It is possible to cause G2 arrest in Aspergillus nidulansby inactivating either p34cdc2 or NIMA. We therefore investigated the negative control of these two mitosis‐promoting kinases after DNA damage. DNA damage caused rapid Tyr15 phosphorylation of p34cdc2 and transient cell cycle arrest but had little effect on the activity of NIMA. Dividing cells deficient in Tyr15 phosphorylation of p34cdc2 were sensitive to both MMS and UV irradiation and entered lethal premature mitosis with damaged DNA. However, non‐dividing quiescent conidiospores of the Tyr15 mutant strain were not sensitive to DNA damage. The UV and MMS sensitivity of cells unable to tyrosine phosphorylate p34cdc2 is therefore caused by defects in DNA damage checkpoint regulation over mitosis. Both the nimA5 and nimT23 temperature‐sensitive mutations cause an arrest in G2 at 42°C. Addition of MMS to nimT23 G2‐arrested cells caused a marked delay in their entry into mitosis upon downshift to 32°C and this delay was correlated with a long delay in the dephosphorylation and activation of p34cdc2. Addition of MMS to nimA5 G2‐arrested cells caused inactivation of the H1 kinase activity of p34cdc2 due to an increase in its Tyr15 phosphorylation level and delayed entry into mitosis upon return to 32°C. However, if Tyr15 phosphorylation of p34cdc2 was prevented then its H1 kinase activity was not inactivated upon MMS addition to nimA5 G2‐arrested cells and they rapidly progressed into a lethal mitosis upon release to 32°C. Thus, Tyr15 phosphorylation of p34cdc2 in G2 arrests initiation of mitosis after DNA damage in A.nidulans.
DNA damage causes a G2 delay of the cell cycle in eukaryotic cells (Hartwell and Weinert, 1989; Carr; 1995; Kaufmann, 1995; Murray, 1995). This G2 delay presumably allows cells enough time to repair damaged DNA before initiation of mitosis. The dependency of initiation of mitosis on completion of DNA damage repair is established by G2/M DNA damage checkpoints. The first direct genetic link between the G2 checkpoint regulation and DNA damage was demonstrated with rad9 mutations in budding yeast. Strains carrying rad9 mutations are deficient in G2 delay after DNA damage, and thus enter lethal premature mitosis in the presence of damaged DNA (Weinert and Hartwell, 1988). Many genes involved in the G2/M DNA damage checkpoint regulation were subsequently identified in both budding and fission yeasts by screening for mutations which uncouple initiation of mitosis from G2 delay after DNA damage (Al‐Khodairy and Carr, 1992; Allen et al., 1994; Weinert et al., 1994). Many of these checkpoint genes have now been cloned. However, the cell cycle targets for the checkpoint regulation to bring about G2 delay in response to DNA damage remain to be established (Carr, 1995; Murray, 1995; Lydall and Weinert, 1996).
The activation of the universally conserved p34cdc2 H1 kinase, the mitosis‐promoting factor (MPF), is central to the timing and initiation of mitosis in all eukaryotic cells (Draetta, 1990; Nurse, 1990; Murray, 1992; Dunphy, 1994; Osmani et al., 1994). The activation of p34cdc2 requires association with the regulatory subunit cyclinB which accumulates during late S‐phase and G2 (Evans et al., 1983; Booher et al., 1989), and the phosphorylation of Thr161 by CDC2‐activating kinase (CAK) (Fesquet et al., 1993; Poon et al., 1993; Solomon et al., 1993). During S and G2 the p34cdc2–cyclinB complex is kept in the inactive pre‐MPF state through Tyr15 phosphorylation of p34cdc2 by Wee1/Mik1/Myt1 tyrosine kinases (Russell and Nurse, 1987; Gould and Nurse, 1989; Lungren et al., 1991; Parker and Piwnica‐Worms, 1992; Mueller et al., 1995). During the G2/M transition the pre‐MPF form of p34cdc2 is rapidly converted into the active MPF form by Tyr15 dephosphorylation catalyzed by the Cdc25 tyrosine phosphatase (Russell and Nurse, 1986; Kumagai and Dunphy, 1991). The major regulatory event for the final activation of p34cdc2 during mitotic initiation thus appears to be the rapid dephosphorylation at Tyr15. Therefore, Tyr15 phosphorylation of p34cdc2 could be a target for negative regulation by the G2/M DNA damage checkpoint control systems.
