Commitment to mitosis is regulated by a conserved protein kinase complex called MPF (mitosis-promoting factor). MPF activation triggers a positive-feedback loop that further promotes the activity of its activating phosphatase Cdc25 and is assumed to down-regulate the MPF-inhibitory kinase Wee1. Four protein kinases contribute to this amplification loop: MPF itself, Polo kinase, MAPK (mitogen-activated protein kinase) and Greatwall kinase. The fission yeast SPB (spindle pole body) component Cut12 plays a critical role in modulating mitotic commitment. In this review, I discuss the relationship between Cut12 and the fission yeast Polo kinase Plo1 in mitotic control. These results indicate that commitment to mitosis is co-ordinated by control networks on the spindle pole. I then describe how the Cut12/Plo1 control network links growth control signalling from TOR (target of rapamycin) and MAPK networks to the activation of MPF to regulate the timing of cell division.
Mitotic commitment is controlled by MPF (mitosis-promoting factor)
Extensive genetic analysis in the fission yeast, Schizosaccharomyces pombe, and biochemical manipulations of frog and sea-urchin eggs led to a core understanding of how eukaryotic cells use a ubiquitous protein kinase complex called MPF to control commitment to cell division. MPF is composed of a regulatory subunit called cyclin B that associates with a small catalytic subunit known either as Cdk1 (cyclin-dependent kinase 1) or Cdc2 (the name of the fission yeast cdk1 gene) [1–4]. As active MPF can promote mitosis from any point in the cell cycle, it is vital that this potent activity is kept in a restrained state until it is appropriate to instigate division. This restraint is imposed by a second kinase, Wee1. Wee1 phosphorylates a tyrosine residue within Cdc2's catalytic site to block MPF activity [5,6]. Once the conditions for division have been satisfied, such as the accumulation of appropriate cell mass, the inhibitory phosphate is removed by a specific phosphatase called Cdc25 [7–10]. Thus the timing of cell division is determined by the balance of the antagonistic activities of Wee1 and Cdc25 towards MPF [1,10].
Positive-feedback control of MPF activation involves Polo kinase
Manipulation of Xenopus egg extracts established that MPF activation promotes a positive-feedback loop that boosts Cdc25 activity and inhibits Wee1 to promote the rapid and complete commitment that is characteristic of all mitoses [11,12]. Consistently, the activation of Cdc25 in the fission yeast actually relies on the prior activation of Cdc2 . Feedback loop modification of Xenopus Cdc25 enabled it to be recognized by a phospho-specific antibody called MPM2 (mitotic protein monoclonal 2) . The persistence of feedback loop activation and MPM2 reactivity after removal of MPF activities from Xenopus egg extracts, or the mutation of MPF consensus phosphorylation sites on Cdc25, indicated that the feedback loop involved at least one more protein kinase [14,15]. The search for this activity led Kumagai and Dunphy  to show that Xenopus Polo kinase, Plx1, bound Cdc25, activated it and converted it into an MPM2 reactive state. Subsequent studies consolidated this view by showing that Plx1 is an integral part of the feedback loop [12,17–21]. Such a role for Polo in the control of mitotic commitment is consistent with the block to mitotic entry after injection of anti-Plk1 antibodies into non-transformed human cells. Importantly, the same antibodies were unable to block mitotic commitment in transformed cells, suggesting that some transformation events can bypass the requirement for Polo during mitotic commitment . This role for Polo kinase in feedback loop control may now be more extensively addressed in human cells with drugs that specifically inhibit Polo . While initial studies with such drugs have focused on transformed cells , analysis of their impact on primary cells is eagerly awaited.
