Condensin HEAT Subunits Required for DNA Repair, Kinetochore/Centromere Function and Ploidy Maintenance in Fission Yeast

Condensin, a central player in eukaryotic chromosomal dynamics, contains five evolutionarily-conserved subunits. Two SMC (structural maintenance of chromosomes) subunits contain ATPase, hinge, and coiled-coil domains. One non-SMC subunit is similar to bacterial kleisin, and two other non-SMC subunits contain HEAT (similar to armadillo) repeats. Here we report isolation and characterization of 21 fission yeast (Schizosaccharomyces pombe) mutants for three non-SMC subunits, created using error-prone mutagenesis that resulted in single-amino acid substitutions. Beside condensation, segregation, and DNA repair defects, similar to those observed in previously isolated SMC and cnd2 mutants, novel phenotypes were observed for mutants of HEAT-repeats containing Cnd1 and Cnd3 subunits. cnd3-L269P is hypersensitive to the microtubule poison, thiabendazole, revealing defects in kinetochore/centromere and spindle assembly checkpoints. Three cnd1 and three cnd3 mutants increased cell size and doubled DNA content, thereby eliminating the haploid state. Five of these mutations reside in helix B of HEAT repeats. Two non-SMC condensin subunits, Cnd1 and Cnd3, are thus implicated in ploidy maintenance.

Condensin contains five essential subunits that are conserved in eukaryotes. Two of them belong to the family of SMC (structural maintenance of chromosomes) proteins, which contain the terminal ATPase domains and the middle hinge domain, interrupted by long coiled coils (Fig. 1A) [14,23,24]. Six SMC proteins are known in eukaryotes, forming three distinct protein complexes, cohesin, condensin, and the SMC5-SMC6 complex [25,26]. In the budding yeast, Saccharomyces cerevisiae, two condensin SMC proteins are called SMC2 and SMC4 [3,14]. In variety of cellular functions. HEAT is an acronym designating four representative proteins, Huntingtin, a protein that is linked to Huntington neurodegenerative disease, elongation factor 3 (EF3), involved in protein translation, protein phosphatase 2A (PP2A) regulatory subunit A, and the target of rapamycin kinase, TOR. In addition to these four proteins, importin β, implicated in protein transport [34,35], ATM (Ataxia telangiectasia/Tel1) kinase, required for DNA damage sensing [36,37], and XMAP215/Dis1/TOG, implicated in chromosome segregation and kinetochore function [38][39][40], contain HEAT repeats. Together, these proteins comprise an impressive group of important cell signaling and regulatory proteins. While the exact role of HEAT repeats is scarcely understood, they are likely to be involved in protein-protein interactions [38,41]. S. pombe condensin non-SMC subunits, Cnd1 and Cnd3, contain 7 and 9 HEAT repeats, respectively [41]. Armadillo (ARM) repeat proteins are structurally related to proteins containing tandem HEAT motifs, and the two protein families probably have a common phylogenetic origin [42,43].
In this study, we attempted to construct mutants of fission yeast condensin non-SMC subunits. We introduced mutations in the genomic cnd + genes by chromosomal integration of mutagenized gene fragments. Candidate integrant strains obtained by transformation were screened for different phenotypes, such as the temperature-sensitive (ts) phenotype or the phenotype hypersensitive to DNA damaging agents such as hydroxyurea (replication inhibitor inducing DNA strand breakage), camptothecin (inhibitor of DNA topoisomerase I), and UV irradiation. Mutation sites were determined by nucleotide sequence determination. To prove that the determined mutation sequences really caused the mutant phenotypes, we also introduced the mutated sequence into the wild-type genome to see whether the same phenotypes appeared in mutated wild-type yeast. We were thus able to obtain and characterize 21 newly isolated mutants of cnd1, cnd2, and cnd3, containing only single amino acid substitutions.
The novelty of our approach for isolating condensin non-SMC mutants was that we did not screen the microscopic condensation-defect, which was used as the selective phenotype in previous studies [17,44]. In this study we isolated mutants that produced ts and/or DNA damagesensitive phenotypes.

