Regulation of Cdc2p and Cdc13p Is Required for Cell Cycle Arrest Induced by Defective RNA Splicing in Fission Yeast*

Screening of cdc mutants of fission yeast for those whose cell cycle arrest is independent of the DNA damage checkpoint identified the RNA splicing-deficient cdc28 mutant. A search for mutants of cdc28 cells that enter mitosis with unspliced RNA resulted in the identification of an orb5 point mutant. The orb5+ gene, which encodes a catalytic subunit of casein kinase II, was found to be required for cell cycle arrest in other mutants with defective RNA metabolism but not for operation of the DNA replication or DNA damage checkpoints. Loss of function of wee1+ or rad24+ also suppressed the arrest of several splicing mutants. Overexpression of the major B-type cyclin Cdc13p induced cdc28 cells to enter mitosis. The abundance of Cdc13p was reduced, and the phosphorylation of Cdc2p on tyrosine 15 was maintained in splicing-defective cells. These results suggest that regulation of Cdc13p and Cdc2p is required for G2 arrest in splicing mutants.

The sequence of cell cycle events is highly regulated to ensure the faithful duplication and segregation of the genome associated with cell division. Various surveillance mechanisms, or checkpoints, coordinate and monitor such events as DNA replication, DNA repair, spindle formation, reorganization of the actin cytoskeleton, and changes in cell size (1)(2)(3)(4). If genotoxic stress results in damage to DNA, for example, checkpoint activation triggers cell cycle arrest before cells enter mitosis to provide sufficient time for the damage to be repaired. In many eukaryotes, the protein kinase Cdc2 (cyclin-dependent kinase 1) controls the onset of mitosis in a manner dependent on various internal and external conditions that include the presence of DNA damage, the status of DNA replication, nutrient availability, and cell size (5). The activity of Cdc2 is determined by the phosphorylation status of its Tyr 15 residue and the availability of cyclin (5). Inhibitory phosphorylation of Cdc2 on Tyr 15 is catalyzed by the tyrosine kinases Wee1 and Mik1, and the dephosphorylation of this residue is mediated predominantly by the tyrosine phosphatase Cdc25 (5).
Splicing of precursor mRNA (pre-mRNA) 3 is essential for the expression of most protein-coding genes in eukaryotes and is mediated by the sequential assembly and rearrangement of small nuclear ribonucleoprotein complexes, or spliceosomes, on the pre-mRNA (6). After completion of the splicing reactions responsible for the excision of each intron, the spliceosome dissociates from the mature mRNA, and the excised introns are rapidly degraded. The prp (pre-mRNA processing) mutants prp5, prp6, prp8, prp11, prp12, prp13, prp14, and prp17 of fission yeast (Schizosaccharomyces pombe) manifest both accumulation of pre-mRNAs and the cell division cycle arrest (cdc) phenotype (7)(8)(9)(10). In budding yeast (Saccharomyces cerevisiae), prp3, prp8, prp17, and prp22 mutants also show the cdc phenotype (11)(12)(13)(14). Furthermore, DBF3 and DBF5 in budding yeast are required not only for DNA replication but also for pre-mRNA splicing (11). In mammalian cells, a component of a 40 S small nuclear ribonucleoprotein-containing complex that is a homolog of fission yeast Cdc5p and budding yeast Cef1 contributes to the G 2 -M transition of the cell cycle (15)(16)(17). A human Dim1 family protein is an evolutionarily conserved U5 small nuclear ribonucleoprotein protein, and its fission yeast ortholog is required for entry into mitosis (18,19). Many genes are repressed during mitosis because of the downregulation of transcription, polyadenylation of RNA, and translation (20), and pre-mRNA splicing has also been found to be targeted for inhibition during mitosis (21). These various observations suggest that regulation of a link between pre-mRNA splicing and the cell cycle is highly conserved. In addition, defects in degradation of the intron lariat result in a marked delay in cell cycle progression (22), further supporting the notion that defective RNA metabolism triggers cell cycle arrest.
