Activation/Proliferation-associated Protein 2 (Caprin-2) Positively Regulates CDK14/Cyclin Y-mediated Lipoprotein Receptor-related Protein 5 and 6 (LRP5/6) Constitutive Phosphorylation*

Low-density lipoprotein receptor-related proteins 5 and 6 (LRP5/6) are co-receptors for Wnt ligands. Upon ligand binding, LRP5/6 undergo glycogen synthase kinase 3 (GSK3)/casein kinase I (CKI)-mediated phosphorylation at multiple PPP(S/T)P motifs in the intracellular domain, which is essential for canonical Wnt signal transduction. On the other hand, in the Wnt-off state, the mitosis-specific CDK14-Cyclin Y kinase complex phosphorylates Ser-1490 of LRP5/6 at G2/M, thereby priming the receptor for Wnt-induced phosphorylation. However, it remains unclear how CDK14/Cyclin Y is recruited to LRP5/6 and whether there are other cofactors involved in this process. Previously, we identified Caprin-2 as a positive regulator of canonical Wnt signaling by promoting GSK3-depedent LRP5/6 phosphorylation upon Wnt stimulation. Here we uncovered that Caprin-2 positively regulates constitutive LRP5/6 Ser-1490 phosphorylation by complexing with CDK14/Cyclin Y. Caprin-2-mediated LRP5/6 phosphorylation is cell cycle-dependent in a pattern similar to that of CDK14/Cyclin Y-dependent LRP5/6 phosphorylation. Moreover, knockdown of Caprin-2 disrupts not only the interaction between CDK14 and Cyclin Y but also the interaction between CDK14/Cyclin Y and LRP6. Overall, our findings revealed an unrecognized role of Caprin-2 in facilitating LRP5/6 constitutive phosphorylation at G2/M through forming a quaternary complex with CDK14, Cyclin Y, and LRP5/6.

The evolutionarily conserved Wnt/␤-catenin signaling plays a pivotal role in myriad physiological and pathological processes, including embryonic development, bone development, tissue homeostasis, neurogenesis, and tumorigenesis (1)(2)(3)(4). At the cellular level, Wnt signals not only promote cell proliferation but also control fate determination or terminal differentiation of postmitotic cells (5). Dysregulation of Wnt signaling may lead to developmental defects and, thereby, underlies a variety of human pathologies such as colorectal cancer, osteoporosis, and neurodegenerative disorders (6).
Plenty of research has suggested a complex interplay between Wnt signaling and the cell cycle. Wnt engagement with the receptors Frizzled and LRP5/6 3 triggers a rapid assembly of a multiprotein complex known as the "LRP signalosome," which is essential for full and sustained pathway activation. The formation of this signalosome is initiated by the recruitment of Axin to LRP5/6 phosphorylated sites (7). The intracellular LRP5/6 phosphorylation domain contains five PPPSPXS motifs. GSK3 and CK1 are responsible for these phosphorylations in the presence of Wnt ligands. Dvl-Frizzled interaction was also required for the recruitment of Axin, GSK3 and CK1, and PIP2-dependent translocation of adenomatous polyposis coli membrane recruitment 1 (Amer1) (also known as WTX) (8 -11). Meanwhile, polymerization of Dvl and Axin facilitates clustering of LRP5/6 in the signalosome. The high local concentration of LRP5/6 further facilitates their phosphorylation mediated by GSK3 and CK1␥ at Ser-1490 and Thr-1479, respectively (12,13). Phosphorylated LRP5/6 inactivate the cytoplasmic ␤-catenin destruction machinery, leading to accumulation and nuclearization of ␤-catenin. This allows for ␤-catenin-dependent expression of various Wnt target genes associated with cell cycle progression, such as c-Myc and Cyclin D1, which directly regulates G 1 /S transition of the cell cycle (14,15). In another scenario, without the presence of Wnt ligands, the activation of LRP5/6 is mediated by the mitosis-specific CDK14-Cyclin Y kinase complex, which phosphorylates LRP6 Ser-1490 during G 2 /M (known as constitutive phosphorylation) to increase the receptiveness of cells for the incoming Wnt signals (16). Therefore, ␤-catenin-dependent Wnt signaling is under cell cycle control and peaks in the G 2 /M phase of the cell cycles. In line with this, the protein levels of ␤-catenin and the Wnt target gene Axin2 also oscillate with the cell cycle, peaking at G 2 /M (17,18). In addition, maximum GSK3 inhibition by constitutive LRP6 phosphorylation also occurs in mitosis, which activates Wnt-dependent stabilization of proteins (Wnt/ STOP) signaling (19). Wnt/STOP signaling acts in a ␤-cateninindependent manner to slow down GSK3-dependent protein degradation and maintains a critical cell size required for cell division (19 -21). Wnt/STOP signaling promotes cell division during early Xenopus embryogenesis and maintains protein homeostasis to promote sperm maturation in the mouse (22,23).
