The Cyclin-dependent Kinase Inhibitor p27Kip1 Is Stabilized in G0 by Mirk/dyrk1B Kinase*

Elevated levels of the cyclin-dependent kinase (CDK) inhibitor p27 block the cell in G0/G1 until mitogenic signals activate G1 cyclins and initiate proliferation. Post-translational regulation of p27 by different phosphorylation events is critical in allowing cells to proceed through the cell cycle. We now demonstrate that the arginine-directed kinase, Mirk/dyrk1B, is maximally active in G0 in NIH3T3 cells, when it stabilizes p27 by phosphorylating it at Ser-10. The phospho-mimetic mutant p27-S10D was more stable, and the non-phosphorylatable mutant p27-S10A was less stable than wild-type when expressed in G0-arrested cells. Following phosphorylation by Mirk, p27 remains a functional CDK inhibitor, capable of binding to CDK2. Mirk did not induce the translocation of p27 from the nucleus in G0, but instead co-localized with nuclear p27. Depletion of Mirk by RNA interference decreased the phosphorylation of p27 at Ser-10 and the stability of endogenous p27. RNAi to Mirk increased cell entry from G0 into G1 as shown by increased expression of proliferating cell nuclear antigen and decreased expression of p27. These data suggest a model in which Mirk increases the amount of nuclear p27 by stabilizing it during G0 when Mirk is most abundant. Mitogen stimulation then causes cells to enter G1, reduces Mirk levels (Deng, X., Ewton, D., Pawlikowski, B., Maimone, M., and Friedman, E. (2003) J. Biol. Chem. 278, 41347-41354), and initiates the translocation of p27 to the cytoplasm. In addition, depletion of Mirk by RNAi in postmitotic C2C12 myoblasts decreased protein but not mRNA levels of p27, suggesting that stabilization of p27 by Mirk also occurs during differentiation.

The cyclin-dependent kinase inhibitor p27 (1, 2) plays a critical role in regulating progression through the cell cycle. Rising levels of p27 block the cell in G 0 /G 1 until mitogenic signals activate G 1 cyclins and initiate proliferation. Posttranslational regulation of p27 by different phosphorylation events is critical in allowing cells to proceed through the cell cycle. CDK2/cyclin E phosphorylates p27 at Thr-187 (3,4) and promotes its degradation by the ubiquitin-proteasome pathway at the G 1 -S checkpoint (5). Mitogenic signals also induce the kinase KIS, which mediates phosphorylation of p27 at Ser-10 (6), facilitating the translocation of p27 to the cytoplasm at later times in G 1 (7,8). However, p27 has been shown to be already phosphorylated at Ser-10 in G 0 /early G 1 before KIS induction (7,8). Furthermore, p27 is solely nuclear in G 0 and does not even bind to its exportin CRM1 until cells enter G 1 following mitogen stimulation (9), implying that phosphorylation at Ser-10 may fulfill another function. In this study we identify the arginine-directed serine/threonine kinase Mirk/ dyrk1B as a kinase active in G 0 which phosphorylates p27 at Ser-10 without inducing its translocation from the nucleus.
Mirk/dyrk1B is a member of the Dyrk/minibrain family of arginine-directed serine/threonine kinases. Mirk is transcriptionally up-regulated by Rho family members (10) and itself functions as a transcriptional activator (11) under control of the MKK3/p38 MAPK signaling system (11,12). Mirk is expressed at elevated levels in some tumor cells and in normal skeletal muscle (13) where it is active in muscle differentiation (10). However, Mirk is widely expressed at low levels, suggesting it performs some general role. In the current study we demonstrate a novel function for Mirk, stabilization of the CDK inhibitor p27 kip1 during the G 0 phase of the cell cycle.

EXPERIMENTAL PROCEDURES
Materials-Antibody to p27 Kip1 was from Transduction Laboratories; antibodies to PCNA, 1 CDK2, ␤-tubulin, and cyclin A were from Santa Cruz Biotechnology, and antibody to the FLAG-epitope was from Sigma. Antibody to the Ser-10-phosphorylated form of p27 was purchased from Zymed Laboratories Inc.. Affinity-purified rabbit polyclonal antibody to a unique sequence at the C terminus of Mirk was raised as described (13). In some experiments affinity-purified rabbit polyclonal antibody to the unique N terminus of Mirk was used (13). GST-ERK2 was purchased from Upstate Biotechnology, inc. Polyvinylidene difluoride transfer paper Immobilon-P was purchased from Millipore. PLUS reagent and LipofectAMINE were from Invitrogen. All radioactive materials were purchased from PerkinElmer Life Sciences, ECL reagents from Amersham Biosciences, and tissue culture reagents from Mediatech (Fisher). All other reagents were from Sigma.
