* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download The study of cell cycle control is entering a new and exciting phase
Cell nucleus wikipedia , lookup
Protein phosphorylation wikipedia , lookup
Cell encapsulation wikipedia , lookup
Endomembrane system wikipedia , lookup
Extracellular matrix wikipedia , lookup
Signal transduction wikipedia , lookup
Programmed cell death wikipedia , lookup
Cell culture wikipedia , lookup
Organ-on-a-chip wikipedia , lookup
Cellular differentiation wikipedia , lookup
Cytokinesis wikipedia , lookup
Cell growth wikipedia , lookup
J. Cell Sci. Suppl. 4, 155-170 (1986) Printed in Great Britain © The Company of Biologists Limited 1986 155 CELL CYCLE REGULATION IN YEAST JA C Q U E L IN E HAYLES a n d PAUL NURSE Imperial Cancer Research Fund, PO Box 123, Lincoln’s Inn Fields, London WC2A3PX, UK INTRODUCTION The study of cell cycle control is entering a new and exciting phase. The classical physiological and cytological approaches to investigating the cell cycle have been supplemented with the powerful techniques of molecular biology and genetics. This is particularly the case for the budding yeast Saccharomyces cerevisiae and the fission yeast Schizosaccharomycespombe. As yeasts are simple eukaryotes, cell cycle regulation is less likely to be confused with the controls involved in growth and differentiation, which are required in multicellular organisms. Thus they provide excellent systems in which to study the basic mechanism of cell cycle regulation and represent good models for understanding eukaryotic cell cycle control. The ease of classical genetic analysis in the two yeasts has resulted in the isolation of mutants in many genes that are required for the cell cycle. Study of such mutants has been useful for formulating hypotheses about how the cell cycle might be controlled and for identifying genes that have a role to play in those controls. These abstract studies have then been followed up using the tools of molecular biology to determine the biochemical basis of the controls involved. In this paper we shall concentrate on one particular gene function, that of cdc28 from S. cerevisiae and the analogous gene cdc2 from Schiz. pombe. This gene function has a pivotal position in cell cycle control. We shall review work that establishes that the cdc2/cdc28 gene function is required in G\ to commit cells to S phase and the rest of the mitotic cell cycle, and then again in Gz, where it determines the cell cycle timing of mitosis. Isolation and characterization of the cdc2 and cdc28 genes have shown that they encode protein kinases that may be regulated by phosphorylation. This has led to the hypothesis that phosphorylation of a specific set of proteins may control the transition from G\ into 5 phase and from Gz into mitosis. The hypothesis has obvious similarities to several models being developed for control of the cell cycle in other eukaryotic cells. However, before considering these matters we want to discuss briefly what we mean by the cell cycle and its control, in order to define more precisely the problems that we think are important to investigate. TH E C ELL CYCLE A N D ITS CONTROL The cell cycle is made up of the events and processes that take place between the birth of the cell and its subsequent division into two daughters. In steady-state growth the mass of the cell will also double during the cell cycle. Therefore, taken jf. Hayles and P. Nurse 156 literally, cell cycle events and processes will include all those that are required for cellular growth as well as those for direct reproduction of the cell. Such a literal interpretation is rather unfocused since it will include processes such as protein and RNA synthesis, which are only indirectly relevant to cell reproduction. So we shall confine our attention to those cell cycle events and processes that are specifically required for cell reproduction and not for cellular growth. This distinction has not always been fully appreciated. For example, the changes that occur when quiescent cells are stimulated are often assumed to be cell-cycle-specific. But, as well as being stimulated to enter into the cell cycle, such cells are also being stimulated to grow. Consequently, many of the changes observed will be associated with the growth of the cell rather than with the cell cycle itself. The two major processes that are specifically required for cell reproduction are S phase and mitosis. These ensure the replication and segregation of the chromosomes and are found in all eukaryotic cell cycles. When considering cell cycle control we run into the problem that the term ‘control’ can mean very different things to different people. It may be used in the sense of ‘required for’. In this sense, any event that is necessary for successful completion of the cell cycle can be considered part of its control. We do not think that this is a very useful way to look at the problem. For us, implicit in the term ‘control’ is the idea of regulation, and we want to identify events that are important for regulation. Three useful ways of identifying regulatory events are to find those acting at: (1) initiation steps, (2) rate-limiting steps and (3) commitment steps. We shall consider each of these in turn. An initiation event is simply the first event that is specifically required for the process under consideration. The difficulty here is knowing when the first event has been found, as there may always be an earlier event that has not been identified. In cell cycle terms we are interested in the initiation events of the whole cell cycle, of S phase and of mitosis. A rate-limiting step is an event that makes a major contribution to the overall rate at which a process is completed. Any event necessary for the cell cycle will become rate-limiting if it is inhibited, but that does not mean that it is normally rate-limiting. It is more informative to speed up an event than inhibit it, since in these circumstances only genuinely rate-limiting steps will advance the whole process. Various treatments such as the use of mutants, medium shifts and addition of cell extracts have been used to advance cells through the cell cycle into 5 phase and mitosis in order to identify major rate-limiting steps. A commitment step is more difficult to identify. It can be defined as a point of no return beyond which no alternative developmental pathways can be initiated until the cell cycle in progress has finished. Thus practical identification of a commitment step requires developmental fates as alternatives to the mitotic cell cycle; then, as cells progress through the cell cycle their ability to undergo these alternatives can be assessed. Problems arise because the various alternatives may become excluded at different times during the cycle, making it difficult to define the true point of no return. 