seminars in CELL BIOLOGY, Vol 6, 1995: pp 73-78

Cyclins of the fission yeast Schixosaccharomyces pombe Daniel Fisher and Paul Nurse

Five cyclin-like genes, cigl, cig2/cycl7, mcs2, pucl und cdcl3, have been discovered in S. pombe to date. It is not yet clear what theirfunctions are or even whether they are all involved with control of the cell cycle. Conflicting data for cigl and cig2/cycl7 have obscured analysis of their function and cigl remains largely uncharacterized, although clues to the role of cig2/cycl7 have emerged. There is genetic data available for the more distant cyclin homologue mcs2, which has an essential although as yet unspecified role. Pucl may be involved in regulation of exit

from the cell cycle. The first cyclin to be discovered, and the best understood, ti cdcl3 which with cdc2 promotes mitosis. Studies of the roles of cdc2 and cdcl3 in the overall ordering of the cell cycle suggest that cdcl3 and fxobably other cyclins are kqr regulators, maintaining the order of S phase and mitosis during the cell cycle.

Key words: S. pombe / cyclins / cell cycle

Structural and functional definitions of cyclins

The original definition of cyclin proteins, as mole- cules which accumulate steadily throughout the cell cycle and which are rapidly degraded following their function at M phase, is probably no longer sufficient. Cyclins have also been proposed to act at other stages of the cell cycle including onset of S phase, the transition from quiescence to the cell cycle of mammalian cells (for review see ref l), and in other aspects of the cell cycle such as spindle pole body duplication and regulation of meiosis (ref 2, also see later). In addition it seems likely that some cyclins will have roles unrelated to control of the cell cycle.

This situation means that a structural rather than a functional definition of a cyclin is required, and in fact many such proteins have been classified as cyclins solely on their amino acid sequence. All cyclins share a region of homology extending over approximately

From the cell Cycle Laboratory, Imperial Cancer Research Fund, PO, Box 123, London, WC2A 3PX, UK

01995 Academic Press Ltd 10434682/95/020073 + 06$8.00/O

200 amino acids, the ‘cyclin box’. Cyclins are usually classified as belonging to a certain subclass based on the degree of homology in the cyclin box. Outside of this region there is little homology between cyclins of different classes, although some conserved motifs have been identifieds which may be important in regulation. Assignment as a cyclin based only on conserved motifs may be misleading. For instance, one study surprisingly showed that members of the ras superfamily have good matches to highly conserved regions of the ‘cyclin box’ - better even than some cyclins - 4 although ras proteins clearly act in a different way from cyclins. Furthermore, in S. pombe, using the Motifs program of the UWGCG package”, for which the cyclin motif is based on a conserved 32 amino-acid pattern, @cl and mcs2 are not identified as cyclins. In general, cyclins are unstable proteins and the sequences which confer this instability reside in the N-terminus of the protein. Conserved amino acid sequences within the cyclin box are required for binding to, and activation of, their associated pro- teins: the cyclin dependent kinases (cdks).5

The presence or absence of some of these motifs in the fission yeast S. Pombe cyclins is indicated in Table 1. Figure 1 shows a relatedness comparison of the five cyclins cloned to date from S. pombe, with those of the rather distantly related budding yeast S.cereuisiae. Three (cdcl3, cigl and cig2/oycl7) are of the B-cyclin family by homology, one (@cl) has limited amine acid homology to Gl cyclins of S. cerzvisiae, while mcs2 is most related to the CCL1 gene although neither of the latter have, as yet, close structural homologues from vertebrate organisms. The dendrogram of Fig- ure 1 places pucl slightly closer to B-type @ins than Gl cyclins, although this is on a consideration of the entire sequence with no additional weighting for specific amino-acid residues which might be more discriminating.

This review will consider some of the lesser known cyclins of S. pmnbe before focussing on the celI cycle control by the cdcl3 protein, and fl’nally consider the general points that have emerged from the study of cyclins in S. pombe.