Elevated levels of Tyr15‐phosphorylated p34cdc2 are found to be associated with G2‐arrested cells after DNA damage in several systems (Kharbanda et al., 1994; O‘Conor et al., 1994; Herzinger et al., 1995; Barth et al., 1996). However, the significance of Tyr15 phosphorylation of p34cdc2 in the G2/M DNA damage checkpoint regulation remains to be established in these systems. The regulatory pathway for p34cdc2 tyrosine phosphorylation/dephosphorylation and its role in cell cycle regulation are best characterized in fission yeast (Russell and Nurse, 1986, 1987; Gould and Nurse, 1989; Lungren et al., 1991). Fission yeast cells unable to tyrosine phosphorylate p34cdc2 advance mitosis and produce small ’wee' cells (Gould and Nurse, 1989; Lungren et al., 1991). These mutant cells enter mitosis even when DNA replication is inhibited (Enoch and Nurse, 1990; Lungren et al., 1991). Thus, Tyr15 phosphorylation of p34cdc2 links initiation of mitosis to completion of DNA replication in fission yeast. However, the role of such phosphorylation in the G2/M DNA damage checkpoint regulation is not well understood. Fission yeast cells deficient in Wee1 function are sensitive to DNA damage by UV and gamma irradiation (Al‐Khodairy and Carr, 1992; Rowley et al., 1992). In addition, Rowley et al. (1992) reported that the DNA damage checkpoint in fission yeast is mediated through the Wee1 protein kinase as fission yeast cells deficient in Wee1 function lack a significant mitotic delay after gamma irradiation. However, Barbet and Carr (1993) subsequently showed that fission yeast cells defective or lacking in Wee1 function have a normal mitotic arrest after DNA damage. The Tyr15 phosphorylation state and H1 kinase activity of p34cdc2 in response to DNA damage, however, were not analyzed biochemically in these studies. Thus, the role of Tyr15 phosphorylation of p34cdc2 in the G2/M DNA damage checkpoint regulation remains to be firmly established in fission yeast through a combination of both biochemical and genetic studies. However, such a role in response to DNA damage has recently been established in human cells (Jin et al., 1996). On the other hand, both DNA replication and DNA damage checkpoints are fully operative in budding yeast cells unable to tyrosine‐phosphorylate p34cdc28 (Amon et al., 1992; Sorger and Murray, 1992; Stueland et al., 1993), conclusively demonstrating that in this system S‐phase and G2/M checkpoint regulations inhibit mitosis by some other mechanism.
In Aspergillus nidulans, coordinate activation of two mitosis‐promoting kinases is required for the initiation of mitosis (Osmani et al., 1991a; Ye et al., 1995). One is the universally conserved p34cdc2 H1 kinase encoded by the nimX gene (Osmani et al., 1994). The regulatory pathway for p34cdc2 activation is also conserved in A.nidulans (Osmani et al., 1991a, 1994; O‘Connell et al., 1992; P.Ramos, X.Ye, R.Fincher, S.Osmani and L.Ellis, in preparation). The other essential mitosis‐promoting kinase is the NIMA kinase which is currently well characterized only in A.nidulans (for review, see Osmani and Ye, 1996). However, indirect evidence indicates that a NIMA‐like mitotic pathway may also exist in fission yeast, Xenopus and human cells (O'Connell et al., 1994; Lu and Hunter, 1995). Unlike most mitotic regulatory functions which regulate p34cdc2 H1 kinase activity, NIMA kinase is required for initiation of mitosis independently of the regulation of p34cdc2 H1 kinase activity (Osmani et al., 1991a).
We recently uncovered two overlapping S‐phase checkpoint mechanisms over the initiation of mitosis in A.nidulans (Ye et al., 1996). The first, the slowing of S‐phase checkpoint, monitors the rate of DNA replication and, when DNA replication is slowed, prevents mitosis temporarily through Tyr15 phosphorylation of p34cdc2. Thus, when DNA replication is slowed, cells deficient in tyrosine phosphorylation of p34cdc2 enter mitosis prematurely. The second, the S‐phase arrest checkpoint, requires both BIME function and Tyr15 phosphorylation of p34cdc2 which in combination inactivate NIMA kinase and cause inhibition of mitosis when DNA replication is actually halted. Thus, cells lacking either Tyr15 phosphorylation of p34cdc2 (Ye et al., 1996) or the function of BIME (Osmani et al., 1988, 1991b; James et al., 1995) have a limited capacity to initiate mitosis when DNA replication is stopped. Only cells lacking both Tyr15 phosphorylation of p34cdc2 and the function of BIME are able to activate NIMA precociously and initiate mitosis effectively from S‐phase in the absence of DNA replication (Ye et al., 1996).
In this study we examined the relationship between Tyr15 phosphorylation of p34cdc2 and G2 delay caused by DNA damage in A.nidulans. We found that A.nidulans cells unable to Tyr15‐phosphorylate p34cdc2 were deficient in the DNA damage‐mediated G2 delay and entered mitosis prematurely after DNA damage. Consequently, only dividing cells deficient in Tyr15 phosphorylation of p34cdc2 were sensitive to DNA damage. Thus, Tyr15 phosphorylation of p34cdc2 in G2 alone links initiation of mitosis to completion of DNA damage repair in A.nidulans.
MMS increases the level of Tyr15‐phosphorylated p34cdc2
Methyl methanesulfonate (MMS) causes effective DNA damage in A.nidulans (Kafer and Mayor, 1986), the addition of 0.01% MMS to germinating wild‐type conidia delaying the first mitosis for more than 4 h. The first mitosis of germinating conidiospores normally occurs between 5.5 and 6.5 h after germination at 32°C, whereas the first mitosis of MMS‐treated germinating conidiospores did not occur until 9–10 h (data not shown, see also Figure 3). This suggests that a DNA damage checkpoint mechanism operates in A.nidulans, the activation of which causes a delay of entry into mitosis.
