Genetic interactions in the control of septation in Schizosaccharomyces pombe

The fission yeast Schizosaccharomyces pombe is an excellent model system for the analysis of the eukaryotic cell divison cycle. Many of the genes defined in fission yeast as important for the regulation and execution of two of the landmark events of the cell cycle, S phase and mitosis, have been functionally conserved through eukaryotic evolution. The control of the third major event of the cell cycle, cytokinesis, remains poorly understood. A number of conditional lethal mutants of S. pombe showing terminal phenotypes consistent with a role in cytokinesis have been identified (Nurse et al., 1976; Minet et al., 1979). These can be divided into three groups. Those that are defective in the organisation of the septum, termed late mutants, defined by mutations in the genes cdc3, cdc4, cdc8 and cdc12, those defective in the regulation of septum formation, defined by a single mutant, cdc16, and those in which the nuclear and growth cycles proceed in the absence of cytokinesis, termed early mutants and defined by the genes cdc7, cdc11, cdc14 and cdc15. The disorganised septa formed by the late mutants suggest that their products are involved in correct placement and/or assembly of this organelle. Conversely, the periodic formation of septa in the absence of cell cycle progression in cdc16 (Minet et al., 1979), and the continued nuclear and growth cycles observed in the early mutants in the absence of cytokinesis (Nurse et al., 1976; Mitchison and Nurse, 1985; Creanor and Mitchison, 1990), suggest that these two classes are involved in the regulation of septum formation. We have therefore focussed on the interac tions amongst them through the examination of the phenotypes of double mutants. Our data are consistent with a physical interaction amongst the products of the cdc7, cdc11, cdcl4 and cdc16 genes that may regulate the activity of the cdc15 gene product.

The media used in this study have been described previously (Gutz et al., 1974; Nurse, 1975). They were Yeast Extract (YE) and Malt Extract (ME), supplemented with adenine, leucine, uracil, histidine and lysine at 50 mg I⁻¹ each, and EMM2 minimal medium supplemented as required. The vitality stain, PhloxinB (Sigma) was added to YE media at a concentration of 5 mg I⁻¹ when required to aid the identification of ts mutants. All mutants were out-crossed to wild type at least twice before use, to eliminate effects resulting from secondary mutations. Crosses were performed according to Gutz et al. (1974) on ME plates at 25°C. After three days, tetrads were dissected using a Leitz micromanipulator. The genotypes of all double mutants were verified by out-crossing to wild type, and where appropriate to the individual mutants. Mutants designated L in Fig. 3 were found to be lethal at all temperatures tested on YE media (19°C to 36°C). They could however be propagated by the addition of 1 M KC1 to YE media, which leads to osmotic suppression of cdc11–136, cdcl6–116 and cdc7–24. Terminal phenotypes were examined following growth in YE+1 M KC1 and shift to YE medium at 25°C, for the equivalent of two generations.

Cells were stained with 4’,6’-diamidino-2-phenylindole (DAPI; Sigma) and Calcofluor (fluorescent brightener no. 1A, Sigma) after fixation with 3% formaldehyde. Photographs were taken using a Reichert-Jung Polyvar microscope with ×100 objective onto Kodak Tri-X Pan film, processed according to the manufacturer’s instructions in D76.

All DNA manipulations were performed according to standard procedures (Sambrook et al., 1989). S. pombe was transformed by the protoplast method (Beach and Nurse, 1981).

The single mutants we have examined display three morphologically distinguishable terminal phenotypes at the restrictive temperature. Mutants in cdc7, cdcll, cdcl4 and cdcl5 do not form septa (Fig. 1A, B), but undergo continued nuclear division cycles and elongate. Although these mutants are classifed as defective in early septation (Nurse et al., 1976), they can be distinguished from each other by a number of criteria. First, cdcl5 mutants elongate to a lesser extent than the other mutants and often swell at one or both ends (compare Fig. 1A and B). Second, examination of number of nuclei per cell at various times after shift to restrictive conditions affords a more sensitive assay to distinguish the phenotypes of the four mutants. After four hours at 36°C, over 60% of cdc7 and cdcll cells have four or more nuclei, while more than 60% of cdcl4 and cdcl5 cells are binucleate (Fig. 2A). The latter strains can be easily distinguished following an eight-hour block. Whereas more than 75% of cdcl4cells have eight or more nuclei, cdcl5 cells remain predominantly binucleate (Fig. 2B). These differences have been used to assess the terminal phenotypes of some of the double mutants discussed below. The cdcl6 mutant shows little elongation and periodically forms multiple, discrete septa following a single mitosis (Fig. 1C; see also Minet et al., 1979).

