- Email: [email protected]
Turning on mitosis
ANDREW W. MURRAY
CELL-CYCLE CONTROL
Turning
on mitosis
Biochemical analysis of enzymes that modify the active p34Cdc2 kinase component of MPF supports the idea that MPF stimulates its own activation during the induction of mitosis. The cell cycle is quantal. Normal cells are either in interphase or mitosis, but never in between. Yet the induction of mitosis requires the accumulation of B-type cyclins, which is a gradual proc8ess. How does the biochemical machinery that regulates the cell cycle convert the continuous accumulation of cyclin into a sharp transition between different states? The first clue came in 1971, during experiments8 that revealed the existence of MPF (maturation promoting factor), the activity that induces cells to enter mitosis or meiosis. Injecting MPF into frog oocytes was found to stimulate the activation of a latent form of MPF, named pre-MPF, suggesting that a positive feedback lloop helps to ensure rapid and irreversible entry into mitosis [ 11. Meanwhile studies on the fission yeast Schizosat:chromycespombe revealed cardidate components of such a positive feedback system, including Weel, a protein that inhibits entry into mitosis, and Cdc25, a protein that stimulates entry into mitosis [ 21. In the last five years the cell cycle has graduated from a black box of genes and physiological observations to biochemical respectability [3]. MPF is the best characterized of a family of protein kinases that regulate events in the cell cycle. These kinases are complexes between a catalytic subunit that is a member of the Cdk (cyclindependent kinase) family and a cyclin. Active MPF is a complex of p34cdc2 (named after the fission yeast cdc2 gene) and cyclin B, in which the p34CdC2subunit is phosphorylated on tyrosine 161. In contrast to this activating modification, phosphorylation of p34CdC2on threonine I4 and tyrosine 15 inhibits MPF activity. Both types of phosphorylation occur only on p34cdc2 molecules that are
The analysis of MPF activation and the demonstration that Wee1 is a tyrosine kinase and Cdc25 a tyrosine phosphatase led to a simple model for the control of MPF activity (Fig.1). In the model, Wee1 phosphorylates p34cdc2 on tyrosine15, inhibiting MPF activation, and Cdc25 removes these phosphates to activate MPF. The balance between the activities of Wee1 and Cdc25 determines the activity of MPF. If Wee1 activity exceeds that of Cdc25, preMPF accumulates and cells remain in interphase. But if Cdc25 gains the upper hand, preMPF is converted into active MPF and the cell enters mitosis. Postulating that the activation of Cdc25 and inactivation of Wee1 are both catalysed by MPF produces a positive feedback loop that would make entry into mitosis rapid and irreversible. In this model, most newly synthesized cyclin B enters the pre-MPF pool, but a small amount escapes into active MPF. At a certain point, the amount of active MPF exceeds a critical threshold, activating the positive feedback loop and triggering entry into mitosis. This model is appealing for many reasons. It fits the genetic data, and mathematical formulations of the model produce, life-like oscillations in MPF activity (G Odell, personal communication). Furthermore, Wee1 does indeed phosphorylate p34cdc2 on tyrosine 15 [4] and Cdc25 does remove this phosphate [5]. Finally, biochemical experiments show that as frog egg extracts enter mitosis the
MO15
Wee1 s
z
Cdc25
117active
complexed to cyclins. The inhibitory phosphorylations are dominant to the activating one: pre-MPF is the inactive p34cdc2cyclin B complex that is phosphorylated on both tyrosine I5 and threonine 161.
Phosphatase I?
