Logo of geneticsGeneticsCurrent IssueInformation for AuthorsEditorial BoardSubscribeSubmit a Manuscript
Genetics. 2006 Jun; 173(2): 635–646.
PMCID: PMC1526528

Consequences of Defective Tubulin Folding on Heterodimer Levels, Mitosis and Spindle Morphology in Saccharomyces cerevisiae


In budding yeast, the essential roles of microtubules include segregating chromosomes and positioning the nucleus during mitosis. Defects in these functions can lead to aneuploidy and cell death. To ensure proper mitotic spindle and cytoplasmic microtubule formation, the cell must maintain appropriate stoichiometries of α- and β-tubulin, the basic subunits of microtubules. The experiments described here investigate the minimal levels of tubulin heterodimers needed for mitotic function. We have found a triple-mutant strain, pac10Δ plp1Δ yap4Δ, which has only 20% of wild-type tubulin heterodimer levels due to synthesis and folding defects. The anaphase spindles in these cells are ∼64% the length of wild-type spindles. The mutant cells are viable and accurately segregate chromosomes in mitosis, but they do have specific defects in mitosis such as abnormal nuclear positioning. The results establish that cells with 20% of wild-type levels of tubulin heterodimers can perform essential cellular functions with a short spindle, but require higher tubulin heterodimer concentrations to attain normal spindle length and prevent mitotic defects.

IN eukaryotic cells, microtubules play important roles in intracellular traffic, morphological differentiation, and cell division. In budding yeast, nuclear and cytoplasmic microtubules are specifically required for chromosome segregation and nuclear positioning during mitosis. Defects in these mitotic functions can lead to aneuploidy and cell death. Regulatory mechanisms that help to ensure proper function and formation occur at multiple steps in microtubule morphogenesis, both before and after a microtubule is formed. The work described here analyzes the consequences of mutations in proteins involved in the steps before microtubule morphogenesis, specifically those that fold α- and β-tubulin polypeptides into functional proteins. Disruption of these genes that are involved in tubulin folding affects the proportion of the proteins capable of forming heterodimer and incorporporating into a microtubule. This work demonstrates that with combined-folding mutations leading to a greatly reduced amount of tubulin heterodimer present (20% that of wild type), the cell forms a shortened spindle that is still capable of segregating chromosomes properly.

The α- and β-tubulin polypeptides must be properly folded to form heterodimers. In budding yeast, the two α-tubulin proteins, Tub1p and Tub3p, and the single β-tubulin protein, Tub2p, are folded by the cytosolic chaperonin, TriC (Ursic and Culbertson 1991; Dunn et al. 2001). In addition, a nonessential complex, genes involved in microtubule biogenesis (GimC)/prefoldin (PFD), acts in conjunction with the chaperonin TriC. The GimC/PFD complex is thought to prevent premature release of unfolded polypeptides and to deliver released but unfolded polypeptides to the chaperonin (Geissler et al. 1998; Vainberg et al. 1998; Leroux et al. 1999). Virtually all of the tubulin polypeptides in wild-type yeast cells under normal conditions are in heterodimers (Archer et al. 1998).

Free β-tubulin is toxic in both Saccharomyces cerevisiae and Schizosaccharomyces pombe (Burke et al. 1989; Weinstein and Solomon 1990; Javerzat et al. 1996). In S. cerevisiae, overexpression that doubles the wild-type β-tubulin levels interferes with microtubule assembly and is lethal, but even a modest excess (20% of the wild-type levels) disrupts microtubule function. Free β-tubulin can arise from direct perturbations of tubulin gene expression, such as deletion of the minor α-tubulin gene TUB3 (Schatz et al. 1986). Surprisingly, α- and β-tubulin expression is also differentially affected by defects in the tubulin-folding pathway. Mutations in the genes encoding components of the GimC/prefoldin complex downregulate tubulin protein levels asymmetrically and so produce a modest excess of β-tubulin and microtubule phenotypes such as sensitivity to the microtubule poison benomyl (Alvarez et al. 1998; Geissler et al. 1998). In this work, we mutate a GimC/prefoldin component PAC10 (also known as GIM2 or PFD3) to cripple the tubulin folding pathway and produce less folded and functional tubulin.