Cell Culture-C2C12 mouse myoblasts and NIH3T3 cells were obtained from the ATCC. NIH3T3 cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal calf serum. C2C12 cells were maintained in growth medium: Dulbecco's modified Eagle's medium containing 20% fetal calf serum, and switched to differentiation medium/Dulbecco's modified Eagle's medium containing 2% horse serum. NIH3T3 and C2C12 cells were transfected with LipofectAMINE/PLUS according to the supplier's manual.
Phosphatase Treatment of Phosphorylated p27-FLAG-tagged p27 was immunoprecipitated from total cell lysates, and the immunocomplex was treated with 1 unit of alkaline phosphatase (Promega) in the presence of protease inhibitors (15).
In Vitro Kinase Assay-The immunoprecipitates were washed five times with lysis buffer and three times with kinase buffer (50 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 0.5 mM DTT), incubated for 15 min at 30°C with 20 l of kinase buffer containing 50 M cold ATP plus 5 Ci of [␥-32 P]ATP and 1 g of purified recombinant p27 protein as substrate, and then analyzed by PAGE and autoradiography. TNT products were produced as detailed previously (13).
GST Fusion Proteins-Mirk and the p27 constructs were subcloned into the pGEX 4T1 vector (Amersham Biosciences) and expressed and purified as described (16). Briefly, 300 ml of LB media was seeded with 10% of overnight culture of JM109 cells carrying the desired expression plasmid and incubated at 37°C for 1-2 h. Protein expression was induced with 0.1 mM isopropyl-1-thio-␤-D-galactopyranoside for 6 h at 24°C, and bacteria were collected by centrifugation and frozen at Ϫ70°C. Pellets, averaging 2 g each, were resuspended in 5 ml of ice-cold STE (10 mM Tris, pH 8.0, 150 mM NaCl, 1 mM EDTA) containing 100 g/ml lysozyme and protease inhibitors (Complete Mini-Tabs Protease Inhibitor Mixture Set, Roche Applied Science) and incubated on ice for 15 min. DTT was added to a final concentration of 5 mM and then Sarkosyl was added to a final concentration of 1.5% (w/v). Bacteria were lysed by sonication using a microtip (Misonix, setting 2.5) for 3 min, using 5-s bursts separated by 5 s of rest. Lysates were clarified by centrifugation at 18,000 ϫ g for 20 min. 5 mM phenylmethylsulfonyl fluoride and an additional 1ϫ of protease inhibitor mixture were added to the clarified lysates, and 13-ml portions were incubated with 1 ml (bed volume) of glutathione-Sepharose 4-B (Amersham Biosciences) at 4°C overnight with end-on rotation. Beads were washed six times with 10 volumes of ice-cold PBS (140 mM NaCl, 2.7 mM KCl, 10 mM NaH 2 PO 4 , 1.8 mM KH 2 PO 4 , pH 7.3). GST fusion proteins used for in vitro kinase assays were eluted from the beads by overnight incubation at 4°C with 1 ml of glutathione elution buffer (75 mM HEPES, pH 7.4, 150 mM NaCl, 10 mM reduced glutathione, 5 mM DTT, 0.1% Triton X-100) per ml of bead volume used. Recombinant p27 used in immune complex kinase reactions was cleaved from GST by incubation with thrombin (Amersham Biosciences, 50 units per ml of bead volume in PBS) for 18 h at room temperature. Thrombin was removed from the preparation by incubation with 1 ml of benzamidine-Sepharose 6-B (Sigma) for 1 h at room temperature.
Immunodetection-Following treatment as indicated and washing twice with PBS, cells were lysed in boiling SDS-PAGE sample buffer (50 mM Tris, pH 6.8, 2% SDS, 10% glycerol). Lysates were boiled again for 5 min and vortexed vigorously. Protein determinations were made using Coomassie protein assay reagent from Pierce. Depending on the experiment, 30 -50 g of cell lysate were blotted onto polyvinylidene difluoride membranes after separation by SDS-PAGE. The blots were blocked in 5% milk in TBST (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, and 0.1% Tween 20) for 1 h at room temperature and incubated for 1 h at room temperature with primary antibody in TBST buffer, 3% milk, and proteins were subsequently detected by enhanced chemiluminescence.
