ABSTRACT
Gene products required for mitotic chromosome separation in the .fission yeast Schizosaccharo- myces pombe are described. They have been identified by two distinct strategies of mutant isolation, followed by gene cloning and immunochemical characterization of gene products. The roles of four representative genes, namely nda3+, nuc2+, top2+ and dis2+, encoding β-tubulin, a nuclear scaffold-like protein, DNA topoisomerase II and type-1 protein phosphatase, respectively, are discussed in regard to the mechanisms and control of chromosome separation.
INTRODUCTION
The eukaryotic cell cycle consists of two major events, one duplicating chromosomal DNA in S phase and the other separating the chromatid DNAs into daughter cells in M phase. Each step has to be very accurate, otherwise the progeny cells would receive impaired genetic information. In contrast to a great deal of knowledge accumulated to understand the fidelty of DNA replication, little is known about the regulatory mechanisms that determine how mitotic chromosomes are correctly disjoined.
In mitosis, the chromosomes and cytoplasmic organelles are divided into each of the two daughter cells. The series of the steps leading to chromosome separation are listed below and are similar in most eukaryotes.
Cytoplasmic microtubules disappear.
The nuclear envelope breaks down.
Chromosomes condense, so that individual chromosomes can be seen in prophase.
The elaborate spindle apparatus forms, and the chromosomes display considerable movement.
The chromosomes are aligned onto the metaphase plate.
The sister chromatids of all the chromosomes concertedly disjoin.
The separated chromosomes move toward the poles in anaphase A concomitant with shrinking kinetochore microtubules, followed by spindle elongation, which increases the distance between the poles in anaphase B.
The spindle disappears, leaving a central midbody region.
Chromosomes decondense and the daughter nuclei form.
Even in unicellular lower eukaryotes such as fungi and yeasts, most of these events occur (an exception is that the nuclear envelope does not break down in fungi and yeasts), indicating that the mechanisms of motisis are universal. Chromosome separation is also an attractive system in which to study the regulation of higher-order chromosome structures in terms of their dynamic changes in organization during the cell cycle. An intricate motile system involving microtubule shortening and sliding is required for chromosome separation. Understanding mitotic chromosome movement will be a prerequisite to the elucidation of the mechanisms underlying the fidelity of chromosome separation.
We have been investigating chromosome separation by identifying genes and their products required for this process. It seems likely that a large number of genes should be implicated in these events. However, we have focused our efforts on representative genes and have investigated their essential roles in chromosome separation.
The organism we have chosen for our investigation is the fission yeast Schizosac- charomyces pombe (Mitchison, 1970). This unicellular eukaryote is convenient for analysis of the cell division cycle and chromosome separation (reviewed by Hirano and Yanagida, 1989). S’, pombe cells divide by fission (not by budding, as is the case for Saccharomyces cerevisiae), and distinct cell cycle stages (G1, S, G2 and M) are identifiable. The G2/M transition is particularly clear. Both haploid and diploid cells are stable. There are three chromosomes per haploid cell (Kohli, 1987; Smith et al. 1987; Fan et al. 1989). The centromeric DNAs are very large and complex (Nakaseko et al. 1986; Chikashige et al. 1989). Cell length is maximal just prior to and during M phase. Upon entry into mitosis, cytoplasmic microtubules disappear (Hagan and Hyams, 1988) and a spindle forms in the nucleus (Fig. 1). Actin accumulates at the growing cell ends in interphase (Marks and Hyams, 1985; Marks et al. 1986), and during mitosis makes a division ring at the site for cytokinesis. The chromosomes are separated into daughter nuclei by spindle elongation.
The factors required for chromosome separation are classified into trans- and exacting factors. The former are proteins, the gene products, which act on all the chromosomes at a certain mitotic step(s) and are required for chromosome separation. The latter are DNA sequences, in particular regions of the chromosomes that are required for the separation of those chromosomes. The defects in transacting protein factors (such as tubulin or DNA topoisomerase II) result in the abnormal behaviour of all the chromosomes. On the other hand, the defects in exacting DNA sequences (such as centromere or telomere) may principally only affect the behaviour of those chromosomes having the defects.
