Identification of Saccharomyces cerevisiae Spindle Pole Body Remodeling Factors

The Saccharomyces cerevisiae centrosome or spindle pole body (SPB) is a dynamic structure that is remodeled in a cell cycle dependent manner. The SPB increases in size late in the cell cycle and during most cell cycle arrests and exchanges components during G1/S. We identified proteins involved in the remodeling process using a strain in which SPB remodeling is conditionally induced. This strain was engineered to express a modified SPB component, Spc110, which can be cleaved upon the induction of a protease. Using a synthetic genetic array analysis, we screened for genes required only when Spc110 cleavage is induced. Candidate SPB remodeling factors fell into several functional categories: mitotic regulators, microtubule motors, protein modification enzymes, and nuclear pore proteins. The involvement of candidate genes in SPB assembly was assessed in three ways: by identifying the presence of a synthetic growth defect when combined with an Spc110 assembly defective mutant, quantifying growth of SPBs during metaphase arrest, and comparing distribution of SPB size during asynchronous growth. These secondary screens identified four genes required for SPB remodeling: NUP60, POM152, and NCS2 are required for SPB growth during a mitotic cell cycle arrest, and UBC4 is required to maintain SPB size during the cell cycle. These findings implicate the nuclear pore, urmylation, and ubiquitination in SPB remodeling and represent novel functions for these genes.


Introduction
The centrosome is the dominant microtubule-organizing center in mammalian cells and is important for chromosome segregation. Centrosomes facilitate the organization of microtubules during interphase, as well as organizing the bipolar spindle during cell division. Each daughter cell must receive only a single centrosome, and duplication must occur only once during the cell cycle to ensure bipolarity. Centrosome abnormalities, including increased number, size, and microtubule nucleation capacity, are a hallmark of many cancer types, and severity of these defects increases during tumor progression [1]. Although multipolarity is often a consequence of centrosome abnormalities in cancer cells, several studies have shown that the amplified centrosomes coalesce and form a bipolar spindle [1,2,3]. This has also been demonstrated in normal cells forced to have a double complement of DNA and centrosomes: retinal pigmented epithelial (RPE1) cells treated with a cytokinesis inhibitor are able to cluster the centrosomes to form a bipolar spindle and proceed through the cell cycle [4]. Minus-enddirected microtubule motor proteins are involved in this clustering process: inhibition of dynein in fibroblasts leads to disassociation of clustered centrosomes [2] and Drosophila kinesin 14 motor protein Ncd is required for focusing of spindle poles [5] and maintaining spindle bipolarity when centrosome amplification is induced [6]. These data demonstrate a cellular response pathway for repairing centrosome and spindle assembly defects.
The spindle pole body (SPB) is the functional equivalent of the mammalian centrosome in Saccharomyces cerevisiae and organizes microtubules for chromosome segregation in mitosis and meiosis. The SPB is not a static structure. Instead, the SPB is remodeled in two ways: by growth, in which new components are added, and by exchange, in which old components are replaced by new components. These changes are cell cycle dependent, with growth occurring late in the cell cycle, and exchange occurring around the time of SPB duplication leading to the parent SPB having a mix of old and new components. Cell cycle arrests have various effects on these remodeling phenotypes. When arrested in G1 with a-factor, the SPB core becomes smaller. Conversely, when cells are arrested at metaphase, the SPB core grows. For example, overexpression of Mps1 kinase, which activates the spindle assembly checkpoint, causes SPBs to double in size. Based on the fact that the SPB is remodeled at discrete times during the cell cycle and in response to checkpoint activation, this process is likely to be important for maintenance of the SPB and possibly for assembly of the spindle. [7] SPB remodeling was observed by tagging the integral SPB component Spc110 with fluorophores and using quantitative fluorescence to determine the level of incorporation or exchange of labeled protein [7]. Determination of the amount of Spc110 using this method is a good measure of the overall SPB core size: comparison of Spc110-YFP fluorescence in tetraploid strains with one to four copies of SPC110::YFP showed that SPB fluorescence is proportional to the amount of Spc110-YFP at the pole [7], Spc110 links c-tubulin to the SPB core and consists of globular domains connected by a long coiled-coil region. Three functional domains have been identified within the protein through mutational analysis for temperature sensitive mutants [8]. The best characterized are spc110-220, spc110-221, and spc110-226, and each of these mutants is defective in one of Spc1109s functions. At the nonpermissive temperature, spc110-221 mutants arrest the cell cycle due to a defect in Spc110-221 attachment to the c-tubulin complex [8]. Mutant spc110-226 cells also lose viability at the nonpermissive temperature due to a weakened connection, in this case between Spc110-226 and the SPB core [9]. Mutant spc110-220 contains several point mutations in the calmodulin-binding domain. At the non-permissive temperature, Spc110-220 is not assembled into the pole efficiently due to defective binding of calmodulin [10].