Condensin non-SMC mutants produced by error-prone mutagenesis
Temperature-sensitive (ts) S. pombe condensin SMC and non-SMC mutants that exhibited chromosome segregation and condensation defects were previously isolated by cytological screening of many ts strains [12,16,17,[44][45][46]. Among them, cut14 (smc2)-Y1 and cnd2 (kleisin)-1 were sensitive to DNA damage [16,17]. Genetic dissections of condensin's non-SMC subunit functions have been scarce. To characterize these subunits in a systematic way, we isolated mutants for all 3 subunits using error-prone mutagenesis [47] (Fig. 2A step1-8). Three plasmids (pBS-CND1, -CND2, and -CND3, Materials and Methods), carrying one of the three cnd + genes, with the hygR/kanR marker and 500 bp downstream sequences (described in Materials and Methods), were constructed and employed in PCR mutagenesis (using a high magnesium concentration, 8 mM). Mutations were introduced at the rate of approximately one nucleotide per kb (estimated by nucleotide sequence determination).
Then mutagenized DNA fragments were used for transformation of S. pombe by chromosomal replacement integration through homologous recombination. Resulting transformants with antibiotic drug resistance markers were plated, and properties of colony formation were manually determined by streaking (in total *18,000 transformant colonies) on agar plates. In addition, each colony was streaked under five different culture conditions: 1) standard culture conditions (26°C without any additional treatment), 2) at 36°C (to examine the ts phenotype through the lack of colony formation at 36°C, the restrictive temperature, for 3 days), 3) 2 mM hydroxyurea (HU) at 26°C (sensitivity to DNA replication inhibitor drug at the permissive temperature), 4) 7.5 μM camptothecin (CPT) at 26°C (sensitivity to type 1 topoisomerase poison), 5) 150 J/m 2 ultraviolet ray (UV) irradiation at 26°C (sensitivity to UV irradiation). Candidate Tsand/or damage sensitive (Ds -) colonies were re-streaked to confirm the phenotypes (Step 5 in Fig. 2A). Screening procedures for non-SMC subunit mutants of S. pombe condensin. A. Procedures consist of eight steps, starting from construction of integration plasmids and ending with verification of the cnd mutant phenotype produced by chromosomal re-integration (see text). B. Twenty-one temperature-and/or DNA damage-sensitive cnd mutant strains containing various single amino acid substitutions were obtained. Their classification according to phenotype is shown in the table (see text). Ts -, temperature sensitive; Ds -, damage sensitive; Ts -Ds -, temperature and damage sensitive.
To verify that the sequence change in each mutant really caused the phenotype, we then newly constructed cnd mutants based on sequencing information by re-integrating the linearized plasmids containing one of the 21 mutant sequences onto the wild-type S. pombe genome. All of the resulting transformants containing chromosomal gene replacements with mutant genes produced the expected phenotypes (Step 8 in Fig. 2A). Further tetrad dissection was conducted to prove that the ts mutation was really caused by a mutation at one locus. The Ts + :Tsmarkers from all mutants segregated 2:2, and the ts phenotype was tightly linked to the antibiotic drug-resistant phenotype. Hence, we concluded that these 21 newly isolated mutants, containing single amino acid substitutions, could cause either Ts -, Ts -Ds -, or Dsphenotypes (Fig. 2B).
Locations of newly identified mutations in the cnd1, 2, and 3 genes are shown in Fig. 3A (a-u), together with information regarding the substitution mutant sequences (Table in Fig. 3B). Locations of the 5 previously isolated non-SMC ts mutants are shown as w1-w5 (explained in the figure legend). In these newly identified 21 mutants, 11 (4 cnd1, 1 cnd2, and 6 cnd3) displayed only the ts phenotypes (black characters). Five (red characters in Fig. 3) showed both ts (36°C) and damage sensitive phenotypes (at 26°C, phenotypic markers are indicated as Ts -Ds -). The remaining five mutants (3 cnd1 and 2 cnd3) are damage-sensitive and not ts. Among the 21 strains, 11 mutations reside within the HEAT repeats (indicated by H and HEAT repeat number in the repeat clusters). One mutant, cnd1-N914Y, is located near the edge of HEAT repeat 4 (indicated as Edge in Fig. 3B). Note that many amino acid substitution mutations (10 of 21) introduce proline (P), which is known as a helix breaker. In addition, glycine (G) also disrupts helices. Arginine residues (R), which possess a bulky side chain, are also plentiful (6 of 21). The significance of these substitutions from wild-type amino acids (9 of 21 are L) to P and R is discussed below, with their implications for the secondary structure of HEAT repeats.

