ΔMEKK3:ER* activation induces a p38α/β2-dependent cell cycle arrest at the G2 checkpoint

Abstract

Whilst many studies have examined the role of the MAP Kinases in regulating the G1→S transition, much less is known about the function of these pathways in regulating other cell cycle transitions. Stimulation of the conditional mutant ΔMEKK3:ER* in asynchronous hamster (CCl39) and rat (Rat-1) fibroblasts resulted in the strong activation of endogenous JNK and p38 but only a weak activation of ERK. Activation of ΔMEKK3:ER* inhibited cell proliferation through a combination of an initial G1 and G2 cell cycle arrest, followed by a delayed onset of apoptosis. When cells were synchronized in S phase with aphidicolin and then released, activation of ΔMEKK3:ER* resulted in the up-regulation of p21CIP1 and a pronounced inhibition of cyclin A/CDK2 and cyclin B1/CDK1 kinase activity. Analysis of mitotic figures indicated that cells failed to enter mitosis, arresting late in G2. ΔMEKK3:ER*-mediated CDK inhibition and G2 arrest did not absolutely require p21CIP1, since both events were observed in Rat-1 cells in which p21CIP1 is transcriptionally silenced due to promoter methylation. Rather, CDK inhibition was associated with a down-regulation of cyclin A and B1 expression. Finally, application of the p38 inhibitor SB203580 partially restored cyclin B associated kinase activity and allowed cells to proceed through mitosis into the next G1 phase, suggesting that activation of the p38α/β2 pathway can promote a G2 cell cycle arrest.

Introduction

Cells that are exposed to chemical stresses, including many commonly used chemotherapeutic agents, undergo cell cycle arrest, allowing the cell to repair the lesion or undergo apoptosis. Cells arrest at specific checkpoints, either late in G1 or late in G2, to prevent replication or segregation of damaged DNA (for reviews see Sherr, 1996; Clarke and Giménez-Albián, 2000). The p53-dependent up-regulation of p21CIP1 (El-Deiry et al., 1993; Macleod et al., 1995; Brugarolas et al., 1995; Deng et al., 1995; Waldmann et al., 1995) is critical for both imposing and maintaining the DNA damage-induced G1 checkpoint (Brugarolas et al., 1995; Deng et al., 1995). p21CIP1 over-expression results in inhibition of cyclin-dependent kinase 2 (CDK2)-containing complexes (Harper et al., 1995; Poon et al., 1996), hypophosphorylation of the Retinoblastoma protein (pRb) and sequestration of E2F transcription factors (Hiebert et al., 1992; Knudsen and Wang, 1997). The importance of this pathway in growth regulation is exemplified by the frequency of inactivating mutations in p53 or components of the Rb pathway in human tumours (Hollstein et al., 1994; Hall and Peters, 1996).

The passage of cells through S phase and into G2 is primarily controlled by cyclin A/CDK2, whilst M phase progression is regulated by the activity of the cyclin B/CDK1 complex. During the normal cell cycle, cyclin B1 is expressed during G2, whereupon it binds to CDK1 (Labbe et al., 1989; Pines and Hunter, 1989). CDK1 subsequently undergoes an activating phosphorylation at T161 catalysed by CAK (CDK-activating kinase) (Solomon et al., 1992; Fesquet et al., 1993). Once active cyclin B1/CDK1 is formed, Wee1 and Myt1 phosphorylate CDK1 on T14 and Y15, which serves to inhibit its kinase activity (Russell and Nurse, 1987; Gould and Nurse, 1989; Meijer et al., 1991; Mueller et al., 1995). The transition between metaphase and anaphase is then triggered by the Cdc25C-catalysed dephosphorylation of T14 and Y15, resulting in the activation of CDK1 (reviewed in Piwnica-Worms, 1999). Further auto-phosphorylation events lead to the relocation of MPF to the nucleus, where it subsequently phosphorylates its target substrates (Hagting et al., 1999). Following DNA damage, Cdc25C is phosphorylated by Chk1 and Chk2, which are in turn activated downstream of the ATM or ATR kinases (Matsuoka et al., 1998; Brown et al., 1999; Liu et al., 2000). Phosphorylation of Cdc25C promotes its sequestration in the cytoplasm by 14-3-3 proteins, resulting in enhanced Y15 phosphorylation and inhibition of CDK1 and arrest of cells at G2 (Jin et al., 1996; Blasina et al., 1997; Lopez-Girona et al., 1999; Yang et al., 1999).

The MAPK (mitogen-activated protein kinase) signalling pathways serve to convey signals into the nucleus, thereby determining whether a cell undergoes division, cell cycle arrest or apoptosis. The Raf→MEK1/2→ERK1/2 pathway regulates cell cycle re-entry (Woods et al., 1997; Balmanno and Cook, 1999; Cook et al., 2000). The SAPKs (stress-activated protein kinases, JNK and p38) are stimulated as part of the biological response to noxious stimuli and are activated by specific MAP Kinase Kinases (MKKs); MKK4 and MKK7 activate JNK whilst MKK3 and MKK6 activate p38. The SAPK-specific MKKs are in turn activated by an array of recently described MKKKs including the MEKKs, ASKs and MLKs (for reviews see Cook et al., 2000; Davis, 2000; Nebrada and Porras, 2000).

JNK and p38 are activated in response to many DNA damaging agents but the precise role of the SAPKs in cell cycle regulation remains unclear. JNK and p38 have been proposed to have both pro and anti-proliferative effects during the G1→S transition (Wang et al., 1999; Wisdom et al., 1999) but less is known about their functions in other stages of the cell cycle. It is likely that this confusion arises because few studies have dissociated SAPK activation from stress-induced cellular damage, which may be inherently growth suppressive. Furthermore, many studies have been performed on serum-starved cells, in G0/G1, which may be pre-disposed to block late in G1 at the first checkpoint they encounter.

To examine the biological effects of the stress kinases in the absence of any stressful lesion we have used a conditional mutant, ΔMEKK3:ER*, which when stimulated with 4-hydroxytamoxifen (4-HT) causes activation of JNK, p38, and to a lesser extent ERK. Activation of ΔMEKK3:ER* in quiescent cells inhibits serum-stimulated cell cycle re-entry (D Todd et al., in preparation). Here we show that activation of ΔMEKK3:ER* in asynchronous cycling cells results in the accumulation of cells with a 4n DNA content which cannot enter mitosis. This G2 arrest appears to be independent of p53 or p21CIP1 expression but requires the activity of the p38α or β2 SAPKs.

Results

Biochemical characterization of ΔMEKK3:ER*

To examine the biological effects of the MAPK and SAPK signalling cascades in the absence of cellular stress we used a conditional kinase (ΔMEKK3:ER*) in which the kinase domain of human MEKK3 is fused ‘in frame’ to a modified form of the oestrogen receptor (ER*). The kinase activity of this chimeric protein is dependent upon the addition of 4-hydroxytamoxifen (4-HT). As a negative control, a catalytically inactive variant (ΔMEKK3:ER*KD) was constructed by replacing Lys 391 (required for ATP binding) with methionine. These constructs were transfected into CCl39 and Rat-1 fibroblasts and clonal cell lines expressing the construct were isolated. Four cell lines were derived from CCl39 cells (CM3.3, CM3.6 and CM3.7 expressing ΔMEKK3:ER* and CM3.KD expressing the catalytically inactive version of ΔMEKK3:ER*) and two from Rat-1 cells (RM3.16 and RM3.21 expressing ΔMEKK3:ER*), and these were used for subsequent experiments.

Initial characterization of ΔMEKK3:ER* was performed in the CCl39 cell lines. Expression of ΔMEKK3:ER* was monitored by Western blotting, with the 9E10 monoclonal antibody directed against the N-terminal Myc epitope tag which detected the fusion protein at the predicted molecular mass of 70 kDa in all CM3 clones, but not in the untransfected parental controls (Figure 1a). A 24 h treatment with 4-HT increased the abundance of the ΔMEKK3:ER* fusion protein (Figure 1a); this has been observed with other kinase:ER* fusions (Samuels et al., 1993) and probably reflects enhanced stability of the de-repressed ΔMEKK3:ER* fusion protein.

