Cell cycle–dependent association of polo kinase Cdc5 with CENP-A contributes to faithful chromosome segregation in budding yeast

Abstract

Evolutionarily conserved polo-like kinase, Cdc5 (Plk1 in humans), associates with kinetochores during mitosis; however, the role of cell cycle–dependent centromeric (CEN) association of Cdc5 and its substrates that exclusively localize to the kinetochore have not been characterized. Here we report that evolutionarily conserved CEN histone H3 variant, Cse4 (CENP-A in humans), is a substrate of Cdc5, and that the cell cycle–regulated association of Cse4 with Cdc5 is required for cell growth. Cdc5 contributes to Cse4 phosphorylation in vivo and interacts with Cse4 in mitotic cells. Mass spectrometry analysis of in vitro kinase assays showed that Cdc5 phosphorylates nine serine residues clustered within the N-terminus of Cse4. Strains with cse4-9SA exhibit increased errors in chromosome segregation, reduced levels of CEN-associated Mif2 and Mcd1/Scc1 when combined with a deletion of MCM21. Moreover, the loss of Cdc5 from the CEN chromatin contributes to defects in kinetochore integrity and reduction in CEN-associated Cse4. The cell cycle–regulated association of Cdc5 with Cse4 is essential for cell viability as constitutive association of Cdc5 with Cse4 at the kinetochore leads to growth defects. In summary, our results have defined a role for Cdc5-mediated Cse4 phosphorylation in faithful chromosome segregation.

INTRODUCTION

Faithful chromosome segregation is essential for the growth and cellular proliferation of organisms because defects in this process result in aneuploidy, which has been observed in human diseases such as cancer, and developmental disorders (Santaguida and Amon, 2015). A key determinant for high-fidelity chromosome segregation is the kinetochore, which is composed of centromeric (CEN) DNA, associated proteins, and a unique chromatin structure (Verdaasdonk and Bloom, 2011; Burrack and Berman, 2012; Musacchio and Desai, 2017). CENs in budding yeast are composed of ∼125 base pairs of unique DNA sequence (Clarke and Carbon, 1980), whereas CENs in other eukaryotes are several mega–base pairs of DNA representing sequence repeats, species-specific satellite arrays, or retrotransposon-derived sequences (Verdaasdonk and Bloom, 2011). Despite the CEN sequence divergence, the role of CEN in chromosome segregation is evolutionarily conserved (Verdaasdonk and Bloom, 2011; Burrack and Berman, 2012). Moreover, many of the ∼70 kinetochore proteins representing different subcomplexes from budding yeast (Westermann et al., 2003; Cho et al., 2010; Biggins, 2013) are functionally conserved (Musacchio and Desai, 2017). For example, CEN identity in eukaryotic organisms is specified by an epigenetic mark in the form of specialized nucleosomes containing Cse4 (CENP-A in humans, Cid in flies, Cnp1 in fission yeast; Sullivan et al., 1994; Stoler et al., 1995; Meluh et al., 1998; Henikoff et al., 2000; Takahashi et al., 2000). In budding yeast, Cse4 contains two distinct domains. The evolutionarily conserved C-terminus histone fold domain (HFD) carries a centromere targeting domain, which is essential for recruitment and incorporation of Cse4 into the CEN chromatin (Meluh et al., 1998; Keith et al., 1999). The N-terminus of Cse4 (∼129 amino acids) interacts with kinetochore proteins such as the components of the COMA complex (Ctf19, Okp1, Mcm21, and Ame1) and facilitates their recruitment to the CEN (Ortiz et al., 1999). Moreover, the N-terminus of CENP-A also directs the targeting of other kinetochore proteins to the CEN (Van Hooser et al., 2001). In addition, posttranslational modifications of Cse4, namely, phosphorylation, ubiquitination, sumoylation, methylation, and acetylation also regulate faithful chromosome segregation (Hewawasam et al., 2010; Ranjitkar et al., 2010; Samel et al., 2012; Au et al., 2013; Boeckmann et al., 2013; Ohkuni et al., 2016; Hoffmann et al., 2018). Previous studies have shown that an evolutionarily conserved Ipl1/Aurora B contributes to phosphorylation of Cse4 (Buvelot et al., 2003; Boeckmann et al., 2013). Using mass spectrometric analysis of Cse4 from wild-type yeast cells, we have previously reported the in vivo phosphorylation of Cse4 sites S22, S33, S40, and S105 (Boeckmann et al., 2013). Moreover, a recent study has confirmed the presence of in vivo phosphorylation of Cse4 on serine 33 and shown that cse4-S33A mutants show reduced levels of Cse4 at CEN when combined with the mutations in histone H2A and H4 (Hoffmann et al., 2018). However, the protein kinase responsible for this modification has not been defined.

Evolutionarily conserved polo-like kinase Cdc5 (Plk1 in humans) regulates several aspects of mitotic cell cycle and chromosome segregation (St-Pierre et al., 2009; Walters et al., 2014; Zitouni et al., 2014) including sister chromatid separation by phosphorylation of Mcd1/Scc1 promoting its proteolytic cleavage by separase (Uhlmann et al., 2000; Alexandru et al., 2001). Cdc5 associates with CEN and cohesin-associated regions (CARs) along chromosome arms in a cell cycle–regulated manner and is required for the removal of cohesins from the CEN chromatin during mitosis (Rossio et al., 2010; Mishra et al., 2016). In addition to cohesins Mcd1/Scc1 and Smc3, several other Cdc5-interacting proteins have been identified, such as protein kinase Swe1, protein phosphatase Cdc14, spindle pole body components Spc72 and Spc110, and the Cdc fourteen early anaphase release (FEAR) network protein Slk19 (Alexandru et al., 2001; Snead et al., 2007; Park et al., 2008; Rahal and Amon, 2008; Roccuzzo et al., 2015; Botchkarev and Haber, 2018). Moreover, Plk1 in human cells has been shown to phosphorylate kinetochore protein Mis18BP1 to facilitate the assembly of newly synthesized CENP-A at the CEN (McKinley and Cheeseman, 2014); however, a homologue of Mis18BP1 has not been identified in budding yeast. Intriguingly, a candidate-based screen using Cdc5 polo-box domain (PBD) as a bait showed an enrichment of kinetochore proteins Cse4 and Tid3 (Snead et al., 2007); however, the molecular significance of the interaction of Cdc5 with Cse4, and Cdc5 substrates that localize exclusively to the kinetochore have not been characterized.

In this study, we show that Cdc5 interacts in vivo with Cse4 in mitotic cells (G2/M) and phosphorylates Cse4 in vitro and in vivo. Cdc5-mediated Cse4 phosphorylation regulates faithful chromosome segregation as evident from the increased frequency of chromosome loss in the nonphosphorylatable cse4 mutant (cse4-9SA) when combined with a deletion of MCM21. Significant reduction in levels of kinetochore protein Mif2 and cohesin Mcd1/Scc1 are observed at CEN chromatin in a cse4-9SA mcm21Δ strain. The constitutive association of Cdc5 with Cse4 at the kinetochore causes growth defects suggesting that cell cycle–regulated interaction of these two proteins restricted to mitosis is essential for cell viability. In summary, we have identified Cse4 as a substrate for Cdc5 and shown that Cdc5-mediated phosphorylation of Cse4 contributes to high-fidelity chromosome segregation.

RESULTS

Cdc5 interacts with Cse4 in vivo in a cell cycle–dependent manner

The budding yeast polo-like kinase, Cdc5, associates with centromeres in mitosis and facilitates the removal of CEN cohesin (Mishra et al., 2016). Cse4 was enriched in a screen to identify proteins that interact with the PBD of Cdc5 used as a bait (Snead et al., 2007). We explored the role of the interaction of Cdc5 with Cse4 in faithful chromosome segregation. Immunoprecipitation (IP) experiments were done to determine whether Cdc5 interacts with Cse4 in vivo. We constructed a strain that expresses HA-tagged Cdc5 and Flag-tagged Cse4. IP was done using protein extracts from logarithmically growing asynchronous cultures (Figure 1, A and B). Western blotting showed that Cdc5 interacts with Cse4 in vivo, whereas no signals were detected in a control experiment using an untagged strain (Figure 1C).

FIGURE 1:

Cdc5 interacts in vivo with Cse4 in a cell cycle–dependent manner. Strains carrying vector control (Untagged, YMB9325), Cdc5-HA (YMB9326), Cse4-Flag (YMB9327), and Cdc5-HA Cse4-Flag (YMB9328) were grown at 30°C to logarithmic phase and synchronized in G1, S, and G2/M stages of the cell cycle. Cell extracts were prepared for immunoprecipitation experiments using α-Flag agarose antibodies. (A) FACS profiles show DNA content in different stages of the cell cycle. (B) Cell cycle stages were determined based on nuclear position and cell morphology by microscopic examination of at least 100 cells for each sample. Different stages of the cell cycle: G1, S phase (S), and mitosis (G2/M). (C) In vivo interaction of Cdc5 with Cse4 is observed in G2/M cells. Immunoprecipitated proteins were analyzed by Western blotting with α-HA (Cdc5), and α-Flag (Cse4) antibodies. IP-Flag represents immunoprecipitated samples.

