An Ancient Yeast for Young Geneticists: A Primer on the Schizosaccharomyces pombe Model System

Gene complement

A central goal of biological research is to describe fully the orchestrated collection of functions and processes that combine to produce living cells. One way to assess progress is by maintaining and interrogating an accurate “parts list” of the contributing gene products and their attributes. The S. pombe parts list is maintained by the model organism database (MOD) PomBase (http://www.pombase.org), with 5054 protein-coding genes currently reported (compared to 5821 for S. cerevisiae). This difference in protein number is partly due to the retention of many paralogous pairs (homologous genes within an organism) in S. cerevisiae following what appeared to have been two rounds of whole-genome duplication (WGD) (Wolfe and Shields 1997; Kellis et al. 2004). Despite a lower number of protein-coding genes, fission yeast has a large number of proteins (currently 338) that are conserved in vertebrates (source: PomBase Query Builder using curated fission yeast/budding yeast and fission yeast/human orthologs) but absent from budding yeast, while only 179 genes are conserved between budding yeast and vertebrates that are absent from fission yeast (source: SGD Yeastmine using Compara human orthologs and PomBase curated fission yeast/cerevisiae orthologs supplemented by manual curation). This is a consequence of the increased lineage-specific gene losses in the budding yeast lineage after fission and budding yeasts diverged from a common ancestor (Aravind et al. 2000; Wood 2006).

Of the 5054 known or predicted fission yeast proteins, at this time, 2154 have a published biological role, 2050 have a biological role inferred from an experimentally characterized ortholog (usually from budding yeast), and 850 have no known biological role [inferred from the absence of Gene Ontology (GO) biological process annotation]. A large number of these uncharacterized proteins are conserved outside the Schizosaccharomyces genus; 182 are conserved in mammals, almost 40% (68) of which have no budding yeast ortholog. Tantalizingly, a number of these genes are orthologs of human disease genes or members of the Catalogue of Somatic Mutations in Cancer (COSMIC) of genes with mutations in cancer genomes (Forbes et al. 2014). For example, SPAC652.01 ortholog BLCAP is associated with bladder cancer, and SPAC6G10.10c ortholog MMTAG2 is associated with multiple myelomas. Characterization of these gene products in fission yeast almost certainly will shed light on the functions of human proteins.

Noncoding RNAs (ncRNAs) include all RNAs other than messenger RNA (mRNA) and are central to a wide range of biological processes, including transcription, translation, gene regulation, and splicing. The ncRNA annotation of the fission yeast genome has increased dramatically in recent years. In addition to the 307 ribosomal RNAs (rRNAs), transfer RNAs (tRNAs), small nuclear RNAs (snRNAs), and small nucleolar RNAs (snoRNAs), there are now 1538 additional ncRNAs annotated in the genome, mainly from RNA sequencing (RNA-Seq) data (Watanabe et al. 2002; Wilhelm et al. 2008; Rhind et al. 2011). Similar extensive intergenic and antisense transcription is detected in budding yeast (Yassour et al. 2010; Goodman et al. 2013; Pelechano et al. 2013). A large number of the ncRNAs identified in fission yeast (694) appear to be antisense to protein-coding genes, and some (e.g., tos1, tos2, and tos3, which are antisense to rec7) carry out a regulatory role during sexual differentiation (Molnar et al. 2001; Watanabe et al. 2002), whereas many others display meiosis-specific expression (Bitton et al. 2011). Of the remainder, only a handful, including BORDERLINE/IRC1 (Cam et al. 2005; Keller et al. 2013), mrp1 (Paluh and Clayton 1995), rrk1 (Krupp et al. 1986), sme2 (Watanabe and Yamamoto 1994), and srp7 (Ribes et al. 1988) are characterized, hinting at a massive potential for future expansion of ncRNA research in fission yeast. Most of the annotated ncRNAs are of low abundance in vegetative cells, similar to tightly repressed meiotic genes, but 187 of them are expressed at ∼1–200 copies/cell (Marguerat et al. 2012).

Chromosome organization

The S. pombe genome size is ∼13.8 Mb compared to the 12.5-Mb genome of S. cerevisiae. Despite similar genome sizes, S. pombe has only 3 relatively large chromosomes of 5.7, 4.6, and 3.5 Mb compared to 16 smaller chromosomes in S. cerevisiae. Indeed, at 3.5Mb, the smallest S. pombe chromosome (chromosome III) is over twice the length of the largest S. cerevisiae chromosome (chromosome I), which is 1.5 Mb.