157 From our discussion so far it would seem that we are dealing with quite different controls. However, this need not be the case. For example, a point of commitment to the cell cycle in G\ might also be a major rate-limiting step of the cell cycle and an initiation event for 5 phase. Indeed if this were the case we would feel encouraged that, imperfect as our individual views of control are, together they are leading us generally in the right direction. As we shall show, this does seem to be the case from studies with the yeasts. Cell cycle regulation in yeast MITOT IC C E L L CYCLES OF THE YEASTS The first parts of the cell cycle of both yeasts are similar to each other and to those of other eukaryotic cells. There is a G\ period that becomes much expanded at very low growth rates. After G\ there is a discrete S phase during which DNA replication takes place using multiple replicons (Newlon, Petes, Hereford & Fangman, 1974). However, the second part of the cell cycle differs between the two yeasts (Nurse, 1985a). Fission yeast is more typically eukaryotic, with a G2 period and mitosis. During mitosis a microtubular spindle forms, chromosomes visibly condense and the cell divides by medial fission. The condensation of Schiz. pombe chromosomes has been dramatically demonstrated by Yanagida and co-workers using a mutant strain defective in tubulin (Umesono, Hiraoka, Toda & Yanigida, 1983). In budding yeast there is no real Gz period. A short mitotic spindle is formed during 5 phase and the cell becomes arrested in mid-mitosis for the rest of the cycle. The spindle then elongates, mitosis is completed and the cell divides by budding. No visible chromosome condensation takes place. Both yeasts differ from most eukaryotes in that the nuclear membrane does not break down during mitosis. As we shall see later the difference in the organization of the two cell cycles is also reflected in their mitotic controls. CE L L CYCLE G E N E S OF T H E YEASTS A large number of conditionally lethal cell cycle mutants have been isolated in the two yeasts that are unable to complete the cell cycle (Hartwell, Mortimer, Culotti & Culotti, 1973; Nurse, Thuriaux & Nasmyth, 1976; Reed, 1980; Nasmyth & Nurse, 1981). Most of these mutants are heat-sensitive although some are cold-sensitive (Moir, Stewart, Osmond & Botstein, 1982; Hiraoka, Toda & Yanigida, 1984). The mutants fall into two classes: those that continue growth and macromolecular synthesis at their restrictive temperature and those that do not. Following our earlier discussion, we consider only those mutants that can continue growth and macromolecular synthesis as being specifically defective in cell cycle progress. By this criterion, about 50 cdc genes have been identified in budding yeast and about 40 genes in fission yeast. These genes define functions that are specifically required for successful completion of the cell division cycle, but only a few can be expected to be involved in cell cycle control. 158 Gi ‘ s t a r t ’ J . Hayles and P. Nurse control The major cell cycle control in budding yeast is located in G\ at start. ‘Start’ was originally identified by the pioneering work of Hartwell, Pringle and co-workers. It is the earliest gene-controlled event of the cell cycle identified to date and, therefore, may be the point of initiation (Pringle & Hartwell, 1981). Traverse of start results in commitment to the cell cycle with respect to the alternative developmental pathway of conjugation (Hartwell, 1974). After start is passed conjugation is not possible until the cell cycle in progress has been completed. The start event does not take place until cells have attained a critical size: at low growth rates daughter cells are smaller at division and so cells must grow for a longer period before passing start (Hartwell & Unger, 1977; Carter & Jagadish, 1978: Lord & Wheals, 1980). Therefore, under these conditions traverse of start is a major rate-limiting step of the cell cycle. We can see that all three ways of identifying regulatory control events can be applied to start. Start gene functions have been investigated by Reed and his co-workers. Four gene functions, cdc28, -36, -37 and -39, are required to traverse start (Reed, 1980). It has been claimed that mutants in other genes also block at or before start but in these cases it has not been firmly established that the mutants can still undergo cellular growth. For this reason they may not be defective specifically in cell cycle progress. The mating pheromones oc and a factor also arrest cells at start to prepare them for conjugation (Bucking-Throm, Duntze, Hartwell & Manney, 1973). Mutants in cdc36 and -39 appear to be constitutive for mating pheromone response. In cells that are unable to mate, e.g. a/ar diploids, the cdc36 and -39 mutants do not show cell cycle arrest. cdc36 and -39 mutants also allow certain ste (sterile) mutants to mate that are normally unable to do so (Shuster, 1982). This evidence suggests that these two genes are involved in the mating pheromone response pathway that arrests cells at start. Mutants in the other two start genes cdc28 and -37 are not influenced by the mating system and so appear to be more central to the start control system. A start control has also been identified in the fission yeast. Mutants in two genes, cdc2 and cdclO, block at a point in the cell cycle during Gi where cells are still able to conjugate if challenged to do so (Nurse & Bissett, 1981; Aves, Durkacz, Carr & Nurse, 1985). But once past this point cells are committed to the mitotic cycle and are unable to conjugate. Traverse of start in fission yeast also appears to be a major rate-limiting step (Fantes, 1977; Nurse & Thuriaux, 1977; Nasmyth, Nurse & Fraser, 1979). In mutants that divide at a reduced cell size, the Gi period before start becomes expanded. A similar expansion is seen when growth rate is reduced by growing cells in chemostat cultures (Nasmyth, 1979). As with budding yeast, passage of start also requires growth to a critical cell size. The cdc2 start gene in fission yeast is functionally similar to the cdc28 start gene in budding yeast (Beach, Durkacz & Nurse, 1982). When the cdc28 gene is introduced into mutants of cdc2 these cells are able to divide and grow at the restrictive temperature. This suggests that the cdc2 and cdc28 genes are functionally inter changeable and therefore perform essentially similar functions within the two organisms. Both genes have been sequenced (Lorincz & Reed, 1984; Hindley & Phear, 1984) and a comparison of their predicted amino acid sequences shows that 159 they are identical at 