Figure 1
figure1

Characterization of ΔMEKK3:ER*-induced signalling in CCl39 cells. (a) CCl39 cells or CM3 clones were treated with 4-HT or ethanol vehicle control for 24 h and whole lysates were immunoblotted with the 9E10 monoclonal antibody to reveal expression of Myc-tagged ΔMEKK3:ER*. (b) Serum-starved CM3.3 cells were stimulated 100 nM 4-HT (closed symbols) or ethanol vehicle control (open symbols) for the indicated times and ERK1, JNK1 or p38α activity was assayed by immune complex kinase assay. (c) Serum-starved CM3.3 and CM3.KD cells were stimulated with 100 nM 4-HT for 2 h and ERK1 (open bars), JNK1 (closed bars) and p38a (hatched bars) were assayed by immune complex kinase assay. (d) Serum-starved CM3.3 cells were treated with increasing concentrations of 4-HT for 2 h and ERK1, JNK1 or p38α activity was assayed by immune complex kinase assay. (e) Serum-starved CM3.3 cells were stimulated with 100 nM 4-HT for the times indicated and whole cell lysates were immunoblotted with the 9E10 monoclonal antibody to detect Myc-tagged ΔMEKK3:ER* and total or phospho-Ser73 c-Jun. In all cases results are shown from a single representative experiment and similar results were obtained in two further independent experiments. In addition, these results were reproducible in other independently derived CM3 clones

To first determine which pathways can be activated by ΔMEKK3:ER*, serum starved CM3.3 cells were used, since this reduced the activity of ERK1/2 to a basal level thereby affording a clearer assessment of the activation of each kinase. The ability of ΔMEKK3:ER* to activate ERK1, JNK1 and p38α was assessed by immune complex kinase assays. CM3.3 cells were treated with either 100 nM 4-HT or 0.04% (v/v) ethanol (as a vehicle control) for varying lengths of time before assay. Analysis revealed that activation of ERK1, JNK1 and p38α was apparent 15–30 min after addition of 4-HT and maximal activation of all three kinases was observed at 1 h (Figure 1b). The activity of all three kinases decreased slightly by 3 or 5 h but was then maintained at the same level for the full 24 h of the time course, whereas treatment with enthanol was ineffective (Figure 1b). Identical kinetics were observed in RM3 cells (K Balmanno, unpublished observations). Treatment of serum-starved CM3.KD cells, expressing the K→M mutant form of ΔMEKK3:ER*, with 4-HT had no effect on ERK, JNK or p38 activity (Figure 1c). 4-HT elicited a dose-dependent activation of ERK1, JNK1 and p38α in CM3.3 cells, with an EC50 of 5–10 nM being observed for all kinases (Figure 1d). Finally, activation of JNK was reflected in the phosphorylation of c-Jun as monitored by the reduced mobility of total c-Jun and the enhanced phosphorylation of c-Jun at Ser73 (Figure 1e). c-Jun phosphorylation was apparent within 30 min of addition of 4-HT whereas increased expression of ΔMEKK3:ER* was not seen until between 5 and 24 h. This suggests that 4-HT-induced de-repression of pre-existing ΔMEKK3:ER* is the primary mechanism for activation. Taken together, these results indicate that in serum-starved cells ΔMEKK3:ER* directly activates all three major MAPK/SAPK pathways with identical kinetics and dose-dependency and the effects of 4-HT are solely attributable to the kinase domain of ΔMEKK3:ER*.

ΔMEKK3:ER* activates p38 and JNK to levels similar to that observed with cellular stresses

Since we wanted to determine the effect of SAPK activation in asynchronous cells, we examined the activation of ERK, JNK and p38 by ΔMEKK3:ER* in sub-confluent cycling cells growing in 10% FBS and made comparisons with various cellular stresses. Stimulation of both CM3.3 and RM3.21 cells with 100 nM 4-HT elicited a robust increase in p38α and JNK1 activity in both cell lines (12–15-fold for CM3.3 compared with 9–24-fold for RM3.21) (Figure 2a,b). This SAPK activation was of a magnitude comparable to that observed in response to UV or anisomycin, two well-characterized stress agonists, and approximately twofold greater than that observed with sodium arsenite. In contrast to the robust response observed in serum-starved cells, treatment with either FBS or activation of ΔMEKK3:ER* resulted in only a modest 2–3-fold increase in ERK1 activity. This discrepancy is simply due to the higher basal level of ERK activity in cycling cells when compared to serum-starved cells. For example, serum starvation causes a five or sixfold reduction in ERK1 activity, so that the resultant fold increases appear correspondingly greater in serum-starved cells than in cycling cells (A Garner, K Balmanno and S Cook, unpublished results).

Figure 2
figure2

Activation of MAPK and SAPK by ΔMEKK3:ER* and cellular stresses in cycling cells. Asynchronous CM3.3 cells (a) and RM3.21 cells (b) growing in 10% (v/v) FBS, were stimulated with 100 nM 4-HT, 10 μg/ml sodium arsenite (Na As), 1 mM hydrogen peroxide or 30 μM anisomycin for 1 h, 10% FBS for 10 min or exposure to 40 mJ UV followed by a 1 h recovery. In all cases ERK1, JNK1 and p38α activities were assayed by immune complex kinase assay. Data is expressed as fold activation relative to the background signal obtained in ethanol (vehicle control) treated cell extracts. Similar results were observed in three independent experiments

Thus, activation of ΔMEKK3:ER* in cycling cells results in the activation of JNK and p38 but causes only a modest further increase in ERK activity. Furthermore, the activation of JNK and p38 is of a magnitude similar to that observed in response to stress agonists, rather than producing a non-physiological super-activation of these pathways.

Activation of ΔMEKK3:ER* inhibits cellular proliferation

The effects of ΔMEKK3:ER* activation upon cellular proliferation were initially investigated by means of a clonogenicity assay. Wild-type CCl39 cells were transfected with the plasmid encoding ΔMEKK3:ER* and the gene for puromycin resistance. These transfected cells were then divided into two populations of equal cell number, plated and selected in puromycin for 14 days in the absence or presence of 1 μM 4-HT. The resulting colonies were visualized by staining with crystal violet which revealed that activation of ΔMEKK3:ER* resulted in a profound inhibition of colony formation (Figure 3a, upper panel). This growth inhibition was entirely dependent upon the kinase activity of ΔMEKK3:ER* as 4-HT-mediated activation of a catalytically inactive version of ΔMEKK3:ER* (KD) had no effect upon colony formation (Figure 3a, bottom panel).

Figure 3
figure3

Activation of ΔMEKK3:ER* inhibits cell proliferation. (a) CCl39 cells were transfected with either pBabePuroΔMEKK3:ER* (top panel) or a catalytically inactive version (KD, bottom panel). Twenty-four hours post-transfection, the cell population was divided, replated and selected in puromycin-containing media±1 μM 4-HT for 2 weeks. Surviving colonies were fixed and stained with crystal violet. (b) Individual, asynchronous CM3 clones (CM3.KD, CM3.3, CM3.6 and CM3.7) were treated with increasing concentrations of 4-HT for 72 h before cellular proliferation was analysed by measuring the absorbance of crystal violet stained protein at 590 nm. Data are expressed as mean±range of duplicate samples. (c) CM3.3, CM3.6 and CM3.7 cells were treated with 4-HT for 2 h and JNK1 and p38α activities were measured by immune complex kinase assay. Data is expressed as fold activation relative to the background signal obtained in ethanol (vehicle control) treated cell extracts. Similar results were observed in three independent experiments

These results were confirmed by studying the effect of increasing concentrations of 4-HT on the asynchronous growth of the CM3.3, CM3.6, CM3.7 and CM3.KD clonal cell lines. Cells were grown for 72 h in the presence of increasing concentrations of 4-HT before determining relative cell numbers by crystal violet staining. Quantitative analysis revealed that the growth of all three CM3 cell lines was inhibited by 4-HT in a dose-dependent manner (Figure 3b). The observed IC50 for inhibition of cell growth in the CM3.3 cell line (3–5 nM) was consistent with the 4-HT dose required for half-maximal activation of ERK, JNK and p38 (Figure 1d) in this cell line. In addition, it was noted that the degree of growth inhibition obtained in each CM3 clone (CM3.3>CM3.6> CM3.7) was proportional to the magnitude of JNK and p38 activation obtained (Figure 3c) with the growth of CM3.KD being unaffected by the addition of 4-HT. These results indicate that activation of ΔMEKK3:ER* by 4-HT causes inhibition of cell proliferation which is proportional to the degree of activation of the downstream signalling pathways, primarily JNK and p38.

Activation of ΔMEKK3:ER* leads to cell cycle arrest followed by apoptosis

Whilst the preceding results unequivocally proved that ΔMEKK3:ER* activation retarded cellular growth, neither of the assays could distinguish between growth arrest or cell death. We therefore monitored the cell cycle distribution of asynchronous, cycling CM3 cells after treatment with 100 nM 4-HT, by staining with propidium iodide and analysing by flow cytometry (Figure 4 and Table 1). Within 5 h of ΔMEKK3:ER* activation the fraction of CM3.3 cells in S phase was reduced (Figure 4, top panel). After 10 h all three cell lines were largely composed of cells with either a 2n or a 4n DNA content (Figure 4 and Table 1). In contrast, 4-HT had no effect upon the cell cycle distribution of CM3.KD cells, indicating that the observed cell cycle arrest was entirely due to the kinase activity of ΔMEKK3:ER*. Thus, ΔMEKK3:ER* activation resulted in the loss of S phase and accumulation of cells in the G1 or G2/M phases of the cell cycle.