Previously, we have identified cytoplasmic activation/proliferation-associated protein 2 (Caprin-2) as a positive regulator in canonical Wnt signaling. Caprin-2 directly binds to LRP5/6 and facilitates GSK3-mediated LRP5/6 phosphorylation (24). The C-terminal CRD of Caprin-2 forms a flexible homotrimer, and this homotrimerization of CRD is required for Caprin-2 to regulate LRP5/6 aggregation and canonical Wnt signaling (25). In this work, we found that Caprin-2 also participates in the regulation of constitutive LRP5/6 phosphorylation. Our findings suggested that Caprin-2 acts as a scaffold protein to potentiate the CDK14-Cyclin Y-LRP6 complex in the G 2 /M phase of the cell cycle, sensitizing cells for incoming Wnt signals when cells enter mitosis.

Results
Caprin-2 Promotes Constitutive LRP6 Ser-1490 Phosphorylation-We previously showed that Caprin-2 was a positive regulator in canonical Wnt signaling through a mechanism of facilitating GSK3␤-mediated LRP5/6 phosphorylation upon Wnt stimulation (24). Interestingly, several studies by Niehrs and co-workers revealed that phosphorylation of LRP5/6 is not exclusively Wnt-induced (12,16). Their results showed that the CDK14-Cyclin Y complex phosphorylates Ser-1490 in the PPP(S/T)P motif of LRP5/6 in G 2 /M phase of the cell cycle, thereby priming LRP5/6 for an instant response to the incom-ing Wnt signals (16). This constitutive Wnt-independent LRP6 phosphorylation also triggers Wnt/STOP signaling, which plays essential roles in controlling daughter cell size and spermatogenes (19). Caprin-2, unlike Dvl, whose association with LRP5/6 requires Wnt stimulation, interacts with LRP5/6 in a Wnt ligand-independent manner (24). To investigate whether Caprin-2 is involved in regulating LRP5/6 constitutive phosphorylation, we first transfected HEK293 cells with HA-tagged Caprin-2 and examined its effects on the levels of phospho-LRP6 (Ser-1490) in the absence of Wnt stimulation. Consistent with our previous report, Caprin-2 overexpression and Wnt-3a treatment showed a synergistic effect on the increase of LRP6 phosphorylation (Fig. 1A); however, in the absence of Wnt stimulation, Caprin-2 also increased the phosphorylation levels of LRP6 (Fig. 1A). To exclude any potential effects because of the possible existence of Wnt proteins in the cell culture medium or secreted by cells per se, we supplemented the medium with DKK1 (Dickkopf 1), a secreted Wnt signaling antagonist that could prevent Frizzled-Wnt-LRP6 complex formation (26 -28). As shown in Fig. 1A, the increase in phosphorylation in LRP6 Ser-1490 induced by Caprin-2 was not affected by the addition of the Dkk1-CM, whereas Wnt-3a-induced LRP6 Ser-1490 phosphorylation was completely blocked by Dkk1-CM treatment. Next, we performed a loss-of-function study by knocking down endogenous Caprin-2 using shRNA lentivirus infection. Two shRNAs were used to avoid siRNA off-target effects. According to our observation, endogenous Caprin-2 is quite stable, and we thus extended the virus incubation time up to 72 h for a maximal knockdown efficiency. As shown in Fig. 1B, knockdown of Caprin-2 significantly decreased both Wnt-3a-stimulated LRP6 phosphorylation and Dkk1-CM-treated Wnt-3a-independent LRP6 phosphoryla- tion. These observations suggested that Caprin-2 plays a role not only in facilitating the Wnt-induced LRP5/6 Ser-1490 phosphorylation but also in regulating constitutive LRP6 Ser-1490 phosphorylation.