One protein that has been previously shown to affect assembly of SPB components is Mlp2, a nuclear pore-associated protein that binds to SPB core components and affects their assembly into the SPB [11]. Deletion of MLP2 leads to formation of smaller SPBs, and combining Mlp2 depletion with spc110-220 exacerbates the assembly defect and is lethal. These data make Mlp2 a likely SPB remodeling factor and implicate nuclear pore proteins in SPB assembly and remodeling.
To identify additional proteins involved in the remodeling process, we developed a system for conditionally inducing SPB remodeling. The remodeling strain contains a version of Spc110 that can be cleaved by TEV protease. Using a synthetic genetic array analysis, we screened for genes required only when cleavage of Spc110 is induced. We hypothesized that remodeling could alleviate SPB damage in these strains either by growth, which could add new Spc110, or by exchange, which could replace damaged Spc110 with functional Spc110. Secondary screens identified four genes required for SPB remodeling. UBC4 is required to maintain SPB size during the cell cycle, and NCS2, POM152, and NUP60 are required for SPB growth during a mitotic cell cycle arrest.

Spc110 cleavage causes a Mad1p/Mad2p-dependent cell cycle delay
Yeast strains with galactose-inducible Spc110 cleavage were constructed and their phenotypes examined. Galactose-inducible TEV protease was introduced into the genome, along with a tandem array of three TEV cleavage sites in the coiled-coil region of Spc110. Upon induction with galactose, TEV protease is produced ( Figure 1A) and cleavage of the majority of Spc110 occurs ( Figure 1B). Spc110 cleavage strains form colonies on plates under induced conditions ( Figure 1C), indicating that the cells are able to assemble functional SPBs even after Spc110 cleavage. These strains have a longer cell cycle than wild type when Spc110 cleavage is induced: the doubling time for KGY54 (SPC110-3xTEV696) is 91 minutes while the doubling time for KGY57 (SPC110-3xTEV696, GAL-TEV) is 115 minutes. The spindle checkpoint is required to maintain viability of Spc110 cleavage strains under induced conditions ( Figure 1D), indicating that a cell cycle delay is necessary to allow for correct spindle assembly in the presence of Spc110 cleavage.

A genetic screen identifies candidate SPB remodeling factors
To identify additional genes required to survive Spc110 cleavage, we performed a synthetic genetic array (SGA) screen for gene deletions that show a growth defect when combined with Spc110 cleavage. The screen was performed in triplicate and the compiled results are shown in Table S1. Top candidate genes fell into a small number of functional categories: microtubule motors, protein modification enzymes, nuclear pore components, mitotic spindle checkpoint/cell cycle regulators, chromatin remodeling factors, and regulators of mRNA levels ( Table 1). Identification of spindle checkpoint genes independently shown to be synthetic lethal with SPB damage (Figure 1D), as well as identification of Spc110 transcriptional regulator HCM1, demonstrate the effectiveness of the SGA screen. A subset of candidate genes was chosen for further study and their genetic interaction with Spc110 cleavage was verified in the W303 background to eliminate false positives ( Table 1). Deletion of nuclear pore component gene MLP2, which was previously shown to have a role in SPB component assembly [11] but was not identified in the SGA screen, did not cause a synthetic growth defect with Spc110 cleavage.