Figure 4
figure4

Activation of ΔMEKK3:ER* induces cell cycle arrest and apoptosis. Asynchronous CM3.3 cells were treated with 100 nM 4-HT for up to 35 h. Cells were fixed and stained with propidium iodide before their cell cycle profiles were analysed by flow cytometry. Representative histograms of number of cells versus DNA content are displayed. The corresponding percentage of cells in each phase of the cell cycle was quantified and is displayed in Table 1. Similar results were observed in up to six independent experiments in all three clones

Table 1 Effect of ΔMEKK3:ER* on the cell distribution of CM3.3, CM3.6, CM3.7 and CM3.KD cells

The accumulation of cells at G1 or G2/M was a persistent response, with cells remaining blocked at these stages for the full 35 h duration of the experiment. However, from approximately 20 h onwards the proportion of cells exhibiting a <2n DNA content (so-called ‘sub G1’ or hypo-diploid cells) increased substantially, concurrent with a decrease in the fraction of cells at G1 or G2/M (Figure 4 and Table 1). This sub-G1 fraction represents cells undergoing apoptosis, since we have observed increases in Annexin-V staining and other markers of apoptosis over the same timecourse (C Weston and S Cook, unpublished observations).

These results demonstrate that activation of ΔMEKK3:ER* in asynchronous cells is sufficient to trigger both cell cycle arrest and apoptosis. The initial response to ΔMEKK3:ER* activation, the loss of cells in S phase and arrest of cells at G1 and G2/M, was apparent after only 5–6 h, whilst the earliest signs of apoptosis were not observed until after 20 h. In this study we have focused on the arrest of cells at G2 or M phase as it was one of the primary effects of ΔMEKK3:ER* activation.

Activation of ΔMEKK3:ER* results in a G2 block

To determine more precisely the location of the G2/M arrest, both CM3.3 and RM3.21 cells were synchronized in S phase using aphidicolin and then released. In the absence of 4-HT these cells progressed through S phase, G2 and mitosis so that 25 h after aphidicolin release the majority of cells had entered the next G1 phase (Figure 5a,b, top panels). In contrast, cells released from aphidicolin in the presence of 4-HT progressed through S phase but arrested with a 4n DNA content so that by 25 h few cells had progressed through to the next G1 (Figure 5a,b, bottom panels). Subsequent experiments revealed that activation of ΔMEKK3:ER* could be delayed for as long as 12 h after aphidicolin release and cells would still arrest with a 4n DNA content (A Garner and S Cook, unpublished observations), suggesting that the arrest was either late in G2 or early in mitosis.

Figure 5
figure5

Activation of ΔMEKK3:ER* prevents cells from initiating mitosis. (a) CM3.3 and (b) RM3.21 cells were blocked with 1 μg/ml aphidicolin for 16 h and then released for the indicated times. At the time of release, cells were either treated with ethanol (vehicle control, top panel) or 100 nM 4-HT (bottom panel). Cells were then harvested, fixed and stained with propidium iodide before being examined by flow cytometry. Representative cell cycle histograms are displayed (A and B). (c) Aphidicolin synchronized RM3.21 cells were released for 7 h before being stimulated with either ethanol or 100 nM 4-HT. Mitotic cells were trapped 3 h post 4-HT stimulation by the addition of 250 ng/ml nocodazole for 14 h. Mitotic index was determined as described in Materials and methods. (d) Corresponding representative photographs of geimsa-stained chromosomes obtained following nocodazole trapping of RM3.21 cells (40× magnification)

To determine if the cells arrested in mitosis, we used nocodazole to trap CM3.3 and RM3.21 cells undergoing mitosis following aphidicolin release and scored the number of mitotic figures. Interestingly, RM3 cells were found to be markedly more sensitive to nocodazole than CM3 cells. This disparity may be due to differences in the Chfr status between these cell lines, resulting in the absence of the recently defined mitotic stress checkpoint in Rat-1 cells (Scolnick and Halazonetis, 2000). Activation of ΔMEKK3:ER* 7 h after release from aphidicolin significantly reduced the number of cells trapped in mitosis, indicating that the 4n arrest took place late in G2 (Figure 5c). This G2 arrest was found to be stable in both cell lines even at time points of up to 48 h (A Garner, unpublished observations). The absence of any condensed chromosomes in arrested RM3.21 cells further supported the finding that ΔMEKK3:ER* activation caused cells to arrest late in G2 before the onset of mitosis (see Figure 5d).

Activation of ΔMEKK3:ER* causes the loss of cyclin A and cyclin B1 and their associated kinase activities

The aphidocolin synchronization and release protocol allowed us to analyse the effects of ΔMEKK3:ER* activation upon the proteins that normally regulate the G2-to-M transition. Western blot analysis of whole cell lysates from CM3.3 cells released from an aphidicolin block revealed that cyclin A levels were initially high but decreased following prolonged ΔMEKK3:ER* activation, whilst CDK2 levels were unaffected (Figure 6a). In the absence of 4-HT, cyclin B1 accumulated as expected when the cells passed through G2 and into mitosis (Figure 6a). Furthermore, a slower migrating form of cyclin B1 was observed after 15 h (Figure 6a) and most likely reflects cyclin B1 phosphorylation (Li et al., 1995; Hagting et al., 1999). However, activation of ΔMEKK3:ER* when cells were released from the aphidicolin block prevented the accumulation of cyclin B1 protein which was normally observed after 10–15 h. Indeed, cyclin B1 levels actually declined to below control levels, and were nearly undetectable 25 h after cells were released from aphidicolin. In addition, we also observed a modest decrease in CDK1 expression levels and an even greater decrease CDK1 Y15 phosphorylation in the presence of 4-HT (Figure 6a). The decrease in total CDK1 levels was variable in magnitude and timing of onset (see for example Figure 6c), whereas the reduction in CDK1 Y15 phosphorylation was highly reproducible. Immunoblots for total ERK1/2 served as a loading control. Thus, activation of ΔMEKK3:ER* causes the loss of cyclin A, cyclin B1 and prevents the phosphorylation of CDK1 at Y15.

Figure 6
figure6

Activation of ΔMEKK3:ER* in CM3.3 cells results in down-regulation of cyclins A and B, inhibition of CDK2 and CDK1 and an increase in p21CIP1. Aphidicolin synchronized CM3.3 cells were released for the specified periods of time in the presence of 100 nM 4-HT or ethanol. (a) Whole cell lysates were immunoblotted for cyclin A, CDK2, cyclin B1, CDK1, phospho-CDK1Y15 and total ERK1/2 as a loading control. (b) Normalized cell lysates were used to immunoprecipitate either cyclin A or cyclin B and the associated kinase activities were assayed as described in the Materials and methods. (c) Whole cell lysates were immunoblotted for p53 and p21CIP1 (upper panel) whilst CDK2 and CDK1 immunecomplexes were immunoblotted for the presence of p21CIP1 and the relevant CDK

To examine the impact of loss of cyclin A and B on the kinase activities of CDK2 and CDK1, the active complexes were immunoprecipitated using antibodies against cyclin A and cyclin B, and their ability to phosphorylate histone H1 assayed. Upon release from the aphidicolin block, the initial kinase activity associated with cyclin A was high as CM3.3 cells passed through S and G2 (Figure 6b). However, upon addition of 4-HT, the cyclin A-associated kinase activity was dramatically inhibited and dropped to virtually zero after 20 h. Activation of ΔMEKK3:ER* completely blocked the increase in cyclin B1-associated histone H1 kinase activity which normally occurs as cells pass from metaphase into anaphase (Figure 6b).

