Report

Requirement for MAPK Activation for Normal Mitotic Progression in Xenopus Egg Extracts

See allHide authors and affiliations

Science  13 Nov 1998:
Vol. 282, Issue 5392, pp. 1312-1315
DOI: 10.1126/science.282.5392.1312

Abstract

The p42 mitogen-activated protein kinase (MAPK) is required for progression through meiotic M phase in Xenopus oocytes. This report examines whether it also plays a role in normal mitotic progression. MAPK was transiently activated during mitosis in cyclingXenopus egg extracts after activation of the cyclin-dependent kinase Cdc2–cyclin B. Interference with MAPK activation by immunodepletion of its activator MEK, or by addition of the MEK inhibitor PD98059, caused precocious termination of mitosis and interfered with production of normal mitotic microtubules. Sustained activation of MAPK arrested extracts in mitosis in the absence of active Cdc2–cyclin B. These findings identify a role for MEK and MAPK in maintaining the mitotic state.

Mitosis is initiated by the activation of Cdc2–cyclin complexes. In Xenopus egg extracts, three mitotic Cdc2–cyclin complexes have been identified; they are activated and inactivated sequentially, beginning with Cdc2–cyclin A1, followed by Cdc2–cyclin B1, and finally Cdc2–cyclin B2 (1). The last of the three, Cdc2–cyclin B2, is inactivated just after nuclear envelope breakdown (NEBD) (1). Chromatin condensation and NEBD persist throughout the remainder of M phase in the absence of active Cdc2. This persistence could be the result of slow reversal of the effects of Cdc2, or these aspects of mitosis could be actively maintained by some regulatory protein other than Cdc2.

Several lines of evidence raise the possibility that p42 MAPK (also called ERK2) participates in mitosis. MAPK activation is required forXenopus oocyte maturation, and the regulation of oocyte maturation and is similar to regulation of mitosis in many important respects (2). Moreover, in sea urchin embryos (3), mammalian cell lines (4), and cyclingXenopus egg extracts (Figs. 1C; 2C; and 3, A and C) (5, 6), MAPKs are activated during mitosis. Finally, MAPKs have been implicated in the spindle assembly checkpoint in extracts and in aXenopus cell line (XTC-2) (5–7), and there is precedent for proteins involved in this checkpoint to be involved in establishing the timing of an unperturbed mitosis (8). However, depletion of p42 MAPK or inhibition of p42 MAPK activation has no effect on the activation or inactivation of Cdc2 in cycling Xenopus egg extracts, which suggests that p42 MAPK might be dispensable for mitotic entry and exit (5).

We examined the role of p42 MAPK in mitosis in cycling extracts, monitoring not only Cdc2 activation and inactivation but also the main morphological hallmarks of mitosis—nuclear envelope breakdown, chromatin condensation, and microtubule dynamics. We prevented mitotic activation of p42 MAPK in cycling extracts by one of two treatments that inhibit MEK, the protein kinase that phosphorylates and activates MAPK: addition of the MEK inhibitor PD98059 (9) or immunodepletion of MEK (10). Both approaches blocked p42 MAPK activation (Fig. 1, A and C). We then tested whether Cdc2 activity cycled normally in the absence of MAPK activation. Activation and inactivation of Cdc2 was similar in control and in PD98059-treated extracts and in mock-depleted and MEK-depleted extracts (Fig. 1, B and D), in agreement with previous reports (5).

Figure 1

Normal activation and inactivation of Cdc2 in extracts after inhibition or immunodepletion of MEK. (A andB) Effects of MEK inhibitor PD98059 on MAPK activation and Cdc2 activity. Extracts were treated with PD98059 or DMSO (8). (C and D) Effects of MEK immunodepletion on MAPK activation and Cdc2 activity. Extracts were depleted of MEK with antibody 662 or were mock-depleted with rabbit IgG (8) and warmed to room temperature to initiate cycling. The resulting MAPK activity was assessed by immune complex kinase assay (A and C), and Cdc2 activity was assessed with histone H1 as substrate (B and D). Indicated times are relative to initiation of cycling.