To examine the potential dependency relationships and interactions among these genes, we performed crosses between them and assessed the phenotypes of the double mutants. The results are summarised in Fig. 3. The double mutants fall into three classes. First, those that show mutual suppression and are viable at the restrictive temperature for either parent; second, those that show synthetic lethality and are inviable at all temperatures; and finally, those that, at the restrictive temperature, show the terminal phenotype of a single parent, suggesting that they are part of a dependent sequence of cell cycle events (Moir and Botstein, 1982; Pringle, 1987). These interactions are described in detail below.

All three alleles of cdc11 in combination with cdc16–116 are viable at the permissive temperature, whereas an allele-specific interaction is observed following shift to the restrictive temperature. The double mutants cdc11–119cdcl6–116 and cdc11–123cdcl6–116 undergo cell cycle arrest, phenotypically resembling a cdc11 mutant, in that cells become elongated and multinucleate. Rarely, however, one or more eccentrically placed septa are formed (data not shown). These data suggest that during cytokinesis the cdc11 gene product is required before the product of the cdc16 gene, suggesting that they form part of a dependent sequence of events. However, the double mutant cdc11–136cdcl6–116 is viable at the restrictive temperature. Analysis of their phenotype at 36°C shows that, in contrast to either parent, the double mutant does not elongate significantly (in contrast to cdc11; see Fig. 1A) and that many of the cells form a single, correctly placed septum (in contrast to both parents). However, while the cells are viable, some septal aberrations remain in terms of both their number and position (Fig. 4). This suppression strongly suggests that the products of these genes interact (Moir et al., 1982; Pringle, 1987).

Subsequently, the interaction of the cdc11 and cdc16genes with the remaining early cytokinesis mutants, cdc7, cdcl4 and cdcl5, were investigated. The doublemutant combinations cdc7–24cdcl4-118, cdc7–24cdcl6–116 and cdcl4–118cdcl6–116 were found to be lethal at all temperatures tested, from 19°C to 36°C, while allelespecific synthetic lethality, rather than suppression, was observed when interactions with cdcll were analysed. The double mutants cdc7-24cdcll–136 and cdc11 136cdcl4–118 were lethal at all temperatures tested. In contrast, cdc7–24cdc11–119, cdc11–119cdcl4–118, cdc7–24cdc11–123 and cdc11–123cdcl4–118 were viable at the permissive temperature, undergoing cell cycle arrest and showing a terminal phenotype characteristic of an early septation mutant at the restrictive temperature (see Fig. 3). The double mutants cdc7–24cdc11–123 and cdc11–123cdc14–118 also showed a reduced restrictive temperature of 29°C. These data are consistent with the existence of an interaction between the products of cdc7 and cdcl4, and those of cdc11 and cdc16.

In order to investigate their terminal phenotypes, synthetic lethal mutants were grown on medium containing 1 M KC1, which suppresses each of the cdc7–24, cdc11–136 and cdc16–116 phenotypes. All the double mutants are still temperature sensitive in this medium. When the restrictive condition was imposed by shift to normal medium at 25°C, the mutants cdc7–24cdc11–136, cdc7–24cdcl6–116, cdcll–136cdcl4–118, cdcl4–118cdcl6–116 and cdc7–24cdcl4–118 became highly elongated and underwent multiple rounds of nuclear division before lysing, characteristic of an early defect in septation, rather than that of cdc16. These data suggest that the cdc16 product acts downstream from the cdc7 and cdcl4 products.