Northern Analysis-20 g of total RNA from each cell line was electrophoresed in a 1.1% agarose-formaldehyde gel, transferred to nylon membranes by downward capillary transfer, and cross-linked by baking. The membranes were hybridized to a labeled p27 cDNA probe containing 600 bp of p27 coding sequence in pCMV5 (1). Probes were labeled with 32 P by random priming. The blot was hybridized overnight at 68°C with at least 10 7 cpm of the labeled probe, washed at room temperature twice for 15 min with 1ϫ SSC-0.1% SDS, then washed for 15 min at 65°C in 0.2ϫ SSC, 0.1% SDS, and autoradiographed.
Band Analysis-Immunoblots were scanned using a Lacie Silverscanner, and densitometry was performed using the IP Lab Gel program (Scanalytics).

RESULTS
Mirk Is a G 0 Kinase That Phosphorylates p27 in Vivo-The serine/threonine kinase Mirk/dyrk1B is widely expressed at low levels in normal tissues (13), suggesting that Mirk may play a general role in either cell physiology or cell growth. In addition, Mirk protein levels are known to vary widely in response to treatment of cells both with mitogens and with anti-proliferative stimuli (13), so we determined whether Mirk exhibited any regulation within the cell cycle. NIH3T3 cells arrest in G 0 /G 1 when grown to confluence and move into S phase in a synchronous wave 10 -14 h after release (17). Protein levels of cyclin A were used to monitor cell cycle progression through late G 1 to S. Mirk protein levels were elevated in density-arrested NIH3T3 cells, and these levels fell sharply when cells transited late G 1 and entered S phase (Fig. 1A). Thus Mirk levels were elevated in G 0 /G 1 and were reduced in cycling cells.
Levels of the CDK inhibitor p27 are elevated when cells reach confluence and arrest in G 0 (18,19), the cell cycle phase in which Mirk activity was greatest, so we tested the hypothesis that Mirk phosphorylated p27 in G 0 . Mirk was immunoprecipitated from NIH3T3 growth-arrested cells and from cells at various times after release, and then its capacity to phosphorylate exogenous p27 was determined by an in vitro kinase assay. The greatest phosphorylation of p27 by Mirk was seen in cells in G 0 /G 1 concordant with the highest Mirk levels 0 -6 h post-release ( Fig. 1B). At later time points, cells had undergone the G 1 /S transition as shown by the increase in cyclin A, and the levels and the capacity of Mirk to phosphorylate p27 decreased.
Mirk Phosphorylates p27 in Vitro and in Vivo at Ser-10 -We next determined the specific site at which Mirk phosphorylated p27. Recombinant Mirk is a constitutively active kinase (13), so its capacity to phosphorylate p27 was assayed using in vitro kinase assays. Mirk phosphorylated the CDK inhibitor p27, with kinase-inactive YF-Mirk serving as the negative control ( Fig. 2). In contrast, recombinant p38␣ and JNK1 had little kinase activity on p27 (data not shown). CDK2 phosphorylation at Thr-187 initiated the ubiquitination and rapid degradation of p27 by the proteasome (3, 4), so we tested whether mutation of the CDK-binding site, the Thr-187 phosphorylation site, or the Ser-10 site, which facilitates the translocation of p27 to the cytoplasm (8,20), altered Mirk activity. The double mutant p27-F62A/F64A, a construct deficient in interaction with CDKs (14), a p27 construct mutant at its conserved C-terminal CDK target site (T187V), and a triple mutant (F62A/F64A/T187V) Mirk is a G 0 kinase that phosphorylates the CDK inhibitor p27. A, cell cycle-dependent expression of Mirk. NIH3T3 cells were maintained at confluent density for 3 days and then replated at 1/3 density in fresh medium. Western blotting (WB) for cyclin A and Mirk was performed on parallel cultures at the times indicated. B, phosphorylation of p27 by Mirk at various points in the cell cycle. NIH3T3 cells were synchronized in G 0 /G 1 by growth to confluent density for 3 days and then released by plating at lower cell density in growth media. Endogenous Mirk was immunoprecipitated (IP), and an immune complex kinase assay was performed on recombinant p27 at the indicated times and analyzed by SDS-PAGE and autoradiography (AR). Western blotting for Mirk and p27 was performed on the immunoprecipitates, and for cyclin A on the total cell lysates (lower panel). IVK, in vitro kinase; CB, Coomassie Blue.
were each phosphorylated by Mirk to a similar extent as wildtype p27 (Fig. 2), indicating these sites were irrelevant to Mirk action in vitro. Moreover, Mirk is a highly selective p27-S10 kinase. Mirk was unable to phosphorylate p27-S10A, whereas it phosphorylated wild-type p27, p27-S178A, and the p27 CDK phosphorylation site mutant T187V to a similar extent in in vitro kinase reactions (Fig. 3). In contrast, ERK2 phosphorylated each of the mutant p27 constructs to an equal extent, somewhat less than wild-type p27. These results show that Mirk has a very defined substrate requirement for the Ser-10 residue in p27 and suggest that Mirk is at least one of the kinases responsible for phosphorylation at Ser-10 in vivo.