To investigate if either (or both) of the mitosis‐producing kinases, p34cdc2 or NIMA, are targets of the DNA damage checkpoint mechanism, we first determined the biochemical effects of addition of MMS to an exponentially growing wild‐type culture. Addition of MMS only slightly inhibited NIMA kinase activity (Figure 1). We previously demonstrated that NIMA is a target for the negative regulation by S‐phase checkpoints via both Tyr15 phosphorylation of p34cdc2 and the function of BIME (Ye et al., 1996). Activation of S‐phase checkpoints by the DNA synthesis inhibitor hydroxyurea (HU) rapidly inactivates the NIMA kinase (Ye et al., 1996). Here, lack of strong inhibition of NIMA by DNA damage suggests that NIMA is not a primary target for the DNA damage checkpoint system.
In contrast, addition of MMS caused rapid Tyr15 phosphorylation of p34cdc2 (Figure 1), the level of phosphorylation increasing markedly 30 min after MMS addition. Even though the level of NIMEcyclinB increased, p34cdc2 H1 kinase activity was reduced after MMS addition (Figure 1). Thus, the DNA damage checkpoint could delay mitosis by inhibition of p34cdc2 through Tyr15 phosphorylation.
Dividing cells deficient in Tyr15 phosphorylation of p34cdc2 are sensitive to UV irradiation and MMS because they enter mitosis prematurely with damaged DNA
If the DNA damage checkpoint is indeed mediated through Tyr15 phosphorylation of p34cdc2, dividing cells unable to undergo such phosphorylation should be sensitive to DNA‐damaging agents, as cells deficient in DNA damage checkpoint regulation would be unable to restrain cell cycle progression and would enter lethal premature mitosis in the presence of damaged DNA. Non‐dividing cells deficient in p34cdc2 Tyr15 phosphorylation, on the other hand, should not be sensitive to DNA‐damaging agents as they would not be able to undergo a lethal premature mitosis but would be able to repair damaged DNA. We recently generated a nimXcdc2AF mutant strain, in which p34cdc2 cannot be tyrosine‐phosphorylated, using a two‐step gene replacement technique after in vitro mutagenesis of nimXcdc2 (Ye et al., 1996). We therefore checked the sensitivity of dividing and non‐dividing nimXcdc2AF cells to UV irradiation. Wild‐type and DNA damage repair‐deficient uvsH4rad18 (Kafer and Mayor, 1986; Yoon et al., 1995) strains were used as controls.
When UV irradiation was applied to quiescent conidiospores, the nimXcdc2AF mutant strain was no more sensitive to UV irradiation than the wild‐type (Figure 2A). The nimXcdc2AF conidiospores must therefore be able to repair DNA damage during the process of germination before they initiate mitosis after entering the cell cycle from G1 arrest. On the other hand, the uvsH4rad18 strain, as previously reported (Kafer and Mayor, 1986), was extremely sensitive to UV irradiation, and was completely killed by a 100 J/m2 dose which only slightly affected the wild‐type and nimXcdc2AF strains (Figure 2A). The data suggest that, unlike the DNA repair‐deficient uvsH4rad18 strain, the nimXcdc2AF mutant has no deficiency in DNA damage repair mechanisms.
In contrast, germinating nimXcdc2AF conidiospores were sensitive to UV irradiation (Figure 2B). In this experiment conidiospores were first allowed to germinate for 4.5 h before UV irradiation was applied, and by which time they had entered the first cell cycle. As the nimXcdc2AF mutant cells attempt the first mitosis earlier than the wild‐type (Ye et al., 1996; see also Figure 3), we checked whether the difference in UV sensitivity between the wild‐type and the nimXcdc2AF mutant strains was caused by the different cell cycle stages at which the UV irradiation was applied. However, when germinating conidiospores of the wild‐type strain were irradiated by UV at various times after germination, their survival rate was very similar to that shown in Figure 2B. Thus, UV sensitivity of dividing nimXcdc2AF mutant cells is likely to be caused specifically by deficient DNA damage checkpoint regulation.
As MMS addition to a growing culture caused rapid Tyr15 phosphorylation and reduced the H1 kinase activity of p34cdc2 (Figure 1), we expected that the nimXcdc2AF mutant cells would also be sensitive to MMS. As shown in Figure 2C and D, the nimXcdc2AF strain was indeed much more sensitive to MMS than the wild‐type, measured either as reduction in survival rate or in colony size in the presence of various concentrations of MMS incorporated into the medium.
To confirm directly that the nimXcdc2AF mutant was deficient in the DNA damage checkpoint control, we determined the kinetics of entry into mitosis of germinating conidiospores in the presence of a low concentration of MMS. In the absence of MMS the chromosome mitotic index (CMI%) of the wild‐type strain began to rise at 5 h and peaked at 6.5 h after germination (Figure 3). Addition of MMS to the germinating medium of the wild‐type strain completely inhibited mitosis for up to 7 h after germination (Figure 3).
As previously demonstrated (Ye et al., 1996), nimXcdc2AF mutant cells entered mitosis earlier than the wild‐type cells (Figure 3). Addition of MMS delayed but did not stop the nimXcdc2AF mutant cells progressing into mitosis (Figure 3). By 7 h after germination, >50% of the nimXcdc2AF mutant cells entered mitosis, whereas none of the wild‐type cells attempted mitosis in the presence of MMS (Figure 3). These data show that the nimXcdc2AF mutant strain is deficient in the DNA damage checkpoint regulation, as it entered mitosis prematurely in the presence of damaged DNA, thus exhibiting increased sensitivity to DNA‐damaging agents.