Strategies for identification of the gene products
We have employed two strategies for identifying the traws-acting factors (Yanagida et al. 1986). Strategy 1 is schematized in Fig. 2. First we isolate >S. pombe mutants which are temperature-sensitive (ts, growth at 26 °C and restrictive at 36 °C) or coldsensitive (cs, growth at 36°C and restrictive at 20°C). Following this we employ a secondary screen to select phenotypically interesting mutations. Selection of the mutant phenotype to be looked for in the secondary screen is crucially important, and requires a great number of assumptions. Using this strategy, we have obtained a number of ts and cs mutants apparently defective in mitotic events, and have identified approximately 50 loci in the genome of S. pombe (Hirano and Yanagida, 1989).
The phenotypes of the mutants are characterized in further detail, and the defective mitotic stages are determined. Following this we clone the genomic DNA sequences that complement the ts or cs phenotype of a mutant by transformation, and determine by chromosome integration whether the cloned DNA is derived from the mutant locus (Rothstein, 1983). Sequencing of the cloned genes (and preferably also the mutant genes) enables us to predict the amino acid sequences of the gene products and search for homology to any known proteins using available databases. As S. pombe introns are short and have well-defined consensus sequences (Mertins and Gallwitz, 1987), it is generally possible to predict the amino acid sequences from nucleotide sequences of genomic DNAs. We raised rabbit antibodies against fusion proteins prepared from bacterial cells. Such antibodies recognize the antigens present in extracts of wild-type S. pombe strains. It is possible to identify or purify the gene products by immunochemical methods, or to determine their intracellular location in the >S. pombe cells by immunofluorescence microscopy. Another useful approach to elucidate protein function is to isolate homologous genes from other organisms. If the nucleotide sequence is well conserved in budding yeast or a mammalian organism (Lee and Nurse, 1987), then it seems reasonable to assume that the gene function can also be expected to be conserved. Thus strategy 1 begins by mutant isolation and can be completed by identification of protein function.
The other strategy (2) is to determine whether a known protein activity is required for chromosome separation. Mutants defective in a particular protein (activity) are isolated by assaying directly for the activity in a large number of mutagenized strains. Alternatively, mutants hypersensitive to specific inhibitors are isolated. If the mutants reveal phenotypes defective in chromosome separation, they are further investigated. Mutants of DNA topoisomerases have been isolated by this strategy (Uemura and Yanagida, 1984). Once the genes that complement mutations are isolated, the following steps are similar to strategy 1.
Genes identified by strategies 1 and 2
Some of the genes identified and cloned according to the procedures described above in our laboratory are listed in Table 1. Those involved in maintaining chromosome organization are also included. The number of amino acid residues for polypeptides predicted from the determined nucleotide sequences, molecular weights determined by immunoblots of SDS-PAGE using antisera against the fusion proteins, and the nature of the gene products are also shown in the Table. We raised a number of antisera and identified 12 different gene products in .S’. pombe extracts.
Phenotypes of chromosome separation
Phenotypes of representative mutants defective in chromosome separation under restrictive conditions are schematized in Fig. 3. The mutant cells are first exponentially grown at a permissive temperature, then transferred to a non-permissive temperature and incubated for the equivalent of two generations. If a highly uniform arrest phenotype is produced, the mutant cells are presumed to be blocked at a specific stage of the cell cycle. Classes of mutants apparently blocked at distinct stages of mitosis have been isolated. In the nda3–311 mutant, the cells are arrested with condensed chromosomes but without the mitotic spindle (Umesóno et al. 1983b). The arrested stage is similar to prophase or prometaphase. Gene cloning and sequencing showed that the nda3+ gene encodes β-tubulin (Hiraoka et al. 1984).
The phenotype of the nda2 mutant is similar to nda3, but a fraction of the rrutant cells at restrictive temperatures proceeds through a metaphase-like stage and are arrested immediately after chromosome separation (Toda et al. 1983; Umesono et al. 1983a). Nucleotide sequencing of the cloned genes that complement nda2 mutations show that the nda2+ gene encodes one of the two α-tubulins (Toda et al. 1984). A possible reason for the phenotypic difference between nda3 and nda2 mutants is described below.