To determine whether the in vivo interaction of Cdc5 with Cse4 is cell cycle regulated, IP experiments were performed using cells synchronized in G1 (α-factor treatment), S (hydroxyurea treatment), or G2/M (nocodazole treatment) stages of the cell cycle. The cell cycle synchronization was confirmed by fluorescence-activated cell sorting (FACS; Figure 1A) and examination of nuclear and cell morphology (Figure 1B). In agreement with previous studies (Charles et al., 1998; Mishra et al., 2016), Cdc5 was expressed in S and G2/M phases of the cell cycle, whereas no protein expression was detected in G1 cells (Figure 1C). IP results showed an in vivo interaction between Cdc5 and Cse4 in G2/M cells (Figure 1C). No interaction of Cdc5 with Cse4 was detected in G1 and S-phase cells despite the expression of Cdc5 in S phase (Figure 1C). As expected, no signals were detected in control experiments performed with an untagged strain (Figure 1C). Taken together, these results provide evidence for cell cycle–regulated in vivo interaction of Cdc5 and Cse4 that occurs in mitotic cells.

Cdc5 phosphorylates Cse4 in vitro

To determine whether Cdc5 phosphorylates Cse4 directly, we performed in vitro kinase assays with radiolabeled ATP using Cdc5 purified from yeast (Ratsima et al., 2011) and Cse4 purified from Escherichia coli. Cse4 was radiolabeled in the presence of Cdc5, whereas no signal was observed from control in vitro assays containing purified histone H3 (Figure 2A). We next performed in vitro kinase assay by incubating purified Cse4 either with wild-type Cdc5 or its kinase inactive Cdc5kd protein (i.e., Cdc5-K110M; Ratsima et al., 2011). Radiolabeled Cse4 was detected in the presence of wild-type Cdc5 but not the kinase inactive cdc5kd protein (Figure 2B) suggesting that the assay specifically reflects Cdc5-mediated kinase activity toward Cse4.

FIGURE 2:

Cdc5 phosphorylates Cse4 in vitro mediated by its kinase domain. (A) Cdc5 phosphorylates Cse4 in vitro. Kinase assays were carried out in vitro using purified Cse4, Cdc5, and radiolabeled ATP at 30°C for 60 min and products were analyzed by SDS gel electrophoresis followed by Coomassie blue staining and autoradiography of radiolabeled proteins. Purified histone H3 with Cdc5 served as control. (B) Phosphorylation of Cse4 is mediated by the kinase domain of Cdc5. In vitro kinase assays were carried out using purified Cse4, Cdc5, or Cdc5kd (K100M, a kinase-dead variant of Cdc5; Ratsima et al., 2011] and radiolabeled ATP as described above.

Cdc5-mediated phosphorylation of Cse4 occurs largely within the N-terminus of Cse4

To identify Cse4 residues phosphorylated by Cdc5, we performed an in vitro kinase assay as described in Figure 2A, and samples were analyzed by liquid chromatography–tandem mass spectrometry (LC-MS/MS). A total of nine phosphorylated serine sites (S9, S10, S14, S16, S17, S33, S40, S105, and S154) were identified (Figure 3A). Except for S154, which is located within the C-terminus histone fold domain, the remaining eight serine residues are largely clustered within the N-terminus of Cse4. Sequence analysis showed that these sites are evolutionarily conserved among different yeast species containing point centromeres (Figure 3B). To explore the physiological effects of the Cdc5-mediated Cse4 phosphorylation, we constructed a phosphorylation-deficient cse4 mutant, in which all nine phosphorylated serines were changed to alanine (cse4-9SA). We examined the ability of cse4-9SA to complement the growth of cse4Δ strain using 5-fluoroorotic acid (5-FOA)–mediated plasmid shuffle assay (Widlund and Davis, 2005; Tukenmez et al., 2016). Strains carrying cse4-9SA grew robustly on 5-FOA plates confirming that cse4-9SA allele can complement the cse4∆ (Figure 3C). As expected, no growth on 5-FOA was observed in cse4∆ strains with a vector used as a negative control (Figure 3C). We next examined the levels of endogenously HA-tagged Cse4 and Cse4-9SA at the CEN in a wild-type strain grown at 25°C. ChIP-qPCR (chromatin immunoprecipitation–quantitative PCR) showed that the CEN levels of Cse4 and Cse4-9SA were not significantly different (Figure 3D; p value > 0.05). No significant enrichment of Cse4 or Cse4-9SA was detected at a negative control non-CEN HML locus (Figure 3D).

FIGURE 3:

Cdc5 phosphorylates Cse4 at its N-terminus in vitro, and contributes to Cse4 phosphorylation in vivo. (A) Cse4 peptides phosphorylated in vitro by Cdc5 were identified by LC-MS/MS. Phosphorylated serines are marked in blue shading. (B) The region containing the phosphorylated serines within the Cse4 (shaded blue) is evolutionarily conserved among yeasts with point centromeres. ClustalW alignment of the Cse4 regions of Sbay = Saccharomyces bayanus, Scer = S. cerevisiae, Sbou = Saccharomyces boulardii, Spas = Saccharomyces pastorianus, Spar = Saccharomyces paradoxus, and Sarb = Saccharomyces arboricola. (C) cse4-9SA mutant is viable. Wild-type strain with CSE4::URA3 (pRB199) was transformed with vector::LEU2 (YMB10341), CSE4::LEU2 (YMB10049), or cse4-9SA::LEU2 (YMB10339). Strains were plated on synthetic medium without or with counterselection for URA3 by 5-FOA and incubated for 7 d at 25°C. (D) The levels of Cse4 and Cse4-9SA are not significantly different at the CEN chromatin. Wild-type (WT; YMB9383) and cse4-9SA (YMB10593) strains were grown in YPD to logarithmic phase at 25°C, and ChIP for endogenously expressed HA-tagged Cse4 or Cse4-9SA was performed using α-HA agarose antibodies. Enrichment of Cse4 or Cse4-9SA at CEN1, CEN3, CEN5, and a negative control (HML) was determined by qPCR and is presented as percentage of input. The average from three biological replicates ± SE is shown. No statistically significant difference was observed between wild-type and cse4-9SA strains (p value ≥ 0.05; Student’s t test). (E) Cse4-9SA protein does not react with αp-Cse4 antibodies. Wild-type strains transformed with GAL1-6HIS-3HA-CSE4 (YMB10426) or GAL1-6HIS-3HA-cse4-9SA (YMB10427) were grown to logarithmic phase of growth in synthetic medium, and gene expression was induced in the presence of galactose plus raffinose (2% each) at 25°C for about four generations of growth. Protein extracts were prepared for affinity purification of Cse4 or Cse4-9SA strains using Ni²⁺-NTA agarose. Eluted proteins were analyzed by Western blotting. Antibodies used were α-HA (Cse4) and αp-Cse4–specific (pCse4) antibodies (Boeckmann et al., 2013). (F) Cdc5 contributes to Cse4 phosphorylation in vivo. FACS profiles show G1 synchronization and release into pheromone-free media to enrich cells in metaphase. Wild-type (YMB10986) and cdc5-99 (YMB10987) strains expressing GAL1-6HIS-3HA-CSE4 (pMB1601) were synchronized in G1 (1.5 μM α-factor) in 1× SC –URA galactose plus raffinose (2% each) for 2 h at 25°C. Cells were collected, washed with water, and released into pheromone-free 1× SC –URA galactose plus raffinose (2% each) at 25 and 37°C for ∼110 min (∼70% cells in metaphase). Protein extracts were prepared and affinity purified as described in E. (G) Cell and nuclear morphology of strains from F post-G1 release into pheromone-free media (∼110 min) showing enrichment of cells in the metaphase stage of the cell cycle. The average from three biological replicates ± SD is shown. (H) Western blotting shows a reduction of Cse4 phosphorylation in cdc5-99 at the nonpermissive temperature (37°C). Affinity-purified proteins from strains grown in F were separated on polyacrylamide gels, and transferred to nitrocellulose membranes. Blots were probed with antibodies: α-HA (total Cse4), and αp-Cse4 antibodies (Boeckmann et al., 2013). Three biological replicates were performed. (I) Quantification of relative phosphorylation of Cse4 from Western blots. Ratio of phosphorylated Cse4 (pCse4) to the total Cse4 levels (Cse4) in wild-type and cdc5-99 strains was calculated. The histogram represents the average of three biological replicates ± SE. **, p < 0.01; Student’s t test.

