Activation of Cdh1‐dependent APC is required for G₁ cell cycle arrest and DNA damage‐induced G₂ checkpoint in vertebrate cells

Chicken Cdh1 cDNA was isolated by PCR using primers specific for human Cdh1 from the chicken cDNAs. Based on the sequences of chicken Cdh1 cDNA, ∼7 kb of the chicken Cdh1 locus was amplified by long‐range PCR using genomic DNA extracted from DT40 cells as a template. Either the histidinol (his) or the blasticidin (bsr) resistance gene was inserted between sequences of 4 and 3 kb length (Figure 1A). Targeted integration of these constructs disrupts the reading frame of the Cdh1 gene at the first WD repeat. Targeted events were examined by PCR, Southern blotting analysis and RT–PCR (Figure 1B, C and D). We isolated two viable Cdh1^−/− clones (clones 1 and 2), and the proliferative properties of Cdh1^−/− clones, in comparison with wild‐type cells, were monitored by using growth curves and cell cycle analysis. The growth curves of the Cdh1^−/− clones were indistinguishable from that of wild‐type cells, which divided approximately every 8 h (Figure 2A). Fluorescence‐activated cell sorting (FACS) analysis showed that the cell cycle distributions were the same in both genotypes (Figure 2B). These data indicate that the Cdh1 gene is dispensable for viability and proliferation of DT40 cells. Low stringency Southern blot analysis using a Cdh1 probe did not detect any significant signal besides the Cdh1 gene (data not shown), suggesting that there is no additional Cdh1‐related gene that may act redundantly, masking phenotypes of knockout cells.

In yeasts (budding and fission) and Drosophila, loss of hct1, ste9 and fzr, which are Cdh1 homologs, results in ectopic retention of mitotic cyclins in the G₁ phase (Schwab et al., 1997; Sigrist and Lehner, 1997; Yamaguchi et al., 1997; Kitamura et al., 1998). We therefore investigated the protein levels of chicken mitotic cyclins, cyclin A and cyclin B, in wild‐type and Cdh1^−/− DT40 cells at different time points after release from nocodazole‐induced metaphase block. FACS analysis showed that both genotypes were arrested in metaphase by nocodazole treatment and progress into G₁ from M phase after release (Figure 3A). Immunoblot analysis revealed that the protein levels of both mitotic cyclins were significantly reduced as the G₁ phase approached in wild‐type cells. In contrast, in Cdh1^−/− cells, the protein levels of mitotic cyclins failed to decline, resulting in ectopic accumulation of cyclin A and cyclin B in the G₁ phase (Figure 3B), which is consistent with previous findings observed in yeast and Drosophila. These data suggest that Cdh1 takes part in the degradation of mitotic cyclins from M to G₁ phase but is dispensable for exit from mitosis in higher vertebrates as well as in yeasts and Drosophila. It is noteworthy that mitotic cyclins are degraded to some extent in Cdh1^−/− cells (Figure 3B), and this Cdh1‐independent degradation of mitotic cyclins may be sufficient for cells to exit mitosis.

Although Cdh1 is dispensable for metaphase–anaphase transition and exit from mitosis, we examined whether loss of Cdh1 affects the duration of the mitotic phase. Wild‐type or Cdh1^−/− cells were synchronized at metaphase by treatment with nocodazole for 10 h. At different times after release from nocodazole block, the frequency of mitotic cells was determined by immunostaining of MPM‐2 antigen, which is generally thought to represent mitotic phosphoproteins (Davis et al., 1983). At 45 min after release from metaphase block, ∼40–45% of Cdh1^−/− cells remained in a mitotic state with positive MPM‐2 staining, compared with only 20% MPM‐2‐positive cells in the wild type (Figure 4A). These observations suggest that Cdh1‐deficient cells delay exit from mitosis. Furthermore, cytological examination of asynchronously growing cultures revealed that Cdh1^−/− cells showed a significant increase in the number of cells in the mitotic phases, particularly in anaphase and telophase (Figure 4B), suggesting that the late mitotic phase is prolonged in Cdh1^−/− cells. Taken together, Cdh1 is not essential for cell viability but plays a regulatory role in transition from late mitotic phase to G₁.