A mutant called nuc2 was isolated by its characteristic arrest phenotype with condensed chromosomes at mitotic metaphase; the short uniform-sized spindle is formed but does not elongate (Hirano et al. 1988). The nucleus is displaced from the center of cell, as seen in nda2 and nda3 cells. The septum is formed but its separation does not occur. The gene product of nuc2+ will be described below.
The phenotype of ts top2 mutants makes a sharp contrast to those described above; most of the mitotic events are not arrested but chromosome separation fails (Ue mura and Yanagida, 1984, 1986). In top2 mutant cells at the restrictive temperature, aberrant chromosomes, which are not fully condensed and are topologically defective, are formed upon entry into mitosis. They are transiently pulled by the spindle, but not successfully separated. Because septum formation and cell separation are not inhibited to top2 cells the undivided nucleus is bisected by the septum producing two dead daughter cells (this phenotype of uncoupled nuclear division and cell separation is called ‘cut’). The mutants cutl, cut2 and cut10 exhibit phenotypes somewhat similar to that of top2 (Hirano et al. 1986). The phenotypes of these mutants are indistinguishable, showing transiently undivided chromosomes (in a configuration which is termed an ‘archery bow’-like structure).
A group of mutants called dis produces an intermediate phenotype between nda3 and top2. Three dis genes (disl, dis2 and dis3) have been identified so far (Ohkura et al. 1988). In these mutants, the spindle is formed and elongates, but chromosomes do not separate. Neither septum formation nor cell separation takes place. Nondisjoined chromosomes are apparently moved to the poles so that they are unequally distributed at the cell ends. The products of the dis+ genes will be described below.
Thus, uncoupled or uncoordinated mitoses are found in top2, cut, dis and also nuc2 mutants. Table 2 shows that spindle dynamics, septum formation and/or cell separation can take place in the absence of chromosome separation.
Role of tubulin genes in chromosome separation
The genome of S. pombe contains two α- (nda2 and atb2) and one /J-tubulin (nda3) genes (reviewed by Yanagida, 1987). Therefore in nda2 mutant cells, a small proportion of functional microtubules (consisting of atb2+ and nda3+ tubulin) may be formed even at the restrictive temperature (Adachi et al. 1986). This may explain the difference in the phenotypes of nda2 and nda3 mutations. In the nda3 mutant, functional microtubules would be completely absent, so that the spindle does not form, causing the arrest at a stage similar to prometaphase. In nda2, an incomplete spindle might be formed, which is able to separate chromosomes but unable to increase the pole to pole distance. As many long microtubules might be required for anaphase B spindle elongation, anaphase A and B would be distinct in regard to their dosage requirements for functional microtubules.
One cs allele nda3-311 is useful for analysis of mitotic events, as the mutant 3- tubulin is reversibly inactivated (Hiraoka et al. 1984). At restrictive temperatures, the cells are uniformly arrested but upon transfer to the permissive temperature, the cells proceed highly synchronously into anaphase; thus cellular /?-tubulin is rapidly reactivated and the spindle is immediately formed, followed by a highly concerted chromosome separation. The rate of spindle elongation is 1μmsec-1, roughly half that of wild-type spindle elongation (Tanaka and Kanbe, 1986; Hagan, I. and Hyams, J., personal communication).
Requirement of topoIl for chromosome condensation and separation
We isolated the ts topi and top2 mutants by strategy 2 (Uemura and Yanagida, 1984). The top2 mutants contain a heat-sensitive topoll enzyme irreversibly inactivated at 37°C. A cs top2 mutant which contained the topoll activity reversibly inactivated at 20°C was also isolated (Uemura et al. 1987a). Gene disruption experiments indicate that the top2 gene is essential, whereas the topi gene is dispensable (Uemura et al. 19876; Shiozaki, K. and Yanagida, M., unpublished). The defect of topi can be substituted by a sufficient amount of topoll.
The ts top2 mutants at the restrictive temperature produce abnormal chromosomes at mitosis; these are transiently extended into filamentous structures along the elongating mitotic spindle, but are not separated (Uemura and Yanagida, 1986). A primary defect in top2 appears to be the formation of aberrant mitotic chromosomes inseparable by the force generated by the spindle apparatus.