Cdc5 contributes to the phosphorylation of Cse4 in vivo

We have previously used α-rabbit polyclonal phospho-Cse4–specific (αp-Cse4) antibodies that did not react with Cse4-4SA in which four serine sites of Cse4 were mutated to alanine (S22A, S33A, S40A, and S105A) to show the occurrence of increased levels of phosphorylated Cse4 at the CEN (Boeckmann et al., 2013). Among the four serine sites, three (S33, S40, and S105) are phosphorylated by Cdc5 in vitro (Figure 3A). Hence, we used αp-Cse4 antibody to investigate the role of Cdc5 in Cse4 phosphorylation in vivo. Western blot analysis of affinity-purified Cse4 showed strong reactivity to αp-Cse4 but no signals were detected with Cse4-9SA suggesting that the nine serine residues in Cse4 contribute to the reactivity of Cse4 with αp-Cse4 antibody (Figure 3E). Because we observed an in vivo interaction of Cse4 and Cdc5 in metaphase (Figure 1), we used the αp-Cse4 antibody to examine the in vivo levels of Cse4 phosphorylation in metaphase cells from wild-type and a well-characterized temperature-sensitive cdc5-99 mutant (St-Pierre et al., 2009). Western blot analysis was done using affinity-purified Cse4 from metaphase cells collected ∼110 min after release from G1 arrest into pheromone-free media at 25 and 37°C (Figure 3, F and G). Our results showed similar levels of expression of Cse4 in wild type and cdc5-99, both at permissive (25°C) and nonpermissive (37°C) temperature of growth (Figure 3H). The levels of p-Cse4 were similar at 25°C between wild type and cdc5-99; however, the levels of p-Cse4 were lower in cdc5-99 than in the wild-type strain at 37°C (Figure 3H). We quantified the fraction of phosphorylated Cse4 and normalized this to the total Cse4 levels for each sample. The level of phosphorylated Cse4 was significantly lower (∼30%) in cdc5-99 than in the wild-type strain at 37°C (Figure 3I). Taken together, these results indicate that Cdc5 contributes to the phosphorylation of Cse4 in vivo.

Cse4 phosphorylation-deficient and mcm21Δ mutants exhibit synthetic defects in chromosome segregation fidelity

With the exception of one serine, eight of the nine Cse4 serine residues that are phosphorylated by Cdc5 are in the N-terminus of Cse4 (Figure 3A). Cse4 interacts in vivo with Ctf19 and Mcm21 (Ortiz et al., 1999; Ranjitkar et al., 2010), and this interaction is mediated by the N-terminus of Cse4 (Chen et al., 2000). Genetic interactions have also been reported for mutants of cse4 with ctf19∆ and mcm21∆ (Samel et al., 2012). Moreover, Mcm21 and Ctf19 have additional roles in maintenance of CEN cohesion (Ng et al., 2009; Hinshaw et al., 2017), a biological process in which Cdc5 is also involved (Rossio et al., 2010; Mishra et al., 2016). Because the Cdc5 target sites in Cse4 are clustered largely within the N-terminus of Cse4, we assayed chromosome segregation in cse4-9SA strains in combination with deletions of MCM21 or CTF19. The loss of a nonessential reporter chromosome fragment (CF) was measured using the colony color assay as described previously (Spencer et al., 1990). The frequency of CF loss is slightly higher in cse4-9SA when compared with the wild-type strain, but the difference is not statistically significant (Figure 4A). The frequency of CF loss in mcm21Δ and ctf19Δ is significantly higher than the wild-type or cse4-9SA strains (Figure 4A). The frequency of CF loss in ctf19Δ and cse4-9SA ctf19Δ mutant is similar and does not differ significantly from each other (p value = 0.3). However, the frequency of CF loss in cse4-9SA mcm21Δ mutant is significantly higher than the mcm21Δ (∼5-fold; p value = 0.0023), cse4-9SA (∼30-fold; p value = 0.0009), and the wild-type (∼50-fold; p value = 0.0008) strains (Figure 4A). These results show that Ctf19-independent events contribute to increased chromosome loss in cse4-9SA mcm21Δ strains but do not rule out a role for Cse4 or Cse4-9SA in the loading of Ctf19 to the CEN chromatin.

FIGURE 4:

Cdc5-mediated phosphorylation contributes to faithful chromosome segregation and modulates the levels of Mif2 and Mcd1/Scc1 at the CEN chromatin. (A) Errors in chromosome segregation are increased in cse4-9SA mcm21Δ strains. Frequency of CF loss in wild-type (YPH1018), cse4-S33A (YMB10984), cse4-9SA (YMB10337), mcm21Δ (YMB10645), cse4-S33A mcm21Δ (YMB10985), cse4-9SA mcm21Δ (YMB10646), ctf19Δ (YMB10647), and cse4-9SA ctf19Δ (YMB10648) strains was determined using a colony color assay as described in Materials and Methods. At least 1000 colonies from three independent transformants were counted, and the average from three biological experiments ± SE is shown. **, p value < 0.01; *, p value < 0.05; ns = statistically not significant; Student’s t test. (B) The CEN levels of Mif2 and Mcd1/Scc1 are reduced in cse4-9SA mcm21Δ strains. FACS profiles show DNA content representing the G2/M stage of the cell cycle. Wild-type (YMB9695), cse4-9SA (YMB10593), mcm21Δ (YMB10740), and cse4-9SA mcm21Δ (YMB10741) carrying Mcd1-GFP were grown in YPD to logarithmic phase at 30°C and synchronized in G2/M with nocodazole. ChIP was performed using α-Mif2 antibodies and α-GFP sepharose beads (Mcd1/Scc1) as described in Materials and Methods. (C) Enrichment of Mif2 at CEN3, CAR (134 and 261), and non-CAR control region (310) on chromosome III was determined by ChIP-qPCR and is presented as the percentage of input. Average values from three biological replicates ± SE are shown. *, p value < 0.05; ns = statistically not significant; Student’s t test. (D) Western blotting showing expression of Mif2 in strains used in ChIP experiments. Antibodies used were α-Mif2 and α-Tub2 (loading control). (E) Enrichment of Mcd1/Scc1 at CEN3, CAR (134 and 261), and non-CAR control region (310) on chromosome III was determined by ChIP-qPCR and is presented as the percentage of input. The average values from three biological replicates ± SE are shown. **, p value < 0.01; *, p value < 0.05; ns = statistically not significant; Student’s t test. (F) Western blotting showing expression of Mcd1/Scc1 in strains used in ChIP experiments. The antibodies used were α-GFP and α-Tub2 (loading control).

We next examined whether defects in phosphorylation of Cse4-S33 affects chromosome segregation when combined with mcm21∆. The rationale for this experiment is based on our identification of Cse4-S33 as a potential Cdc5 phosphorylation site and a recent study, which showed that phosphorylation-deficient cse4-S33A and mutations in histones H2A and H4 exhibit synthetic defects in CEN deposition of Cse4 (Hoffmann et al., 2018). The frequency of CF loss in cse4-S33A is statistically similar to that observed for wild-type or cse4-9SA strains (p value = 0.85). However, the frequency of CF loss in cse4-S33A mcm21Δ mutant is significantly higher than the mcm21Δ (∼2-fold; p value = 0.021), but is significantly lower than cse4-9SA mcm21Δ (p value = 0.0094) strains (Figure 4A). Taken together, these results support a role for phosphorylation of Cse4 in faithful chromosome segregation.

Phosphorylation of Cse4 regulates the CEN association of kinetochore protein Mif2 and cohesin component Mcd1/Scc1

Our results for increased chromosome loss in cse4-9SA mcm21Δ strains prompted us to examine the role of Cse4 phosphorylation in kinetochore structure. Hence, we examined the levels of CEN-associated kinetochore protein Mif2, the yeast orthologue of mammalian CENP-C, which contributes to localization of Cse4 at the CEN, maintenance of spindle integrity, and cohesin-based partitioning mechanisms at the kinetochore (Brown et al., 1993; Meluh and Koshland, 1995; Cohen et al., 2008; Ho et al., 2014; Tsabar et al., 2016). ChIP experiments were performed to determine the enrichment of Mif2 at CEN and CARs: peri-CEN (134), chromosomal arm (261), and negative control region (310) on chromosome III in cells synchronized with nocodazole in the G2/M stage of the cell cycle (Figure 4B). No significant enrichment of Mif2 was detected at CARs located at peri-CEN (134), chromosomal arm (261), or a negative control region (310; Figure 4C). ChIP-qPCR revealed mildly lower levels of CEN-associated Mif2 in cse4-9SA than the wild-type strain (Figure 4C). However, Mif2 levels at CEN were significantly lower in cse4-9SA mcm21Δ than the wild-type or single mutant strains (Figure 4C). Western blotting revealed that the reduction in CEN-associated Mif2 in the cse4-9SA mcm21Δ strain was not due to a reduction in the levels of Mif2 (Figure 4D). On the basis of these results, we conclude that phosphorylation of Cse4 regulates CEN association of Mif2 in the absence of Mcm21.

