The Protein Kinase p90 Rsk as an Essential Mediator of Cytostatic Factor Activity

See allHide authors and affiliations

Science  12 Nov 1999:
Vol. 286, Issue 5443, pp. 1362-1365
DOI: 10.1126/science.286.5443.1362


Persistent activation of p42 mitogen-activated protein kinase (p42 MAPK) during mitosis induces a “cytostatic factor” arrest, the arrest responsible for preventing the parthenogenetic activation of unfertilized eggs. The protein kinase p90 Rsk is a substrate of p42 MAPK; thus, the role of p90 Rsk in p42 MAPK-induced mitotic arrest was examined. Xenopus laevis egg extracts immunodepleted of Rsk lost their capacity to undergo mitotic arrest in response to activation of the Mos–MEK-1–p42 MAPK cascade of protein kinases. Replenishing Rsk-depleted extracts with catalytically competent Rsk protein restored the ability of the extracts to undergo mitotic arrest. Rsk appears to be essential for cytostatic factor arrest.

Masui identified two hypothetical M-phase regulators in his classic studies of Rana pipiensoocyte maturation. The first, maturation-promoting factor (MPF), was an activity present in mature oocytes that was able to induce immature oocytes to mature even in the absence of protein synthesis (1). MPF ultimately proved to be a complex of the universal M-phase regulators Cdc2 and cyclin B (2). Cytostatic factor (CSF) was defined as an activity present in mature oocytes that induced mitotic arrest when injected into cleaving embryos (1). The underlying hypothesis was that CSF activity is responsible for the maintenance of mature oocytes in their normal metaphase arrest state.

Studies over the past decade have identified the proto-oncoprotein Mos as CSF and the protein kinases MEK and p42 MAPK as essential mediators of CSF activity (3). The introduction of Mos mRNA (4) or protein (5), constitutively active MEK (3), or thiophosphorylated, active p42 MAPK (6) into Xenopus laevis embryos or cell-free cycling extracts (7, 8) causes a metaphase arrest. Depletion of Mos from extracts of mature oocytes depletes the extracts' CSF activity (4). In addition, injection of a neutralizing MEK antibody prevents Mos from causing a CSF arrest (9), as do a pharmacological inhibitor of MEK activation (10) and the CL100 MAPK phosphatase (7). Moreover, in Mos-deficient mice oocytes do not arrest properly in metaphase of meiosis II and therefore do not develop into normal fertilizable eggs (11). These findings demonstrate the importance of CSF and Mos in normal reproduction. The Mos-MEK-p42 MAPK cascade arrests embryos in mitosis at least in part by preventing the destruction of B-type cyclins (3), although the cascade also functions to maintain the mitotic state even after cyclins have been degraded (12, 13).

Members of the p90 Rsk family of protein kinases are activated by p42 MAPKs in diverse biological contexts and may be important MAPK effectors (14). Rsks have been implicated in MAPK-induced transcriptional changes (15), and mutations in the human Rsk-2 gene are associated with Coffin-Lowry syndrome, an X-linked syndrome characterized by severe psychomotor retardation and facial and digital dysmorphisms (16).

Two closely related Rsk isoforms, Rsk-1 and Rsk-2, are present and active in CSF-arrested Xenopus eggs (17, 18). We initially focused on Rsk-2 as a possible mediator of CSF arrest because it is the more abundant isoform (18). We chose cycling Xenopus egg extracts (19) for our studies. Cycling extracts respond to the properly timed addition of recombinant Mos or activated MEK by undergoing a classical CSF arrest, with high Cdc2 activity, dissolved nuclear envelopes, and condensed chromatin aligned on metaphase spindles (7, 8,20). Moreover, extracts offer the possibility of abolishing Rsk function by immunodepletion and restoring Rsk function by addition of recombinant Rsk protein.

We determined whether Rsk-2, like p42 MAPK (12, 20), was activated during mitosis in cycling extracts. Rsk-2 was transiently activated (Fig. 1A) concomitantly with p42 MAPK (Fig. 1B) (12). Like p42 MAPK activation, Rsk-2 activation lagged behind that of Cdc2 and preceded mitotic exit (Fig. 1, A and B). Cycling extracts responded to an appropriately timed addition of recombinant Mos (21) by entering a CSF arrest state with active Cdc2, condensed chromatin, and disassembled nuclear envelopes. This CSF arrest was accompanied by increased phosphorylation of Rsk-2 as indicated by a decreased electrophoretic mobility (Fig. 1C), and activation of the enzyme as assessed by immune complex kinase assays (22). Similar results were found for Rsk-1 (22). Thus, Rsks are potential mediators of CSF activity.