Fission yeast has large modular centromeres, more reminiscent of those of higher organisms than of the 125-bp element sufficient for centromere function in S. cerevisiae. Fission yeast centromeres (cen1, cen2, and cen3) are approximately 40, 69, and 110 kb. The centromere structure comprises a nonconserved central core sequence flanked by variable numbers of outer repeats. Many of the proteins that bind to S. pombe centromeres are conserved in mammals but absent from budding yeast, including Swi6 and Chp1 and the RNA interference (RNAi) pathway components (Lorentz et al. 1994; Ekwall et al. 1995; Doe et al. 1998; Volpe et al. 2002). The conservation of centromere features, including size, structure, and multilayered organization, makes fission yeast a valuable model for the study of eukaryotic chromatin remodeling and centromere function.

Subtelomeric regions in both fission yeast and budding yeast are reversibly transcriptionally silent (Aparicio et al. 1991; Nimmo et al. 1998; Halme et al. 2004) and are enriched for genes encoding membrane transporters, stress-response proteins, and expanded families of species-specific cell surface glycoproteins (Goffeau et al. 1996; Robyr et al. 2002; Wood et al. 2002; Hansen et al. 2005). In fission yeast, these regions are very similar to each other (approximately 50–60 kb of the sequence immediately proximal to the telomeric repeats has ∼99% identity among telomeres for most of the region). A similar localization of contingency genes involved in antigenic variation for immune evasion has been observed in parasitic microbes (Barry et al. 2003). Subtelomeric regions therefore may provide the ideal genomic location for eukaryotic cells to test and select for novel species-specific genes required for stress and surface variation in response to changing environmental conditions, and the over-representation of environmentally regulated genes in subtelomeres may be applicable to eukaryotes in general.

Finally, the S. pombe genome contains 262 sequences comprising partial or complete copies of mobile genetic elements (Bowen et al. 2003) derived from LTR retrotransposons. Thirteen full-length active copies of Tf2 exist in the classic laboratory strains, while active copies of Tf1 only occur in wild isolates of S. pombe (Levin et al. 1990; Levin and Boeke 1992; Hoff et al. 1998). Although these are expressed at low levels under standard laboratory conditions, forced induction of retrotransposon transcription leads to integration of new copies into promoters of RNA pol II–transcribed genes (Guo and Levin 2010). Interestingly, Tf1 integration favors stress-response gene promoters, and once integrated, the Tf1 generally increases expression of the adjacent gene (Guo and Levin 2010; Feng et al. 2013). Thus, these mobile elements may serve as a source of genetic variability to promote adaptation to new conditions of environmental stress.

Mitotic cell cycle

Mitchison’s adoption of S. pombe as an ideal organism in which to study the cell cycle (Mitchison 1990) was based on the modes of growth and division of the cells. S. pombe cells are cylindrical in shape with roughly hemispherical ends and are about 3.5 µm in diameter during exponential growth on nutrient-rich medium. Newborn cells are about 8 µm long and grow continuously through most of the cell cycle without width change. This means that a good estimate of the age of a cell (time elapsed since birth) can be obtained simply by measuring its length. When cells reach about 15 µm in length, they undergo a closed mitosis (as with most fungi, the nuclear envelope does not break down) with the nucleus close to the center of the cell. Shortly after nuclear division, a transverse septum is laid down medially, which is then cleaved to produce two daughters (Figure 2A and supporting information, File S1). The symmetric mode of division means that the two daughter cells are nearly equal in size, similar to cultured mammalian cells, but quite different from the asymmetric division of budding yeast. The nuclear DNA replicates very soon after mitosis (Nasmyth et al. 1979). Thus, the short G₁ phase takes place between mitosis and cell separation, with two unreplicated nuclei present in the cell until the septum is cleaved. S phase is coincident with the presence of the septum, so newborn daughter cells are already in G₂, containing fully replicated chromosomes. After cytokinesis, cells remain in a G₂ phase, which occupies about three-quarters of the cell cycle, leading up to the next mitosis. As such, virtually every haploid cell in a growing population contains two copies of the chromosomal complement. Diploid S. pombe cells are noticeably wider (∼4.4 µm) than haploid cells and undergo cytokinesis at approximately 24 µm in length (Figure 2B and File S2), making them almost double the volume of haploid cells.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

S. pombe cells and asci. (A) A haploid cell containing a fission plate. (B) Diploid cells. (C) Zygotic asci formed by the mating of two haploid cells. (D) Azygotic asci formed when diploid cells go through meiosis.

