The rsf12 mutation was isolated in a synthetic lethal screen for genes functionally interacting with Swi4. RSF12 is CLB5. The clb5 swi4 mutant cells arrest at G2/M due to the activation of the DNA-damage checkpoint. Defects in DNA integrity was confirmed by the increased rates of chromosome loss and mitotic recombination. Other results suggest the presence of additional defects related to morphogenesis. Interestingly, genes of the PKC pathway rescue the growth defect of clb5 swi4, and pkc1 and slt2 mutations are synthetic lethal with clb5, pointing to a connection between Clb5, the PKC pathway, and Swi4. Different observations suggest that like Clb5, the PKC pathway and Swi4 are involved in the control of DNA integrity: there is a synthetic interaction between pkc1 and slt2 with rad9; the pkc1, slt2, and swi4 mutants are hypersensitive to hydroxyurea; and the Slt2 kinase is activated by hydroxyurea. Reciprocally, we found that clb5 mutant is hypersensitive to SDS, CFW, latrunculin B, or zymolyase, which suggests that, like the PKC pathway and Swi4, Clb5 is related to cell integrity. In summary, we report numerous genetic interactions and phenotypic descriptions supporting a close functional relationship between the Clb5 cyclin, the PKC pathway, and the Swi4 transcription factor.
CELL cycle progression is governed by the periodic activation and inactivation of different cyclin-dependent kinases (CDK) (Morgan 1997; Roberts 1999). In Saccharomyces cerevisiae, Cdc28 is the CDK responsible for the major cell cycle transitions through its association with nine different cyclins: Cln1–3 and Clb1–6 (Andrews and Measday 1998; Mendenhall and Hodge 1998; Miller and Cross 2001). At the end of G1, in a process called Start (restriction point in mammalian cells), the yeast cells commit to a new round of cell division when external and internal conditions are appropriate, switching on the events associated with cell cycle initiation: DNA replication, bud morphogenesis, and duplication of the spindle pole body (Wittenberg and Flick 2003). Execution of Start depends on the initial activation of Cdc28 by the Cln3 cyclin, which in turn, triggers a program of G1/S gene expression. More than 200 genes are periodically transcribed at this point of the cell cycle (Cho et al. 1998; Spellman et al. 1998), leading to the accumulation of proteins that are required for subsequent processes. The transcribed genes include the CLN1, CLN2, CLB5, and CLB6 cyclin genes. The synthesis of Cln1 and Cln2 results in an abrupt accumulation of Cln1,2-Cdc28 kinase activities, which trigger budding and duplication of the spindle pole body (Gulli and Peter 2001; Haase et al. 2001; Moffat and Andrews 2004). Clb5 and Clb6, the primary regulators of the initiation of DNA replication (Epstein and Cross 1992; Schwob and Nasmyth 1993; Donaldson et al. 1998; Weinreich et al. 2001), are also synthesized at this point; however, the Clb5,6-Cdc28 complexes are inactive due to the action of the CDK inhibitor Sic1 (Schwob et al. 1994). Only when Sic1 is degraded after phosphorylation by Cln-Cdc28 does the Clb5,6-Cdc28 kinase activity appear to initiate DNA replication (Schneider et al. 1996; Tyers 1996).
Expression of the G1/S-specific genes is mediated by the SBF and MBF transcription factors, which are composed of a common subunit, Swi6, and of different sequence-specific DNA-binding subunits, Swi4 and Mbp1, respectively (reviewed in Breeden 1996, 2003). This transcriptional system is essential for cell viability, as deduced from the lethality of the swi4 swi6 and swi4 mbp1 double mutants. SBF controls the expression of CLN1 and CLN2 genes as well as many genes encoding proteins involved in morphogenesis. MBF regulates periodic expression of the CLB5 and CLB6 genes as well as many genes involved in DNA metabolism. Extensive crosstalk exists between SBF and MBF, although SBF appears to play a dominant role in Start, since the loss of Swi4 causes severe cell cycle defects whereas inactivation of Mbp1 is of little consequence. The binding of SBF (and probably of MBF, although not proven) to the target promoters at early G1 is not sufficient to activate transcription (Cosma et al. 2001). The repressor Whi5 binds and inhibits SBF, and only at the end of G1, in response to cell growth, does Cln3-Cdc28 antagonize the Whi5 function, leading to G1/S-specific transcriptional activation (Costanzo et al. 2004; de Bruin et al. 2004).
