|Home | About | Journals | Submit | Contact Us | Français|
The integrity of mitosis is dependent upon strict regulation of microtubule stability and dynamics. Although much information has been accumulated on regulators of the microtubule cytoskeleton, our knowledge of the specific pathways involved is still limited. Here we designed genetic screens to identify regulators of microtubule stability that are dispensable in the wild type yet become essential under microtubule-disrupting conditions. We found that the transcriptional cofactor Swi6p and activator Swi4p, as well as the G2/M-specific cyclin Clb2p, are required in a microtubule-destabilizing environment. Swi6p and Swi4p can combine as a transcriptional complex, called the SBF complex (SBF for Swi4/6 cell cycle box [SCB]-binding factor) that is functionally homologous to the metazoan DP1/2-E2F complex and that controls the G1/S transition through the genes it regulates. We show that Swi6p's contribution to microtubule stability can be either dependent or independent of the SBF complex. The SBF-dependent pathway requires downregulation of SBF complex levels and may thereby reroute the transcriptional program in favor of greater microtubule stability. This pathway can be triggered by overexpression of Fcp1p, a phosphatase in the general transcription machinery, or by expression of an allele of SWI6 that is associated with reduced transcription from SBF-controlled promoters. The SBF-independent pathway is activated by a constitutively nuclear allele of Swi6p. Our results introduce novel roles in microtubule stability for genes whose participation in the process may be masked under normal conditions yet nonetheless acquire a dominant role when microtubule stability is compromised.
The fidelity of chromosome segregation in mitosis is essential for the accurate propagation of genetic information to daughter cells, and it is dependent upon timely, coordinated changes in the microtubule (MT) cytoskeleton. In Saccharomyces cerevisiae, chromosome segregation is achieved in four microtubule-dependent steps. The first step is spindle assembly, which involves the migration of duplicated spindle pole bodies (SPBs) to form a bipolar spindle. This process requires the plus-end-directed activity of the kinesin-like motors of the BimC family, Cin8p and Kip1p, that cross-link and slide antiparallel microtubules of the spindle midzone at the G2/M transition (27, 47). The second step is orientation of the mitotic spindle at the site of cytokinesis, which is mediated by cytoplasmic MTs and requires their transient, dynamic interactions with the cell cortex through dynein-dependent (51) and dynein-independent (50, 4) pathways. The third step is chromosome movement along kinetochore MTs through their depolymerization in anaphase A (46), and the fourth step is complete chromosome separation in the process of anaphase B spindle elongation. This last process is powered by Cin8p and Kip1p that cross-link and push apart polar MTs, by polymerization of the same MTs in the midzone driven by a poleward flux of tubulin subunits, and by a pulling force generated by dynein on cytoplasmic MTs (49).
These changes in the MT architecture are dependent on tight regulation on different levels. Most directly, motor proteins and MT-associated proteins (MAPs) influence MT polymerization, stability, and dynamics (31, 18, 7, 29, 11), thereby affecting processes the specific MT state facilitates. For example, the temperature-sensitive double motor cin8-3 kip1Δ mutant is sensitive to various MT-destabilizing drugs. At elevated temperatures, spindle assembly at G2/M is compromised (20, 42, 49). Introduction of the tub2-402 allele, which hyperstabilizes MTs (28), suppressed the sensitivities both to elevated temperatures and to MT-destabilizing drugs that are associated with this double mutant (39), suggesting destabilized MTs in this background.
Another level of regulation is by upstream cyclin/Cdk complexes whose periodic expression drives specific cell cycle events. Late in G1 phase, the Cln3p-Cdc28p protein kinase complex activates two transcription complexes, the MBF complex (MBF for MluI cell cycle box [MCB]-binding factor) and the SBF complex (SBF for Swi4/6 cell cycle box [SCB]-binding factor), and these in turn promote the transcription of a number of genes important for budding and DNA synthesis (10, 26). At G2/M, the Clb2p-Cdc28p complex represses the activity of SBF, returning the expression of SBF-regulated genes to low levels (1). MT regulation by cyclin/cdk complexes may manifest indirectly: Clb2p-Cdc28p, for example, contributes to the stability or localization of motors posttranslationally (8, 12). It is therefore not surprising that certain Clb/Cdk mutants share the same phenotype exhibited by cin8-3 kip1Δ double mutants grown at the elevated temperature, being unable to assemble a bipolar spindle due to a failure to segregate duplicated SPBs (20, 42, 49). These reports reinforce the ties that exist between upstream and downstream MT regulators.