The activity of CDK's is also regulated by CDK inhibitors (CDKI's) which either disrupt the CDK-cyclin complex or bind within the CDK catalytic core. p21CIP1 is known to be up-regulated by cellular stresses such as H2O2 and DNA damaging agents in a p53-dependent manner (El-Deiry et al., 1993; Waldmann et al., 1995). Whilst p21CIP1 is well known as a regulator of the G1→S transition, several studies have recently suggested that it may also be involved in regulating the G2 checkpoint (Dulic et al., 1998; Medama et al., 1998; Passlaris et al., 1999). We therefore examined whether ΔMEKK3:ER* could up-regulate the expression of p21CIP1. In the absence of 4-HT little or no p21CIP1 was expressed, indicating that aphidicolin treatment and release did not stress the cells sufficiently to induce p21CIP1 (Figure 6c). Activation of ΔMEKK3:ER* yielded a striking increase in p21CIP1 protein levels (Figure 6c), whilst p27KIP1 was unaffected (A Garner, unpublished observations). Examination of the cyclin A/CDK2 complexes by Western blotting revealed a large amount of p21CIP1 co-immunoprecipitating with CDK2 under these conditions (Figure 6c). Since p21CIP1 is an effective inhibitor of CDK2 these results would be consistent with ΔMEKK3:ER* inhibiting CDK2 by up-regulating the expression of p21CIP1. Under the same conditions we failed to detect p21CIP1 in CDK1 immunecomplexes by Western blotting (see Figure 6c). Collectively, these data suggest that inhibition of cyclin B/CDK1 activity in CM3.3 cells is not likely to be due to p21CIP1 association (Figure 6c) or inhibitory CDK1 phosphorylation, which actually decreases as CDK1 levels decrease (Figure 6a).

p21CIP1 expression can be induced by p53-dependent pathways (El-Deiry et al., 1993; Waldmann et al., 1995) but also by p53-independent pathways, including the Ras-dependent Raf-MEK-ERK cascade (Woods et al., 1997). The expression of p21CIP1 appeared to be independent of p53 in CM3 cells as little increase in p53 levels was observed upon the addition of 4-HT (see Figure 6c). To further characterize the apparent p53-independent nature of this p21CIP1 expression, the p53 status of these cell lines was assessed by treatment with Doxorubicin (Dox), a known activator of the p53→p21CIP1 pathway. The requirement for p53 function in Dox-induced p21CIP1 expression and G1 cell cycle arrest was first defined using HCT116 parental cells (p53+/+) and the p53 deficient (p53−/−) derivatives (Bunz et al., 1998). As expected treatment of parental HCT116 cells with Doxorubicin caused an increase in p53 and p21CIP1 (Figure 7a) and a loss of S phase with cells arresting at G1 and G2 (Figure 7b). In contrast, Dox-treated HCT116 p53−/− cells failed to express p21CIP1 and only accumulated in G2, confirming that a functional p53→p21CIP1 pathway is required for Dox-induced G1 arrest (Figure 7a,b). Dox treated CM3.3 cells also failed to increase p53 or p21CIP1 levels and only accumulated at G2 with a 4n DNA content. However, the CM3.3 cells were still capable of expressing p21CIP1 levels (Figure 7a) and arresting at G1 and G2 (Figures 4 and 5) in response to ΔMEKK3:ER*. These results suggest that whilst the p21CIP1 locus is intact and responsive to ΔMEKK3:ER* and the G1 and G2 checkpoint machinery is intact, CM3.3 cells either lack functional p53 or another component of the DNA damage regulated p53 pathway.

Figure 7
figure7

CM3.3 cells fail to induce p53 or p21CIP1 in response to DNA damage. (a) Asynchronous HCT116 (WT), HCT116 (p53−/−) and CM3.3 cells were treated with vehicle control (C), 1 μM Doxorubicin (D) or 100 nM 4-hydroxytamoxifen (4-HT) for 18 h. Whole cell lysates were fractionated by SDS–PAGE and subjected to immunoblotting with antibodies for p53 and p21CIP1. (b) Cycling HCT116 or CM3.3 cells were treated with vehicle or Dox for 18 h and were then stained with propidium iodide prior to analysis of cell cycle profiles

The ΔMEKK3:ER* induced G2 arrest proceeds in the absence of p21CIP1

Although p21CIP1 is known to inhibit CDK2 activity, its role in G2 checkpoint control remains poorly defined. To determine if p21CIP1 was absolutely required for G2 arrest we took advantage of the recent report that Rat-1 cells fail to express p21CIP1 in response to genotoxic stresses due to methylation and silencing of their p21CIP1 promoter (Allan et al., 2000). Treatment of RM3.21 cells with 4-HT resulted in activation of all three classes of MAPK cascade in a similar fashion to CM3.3 (Figure 2b and unpublished results). However, when directly compared with CM3.3 cells, activation of ΔMEKK3:ER* failed to result in the up-regulation of p21CIP1 in aphidicolin-synchronized RM3.21 cells (Figure 8a). This was not due to any inability of the antibody to detect rat p21CIP1, since the protein was readily detectable when cells were treated with the DNA methyltransferase inhibitor 5-Aza-2′-deoxycytidine, confirming that the p21CIP1 promoter is methylated and silenced in Rat-1 cells (Figure 8b). Despite the lack of inducible p21CIP1 in RM3 cells, activation of ΔMEKK3:ER* still inhibited cyclin A- and cyclin B1-associated kinase activity (Figure 8c) and still caused a G2 arrest (Figure 5b,c,d). Thus, the ability of ΔMEKK3:ER* to inhibit CDK2 and CDK1 activity and promote a G2 cell cycle arrest proceeds normally in Rat-1 cells and so is not absolutely dependent upon p21CIP1.

Figure 8
figure8

RM3.21 cells display ΔMEKK3:ER* mediated CDK inhibition in the absence of p21CIP1. (a) RM3.21 and CM3.3 cells were simultaneously released from S phase in the presence of 100 nM 4-HT for 15 and 25 h. The relative induction of p21CIP1 in the two cell lines was compared by immunoblotting equal quantities of whole cell lysates for p21CIP1; total ERK1/2 served as a loading control. (b) RM3.21 cells were treated with the indicated concentrations of 5-aza-2′-deoxycytidine and whole cell lysates were immunoblotted for expression of p21CIP1 and total ERK1/2. (c) RM3.21 cells were released from S phase in the presence of 100 nM 4-HT or ethanol for the indicated timepoints and cyclin A- and cyclin B1-associated kinase activities determined as described in the Materials and methods section. Results are taken from a single experiment representative of three giving similar results

The ΔMEKK3:ER* mediated G2 arrest is dependent on p38α/β2 activity

As ΔMEKK3:ER* can activate all three major classes of MAPK cascades, selective pharmacological inhibitors of the ERK or p38α/β2 pathways were used to try and dissect which of the pathways were responsible for the G2 arrest (Favata et al., 1998; Cuenda et al., 1995; Enslen et al., 1998). Ten μM U0126, an inhibitor of MKK1/2 which prevents ERK activation, had little effect upon the ΔMEKK3:ER*-induced G2 arrest (Figure 9a,b) despite reducing ERK phosphorylation to control levels (Figure 9c). Pre-treatment of CM3.3 cells with 5 μM SB203580 allowed a significant proportion of cells to bypass the ΔMEKK3:ER* mediated G2 checkpoint, progress through mitosis and into the next G1 (Figure 9a,b). The ability of SB203580 to reverse the ΔMEKK3:ER*-induced G2 arrest was also reflected in rescue of cyclin B-associated kinase activity and cyclin B levels (Figure 9d). Activation of ΔMEKK3:ER* reduced total cyclin B-associated kinase activity to 10.1% of control values, and this was partially rescued by SB203580 (39.5% of control values, significantly different by student's t-test, P=0.002). This was also accompanied by a partial rescue of cyclin B levels (Figure 9d, lower panel). The disparity between the nearly complete rescue of cells from G2 arrest (Figure 9a,b) and the partial rescue of cyclin B associated kinase activity probably reflects the fact that cells progressively lose synchrony following release from aphidicolin, and the G2→M transition is relatively brief. Consequently, assay of cyclin B-associated kinase activity in a population of cells will include those that have been rescued by SB203580 and are transiting the G2/M boundary at the time of assay, but also those that haven't reached the G2/M transition yet and those which have been rescued by SB, have passed through M phase into G1 and in which cyclin B/CDK1 activity has declined. Similar rescue of the G2 arrest was seen in RM3.21 cells, where SB203580 restored the mitotic index (Figure 9e) and allowed cells to progress into G1 (A Garner and S Cook, unpublished observations).

Figure 9
figure9

ΔMEKK3:ER* induced G2 block is reversed by SB203580 irrespective of p21CIP1 status. (a) CM3.3 cells released from S phase were immediately treated with either 5 μM SB203580 (SB) or 10 μM U0126 (UO). Thirty min later 100 nM 4-HT was administered to these cells for the designated times and the cell cycle profiles were compared using PI staining and flow cytometry. (b) The number of cells in G2/M or G1 from (a) is quantified and shown graphically. (c) CM3.3 cells were left cycling (Cyc), blocked with aphidicolin (Aph) or released from aphidicolin for 20 h in the presence or absence of 4-HT, U0126 or SB203580 as indicated. Whole cell lysates were immunoblotted with antibodies specific for p21CIP1 or phospho-ERK1/2, ERK1/2, phospho-Ser63-cJun and phospho-p38. (d) CM3.3 cells were released from an aphidicolin block and left untreated (Con), or treated with 4-HT with or without SB203580. After 15 h cells were lysed and assayed for cyclin B-associated histone kinase activity. The data are the mean±s.d. of three independent determinations and the asterisk * indicates a significant difference by t-test (P=0.002). In addition, whole cell lysates were also immunoblotted for the presence cyclin B. (e) The effect of 5 μM SB203580 pre-treatment on the ΔMEKK3:ER* derived G2 arrest observed in RM3.21 cells was determined by quantifying the mitotic index

Western blot analysis, performed in parallel, revealed that SB203580 alone did not inhibit the ΔMEKK3:ER*-induced p21CIP1 expression, although the combination of U0126 and SB203580 resulted in a partial reduction (Figure 9c). Thus, there was no simple pharmacological correlation between the ΔMEKK3:ER*-induced G2 arrest (blocked by SB203580 alone) and p21CIP1 expression (not blocked by SB203580 alone). Western blot analysis of c-Jun phosphorylation (both S63 and S73) performed as a specificity control revealed that in contrast to previous reports (Whitmarsh et al., 1997; Le-Niculescu et al., 1999), SB203580 used in this study had no effect upon JNK activity.