In addition to systematic false positives, the SGA screen might have identified genes that alter the level of Spc110 or TEV protease. Reduced Spc110 or increased TEV protease could exacerbate the Spc110 cleavage phenotype and cause the cells to die. Spc110 levels were quantified by Western blotting in strains containing a single gene deletion from the subset of candidate genes mentioned above (data not shown). Only sac3D, kar3D and lsm7D caused reduced Spc110 levels, which might account for their appearance in the SGA screen data. TEV protease levels were also quantified by Western blotting in strains containing a single gene deletion and the GAL-TEV gene (data not shown). The only gene deletion mutant that had a significant increase in TEV protease production was ubc4D.
Deletion mutant crosses to spc110 temperature sensitive mutants identify candidate SPB remodeling factors with a specific defect in SPB assembly Deletion of genes involved in assembly should show allele specific defects with SPC110 mutant spc110-220, which was previously determined to be defective in assembly [8]. Each of the candidate gene deletion mutants were mated to a strain carrying spc110-220 and to strains carrying either of two alleles not involved in assembly: spc110-221 and spc110-226 (Figures S1-S16). A summary of the growth phenotypes of these double mutants is compiled in Table 2. Deletion mutants that had the strongest synthetic growth defect in combination with spc110-220 are cik1D, jnm1D, ncs2D, ppm1D, ubc4D, pom152D, and nup60D. The allele specific defect with spc110-220 that was previously shown for mutant mlp2D [11] was also verified. These candidates are likely to have a role in SPB component assembly based on their specific defect in combination with an assembly mutant.

GAL-MPS1 metaphase arrest identifies candidate SPB remodeling factors with a defect in SPB growth
Six deletion mutants that had a specific growth defect in combination with the spc110-220 assembly mutant were further characterized for an SPB remodeling phenotype: jnm1D, ncs2D, ppm1D, ubc4D, pom152D, and nup60D. The seventh mutant strain, cik1D, grows very poorly and was not characterized further. Candidate SPB remodeling gene deletion strains with SPC110-GFP and galactose-inducible MPS1 were imaged during normal asynchronous growth and during a GAL-MPS1 metaphase arrest (.90% large buds). In wild type cells, arresting the cell cycle at metaphase using a GAL-MPS1 arrest causes SPBs to double in size [7]. We therefore quantified the amount of SPB fluorescence and examined the distribution of fluorescence for each strain to identify mutants with an impaired ability to increase SPB size ( Figure 2). Cells with ncs2D, pom152D, or nup60D showed a significant defect in SPB growth during metaphase arrest compared to wild type cells (pvalue ,1610 25 using a Kolmogorov-Smirnov test to compare mutant distributions to wild type in three replicates).
Deletion of UBC4, but not UBC5, causes a defect in SPB size regulation Ubc4 and Ubc5 are ubiquitin-conjugating E2 enzymes that have 77% sequence homology and can functionally complement one another [12]. UBC4, but not UBC5, was identified in the SGA screen, and the ubc5D mutant shows no growth defect when combined with Spc110 cleavage ( Figure 3A). Combining spc110-220 with ubc5D resulted in a milder growth phenotype than with ubc4D ( Figure 3B), and SPBs in the ubc4D mutant, but not ubc5D, have a wider fluorescence distribution than wild type during asynchronous growth with a greater number of large SPBs ( Figure 3C). The percentage of large SPBs (Spc110::GFP fluorescence value greater than the wild type mean plus one standard deviation) in asynchronously growing wild type cells is 16.1%60.2% while the percentage of large SPBs in ubc4D cells is 47.2%68.7%. Large SPBs in ubc5D cells make up 13.6%61.0% of the population, which is similar to the percentage found in wild type SPB populations. A comparable increase in the percentage of large SPBs in ubc4D cells was seen when SPB core component Spc42 was tagged with GFP (data not shown). The altered distribution of SPB fluorescence seen in ubc4D cells does not result from an increase in the number of sideby-side SPBs or collapsed spindles because a similar size distribution is seen when the subset of SPBs that are in metaphase pairs is plotted ( Figure 3C). Moreover, the ubc4D mutant does not have an impaired ability to remove core components in G1. Arresting wild-type cells in G1 with a-factor causes a 40% decrease in SPB size [7] and leads to a shift in the peak of fluorescence distribution to lower values  ( Figure 3C). The ubc4D mutant shows a shift in peak fluorescence after the arrest that is comparable to wild type ( Figure 3C). Asynchronously growing wild type, ubc4D, and ubc5D cells were also imaged and the amount of Spc110::GFP or Spc42::GFP was quantified for each strain (Table 3). Mutant ubc4D SPBs contained high amounts of Spc42::GFP and Spc110::GFP compared to wild type and ubc5D, which is consistent with ubc4D cells having a defect in SPB size regulation.

