The Yes-associated protein, YAP, is a downstream effector of the Hippo pathway of cell-cycle control that plays important roles in tumorigenesis. Hippo-mediated phosphorylation YAP, mainly at S127, inactivates YAP function. In this study, we define a mechanism for positive regulation of YAP activity that is critical for its oncogenic function. Specifically, we found that YAP is phosphorylated in vitro and in vivo by the cell-cycle kinase CDK1 at T119, S289, and S367 during the G2–M phase of the cell cycle. We also found that ectopic expression of a phosphomimetic YAP mutant (YAP3D, harboring T119D/S289D/S367D) was sufficient to induce mitotic defects in immortalized epithelial cells, including centrosome amplification, multipolar spindles, and chromosome missegregation. Finally, we documented that mitotic phosphorylation of YAP was sufficient to promote cell migration and invasion in a manner essential for neoplastic cell transformation. In support of our findings, CDK1 inhibitors largely suppressed cell motility mediated by activated YAP-S127A but not the phosphomimetic mutant YAP3D. Collectively, our results reveal a previously unrecognized mechanism for controlling the activity of YAP that is crucial for its oncogenic function mediated by mitotic dysregulation. Cancer Res; 73(22); 6722–33. ©2013 AACR.
Studies from Drosophila have defined the Hippo signaling pathway (1). Genetically engineered mouse models demonstrated that the Hippo pathway is highly conserved in mammals and controls organ size, tumorigenesis, cell contact inhibition, and stem cell self-renewal by regulating cell proliferation and apoptosis (2–4). The core of the Hippo pathway is a kinase cascade including the tumor suppressors Mst1/2 (Hippo in Drosophila), Lats1/2 (Warts in Drosophila), and the oncoproteins Yes-associated protein (YAP)/transcriptional coactivator with PDZ-binding motif (TAZ; Yorkie in Drosophila). Mst1/2, in complexes with WW45 (Salvador in Drosophila), phosphorylates and activates Lats1/2. Activated Lats1/2 (together with Mob1) phosphorylates and inactivates the transcriptional coactivators YAP/TAZ. The transcription factors TEAD1-4 (Scalloped in Drosophila) are the main mediators for YAP's oncogenic function (5, 6). Recent work has identified lysophosphatidic acid, sphingosine-1-phosphate as ligands, and G-protein–coupled receptors as receptors for the Hippo pathway (7–9). YAP is phosphorylated at S127 and S381 by Lats1/2 kinases and inactivated by cytoplasmic sequestration and ubiquitination-dependent degradation (10–13).
YAP promotes tumorigenesis in several types of cancers, including hepatocellular carcinoma (13) and skin cancer (14). It is not surprising that YAP is amplified or overexpressed/hyperactivated in many types of human cancers (13, 15). Accordingly, genetic ablation of upstream tumor suppressors (e.g., Mst1/2, WW45, and Mob1a/1b) in mice also leads to the formation of many types of tumors including hepatocellular carcinoma (16–20).
Although extensive studies have demonstrated the important roles for the Hippo pathway in tumorigenesis, the underlying mechanisms are still unclear. Interestingly, recent studies demonstrated that several key members of the Hippo pathway, such as Mst1/2, Lats1/2, WW45, and Mob1, are involved in regulating mitosis (21). Aberration of mitosis often causes genome instability/aneuploidy and subsequent tumor formation. Thus, the Hippo pathway may contribute to cancer development by regulating mitosis-related events.
We recently reported that KIBRA (an upstream regulator of the Hippo pathway) is required for chromosome alignment and proper microtubule organization during mitosis (22). Furthermore, KIBRA is phosphorylated by mitotic kinases Aurora and CDK1 during mitosis (23, 24). These studies prompted us to examine whether YAP, the most critical growth mediator of the Hippo pathway, is regulated during mitosis. We show that YAP is also hyperphosphorylated during the G2–M phase. We further characterized the phosphorylation sites, the corresponding kinase, and the functional significance of the phosphorylation. Our data reveal a new layer of regulation for YAP activity, implicating that YAP exerts its oncogenic function through dysregulation of mitosis.