When subjected to nutrient starvation in the absence of a mating partner, haploid cells exit the cell cycle and enter stationary phase. One unusual feature of S. pombe is that while they enter stationary phase from G₁ on nitrogen starvation, they do so from G₂ upon glucose starvation (Costello et al. 1986). If a mating partner is present, haploid cells shift from vegetative growth to a sexual cycle (described in the next section). Depending on their mating-type constitution (see next section), diploid cells enter either stationary phase or the sexual cycle in response to nutrient starvation.

Sexual cycle and mating-type system

In addition to vegetative growth, S. pombe has a sexual cycle. Two haploid cells of appropriate mating types (see later) respond to nutrient starvation by mating (conjugation) (Egel and Egel-Mitani 1974; Egel 1989). This involves adhesion of the cells followed by digestion of the walls separating them to form a single fusion cell containing the two parent nuclei (Figure 3 and File S3). The nuclei then fuse to form a zygote, a single cell containing a single diploid nucleus. Normally, the nucleus enters the meiotic pathway soon after zygote formation, and four haploid nuclei are generated. Subsequently, a spore wall forms around each nucleus to produce an ascus consisting of four spores within the shell of the zygote (Figure 2C and Figure 3H). These zygotic asci typically have a bent shape that reflects the angle between the two cells that fused to generate the zygote (Figure 2C). In contrast, azygotic asci derived from vegetative diploid cells are short and linear (Figure 2D). Whether zygotic or azygotic, the ascus wall lyses on exposure to growth medium, and the spores germinate and develop into vegetative cells that return to mitotic growth.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Sexual development in S. pombe. Cells of the homothallic h⁹⁰ mating type were plated onto a thin EMM agar pad at 30° and photographed at (A) 0 min, (B) 184 min, (C) 204 min, (D) 254 min, (E) 286 min, (F) 334 min, (G) 446 min, and (H) 560 min. Because h⁹⁰ cells can undergo mating-type switching, which occurs in only one of the two daughter cells, two pairs of daughter are seen to have mated and formed asci.

Truly wild-type S. pombe strains are of the homothallic mating type referred to as h⁹⁰, in that they possess three copies of information at the mating-type (mat) locus, one of which is expressed and two of which are transcriptionally silenced (Figure 4). S. pombe h⁹⁰ strains can transfer genetic information from either silent locus to the expressed locus. This produces a mixture of cells expressing one or the other mating type within a clone to allow mating to take place. This may be beneficial to surviving starvation conditions in the wild by allowing for the creation of spores. S. pombe cells also can be heterothallic such that they require a mating partner of the opposite mating type (Leupold 1950).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Schematic of the S. pombe mating-type locus.

Diploids

After cells mate, the diploid zygote generally undergoes meiosis and sporulation immediately to produce a zygotic ascus. This depends on the mating-type composition of the diploid: h⁻ and h⁺ haploid parents will produce an h⁻/h⁺ diploid that is meiosis (and sporulation) competent, whereas a diploid that is homozygous for either the h⁺ or h⁻ mating type is not able to enter meiosis. Mating within an h⁹⁰ culture or of h⁹⁰ cells with h⁻ or h⁺ cells will also produce a meiosis-competent diploid. The normally transient diploid state of a cell can be “rescued” by transfer to nutrient-rich growth medium. A fraction of such diploid cells will propagate vegetatively as diploids, though rearrangements at the mat locus and/or spontaneous meiosis and sporulation are frequent events. The practice of making and using diploid S. pombe strains is discussed further later.

Growth media

It is common for people who have experience with S. cerevisiae to simply use the same growth medium when starting to work with S. pombe. We strongly recommend against the use of YPD (the rich medium used for S. cerevisiae) for S. pombe. Rich medium for S. pombe (YES) is made with only yeast extract, glucose, and trace amounts of histidine, adenine, uracil, lysine, and leucine (Gutz et al. 1974). The peptone in YPD induces lysis of S. pombe ura4⁻ mutants (Matsuo et al. 2013) and creates a stress in all strains such that they are induced to mate while growing in what should be a rich medium. Thus, YPD does not provide a permissive growth condition for S. pombe. In addition, the standard defined medium used for S. cerevisiae, SC or SD medium, is inferior for S. pombe growth to Edinburgh Minimal Medium (EMM) (Gutz et al. 1974), although the basis of this growth difference is not known.