Ser-10 is the major phosphorylation site for p27 in vivo, accounting for about 70% of the total amount of p27 phosphorylation as determined by metabolic labeling with [ 32 P]orthophosphate (7). Phosphorylation of p27 at Ser-10 has been shown, by use of a p27-S10D phospho-mimetic mutant, to facilitate p27 translocation to the cytoplasm in response to mitogens, whereas mutation of Ser-10 to alanine prevents export to the cytoplasm (8,20). In G 0 , p27 is solely nuclear and does not even bind to its exportin CRM1 until cells enter G 1 following mitogen stimulation (9). These studies indicate that phosphorylation of p27 at Ser-10 is a major control point in the cellular mitogenic response. The mitogen-induced kinase KIS has been identified as one kinase that mediates phosphorylation of Ser-10 in vivo (6). However, p27 has been shown to be phosphorylated already at Ser-10 to some extent in G 0 /early G 1 before KIS induction (7,8). Mirk activity is highest in densityarrested NIH3T3 cells in G 0 (Fig. 1), and it is rapidly downregulated by serum mitogens (13). These properties suggest that Mirk could phosphorylate p27 at Ser-10 during G 0 , before mitogen induction of KIS.
To confirm that Mirk phosphorylates p27 in vivo, either wild-type or kinase-inactive KR-Mirk was co-transfected with either wild-type or p27-S10A FLAG epitope-tagged expression constructs into NIH3T3 cells. Following expression, the con-structs were labeled for 16 h with 32 P i , and then the 32 P-labeled p27 was immunoprecipitated with anti-FLAG antibody and analyzed by PAGE and autoradiography. The amount of 32 P i incorporated into wild-type p27 was severalfold higher when wild-type Mirk was co-expressed (Fig. 4, 3rd lane), whereas little incorporation was seen when wild-type p27 was co-expressed with kinase-inactive Mirk or vector control. In contrast, co-expression of wild-type Mirk did not increase the low basal level of 32 P i incorporated into mutant p27-S10A (Fig. 4, last 3 lanes). Western blotting for transfected p27 using anti-FLAG antibody demonstrated equal expression of all constructs (Fig. 4, lower lanes). These metabolic labeling data confirm that Mirk phosphorylates p27 in vivo at Ser-10.
Phosphorylation can impede protein mobility on SDS-PAGE. Mirk phosphorylation of p27 in vivo reduced its electrophoretic mobility. Constructs for wild-type p27 and either wild-type Mirk, kinase-inactive YF-Mirk, or the pcDNA(HisA) vector were co-transfected into NIH3T3 cells and allowed to express, and the mobility of p27 was determined by Western blotting (Fig. 5A, 1st 3 lanes). Only the p27 co-transfected with wildtype Mirk exhibited a decrease in mobility. When wild-type Mirk was co-expressed with the mutant p27 constructs S10A, T170A, S178A, or T187V, all of the mutants except S10A exhibited the mobility shift. The p27-T187V construct is itself a little longer than wild-type p27 so its shift was confirmed by co-transfection with the HisA epitope-tagged vector (Fig. 5, last  3 lanes on right).
Phosphatase treatment demonstrated that the p27 mobility shift was caused by phosphorylation (Fig. 5B). Immunoprecipitated FLAG epitope-tagged p27 constructs (wild-type or S10A) were incubated with calf intestinal alkaline phosphatase before Western blotting. Treatment with the phosphatase eliminated the p27 mobility shift induced by co-transfected wild-type Mirk, thus demonstrating definitively that Mirk phosphorylates p27 in vivo solely at Ser-10.
p27 Phosphorylated at Ser-10 by Mirk Binds to CDK2-We next determined whether p27 phosphorylated at Ser-10 could function as a CDK inhibitor. FLAG-p27 and Mirk were coexpressed in NIH3T3 cells. Following phosphorylation by Mirk in vivo, p27 was still capable of binding to CDK2 (Fig. 6A). The FLAG-p27 bound to immunoprecipitated CDK2 exhibited decreased mobility on SDS-PAGE following co-expression with Mirk (Fig. 6A, arrows), indicating that it had been phosphorylated. The equal amount of immunoglobulin heavy chain (Hchain) indicates equal amounts of antibody were found in the immunoprecipitates. These data demonstrate that p27 phosphorylated by Mirk in G 0 was still capable of functioning as a CDK inhibitor.