Inactive
Inactive
Active
I Fig. 1. The steps involved in MPF activation. Newly synthesized cyclin B binds to unphosphorylated ~34 cdc2 to form an p34cdc2 can be phosphorylated on tyrosine15 (Y) by Wee1 and on threonine 161 0 by a kinase named M015; these reversed by Cdc25 and an unknown phosphatase (possibly phosphatase I), respectively. Molecules phosphorylated (pre-MPF) accumulate in C2 and are converted into active MPF by Cdc25. The activation of MPF is accompanied by and Wee1 inhibition, suggesting that MPF can stimulate its own activation,
@ Current
Biology
1993, Vol 3 No 5
inactive complex. modifications are on both residues Cdc25 activation
291
292
Current
Biology
1993, Vol 3 No 5
rate of tyrosine phosphorylation on p34CdC2declines and the rate of tyrosine Idephosphorylation increases. The past year has seen a number of papers that investigate the biochemical regulation of MPF activation by Cd&5 and Weel, resulting in the scheme shown in Figure 2. In frog extracts the amount of Cdc25 remains constant throughout the cell cycle, but its activity and phosphory lation state do not: in interphase, Cdc25 is weakly phosphorylated and has low phosphatase activity, but in mitosis it is heavily phosphorylated and highly active [6,71. This change reflects an increase in the activity of the enzymes that phosphorylate Cdc25 as well as a decrease in the activity of thos,e that dephosphorylate it. What are these activities? The ability of okadaic acid to prevent Cdc25 dephosphotylation suggests that the Cdc25specific phosphatase is a form of protein phosphatase 24, most likely an activity named INH that was purified as an inhibitor of MPF activation by Lee et al. [$]. In interphase extracts, okadaic acid treatment can drive Cdc25 into its fully phosphotylated and active form, providing a simple explanation for the ability of this compound to induce premature entry into mitosis in a variety of organisms.
0
Cdc25
Okadaic acid
?
Nim 1
Fig. 2. The regulation of Cdc25 and Wee1 activity. The balance of Wee1 and Cdc25 activity controls the activation of MPF. In its active form Wee1 inhibits fvtPF activation, whereas active Cdc25 stimulates conversion of ore-MPF into MPF. Cdc25 is activated by phosphorylation catalised by MPF, and inactivated by protein phosphatase 2A. Wee1 is inactivated by phosphorylation catalysed by Niml; it is not known whether MPF can activate Niml.
Is MPF the Cdc2Qpecific kinase? A recent paper suggests that the answer may be yes. Hoffman et all [9] have found that mitotic extracts of tissue culture cells contain Cdc25specific kinase activity, but interphase extracts do not. The
suggesting that MPF - or an activity physically associate& with it - can phosphoryIate and activate Cdc25. In co$ trast, p34CdC%yclin A complexes, which are active in g< and G2 phases of the cell cycle, cannot phosphoiylajg Cdc25: this makes sense, because if cyclin A-contain&$ complexes could activate Cdc25, they would indire+ induce the conversion of pre-MPF into MPF, leading to: premature mitosis. This difference between cyclin A and cyclin B also supports the widespread belief that in addi- .’ tion to activating p34cdc2 and its relatives, cyclins help to confer substrate specificity on the resulting protein kinase complexes. What inactivates Wee1 as cells enter mitosis? A combination of genetic and biochemical evidence suggests that, at least in fission yeast, the guilty party is not MPF. The fission yeast nirnl gene (also known as c&1) was identified both as a gene whose overexpression suppresses cdc25fi mutations and by mutations that alter the response of cells to starvation [lo]. By genetic criteria, Niml acts as a negative regulator of Weel, and Coleman et al. [ 111 and Wu and Russell [12] have now shown that Niml performs this function by phosphorylating and inactivating Weei. Incubating purified Wee1 and Nina1 with each other leads to extensive phosphorylation of Wee1 and prevents Wee1 from phosphorylating ~34~d-clin B complexes on tyrosine 