Caprin-2-mediated LRP6 Phosphorylation Peaks at G 2 /M-Because constitutive CDK14/Cyclin Y-mediated LRP6 phosphorylation is under the control of the cell cycle, peaking at G 2 /M (16), we next set out to examine whether Caprin-2-induced LRP6 constitutive phosphorylation is also related to the cell cycle. We first incubated HeLa cells with nocodazole to arrest the cell cycle at G 2 /M phase. Nocodazole-treated cells enter mitosis but cannot form metaphase spindles because microtubules cannot polymerize. Flow cytometry analysis confirmed the increase in mitotic cells in nocodazole-treated cells. Consistent with the previous report, endogenous Ser-1490 phosphorylation levels exhibited an apparent increase in G 2 /M-arrested cells; however, knockdown of Caprin-2 blocked this increase in LRP6 Ser-1490 phosphorylation at G 2 /M ( Fig.  2A). To further dissect the role of Caprin-2 in cell cycle-dependent LRP6 phosphorylation, we performed a double thymidine block and release experiment in HeLa cells. Parallel FACS analysis confirmed cell cycle phase enrichment. As shown in Fig. 2B, LRP6 Ser-1490 phosphorylation peaked during G 2 /M, and so did the expression of the endogenous Cyclin Y. Of note, the expression of Caprin-2 also oscillated during the cell cycle and exhibited a peak in G 2 /M, implying a possible functional correlation between Caprin-2 and Cyclin Y. By contrast, knockdown of Caprin-2 not only decreased the overall phosphorylation levels of LRP6 but also compromised the oscillation pattern of LRP6 phosphorylation through the cell cycle (Fig. 2B). We also monitored the expression levels of Caprin-2 in drug-synchronized HeLa cells and observed a clear enrichment of both Caprin-2 and CDK14/Cyclin Y in nocodazole-treated cells (Fig.  2C). Moreover, consistent with the results from the double thymidine block and release assay, Caprin-2 knockdown led to decreased LRP6 phosphorylation through cell cycle phases in drug-synchronized HeLa cells (Fig. 2C). Taken together, these results suggested that Caprin-2-mediated LRP6 Ser-1490 phosphorylation is subject to cell cycle control and that Caprin-2 may synergize with CDK14/Cyclin Y to boost LRP6 Ser-1490 phosphorylation in G 2 /M.
Caprin-2 Interacts with CDK14/Cyclin Y-In a previous effort to identify the potential binding partners for Caprin-2, we carried out affinity purification coupled to mass spectrometry and found many CDKs in the list of candidate Caprin-2-interacting proteins (data not shown). We then asked whether Caprin-2 interacts with CDK14/Cyclin Y. Cyclin Y is localized to the plasma membrane via N-terminal myristoylation, and CDK14 is recruited to the plasma membrane by interacting with Cyclin Y. The subcellular distribution of Caprin-2 has so far not been clearly addressed. Thus, we first performed a subcellular fraction assay to examine the distribution of Caprin-2. We isolated the membrane, cytoplasmic, and nuclear proteins of HeLa cells and found that most of Caprin-2 proteins were in the membrane fraction, with very few distributed in the cytoplasmic and nuclear fractions (Fig. 3A). The co-localization of Caprin-2, CDK14, and Cyclin Y to the plasma membrane suggested that they may collaborate with each other in regulating constitutive LRP6 Ser-1490 phosphorylation. Next we set out to  3B). We then co-transfected GFP-tagged CDK14 or Cyclin Y with HA-tagged full-length Caprin-2 or the three Caprin-2 fragments into HEK293T cells. At 24 h post-transfection, whole-cell lysates were used in a co-immunoprecipitation experiment using anti-GFP antibody. As shown in Fig. 3, C and D, CDK14 could be co-immunoprecipitated by the full-length Caprin-2, Ca2-M, and Ca2-C but not by Ca2-N, whereas Cyclin Y was co-immunoprecipitated with the full-length Caprin-2 and Ca2-N but not with Ca2-M or Ca2-C. These results indicate that Caprin-2 interacts with both CDK14 and Cyclin Y but through different regions. To determine whether Caprin-2 interacts with CDK14 or Cyclin Y directly, we performed an in vitro GST pulldown experiment using bacterially expressed recombinant proteins. We failed to obtain the full-length Caprin-2 and Ca2-M samples because of their extremely low solubility when expressed in Escherichia coli. We purified His 6tagged Ca2-N and Ca2-C and performed a GST pulldown assay using purified GST-tagged CDK14 or Cyclin Y. As shown in Fig.  3, E and F, GST-CDK14, but not the control GST, pulled down His 6 -Ca2-C, whereas GST-Cyclin Y pulled down His 6 -Ca2-N. These results are consistent with the observation from the co-immunoprecipitation assay, suggesting that Caprin-2 binds directly to both CDK14 and Cyclin Y through different regions.