Mirk Phosphorylation of p27 Blocked by RNA i to Mirk-The capacity of endogenous Mirk to phosphorylate p27 at Ser-10 FIG . 2. In vitro phosphorylation of p27 constructs. GST-p27 constructs were prepared of wild-type (WT) p27, p27 doubly mutated at the CDK-binding site (F62A/F64A), p27 mutated at the CDK2 phosphorylation site (T187V), and a p27 construct triply mutated (F62A/F64A/ T187V). Protein preparations of wild-type Mirk and kinase-inactive Mirk were prepared by coupled in vitro transcription and translation, immunoprecipitated, and used in in vitro kinase reactions. The phosphorylated p27 molecules were separated by SDS-PAGE and autoradiographed. Western blotting was performed to confirm that equal amounts of Mirk were present in each reaction mixture, whereas Coomassie Blue (CB) staining confirmed that equal amounts of p27 constructs were tested. Controls included kinase-inactive YF Mirk (5th to 8th lanes) and GST alone (not shown). Mirk protein levels were determined by Western blotting with affinity-purified antibody directed toward a unique sequence within the C terminus of Mirk. Wild-type Mirk was active in each reaction mixture as shown by its autophosphorylation (upper panel) .   FIG. 3. Mirk phosphorylates p27 at Ser-10. GST-p27 constructs were prepared of wild-type (WT) p27, p27 mutated to S10A, S178A, or T187A. Purified recombinant GST-Mirk or GST-ERK2 was added to each p27 construct in an in vitro kinase reaction. The phosphorylated p27 molecules were separated by SDS-PAGE and autoradiographed. Coomassie Blue (CB) staining confirmed equal amounts of p27 constructs were tested. was further evaluated by expression of small interfering RNA to Mirk using the pSilencer vector. Two sequences within the Mirk coding region were targeted to murine Mirk sequences and termed si1 and si3. For the RNA i control, a mutant sequence si2 was used. NIH3T3 cells were co-transfected in serum-free medium with a 10 to 1 ratio (g/g) of RNA i plasmid to pEGFP plasmid, and with an expression plasmid for p27 to enable the phosphorylation of p27 to be readily detected. Cell lysates contained roughly equal levels of green fluorescent protein, demonstrating similar transfection efficiencies (not shown), so lysates were examined for Mirk abundance by Western blotting. Two days after transfection with plasmids expressing either si1 or si3, a decrease in both endogenous Mirk protein levels and in phosphorylation of p27 at Ser-10 was detected by Western blotting. Little effect on either Mirk or p27 phosphorylation was seen with the si2 RNA that was used as an small interfering RNA control (Fig. 6B). A plot of the Mirk/ tubulin ratio and the ratio of p27 phosphorylated at Ser-10 to total exogenous p27 from Fig. 6B and a duplicate experiment demonstrated that RNA i si1 and si3 both reduced endogenous Mirk levels and reduced exogenous p27-S10 phosphorylation by half after 48 h (Fig. 6D). Thus there was a concordance between reduction of Mirk protein levels and reduction of p27 phosphorylation at Ser-10.
Mirk Phosphorylation of p27 Increases Its Stability in Vivo-Phosphorylation of p27 at Ser-10 has been reported to stabilize p27 (7). We confirmed this observation by measuring the halflife of non-phosphorylatable p27-S10A, phospho-mimetic p27-S10D, and wild-type FLAG-p27 constructs in NIH3T3 cells growth-arrested in G 0 by serum deprivation. Translation arrest by addition of cycloheximide followed by Western blotting confirmed that the p27-S10D phospho-mimetic construct was the most stable (Fig. 6E). Reducing Mirk levels with RNA i also decreased the stability of endogenous p27 (Fig. 6F).