Defects in chromosome segregation caused by Tsmutations
In liquid cultures, the above 16 ts mutants (11 Tsonly and 5 Ts -Ds -) were cultured first at 26°C and then shifted to 36°C for 0-8 hr. Cells were stained with DAPI (4,6-diamidino-2-phenylindole), a fluorescent probe for DNA, and were observed under a fluorescence microscope. Mutant cells collected from cultures after 2-4 h at 36°C showed severe defects in chromosome segregation. Micrographs from cnd1-L193P (a in Fig. 3A), -L685P (c), cnd2-A708V (i) and cnd3-L662P (s) show defects in condensation and segregation of mitotic chromosomes ( Fig. 4A for 2.5 hrs at 36°C; S1 Fig. for 5 hrs at 36°C). ts cnd1, cnd2, and cnd3 mutants examined so far at 36°C in liquid cultures, showed that defects in mitotic chromosome segregation revealed by DAPI staining were indistinguishable from those found in previously isolated ts condensin mutants of S. pombe [12,16,17]. phi-shaped nuclear chromatin is the hallmark of condensation deficiency. Damage-sensitive cnd1-G984R that was not ts was cultured at 26°C in the presence or absence of 2 mM HU (right panel in Fig. 4A). In the presence of HU, basically identical condensation and segregation defects, as in ts mutants, were observed. In the absence of HU (no treatment), no defects were observed.
Cell viability (%) and cell density (cells/mL) of cultures were measured for cnd1-H1133P, cnd2-A708V, and cnd3-L269P mutants by spreading cells on YPD plates and incubating them at 26°C for 4 days (

Many damage sensitive mutations reside in HEAT repeats
Among the 8 cnd1 mutants, 4 mutations reside in the HEAT repeats (Fig. 3A). Three of the 4 are damage-sensitive. The other damage-sensitive mutant cnd1-N914Y is in close proximity to the C-terminal HEAT repeat 4. cnd1-L1037P (g in Fig. 3A), -G984R (f) and -M951R (e) are sensitive to HU and CPT, but not to UV (Fig. 4D).
Among the 12 cnd3 mutants isolated, seven are located in HEAT repeats. All five damagesensitive mutants are located in HEAT repeats (Fig. 3A). In Fig. 4D, sensitivities of cnd3-K627E (p), -W717R (u) and -L269P (l) to HU, CPT and UV are shown. In total, 11 cnd1 and cnd3  DAPI-stained S. pombe wild-type (WT) and mutants cells cultured at 26°C and 36°C for 2.5 h. Cells displaying severe defects in chromosome segregation were observed in mutant cells at 36°C. More than 100 mitotic cells for each mutant strain were counted. Almost all (95*100%) mitotic mutant cells observed after 2.5 hrs at 36°C showed severe defects of chromosome segregation as shown here. Mutant cells (cnd1-G984R) cultured in 2 mM HU for 2 h, which showed severe segregation defects at 26°C, are also shown (see text). Bars are 10 μm. B and C. In YPD liquid medium at the restrictive temperature, 36°C, ts mutants ceased to divide within 4 h and lost viability. Asynchronous cultures of WT and ts mutants (cnd1-H1133P, cnd2-A708V and cnd3-L269P) in the YPD liquid medium were shifted from 26°C to 36°C. Cell density (per mL) was determined and viability was measured after a temperature shift to 36°C. D. S. pombe wild-type (WT), and cnd1 and cnd3 mutant cells were spotted on solid agar plates in the presence or absence of HU, CPT, and mutants are located in HEAT repeats. Among them, 8 are damage-sensitive. In contrast, for the 10 ts mutants (Ts -), which are not damage-sensitive, only 3 (3/10) are located in HEAT repeats. Interestingly, cnd3 mutants are broadly sensitive to damage, as they were sensitive to UV as well as to HU and CPT. UV damage repair requires excision of thymine dimers formed at the lesion followed by homologous DNA recombination and gap-filling replication [48,49].