In wild-type cells under normal conditions, phenotypes are often not manifested because other proteins or pathways act redundantly. A general example is KSS1, a gene whose deletion in wild-type cells has no phenotypic consequence on the mating pathway and therefore was originally thought not to be involved in the process. It was later shown that it can fill in for another protein, Fus3p, which is functionally redundant with Kss1p, when Fus3p is not present (48). Similarly, in this study we aimed to identify proteins that under normal conditions have no apparent role in MT stability, but when the stability of MTs is compromised, their involvement becomes essential.
We previously conducted a genetic screen to identify proteins that when overexpressed can correct the temperature sensitivity and the MT destabilization phenotype associated with cin8-3 kip1Δ cells (54). Here we describe pathways by which Fcp1p, one of the overexpression suppressors, increases MT stability in this background. The underlying mechanisms were elucidated through the identification of proteins that Fcp1p's effect is dependent upon or that can substitute for it altogether. On the basis of these results, we propose novel pathways that regulate the stability of microtubules.
Saccharomyces cerevisiae strains used in this study are listed in Table 1. All strains are either S288C derivatives or were backcrossed at least seven times to an S288C background. Strain DEY2169 was constructed by crossing strain MAY2169 with strain YNN282 (Yeast Genetic Stock Center, University of California, Berkeley), followed by tetrad dissection and analysis as previously described (32). Medium preparation and yeast genetic and transformation techniques were essentially as previously described (32). Sensitivity to thiabendazole (TBZ) was tested on yeast extract-peptone-dextrose (YPD) agar medium to which the desired amount of TBZ (MP Biomedicals Inc.) was added from 10 mg/ml stock in dimethylformamide (DMF). Serial dilution experiments were repeated three times.
Details of the screen for multicopy suppressors were previously described (54). In brief, after transformation of cin8-3 dyn1Δ cells with the YEp24-based yeast genomic library, colonies that grew at both the permissive temperature (26°C) and the nonpermissive temperature (35°C) upon replica plating were picked. Plasmids were extracted from these colonies, the genomic insert containing the suppressor was sequenced, and the gene responsible for the suppression was isolated. One of these suppressors, the FCP1 gene, was subcloned into the XbaI and SmaI sites of the 2μm plasmid YEplac112 and was used to transform cin8-3 kip1Δ cells.
cin8-3 kip1Δ double mutants carrying the YEplac112-FCP1 plasmid were transformed with a NotI-digested Tn7-derived insertional library (41). Gene disruptions were generated by homologous recombination as the truncated genomic fragments replaced their native chromosomal locus. In our cin8-3 kip1Δ background, such transformations generated triple mutants. Cells containing the transposon were grown at 26°C, replica plated, and incubated at the nonpermissive temperature, 35°C. Since all cells originally grew at the nonpermissive temperature (due to FCP1 overexpression), colonies that ameliorated this effect subsequent to insertional mutagenesis and grew at 26°C but not at 35°C were selected for further analysis.
To identify genes that ameliorate the effect of FCP1 overexpression, we isolated genomic DNA from colonies that lost the ability to grow at the nonpermissive temperature, as described previously (32). Thermal asymmetric interlaced (TAIL) PCR (24) was used to amplify flanking sequences adjacent to the insertions. Degenerate PCR primers were as described previously (24), and specific primary and nested primers were designed on the basis of the inserted Tn7 sequences. Sequence data from both ends of the insert were compared against the yeast genome database to determine the identification of the genomic sequences it contains. DNA from transformant colonies was subjected to Southern blot analysis in order to exclude multiple Tn7 insertions.
Deletion strains from the Saccharomyces Deletion Project were propagated on YPD plates containing 200 μg/ml Geneticin (Gibco, Invitrogen). DNA was isolated from these strains as described above, and the genes containing the KanMX4 cassettes were PCR amplified using the primers shown in Table 2. PCR products containing the deletion cassettes were verified by gel electrophoresis and transformed into haploid cin8-3 kip1Δ double mutants carrying the 2μm plasmid YEplac112-FCP1. DNA was subsequently purified from Geneticin-resistant triple mutants, and a confirmation primer (a few hundred base pairs upstream or downstream of the deletion) together with one of the original primers (Table 2) was used to confirm the correct integration.