Additional kinases in MPF feedback control
Other recent studies in Xenopus have indicated that MAPK (mitogen-activated protein kinase) and Greatwall kinases also act in the feedback loop [25,26]. Greatwall apparently acts independently of Polo, MPF and MAPKs and has been proposed to oppose a phosphatase activity within the feedback loop . This phosphatase may well be PP2A (protein phosphatase 2A) as PP2A has been linked, by a number of studies, to feedback loop control in Xenopus [12,18,27]. Furthermore, while deletion of both PP2A genes in the fission yeast is lethal, deletion of one alone accelerates commitment to mitosis and can partially suppress the mitotic commitment defect of cdc25.22 . Thus PP2A is a good candidate for the Greatwall antagonist. In Drosophila, Greatwall (gwl) antagonizes Polo kinase and recessive gwl mutants delay mitotic commitment, suggesting that Greatwall involvement in feedback control is likely to be conserved across species [29,30]. Thus there is much yet to learn about feedback loop amplification and it will be important to test the insights gained in one system in others to differentiate between universal elements of cell cycle control and specific adaptations to the demands of particular lifestyles.
The fission yeast SPB (spindle pole body) and MPF control
Our poor understanding of the sequence motifs that encode tyrosine phosphatase activity in the 1980s led to a significant delay between the realization that the fission yeast Cdc2 was controlled by tyrosine phosphorylation and the demonstration that Cdc25 was the phosphatase that removed this modification [7,31]. The most plausible model during this period was that Cdc25 alone was not the tyrosine phosphatase, but was either an activator or regulatory subunit of an unidentified phosphatase. Such a view prompted Hudson et al.  to search for the gene encoding this phosphatase by isolating mutants that suppressed the temperature-sensitive growth defect of cdc25.22 cells. They identified three mutations in the same gene that they named stf1+ for suppressor of twenty-five. Interestingly, stf1.1 mutation supported ten cell divisions of cells completely lacking the cdc25+ gene . As Cdc25 was soon shown to be the phosphatase and all three stf1 mutations were dominant mutations, stf1+ retreated from the spotlight.
cut12.1 cells are unable to form a mitotic spindle at the restrictive temperature, as one of the two SPBs fails to nucleate microtubules, suggesting that Cut12 influences SPB function . Molecular characterization established that cut12+ encoded an essential SPB component and the unanticipated realization that cut12+ was allelic to stf1+. The cut12.1 mutation introduces a stop codon 13 amino acids before the end of the ORF (open reading frame), whereas the stf1.1 mutation substitutes a glycine for valine at residue 71 of the same ORF (because cut12.1 is recessive and stf1 mutations are dominant, we elected to refer to the gene as cut12+ and the stf1.1 mutation as cut12.s11) . Thus a dominant mutation in an essential SPB component was able to compensate for loss of Cdc25 activity.
Further genetic analysis consolidated the link between Cut12 and MPF activation. The introduction of a cdc25.22 mutation into a cut12.1 background exacerbated the cut12.1 defect, leading to death at a temperature that was normally permissive for either single mutant. Cells died with all microtubules emanating from just one of the two SPBs . Furthermore, simply raising Cdc25 levels suppressed the cut12.1 defect . This reciprocal relationship between cut12 and cdc25 strongly suggests that the primary function of Cut12 is as a cell cycle regulator and that the SPB activation defect is a secondary consequence of inappropriate regulation of MPF. Within this context, the observation that MPF is recruited to SPBs in G2 [35,36] suggests that Cut12 would regulate SPB activation of MPF in a localized manner on commitment to mitosis and that the cut12.1 mutation compromises this localized activity. Dominant activation of Cut12 function by the stf mutations is, however, not locally constrained, as it clearly has the ability to influence the global activation of MPF throughout the cell.
Cut12 influences Polo kinase control of MPF
When considering models to account for suppression of cdc25.22 by cut12.s11, we were influenced by two observations. First, inactivation of Wee1 alone is sufficient to overcome the need for Cdc25 . Secondly, Wee1 is heavily phosphorylated during feedback activation [37–39]. We therefore tested the possibility that Cut12 may be influencing the feedback loop controls that normally promote the inactivation of Wee1. This inactivation of Wee1 would then overcome any need for Cdc25 . In this scenario, once Wee1 has been shut down, any newly synthesized cyclin B would form a complex with unphosphorylated Cdc2 and so not rely on Cdc25 to remove an inhibitory phosphate for activation. As Polo kinase is a key part of these controls in higher systems, we asked whether cut12 status influenced the function of the fission yeast Polo kinase, Plo1 .