During embryogenesis in Drosophila, fzr, a homolog of Cdh1, was shown to be required for removal of mitotic cyclin during G₁ when proliferation of embryonic epidermal cell stops. In fzr‐deficient embryos, epidermal cells progress through an extra division cycle without stopping in G₁ after the terminal mitosis (mitosis 16). This defect is found to be caused by failure to degrade mitotic cyclins due to loss of fzr (Sigrist and Lehner, 1997). Since our Cdh1^−/− DT40 cells appeared to have an unscheduled accumulation of mitotic cyclins in the G₁ phase (Figure 3B), we attempted to investigate whether loss of Cdh1 function affects G₁ arrest in higher vertebrates, as well as in Drosophila. During Drosophila embryogenesis, p27^DAP, an inhibitor of cyclin E–cdk2, was shown to be expressed before mitosis 16 and to induce G₁ arrest after completion of cycle 16 in epidermal cells (Lane et al., 1996). Unscheduled accumulation of mitotic cyclin by loss of fzr, therefore, appears to overcome the p27^DAP‐induced G₁ arrest, resulting in progression through an extra division cycle (Du and Dyson, 1999). p27^Kip1, a vertebrate homolog of p27^DAP, inhibits the activity of cyclin E–cdk2, leading to a block of the G₁–S transition in vertebrate cells (Polyak et al., 1994a). Accordingly, we tested the possibility that Cdh1‐deficient cells abrogate G₁ arrest induced by the expression of p27^Kip1. (In the subsequent experiments, both Cdh1^−/− clones behaved identically. Therefore, the results obtained from clone 1 are reported below.) We employed rapamycin, which is known to activate p27^Kip1 expression (Nourse et al., 1994), to induce G₁ arrest in DT40 cells. The p27^Kip1 protein levels were up‐regulated in both wild‐type and Cdh1^−/− DT40 cells after 72 h of rapamycin treatment (Figure 5A). FACS analysis revealed that rapamycin treatment for 72 h induced G₁ arrest in wild‐type cells (Figure 5B and C). In contrast, Cdh1^−/− DT40 cells showed a significant decrease in the G₁ fraction at 72 h of rapamycin treatment, and a corresponding increase in the number of cells with S phase DNA content (Figure 5B and C). While pRb was not phosphorylated in wild‐type cells treated with rapamycin, Cdh1^−/− cells dominantly expressed the phosphorylated form (Figure 5A). In addition, Cdh1^−/− cells expressing the human Cdh1 transgene arrested in G₁ in the presence of rapamycin, similarly to wild‐type cells (Figure 5B). All these findings indicate that loss of Cdh1 results in abrogation of rapamycin‐induced G₁ arrest. Furthermore, Cdh1^−/− cells treated with olomoucine, which is a specific cdk2 kinase inhibitor, were induced to undergo G₁ arrest in the presence of rapamycin (Figure 5C). This result suggests that the abrogation of G₁ arrest in Cdh1‐deficient cells is caused by unscheduled activation of cdk2 kinase.

The expression of p27^Kip1 induces G₁ arrest by inhibiting the activity of the cyclin E–cdk2 complex, which is essential for initiation of DNA replication (Polyak et al., 1994b). It has also been shown that p27^Kip1 interacts with the cyclin A–cdk2 complex and inhibits its kinase activity (Blain et al., 1997). Therefore, it can be postulated that p27^Kip1 is sequestered by the cyclin A–cdk2 complex ectopically accumulated during the G₁ phase in Cdh1^−/− cells, resulting in activation of the cyclin E–cdk2 complex and leading to abrogation of the rapamycin‐induced G₁ arrest. To investigate this, cyclin E and cyclin A immunoprecipitated with p27^Kip1 were examined in wild‐type and Cdh1^−/− cells that were treated with rapamycin for 72 h. While a large amount of cyclin E was associated with p27^Kip1 in wild‐type cells, cyclin E was reduced in p27^Kip1 immunoprecipitates in Cdh1^−/− cells (Figure 5D). Conversely, association between cyclin A and p27^Kip1 was markedly increased in Cdh1^−/− cells. Furthermore, the cyclin E‐associated kinase activity was significantly higher in the lysate derived from Cdh1^−/− cells than in that from wild‐type cells (Figure 5D). These findings suggest that the cyclin E–cdk2 complex dissociated from p27^Kip1, which is sequestered by ectopically accumulated cyclin A–cdk2 complex, leads to abrogation of G₁ arrest in Cdh1^−/− cells.