62% of their amino acid positions. This suggests that the two genes are related by common descent. The two yeasts are not closely related: sequence comparisons suggest that they may have diverged as much as 1000 million years ago, close to the origin of eukaryotic life (Huysmans, Dams, Vandenberghe & de Wachter, 1983). Therefore, any function that is conserved in both yeasts is likely to be found in other organisms and it is possible that a similar control mechanism exists in all eukaryotic cells. Cell cycle regulation in yeast g2 m ito tic c o n tr o l A second major cell cycle control has been identified in the fission yeast (Nurse, 1975). This acts during the Gz period and determines the cell cycle timing of mitosis. The control was identified as the major rate-limiting step in G2 by isolating ‘wee’ mutants, which are advanced through Gz and undergo mitosis and cell division prematurely at a reduced size. The mutants define two genes weel and cdc2 (Nurse & Thuriaux, 1980). The cdc2 gene is required twice during the cell cycle, in G\ at start, as discussed earlier, and then in Gz at the mitotic control. Recessive temperature-sensitive lethal cdc2 mutants become blocked in Gi or Gz, depending on where they are in the cell cycle when shifted to the restrictive temperature (Nurse & Bissett, 1981). In contrast, dominant wee cdc2 and partially recessive weel~ mutations advance cells through Gz- The weel gene codes for an inhibitor and the cdc2 gene for an activator in the mitotic control (Nurse & Thuriaux, 1980). Changes in the medium or growth rate of the cells modulate the mitotic control. Shifting cells into poor medium transiently advances cells through Gz as do wee mutants (Fantes & Nurse, 1977). This can be explained if a critical cell size is required before the mitotic control can be passed, which is modulated downwards in poor growth medium. As a consequence, cells undergo cell division at a small size and are too small to pass start in the subsequent cell cycle. The G\ period therefore becomes expanded before the start point of commitment. With further nutrient deprivation these uncommitted cells will undergo conjugation and sporulation or enter stationary phase. In budding yeast there does not appear to be an equivalent of this control because as explained earlier there is no G2 period. The cdc28 gene is required for mitosis (Piggott, Rai & Carter, 1982) but its mitotic function is completed very soon after start, consistent with the early initiation of mitosis. There is an approximately constant time from S phase to the completion of mitosis and thus a certain period of time may be required to complete all the events of mitosis (Carter & Jagadish, 1978). It is this that determines when mitosis is completed. cdc2/cdc28 g e n e f u n c t i o n The cdc2/cdc28 gene has a remarkable role in cell cycle control of the two yeasts. It is involved in the G \/S phase transition at start and then again at the G2/M transition of the mitotic control in fission yeast. Several powerful molecular th e J . Hayles and P. Nurse 160 biological techniques have been used to investigate this gene and we shall now summarize what these recent studies have shown concerning the molecular role of this gene function in the cell. Plasmids containing the cdc2 and cdc28 genes have been isolated from gene banks by virtue of their ability to allow mutants of cdc2 and cdc28 to grow at their restrictive temperatures (Nasmyth & Reed, 1980; Beach et al. 1982). The sequence of the cdc28 gene reveals that it contains a single open reading frame encoding a protein of 298 amino acids (Lorincz & Reed, 1984). A similar protein of 297 residues is predicted from the sequence of the cdc2 gene, although in this case four introns have to be postulated (Hindley & Phear, 1984). The two proteins have a 20% identity with the src family of oncogene products and with cyclic-AMP-dependent protein kinase. Most of this similarity can be accounted for by the presence of two active sites within the protein: an ATP-binding site of the sort found in protein kinases and a phosphorylation site (Nurse, 1985a). Therefore, it is likely that the cdc2/28 gene function is a protein kinase that can become phosphorylated. Because of its similarity to the src family of oncogenes the cdc2/28 gene could be considered a primitive oncogene. We do not think that this is likely since most of the similarity is due to the presence of the two active sites, and once these sites are removed from the comparison there is very little further similarity. Antibodies have been raised against a afc25-/3-galactosidase fusion, which identifies a budding yeast protein of the correct relative molecular mass (M r) of 36000 (Reed, 1984). This protein has been shown to be the product of the cdc28 gene, since in strains containing the cdc28 gene on a multicopy number plasmid the protein is increased in amount (Reed, Hadwiger & Lorincz, 1985). The cdc28 gene product is phosphorylated and also has protein kinase activity. Exponentially growing cells labelled with 32PC incorporate the isotope into the p36 protein. The protein kinase activity has been assayed by following the transfer of phosphate to a protein of Mr = 40000, which is co-immunoprecipitated with the cdc28 gene product. The 40000 Mr protein may be the natural substrate for the cdc28 protein kinase. Extracts made from temperature-sensitive cdc28 mutants also have temperature-sensitive activity in vitro, proving that the protein kinase activity is definitely associated with the cdc28 gene product. The protein kinase activity is magnesium-dependent and apparently zinc-activated. Antibodies have also been raised against the cdc2 gene product by using synthetic peptides (V. Simanis & P. Nurse, unpublished data). These identify a protein from fission yeast with a relative molecular mass of 34 000. This protein is overproduced in cells with an increased level of cdc2 transcripts produced by fusion of the cdc2 gene with the strong alcohol dehydrogenase (adh) promoter. The protein is phosphory lated like the cdc28 gene product and also has protein kinase activity. The kinase activity is divalent-cation-dependent but in contrast with cdc28 is not activated by zinc. These data confirm that the cdc2lcdc28 gene product is indeed a protein kinase that becomes phosphorylated. The cdc2 gene is not regulated at the level of transcription (Durkacz, Carr & Nurse, 1986). No changes in steady-state transcript level are seen during the cell >4 161 cycle or when cells undergo the transition from rapid growth into quiescence. Nor are there any changes at the level of RNA processing; all four introns appear to be used all the time. Therefore, changes in cdc2 transcript level or its processing do