Addition of MMS, however, did cause a delay of progression into mitosis in the nimXcdc2AF germinating conidiospores (Figure 3). Such delay may be caused by non‐specific cytotoxic effects of MMS, although addition of MMS may have activated additional checkpoint mechanisms. As germinating conidiospores enter the cell cycle from G1, candidates for such additional checkpoint mechanisms are the G1/S checkpoints which transiently delay entry into S‐phase after DNA damage, as demonstrated in the budding yeast (Siede et al., 1993, 1994; Lydall and Weinert, 1996).
We demonstrated previously that Thr14 phosphorylation cooperates with Tyr15 phosphorylation in the negative regulation of p34cdc2 in response to the S‐phase checkpoint control (Ye et al., 1996). Thus, the strain bearing both T14A and Y15F mutations of nimXcdc2 demonstrates greater sensitivity to HU than does the strain with only the Y15F mutation (Ye et al., 1996). To investigate whether Thr14 phosphorylation of p34cdc2 also had a role in the DNA damage checkpoint regulation, we compared the UV sensitivity of germinating conidiospores derived from various mutant strains (Ye et al., 1996). The strain bearing the T14A mutation alone showed a wild‐type level of sensitivity to UV (Figure 2E). Unlike the sensitivity to HU, the strain bearing both T14A and Y15F mutations did not have more UV sensitivity than did the strain with the Y15F mutation alone, and similar results were obtained using MMS as the DNA‐damaging agent (data not shown). Thus, Thr14 phosphorylation of p34cdc2 appears to have no role in the DNA damage checkpoint regulation, at least in response to UV and MMS. As previously demonstrated for HU sensitivity of the strains (Ye et al., 1996), MMS sensitivity caused by the Tyr15 mutation of p34cdc2 is also dominant over the endogenous wild‐type nimXcdc2. The strain with ΔankAwee1 was also found to be UV (Figure 2E) and MMS sensitive (data not shown). Thus the DNA damage checkpoint in A.nidulans in response to UV and MMS is likely to be mediated through Tyr15 phosphorylation of p34cdc2.
DNA damage by MMS delays entry into mitosis during nimT23cdc25 block–release by delaying Tyr15 dephosphorylation of p34cdc2
The major DNA damage checkpoint in both budding and fission yeasts operates during the G2/M transition preventing initiation of mitosis in the presence of damaged DNA (Hartwell and Weinert, 1989; Carr, 1995; Lydall and Weinert, 1996). To address the role of Tyr15 phosphorylation of p34cdc2 in the G2/M DNA damage checkpoint control specifically, we utilized G2‐specific temperature‐sensitive mutations to synchronize cells at G2 by temperature upshift before causing DNA damage with MMS, and then observed the relationship between Tyr15 phosphorylation of p34cdc2 and the initiation of mitosis upon return of the cells to permissive temperature.
Inactivation of NIMTcdc25, a homolog of the fission yeast p34cdc2‐specific tyrosine phosphatase Cdc25, blocks cells at G2 (Figure 4B, −MMS). At the arrest point, p34cdc2–cyclinB accumulates as inactive pre‐MPF in which p34cdc2 is phosphorylated at both Tyr15 and Thr161 (Figure 4A and C, G2 sample). In exponentially‐growing cells, p34cdc2 is present in two states, seen as two bands on Western blots (Figure 4A and C, EX. cells). The faster‐migrating band of p34cdc2 is associated with NIMEcyclinB and is phosphorylated at Thr161 or at both Thr161 and Tyr15 (Osmani et al., 1994; Ye et al., 1996). Upon release to permissive temperature, cells entered a rapid and synchronous mitosis with peak CMI at 10–15 min (Figure 4B, −MMS; Ye et al., 1995) as p34cdc2 became rapidly tyrosine‐dephosphorylated and activated. As previously reported (Ye et al., 1995), NIMA and NIMEcyclinB accumulated at the nimT23cdc25 G2 arrest point, and the NIMA kinase became fully activated and hyperphosphorylated after activation of p34cdc2 by tyrosine dephosphorylation (Figure 4A). As cells progressed through mitosis, both NIMA and NIMEcyclinB were degraded, leading to the down‐regulation of the two mitosis‐promoting kinases. When NIMEcyclinB was degraded, some p34cdc2 became dephosphorylated at Thr161 and was thus converted to the non‐phosphorylated slower‐migrating band on Western blot (Figure 4A and C).
