TTK/hMps1 participates in the regulation of DNA damage checkpoint response by phosphorylating CHK2 on threonine 68.

CHK2/hCds1 plays important roles in the DNA damage-induced cell cycle checkpoint by phosphorylating several important targets, such as Cdc25 and p53. To obtain a better understanding of the CHK2 signaling pathway, we have carried out a yeast two-hybrid screen to search for potential CHK2-interacting proteins. Here, we report the identification of the mitotic checkpoint kinase, TTK/hMps1, as a novel CHK2-interacting protein. TTK/hMps1 directly phosphorylates CHK2 on Thr-68 in vitro. Expression of a TTK kinase-dead mutant, TTK(D647A), interferes with the G(2)/M arrest induced by either ionizing radiation or UV light. Interestingly, induction of CHK2 Thr-68 phosphorylation and of several downstream events, such as cyclin B1 accumulation and Cdc2 Tyr-15 phosphorylation, is also affected. Furthermore, ablation of TTK expression using small interfering RNA results not only in reduced CHK2 Thr-68 phosphorylation, but also in impaired growth arrest. Our results are consistent with a model in which TTK functions upstream from CHK2 in response to DNA damage and suggest possible cross-talk between the spindle assembly checkpoint and the DNA damage checkpoint.

To ensure that critical events of cell division, such as DNA replication and chromosome segregation, are faithfully and precisely executed, eukaryotic cells employ checkpoints to control the order and timing of cell cycle progression (1,2). DNA damage or other environmentally induced genotoxic stress triggers surveillance pathways, leading to checkpoint activation. The DNA damage checkpoints enable cells to delay cell cycle progression or, alternatively, to undergo apoptosis (3). Defects in checkpoint genes allow cells to grow and divide out of control, which often leads to increased sensitivity to damaging agents, as well as genome instability, and predisposition to cancer (4,5). Accordingly, knowledge of the DNA damage response pathways and participating proteins is extremely important for understanding tumorigenesis and for the development of effective therapy for cancer.
CHK2/hCds1, an evolutionary conserved protein kinase, was identified by several groups as a mammalian homolog of the Saccharomyces cerevisiae Rad53 and Schizosaccharomyces pombe Cds1 (6 -10). Rad53 and Cds1 are required for responses to DNA damage and stalled replication (11,12). Likewise, the mammalian CHK2 responds to various types of DNA damage and interference with DNA replication (1,13). Because of its predominantly nuclear localization, CHK2 is thought to act as an effector protein in DNA damage checkpoint pathways. Through its functional links with other proteins, the signals resulting from DNA lesions can be transduced instantly and amplified rapidly to effect diverse cellular responses, including growth arrest, DNA repair, and programmed cell death, by a cascade of phosphorylation and dephosphorylation events. On exposure to genotoxic stress, CHK2 becomes fully activated through a chain of reactions, including the initial wave of phosphorylation carried out by ATM or ATR kinases at sites within the SQ/TQ cluster domain (SCD), 1 preferentially at Thr-68 (14 -17), followed by forkhead-associated domain-mediated oligomerization and autophosphorylation on residues Thr-383 and Thr-387, located in the activation loop segment within the catalytic domain, and on Ser-516 in the C terminus (18 -21). The kinases, ATM and ATR, which are members of the phosphatidylinositol 3-kinase family, are two major upstream controllers of CHK2 (22)(23)(24). ATM is specifically activated by ionizing radiation (IR) and other sources of DNA double strand breaks such as those generated spontaneously during DNA replication or by radiomimic drugs, whereas ATR is suggested to function in parallel with ATM in responding to stalled replication forks resulting from hydroxyurea treatment or UV light-induced DNA damage. As a result, after DNA damage, CHK2 is phosphorylated and thus activated in either an ATMdependent or ATR-dependent manner, depending on the type of damage.
When DNA damage occurs, activated CHK2 targets several cell cycle modulators, such as p53, Brca1, and two Cdk-activating phosphatases, Cdc25A and Cdc25C. These events ultimately lead to cell cycle arrest, DNA repair, or programmed cell death. Activated CHK2 phosphorylates Cdc25A on Ser-123, leading to ubiquitination and proteasome-mediated degradation of Cdc25A. This results in immediate G 1 /S cell cycle blockade and stalled S phase progression (25). Furthermore, CHK2 is involved in maintaining G 2 /M arrest by phosphorylating the mitosis-promoting Cdc25C phosphatase on Ser-216. The phos-phorylation enhances the association of Cdc25C with 14-3-3, resulting in Cdc25C inactivation (9, 26 -29). CHK2 also phosphorylates p53 at several sites, including Ser-20 in its transactivation domain, thus contributing to its stabilization and enhancement of its transcriptional activity toward its downstream target genes (30 -34). In addition to growth arrest, CHK2 regulates p53-mediated apoptosis in both an ATM-dependent and an ATM-independent manner (33,35,36). Recently, CHK2 was also shown to regulate apoptosis through a p53-independent pathway by phosphorylating the promyelocytic leukemia protein (37). Moreover, CHK2 phosphorylates Brca1, which is involved in DNA double strand break repair, further implicating CHK2 in DNA repair (38). In addition to its functions in the regulation of tumor suppressors, CHK2 itself seems to be a tumor suppressor. Mutations in CHK2 occur in Li-Fraumeni families with wild-type p53 (39). Moreover, CHK2 mutation carriers have an increased risk of breast cancer (40). To summarize, CHK2 is a checkpoint protein kinase that coordinates cell cycle responses and plays important roles in maintaining genomic stability.