not appear to be responsible for regulating either progress through the cell cycle or the transition between rapid proliferation and quiescence. Even when cdc2 transcription is increased 30-fold over normal there is no disturbance of normal regulation of the mitotic cycle. The cdc2 antibody has been used to monitor changes in protein level and phosphorylation state of the cdc2 gene product (V. Simanis & P. Nurse, unpublished data). Surprisingly, no changes in either steady-state protein level or overall phosphorylation state are found during the cell cycle. However, a dramatic change is observed when cells become quiescent upon shift into nitrogen starvation conditions. The level of cdc2 protein does not change, but it becomes rapidly dephosphorylated and there is a dramatic reduction in the level of its protein kinase activity. Therefore, exit of cells from the mitotic cell cycle into a quiescent state may well be regulated by modulating the cdc2 protein kinase activity by changes in phosphorylation. The reduction in protein kinase activity would then prevent the cell from initiating any further mitotic cell cycles. The regulation of the phosphorylation state of the cdc2 gene product may be the key to understanding how the cell cycle is initiated. Also, identification of substrates should be revealing about what the cdc2/28 gene protein does and how it mediates the G \/S phase transition at start and the Gz/M transition at the mitotic control (Dudani & Prasad, 1984). Cell cycle regulation in yeast cdc2(28 G EN E F U N C T IO N Gene products that interact with the cdc2/28 gene function are potential regulators and substrates. Two gene functions that interact closely with the cdc2 function at the mitotic control point are those of weel and cdc25 (Fantes, 1979; Nurse & Thuriaux, 1980; Fantes, 1983). The weel gene codes for an inhibitor in the mitotic control as discussed earlier. The cdc25 gene has been implicated in the control because mutants in weel and wee mutants of cdc2 are able to suppress a defective cdc25 gene function (Fantes, 1979). Normally, mutants of cdc25 are unable to undergo mitosis but in a wee mutant background they are able to do so. The suppressor effects on cdc25 suggest a close relationship between cdc25, weel and cdc2, and several models have been proposed to account for their close interaction. Direct evidence in favour of cdc25 being involved in the mitotic control has recently been obtained after cloning the gene (Russell & Nurse, 1986). When the cdc25 gene is fused behind the strong adh promoter, thus increasing its transcription, cells are advanced through Gz into mitosis and cell division at a reduced size. This establishes that the cdc25 gene function can be rate-limiting for the timing of mitosis. The cdc25 gene function is also completely dispensable in a weel mutant background. Deletion of the chromosomal copy of the cdc25 gene using recombinant DNA procedures is lethal in a weel+ strain whereas a weel~ strain is still viable. The similarity of the phenotype of cdc2 wee mutants, weel mutants and G E N E S IN T E R A C T I N G W I T H THE 162 J . Hoyles and P. Nurse of overproducing cdc25 suggests that the weel and cdc25 gene products are regulators of the cdc2 protein kinase. The cdc25 gene codes for an inducer and the weel gene for an inhibitor. If cdc25 is overproduced in a weel mutant then the mitotic control is grossly disturbed, leading to various mitotic catastrophes and cell death. These data are all consistent with the following model. The cdc2 gene product is considered to be in two states, an active one and an inactive one. The cdc25 gene product converts the inactive form into the active form and the weel gene product converts the active form into the inactive form. The balance of these two gene product activities determines when the cdc2 gene product becomes sufficiently active to initiate mitosis. Another gene product appears to interact directly with the cdc2 protein at both its G\ and its Gz points of action. This gene, called sucl (for swppressor of cdc2) was identified by mutants that allow temperature-sensitive lethal mutants of cdc2 to grow at their restrictive temperature (Hayles, Beach, Durkacz & Nurse, 1986). The gene has been cloned by exploiting the fact that overproduction of sucl+ transcripts from a high copy-number plasmid enables certain cdc2 mutants to divide. The sup pression is also seen when the sucl+ gene is fused to the strong adh promoter. Overproduction of the sucl+ transcript, and therefore presumably of the sucl+ gene product, in a cdc2+ strain causes a delay of mitosis. This suggests that the sucl+ gene product has a direct effect in its own right upon cell cycle progress and may physically interact with the cdc2 protein. Further evidence in favour of this hypothesis is that sucl suppression of cdc2 mutants is allele-specific. The role that the sucl gene product may play is as yet unclear but it is an obvious candidate for a protein that directly interacts with the cdc2 protein. Various classes of mutants appear to act at a similar time to the start genes already mentioned. They include cdc37 in budding yeast and cdclO in fission yeast. The whil and whi2 (Sudbery, Goodey & Carter, 1980) mutants in budding yeast are advanced into start at a reduced cell size so these may be important for regulating cdc28 gene function. Mutants of the rani/pat 1 gene (Nurse, 19856; lino & Yamamoto, 1985a,b) in fission yeast may also function close to the cdc2 point of action at start. These mutants undergo conjugation and sporulation in conditions under which cells should be progressing through the mitotic cell cycle. They appear to be defective in the monitoring process that determines whether a cell should become committed to conjugation and sporulation or mitosis. Study of all these various mutants may turn out to be revealing about how cdc2/28 gene function is regulated. A N A L O G O U S CO NT RO L S IN OTHER E UKARYOTES The work we have described indicates that there are two major cell cycle controls in the yeasts that both involve the cdc2/28 protein kinase function: a commitment control acting at the G \/S boundary and, at least in fission yeast, a mitotic control acting at the Gz/M boundary. We will now consider whether there are any analogous controls present in other eukaryotic cells. Cell cycle regulation in yeast 163 Many workers studying animal cells have proposed that the major point of cell cycle regulation occurs in late G\ (Pardee, Dunbrow, Hamlin & Kletzien, 1978). These proposals have been based mainly on the observations that arresting cell proliferation by using a variety of conditions such as serum starvation, nutrient depletion or high cell density leads to G\ arrest. A