MMS was added to the nimT23cdc25 G2‐arrested cells for 75 min to cause transient DNA damage and then removed from the culture by fresh medium exchange as cells were down‐shifted to release the G2 block. Addition of MMS to nimT23cdc25 G2‐arrested cells markedly delayed entry into mitosis after return to 32°C (Figure 4B, + MMS). This mitotic delay was correlated with delayed activation of p34cdc2 by tyrosine dephosphorylation (Figure 4C) and p34cdc2 remained Tyr15‐phosphorylated for 1 h compared with only 5 min without MMS addition (Figure 4A). NIMEcyclinB accumulated at the nimT23cdc25 G2 arrest point and remained stable during the MMS‐induced mitotic delay (Figure 4C). As p34cdc2 became transiently activated by Tyr15 dephosphorylation between 60 and 70 min after release (Figure 4C) cells concurrently underwent a largely synchronous mitosis with peak CMI at 70 min (Figure 4B). Then, the level of NIMEcyclinB was reduced, and p34cdc2 kinase activity was down‐regulated, as cells progressed through mitosis (Figure 4B and C). These data clearly demonstrate a strong correlation between G2 delay and Tyr15 phosphorylation of p34cdc2 after DNA damage caused by addition of MMS.
Inactivation of p34cdc2 by Tyr15 phosphorylation after DNA damage by MMS prevents rapid entry into mitosis upon release of the nimA5 mutation
Previous studies have demonstrated that NIMA is required for mitotic initiation by a mechanism that does not involve activation of p34cdc2 H1 kinase activity (Osmani et al., 1991a; Ye et al., 1996). For example, inactivation of NIMA does not prevent full activation of p34cdc2 H1 kinase activity and inactivation of NIMA prevents mitotic initiation even when p34cdc2 is fully activated and cannot be Tyr15 phosphorylated (Ye et al., 1996; Figures 5A and 6). In fact, upon release of cells from the nimA5 G2 arrest point, the levels of p34cdc2 H1 kinase activity are seen to decrease as cells are entering mitosis (Figure 5A). We therefore asked whether the activated p34cdc2 present at the nimA5 G2 arrest point is inactivated upon DNA damage by MMS. As can be seen in Figure 5B, and as expected, p34cdc2 activity increased when cells were arrested in G2 due to inactivation of NIMA. However, upon addition of MMS, p34cdc2 activity was severely inhibited and it was Tyr15‐phosphorylated (Figure 5B). Normally, release of cells from the nimA5 G2 arrest point allows cells rapidly to enter mitosis (Figure 5A). However, after MMS addition there was a long G2 delay as p34cdc2 remained Tyr15‐phosphorylated and largely inactivated. These cells eventually entered a partially synchronous mitosis as the level of Tyr15 phosphorylation decreased (Figure 5B). During the G2 delay caused by MMS addition, the level of NIMEcyclinB increased and virtually all p34cdc2 was phosphorylated on the activating Thr161 site (faster‐migrating band on NIMXcdc2 Western blot, Figure 5B). Thus, the G2 delay after DNA damage is not mediated through regulation of either NIMEcyclinB levels or CAK activity but instead correlates well with Tyr15 phosphorylation of p34cdc2.
In summary, nimT23cdc25 and nimA5 mutations arrest cells in G2 at restrictive temperature either before (nimT23cdc25) or after (nimA5) activation of p34cdc2 and generate rapid synchronous mitosis upon release into permissive temperature. Addition of MMS to nimT23cdc25 and nimA5 G2‐arrested cells caused a marked delay of entry into mitosis after release of the G2 arrests. This mitotic delay caused by MMS was correlated with Tyr15 phosphorylation of p34cdc2.
Cells unable to tyrosine‐phosphorylate p34cdc2 lack the G2 delay after DNA damage and enter mitosis prematurely
To demonstrate directly whether Tyr15 phosphorylation of p34cdc2 was responsible for the DNA damage‐induced G2 delay, we repeated the nimA5 block–release experiments with or without MMS addition using nimA5 + ΔankAwee1 and nimA5 + nimXcdc2AF double mutants. As expected, the double mutants were still arrested in G2 at the restrictive temperature for the nimA5 mutation, although p34cdc2 cannot be Tyr15‐phosphorylated (Ye et al., 1996; Figure 6). Upon release into permissive temperature, the double‐mutant strains both rapidly underwent a similar synchronous mitosis (Figure 6, −MMS). This G2 arrest using the nimA5 mutation therefore enabled us to analyze directly the consequence of lack of p34cdc2 tyrosine phosphorylation during a synchronized mitosis in the presence of damaged DNA.
Addition of MMS to nimA5 G2‐arrested cells markedly delayed entry into mitosis (Figure 6). If the ankAwee1 gene was deleted, this G2 delay was partially overcome (Figure 6) and a partially synchronous mitosis of the nimA5 + ΔankAwee1 strain occurred ∼30 min earlier than the nimA5 single mutant strain. However, if the putative inhibitory phosphorylation sites, T14 and Y15, of p34cdc2 (nimXcdc2) were mutated to non‐phosphorylatable A and F residues respectively (nimXcdc2AF), this G2 delay was then completely overcome (Figure 6). Furthermore, p34cdc2 H1 kinase activity remained high after MMS addition in the nimXcdc2AF mutant cells (data not shown). Thus, the G2 delay after DNA damage was mediated through Tyr15 phosphorylation of p34cdc2.
The partial override of the MMS‐induced G2 delay by ΔankAwee1 suggests the existence of a mik1 homolog in A.nidulans. This likelihood is further indicated as ΔankAwee1 only partially complements the nimT23cdc25 mutation, whereas non‐Tyr15‐phosphorylated p34cdc2 (nimXcdc2Y15F or nimXcdc2AF) completely complements nimT23cdc25 (P.Ramos, X.Ye, R.Fincher, S.Osmani and L.Ellis, in preparation).