73

D. Fisher and I? Nurse

Table 1. Some properties of S. ponhe cyclins

Gene Closest structural homologue

Class CC regulation Conserved structural motifs Ref

Dest PEST CAMP SP IN

cdcl3 nimE A.nidulans B Protein, G2/M J J J JJ 23-26 cig2/cycl7 cigl pIU1 mcs2

cdcl3 S.pom6e B RNA, -N, Gl/S J cLB3/4 S.cmisiae No

CL.Nl C.ahicans G:,B RNA,-N f J CCL.1 S.cermisiae C/H No %

no apparent regulation (No). Peak of expression of regulated at the transition from the G2 phase of the cell cycle to mitosis;

consensus published in ref 3 t for PEST sequences of known cyclin proteins; Motifs, presence or absence of motif

exce P.

s7; Dest! destruction box; SP, putauve cdc2 phosphorylation site; CAMP, putative cAMP-kinase consensus phosphory auon site; W, tyrosine-contaimng putative amphipathic helix common to many cyclins. A space indicates that there is no clearly observable match to the consensus. The closest structural homologue was obtained in each case by searching the GenFMBL databank using the TFASTA program of the UWGCG packages*. Ref. literature reference.

mcst, upstream or downstream of cdc2?

The function of mcs2 is, at present, unresolved although it may have a role in maintaining cdc2 kinase activity. The original mcs2-75 mutation was described in a screen for extragenic suppressors of the mitotic catastrophe phenotype of’ a cdc2-3w weel-50 strain (the term ‘mitotic catastrophe’ describes the lethal mitosis caused by advancement of mitosis to a point where DNA replication is incomplete). The mcs2 cyclin was cloned by complementation of the synthetic lethality of an mcs2-75 cdc2-3w cdc25-22 triple mutant Mcs2 is an essential cyclin6 which is present in the nucleus throughout the cell cycle, and the abundance of the protein does not vary, unlike mitotic cyclins and some Gl cyclins. The deletion causes an arrest in the second or third division following spore germination, with cells displaying division septa and multiple compact nuclei, although it is probable that these cells are no longer in mitosis. Some mcs2 protein can be detected after germination of deleted cells6 which could explain why the deletion does not arrest in the first cycle following germination if enough of the protein is carried over during sporulation to sustain several cell cycles (given that it is not specifically degraded at one point of the cycle). Alternatively, msc2 may be essential for some process other than the cell cycle. The allele-specific genetic interactions of mcs2 with cdc2 and other known cell- cycle genes, and the septated cells resulting from the deletion do, however, argue that mcs2 is involved somehow in the cell cycle. The genetic interactions might suggest that mcs2 is an upstream activator of the mitotic function of cdc2, while the terminal pheno- type of the mcs2 deletion places it downstream. This could be explained by the proposed feedback loop between septation activation and cdc2 activity.’ Before

cytokinesis occurs cdc2 activity is maintained by a feedback control dependent on the cdcl6 gene, which itself is required following septation. In other words, a protein which acts downstream of cdc2 can be required for maintenance of cdc2 activity itself.