We also tested whether nuclear envelope breakdown and re-formation and chromatin condensation and decondensation were altered in the absence of MAPK activation (11). We added low concentrations of demembranated sperm (500 per microliter) (12) to MAPK-inhibited and control cycling extracts, allowed nuclei to form, and took portions at various times to assess chromatin condensation by 4′,6-diamidino-2-phenylindole (DAPI) staining and NEBD by phase-contrast microscopy. Both the control and PD98059-treated extracts underwent chromatin condensation and nuclear envelope breakdown 50 min after cycling was initiated (Fig. 2, A and B). However, the extracts in which MEK was inhibited exited mitosis prematurely (Fig. 2, A and B). The chromatin had decondensed and nuclear envelopes re-formed by 60 min in the PD98059-treated extract but not until 75 min in the control extract. Thus the duration of mitosis (taken here to be the interval between NEBD and re-formation) in the MEK-inhibited extract was less than half that in the control extract. Premature mitotic exit was also observed with MEK-depleted extracts but not with mock-depleted extracts (Fig. 2C), and adding purified recombinant MEK to the MEK-depleted extracts restored mitosis to a normal length (Fig. 2D). These results indicate that MEK activation is necessary to maintain the mitotic state for a normal period of time. Because p42 and p44 MAPK are the only known substrates of MEK, and only p42 is present in egg extracts, these findings implicate p42 MAPK in maintenance of the mitotic state.

Figure 2

Premature mitotic exit after inhibition of MEK. (A) Nuclear morphology in control and PD98059-treated extracts. Nuclei were stained with DAPI and observed by fluorescence microscopy. Scale bar, 20 μm. (B) NEBD in control DMSO-treated extracts and PD98059-treated extracts. (C) NEBD mock-depleted and MEK-depleted extracts. (D) Restoration of normal mitotic duration in MEK-depleted extracts supplemented with recombinant MEK. Data are averages from two (add back) or three (mock-depleted, MEK-depleted) experiments. Error bar represents one standard deviation for the MEK-depleted extracts. For other extracts the duration of mitosis did not vary measurably from experiment to experiment. NEBD was assessed by phase-contrast microscopy. In (A) through (C), indicated times are relative to initiation of cycling.

When MAPK is artificially activated before Cdc2 is activated, it can inhibit cyclin degradation and hence prolong Cdc2 activation and mitosis (13). However, the present results (Figs. 1 and 2) suggest that p42 MAPK may be able to sustain mitosis even after Cdc2 is inactivated. To test this idea further, we established conditions for inducing sustained activation of p42 MAPK in a cycling extract at about the time when transient MAPK activation normally occurs. We activated MEK and MAPK in the extract by adding recombinant malE-Mos (14), a MEK kinase, and timed the Mos addition to be too late to prevent cyclin destruction (Fig. 3, A and B).

Figure 3

Effect of prolonged mitotic activation of MAPK on Cdc2 activity and mitotic exit. Recombinant malE-Mos (a Mos–maltose-binding protein fusion protein; 300 nM final concentration) or buffer was added to cycling Xenopus egg extracts 40 min after initiation of cycling. (A) MAPK immunoblot. Lower band represents nonphosphorylated MAPK. Upper band represents phosphorylated MAPK. (B) Cdc2 H1 kinase activity. (C) NEBD assessed by phase-contrast microscopy. (D) Chromatin condensation, assessed by DAPI staining and fluorescence microscopy. Scale bar, 20 μm. Indicated times are relative to initiation of cycling.

Under these circumstances Mos treatment had no effect on the timing of Cdc2 activation (Fig. 3B), NEBD (Fig. 3C), or chromatin condensation (Fig. 3D), all of which occurred by 60 min in both control and Mos-treated cycling extracts. Moreover, Cdc2 inactivation occurred normally in both control and Mos-treated extracts (Fig. 3B). The inactivation of Cdc2 coincided with mitotic cyclin degradation as determined by [35S]methionine labeling, gel electrophoresis, and autoradiography (15).

Control extracts returned to interphase by 80 min, as indicated by decondensed chromatin and intact nuclear envelopes (Fig. 3, C and D). In contrast, Mos-treated extracts remained arrested in mitosis for the duration of the experiment despite their lack of Cdc2 activity (Fig. 3, C and D). Thus, active p42 MAPK can maintain the mitotic state in the absence of active Cdc2–cyclin B. This finding implies a role for p42 MAPK in maintaining mitosis and indicates that inactivation of p42 MAPK may be required for exit from mitosis.