Discussion
SPB remodeling has been shown to occur at discrete times during the cell cycle. However, very little is known about the process of remodeling and the proteins involved in regulating and facilitating SPB growth and component exchange. In this study, we have identified several candidates for involvement in the SPB remodeling process. Proteins identified include microtubule motors, protein modification enzymes, and nuclear pore proteins.
Many of the yeast microtubule motors were identified and had a synthetic growth defect with Spc110 cleavage. This information, coupled with previous studies on motors in other organisms, suggests that motors play a role in assembly of the spindle and specific SPB components. Dynein and Ncd (the Kar3p homolog) have previously been shown to move microtubule bundles to the centrosome in Drosophila [5], and dynein has also been shown to transport pericentrin (the Spc110p homolog) and c-tubulin to the centrosome in mammalian cells [13]. While deletion of the motors identified in our study did not lead to a defect in SPB growth during metaphase arrest, further characterization of these proteins and their role in SPB remodeling could shed light on the process of spindle assembly.
Our secondary screens identified four proteins that regulate SPB size: Ncs2, Nup60, Pom152, and Ubc4. Loss of Ncs2, Nup60, or Pom152 led to an impaired ability to increase SPB size during metaphase arrest, implicating these proteins in SPB component assembly. Ncs2 is involved in the ubiquitin-related modifier Urm1 pathway and is necessary for thiolation of Lys(UUU) and Glu(UUC) tRNAs [14,15]. Ncs2 has no known association with SPB proteins. However, our results suggest involvement of the urmylation pathway in regulation of SPB size. Components of the urmylation pathway have been previously shown to have genetic interactions with nuclear pore component NUP133: deletion of URM1 or UBA4 (Urm1 activator) leads to a synthetic growth defect when combined with NUP133 deletion [16]. Furthermore, deletion of NUP133 is synthetic lethal with deletion of another nuclear pore component gene, NUP60 [17]. We have shown that Nup60 and Pom152 are necessary for surviving Spc110 cleavage and for SPB growth during metaphase arrest. The only protein previously described as having a role in assembly of SPB components is nuclear pore protein Mlp2. We found that mlp2D does not have a synthetic growth defect when combined with Spc110 cleavage and therefore was not found in our SGA screen. However, attachment of Mlp2 to the nuclear pore is mediated by Nup60 [18]. Pom152 has previously been shown to form a complex with Ncd1 and Pom34 [19], which assembles to form a ring around the nuclear membrane structure of the pore [20]. Additionally, deletion of POM34 or POM152 disrupted the function of essential SPB duplication regulator, Mps2 [21]. These data, combined with our data on Ncs2, Nup60, and Pom152, further implicate the nuclear pore in proper assembly of the SPB and suggest that the urmylation pathway may act in conjunction with nuclear pore components to regulate SPB size. Table 2. The highest temperature that supports normal growth (uC) for haploid progeny of crosses between SPB remodeling candidate gene deletion strains and spc110 mutants.