Mcs2 was originally found to be most homologous (about 20%) to human cyclin C (ref 6), the function of which is unknown. However, the discovery of another more homologous cyclin C (>70%) in Drosophila4 suggests that the cyclin C family may be more closely related in different organisms and that mw2 may not be a true Ctype cyclin. The CCL1 gene of S. cereuisiae shares the closest sequence homology (34% identity, 60% with conservative changes) of any protein, although it is not known whether the essential functions of the CCL1 protein or its inter- acting protein kinase, KIN28, are cell-cycle related.’ Recently cyclin H, the cyclin which interacts with the cdc2 activating kinase (CAK), was found to have mcs2 as its closest relative (about 20% overall and 35% within the cyclin box).’ Perhaps mcs2 might fulfill a similar role in S. pombe, in which case one would expect it to interact with a UK-type protein kinase. Mcs2 does interact with a protein kinase which is not cdc2, and which has a different substrate specific- ity - instead of phosphorylating histone Hl, the mcs2 interacting kinase acts efficiently on casein. A kinase encoded by cskl, a distant cdk homologue, suppresses the mzs2 deletion in high copy, although the cskl deletion is not lethal and reduces, but does not abolish, mcs2 associated kinase activity. These data suggest that neither the true CDK partner of mcs2 nor the cyclin partner, if there is one, of cskl have yet been found. It will be interesting to see whether the mcsB associated kinase is a homologue of M015, the CAK kinase subunit. Another mcs mutant, m&i-13 shares a very similar spectrum of genetic interactions to mcs2-

74

Cyclins of fission yeast

75 and it may well be that mcs6 will turn out to be a cdk or cyclin.”

Cig 1 and cig2/cyc17, Gl cyclim?

Cigl and cig2/cycl7 are cyclins most closely related to CLB3 and CLB4 genes of S. cerevisiae and cdcl3, a S. pombe B-type cyclin, although of the two only cig2/ cycl7 has a destruction box and putative ubiquitin interacting motX3 The latter two motifs are thought to confer specific degradation to the protein and it is

CLB6

CLB5

CLB4

CLB3

cigl

CLB2

CLBI

cig2kyc 17

cdcl3

pucl

mcs2

CCL 1

HCS26

ORFD

CLGl

CLN2

CLNl

WHll

Figure 1. Dendrogram of S. pombe (bold) and S. cereuis& (upper case) cyclins clustered by the pileup program of the UWGCG package. ‘s The dendrogram indicates similarity on the horizontal axis (proportional to the length of the branch line before bifurcation) but does not indicate evolutionary relationship nor functional similarity.

true that the cig2/cycl7 gene, and probably protein, is cell cycle regulated but cigl is not.

Cigl’ ’ was cloned by PCR and originally proposed to be a Gl cyclin due to a reported extension of Gl in a cigl deletion, although recent work has shown this to be incorrect’*“s. cigl mRNA is constant through- out the cell cycle. The cigl deletion is not lethal and no clear role for cigl in the cell cycle has thus far been found, although this may due to redundancy with other cyclins. A mammalian structural homologue of cigl, cyclin G, is of unknown function.t4

Cig2/cycl7 was cloned by complementation of the triple CLN deletion of S. cere-otiiae by two groups’5*‘6 and as a multicopy suppressor (cycl7+) of the temperature sensitivity of the meiotic kinase mutant, patl-114, by a third group. ” Originally, cig2/cycl7was proposed to function along with cdcl3 at G2/M by virtue of synthetic lethality with cdc25-22 and cdc2-3w which act at G2/M, rescue of the mitotic catastrophe of mutants which lack cdc2 tyrosine phosphorylation, and a synthetic phenotype with cdcl3-11 7.15 However, these synthetic interactions have not been repeat- able.16*‘s A cigl cig2/cycl7 double disruption was claimed by one group” to have a synthetic phenotype with slower growth rates. Cells were also reported as having two 1C nuclei per cell (although the large 4C peak in the DNA flow cytometry profile of the latter would not rule out the nuclei being 2C). This, also, has not been demonstrable by all groups.“*” It is possible that some of these discrepancies are due to the different cig2/cycl7 null constructs that have been used: a LEU2 insertion,” sup3-5 insertion or partial deletionr6 or URA4 complete deletion.” Another possible explanation is that the behaviour is different depending on the history of the cell-for instance, whether the cells have been growing vegetatively or have been starved or just sporulated. This is possible given the diversity of functions that have been attributed to cyclins.