15. These experiments raise the crucial question of how general the control of Wee1 activity by Niml is, and what regulates Niml activity during the cell cycle. So far there is no evidence that homologs of Niml exist in other eukaryotes, though the conservation of the cell-cycle machinery suggests that they will soon be found. The obvious candidate for an activator of Niml is MPF itself. An alternative is that Niml is activated by some other pathway that depends on the accumulation ofcyclin, or some other parameter that vanes during the cell cycle. What other factors regulate the activation of MPF and entry into mitosis? In particular, are the activities of the enzymes that catalyse other modifications of p34Cdc2regulated during the cell cycle? One such modification is the phosphorylation of threonine 14, which occurs to different extents in different organisms and, like tyrosine15 phosphorylation, inactivates MPF. The most recent evidence suggests that threonine modification is not due to Weel, raising the possibility of further regulatory complexity [4]. There is no evidence that the activity of the threonine 161 kinase, named Cdk activating kinase (UK), vanes during the cell cycle [13]. The successful purification of this kinase, and its identification as the previously characterized kinase named MO15 ([ 143, and T Hunt and J Shut&worth, personal communication) suggests that it will soon be possible to reconstitute the mechanisms that regulate the activation of MPF from purified components. Once we have left the test tube behind, can we be certain that the activation of a small amount of MPF initiates a positive feedback loop that drives cells into mitosis? Not only can we not be certain, but experiments in frog egg extracts have generated evidence against this widely held belief. Norbury et al. [ 151 investigated the role of
DISPATCH
threonine 14 and tyrosine15 phosphorylation in MPF activation by synthesizingmutant forms of human p34cdc2in frog extracts.In the extractsused by Norbury et al. there is a substantiallag before the appearanceof kinaseactivity associatedwith either endogenousp34CdC2 or the newly synthesizedhuman p34CdC2. As expected, double mutant human p34CdC2, carrying mutations that prevent phosphorylation on both threonine and tyrosine, is activated without a lag. If the activationof MPF is autocatalytic,this earlyactivationof human p34CdC2 should induce a positive feedback loop that would activatethe much larger population of endogenouslp34cdC2 in the extract. Strikingly, this behavior is not seen:the lag before the activationof endogenousp34CdC2 is undiminished. We are left with a paradox. On the one hand there is biochemical evidencefor at least one of the steps in the model for the autocatalyticactivation of MPF. Furthermore, the replacementof wild-type p34CdC2 with mutants that cannot be phosphotylated on tyrosine15 or threonine 14 clearly leads to precocious and disastrousentry into mitosis. On the other hand, an experiment that should show that MPF can induce its own activation in a physiologicalsetting dramaticallyfails to do so. One way of resolving this paradox is to postulate that cells have mechanisms that defend them against perturbations that would cause partial MPF activation. For example, when p34cdc2associateswith cyclin, phosphorylation of threonine1611by the Cdk-activatingkinase before phosphotylation of tyrosine15 by Wee1 will lead to the transient appearanceof MPF. In the straightforward autocatalyticmodel such a temporal discrepancywould lead to premature entry into mitosis and incur the risk of genetic damage.These considerationsremind us that although mathematicalmodels and the simple biochemical circuits they are based on can explain how the cell cycle ought to work, these elegant schemesmust always be carefullymeasuredagainstthe behavior of real cells. References
CL: Cytopiasmic control of nuclear behavior 1. MAsul Y, MARKETT during meiotic maturation of frog oocytes. / Exp Zool 1971, 177~12945.
