., 2003).

Therefore, we tested CK1 inhibitors for their ability to stabilize Cdc25A-FLuc protein. D4476 was chosen because it is a more potent and selective CK1 inhibitor than either IC261 or CKI-7 (Bain et al., 2007; Rena et al., 2004) and IC261 was also tested. IC261 shows selectivity for CK1δ and CK1ε over CK1α (Behrend et al., 2000). As seen in Fig. 2B, enhanced bioluminescence indicative of Cdc25A-FLuc stabilization was observed in D4476-treated cells but not in IC261-treated cells. These results suggest that CK1 family members other than CK1δ and CK1ε may function to negatively regulate Cdc25A stability in vivo. To test this hypothesis more directly, levels of endogenous Cdc25A were monitored in CK1α-depleted cells. As seen in Fig. 2C, an approximate 2-fold increase in endogenous Cdc25A protein was observed in CK1α-depleted cells. In addition, the half-life of Cdc25A was increased in cells knocked down for CK1α (Fig. 2D). In contrast, endogenous Cdc25A was not stabilized in cells depleted for CK1δ, ε or γ1 (Fig. S3).

A domain bordered by amino acids 76 to 88 has been shown to regulate the ubiquitin-mediated proteolysis of Cdc25A during interphase and to contain a novel β-TrCP recognition motif or phosphodegron (Fig. 3A) (Busino et al., 2003; Jin et al., 2003). The novel phosphodegron in Cdc25A utilizes pS79 and pS82, rather than pS82 and pS88, for β-TrCP binding (Jin et al., 2003). Six potential phosphorylation sites reside within this 14 amino acid domain. Cdc25A has been shown to be phosphorylated on S76 and T80 and phosphorylation of S79, S82 and S88 have been inferred based on mutagenesis studies and phosphopeptide competition experiments (Goloudina et al., 2003; Hassepass et al., 2003; Jin et al., 2003; Kang et al., 2008). Antibodies specific for known and predicted phosphorylation sites were generated (Fig. 3B) and used to demonstrate that Cdc25A is phosphorylated on S79, S82 and S88 in addition to S76 and T80 (Fig. 3C). Specificity was verified by testing each phospho-specific antibody for its ability to recognize WT Cdc25A but not the corresponding alanine-mutant protein by Western blotting (Fig. 3C). It was also determined that neither S79 phosphorylation nor substitution of alanine for serine at position 79 interfered with the ability of the phospho-S82 antibody to recognize Cdc25A when it is phosphorylated on S82 (Fig. S1) and neither S76 phosphorylation nor substitution of alanine for serine at position 76 interfered with the ability of the phospho-S79 antibody to recognize Cdc25A when it is phosphorylated on S79 (Fig. S2).

Kinase assays were performed to determine if CK1 is able to directly phosphorylate Cdc25A on S82 in vitro (Fig. 4A). Wild-type Cdc25A and phosphorylation-site mutants were purified as GST fusion proteins from bacteria and incubated with purifed CK1. Phosphorylation of S82 was monitored by Western blotting with the phospho-S82 antibody. As seen in Fig. 4A, CK1 phosphorylated Cdc25A on S82 (lane 1) and substitution of alanine for serine at position 88 did not effect S82 phosphorylation (lane 3).

If CK1α phosphorylates Cdc25A on S82 in vivo, changes in level or activity of CK1α should affect S82 phosphorylation. Indeed, enhanced S82 phosphorylation of endogenous (Fig. 4B) and ectopic (Fig. S4A) Cdc25A was observed in cells overproducing CK1α. Overproduction of CK1α did not enhance phosphorylation of Cdc25A on S76, T80, S88 or S123 (data not shown). Reduced S82 phosphorylation was detected in cells depleted for CK1α (Fig. S4B). Similarly, inhibition of CK1 kinase activity also decreased S82 phosphorylation of Cdc25A (Fig. S4C). Whereas 50 μM D4476 was sufficient to reduce S82 phosphorylation in vivo, concentrations of IC261 as high as 100 μM had minimal effects on S82 phosphorylation. Conversely, S82 phosphorylation was not affected in cells treated with either TBB (CK2 inhibitor) or KN-93 (CaMKII inhibitor) (Fig. S4D).