good example of this type of analysis is provided by the work of Pardee and his co-workers (Pardee, 1974), who put forward the hypothesis that there is a restriction point, R, in late G\, which can only be passed in conditions favourable for cell proliferation. When conditions are not favourable cells become arrested at the R point and cannot enter the mitotic cycle. Transformed cells may be defective in R point control, failing to arrest correctly in unfavourable conditions. Such an R point in G\ has obvious similarities to the start commitment control in the yeasts. There are other features about G\ arrest points in animal cells that also show similarities. It has been shown that in certain cases of differentiation cells leave the mitotic cycle from a specific point in Gy (Scott, Florine, Willie & Yun, 1982). In an in vitro adipocyte differentiation system, cells become committed to differ entiation from a point in G\ called Gd- Once past this point they are unable to differentiate until they have completed the mitotic cycle in progress. However, it is unlikely that a general argument of this sort can be made, since many cell types undergoing differentiation still continue to proliferate. Traverse of a G\ arrest point is also rate-limiting for entry into S phase in certain growth conditions; the G\ period usually becomes much extended at low growth rates (Darzynkiewicz, Traganos & Melamed, 1980). The major rate-limiting step may be the attainment of a critical cell size as in the yeasts, but the evidence for mammalian cells is conflicting and if there is a relationship between cell size and entry into the mitotic cycle it must be of a much looser kind and easily dissociated (Shields et al. 1978). What is clear from cell fusion experiments, though, is that factors accumulate during the G\ period that determine the rate at which cells initiate S phase (Rao, Sunkara & Wilson, 1977; Rao & Johnson,1974). Mitotic controls have received much less attention in mammalian cells, although some cell types do become arrested at some point in G2 (Gelfant, 1962; Pederson & Gelfant, 1970). These cells are only a small proportion of the total population, but when stimulated to divide they undergo mitosis without first entering 5 phase. Mammalian cell fusion experiments also suggest that there are inhibitory factors that can block cells in G2 before mitosis (Rao & Johnson, 1974). There have been more extensive studies of mitotic control in the slime mould Physarum plasmodium. Fusion of naturally synchronous plasmodia at different times of G2 have indi cated that mitotic inducing factors accumulate during the G2 period (Rusch, Sachsenmaier, Behrens & Gruter, 1966). Such factors have been purified and found to advance nuclei into mitosis (Loidl & Grobner, 1982). The factors can be mimicked by Ehrlich ascites cell extracts containing protein kinase activity able to phosphorylate histone HI in vitro (Inglis et al. 1976), suggesting a role for protein phosphorylation in the mitotic control. A formally similar approach has been used recently by Kirschner and his co-workers to investigate mitotic control in Xenopus 164 J . Hayles and P. Nurse eggs. They have investigated the role of maturation promoting factor (MPF) in mitosis and meiosis. MPF was first identified as a factor required for maturation of Xenopus oocytes; the maturation process involves the transition of G -arrested oocytes into meiotic nuclear division. MPF also varies cyclically during the mitotic cycle in early Xenopus embryos, peaking at mitosis (Gerhart, Wu & Kirschner, 1984; Newport & Kirschner, 1984). This suggests that MPF may be a component of a regulatory system controlling the initiation of mitosis. Preparations containing MPF activity have been isolated from a variety of organisms including mitotically arrested yeast and mammalian cells (Weintraub et al. 1982; Sunkara, Wright & Rao, 1979), and so this regulatory system may be widespread amongst eukaryotic cells. It is also possible that MPF may involve phosphorylation, since its activity is stabilized by phosphatase inhibitors, and a fraction that is 20-fold enriched in MPF has protein kinase activity (Wu & Gerhart, 1980). Other evidence also indicates that protein phosphorylation may play an important role in regulating progress through the eukaryotic cell cycle. A number of pro teins have been identified that only become phosphorylated during mitosis (Sahasrabuddhe, Adlakha & Rao, 1984; Davis, Tsao, Fowler & Rao, 1983; Vandre, Davis, Rao & Borisy, 1984). Antibodies have been raised against these proteins and a subset of these have been shown to be associated with microtubular organizing centres, which are responsible for organizing the mitotic spindle. A variety of other proteins including histones also undergo changes in phosphorylation during the cell cycle (Ottaviano & Gerace, 1985; Piras & Piras, 1975; Bradbury, Inglis & Matthews, 1974; Gurley, Walters, Barham & Deaven, 1978: Yamashita, Nishimoto & Sekiguchi, 1984). Taken together these data establish that there are G\ and Gz control points in other eukaryotic cells, including mammalian cells, which may be analogous to the two controls identified in the two yeasts. Furthermore, phosphorylation has been implicated in certain cases, suggesting that protein kinases that are possibly similar to that encoded by cdc2l28 may have an important role to play in the controls. 2 cdc2/28 F U N C T IO N We conclude this review by speculating on the molecular basis of the cdc2/28 function and its regulation. The best clue for regulation obtained so far is the observation that phosphorylation and protein kinase activity of the cdc2 protein is much reduced when cells become quiescent and accumulate before start. Therefore phosphorylation and activation of the cdc2 protein kinase may regulate entry into the mitotic cycle and the transition from G\ into S phase. This raises the question of how the phosphorylation of the cdc2/28 protein is controlled. One exciting possibility is provided by recent studies of the ras oncogene homologues and cyclic-AMPdependent protein phosphorylation in budding yeast. Mutants with reduced levels of cyclic-AMP-dependent protein phosphorylation accumulate in G\ before start (Matsumoto, Uno, Oshima & Ishikawa, 1982). These observations have led to the model that cyclic-AMP-dependent protein phosphorylation may regulate the R E G U L A T I O N OF T H E 165 initiation of the mitotic cell cycle. Furthermore, the control of phosphorylation may be determined by the ras oncogene homologues modulating adenylate cyclase activity and thus cyclic AMP levels in the cell (Toda et al. 1985). An obvious candidate for activation by this phosphorylation system is the cdc2/28 function, which