Figure 1

Mitotic activation of Rsk-2. (A) Cell cycle–regulated activation of Rsk-2 and Cdc2 in cycling extracts. A cycling extract was prepared (19) and warmed to room temperature to initiate cycling. Demembranated sperm chromatin (500 sperm per microliter) was added to allow monitoring of progression into and out of mitosis. The times indicated are times after warming. Rsk-2 activity was assessed by immune complex kinase assay with S6 peptide as the phosphoacceptor. Cdc2 activity was assessed by histone H1 phosphorylation and quantified by PhosphorImager analysis (Molecular Dynamics). The duration of mitosis was determined by phase contrast microscopy. (B) Cell cycle–regulated activation of p42 MAPK. Portions of the same cycling extract were subjected to SDS-PAGE and immunoblotting with antibodies to active MAPK (New England Biolabs) or p42 MAPK antiserum X15. (C) Phosphorylation of p42 MAPK and Rsk-2 in Mos-treated cycling extracts. Purified Mos was added (final concentration, 1 μM). Phosphorylation of Rsk-2 and p42 MAPK was detected by shifts in their electrophoretic mobilities on immunoblots. The antisera used were from Santa Cruz Biotechnology (Rsk-2) and our laboratory (X15).

We were able to deplete essentially all of the Rsk-2 fromXenopus egg extracts with a single round of immunodepletion, with no measurable effect on the levels of p42 MAPK and only partial depletion of Rsk-1 (Fig. 2A) (23). We then examined whether Rsk-2 depletion compromised the extract's ability to undergo a CSF arrest. We prepared mock-depleted (Fig. 2, B to D) and Rsk-2–depleted (Fig. 2, E to M) extracts, and supplemented the Rsk-2–depleted extracts with either no added p90 Rsk-2 (Fig. 2, E to G) or physiological concentrations of catalytically inactive KR Rsk-2 (Fig. 2, H to J) (23) or wild-type Rsk-2 (Fig. 2, K to M).

Figure 2

Requirement of Rsk-2 for the Mos-induced mitotic arrest. (A) Specificity of Rsk-2 immunodepletion. Extracts were mock-depleted or immunodepleted of Rsk-2. The extracts were subjected to immunoblotting to detect Rsk-2 (left), p42 MAPK (middle), and Rsk-1 (right). (B toM) Effect of Rsk-2 immunodepletion on the response of cycling extracts to Mos. (B, E, H, and K) Rsk-2 and p42 MAPK phosphorylation in Mos treated extracts. Rsk-2 protein was depleted and replenished (final concentration 100 nM) as indicated. Mos (final concentration, 1 μM) and sperm chromatin (500 per microliter) were added when cycling was initiated. Rsk-2 and p42 MAPK phosphorylation was assessed by immunoblotting with antibodies to Rsk-2 (top), active MAPK (middle), or total p42 MAPK (bottom). (C, F, I, and L) Cdc2 activity in Mos-treated extracts. Cdc2 activity was assessed with histone H1 as substrate and quantitated by PhosphorImager analysis (Molecular Dynamics). Mitotic progression was assessed morphologically by phase contrast microscopy and DAPI (4′,6′-diamidino-2-phenylindole) staining. (D, G, J, and M) Nuclear morphol- ogy of Mos-treated extracts. Nuclei were stained with DAPI and observed by fluorescence microscopy. Bar, 10 μm.

The mock-depleted extract (Fig. 2, B to D) responded to added Mos by undergoing a typical CSF arrest. The extract entered mitosis by 30 min, with increased Cdc2 activity (Fig. 2C), condensed chromatin and dissolved nuclear envelopes (Fig. 2D), increased Rsk-2 phosphorylation (Fig. 2B), and activated p42 MAPK (Fig. 2B). Cdc2 activity remained increased for the duration of the experiment (Fig. 1C) and as a consequence the extract remained in mitosis (Fig. 2, B and D).

In contrast, the Rsk-2–depleted extract (Fig. 2, E to G) did not undergo CSF arrest in response to Mos. Instead, the extract entered and exited mitosis normally, with the mitotic activation of Cdc2 followed by its rapid inactivation (Fig. 2F). The Rsk-2–depleted, Mos-treated extract reformed nuclei with decondensed chromatin and intact nuclear envelopes about 20 min after nuclear envelope breakdown (Fig. 2G). Mock-depleted and Rsk-2–depleted extracts incubated without added Mos cycled with the same 20-min duration of mitosis (22). Thus, Rsk-2–depleted extracts lost their ability to undergo a CSF arrest in response to added Mos (24).