Recent work demonstrates an early step in the folding of β-tubulin that is important for the formation of β-tubulin competent to form tubulin heterodimer with α-tubulin or, if not in heterodimer, to become toxic to the cell (Lacefield and Solomon 2003). That step is defined by the gene PLP1; plp1Δ suppresses the toxicity of β-tubulin not by altering levels of tubulin but by affecting the formation of β-tubulin that is functional and toxic. Thus, plp1Δ substantially suppresses the microtubule phenotypes of both tub3Δ and GimC/prefoldin mutant cells and even partially suppresses the lethality of high-level overexpression of β-tubulin. Deletion of PLP1 in otherwise wild-type cells produces no detectable microtubule or growth phenotypes. However, only ∼65% of the α- and β-tubulin in plp1Δ cells is in heterodimer (Lacefield and Solomon 2003), a level sufficient to support normal growth (Katz et al. 1990). On the basis of these phenotypes, the step in tubulin folding defined by Plp1p is distinct from the function of the cytosolic chaperonin TriC, the GimC/PFD complex, or the putative tubulin heterodimerization cofactors (Lacefield and Solomon 2003).

The experiments described here use mutations affecting tubulin folding, pac10Δ and plp1Δ, to reduce the levels of tubulin heterodimers in the cell. A strain carrying these two mutations is deficient for β-tubulin and compensates for this deficiency by becoming aneuploid to increase the copy number of the β-tubulin gene. Increasing β-tubulin copy number could normally not be tolerated due to the toxic effects of excess β-tubulin (Burke et al. 1989; Weinstein and Solomon 1990). We found that this strain with a third mutation, of a putative transcription factor YAP4 (also named CIN5), has very low tubulin heterodimer levels and does not have an extra copy of β-tubulin. The results define the minimum level of tubulin needed to form a normal spindle. The results also demonstrate the cells' responses to a range of tubulin levels.


Strains, media, and cell morphology:

All yeast strains (Table 7) are derivatives of FSY182, FSY183, and FSY184 (Weinstein and Solomon 1990). We used standard yeast manipulation methods and media (Sherman et al. 1986; Guthrie and Fink 1991; Solomon et al. 1992). Budding morphology was determined by counting the number of large-budded, medium-budded, small-budded, and unbudded cells in a log phase asynchronous culture. Over 600 cells were counted for each strain. Nuclear positioning was analyzed as described previously (Schatz et al. 1988). Over 200 nuclear positions in large-budded cells were counted for each strain. For growth curves, cultures were initiated at 0.25 × 107 cells/ml and then grown in YPD at 30° and counted every 2 hr for 12 hr total. Cold sensitivity was measured by plating 1:5 serial dilutions of a saturated yeast culture onto plates that were then incubated at 30° for 3 days and 15° for 6 days. Quantitative cold sensitivity was measured by plating 300 colonies per plate and counting the number of cells that grew at 30° and 15° after 10 days.

Strain list

Gene disruptions:

YAP4 was deleted using a PCR-based method (Longtine et al. 1998). Deletions of PAC10 (Alvarez et al. 1998) and PLP1 were previously described (Lacefield and Solomon 2003). The pac10Δ plp1Δ double mutant was made in two ways: by knocking out PLP1 in a pac10Δ mutant with a plasmid containing PAC10 and then losing the plasmid and by knocking out PLP1 in a heterozygous diploid and dissecting tetrads. MAD2 was disrupted both by the PCR-based method (Longtine et al. 1998) and with the URA3-containing construct pRC10-1 (a gift from A. Murray). This construct was digested with XhoI and HindIII and transformed into the appropriate diploid strains. The TUB2 gene was integrated into SSY262 (pac10Δ plp1Δ yap4Δ) by integrating the vector pRS305-TUB2 that contains 622 bp of the 5′-UTR, the TUB2 ORF, and 367 bp of the 3′-UTR. This vector was digested with BsmI to integrate. The integration was checked by PCR, and Western blots were carried out to ensure that the levels of Tub2p were doubled.