The precise duplication of the genetic material is critical for successful cell division. DNA replication is tightly controlled by a two-step mechanism, which guarantees that all the genome is replicated only once per cell cycle (reviewed in Lei and Tye 2001; Bell and Dutta 2002; Diffley and Labib 2002). In a first step, during the G1 phase, competent prereplicative complexes (pre-RC) are assembled by the binding of Cdc6 and Cdt1 to the origin replication complex (ORC)-bound origins, which direct the recruitment of the helicase Mcm2-7 complex to the replication origins. Then, during the S phase, initiation of DNA synthesis is triggered by CDK activity. As mentioned above, Clb5,6-Cdc28 are the main kinase activities that drive the activation of origins in S. cerevisiae. Clb5 plays a more predominant role than Clb6 since it promotes the timely activation of early and late origins, whereas Clb6 is able to activate only early origins (Donaldson et al. 1998). Although DNA replication is not blocked in the absence of these cyclins (due to the action of other Clb cyclins), the timely onset and progression of DNA replication is severely affected (Epstein and Cross 1992; Schwob and Nasmyth 1993; Schwob et al. 1994; Donaldson et al. 1998). Firing of the origins destroy the pre-RC complexes and high Clb-Cdc28 activity blocks the formation of new pre-RC complexes by different mechanisms acting on Cdc6, the minichromosome maintenance (MCM) complex, and the ORC complex (Labib et al. 1999; Drury et al. 2000; Nguyen et al. 2000, 2001; Weinreich et al. 2001; Mimura et al. 2004; Wilmes et al. 2004). Only after mitotic exit, when CDK activity vanishes, are new competent origins allowed to assemble. The proper control of replication origins, both in timely firing and in preventing aberrant initiation, is crucial to avoiding DNA damage (Lengronne and Schwob 2002; Sidorova and Breeden 2002; Tanaka and Diffley 2002; Watanabe et al. 2002; Wilmes et al. 2004). In addition, checkpoint mechanisms help to guarantee DNA integrity. The DNA-damage and the DNA replication checkpoints respond to DNA lesions or defects in the replication forks, halting cell cycle progression to avoid genomic instability (reviewed in Melo and Toczyski 2002; Osborn et al. 2002).
The S. cerevisiae protein kinases C (PKC) pathway is essential to maintain cell integrity and to coordinate polarized growth and cell proliferation (reviewed in Heinisch et al. 1999; Hohmann 2002). The pathway is activated under many conditions that cause cell surface stress, such as growth at high temperature, a hypo-osmotic shock, periods of polarized growth such as budding or mating, actin perturbation, or the presence of compounds or mutations that interfere with cell wall synthesis. PKC pathway activation is essential to maintain cell integrity under these conditions. Cell surface stress is detected by sensor proteins (Wsc1-4, Mid2, Mtl1) that activate the small GTP-binding protein Rho1, which in turn binds and activates the Pkc1 kinase. Pkc1 regulates a MAP kinase cascade consisting of the MAPKKK Bck1, the MAPKK Mkk1 and Mkk2, and the MAPK Slt2. Activation of Slt2 induces the expression of genes involved in cell wall biosynthesis. Most of this transcriptional response is mediated by the Rlm1 transcription factor (Jung and Levin 1999; Garcia et al. 2004). In addition, the PKC pathway regulates the changes in actin cytoskeleton polarization that occur in response to the cell surface stress (Delley and Hall 1999). Diverse studies have related the PKC pathway to the G1/S transition (Igual et al. 1996; Marini et al. 1996; Zarzov et al. 1996; Gray et al. 1997), a period associated with highly polarized growth, and in fact, the SBF transcription factor has been established as a target of Slt2 kinase (Madden et al. 1997; Baetz et al. 2001).
In this study, we characterized a synthetic lethal interaction between the swi4 and the clb5 mutations. Overexpression of genes in the PKC pathway suppressed the growth defect of the clb5 swi4 mutant, and furthermore we identified a synthetic interaction between mutations in the PKC pathway and clb5. Other genetic interactions and phenotypic similarities support a close functional connection between the Clb5 cyclin, the PKC pathway, and the SBF transcription factor in the control of such as diverse cellular processes as DNA integrity and cell morphogenesis.
MATERIALS AND METHODS
Strains and growth conditions:
The yeast strains used in this study are shown in Table 1. Strains containing the swi4∷LEU2 and clb5∷URA3 disruption cassettes were obtained by using plasmids Bd194 (from L. Breeden) or 2718 (from E. Schowb), respectively. Strains containing rad9∷HIS3, rad24∷URA3, clb1∷URA3, clb2∷LEU2, clb3∷TRP1, clb4∷HIS3, and swe1∷kanr alleles were obtained by PCR amplification of the specific disruption cassettes using as template genomic DNA from CD102-1a, CD104-2c, CD116-8d (from B. Futcher), rad9rad24, or swe1 (from M. A. de la Torre) strains. clb6∷kanMX6 strains were constructed by integrating a DNA fragment amplified from plasmid pFA6-kanMX6 (from J. R. Pringle). Tagging of the Clb2 protein at the C terminus with three copies of the hemagglutinin (HA) epitope was achieved by integrating a DNA fragment amplified from plasmid pFA6a-3HA-kanMX6 (from J. R. Pringle). The MATa LEU2/MATα leu2 diploid strains were obtained by crossing the appropriate MATa LEU2 and MATα leu2 haploids strains derived from W303-1a and BAM1-11A; when required, the mating type was switched by using plasmid pGAL-HO (from L. H. Johnston) and the tTA-LEU2 gene was integrated at the LEU2 locus by using plasmid pCM87 (from E. Herrero). Yeast cells were grown on standard yeast peptone dextrose (YPD) extract or on synthetic dextrose minimal media supplemented as required. For growth assays, 10-fold serial dilutions in growth medium were prepared from exponentially growing cultures (usually 2–8 × 106 cells/ml) of the different strains, 5 μl of each dilution were then spotted onto the appropriate medium, and plates were incubated at the indicated temperature. Where indicated, 5 mm caffeine, 200 mm hydroxyurea, 2–10 ng/ml rapamycin (depending on the strain background), 0.003 or 0.01% SDS, 3 μg/ml calcofluor white, 40 or 50 mm latrunculin B, or 1 m sorbitol were added to the growth medium.