Materials and Methods
YAP-S381A, YAP with mutations in the WW domains, and YAP-5SA plasmids were purchased from Addgene. pcDNA-YAP (no tag) expression constructs have been described (13). To make HA-tagged human YAP, full-length YAP cDNA (IMAGE clone 5747370, contains two WW domains and 504 amino acids) was subcloned into the pcDNA3.1+HA vector (25). To make the retroviral-mediated YAP expression construct, the above cDNA was cloned into MaRX™IV vector (25). Point mutations were generated by the QuickChange Site-Directed PCR mutagenesis kit (Stratagene) and verified by sequencing.
Cell culture and transfection
HEK293T, HeLa, MCF-7, and MCF10A cell lines were purchased from American Type Culture Collection (ATCC). The cell lines were authenticated at ATCC and were used at low (<20) passages. MCF-10A cells were cultured as described (10). H2229, H1277, T47D, MDA-MB-231, and HCT-116 cell lines were maintained in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% FBS. Attractene and HiPerFect (Qiagen) were used for transient overexpression and siRNA transfections, respectively, following the manufacturer's instructions. Nocodazole (100 ng/mL for 16–20 hours) and Taxol (0.1 μmol/L for 16 hours) were used to arrest cells in the G2–M phase unless otherwise indicated. Etoposide and doxorubicin were also from Sigma. YAP siRNA was synthesized by GenePharma based on the following target sequence (YAP-1: 5′-CAGGTGATACTATCAACCAAA-3′; YAP-2: 5′-GACCAATAGCTCAGATCCTTT (selected by Invitrogen online software). All other chemicals were either from Sigma or Thermo Fisher.
VX680 (Aurora-A, -B, -C inhibitor), ZM447439 (Aurora-B, -C inhibitor), and BI2536 (Plk1 inhibitor) were obtained from Selleck Chemicals. U0126 [MAP–ERK kinase (MEK)-extracellular signal—regulated kinase (ERK) inhibitor], SB203580 (p38 inhibitor), LY294002 (PI-3K inhibitor), rapamycin (mTOR inhibitor), and SP600125 (JNK inhibitor) were from LC Laboratories. RO-3306 (CDK1 inhibitor) and roscovitine (CDKs inhibitor) were from ENZO Life Sciences. MK5108 (Aurora-A inhibitor) was from Merck. SB216763 (GSK3β inhibitor) and anisomycin (JNK activator) were from Sigma.
Recombinant protein purification
The glutathione S-transferase (GST)–tagged proteins were bacterially expressed and purified on GSTrap FF affinity columns (GE Healthcare) following the manufacturer's instructions. To make His-tagged human YAP, full-length YAP cDNA was subcloned into the pET-21c vector (Novagen/EMD Chemicals). The proteins were expressed and purified on HisPur™ Cobalt spin columns (Pierce) following the manufacturer's instructions.
Identification of phosphorylation sites by mass spectrometry
Endogenous YAP from HeLa cells treated with Taxol was immunoprecipitated and stained with Coomassie blue. The upshifted YAP bands were sliced and in-gel digested as described (26). Nano LC-MS/MS (nanoscale liquid chromatography followed by tandem mass spectrometry) was performed with an in-house built nanoLC system (27) coupled with an LTQ (linear ion trap)-Velos mass spectrometer (Thermo Scientific). Survey full scan MS spectra (from m/z 375 to 1,700) were acquired in the LTQ-Velos with a resolution of 6,000. The 20 most intense ions (depending on signal intensity) were sequentially isolated for fragmentation in the linear ion trap by collision-induced dissociation. The capillary was maintained at 200°C, and the spray voltage was kept at 2.3 kV. DeconMSn was used to determine and refine the monoisotopic mass and charge state of parent ions from the LTQ-Velos raw data, and to create a peak list of these ions in .mgf format. The .mgf files were searched against the human international protein index (IPI) protein sequence database (Version 3.52), which contained the normal IPI human proteins, commonly observed contaminants, and the reverse sequences of all proteins with the OMSSA search engine (version 2.1.9; National Center for Biotechnology Information; ref. 28). In the database search, carbamidomethylation of cysteine was set as the fixed modification. Oxidation of methionine, and phosphorylation of serine, threonine, and tyrosine were set as the variable modifications. The precursor tolerance was set as 1.5 Da and MS2 tolerance was 0.5 Da. The E value cutoff was set at 0.1. The false discovery rate was made 1% by filtering on the E-value of all forward and reversed peptide identifications. The spectra were manually checked.