Reciprocal temperature-shift experiments using the double mutants of either ts or cs top2 and cs nda3 indicated that topoll is required for chromosome condensation and separation (Uemura et al. 1987a). A cs top2-cs-nda3 double mutant at 20°C shows long, entangled chromosomes, which condense and separate upon shift to the permissive temperature. If spindle formation is prevented at the permissive temperature, the chromosomes condense but do not separate. Thus topoll appears to be required for final chromosome condensation. Moreover pulse-shift experiments show that topoll is required for chromatid disjunction. Experiments with ts top2- cs-nda3 cells show that topoll is also required for chromosome separation in anaphase: inactivation of topoll and activation of /I-tubulin allow normal spindle formation but result in ‘streaked’ chromosomes. The topoll activity of decatenation/ catenation and knotting/unknotting of DNA appears to become essential during mitosis. DNA relaxing activity is abundant in the top2 mutant cells due to the presence of topo I.
In yeasts, topoll is not essential for replication and transcription if the amount of topol is sufficiently high (reviewed by Yanagida and Wang, 1987). If either topol or topoll is absent, replication and rRNA synthesis proceeds normally. Studies of the double mutant topl-top2 indicate that DNA relaxing activity is essential for replication, rRNA transcription and nucleolar organization. The effect of combining ts nucl (defective in the largest subunit of RNA polymerase I) and ts topl-top2 shows that the topol and II enzymes are required for folding of rDNA and RNA polymerase I molecules into the assembly of nucleolar genes to allow their function (Hirano et al. 1989).
The amino acid sequence of topoll, deduced from the cloned top2+ gene, reveals homology to prokaryotic type II DNA topoisomerase II, gyrase (Uemura et al. 1986). The NH2 terminal domain of topoll is similar to the ATP binding B-subunit of gyrase, the central COOH region resembles the DNA binding A-subunit and the COOH terminal domain consists of highly charged residues. The topoll sequences of other organisms such as budding yeast and human also show these three domains.
We are dissecting essential and non-essential domains in topoll to try to understand the role of the enzyme in mitosis. The mechanism of mitotic activation of the enzyme is particularly interesting. For this purpose, we are attempting to identify an alteration in the enzyme structure required for its mitotic activity. A possible one is phosphorylation and dephosphorylation of topo II.
Role of cut+ genes and their products
The phenotypes of cutl, cut2 and cutlO mutants are similar, suggesting that their gene functions are related (Hirano et al. 1986; Uzawa, S., unpublished results). Consistently, the cloned cutl+ gene is able to complement not only cutl but also cut2 and cut10 mutants. The cloned cut2+ and cut10+ genes, however, do not complement cutl mutants. The product of the cutl+ gene has recently been identified (Uzawa, S. et al. to be published). It is a minor nuclear protein of 210K (K = 103MR) containing a consensus sequence for ATPase. The nucleotide sequence suggests that the cutl+ gene product is a novel DNA-dependent ATPase, essential for chromosome separation.
Role of the nuc2 gene as a nuclear component
Mutant ts nuc2 cells enter mitosis with normal timing under restrictive conditions, and are arrested at a metaphase-like stage (Hirano et al. 1988). The chromosomes are condensed but do not separate. A short uniform-sized spindle forms but does not elongate (Fig. 4). The chromosomes are arranged so that a plate-like structure is formed through the center of which the spindle runs, perpendicular to the plate. It is important to determine whether chromosome structures or the spindle is defective in the nuc2 mutant.