We followed standard procedures for immunoblotting (Solomon et al. 1992), using isoform specific anti-α-tubulin (Tub1p) antibody no. 345 (Schatz et al. 1987) and anti-β-tubulin antibody no. 206 (Bond et al. 1986) at a dilution of 1/3500. α- and β-tubulin protein levels were normalized to carboxypeptidase Y (CPY) using anti-CPY antibody 1410 at a dilution of 1:5000 (gift of H. Ploegh). Immunoblot detection was performed as previously described (Abruzzi et al. 2002), and by using the ECL Plus chemiluminescence reagent kit (Amersham Biosciences, Arlington Heights, IL) and fluorescence imaging (Storm System; Molecular Dynamics, Sunnyvale, CA), and quantitated by ImageQuant Macintosh software. Western blots were repeated at least four times and results were averaged.

Gel filtration chromatography:

We used FSY183 (wild-type haploid), FSY846 (pac10∷HIS3), SSY38 (plp1∷hisG haploid), SSY262 (pac10∷HIS3 plp1∷hisG yap4∷kanR haploid), and MMY327 (pac10∷HIS3 plp1∷hisG yap4∷kanR + TUB2) cells to determine the state of tubulin polypeptides in each strain. Gel filtration chromatography was performed as described previously (Abruzzi et al. 2002). Briefly, cell extracts were obtained through French Press cell lysis and applied to a Sephacryl S-300HR column. One-milliliter fractions were collected and analyzed by Western blot for the presence of α- and β-tubulin. The percentage of tubulin present in heterodimer was calculated by adding the total of the tubulin present in all of the heterodimer fractions divided by the total of tubulin in all fractions. The columns were repeated at least two times. The amount of heterodimer present in the mutant cells compared to wild type was determined as follows: the amount of tubulin present in each of the heterodimer fractions was added together and then divided by the total amount of tubulin present in all of the fractions. This percentage of tubulin in heterodimer was then multiplied by the percentage of tubulin present in the strain compared to wild type to give the percentage of tubulin in heterodimer compared to wild type.

Synthetic interactions:

To determine if the spindle checkpoint protein Mad2p is needed for normal survival, SSY339 (pac10∷HIS3/+; plp1∷hisG/plp1∷hisG; yap4∷kanR/+; mad2∷URA3/+), MMY363 (pac10∷HIS3/+; plp1∷hisG/plp1∷hisG; yap4∷kanR/yap4∷hygR; mad2∷URA3/+), and SSY365 (pac10Δ∷HIS3/+; mad2Δ∷kanR/+) cells were sporulated. Genotypes were determined through analysis of auxotrophic markers. One hundred six spores were analyzed for SSY339, 196 for MMY363, and 100 for SSY365. No spores that made colonies larger than pin sized (smaller than small colonies) were the genotype of pac10Δ plp1Δ mad2Δ. Only 5 spores that made colonies larger than pin sized were of the genotype pac10Δ plp1Δ yap4Δ mad2Δ. Only 9 spores that were recovered that made colonies larger than pin sized were of the genotype pac10Δ mad2Δ. A χ2-test was used to determine if the pac10Δ plp1Δ yap4Δ mad2Δ and pac10Δ mad2Δ colonies were statistically significant. The χ2-test gave a P-value of ≪0.05% for each and the null hypothesis that each genotype is equally likely was rejected. The small number of colonies recovered likely acquired suppressor mutations.

To determine if the spindle position checkpoint protein Bub2p was required, SSY396 (pac10∷HIS3/+; plp1∷hisG/plp1∷hisG; yap4∷kanR/+; bub2∷URA3/+) and SSY399 (pac10Δ∷HIS3/+; bub2Δ∷kanR/+) cells were sporulated. Ninety-four spores were analyzed for SSY396 and 50 spores for SSY399 for genotypes. All genotypes were recovered in equivalent numbers.