To determine the consequence of the premature mitosis caused by nimXcdc2AF after DNA damage in the above block–release experiments, we compared the viability of the single nimA5 and the nimA5 + nimXcdc2AF double‐mutant cells. As shown in Figure 7, cells of either the nimA5 or nimA5 + nimXcdc2AF double‐mutant strains remained viable after nimA5 block–release in the absence of MMS treatment. With the addition of MMS, the viability of the single nimA5 mutant cells was slightly reduced. In contrast, a major reduction in viability of the double‐mutant cells was observed. Thus, A.nidulans cells unable to tyrosine‐phosphorylate p34cdc2 are deficient in the G2/M DNA damage checkpoint control. The mutant cells undergo premature lethal mitosis in the presence of damaged DNA and thus exhibit increased sensitivity to DNA‐damaging agents.
In this study we have demonstrated that A.nidulans cells have a DNA damage checkpoint mechanism operating during the G2/M transition to restrain progression into mitosis if cellular DNA is damaged. Furthermore, we have established that this G2/M DNA damage checkpoint is mediated through Tyr15 phosphorylation of p34cdc2 to bring about G2 delay of entry into mitosis. A role for the inhibitory phosphorylation of p34cdc2 in radiation‐induced G2 arrest has also recently been demonstrated in human cells (Jin et al., 1996). In A.nidulans the NIMA kinase is also required for the initiation of mitosis and has a role in the S/M checkpoint regulation (Ye et al., 1996). However, we found that NIMA does not have a role in the G2/M DNA damage checkpoint regulation. The relationship between the G2/M DNA damage checkpoint regulation and Tyr15 phosphorylation of p34cdc2 is shown in Figure 8.
The cellular level of Tyr15‐phosphorylated p34cdc2 is regulated by two opposing enzymatic activities. The Wee1 tyrosine kinase specifically phosphorylates Tyr15, and Cdc25 tyrosine phosphatase specifically dephosphorylates Tyr15 of p34cdc2. In A.nidulans the Tyr15 phosphorylation/dephosphorylation of p34cdc2 (NIMXcdc2) is carried out by ANKAwee1 and NIMTcdc25, homologs of fission yeast Wee1 and Cdc25 respectively (O'Connell et al., 1992; P.Ramos, X.Ye, R.Fincher, S.Osmani and L.Ellis, in preparation). The activities of Wee1 and Cdc25 are cell cycle‐regulated (Izumi et al., 1992; Kumagai and Dunphy, 1992; Tang et al., 1993; for a review, see Maller, 1994). During interphase, the Wee1 kinase activity is high, whereas Cdc25 phosphatase activity is low. Thus, p34cdc2 H1 kinase accumulates in the Tyr15‐phosphorylated, inactive pre‐MPF form during interphase. During the G2/M transition Cdc25 phosphatase activity is abruptly activated and Wee1 kinase inactivated, thus leading to rapid Tyr15 dephosphorylation and activation of p34cdc2 H1 kinase. It is believed that Wee1, Cdc25 and p34cdc2 H1 kinase are all in a feedback loop in which activation of p34cdc2 further activates itself through feedback activation of Cdc25 and inhibition of Wee1 (King et al., 1994; Maller, 1994). However, this regulatory circuit of p34cdc2 activation is apparently not essential during a normal cell cycle progression in A.nidulans as cells deficient in Tyr15 phosphorylation of p34cdc2 are viable (Ye et al., 1996 and the present study).
At present, it is not understood how the DNA damage checkpoint in A.nidulans regulates p34cdc2 activation to keep p34cdc2 Tyr15‐phosphorylated during response to DNA damage. This could be achieved by either activating the ANKAwee1 kinase or inactivating the NIMTcdc25 phosphatase, or both. Certainly ANKAwee1 kinase is required for the G2/M DNA damage checkpoint regulation because ANKAwee1 is the major tyrosine kinase that phosphorylates p34cdc2 at Tyr15 (P.Ramos, X.Ye, R.Fincher, S.Osmani and L.Ellis, in preparation). Moreover, ΔANKAwee1 cells are DNA damage checkpoint‐deficient and are sensitive to DNA‐damaging agents. However, whether the DNA damage checkpoint regulates ANKAwee1 kinase activity directly is currently not known.
The results of nimT23cdc25 block–release with MMS addition suggest that NIMTcdc25 may be inactivated by the DNA damage checkpoint. Normally NIMTcdc25 is very rapidly activated upon release from nimT23cdc25 G2 arrest as p34cdc2 becomes Tyr15‐dephosphorylated and activated within 5 min after release (Figure 4A). However, p34cdc2 remains highly Tyr15‐phosphorylated for 1 h after release if MMS is added to cause transient DNA damage to the nimT23cdc25 G2‐arrested cells (Figure 4C). If activation of NIMTcdc25 is indeed prevented during this temperature block–release in response to DNA damage, then p34cdc2 would remain Tyr15‐phosphorylated. The inactivation of Cdc25C phosphatase activity has been implicated in G2 arrest in response to DNA damage in human cells (O'Conor et al., 1994; Barth et al., 1996). However, we cannot exclude the possibility that high activation of ANKAwee1 after DNA damage could act to counteract the normal activation of NIMTcdc25.