2. RUSSELL P, NURSEP: Negative regulation of mitosis by weel+, a gene encoding a protein kinase homolog. Cell 1987, 49559-567. 3. MURRAYA, HUNT T: The ceiI cycle: an introduction. W. H. Freeman: New York; 1333. 4. PARI(ER LL, PIWNIC&WORMS H: Inactivation of the p34cdc2-~clin B complex by the human WEE1 tyrosine kinase. Science1992, 257~1955-1957. 5. STRAUSFIELD U, IABBEJC, FESQUETD, CAVAWREJC, PICARDA, SADHUK, RUSSELL P, DOREEM: Dephosphorylation and activation of a p34cdc2/cyclin B complex in vitro by human CDC25 protein. Nature 1991, 351:242-245. 6. KIJMAGIA, DUNPHYWG: Regulation of the cdc25 protein during the cell cycle in Xenopus extracts. CeZZ1992, 70:133-151. 7. IZUMIT, WALKERDH, MALIERJL: Periodic changes in the phosphorylation of the Xenopus cdc25 phosphotase regulate its activity. Mel Biol Cell 1992, 3~927-939. 8. LEE TH, SOLOMON MJ, MUMBYMC, KIR~CHNFZR MW: INH, a negative regulator of MPF, is a form of protein phosphatase 2A. Cell 1991, C&415-423. 9. HOFFMANI, CLARKEPR, MARCO%MJ, KAR~ENTIE, DRAET~AG: Phosphorylation and activation of human cdc25-C by cdc2cvciin B and its involvement in the self amphcation of MPF at mitosis. LTMBOJ 1993, 12:53-63. 10. RUSSELL P, NURSEP: The mitotic inducer niml+ functions in a regulatory network of protein kinase homologs controlling the initiation of mitosis. Cell 1987, 49:56V-576. 11. COLEMANTR, TANG 2, DUNPHVWG: Negative regulation of the Wee1 protein khtase by direct action of the Nhnl/Cdrl mitotic inducer. Cell 1993, 731.20. 12. Wu L, RUSSELL P: Nil kinase promotes mitosis by inactivating Wee1 tyrosine kinase. Nature 1993, in press. MJ, LEET, KIR.SCHNER MW: The role of phosphot-yla13. SOLOMON tion in ~34~~~~ activation: idemmcation of an activating kinase. Mot Biol CeU 1992, 3327. 14. SHUT~LEWORTH J, GODFREYR, COLMAN A: p4OMO15, a cdc2-related protein kinase involved in negative regulation of meiotic maturation of Xenopus oocytes. EMBO J 1990, 9:3233-3240. C, BEOWJ, NURSEP: Regulatory phosphorylation of 15. NORBURY the p34cdcz protein kinase in vertebrates. Eh4BO J 1991, 10:3321-3329.
Andrew W. Murray, PhysiologyDepartment, Box 0444, University of California, San Francisco,California 94143. 0444, USA
293
CELL-CYCLE CONTROL
Turning
on mitosis
Biochemical analysis of enzymes that modify the active p34Cdc2 kinase component of MPF supports the idea that MPF stimulates its own activation during the induction of mitosis. The cell cycle is quantal. Normal cells are either in interphase or mitosis, but never in between. Yet the induction of mitosis requires the accumulation of B-type cyclins, which is a gradual proc8ess. How does the biochemical machinery that regulates the cell cycle convert the continuous accumulation of cyclin into a sharp transition between different states? The first clue came in 1971, during experiments8 that revealed the existence of MPF (maturation promoting factor), the activity that induces cells to enter mitosis or meiosis. Injecting MPF into frog oocytes was found to stimulate the activation of a latent form of MPF, named pre-MPF, suggesting that a positive feedback lloop helps to ensure rapid and irreversible entry into mitosis [ 11. Meanwhile studies on the fission yeast Schizosat:chromycespombe revealed cardidate components of such a positive feedback system, including Weel, a protein that inhibits entry into mitosis, and Cdc25, a protein that stimulates entry into mitosis [ 21. In the last five years the cell cycle has graduated from a black box of genes and physiological observations to biochemical respectability [3]. MPF is the best characterized of a family of protein kinases that regulate events in the cell cycle. These kinases are complexes between a catalytic subunit that is a member of the Cdk (cyclindependent kinase) family and a cyclin. Active MPF is a complex of p34cdc2 (named after the fission yeast cdc2 gene) and cyclin B, in which the p34CdC2subunit is phosphorylated on tyrosine 161. In contrast to this activating modification, phosphorylation of p34CdC2on threonine I4 and tyrosine 15 inhibits MPF activity. Both types of phosphorylation occur only on p34cdc2 molecules that are