Ribonuclease protection assay:

Total yeast RNA was prepared using a modified hot phenol extraction procedure (Ausubel et al. 1989). Antisense RNA probes for TUB1, TUB2, and CPY RNAs were obtained, using the Maxiscript in vitro transcription kit (Ambion, Austin, TX). The RPA11 kit (Ambion) was used to perform the assay, using 10 μg of yeast RNA. The protected fragments were run on a urea polyacrylamide denaturing gel. The gel was dried and analyzed using phosphor imaging (Storm System, Molecular Dynamics) and quantitated using ImageQuant Macintosh software.

Immunofluorescence and spindle measurements:

Cells were fixed and prepared for immunofluorescence as previously described (Schatz et al. 1988). Imaging was performed using a Deltavision deconvolution microscope on a Nikon TE200 base. All image processing and quantitation were performed in SoftWorx (Applied Precision). Late anaphase/telophase spindles were measured in large-budded cells with two divided nuclei, one in mother and one in bud. At least 40 cells per genotype were analyzed for spindle length.

Chromosome loss assays:

To construct wild-type, plp1Δ, pac10Δ, yap4Δ, pac10Δ plp1Δ, and pac10Δ plp1Δ yap4Δ strains that also carry the ade2-101 mutation and a linear minichromosome containing the ochre suppressor SUP11 and URA3, we created a diploid heterozygous for all of the mutations and containing the minichromosome. The diploid was then sporulated and haploids containing the various mutations were obtained using selectable markers and PCR. The assay was performed as described in Spencer et al. (1990) with slight variations. We measured chromosome loss events in the first cell division by counting colonies that were at least one-half red. We grew the cells in selective media overnight and plated them to YPD at 30° and 18° and 15°. The cells growing at 18° and 15° were incubated for 10 days at that temperature and then transferred to 30° for 2 days to allow color accumulation. Over 800 colonies per strain were analyzed.

Southern blots:

Southern blots compared two genes on chromosome VI, TUB2 and ACT1, to another gene on chromosome XIII, CPY. Genomic DNA was prepared as described (Solomon et al. 1992) for strains FSY183 (wild type), FSY846 (pac10Δ), SSY38 (plp1Δ), SSY43 (pac10Δ plp1Δ), and SSY262 (pac10Δ plp1Δ yap4Δ). Genomic DNA was digested with EcoRI and NdeI and probed. The probes were made by PCR amplification of genomic regions and labeled with 32P using the Megaprime DNA labeling kit (Amersham Biosciences). The TUB2 primers used for PCR amplification were TGTTTTATTTATTTCAACCTGGGCCT and TTTGGTTACCACACTGACCTGTCG. The ACT1 primers used for PCR amplification were GGGTTTGTTTGATCCTTTCCTTCC and GCAATCGATGTTAGTACATGAGAC. The CPY primers were CATACGCTATGAAAGCATTCACCA and CGGTTAGTGAAGAACAACCTGGAC. Southern blots were performed as previously described (Sambrook et al. 1989; Abruzzi et al. 2002) with minor changes. Southern blots were analyzed using phosphor imaging (Storm System, Molecular Dynamics) and quantitated using ImageQuant Macintosh software. Genomic DNA preparations and Southern blots were repeated and the two results averaged for a value of 1.9-fold increase in the genes on chromosome VI.


Mutations affecting tubulin heterodimer levels:

We are interested in cellular regulation of the pathways leading to tubulin heterodimer competent to form microtubules. These experiments have helped to define interactions between tubulin expression and folding pathways and to characterize the cells' response to extremely low levels of tubulin.