At the nimA5 G2 arrest point, p34cdc2 is already activated by Tyr15 dephosphorylation (Figure 5A; Osmani et al., 1991a). According to the positive feedback loop scheme of p34cdc2 activation (King et al., 1994; Maller, 1994), activated p34cdc2, as seen at the nimA5 G2 arrest point, would generate fully activated NIMTcdc25 but inactive ANKAwee1. If this regulatory scheme is active in A.nidulans then the ANKAwee1 kinase would be inactive at the nimA5 G2 arrest point as p34cdc2 is activated. DNA damage would then have to lead to some activation of ANKAwee1, or an equivalent kinase, in order to cause Tyr15 phosphorylation of p34cdc2. In addition to the studies presented here, DNA damage by MMS can also effectively cause Tyr15 phosphorylation and inactivation of p34cdc2 in cells blocked in mitosis (X.Ye, R.Fincher, A.Tang and S.Osmani, unpublished data). This suggests that in A.nidulans checkpoint systems that monitor successful progression through mitosis could also be mediated through inactivation of p34cdc2 by Tyr15 phosphorylation. It will be of interest to see how ANKAwee1 and NIMTcdc25 are regulated in response to DNA damage at different periods of the cell cycle to ascertain their respective roles in the inhibition of p34cdc2 at different cell cycle stages.
We recently demonstrated that Tyr15 phosphorylation of p34cdc2 has a role in two overlapping S‐phase checkpoint mechanisms in A.nidulans, one involving both Tyr15 phosphorylation of p34cdc2 and the function of BIME (Ye et al., 1996). How is Tyr15 phosphorylation of p34cdc2 involved in both the S‐phase checkpoint and the G2/M DNA damage checkpoint regulation? The simplest explanation is that although cells may use different mechanisms to detect unreplicated DNA and DNA damage, the signals generated in response to such detections may converge upon the same cell cycle target, that is, inhibition of p34cdc2 by Tyr15 phosphorylation. In addition, these two checkpoint mechanisms may actually overlap as they both monitor, and respond to, the state of DNA, and therefore may share common components of a checkpoint signal transduction pathway leading to Tyr15 phosphorylation of p34cdc2. This hypothesis is supported by the observations that A.nidulans cells bearing mutations in uvsB or uvsD, which were originally isolated as UV irradiation‐sensitive mutants (Jansen, 1970; Fortuin, 1971), are deficient not only in S‐phase checkpoint control but also in the G2/M DNA damage checkpoint control (X.Ye, A.Tang, R.Fincher and S.Osmani, unpublished results). Similarly, several genes have been identified in fission yeasts that are required for both S‐phase and G2/M checkpoint controls (Carr, 1995) and these functions may also influence the Tyr15 phosphorylation state of p34cdc2, although this has not as yet been tested. Although tyrosine phosphorylation of p34cdc28 has no role in S‐phase or DNA damage checkpoint controls in budding yeast, several genes which have overlapping functions in both S‐phase and G2/M DNA damage checkpoint controls have also been identified in this system (Murray, 1995; Lydall and Weinert, 1996). The target of these genes to cause cell cycle arrest after DNA damage or inhibition of DNA replication remains to be identified.
In addition to identifying the G2/M DNA damage checkpoint system mediated through Tyr15 phosphorylation of p34cdc2, the present results also suggest that A.nidulans cells may have a p34cdc2 Tyr15 phosphorylation‐independent DNA damage checkpoint system operating in earlier stages of the cell cycle. This hypothesis is based on the observation that although non‐Tyr15‐phosphorylated p34cdc2 mutant cells completely overcome the MMS‐induced G2 delay (Figure 6), the same mutant cells show a delay of progression into mitosis when germinated from G1 in the presence of MMS (Figure 3). There are at least two possible levels of control that may mediate this delay after DNA damage that do not rely on Tyr15 phosphorylation of p34cdc2. In mammalian cells, p53 is known to be required for the DNA damage‐induced G1 arrest, partly through a mechanism involving p53‐dependent induction of the CDK inhibitor p21 (Kaufmann, 1995; Lydall and Weinert, 1996). In addition, DNA damage may cause a slowing or arrest of DNA replication which could then impose S‐phase checkpoint regulation over mitosis (Lamb et al., 1989; Paulovich and Hartwell, 1995). If DNA damage does lead to impaired DNA replication in A.nidulans it is unlikely that this would lead to our observed delay in mitosis in the nimXcdc2AF mutant strain germinated in the presence of MMS, as the slowing of S‐phase checkpoint over mitosis is defective in this strain (Ye et al., 1996). If germination in MMS leads to an arrest of the initiation of S‐phase, or causes its delay, then the arrest of S‐phase checkpoint could be operative as this level of control involves not only Tyr15 phosphorylation of p34cdc2 but also the function of bimE (Ye et al., 1996). Future studies will determine if bimE plays a role in DNA damage checkpoint control in G1 and S‐phase.