S82 phosphorylation has been shown to mediate β-TrCP binding, which in turn, promotes ubiquitin-mediated proteolysis of Cdc25A (Busino et al., 2003; Jin et al., 2003). Thus, CK1α overproduction was expected to enhance interactions between Cdc25A and β-TrCP. As seen in Fig. 4C, an increase in association between β-TrCP and Cdc25A was observed in CK1α-overexpressing cells comparing lanes 3 and 4. Enhanced β-TrCP binding is predicted to promote Cdc25A turnover and indeed, lower levels of Cdc25A were also observed in CK1α-overproducing cells (Fig. 4D) and CK1α knockdown increased the half-life of Cdc25A (Fig. 2D). Taken together, these results provide strong evidence that CK1α directly regulates phosphorylation of Cdc25A on S82 and promotes β-TrCP binding in vivo.

Ionizing radiation has been shown to increase the phosphorylation of Cdc25A on S82 (Busino et al., 2003), which in turn promotes Cdc25A proteolysis. We asked if inhibition of CK1 would reduce S82 phosphorylation and extend the half-life of Cdc25A in irradiated cells. As seen in Fig. 4E, S82 phosphorylation of Cdc25A was significantly reduced after D4476 treatment in both mock-irradiated (lane 2) and irradiated (lane 4) cells. D4476 did not affect levels of S76 or S88 phosphorylation under either condition. These experiments were conducted in the absence of cycloheximide so levels of ectopic Cdc25A did not change significantly upon irradiation. In the presence of cycloheximide, D4476 was observed to extend the half-life of both endogenous and ectopic Cdc25A (Cdc25A-FLuc) in irradiated cells (Fig. 4F). These results indicate that CK1 regulates S82 phosphorylation in normal cycling cells and in cells experiencing genotoxic stress.

CK1, like GSK-3β, requires a priming phosphate in its substrates, which is introduced by a priming kinase. CK1 often phosphorylates the serine residue (underlined) in the motif pS-X-X-S where pS is phosphoserine (Flotow et al., 1990). Phosphorylation of S79 could serve as a priming phosphorylation site for S82 phosphorylation and we determined that S79 is phosphorylated in vivo (Fig. 3C). Interestingly, sequences inclusive of and surrounding S79 also conform to one of the CK1 phosphorylation consensus motifs (Fig. 3A). Thus, kinase assays were performed to determine if CK1α was also capable of phosphorylating S79. As seen in Fig. 5A, CK1α phosphorylated Cdc25A on both S79 and S82 by not S88 in vitro. Importantly, CK1α was unable to phosphorylate S82 when serine 79 was changed to alanine.

To determine if CK1α also regulates S79 phosphorylation in vivo, CK1α was overproduced with Flag-Cdc25A and as seen in Fig. 5B, enhanced phospho-S79 levels was observed in the presence of ectopic CK1α. In addition, D4476-treatment resulted in reduced S79 phosphorylation in vivo (Fig. 5C). These results indicate that CK1α phosphorylates Cdc25A on both S79 and S82 in vivo. Taken together, our data is supportive of a model whereby S76 phosphorylation by either Chk1 or GSK-3β primes Cdc25A for phosphorylation on S79 by CK1α and this, in turn, primes Cdc25A for S82 phosphorylation also by CK1α.