Ten-milliliter log-phase cultures were centrifuged, and pellets were resuspended in 1 ml ice-cold Tris-EDTA (TE) buffer. Total protein lysate was prepared by vortexing cells with glass beads in 50 μl yeast extraction buffer (0.6% SDS and 10 mM Tris [pH 7.4] with the addition of 1 μg/ml leupeptin [Sigma], 2 μg/ml aprotinin [Sigma], 1 μg/ml pepstatin [Sigma], and 1 mM phenylmethylsulfonyl fluoride [PMSF] [Sigma]). Sample buffer (50 μl) (125 mM Tris-HCl, 20% glycerol, 4.1% SDS, 4% β-mercaptoethanol) was then added, the solution was spun, and the supernatant was boiled for 5 min. A 5-μl sample was mixed with an equal volume of Laemmli Sample Buffer (Bio-Rad) and loaded on 4 to 12% gradient NuPAGE Novex Bis-Tris minigel (Invitrogen), according to the manufacturer's instructions. EZ-Run prestained protein marker (Fisher Scientific) was loaded as well. Western blotting was performed as described previously (53) using Hybond enhanced chemiluminescence (ECL) nitrocellulose membrane (Amersham Biosciences). Rabbit anti-Swi6 antibody (a gift from L. Breeden) and peroxidase-linked goat anti-rabbit secondary antibodies (GE Healthcare) were both used at a 1:5,000 dilution. Rabbit anti-Swi4 antibody (a gift from B. Andrews) and peroxidase-linked goat anti-rabbit secondary antibodies were used at a dilution of 1:5,000 and 1:10,000, respectively. Mouse anti-3-phosphoglycerate kinase (anti-PGK) antibody (Invitrogen) was used as a control to normalize the results. Peroxidase-linked anti-mouse IgG secondary antibody (GE Healthcare) was used at a 1:5,000 dilution. To visualize the results, blots were developed using Amersham's ECL detection reagents (GE Healthcare) in accordance with the manufacturer's instructions and detected using Fujifilm Imager LAS-4000. Fluorescence quantification and analysis were done using Fujifilm Multi Gauge software. Western blots were repeated three times.
Data were expressed as means ± standard errors of the means (SEM). Data were analyzed using analysis of variance coupled with Holm-Sidak test for multiple pairwise comparison. Data analysis was performed using SigmaStat (Chicago, IL). P values of <0.05 were considered statistically significant.
FCP1 on an overexpression vector was identified in our original screen (54) as a strong suppressor of the temperature and thiabendazole (TBZ) sensitivities of cin8-3 kip1Δ mutants. It restored growth at the restrictive temperature (35°C) and conferred resistance to the MT-destabilizing drug TBZ to levels exceeding that of the isogenic wild-type strain (Fig. 1A and B). cin8-3 kip1Δ cells are also sensitive to the MT-destabilizing drug benomyl and its derivative methyl-benzimidazole-2-yl carbendazim (MBC) (39) (data not shown). FCP1 overexpression rendered both double mutant and wild-type cells more resistant to these latter drugs (not shown). TBZ, however, was the preferred drug for this study, as it showed pronounced differences in the growth profiles for various mutations in Swi6p, a protein that is the primary focus of this work.
Fcp1p (transcription factor IIF [TFIIF]-stimulated C-terminal domain [CTD] phosphatase) is a phosphatase in the general transcription machinery. It specifically dephosphorylates Ser2 of the tandem heptapeptide repeat Tyr1-Ser2-Pro3-Thr4-Ser5-Pro6-Ser7 in the C-terminal domain of the largest subunit (Rbp1p) of RNA polymerase II (RNAPII) (36). This conserved heptapeptide repeats 26 times in yeast and up to 52 times in metazoan CTDs (13). Upon transcriptional termination, Fcp1p-mediated CTD dephosphorylation is required to recycle the polymerase at the end of each round of transcription, an action that primes it for the next transcription cycle (16).