In this study, using a yeast two-hybrid screen, we identified TTK as a novel CHK2-interacting protein and demonstrated that it binds directly to CHK2, both in vitro and in vivo. We also established a Tet-Off (Clontech) cell line that expressed kinasedead (KD) TTK (D647A) under the control of tetracycline to investigate the roles of TTK in cell cycle checkpoints. We found that expression of TTK-KD impaired G 2 /M arrest and CHK2 Thr-68 phosphorylation induced by DNA damage. Similarly, disruption of TTK expression by small interfering RNA (siRNA) also caused a reduction in CHK2 Thr-68 phosphorylation. Finally, we present evidence that TTK exerts its effects possibly by directly phosphorylating CHK2 Thr-68. Together, our data support a model in which TTK functions upstream from CHK2 in response to DNA damage and participates in DNA damage checkpoints through its effects on CHK2.

EXPERIMENTAL PROCEDURES
Yeast Two-hybrid Screen-For the yeast two-hybrid screen, the Nterminal domain of CHK2 (residues 1-223) was amplified by PCR and cloned into pDBTrp (Invitrogen) between the SalI and StuI sites. This was used as bait to screen a human testis cDNA library cloned in the pACT2 vector (Clontech). The yeast strains and assay conditions were those used in the ProQuest TM two-yeast system (Invitrogen).
For IR treatment, cells were exposed to 8 Gy, or in the case of U2OS cells, 4 Gy, x-rays using the Torrex 150D inspection system (EG&G), then incubated for the indicated period. For UV irradiation, PBSwashed cells were irradiated in uncovered tissue culture dishes with 254-nm light in a UV cross-linker (Spectronics). The culture medium was then added back and the cells incubated for the indicated period.
Plasmids-The pcDNA3 vector (Invitrogen) modified by insertion of sequences encoding a HA tag was used to generate the expression construct for full-length TTK (1-857). pcDNA3-HA-TTK (D647A) was made from the pcDNA-HA-TTK vector by PCR-based site-directed mutagenesis. For Tet-Off expression, a 3.0-kb TTK fragment including 500 bp of the 3Ј-untranslated region was cloned into the EcoRI site of the pTRE vector (Clontech).
For mammalian expression of CHK2, the cDNA was cloned into the BamHI and KpnI sites of the pXJ-myc vector (a kind gift from V. Yu, Institute of Molecular and Cell Biology, Singapore) and expressed as a myc-tagged protein.
For expression in Escherichia coli, cDNAs coding for full-length TTK and truncation mutants were cloned into pGEX4T vectors (Amersham Biosciences) for expression as glutathione S-transferase (GST) fusion proteins. Full-length CHK2 or its truncation mutants were expressed from pRSET vectors (Invitrogen) as His-tagged fusion proteins.
Recombinant Proteins-GST fusion proteins were expressed in E. coli strain DH5␣. After induction with 1 mM isopropyl-␤-D-thiogalactopyranoside, the bacteria were lysed by sonication in sonication buffer (50 mM NaH 2 PO 4 , 300 mM NaCl, and 20% glycerol) containing 1% Triton X-100 and 1 mg/ml lysozyme, then the lysates were incubated with glutathione-Sepharose beads (Amersham Biosciences) and the bound proteins eluted with 20 mM reduced glutathione.
WCEs were boiled for 5 min in protein sample buffer (0.7 M ␤-mercaptoethanol, 4% SDS, 0.17 M Tris, pH 6.8, 0.17 mg/ml bromphenol blue, 10% glycerol), and a portion (20 -30 g of protein) was analyzed on SDS-polyacrylamide gels. The resolved proteins were wet transferred to PROTRAN nitrocellulose membrane (Schleicher & Schuell), then the blots were blocked for 30 min at room temperature in 1% milk in PBS containing 0.05% Tween 20 and incubated with primary antibody followed by appropriate secondary horseradish peroxidase-conjugated antibody. After washing, the blots were treated with chemiluminescent reagents (Pierce) and exposed to BioMax films (Kodak).