In vitro kinase assay
1–2 μg of His-YAP was incubated with 10 U recombinant CDK1/cyclin B complex (New England Biolabs) or 100 ng CDK1/cyclin B (SignalChem) or HeLa cell total lysates (treated with DMSO or Taxol) in kinase buffer (23) in the presence of 5 μCi γ-32P-ATP (3,000 Ci/mmol; PerkinElmer). MEK1, ERK1, and p38α active kinases were purchased from SignalChem. Myelin basic protein (Sigma) was used for positive control. The samples were resolved by SDS-PAGE, transferred onto polyvinylidene difluoride (Millipore), and visualized by autoradiography followed by Western blotting or detected by phospho-specific antibodies.
The YAP antibodies from Abnova (H00010413-M01) and Abcam (52771) were used for immunoprecipitation of endogenous YAP and for Western blotting, respectively, throughout the study. Rabbit polyclonal phospho-specific antibodies against YAP S367, S289, and T119 were generated and purified by AbMart. HA antibodies were from Sigma. Anti-β-actin, anti-ERK1/2, and anti-cyclin B antibodies were from Santa Cruz Biotechnology. Anti-Aurora-A, anti-GST, anti-His, anti-Mst1, anti-Mst2, anti-Lats1, and anti-Lats2 antibodies were from Bethyl Laboratories. Anti-phospho-Aurora-A,B,C, anti-phospho-S10 H3, anti-phospho-T202/Y204 ERK1/2, anti-phospho-S127 YAP, anti-phospho-T180/Y182 p38, anti-phospho-c-Jun, anti-phospho-Mst1/2, anti-phospho-Lats1/2, anti-phospho-S345 Chk1, anti-p38, anti-WW45, anti-TAZ, anti-NF2, anti-Mob1, and anti-Cdc2 antibodies were from Cell Signaling Technology. Anti-Plk1 and anti-phospho-T210 Plk1 antibodies were obtained from Biolegend. Anti-α-tubulin (Abcam), anti-β-tubulin (Sigma), and anti-γ-tubulin (Biolegend) antibodies were used for immunofluorescence staining.
Immunoprecipitation, Western blot analysis, and lambda phosphatase treatment
Immunoprecipitation, Western blotting, and lambda phosphatase treatment assays were done as previously described (23).
Immunofluorescence staining and confocal microscopy
Cell fixation, permeabilization, fluorescence staining, and microscopy were done as previously described (22). For peptide blocking, a protocol from Abcam website was used. Briefly, the phospho-YAP antibodies were first neutralized by an excess of immunizing (phosphorylated) peptides (1 μg/mL for 1 hour at room temperature). The antibody (containing the phosphopeptide) was then used for staining in parallel with staining using antibodies with no peptide or nonphosphopeptide.
Colony formation, cell migration, and invasion assays
Colony formation assays in soft agar were performed as described (13). In vitro analysis of invasion and migration was assessed using the BioCoat invasion system (BD Biosciences) and Transwell system (Corning), respectively, according to the manufacturer's instructions. The invasive and migratory cells were stained with ProLong® Gold Antifade Reagent with 4′,6-diamidino-2-phenylindole (DAPI). The relative invading and migrating rates were calculated by the number of cells invading and migrating through the membrane divided by the number of cells that invaded and migrated in the control group.
Statistical significance was performed using a two-tailed, unpaired Student t test.