Zymolyase digestion sensitivity assays:
Cells from exponentially growing cultures were resuspended at an OD600 of 1 in water containing 0.1% β-mercaptoethanol and 5 mg/ml of zymolyase 20T. Samples were incubated at 30° and the OD600 was determined at 0, 5, 10, 15, 20, 25, 35, and 40 min.
Analysis of chromosome III loss and mitotic recombination:
The rates of chromosome III loss and mitotic recombination were determined by using heterozygous MATa LEU2/MATα leu2 diploid strains essentially as described in Gerring et al. (1990) and Sugimoto et al. (1995). A total of 1 × 106 cells of each diploid strain to be analyzed were incubated with 1 × 106 cells of a MATa leu2 haploid tester in a final volume of 100 μl of YPD at 30° for 6 hr. After the incubation, the mating mixtures were plated on medium selecting for prototrophic triploids. Wild-type, swi4ts, and clb5 MATα haploid strains were used as a control to determine mating efficiency, which also revealed that no mating deficiency associates with the swi4ts and the clb5 mutations at 30°. The analysis of leucine prototrophy in the triploids enabled us to distinguish diploid cells that had mated with the tester due to MATα homozygosis by mitotic recombination (LEU2) from those that had mated because of missegregation of chromosome III (leu2). The rate of mitotic recombination was obtained by calculating the total number of LEU2 triploids, whereas the rate of chromosome loss was obtained by calculating the total number of leu2 triploids. The assay was performed at least twice using independent colonies.
Indirect immunofluorescence, fluorescence-activated cell sorter (FACS) analysis, Western blot analysis, immunoprecipitation, and kinase activity assay were carried out as described previously (Queralt and Igual 2003, 2004). Antibodies used were anti-HA 12C5A antibody (Roche) for Western detection of Clb2-HA, anti-HA 3F10 antibody (Roche) for immunoprecipitation of Clb2-HA, phospho p44/42 MAPK antibody (Cell Signaling Technology) for Western detection of phosphorylated Slt2, and anti-α-tubulin antibody (Serotec, Oxford, UK) for staining of the spindle.
The rsf12 mutation is synthetic lethal with swi4Δ:
The crucial role played by SBF in cell cycle initiation led us to set up a synthetic lethal screen for genes functionally interacting with SWI4 (Igual et al. 1997). We isolated conditional lethal mutants unable to grow at 37° in a swi4Δ background. One of these mutations, rsf12, is described here.
Backcross of rsf12 swi4Δ to the parental swi4Δ strain indicated that the thermosensitive phenotype was due to a single recessive mutation. The rsf12 swi4Δ strain was also crossed to the wild-type strain W303-1A. The analysis of the dissected tetrads demonstrated a clear synthetic interaction between rsf12 and swi4Δ: whereas in the swi4Δ spore clones the conditional phenotype segregates 2:2, in the presence of wild-type SWI4 only 2 of 53 segregants retained their temperature sensitivity. Thus, the conditional lethal phenotype is apparent only in a swi4Δ background.
rsf12 is allelic to CLB5:
To identify the RSF12 gene, a centromeric-based genomic library was screened for plasmids that suppress the thermosensitive growth of the rsf12 swi4Δ mutant. Two clones were isolated. The gene encoding the suppressor activity was identified as the CLB5 gene. To explore this further, the rsf12 swi4Δ mutant strain was transformed with a centromeric plasmid containing the CLB5 gene. As shown in Figure 1, the plasmid rescued the conditional lethal phenotype of rsf12 swi4Δ.
To prove the allelism of rsf12 with CLB5, the rsf12 swi4Δ strain was crossed with a clb5∷URA3 mutant strain. Tetrad analysis showed that the thermosensitivity always segregated in opposition to URA3. Thus, the 43 swi4 ura3 spores, which did not contain the URA3 disrupted clb5 gene, were all temperature sensitive and must have contained rsf12. This confirms the allelism of CLB5 and rsf12, which, henceforth, will be referred to as clb5-12. The data from this cross also confirmed the synthetic lethality of the swi4 and clb5 mutations, since no clb5Δ swi4Δ segregant was recovered in 72 dissected tetrads. It was possible to infer the clb5Δ swi4Δ genotype in 44 nonviable spores. Hence, the deletion of both swi4 and clb5 results in a synthetic lethality.
SBF, but not MBF, is essential in the absence of Clb5:
Swi4, along with Swi6, forms the heterodimeric transcription factor SBF. To determine whether the SWI4 interaction with CLB5 is dependent on the whole SBF, a genetic cross between clb5-12 swi4Δ and swi6Δ strains was carried out. Whereas 51 of 52 of the SWI4 SWI6 segregants grew at 37°, the thermosensitivity segregated 2:2 in the presence of either swi4Δ (15 of 35) or swi6Δ (18 of 34) spore clones. These results confirm that swi6Δ is synthetic lethal with clb5-12, just as swi4Δ is. This is consistent with a functional SBF being essential in cells bearing the clb5-12 mutation.