All combinations of cdcl5 with cdc7, cdc11, cdc14 and cdc16 are viable at permissive temperature, and at restrictive temperature undergo cell cycle arrest. The double mutants of either allele of cdcl5 with cdc7–24, cdcl4–118, cdc11–119, cdc11–123 or cdc11–136 become highly elongated and show no signs of making septa or swelling. In addition, analysis of the number of nuclei per cell at various times after shift indicates that the double mutants show the phenotype of the cdc7, cdc11 or cdc14 rather than the cdc16 parent. After a four-hour block, cdc7cdcl5 and cdc11 cdc16 both showed predominantly tetranucleate cells, reminiscent of the cdc7 and cdc11 parents at a similar time after shift (Fig. 2A, C). The double mutant cdc14cdcl5 was also predominantly tetranucleate after four hours (Fig. 2B). Since a longer incubation is required to distinguish between cdcl4 and cdcl5 (see above), the phenotype of the double mutant cdcl4cdcl5 was examined at eight hours after shift. The majority of cells contained eight or more nuclei, reminiscent of cdcl4 rather than cdcló mutants (Fig. 2B). In contrast, and as previously reported (Minet et al., 1979), double mutants of either cdc.15–127 or cdcl5–140 with cdcl6–116 showed a cdc.15 phenotype (not shown). These findings imply that cdcl5 functions upstream from cdcl6, but downstream from cdc7, cdc11 and cdcl4.

In order to study the proteins encoded by these genes, we have initiated a molecular analysis of them. A cdc7–24ura4D18 mutant was transformed with a multicopy library and colonies capable of growth at the restrictive temperature identified. A plasmid that could rescue the cdc7–24 mutation following réintroduction into cells was obtained. The plasmid pcdc7.8 was shown by integration and tetrad dissection to encode the cdc7⁺ gene. Crosses of a stable integrant of the plasmid (SP964) were performed with two strains, SP217 (ura4D18 h^+N) and SP762 (cdc7-24 ura4D18 h^+N). In the first cross, dissection of 38 tetrads indicated that all the progeny were cdc⁺, establishing tight linkage of the integrated plasmid to the cdc7 locus. In the second cross, none of the progeny from dissection of 27 tetrads was cdc⁻ura⁺ or cdc⁺ura⁻, establishing that the ura4⁺ gene on the integrated plasmid was linked to the cdc⁺ activity. Together, these crosses establish that the plasmid pcdc7.8 had integrated into the cdc7 locus and therefore carries the cdc7⁺ gene. Introduction of the plasmid into cdc11 mutants showed that this clone could complement cdc11–119, but not cdc11–136, while complementation of cdc11–123 was poor. The allele specificity of the rescue is consistent with an interaction between the products of the cdc11 and cdc7 genes, strongly supporting the genetic data described above.

We are presently sequencing the cdc.7⁺ gene to begin to study the role of its product.

We have investigated the dependency relationships and interactions amongst genes implicated in the control of septation in the fission yeast Schizosaccharomyces pombe, by examination of the phenotypes of double mutants under restrictive conditions. This approach permits the ordering of genes within a dependent pathway, as well as the identification of interactions between products through the observation of synthetic lethality or suppression (Moir and Botstein, 1982; Moir et al., 1982; Hennessey et al., 1991).

We have found an allele-specific suppression between cdc11–136 and cdc16–116 and, in addition, lethal interactions between cdc7, cdc!4, cdcl6 and the other two alleles of cdc11, suggesting that the products of these four genes interact. Examination of the phenotypes of double mutants of cdcl5 with the other four genes suggests that cdcl5 functions downstream from cdc7, cdc11 and cdcl4, yet before cdcl6.