Mirk Knockdown by RNA Interference Increases Cell Cycling-These data implied that depleting endogenous Mirk in G 0 would destabilize p27 and might allow some cells to "leak" into G 1 . To test this hypothesis, NIH3T3 cells were transfected for 24 h with either pSilencer expressing si1 or control. Cells were cultured serum-free for another 2 days to block the untreated cells in G 0 and to determine whether cells depleted of Mirk could enter G 1 in the absence of mitogens. Depletion of Mirk blocks the stabilization of p27 (Fig. 6F) and should decrease p27 levels. RNA i to Mirk decreased Mirk levels to a mean of 37% of control values and decreased p27 levels by half (Fig. 6C, mean of three measurements Ϯ S.E.), confirming the hypothesis. Movement from G 0 to G 1 was assayed by measurement of the abundance of the proliferation marker PCNA by Western blotting. RNA i to Mirk increased the abundance of PCNA 33% Ϯ 7 (mean of three experiments, Fig. 6C). Thus depleting the G 0 -arrested cell of Mirk destabilized p27 and moved the cell into G 1 .
Mirk Co-localizes with p27 in Vivo-Phosphorylation of p27 at Ser-10 has been shown to facilitate p27 translocation from the nucleus to the cytoplasm in response to mitogens, whereas mutation of Ser-10 to alanine prevents export to the cytoplasm (8,20). Mirk phosphorylates p27 at Ser-10 in G 0 , yet p27 remains localized in the nucleus until the middle of G 1 . We tested the hypothesis that phosphorylation at p27-S10 was not sufficient to initiate p27 translocation in G 0 by overexpressing Mirk-EGFP or kinase-inactive YF-Mirk-EGFP together with either p27-DsRed or p27-S10A-DsRed in G 0 -arrested serumstarved NIH3T3 cells. Fluorescent photomicrographs of Mirk alone, p27 alone, and a merged green/red fluorescence were made (Fig. 7). Mirk and YF-Mirk were predominantly nuclear with some diffuse presence in the cytosol, whereas p27 and p27-S10A were localized solely in the nucleus. When the fields were merged, p27, either wild-type or S10A, and nuclear Mirk, wild-type or kinase-inactive, strongly overlapped. Therefore, phosphorylation of p27-S10 by overexpressed Mirk was not sufficient to induce translocation of p27 from the nucleus to the cytoplasm in G 0 . Phosphorylation of p27-S10 by Mirk in G 0 stabilizes p27 and increases its abundance rather than facilitating its translocation out of the nucleus. These data are compatible with the hypothesis that Mirk aids in maintaining cell arrest through phosphorylation of p27 at Ser-10, which occurs in G 0 when Mirk is most abundant. Mitogen stimulation then causes cells to enter G 1 , reduces Mirk levels (13), and induces the KIS kinase that phosphorylates p27 during G 1 (6).
Induction of Mirk during Myoblast Differentiation Enables p27 to Be Stabilized in Post-mitotic Myoblasts-We used RNA i in differentiated C2C12 myoblasts to examine the effect of Mirk expression on p27 stability in a cell system in which Mirk functioned physiologically in differentiation. We reported recently (10) that depletion of endogenous Mirk from differenti- FIG. 4. In vivo phosphorylation of p27 at Ser-10 by co-transfected Mirk. Mirk, kinase-inactive KR-Mirk, or HisA pcDNA vector were co-transfected with either wild-type (WT) FLAG-p27 or mutant FLAG-p27-S10A into NIH3T3 cells. Cells were labeled for 16 h with [ 32 P]orthophosphate, and then lysates were immunoprecipitated with anti-FLAG antibody and collected with protein G-agarose. The immunoprecipitates were analyzed by SDS-PAGE and autoradiography and then Western-blotted (WB) for the FLAG epitope.
FIG. 5. In vivo phosphorylation of co-transfected p27 by Mirk decreases p27 mobility. A, NIH3T3 cells were co-transfected with wild-type (WT) FLAG-p27 or mutant FLAG-p27 constructs together with wild-type HisA-Mirk, kinase-inactive HisA-YF-Mirk, or empty HisA vector. The p27-T187V construct was slightly longer in the C terminus than wild type, as shown by DNA sequencing, and displayed decreased mobility. The lysates were analyzed by Western blotting (WB) to the FLAG epitope. The long arrow indicates p27 with normal mobility, and the shorter arrow indicates p27 with reduced mobility due to phosphorylation. (C), indicates expression in complete medium. B, phosphatase treatment. NIH3T3 cells were co-transfected with wildtype FLAG-p27 or mutant FLAG-p27 constructs together with wildtype Mirk or kinase-inactive YF-Mirk, and the p27 constructs were then immunoprecipitated with anti-FLAG antibody. The immunoprecipitates were incubated with calf intestinal alkaline phosphatase (CIAP) and then analyzed by Western blotting. The amount of antibody used in the immunoprecipitates was identical as shown by the abundance of heavy chain (Hchain) in each lane.