Caprin-2 Scaffolds the CDK14-Cyclin Y-LRP6 Complex-CDK14 and Cyclin Y are classic CDK/Cyclin partners. Cyclin Y is the regulatory subunit of CDK14, which is important for the kinase activity of CDK14. Our results above showed that Caprin-2 can interact directly with both CDK14 and Cyclin Y. We then asked whether Caprin-2 affects the CDK14-Cyclin Y interaction. We co-transfected CDK14 and Cyclin Y with or without Caprin-2 into HEK293T cells. As expected, Caprin-2 strengthened the interaction between CDK14 and Cyclin Y (Fig. 4A), which is also evidenced by the knockdown experiment using Caprin-2 shRNA (Fig. 4B). To examine whether Caprin-2-CDK14-Cyclin Y complex formation is subject to regulation by the cell cycle, we performed endogenous co-IP in different cell cycle phases. As shown in Fig. 4C, Caprin-2 and CDK14 co-immunoprecipitated by Cyclin Y were increased at G 2 /M. This result further confirmed that Caprin-2, CDK14, and Cyclin Y form a complex that is potentiated at G 2 /M. On the other hand, CDK14 and Cyclin Y both associate with LRP5/6, which is required for CDK14/Cyclin Y-mediated LRP5/6 constitutive phosphorylation (16). We then tested whether Caprin-2 plays a role in the interaction between CDK14/Cyclin Y and LRP6. We first examined the interactions between endogenous CDK14/Cyclin Y and LRP6. We immunoprecipitated endogenous LRP6 from 293 cell lysates using anti-LRP6 antibody and detected the presence of endogenous CDK14/Cyclin Y in the immunoprecipitates. Of note, these endogenous interactions of either LRP6-CDK14 or LRP6-Cy- Co-IP was performed as described in A. C, endogenous interaction of Caprin-2/CDK14/Cyclin Y increased at G 2 /M. HeLa cells were synchronized by double thymidine block, and the cells lysates from G 1 /S-or G 2 /M-synchronized cells were immunoprecipitated by Cyclin Y-specific antibody, followed by Western blotting analysis with the indicated antibodies. D, endogenous Caprin-2/CDK14/Cyclin Y/LRP6 complexes were analyzed in HEK293 cells. Immunoprecipitation was performed with an anti-LRP6 polyclonal antibody. IgG was used as a control. Before co-immunoprecipitation, cells were treated with shCaprin-2 lentivirus for 56 h and nocodazole for another 16 h. E, model for Caprin-2 modulating constitutive LRP6 phosphorylation. DECEMBER 16, 2016 • VOLUME 291 • NUMBER 51 clin Y could only be observed in cells arrested at G 2 /M by nocodazole but not in unsynchronized cells. However, knockdown of endogenous Caprin-2 abolished both LRP6-CDK14 and LRP6-Cyclin Y interactions (Fig. 4D). These results, combined with those above, demonstrated that Caprin-2 serves as a scaffold to mediate the formation of the CDK14-Cyclin Y/LRP6 complex, in which LRP6 is phosphorylated by CDK14/Cyclin Y, thereby sensitizing cells for the incoming Wnt signals when cells enter mitosis (Fig. 4E).

Discussion
We previously identified Caprin-2 as a new LRP5/6-binding protein that promotes Wnt-induced LRP5/6 phosphorylation. Here we revealed a new role of Caprin-2 in the regulation of the Wnt-independent LRP5/6 phosphorylation. Our findings showed that Caprin-2 acts as a scaffold protein to promote CDK14-Cyclin Y-LRP5/6 complex formation at G 2 /M, thereby positively regulating Wnt-independent phosphorylation of LRP5/6. Because Caprin-2 interacts with LRP6 independent of Wnt stimulation (24), Caprin-2 could be a constitutive binding partner of LRP5/6, which, in the Wnt-off state, promotes CDK14-Cyclin Y-mediated LRP6 phosphorylation, whereas, in the Wnt-on state, it acts synergistically with Wnt signals to promote GSK3␤-mediated LRP5/6 phosphorylation. Therefore, Caprin-2 regulates both constitutive and Wnt-induced LRP5/6 phosphorylation, contributing to a full and quick activation of LRP5/6. It is possible that other signals may be required in the Wnt-off state to regulate the assembly of the Caprin-2-CDK14-Cyclin Y complex at G 2 /M. This complex may also include additional components, allowing for fine-tuning of the constitutive LRP5/6 phosphorylation.