In addition to SBF, another transcription factor, MBF (composed of Mbp1 and Swi6), is involved in gene regulation at the G1/S transition. Mbp1 is related to Swi4 and SBF and MBF show functional overlap. However, the overexpression of MBP1 does not rescue the thermosensitivity of the clb5-12 swi4Δ mutant (data not shown). Moreover, the genetic analysis of a cross between clb5-12 swi4Δ and mbp1Δ strains did not reveal any interaction between mbp1Δ and clb5-12: whereas the thermosensitivity segregated 2:2 in the swi4Δ spore clones, all the mbp1Δ spore clones grew readily at 37°. These observations indicate that specifically SBF, and not MBF, is essential in the absence of Clb5.
SBF regulates the transcription of the G1 cyclin genes CLN1 and CLN2; therefore it was possible that the phenotype of clb5-12 swi4Δ could be due to a defect in Cln function. However, multicopy plasmids containing the CLN1 or CLN2 gene, the ectopic expression of CLN2 from the mild Schizosaccharomyces pombe adh1 promoter or the strong tetO2 promoter, or the dominant allele CLN2-4 did not rescue the clb5-12 swi4Δ defect (Figure 1 and data not shown). This result indicates that the phenotype of the clb5-12 swi4Δ mutation is not caused by a defect in CLN activity and probably is related to a defective expression of some other genes regulated by SBF.
SWI4 genetically interacts with CLB5 but not with the other CLB genes:
We wondered whether the genetic interaction with SWI4 is a specific property of the CLB5 gene or whether it is shared with the other CLB cyclin genes. To answer this question, the viability of the different clb swi4 double-mutant strains was analyzed. Viable clb1Δ swi4Δ, clb3Δ swi4Δ, clb4Δ swi4Δ, and even clb3Δ clb4Δ swi4Δ mutant spores were obtained from crosses of swi4Δ with clb1Δ or clb3Δ clb4Δ strains. In addition, the CLB1-6 genes were deleted in a strain carrying a thermosensitive allele of SWI4. Figure 2 shows that the clb1Δ swi4ts, clb2Δ swi4ts, clb3Δ swi4ts, clb4Δ swi4ts, and the clb6Δ swi4ts double mutants, in contrast to clb5Δ swi4ts, grew properly at 37°. In conclusion, the synthetic lethality between swi4 and clb5 is not reproduced with the mutations of the other CLB genes, including the partially redundant CLB6 gene, and therefore it must reflect a specific characteristic of the Clb5 cyclin.
clb5-12 swi4Δ cells arrest at the G2/M phase at the restrictive temperature:
In a first attempt to understand the functional interaction between SBF and CLB5, the terminal phenotype of the clb5-12 swi4Δ mutant strain at the restrictive temperature was analyzed. When exponentially growing cells of the clb5-12 swi4Δ double mutant were shifted from 28° to 37°, a cell cycle arrest occurred with little increase in cell number (data not shown). Morphological observation revealed a high proportion (90%) of budded cells, mainly “dumbbell” cells (cells with a bud of a size similar to that of the mother cell) (Figure 3). FACS analysis indicated that most of the cells accumulate mainly with a 2C DNA content. Chromosomes were not segregated in the large-budded cells, which presented a single nucleus and a short spindle located near or at the bud neck in most of the cases. In summary, the clb5-12 swi4Δ mutant manifested a defect in cell cycle progression with a major arrest at the G2/M phase.
We next investigated whether the arrest at G2/M is specific for the clb5-12 allele. The terminal phenotype of clb5Δ swi4ts cells at the restrictive temperature of 37° indicated a blockage in cell cycle progression with the accumulation of cells with a large bud (85%), a 2C DNA content, and a single nucleus migrated to the neck (data not shown). Thus, as occurred with the clb5-12 swi4Δ strain, the clb5Δ swi4ts cells were arrested at the G2/M phase, suggesting that this is a general phenotypic trait of any clb5 swi4 double-mutant strain.
The G2/M transition is governed by the mitotic cyclins, mainly by Clb2. However, the level of the Clb2p protein or its associated kinase activity is not importantly affected in the clb5-12 swi4Δ strain when compared to the parental swi4Δ strain (Figure 3, C and D). Thus, the arrest at G2/M is not due to a defect in the mitotic CDK activity and must be caused by another mechanism. An alternative explanation of the G2/M arrest could imply the activation of a checkpoint mechanism. Both the DNA-damage and the morphogenesis checkpoints have been described to affect G2/M phase progression. To investigate if any of these mechanisms is responsible for the terminal phenotype of the clb5-12 swi4Δ mutant strain, the SWE1 gene involved in the morphogenesis checkpoint (reviewed in Lew 2003), or the RAD9 and RAD24 genes involved in the DNA-damage checkpoint (reviewed in Lowndes and Murguia 2000) were inactivated in the clb5-12 swi4Δ strain and the terminal phenotype at the restrictive temperature was analyzed. In the case of the clb5-12 swi4Δ swe1Δ triple mutant, cells arrest at G2/M with the same terminal phenotype as that of the clb5-12 swi4Δ mutant strain, i.e., 90% of budded cells, mainly dumbbell cells, with replicated DNA and a single nucleus migrated to the neck (Figure 4 and data not shown). The same result was observed in a clb5Δ swi4ts swe1Δ strain (data not shown). This result demonstrates that the G2/M arrest was not caused by the activation of the morphogenesis checkpoint (this is consistent with the presence of Cdc28-Clb2 kinase activity since the morphogenesis checkpoint acts by inactivating the CDK). On the other hand, when the RAD9 and the RAD24 genes were disrupted, clb5-12 swi4Δ rad9Δ and clb5-12 swi4Δ rad9Δ rad24Δ cells remain unviable; yet remarkably, they are distributed through all the cell cycle phases, which indicates that the G2/M arrest was overridden (Figure 4). The same result was observed in a clb5Δ swi4ts rad9Δ mutant strain. This result demonstrates that the blockage of cell cycle progression at G2/M in the clb5 swi4 mutant cells is caused by the activation of the DNA-damage checkpoint.