In order to identify domains, and therefore a specific function of Fcp1p that is necessary for the suppression of the cin8-3 kip1Δ phenotype, we assayed the effects of overexpression of various fcp1 mutant genes on our strain. We separately generated amino acid substitutions of the first and second conserved aspartatic acids of the ΨΨΨDXDX(T/V)ΨΨ motif (where Ψ is any aromatic amino acid), which are central to catalysis (33). In vivo, mutation of either aspartate abrogates the ability to rescue the lethality associated with a chromosomal deletion of FCP1, even though the mutant proteins are expressed similar at levels to that of the wild-type protein (36). We also generated a deletion of the C-terminal region of FCP1 (amino acids 626 to 732), a region that was implicated in binding the RAP74 subunit of TFIIF and TFIIB, both of which regulate Fcp1p's activity (37). This deletion abolishes the protein's interaction with the RAP74 subunit of TFIIF and with TFIIB yet maintains normal CTD phosphatase function in vitro (2).
We found that the integrity of both the phosphatase domain and the C-terminal region was necessary for the suppression of the temperature sensitivity of cin8-3 kip1Δ cells when FCP1 was overexpressed (Fig. 1C). These findings of the requirement of Ser2 dephosphorylation by Fcp1p and the C-terminal region of Fcp1p for suppression were further corroborated when we overexpressed SSU72 in the same strain and failed to observe suppression of the mutant phenotype. SSU72 codes for a phosphatase specific for Ser5 of the same consensus sequence of the CTD of RNAPII and lacks a functionally distinct C-terminal domain (40).
The results of the mutational analysis of Fcp1p led us to conclude that the effect of FCP1 overexpression on MT stability may be indirect, as it requires the normal function of Fcp1p in the context of general transcription. We therefore hypothesized that FCP1 overexpression, through aberrant transcription, triggers a putative pathway that results in greater MT stability. To test this possibility, we designed a secondary screen with a Tn7 insertional library (see Materials and Methods) to search for genes that are needed for the effect of FCP1 overexpression in the cin8-3 kip1Δ background to manifest. Transposon-mediated truncation of SWI6 in the cin8-3 kip1Δ [FCP1-2μm] background restored the sensitivity of the double mutant to the elevated temperature and exacerbated its sensitivity to the microtubule-destabilizing drug TBZ (data not shown). We confirmed these results with a targeted deletion of SWI6 to verify that it is this specific deletion and not other unforeseen elements of the screen that is responsible for ameliorating the effect of FCP1 overexpression on the temperature and TBZ sensitivity of cin8-3 kip1Δ cells (Fig. 2A and B).
Furthermore, to exclude the possibility that a SWI6 deletion independently introduces added temperature sensitivity to the already compromised strain, we introduced to the cin8-3 kip1Δ swi6Δ triple mutant cells wild-type CIN8 on a plasmid. A rescue of the mutant phenotype (Fig. 2A) suggested that the effect of Swi6p on MT stability is pertinent specifically to the cin8-3 kip1Δ genetic background.
Figure 2B also demonstrates that the single swi6 deletion increases the sensitivity of wild-type cells to TBZ to approximately the same degree as that of the cin8-3 kip1Δ double mutant and that FCP1 overexpression does not increase the resistance to TBZ of the swi6 single mutant. Thus, FCP1 overexpression may act through Swi6p to stabilize microtubules, as it does not bypass the need for Swi6p under microtubule-destabilizing conditions.
To investigate the role of Swi6p in our system, we assayed genes whose products physically interact with Swi6p for their ability to phenocopy the effect of a SWI6 deletion. Swi6p is a transcription cofactor that forms heteromeric complexes with the DNA-binding proteins Swi4p and Mbp1p to mediate transcription at the G1/S transition (26). It is the regulatory or trans-activator subunit of these complexes. SBF (SCB-binding factor) is the complex containing Swi4p and Swi6p that enhances transcription of G1 cyclin genes via SCB elements (Swi4/6 cell cycle box; CACGAAA). MBF (MCB-binding factor) consists of Mbp1p and Swi6p. This complex enhances the transcription of S-phase cyclin genes as well as genes involved in DNA synthesis and repair. It acts via MCB elements (MluI cell cycle box; ACGCGTNA) and leads to the initiation of a complex transcriptional cascade required for coordinated cell cycle progression (10).
Because of the involvement of Swi6p in both these complexes, we specifically looked into the potential implication of these complexes in the suppression. To do this, we first generated cin8-3 kip1Δ swi4Δ and cin8-3 kip1Δ mbp1Δ triple mutants to see whether these deletions similarly block the effect of FCP1 overexpression on the temperature sensitivity of our strains.