The nucleotide sequence of the cloned nuc2+ gene predicts a 76K protein with several internal repeats (Hirano et al. 1988). Gene disruption indicates that the nuc2+ gene is essential. To identify the nuc2+ gene product, antisera against fusion proteins were made. Immunoblotting detects a polypeptide (apparent MW 67K, designated p67) in wild-type extracts of S. pombe. (In extracts of the mutant prepared after incubation at the restrictive temperature, a polypeptide (MW 76K) in addition to p67 is found.) The amount of p67 in wild-type extracts is greatly increased by introduction of multicopy plasmids carrying the nuc2 + gene. Plasmids carrying the nuc2+ gene with a NH2 domain deleted produce a shorter polypeptide with the expected MW, therefore, the bands detected by immunoblotting should represents the nuc2 + gene product. To determine its cellular location, homogenates were run in a Percoll gradient and each fraction was analyzed by immunoblolting. Results indicate that most of the p67 is present in nuclear fractions. (Immunofluorescence microscopy has failed to localize the nuc2+ protein in wild-type cells, )
p67 is an insoluble nuclear protein, and behaves as a nuclear scaffold-like protein (Hirano et al. 1988). It is insoluble in 2M-NaCl, 25 mM-LIS and 2% Triton but is soluble in 8M-urea. The p76 protein found in mutant extracts, however, is soluble. Recently we found that the nuc2 protein made in Escherichia coli binds tightly to DNA (Hirano, T., unpublished result). This DNA binding activity is localized to a small region of the polypeptide. Furthermore, weak but significant homology was found to certain gene products of S. cerevisiae related to mitosis and the induction of gene expression (R. S. Sikorski, M. Goebl, M. Boguski and P. Hieter, personal communication). Therefore, it is most likely that the nuc2+-related proteins are universally present in eukaryotes, and constitute a new class of gene family. A potential molecular function of nuc2 and related genes would be in altering chromosome structures by interacting simultaneously with chromosome DNA and the nuclear scaffold. It is of great interest to see whether the nuc2 protein associates with specific sites in the chromosomes.
Isolation of dis mutants and their phenotype
The dis mutants were isolated by screening approximately 1000 cs strains. They show basically the same cytological phenotype, but are classified into ;hree complementation groups, designated disl, dis2 and dis3 (Ohkura et al. 1988). The dis mutant cells were first grown exponentially at permissive temperature and separated in a sucrose gradient by centrifugation to select early G2 phase cells. Then, the synchronized cells were incubated at the restrictive temperature. The dis mutant cells became lethal during mitosis, suggesting that the dis+ genes are essential in mitosis. The terminal phenotype shows the asymmetric distribution (3: 0 or 2:1) of the nondisjoined, condensed chromosomes at the two ends of cell. If the above cells are incubated at permissive temperature, mitotic chromosome separation proceeds normally. The dis mutants are not defective in DNA replication.
The phenotypes of the dis mutants are pleiotropic, as listed below (Ohkura et al. 1988).
(1) Chromosomes strongly condense under restrictive conditions. The extent of condensation appears to be much greater than that of the wild type.
(2) The early spindle structures seen in the cells at the restrictive temperature lack the thick ends normally seen in the wild-type early spindle structure and considered to be kinetochore microtubules (Hiraoka et al. 1984; Hagan and Hyams, 1988). Thus dis mutants may be impaired in kinetochore structure.
(3) The intermediary segregating form (called the U-form) of chromosomes seen in the wild type (Toda et al. 1981) is never seen in dis mutant cells. Instead the chromosomal domain appears to be split into three subdomains (presumably corresponding to the three chromosomes) after the spindle forms. Therefore sister chromatid separation is defective in the dis mutants.
(4) Non-disjoined chromosomes move to the cell ends with either a 2:1 or 3:0 distribution.
(5) The nucleolus does not separate.
(6) Terminal degradation of the mitotic spindle is not complete. The parts associated with the condensed chromosomes remain.
(7) Decondensation of chromosomes does not occur.
(8) Artificial minichromosomes are lost with a high frequency at the permissive temperature.
(9) Homozygous diploids cannot be made at any temperature.
(10) All of the isolated dis mutants are hypersensitive to caffeine. Mutant cells in the presence of caffeine at permissive temperature show a phenotype similar to that expressed at non-permissive temperature.
These phenotypes indicate that the dis+ genes play a fundamental role in chromosome behaviour. Genes that complement the cs dis phenotype have been cloned by complementation (Ohkura et al. 1988). Interestingly, there are six clones that complement disl mutants while four clones complement dis2. Only one clone complements dis3. There appears to be a complex interacting system to execute the dis+ gene functions.
Genes that complement dis2 and their products
One of the four genomic sequences that complements dis2-11 directs integration to a chromosomal site which is tightly linked to the dis2 locus, suggesting that it contains the dis2+ gene. The other three genes are not linked to dis2, and are designated sds21 +, sds22+ and sds23+. Interestingly, the dis2+ and sds21+ genes complement both cs and caffeine hypersensitivity, whereas the other two complement only the cs phenotype (Ohkura et al. 1989).