YAP is phosphorylated during antimitotic drug-induced G2–M arrest
To further explore whether other members of the Hippo pathway are regulated during mitosis, we treated HeLa cells with Taxol or nocodazole (both agents arrest cells in G2–M) and systematically examined the responses of the Hippo pathway during G2–M arrest. As shown in Fig. 1A, the most prominent change we observed was the dramatic mobility upshift of YAP and, to a lesser extent, of TAZ. The phosphorylation level of YAP at S127 (the major phosphorylation site regulated by the Hippo pathway) was not altered, suggesting that the mobility shift of YAP was not likely due to the phosphorylation at S127. Taxol or nocodazole treatment did not cause any evident change in the expression levels for NF2, Mst1/2, WW45, Mob1, or in the activity of Mst1/2 and Lats1/2 revealed by phospho antibodies (Fig. 1A). Consistent with previous reports, we detected a mobility upshift of Lats1 (due to mitotic phosphorylation; refs. 29, 30) and a significant increase of Lats2 expression (31) during Taxol or nocodazole treatment (Fig. 1A). The mobility upshift of YAP was also evident in breast (MCF-7, T47D, and MDA-MB-231), colon (HCT-116), and lung (H1299 and H2227) cancer cell lines during Taxol-arrested mitosis (Fig. 1B). The shift was detectable as early as 4 hours after 100-nmol/L Taxol treatment, and only 10 nmol/L of Taxol (for a 16-hour treatment) was sufficient to induce the shift in HeLa cells (Fig. 1C). Taken together, YAP mobility is significantly retarded during antimitotic drug-induced G2–M arrest.
YAP is a phosphoprotein whose mobility is retarded on SDS-polyacrylamide gels when phosphorylated (10, 13). Lambda phosphatase treatment completely converted all slow-migrating bands to fast-migrating bands, indicating that the mobility shift of YAP during G2–M is caused by phosphorylation (Fig. 1D). To further test whether YAP phosphorylation is specific to mitosis, we collected mitotic cells by mechanical shake-off from Taxol-treated cells. As shown in Fig. 2A, mitotic cells expressed exclusively phosphorylated YAP, while YAP is not phosphorylated/shifted in attached nonmitotic cells, suggesting that YAP phosphorylation is specifically associated with G2–M cell-cycle arrest.
Previous reports showed that YAP was phosphorylated by c-Abl, p38, and JNK kinases in response to the DNA damage (32–34). We further explored whether DNA-damaging agents can cause such a mobility shift of YAP. As shown in Fig. 2B, in contrast to YAP shift during mitosis, treatment with doxorubicin, etoposide, UV, or ionizing radiation (IR) failed to cause any obvious change on the YAP mobility. Increased activity of Chk1 revealed by phospho-S345 antibody indicated that these treatments were effective. These data further suggest that the mobility shift/phosphorylation of YAP is a specific response during cell cycle G2–M.
Identification of the corresponding kinase(s) for YAP phosphorylation
Upon treatment of antimitotic agents, several mitotic kinases, including Aurora and Plk1, are activated during mitosis (35). Inhibitors for these kinases are widely used. Inhibition of Aurora-A (with MK5108), or Aurora-B, C (with ZM447439), or Aurora-A, B, C (with VX-680) kinases did not alter the YAP phosphorylation. The treatments with these inhibitors were effective, as revealed by the phospho-Aurora antibody (Supplementary Fig. S1A). Addition of BI2536 (an inhibitor of Plk1 kinase) did not reverse the YAP mobility shift/phosphorylation either (Supplementary Fig. S1B). It has been reported that MEK-ERK signaling is also activated during Taxol treatment (36). However, in our hands, MEK-ERK activity was strongly inhibited upon treatment with Taxol (Supplementary Fig. S1C; refs. 24, 37). Thus MEK-ERK kinases are not likely the kinases responsible for YAP phosphorylation under antimitotic drug-induced mitosis, and treatment with U0126 (an inhibitor for MEK-ERK kinases) did not affect the YAP phosphorylation (Supplementary Fig. S1C). In addition, inhibition of PI-3K (with LY294002), mTOR (with rapamycin), MAPK-p38 (with SB203580), and GSK-3β (with SB216763) did not affect the phosphorylation of YAP induced by Taxol treatment (Supplementary Fig. S1D).