clb5 swi4 mutant cells have defects in the maintenance of DNA integrity:
The previous results suggest that the clb5 swi4 cells have accomplished the bulk of DNA replication but they have DNA lesions, which activate the checkpoint response. To directly test the presence of DNA damage, the rate of chromosome loss and mitotic recombination were determined in wild type, swi4ts, clb5Δ, and clb5Δ swi4ts strains. It was known that cells with defects in DNA metabolism manifest elevated frequencies of both chromosome loss and mitotic recombination, whereas cells with defects in chromosome segregation show only an elevated chromosome loss. The rates of chromosome loss and mitotic recombination obtained for the wild-type strain (Table 2) were similar to that previously described in similar experiments (Hartwell and Smith 1985; Gerring et al. 1990). Strikingly, the clb5Δ swi4ts mutant strain showed a dramatic increase in the frequency of both chromosome loss (387 times higher) and mitotic recombination (252 times higher). This strongly suggests that the inactivation of CLB5 and SWI4 causes defects in DNA integrity. It is noteworthy to point out that neither chromosome loss nor mitotic recombination rates are importantly affected in the swi4ts or clb5Δ single mutant strains. This is consistent with the synthetic lethality between the swi4 and clb5 mutations described above.
Next, the requirement of a functional DNA-damage checkpoint for viability in clb5 and swi4 mutant strains was investigated. Disruption of the RAD9 gene in two different swi4 mutant strains (BY604 and BAM1-11a) demonstrates that swi4 rad9 double-mutant cells were viable. When the rad9Δ mutation was introduced in a clb5Δ mutant strain, the clb5Δ rad9Δ double-mutant cells were unable to grow at 37°, a temperature at which the rad9Δ and clb5Δ single mutants grow readily (Figure 5A). Thus, in the absence of Clb5, a functional DNA-damage checkpoint is required for viability at high temperature, most probably due to the generation of some kind of DNA lesion by the lack of Clb5.
In addition, the sensitivity of the clb5Δ and swi4Δ mutant strains to caffeine was tested. Caffeine is a potent inhibitor of ATM and ATR kinases, and in fact it abrogates the DNA integrity checkpoint responses in mammalian cells (Kaufmann et al. 2003), fission yeast (Wang et al. 1999; Moser et al. 2000), and budding yeast (Vaze et al. 2002); in addition, caffeine alters the timing of DNA replication origin firing in Xenopus (Marheineke and Hyrien 2004; Shechter et al. 2004). As shown in Figure 5B, the inactivation of either SWI4 or CLB5 results in hypersensitivity to caffeine. This result is consistent with the presence of defects in DNA integrity in cells harboring the clb5 or the swi4 mutations.
The sensitivity of clb5Δ and swi4Δ mutant strains to hydroxyurea was also tested. Hydroxyurea blocks DNA replication, and mutations in genes involved in DNA metabolism cause hypersensitivity to this drug (Sugimoto et al. 1995). As shown in Figure 5B, neither the clb5 nor the swi4 strains were able to grow properly in the presence of 200 mm hydroxyurea, as opposed to that observed for the wild-type strain. This result reinforces the conclusion that Swi4 and Clb5 take part in DNA metabolism.