As shown in Fig. 2A, cin8-3 kip1Δ swi4Δ cells, but not cin8-3 kip1Δ mbp1Δ cells, lost the ability to grow at the nonpermissive temperature even in the presence of a multicopy plasmid carrying FCP1. Growth of cin8-3 kip1Δ swi4Δ cells at the elevated temperature was restored only upon introduction of a plasmid-borne wild-type copy of CIN8, suggesting that a swi4 deletion does not independently introduce sensitivity to the strain. These results suggest that the SBF complex is a likely candidate through which MT stability is conferred. Cells in which swi4 alone was deleted or cells in which swi4 was deleted in combination with cin8-3 kip1Δ are not sensitive to TBZ; however, the deletion does exacerbate the sensitivity of the double mutant to other microtubule-destabilizing drugs, such as benomyl and its derivative MBC (not shown). The mechanisms of action of the three drugs employed in our screens are similar—they all work by sequestering tubulin subunits, preventing polymerization (14, 55). Variation in the sensitivities of different strains to different drugs was previously attributed to ATP-binding cassette (ABC) transporters that are known to be responsible for drug resistance in fungi (6, 43, 44).
We next performed Western blot analysis to further investigate whether the SBF complex has a role in the suppression. We found that in wild-type cells, FCP1 overexpression does not significantly alter the levels of Swi4p and Swi6p (Fig. 3). cin8-3 kip1Δ cells, however, have elevated levels of these proteins compared to the wild type. When FCP1 is overexpressed in the double mutant, the levels of the two proteins return to roughly wild-type levels. Furthermore, though Swi4p and Swi6p are present in different amounts in cells, overexpression of FCP1 in the cin8-3 kip1Δ background has the outcome of downregulating the proteins similarly (1.61- and 1.55-fold, respectively) relative to double mutants that do not overexpress FCP1.
These results demonstrate that the levels of Swi6p and Swi4p in the cell are affected when MT stability is compromised, as in the cin8-3 kip1Δ background. This compromised condition may be sensed, resulting in a change in transcription patterns. More importantly, and in line with Fcp1p's role in transcription, the results suggest that in cin8-3 kip1Δ cells, FCP1 overexpression leads to a reduction in the levels of the SBF complex specifically, as its components Swi4p and Swi6p are downregulated comparably. Taken together with the results shown in Fig. 2, the data reveal a correlation between SBF complex levels and MT stability and present the possibility that curtailing transcription from SBF promoters supports MT stability.
The results in Fig. 2 and and33 that point to the SBF complex as an intermediary through which MT stability is reconciled led us to ask whether alteration of transcription from SBF promoters is needed for the suppression of the TBZ sensitivity of our mutants. To examine this, we obtained two alleles of SWI6, a constitutively nuclear (swi6-S160A) allele, whose expression does not reveal major changes in SBF-driven transcription, and a predominantly cytoplasmic (swi6-S160D) allele, which was shown to result in a marked reduction of transcription from SBF-regulated promoters (52). Protein expression from plasmids carrying these alleles was previously shown to be comparable to protein expression from a plasmid carrying the wild-type SWI6 gene (52).
We expressed these alleles in the compromised cin8-3 kip1Δ swi6Δ [FCP1-2μm] strain. As shown in Fig. 4A, introduction of a wild-type copy of SWI6 on a centromeric plasmid resulted in the predicted complementation of the temperature-sensitive phenotype. Introduction of other alleles of SWI6 on a plasmid also resulted in complementation. When assayed for growth on a medium containing TBZ, cin8-3 kip1Δ swi6Δ [FCP1-2μm] cells carrying wild-type SWI6 or swi6-S160A on a plasmid resulted in resistance to the microtubule-destabilizing drug at a level roughly equivalent to that of cin8-3 kip1Δ double mutants overexpressing FCP1 (compare Fig. 4A with Fig. 1B). Interestingly, cells carrying the Swi6-S160D allele demonstrated even higher resistance to TBZ. This led us to suggest that an additive effect may be in place when both FCP1 overexpression and Swi6-S160D are present in the cin8-3 kip1Δ swi6Δ background, resulting in such high levels of resistance. If this is the case, then the swi6-S160D allele may be able to mediate sufficient suppression in the absence of FCP1 overexpression.