could then proceed to initiate the cell cycle. This is an intriguing possibility since it would provide a direct link between the ras oncogene and the control of cell proliferation. However, the situation may be more complex than this. The mutants defective in cyclic-AMP-dependent protein phosphorylation that accumulate in G\ are also defective in overall growth of the cell (Matsumoto et al. 1982). As explained earlier, cells that grow at a reduced rate often accumulate in G\. Therefore, the G\ block point of these mutants may simply be a pleiotropic effect very distant from the initial biochemical defect. The critical experiment in these studies concerns the bcyl gene, mutants of which result in a deficiency of the regulatory subunit of cyclic-AMPdependent protein kinase. These mutants have a cyclic-AMP-dependent protein kinase, which is now cyclic-AMP-independent, and consequently have a constitutive kinase activity. If the model is correct these mutants should undergo cell pro liferation even when conditions are not appropriate, for example under nutrient deprivation. Mutants that behave in this way do exist, for example whi2 mutants, continue to proliferate even when they have run out of nutrients. They continue to bud and divide and produce very small cells, bcyl mutants do not seem to have this phenotype. It has been claimed that on starvation bcyl mutants continue budding and eventually arrest in a budded state (Matsumoto, Uno & Ishikawa, 1983). However, if the results are examined more carefully it appears that cells do not continue to bud, they simply fail to complete the mitotic cycle that is in progress when shifted into nutrient starvation conditions. This can be seen by the low increase in cell number observed in these mutants when shifted into nutrient starvation conditions. They fail to complete the cell cycle in progress and thus arrest as budded cells. Therefore, an alternative interpretation of the data is that the cells are sick and are unable to complete the cell cycle in progress in starvation conditions. Until these points are clarified the roles of cyclic AMP and the ras gene product in cell cycle regulation, and thus their effects on the cdc2/28 gene function and overall phosphorylation, will remain uncertain. Overall phosphorylation of the cdc2/28 protein is unlikely to be responsible for regulating the mitotic control since phosphorylation levels are unchanged during G2 and mitosis. Of course, it is possible that there may be a more subtle change in phosphorylation not detected to date, such as different amino acid residues becoming modified. Alternatively, the cdc2)28 gene function could be regulated by substrate availability. Changes in compartmentation of the cdc2/28 protein or its substrates may be responsible for determining the cell cycle timing of mitosis. Cell cycle regulation in yeast cdc2/28 F U N C T IO N What could the single cdc2/28 gene function be doing that facilitates both the transitions from G\ into S and from Gz into M ? Many different changes occur as cells M O L E C U L A R BASIS OF T H E 166 J . Hayles and P. Nurse progress through the cell cycle and it is difficult to identify those processes that may be important. We suggest that the most important processes that occur are those central to DNA replication and chromosome segregation. One such process is a change in the chromatin condensation state. Chromatin must decondense to become available for replication at 5 phase, and then condense in mitosis to permit segregation of the chromosomes. The hypothesis has been suggested that there is a cycle of chromatin condensation and decondensation (Mazia, 1963): when chroma tin is maximally decondensed 5 phase occurs and when it is maximally condensed mitosis takes place. Decondensation of chromatin may make the DNA available to a constantly present replication apparatus giving rise to a discrete S phase (Newport & Kirschner, 1984). A second important process concerns alterations in the cytoskeleton. Changes occur in the cytoskeleton during the cell cycle, the most dramatic being the disappearance of cytoplasmic microtubules at the Gz/M transition and the formation of a mitotic spindle, which is necessary for accurate segregation of the sister chromatids. A transient disassembly of microtubules has also been suggested as a necessary requirement for the initiation of 5 phase (Otto et al. 1979; Friedkin, Legg & Rozengurt, 1979; Crossin & Carney, 1981; Thyberg, 1984). Therefore, there may well be changes in both chromatin and cytoskeletal organization at the G\/S and Gz/M transitions, and we suggest that these changes may be brought about by the cdc2/28 protein kinase phosphorylating specific proteins. Obvious candidates for these proteins are those associated with chromatin and the cytoskeleton; for example, both histones and microtubular associated proteins (MAPs) are believed to be important for organization of chromatin and the cytoskeleton. Several of the histones undergo substantial changes in phosphorylation during the cell cycle and may be partly responsible for the state of chromatin condensation (Dolby, Belmont, Borun & Nicolini, 1981; Gurley et al. 1978; Bradbury et al. 1974; Hanks, Rodriquez & Rao, 1983). Both MAPs 1 and 2 can become phosphorylated (Herrmann, Dalton & Wiche, 1985) and their level of phosphorylation may be important for levels of microtubular disassembly. The cdc2/28 protein when activated could phosphorylate a series of proteins within the cell which could bring about changes in chromatin and cytoskeletal organization. The reason for the different behaviour at G\/S and Gz/M is not clear but could be due to different availability of substrates. The cdc2/28 protein at the apex of a protein phosphorylation cascade can potentially control the two major transitions involved in cell cycle control. We propose that this is how the cell cycle is regulated in yeast and knowledge gained from these studies may provide a useful framework for studying cell cycle regulation in higher eukaryotes. We thank Steve Aves, Tony Carr, Iain Hagan, Pete Magee, Paul Russell and Viesturs Simanis for helpful discussions and for supplying unpublished data. Cell cycle regulation in yeast 167 REFERENCES A ves, S. J., D u rk a c z , B. W., C a rr, A. & N u rse , P. (1985). Cloning, sequencing and transcriptional control of the Schizosaccharomyces pombe cdclO “start” gene. EMBO J. 4, 457-463. B each , D ., D u rk a c z , B. & N u rse , P. (1982). Functionally homologous cell cycle control genes in budding and fission yeast. Nature, Lond. 300, 706-709. B ra d b u ry , E. M ., In g lis , R. J. & M a tth e w s, H. R. (1974). Control of cell division by very lysine rich histone (FI) phosphorylation. Nature, Lond. 