Genetic interactions between CLB5 and the PKC pathway:
A 2μ-plasmid yeast genomic library was searched for genes able to suppress the lethality of clb5-12swi4Δ. One of the suppressors was identified as the ROT1 gene. A mutation in the ROT1 gene was isolated as a suppressor of a tor2 mutant strain (Bickle et al. 1998). The Tor2 protein kinase is involved, among other cellular processes, in actin cytoskeleton polarization through the regulation of the PKC pathway (the TOR, target of rapamycin, signaling pathway is reviewed in Crespo and Hall 2002). The Rot1 molecular function is unknown, but the rot1 mutant manifests cell wall defects (Bickle et al. 1998; Machi et al. 2004). These observations led us to investigate whether the overexpression of genes from the PKC pathway would suppress the lethality of the clb5-12 swi4Δ mutant strain. Plasmids containing the TOR2, RHO1, PKC1, or SLT2 genes were able to rescue growth at the restrictive temperature of the clb5-12 swi4Δ cells (Figure 6). It was already known that an interaction exists between the SBF transcription factor and the PKC pathway (Igual et al. 1996; Madden et al. 1997; Baetz et al. 2001), but our results also point to a functional connection between the Clb5 cyclin and the PKC pathway. To investigate this further, a synthetic interaction assay between mutations in the PKC pathway and CLB5 was carried out. First, the CLB5 gene was inactivated in the pkc1-8 (JC6-3a) and slt2Δ (DL454) strains. When growth at high temperature was analyzed, neither the clb5Δ pkc1-8 nor the clb5Δ slt2Δ double mutants were able to grow, in contrast to that observed in the clb5Δ, pkc1-8 and slt2Δ single-mutant strains (Figure 7A). The defective growth was rescued by the reintroduction of a functional copy of the CLB5 or the PKC1 or SLT2 genes. These results indicate a synthetic interaction between clb5 and pkc1 or slt2 mutations and suggest that Clb5 and the PKC pathway share some common essential functions. The interaction between Clb5 and the PKC pathway is a specific characteristic of this cyclin, since all the clbΔ slt2Δ and clbΔ pkc1-8 double-mutant cells, other than clb5Δ slt2Δ and clb5Δ pkc1-8, were viable at 37° (data not shown).
To test the interaction with Tor kinases, sensitivity to rapamycin (a inhibitor of Tor function) was analyzed. The growth of the clb5Δ mutant strain, as well as that of the swi4Δ, pkc1Δ, and slt2Δ mutant strains, was significantly impaired in a rapamycin-supplemented medium (Figure 7B). Hypersensitivity of clb5 and slt2 mutant strains to rapamycin has been previously identified in a genome-wide analysis (Chan et al. 2000). Thus, CLB5, SWI4, and the PKC pathway are crucial for proper growth in the absence of Tor function. This result reinforces the conclusion that Clb5, the SBF transcription factor, and the PKC pathway are functionally connected.
The slt2 and pkc1 mutant strains manifest defects in DNA integrity:
The presence of DNA damage in the clb5-12 swi4Δ mutant and the suppression of its conditional lethality by the overexpression of genes from the PKC pathway suggest the possibility that this pathway is somehow involved in DNA metabolism. In a first attempt to investigate this possibility, the RAD9 gene was deleted in the slt2Δ (SEY6211slt2) and pkc1-8 (JC6-3a) mutant strains. Interestingly, the inactivation of the RAD9 gene results in lethality of the slt2Δ and pkc1-8 mutants at high temperature (Figure 8A), indicating that a functional DNA-damage checkpoint is necessary for the viability of slt2 and pkc1 mutants at high temperature. Furthermore, the pkc1Δ (GPY115) and the slt2Δ (DL454) mutant cells could not grow in the presence of 200 mm hydroxyurea, a concentration that does not impede the growth of the isogenic wild-type cells (Figure 8B). Moreover, the blockage of DNA replication by the addition of hydroxyurea induces the activation of the PKC pathway as deduced from the dramatic increase in the phosphorylation state of the Slt2 MAP kinase (Figure 8C and Figure 10C). It has been reported that the Slt2 kinase is activated during periods of polarized cell growth (Zarzov et al. 1996), so it was possible that the hydroxyurea-induced activation could be an indirect effect due to the cell cycle regulation of Slt2 activity. However, there is no accumulation of cells at the early stage of bud formation during the course of the experiment (Figure 8C). Furthermore, mutant cells defective in polarized cell growth like the cdc42-1 cells, which accumulate at the restrictive temperature as unbudded cells but without blocking DNA replication or chromosome segregation (Adams et al. 1990 and data not shown), are able to activate Slt2 in response to hydroxyurea. These results indicate that activation of Slt2 is a direct response to the presence of hydroxyurea and confirm that the PKC pathway is involved in the maintenance of DNA integrity.
The clb5 mutant strain manifests defects in cell integrity:
The major function of the PKC pathway is the maintenance of cell integrity. We therefore wondered whether the genetic interaction between CLB5 and the PKC pathway described above could also reflect a role for Clb5 in cell morphogenesis. In support of this possibility, we have observed that the clb5-12 swi4Δ rad9Δ and clb5-12 swi4Δ rad9Δ rad24Δ mutant cells show abundant cell lysis when they override the G2/M arrest. This result suggests the presence of morphogenetic defects in the clb5-12 swi4Δ mutant cells, in addition to the defects in DNA integrity described above. To get a proof of the involvement of Clb5 in morphogenetic processes, we tested the sensitivity of clb5Δ mutant cells to compounds that affect the cell wall, such as SDS and calcofluor white (CFW), and to the actin-polymerization inhibitor latrunculin B. Strains with defects in the maintenance of cell integrity, including mutant strains in the PKC pathway and the swi4 mutant strain, are hypersensitive to these compounds. Similarly, growth of the clb5Δ mutant strain was severely impaired in the presence of SDS, CFW, or latrunculin B (Figure 9A). In addition, as occurs in slt2Δ mutant cells, the cell wall of the clb5Δ mutant cells was more sensitive to digestion with zymolyase than the cell wall of wild-type cells (Figure 9B). Interestingly, the clb6 and the clb2 mutant strain does not manifest a growth defect in the presence of SDS and latrunculin B (Figure 9C), suggesting that the hypersensitivity of clb5 cells to these compounds is not an indirect effect due to defects in DNA replication or hyperpolarized growth. All these results strongly suggest that Clb5 is involved in morphogenesis, controlling the cell wall metabolism and the maintenance of cell integrity.