In the normal cell cycle, cyclin A and cyclin B begin to accumulate from early S phase and G₂ phase, respectively. We tested whether loss of Cdh1 affected levels of mitotic cyclins in S and G₂ and cell cycle progression from S to mitotic entry. Wild‐type or Cdh1^−/− DT40 cells were synchronized in early S phase by an aphidicolin block. At different times after release from the S block, the protein levels of cyclin A and cyclin B and the mitotic index were determined. Immunoblot analyses showed that the expression of mitotic cyclins from S through G₂ and early M phases was comparable in wild‐type and Cdh1^−/− cells (data not shown). Furthermore, the rate of mitotic entry of Cdh1^−/− cells after release from S block was similar to that of wild‐type cells (Figure 6A). These data suggest that Cdh1 deficiency does not impair transition from S phase to mitotic entry in normal cell cycle progression.

Although our original goal in this work was to understand the involvement of Cdh1 in normal cell cycle progression, our results identify a new role for Cdh1 in cell cycle checkpoints responsive to damaged DNA. Cells released from early S phase block were exposed to 10 Gy of X radiation at mid‐S phase, and the cell cycle progression of wild‐type and Cdh1^−/− cells after irradiation was assessed by counting the rate of mitotic entry and by FACS analysis of DNA. Morphological examination revealed that both genotypes entered mitosis within 4 h after release from S block without irradiation (Figure 6A). Following treatment with 10 Gy of X radiation, the vast majority of wild‐type cells appeared to arrest in the G₂ phase until 6 h after irradiation, suggesting that wild‐type cells have the damaged DNA‐induced G₂ checkpoint function (Figure 6B, upper panel). However, substantially more Cdh1^−/− cells entered mitosis relative to wild‐type cells at 5 h after irradiation, indicative of a defect in the G₂ checkpoint. Moreover, in FACS analysis, Cdh1^−/− cells showed a relative decrease in the G₂ fraction as early as 8 h after irradiation and a corresponding increase in the number of cells having sub‐G₁ DNA content, which is characteristic of dead cells (Figure 6C). These findings suggest that the massive cell death of Cdh1‐deficient cells after irradiation was due to premature entry into mitosis with DNA damage and, thereby, that Cdh1 plays a crucial role in the damaged DNA‐induced G₂ checkpoint.

DNA strand damage caused by UV‐irradiation also activates the G₂ checkpoint, but molecular mechanisms involved in regulating G₂ delay after UV‐irradiation are different from those after X‐irradiation (Bulavin et al., 2001). In contrast to X‐irradiation, UV‐irradiation induced a G₂ delay in both genotypes (Figure 6B, lower panel), suggesting that Cdh1 deficiency has no effect on the UV‐induced G₂ checkpoint.

The activity of APC is largely dependent on its interaction with Cdc20 or Cdh1 (Kramer et al., 2000). The phosphorylation of Cdh1 is known to prevent binding to and activation of the APC. Hct1, a budding yeast homolog of Cdh1, was shown to be inactivated through phosphorylation by maturation‐promoting factor (MPF) and activated through dephosphorylation by Cdc14 from late mitosis to the end of the G₁ phase (Visintin et al., 1998; Jaspersen et al., 1999). Moreover, it has been suggested that interaction of Cdh1 and APC (active Cdh1–APC) is maintained until late in the G₁ phase and diminished by the phosphorylation of Cdh1 with cdk2 from S phase in mammalian cells (Lukas et al., 1999; Kramer et al., 2000). These previous observations indicate that Cdh1–APC activity in normal cells is down‐regulated by the dissociation of Cdh1 from the APC during S phase through to mitotic entry in the normal cell cycle. Based on our findings obtained using Cdh1^−/− DT40 cells, we investigated whether APC is activated in the G₂ phase through interaction with Cdh1 in response to DNA damage. Cdh1^−/− cells expressing the human Cdh1 transgene were synchronized in early S phase by treatment with aphidicolin. The cells released from early S phase block were exposed to X‐irradiation, and interaction between Cdh1 and APC was investigated. In the cells not treated with X‐irradiation, Cdh1 had little association with Cdc27, a component of the APC complex, during G₂ phase (Figure 7A), which is consistent with findings reported previously (Kramer et al., 2000). In contrast, when cells were exposed to X‐irradiation at 1 h after release from S block, Cdh1 was induced to interact with Cdc27 in the G₂ phase (Figure 7A). Interaction between Cdh1 and the APC complex induced by DNA damage was observed not only in DT40 cells but also in both rodent fibroblast (Rat1 cell) and a human cancer cell line (HeLa cell) (Figure 7B).