Ploidy and meiosis

The difficulties of creating and maintaining stable diploid strains of S. pombe were mentioned earlier. S. cerevisiae cells mate in nutrient-rich conditions but require a strong starvation signal to induce meiosis. In contrast, S. pombe cells require starvation conditions to mate, but nearly all diploid zygotes sporulate immediately (Figure 3 and File S3). While this is advantageous for tetrad dissection and strain construction, it poses a challenge for experiments that require diploid strains, such as dominance and complementation tests. Diploid strains are also used to introduce a change into the genome that may be lethal or deleterious in a haploid strain, such as a gene deletion. By altering one allele in a diploid, one can first confirm that the change has been introduced before sporulating the diploid to determine the consequence of the change in haploid strains.

Two methods exist for constructing and maintaining S. pombe diploids. First, the ade6-M210 and ade6-M216 alleles both confer adenine auxotrophy, but they complement intragenically such that the heterozygous diploid (ade6-M210/ade6-M216) grows in medium lacking adenine. Since these mutations are in the same gene, even when some diploid cells enter meiosis and produce spores, these haploid spores are virtually all adenine auxotrophs because of the tight linkage between the ade6-M210 and ade6-M216 alleles. This allows one to use Ade⁺ selection to maintain the diploid strain in the absence of contaminating Ade⁺ haploid progeny. Second, strains expressing the mat2-102 allele can mate with either an h⁺ or h⁻ heterothallic strain to generate a diploid. The diploid with h⁺ immediately enters meiosis in the normal way, and sporulation ensues. However, in the diploid with h⁻, neither haploid parent expresses a functional mat-Pm gene, which is required for meiosis, so the diploid nucleus persists, generating a stable diploid line. This allows one to select for diploid cells by using two haploid parents carrying complementary selectable markers (e.g., two different auxotrophies). Because meiosis is completely suppressed, there is no possibility of haploid progeny arising by meiotic recombination that would share the same phenotype as the diploid. Spontaneous mitotic haploidization following chromosome loss can occur. This haploidization can be stimulated by exposure to a low concentration of the microtubule-destabilizing drug benomyl (Alfa et al. 1993). Since it occurs in the absence of meiotic recombination, this method can be used to map mutations to one of the three S. pombe chromosomes, with the caveat that a mitotic recombination event could produce recombinant chromosomes in rare haploid segregants. One can distinguish between the original diploid strain and the viable haploid derivatives by plating for colonies on medium containing the vital stain Phloxine B, which stains dead cells (Gutz et al. 1974; Tange and Niwa 1995). Diploid colonies contain many more dead cells than haploids, so they are dark red on Phloxine B–containing plates, while haploid colonies are light pink. Phloxine B is also be used to detect rare diploids that arise in a culture, presumably as a result of chromosomal nondisjunction. These diploids grow faster than the haploid parent and will overtake a culture if one does not initiate an experiment from a single colony. Such diploids are mating competent but produce mostly inviable aneuploid progeny when crossed with a haploid strain owing to the creation of a triploid zygote.

Homologous recombination

One of the key strengths of both S. pombe and S. cerevisiae for molecular genetics studies is the ability to introduce DNA into the chromosome via homologous recombination during vegetative growth. This allows one to delete genes, construct translational fusions, or even construct specific mutant alleles. However, the relative ratio of homologous vs. nonhomologous insertion events is lower for S. pombe than for S. cerevisiae. The relatively high proportion of nonhomologous events in S. pombe, in which a small linear DNA molecule integrates at a site in the genome other than at the location targeted by the sequences at the ends of the molecule, can create difficulties. While different laboratories take different approaches to increase the efficiency of targeted DNA insertions, it is generally sufficient to create a PCR product carrying the transforming DNA that has 60 bp of targeting sequences flanking the product (Bähler et al. 1998).

RNAi

Unlike S. cerevisiae, S. pombe has retained through evolution the genes that encode the RNAi machinery, including the Dcr1 DICER ribonuclease, the Rdp1 RNA-dependent RNA polymerase, and the Ago1 Argonaute family member. Research on RNAi in S. pombe has not only contributed to our understanding of its mechanism and consequences but also helps to explain other biological differences between fission and budding yeasts. Loss of any RNAi genes leads to an increase in the presence of transcripts from the heterochromatic region of S. pombe centromeres as well as defects in chromosome segregation during mitosis (Volpe et al. 2002, 2003). Thus, the absence of RNAi in budding yeast may be linked to loss of large, repetitive centromeres as found in S. pombe, which, in turn, may explain the relatively small size of S. cerevisiae chromosomes that appear to segregate in mitosis by attachment to a single microtubule (Peterson and Ris 1976). RNAi also appears to play a modest role in regulating expression of Tf2 retrotransposons in S. pombe (Hansen et al. 2005) and the more abundant retrotransposons found in the related fission yeast Schizosaccharomyces japonicus (Rhind et al. 2011). Because RNAi regulates retrotransposon expression in higher organisms (Tabara et al. 1999), it has been suggested that the loss of RNAi in S. cerevisiae could have led to the increased expression of reverse transcriptase activity that ultimately led to the loss of introns in the genomic DNA (Aravind et al. 2000). This might explain why S. pombe is more like metazoan cells than S. cerevisiae with regard to both the presence of introns and the retention of RNAi.