Protein-Protein Interaction Assays-For the GST pull-down assay, bacterial lysates containing GST fusion proteins were prepared by sonication as described above, then incubated with glutathione-Sepharose beads for 1 h at 4°C. After two washes with sonication buffer and one wash with coimmunoprecipitation buffer (10 mM Tris, pH 7.5, 1 mM EDTA, 150 mM NaCl, 10% glycerol, and 0.15% Nonidet P-40), the beads were incubated for 1 h at room temperature in coimmunoprecipitation buffer with 35 S-labeled in vitro translated CHK2. After washes with coimmunoprecipitation buffer, bound CHK2 were eluted using 10 mM reduced glutathione and analyzed by SDS-PAGE followed by autoradiography.
For coimmunoprecipitation experiments, 293T cells were plated on 60-mm dishes and transiently transfected by calcium phosphate precipitation. At 48 h post-transfection, cells were lysed in TEGN buffer containing 1 mM PMSF, protease inhibitors, 10 mM NaF, 10 mM ␤-glycerophosphate, 5 nM microcystin LR, and 1 mM dithiothreitol. The WCEs were mixed with an equal volume of TEG buffer (10 mM Tris, pH 7.5, 1 mM EDTA, and 20% glycerol), 15 l of 50% protein G beads (Pierce), and antibodies. The mixtures were rocked for 2-3 h at 4°C, then washed three times with 1:1 TEG/TEGN buffer. Bound proteins were analyzed by SDS-PAGE followed by Western blotting.
Flow Cytometry-For FACS analysis, both attached and floating cells were harvested, resuspended in cold PBS, fixed by adding 10 volumes of cold (Ϫ20°C) methanol, and stored at Ϫ20°C for at least 1 h. Cells were rehydrated with cold PBS and stained for 30 min at 37°C with 60 g/ml propidium iodide (Sigma) in PBS containing 50 g/ml RNase A. The DNA content was analyzed using a FACS Calibur flow cytometer (BD Biosciences).
In Vitro Kinase Assay and Immunoprecipitation/Kinase Assay-For the in vitro kinase assay, purified recombinant full-length wild-type (WT) or KD GST-TTK fusion proteins were incubated with purified His-CHK (D347A) for 15 min at 30°C in kinase buffer (20 mM Tris, pH 7.5, 10 mM MgCl 2 , 2 mM MnCl 2 , 1 mM PMSF, and 1 mM dithiothreitol) supplemented with a mixture of protease inhibitors, 10 mM NaF, 5 nM microcystin LR, and 50 M ATP. The reaction was stopped by the addition of a 0.5 volume of 3ϫ protein sample buffer and the proteins resolved by SDS-PAGE followed by detection with anti-CHK2-phosphothreonine 68 rabbit polyclonal antibody (2661, Cell Signaling Technology).
For immunoprecipitation/kinase assays, TTK was first immunoprecipitated from LNCaP or MCF7 cell lysates with anti-TTK (sc-540), washed twice in 1:1 TEG/TEGN buffer, once in 1:1 TEG/TEGN buffer containing 0.7 M LiCl, and twice in kinase buffer. Kinase assay was then performed as described above using His-CHK2-D347A as substrate.

TTK Interacts with CHK2 Both in Vitro and in Vivo-
To delineate the CHK2-mediated signaling pathways, we conducted a yeast two-hybrid screen using the CHK2 N-terminal domain (1-243) as bait to isolate novel CHK2-interacting proteins. By screening a human testis cDNA library, we obtained several positive clones. One of the clones, d2, encoded 470 amino acid residues encompassing the C-terminal portion (amino acids 387-857) of TTK (42), the human homolog of the yeast dual specific protein kinase, Mps1. Human Mps1, like its yeast counterpart, plays an essential role in the spindle assembly checkpoint (45,54) To determine whether TTK interacted directly with CHK2, an in vitro GST pull-down assay was performed using 35 Slabeled in vitro translated CHK2 and bacterially produced GST fusion constructs of TTK (Fig. 1A). As shown in Fig. 1B, the originally cloned d2 (387-857) construct and the d2C (521-857) construct, but not GST, brought down CHK2 in vitro. The KD mutant d2 (D647A) also bound to CHK2, demonstrating that the kinase activity of TTK was not essential for the interaction in vitro. The interaction, although specific, was not strong. It is possible that bacteria-expressed recombinant d2 protein lacks additional modification or was not properly folded for optimal interaction.