Many cell cycle regulators have been identified in genetic screens for cdc mutants in fission yeast (5). Some cdc mutants, including cdc1 (subunit of DNA polymerase ␦), cdc17 (DNA ligase), and cdc18 (homolog of budding yeast Cdc6), are defective in DNA metabolism, whereas others (cdc2, cdc13, or cdc25) are defective in the cell cycle machinery itself (5). In addition, two mutants, cdc5 and cdc28, have been found to be deficient in pre-mRNA splicing. Whereas cdc5 ϩ encodes a Myb-related protein, cdc28 ϩ is allelic with prp8 ϩ and encodes a DEAH box-containing an RNA helicase (9,16).
Many cdc mutants with defects in DNA replication arrest in a manner dependent on the DNA damage checkpoint (5), whereas cdc mutants with defects in the cell cycle machinery arrest independently of this checkpoint (23). We have now screened for cdc mutants that arrest independently of the DNA damage checkpoint to reveal new mechanisms of cell cycle arrest. We found that pre-mRNA splicing-deficient cdc28 cells arrest independently of the DNA damage checkpoint, and we investigated the mechanism of cell cycle arrest in these cells.

MATERIALS AND METHODS
Yeast Strains, Media, and Genetic Methods-Complete medium (YES, yeast extract plus supplement) and minimal medium (EMM, Edinburgh minimal medium) were prepared and standard methods were performed as described (24). The procedures for gene disruption and COOH-terminal tagging of proteins with HA or the Myc epitope were described previously (25). Transformation of S. pombe was performed by the lithium method (26). For microscopic analysis, cells were fixed with 70% ethanol and stained with 4Ј,6-diamidino-2-phenylindole (DAPI) as described (27).
Isolation of Mutants with Abrogated G 2 Arrest in cdc28 Cells-cdc28 cells were mutagenized with 1-methyl-3-nitro-1-nitrosoguanidine (28,29), spread on YES plates, and incubated for 4 days at 24°C. Colonies (8 ϫ 10 4 ) were replica-plated on YES medium containing PhloxinB (0.01 mg/ml) and incubated at 34°C, and strains that grew poorly or died were collected. Cells that did not show the cdc phenotype were selected and crossed with the wt strain. Among those that were not linked to wee1 ϩ , we focused on one mutant designated 3c13. We transformed the double 3c13 cdc28 mutant with an S. pombe genomic library and found that orb5 ϩ suppressed the phenotype of 3c13 but not that of cdc28. We crossed the 3c13 mutant with the orb5-19 mutant, which was isolated as a morphology mutant defective in reestablishment of polarized growth (30). Among 5000 cells examined, there were no wt cells, indicating that orb5 ϩ is the gene affected by the 3c13 mutation. Indeed, we detected two point mutations in the open reading frame of orb5 ϩ in 3c13 cells; Glu 59 and Gly 239 were changed to lysine and glutamic acid, respectively.
RNA Isolation and Northern Blot Analysis-Total RNA (5 g) extracted from cells with the use of Isogen (Wako) was denatured in loading buffer and fractionated by electrophoresis on a 1% agarose gel. The RNA molecules were transferred to a Hybond-Nϩ nylon membrane (Amersham Biosciences) and cross-linked to the membrane by ultraviolet irradiation. Probes (DNA fragments containing the coding regions of tfIId ϩ , cdc13 ϩ , cdc25 ϩ , or ade6 ϩ genes) were labeled with 32 P with the use of a random primer labeling kit (Amersham Biosciences). Northern blot analysis was performed as described (9).
Preparation of Synchronous Cultures-For synchronization in G 1 , exponentially growing cells in EMM medium were washed several times with and resuspended at a density of 2 ϫ 10 7 cells/ml in EMM medium without a nitrogen source. After incubation for 14 -15 h at 24°C, the cells were released into YES medium or, for those harboring plasmids, into EMM medium at 36.5°C. For synchronization in G 2 , cells grown to log phase at 24°C were layered on top of a 7-30% lactose gradient and centrifuged for 8 min at 1000 rpm and room temperature. Small cells recovered from the top of the gradient were inoculated into YES medium, incubated (or not) at 42°C for 50 min to inactivate ts protein efficiently, and then incubated at 36.5°C for various times. Cells collected at intervals of 10 -15 min were fixed in ethanol and stained with DAPI for monitoring of cell cycle progression.