Genetic interactions of non-SMC subunit mutants with SMC mutants
To understand the role of non-SMC subunits in the condensin complex, genetic interactions between non-SMC and SMC mutants were examined (Fig. 4E). All 7 mutants (3 of cnd1, 1 of cnd2, and 3 of cnd3) caused synthetic defects (SD, poor colony formation) with cut14-208 (S861P, mutated in the coiled coil). Two of them were synthetic defects involving cut3-477 (S1147P, mutated in the coiled coil); the other 5, as shown using tetrad analysis, failed to produce double mutants (designated "synthetic lethal SL"). None of them can produce double mutants when combined with cut14-Y1 (the mutation site of cut14-Y1 is L543S). These data strongly suggest that non-SMC subunits work in a cooperative or parallel fashion with SMC subunits to form the condensin complex and to execute its functions. Results are consistent with the finding that double mutants cnd2-A708V cut14-208 and cnd2-A708V cut3-477 hardly formed colonies at the semi-restrictive temperature (30°C) (Fig. 4E, right panel) (red arrowheads). Newly isolated cnd2-A708V located at the C-terminus is ts, but not damage-sensitive, in contrast to the previously isolated N-terminal mutation cnd2-1, which is sensitive to DNA damage [17].
Defective kinetochore segregation in the cnd3-L269P mutant, which is hypersensitive to a tubulin poison We found that one of the cnd3 Ts -Dsmutants, L269P, with a mutation in the N-terminal (6 th ) HEAT repeat, was sensitive to a tubulin poison (10 μg/ml TBZ) (Fig. 5A). Amino acid sequence comparison indicated that residue L269 (gray color residue in the right panel of Fig. 5B) is highly conserved among homologs from other organisms, such as fungi, flies, and humans. L269 is present in helix A [34] of the 6 th HEAT repeat so that the introduction of a helix-breaking proline residue instead of leucine into the helix should cause significant structural perturbation. TBZ sensitivity was found only for cnd3-L269P among 21 condensin mutants isolated.
We observed typical condensin mutant defects in chromosome segregation producing phi (φ)-shaped chromosomes in cnd3-L269P mutants at 36°C, the restrictive temperature, in the absence of TBZ (Fig. 5C). No defect was found at 26°C. On the other hand, cnd3-L269P mutants cultured at 26°C, the permissive temperature, for 4 h in the presence of TBZ (added to the medium to a final concentration of 10 μg/mL) produced severe segregation defects (Fig. 5D, +TBZ). Mutant cells stained with DAPI revealed large (indicated by light blue arrow) and small (light blue arrowhead) daughter nuclei in a high frequency (45/68 mitotic cells), which are a typical phenotype of centromere mutants, like mis6-302 and mis12-537 [50]. In this mutant strain, the cen2-GFP probe used previously [51] was introduced (described in the legend of Fig. 5D). Cen2-GFP visualization further showed a high frequency (7/20 mitotic cells) of cen2 mis-segregation (2:0 rather than 1:1), in which two sister centromere GFP dots went into UV irradiation. E. Synthetic genetic interactions between non-SMC ts mutants and SMC mutants (cut14-208, cut3-477 and cut14-Y1  The mutation site is shown in the structure of a HEAT repeat determined for human importin β [34]. In HEAT, Cnd3 L269 corresponds to the position of importin β L416 residue located in helix A of the HEAT unit. The change to a helix-breaking proline strongly perturbs the helical structure. C. Mutant cnd3-L269P cells show severe mitotic defects at 36°C in the absence of TBZ. The bar is 10 μm. D. In the presence of TBZ, defective chromosome segregation seen as unequal segregation (2:0) of cen2 dots instead of equal segregation (1:1) [51] was observed at 26°C. Large and small daughter nuclei were indicated by light blue arrows and arrowheads, respectively. The bar is 10 μm. Enlarged micrograph for cen2-GFP unequal separation, indicated by the red arrowhead. Dot-like intense cen2-GFP signals were obtained by association of LacI repressor-GFP protein with the LacO operon repeats present in the cen2 region [57]. LacI-GFP containing nuclear localization signal is preferentially localized in the nucleus when it is unbound to the centromere. E. Synthetic genetic interactions of cnd3-L269P with the same daughter cell (Fig. 5D, yellow arrowheads and a red arrowhead in the enlarged micrograph). In the absence of TBZ, large-small and 2:0 missegregation was hardly observed.