The best evidence for the function of cig2/cycl7 has come from Obara Isihara et aZ.,” who found that arrest in Gl in response to nutrient limitation occurs much more rapidly in a cig2/cycl7 deletion. In addition, cig2/cycl7 mRNA, which normaRy peaks at Gl/S of the celI cycle ‘w’ is strongly induced during nitrogen starvation and conjugation, and this induc- tion requires the start gene res2t. Normal cell cycle regulation of expression is dependent on resl t. Conjugation is markedly elevated in the cig2/cycl7 deletion, and overproduction of the wild type gene suppresses conjugation. Conjugation is also sup pressed by overproducing the resl and res2 genes

75

D. Fisher and l? Nurse

which interact with cdcl0 to form Gl transcription complexes operating predominantly in the mitotic and meiotic cycles, respectively. Cig2/cycl7 is pro- posed to be downstream of these two genes since the suppression of conjugation by overproduction of resl and res2 is dependent on cig2/cycl7+. This situation may be analogous to the regulation of Gl cyclins of S. cerevisiae in which the homologues of cdcl0 and resl and ~2, SW6 and SW4 respectively, activate tran- scription of the cydins CLNl and CLN2. In S. cerevisine, B-type cyclins have been found to be involved in Sphase as well as the G2 to M transition.‘g-21

Thus, cig2/qcl7 appears to act as an inhibitor of mating during the mitotic cycle, and is transcription- ally upregulated during Gl and in nitrogen starvation. However, it is not clear whether in S. pombe cig2/cycl7 promotes Gl or only inhibits mating during the mitotic cycle, and it is possible that these two functions are not separable. Given the viability of a cigl cig2/cycl7 double deletion (and a triple deletion with pucl), it would seem that another cyclin, required for Gl, is likely to be found in S. pombe.

pucl, not really a Gl cych?

Pucl was originally cloned by restoring resistance to u-factor arrest of a WHIl (CLN3) deletion of S. cerevisiae and it rescues the lethality, due to Gl arrest, of the triple CLN deletion.** The protein has some similarities to the Gl cyclins of S. cereoisiae including 10 out of 12 amino acids that distinguish this class from other cyclins, and is slightly more similar to A-type cyclins than to other classes (B, C, D and E).** However, subsequent work has failed to identify a role for @cl analogous to that of the Gl cyclins of S. ~erevisia~~ In S. pombe no synthetic effect of a pucl deletion has been found with any known cell cycle genes. Overexpression of the full length protein or a truncated protein lacking the N-terminal PEST sequences interferes with cdcl3 function in G2 and is synthetically lethal with c&13-117, but so far only under these circumstances has any physical inter- action been shown between pucl and cdc2.

However, pucl does show some effects similar to those of cig2/cyclZ the message is upregulated in nitrogen starvation (although not to levels such that the protein can be seen to iinteract with cdc2), and o&expression rescues the patl-114 mutant by sup- pressing the haploid meiosis at the restrictive tem- perature. In addition, the pucl deletion confers a

slight acceleration of Gl arrest and meiosis with low levels of nitrogen source. ‘s However, all of the effects of deletion or overexpression are smaller than those of cig2/cycl7 and it seems possible that, if the G2 phenotype of pucl overexpression is caused by an unphysiological interference with cdcl3, then the slight but significant Gl phenotypes of the @cl deletion might be due to similar effects with cig2/ cycl7. In this case the physiological role of pucl would have yet to be found. Alternatively, of course, @cl may be redundant with other cyclins involved with exit from the meiotic cycle, or Gl progression.