Adding back physiological concentrations of a catalytically inactive KR Rsk-2 (25) protein did not restore the ability of the extract to undergo a CSF arrest (Fig. 2, H to J). The extract still underwent a transient mitosis in the presence of Mos (Fig. 2I), with inactivation of Cdc2 (Fig. 2I) followed by reformation of interphase nuclei with intact nuclear envelopes and decondensed chromatin (Fig. 2, I and J). However, addition of purified recombinant wild-type Rsk-2 did restore the ability of the extract to undergo a CSF arrest in response to Mos (Fig. 2, K to M). When treated with Mos, the extract to which Rsk-2 had been restored underwent a sustained mitotic arrest (Fig. 2, L and M) with increased Cdc2 activity (Fig. 2L); it behaved indistinguishably from the mock-depleted extract (Fig. 2, K to M and B to D). Recombinant Rsk-2 had no effect on the duration of mitosis in the absence of added Mos (18). Moreover, neither the depletion of Rsk-2 nor the addition of KR Rsk-2 or wild-type Rsk-2 had any appreciable effect on Mos-induced p42 MAPK phosphorylation (Fig. 2, B, E, H, and K). Taken together, these data establish Rsk-2 as an essential mediator of the CSF activity of the Mos–MEK-1–p42 MAPK cascade (26).

Activation of the p42 MAPK cascade can sustain the mitotic state even after Cdc2 activity has dropped to low interphase levels (12, 13). The suppression of mitotic exit by p42 MAPK may be important for establishing normal mitotic timing. Accordingly, we examined whether Rsk-2 was required for p42 MAPK-induced suppression of mitotic exit. Even when the addition of Mos was too late to prevent Cdc2 inactivation, it still suppressed mitotic exit (Fig. 3), in agreement with previous reports (12). Depleting Rsk-2 eliminated the ability of Mos to suppress mitotic exit, and wild-type Rsk-2 but not KR Rsk-2 restored the suppression of mitotic exit (Fig. 3). Therefore, Rsk-2 is essential for both suppression of Cdc2 inactivation and post-inactivation suppression of mitotic exit. Any direct effects p42 MAPK might have on mitotic substrates [such as CENP-E (27)] are evidently insufficient to maintain the mitotic state in the absence of Rsk function.

Figure 3

Rsk-2–dependent sustained mitosis after cyclin destruction. The Mos protein (1 μM) was added when cycling was initiated. Cdc2 activity was monitored by histone H1 kinase assay. Mitotic progression was assessed by phase contrast microscopy and DAPI staining.

Finally, we asked whether Rsk-2 is required for the maintenance of mitotic arrest, or just for initiation of mitotic arrest. We prepared CSF-arrested extracts from Xenopus eggs, subjected them to mock depletion or Rsk-2 depletion, and examined whether the depletion caused the extracts to exit their CSF arrest. Both the mock-depleted and Rsk-2–depleted extracts maintained high mitotic levels of Cdc2 activity for at least 60 min, and responded normally to Ca2+ by rapidly inactivating their Cdc2 (Fig. 4). Thus, although Rsk-2 function is essential for the initiation of mitotic arrest, it appears to be dispensable for maintenance of the arrest. Alternatively, it is possible that eggs possess a back-up CSF-like activity that is absent from cycling egg extracts.

Figure 4

Maintenance of CSF arrest in the absence of Rsk-2. CSF-arrested egg extracts were prepared and subjected to mock depletion or Rsk-2 depletion. Calcium (400 μM) was added at 65 min to release the extracts from their arrest state. Amounts of Rsk-2 were assessed by immunoblotting. Cdc2 activity was monitored by H1 kinase assay.

Our findings demonstrate that the Rsk protein kinase is essential for cytostatic factor arrest. Rsk-2 is required for Mos–MEK-1–p42 MAPK–induced suppression of Cdc2 inactivation and also for suppression of mitotic exit subsequent to Cdc2 inactivation. As shown in the accompanying report (26), an activated form of Rsk-1 is capable of causing a CSF arrest; thus, Rsk activation appears to be both necessary and sufficient for CSF arrest. Rsk proteins are therefore critical targets of p42 MAPK in the regulation of cell cycle progression and the development of fertilizable eggs.

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


Stay Connected to Science

Navigate This Article