The level of heterodimer depends upon levels of tubulin polypeptide expression as well as the proper folding of those polypeptides so that they can form heterodimer. Deletion of nonessential genes that modulate tubulin folding can affect both of those processes. For example, deletion of the gene encoding the prefoldin component Pac10p leads to lower levels of both α- and β-tubulin (Figure 1) (Geissler et al. 1998; Lacefield and Solomon 2003). In contrast, cells lacking Plp1p, which participates in the folding of β-tubulin, express both tubulin polypeptides at normal levels (Figure 1) (Lacefield and Solomon 2003). Cells with deletions in both PAC10 and PLP1 combined with a deletion of a putative transcription factor YAP4 have dramatically lower levels of tubulin polypeptides. But, yap4Δ cells have normal levels (Figure 1). Both PAC10 and PLP1 were knocked out together to determine the consequence of the two folding defects. We also tested the role of Yap4p on the basis of the results of other screens ongoing in our lab (S. Lacefield and F. Solomon, unpublished results). The anomalous results for pac10Δ plp1Δ double-mutant cells, which have normal levels of β-tubulin protein, are explained in a subsequent section.

Figure 1.Figure 1.
Tubulin protein levels in different mutant backgrounds. (A) A summary of α-tubulin (Tub1p) and β-tubulin (Tub2p) protein levels from wild-type, yap4Δ, plp1Δ, pac10Δ, pac10Δ plp1Δ, pac10Δ ...

Previous work demonstrates that tubulin polypeptides in mutant cells can be in the form of heterodimer or in the form of stable aggregates (Abruzzi et al. 2002; Lacefield and Solomon 2003). Gel filtration chromatography can distinguish heterodimer from aggregate (Figure 2). By this assay, 93% of tubulin in wild-type cells elutes from the column as heterodimer. The 7% of the tubulin that elutes at the void volume, in the position of aggregated tubulin, likely represents an artifact of the sample preparation technique, since all of the α-tubulin co-immunoprecipitates with β-tubulin from extracts of wild-type cells (Archer et al. 1998). In contrast, pac10Δ, plp1Δ, and pac10Δ plp1Δ yap4Δ cells all show much larger and significant amounts of heterodimer in aggregates.

Figure 2.
Proportion of tubulin in heterodimer. Cell extracts from wild-type, plp1Δ, pac10Δ, pac10Δ plp1Δ yap4Δ, and pac10Δ plp1Δ yap4Δ + TUB2 cells were analyzed by gel filtration chromatography. ...

The fraction of tubulin in heterodimer, and the expression levels of the α- or β-tubulin polypeptides, permits evaluation of the heterodimer content of these strains relative to wild type (Table 1). Both pac10Δ and plp1Δ reduce heterodimer levels significantly—to 35 and 65% of normal levels, respectively. Most striking is the pac10Δ plp1Δ yap4Δ triple mutant, which contains only ∼20% of wild-type levels. The survival of cells with such low levels of tubulin heterodimer is surprising because of the essential functions microtubules play during cell division. A detailed characterization of the properties of these cells with low tubulin levels is presented in a subsequent section.

The percentage of heterodimer is low in pac10Δ plp1Δ yap4Δ

Diminution of tubulin levels is primarily due to differences at the level of the protein:

The data above demonstrate diminished levels of tubulin expression in strains containing mutations in genes encoding proteins active in tubulin folding and, potentially, gene expression. One explanation of these results is that improperly folded and therefore undimerized tubulin polypeptides are unstable and so are degraded. Alternatively, it is possible that these genes, singly or in combination, affect tubulin expression at the level of synthesis. To distinguish the contributions of these two mechanisms, we measured α- and β-tubulin RNA using RNAse protection assay (Figure 3).

Figure 3.Figure 3.
Tubulin RNA levels in different mutant backgrounds. (A) A summary of α-tubulin (TUB1) and β-tubulin (TUB2) RNA levels from wild-type, plp1Δ, yap4Δ, pac10Δ, pac10Δ plp1Δ, and pac10Δ plp1Δ ...