FIG. 6. Phosphorylation of p27 at Ser-10 by Mirk stabilizes p27, allows binding to CDK2, and is blocked by RNA i to Mirk, which enhances mitogen-free cell cycling. A, p27 phosphorylated by Mirk in vivo remains capable of binding to CDK2. NIH3T3 cells were co-transfected with wild-type FLAG-p27 together with wild-type Mirk or vector control, immunoprecipitated (IP) with antibody to CDK2, and the amount of FLAG-p27 associated with CDK2 assayed by Western blotting (WB) the immunoprecipitates. The amount of immunoglobulin heavy chain (H-chain) by cross-reactivity in the immunoprecipitates is shown as an internal control. The lowest panel shows the input amount of FLAG-p27, as determined by Western blotting the lysates for the FLAG epitope. B, depletion of endogenous Mirk by two different RNA i sequences decreases phosphorylation of p27 at Ser-10. NIH3T3 cells were co-transfected for 16 h in serum-free medium with the pSilencer vector encoding RNA i sequences si1 or si3, or mutant control si2, pcDNA-EGFP (10:1 ratio), and FLAG-p27. After an additional 24 h, lysates were examined by Western blotting for Mirk, p27, and tubulin as a blotting control, and the phosphorylated form of p27-S10. C, depletion of endogenous Mirk by RNA i enhances cell cycling. NIH3T3 cells were transfected for 24 h in serum-free medium with the pSilencer vector encoding either si1 or control, and then medium was changed to serum-free and culture continued for 48 h. Lysates were examined by Western blotting for Mirk, p27, PCNA, or ␤-actin. After normalization to ␤-actin, the percent of the proteins compared with control values were calculated (mean Ϯ S.E., n ϭ 3). Transfection efficiency, as monitored by co-transfection with EGFP, was equivalent in si1-treated and control cultures and averaged 30%. D, the amount of Mirk normalized to tubulin and the amount of p27 phosphorylated at Ser-10 to total p27 from the experiment shown in B and a duplicate experiment are shown (ϮS.D. if Ͼ5%). E, exogenous p27 is stabilized by phosphorylation at Ser-10. NIH3T3 cells were co-transfected for 16 h with wild-type (WT) FLAG-p27, FLAG-p27-S10A, or FLAG-p27-S10D and then cultured for 2 days in serum-free medium with cycloheximide added at 50 g/ml for the last 0 -12 h, as noted. The abundance of the FLAG epitope normalized to tubulin is shown for two experiments. The mean Ϯ S.D. for the values shown is 4%. F, endogenous p27 is stabilized by phosphorylation at Ser-10. NIH3T3 cells were co-transfected for 16 h in serum-free medium with the pSilencer vector encoding the RNA i sequence si1 and pcDNA-EGFP (10:1 ratio). After an additional 24 h, lysates were examined by Western blotting for Mirk, endogenous p27, and tubulin as a blotting control. ating C2C12 myoblasts using RNA interference prevented the induction of myogenin and various contractile proteins and blocked the fusion of myoblasts into myocytes. We found that reduction of Mirk levels by RNA interference in differentiating myoblasts caused a 13-fold decrease in p27 protein levels ( Fig.  8) with only a 20% decrease in p27 mRNA levels (not shown). The minimal effect of Mirk on p27 mRNA levels was also seen by microarray analysis (data not shown) and confirmed our earlier results (21), which showed that Mirk did not modulate the activity of a p27 promoter construct. These data indicated that Mirk plays a post-translational role in maintaining p27 stability in post-mitotic myoblasts as well as NIH3T3 cells. DISCUSSION The current study resolves some apparently conflicting experimental results in the literature. It was clearly demonstrated previously (7) that phosphorylation of p27 at Ser-10 confers increased stability on p27, data which we have duplicated in the current study using NIH3T3 cells arrested in G 0 . Other investigators (8,9) have presented equally strong data that phosphorylation of p27 at Ser-10 is required for p27 to bind to the exportin CRM1 and to be translocated to the cytoplasm during G 1 where p27 is proteolyzed by the proteasome, and thus shows less stability. Our results indicate that phosphorylation of p27 at Ser-10 has two outcomes which are cell cycle-dependent. Phosphorylation of p27 at Ser-10 during G 0 by Mirk/dyrk1B stabilizes p27 and maintains p27 within the nucleus where it can bind to CDK2. In contrast, phosphorylation of p27 at Ser-10 by the KIS kinase in G 1 enables p27 to bind to CRM1 and to be transported into the cytoplasm for destruction (6). In support of our conclusions, other investigators (9) have