To test this hypothesis, triple mutants were cured of the multicopy plasmid carrying FCP1. Figure 4B reveals that while plasmids carrying the wild-type SWI6 or its S160A version no longer supported growth at TBZ concentrations that compromised the double mutant, swi6-S160D, although less resilient at extremely high TBZ concentrations in the absence of FCP1 overexpression, is nonetheless able to restore growth in the cells to a level similar to that of the double mutant overexpressing FCP1 (compare Fig. 4B with Fig. 1B).
These results suggest that the effects of these swi6 alleles on MT stability is independent of FCP1 overexpression and may substitute for it to achieve the same suppression. Furthermore, because of its reported reduced transcription from SBF promoters (52), the suppression caused by the S160D allele of SWI6 proves that not only is modulation of the levels of the proteins that make up the SBF complex needed (as suggested in Fig. 3) but that a functional SBF complex must be in place to exert its effect on transcription of the G1/S genes it is known to regulate.
In the presence of FCP1 overexpression, all alleles of SWI6 tested have the capacity to complement the temperature-sensitive phenotype (Fig. 4A). However, in the absence of FCP1 overexpression, only the swi6-S160A allele is able to suppress the temperature sensitivity of our mutants (Fig. 4B).
The fact that rescue of the temperature sensitivity in the absence of FCP1 overexpression requires the S160A allele while resistance to TBZ requires the S160D allele of SWI6 is explained as follows: when damage to microtubules is global (i.e., when both nuclear and cytoplasmic MTs are affected, as is the case when cells are grown in the presence of TBZ), downregulating G1/S transcription from SBF-controlled promoters through expression of the swi6-S160D allele can suppress the MT defect. On the other hand, when the damage is limited to spindle MTs, as is the case for cin8-3 kip1Δ cells grown at the elevated temperatures (20, 42, 49), the constitutive presence of Swi6p in the nucleus can cause some suppression. There may be a dose effect in that Swi6p-S160D, which is mostly cytoplasmic, is not able to achieve the same effect at elevated temperatures when the damage is solely to spindle MTs. Since expression of the swi6-S160A allele does not alter transcription from SBF-regulated promoters (52), this suggests that there may be another role for Swi6p in MT stability that is distinct from its role in the SBF complex.
To explore the possibility that regulators of the SBF complex are also involved in FCP1's suppression and therefore relevant to our background, we screened different cyclins for their ability to mitigate the effect of FCP1 overexpression on MT stability in the cin8-3 kip1Δ background. The rationale for looking into cyclins is that Cin8p, Kip1p, Fcp1p, Swi6p, and Swi4p were all previously shown to physically interact with Cdc28p, the catalytic subunit of the main cyclin-dependent kinase (1, 8, 22, 56), which pairs up with different cyclins to regulate cell cycle progression in yeast. Furthermore, strains containing certain cyclin or CDC28 deletions are incapable of assembling a bipolar spindle, a phenotype shared with our double mutant (20).
As the periodic expression of cyclins serves to limit the window of action of Cdc28p to the proper time in the cell cycle (17), we generated triple mutants, strains with cin8-3 kip1Δ mutations and mutations of different cyclins in order to test for a stage-specific dependency of the suppression (Fig. 5A). Deletions of different cyclins were introduced individually into cin8-3 kip1Δ cells carrying FCP1 on a multicopy plasmid. As before, we looked for a deletion that abrogates FCP1's suppression of the temperature and TBZ sensitivities and for complementation by a plasmid carrying wild-type CIN8. CLB2 was the only gene whose deletion satisfied these parameters (Fig. 5). Incidentally, CLB2 codes for a B-type cyclin that activates Cdc28p to promote the transition from G2 to M phase (23), which corresponds to the timing of cin8-3 kip1Δ arrest following G1 synchronization and release at the restrictive temperature (27, 47, 49). We also tested for the ability of the three mutations to antagonize FCP1's suppression (i.e., stabilize MTs when deleted and when FCP1 overexpression is absent), but none of the deletions above restored growth without FCP1 overexpression (not shown).
Surprisingly, deletion of CLN3 or WHI5, both of which are known to regulate the SBF complex at the G1/S transition (15), had no effect on the Fcp1p-mediated rescue of the double mutant phenotype. This corroborates earlier indications that Swi6p has a function in MT stability that is independent of the SBF complex.