247, 257-261. B u ckin g-T h ro m , E., D u n tz e , W., H a r tw e ll, L. H . & M an n ey , T . R. (1973). Reversible arrest of haploid yeast cells at the initiation of DNA synthesis by a diffusible sex factor. Expl Cell’Res. 76, 99-110. C a r te r , B. L. A. & Ja g a d ish , M. N. (1978). The relationship between cell size and cell division in the yeast Saccharomyces cerevisiae. Expl Cell Res. 112, 15-24. C ro ssin , K. L. & C a rn ey , D. H. (1981). Evidence that microtubule depolymerization early in the cell cycle is sufficient to initiate DNA synthesis. Cell 23, 61—71. D a rz y n k ie w ic z , Z., T ra g a n o s , F. & M elam ed , M . R. (1980). New cell cycle compartments identified by multiparameter flow cytometry. Cytometry 1, 98-109. D avis, F. M., T s a o , T. Y., F o w le r, S. K & R ao, P. N. (1983). Monoclonal antibodies to mitotic cells. Proc. natn. Acad. Sci. U.SA. 80, 2926-2930. D o lb y , T. W., B elm o n t, A ., B o ru n , T. W & N ico lin i, C. (1981). D N A replication, chromatin structure and histone phosphorylation altered by theophylline in synchronised HeLa S3 cells. J. Cell Biol. 89, 78-85. D u d a n i, A. K. & P ra s a d , R. (1984). A hypothesis for the possible involvement of microtubules and protein kinase in the mechanism of action of cdc28 gene product of Saccharomyces cerevisiae. FEBS Lett. 172, 139—141. D u rk a c z , B., C a rr, A. & N u rse , P. (1986). Transcription at the cdc2 cell cycle control gene of the fission yeast Schizosaccharomyces pombe. EMBOJ. 5 (in press). F a n te s , P. A. (1977). Control of cell size and cycle time in Schizosaccharomyces pombe. J. Cell Sci. 24, 51-67. F a n te s , P. A. (1979). Epistatic gene interactions in the control of division in fission yeast. Nature, Lond. 279, 428-430. F a n te s , P. A. (1983). Control of timing of cell cycle events in fission yeast by the weel+ gene. Nature, Lond. 302, 153-155. F a n te s , P. & N u rs e , P. (1977). Control of cell size in fission yeast by a growth modulated size control over nuclear division. Expl Cell Res. 107, 377-386. F rie d k in , M., L eg g , A. & R o z e n g u rt, E. (1979). Antitubulin agents enhance the stimulation of DNA synthesis by polypeptide growth factors in 3T3 mouse fibroblasts. Proc. natn. Acad. Sci. U.SA. 76, 3909-3912. G e r h a r t, J., Wu, M. & K irs c h n e r, M. W. (1984). Cell cycle dynamics of a M phase specific cytoplasmic factor inXenopus laevis.J. Cell Biol. 98, 1247-1255. G elfa nt , S. (1962). Initiation of mitosis in relation to the cell division cycle. Expl Cell Res. 26, 395-403. G u rle y , L. R., W a lte r s , R. A., B arham , S. S. & D eav en , L. L. (1978). Heterochromatin and histone phosphorylation. Expl Cell Res. I l l , 373—383. H a n k s, S. K., R o d riq u e z , L. V. & R ao P. N. (1983). Relationship betw een histone phosphorylation and prem ature chrom osom e condensation. Expl Cell Res. 148, 293-302. H a r tw e ll, L. H . (1974). Saccharomyces cerevisiae cell cycle. Bad. Rev. 38, 164-198. H a r tw e ll, L. H ., M o rtim er, R. K., C u lo tti, J. & C u lo tti, M . (1973). Genetic control of cell division in yeast: Genetic analysis of cdc mutants. Genetics 74, 267-286. H a r tw e ll, L. H . & U n g e r, M. W. (1977). Unequal division in Saccharomyces cerevisiae and its implications for the control of cell division. J. Cell Biol. 75, 422-435. H a y le s , J., B each , D ., D u rk a c z , B. & N u rse , P. (1986). The fission yeast cell cycle control gene cdc2 isolation of a sequence sucl that suppresses cdc2 mutant function. Molec. gen. Genet. (in press). : 168 jf. Hayles and P. Nurse H e rrm a n n , H ., D a lto n , J. M. & W iche, G . (1985). Microheterogeneity of microtubule associated proteins MAPI and MAP2 and differential phosphorylation of individual subcom ponents.^. biol. Chem. 260, 5797-5803. H in d le y , J. & P h e a r, G. A. (1984). Sequence of the cell division gene CDC2 from Schizosaccharomyces pombe; patterns of splicing and homology to protein kinases. Gene 31, 129-134. H ira o k a , Y ., T o d a , T . & Y an ag id a , M. (1984). NDA3 gene of fission yeast encodes /3 tubulin: a cold sensitive nda3 mutation reversibly blocks spindle formation and chromosome movement in mitosis. Cell 39, 349-358. H u ysm an s, E., D am s, E., V an d en b erg h e, A. & D e W a c h te r, R. (1983). The nucleotide sequences of the 5S RNAs of four mushrooms and their use in studying the phylogenetic positions of basidiomycetes among the eukaryotes. Nucl. Acids Res. 11, 2871-2880. Iin o , Y. & Y am am oto, M. (1985a). Mutants of Schizosaccharomyces pombe which sporulate in the haploid state. Molec. gen. Genet. 198, 416-421. Iin o , Y. & Y am am oto, M. (19856). Negative control for the initiation of meiosis in Schizosaccharomyces potnbe. Proc. natn. Acad. Sci. U.SA. 82, 2447-2451. I n g lis , R. J., L a n g a n , T. A., M a tth e w s, H . R., H a rd ie , D . G. & B ra d b u ry , E. M . (1976). Advance of mitosis by histone phosphokinase. Expl Cell Res. 9 7 418—425. L o id l, P. & G ro b n e r, P. (1982). Acceleration of mitosis induced by mitotic stimulators of Physarum polycephalum. Expl Cell Res. 137, 469—472. L o rd , P. G . & W h e a ls, A. E. (1980). Asymmetrical division of Saccharomyces cerevisiae. J. Bacteriol. 142, 808—818. L O rin cz, A. T. & R eed , S. I. (1984). Primary structure homology between the product of yeast cell division control gene CDC28 and vertebrate oncogenes. Nature, Lond. 307, 183-185. M atsu m o to , K., U n o , I. & Ish ik aw a, T . (1983). Control of cell division in Saccharomyces cerevisiae mutants defective in adenylate cyclase and cAM P dependent protein kinase. Expl Cell Res. 146, 151-161. M a tsu m o to , K ., U n o , I., O shim a, Y. & Ish ik aw a, T . (1982). Isolation and characterisation of yeast m utants deficient in adenylate cyclase and cA M P-dependent protein kinase. Proc. natn. Acad. Sci. U.SA. 79, 2355-2359. M a z ia , D. (1963). Synthetic activities leading to mitosis. J. cell. comp. Physiol. 1 (suppl. 1), 123-140. M o ir, D., S te w a rt, S. E., O sm ond, B. C. & B o tste in , D. (1982). Cold-sensitive cell division cycle mutants of yeast: Isolation, properties and pseudoreversion studies. Genetics 100, 547-563. N asm y th , K. (1979). A control acting over the initiation of DNA replication in the yeast Schizosaccharomyces pombe. J. Cell Sci. 36, 215-233. N asm y th , K. & N u rs e , P. (1981). Cell division cycle mutants altered in D N A replication and mitosis in the fission yeast Schizosaccharomyces pombe. Molec. gen. Genet. 