CDK1 is a master regulator of cell cycle and is activated during normal and drug-arrested G2–M (35, 38). JNK1/2 kinases are also activated upon Taxol treatment (Fig. 2C, revealed by increased p-c-Jun levels). We tested whether CDK1 and/or JNK1/2 kinases are responsible for YAP phosphorylation. As shown in Fig. 2C (left panel), both RO3306 (a CDK1 inhibitor) and SP600125 (a JNK1/2 inhibitor) completely reverted the mobility shift of YAP. These drugs are known to induce mitotic exit as shown by the complete loss of phospho-Aurora and degradation of cyclin B. We treated the cells with MG132 along with RO3306 or SP600125 to prevent cyclin B degradation and cells exiting from mitosis. Under these conditions, RO3306, but not SP600125, was still able to completely inhibit YAP phosphorylation (Fig. 2C). In the presence of RO3306, YAP was no longer phosphorylated even when JNK1/2 kinases were strongly activated (Fig. 2C, lanes 3 and 7), suggesting that CDK1, but not JNK1/2 kinases, is likely to be responsible for YAP phosphorylation. Furthermore, another CDK1 inhibitor (roscovitine) also completely inhibited YAP phosphorylation (Fig. 2D). Taken together, these data strongly suggest that YAP phosphorylation induced by Taxol treatment is CDK1 dependent and is independent of JNK1/2 kinases.
CDK1 phosphorylates YAP in vitro
To determine whether CDK1 kinase can directly phosphorylate YAP, we performed in vitro kinase assays with His-tagged YAP as substrates. Figure 2E shows that Taxol-treated mitotic lysates robustly phosphorylated YAP and that CDK1 depletion greatly reduced phosphorylation of His-YAP (top row, compare lane 4 with lane 3). As expected, purified CDK1/cyclin B complex phosphorylated His-YAP in vitro (Fig. 2F). These results indicate that CDK1 directly phosphorylated YAP in vitro. YAP is not a suitable substrate for MEK1, ERK1, and MAPK-p38α kinases in vitro (Supplementary Fig. S1E; ref. 34).
Identification of phosphorylation sites on YAP
Next, we set out to map the phosphorylation site on YAP. YAP was immunoprecipitated from Taxol-treated HeLa cells, Coomassie stained (Supplementary Fig. S2A, inset), and the shifted bands were excised and subjected to Mass/OB-LCA. The following four sites were identified: Serine-109, Threonine-119, Serine-289, and Serine-367 (Fig. 3A; Supplementary Fig. S2A–S2C). S109 is one of the Hippo-mediated phosphorylation sites (10, 12). The rest of the three sites all fit the proline-directed consensus sequence (Fig. 3A) of CDK1-phosphorylation sites (39). Interestingly, all these three sites have been identified as mitotic phosphorylation sites from large-scale proteomic studies (40, 41).
We next examined whether these sites affect the mobility of YAP during Taxol treatment. YAP-mutated S367 or T119 (to alanine) had a reduced mobility shift when compared with wild-type YAP (Fig. 3B, compare lanes 6 and 4 with lane 2). S289A mutation had no effect on YAP mobility induced by Taxol (Fig. 3B, compare lane 8 with lane 2). No further decrease on YAP mobility was observed when T119 and S289, or S289 and S367 were mutated to nonphosphorylatable alanine (Fig. 3B, compare lane 10 with lane 6; lane 14 with lane 4). However, double mutation of T119A and S367A or triple mutation of all three sites largely abolished the mobility shift of YAP, suggesting that T119 and S367 are the main sites responsible for mobility shift of YAP upon Taxol treatment (Fig. 3B).
For comparison, we also tested whether some other known phosphorylation or binding sites are involved in the YAP mobility shift induced by Taxol treatment. Mutating the Hippo phosphorylation sites (S127, S381, and 5SA: S381A/S164A/S127A/S109A/S61A; refs. 10, 12, 13) or TEAD binding site (S94; ref. 5) or the c-Abl phosphorylation site (Y407; ref. 33) did not affect the YAP shift induced by Taxol in HEK293T cells (Supplementary Fig. S2D). The WW domain mutations (W199A/P202A and/or W258/P261A) did not affect the Taxol-induced YAP shift either (data not shown). Together, our data identified novel phosphorylation of YAP during Taxol-arrested G2–M.