Previous studies have shown that cse4 and mcm21Δ strains exhibit reduced levels of cohesin at the CEN and peri-CEN chromatin (Weber et al., 2004; Ng et al., 2009). Deletion of MCM21 results in the failure of Ctf19 loading onto de novo kinetochores suggesting that Mcm21 is required for the assembly or productive association of Ctf19 complex at the kinetochores (Lang et al., 2018). Moreover, defects in levels of CEN cohesin have been linked with altered kinetochore function (Brooker and Berkowitz, 2014). Notably, Cdc5 regulates the removal of CEN cohesin (Alexandru et al., 2001; Mishra et al., 2016). On the basis of these results, we postulated that defects in Cdc5-mediated phosphorylation of Cse4 may affect CEN association of cohesins in an mcm21Δ strain. ChIP experiments were performed to examine the enrichment of cohesin component Mcd1/Scc1 at CEN and CARs in mitotic cells. Mcd1/Scc1 enrichment at chromosomal arm region (261) was similar, and was not significantly different among the strains. No significant enrichment of Mcd1/Scc1 was detected at a negative control region (310; Figure 4E). Enrichment of Mcd1/Scc1 at CEN and CARs located at peri-CEN (134) and chromosomal arm (261) was observed in wild-type and cse4-9SA strains (Figure 4E). Significantly reduced levels of Mcd1/Scc1 were observed at CEN and peri-CEN (134) in a mcm21Δ strain (Figure 4E), and this reduction was further exacerbated in cse4-9SA mcm21Δ strain in comparison to wild-type, cse4-9SA, or mcm21Δ strains (Figure 4E). A reduction in enrichment of Mcd1/Scc1 at CEN and peri-CEN in cse4-9SA mcm21Δ strain was not due to a reduction in the protein expression of Mcd1/Scc1 (Figure 4F). On the basis of these results, we conclude that phosphorylation of Cse4 affects the CEN association of Mif2 and cohesins during mitosis.

Centromeric association of Cdc5 regulates CEN-associated Cse4 and structural integrity of kinetochores

Cdc5 associates with CEN chromatin during mitosis (Mishra et al., 2016), which correlates with the increased levels of phosphorylated Cse4 at the CEN (Boeckmann et al., 2013). On the basis of these results, we posit that the absence of Cdc5 from the CEN chromatin may exhibit alterations in levels of CEN-associated Cse4 and defects in the structural integrity of kinetochores. To address this hypothesis, we assayed the CEN association of Cdc5 and Cdc5-99 mutant grown at permissive (25°C) and nonpermissive (37°C) temperatures (St-Pierre et al., 2009). Western blot analysis showed similar levels of expression of Cdc5 and Cdc5-99 at the permissive temperature of 25°C and after a shift to the nonpermissive temperature of 37°C (Figure 5A). ChIP-qPCR showed that the enrichment of Cdc5 and Cdc5-99 at CEN chromatin (CEN1, CEN3, and CEN5) is not significantly different at 25°C (p value > 0.05). However, reduced levels of Cdc5-99 were observed at CEN chromatin (approximately five- to ninefold) at 37°C (Figure 5B). There was no significant enrichment of Cdc5 or Cdc5-99 at the non-CEN negative control region (6K120) relative to that observed at the CEN (Figure 5B). Overall, these results show that mutant Cdc5-99 cannot associate with CEN at the nonpermissive temperature.

FIGURE 5:

Loss of Cdc5 from CEN correlates with the reduction in CEN-associated Cse4 and defects in structural integrity of kinetochores. (A) Expression of Cdc5 is not affected in cdc5-99 mutant grown at the nonpermissive temperature (37°C). Wild-type (YMB9431) and cdc5-99 (YMB9432) were grown to logarithmic phase at 25°C and shifted to the nonpermissive temperature (37°C) for 2.5 h. Whole cell extracts were prepared and Western blots were done using α-Cdc5 and α-Tub2 (loading control) antibodies. (B) Cdc5-99 does not associate with CEN at the nonpermissive temperature (37°C) in cdc5-99 strain. ChIP was performed in strains as described in A using α-Cdc5 antibodies. Enrichment of Cdc5 at CEN1, CEN3, CEN5, and a negative control (6K120) was determined by qPCR and is presented as the percentage of input. The average from three biological replicates ± SE is shown. **, p value < 0.01; ns = statistically not significant; Student’s t test. (C) Cdc5 regulates the levels of Cse4 at the CEN. Wild-type (YMB9383) and cdc5-99 (YMB9175) were grown in YPD to logarithmic phase at 25°C and shifted to the nonpermissive temperature (37°C) for 6 h. ChIP for HA-tagged Cse4 was performed using α-HA agarose antibodies. Enrichment of Cse4 at CEN1, CEN3, CEN5, and a negative control (HML) was determined by qPCR and is presented as the percentage of input. The average from three biological replicates ± SE is shown. *, p value < 0.05; ns = statistically not significant; Student’s t test. (D) Cdc5 is required for the structural integrity of kinetochores. Wild-type (KBY2012) and cdc5-99 (YMB9367) were grown in YPD to logarithmic phase at 25°C and shifted to nonpermissive temperature (37°C) for 6 h. Nuclei were extracted and incubated with 100 units of DraI restriction endonuclease at 37°C for 30 min as described in Materials and Methods. DraI accessibility at CEN3 and ADP1 (control) chromatin is shown. The average from three biological experiments ± SE is shown. *, p value < 0.05; ns = statistically not significant; Student’s t test. Right inset: schematic modified from our previous study (Mishra et al., 2013) for CEN3 and ADP1 regions examined for DraI accessibility.

We next examined the effect of loss of CEN association of Cdc5-99 on the levels of endogenously HA-tagged Cse4 at the CEN using wild-type and cdc5-99 strains grown at 25°C and after a shift to 37°C. ChIP-qPCR showed that the levels of CEN-associated Cse4 in wild-type and cdc5-99 strains at 25°C are not statistically different (p value > 0.05). However, enrichment of Cse4 at the CEN was reduced significantly in cdc5-99 (1.29% of input at CEN1, 1.33% at CEN3, and 1.31% at CEN5) compared with the levels observed in a wild-type strain (2.49% at CEN1, 2.26% at CEN3, and 2.42% at CEN5) at 37°C (Figure 5C). No significant enrichment of Cse4 was detected at the non-CEN HML locus used as a negative control (Figure 5C).

We reasoned that the reduced levels of CEN-associated Cse4 in cdc5-99 strain at 37°C (Figure 5C) may affect the structural integrity of kinetochores. Hence, we used DraI restriction enzyme accessibility assay as described previously to measure the structural integrity of the kinetochore (Saunders et al., 1990; Mishra et al., 2013). DNA was extracted from nuclei prepared from wild-type and cdc5-99 strains grown at 25 and 37°C after treatment with 100 units of DraI. We quantified the levels of DraI accessibility by qPCR using primers flanking CEN3 or a non-CEN control ADP1 region (Mishra et al., 2013). The CEN3 chromatin in cdc5-99 strain was significantly more susceptible to DraI digestion (approximately twofold) at 37°C than that observed at 25°C (Figure 5D). No significant increased DraI accessibility of CEN3 chromatin was observed in a wild-type strain at 25 or 37°C (∼1.1–1.8%), which was similar to that observed for cdc5-99 strain at 25°C (∼1.4–1.9%; Figure 5D). The ADP1 chromatin showed low sensitivity to DraI treatments (∼1.0–1.3%) and no significant differences in the accessibility of DraI to ADP1 region were observed between wild-type and cdc5-99 strains grown at 25 or 37°C (Figure 5D). These results show that CEN association of Cdc5 regulates the structural integrity of kinetochores.

Cell cycle–regulated interaction of Cdc5 with Cse4 is required for cell growth

In vivo interaction between Cdc5 and Cse4 is detectable only in mitotic cells (G2/M; Figure 1C). Hence, we sought to understand the physiological significance of cell cycle–dependent association of Cdc5 with Cse4. We postulated that constitutive association of Cdc5 with Cse4 throughout the cell cycle may affect cell growth. Hence, we used the synthetic physical interaction (SPI) assay (Olafsson and Thorpe, 2015) to examine the effect of constitutive association of Cdc5 with Cse4 at kinetochores. Wild-type Cdc5 protein was linked to the sequence encoding a GFP-binding protein (GBP; Rothbauer et al., 2008), which also carries a tag representing red fluorescent protein (RFP). Plasmids expressing Cdc5-GBP or control-GBP (vector carrying GBP domain) were transformed into strains carrying Cse4-GFP, Cep3-GFP, or non-GFP controls. Microscopic examinations of cells confirmed the colocalization of Cdc5 with Cse4-GFP or Cep3-GFP (Figure 6, A and B). Cep3 is an essential kinetochore protein that binds the CDEIII region of the CEN (Lechner and Carbon, 1991; Strunnikov et al., 1995) and is ∼44 nm away from Cse4 at the metaphase kinetochores (Haase et al., 2013). In control experiment with GBP-RFP, only one or two foci of Cse4 were observed in a cell; however, constitutive association of Cdc5 with Cse4 causes an alteration in its localization pattern as evident from the multiple and diffused Cse4 foci (Figure 6A). Constitutive association of Cdc5 with inner kinetochore protein Cep3, which was used as a control, does not exhibit altered localization phenotype (Figure 6B).