Cell cycle

S. pombe and S. cerevisiae have both contributed greatly to our understanding of events and processes that determine and regulate passage through the eukaryotic cell cycle. It is instructive to compare the cell cycles of the two yeasts to appreciate both their similarities and their differences. S. cerevisiae cells start their cycles in G₁ as unbudded cells with unreplicated (1c) DNA content. Prior to “start,” a point in G₁, cells are uncommitted to the mitotic cycle or other developmental pathways (Hartwell et al. 1974), while completion of “start” commits the cell to the ongoing mitotic cycle. The G₁ phase lasts about one-quarter of the cell cycle and is followed by the S phase, a further third of the cycle. There is arguably no true G₂ phase because formation of the mitotic spindle starts during late G₁ or early S phase and does not depend on completion of S phase. Thus, the spindle persists through most of the cell cycle, though its final elongation, which separates the chromosomes, takes place well after the end of S phase. Following mitosis, the (larger) mother and (smaller) daughter cells separate, the mother embarking on a new cycle and the daughter becoming a mother cell, giving rise to a new bud. In contrast, S. pombe cells complete their short G₁ and S phases very soon after mitosis so that by the time of cell division, virtually every cell is already in G₂, and this phase comprises about three-quarters of the cycle (Nasmyth et al. 1979). The mitotic spindle is present for only a short period, as in most eukaryotes, and a symmetric cell division ensues. Other genetic (Nurse 1975) and physiologic studies (Fantes and Nurse 1977, 1978) indicated the presence of a major regulatory point between G₂ and M controlling entry into mitosis.

Initial studies of the cell cycles of the two yeasts did not reveal any similarities in the underlying control mechanisms. Indeed, for a considerable time it was believed that the budding yeast cell cycle was controlled only in G₁, while the fission yeast cell cycle was regulated at the G₂/M boundary. Therefore, it was surprising when S. pombe was shown to have a G₁ control point despite the shortness of this phase of the cell cycle (Nurse and Bissett 1981). Furthermore, the S. pombe G₁ and G₂ control points turned out to share a key component, the Cdc2 gene product, later shown to be a protein kinase (Simanis and Nurse 1986). The S. cerevisiae homolog of Cdc2, Cdc28, was known to be required for G₁ progression and was subsequently determined also to be needed at later stages of the cycle. Cdc2 and Cdc28 were shown to be functional homologs in that the CDC28 gene could rescue the defect of a cdc2 mutant and vice versa (Beach et al. 1982). Given the major differences between the cell cycles of the two yeasts and the large evolutionary distance between them, Nurse argued that if two disparate yeasts shared the Cdc2/28 function, other eukaryotes such as humans also might do so. This idea was proven by the isolation of a human Cdc2 homolog based on its ability to rescue the cell cycle defect of a S. pombe cdc2 mutant (Lee and Nurse 1987). This discovery was the basis of the awarding of the Nobel Prize to Nurse a decade later, shared with Lee Hartwell for his pioneering studies of cell cycle genes in S. cerevisiae and with Tim Hunt for his identification of the cyclin partner to Cdc2.

Coping with DNA damage during the cell cycle

Cells face intermittent threats to their survival and fitness. One of these is damage to nuclear DNA, which can arise spontaneously or following exposure to one of several environmental agents. Particularly damaging is ionizing radiation, which can cause double-strand DNA breaks. If left unrepaired, these breaks can lead to loss of large chromosomal regions and cell death. A very effective way to repair a broken chromosome is to copy the information from the corresponding sequence on another DNA molecule. This works well for diploid organisms because there are always at least two copies of a given chromosomal region present, but a haploid cell in G₁ phase has only a single copy of most sequences. Two quite different strategies have evolved for dealing with double-strand breaks in the two yeasts, and they relate to the different lifestyles of the yeasts.