Having determined the complex formed by APC and Cdh1 after DNA damage, we asked whether the Cdh1–APC complex becomes catalytically active in response to DNA damage. First, we performed an in vitro ubiquitylation assay using human Cdc20 as substrate. Human Cdc20 contains a KEN box, which was shown to be the Cdh1‐specific targeting motif (Pfleger and Kirschner, 2000) and, thereby, is considered to be a specific substrate for Cdh1–APC. APC immunoprecipitated with anti‐Cdc27 antibody from cells in normal G₂ phase showed no activity to ubiquitylate human Cdc20 (Figure 7C, lanes 5–8), which is consistent with results reported previously (Kramer et al., 2000). In contrast, APC purified from cells arrested in the G₂ phase after X‐irradiation ubiquitylated human Cdc20 in a time‐dependent manner (Figure 7C, lane 1–4), suggesting that Cdh1–APC becomes catalytically active by DNA damage induced by X‐irradiation. Furthermore, we attempted to investigate the in vivo ubiquitylation of the Cdh1–APC complex after X‐irradiation in the G₂ phase. HeLa cells stably expressing human ubiquitin were established and used to examine the ubiquitylation of the Cdh1‐associated complex. Ubiquitylation of the complex immunoprecipitaed with hemagglutinin (HA)‐tagged Cdh1 was very low in cells in the G₂ phase without DNA damage (Figure 7D, left lane). However, Cdh1‐associated complex in X‐irradiated cells in the G₂ phase showed significantly increased ubiquitylation (Figure 7D, right lane). All these findings suggest that Cdh1–APC is activated at the X‐irradiation‐induced G₂ checkpoint.

Our findings indicate the possibility that activation of Cdh1–APC by DNA damage may induce degradation of regulators for G₂–M transition, resulting in G₂ arrest. In HeLa and other cells, it is known that the expression of cyclin B is down‐regulated in S and G₂ phase in response to DNA damage (Muschel et al., 1993). To investigate whether abrogation of the damaged DNA‐induced G₂ checkpoint in Cdh1^−/− cells is due to an inability to degrade mitotic cyclins, the protein levels of mitotic cyclins after X‐irradiation were examined. Immunoblot analyses showed that the levels of mitotic cyclins did not change significantly but remained similarly stable in both wild‐type and Cdh1^−/− cells (Figure 7E). These data suggest that the degradation of mitotic cyclins mediated by the Cdh1–APC complex plays little role in the G₂ checkpoint after DNA damage.

In the present study, we have generated Cdh1‐deficient cells using DT40 chicken B lymphocytes. Cdh1^−/− DT40 cells appeared to be exiting mitosis and proliferating normally despite the fact that expression of cyclin A and cyclin B in these cells in M and G₁ phases was significantly elevated. These findings suggest that Cdh1 plays a major role in proteolysis of mitotic cyclins from the M to G₁ phases but is dispensable for viability and proliferation of higher vertebrate cells, which is similar to observations in yeast and Drosophila (Schwab et al., 1997; Sigrist and Lehner, 1997). In higher vertebrate cells, cyclin B destruction has been believed to be required for exit from mitosis. However, recent observation has clearly demonstrated that mammalian cells overexpressing cyclin B1 are capable of exiting mitosis (Jin et al., 1998), which appears consistent with our result.

Our results show that mitotic cyclins accumulate significantly but are partially degraded at the end of mitosis in Cdh1^−/− cells, suggesting that not only Cdh1 but also another molecule, such as Cdc20, is involved in the proteolysis of mitotic cyclins in the mitotic phase. In budding yeast, Hct1/Cdh1 and Cdc20 have been shown to target some common molecules, and the main difference between them is the time during the cell cycle when each is active (Yeong et al., 2000). Furthermore, Cdc20–APC can only be activated under conditions in which Hct1/Cdh1–APC is inactive, and this dependency helps to ensure that the events of mitosis occur in the proper sequence (Lim et al., 1998; Visintin et al., 1998; Shirayama et al., 1999; Yeong et al., 2000). Therefore, collectively, these previous data suggest that Cdc20 and Hct1/Cdh1 target distinct, but overlapping sets of proteins for degradation in crucial mitotic events, such as separation of sister chromatids and exit from mitosis. It was found recently that in early Xenopus embryos, Cdh1 is not expressed (Lorca et al., 1998), indicating that Cdh1 is not required for cyclin oscillations in early embryonic cell cycles. Similarly, Cdh1 is not expressed in early fly embryos, where Cdc20 is required for cyclin destruction (Sigrist and Lehner, 1997). Additionally, vertebrate Cdc20–APC and Cdh1–APC both ubiquitylate mitotic cyclins in vitro (Fang et al., 1998). Thus, it is reasonable to speculate that Cdc20 promotes not only the initiation of anaphase but also the partial destruction of mitotic cyclins, which is sufficient for exit from mitosis, in Cdh1^−/− DT40 cells. Our observations suggest that proteolysis of mitotic cyclins is regulated by two steps, mediated by Cdc20 and Cdh1, in higher vertebrate cells.