The analysis of MPF activation and the demonstration that Wee1 is a tyrosine kinase and Cdc25 a tyrosine phosphatase led to a simple model for the control of MPF activity (Fig.1). In the model, Wee1 phosphorylates p34cdc2 on tyrosine15, inhibiting MPF activation, and Cdc25 removes these phosphates to activate MPF. The balance between the activities of Wee1 and Cdc25 determines the activity of MPF. If Wee1 activity exceeds that of Cdc25, preMPF accumulates and cells remain in interphase. But if Cdc25 gains the upper hand, preMPF is converted into active MPF and the cell enters mitosis. Postulating that the activation of Cdc25 and inactivation of Wee1 are both catalysed by MPF produces a positive feedback loop that would make entry into mitosis rapid and irreversible. In this model, most newly synthesized cyclin B enters the pre-MPF pool, but a small amount escapes into active MPF. At a certain point, the amount of active MPF exceeds a critical threshold, activating the positive feedback loop and triggering entry into mitosis. This model is appealing for many reasons. It fits the genetic data, and mathematical formulations of the model produce, life-like oscillations in MPF activity (G Odell, personal communication). Furthermore, Wee1 does indeed phosphorylate p34cdc2 on tyrosine 15 [4] and Cdc25 does remove this phosphate [5]. Finally, biochemical experiments show that as frog egg extracts enter mitosis the
MO15
Wee1 s
z
Cdc25
117active
complexed to cyclins. The inhibitory phosphorylations are dominant to the activating one: pre-MPF is the inactive p34cdc2cyclin B complex that is phosphorylated on both tyrosine I5 and threonine 161.
Phosphatase I?
Inactive
Inactive
Active
I Fig. 1. The steps involved in MPF activation. Newly synthesized cyclin B binds to unphosphorylated ~34 cdc2 to form an p34cdc2 can be phosphorylated on tyrosine15 (Y) by Wee1 and on threonine 161 0 by a kinase named M015; these reversed by Cdc25 and an unknown phosphatase (possibly phosphatase I), respectively. Molecules phosphorylated (pre-MPF) accumulate in C2 and are converted into active MPF by Cdc25. The activation of MPF is accompanied by and Wee1 inhibition, suggesting that MPF can stimulate its own activation,
@ Current
Biology
1993, Vol 3 No 5
inactive complex. modifications are on both residues Cdc25 activation
291
292
Current
Biology
1993, Vol 3 No 5
rate of tyrosine phosphorylation on p34CdC2declines and the rate of tyrosine Idephosphorylation increases. The past year has seen a number of papers that investigate the biochemical regulation of MPF activation by Cd&5 and Weel, resulting in the scheme shown in Figure 2. In frog extracts the amount of Cdc25 remains constant throughout the cell cycle, but its activity and phosphory lation state do not: in interphase, Cdc25 is weakly phosphorylated and has low phosphatase activity, but in mitosis it is heavily phosphorylated and highly active [6,71. This change reflects an increase in the activity of the enzymes that phosphorylate Cdc25 as well as a decrease in the activity of thos,e that dephosphorylate it. What are these activities? The ability of okadaic acid to prevent Cdc25 dephosphotylation suggests that the Cdc25specific phosphatase is a form of protein phosphatase 24, most likely an activity named INH that was purified as an inhibitor of MPF activation by Lee et al. [$]. In interphase extracts, okadaic acid treatment can drive Cdc25 into its fully phosphotylated and active form, providing a simple explanation for the ability of this compound to induce premature entry into mitosis in a variety of organisms.
0
Cdc25
Okadaic acid
?
Nim 1
Fig. 2. The regulation of Cdc25 and Wee1 activity. The balance of Wee1 and Cdc25 activity controls the activation of MPF. In its active form Wee1 inhibits fvtPF activation, whereas active Cdc25 stimulates conversion of ore-MPF into MPF. Cdc25 is activated by phosphorylation catalised by MPF, and inactivated by protein phosphatase 2A. Wee1 is inactivated by phosphorylation catalysed by Niml; it is not known whether MPF can activate Niml.