This study suggests novel roles in MT stability for proteins that were not previously associated with this function. It opens new avenues of research into the exploration of the specific mechanisms by which upstream cell cycle regulators affect the stability of the microtubule cytoskeleton.
FCP1 is an essential gene in both budding and fission yeast and is conserved among eukarya (2). Domain analysis of Fcp1p revealed that both the phosphatase activity and functional targeting to the transcription machinery are required for the suppression of the temperature sensitivity of cin8-3 kip1Δ cells (Fig. 1C). These data, together with the established role of Fcp1p in global gene transcription, suggest that the effect of FCP1 overexpression on MT stability may not be direct (i.e., mediated through physical interaction with components of the cytoskeleton) but instead may be indirect, possibly involving different pathways that act in conjunction or in parallel. Beyond the role of Fcp1p in transcription, it may be a key player in shaping specific phosphorylation patterns in its substrate, the C-terminal domain of the large subunit of RNAPII, through dephosphorylation of Ser-2 in the heptad repeats of the CTD. The significance of the phosphorylation state of the CTD of RNAPII is iterated in the proposed “CTD code” (5), which was suggested to convey information to CTD-binding proteins, some of which recognize particular phosphorylation patterns and trigger specific cellular responses (19).
How can overexpression of FCP1 suppress defects associated with a compromised microtubule cytoskeleton? Mutations in FCP1 increase phosphorylation of Ser2 residues of the CTD of RNAPII and rapidly shut down mRNA synthesis (30, 37). Conversely, FCP1 overexpression in Schizosaccharomyces pombe results in increased levels of the hypophosphorylated (IIa) form of the CTD relative to control strains (35). By overexpressing FCP1 in our strains, we may be mimicking massive dephosphorylation of the CTD. It was demonstrated in Xenopus extracts that upon fertilization, the CTD undergoes fast and massive dephosphorylation that is attributable to the Xenopus orthologue of FCP1 (45). This dephosphorylation contributes to the preparation of the transcriptional machinery for zygotic genome activation, as it is accompanied by a shift between different gene expression programs for the fertilized oocyte. The transition results in the activation of intracellular signals, some of which regulate changes in the microtubule cytoskeleton, including remodeling of the MT architecture that is required for expulsion of half the chromosomes into the polar body and for the establishment of a cytoplasmic microtubule network whose responsibility is to promote migration of the male and female pronuclei (21). Likewise, FCP1 overexpression in our genetic background may specify a transcriptional program that is accompanied by changes in MT stability.
The transcriptional program specified by FCP1 overexpression that is relevant for MT stability converges with Swi6p, Swi4p, and Clb2p, because deletion of these genes, separately, in the cin8-3 kip1Δ background abrogates the FCP1-mediated suppression of the temperature sensitivity and the sensitivity to MT-destabilizing drugs that are associated with these cells.
The relevance of Clb2p for MT-dependent processes was previously characterized. Cells in which CLB2 has been deleted have a diminished capacity to separate the SPBs (20), a phenotype in common with cin8-3 kip1Δ double mutants grown at the nonpermissive temperature after release from α-factor-induced G1 arrest (27, 47, 49). Additionally, the cyclin/Cdk complex Clb2/Cdc28 was shown to promote SPB separation and spindle assembly via Cin8p and Kip1p. The motors are phosphorylated directly by the Clb2/Cdc28 complex in vitro and in vivo, and this phosphorylation plays a role in promoting SPB separation and spindle assembly (8). In another study, it was proposed that Clb2/Cdc28 kinase activity regulates SPB separation indirectly by preventing the degradation of the motors (12).
Still, there may be paths to MT stability mediated through Clb2p that were not previously considered, paths that become dominant when the MT cytoskeleton is compromised. For example, the Clb2/Cdc28 complex, by virtue of its interaction with and phosphorylation of Swi4p, was shown to switch off SBF-dependent transcription at the G2/M phase (1, 38), a time when cin8-3 kip1Δ cells become arrested at the restrictive temperature (27, 47). In our study, the presence of FCP1 on a multicopy plasmid correlated with a proportional reduction in the levels of Swi6p and Swi4p relative to cells not overexpressing FCP1 (Fig. 3), suggesting that the SBF complex is being downregulated, and expression of the swi6-S160D gene, which leads to reduced activity from SBF promoters (52), was able to substitute for FCP1 overexpression in suppressing the TBZ sensitivity. Because of this similar effect of Clb2p at G2/M, of FCP1 overexpression, and of constitutive expression of swi6-S160D on the SBF complex and on MT-dependent processes, it is not unreasonable to propose that when MT stability is compromised, downregulation of the SBF complex alleviates the impairment. This downregulation, in turn, may lead to rerouting of the transcriptional program in favor of greater MT stability.