Although Cdh1 is dispensable for viability of DT40 cells, the length of time between anaphase and the end of mitosis was found to be prolonged in Cdh1^−/− DT40 cells. This result suggests that Cdh1–APC regulates the progression of late mitosis despite the fact that it is not essential for exiting mitosis.

In cultured human cells, ectopic overexpression of cyclin A can drive premature entry into S phase and causes a decrease in the percentage of cells in G₁ (Resnitzky et al., 1995). These results demonstrate that the expression of cyclin A is rate limiting for the G₁–S transition. However, our observations showed that Cdh1^−/− DT40 cells have a normal G₁ phase and do not enter S phase prematurely despite the high expression of both cyclin A and cyclin B. These findings suggest that the accumulation of mitotic cyclins due to loss of Cdh1 does not affect the duration of the G₁ phase in normal exponential conditions. This discrepancy may be explained by a difference in the expression levels of cyclin A. The level of cyclin A induced by the exogenous expression system in the study of Resnitzky et al. (1995) appears to be much higher than physiological levels. Thus, it is possible that the levels of mitotic cyclins accumulated at G₁ phase in Cdh1‐deficient cells are not enough to induce premature entry into S phase.

In fzr‐deficient Drosophila embryos, epidermal cells progress through an extra division cycle without stopping in G₁ after terminal mitosis. This defect is believed to be caused by failure to decrease mitotic cyclins (Sigrist and Lehner, 1997). Furthermore, srw1‐mutant fission yeasts also do not arrest at G₁ under conditions of nitrogen starvation (Yamaguchi et al., 1997). Similarly, we found that Cdh1^−/− DT40 cells abrogate the G₁ arrest induced by rapamycin, which is an activator of p27^Kip1, following entry into S phase. p27^Kip1 is known to induce G₁ arrest by inhibiting the activity of the cyclin E–cdk2 complex. p27^Kip1 has also been shown to interact with cyclin A–cdk2. Since the induction of p27^Kip1 expression by rapamycin was similar in both wild‐type and Cdh1^−/− DT40 cells (Figure 5A), one mechanistic explanation as to how inactivation of Cdh1 can cause an abrogation of G₁ cell cycle arrest is by sequestration of p27^Kip1 by accumulated cyclin A–cdk2 complex to activate cyclin E–cdk2. Four lines of evidence presented here support this hypothesis. First, inhibition of cdk2 kinase activity by treatment with olomoucine led Cdh1^−/− cells to arrest at the G₁ phase in the presence of rapamycin (Figure 5C). Secondly, the amounts of cyclin E associated with p27^Kip1 were decreased in Cdh1^−/− cells. Thirdly, p27^Kip1 interacted predominantly with ectopically accumulated cyclin A in Cdh1^−/− cells. Finally, the level of cyclin E‐associated kinase activity was dramatically elevated in Cdh1^−/− cells (Figure 5D). In fission yeast, cdk1 activated by B‐type cyclin can promote G₁–S progression in the absence of G₁ cyclin (Fisher and Nurse, 1996). Therefore, the possibility cannot be ruled out that cdk1 activated with ectopically accumulated cyclin B may promote G₁–S transition in rapamycin‐treated Cdh1^−/− cells. It is also possible that high levels of cyclin A promote transition of Cdh1^−/− cells from G₁ to S phase. In any case, our findings suggest that activation of Cdh1 plays a crucial role in establishment of G₁ arrest by cooperating with inhibitors of G₁ cyclin‐dependent kinases.

Recent reports have shown that Cdh1–APC is inactivated from S phase until the mid‐mitotic phase through phosphorylation by cyclin A–cdk2 or cyclin B–cdk1 in mammalian cells (Lukas et al., 1999). These data indicate that Cdh1–APC is not required for progression of the cell cycle from S phase to mitotic entry, which is consistent with our finding that Cdh1 deficiency does not affect the cell cycle progression from S phase to the mitotic phase in physiological conditions. However, we found that Cdh1^−/− cells failed to maintain X‐irradiation‐induced arrest in the G₂ phase of the cell cycle. This finding strongly suggests that the activation of Cdh1–APC plays a role in DNA damage‐induced G₂ checkpoint function. In fact, we demonstrate here that X‐irradiation induces interaction of Cdh1 with APC in the G₂ phase (Figure 7A and B) and consequently activates the Cdh1–APC (Figure 7C and D). These findings indicate that Cdh1–APC is activated at G₂ by X‐irradiation‐induced DNA damage.