HeLa cells were cultured in Dulbecco's modified Eagle's medium (DMEM, Invitrogen) supplemented with 10% bovine growth serum (BGS, Hyclone), 1 mM glutamine, and 100 U/ml each of penicillin and streptomycin. HeLa Tet-on cells (BD Biosciences-Clontech) were cultured in DMEM supplemented with 10% Tet system approved FBS (Clontech), 1 mM glutamine, 100 U/ml each of penicillin and streptomycin, and Geneticin (Clontech, 100 μg/ml). HeLa Tet-on Cdc25A-FLuc stable cells were cultured in the above media plus hygromycin B (200 μg/ml, Invitrogen). MDA-MB-231 cells were cultured in DMEM supplemented with 10% BGS, 1mM glutamine, 1mM sodium pyruvate, and 100 U/ml penicillin and streptomycin. MG132 and cycloheximide (Sigma Chemical Co.) were dissolved in DMSO and 1× PBS respectively. Kinase inhibitors used in this study include D4476, IC261, Gö6976, TBB and KN-93 (Calbiochem), UCN-01 (Sigma Chemical Co.) and AZD7762 (Zabludoff et al., 2008) (Axon Medchem BV). D4476 was prepared as described (Rena et al., 2004).

Monoclonal anti-Cdc25A antibodies (Ab-3, Neomarkers, and Ab2357, Abcam) were used to detect human Cdc25A. Other antibodies used in this study include: anti-actin (Sigma Chemical Co), anti-Myc (A-14 and 9E-10, Santa Cruz Biotechnology), anti-β-Catenin (BD Transduction), anti-CK1α (C-19, Santa Cruz Biotechnology), anti-phospho-S45 β-Catenin (Biosource), anti-Chk1 (G4, Santa Cruz Biotechnology), and anti-pS317 Chk1 (Cell Signaling Technology). Antibodies specific for Cdc25A phosphorylated on T80 and S123 were generated by immunizing rabbits with the coupled phosphopeptides C-GSSES-pT-DSGFC and C-LKRSH-pS-DSLD, respectively (Kang et al., 2008). Antibodies specific for Cdc25A phosphorylated on S76 (C-NSNLQRMG-pS-SEST), S79 (MGSSE-pS-TDSGFC), S82 (C-SESTD-pS-GFCLD) and S88 (C-GFCLD-pS-PGPLD) were generated by immunizing rabbits with the indicated phosphopeptides followed by affinity purification (Abgent). Bound primary antibodies were reacted with horseradish peroxidase (HRP)-conjugated goat anti-mouse antibody (Jackson), HRP-goat-anti-rabbit antibody (Zymed), or HRP-rabbit anti-goat antibody (Zymed) and detected by ECL (GE Healthcare).

Cell lysates prepared from control cells and cells overproducing Flag-Cdc25A(WT) were resolved by SDS-PAGE and transferred to nitrocellulose following by incubation in Blocking buffer (5% milk/1× TBST). HPLC purified peptides (Tufts University Core Facility, Boston, MA) were incubated with 2 μg/ml of phospho-antibodies at concentrations between 4 -40 μg/ml in 1× TBST overnight at 4°C. Western blotting was then performed.

Generation of pcDNA3-Myc-Cdc25A-FLuc: the stop codon of Cdc25A was eliminated and an Xba I site was inserted at the C-terminus of Cdc25A. The resulting plasmid pcDNA3-Myc-Cdc25A (Xba I) was digested with Xba I and Xho I and ligated to sequences encoding codon-optimized firefly luciferase (FLuc) that had been isolated as an Xba I-Xho I fragment from the pGL-3 control plasmid (Promega). Sequences encoding Cdc25A-FLuc but lacking the Myc-tag were isolated from pcDNA3-Myc-Cdc25A-FLuc and subcloned into pTRE-Tight, a tetracycline-regulated expression plasmid (Clontech) to generate pTRE-Tight-Cdc25A-FLuc. Cdc25A serine-to-alanine mutants were generated using the QuikChange XL site-directed mutagenesis kit (Stratagene) and all the mutations were verified by DNA sequencing. Plasmids encoding CK1α and Myc-β-TrCP1 were provided by Dr. Jiandong Chen (Chen et al., 2005) and Dr. Binhua P. Zhou (Zhou et al, 2004), respectively.