cdcl3, the prototype cych in S. pombe

Cdcl3 might be considered to be the only cyclin in fission yeast whose function has been ascertained. Certainly cdcl3 is a mitotic B-type cyclin which is required for activation of cdc2 at the G2/M bound-

ary. 23-26 Cdc2 expressed from a plasmid can rescue c&l3 ts and cs mutants25 and a deletion of cdcl3 shows no aspects of mitosis.*@j B-type cyclins interact with the cdc25 phosphatase which activates cdc2.” Two tem- perature sensitive mutations of cdcl3, cdcll117 and cdc13-9, arrest the cell with certain aspects of mitosis completed, including condensation of chromosomes and formation of a septum. However, by other criteria, cells have not entered mitosis since they have an interphase array of microtubules rather than a mitotic spindle.*s In these cells cdc2 kinase is active,*” although the specific activity might not be sufficient to bring about all aspects of mitosis. Another inter- pretation is that cdcl3 may be required for a further function during mitosis, possibly including the rear- rangement of the G2 microtubular network into a mitotic spindle. Of interest in this context is the further function proposed for cdcl3 of tar-getting cdc2 to the nucleus24r28 or to regions within the nucleus such as the nucleolus2’ and the spindle pole bodies.” It is also noteworthy that, in different systems, cyclins seem to modulate substrate specificity of the cdc2 kinase (for example ref 31).

Given that there is only a single CDK, encoded by cdc2, which is required for both S phase and mitosis in fission yeast, the cdc2 kinase can presumably control both these major events of the cell cycle by interacting with different cyclin proteins. This would allow a binary switch to maintain the exclusivity of S phase and M phase, controlled by mutually exclusive cyclins. There is certainly evidence for such control in S. cerevtiaewhere Gl cyclins act to repress G2 cyclins and

76

Cyclins ofjission yeast

vice versa.32 It is significant that mutations which destabilize the cdc2-cdcl3 interaction can lead to inappropriate S phase,sss4 while deletion of c&l3 causes cells to become locked into a cycle of recurring DNA replication without mitosis.34 The latter result suggests that the presence or absence of a complex between cdc2 and cdcl3 may be the only determinant of the decision whether to enter mitosis or to perform S-phase. In support of this view overexpression of cdcl3 during Gl can lead to a lethal mitosis if there is sufficient activatable cdc2.34 It might also be expected that overexpression of Gl cyclins would also lead to DNA rereplication.

Cdc2 acts at S phase as well as mitosis, while a cdcl3 cdc2 complex in G2 promotes mitosis but a lack of cdcl3-cdc2 complex in G2 promotes S phase. Thus the cdcl3-cdc2 complex normally prevents S phase. This prevention of S phase by cdcl3-cdc2 complexes could be explained in two alternative ways. Firstly, it might be that the cdcl3 component prevents associa- tion of cdc2 with Sphase substrates, and, conversely, Gl cyclins might not be able to interact well with mitotic substrates, i.e. there is a qualitative difference between Gl and G2 cyclins. Replacement of a G2 cyclin with a Gl cyclin would then promote S phase in lieu of mitosis and vice versa. Alternatively, the normal completion of S phase might coincide with an increase in cdc2 kinase activity due to an association which can now take place between cdc2 and cdcl3. This kinase activity itself could be enough to restrain S phase and a drop in activity might be required before S phase can take place again, before reactiva- tion of cdc2 brought about by Gl cyclins. This would normally happen at the end of mitosis when cdcl3 is degraded. Thus the difference between Gl and G2 cyclins would be one of quantitative regulation, such that the level of protein kinase activity during one phase of the cell cycle would specify the following event.

Cyclins, the universal regulators?

Protein phosphorylation is a universal mechanism for cell regulation, and specific protein kinases and phosphatases have proven to be universal cell cycle regulators (For reviews see refs 35,36). The cyclins, it seems, add another dimension of flexibility to this control, by directing the action of cyclindependent kinases. While in general they are structurally similar in regions which directly act on the cdks, they perform a diversity of functions which will doubtless

be attributed to other parts of the protein. They are vital for control of the cell cycle and, in some cases, may be redundant in one process by virtue of the need to integrate multiple signals at one point. For instance, in Gl the signals for correct growth and nutrient availability act to promote S phase of the mitotic cycle while pheromones and other signals may act in opposition. Once commitment to S phase has been made it is necessary to repress alternative pathways leading to stationary phase arrest or meiosis. Cyclins are probably involved at all of these points and it is notable that in S. ponrbe, as in all other organisms to date, no single cyclin has been found to be indispensable for S phase.