A pair of haploid S. cerevisiae cells will mate with each other independently of the nutritional environment, provided that they are of opposite mating type, to form a mitotically stable diploid cell that proliferates indefinitely, as long as nutrients are available. Following starvation, the diploid undergoes a meiotic division to generate four ascospores, two of each mating type. On return to growth medium, the ascospores germinate and, unless artificially separated, mate with one another, once again generating diploids. Thus, although laboratory S. cerevisiae strains can be maintained as haploids, this is not the natural state for the organism. The diploid state ensures that the cell always has two copies of each genomic sequence and can deal effectively with broken chromosomes by copying the intact version of the damaged sequence from the chromosomal homolog. Thus, S. cerevisiae cells can have a long G₁ phase and still survive an otherwise lethal DNA-damaging event.

However, S. pombe cells will not mate with one another unless starved. Following starvation for (usually) nitrogen, cells of opposite mating types form a zygote that generally enters meiosis directly, with no extended diploid phase. The four resulting haploid ascospores each will proliferate as a haploid clone. Thus, the default lifestyle of S. pombe is as a haploid; diploid strains are unstable and difficult to propagate in the laboratory. So how do haploid S. pombe cells survive a double-strand DNA break, given that they have no sister chromosome to copy? The answer is that while the cells only have a single copy of each chromosome, the cells replicate their DNA and enter G₂ very early in the cell cycle so that the chromosomes are composed of two sister chromatids. Breakage of one chromatid therefore can be repaired using the intact chromatid as a template. This shows how these yeasts achieve the same goal via two distinct mechanisms.

Cell cycle

S. pombe research has played a unique and highly prominent role in understanding the cell cycle. Nurse combined Mitchison’s interest in the cell cycle with Leupold’s genetic approaches to identify genes that play essential roles in cell cycle progression (Nurse 1975; Nurse et al. 1976). The best known and most important outcome of this work was the cloning of the cdc2⁺ cyclin-dependent kinase gene (Beach et al. 1982) and the critical demonstration that the human homolog could supply this function to a cdc2^ts temperature-sensitive mutant (Lee and Nurse 1987). The mutant screens that identified cdc2⁺ also led to the discovery of several genes whose products interact with or act on the Cdc2 kinase, such as Cdc25 [the phosphatase that activates Cdc2 by dephosphorylation of a crucial phosphotyrosine residue (Russell and Nurse 1986)] and Cdc13 (Nurse et al. 1976; Fisher and Nurse 1996). A separate study led to the isolation of mutants with reduced cell size, nearly all of which had a defective wee1 gene, later shown to encode the tyrosine kinase that inhibits Cdc2 activity until cells are ready to pass from G₂ to M phase (Nurse 1975; Nurse and Thuriaux 1980; Russell and Nurse 1986, 1987; Gould and Nurse 1989). Furthermore, the importance of cell size in cell cycle control was supported by the direct demonstration that cell size at division is homeostatically controlled. A cell that is larger or smaller than normal at birth will undergo division at close to the normal size, so nearly all the deviation from normal is corrected within the same cycle (Fantes 1977).

Strains carrying mutations in other cdc genes facilitated studies of important processes in addition to entry to mitosis: because they displayed defects in DNA replication, septum formation/septation, and cytokinesis. For example, Cdc1, Cdc6/Pol3, and Cdc27 are subunits of DNA polymerase delta (MacNeill et al. 1996), while Cdc18, Mcm2 (Cdc19), Mcm4 (Cdc21), Cdc23, and Cdc24 all play roles in either regulation of S phase or DNA replication itself (Coxon et al. 1992; Kelly et al. 1993; Forsburg and Nurse 1994; Aves et al. 1998; Gould et al. 1998). Cdc7, Cdc11, Cdc14, and Cdc16 are required for function of the septation initiation network (SIN) that monitors the completion of mitosis to regulate the start of cytokinesis (McCollum and Gould 2001). Finally, Cdc3 (profilin), Cdc8 (tropomyosin), Cdc12 (formin), and Cdc15 are all required for cytokinesis (Fankhauser et al. 1993; Balasubramanian et al. 1994; Chang et al. 1997). Thus, the study of mutations that cause cell cycle defects in S. pombe formed the basis of many distinct areas of research.

In addition to requiring dedicated gene products, progress through the cell cycle is influenced by both intrinsic and extrinsic conditions. There exist several cell cycle checkpoints at which the cell determines whether or not conditions have been met to allow progression to the next phase of the cell cycle. Several related DNA checkpoints act at various cell cycle stages (Carr 2004), while the spindle checkpoint acts to regulate exit from mitosis (Hauf 2013). Similar checkpoints exist in all eukaryotes, and many of the components are conserved.