How can Cdh1–APC induce G₂ arrest in response to DNA damage? One obvious possibility is that Cdh1–APC targets regulators for G₂–M transition. However, our findings showed that the levels of mitotic cyclins in Cdh1^−/− cells after DNA damage were similar to those in wild‐type cells (Figure 7E). These data imply that cyclin A and cyclin B cannot be substrates for Cdh1–APC when it is activated irregularly by DNA damage at G₂. Recent reports have demonstrated that Cdc20 strictly recognizes a substrate with a well‐defined D‐box, whereas Cdh1–APC recognizes both D‐box and non‐D‐box substrates (Pfleger and Kirschner, 2000) for destruction. Therefore, the regulation of the interaction between Cdh1–APC and those diverse substrates may be dependent on the cell cycle phases and events. It seems likely that critical molecules for promoting the G₂–M transition, besides mitotic cyclins, are potential substrates of activated Cdh1–APC in the DNA damage checkpoint and that the degradation of those substrates by the activated Cdh1–APC induces G₂ arrest.

We demonstrate that UV radiation induced G₂ delay not only in wild‐type cells but also in Cdh1^−/− cells (Figure 6B), suggesting that Cdh1 deficiency has no effect on the UV‐induced G₂ checkpoint. Recently, it has been reported that molecular mechanisms regulating UV‐induced G₂ delay are distinct from those induced by X radiation (Bulavin et al., 2001). Therefore, this result shows that Cdh1 is involved in the X‐irradiation‐induced G₂ checkpoint pathway but not in the UV‐induced G₂ checkpoint pathway.

The maintenance of genomic integrity after DNA damage depends on the coordinated action of the DNA repair and checkpoint systems. The failure of a checkpoint leads to genomic instability and predisposition to cancer (Lengauer et al., 1998). Since our findings suggest that activation of Cdh1–APC is required for both cdk inhibitor‐dependent G₁ arrest and DNA damage‐induced G₂ arrest, it is conceivable that abnormal regulation of Cdh1 activation may involve failure in normal cell cycle progression, leading to genetic and chromosomal instabilities, which result in development of malignant tumors. Investigation of this possibility is currently underway using various cancer cells.

Our findings also have potential relevance for the treatment of cancer. Abrogation of the DNA damage‐induced G₂ checkpoint sensitizes cells to genotoxic stress. This suggests that a specific inhibitor of DNA damage‐induced Cdh1 activation could make radiotherapy and/or chemotherapy more effective.

DT40 cells were cultured in RPMI1640 medium supplemented with 10⁻⁵ M β‐mercaptoethanol, penicillin/streptomycin, 10% fetal calf serum and 1% chicken serum (Sigma) at 37°C. Rat1 and HeLa cells were cultured in Dulbecco's modified Eagle's medium (DMEM)‐F12 supplemented with penicillin/streptomycin and 10% fetal calf serum.

Chicken Cdh1 cDNA was amplified by RT–PCR using mRNA extracted from DT40 cells with primers for human Cdh1. Using this cDNA, ∼9 kb of the chicken genomic cdh1 locus was isolated from DT40 genomic DNA using long‐range PCR. Chicken Cdh1 disruption constructs, Cdh1‐Bsr (blasticidin) and Cdh1‐His (histidinol), were made by replacing ∼1 kb of the genomic sequence of Cdh1 with Bsr and His selection marker cassettes (Figure 1A).

For gene targeting of the chicken Cdh1 locus, DT40 cells (10⁷) were suspended in 0.5 ml of phosphate‐buffered saline (PBS) containing 30 μg of linearized targeting vector for each transfection and electroporated with a Gene Pulser apparatus (BTX) at 210 V and 475 μF. Following electroporation, cells were transferred into 20 ml of fresh medium and incubated for 24 h. Cells were then resuspended in 80 ml of medium containing 30 μg/ml blasticidin‐S (Calbiochem) or 1 mg/ml histidinol (Sigma) and divided into four 96‐well plates. After 7–10 days, drug‐resistant colonies were transferred to 24‐well plates. Genomic DNA was extracted from each expanded clone, and successful targeted recombination was identified by Southern blot and PCR analyses. To generate Cdh1^−/− cells expressing the human Cdh1 transgene, we transfected cells with the pUHG plasmid containing human Cdh1 and selected stable transformants with 2.5 mg/ml hygromycin (Calbiochem).