Is MPF the Cdc2Qpecific kinase? A recent paper suggests that the answer may be yes. Hoffman et all [9] have found that mitotic extracts of tissue culture cells contain Cdc25specific kinase activity, but interphase extracts do not. The
suggesting that MPF - or an activity physically associate& with it - can phosphoryIate and activate Cdc25. In co$ trast, p34CdC%yclin A complexes, which are active in g< and G2 phases of the cell cycle, cannot phosphoiylajg Cdc25: this makes sense, because if cyclin A-contain&$ complexes could activate Cdc25, they would indire+ induce the conversion of pre-MPF into MPF, leading to: premature mitosis. This difference between cyclin A and cyclin B also supports the widespread belief that in addi- .’ tion to activating p34cdc2 and its relatives, cyclins help to confer substrate specificity on the resulting protein kinase complexes. What inactivates Wee1 as cells enter mitosis? A combination of genetic and biochemical evidence suggests that, at least in fission yeast, the guilty party is not MPF. The fission yeast nirnl gene (also known as c&1) was identified both as a gene whose overexpression suppresses cdc25fi mutations and by mutations that alter the response of cells to starvation [lo]. By genetic criteria, Niml acts as a negative regulator of Weel, and Coleman et al. [ 111 and Wu and Russell [12] have now shown that Niml performs this function by phosphorylating and inactivating Weei. Incubating purified Wee1 and Nina1 with each other leads to extensive phosphorylation of Wee1 and prevents Wee1 from phosphorylating ~34~d-clin B complexes on tyrosine 15. These experiments raise the crucial question of how general the control of Wee1 activity by Niml is, and what regulates Niml activity during the cell cycle. So far there is no evidence that homologs of Niml exist in other eukaryotes, though the conservation of the cell-cycle machinery suggests that they will soon be found. The obvious candidate for an activator of Niml is MPF itself. An alternative is that Niml is activated by some other pathway that depends on the accumulation ofcyclin, or some other parameter that vanes during the cell cycle. What other factors regulate the activation of MPF and entry into mitosis? In particular, are the activities of the enzymes that catalyse other modifications of p34Cdc2regulated during the cell cycle? One such modification is the phosphorylation of threonine 14, which occurs to different extents in different organisms and, like tyrosine15 phosphorylation, inactivates MPF. The most recent evidence suggests that threonine modification is not due to Weel, raising the possibility of further regulatory complexity [4]. There is no evidence that the activity of the threonine 161 kinase, named Cdk activating kinase (UK), vanes during the cell cycle [13]. The successful purification of this kinase, and its identification as the previously characterized kinase named MO15 ([ 143, and T Hunt and J Shut&worth, personal communication) suggests that it will soon be possible to reconstitute the mechanisms that regulate the activation of MPF from purified components. Once we have left the test tube behind, can we be certain that the activation of a small amount of MPF initiates a positive feedback loop that drives cells into mitosis? Not only can we not be certain, but experiments in frog egg extracts have generated evidence against this widely held belief. Norbury et al. [ 151 investigated the role of
DISPATCH
threonine 14 and tyrosine15 phosphorylation in MPF activation by synthesizingmutant forms of human p34cdc2in frog extracts.In the extractsused by Norbury et al. there is a substantiallag before the appearanceof kinaseactivity associatedwith either endogenousp34CdC2 or the newly synthesizedhuman p34CdC2. As expected, double mutant human p34CdC2, carrying mutations that prevent phosphorylation on both threonine and tyrosine, is activated without a lag. If the activationof MPF is autocatalytic,this earlyactivationof human p34CdC2 should induce a positive feedback loop that would activatethe much larger population of endogenouslp34cdC2 in the extract. Strikingly, this behavior is not seen:the lag before the activationof endogenousp34CdC2 is undiminished. We are left with a paradox. On the one hand there is biochemical evidencefor at least one of the steps in the model for the autocatalyticactivation of MPF. Furthermore, the replacementof wild-type p34CdC2 with mutants that cannot be phosphotylated on tyrosine15 or threonine 14 clearly leads to precocious and disastrousentry into mitosis. On the other hand, an experiment that should show that MPF can induce its own activation in a physiologicalsetting dramaticallyfails to do so. One way of resolving this paradox is to postulate that cells have mechanisms that defend them against perturbations that would cause partial MPF activation. For example, when p34cdc2associateswith cyclin, phosphorylation of threonine1611by the Cdk-activatingkinase before phosphotylation of tyrosine15 by Wee1 will lead to the transient appearanceof MPF. In the straightforward autocatalyticmodel such a temporal discrepancywould lead to premature entry into mitosis and incur the risk of genetic damage.These considerationsremind us that although mathematicalmodels and the simple biochemical circuits they are based on can explain how the cell cycle ought to work, these elegant schemesmust always be carefullymeasuredagainstthe behavior of real cells. References
CL: Cytopiasmic control of nuclear behavior 1. MAsul Y, MARKETT during meiotic maturation of frog oocytes. / Exp Zool 1971, 177~12945.