DNA checkpoints detect damaged or unreplicated DNA and block cell cycle progress until the damage is repaired or replication is completed, when progress resumes. In S. pombe, the best understood is the G₂/M checkpoint that prevents entry into mitosis while damaged DNA is present. Mutants defective in this checkpoint function have been isolated in both budding and fission yeast by their sensitivity to DNA-damaging agents such as radiation or agents that block replication such as hydroxyurea (Nasim and Smith 1975; Enoch et al. 1992). In outline, the checkpoint system works as follows: sites of damaged DNA are recognized by binding of checkpoint protein complexes, which activates a cascade of signaling components, ultimately preventing entry into mitosis. Gene products involved in binding to damaged or unreplicated DNA include the Rad3 protein kinase (Bentley et al. 1996) and its accessory factor Rad26 (Edwards et al. 1999). Further complexes involved in binding to sites of damage contain Rad17 (Griffiths et al. 1995), Hus1 (Kostrub et al. 1997), Rad1 (Sunnerhagen et al. 1990), Rad9 (Murray et al. 1991), and other proteins. Rad3 then interacts with other components, including the effector kinase Chk1 (Walworth et al. 1993; Al-Khodairy et al. 1994), whose activation results in increased tyrosine phosphorylation of Cdc2, which prevents entry into mitosis. The increase in Cdc2 phosphorylation appears to involve both downregulation of the Cdc25 phosphatase and upregulation of Wee1 (Furnari et al. 1997; Boddy et al. 1998; Raleigh and O’Connell 2000).

DNA checkpoints that act at other cycle stages—e.g., to check that the DNA is properly replicated in S phase—share many of the components and complexes involved in the G₂/M checkpoint, but there are also elements unique to one or other system. For instance, the effector kinase in the S-phase checkpoint is Cds1 (Murakami and Okayama 1995; Raleigh and O’Connell 2000) rather than Chk1 (Furnari et al. 1997; Boddy et al. 1998). Many of these components and mechanisms are conserved in mammalian cells, along with the functions of many more proteins involved in the detection and responses to DNA damage or stalled DNA replication (Sancar et al. 2004).

A significant difference between fission and budding yeast is that the final target of the G₂/M checkpoint in budding yeast is not Cdc28 (the functional homolog of Cdc2) but Pds1, a protein whose destruction is required for exit from metaphase and entry into anaphase (Sanchez et al. 1999). The controls in mammalian cells are in this respect similar to those in S. pombe. The G₂/M checkpoint is more complex than in S. pombe, but a major part of the G₂ delay following DNA damage is due to a reduction in Cdc25 activity. This leads to an increased level of Cdc2/Cdk1 tyrosine phosphoryation on the equivalent residue to that of S. pombe Cdc2 and hence its inactivation. In response to DNA damage in G₁, Cdc25A is phosphorylated and degraded so that it is unable to activate Cdk2–cyclin E, the cyclin complex required to initiate S phase (Sancar et al. 2004).

The observation that the nutritional composition of the growth medium influences cell cycle progress is an old one (Fantes and Nurse 1977), but until recently, the mechanism has not been clear. The fortuitous isolation of a mutant in which cell size was greatly affected by nutritional status (Ogden and Fantes 1986) led to the discovery and elucidation of the Sty1/Spc1 stress-activated MAP kinase signaling pathway that affects the rate of passage through the G₂/M control. This nutritional control involves signaling through the TOR system and the Polo-like kinase Plo1, which acts indirectly to regulate the tyrosine phosphorylation status and hence activity of Cdc2 (Petersen and Nurse 2007; Hartmuth and Petersen 2009). In addition to influencing progress at the G₂/M boundary, the Sty1/Spc1 pathway transduces nitrogen and glucose starvation stress as well as osmotic, oxidative, and heat stress to generate a spectrum of cellular responses (Chen et al. 2003).

Chromosome dynamics

S. pombe also has been a key model for the study of biological mechanisms in the area of chromosome dynamics. Possibly owing to the large sizes of the S. pombe chromosomes, mechanisms associated with chromosome replication, recombination, and segregation during meiosis and mitosis are areas in which S. pombe research has played a central role. Mitsuhiro Yanagida and coworkers identified proteins required for proper chromosomal pairing and segregation in mitosis by isolating cut mutants that formed a septum prior to chromosome segregation (Hirano et al. 1986; Samejima et al. 1993). Yanagida’s contributions to this field have been recognized with several awards, including the Imperial Prize and Japan Academy Prize in 2003 and the Order of Culture Prize (Japan) in 2011. He was elected as a foreign member of the Royal Society (United Kingdom) in 2000 and a foreign associate of the National Academy of Sciences (United States) in 2011.