To evaluate the targeted allele, genomic DNA from DT40 cells was amplified using the three sets of primers (p2–p4, p1–p3 and p1–p5) illustrated in Figure 1(A). Genomic DNA was prepared using InstaGene Matrix (Bio‐Rad). mRNA was isolated using a Micro‐FastTrack 2.0 kit (Invitrogen) and reverse‐transcribed using the Superscript cDNA synthesis system as described by the suppliers (Gibco‐BRL). Full‐length chicken Cdh1 cDNA was amplified by PCR with 35 cycles of 94°C (1 min), 60°C (1 min) and 72°C (1 min). Primers for the housekeeping gene glyceraldehyde‐3‐phosphate dehydrogenase (GAPDH) were used along with the primers for the amplification of Cdh1 cDNA.

Proteins were detected by anti‐chicken cyclin A antibody (Transduction Laboratories), anti‐chicken cyclin B2 antibody (Gallant and Nigg, 1992), anti‐cyclin E antibody (Santa Cruz), anti‐Rb antibody (Transduction Laboratories), anti‐ubiquitin antibody (Sigma), anti‐chicken Ku70 antibody (Takata et al., 1998) and anti‐Cdh1 antibody (Wang et al., 2000).

To analyze the intracellular DNA content, cells were fixed in 70% methanol at −20°C for several hours. Cells were centrifuged at 2000 r.p.m. and resuspended in PBS containing RNase A (Sigma) at 0.1 mg/ml. Samples were incubated at 37°C for 15 min, and propidium iodide (Sigma) was added to a final concentration of 25 μg/ml. Samples were processed by a FACScan (Becton Dickinson) using Cell Quest software (Becton Dickinson). For determination of the bromodeoxyuridine (BrdU) incorporation, cells were labeled for 10 min with 20 μM BrdU (Amersham). They were harvested and fixed in 70% methanol at −20°C for several hours. Then, cells were stained with an anti‐BrdU antibody as described previously (Sonoda et al., 1998).

In DT40 cells, to synchronize cells at the G₁, early S phase or metaphase, cells were cultured in the presence of 100 nM rapamycin (Calbiochem) for 72 h, 1 μg/ml aphidicolin (Wako) for 10 h or 0.5 μg/ml nocodazole (Sigma) for 10 h, respectively. Rat1 cells were synchronized at early S phase by 1 μg/ml aphidicolin treatment for 12 h. HeLa cells were synchronized at the beginning of S phase by means of a double thymidine/aphidicolin block. Briefly, HeLa cells were incubated with 2.0 mM thymidine (Sigma) for 24 h, released for 8 h and then treated with 1 μg/ml aphidicolin for a further 16 h. Cells were exposed to 10 Gy of X‐rays at 1 h (DT40 and Rat1) or 3 h (HeLa) after release from S block.

Cells with mitotic condensed chromatin were visualized by staining with aceto‐orcein (Merck) in 60% acetic acid (Hirota et al., 2000). Cells at anaphase and telophase are easily identified by aceto‐orcein staining. To avoid the morphological confusion between mitotic cells and apoptotic cells in the X‐ and UV‐irradiation experiments, mitotic indices were determined by counting cells with mitotic spindles. To detect the mitotic spindles, DT40 cells were fixed with 100% methanol for 10 min, washed with PBS twice and incubated in a blocking buffer (1% bovine serum albumin in PBS) for 1 h at room temperature. The cells were incubated with the monoclonal antibody against α‐tubulin (Sigma) for 20 h at 4°C, washed with PBS and incubated with fluorescein isothiocyanate (FITC)‐conjugated anti‐mouse IgG (Biosource) for 4 h at room temperature.

The proportion of cells in the mitotic phase was also determined by FACS analysis using anti‐MPM‐2 antibody (Upstate Biotechnology). Briefly, cells were harvested and fixed in 70% methanol at −20°C for several hours. Fixed cells were stained for 1 h at room temperature with anti‐MPM‐2 antibody, washed three times in PBS and stained with FITC‐conjugated anti‐mouse IgG (Biosource). Stained cells were analyzed with FACScan using Cell Quest software.