The data presented above lead us to propose the following model for the genetic control of septation in fission yeast, depicted in Fig. 5. The initiation of septation is postulated to require the activation of a complex that contains the products of the cdc7, cdc11 and cdcl4 genes. It should be noted that there may be other gene products in the complex in addition to those identified through the analysis of temperature-sensitive mutants. The genetic interactions we observe among the cdc7, cdc11, cdc14 and cdc16 genes lead us to suggest that the product of the cdc16 gene is also a part of the complex. This is implied by the allele-specific suppression between cdc11–136 and cdc16–116, which suggests that these proteins interact. By analogy with studies in other eukaryotes we propose that this complex is activated at the onset of anaphase (Girbardt, 1979; Rappaport, 1986). Studies of changes in the actin cytoskeleton during the fission yeast cell cycle (Marks and Hyams, 1985; Marks et al., 1986), show that the appearance of the equatorial actin ring, which precedes the deposition of cell wall material, does not occur prior to the onset of anaphase and that the formation of the septum is not seen until anaphase is complete. Consist ent with this, 5. pombe mutants that have initiated anaphase but cannot complete mitosis, such as cut, dis, topi and cdc!3, initiate septum formation (Hirano et al., 1986; Marks et al., 1986; Uemura et al., 1987; Hagan and Hyams, 1988; Ohkura et al., 1988; Uzawa et al., 1990). The nature of the signal that activates septum formation is currently unclear, but it is tempting to speculate that it might be given by the inactivation of MPF (Nurse, 1990), which occurs at this point. We are investigating this possibility.

Our genetic analysis of the role of the cd16 gene is consistent with earlier cytological observations, from which it was proposed that the cdc16 gene inhibits the formation of more than a single septum per cell cycle (Minet et al., 1979). We suggest that the completion of the septum generates a signal that alters the activity of the cdc16 product, leading to inactivation of the complex. In the absence of functional cdc16 product, the complex remains active and septum formation is reinitiated, leading to the periodic synthesis of additional septa. It is unlikely that the cdc16 product is the signal that is generated by the completion of the septum, as this is not consistent with observed interactions of cdcl6 with early septation mutants (suppression of cdc11 and the lethal interactions with cdc7 and cdc14), since in these cells no septa are formed. We also propose that inactivation of the complex is necessary to proceed to cell separation, since cdc16 mutants do not utilise the septa that are made. The nature of the signal that defines the position and timing of subsequent septa in cdcl6 mutants is unclear.

In order to resolve the apparent contradiction between the finding of interactions between cdc11 and cdc16 and the temporal order of their functions with respect to cdcl5 (see Results), we propose that the cdc15 gene encodes a function that is necessary for septum formation and is dependent upon the action of the cdc7, cdc11 and cdcl4 products for activation. Consistent with this, we observed that double mutants of cdcl5 with cdc7, cdc11 and cdc14, show the terminal phenotype characteristic of an early septation mutant. Mutants, such as cdcl6, which manifest their effects upon the termination of septum formation will thus be masked in a cdc15 background, in which a septum is not formed. In agreement with this, we find that the phenotype of a cdc15cdc16 double mutant is that of a cdcl5 mutant. We currently favour the placement of cdcl5 function downstream from the other early septation mutants, as we have not observed either allele-specific suppression or lethality in any of the double mutants. However, the possibility that the cdc15 product also forms a part of the putative cdc7-cdc11-cdc14 complex cannot be excluded by our present data. Cloning of the early septation genes and analysis of their products should permit us to distinguish between these alternatives. Since three of the five genes we have analysed here are defined by single alleles, it seems likely that there will be a number of other genes that are involved in the process of septation and its control that have not been identified genetically. We are presently undertaking a pseudoreversion analysis of the known genes in order to identify novel elements involved in the process. Recent studies in higher eukaryotic cells have identifed a new “organelle”, which appears at cell division, called the telophase disc (Andreasson et al., 1991). In view of the functional conservation of a number of genes that are involved in mitosis and DNA synthesis throughout eukaryotic evolution, it is tempting to speculate that if such a structure is also present in lower eukaryotes, then the products of the cdc7, cdc11, cdcl4, cdcl5 and cdcl6 genes are components of it. The production of antisera to the products of these genes will allow us to investigate this possibility.

We thank Prof. B. Hirt for his support and encouragement, M. Allegrini and P. Dubied for artwork and photography, Alexandre Reymond for constructive and entertaining discussions, and Eva Bucheli for technical assistance. We also thank Willy Krek for constructive criticisms of the manuscript and Tony Carr (University of Sussex, UK) for a fission yeast plasmid DNA library. This work was supported in part by Swiss National Science Foundation grant no. 31-26285.89, Swiss Cancer League grant no. FOR.428.90.1 and by an EEC Postdoctoral Fellowship to J.M.