To confirm that the arrest of cdc28 cells is independent of the DNA damage checkpoint, we constructed a cdc28 rad3 tel1 mutant. Most of these cells did not enter mitosis and arrested with the cdc phenotype at the restrictive temperature (Fig. 1a), verifying that cdc28 cells arrest independently of the DNA damage checkpoint.
Isolation of Mutants That Progress through Mitosis in the cdc28 Background-To clarify the mechanism of cell cycle arrest in cdc28 cells, we searched for mutants that progress through mitosis with defective pre-mRNA splicing. From 80,000 independent colonies, we isolated 10 mutants that failed to show the cdc phenotype and manifested reduced viability at the semirestrictive temperature. Genetic crosses and phenotypic analyses revealed that four of these mutants were linked to wee1 ϩ FIGURE 1. Cell cycle delay in cdc28 cells requires orb5 ؉ , wee1 ؉ , and rad24 ؉ . a-c, cells of the indicated genotypes were grown to late log phase, washed, and cultured in nitrogen-free medium for 14 -15 h at 24°C. They were then transferred to YES medium at time 0 and incubated at 36.5°C for the indicated times, after which the proportion of cells with two nuclei was determined. At 6 h after the temperature shift, samples were also processed for staining with DAPI. Representative micrographs of the cells are shown in a and b. d, cells of the indicated genotypes were transformed with pcL-cdc2 ϩ , pcL-cdc2-F15 or the empty vector, grown to log phase at 24°C. They were then transferred to 36.5°C at time 0. After 8 h the proportion of the cut phenotype was determined. e, cells of the indicated genotypes were transformed with pcL-cdc13 ϩ or the empty vector, grown to late log phase, washed, and cultured in nitrogen-free EMM medium for 14 -15 h at 24°C. They were then transferred to EMM medium containing nitrogen at time 0 and incubated at 36.5°C for the indicated times, after which the proportion of binucleate cells was determined.
(32) and one was to orb5 ϩ (see "Materials and Methods"). The orb5 ϩ gene encodes a catalytic subunit of casein kinase II (CK2), which is a highly conserved serine-threonine kinase (33). The orb5-3c13 single mutant exhibited no obvious defect in cell cycle progression (Fig. 1b), although its growth rate was slightly reduced at high temperature, whereas complete deletion of orb5 ϩ was lethal. 4 Ckb1p is a regulatory subunit of CK2 in fission yeast (34). Like cdc28 orb5 cells, cdc28 ckb1 cells also entered mitosis, 4 suggesting that fission yeast CK2 is required for cell cycle arrest in splicing mutants. orb5 ϩ , wee1 ϩ , and rad24 ϩ Are Required for Cell Cycle Delay in cdc28 Cells-Given that we identified both wee1 ϩ and orb5 ϩ as potential key regulators of cell cycle arrest in cdc28 cells, we investigated the phenotype of the double mutants further. We synchronized cdc28, cdc28 orb5, and cdc28 wee1 cells in G 1 by nitrogen deprivation and released them into rich medium at the restrictive temperature (36.5°C). At 4 h, the frequency of binucleate cells was increased for wt, cdc28 orb5, and cdc28 wee1 cells but not for cdc28 cells (Fig. 1, a and b). These results suggested that wee1 ϩ and orb5 ϩ are required for inhibition of the entry of cdc28 cells into mitosis. The rad24 ϩ gene is required for the DNA damage checkpoint (1, 5) as well as for the cell cycle arrest induced in fission yeast by accessory protein R (Vpr) of human immunodeficiency virus (35). We found that cdc28 rad24 cells progressed through mitosis (Fig.  1b). Taken together these results indicate that orb5 ϩ , wee1 ϩ , and rad24 ϩ are required for cell cycle arrest in cdc28 cells.
To examine the relation between orb5 ϩ and wee1 ϩ , we constructed a cdc28 orb5 wee1 mutant. The proportion of cells that entered mitosis was greater for the triple mutant than for the cdc28 orb5 or cdc28 wee1 double mutants (Fig. 1c), suggesting that orb5 ϩ and wee1 ϩ have an additive effect on entry into mitosis in cdc28 cells. Whereas cdc28 mik1 cells showed a cdc phenotype similar to that of cdc28 cells, 4 cdc28 mik1 wee1 cells entered mitosis more frequently than did cdc28 wee1 cells (Fig. 1c). In addition, expression of cdc2-F15, which encodes a Cdc2p mutant in which Tyr 15 is replaced with phenylalanine induced the cut phenotype in cdc28 cells grown at the restrictive temperature (36.5°C) but not in wild-type cells (Fig. 1d). These results suggest that phosphorylation of Cdc2p on Tyr 15 is required for the arrest of cdc28 cells.