Cumulatively, these results suggest that the ΔMEKK3:ER* mediated G2 cell cycle arrest is dependent upon the p38α/β2 isoforms (p38δ/γ are unaffected by SB203580 (Cuenda et al., 1995; Enslen et al., 1998)), can proceed in the absence of p21CIP1 induction and is associated with the down-regulation of cyclin A, cyclin B1 and their associated kinases.

Discussion

Activation of ΔMEKK3:ER* promotes a persistent G2 cell cycle arrest

Activation of ΔMEKK3:ER* in serum-starved cells resulted in the rapid, robust activation of the ERK, JNK and p38 pathways, indicating that ΔMEKK3:ER* allows conditional and titratable activation of the same MAPK pathways that full-length MEKK3 activates in transient assays (Deacon and Blank, 1997, 1999). In asynchronous, cycling cells, activation of ΔMEKK3:ER* resulted in a strong stimulation of JNK1 and p38α, with only a small increase in the basal level of ERK1 activity; this apparent disparity simply reflects the already high level of ERK activity in cycling CCl39 cells. Furthermore, ΔMEKK3:ER* activated JNK and p38 to levels similar to that observed with cellular stresses rather than causing a super-activation of these pathways. It is important to emphasize that we have used ΔMEKK3:ER* solely as a tool to activate a defined set of signalling pathways (predominantly JNK and p38 under these conditions) in the absence of any primary cellular stress or damage, and examine the impact of these pathways on cell cycle progression. Whether full-length MEKK3 is involved in cell cycle regulation is not known but the generation of MEKK3−/− mice (Yang et al., 2000) may help to address this in the future.

Activation of ΔMEKK3:ER* in cycling cells promoted a cell cycle arrest at G1 and G2 within 5–6 h; apoptosis was not apparent until after 20 h of ΔMEKK3:ER* activation. Thus, the initial effect of activating ΔMEKK3:ER* in asynchronous cells is cell cycle arrest rather than cell death. The delay in onset of apoptosis may simply reflect the time required for the up-regulation of pro-apoptotic genes and studies are currently underway to address this.

In mammalian systems, the accurate transmission of genetic information into daughter cells is regulated by the DNA damage checkpoint, acting at the end of G2 (reviewed in Elledge, 1996), the Chfr-dependent checkpoint at the end of prophase (Scolnick and Halazonetis, 2000) and the MAD/BUB-dependent spindle checkpoint at the end of metaphase (reviewed in Clarke and Giménez-Abián, 2000). Since activation of ΔMEKK3:ER* inhibited activation of CDK2 and CDK1 and prevented chromosome condensation in both CM3 and RM3 cells we believe that ΔMEKK3:ER* is triggering a G2 arrest analogous to the DNA damage checkpoint, thus precluding cells from entering mitosis. Whilst it has previously been shown that a related MEKK3-ER fusion can inhibit the initiation of DNA synthesis via the activation of ERK, JNK and p38 (Ellinger-Ziegelbauer et al., 1999; D Todd et al., unpublished observations), this is the first description of a MAPKKK being sufficient to prevent entry into mitosis. A previous study did not observe a significant G2 block upon activation of MEKK3-ER (Ellinger-Ziegelbauer et al., 1999) but those experiments were performed in quiescent cultures, which would arrest primarily in G1 (Ellinger-Ziegelbauer et al., 1999; D Todd et al., unpublished observations) and therefore not even reach G2.

Activation of ΔMEKK3:ER* inhibited CDK2 activity but didn't affect the normal progression of CM3 and RM3 cells through S phase. CDK2 is traditionally thought to be required for both S phase initiation and progression in mammalian cells (Girard et al., 1991; Ohtsubo et al., 1995, reviewed in Sherr, 1996). However, several recent reports have indicated that inhibition of cyclin A/CDK2 by the expression of CDKI's has no effect upon S phase once it is underway (Bates et al., 1998; Medama et al., 1998). It now seems that the primary role of cyclin A/CDK2 is to co-ordinate the orderly progression from DNA synthesis through to mitosis whilst ensuring that it only occurs once per cell cycle (Peterson et al., 1999).

The ΔMEKK3:ER*-induced G2 block requires p38α/β2 activity

In Xenopus oocytes the ERK pathway plays an essential role in both the meiotic G2/M transition and the metaphase arrest of mature oocytes (Haccard et al., 1993; Kosako et al., 1994). However, in both RM3 and CM3 cells U0126 had little effect on the ΔMEKK3:ER*-induced G2 arrest, suggesting that the ERK pathway plays a minor role in this case. In contrast, SB203580 was able to prevent the G2 block seen in CM3 cells (Figure 9a) and restored mitotic figures to control levels in 4-HT-treated RM3 cells (Figure 9c). Hence, activation of p38α and/or β2 is required to maintain the ΔMEKK3:ER*-induced G2 block. It has not been possible to address the role of JNK in this ΔMEKK3:ER*-induced G2 arrest due to the lack of suitable cell-permeant and selective inhibitors of the pathway. However, we have developed a similar conditional version of MEKK1 which selectively activates the JNK pathway and this can cause CCl39 cells to arrest at G2 (S Molton, manuscript submitted), suggesting that JNK activation can also impair the G2→M transition.

Recently, p38α/β2 have been implicated in control of the spindle assembly checkpoint in Xenopus oocytes (Takenaka et al., 1998), whilst the budding yeast p38 homologue, Hog1p, mediates a G2 delay in response to osmotic stress (Alexander et al., 2001). In addition, p38γ is required for a γ-radiation mediated G2 arrest in mammalian cells (Wang et al., 2000b). Our own results are more in accord with the latter study and suggest that activation of the different p38 family members in response to cellular stress may act to prevent cells from entering mitosis by preventing activation of cyclin B/CDK1. Since there is a significant overlap in the substrate specificity of the different p38 isoforms, the isoform responsible for the arrest may vary depending on the activating signal and the cellular isoform expression profile. Taken together with other reports our studies raise the possibility that p38 activation by cellular damage can trigger an evolutionarily conserved response, the ultimate outcome of which is to prevent cell division.

The ΔMEKK3:ER*-induced G2 arrest does not require p21CIP1

CDK2 kinase activity is primarily controlled by its association with cyclins E or A, and by inactivating CDKI's such as p21CIP1 (Harper et al., 1993; Xiong et al., 1993; Gu et al., 1993) and p27KIP1 (Polyak et al., 1994; Toyoshima and Hunter, 1994). p21CIP1 plays a central role in damage induced G1 arrest, primarily through the inhibition of G1 CDK activity (Harper et al., 1995; Poon et al., 1996; reviewed in Sherr and Roberts, 1995), but has also been implicated in promoting a G2 arrest in several recent studies by inhibition of CDK2 or CDK1 (Li et al., 1994; Guardagno and Newport, 1996; Bates et al., 1998; Dulic et al., 1998; Medama et al., 1998; Furono et al., 1999). Indeed, when ΔMEKK3:ER* was activated in CM3 cells we observed a striking increase in p21CIP1 levels (in the absence of any p53 accumulation) and the recruitment of p21CIP1 into cyclin A-containing complexes concomitant with the loss of CDK2 activity. In contrast, cyclin B1-containing complexes contained little or no associated p21CIP1, which is consistent with reports that p21CIP1 is principally targeted to cyclin A/CDK2 complexes (Harper et al., 1995) and is a poor inhibitor of CDK1 (Furono et al., 1999).

p21CIP1 expression can be induced by p53 and by a p53-independent pathway which can be activated by high levels of persistent ERK activation (Woods et al., 1997). In CM3 cells, doxorubicin failed to increase p53 or p21CIP1 levels and also failed to cause the characteristic DNA damage-induced G1 arrest, suggesting that p53 may be mutant in CCl39 cells. Despite this, ΔMEKK3:ER* caused a strong increase in p21CIP1 in the absence of any increase in p53 levels. The increase in p21CIP1 could be reduced by the combination of U0126 and SB203580, suggesting that the ERK and p38 pathways cooperate to promote p21CIP1 expression independently of p53 in response to activation of ΔMEKK3:ER*. Despite the ability of U0126 to reduce phospho-ERK to control levels, and the clear efficacy of SB203580 in preventing the G2 arrest, p21CIP1 expression was only partially inhibited by the combination of these drugs, suggesting that other pathways may also be important. Indeed, JNK can phosphorylate and activate Ets-2, c-Jun, JunB and ATF2, all of which have been proposed as regulators of the p21CIP1 promoter (Kardassis et al., 1999; Smith et al., 2000). Unfortunately the lack of selective inhibitors of either JNK or p38γ/δ prevents us from further investigating the relationship between p21CIP1 expression and SAPK activation at present.