FIGURE 6:

Cell cycle–regulated interaction of Cdc5 with Cse4 is required for cell growth. A synthetic physical interaction (SPI) assay was performed using plasmids expressing Cdc5-GBP, cdc5kd-GBP (kinase-dead version), or GBP alone, which were introduced into Cse4-GFP (internally tagged), Cep3-GFP, and non-GFP strains. (A) The cells from the SPI screen were grown overnight in 1× SC –Leu +Ade with 2% galactose medium at 23°C and imaged using fluorescence microscopy. The GBP-RFP and Cdc5-GBP-RFP signal colocalizes with the GFP signal. Cells with Cse4-GFP and Cdc5-GBP-RFP show multiple Cse4-GFP foci in contrast to Cse4-GFP cells containing GBP-RFP control. (B) Cep3-GFP cells containing either Cdc5-GBP-RFP or GBP-RFP control show normal kinetochore foci; each image is 20.6 µm square. (C) Representative images of the scanned plates from the SPI screen show 16 replicates for each strain (rows) and plasmid (columns) combination. (D) The colony sizes in C were measured and log growth ratios plotted for the GFP and wild-type strains with pCUP1-GBP as controls for each comparison. Error bars indicate SD from the mean. **, p value < 0.01; Student’s t test. (E) The forced association of Cdc5 with Cse4 does not arrest cells at a specific cell cycle stage. Tenfold serial dilutions of wild-type and CSE4-YFP (T664) strains carrying the GAL1-CDC5 (pHT573), GAL1-CDC5∆C-GBP (pHT580), GAL1-cdc5kd∆C-GBP (pHT581), and GAL1-Vector (pHT103) plasmids were spotted onto 1× SC –Leu media containing either 2% glucose (expression OFF) or 2% galactose (expression ON), and grown at 30°C for 2 d. (F) Quantification of the cell cycle stages of the CSE4-YFP (T664) strain carrying either the GAL1-Cdc5∆C-GBP or the GAL1-Vector and GAL1-cdc5kd∆C-GBP control plasmids after growing to logarithmic phase in 1× SC –Leu 2% raffinose media, and then swapped to 1× SC –Leu 2% galactose media for 4 h. The cell cycle stage was assessed by fluorescence microscopy and each cell was counted and given the following cell cycle category: nonbudded cells were categorized as G1 cells, small-budded as S/G2, large-budded cells with two Cse4-YFP foci in the bud neck as metaphase (M), and large-budded cells with completely separated Cse4-YFP foci in the mother and daughter as anaphase/telophase cells. No statistical difference was found between Cdc5∆C-GBP to either control as evaluated by Fisher’s exact test. Error bars indicate 95% confidence interval. (G) cse4-9SD mutant is unable to complement the growth defect of cse4∆ strain. Wild-type strain with CSE4::URA3 (pRB199) was transformed with vector::LEU2 (YMB10341), CSE4::LEU2 (YMB10049), cse4-9SA::LEU2 (YMB10339), or cse4-9SD::LEU2 (YMB10340). Strains were streaked on synthetic medium without or with counterselection for URA3 by 5-FOA and incubated for 6 d at 25°C.

We next performed growth assays using selective ploidy ablation technology (Reid et al., 2011) with 16 replicates per strain to examine the effect of constitutive association of Cdc5 on Cse4. Plates were scanned and growth measurements were determined using the ScreenMill software (Dittmar et al., 2010). Colony sizes were quantified and compared among strains as described previously (Olafsson and Thorpe, 2015). We observed that constitutive association of Cdc5 with Cse4-GFP causes growth defects, whereas no growth inhibition was observed with inner kinetochore protein, Cep3-GFP, or non-GFP control strains (Figure 6, C and D). The growth defects were mediated by the kinase activity of Cdc5 because strains expressing kinase inactive cdc5kd exhibited growth phenotypes similar to Cep3-GFP or non-GFP control strains (Figure 6, C and D). To determine whether constitutive association of Cdc5 with Cse4 causes arrest at a particular cell cycle stage, we created conditionally expressed Cdc5 in which the polo-box domain (PBD) was replaced with GBP (Cdc5ΔC-GBP) and its kinase inactive mutant (cdc5ΔCkd-GBP) under the control of GAL1 promoter. Cdc5ΔC-GBP was used because overexpression of full-length Cdc5 is lethal. GALCdc5ΔC-GBP and GALcdc5ΔCkd-GBP were expressed in a wild-type and Cse4-YFP (yellow fluorescent protein) strains. Consistent with the results of the SPI assay, constitutive expression of Cdc5ΔC-GBP causes lethality in Cse4-YFP strain, but not in a wild-type strain (Figure 6E). Hence, we assayed the cell cycle stages of GALCdc5ΔC-GBP in Cse4-YFP strain after 4 h of growth on galactose. The cell cycle stages categorized based on nuclear position and cell morphology showed that constitutive association of Cdc5 with Cse4 does not cause accumulation of cells in any specific cell cycle stage (Figure 6F). The cell cycle distribution of strains expressing GALCdc5ΔC-GBP is not statistically different from GALcdc5ΔCkd-GBP or an empty vector control (Figure 6F). To further determine the biological significance of constitutive phosphorylation of Cse4, we constructed the phosphomimetic cse4 mutant, in which all nine phosphorylated serines were changed to aspartic acid (cse4-9SD), and examined its ability to complement the growth of a cse4Δ strain after loss of CSE4/URA3 plasmid by counterselection on medium containing 5-FOA. Strains with Cse4 and cse4-9SA grew robustly on plates containing 5-FOA, whereas strains carrying cse4-9SD did not exhibit growth on 5-FOA plates after 6 d of incubation at 25°C (Figure 6G). Taken together, these results show that constitutive association of Cdc5 with Cse4 is detrimental to cell growth and define a physiological role for cell cycle–regulated association of Cdc5 with Cse4.

DISCUSSION

Polo-like kinase Cdc5 and its homologues regulate different stages of the mitotic and meiotic cell cycle and high-fidelity chromosome segregation (Zitouni et al., 2014). Cdc5 in budding yeast associates with the CEN chromatin during mitosis (Mishra et al., 2016); however, kinetochore-specific substrates for Cdc5 and the physiological role of Cdc5-mediated phosphorylation of kinetochore proteins have not been characterized. In this study, we have identified Cdc5 as a kinase for Cse4 and defined a role for Cdc5-mediated Cse4 phosphorylation in faithful chromosome segregation. Our results have shown that 1) Cdc5 interacts in vivo with Cse4 in mitotic cells, 2) Cdc5 phosphorylates Cse4 in vitro, 3) Cdc5 contributes to phosphorylation of Cse4 in vivo, 4) mutations that abrogate Cdc5-mediated phosphorylation of Cse4 (cse4-9SA) lead to increased chromosome loss, reduction in kinetochore protein Mif2, and cohesin Mcd1/Scc1 at the CEN chromatin, 5) constitutive association of Cdc5 with Cse4 at the kinetochore causes growth defects, and 6) mutations that mimic phosphorylation (cse4-9SD) lead to loss of viability. We propose that the cell cycle–regulated association of Cdc5 with Cse4 regulates phosphorylation of Cse4 for the structure and function of the kinetochore and cell viability.

In vitro assay showed that the kinase domain of Cdc5 mediates the phosphorylation of Cse4. The failure of Cdc5 to phosphorylate histone H3 implies that in vitro phosphorylation observed is specific to Cse4. Mass spectrometric analysis revealed that nine of the eight serine residues in Cse4 phosphorylated by Cdc5 are within the N-terminus domain of Cse4 (S9, S10, S14, S16, S17, S33, S40, S105). We previously showed in vivo phosphorylation of Cse4 serine sites: S22, S33, S40, and S105 using mass spectrometric analysis of Cse4 from wild-type yeast cells. The phosphorylation of S40 and S105 is regulated by Aurora B kinase Ipl1 in vitro (Boeckmann et al., 2013). Phosphorylation of Cse4 site S33 has been linked with its CEN deposition as reduced levels of Cse4 were detected in histone H2A and H4 mutants with phosphorylation-deficient cse4-S33A (Hoffmann et al., 2018); however, the kinase responsible for this phosphorylation has not been identified. Our in vitro kinase assay revealed that S33 of Cse4 is a target site for Cdc5 phosphorylation. Moreover, biochemical assays showed that Cdc5 contributes to the phosphorylation of Cse4 in vivo. For example, using αp-Cse4 antibody, we observed a reduction in phosphorylated Cse4 in metaphase cells of a temperature-sensitive cdc5-99 mutant (St-Pierre et al., 2009). It is notable that a fraction of Cse4 can still be phosphorylated in cdc5-99 mutant suggesting that this may be mediated by the Ipl1 kinase as reported previously (Boeckmann et al., 2013). Together, these data show that phosphorylation of Cse4 is facilitated by at least these two kinases. It is possible that Cdc5 and Ipl1 may regulate differential phosphorylation of Cse4 in response to geometric or conformational changes at the kinetochores during the cell cycle (Pearson et al., 2004; Yeh et al., 2008; Verdaasdonk et al., 2012). This conclusion is consistent with previous reports for multiple protein kinases coordinatively modifying a substrate in response to cell cycle dynamics. For example, Cdc28 and Cdc5 work synergistically for the phosphorylation of Swe1 and condensin in budding yeast (Asano et al., 2005; St-Pierre et al., 2009; Robellet et al., 2015). Moreover, cyclin-dependent kinase (Cdk), meiosis-specific kinase (Ime2), and Cdc5 block DNA replication between the two meiotic divisions by phosphorylation of several components involved in helicase loading and an essential helicase-activation protein Sld2 (Phizicky et al., 2018). Notably, Cdk and Cdc7 kinases function in a concerted manner in phosphorylation of Mcm2 in human cells (Cho et al., 2006). Phosphorylation of S26 and S40 of Mcm2 by both Cdk and Cdc7 kinases have been implicated in DNA replication (Cho et al., 2006).