Cells were washed with PBS and used immediately, or were frozen in liquid nitrogen and stored at −80°C. The cells were then lysed in 0.5% Triton X‐100 lysis buffer (20 mM Tris–HCl pH 7.7, 100 mM KCl, 50 mM sucrose, 0.1 mM CaCl₂, 1 mM MgCl₂, 0.5% Triton X‐100, 20 mM β‐glycerophosphate, 1 mM Na₃VO₄ and a cocktail of protease inhibitors). The extracts were centrifuged at 14 000 r.p.m. for 15 min at 4°C, and the supernatants were used for immunoprecipitation. Aliquots of supernatants were incubated for 2 h at 4°C with a monoclonal anti‐p27^Kip1 antibody (Transduction Laboratories), anti‐Cdc27 antibody (Santa Cruz) and anti‐HA antibody (Boehringer Mannheim), respectively. Cells were incubated for another 2 h after addition of protein G plus protein A–agarose beads (Oncogene Science). The washed immunoprecipitants and the supernatant were collected for western blotting.

Cells were washed with PBS twice and suspended in 120 μl of lysis buffer (20 mM Tris–HCl pH 7.5, 10 mM EDTA, 100 mM NaCl, 1% Triton X‐100, 1 mM NaF, 1 mM β‐glycerophosphate and a cocktail of protease inhibitors). A 100 μl aliquot of lysates and 5 μl of an anti‐cyclin E antibody were mixed and incubated for 1 h at 4°C. Then, a 20 μl aliquot of protein G plus protein A–agarose beads was added. After gentle agitation for 30 min at 4°C, beads were washed four times in lysis buffer, the second wash containing 1 M NaCl, and twice in kinase buffer [25 mM MOPS pH 7.2, 15 mM MgCl₂, 5 mM EDTA, 1 mM dithiothreitol (DTT), 60 mM β‐glycerophosphate, 0.1 mM sodium orthovanadate and a cocktail of protease inhibitors]. Pelleted beads were mixed with an equal volume of a reaction mix containing Triton X‐100 (1%), [γ‐³²P]ATP and histone H1 (1 mg/ml) and incubated for 15 min at 30°C. The reaction was chilled and stopped by addition of a sample buffer. The samples were loaded on an SDS–polyacrylamide gel and visualized by autoradiography.

The N‐terminal half of human Cdc20 cDNA encoding amino acid residues 1–250 was inserted into pET23d and expressed in NovaBlue (DE3) (Novagene). The His‐tagged recombinant Cdc20 protein was purified and biotinylated (Pierce). The in vitro ubiquitination assay was carried out as described previously (Kotani et al., 1998, 1999). In brief, synchronized HeLa cells were treated with or without X‐irradiation at the G₂ phase, and the cell extracts harvested 8 h after the release were immunoprecipitated with anti‐Cdc27 antibody. Immunoprecipitants were incubated with 2 μg of His‐Cdc20 in 30 μl of 5 mM Tris pH 7.6, 0.5 mM MgCl₂, 2 mM ATP, 2 mM DTT, 2 mM creatinine phosphate, 1 μg/ml creatine, phosphokinase, 0.2 mg/ml bovine ubiquitin (Sigma), 40 μg/ml mouse recombinant E1 and 50 μg/ml human recombinant hE2‐C. After the reaction, His‐Cdc20 was purified using ProBond resin, subjected to SDS–PAGE and transferred onto a nitrocellulose membrane followed by detection with streptavidin–horseradish peroxidase.

HeLa cells were transfected with myc‐tagged ubiquitin expression plasmid (pcDNA3/myc‐Ub), and neomycin‐resistant clones were isolated and cultured in the presence of 1 mg/ml geneticin (Gibco‐BRL). To synchronize the myc‐ubiquitin‐expressing HeLa cells at early S phase, asynchronously growing cells were treated with 2 mM thymidine for 16 h. Cells were then released from the block by changing to complete growth medium containing 24 mM each of thymidine and deoxycytidine. At 6 h after the release, cells were transfected with Cdh1/pCGN‐HA using FuGENE 6 (Roche) reagent. After 2 h, thymidine was added to the medium to a final concentration of 2 mM and cells were cultured for another 16 h. Before harvest at 8 h after the release, 10 mM MG132 (Calbiochem) was added to the medium and incubated for 1 h. An immunoprecipitation analysis with anti‐HA antibody was performed as described above. The samples were subjected to SDS–PAGE followed by immunoblot analysis using anti‐ubiquitin antibody (Sigma).