Sequential phosphorylation of the five PPPSPXS motifs of LRP5/6 is the key event for the initiation of Wnt signaling. These phosphorylated sites provide docking sites for Axin. Axin recruitment to the phosphorylated LRP5/6 inactivated the ␤-catenin destruction complex, ultimately liberating ␤-catenin from the destiny of degradation. The phosphorylation of the five PPPSPXS motifs is hierarchically controlled. It is well known that Ser-1490 is phosphorylated by GSK3 upon Wnt stimulation, which primes the subsequent phosphorylation by CK1. But Ser-1490 is also constitutively phosphorylated by CDK14/Cyclin Y. The constitutive Ser-1490 phosphorylation of LRP5/6 may provide a weak binding site for the initial recruitment of Axin, thereby conferring LRP5/6 an easily activated status. We speculate that constitutive Ser-1490 phosphorylation may primesLRP5/6 to be phosphorylated by GSK3 and CK1 at more sites of the five PPPSP motifs upon Wnt engagement.
Caprin-2 contains a C1q-related domain at its C terminus (Cap2_CRD), and we previously determined its three-dimensional structure, which exhibits a trimeric assembly with a typical jelly roll topology shared by both C1q and TNF family proteins (25). Our previous study showed that trimerization of Cap2_CRD is not essential for Caprin-2 to interact with LRP5/6; however, it is required for Caprin-2 to regulate LRP5/6 phosphorylation upon Wnt stimulation (24). One possibility is that Caprin-2 constitutively interacts with LRP5/6 and recruits CDK14/Cyclin Y to LRP5/6 in G 2 /M for maintaining constitu-tive phosphorylation of LRP5/6, whereas Wnt stimulation may induce CRD domain-mediated aggregation of Caprin-2, which could facilitate LRP5/6 aggregation as well as the subsequent formation of the LRP5/6 signalosome.
In addition to priming Wnt-induced LRP5/6 phosphorylation, CDK14/Cyclin Y-mediated LRP5/6 phosphorylation also triggers Wnt/STOP signaling, which leads to GSK3 inactivation and thus protects proteins from GSK3-dependent degradation (19). Wnt/STOP signaling plays key roles in maintaining the size of daughter cells and the faithful execution of mitosis and has been found recently to be the dominant mode of Wnt signaling in several cancer cell lines (19,29). Considering the essential role of Caprin-2 in CDK14/Cyclin Y-mediated LRP5/6 phosphorylation, it is very likely that Caprin-2 may also function in Wnt/STOP signaling. The role of Caprin-2 in Wnt/ STOP signaling and whether this role may correlate with the generation and progression of some Wnt/STOP signaling-related cancers remains to be determined.
Based on current literature, both CDK14 and Cyclin Y were highly expressed in postmitotic tissues like the testis and brain (30,31). Coincidentally, Caprin-2 is also highly expressed in the testis and brain (32,33). These results imply a functional interaction between CDK14/Cyclin Y and Caprin-2 in these tissues. The majority of the cells in the brain and testis are not active in mitosis. The expression pattern of CDK14/Cyclin Y predominant in the brain and testis is consistent with the previous finding that the function of CDK14/Cyclin Y is different from the classic CDK-Cyclin complex (22, 34 -36). Whether Caprin-2 exerts any physiological function in the testis or brain together with CDK14/Cyclin Y warrants further studies in a model organism, such as Caprin-2 knockout mice.
Cell Culture, Transfection, and Reporter Gene Assay-HEK293T, HEK293, and HeLa cells were maintained in Dulbecco's modified Eagle's medium (Gibco) plus 10% fetal bovine serum (Gibco). Cells were seeded in plates 24 h before transfection, and plasmids were transfected using Lipo2000 reagent (Invitrogen) according to the instructions of the manufacturer. For the reporter gene assay, HEK293T cells in a 24-well plate were transfected with 250 ng of plasmids in total for each well, including 5 ng of TOPFlash reporter plasmid and 10 ng of GFP Caprin-2 Regulates LRP5/6 Constitutive Phosphorylation plasmid as the transfection control. After 18 h of transfection, cells were treated with Wnt3a-conditioned medium (Wnt3a-CM), Dkk1-conditioned medium (Dkk1-CM), or control medium (Ctrl-CM) for an additional 6 h. Cells were then lysed, and luciferase assays were performed. The luciferase activities presented were normalized against the levels of GFP expression as described previously (37).