Our study revealed that the in vivo interaction of Cdc5 and Cse4 is cell cycle regulated and occurs in mitotic cells. The mitotic interaction of Cdc5 with Cse4 is coincident with the cell cycle–regulated association of Cdc5 with CEN chromatin in metaphase and early anaphase cells, but lack of enrichment in telophase and G1 cells (Mishra et al., 2016). Notably, the mitotic interaction of Cdc5 with Cse4 also correlates with the increased levels of phosphorylated Cse4 observed at CEN in cells arrested in the G2/M stage of the cell cycle but not in G1 (Boeckmann et al., 2013). Taken together, our results show that the phosphorylation pattern of Cse4 overlaps with the CEN association and activity of Cdc5 kinase during mitosis (Charles et al., 1998; Alexandru et al., 2001; Hornig and Uhlmann, 2004; Mishra et al., 2016). The cell cycle–dependent phosphorylation of Cse4 is physiologically important because constitutive phosphorylation of Cse4 is detrimental for cell growth as cse4-9SD phosphomimetic mutant cannot rescue the growth of a cse4∆ strain. Consistent with this hypothesis, we have shown that constitutive association of Cdc5 with Cse4 results in growth defects. We propose that cell cycle–regulated association of Cdc5 facilitates dynamic phosphorylation of Cse4 for the maintenance of proper kinetochore structure and faithful chromosome segregation. Previous studies have shown that dynamic phosphorylation of kinetochore proteins, such as Cse4, Dam1, Ndc80, Dsn1, and Ask1 destabilizes defective kinetochore to promote biorientation by interaction with microtubules (Cheeseman et al., 2002; Westermann et al., 2003; Akiyoshi and Biggins, 2010; Boeckmann et al., 2013; Jin et al., 2017).

A defect in Cse4 phosphorylation (cse4-9SA) causes increased errors in chromosome segregation when combined with mcm21Δ indicating a role for Cse4 phosphorylation in the maintenance of kinetochore integrity during mitosis. This is not surprising given that cse4-4SA and cse4-S33A exhibit phenotypic defects only when combined with dam1 and hhf1 mutants, respectively (Boeckmann et al., 2013; Hoffmann et al., 2018). Moreover, both Cse4 and Mcm21 have roles in CEN structure function, spindle biorientation, and maintenance of CEN cohesion (Meluh et al., 1998; Ng et al., 2009; Pekgoz Altunkaya et al., 2016; Tsabar et al., 2016; Mishra et al., 2018). Our results showing significantly reduced levels of Mif2 at CEN in cse4-9SA mcm21Δ compared with the single mutant (cse4-9SA or mcm21Δ) further supports a role for phosphorylation of Cse4 in the assembly of CEN chromatin and kinetochore function. As CEN localization of Mif2 requires Cse4 (Ho et al., 2014), the reduced levels of Mif2 at CEN in cse4-9SA mcm21Δ strain may be a reflection of altered association of cse4-9SA at the CEN. A previous study has shown that Mif2 and Cse4 are required for the association of cohesins at the centromeres (Eckert et al., 2007). Moreover, deletion of MCM21 affects the assembly of Ctf19 complex at the kinetochores (Lang et al., 2018). In agreement with these reports, our results showed reduced levels of Mcd1/Scc1 at the CEN in cse4-9SA mcm21∆ strains that exhibited a reduction in Mif2 at the CENs. We propose that Cdc5-mediated phosphorylation of Cse4 contributes to faithful chromosome segregation.

In summary, we have shown that Cdc5 interacts with Cse4 in vivo in a cell cycle–dependent manner, and this interaction is essential for cell viability. We provide the first evidence for a functional role for Cdc5-mediated phosphorylation of Cse4 in faithful chromosome segregation. It is notable that Plk1 (Cdc5 homologue in humans) phosphorylates kinetochore protein Mis18BP1, which in turn promotes the CEN assembly of newly synthesized CENP-A (McKinley and Cheeseman, 2014). However, it remains unexplored whether CENP-A is a direct substrate for Plk1 in human cells. Identification and characterization of additional Plk1 substrates at the human kinetochores will allow us to better understand the role of epigenetic modifications, such as phosphorylation in the assembly of a functional kinetochore for chromosomal stability.

MATERIALS AND METHODS

Yeast strains, plasmids, and growth conditions

Yeast strains were grown in yeast peptone dextrose medium (1% yeast extract, 2% bacto-peptone, 2% glucose; YPD) or in synthetic medium with supplements to allow for the selection of plasmids being used. Yeast strains and plasmids are listed in Table 1.

TABLE 1:

Strains and plasmids used in this study.

(A) Saccharomyces cerevisiae strains	Genotype	Reference
YMB9325	MATa ura3-52 lys2-801 ade2-101 trp1Δ63 his3Δ200 leu2Δ1 TRP1::CEN URA3::CEN	This study
YMB9326	MATa ura3-52 lys2-801 ade2-101 trp1Δ63 his3Δ200 leu2Δ1 CDC5-3HA::TRP1 URA3::CEN	This study
YMB9327	MATa ura3-52 lys2-801 ade2-101 trp1Δ63 his3Δ200 leu2Δ1 TRP1::CEN GAL1-FLAG-CSE4::URA3	This study
YMB9328	MATa ura3-52 lys2-801 ade2-101 trp1Δ63 his3Δ200 leu2Δ1 CDC5-3HA::TRP1 GAL1-FLAG-CSE4::URA3	This study
YMB10341	MATα cse4∆::kanMX pRS416-CSE4 (pRB199) GAL1-Vector::LEU2	TianYi Zhang, National Cancer Institute, Bethesda, MD
YMB10049	MATα cse4∆::kanMX pRS416-CSE4 (pRB199) GAL1CSE4-3HA::LEU2	TianYi Zhang
YMB10339	MATα cse4∆::kanMX pRS416-CSE4 (pRB199) GAL1cse4-9SA-3HA::LEU2	This study
YMB10340	MATα cse4∆::kanMX pRS416-CSE4 (pRB199) GAL1cse4-9SD-3HA::LEU2	This study
YPH1018	MATα ura3-52 lys2-801 ade2-101 trp1∆63 his3∆200 leu2∆1 CFIII (CEN3L.YPH278) HIS3 SUP11	Phil Hieter, University of British Columbia, Vancouver, BC, Canada
YMB10337	MATα ura3-52 lys2-801 ade2-101 trp1∆63 his3∆200 leu2∆1 CFIII (CEN3L.YPH278) HIS3 SUP11 cse4-9SA-3HA::URA3	This study
YMB10645	MATα ura3-52 lys2-801 ade2-101 trp1∆63 his3∆200 leu2∆1 CFIII (CEN3L.YPH278) HIS3 SUP11 mcm21∆::kanMX	This study
YMB10646	MATα ura3-52 lys2-801 ade2-101 trp1∆63 his3∆200 leu2∆1 CFIII (CEN3L.YPH278) HIS3 SUP11 mcm21∆::kanMX cse4-9SA-3HA::URA3	This study
YMB10647	MATα ura3-52 lys2-801 ade2-101 trp1∆63 his3∆200 leu2∆1 CFIII (CEN3L.YPH278) HIS3 SUP11 ctf19∆::kanMX	This study
YMB10648	MATα ura3-52 lys2-801 ade2-101 trp1∆63 his3∆200 leu2∆1 CFIII (CEN3L.YPH278) HIS3 SUP11 ctf19∆::kanMX cse4-9SA-3HA::URA3	This study
YMB10984	MATα ura3-52 lys2-801 ade2-101 trp1∆63 his3∆200 leu2∆1 CFIII (CEN3L.YPH278) HIS3 SUP11 cse4-S33A-3HA::NAT	This study
YMB10985	MATα ura3-52 lys2-801 ade2-101 trp1∆63 his3∆200 leu2∆1 CFIII (CEN3L.YPH278) HIS3 SUP11 mcm21∆::kanMX cse4-S33A-3HA::NAT	This study
YMB9695	MATa MCD1-GFP leu2-3112 ura3-52 his3-11,15 bar1 GAL+ SPC29-RFP::Hyg	Mishra et al., 2016
YMB10593	MATa MCD1-GFP leu2-3112 ura3-52 his3-11,15 bar1 GAL+ SPC29-RFP::Hyg cse4-9SA-3HA::URA3	This study
YMB10740	MATa MCD1-GFP leu2-3112 ura3-52 his3-11,15 bar1 GAL+ SPC29-RFP::Hyg mcm21∆::HIS3	This study
YMB10741	MATa MCD1-GFP leu2-3112 ura3-52 his3-11,15 bar1 GAL+ SPC29-RFP::Hyg mcm21∆::HIS3 cse4-9SA-3HA::URA3	This study
YMB9431	MATα ura3-1 leu2-3112 his3-11,15 trp1-1 ade2-1 can1-100 Smc3-GFP::URA3	This study
YMB9432	MATα ura3-1 leu2-3112 his3-11,15 trp1-1 ade2-1 can1-100 Smc3-GFP::URA3 cdc5-99::HIS3MX	This study
YMB9383	MATa ade2-1 ura3-1 his3-11,15 leu2,3–112 can1-100 CSE4-3HA::NAT	This study
YMB9175	MATa ade2-1 ura3-1 his3-11,15 trp1-1 leu2,3–112 can1-100 CSE4-3HA::NAT cdc5-99::HIS3MX	This study
KBY2012	MATa trp1Δ63 leu2Δ ura3-52 his3 Δ 200 lys2-8Δ1 CSE4GFP::TRP1 (pKK1) SPC29CFP::kanMX	Haase et al., 2013
YMB9367	MATa trp1Δ63 leu2Δ ura3-52 his3 Δ 200 lys2-8Δ1 CSE4GFP::TRP1 (pKK1) SPC29CFP::kanMX cdc5-99::HIS3MX	This study
W8164-2B	MATα CEN1-16::Gal-Kl-URA3	Reid et al., 2011
CEP3-GFP strain	MATahis3∆1 leu2∆0 met15∆0 ura3∆0 CEP3-GFP::HIS3	Huh et al., 2003
T548	MATa his3∆1 leu2∆0 met15∆0 ura3∆0 CSE4-GFP (internal)::HIS3MX6	This study
BY4742	MATα his3Δ1 leu2Δ0 met15Δ0 ura3Δ0	Resgen
T664	MATα his3∆1 leu2∆0 lys2∆0 ura3∆0 CSE4-YFP (internal)::HIS3MX6	This study
YMB10426	MATa ura3-1 leu2-3112 his3-11,15 trp1-1 ade2-1 can1-100 GAL1-6HIS-3HA-CSE4::LEU2 (pMB1515)	This study
YMB10427	MATa ura3-1 leu2-3112 his3-11,15 trp1-1 ade2-1 can1-100 GAL1-6HIS-3HA-cse4-9SA::LEU2 (pMB1847)	This study
YMB10986	MATa ade2-1 ura3-1 his3-11,15 trp1-1 leu2,3-112 can1-100 GAL1-6HIS-3HA-CSE4::URA3 (pMB1601)	This study
YMB10987	MATa ade2-1 ura3-1 his3-11,15 trp1-1 leu2,3-112 can1-100 cdc5-99::HIS3MX6 GAL1-6HIS-3HA-CSE4::URA3 (pMB1601)	This study