In addition to centromere mis-segregation, we found synthetic genetic interactions between cnd3-L269P and mis6-302 (a centromere/kinetochore mutant [50,52]) (Fig. 5E). Mis6 encodes a centromere protein essential for equal chromosome segregation. The synthetic lethal phenotype was also obtained with Δmad2 deletion mutant [53]. Mad2 is an essential component for the spindle assembly checkpoint, SAC [54,55]. The double mutant between cnd3-L269P and Δmad2 was lethal at both the permissive and semi-permissive temperatures (26-30°C). Note that the Δmtc1 deletion mutant implicated in mitotic spindle microtubule organization, was also additive with the defect caused by cnd3-L269P, as the double mutant hardly formed colonies at 26°and 30°C [40,56]. Together, these results strongly suggest that the role of Cnd3 L269 in the HEAT repeat is to assist or orchestrate centromere/kinetochore segregation and spindle assembly checkpoint (SAC) control. In addition, we found that mitotically arrested cold-sensitive mutants dis1-203 and dis2-11 were rescued when they were combined with the cnd3-L269P mutant at 22°C (Fig. 5F), indicating that the aberrant mitotic arrest in dis1-203 and dis2-11 probably due to activation of SAC, was compromised by mutation of cnd3-L269P.

Increased DNA content from a haploid to a diploid level in six non-SMC mutants
During the course of genetic analysis, we observed that spores from 6 of 21 strains did not germinate, when they were crossed with the wild-type haploid (h + leu1). Tetrads were analyzed: Four spores formed, but they were not viable. The six strains were cnd1-N914Y, -M951R, -G984R, cnd3-C101R, -S546R, and -D584G. Five were HEAT mutants. The remaining mutant, cnd1-N914Y, resides close to one HEAT (H4) repeat. None of the mutants involved proline substitution, suggesting that the effects of these mutations on protein conformation might not be severe. Four of five HEAT mutations changed from residues with small side chains to the bulky arginine, causing significant conformation perturbation. Two of six showed the ts kinetochore mutant, mis6-302, spindle checkpoint mutant, Δmad2, and XMAP215-like microtubuleinteracting protein mutant, Δmtc1/alp14. F. Cold sensitivity of dis1-203 and dis2-11 mutants was rescued when they were combined with cnd3-L269P mutation. See text. phenotype, while the majority (5/6) were damage-sensitive (either Ts -Dsor Ds -). The inviable spore phenotype was observed at 26°C for all six strains above.
We performed fluorescence-activated cell sorting (FACS) analysis (Experimental procedures). S. pombe wild-type and mutant cells were cultured and when they were growing Non-SMC mutants, cnd1-L685P and cnd2-A708V, are rescued by trichostatin A, a histone deacetylase inhibitor. A. cnd1-L685P and cnd2-A708V are ts and are unable to form colonies at 33°C, a semi-permissive temperature, but they form normal colonies in the presence of trichostatin A (TSA), an inhibitor of histone deacetylase. cnd3-L269P, on the other hand, was slightly hyper-sensitive to TSA at 26°C. The mutant apc10-666 is the control strain that was strongly rescued by TSA at 33°C [58]. B. Consistent with the results above, mitotic defects of cnd1-L685P (33°C in the absence of TSA for 5 hrs) were not observed in the culture at 33°C in the presence of 12.5μg/ml TSA for 5hrs (right panel). The bar is 10 μm. exponentially at 26°C, cells were harvested, and prepared for FACS analysis (Materials and Methods) (Fig. 7A). Unexpectedly, the amount of DNA in these six mutants was found to be basically equal to that of diploid cells (4C) (approximately 80% of growing S. pombe cells are in G2 phase). Because the majority (*80%) of these mutant cells contained single nuclei (Fig. 7B), similar to values for wild-type haploid (2C) and diploid (4C DNA) cells, binuclear cells were unlikely for these mutant strains.
Cells grown exponentially at 26°C were then stained with DAPI and observed using fluorescence microscopy. Cells were often elongated and cell width exceeded that of wild-type cells. Quantitative cell length and width data for 100 cells in each of three strains that increased DNA content are shown with wild-type haploid and diploid controls (Fig. 7C and D). DAPIstained micrographs of two strains are shown in Fig. 7E. To confirm that these mutant cells are definitively larger than the wild-type haploid cells, we employed the wild-type haploid expressing red RFP-tagged histone H2A, which was cultured independently and then mixed with cnd3-S546R cells that did not contain RFP. Only small cells showed the H2A-RFP positive nucleus (Fig. 7F). This result with the internal control firmly demonstrated that cnd3-S546R mutant cells were bigger than haploid wild type.