Conditioned Medium-Wnt3a-CM and L cell control CM were prepared as described previously (7). For Dkk1 CM, HEK293T cells in 10-cm dishes were transfected with 10 g of pcDNA3-Myc-Dkk1 using Lipo2000 according to the specifications of the manufacturer. The media were exchanged with fresh Dulbecco's minimum essential medium containing 10% fetal bovine serum 24 h after transfection. After further 48 h of incubation, the media were collected, centrifuged to remove cell debris, and stored at Ϫ80°C.
Co-immunoprecipitation (Co-IP) and Western Blotting Assays-At 24 h post-transfection, cells were harvested and lysed in lysis buffer (50 mM Tris-HCl (pH 8.0), 1% Triton X-100, 150 mM NaCl, and 5 mM EDTA with protease and phosphatase inhibitors). The lysates were centrifuged for 15 min at 13,000 rpm at 4°C. The supernatants were incubated with antibodies and protein A/G Plus-agarose (Santa Cruz Biotechnology, Inc.) for 3 h at 4°C. Then, samples were washed three times with lysis buffer and denatured in SDS sample buffer for 10 min at 95°C. Proteins were separated by SDS-PAGE and blotted onto PVDF membranes (Millipore). Membranes were blocked with 5% nonfat dry milk for 1 h and then incubated with primary antibodies for 1 h at room temperature or overnight at 4°C. After being washed, membranes were incubated for 1 h at room temperature by using the appropriate horseradish peroxidase-conjugated secondary antibodies (Thermo Scientific) for 1 h at room temperature. The results were visualized by FujiFilm Las 4000 (FujiFilm).
GST Pulldown Assay-Recombinant proteins (GST-or Histagged) were expressed in E. coli host strain BL21(DE3). The cells were collected, resuspended in lysis buffer, and lysed using a high-pressure homogenizer. GST or GST-tagged fusion protein was incubated with glutathione-agarose beads (Sigma) for 3 h, and then His tagged proteins were added and mixed for another 3 h. The beads were washed three times with the lysis buffer. The samples were treated with SDS loading buffer and analyzed by Western blotting using anti-GST and anti-His monoclonal antibodies.
Cell Fractionations-Cells grown in 6-well plates were harvested with a cell scraper into 1.5 ml of PBS, followed by centrifugation at 700 ϫ g for 10 min. The pelleted cells were resuspended in buffer A (10 mM HEPES (pH 7.9), 1.5 mM MgCl 2 , 10 mM KCl, 0.5 mM DTT, 10 mM NaF, 2 mM Na 3 VO 4 , 1 mM pyrophosphoric acid, and Complete TM protease inhibitor) and incubated on ice for 10 min. Cells were then passed through a 0.4-mm needle point to break down the cell membrane and centrifuged at 700 ϫ g at 4°C for 10 min. The supernatant was collected and centrifuged at 100,000 ϫ g at 4°C for 1 h. The supernatant from the ultracentrifugation was collected as the cytosolic fraction. For nuclear protein extraction, the pellet from the 700 ϫ g centrifugation was washed with buffer A, resuspended in buffer C (20 mM HEPES (pH 7.9), 1.5 mM MgCl 2 , 420 mM NaCl, 0.2 mM EDTA, 10 mM NaF, 2 mMNa 3 VO 4 , 1 mM pyrophosphoric acid, and Complete TM protease inhibitor) and incubated on ice for 30 min. Nuclear extracts were recovered from the supernatants after centrifugation at 100,000 ϫ g at 4°C for 1 h.
Cell Cycle Assay-To arrest cells at the G 2 /M stage, cells were treated with nocodazole (100 ng/ml) for 16 h. G 0 phase cells were obtained by serum starvation for 50 h in 0.2% serum. For the double thymidine block, cells were incubated in 2 mM thymidine twice for 17 h of exposure time separated by a 9-h recovery interval. G 1 /S-arrested cells were then allowed to progress through the cell cycle by removing thymidine. Cell cycle states were confirmed by flow cytometry after trypsinization of cells, fixation in 70% ethanol, and staining with 50 mg/ml propidium iodide for 30 min at 37°C. Cells (at least 10,000) were analyzed on a FACSCalibur (BD Biosciences).