Regardless of the mechanism for p21CIP1 up-regulation, several observations lead us to suggest that p21CIP1 is not absolutely required for the G2 arrest or even the inhibition of CDK2 and CDK1 observed upon activation of ΔMEKK3:ER*. First, both a sustained G2 arrest and loss of CDK activity were observed in RM3 cells, which fail to express p21CIP1 due to methylation of the p21CIP1 promoter (Allan et al., 2000). Second, even in CM3 cells, activation of ΔMEKK3:ER* could be delayed until just prior to mitosis and still elicit a G2 block, implying a more direct effect, perhaps independent of de novo induction of regulators such as p21CIP1. Finally, in CM3 cells the ΔMEKK3:ER*-induced G2 arrest was prevented by SB203580 alone, whereas p21CIP1 expression was not. Therefore, in agreement with others (Passlaris et al., 1999) we believe that cells possess multiple redundant mechanisms triggered by cellular stress or SAPK activation that are sufficient to arrest cells late in G2.

ΔMEKK3:ER* promotes down-regulation of cyclin A and cyclin B

An alternative, p21CIP1-independent, mechanism for SAPK-mediated CDK inhibition is suggested by the concomitant decrease in the expression of cyclin A and cyclin B1. The loss of these controlling subunits would certainly account for both the reduction in CDK activity and the lack of any CDK1 Y15 inhibitory phosphorylation, since Wee1/Myt1 can only phosphorylate this inhibitory site once the cyclin B/CDK1 complex is assembled (Parker et al., 1995; Booher et al., 1997). There are several reported precedents for this. Decreased levels of mitotic cyclins have previously been observed in response to treatment with ionizing radiation (Muschel et al., 1991), UVB (Petrocelli and Singerland, 2000), cis-platin (Spitkovsky et al., 1997), polycyclic aryl hydrocarbons (Guo et al., 2000) and UV-C (Poon et al., 1996) thus preventing mitotic initiation. Since JNK and p38 are also activated by these stresses we speculate that their activation may regulate the intracellular level of mitotic cyclins following cellular damage. Certainly SB203580 caused a partial rescue of cyclin B levels, suggesting that p38 is involved in the loss of cyclin B.

Cellular stresses can down-regulate mitotic cyclins through decreases in promoter activity (Innocente et al., 1999), decreased message stability (Maity et al., 1995; Wang et al., 2000a) and an increase in protein turnover (Guo et al., 2000) via the proteasome (Glotzer et al., 1991; King et al., 1996; Lukas et al., 1999). The ubiquitin ligase subunits, Cdc20 and Cdh1, are postulated to direct cyclins A and B respectively for ubiquitination, thereby flagging them for degradation by the proteasome (Dawson et al., 1995; Sigrist et al., 1995; Sudakin et al., 1995; Kotani et al., 1999). Whilst preliminary analysis has failed to show an increase in association between the proteasome subunit Cdc27 and Cdh1 upon activation of ΔMEKK3:ER* stimulation (data not shown), we cannot rule out the possibility that ΔMEKK3:ER* is subtly controlling the activity of this complex, or that other ubiquitin ligases are involved.

In conclusion, activation of ΔMEKK3:ER* serves as a useful tool with which to rapidly activate a defined set of signalling pathways, predominantly JNK and p38 in cycling cells, in the absence of any primary cellular stress. Activation of ΔMEKK3:ER* causes a sustained G2 arrest which requires the p38α/β2 pathway, but is apparently independent of p21CIP1 induction. It is hoped that ΔMEKK3:ER* and similar conditional mutants will prove to be useful tools in examining the role of the SAPK pathways in regulating the G2→M transition.

Materials and methods

Materials

Cell culture reagents were purchased from Gibco Life Technologies, SB203580 was from Calbiochem and U0126 was generously donated by James Trzaskos, DuPont Pharmaceuticals. [γ-32P]ATP was routinely purchased from Amersham. The following antibodies were used throughout this study: ERK1, JNK1, p38α from the Babraham Microchemical Facility; CDK2 (sc-163) from Santa Cruz; cyclin A (MS-384-P) and p53 (MS-104-P) from NeoMarkers; p21CIP1 (556431) from Pharmingen; CDK1 (9112) and phospho-CDK1 (9111S) from New England Biolabs; cyclin A (E72, used for immunoprecipitations) and cyclin B1 (V92) were kindly supplied by Tim Hunt, ICRF, London, UK. Horseradish peroxidase-conjugated secondary antibodies were from Bio-Rad. Unless indicated otherwise in the text, all other chemicals were purchased from Sigma and were of the highest grade available. UV treatment of cells at 254 nm was performed with a Stratalinker 2400 (Stratagene).

Plasmids

PCR oligonucleotides were used to amplify the catalytic domain of human MEKK3 (ΔMEKK3, amino acids 340–626) from a template plasmid, pCEVMEKK3CD-hERLBD, kindly provided by Ulrich Siebenlist. The 5′ oligonucleotide contained a start codon and Myc epitope tag sequence (MEQKLISEEDL), which was introduced in-frame upstream of the ΔMEKK3 sequence following PCR amplification. These oligonucleotides allowed for the introduction of a BamHI site at the 5′-end and an EcoRI site at the 3′-end of ΔMEKK3. The PCR product was digested and cloned into the respective sites within the expression vector pBP3:hbER*, which was kindly provided by Martin McMahon. The resulting construct, ΔMEKK3:ER*, encodes a conditional kinase comprising the catalytic domain of MEKK3 (ΔMEKK3) fused in-frame to a modified form of the hormone binding domain of the oestrogen receptor (ER*), which can be selectively de-repressed by 4-hydroxytamoxifen (4-HT). A catalytically inactive variant, pΔMEKK3:ER*KD, was constructed by replacing the conserved lysine residue in sub-domain II (required for ATP binding) with methionine (K391→M) using site-directed mutagenesis. The sequence of both the wild-type and catalytically inactive constructs was confirmed by sequencing the entire length of the construct on both strands. Details of sequence and oligos used are available upon request.

Cell culture and synchronization

The construction of CCl39 or Rat-1 cells expressing ΔMEKK3:ER* was performed by transfection with Qiagen superfect reagent, limiting dilution in 6 μg ml−1 puromycin and ring cloning. The resulting CM3 and RM3 clones were routinely cultured in phenol red-free Dulbecco's modified Eagle medium (DMEM) supplemented with 4.5 mg/ml glucose, 100 U ml−1 penicillin, 100 μg ml−1 streptomycin, 2 mM glutamine, 10% (v/v) foetal bovine serum (FBS) and 2 μg ml−1 puromycin.

Cells were synchronized in S phase by initially growing until confluent, replating on 10 cm dishes at a density of 1.25×105 cells/ml and allowing to attach for 6 h. The cells were then arrested in S phase by treatment with the DNA polymerase inhibitor aphidicolin (1 μg/ml) for 14 h. Cells were released back into cycle by washing four times with complete media. Unless indicated otherwise, cells were stimulated with either 100 nM 4-hydroxytamoxifen (4-HT) or ethanol (vehicle control) for the times indicated in the figure legends. Where additionally required, cells were treated with 5 μM SB203580 or 10 μM U0126 30 min prior to 4-HT treatment.

Analysis of cell cycle distribution and apoptosis

At the indicated times, cells were harvested by trypsinization, washed with phosphate buffered saline (PBS), and fixed in ice-cold 70% (v/v) ethanol/PBS for 30 min on ice. Following centrifugation the cell pellets were resuspended in PBS containing RNase A (0.1 mg/ml) and stained with propidium iodide (PI, 50 μg/ml) for 30 min at 37°C. Flow cytometry was performed on a FACSCalibur flow cytometer (Becton Dickinson).

Cell survival assays

Clonogenic survival was measured by transfecting plated CCl39 cells with either ΔMEKK3:ER* or ΔMEKK3:ER*KD using Superfect (Qiagen) according to the manufacturer's instructions. Twenty-four hours post-transfection the cell population was divided and replated. Cells were then selected for 2 weeks in puromycin media in the presence or absence of 4-HT (1 μM) before being fixed with acetic acid : methanol (1 : 3, v/v) and stained with 0.1% (w/v) crystal violet.

Cellular proliferation was measured by plating the different CM3 clones in 24-well plates at a density of 1×104 cells/ml. Following attachment, the cells were treated with increasing doses of 4-HT and the amount of growth assessed 72 h later. Relative cell numbers were quantified by staining protein with crystal violet as described above and solubilizing the resulting stained cells with 10% (v/v) acetic acid. Absorbances were then read at 590 nm on the spectrophotometer.