Deletions whose strongest effect is on spc110 assembly mutant spc110-220  Figure 2. GAL-MPS1 metaphase arrest identifies deletion mutants with a defect in SPB growth. Deletion mutants containing SPC110-GFP and GAL-MPS1 were grown overnight on YP RAFF plates at room temperature, then struck out onto YP RAFF/GAL plates (time zero). Samples were taken at time zero and fixed for 15 minutes in 3.7% formaldehyde at 30u in a roller drum. Plates were then incubated at room temperature for six hours to induce mitotic arrest (.90% large buds) and samples were taken and fixed in formaldehyde as above for imaging. Fluorescence was quantified for in-focus SPBs and then normalized using the photosensor value of the microscope. The distribution of GAL-MPS1 arrested SPB fluorescence in each mutant strain was compared to the wild type distribution using a Kolmogorov-Smirnov test. ncs2D, pom152D, and nup60D arrested distributions were consistently different from wild type (p-value ,1610 ' -5). Histograms of SPB fluorescence values from asynchronously growing (time zero) and metaphase-arrested (six hour time point) yeast from a representative experiment are shown for these strains. The data was normalized by population size and the best fit Gaussian curve was fit to each asynchronous distribution. Histogram bins that fell below the signal to noise cutoff chosen during image analysis were excluded when fitting the Gaussian curves. The wild type asynchronous best fit curve is overlaid on the mutant asynchronous distributions in red for comparison. The wild type arrested SPB fluorescence distribution is overlaid onto the arrested mutant distribution histograms in light grey for comparison. doi:10.1371/journal.pone.0015426.g002 Our screen also identified Ubc4, the ubiquitin-conjugating enzyme (E2). In mammalian cells, tumor suppressor BRCA1 uses a Ubc4 homolog as one of its ubiquitin E2 ligases for conjugating ubiquitin to target proteins [22]. It has also been shown that BRCA1-dependent ubiquitination is important in regulating centrosome number [23], and centrosome amplification is a hallmark of cancer. Our results show that Ubc4 but not its close relative, Ubc5, regulates the size of SPBs during asynchronous growth. Deletion of UBC4 leads to disruption of SPB size regulation as indicated by increased levels of Spc110 and Spc42 in the poles. Ubiquitination of target proteins by Ubc4 could regulate SPB size by altering levels of SPB proteins or by affecting their incorporation into the pole, thereby changing the nucleation capacity of the SPB. BRCA1 regulates centrosome nucleation activity through ubiquitination of c-tubulin and a centrosome adaptor component [24], and our data implicate Ubc4 in a conserved centrosome regulation pathway in yeast.

Materials and Methods
Media YPD and SD media was prepared as previously described [25]. YP raffinose plates (YP RAFF) contain 2% raffinose, YP galactose plates and liquid media (YP GAL) contain 2% galactose, and YP raffinose/galactose (YP RAFF/GAL) plates contain 2% raffinose and 2% galactose. YPD NAT plates were made by spreading 30 ml of 10 mg/ml clonNAT (Werner BioAgents, Jena, Germany) solution onto YPD plates. YPD HYG plates are standard YPD supplemented with 0.6 mg/ml hygromycin B. Media used in the SGA screens were made as previously described [26] except where tailored to fit the Spc110 cleavage strains as noted below.

Plasmids
Plasmids used in this study are listed in Table S2.
SPC110 three TEV protease cleavage site tandem array (SPC110-3xTEV696) plasmids. QuickChange Site Directed Mutagenesis (Stratagene, La Jolla, CA) was performed on pHS31 [8] to create a BamH1 site at SPC110 base pair 2085 (corresponding to amino acid 696), resulting in plasmid pKG2. DNA oligos with sequence for three TEV cleavage sites and flanking BamH1 sites were constructed. Sense and missense oligos (59-GATCCGAAAATTTATATTTTCAAGGTGAAAATTTAT-ATTTTCAAGGTGAAAATTTATATTTTCAAG-39 and 59-G-ATCCTTGAAAATATAAATTTTCACCTTGAAAATATAAA-TTTTCACCTTGAAAATATAAATTTTCG-39) were annealed and ligated to a BamH1 digest of pKG2 to create pKG7. Nco1 and Sac1 digest of pAG25 [27] yielded a fragment containing the nourseothricin (NAT1) resistance cassette (natMX4), and this fragment was ligated to the Nco1, Sac1 large fragment of pFA6a-3HA-kanMX6 plasmid [28] to create pKG9, a plasmid for tagging genes with an HA tag and NAT1 selectable marker. The HA-natMX4 cassette was amplified from pKG9 and integrated into pKG7 to create a plasmid containing SPC110-3xTEV696 with an HA tag and a NAT1 selectable marker (pKG16). This plasmid was then converted to an integrating plasmid by ligating the pRS306 [29] large AlwN1 fragment to the pKG16 large AlwN1 fragment, resulting in the pKG17 plasmid used for Spc110 cleavage strain construction.