We next examined whether overexpression of cdc13 ϩ induced cdc28 cells to enter mitosis (Fig. 1e). Although overexpression of cdc13 ϩ did not affect the frequency of M phase entry in wt cells, it partially alleviated the arrest of cdc28 cells. These results suggest that the regulation of Cdc2p both by phosphorylation on Tyr 15 and by Cdc13p contributes to the arrest of cdc28 cells.
orb5 ϩ , wee1 ϩ , and rad24 ϩ Are Required for Cell Cycle Arrest in prp12 or dbr1 Cells-To test whether orb5 ϩ , wee1 ϩ , and rad24 ϩ are required for cell cycle arrest in another pre-mRNA splicing-deficient mutant, we examined prp12 cells. Prp12p is required for formation of a functional U2 small nuclear ribonucleoprotein (7). We found that prp12 orb5, prp12 wee1, and prp12 rad24 cells entered mitosis, whereas most prp12 cells did not (Fig. 2a), suggesting that orb5 ϩ , wee1 ϩ , and rad24 ϩ are all required for cell cycle arrest in prp12 cells.
We next investigated whether orb5 ϩ , wee1 ϩ , and rad24 ϩ are required for cell cycle delay caused by a defect in RNA metabolism other than impaired pre-mRNA splicing. In fission yeast, dbr1 cells accumulate introns to high levels, grow slowly, and exhibit the cdc phenotype (22). Like the splicing mutants, dbr1 cells showed a marked delay in entry into mitosis, whereas dbr1 orb5, dbr1 wee1, or dbr1 rad24 cells entered mitosis more frequently (Fig. 2b). These results thus indicate both that orb5 ϩ , wee1 ϩ , and rad24 ϩ are required for cell cycle delay in dbr1 cells as well as that accumulated introns induce cell cycle delay by a mechanism similar to that operative in splicing mutants.
Contribution of orb5 ϩ , wee1 ϩ , rad24 ϩ , and cdc13 ϩ to Cell Survival in cdc28, prp12, or dbr1 Cells-If orb5 ϩ , wee1 ϩ , and rad24 ϩ contribute to cell cycle arrest in splicing mutants, it would be expected that cell survival would decrease when the arrest is abrogated. To test this prediction, we determined the survival of G 1 -synchronized cells after shifting the incubation temperature to 36.5°C for 8 h. The viability of cdc28 orb5, cdc28 wee1, or cdc28 rad24 cells decreased markedly as a result of incubation at the restrictive temperature, whereas that of the corresponding single mutants remained largely unchanged (Fig. 3a). Similarly, the viability of prp12 orb5, prp12 wee1, or prp12 rad24 mutants as well as that of dbr1 orb5, dbr1 wee1, or dbr1 rad24 cells was decreased by incubation at 36.5°C, whereas that of prp12 and dbr1 single mutants remained constant. In addition, overexpression of cdc13 ϩ also reduced the viability of cdc28 cells (Fig. 3b). These results suggest that the mech-  were grown to late log phase, washed, and cultured in nitrogen-free medium for 14 -15 h at 24°C. They were then transferred to YES medium at time 0 and incubated at 36.5°C for the indicated times, after which the proportion of cells with two nuclei was determined. SEPTEMBER 23, 2005 • VOLUME 280 • NUMBER 38 anism of cell cycle arrest regulated by orb5 ϩ , wee1 ϩ , rad24 ϩ , and cdc13 ϩ is important for cell survival in RNA metabolism mutants. orb5 ϩ Is Not Required for the DNA Replication and DNA Damage Checkpoints-To investigate whether orb5 ϩ is important for S phase arrest induced by inactivation of DNA replication proteins, we constructed double mutants of orb5 and several cdc mutants that arrest in S phase (cdc1, cdc17, cdc18, orp1, or swi7) (5,28,36). Orp1p is the largest subunit of the origin recognition complex, and Swi7p is the catalytic subunit of DNA polymerase ␣. The cdc1 orb5, cdc17 orb5, cdc18 orb5, orp1 orb5, and swi7 orb5 mutants manifested the cdc phenotype similar to that of the corresponding single cdc mutants and most of the double mutant cells did not enter mitosis (Fig. 4, a and b), suggesting that orb5 ϩ is not required for cell cycle arrest in the single cdc mutants. In contrast, a substantial proportion of cdc1 wee1, cdc17 wee1, cdc18 wee1, orp1 wee1, swi7 wee1, cdc1 rad24, cdc17 rad24, cdc18 rad24, and orp1 rad24 cells entered mitosis, suggesting that wee1 ϩ and rad24 ϩ contribute to cell cycle arrest in these mutants.