By nucleotide sequence determination, the dis2+ and sds21+ genes are found to encode highly homologous proteins (calculated size approximately 37K). Antiserum raised against the fused dis2+ detects two polypeptides of 37K and 39K in the extracts of wild-type 5. pombe. The intensity of the major p37 band was increased by the presence of a multicopy plasmid carrying the dis2+ gene, while the minor p39 band increases its intensity by a multicopy plasmid carrying the sds21+ gene. Thus we conclude that dis2+ and sds21+ genes encode p37 and p39, respectively. By immunochemical analyses, p37 and p39 are found to be preferentially enriched in the nucleus; they exist as oligomers and are solubilized by 0.4M-NaCL Immunofluorescence microscopy shows that the nucleus is intensely stained, and granules are seen in the cytoplasm.
The dis3+ gene was cloned by complementation. By nucleotide sequence determination and immunochemical analyses, we found that it encodes a protein whose behaviour is highly similar to that of the dis2+ protein (Kinoshita, N., unpublished result).
Genes similar to dis2+ have been cloned from other organisms. By hybridization, the sequences highly homologous to dis2+ were obtained from S. cerevisiae and mouse (Ohkura et al. 1989). The predicted amino acid sequences of the budding yeast and mouse clones are more than 80 % identical to that of dis2 +. Thus the dis2 +gene, and concomitantly its function, is highly conserved from yeasts to mammals. It was found that the dis2+ protein is highly homologous to a rabbit protein phosphatase 1 (Ohkura et al. 1989). The significance of this finding is described in the following section.
The sds22+ gene has been cloned, and its nucleotide sequence shows that the sds22+ gene product consists of a series of leucine-rich repeats (Ohkura, H., unpublished result). Characterization of this protein by immunoblotting using antiserum raised against a sds22+ fusion protein indicates that this leucine-rich protein exists in the nucleus.
Roles of trans-acting protein factors in chromosome separation
Our results are summarized in Fig. 5. These gene products are essential for chromosome separation. Essentiallity of the genes can be examined only through genetical investigation. These may represent only a small fraction of the genes required for chromosome separation.
It is still a surprise that tubulin mutants show cell cycle arrest phenotypes, considering that microtubules are involved in a wide range of cellular functions. Our results indicate that certain tubulin gene mutations cause specific arrest at a mitotic stage by the inability of the cells to form a mitotic spindle. There may be other cell cycle-associated genes that act on the mitotic spindle. Their mutant phenotypes would possibly be expected to be similar to those of tubulin mutants. Therefore, we have isolated such mutants and are characterizing them, hoping that some of them are related to microtubule-associating or interacting proteins.
There may also be DNA-interacting enzymes present other than topoll that are required for condensation and segregation of sister chromatids. An example would be the cutl+ protein. The cutl mutants exhibit phenotypes similar to those of top2 mutants, and the cutl+ protein seems to encode a 210K potential DNA-dependent ATPase. Other mutations involved in the covalent bond changes in mitotic chromosome DNAs may show similar phenotypes. The defects caused by top2 and cut are perhaps not recognized by the regulatory systems for spindle dynamics and cytokinesis.
The nuclear scaffold-like protein nuc2+ is required for chromosome separation. It is unknown why the nuc2 mutant is blocked at a stage similar to metaphase. A hypothesis is that the nwc2+ gene product is required for a gross structural alteration of chromosome structures during the metaphase-anaphase transition. It will be interesting to see whether the nuc2 protein interacts with specific sequences of DNA, or nuclear scaffold, or both. Further investigation is necessary to understand the role of the nuc2+ protein, especially with regard to its subdomain and repeat units.
The dis2+ protein is highly homolgous to rabbit protein phosphatase-1, and therefore the dis phenotype may be related to dephosphorylation. Approximately 82% and 74% of the residues of dis2+ and sds21+ proteins, respectively, are identical to rabbit protein phosphatase 1 (PP-1; Berndt et al. 1987; Cohen, 1988). These values of similarity are very high, considering the evolutionary distance between yeast and rabbit. The gene disruption experiments indicate that single disruptants are viable, where the double disruption is lethal; dis2+ and sds21+ are functionally overlapped, although they may not be identical. The disruption of dis2+may be substituted by sds21+ and vice versa.