CDK1/cyclin B complex phosphorylates YAP at T119 and S289 in vitro and in cells
We have generated phospho-specific antibodies against T119, S289, and S367. Using these antibodies, we demonstrated that CDK1 robustly phosphorylated YAP at T119, S289 and at S367 as well in vitro (Fig. 3C). Addition of RO3306 abolished the phosphorylation (Fig. 3C). To explore whether T119 and S289 are also phosphorylated within cells during Taxol-induced G2–M arrest, we transfected YAP or corresponding nonphosphorylatable mutants into cells, treated the cells with Taxol, and determined levels of phosphorylation by phospho antibodies. Taxol treatment significantly increased the phosphorylation of T119 and S289, and the signal was abolished by mutating T119 or S289 to alanine (Fig. 3D). Taxol treatment also significantly increased the phosphorylation of T119 and S289 in immunoprecipitated endogenous YAP (Fig. 3E). As expected, no signal was detected in control (IgG) immunoprecipitates, suggesting that these antibodies specifically recognize phosphorylated YAP. Lambda phosphatase treatment completely abolished the signal, further confirming the specificity of the phospho-specific antibodies (Fig. 3F). Using inhibitors for CDK1 kinase, we demonstrated that phosphorylation of YAP T119 and S289 is CDK1 kinase dependent (Fig. 3F). Taken together, these results indicate that YAP is phosphorylated at T119 and S289 in cells during antimitotic drug-induced G2–M arrest in a CDK1-dependent manner.
Phosphorylation of YAP occurs in cells during normal mitosis
To determine whether mitotic phosphorylation of YAP occurs during unperturbed/normal mitosis, we collected samples from a double thymidine block and release (22) and determined the phospho levels of YAP during different cell-cycle phases. Figure 4A shows that phosphorylated YAP was readily detected in cells 8–10 hours after being released from double thymidine block, which is coincident with increased levels of cyclin B and phospho-H3 S10 (both of which are mitotic markers). The signal was diminished when cells exit mitosis (Fig. 4A, lane 4). These data strongly suggest that YAP is phosphorylated at T119 and S289 during normal mitosis.
To further investigate the dynamics of YAP phosphorylation in cells, we performed immunofluoresence microscopy with these phospho-specific antibodies. A strong signal was detected in nocodazole-arrested prometaphase cells for both antibodies against S289 and T119 (Supplementary Fig. S3A and S3B, top, red arrows). Very low or no signal was detected in interphase cells (Supplementary Fig. S3A and S3B, yellow arrows). Importantly, phosphopeptide, but not nonphosphopeptide incubation completely blocked the signal, suggesting that these antibodies specifically detect phosphorylated YAP (Supplementary Fig. S3A and S3B, bottom). The specificity of the antibodies was further confirmed by siRNA knockdown of YAP (Supplementary Fig. S3C and S3D). We found that the p-YAP-S289 signal was clearly increased in prophase and peaked in prometaphase/metaphase. The signal was then weakened during metaphase to anaphase transition and further diminished in telophase and cytokinesis (Fig. 4B). Similar staining patterns were generated with p-YAP-T119 antibody staining (Fig. 4C; Supplementary Fig. S4). These data further demonstrate that mitotic phosphorylation of YAP occurs dynamically in cells. Addition of RO3306 or roscovitine largely abolished the signals detected by p-YAP S289 and T119 antibodies in mitotic cells, further indicating that the phosphorylation is CDK1 dependent (Fig. 4D).