It is not yet clear to what extent cyclins are redundant. For example, in S. cerez~tin’ae, at least 13 cyclin like genes exist although no single cyclin is essential for the cell cycle. In S. pombe five have so far been found although it is almost certain that more will be cloned. It may well be that what appears, at the moment, to simply be redundancy, will turn out to be a further level of more specific or subtle controls.

Acknowledgements

We thank Matthew O’Connell and Jacky Hayles for their helpful discussions and comments on the manuscript, and the ICRF and The Wellcome Trust for support to D.F.

References

1. 2.

3.

4.

5.

6.

7.

8.

9.

Sherr CT (1993) Mammalian Cl cvclins. Cell 73:1059-1065 Nasmyth K (1993)Control of the yeast cell cycle by the CDC28 protein kinase. Curr Opin Cell Biol 5:166-179 Nugent JHA, Alfa CE, Young T, Hyams JS (1991) Conserved structural motifs in cyclins identified by sequence analysis. J Cell Sci 99669-674 Leopold P, O’Farrell PH (1991) An evolutionarily conserved cyclin homolog from drosophila rescues yeast deficient in Gl cydms. Cell 661207-1216 Lees EM, Harlow E (1993) Sequences within the conserved cyclin box of human @m-A are su8icient for biding to and activation of cdc2 kinase. Mol Cell Biol 15:119&1201 Mob L, Beach D (1993) Characterization of the fission yeast mcr.2 cychn and its associated protein-kinase activity. EMBO J 121723-1732 Fankhauser C, Marks J, Raymond A, Siianis V (1993) The S. pombc c&I6 gene is required both for maintenance of .p34*” kinase activity and regulation of septum formation: a link between mitosis and cytokinesis. EMBO J 122697%2704 Way JG, Simon M, Faye G (1993) The KIN28 protein~titieis associated with a cyclin in Sacchmvmycss cetdsh. J. Molf B@l 234307-310 Fisher BP, Morgan DO (1994) A novel cychn a&o+& .k$tb M015/CDK7 to form the CDKactivating ~l&ase?~&il 787X3-724

77

D. Fisher and I! Nurse

10. Mob L, Booher R, Young P, Beach D (1989) c&2 and the regulation of mitosis: six interacting mcs genes. Genetics 122:773-782

11. Bueno A, Richardson H, Reed SI, Russell P (1991) A fission yeast Etype cyclii functioning early in the cell cycle. Cell 66149-159

12. Bueno A. Russel P (1993) A fission Yeast .Btvne cvclin

13. functioning early in the cell ‘cycle (Erratum). Cell 73:1056 Forsburg S, Nurse P (1994) Analysis of the Schi.zasacchammyccs pornbe cyclin wl: evidence for a role in celi cycle exit. J Cell Sci 107:601-613

14.

15.

Tamura K, Kanaoka Y, Jinno S, Nagata A, Ogiso Y, Shimiau K, Hayakawa T, Nojhna H, Okayama H (1993) Cyciin-G-A new mammalian cyclin with homology to fission yeast cigl. Onto- gene 82113-2118 Bueno A, Russell P (1993) 2 fission yeast B-type cyclins, cig2 and c&13, have diierent functions in mitosis. Mol Cell Biol 13:228&2297

16. Connolly T, Beach D (1994) Interaction between the cigl and cig2 Btype cyclins in the fission yeast cell cycle. Mol Cell Biol 14768-776

17.

18.