We found that the amount of DAPI-stained mutant nuclear chromatin was also greater than that of control wild-type cells, when cells were grown at 26°C. The average size of the nuclear chromatin region and size distributions for each of the three strains are summarized in the Table of Fig. 7G, top and bottom panels, respectively. The size was greater than that of wild-type haploid nuclear chromatin, even bigger than that of diploid nuclear chromatin. Three mutants revealed a volume of nuclear chromatin roughly 2.7-2.8-fold bigger than that of control wild-type haploid cells. In contrast, other cnd mutants cells with haploid DNA content were not big, and their cell sizes resembled those of wild-type haploids.

Discussion
Condensin plays a variety of roles in chromosome dynamics, but its genetic dissection has been hampered due to the lack of mutants that cause defects in its diverse and unrecognized functions. Previous mutant screening of S. pombe SMC and non-SMC mutants [17,[44][45][46] was restricted, as only cytological assays were employed to identify phenotypes with impaired chromosome condensation and segregation. In the present study, screening was targeted to obtain mutants in the three non-SMC subunit genes, assayed comprehensively (e.g., ts phenotype or drug-sensitivity assay). By such means, we were able to isolate and characterize 21 distinct, single-amino acid substitution mutants (8 cnd1, 1 cnd2 and 12 cnd3) for the three non-SMC subunit genes, the isolation procedures of which are described in detail in Fig. 2A. These mutants may cover diverse roles of condensin non-SMC subunits, and shed light on functions hitherto unknown for non-SMC proteins. The role of non-SMC subunits has been scarcely understood in any organism.
Sixteen mutants (5 cnd1, 1 cnd2 and 10 cnd3) produced the ts phenotype. At the restrictive temperature (36°C), these mutant cells failed in chromosome segregation and condensation in mitosis. The ts phenotype was indistinguishable from that of SMC ts mutants of S. pombe [12]. This supports the notion that severe functional loss of any of the three non-SMC subunits results in failure of holocondensin's essential mitotic function. In contrast, five mutants (3 cnd1, 2 cnd3) produced a novel, non-ts, non-mitotic phenotype, but are sensitive to HU, CPT and/or UV. Curiously, none of them is proline substituted with the exception of cnd1-L1037P. Four of the five non-ts mutations reside in HEAT repeats. Defects in this group of damage-sensitive mutations are apparently relatively mild so that mutant holocondensin may retain its mitotic function. Eleven mutations reside in HEAT repeats. Their locations are assembled in one repeat ( Fig. 8; b, e, f and g, for Cnd1, j, k, l, m, n, o and p for Cnd3) (see also Fig. 3). None of them is located in the turn. The majority (8/11) are found in helix B, suggesting that helix B might play a key role in Cnd1 and Cnd3 proteins. Indeed, five mutants with doubled DNA content were found in helix B (Fig. 8, magenta). The remaining cnd1-N914Y did not occur in the HEAT repeat, but was situated close to the amino-terminus of HEAT 4. Three mutations in helix A (cnd1-L1037P, cnd3-L126P, cnd3-L269P) are DNA damage-sensitive at the permissive temperature, but they are small haploid-like cells. They are all proline substitutions. One possible explanation is that helices A and B interact with different protein species. Alternatively, the roles of helix A and B may be distinct in determining the 3D conformation. It would be interesting to introduce the same mutations in higher eukaryotic homologs and to study the phenotype if it does not cause lethal phenotypes. Known mutations in condensin non-SMC genes either cause disease, such as human microcephaly (MCPH1, [20]; NCAPG2, [59], or bovine body size increase [21,22]. A striking mutant phenotype obtained in this study is that cells of six mutants (cnd1-N914Y, -M951R, -G984R, cnd3-C101R, -S546R and D584G) were larger than and contained twice the DNA of wild-type haploids (from 2C to 4C), suggesting that these mutant cells maintained a diploid-like state. DNA contents were measured for cells dividing normally at 26°C. When crossed with wild-type haploids, most spores formed were inviable, consistent with the notion that these mutant cells behave like diploid cells. Mutation sites of these mutants are either within or close to HEAT repeats. Interestingly, none of the mutations involved proline, and the majority involved arginine substitutions (4/6). We suppose that arginine replacement may not perturb protein conformation as much as proline substitution. Alternatively, in changing to a diploid-like state, a particular change in protein-protein interaction might be required in these mutants.