Cell Cycle Arrest and RNA Splicing
We next tested whether orb5 ϩ is required for the arrest induced either by hydroxyurea, which inhibits DNA replication, or by methylmethane sulfonate, which induces DNA alkylation. Like wt cells, orb5 cells showed the cdc phenotype after treatment with either of these agents (Fig. 4c). These data suggested that orb5 ϩ is not required for operation of the DNA replication and DNA damage checkpoints.
cdc28 orb5, cdc28 wee1, and cdc28 rad24 Cells Progress through Mitosis with Unspliced mRNA-It was possible that cdc28 orb5, cdc28 wee1, and cdc28 rad24 cells entered mitosis by suppressing the splicing defect of cdc28 cells. To test this possibility, we performed Northern blot analysis of the tfIId ϩ gene, which encodes a TATA box-binding factor and contains three introns (37). After incubation of cells for 4 or 8 h at 36.5°C, substantial amounts of tfIId ϩ pre-mRNA had accumulated in cdc28, cdc28 orb5, cdc28 wee1, and cdc28 rad24 cells but not in wt cells (Fig. 5). These results indicate that orb5, wee1, and rad24 do not suppress the splicing defect of cdc28 cells and that cdc28 orb5, cdc28 wee1, and cdc28 rad24 cells enter and progress through mitosis with unspliced mRNA.
orb5 ϩ , wee1 ϩ , and rad24 ϩ Are Required Specifically for G 2 -M Arrest in cdc28 Cells-To confirm that orb5 ϩ , wee1 ϩ , and rad24 ϩ are required specifically for G 2 -M arrest in cdc28 cells, we synchronized various mutants in early G 2 phase by lactose gradient centrifugation. The temperature was shifted to 36.5°C at time 0, and G 2 -M progression was  Viability of cdc28, prp12, or dbr1 cells that enter mitosis. Cells of the indicated genotypes (a) or wild-type or cdc28 cells transformed with pcL-cdc13 ϩ or the empty vector (b) were grown to late log phase, washed, and cultured in nitrogen-free medium for 14 -15 h at 24°C. They were then transferred to YES medium (a) or EMM medium (b) at time 0 and incubated (or not) at 36.5°C for 8 h (a) or 12 h (b). Cells were plated on YES medium (a) or EMM medium (b) and then cultured for 5 days at 24°C. Viability was expressed as a percentage of the value for wt cells at time 0 (a) or for the respective cells at time 0 (b). monitored (Fig. 6a). Whereas wt and orb5 cells entered mitosis at ϳ120 min after the temperature shift, wee1 cells did so much earlier (32). In contrast to cdc28 cells, cdc28 orb5 and cdc28 wee1 cells entered mitosis, although progression through mitosis was delayed in cdc28 orb5 cells. These results suggested that both orb5 ϩ and wee1 ϩ are required for G 2 arrest in cdc28 cells. Similar to cdc28 wee1 cells, cdc28 rad24 cells entered mitosis earlier than did wt cells, 4 suggesting that rad24 ϩ is also required for G 2 arrest in cdc28 cells and that Rad24p is a rate-limiting negative regulator of the G 2 -M transition. However, it remains to be determined that Wee1, Orb5, and Rad24 are transducers or targets of the G 2 arrest in splicing mutants.