Immunecomplex MAPK/SAPK assays

At the required times, cells were washed with PBS and harvested in lysis buffer (20 mM Tris [pH 7.5], 137 mM NaCl, 1 mM EGTA, 1% (v/v) Triton X-100, 10% (v/v) glycerol, 1.5 mM MgCl2, 1 mM Na3VO4, 1 mM PMSF, 20 μM leupeptin, 10 μg/ml aprotinin and 50 mM NaF). Cell extracts were snap frozen, thawed and cleared by centrifugation. Supernatant protein concentration was measured by Bradford protein assay (Bio-Rad) and active kinase complexes were immunoprecipitated using ERK1, JNK1 and p38α antibodies. Subsequently, the immune complexes were bound to protein A-sepharose and the beads washed twice with lysis buffer. ERK1 samples were washed once in ERK buffer (30 mM Tris [pH 8], 20 mM MgCl2 and 2 mM MnCl2) whilst the JNK1 and p38α samples were washed with (20 mM HEPES [pH 7.5], 20 mM β-glycerophosphate, 10 mM MgCl2, 1 mM DTT and 50 μM Na3VO4). Kinase activities were assessed by incubating the aspirated beads in 30 μl of their respective kinase buffers supplemented with 10 μM ATP, 3 μCi [γ-32P]ATP and either 2.5 μg GST-c-Jun (JNK substrate) or 2 μg His6-MAPKAP-K2 (p38α substrate) or 7 μg myelin basic protein (ERK substrate) for 30 min at 30°C. Reactions were terminated by boiling the samples in SDS–PAGE sample buffer before the samples were resolved by SDS–PAGE. Incorporation of [32P] into the respective substrates was quantified by PhosphorImager (Bio-Rad).

Immunecomplex CDK assays

To ascertain the kinase activity associated with either cyclin A or cyclin B, cells were washed and harvested in CDK lysis buffer (50 mM Tris [pH 7.5], 150 mM NaCl, 20 mM EGTA, 50 mM NaF, 1 mM Na3VO4, 0.1% (v/v) NP40, 1 mM PMSF, 20 μM leupeptin and 10 μg/ml aprotinin). Supernatants were extracted and the CDK complexes immunoprecipitated using cyclin A (E72) and cyclin B1 (V92) antibodies and protein G-sepharose. The complexes were washed twice with CDK lysis buffer containing 500 mM NaCl and twice in β-glycerophosphate buffer (80 mM β-glycerophosphate, 20 mM EGTA and 15 mM MgCl2). CDK activity was measured by incubating the beads with 10 μl of β-glycerophosphate buffer supplemented with 100 μM ATP, 0.3 μCi of [γ-32P]ATP and 1 μg of Histone H1 (Calbiochem) for 15 min at 37°C. The reactions were terminated, resolved by SDS–PAGE and the amount of phosphorylation quantified by PhosphorImager.

Western blotting

At the appropriate time points, cells were washed and harvested as described earlier (see Immunecomplex MAPK/SAPK assays). Equal quantities of cell extracts were resolved by SDS–PAGE and transferred onto Immobilon P membranes (Millipore). Filters were then blocked for at least 1 h in 0.1% (v/v) Tween-20/PBS containing either 5% (w/v) powdered milk or bovine serum albumin (BSA) before being probed with the desired antibodies (see Materials) diluted in the same buffer at the recommended concentrations. Immunoreactive proteins were visualized with the enhanced chemiluminescence (ECL) system (Amersham). Where reprobing of the same filter was necessary (e.g. when determining the phosphorylation status of a protein) the filter was incubated in strip buffer (62.5 mM Tris [pH 6.8], 2% (w/v) SDS, 100 mM β-mercaptoethanol) for 30 min at 50°C before being washed and blocked as described.

Mitotic index assay

Seven hours post-release from an aphidicolin induced block (see Cell culture and synchronization) CM3 and RM3 cells were stimulated with 100 nM 4-HT. Three hours after 4-HT treatment, cells undergoing mitosis were trapped by the addition of 250 ng/ml nocodazole for 14 h. At the appropriate time, cells were trypsinized and resuspended in 0.075 M KCl. Following a 30 min incubation at room temperature the cells were fixed in ice-cold Carnoys solution (acetic acid : methanol (1 : 3, v/v)) before staining with 0.02% (v/v) Geimsa. Geimsa stained chromosomes were inspected by light microscopy and the mitotic index scored.

References

  1. Alexander MR, Tyers M, Perret M, Craig BM, Fang KM, Gustin MC . 2001 Mol. Biol. Cell 12: 53–62

  2. Allan LA, Duhig T, Read M, Fried M . 2000 Mol. Cell. Biol. 20: 1291–1298

  3. Balmanno K, Cook SJ . 1999 Oncogene 18: 3085–3097

  4. Bates S, Ryan KM, Phillips AC, Vousden KH . 1998 Oncogene 17: 1691–1703

  5. Blasina A, Paegle ES, McGowan CH . 1997 Mol. Biol. Cell 8: 1013–1023

  6. Booher RN, Holman PS, Fattaey A . 1997 J. Biol. Chem. 272: 22300–22306

  7. Brown AL, Lee CH, Schwarz JK, Mitiku N, Piwnica-Worms H, Chung JH . 1999 Proc. Natl. Acad. Sci. USA 96: 3745–3750

  8. Brugarolas J, Chandrasekaran C, Gordon JC, Beach D, Jacks T, Hannon GJ . 1995 Nature 377: 552–556

  9. Bunz F, Dutriaux A, Lengauer C, Waldman T, Zhou S, Brown JP, Sedivy JM, Kinzler KW, Vogelstein B . 1998 Science 282: 1497–1501

  10. Clarke DJ, Giménez-Abián JF . 2000 BioEssays 22: 351–363

  11. Cook SJ, Balmanno K, Garner A, Millar T, Taverner C, Todd D . 2000 Biochem. Soc. Trans. 28: 233–240

  12. Cuenda A, Rouse J, Doza YN, Meier R, Cohen P, Gallagher TF, Young PR, Lee JC . 1995 FEBS Lett. 364: 229–233

  13. Davis RJ . 2000 Cell 103: 239–252

  14. Deacon K, Blank JL . 1997 J. Biol. Chem. 272: 14489–14496

  15. Deacon K, Blank JL . 1999 J. Biol. Chem. 274: 16604–16610

  16. Dawson I, Roth AS, Artavanis-Tsakonas S . 1995 J. Cell Biol. 129: 725–737

  17. Deng C, Zhang P, Harper JW, Elledge SJ, Leder P . 1995 Cell 82: 675–684

  18. Dulic V, Stein GH, Far DF, Reed SI . 1998 Mol. Cell. Biol. 546: 546–557

  19. El-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons R, Trent JM, Lin D, Mercer WE, Kinzler KW, Vogelstein B . 1993 Cell 75: 817–825

  20. Elledge SJ . 1996 Science 274: 1664–1672

  21. Ellinger-Ziegelbauer H, Kelly K, Siebenlist U . 1999 Mol. Cell. Biol. 19: 3857–3868

  22. Enslen H, Raingeaud J, Davis RJ . 1998 J. Biol. Chem. 273: 1741–1748

  23. Favata MF, Horiuchi KY, Manos EJ, Daulerio AJ, Stradley DA, Feeser WS, Van Dyk DE, Pitts WJ, Earl RA, Hobbs F, Copeland RA, Magolda RL, Scherle PA, Trzaskos JM . 1998 J. Biol. Chem. 273: 18623–18632

  24. Fesquet D, Labbe JC, Derancourt J, Capony JP, Galas S, Girard F, Lorca T, Shuttleworth J, Doree M, Cavadore JC . 1993 EMBO J. 12: 3111–3121

  25. Furono N, den Elzen N, Pines J . 1999 J. Cell Biol. 147: 295–306

  26. Girard F, Strausfeld U, Fernandez U, Lamb NJ . 1991 Cell 67: 1169–1179

  27. Glotzer M, Murray AW, Kirschner MW . 1991 Nature 349: 132–138

  28. Gould KL, Nurse P . 1989 Nature 342: 39–42

  29. Gu Y, Turck CW, Morgan DO . 1993 Nature 366: 707–710

  30. Guardagno TM, Newport JW . 1996 Cell 84: 73–82

  31. Guo N, Faller DV, Vazir C . 2000 J. Biol. Chem. 275: 1715–1722

  32. Haccard O, Sarcevic B, Lewellyn A, Hartley R, Roy L, Izumi T, Erikson E, Maller JL . 1993 Science 262: 1262–1265

  33. Hagting A, Jackman M, Simpson K, Pines J . 1999 Curr. Biol. 9: 680–689

  34. Hall M, Peters G . 1996 Adv. Cancer Res. 68: 67–108

  35. Harper JW, Elledge SJ, Keyomarsi K, Dynlacht B, Tsai LH, Zhang P, Dobrowolski S, Bai C, Connel-Crowley CL, Swindell E, Fox MP, Wei N . 1995 Mol. Biol. Cell 6: 387–400