The surprising observation that both the Clb5 cyclin and the PKC pathway are involved in such diverse processes as DNA metabolism and the maintenance of cell integrity led us to assay whether a direct regulation among these pathways could exist. As shown in Figure 10A, the level of Clb5 protein is not affected by the inactivation of the PKC pathway or the TOR kinases. Reciprocally, activation of Slt2 by heat shock, SDS, caffeine, or hydroxyurea is not affected by the inactivation of Clb5 (Figure 10, B and C). These results indicate that there is not a direct connection between the PKC pathway and the Clb5 activities, which otherwise is consistent with the synthetic lethality between clb5 and pkc1 or slt2 mutations described above.
The characterization of the clb5-12 swi4Δ mutant has permitted the identification of a functional interaction between the SBF (Swi4-Swi6) transcription factor and the Clb5 cyclin. Clb5 is the main cyclin involved in the control of the initiation of DNA replication (Epstein and Cross 1992; Schwob and Nasmyth 1993; Schwob et al. 1994; Donaldson et al. 1998; Wilmes et al. 2004). Furthermore, Clb5 is involved in the correct positioning of SPB (Segal et al. 2000) in the control of mitotic exit (Shirayama et al. 1999) and it participates in the metaphase checkpoint (Meyn and Holloway 2000). With regards to SBF, it was identified as a transcription factor involved in Start control through the regulation of the periodic expression at the G1/S transition of many genes, among them the CLN1 and CLN2 cyclin genes (Nasmyth and Dirick 1991; Ogas et al. 1991) and many genes involved in morphogenesis (Igual et al. 1996; Iyer et al. 2001). Our results reflect that a connection exists between Clb5 and SBF. It is known to date that SBF indirectly affects Clb5 activity since it regulates the synthesis of the Cln1 and Cln2 cyclins, which in turn induce the degradation of Sic1, an inhibitor of the Clb-Cdc28 kinases (Schneider et al. 1996; Tyers 1996). Nonetheless, the overexpression or ectopic expression of CLN1 and CLN2 was not capable of recovering the lethality of the clb5-12 swi4Δ or the clb5Δ swi4ts mutants, which suggests that the synthetic interaction between swi4 or swi6 and clb5 is related to a defect in functions other than the Cln cyclins.
The terminal phenotype of the clb5 swi4 mutant at the restrictive temperature is characterized by a cell cycle arrest at G2/M as a consequence of the activation of the DNA-damage checkpoint mechanism. We have observed an increase in the rates of chromosome loss and mitotic recombination in the clb5 swi4 mutant strain, which strongly supports the occurrence of DNA damage. Interestingly, the analysis revealed that this is a specific phenotype of the coexistence of both the clb5 and the swi4 mutation. This, along with the synthetic interaction detected between clb5 and swi4 or swi6 mutations, reinforces the hypothesis that Clb5 and SBF perform a common essential function related to DNA metabolism. Other results like the hypersensitivity of swi4 and clb5 cells to hydroxyurea or caffeine or the conditional synthetic lethality between the clb5 and rad9 mutations support a role for SBF and Clb5 in the maintenance of DNA integrity. How can Clb5 and SBF affect DNA integrity? It has been suggested that Clb5 (Meyn and Holloway 2000) and Swi6 (Sidorova and Breeden 1997) take part in the arrest of cell cycle progression in response to DNA damage. A defective checkpoint mechanism could certainly lead to lesions in the DNA. However, we observed a functional DNA-damage checkpoint in the clb5 swi4 mutant strain as deduced from the Rad9, Rad24-dependent arrest of cell cycle progression. Another possibility is that Clb5 and SBF might play a role in DNA repair; this cannot be ruled out, yet to our knowledge, no data support this possibility. On the other hand, the hypersensitivity of the clb5 and swi4 mutant strains to hydroxyurea, which blocks the advance of the replication forks, and to caffeine, which at least in Xenopus affects the control of the DNA replication origins firing (Marheineke and Hyrien 2004; Shechter et al. 2004), suggests that the loss of Clb5 and SBF function affects DNA replication. It has been described that defects in the timing of DNA replication origin firing give rise to genomic instability (Lengronne and Schwob 2002; Sidorova and Breeden 2002; Tanaka and Diffley 2002). In addition, some mutant strains in Orc or Cdc6 proteins experience some level of DNA damage, possibly resulting from aberrant replication fork initiation (Watanabe et al. 2002; Wilmes et al. 2004). Clb5 is crucial for the correct timing and duration of DNA replication (Epstein and Cross 1992; Schwob and Nasmyth 1993; Donaldson et al. 1998) and, at the same time, for preventing reinitiation specifically at origins that have already been fired (Wilmes et al. 2004), so it is possible that dysfunctions in the control of DNA replication origins due to the absence of Clb5 might affect DNA integrity. In fact, O. M. Aparicio's group has recently suggested the presence of DNA damage in clb5 mutant cells, probably as a consequence of a replication stress caused by a reduced replication origin usage (Gibson et al. 2004). With regard to SBF, the overexpression of a deregulated truncated form of Swi4 alters the regulation of origin firing (Sidorova and Breeden 2002). It is accepted that SBF activates the expression of genes involved in budding and cell wall biosynthesis, whereas genes involved in DNA repair and replication are regulated by MBF. Nonetheless, genomic studies (Iyer et al. 2001; Simon et al. 2001; Horak and Snyder 2002) have suggested that SBF may also participate in regulating the expression of some genes involved in DNA metabolism and repair (such as RNR1), genes that encode chromatin proteins (such as those coding for H1, H2A, and H2B histones), or the CDC6 gene involved in the formation of competent prereplication complexes at origins. The latter is particularly interesting since it might suggest that Clb5 and Swi4 collaborate in the regulation of replication origins, a control necessary to avoid DNA damage.