To validate the interaction between TTK and CHK2 in mammalian cells, 293T cells were transiently transfected with expression vectors encoding HA-tagged, full-length TTK together with myc-tagged CHK2. WCEs were then immunoprecipitated with anti-HA antibodies, and binding of CHK2 was detected by Western blotting with anti-myc antibody. In addition, we also examined the requirement for kinase activity for the interaction using different combinations of WT and KD mutants (Fig.  1C). The results showed that anti-HA antibody pulled down myc-CHK2 in cells coexpressing HA-TTK (Fig. 1C, lanes 3 and  4 cf. lanes 8 and 9), demonstrating that the two proteins form complexes in vivo. The interaction did not require the kinase activity of CHK2 because the KD CHK2 mutant (D347A) also interacted with WT TTK (Fig. 1C, lane 4). Whether HA-

FIG. 1. TTK interacts with CHK2 in vitro and in vivo.
A, schematic diagram of full-length TTK and two truncated constructs (d2, and d2C), with the first and last amino acids shown in parentheses. For in vitro binding assays, these cDNAs were cloned into pGEX vectors and expressed as GST fusion proteins. B, GST pull-down assays were performed by incubating 35 S-labeled in vitro translated CHK2 with GSTfused d2C, d2, or d2KD (D647A). The proteins pulled down by glutathione-Sepharose beads were analyzed by SDS-PAGE followed by autoradiography (right panel). The silver-stained input GST-fused proteins are shown in the left panel. C, overexpressed CHK2 and TTK interact in 293T cells. 293T cells were transfected with HA-tagged TTK-WT or TTK-KD alone or together with myc-tagged CHK2-WT or CHK2-KD, then lysates were immunoprecipitated (IP) with anti-HA antibody and analyzed by Western blotting using anti-myc antibody to detect CHK2. D, coimmunoprecipitation of endogenous CHK2 and TTK. HeLa cells were untreated or treated with 8 Gy x-rays and then collected at the indicated times. CHK2 was immunoprecipitated from HeLa WCEs using anti-CHK2, and TTK in the immune complex was then detected by Western blotting using anti-TTK.

TTK/hMps1 in DNA Damage Checkpoint
TTK-KD coimmunoprecipitated with myc-CHK2 was inconclusive because the protein consistently expressed at a level much lower than HA-TTK-WT when coexpressed with myc-CHK2-WT (Fig. 1C, compare lane 3 with lane 6).
The interaction between endogenous CHK2 and TTK was also examined by immunoprecipitating endogenous CHK2 from HeLa lysates followed by Western blotting against endogenous TTK. Fig. 1D shows that the two indeed interacted in cells, and the interaction was slightly enhanced by x-rays, after normalizing to the amount of precipitated CHK2.
TTK Phosphorylates CHK2 Thr-68 in Vitro-To understand the functional relationship between TTK/hMps1 and CHK2, we first examined whether TTK/Mps1 can directly phosphorylate CHK2 in vitro. To this end, in vitro kinase assays were performed by incubating bacteria-expressed, purified GST-d2C The results showed that GST-d2, which contains the kinase domain of TTK, phosphorylated CHK2 in vitro and that sites on both the N-and C-terminal domains of CHK2 were phosphorylated ( Fig. 2A, lanes 10 -12).
No phosphorylation was seen when the d2KD mutant was used (lanes 13-15). The d2C construct, which is slightly shorter than d2 in the N terminus but still contains an intact kinase domain, was less efficient in phosphorylating CHK2 ( Fig. 2A, lanes 7-9), suggesting that sequences in addition to the kinase domain are required for the full activity of TTK. Notably, TTK phosphorylated CHK2 on Thr-68, as demonstrated by in vitro kinase assay followed by Western blotting using antibody that specifically In vitro kinase assays were performed by incubating purified GST-d2C, GST-d2, or GST-d2KD (D647A) with purified His-tagged full-length CHK2 or the truncated fragments, CHK2N (amino acids 1-243) or CHK2C (244 -543), in the presence of [␥-32 P]ATP, then the products were separated on SDS gels and phosphorylation visualized by autoradiography. B, TTK specifically phosphorylates CHK2 on Thr-68 in vitro. In vitro kinase assays were performed by incubating purified recombinant full-length GST-WT or KD TTK with purified His-CHK2 in the presence of unlabeled ATP. Phosphorylation was detected by Western blotting using anti-phospho-Thr-68-specific CHK2 antibody. C, in vitro kinase assays were conducted as described in B, and phosphorylation on CHK2 Ser-19 or Thr-68 was detected with the corresponding antibodies. Lysates prepared from x-ray-irradiated 293T cells were used as positive controls. D and E, immunoprecipitated (IP) endogenous TTK phosphorylates CHK2 on Thr-68. Immunoprecipitation/kinase assays were performed using TTK immunoprecipitated from LNCaP (D) or MCF7 (E) cells that were either untreated (Ϫ) or treated (ϩ) with 8 Gy x-rays and harvested 30 min after. Normal rabbit IgG (NR, lane 1) was used as an immunoprecipitation control. The amounts of Thr-68 phosphorylation on the added substrate (His-CHK2) in lanes 2 and 3 were quantified by densitometry and are shown in F. Although the overall phosphorylation as revealed by 32 P incorporation was unchanged before and after IR, the specific phosphorylation on Thr-68 was enhanced by IR. G, expression of TTK-KD in cells diminishes the induction of Thr-68 phosphorylation on endogenous CHK2 by DNA damage. 293T cells were transiently transfected with HA-tagged TTK-KD or mock-transfected for about 45 h before irradiation with x-rays (left panels) or UV light (right panels). Induction of CHK2 Thr-68 phosphorylation was compared at the indicated time points by Western blotting.