Phosphorylation of Cdc2p on Tyr 15 and the Amount of Wee1p Are Maintained in cdc28 Cells-Phosphorylation of Cdc2p on Tyr 15 and the amounts of both the kinase Wee1p and the B-type cyclin Cdc13p are key determinants of the onset of mitosis (5,23). We monitored the amounts of the Tyr 15 -phosphorylated form of Cdc2p, total Cdc2p, Wee1p, Cdc13p, and Orb5p in wt and cdc28 cells synchronized by lactose gradient centrifugation (Fig. 6b). The abundance of Cdc2p remained virtually unchanged during the incubation of wt or cdc28 cells for up to 165 min. The amount of the Tyr 15 -phosphorylated form of Cdc2p decreased transiently during nuclear division in wt cells but was maintained at a high level in cdc28 cells. The level of Cdc13p gradually increased during nuclear division and then decreased in wt cells, whereas it maintained a    15 and of a low level of Cdc13p in cdc28 cells at G 2 arrest. Cells of the indicated genotypes were grown to mid log phase in YES medium at 24°C, and those in early G 2 phase were then collected by lactose gradient centrifugation. The cells were subjected to heat shock at 42°C for 50 min before incubation at 36.5°C for the indicated times. Samples were then subjected either to DAPI staining for determination of the percentage of cells entering mitosis (a) or to immunoblot analysis with antibodies specific for Cdc2p, for Cdc2p phosphorylated on Tyr 15 , for Cdc13p, or for ␣-tubulin (control) or, in the case of cells expressing hemagglutinin epitope (HA)-tagged Wee1p or Myc epitope-tagged Orb5p, with antibodies to HA or to Myc (b). low level in cdc28 cells. The abundance of Wee1p decreased on entry of wt cells into mitosis but was maintained at a high level in cdc28 cells. The amount of Orb5p remained virtually constant in both wt and cdc28 cells. In addition, the kinase activity of Orb5p was similar in wt and cdc28 cells and did not change during cell incubation. 4 These results suggest that the maintenance of Cdc2p phosphorylation on Tyr 15 , of a high level of Wee1p, and of a low level of Cdc13p is important for G 2 arrest in cdc28 cells, consistent with our findings that mutation of wee1 ϩ or overexpression of cdc13 ϩ induced the entry of cdc28 cells into mitosis.
Changes in Cdc2p Phosphorylation on Tyr 15 and in the Abundance of Cdc13p and Wee1p in cdc28 Cells Undergoing Mitosis-Given that orb5 ϩ and wee1 ϩ are required for G 2 arrest in cdc28 cells, we monitored the amounts of Cdc2p, the Tyr 15 -phosphorylated form of Cdc2p, and Cdc13p in cdc28 orb5, cdc28 wee1, and cdc28 rad24 cells 4 synchronized by lactose gradient centrifugation (Fig. 6b). The abundance of Cdc2p did not change during the incubation of either mutant. The amount of the Tyr 15 -phosphorylated form of Cdc2p remained unchanged in cdc28 orb5 cells, probably because the percentage of cells in mitosis at any one time is low.
Orb5p Is Required to Maintain Cdc13p at a Low Level in cdc28 Cells-To clarify the relative contributions of orb5 ϩ , wee1 ϩ , and mik1 ϩ to the regulation of Cdc2p activity, we constructed several double, triple, and quadruple mutants and monitored the level of Cdc2p phosphorylation on Tyr 15 and the abundance of Cdc13p after growth of the cells to log phase and subsequent incubation at 36.5°C. The amount of Cdc13p was decreased at 4 h after the temperature shift in cdc28, cdc28 wee1, cdc28 wee1 orb5, cdc28 wee1 mik1, and cdc28 wee1 mik1 orb5 cells, whereas it remained high in wt, cdc28 orb5, and cdc28 mik1 orb5 cells (Fig. 7). The level of Cdc13p was higher in cdc28 wee1 orb5 and cdc28 wee1 mik1 orb5 cells than in cdc28 wee1 mik1 cells. These results suggest that orb5 ϩ is required to maintain a low level of Cdc13p. The down-regulation of Cdc13p in cdc28 wee1 cells was unexpected. Although the mechanism by which Wee1p maintains the abundance of Cdc13p is unknown, it is possible that the Tyr 15 -dephosphorylated form of Cdc2p regulates its own activity by reducing the level of Cdc13p.