The defect in protein phosphatases may well cause a much more general phenotype in cellular function as the enzymes are involved in many cellular processes. However, the arrested phenotype of dis2 is apparently highly uniform and specific for mitosis. Genes similar to PP-1 were also isolated from <S. pombe and Aspergillus nidulans using cell cycle-specific phenotypes. The gene bwsl+ of S. pombe, which reverses weel mutant suppression of the cdc25 mutation, is identical to the dis2+ gene (Booher and Beach, 1989). The phenotype of A. nidulans bimG mutant is similar to that of the dis mutants, and the gene complementing bimG is highly homologous to PP-1 (Doonan and Morris, 1989). Therefore the role of PP-1 may be specifically implicated in mitotic control. It has been pointed out that protein phosphatases would be expected to play an important role in mitosis (Foulkers and Maller, 1982; Cyert and Kirschner, 1988). However, only four types of protein phosphatases have been described in mammalian systems (Ingebritsen and Cohen, 1983). This is in contrast to the large number of protein kinases in which mutations often cause a cell cycle stage-specific phenotype. There may be many more protein phosphatases awaiting discovery or there may be some unknown mechanisms which attribute a wide variety of substrate specificities to a restricted number of protein phosphatases.
How does a mutation in the structural gene for protein phosphatase 1 cause the block of chromosome disjoining? Current hypotheses are: (1) alteration in the dephosphorylation pattern causes hypercondensation of chromosomes which are unable to be disjoined. In this case, mitotic induction is too strong to be properly regulated; (2) coordinate coupling of PP-1 with mitotic kinases is broken, so that chromosome separation is blocked while chromosomes condense and the spindle is made and elongates, and (3) temporally-regulated expression of type 1 protein phosphatase is essential for correct segregation of chromosomes so that mutations impaired in such temporal control results in the dis phenotype.
The dis2+ might be implicated in the cAMP cascade pathway because dis2-\Acs is weakly complemented by the 5. cerevisiae PDE2 (cAMP phosphodiesterase gene; Ohkura et al. 1988). The mutation can be rescued by apparently decreasing the intracellular cAMP concentration. It is clear that the expression of dis2+/PP-1 is under the control of a highly complex interacting system.
In conclusion, we have identified some of the gene products required for mitotic chromosome separation in the fission yeast S. pombe by mutation, gene cloning, nucleotide sequence and immunological methods. They include; a-, β-tubulin, DNA topoisomerase II, a nuclear scaffold-like protein nuc2+ and type 1 protein phosphatase. Interestingly, none of them apparently has a unique role in mitosis, but instead, they have multiple cellular roles. However, mutations which cause defects at specific steps in chromosome separation can be identified in these genes. This apparent contradiction may be understood by the fact that chromosome separation is a highly intricate and complex system that involves the coordinated function of a number of inter-related gene products. For example, top2 mutations specifically block chromosome condensation and separation, because functionally overlapping topo I can substitute most of the topo II functions in other steps of the cell cycle. A topo II function which cannot be substituted by any other gene product becomes essential during chromosome separation. Certain gene products have to be activated, modulated or inactivated during mitosis with correct temporal control. Therefore the study of chromosome separation requires an understanding of many gene functions involved in cell growth control, transcription or DNA replication. Examples are shown in this paper. Mutations affecting such mitosis-specific modulations will cause a defect in chromosome separation. Genetic analyses of chromosome separation have revealed a number of regulatory mechanisms, the coordinate execution of which can be uncoupled. A major conclusion from our studies to date is that there are at least four potentially independent pathways in mitosis in the fission yeast, namely chromosome separation, spindle dynamics, septum formation and cytokinesis.
ACKNOWLEDGEMENTS
The author is grateful to members of his laboratory for quoting their unpublished results and to fain Hagan for reading the manuscript. The work done in the author’s laboratory was supported by grants from the Mitsubishi Foundation, the Ministry of Education, Science and Culture and the Science and Technology Agency of Japan.