Because TTK/hMps1 interacted with and phosphorylated CHK2 on Thr-68, we postulated that overexpression of a TTK-KD mutant (T647A) in cells would down-regulate CHK2 Thr-68 phosphorylation by competing with endogenous TTK-WT. Indeed, 293T cells transiently transfected with TTK-KD exhibited reduced IR and UV light-induced CHK2 Thr-68 phosphorylation (Fig. 2G). Taken together, our data indicate that phosphorylation on CHK2 Thr-68 is possibly contributed at least in part by TTK in vivo.
Expression of TTK-KD Diminishes the G 2 /M Arrest Induced by DNA Damage-CHK2 plays important roles in cell cycle control in response to DNA damage. TTK/hMps1, although important for the mitotic spindle checkpoint (45), has never been implicated in the cellular response to DNA damage. To explore the roles of TTK in DNA damage-induced checkpoints, two stable cell lines, HeLa 10-3 and HeLa 6, were established, which had been transfected, respectively, with a control empty vector and a vector that expresses the HA-tagged TTK-KD mutant (T647A) on withdrawal of tetracycline (Fig. 3, A-C). DNA damage-induced changes in cell cycle progression after exposure to IR or UV light were monitored by flow cytometry.
As shown in Fig. 3D, on removal of tetracycline from the medium, HeLa 10-3 cells displayed normal G 2 /M arrest after

TTK/hMps1 in DNA Damage Checkpoint
x-ray irradiation, with the G 2 /G 1 ratio reaching a maximum at 12 h after exposure (Fig. 3E, X-ray). In contrast, HeLa 6 cells, expressing TTK-KD, escaped G 2 /M block and reentered G 1 phase (Fig. 3D, X-ray), showing a G 2 /G 1 ratio 43% lower than the control 10-3 cells at 12 h after irradiation (Fig. 3E, X-ray). Similarly, HeLa 6 cells were incapable of arresting at G 2 /M after UV irradiation (Fig. 3D, UV), possibly because the majority of cells were retained in S phase. Consequently the G 2 /M increase of HeLa 6 cells after UV was significantly below that of 10-3 cells at all three time points examined (Fig. 3E, UV). These results indicate that HeLa 6 cells expressing TTK-KD are defective in G 2 /M arrest after DNA damage.
Expression of TTK-KD Impairs CHK2-mediated Checkpoint Signaling after DNA Damage-Because our results showed that TTK and CHK2 interacted in vivo, it was therefore possible that TTK-KD functioned by blocking the DNA damageinduced cell signaling to CHK2 in HeLa 6 cells. To address this possibility, several key proteins involved in the DNA damage checkpoint pathway were examined by Western blotting. Consistent with the defective G 2 /M arrest, DNA damage-induced (IR or UV) phosphorylation of CHK2 Thr-68, a condition required for full activation of CHK2, was reduced in HeLa 6 cells (pT68; Fig. 3, A and C). It was noted that the levels of CHK2 protein were sometimes lowered in HeLa 6 cells especially after DNA damage (Fig. 3, A and C), a condition possibly related to lower CHK2 activity as a result of reduced Thr-68 phosphoryl-ation. A similar observation was made in TTK siRNA-transfected U2OS cells (see Fig. 5A and 5B, below). Furthermore, events downstream of CHK2 were also affected, such that Tyr-15 phosphorylation of Cdc2, which inactivates Cdc2 and results in G 2 /M arrest, was also decreased in HeLa 6 cells in response to IR and UV light treatment (pY15; Fig. 3A, lanes 6  and 8; Fig. 3C, lanes 2, 3, 5, and 6). Moreover, accumulation of cyclin B1, a protein degraded upon mitotic exit, was reduced in HeLa 6 cells after IR treatment (Fig. 3B), indicative of a failure in G 2 /M arrest.