GAL-TEV was amplified by PCR from plasmid 118 (gift from Frank Uhlmann) and ligated into pCR Blunt II-TOPO (Invitrogen Corporation, Carlsbad, CA) to create pKG10. pRS306 was then digested with Not1 and Xho1 and ligated to the small Not1, Xho1 fragment of pKG10 to form pKG12, an integrating plasmid containing GAL-TEV with a URA3 marker.
GAL2 plasmid. Wild type GAL2 was PCR amplified from HSY2-12C [30] and ligated into pCR Blunt II-TOPO to create pKG11. pRS306 was digested with Not1 and Sac1 and ligated to the small Not1, Sac1 fragment of pKG11 to create pKG13, an integrating plasmid containing GAL2 with a URA3 marker. QuickChange Site Directed Mutagenesis was performed on pRS315 [29] to create a BsrG1 site for subsequent removal of the LEU2 gene. The BsrG1 fragment of the resulting pKG14 plasmid was then ligated into the BsrG1 site in pKG13, resulting in plasmid pKG15. The BsrG1 site is located in the genomic sequence directly downstream of GAL2. pKG15 was checked by restriction digest to ensure the proper orientation.

Strains
Strains used in this study are listed in Table S3. Spc110 cleavage strain (W303 background). The SPC110-3xTEV696 fragment was amplified from pKG17 and transformed into GZY7-5B (gift from Gefeng Zhu). Selecting for integrants on YPD NAT plates resulted in strain KGY54, and correct integration of the TEV cleavage sites was confirmed by sequencing. This strain has three TEV protease cleavage sites at amino acid 696 in the only copy of SPC110. GAL-TEV containing plasmid pKG12 was digested with Nco1 for integration at the URA3 locus and stable integrants were selected for on SD -ura dropout media. The resulting Spc110 cleavage strain is KGY57.  Figure 3. Deletion of UBC4, but not UBC5, causes a defect in SPB size regulation. (A) ubc4D, but not ubc5D, has a synthetic growth defect when combined with Spc110 cleavage. Strains were grown on YP RAFF/GAL plates at room temperature. (B) ubc4D has a more pronounced growth defect when combined with assembly mutant spc110-220 than ubc5D. Strains were grown on YPD plates at 25u and 37u. (C) Asynchronously growing ubc4D, but not ubc5D, cells show a defect in SPB size regulation. Strains were grown overnight at 23u in YPD liquid media to a cell density of 25 Klett units. a-factor was added to a concentration of 7.56 mg/ml at time zero and samples were taken and fixed in formaldehyde for imaging of the asynchronous cultures. Strains were then incubated at 23u for 3.5 hours (1.5 generations) and samples were taken from the G1 arrested cells and fixed in formaldehyde for imaging. SPB fluorescence of asynchronously growing and G1 arrested cells was measured and plotted as described. Best fit Gaussian curves were fit to each distribution and the wild type fit is overlaid in red on the ubc4D and ubc5D histograms for comparison. SPBs in the asynchronous population that were part of metaphase pairs were isolated and their SPB fluorescence values were also plotted. The distribution of SPB fluorescence after a-factor arrest shows a similar shift in wild type, ubc4D, and ubc5D cells, indicating that the mutant strains are able to reduce SPB size in a-factor as well as wild type. doi:10.1371/journal.pone.0015426.g003 Spc110 cleavage strain for synthetic genetic array (SGA) screening (S288C background). Strain Y7029 was transformed with the Not1, Sac1 fragment of pKG15, which contains wild type GAL2. GAL2 is a plasma membrane galactose permease that is defective in S288C and might be necessary for full activation of galactose-inducible promoters [31]. The resulting strain, KGY39, was then transformed with the SPC110-3xTEV696 cassette as described in the Spc110 cleavage strain (W303 background) section above to make strain KGY53, and correct integration of the cleavage sites was confirmed by sequencing. The GAL-TEV::URA3 fragment was amplified from pKG12 and integrated in a region near the URA3 locus between TIM9 and RPR1 because URA3 and its flanking sequence are deleted in Y7029. Stable integrants were isolated by selection on SD -ura plates resulting in the strain KGY61, which was used in the SGA screens. Gene deletion strains. Gene deletions were made by first PCR amplifying a hygromycin B cassette (hphMX4) from pAG32 [27] using primers that had ends homologous to the flanking DNA of each gene to be deleted. The cassette was then transformed into the diploid strain BSY9 [32] or KGY315. The resulting transformants were dissected and scored for growth on YPD HYG plates. The deletions were checked by PCR to ensure replacement of the target gene with the cassette.