The specific model we addressed was that the dominant cdc25.22-suppressing cut12.s11 mutation inappropriately promoted Plo1 activity to override the requirement for Cdc25 (by shutting down Wee1). A large body of data now supports this model. Plo1 and Cut12 associate with one another in two-hybrid and co-immunoprecipitation assays . Plo1 is normally only recruited to mitotic and late G2 SPBs. In cells that harbour the cdc25.22-suppressing cut12.s11 mutation however, Plo1 associates with many more interphase SPBs . Having established that Plo1 kinase is activated downstream of MPF (i.e. it is part of the feedback loop), we found that the cut12.s11 mutation boosted Plo1 activity in both interphase and mitosis and that Plo1 activity was barely detectable in mitotically arrested cut12.1 cells [41,43]. Furthermore, we were able to use MPM2 antibodies to specifically study Plo1-dependent kinase activity on the SPBs. These experiments confirmed that Plo1 activity on the SPBs was enhanced by cut12.s11. Genetic experiments established that cut12.s11 suppression of cdc25.22 relied on Plo1 activity and dominant activation of Plo1 in a wild-type cut12+ background enabled cdc25.22 cells to enter mitosis. We therefore believe that interplay between Cut12 and Plo1 constitutes part of the normal controls that govern entrance to mitosis in the fission yeast .
The centrosome and MPF control
Evidence from other systems suggests that regulation of mitotic control from the spindle pole may be a universal aspect of MPF activation, as activation of MPF on centrosomes precedes MPF activation at any other location within human cells  and the addition of centrosomes to Xenopus egg extracts accelerates MPF activation . Triggering mitotic commitment at spindle poles may enable a limited number of molecules from distinct signalling pathways to converge upon a common downstream switch. Relatively few molecules could then integrate signals from diverse networks to generate a unified signal that would promote division with the appropriate timing for the particular conditions at any given time. The impact of phosphorylation of Ser-402 of Schizosaccharomyces pombe Plo1 consolidates this view by demonstrating that changes in the environment alter the affinity of Plo1 kinase for the SPB to change the timing with which cells commit to division [46,47].
The MAPK SRP (stress-response pathway) influences SPB recruitment of Plo1 to control mitotic commitment
Plo1 Ser-402 phosphorylation is enhanced in late G2 to reach a peak during mitosis. As stated above, Plo1 is prematurely recruited to the SPB of cells harbouring the cdc25.22-suppressing cut12.s11 mutation . Mutation of Ser-402 to alanine to block the phosphorylation (plo1.S402A) abolished both premature recruitment to the SPB and the ability of cut12.s11 to suppress cdc25.22 lethality. In contrast, mutation of Ser-402 to a glutamic acid residue (plo1.S402E; to mimic phosphorylation) led to the premature recruitment of Plo1 to the SPB of a cut12+ strain. Furthermore, mutation of Ser-402 to alanine blocked the suppression of cdc25.22 by a constitutively active Plo1 in a cut12+ background. Thus phosphorylation at Ser-402 regulates the affinity of Plo1 for the SPB and is required for suppression of cdc25.22 .
The fortuitous observation that the heat shock that accompanies a temperature shift from 25 to 36°C stimulated Ser-402 phosphorylation  led us to establish that phosphorylation of Ser-402 was regulated by the activity of the Sty1/Spc1 MAPK SRP. Deletion of sty1+ (sty1.Δ) completely blocked Ser-402 phosphorylation. Given the correlation between Ser-402 status, recruitment to the SPB and the suppression of cdc25.22 by cut12.s11, we asked whether boosting signalling through the SRP could mimic cut12.s11 in suppressing cdc25.22. In this scenario, SRP signalling would promote the phosphorylation of Ser-402 to enhance the affinity for the SPB and this would then lead to suppression of cdc25.22. Encouragingly, hyperactivation of Sty1 by constitutive activation of the upstream Sty1 kinase, Wis1 (with the wis1.DD mutation), suppressed cdc25.22. Furthermore, this suppression was blocked by plo1.S402A .