Taken together, our results strongly suggest that expression of the TTK-KD mutant attenuates DNA damage-induced G 2 /M arrest by blocking CHK2 Thr-68 phosphorylation and activation.
Ablation of TTK Expression by siRNA Diminishes CHK2 Thr-68 Phosphorylation and G 2 /M Checkpoint-Because the KD mutant works by competing with endogenous TTK, we also sought to disrupt the expression of endogenous TTK by siRNA. HeLa 10-3 cells were transiently transfected with TTK siRNA TK1, and for comparison, CHK2 siRNA 2-3. The levels of endogenous TTK and CHK2 were effectively down-regulated 40 -48 h after transfection (Fig. 4A, lanes 2 and 3). Cells were then irradiated with x-rays or UV light, and after a period of recovery, collected for Western analysis (Fig. 4B) and flow cytometry (Fig. 4C). As in the TTK-KD-expressing cells, the initial induction of CHK2 Thr-68 phosphorylation was reduced TTK/hMps1 in DNA Damage Checkpoint significantly 30 min after IR in TTK siRNA (TK1)-transfected cells (Fig. 4B, lanes 2 and 8). The reduction was even more pronounced after UV irradiation, with Thr-68 phosphorylation barely detected in TK1-transfected cells (Fig. 4B, lanes 11 and  12). We also examined one of the CHK2 targets, Cdc25A, which is degraded rapidly after DNA damage (25,56). Consistent with previous observations (25,56), the levels of Cdc25A were transiently reduced after x-rays and UV irradiation in control transfected cells (Fig. 4B, lanes 1-6). The reduction after IR, a result of CHK2 activation (25), was blocked in TK1-transfected cells, consistent with decreased CHK2 activity probably as a result of reduced Thr-68 phosphorylation (Fig. 4B, lane 8). In contrast, the destruction of Cdc25A after UV irradiation, an effect mediated by CHK1 (56), was not rescued in TK1-transfected cells (Fig. 4B, lanes 11 and 12).
The cell cycle distribution of siRNA-transfected cells was analyzed by FACS. In control scrambled siRNA-transfected cells, the G 2 /M population increased by about 2-fold at 16 h after x-ray irradiation (Fig. 4C). In contrast, the G 2 /M increase in TK1-transfected cells was more transient and barely maintained, such that it peaked briefly at 12 h then subsided at 16 h to a level similar to the untreated control (Fig. 4C). UV irradiation resulted in visible increase of sub-G 1 population in addition to G 2 /M accumulation in control transfected 10-3 cells (Fig.  4C). In TK1-transfected cells, the increase in sub-G 1 population was more evident, and the G 2 /M arrest was barely detected (Fig. 4C). Because our data suggest that CHK2 and TTK may lie in the same signaling pathway, we also transfected HeLa 10-3 cells with CHK2 siRNA for comparison. Although CHK2 has been shown to be required for DNA damage-induced check-point response in chk2-null murine cells (33 35, 36, 57), its roles in human cells have not been characterized in detail. Interestingly, the results with CHK2 siRNA transfection closely mimicked those with TTK-targeting siRNA (Fig. 4C), with the exception that a significantly higher percentage of sub-G 1 population was observed after UV irradiation when CHK2 was ablated (Fig. 4C). Note that the difficulty in maintaining the G 2 /M arrest has also been observed in chk2-null ES cells after IR (57). These results are consistent with a model in which TTK and CHK2 lie in the same signaling pathway in response to DNA damage and suggest a role of TTK/hMps1 in DNA damage-induced cell cycle checkpoint.
To demonstrate that our observation regarding the role of TTK/hMps1 is not limited to HeLa cells, siRNA experiments were also conducted using another cell line, U2OS, which possesses functional p53 and a normal checkpoint response (25,56). As shown in Fig. 5A, the introduction of TK1 siRNA blocked the expression of TTK and reduced the CHK2 Thr-68 phosphorylation by 80% at 30 min after IR. The blockade of CHK2 Thr-68 phosphorylation by TTK siRNA was even more evident after UV irradiation (Fig. 5B). These data are in line with our results obtained in HeLa (Fig. 3, A and C, and Fig. 4B) and TTK-KD-transfected 293T cells (Fig. 2G) and provide further support for the notion that TTK, regardless of its cellular context, participates in the induction of CHK2 Thr-68 phosphorylation after DNA damage.
Next, we also examined the effects of TTK-targeting siRNA on cell cycle progression after IR in U2OS cells. Control siRNAtransfected cells responded normally to IR, showing a 90% increase in G 2 /M population and a significant reduction in S TTK/hMps1 in DNA Damage Checkpoint phase (Fig. 5D). Compared with the control, the TK1 siRNAtransfected cells displayed only modest changes in G 2 /M (48% increase) and S population after IR (Fig. 5D).