clearly shown that association of p27 with FIG. 8. RNA i to Mirk decreases p27 abundance in postmitotic differentiated myoblasts. C2C12 cells were co-transfected with an expression plasmid for Mirk RNA i and EGFP, or vector DNA and EGFP, selected by cell sorting for EGFP, placed in growth medium for 1 day, and then switched to differentiation medium for 2 days to induce Mirk that is expressed at very low levels in proliferating myoblasts. Vc, vector; si, small interfering RNA to Mirk. Cell lysates were examined by Western blotting for Mirk and p27 Kip1 . NS, nonspecific cross-reacting protein to demonstrate equal loading and transfer. FIG. 7. Co-localization of Mirk and p27. Mirk-EGFP or kinase-inactive YF-Mirk-EGFP were transfected with either p27-DsRed or p27-S10A-DsRed into serum-starved NIH3T3 cells for 16 h and allowed to express in serum-free medium for 24 h. Fluorescent photomicrographs of Mirk alone and p27 alone were made using a simultaneous collection camera, and a merged green/red fluorescence photomicrograph was made. Fluorescence was visualized on a Nikon Eclipse E800 fluorescent microscope (Melville, NY). Images were captured with a Hamamatsu ORCA-ER digital camera (Bridgewater, NJ) and processed with Simple PCI and Adobe Photoshop 5.5 software. CRM1 is minimal in G 0 and increases markedly during the G 1 to S phase progression.
Why are two different kinases needed to perform the same phosphorylation? Mirk and KIS exhibit markedly different transcriptional regulation so they do not occupy the same phase of the cell cycle. Cellular mitogens transcriptionally down-regulate Mirk/dyrk1B (10), whereas mitogens induce the transcription of KIS (6). Mirk levels and activity are highest in G 0 (this study) and decline during G 1 when KIS levels and activity are rising. Therefore, KIS is not present during G 0 to phosphorylate p27. The current study has shown that p27 is not transported to the cytoplasm even after phosphorylation at Ser-10 by Mirk. Thus phosphorylation at Ser-10 is not sufficient to enable p27 to bind to CRM1 during G 0 . Possibly Mirk and p27 occupy a different part of the nucleus than CRM1 during G 0 . Mirk is found in 670-kDa complexes within the nucleus together with the Ran-binding protein RanBPM (12,22). Jab1 functions as an adaptor between CRM1 and p27 (23). The Mirk-containing nuclear complexes have not been completely characterized, but we found in an earlier study that they do not contain Jab1 (22), which is found within the 450-kDa signalosome and smaller complexes (22,23).
The serine/threonine kinase Mirk/dyrk1B is a RhoA-induced gene that functions in muscle differentiation (10) but is widely expressed at low levels, suggesting that it participates in some general cellular control mechanisms. In the current study we have identified one such Mirk function, phosphorylation of p27 to assist the maintenance of the cell in G 0 . Other members of the Dyrk/minibrain family include Dyrk1A, which has been implicated in neuronal development through its phosphorylation of cAMP-response element-binding protein (24,25). Dyrk3 functions in late erythroid progenitor cells to inhibit programmed cell death through activation of the cAMP-response element response pathway (26,27). The related kinase HIPK2 promotes the pro-apoptotic transcriptional function of the tumor suppressor protein p53 (28,29).
Although found at only low levels in most normal tissues, Mirk is expressed at high levels in skeletal and cardiac muscle (13,30). The role of Mirk in muscle differentiation is likely to be multifactorial, involving broad, indirect effects of Mirk acting as a transcriptional activator, as well as more direct effects of Mirk functioning as a substrate-specific kinase. Mirk has been shown to function, in a kinase-dependent manner, as a coactivator of the transcription factor hepatocyte nuclear factor 1␣ (HNF1␣) (11). Mirk binds to HNF1␣ through its cofactor, DCoH, which binds as a dimer to the unstable HNF1␣ dimer, thus enabling effective binding of the tetrameric complex to DNA. We speculate that Mirk may act as a co-activator of one of the transcription factors for the myogenin gene. We have also reported recently (22) that Mirk functions to inhibit cell migration. This study identifies an additional function of Mirk/ dyrk1B, specific phosphorylation of the CDK inhibitor p27 at Ser-10. p27 acts as a brake on the proliferation program. Mirk phosphorylation of p27 may assist differentiating myoblasts to arrest in G 0 by stabilizing p27.