Regulation of Cdc13p at the Protein and mRNA Levels in RNA Metabolism Mutants-Given that either overexpression of cdc13 ϩ induced cdc28 cells to enter mitosis, we monitored the amounts of Cdc13p in several splicing-defective mutants after a temperature shift to 36.5°C. The abundance of Cdc13p decreased in cdc5, prp12, and prp13 cells but not in cdc25 and swi7 cells (Fig. 8a). The reduced level of Cdc13p in the mutant cells was not simply attributable to cell cycle arrest, given that the amount of this protein was maintained at a high level in G 2 -arrested cdc25 cells and S phase-arrested swi7 cells. Similarly, prp11 and dbr1 cells, but not wt cells, manifested down-regulation of Cdc13p in response to a temperature shift to 18°C (Fig. 8b). The low abundance of Cdc13p in cdc28 cells appeared to be attributable, at least in part, to a low level of cdc13 ϩ mRNA whose mobility was retarded at 4 and 8 h (Fig. 8c). However, the amount of cdc13 ϩ mRNA was also low in cdc28 orb5, cdc28 wee1, and cdc28 rad24 cells, suggesting that maintenance of the Cdc13p level in cdc28 orb5 cells is not attributable to the up-regulation of cdc13 ϩ mRNA.

DISCUSSION
Our investigation into the mechanism of cell cycle arrest in pre-mRNA splicing mutants of fission yeast has revealed the following. 1) These mutant cells arrest in a manner dependent on Cdc2p but not on proteins that mediate the DNA damage checkpoint, consistent with previous observations with cdc5 cells (16). 2) The arrest correlates with the maintenance both of Cdc2p phosphorylation on Tyr 15 and of a low abundance of the B-type cyclin Cdc13p. 3) Wee1p, Rad24p, and Orb5p  Samples collected at the indicated times thereafter were subjected either to immunoblot analysis with antibodies specific for the indicated proteins (a, b) or to Northern blot analysis with probes specific for cdc13 ϩ or ade6 ϩ transcripts (c).
are required for maintenance of the viability of the splicing mutants. 4) The reduced levels of Cdc13p in cdc28 cells are attributable at least in part to down-regulation of mRNAs. 5) Orb5p is required for cell cycle arrest in the splicing mutants but not for the DNA replication or DNA damage checkpoints.
There are at least two possible mechanisms by which defects in pre-mRNA splicing might halt cell cycle progression. One possibility is that proteins essential for entry of cells into mitosis are not expressed as a result of the splicing defect; 43% of the genes in fission yeast contain introns (38). At least three proteins are essential for mitotic entry: Cdc2p, Cdc13p, and Cdc25p. Cdc2p appeared to be expressed normally in the cdc28 mutant, even though the cdc2 ϩ gene contains introns. In contrast, the amount of Cdc13p was decreased in the mutant, probably because of the low level of the corresponding mRNA, even though the cdc13 ϩ gene does not contain introns. Similarly, the amounts of both Cdc25p and cdc25 ϩ mRNA did not vary significantly in cdc28 cells, 4 again even though the cdc25 ϩ gene does not contain introns. The reduced expression of Cdc13p in the mutant cells is thus not directly attributable to defective pre-mRNA splicing. Rather, transcription of cdc13 ϩ gene or degradation of the mature mRNAs is likely to be affected indirectly by the splicing defect. In addition, the slower migrating form of cdc13 ϩ mRNA was observed in the cdc28 mutant. The reason for this slow migration is unknown at present, but it is possible that pre-mRNA containing no intron is also processed by the splicing machinery, which may affect transcription initiation site or polyadenylation. However, we cannot exclude the possibility that the mechanism of G 2 arrest in splicing mutants might be because of the reduced expression of other genes essential for G 2 -M progression. Screening for those genes that suppress G 2 arrest of splicing mutants using cDNA library would help the elucidation of this mechanism. The other possible explanation for the cell cycle arrest induced by defective pre-mRNA splicing is that a regulatory system such as the DNA damage checkpoint is responsible. It is likely that such a checkpoint-like system operates in fission yeast, given that cdc28 cells with orb5, wee1, or rad24 mutations were found to enter mitosis with unspliced pre-mRNA. In budding yeast, cell cycle arrest in cef1 cells, which are defective in pre-mRNA splicing, results from inefficient splicing of the TUB1 gene, which encodes the major isoform of ␣-tubulin (39). Similar to cef1 cells, the cell cycle arrest apparent in other mutants of budding yeast with defects in pre-mRNA splicing is because of activation of the spindle checkpoint (40). Cell cycle arrest in cdc28 cells of fission yeast was not attributable to inefficient splicing of a tubulin gene, given that mutation of spindle checkpoint proteins (mad2 or bub1) or the expression of ␣-tubulin cDNA (nda2 ϩ or nda3 ϩ ) failed to suppress G 2 arrest in these cells. 4 The mechanisms of cell cycle arrest induced by defects in pre-mRNA splicing thus appear to differ between fission yeast and budding yeast.