TTK/hMps1 is known to be essential for proper execution of the spindle checkpoint and normal mitotic progression (45,58,59). To exclude the possibility that the effects we observed using TK1 siRNA might be a result of a defective mitotic checkpoint rather than a compromised DNA damage checkpoint, we then sought to inactivate the mitotic checkpoint directly by transfection of siRNA targeting MAD2, a spindle checkpoint player downstream of TTK responsible for the inhibition of the anaphase promoting complex (60). In this case, however, the levels of CHK2 Thr-68 phosphorylation in MAD2knock-down cells were similar to the control transfected cells after either IR or UV (Fig. 5C), suggesting that the reduced CHK2 Thr-68 phosphorylation in TTK-ablated cells is not a result of failure in the mitotic spindle checkpoint. Consistent with the induction of CHK2 Thr-68 phosphorylation, IR induced a 75% G 2 /M increase in MAD2-knock-down cells (Fig.  5D), a level significantly higher than that of TTK-ablated cells (48% increase). The defects in growth arrest can also be seen in the G 2 /G 1 ratios, with the control and MAD2-knock-down cells showing a 3-fold increase but the TTK-knock-down cells showing only a 2-fold increase after IR (Fig. 5E). Note that unlike HeLa cells, the cell cycle distribution was perturbed in both TTK-and MAD2-knock-down U2OS cells even before DNA damage, possibly because of their defects in controlling the mitotic progression (58). Nevertheless, our results indicate that a defective spindle assembly checkpoint does not prevent the execution of DNA damage response, and TTK, in addition to its roles in the mitotic checkpoint, may participate in the regulation of DNA damage checkpoint. DISCUSSION In this study, we have shown that TTK/hMps1 interacts with CHK2, both in vitro and in vivo. In addition, we demonstrated that TTK was required for proper G 2 /M arrest after DNA damage in HeLa and U2OS cells. Furthermore, TTK phosphorylated CHK2 on Thr-68. Both the expression of the TTK-KD mutant and the silencing of TTK expression using siRNA resulted in reduced CHK2 Thr-68 phosphorylation.
TTK/hMps1 Is a Novel Non-phosphatidylinositol 3-Kinase Protein Kinase That Phosphorylates CHK2 Thr-68 and Regu-lates the CHK2-mediated Checkpoint Response after DNA Damage-Although ATM and ATR are generally considered as the primary upstream controllers of CHK2, the question of whether other kinases also phosphorylate CHK2 and regulate its activation is unanswered. ATM and ATR phosphorylate overlapping but distinct sites in the SCD of CHK2 after DNA damage (15). ATM primarily controls the response to double strand breaks, whereas ATR responds to other types of damage. Despite the close relationship between CHK2 and ATM in responding to IR, evidence provided by gene knock-out studies in mice suggests that they do not lie exactly in a linear pathway (35). In fact, CHK2 is phosphorylated on Thr-68 in AT fibroblasts (20). The study described in this report provides an additional mechanism for the regulation of CHK2. Our results showed that TTK phosphorylated CHK2 on Thr-68 in vitro. Furthermore, both the expression of the TTK-KD mutant and the inhibition of TTK expression using siRNA resulted in reduced CHK2 Thr-68 phosphorylation after DNA damage, suggesting that TTK may also function as a CHK2 Thr-68 kinase in vivo. More importantly, the reduced CHK2 Thr-68 phosphorylation was accompanied by a decreased checkpoint response and defective cell cycle arrest, suggesting that TTK is involved in the regulation of cell cycle progression after DNA damage.
Our results support a model in which TTK functions upstream of CHK2 (Fig. 6) and is involved in the DNA damage checkpoint pathways, perhaps acting as a sensor.