Our results are consistent with the notion that these mutant cells behave like diploid cells, (4C DNA in G2 phase). Meiosis between haploid wild type and these mutant cells was abnormal, resulting in barely viable spores. Massive aneuploidy seemed to have occurred due to the triploid meiosis between haploid and diploid [60]. Certain fission yeast genes are implicated in diploidization. The cdc2-tws1 mutation [60,61], cut8-563 [62], and myo2 [63] are known to cause diplodization. How condensin cnd1 and cnd3 mutations cause a diploid-like state remains to be investigated. Because meiosis was not involved in generating these mutations, defects in meiosis II are unlikely. Diploidization of cut8-563 occurs at the restrictive temperature (36°C), and 60-70% of the cells surviving after two generations are diploid. Cut8 is located in the nuclear envelope and tethers proper localization of 26S proteasome along the nuclear membrane [64,65]. The nuclear envelope architecture might be implicated in diploidization, since the volume of the nucleus should be related to maximal and minimal DNA content in cells. Non-SMC cnd1 and cnd3 mutations seemed to have more chromatin than wild-type diploid nuclei. These mutations might have caused the expansion of nuclear and chromosome scaffold (e.g., [66]), which might accompany the increase of DNA content. Very recently it was shown that nuclear architecture and nuclear size of human cells in interphase are disrupted by RNAi depletion of condensin SMC and non-SMC subunits [67]. The increased nuclear size is consistent with the present results.
Phenotypes produced by cnd3-L269P are of considerable interest regarding the link between Cnd3/Ycg1/CAP-G and mitotic regulation involving kinetochores/centromeres and spindle assembly checkpoint (SAC). S. pombe GFP-tagged condensin was previously shown to be localized at centromeres/kinetochores during mitosis [68]. The mitotic phenotype of cnd3-L269P mutants in the presence of TBZ showed large and small nuclei and unequal segregation of cen2-GFP, which are hallmarks of centromeric mutants. Cnd3 might directly contribute to the kinetochore function of condensin.
Mad2, located at mitotic kinetochores, is required for the spindle assembly checkpoint (SAC), and Δmad2 deletion mutants are sensitive to TBZ [53]. The strong synthetic phenotype of cnd3-L269P Δmad2 double mutants is consistent with the hypothesis that Cnd3 may directly or indirectly interact with Mad2 for proper arrangement of mitotic kinetochores. Mad2 was reported to physically interact with centromere protein Mis6 [69]. The additive interaction between cnd3-L269P and centromeric mutant mis6-302 is also consistent with the notion that Cnd3 plays an important role in the interaction between condensin and kinetochore. Dis1 and Mtc1/Alp14 (a homolog of human TOG, frog XMAP215 and budding yeast Stu2) implicated in proper chromosome segregation, interact with kinetochore microtubules and contain HEAT repeats [38,41,70]. We show in this study that the cold-sensitive phenotype of dis1-203 (similar to Mtc1/Alp14) and dis2-11 (type 1 protein phosphatase catalytic subunit) mutants is partly rescued by cnd3-L269P, while Δmtc1/alp14 mutant interacts negatively. Dis1 and Mtc1/Alp14 appear to function in an opposing way toward cnd3-L269P at the kinetochore during mitosis. To our knowledge, only one TBZ-sensitive condensin mutant was isolated recently [71]. Two types of condensin, I and II, exist in higher eukaryotes, while fungi, such as budding yeast and fission yeast contains only one condensin. Present and previous results show that fungal condensin does interact with centromeres/kinetochores perhaps through Cnd3/Ycg1/CAP-G, one of three non-SMC subunits.