Deletion of sty1+ greatly delays the timing with which cells commit to mitosis [48,49]. Furthermore, SRP signalling has been proposed to link environmental signals to cell cycle control, as SRP signalling is stronger in minimal medium than in rich medium . We therefore asked whether SRP signalling modulated the timing of cell division by modulating the phosphorylation of Ser-402. plo1.S402E reduced the length at division of sty1.Δ cells from 21.6±2 μm to 17.5±1.6 μm, whereas mutation to alanine had no impact . This reduction in cell length shows that Ser-402 control constitutes a significant part of the mechanism by which SRP signalling regulates mitotic commitment.
TOR (target of rapamycin) signalling modulates phosphorylation of Plo1 Ser-402 to couple mitotic commitment to changes in nutrient quality
The significance of exploiting Ser-402 signalling to regulate mitotic commitment has been firmly established by Petersen and Nurse . Fantes and Nurse  have established that a shift from a rich to a poor nitrogen source led to an immediate acceleration of commitment to mitosis after which cells adjusted to the new environment and divided at a new rate. Petersen and Nurse  re-examined this control in light of the cues from other systems that TOR signalling leads to environmental regulation of growth control. They found that TOR signalling was reduced by the shift from a rich to a poor nitrogen source and that it was this reduction in TOR signalling that was responsible for the acceleration of mitotic commitment. Simple inhibition of TOR kinases while maintaining cells in a good nitrogen source was also sufficient to reduce cell length at division. Significantly, TOR inhibition reduced the size at division by the same degree as that seen on shift to a poor nitrogen source. A reciprocal shift from a poor to a rich nitrogen source delays mitotic commitment. Again the delay to mitotic commitment invoked by the enrichment of the nitrogen source relied on TOR signalling. The search for the mechanism by which TOR signalling was modulating the timing of commitment to mitosis led to an SRP protein phosphatase called Pyp2. As Pyp2 dephosphorylates Sty1 , any alteration of Pyp2 levels/activity will alter Sty1 activity (and thus in turn influence phosphorylation on Ser-402 of Plo1). Indeed, Pyp2 levels declined after a medium shift from a good to a poor nitrogen source. The concomitant rise in Sty1 activity enhanced Plo1.S402 phosphorylation to promote the affinity of Plo1 for the SPB and thus accelerated the rate of cell division. The importance of this signal transduction to Ser-402 of Plo1 was demonstrated by the inability of plo1.S402A cells to alter their rate of division in response to a shift from a rich to a poor nitrogen source. Thus Plo1-dependent signalling events on the SPB connect mitotic commitment to nutrient-dependent changes in TOR signalling to control the rate of cell division .
Plo1 Ser-402 phosphorylation drives recovery of division following stress
A second instance in which Plo1 recruitment to the SPB via S402 phosphorylation plays a critical role in controlling mitotic commitment was revealed by subjecting cells to heat and centrifugation stress . This stress activates SRP signalling to block commitment to mitosis, allowing time for the damage invoked by the stress to be repaired before division. Once repair is complete, cells re-enter the cell cycle and divide. The considerable delay in a return to the cell cycle of plo1.S402A mutants after such a stress established that re-entry into the cell cycle is not a default process that simply occurs once SRP signalling has abated below a set level; rather, it requires active signalling through Plo1 Ser-402 .
These experiments establish that Plo1-dependent signalling events on the SPB control the timing of mitotic commitment in at least two distinct responses to changes in the cellular environment. It is likely that further analysis will reveal additional pathways that modulate Plo1 itself, Cut12 or interacting proteins to control the timing of mitotic commitment.
The genetic link between an SPB component and cell cycle control forged by the isolation of the stf alleles by Paul Young's group has provided a conduit to an entirely unforeseen, yet central aspect of cell cycle control . The activation of MPF at centrosomes in human cells and the association of critical cell cycle regulators such as Chk1 with centrosomes  suggest that the pathways that we are unravelling in yeast are likely to represent conserved signalling networks.
British Yeast Group Meeting 2008: Independent Meeting held at National University of Ireland Maynooth, Maynooth, Co. Kildare, Ireland, 18–20 March 2008. Organized and Edited by Gary Jones (National University of Ireland Maynooth, Ireland).
I thank Janni Petersen and Agnes Grallert for a critical reading of the paper and discussions. My group is supported by Cancer Research UK (CRUK) grant number C147/A6058.