In conclusion, A.nidulans cells have a G2/M DNA damage checkpoint mechanism and the major cell cycle target for this checkpoint regulation is Tyr15 phosphorylation of p34cdc2. In response to DNA damage the checkpoint control rapidly inhibits cell cycle progression into mitosis by inactivating p34cdc2 H1 kinase through Tyr15 phosphorylation, presumably allowing time for DNA damage repair in G2 before initiation of mitosis. Having established the cell cycle targets for the S‐phase checkpoints (Ye et al., 1996) and for the DNA damage checkpoint we can now begin to delineate the signal transduction pathways leading to Tyr15 phosphorylation of p34cdc2 in response to inhibition of DNA replication and to DNA damage in A.nidulans.
Materials and methods
Aspergillus nidulans strains and general techniques
A.nidulans strains used in this study were R153 (pyroA4; wA3); SO53 (nimT23cdc25; wA2); SO54 (nimA5; wA2); ΔAnkAwee1(ΔankAwee1; pyrG89; pyr4+; pyroA4; wA3); FRY2 (nimA5; ΔankAwee1; pyrG89; pyr4+; yA2); FRY20 (pNIG6‐nimXcdc2AF; pyr4+ pyroA4; pyrG89; wA3); FRY20‐1 (nimXcdc2AF; pyroA4; pyrG89; wA3); FRY24 (pNIG6‐nimXcdc2Y15F; pyr4+; pyroA4; pyrG89; wA3); FRY25 (pNIG6‐nimXcdc2T14A; pyr4+; pyroA4; pyrG89; wA3); AT27 (nimXcdc2AF; nimA5; pyroA4; riboA2; wA3); A329 (adE20; biA1; wA3; uvsH4; methG1; pyroA4). Media and general techniques for culture, protein extraction, protein immunoprecipitation, NIMA and p34cdc2 kinase assays, Western blotting, and DAPI staining for chromosome mitotic index determination were as previously described (Osmani et al., 1987, 1991a, 1994; Oakley and Osmani, 1993; Ye et al., 1995).
Sensitivity test to UV irradiation and MMS
Non‐dividing and dividing cells deficient in Tyr15 phosphorylation of p34cdc2 were tested for sensitivity to UV irradiation. Conidiospores (dormant in a quiescent G0 state) were suspended in 0.2% Tween‐20 and plated out on YAG plates (250 spores/plate). The plates were then irradiated immediately with UV using a microprocessor‐controlled UV crosslinker (FB‐UVXL‐1000; Fisher Biotech, Pittsburgh, PA, USA) to determine UV sensitivity of non‐dividing cells. To determine UV sensitivity of dividing cells, conidiospores on YAG plates were first allowed to germinate for 4.5 h at 32°C before UV irradiation. By this time the germinated spores had entered the cell cycle and were about to undergo the first mitosis. UV‐irradiated germlings were then incubated in a 32°C incubator for 2 days for colony formation. The survival rate after UV irradiation was determined as a percentage of colonies produced by control conidiospores without UV irradiation.
In MMS (Aldrich, St Louis, MO, USA) sensitivity tests, various concentrations of MMS were incorporated into YAG medium. To determine the survival rate in the presence of MMS, conidiospores (250 spores/plate) were plated out on YAG plates containing MMS. After a 2 day incubation at 32°C, the survival rate was determined as the percentage of colonies produced by control conidiospores on YAG plates containing no MMS. MMS sensitivity was also measured by reduction in colony growth. In this case conidiospores were spot‐inoculated with tooth‐picks onto YAG plates containing MMS.
nimT23cdc25 and nimA5 temperature block–release
nimT23cdc25 and nimA5 mutant cells were first grown to early log phase at 32°C and were then rapidly upshifted to the restrictive temperature of 42°C. After G2 arrest at 42°C, cells were released into synchronous mitosis by temperature downshift to 32°C. To determine the effect of MMS on the rapid synchronous mitosis generated by nimT23cdc25 and nimA5 block–release, MMS (0.04%) was added to the G2‐arrested cells 2.5 h after temperature upshift to 42°C. After incubation at 42°C for a further 75 min, MMS was removed from the culture by medium exchange and cells were then downshifted to permissive temperature of 32°C.
Cell viability assay after nimA5 block–release
nimA5 mutant conidiospores were germinated at 42°C for 8 h in YG medium containing 0.1% agar to prevent clumping. The germinating spores were arrested in the first cell cycle at the nimA5 G2 arrest point. MMS (0.04%) was added to the nimA5 G2‐arrested germlings for 75 min to cause DNA damage and was then removed from the culture by medium exchange with fresh YG containing no agar. The germlings were then released from nimA5 G2 arrest into the permissive temperature of 32°C. The germlings were resuspended in 0.2% Tween‐20, plated out (250 germlings/plate), and incubated at 32°C for colony formation. Cell viability was assessed as the percentage of colonies produced by 250 conidiospores without nimA5 temperature block–release and MMS addition.
We thank Dr L.Ellis and Dr P.Ramos (W.M.Keck Center for Genome Informatics, Institute of Biosciences and Technology, Texas A&M University) for providing the ΔankAwee1 strain. We also thank Elizabeth Oakley and Dr B.R.Oakley for critically reading the manuscript. This work was supported by NIH grant GM42564 and by funds from the Geisinger Clinic Foundation.
- Copyright © 1997 European Molecular Biology Organization