In summary, we isolated 21 fission yeast mutants defective in condensin's three non-SMC subunits. Beside previously known mitotic functions in chromosome condensation and segregation and in DNA repair, we identified two novel condensin functions. First, HEAT repeatcontaining subunits, Cnd1 and Cnd3, are involved in maintaining the haploid genome state. Each of the three cnd1 and the three cnd3 mutations resided in or near the HEAT repeats caused an increase in cell size and doubled DNA content. It remains to be determined whether the size increase of nucleus is the primary event caused by these diploid HEAT-repeat mutations. These mutants normally proliferate, but do not form viable spores in meiotic segregation when crossed with wild-type haploids, probably due to triploid meiosis producing massive aneuploidy. Condensin thus seems to be required for maintaining the ploidy level via its HEAT-containing subunits. Secondly, we show that one cnd3-L269P mutant is defective in centromere/kinetochore function and spindle assembly checkpoint control. Consistently, this mutant, bearing an altered residue in the conserved HEAT repeat is hypersensitive to TBZ, a tubulin poison. These results together demonstrate novel aspects of the role of non-SMC, HEAT-containing subunits of condensin. The non-SMC condensin subunits Cnd1 and Cnd3 may play an unexpected role in cell growth and/or ploidy level stability.

Error-prone PCR for mutagenesis
Error-prone PCR was performed as described previously [47]. Briefly, the DNA fragment containing the cnd + gene, the hygR/kanR marker and a 500 bp downstream sequence was amplified from integration plasmids by PCR under error-prone conditions: The reaction buffer used contained 0.8 mM dNTP mixture (equimolar solution of dATP, dCTP, dGTP and dTTP), 8 mM Mg 2+ , 2 μM forward primer, 2 μM reverse primer, TaKaRa Ex Taq HS: 1.25 U/ 50 μL. PCR was repeated cycling of three steps for 30 cycles: denature DNA at 98°C for 10 sec, primer annealing at 50°C for 30 sec and extension at 72°C for 6 min.
Screening of temperature-sensitive and/or damage-sensitive mutants Amplified DNA fragments of cnd1, cnd2, and cnd3 prepared by error-prone PCRs were used for transformation of the S. pombe haploid wild-type strain 972 h -. Hygromycin (or G418 for cnd2)-resistant clones were picked up. Transformants were plated at 26°C, 36°C in the absence of DNA-damaging reagents, 26°C in 2mM hydroxyurea (HU), 26°C in 7.5μM camptothecin (CPT) and 26°C after 150J/m 2 ultraviolet irradiation (UV) for 3 days. Colony formation was compared with that of the control culture at 26°C in the absence of DNA damage. Tsand damage-sensitive (Ds -) mutants were selected. Mutation sites in isolated strains were determined by Sanger dideoxy sequencing of the genomic cnd regions after PCR amplification. Twentyone strains contained single, non-synonymous mutations. Subsequently, mutations were introduced into the wild-type strain by site-directed mutagenesis as described previously [75] and phenotypes produced were examined for reproducibility. Briefly, complementary pairs of oligonucleotide DNAs with mutations were used as primers, followed by two rounds of PCR. Chromosome integration into endogenous cnd + loci was followed, and the resulting integrant strains were tested for ts and/or damage sensitive phenotypes.

Microscopy
Each ts mutant strain was first grown in YPD medium at the permissive temperature (26°C), and then shifted to the restrictive temperature (36°C). DAPI staining was performed as described previously [77], and cells were observed with a BZ9000 microscope (Keyence, Japan). To visualize localization of centromere 2 (cen2), the cnd3-L269P mutant strain was crossed with the GFP-tagged cen2-GFP strain [51]. DAPI-stained nuclear chromatin and cen2-GFP centromere signals were observed with the BZ9000. Cell length and width were measured using the "Between 2-points" tool of the "Measurement module" in BZ-II Analyzer software (Keyence, Japan). Radii of nuclei were measured using the "Radius" tool in the software.

FACScan analysis
Procedures described [78,79] were followed to determine the average nuclear DNA content. Haploid and diploid wild-type cells were used as controls. Cells (1 × 10 7 ) were stained with propidium iodide (final concentration, 12.5 μg/ml), digested with RNase I, and analyzed with a FACScan (Beckton-Dickinson).  Table. Summary of initial-screening for S. pombe condensin non-SMC mutants. At Step 6 ( Fig. 2A), after streaking a large number of transformants and nucleotide sequencing of 130 strains, we obtained a total of 77 S. pombe chromosome integrant strains, which contained single or multiple mutations in the cnd1, cnd2, or cnd3 gene and displayed either Ts -, Ts -Ds -, or Dsphenotype. (TIF)