pTRE-Tight-Cdc25A-FLuc was co-transfected with a selection plasmid pTK-Hyg (Clontech) into HeLa Tet-on cells using Lipofectamine 2000 (Invitrogen). Twenty-four hours later, cells were trypsinized and seeded into 10-cm tissue culture dishes. Selection media containing hygromycin B was added 24 h later. Media was changed every 3 days until the drug-resistant colonies became visible. Cells were incubated with D-Luciferin (150 μg/ml in PBS, BIOSYNTH International) at 37°C for 10 min followed by imaging using an IVIS 100 imaging system (Xenogen). Luciferase activity (photons/sec/cm²/steradian) was quantitated using Living Image Software (Xenogen). Those colonies that scored positive for bioluminescence were eliminated. Cells were then rinsed in PBS and incubated in culture media containing 2 μg/ml of doxycycline (Dox, Clontech). Twenty-four hours later, D-Luciferin was added to the media and colonies were re-imaged. Colonies that displayed Dox-inducible bioluminescence were isolated and expanded for further analysis. The stable cell line was maintained in selection medium containing hygromycin B and geneticin.

Wild-type and mutant forms of Cdc25A were produced in bacteria as GST-fusion proteins and purified as described previously (Chen et al., 2003). The elution of GST-Cdc25A proteins was performed as described (Kang et al., 2008). 100 ng CK1 (GST-CK1α or GST-CK1δ from Sigma Chemical Co.) was diluted in kinase reaction buffer (50 mM Tris-HCl, pH 7.4, 2 mM DTT, 10 mM MgCl₂ and 120 μM ATP) followed by the addition of 1 μg soluble GST-Cdc25A. Reactions were incubated at 30°C for 30 min, boiled in 1× loading buffer at 95°C for 10 min and then resolved by SDS-PAGE followed by Western blotting.

Approximately 1 × 10⁶ HeLa cells were seeded in p60 plates 16-20 h before transfection. Cells at 95% confluence were transfected with 2 to 4 μg plasmids using Lipofectamine 2000. 24 h after transfection, cells were harvested or subjected to the indicated treatments prior to harvest and analysis. For RNAi transfection, HeLa cells (3× 10⁵) were seeded in p60 dishes 20 h before transfection. Cells were transfected with 50 nM CK1α, GSK-3β (Dharmacon) or Chk1 siRNA (Zhao et al., 2002) when they reached 20-30% confluency using DharmaFECT 1 or 3 (Dharmacon). Forty-eight hours after transfection, cells were treated as indicated and then harvested for Western blotting. Scrambled siRNA or luciferase siRNA (GL3, Dharmacon) was used as transfection controls. For half-life experiments, siRNA-transfected cells were incubated with 10 μg/ml cycloheximide 48 h after siRNA-transfection. Cells were harvested at various times after cycloheximide treatment and analyzed by Western blotting. MG132 additions were made 48 h after siRNA transfection and cells were harvested 2 h later.

Cells were lysed in mammalian cell lysis buffer (MCLB) [50 mM Tris-HCl pH 8.0, 2 mM DTT, 5 mM EDTA, 0.5% NP-40, 100 mM NaCl, 1 mM microcystin (Sigma Chemical Co), 1 mM sodium orthovanadate, 2 mM PMSF] supplemented with protease inhibitor cocktail (Sigma Chemical Co.) and phosphatase inhibitor cocktail (Calbiochem). Clarified lysates were resolved directly by SDS-PAGE and transferred to nitrocellulose membranes for Western blotting. Alternatively, cell lysates containing 0.5 to 1 mg of total cell protein were pre-cleared with protein A beads (Santa Cruz Biotechnology) for 1 h at 4°C. Pre-cleared lysates were incubated with anti-Flag agarose (Sigma Chemical Co.) or anti-Myc agarose (9E-10, Santa Cruz Biotechnology) for 1 h at 4°C. Immunoprecipitates were then washed 6 times with MCLB. Proteins were visualized using the ECL reagent (GE Healthcare Life Sciences), and densitometry analysis was performed using ImageJ software (Abramoff et al., 2004).