  36. Harper J, Adami WGR, Wei N, Keyomarsi K, Elledge SJ . 1993 Cell 75: 805–816

  37. Hiebert SW, Chellappan SP, Horowitz JM, Nevins JR . 1992 Genes Dev. 6: 177–185

  38. Hollstein M, Rice K, Greenblatt MS, Soussi T, Fuchs R, Sorlie T, Hovig E, Smith-Sorensen B, Montesano R, Harris CC . 1994 Nuc. Acids Res. 22: 3551–3555

  39. Innocente SA, Abrahamson JLA, Cogswell JP, Lee JM . 1999 Proc. Natl. Acad. Sci. USA 96: 2147–2152

  40. Jin P, Gu Y, Morgan DO . 1996 J. Cell Biol. 134: 963–970

  41. Kardassis D, Papakosta P, Pardali K, Moustakas A . 1999 J. Biol. Chem. 274: 29572–29581

  42. King RW, Deshaies RJ, Peters RM, Kirschner MW . 1996 Science 274: 1652–1659

  43. Knudsen ES, Wang JYJ . 1997 Mol. Cell. Biol. 17: 5771–5783

  44. Kosako H, Gotoh Y, Nishida E . 1994 EMBO J. 13: 2131–2138

  45. Kotani S, Tanaka H, Yasuda H, Todokoro K . 1999 J. Cell Biol. 146: 791–800

  46. Labbe JC, Capony JP, Caput D, Cavadore JC, Derancourt J, Kaghad M, Lelias JM, Picard A, Doree M . 1989 EMBO J. 8: 3053–3058

  47. Le-Niculescu H, Bonfoco E, Kasuya Y, Claret FX, Green DR, Karin M . 1999 Mol. Cell. Biol. 19: 751–763

  48. Li J, Meyer AN, Donoghue DJ . 1995 Mol. Biol. Cell 6: 1111–1124

  49. Li Y, Jenkins CW, Nichols MA, Xiong Y . 1994 Oncogene 9: 2261–2268

  50. Liu Q, Guntuku S, Cui XS, Matsuoka S, Cortez D, Tamai K, Luo G, Carattini-Rivera S, DeMayo F, Bradley A, Donehower LA, Elledge SJ . 2000 Genes Dev. 14: 1448–1459

  51. Lopez-Girona A, Furnari B, Mondesert O, Russell P . 1999 Nature 397: 172–175

  52. Lukas C, Storgaard Sørensen C, Kramer E, Santoni-Rugiu E, Lindeneg J, Peters J-M, Bartek J, Lukas J . 1999 Nature 401: 815–818

  53. Macleod KF, Sherry N, Hannon G, Beach D, Tokino T, Kinzler K, Vogelstein B, Jacks T . 1995 Genes Dev. 9: 935–944

  54. Maity A, McKenna WG, Muschel RJ . 1995 EMBO J. 14: 603–609

  55. Matsuoka S, Huang M, Elledge SJ . 1998 Science 282: 1893–1897

  56. Medama RH, Klompmaker R, Smits VAJ, Rijsken G . 1998 Oncogene 16: 431–441

  57. Meijer J, Azzi L, Wang JY . 1991 EMBO J. 10: 1545–1554

  58. Mueller PR, Coleman TR, Kumagai A, Dunphy WG . 1995 Science 270: 86–90

  59. Muschel RJ, Zhang HB, Iliakis G, McKenna WG . 1991 Cancer Res. 51: 5113–5117

  60. Nebreda AR, Porras A . 2000 Trends Biochem. Sci. 25: 257–260

  61. Ohtsubo M, Theodoras AM, Schumacher J, Roberts JM, Pagano M . 1995 Mol. Cell. Biol. 15: 2612–2624

  62. Passlaris TM, Benanti JA, Gewin L, Kiyono T, Galloway DA . 1999 Mol. Cell. Biol. 19: 5872–5881

  63. Parker LL, Sylvestre PJ, Byrnes MJ, Liu F, Piwnica-Worms H . 1995 Proc. Natl. Acad. Sci. USA 92: 9638–9642

  64. Peterson BO, Lukas J, Storgaard Sørensen C, Bartek J, Helin K . 1999 EMBO J. 18: 396–410

  65. Petrocelli T, Singerland J . 2000 Oncogene 19: 4480–4490

  66. Pines J, Hunter T . 1989 Cell 58: 833–846

  67. Piwnica-Worms H . 1999 Nature 401: 535–537

  68. Polyak K, Lee MH, Erdjument-Bromage H, Koff A, Roberts JM, Tempst P, Massagué J . 1994 Cell 78: 59–66

  69. Poon RYC, Jiang W, Toyoshima H, Hunter T . 1996 J. Biol. Chem. 271: 13283–13291

  70. Russell P, Nurse P . 1987 Cell 49: 559–567

  71. Samuels ML, Weber MJ, Bishop JM, McMahon M . 1993 Mol. Cell. Biol. 13: 6241–6252

  72. Scolnick DM, Halazonetis TD . 2000 Nature 406: 430–435

  73. Sherr CJ . 1996 Science 274: 1672–1677

  74. Sherr CJ, Roberts JM . 1995 Genes Dev. 9: 1149–1163

  75. Sigrist S, Jacobs H, Stratmann R, Lehner CF . 1995 EMBO J. 14: 4827–4838

  76. Smith JL, Schaffner AE, Hofmeister JK, Hartman M, Wei G, Forsthoefel D, Hume DA, Ostrowski MC . 2000 Mol. Cell. Biol. 20: 8026–8034

  77. Solomon MJ, Lee T, Kirschner MW . 1992 Mol. Biol. Cell 3: 13–27

  78. Spitkovsky D, Schulze A, Boye B, Jansen-Durr P . 1997 Cell Growth Differ. 8: 699–710

  79. Sudakin V, Ganoth D, Dahan A, Heller H, Hershko J, Luca FC, Ruderman JV, Hershko A . 1995 Mol. Biol. Cell 6: 185–198

  80. Takenaka K, Moriguchi T, Nishida E . 1998 Science 280: 599–602

  81. Toyoshima H, Hunter T . 1994 Cell 78: 67–74

  82. Waldmann T, Kinzler KW, Vogelstein B . 1995 Cancer Res. 55: 5187–5190

  83. Wang S, Nath N, Minden A, Chellappan S . 1999 EMBO J. 18: 1559–1570

  84. Wang W, Caldwell MC, Lin S, Furneaux H, Gorospe M . 2000a EMBO J. 19: 2340–2350

  85. Wang X, McGowan CH, Zhao M, He L, Downey JS, Fearns C, Wang Y, Huang S, Han J . 2000b Mol. Cell. Biol. 20: 4543–4552

  86. Whitmarsh AJ, Yang SH, Su MS-S, Sharrocks AD, Davis RJ . 1997 Mol. Cell. Biol. 17: 2360–2371

  87. Wisdom R, Johnson RS, Moore C . 1999 EMBO J. 18: 188–197

  88. Woods D, Parry D, Cherwinski H, Bosch E, Lees E, McMahon M . 1997 Mol. Cell. Biol. 17: 5598–5611

  89. Xiong Y, Hannon GJ, Zhang H, Casso D, Kobayashi R, Beach D . 1993 Nature 366: 701–704

  90. Yang J, Winkler K, Yoshida M, Kornbluth S . 1999 EMBO J. 18: 2174–2183

  91. Yang J, Boerm M, McCarty M, Bucana C, Fidler IJ, Zhuang Y, Su B . 2000 Nat. Genet. 24: 309–313

Download references

Acknowledgements

We would like to thank Geoff Morgan for maintenance of the Babraham Institute Flow Cytometry Facility and for his technical advice, Adele Murrell for advice on counting mitotic indices and members of the Cook group for interesting discussions and comments. We are grateful to Tim Hunt for the generous provision of cyclin A and B antibodies, Jiri Lukas for Cdh1 antibodies, James Trzaskos (DuPont) for providing U0126 and Bert Vogelstein for providing WT and p53−/− HCT116 cells. This work was supported by Cancer Research UK (SP2458/0201), the BBSRC Science of Ageing Initiative (202/SAG10012) and a Competitive Strategic Grant from the BBSRC. CR Weston was funded by a MRC PhD studentship and SJ Cook is a Senior Cancer Research Fellow of Cancer Research UK.

Author information

Correspondence to Simon J Cook.

Rights and permissions

Reprints and Permissions

About this article

Keywords

  • MEKK3
  • p38
  • G2
  • cyclin
  • CDK
  • p21CIP1

Further reading