(B) Plasmids	Description	Reference
p344	CDC5-3HA::TRP1	D. D’Amours, ­University of ­Otttawa, Ottawa, ON, Canada
pRB199	CSE4-3HA::URA3	R. Baker, University of Massachusetts Medical School, Worcester, MA
pHT4	pCUP1-GBP-RFP LEU2	Olafsson and Thorpe, 2015
pHT425	pCUP- CDC5-GBP LEU2	This study
pHT442	pCUP1-cdc5kd-GBP LEU2	This study
pHT103	pGAL1-empty LEU2	Olafsson and Thorpe, 2016
pHT573	pGAL1-CDC5 LEU2	This study
pHT580	pGAL1-CDC5∆C-GBP LEU2	This study
pHT581	pGAL1-cdc5kd∆C-GBP LEU2	This study
pMB1515	pGAL1-6HIS-3HA-CSE4 LEU2	This study
pMB1847	pGAL1-6HIS-3HA-cse4-9SA LEU2	This study
pMB1601	pGAL1-6HIS-3HA-CSE4 URA3	This study

Chromosome segregation assay

The fidelity of chromosome segregation was measured using a colony color assay as described previously (Spencer et al., 1990). In this assay, the loss of a nonessential reporter CF leads to red sectors in an otherwise white colony. Wild-type, cse4-S33A, cse4-9SA, mcm21Δ, cse4-9SA mcm21Δ, cse4-S33A mcm21Δ, ctf19Δ, and cse4-9SA ctf19Δ strains carrying CF were grown in medium selective for the CF to the logarithmic phase, and plated on complete synthetic medium with limiting adenine at 33°C to allow the loss of CF. About 1000 colonies of three transformants were examined for each strain. The frequency of CF loss was determined by counting the colonies that were at least half red representing the loss of the CF during the first mitotic cell division cycle.

Chromatin immunoprecipitation and qPCR

Chromatin immunoprecipitation (ChIP) experiments were performed with three biological replicates following the procedure as described previously (Mishra et al., 2007, 2011). Antibodies used to capture protein–DNA complexes were α-GFP sepharose (ab69314; AbCam), α-Mif2 (a gift from Pam Meluh, Johns Hopkins University), α-Cdc5 (custom made by the D’Amours laboratory; Ratsima et al., 2011; Robellet et al., 2015), and α-HA agarose (A2095; Sigma Aldrich). ChIP-qPCR was performed in a 7500 Fast Real-Time PCR System using Fast SYBR-Green Master Mix (Applied Biosystems, Foster City, CA) with the following conditions: 95°C for 20 s, followed by 40 cycles of 95°C for 3 s and 60°C for 30 s. The enrichment was determined as percent input using the _ΔΔC_T method (Livak and Schmittgen, 2001). Primer sequences are listed in Table 2.

TABLE 2:

Primers used in this study.

Locus	Forward (5′-3′)	Reverse (5′-3′)	Reference
CEN1	CTCGATTTGCATAAGTGTGCC	GTGCTTAAGAGTTCTGTACCAC	Choy et al., 2011
CEN3	GATCAGCGCCAAACAATATGG	AACTTCCACCAGTAAACGTTTC	Choy et al., 2011
CEN5	AAGAACTATGAATCTGTAAATGACTGATTCAAT	CTTGCACTAAACAAGACTTTATACTACGTTTAG	Choy et al., 2011
6K120	AACGTCACTTTTTTTCCAGGG	GCAAAGCTAGCTAACGAACAA	Mishra et al., 2016
HML	CACAGCGGTTTCAAAAAAGCTG	GGATTTTATTTAAAAATCGAGAGG	Choy et al., 2011
CEN3-DraI	TTGATGAACTTTTCAAAGATGAC	GTCAACGAGTCCTCTCTGGCTA	Choy et al., 2011
ADP1	ATCCAAATGTGCTCAAGATAGTAGC	CACCAAACAACATTTACTAGCAGTG	Mishra et al., 2013
134	CCGATGGTTAGGATTTCCAACG	GGTTTTCAGAACAGAATGGGGC	Eckert et al., 2007; Ng et al., 2009
261	TTGCCACAGCCACAGATATAACTG	GATGGACAAAGCGTTGTATCCG	Eckert et al., 2007; Ng et al., 2009
310	TCTCGGAATTTATCATGACCCAT	AAACCCTGCACACATTTCGT	Laloraya et al., 2000

Cell cycle synchronization, IP, and Western blotting

Strains were grown to logarithmic phase at 30°C in synthetic complete (SC) medium lacking tryptophan (1x SC –Trp) and further incubated for 2 h to synchronize cells in G1 (3 μM alpha factor treatment), S (0.2 M hydroxyurea treatment), and G2/M (20 μg/ml nocodazole treatment) stages of the cell cycle. Cells were collected, washed with water, and grown for 1 h in SC –Trp with galactose + raffinose (2% each) medium to induce the expression of Flag-tagged Cse4 expressed from the GAL1 promoter. Culture media also contained the chemicals described above to keep the cells in the G1, S, and G2/M stages of the cell cycle. Samples were collected for nuclear morphology, DNA content, and IP analyses. IP experiments were performed using α-Flag agarose antibodies (A2022; Sigma Aldrich) as described previously (Mishra et al., 2011, 2018). Whole cell extracts were prepared with the trichloroacetic acid method (Kastenmayer et al., 2005), and quantified using Bio-Rad DC protein quantitation assay (Bio-Rad Laboratories, Hercules, CA). Protein samples were resolved on SDS–PAGE and transferred to nitrocellulose membrane. Primary antibodies used for Western blotting were α-HA (H6908; Sigma Aldrich), α-Flag (F3165; Sigma Aldrich), α-GFP (11814460001; Roche), α-Mif2 (a gift from Pam Meluh), and α-Tub2 (custom made by the Basrai laboratory). Secondary antibodies: HRP-conjugated sheep α-rabbit immunoglobulin G (IgG) (NA934V) and HRP-conjugated sheep α-mouse IgG (NA931V) were obtained from Amersham Biosciences (United Kingdom).