The p42 MAPK associates with microtubules (16) and in particular with the mitotic spindle (3, 4, 7). We therefore examined whether microtubule dynamics were altered in cycling extracts in which MEK was inhibited. We prepared cycloheximide-treated interphase egg extracts devoid of cyclins A1, B1, and B2 (17) and mock-depleted them, immunodepleted them of MEK, or immunodepleted them and then added back purified recombinant MEK R4F (an activated form of human MEK) (18) to yield a physiological concentration of MEK (Fig. 4A). We then drove the extracts into a stable mitotic state by addition of purified nondegradable sea urchin cyclin B. This yielded large amounts of Cdc2–cyclin B activity in all three types of extracts (Fig. 4B). The active Cdc2–cyclin B brought about activation of p42 MAPK in the mock-depleted extract (Fig. 4B, lane 2), demonstrating that p42 MAPK can be activated downstream of Cdc2–cyclin B in this system. MEK depletion blocked Cdc2–cyclin B-induced activation of p42 MAPK, and adding back recombinant MEK R4F to the MEK-depleted extract restored activation of p42 MAPK (Fig. 4B).

Figure 4

Effect of MEK inhibition on aster formation. Mock-depleted or MEK-depleted interphase extracts were prepared from electrically activated cycloheximide-treated eggs and were treated with nondestructible sea urchin Δ90 cyclin B (to drive them into a stable mitotic state) or with no cyclin (to keep them in interphase). (A) Amounts of MEK in a mock-depleted extract, a MEK-depleted extract, and a MEK-depleted extract supplemented with human MEK R4F. Immunoblot with an antiserum (662) raised against a peptide that is identical in the Xenopus and human MEK proteins. (B) Cdc2 H1 kinase activity (upper autoradiogram) and MAPK immune complex kinase activity (lower autoradiogram) in an interphase extract (lane 1); a mock-depleted, mitotic extract (lane 2); a MEK-depleted, mitotic extract (lane 3); and a MEK-depleted mitotic extract supplemented with MEK R4F (lane 4). (C) Sperm-nucleated aster formation in mitotic Δ90 cyclin B–treated extracts with or without normal MAPK activity. A mock-depleted extract (left), MEK-depleted extract (center), and MEK-depleted extract supplemented with MEK R4F (right) are shown. DAPI staining of the sperm is shown in blue; rhodamine-labeled tubulin is shown in red. Scale bar, 20 μm. (D) Long interphase-like microtubules in interphase extracts with normal low MAPK activity (left) or high MAPK activity induced by Mos (right). Scale bar, 20 μm.

We added rhodamine-labeled bovine brain tubulin and demembranatedXenopus sperm to the three cyclin-treated extracts and assessed the size of the microtubule asters nucleated from the sperm centrioles (19). Larger microtubule asters were present in the MEK-depleted mitotic extracts than in the mock-depleted extract, despite the fact that both extracts had high levels of Cdc2 activity (Fig. 4C). Mitotic extracts treated with the MEK inhibitor PD98059 also produced large asters (15). The addition of purified recombinant MEK protein to MEK-depleted extracts restored production of normal small mitotic asters (Fig. 4C). Taken together, these data indicate that p42 MAPK activity is required for normal mitotic microtubules.

To determine whether MAPK activation was sufficient to produce mitotic microtubule asters in the absence of Cdc2 activity, we compared microtubule asters in interphase extracts and Mos-treated interphase extracts (which contain high p42 MAPK activity and low Cdc2 activity). Both extracts supported the formation of large asters (Fig. 4D), and extensive networks of microtubules could be seen in the absence of sperm in both extracts. Thus MAPK activation is not sufficient to initiate mitotic microtubule dynamics. Evidently, both Cdc2 and MAPK function in initiating or maintaining mitotic microtubules (20).

Our results demonstrate that p42 MAPK is activated downstream of Cdc2 during mitosis in Xenopus egg extracts, that the mitotic state is actively maintained during the period after Cdc2 inactivation, and that p42 MAPK activity is essential for this maintenance. During meiosis in oocytes, p42 MAPK has a similar function, suppressing a return to interphase during the interval between meiosis I and meiosis II (21, 22). Thus this versatile, evolutionarily ancient protein kinase, implicated in diverse responses to extracellular signals, is also an important component of the cell cycle clock.

  • * To whom correspondence should be addressed. E-mail: ferrell{at}cmgm.stanford.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article