Because S. pombe centromeres are large, repetitive structures, similar to those of mammalian cells and unlike S. cerevisiae centromeres, research on chromatid cohesion at the centromeres (Watanabe and Nurse 1999; Kitajima et al. 2004), kinetochores and their interactions with the mitotic spindle (Goshima et al. 1999), and centromere clustering and movement during mitosis (Funabiki et al. 1993) have been central to developing our understanding of these processes in metazoan chromosomes. For example, two S. pombe Shugoshin homologs, Sgo1 and Sgo2, connect the spindle checkpoint proteins to sister chromatid cohesion (Kitajima et al. 2004). This work has prompted studies of vertebrate Shugoshin-like proteins to show that they are involved in the same interactions as in S. pombe (Salic et al. 2004; Kitajima et al. 2005) and in chromosomal instability associated with various forms of cancer (Kahyo et al. 2011; Yamada et al. 2012).

Similarly, studies of the structure and behavior of telomeres at the ends of these large chromosomes have led to important discoveries about teleomeric proteins (Cooper et al. 1997; Baumann and Cech 2001; Chikashige and Hiraoka 2001; Martin-Castellanos et al. 2005) involved in controlling the clustering and positioning of telomeres during meiosis that contribute to proper chromosome segregation as well as the pairing of chromosomes for homologous recombination.

In addition, S. pombe continues to be a key model organism for the study of DNA replication, with many of the original cdc genes being involved in various aspects of DNA replication. Current studies are examining the locations of origins of replication, the timing of origin firing, the regulation of replication fork movement, and the response to events such as replication fork stalling. S. pombe origins of replication are somewhat like those of higher eukaryotes in their lack of a conserved sequence element (Mojardin et al. 2013), thus making S. cerevisiae an outlier in this regard. Along with studies of DNA replication mechanisms, it has been shown that gene expression is affected by whether a DNA replication fork passes through the gene in the sense or antisense direction and that this is controlled by the location of surrounding origins, the timing of origin firing, and the presence of binding sites for the Sap1 or Reb1 proteins that block replication fork progression (Sanchez-Gorostiaga et al. 2004; Krings and Bastia 2005; Mejia-Ramirez et al. 2005). Replication fork arrest is also used to control the direction of replication through a site in the mating locus (Figure 4) that carries an imprint that controls which daughter cell will undergo a mating switch event (Dalgaard and Klar 2001). Further studies of replication origins, DNA replication, and the interplay between DNA replication and mating-type switching or gene expression in S. pombe undoubtedly will lead to new insights that will inform our understanding of mammalian molecular mechanisms.

Chromatin, epigenetics, and gene expression

Both S. pombe and S. cerevisiae have been important organisms for the study of epigenetic control of gene expression. Expression of genes present at the mat2-P and mat3-M mating-type loci (Figure 4) and in telomeric and subtelomeric regions is silenced by a variety of mechanisms that ultimately affect the chromatin in these regions (Egel et al. 1984; Thon and Klar 1992; Thon et al. 1994; Allshire et al. 1995; Ekwall et al. 1996). In addition, the large heterochromatic centromeres of S. pombe have provided researchers with a third region in which to study silencing mechanisms (Allshire et al. 1994, 1995). Studies of heterochromatin formation at S. pombe centromeres have helped to establish a “histone code” that associates specific posttranslational modifications of histones with the transcriptional state of the associated chromatin (Jenuwein and Allis 2001). In S. pombe, Clr4 methylation of lysine 9 of histone H3 leads to the recruitment of the chromodomain protein Swi6 to nucleosomes containing the H3K9 mark (Nakayama et al. 2001). The establishment of transcriptionally silent heterochromatin also requires the function of the RNAi RITS complex (Volpe et al. 2002; Verdel et al. 2004), and Clr4, Swi6, and components of the RITS complex have been found to colocalize throughout the genome at heterochromatic regions (Cam et al. 2005). For an in-depth review of the many studies into the mechanisms and consequences of chromatin remodeling at S. pombe centromeres, see Allshire and Ekwall (2015). Finally, ncRNAs can affect chromatin remodeling and gene expression in an RNAi-independent manner. For example, transcriptional activation of the glucose-repressed fbp1⁺ gene involves the transient and stepwise expression of three ncRNAs that facilitate chromatin remodeling in the promoter region and the eventual expression of the mRNA (Hirota et al. 2008). Given the complexity of chromatin remodeling mechanisms at various regions of the S. pombe genome and the importance of this to both regulated gene expression and the maintenance of heterochromatic regions of the genome, this will remain a vibrant area of research for years to come.