In vitro kinase assay and mass spectrometry

In vitro kinase assay and mass spectrometry were carried out using Cse4 produced and purified from E. coli as described previously (Luger et al., 1997; Boeckmann et al., 2013). Wild-type Cdc5 and its kinase-dead derivative (K100M) were purified from yeast as previously described (Ratsima et al., 2011). In vitro kinase assays were performed using radiolabeled ATP in 20-μl reaction volume containing 0.5 μg Cse4, 40 ng Cdc5, 2 mM dithiothreitol, 1 mM MgCl₂, 25 mM Tris-HCl, pH 7.5, 100 μM ATP, and 1 μCi of [γ-³²P]ATP. Control reactions were performed using purified histone H3 with Cdc5. Reactions were incubated at 30°C for 60 min, stopped with 5 μl of 4× NuPAGE LDS loading buffer (Life Technologies, Grand Island, NY), boiled for 5 min at 95°C, and were run on 4–12% Bis-Tris SDS–PAGE (Invitrogen). Gels were stained with Coomassie blue, and radiolabeled proteins were visualized using a Storm Detector Model 860 (Molecular Dynamics). For mass spectrometry, in vitro kinase assay with and without Cdc5 were performed as described previously (Boeckmann et al., 2013). Reactions were analyzed on 4–12% Bis-Tris SDS–PAGE (Invitrogen), and Cse4 bands were excised and subjected to mass spectrometry following the procedures described previously (Waybright et al., 2008; Boeckmann et al., 2013). The Saccharomyces cerevisiae proteome database (www.expasy.org) was used for data analysis.

In vivo assay for phosphorylation of Cse4

The levels of Cse4 phosphorylation in vivo were determined using procedures and αp-Cse4 antibodies as described previously (Boeckmann et al., 2013). Wild-type and cdc5-99 strains carrying 6HIS-3HA-CSE4 expressed from GAL1 promoter were grown in 1× SC –URA with 2% glucose media at 25°C. Cells were washed with water and inoculated into 1× SC –URA with galactose + raffinose (2% each) to induce the expression of 6HIS-3HA-CSE4 and 1.5 µM α-factor to synchronize cells in G1. Cells were collected, washed with water, and released into pheromone-free media (1× SC –Ura with 2% galactose + raffinose) at 25 and 37°C. Cell cycle progression was monitored by microscopic examination of nuclear and cell morphology. Samples for FACS and affinity pull down were collected ∼110 min after G1 release when the majority of cells were at the metaphase (∼70%) stage of the cell cycle. Cells were dissolved in lysis buffer (6 M guanidine chloride, 0.5 M NaCl, 0.1 M Tris, pH 8.0) and whole cell extracts were prepared using a FastPrep-24 5G bead beater (40 s, 10 times, 1 min interval between bursts; MP Biomedicals) at 4°C. Whole cell extracts were clarified by centrifugation and incubated with nickel-charged superflow NTA agarose (Qiagen, Valencia, CA) for 16 h at 4°C. Beads were centrifuged and washed once with lysis buffer, followed by three washes with washing buffer (100 mM Tris-Cl, pH 8.0, 20% glycerol, 1 mM phenylmethylsulfonyl fluoride [PMSF]; 5 min each wash at the room temperature). The bound protein was eluted by boiling at 100°C for 10 min in 2× Laemmli buffer with 200 mM imidazole. Protein samples were resolved by SDS–PAGE on 4–12% Bis-Tris SDS–PAGE and transferred to nitrocellulose membranes. Blots were washed with 1× TBST (Tris-buffered saline plus 0.1% Tween 20) three times for 5 min and blocked for 15 min in 1× TBST containing 5% skimmed milk. Western blot analysis was done using primary antibodies: α-HA (1/1000 dilutions; 12CA5; Roche) or αp-Cse4 (1/250 dilutions; Boeckmann et al., 2013) in 1× TBST with 5% milk. The secondary antibodies used were HRP-conjugated sheep α-rabbit IgG (NA934V) and HRP-conjugated sheep α-mouse IgG (NA931V). Signal intensities from Western blots were quantified using ImageJ (Schneider et al., 2012).

Extraction of nuclei and DraI accessibility assays

Extraction of nuclei and DraI accessibility experiments were as described previously (Mishra et al., 2013) with some modifications. Briefly, cells were dissolved in spheroplasting buffer (20 mM HEPES, pH 7.4, 1.2 M sorbitol, 0.5 mM PMSF), added β-mercaptoethanol (5 μl/ml cell suspension; Sigma Aldrich) and Zymolyase 100T (0.04 mg/ml cell suspension; MP Biomedicals), and incubated at 37°C for spheroplast preparation. Spheroplasting was monitored by measuring OD₈₀₀ in 1% SDS, and reactions were stopped by washing in postspheroplasting buffer (20 mM Pipes, pH 6.8, 1.2 M sorbitol, 1 mM MgCl₂, 1 mM PMSF) when >90% spheroplasting was achieved. Spheroplasts were lysed in 20 mM Pipes (pH 6.8), 18% Ficoll 400, 0.5 mM MgCl₂, and 1 mM PMSF. Nuclei were released by vortexing for 10 min at 4°C and harvested by centrifugation through a glycerol/Ficoll gradient cushion (20% glycerol, 20 mM Pipes, pH 6.8, 7% Ficoll 400, 0.5 mM MgCl₂, 1 mM PMSF). Nuclei were washed and resuspended in DraI buffer (1.0 M sorbitol, 20 mM Pipes, pH 6.8, 0.1 mM CaCl₂, 1 mM PMSF, 0.5 mM MgCl₂). The nuclei concentrations were determined by measuring OD₂₆₀ in alkaline SDS buffer (0.2 N NaOH, 1% SDS). Equal volumes of nuclei (100 μl) from each sample were prewarmed for 5 min at 37°C followed by the addition of 100 units of DraI (New England BioLabs) for 30 min. Restriction digestion was stopped by adjusting aliquots to 2% SDS, 20 mM EDTA. DNA was isolated after extraction with phenol, chloroform, and isoamyl alcohol (twice), treated with RNase A and proteinase K, followed by extraction with chloroform. DNA was precipitated in ethanol at −20°C, collected by centrifugation, dissolved in 1× TE (pH 8.0), and was used in qPCR to determine the susceptibility of CEN3 chromatin to digestion by DraI using Fast SYBR-Green Master Mix (Applied Biosystems, CA) and PCR primers flanking CEN3 and a control region ADP1 (Mishra et al., 2013). The amplification conditions for CEN3 were initial denaturation at 95°C for 20 s followed by cycling of 95°C for 3 s and 60°C for 30 s (data acquisition step); and for ADP1 the amplification conditions were initial denaturation at 95°C for 30 s followed by cycling of 95°C for 15 s, 54°C for 15 s, and 68°C for 60 s (data acquisition step) in a 7500 Fast Real-Time PCR System (Applied Biosystems, CA). The fraction of DNA cleaved by DraI was determined by normalization of C_T values to those obtained from the no DraI control.

SPI and microscopy

SPI screens were performed as previously described (Olafsson and Thorpe, 2015, 2018). Briefly, a universal donor strain, which contains conditional GAL-CEN centromeres, was transformed separately with the control and experimental plasmids (expressing either Cdc5-GBP, cdc5kd-GBP [a kinase-dead version], or GBP alone; all under the control of a constitutive CUP1 promoter). These universal donor strains were then mated with members of the GFP collection arrayed with 16 replicates on 1536-colony rectangular agar plates using a pinning robot (ROTOR robot; Singer Instruments, UK). The resulting diploids were put through a series of sequential selection steps to maintain the plasmid, while destabilizing and then removing the chromosomes of the universal donor strain. The resulting plates were scanned using a desktop flatbed scanner (Epson V750 Pro; Seiko Epson Corporation, Japan). Colony sizes were assessed and the resulting data analyzed using the ScreenMill suite of software (Dittmar et al., 2010). Flourescence imaging was performed on yeast cells embedded in 0.7% low-melting-point agarose dissolved in growth medium. The cells were imaged with a Zeiss Axioimager Z2 microscope using a 63× 1.4 NA oil immersion lens, illuminated with a Zeiss Colibri LED light source (GFP = 470 nm, RFP = 590 nm). Bright-field contrast was enhanced using differential interference prisms. Images were captured using a Flash 4.0 LT CMOS camera with 6.5 µm pixels binned 2 × 2 (Hamamatsu Photonics, Japan).

Abbreviations used:

CAR

cohesin-associated region

CEN

centromere

CF

chromosome fragment

ChIP

chromatin immunoprecipitation

FACS

fluorescence-activated cell sorting

FEAR

Cdc fourteen early anaphase release

FOA

5-fluoroorotic acid

GBP

GFP-binding protein

GFP

green fluorescent protein

IP

immunoprecipitation

LC-MS/MS

liquid chromatography–tandem mass spectrometry

PBD

polo-box domain

qPCR

quantitative PCR

RFP

red fluorescent protein

SPI

synthetic physical interaction

YFP

yellow fluorescent protein.