Several lines of evidence led us to propose the notion of an SBF-independent route to MT stability. The first is the suppression of the temperature sensitivity of cin8-3 kip1Δ swi6Δ cells by constitutive expression of the S160A allele of SWI6 (Fig. 4B), an allele which does not lead to major changes in cell cycle-specific activation or repression of transcription of SBF-regulated promoters (52). The second is the fact that deletion of regulators of the SBF complex did not achieve the same abrogation of FCP1 suppression, as did a deletion of SWI4 or SWI6 (Fig. 5A).
Swi6p shuttles in and out of the nucleus in a cell cycle-dependent manner, with nuclear import occurring concomitantly with dephosphorylation of Ser-160 in late mitosis and with nuclear export peaking from late G1 to M phase and requiring phosphorylation of Ser-160 (22). Activation or repression of Swi6p-regulated genes is independent of Swi6p's phosphorylation and hence, its localization (52). Aspartate substitution of serine 160 was shown to significantly impair nuclear localization, causing Swi6p to remain predominantly cytoplasmic throughout the cell cycle (though Swi6p is not completely excluded from the nucleus), whereas an alanine substitution at the same position leads to constitutive nuclear accumulation of Swi6p (52).
Interestingly, there is another convergence point between Clb2p and Swi6p which is independent of the SBF complex but dependent on the principal mitotic exit regulator Cdc14p. Clb2/Cdc28 was shown to lead to the liberation of Cdc14p from nucleolar sequestration in early anaphase (3). At anaphase onset, when the spindle starts to elongate, the phosphatase activity of Cdc14p was shown to be required for nuclear microtubule stabilization (25) and for the localization of spindle midzone proteins like Cin8p (34). When CLB2 is deleted, cells are defective for early anaphase release of Cdc14p (3), and cdc14-1 clb2::LEU2 double mutants do not start anaphase and arrest with unseparated spindle pole bodies (57), an arrest phenotype shared with cin8-3 kip1Δ cells cultured at the nonpermissive temperature after release from G1 synchronization (20). Recently, it was shown that CDC14 overexpression can bring about SPB separation in cin8-3 kip1Δ cells (9). In an unrelated study, Cdc14p was shown to be able to trigger nuclear import of Swi6p in vivo and to dephosphorylate Swi6p at serine 160 in vitro (22), a condition mimicked with the swi6-S160A (nonphosphorylatable) nuclear allele we employed.
We therefore propose that in cin8-3 kip1Δ cells, Swi6p may have the potential to mediate the role of Cdc14p in nuclear MT stabilization through a similar mechanism to that by which Cdc14p leads to an increase in nuclear MT stability in wild-type cells (25). Cdc14p normally recruits Swi6p to the nucleus only in late M phase (22), but constitutive nuclear localization of Swi6p (through expression of the S160A allele) may be able to stabilize spindle MTs in the compromised cin8-3 kip1Δ background. This is supported by our finding that the nuclear version of Swi6p restores growth at the nonpermissive temperature independently of FCP1 overexpression (Fig. 5B).
Our results demonstrating that certain SWI6 alleles play a part in MT stability raise the possibility that under normal conditions, transient expression of different phosphorylated forms and their restriction to specific cell cycle phases serves to limit the functional diversity of Swi6p. Constitutive expression of modified forms of Ser-160 may bring out Swi6p's MT-stabilizing potential.
This work was supported by grants from the National Institutes of Health (grant GM065885) and the Professional Staff Congress of the City University of New York.
We thank P. Lipke, S. Krishnamurthy, A. Kumar, L. Breeden, B. Andrews, and C. Forest for donating strains, plasmids, antibodies, primers, and other materials for this work, P. Lipke and R. Tal for critical comments on the manuscript, and R. Tal and E. Korolyev for technical assistance.
Published ahead of print on 28 October 2011.