2. RUSSELL P, NURSEP: Negative regulation of mitosis by weel+, a gene encoding a protein kinase homolog. Cell 1987, 49559-567. 3. MURRAYA, HUNT T: The ceiI cycle: an introduction. W. H. Freeman: New York; 1333. 4. PARI(ER LL, PIWNIC&WORMS H: Inactivation of the p34cdc2-~clin B complex by the human WEE1 tyrosine kinase. Science1992, 257~1955-1957. 5. STRAUSFIELD U, IABBEJC, FESQUETD, CAVAWREJC, PICARDA, SADHUK, RUSSELL P, DOREEM: Dephosphorylation and activation of a p34cdc2/cyclin B complex in vitro by human CDC25 protein. Nature 1991, 351:242-245. 6. KIJMAGIA, DUNPHYWG: Regulation of the cdc25 protein during the cell cycle in Xenopus extracts. CeZZ1992, 70:133-151. 7. IZUMIT, WALKERDH, MALIERJL: Periodic changes in the phosphorylation of the Xenopus cdc25 phosphotase regulate its activity. Mel Biol Cell 1992, 3~927-939. 8. LEE TH, SOLOMON MJ, MUMBYMC, KIR~CHNFZR MW: INH, a negative regulator of MPF, is a form of protein phosphatase 2A. Cell 1991, C&415-423. 9. HOFFMANI, CLARKEPR, MARCO%MJ, KAR~ENTIE, DRAET~AG: Phosphorylation and activation of human cdc25-C by cdc2cvciin B and its involvement in the self amphcation of MPF at mitosis. LTMBOJ 1993, 12:53-63. 10. RUSSELL P, NURSEP: The mitotic inducer niml+ functions in a regulatory network of protein kinase homologs controlling the initiation of mitosis. Cell 1987, 49:56V-576. 11. COLEMANTR, TANG 2, DUNPHVWG: Negative regulation of the Wee1 protein khtase by direct action of the Nhnl/Cdrl mitotic inducer. Cell 1993, 731.20. 12. Wu L, RUSSELL P: Nil kinase promotes mitosis by inactivating Wee1 tyrosine kinase. Nature 1993, in press. MJ, LEET, KIR.SCHNER MW: The role of phosphot-yla13. SOLOMON tion in ~34~~~~ activation: idemmcation of an activating kinase. Mot Biol CeU 1992, 3327. 14. SHUT~LEWORTH J, GODFREYR, COLMAN A: p4OMO15, a cdc2-related protein kinase involved in negative regulation of meiotic maturation of Xenopus oocytes. EMBO J 1990, 9:3233-3240. C, BEOWJ, NURSEP: Regulatory phosphorylation of 15. NORBURY the p34cdcz protein kinase in vertebrates. Eh4BO J 1991, 10:3321-3329.
Andrew W. Murray, PhysiologyDepartment, Box 0444, University of California, San Francisco,California 94143. 0444, USA
293