Phosphomimetic YAP induces mitotic abnormalities in immortalized cells
Some of the Hippo components regulate mitotic events including chromosome alignment, centrosome duplication, and microtubule dynamics (21). We next examined whether YAP or its phosphorylation is able to trigger mitotic defects. The immortalized epithelial cell line MCF10A, stably expressing vector, YAP, and YAP3D (a phosphomimetic mutant), were used for this purpose (Fig. 5A). As expected, immunofluoresence staining with α-tubulin and γ-tubulin showed normal microtubule/spindle formation and centrosome number during mitosis in control cells (Fig. 5B). Overexpression of wild-type YAP is not sufficient to cause significant mitotic defects in MCF10A cells (Fig. 5D–F). In contrast, mitotic abnormalities were detected in a significant higher percentage of cells expressing YAP3D (3–4-fold). There is a threefold increase of number of YAP3D-expressing cells with disorganization of microtubules and formation of multipolar spindles when compared with control cells (Fig. 5B and D). We also found the percentage of cells with more than two centrosomes (γ-tubulin staining) significantly increased in YAP3D-expressing cells (Fig. 5B and E). As expected, massive chromosome misalignment, chromosome lagging, and chromosome missegregation were observed in about 25% of YAP3D-expressing cells (Fig. 5B, C, and F). These data suggest that ectopic expression of phosphomimetic YAP, but not wild-type YAP, is sufficient to trigger mitotic abnormalities in immortalized epithelial cells.
Mitotic phosphorylation of YAP is required for cellular transformation
Overexpression of YAP transforms MCF10A cells (15). We further examined the biologic significance of mitotic phosphorylation of YAP using these cell lines stably expressing YAP or YAP mutants (Fig. 6A). MCF10A cells expressing YAP-S127A formed colonies in soft agar, however, MCF10A-YAP4A (S127A/T119A/S289A/S367A) cells failed to produce any obvious colonies (Fig. 6B and C). YAP3D overexpression is not sufficient to stimulate anchorage-independent growth in soft agar (data not shown). YAP or YAP-3A (T119A/S289A/S367A) overexpression failed to produce colonies in soft agar (Fig. 6B and data not shown). Together, these data strongly suggest that mitotic phosphorylation is required for YAP-mediated cellular transformation in MCF10A cells.
Mitotic phosphorylation of YAP promotes cell migration and invasion
Recent reports showed that YAP/YAP-S127A also promotes migration and invasion in vitro (42) and metastasis in vivo (43). Interestingly, centrosome amplification has been correlated with cancer invasiveness and enhances migration and invasion of malignant cells via the modulation of microtubule cytoskeleton (44). We, therefore, tested whether mitotic phosphorylation of YAP is involved in cell motility. As expected, ectopic expression of YAP and YAP-S127A increased migration (Fig. 6D) and invasion (Fig. 6E) of MCF10A cells. Mutating CDK1-mediated phosphorylation sites to alanine (YAP-4A) dramatically suppressed YAP-S127A-mediated migration (Fig. 6D) and invasion (Fig. 6E). Interestingly, cells expressing YAP3D possess much higher migration and invasion activity than cells expressing wild-type YAP (Fig. 6D and E). These data suggest that mitotic phosphorylation of YAP promotes cell motility in immortalized epithelial cells.
The above observations indicate that elimination of phosphorylation of YAP by CDK1 inhibitors reduces YAP-, but not YAP3D-mediated cell motility. Indeed, addition of RO3306 almost completely suppressed YAP-S127A-driven migration (Fig. 6F) and invasion (Fig. 6G). Importantly, YAP3D-mediated migration and invasion was not affected by the presence of RO3306 (Fig. 6F and G). Collectively, these data strongly indicate that YAP promotes migration and invasion in a CDK1-phosphorylation dependent manner.
Although the role of the Hippo pathway in tumorigenesis has been firmly demonstrated in several types of cancers, the underlying mechanisms are less clear. Recent reports support the notion that Hippo pathway plays critical roles in maintaining normal mitosis and that inactivation of key members of the Hippo pathway (including Lats2, Mst1/2, Mob1, WW45) leads to mitotic defects in multiple processes including centrosome maturation and disjunction, chromosome alignment, and cytokinesis (19, 45, 46). Mitotic aberrations cause aneuploidy or chromosome instability (47), which is often associated with tumorigenesis (48). The current study identified novel phosphorylation of YAP during mitosis and the mitotic phosphorylation controls YAP's oncogenic activity. Importantly, YAP-3D (a phosphorylation mimetic mutant), but not wild-type YAP, drives mitotic defects (Fig. 5). Thus, our data suggest that there may be a positive layer of regulation for YAP activity during tumorigenesis, and highlight a previously unrevealed mechanism through which YAP exerts its oncogenic function (Fig. 6H). Interestingly, increased CDK1 activity also promotes defects in various mitosis-related processes and aneuploidy, and CDK1 overexpression is often observed in many types of cancers (49). Therefore, it is also possible YAP is one of the critical substrates of CDK1 that mediates CDK1-driven mitotic defects. There are several important questions that need to be answered in the future. For example, does CDK1 phosphorylation of YAP impact its transcriptional activity? Does this phosphorylation affect YAP's binding partners, including the transcription factors TEAD1-4? Furthermore, future studies are required to address the in vivo relevance of CDK1 phosphorylation of YAP and whether this phosphorylation occurs in human patients with cancer. Addressing these unanswered questions should further strengthen the biologic significance of CDK1 phosphorylation of YAP.