Apart from TTK, recent reports indicate that Plk1 and Plk3, two members of the Polo-like kinase (Plk) family, may also be involved in the phosphorylation of CHK2 (61,62). Plk1, the mammalian ortholog of yeast Cdc5, regulates many critical events in mitosis, including mitotic entry, centrosome maturation and separation, the metaphase-anaphase transition, mitotic exit, and the onset of cytokinesis (63). Plk1 phosphorylates recombinant CHK2 on Thr-68 in vitro, and coexpression of Plk1 and CHK2 increased Thr-68 phosphorylation of CHK2 (62). Plk3 also phosphorylates CHK2 in vitro; this also occurs preferentially in the N-terminal SCD, but not at Thr-68 (61). Curiously, Plk1 and Plk3 respond to DNA damage differently, with Plk1 being inhibited (64) and Plk3 activated (61,65). Obviously, there are mechanistic questions to be answered before their roles in DNA damage checkpoint can be defined. Nevertheless, these studies in conjunction with ours point to FIG. 6. Model depicting the possible role of TTK in DNA damage-induced checkpoint responses. After DNA damage, TTK phosphorylates CHK2 on Thr-68, which leads to activation of CHK2 and phosphorylation of the downstream targets, Cdc25 and p53. The phosphatase activity of Cdc25C is inactivated by CHK2 phosphorylation, leading to accumulation of Cdc2 Tyr-15 phosphorylation and Cdk inactivation. CHK2 may also regulate the activity of p53 and consequently the expression of its downstream target genes, such as p21 and AIP1, leading to either growth arrest or apoptosis. the emerging issues of how mitotic cells respond to DNA damage and how components of checkpoint pathways and the mitotic machinery interact functionally with one another. Perhaps the immediate question to be addressed is the role of CHK2 in mitotic cells. Although CHK2 plays a pivotal part in DNA damage-induced cell cycle arrest and apoptosis, whether it is also involved in the regulation of mitosis remains to be determined.
Mechanistically, how TTK contributes to the DNA damage checkpoint response remains unclear, and several hypothetical models could be envisaged. One is that TTK may cooperate with ATM in the activation or maintenance of DNA damage checkpoints. Although in the present study TTK was found to be a CHK2-interacting protein, the possibility that it may also work through ATM to activate CHK2 indirectly in vivo cannot be excluded. Alternatively, because TTK siRNA did not completely block CHK2 Thr-68 phosphorylation after IR, TTK may act independently via a parallel, partly redundant pathway. The actual regulatory mechanism awaits further investigation.
TTK Functions in Two Distinct Checkpoint Pathways and May Provide a Lateral Link between DNA Damage Checkpoints and Spindle Assembly Checkpoints-Given the role of TTK/ hMps1 in the DNA damage checkpoint demonstrated in our study and its previously identified roles in spindle assembly checkpoint, it may appear puzzling how a protein can function in two seemingly distinct pathways. Mps1 has been found in kinetochores and centrosomes and is required for proper execution of spindle assembly checkpoint and more controversially for centrosome duplication. Stucke et al. (45) showed that in U2OS cells, hMps1/TTK localizes to kinetochores, but not to centrosomes, and that it is necessary for the spindle assembly checkpoint, but not for centrosome duplication. In contrast, Fisk et al. (53) showed that mMps1 associates with both kinetochores and centrosomes and is required for centrosome duplication. In any case, the importance of TTK in the control of mitosis, from yeast to mammals, is indisputable. What has been overlooked is the role of the majority of TTK, which is located outside of the centrosomes, especially in interphase cells. It is possible that this TTK functions as a DNA damage sensor to activate CHK2 directly, in a manner independent of its role in the mitotic spindle checkpoint. This notion is substantiated by our siRNA experiments showing that knockdown of MAD2, a spindle checkpoint regulator downstream from TTK, had no effect on either the induction of CHK2 Thr-68 phosphorylation or G 2 /M arrest (Fig. 5). Nevertheless, for the reasons described below, we cannot exclude the possibility that TTK may also act as a DNA damage sensor in mitotic cells.
It is noteworthy that Thr-68-phosphorylated CHK2, previously found in nuclear foci at sites of DNA damage (66), was recently discovered to be present in centrosomes in the absence of DNA damage (62). This finding places CHK2 and TTK in the same cellular compartment and provides a physical link for their functional interaction. It also raises the question regarding the role of CHK2 in mitosis and whether it is present in the centrosome purely as a "bystander" or whether it has additional roles in the regulation of mitotic events. On a related note, we consistently observed a marked increase in the levels of TTK when it was coexpressed with WT, but not KD, CHK2 (Fig. 1C). Although the significance of such an increase is unclear, it suggests the possible existence of a positive feedback loop between TTK and CHK2. It is tempting to speculate that a possible molecular interplay exists between the DNA damage and the spindle assembly checkpoint pathways during the progression of cell cycle. In fact, Mikhailov et al. (67) have demonstrated that DNA damage during mitosis in some human cells delays the metaphase/anaphase transition, notably in an ATMindependent manner, through the spindle assembly checkpoint. The DNA damage checkpoint has always been connected with the G 1 , S, or G 2 phase of the cell cycle. However, the fact that Plk1 activity is inhibited by DNA damage, thus preventing mitotic exit (64), also suggests that the DNA damage checkpoint is functional in mitotic cells. Whether TTK/hMps1 plays any role in such control is certainly intriguing, especially in light of the data presented in this report. Although the detailed mechanism of cross-talk is still unclear, our study provides a possible base for functional interaction between different checkpoints that together maintain the genome stability.