We have shown that the overexpression of different genes of the PKC pathway suppresses the lethality of the clb5-12 swi4Δ mutant and that the clb5 mutation and mutations in genes of the PKC pathway show a conditional synthetic lethality. Given the defects in DNA integrity of the clb5 swi4 mutant and the known role of Clb5 in DNA replication, the detected genetic interactions could suggest that the PKC pathway participates in DNA metabolism. The same hypothesis had been suggested by Huang and Symington (1994) on the basis of the elevated rate of mitotic recombination observed in a pkc1 mutant strain. In support of a function of the PKC pathway in DNA metabolism, we have described a synthetic interaction between pkc1 and slt2 mutations with the rad9 mutation and the hypersensitivity of the pkc1 and slt2 mutant strains to hydroxyurea. In addition, the hypersensitivity of PKC pathway mutants to caffeine (Costigan et al. 1994) may be reevaluated in light of the recently identified effect that caffeine has on both the DNA integrity checkpoint and the regulation of DNA replication origins (Moser et al. 2000; Vaze et al. 2002; Kaufmann et al. 2003; Marheineke and Hyrien 2004; Shechter et al. 2004). The question of how the PKC pathway could affect DNA metabolism arises. We have observed that the Slt2 MAP kinase is activated by the addition of hydroxyurea. This could suggest that activation of the PKC pathway is involved in the checkpoint response to defects in DNA integrity. Recently published results could give clues about the nature of possible PKC-pathway-dependent mechanisms. It has been described that DNA damage leads to an increase in dNTP levels and that this increase is important to cells to survive the damage (Chabes et al. 2003). Considering that the PKC pathway is related to nucleotide biosynthesis through the regulation of the CTP synthetase (Yang et al. 1996), it would be interesting to investigate whether it could be involved in the control of dNTPs pools in response to DNA damage. On the other hand, the mammalian PKCδ is responsible for constitutive and DNA-damage-induced phosphorylation of hRad9, a protein that is involved in checkpoint signaling (Yoshida et al. 2003). It is tempting to speculate that the PKC pathway might regulate Ddc1, the S. cerevisiae homolog to hRad9, or other proteins involved in the checkpoint response.
In addition to the defects in DNA integrity, the characterization of the clb5-12 swi4Δ mutant revealed the presence of morphogenetic defects, as deduced by the cell lysis observed when the DNA-damage-induced arrest is overcome. This suggests that Clb5 could be involved in morphogenesis. In fact, the hypersensitivity of the clb5 mutant strain to SDS, latrunculin B, or to zymolyase strongly supports that Clb5 is involved in the control of actin cytoskeleton and cell integrity. The participation of Clb5 in the control of cell integrity is a new function for this cyclin, which has been previously involved in many processes throughout the cell cycle: DNA replication initiation, the correct positioning of SPB, control of mitosis exit, and the metaphase checkpoint mechanism. It is also important to highlight that the functional interaction with SBF and the PKC pathway is a specific characteristic of the Clb5 cyclin, and it is not observed with other B cyclins, not even with Clb6, which is functionally related to Clb5.
Our work has enabled us to identify a close functional relationship among the SBF transcription factor, the Clb5 cyclin, and the PKC pathway in the metabolism of DNA and in the maintenance of cell integrity. The fact that they are involved in such diverse processes is apparently surprising, and it raises an intriguing question: Is there any relationship between actin cytoskeleton and DNA integrity? The characterization at a molecular level of the functional connection between the Clb5 cyclin, the SBF transcription factor, and the PKC pathway will help to answer this question.
We especially thank L. H. Johnston for his collaboration in this work. We are also very grateful to M. C Bañó for her help and to M. A. de la Torre, D. Levin, M. Molina, J. R. Pringle, E. Schwob, and Y. Takai for kindly supplying plasmids and strains. This work was supported by grant no. BMC2001-0446-CO2-02 and BFU2004-05763-CO2-02 from the Ministerio de Ciencia y Tecnología of the Spanish Governement and grant no. GRUPOS04/23 from Generalitat Valenciana.
↵1 Present address: Chromosome Segregation Laboratory, Cancer Research UK London Research Institute, 44 Lincoln's Inn Fields, London WC2A 3PX, United Kingdom.
Communicating editor: L. S. Symington
- Received April 29, 2005.
- Accepted August 4, 2005.
- Copyright © 2005 by the Genetics Society of America