In mammalian cells, expression of a dominant negative mutant of human Cdc5, a homolog of fission yeast Cdc5p, slowed G 2 progression and delayed entry into mitosis, whereas overexpression of wild-type human Cdc5 shortened the G 2 phase (15), suggesting that the link between cell cycle regulation and pre-mRNA splicing is conserved among eukaryotes. Consistent with this notion, cyclin E associates with components of the pre-mRNA splicing machinery in mammalian cells (41) and Cdc2p of mammalian cells phosphorylates the splicing factor SF2 (ASF) (42).
The mechanism of cell cycle arrest caused by defects in pre-mRNA splicing in fission yeast differs from that caused by inhibition of DNA replication or by DNA damage in several respects. First, proteins that mediate the DNA replication or DNA damage checkpoints are not required for the arrest because of deficient splicing; cdc28 rad3 tel1 cells thus failed to enter mitosis. Rad1p is also not required for the arrest in cdc5 cells (16). Second, Orb5p is required for cell cycle arrest in splicing mutants but not for the operation of the DNA replication or DNA damage checkpoints. Cell cycle progression in cdc28 orb5 cells was not identical to that in cdc28 wee1 cells. The frequency of the cut phenotype was low in the former cells, even though they progressed through mitosis with unspliced mRNA, 4 probably because DNA replication was completed, and DNA was not damaged and because the orb5 mutation, unlike the wee1 mutation, did not advance cell cycle progression. Third, a level of Cdc13p is negatively correlated with cell cycle arrest in splicing mutants, given that overexpression of cdc13 ϩ abrogated arrest, and Cdc13p is maintained at a low concentration in cdc28 cells. In contrast, the abundance of Cdc13p is not affected markedly and overexpression of cdc13 ϩ does not induce mitosis in cells in which the DNA replication or DNA damage checkpoint is activated (43).
There are several common characteristics of cell cycle arrest induced by defective pre-mRNA splicing or by activation of the DNA replication or DNA damage checkpoints. The phosphorylation of Cdc2p on Tyr 15 is required for arrest in all these instances (5). Furthermore, Rad24p is required for cell cycle arrest in response either to defective splicing or to DNA damage (1,5). Budding yeast with a defect in synthesis of the cell wall also halts cell cycle progression at G 2 (44). This cell wall checkpoint requires down-regulation of B-type cyclin, similar to that observed in the splicing mutants of fission yeast.
The Vpr protein of human immunodeficiency virus-type 1 induces G 2 arrest in both human and fission yeast cells through phosphorylation of Cdc2p on Tyr 15 (35,45). This cell cycle arrest requires Wee1p and Rad24p but not proteins that mediate the DNA damage or replication checkpoints. The mechanism of Vpr-induced arrest is similar to that apparent in the splicing mutants. The proteins with which Vpr interacts, which remain to be identified, might thus also mediate cell cycle arrest in splicing mutants.
As is the case for the DNA replication and DNA damage checkpoints, abrogation of cell cycle arrest reduced the viability of splicing mutants. Similar to transcription, polyadenylation, and translation, splicing is a target for mitotic inhibition (20,21). Such controls likely ensure that inappropriate gene expression does not occur during mitosis, thereby preventing aberrant proteins or RNAs from acting in a dominant negative manner to inhibit normal gene function. Progression of cells into M phase with unspliced RNA may therefore directly result in a loss of viability in splicing mutants. These facts imply that cells are equipped with the ability to arrest in G 2 in the presence of unspliced RNA.