Fluorescence microscopy and image analysis
Fluorescently labeled strains were mounted on a 1% agarose in S media pad and SPBs were imaged using a DeltaVision Core Restoration Microscopy System (Applied Precision, Issaquah, WA) that incorporates an Olympus U-plan Apo 100X oil objective (NA, 1.35). GFP filter sets (ex. 470/40, em. 525/50) were from Chroma Technology. Images were captured using a Photometrics Coolsnap HQ camera (Photometrics, Pleasanton, CA) and analyzed using the Fluorcal software program [34] to identify SPBs that were in focus in a single focal plane. Fluorescence intensity was calculated by determining the integrated intensity in a 565 pixel square around each SPB and subtracting background fluorescence.

Analysis of SPB fluorescence data
Histograms of SPB fluorescence intensity values were constructed and Gaussian curves were fit to the data using Igor Pro version 6.12 (WaveMetrics, Portland, OR). Kolmogorov-Smirnov tests were performed using the KS version 2.0 for NeuroMatic version 2.00 Igor Pro procedure.  Figure S1 spc110 mutants crossed with cik1. Haploids with the genotypes indicated on the left were grown on YPD plates and incubated at the temperatures indicated. (TIF) Figure S2 spc110 mutants crossed with kar3. Haploids with the genotypes indicated on the left were grown on YPD plates and incubated at the temperatures indicated. (TIF) Figure S3 spc110 mutants crossed with vik1. Haploids with the genotypes indicated on the left were grown on YPD plates and incubated at the temperatures indicated. (TIF) Figure S4 spc110 mutants crossed with jnm1. Haploids with the genotypes indicated on the left were grown on YPD plates and incubated at the temperatures indicated. (TIF) Figure S5 spc110 mutants crossed with dyn3. Haploids with the genotypes indicated on the left were grown on YPD plates and incubated at the temperatures indicated. (TIF) Figure S6 spc110 mutants crossed with rts1. Haploids with the genotypes indicated on the left were grown on YPD plates and incubated at the temperatures indicated. (TIF) Figure S7 spc110 mutants crossed with ncs2. Haploids with the genotypes indicated on the left were grown on YPD plates and incubated at the temperatures indicated. (TIF) Figure S8 spc110 mutants crossed with ppm1. Haploids with the genotypes indicated on the left were grown on YPD plates and incubated at the temperatures indicated. (TIF) Figure S9 spc110 mutants crossed with ubc4. Haploids with the genotypes indicated on the left were grown on YPD plates and incubated at the temperatures indicated. (TIF) Figure S10 spc110 mutants crossed with sac3. Haploids with the genotypes indicated on the left were grown on YPD plates and incubated at the temperatures indicated. (TIF) Figure S11 spc110 mutants crossed with pom152. Haploids with the genotypes indicated on the left were grown on YPD plates and incubated at the temperatures indicated. (TIF) Figure S12 spc110 mutants crossed with nup60. Haploids with the genotypes indicated on the left were grown on YPD plates and incubated at the temperatures indicated. (TIF) Figure S13 spc110 mutants crossed with mlp2. Haploids with the genotypes indicated on the left were grown on YPD plates and incubated at the temperatures indicated. (TIF) Figure S14 spc110 mutants crossed with dot1. Haploids with the genotypes indicated on the left were grown on YPD plates and incubated at the temperatures indicated. (TIF) Figure S15 spc110 mutants crossed with lsm7. Haploids with the genotypes indicated on the left were grown on YPD plates and incubated at the temperatures indicated. (TIF) Figure S16