Compelling evidence clearly indicates that supernumerary centrosome is one of the hallmarks of cancer and has been correlated with metastatic progression (44, 50). Cell migration and invasion are critical processes for metastasis, which accounts for the majority of deaths related to cancer. YAP hyperphosphorylation induces supernumerary centrosomes (Fig. 5) and promotes migration and invasion (Fig. 6). These data suggest a novel mechanism in which YAP promotes migration and invasion through centrosome amplification in a CDK1-phosphorylation–dependent manner. YAP plays important roles in cancer development and metastasis in several types of cancers and remains as an attractive target for cancer therapy (3). We found that inhibition of CDK1 activity substantially impaired YAP-driven migration and invasion (Fig. 6F and G). Because R-roscovitine (Seliciclib) has been in phase I clinical trials with low toxicity (51), our findings support the feasibility of using CDK1 inhibitors in human cancers especially for those in which the Hippo-YAP signaling is dysregulated.
Upon treatment with antimitotic agents, one of the prominent responses of the Hippo pathway is the marked increase in the phosphorylation of YAP (Fig. 1) and KIBRA (23). Interestingly, the Hippo pathway core components Mst1/2 and Lats1/2 have also been shown to be regulated during mitosis. Autophosphorylation (kinase activity) of Mst and Lats is increased in U2OS cells upon nocodazole treatment (52). However, we could not detect any significant change of Mst or Lats activity (by phospho antibodies) in HeLa and MCF-7 cells (Fig. 1A and data not shown), suggesting that the mitotic activation of Mst and Lats is cell-type specific. Lats2 was also phosphorylated by Aurora-A at the centrosomes during mitosis (53). YAP is a direct substrate of Lats1/2 (10, 12, 13) and KIBRA associates with Lats1/2 (25). Interestingly, CDK1 also phosphorylates KIBRA (24) and Lats1 (30) during mitosis, and one question is whether or how KIBRA/Lats1/2 is involved in the regulation of mitotic phosphorylation of YAP. Furthermore, YAP and TAZ are paralogs and have similar and distinct functions (2, 4). We also found that TAZ is upshifted upon Taxol or nocodazole treatment (Fig. 1A). Although all three phosphorylation sites on YAP are conserved in vertebrates, however, surprisingly, none of the phosphorylation sites of YAP exists on TAZ. We are currently investigating how TAZ is regulated and whether TAZ also plays a role in mitosis similar to YAP.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: S. Yang, J. Dong
Development of methodology: S. Yang
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Yang, L. Zhang, M. Liu, S.-J. Ding
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Yang, L. Zhang, M. Liu, S.-J. Ding, J. Dong
Writing, review, and/or revision of the manuscript: S. Yang, M. Liu, S.-J. Ding, J. Dong
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Yang, L. Zhang, R. Chong, Y. Chen
Study supervision: S. Yang, J. Dong
The authors thank Dr. Nick George for designing siRNA oligos targeting YAP and Tom Dao for assistance with confocal microscopy at the imaging core facility at Nebraska Center for Cellular Signaling. The authors also thank Drs. Joyce Solheim, Robert Lewis, and Keith Johnson for critical reading and comments on the article.
This work was supported in part by grant 8P20 GM103489 from the National Institutes of Health (J. Dong).