182, 119—124. N asm y th , K., N u rse , P. & F ra s e r, R. S. S. (1979). The effect of cell mass on the cell cycle timing and duration of 5-phase in fission yeast. J. Cell Sci. 39, 215-233. N asm y th , K . A. & R eed, S. I. (1980). Isolation of genes by com plem entation in yeast: M olecular cloning of a cell cycle gene. Proc. natn. Acad. Sci. U.SA. 77, 2119-2123. N e w lo n , C. S., P e te s, T . D ., H e re fo rd , L . M. & F an gm an, W . L . (1974). Replication of yeast chromosomal D N A . Nature, Lond. 247, 32-35. N e w p o rt, J. W. & K irs c h n e r, M. W. (1984). Regulation of the cell cycle during earlyXenopus development. Cell 37, 731-742. N u rs e , P. (1975). Genetic control of cell size at cell division in yeast. Nature, Lond. 256, 547-551. N u rs e , P. (1985a). Cell cycle control genes in yeast. Trends Genet. 1, 51-55. N u rs e , P. (19856). M utants of the fission yeast Schizosaccharomyces pombe w hich alter th e shift betw een cell proliferation and sporulation. Molec. gen. Genet. 198, 497-502. N u rs e , P. & B isse tt, Y. (1981). Gene required in G\ for commitment to cell cycle and in Gz for control of mitosis in fission yeast. Nature, Lond. 292, 558-560. N u rs e , P. & T h u ria u x , P. (1977). Controls over the timing of DNA replication during the cell cycle of fission yeast. Expl Cell Res. 107, 365-375. N u rs e , P. & T h u ria u x , P. (1980). Regulatory genes controlling mitosis in the fission yeast Schizosaccharomyces pombe. Genetics 96, 627-637. , 169 Cell cycle regulation in yeast N u r se , P., T h uriaux , P. & N asmyth , K. (1976). Genetic control of cell division cycle in fission yeast Schizosaccharomyces pombe. Molec. gen. Genet. 146, 167-178. Y. & G , L. (198S). Phosphorylation of nuclear lamins during interphase and mitosis. J. biol. Chem. 1, 624-632. O tto , A. M., Z um be , A ., G ibson , L., K u bler , A. M. & D e A su a , L. J. (1979). Cytoskeletondisrupting drugs enhance effect of growth factors and hormones on initiation at D N A synthesis. Proc. natn. Acad. Sci. U.SA. 76, 6435-6438. P ardee , A. B. (1974). A restriction point for control of normal animal cell proliferation. Proc. natn. Acad. Sci. U.SA. 71, 1286-1290. P a rdee , A. B., D u nbrow , R., H am lin , J. L. & K letzien , R. F. (1978). Animal cell cycle. A. Rev. Biochem. 47, 715-750. P ed erson , T . & G elfant , S. (1970). G2 -population cells in mouse kidney and duodenum and their behaviour during the cell division cycle. Expl Cell Res. 59 32-36. P iggott , J. R ., R a i , R. & C arter , B. L. A. (1982). A bifunctional gene product involved in two phases of the yeast cell cycle. Nature, Lond. 298, 391-394. P ira s, R. & P ira s, M. M. (1975). Changes in microtubule phosphorylation during cell cycle of HeLa cells. Proc. natn. Acad. Sci. U.SA. 72, 1161-1165. P r in g le , J. R. & H artw ell , L. H. (1981). The Saccharomyces cerevisiae cell cycle. In The Molecular Biology of the Yeast Saccharomyces - Life Cycle and Inheritance (ed. J. N. Strathern, E. W. Jones & J. R. Broach), pp. 97-142. New York: Cold Spring Harbor Laboratory Press. R a o , P. N. & J ohn son , R. T. (1974). Regulation of cell cycle in hybrid cells. In Control of Proliferation in Animal Cells, Cold Spring Harbor Conf. Cell Prolif. (ed. B. Clarkson & R. Baserga), vol. 1, pp. 785-800. New York: Cold Spring Harbor Laboratory Press. R a o , P. N., S unkara , P. S. & W il so n , B. A. (1977). Regulation of DNA synthesis: age dependent cooperation among G\ cells upon fusion. Proc. natn. Acad. Sci. U.SA. 74, 2869-2873. R e e d , S. I. (1980). The selection of S. cerevisiae mutants defective in the start event of cell division. Genetics 95, 561—577. R e e d , S. I. (1984). Preparation of product specific antisera by gene fusion: antibodies specific for the product of yeast cell division cycle gene CDC28. Gene 20, 255—265. R e e d , S. I., H adw iger , J. A. & L ò rincz , A. T. (1985). Protein kinase activity associated with the product of the yeast cell division cycle gene CDC28. Proc. natn. Acad. Sci. U.SA. 82, 4055-4059. R u sc h , H. P., S achsenm aier , W., B eh ren s , K . & G ruter , V. (1966). Synchronization of mitosis by the fusion of the plasmodia of Physarum polycephalum. J. Cell Biol. 31, 204-209. R u sse l l , P. & N u r se , P. (1986). cdc25+ functions as an inducer in the m itotic control of fission yeast. Cell (in press). S ahasrabu ddhe , C. G., A d lakha , R. C. & R ao , P. N. (1984). Phosphorylation of non-histone proteins associated with mitosis in HeLa cells. Expl Cell Res. 153, 439-450. S c o tt, R. E., F lo rin e , D. L., W illie , J. J. Jr & Y u n, K. (1982). Coupling of growth arrest and differentiation at a distinct state in the Gi phase of the cell cycle: Gd- Proc. natn. Acad. Sci. U.SA. 79, 845-849. OTTAVIANO, erace , S h ie l d s , R ., B rooks , R ., R id d l e , P. N ., C apellaro , D . F. & D elia , D . (1978). Cell size, cell cycle and transition probability in m ouse fibroblasts. Cell 15, 469-474. S huster , J. R. (1982). Mating-defective ste mutations are suppressed by cell division cycle start mutations in Saccharomyces cerevisiae. Molec. Cell Biol. 2, 1052-1063. S udbery , P. E., G oodey , A. R. & C arter , B. L . A. (1980). Genes which control cell proliferation in the yeast Saccharomyces cerevisiae. Nature, Lond. 288, 401-404. S unkara , P. S., W right , D. A. & R ao , P. N. (1979). Mitotic factors from mammalian cells induce germinal vesicle breakdown and chromosome condensation in amphibian oocytes. Proc. natn. Acad. Sci. U.SA. 76, 2799-2802. T o d a , T ., U n o , I., Ish ik aw a, T ., P o w ers, S., K a ta o k a , T ., B ro ck, D ., C am ero n, S., B ro a c h , J., M atsu m o to , K . & WlGLER, M . (1985). In yeast RAS proteins are controlling elements of adenylate cyclase. Cell 40, 27-36. 170 J . Hayles and P. Nurse T hyberg , J. (1984). T h e m icrotubular cytoskeleton and the initiation of D N A synthesis. Expl Cell Res. 155, 1 -8 . U m esono, K ., H ira o k a , Y ., T o d a , T . & Y an ag id a , M . (1983). V isualisation of chrom osom es in m itotically arrested cells of the fission yeast Schizosaccharomyces pombe. Curr. Genet. 7, 123-128. V a n d re , D . D ., D av is, F. M., R ao , P. N. & B orisy, G. G. (1984). Phosphoproteins are components of mitotic microtubule organising centres. Proc. natn. Acad. Sci. U.SA. 81, 4439-4443. W e in tra u b , H ., B u sc a g lia , M ., F e rre z , M ., W e ille r, S., B o u le t, A., F a b re , F . & B a u lie u , E . E . (1982). Mise en evidence d’une activité “M P F ” chez Saccharomyces cerevisiae. C.R. hebd. Séanc. Acad. Sci., Paris (series III) 295, 787-790. Wu, M. & G e r h a r t, J. C. (1980). Partial purification and characterisation of Maturation Promoting Factor from eggs oiXenopus laevis. Devi Biol. 79, 465-477. Y am ash ita, K., N ishim oto, T. & S ekiguchi, M. (1984). Analysis of proteins associated with chromosome condensation in Baby Hamster kidney cells. J. biol. Chem. 259, 4667-4671.