Obara-Ishiiara T Okayama H (1994) A B-type cyclin negatively regulates conjugation via interacting with cell cycle ‘start’ genes in fission yeast. EMBO J 13:1863-1872 Bueno A, Russell P (1994) 2 fission yeast B-type cyclins, cigz and cdcl3, have different functions in mitosis (Erratum). Mol Cell Biol l&869

19. Schwab E, Nasmyth K (1993) CLB5 and CLBG, a new pair of B cyciins involved in DNA replication in Sacchammyces weuisiae. Genes Dev 7:1160-1175

20. Richardson H, Lew DJ, Henze M, Sugimoto K, Reed SI (1992) Cyclin-B homologs in Sacchuromyces cerevicioe function in S-phase and in G2. Genes Dev 62021-2034

21. Kuhne C, Linder P (1993) A new pair of Btvpe cvclins from sacchammyces c-emdae that function early in a the’ cell cycle. EMBO I 123437-3447

22. Forsbu& SL, Nurse P (1991) Identification of a Gl-Lype cychn pucli in the fission yeast Schizosatchawmy~ pmn& Nature 351:245-248

23. Hagan I, Hayles J. Nurse P (1988) Cloning and sequencing of the @in-related c&13+ gene and a cytological study of its role in fission yeast mitosis. J Cell Sci 91:587-595

24. Booher RN, AIfa CE, Hyams JS and Beach DH (1989) The tission yeast cdc2/cdcl3/sucl protein kinase: regulation of catalytic activity and nuclear localization. Ceil 58:485-497

25.

26

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

Booher R, Beach D (1988) Involvement of c&13+ in mitotic control in Schitosacchammyces pornbe possible interaction of the gene product with microtubules. EMBO J. 72321-2327 Moreno S, Hayles J, Nurse P (1989) Regulation of ~34’~” protein kinase during mitosis. Cell 58361-372 Zheng XF, Ruderman JV (1993) Functional-analysis of the P-box, a domain in cyclin-B required for the activation of cdc25. Cell 75:155-164 Aifa CE, Booher R, Beach D, Hyams JS (1989) Fission yeast cyclin: subcellular localisation and cell cycle regulation. J Cell Sci Suppl. 12:9-19 Gallagher IM, Aifa CE, Hyams JS (1993) P63(cdc’s), a B-type cyclin, is associated with both the nucleolar and chromatin domains of the fission yeast nucleus. Mol Biol Cell 41087-1096 Alfa CE, Ducommun B, Beach D, Hyams, JS (1990) Distinct nuclear and spindle pole body population of cyclincdc:! in fission yeast. Nature 347680-682 Peeper DS, Parker LL, Ewen ME, Toebes M, Hall FL, Xu M, Zantema A, Vandereb, AJ, Piwnicaworms H (1993) A-type and Btype cyclins diierentiaily modulate substrate-specificity of cyclin cdk complexes. EMBO J 12:1947-1954 Amon A, Tyers M, Futcher B, Nasmyth K (1993) Mechanisms that help the yeast cell cycle clock tick: G2 cyclins transcription- ally activate G2 cyclins and repress Gl cyclins. Cell 74993-1007 Broek D, Bartlett R, Crawford K, Nurse P (1991) Involvement of P34=d’2 . in establishing the dependency of S phase on mitosis. Nature 349:388-393 Hayles.1, Fisher D, Woollard A, Nurse P (1994) Temporal order of S-phase and mitosis in fission yeast is determined-by the state of the p34’d’*/mitotic B cydin complex. Cell 78:813-822 Pines J-(1993) Cyclins and’cyciindependent kinases: take your partners. Trends Biochem Sci 18:195-197 Labib K, Nurse P (1993) Bring on the phosphatases. Curr Biol 3164-166 Rogers S, Wells R, Rechsteiner M (1986) Amino-acid sequences common to rapidly degraded proteins: the PEST hypothesis. Science 234364-368 Program Manuai for the Wisconsin Package, Version 8, September 1994, Genetics Computer Group, 575 Science Drive, Madison, Wisconsin, USA 53711