As expected, the mRNA levels in both plp1Δ cells and yap4Δ cells—like the tubulin protein levels in those cells—are comparable to those in wild type. Surprisingly, the pac10Δ single mutant, which causes a reduction in tubulin polypeptides, also shows a reduction in the levels of RNA. However, the reduced mRNA levels can account for only a part of the loss of tubulin polypeptides. Similarly, the pac10Δ plp1Δ yap4Δ triple mutant also has modestly reduced tubulin RNA levels. Again, the low level of tubulin polypeptides cannot be fully explained by this minor decrease in mRNA. The data in Figure 3 demonstrate that the β-tubulin mRNA levels in the pac10Δ plp1Δ mutant cells are, like the polypeptide levels, anomalous. These results are explored further in the next section.

We do not know the mechanism responsible for reduced mRNA levels in pac10Δ cells and in the pac10Δ plp1Δ yap4Δ cells. Tests for altered mRNA half-life, using RNA synthesis inhibitors, were negative (data not shown), but the downregulation is too small to make further experiments testing transcript levels interpretable.

The gel filtration experiments shown in Figure 2 and in previous experiments (Abruzzi et al. 2002; Lacefield and Solomon 2003) provide no evidence for α- or β-tubulin polypeptides eluting in the position expected for the monomeric protein. Therefore, these results suggest that undimerized tubulin polypeptides have two possible fates: they may aggregate and so elute in the void volume of gel filtration columns, or they may likely be degraded. The factors that determine which one of these fates is realized are currently unknown. Our mutants show a combination of both aggregation and degradation of the tubulin polypeptides, but clearly, the decreased levels of tubulin cannot be explained by a difference in expression levels.

A duplication of chromosome VI accounts for the increase of β-tubulin mRNA and protein in pac10Δ plp1Δ cells:

The twofold increase of β-tubulin RNA in pac10Δ plp1Δ cells relative to pac10Δ cells is the consequence of a duplication of chromosome VI, which carries the TUB2 gene. This strain was constructed several times using two different methods (see materials and methods). In each case, Southern blotting demonstrates that the amounts of both TUB2 and ACT1, also on chromosome VI, are approximately twofold higher than that of CPY, which is on chromosome XIII (chromosome XIII also contains TUB1 and TUB3) (Table 2). A duplication of chromosome VI normally cannot be tolerated because excess TUB2 is lethal in wild-type cells. However, the absence of Plp1p in these cells confers upon them resistance to free β-tubulin (Lacefield and Solomon 2003). The duplication of the chromosome containing TUB2 suggests that these cells may be limited for β-tubulin and the strain may have an advantage when the extra chromosome is present.

The chromosome containing TUB2 is duplicated in pac10Δ plp1Δ cells but not in pac10Δ plp1Δ yap4Δ cells

β-Tubulin is limiting for α-tubulin levels in cells lacking Pac10p, Plp1p, and Yap4p:

Although α-tubulin RNA levels in pac10Δ plp1Δ cells and pac10Δ plp1Δ yap4Δ cells are similar, α-tubulin protein is approximately twofold higher in pac10Δ plp1Δ cells (Figure 1). The difference in α-tubulin levels could be due to: (1) increased synthesis of β-tubulin due to the chromosome duplication described above, (2) increased synthesis of a different protein on the duplicated chromosome VI that functions to stabilize α-tubulin, or (3) a function of Yap4p that directly or indirectly stabilizes α-tubulin polypeptides.

To distinguish among these possibilities, we integrated a second copy of β-tubulin into the genome of pac10Δ plp1Δ yap4Δ cells. We found that this extra copy of the TUB2 gene, as expected, approximately doubles the expression level of β-tubulin protein (from 40 to 79% of wild type). It also is sufficient to approximately double α-tubulin polypeptide levels in the triple mutant: from 25 to 57% (Figure 1). These increases in tubulin expression can be explained if properly folded β-tubulin is limiting for heterodimer formation in the triple mutant and if undimerized α-tubulin is subject to degradation. Thus, increased β-tubulin expression leads to more properly folded β-tubulin, allowing increased heterodimer formation and stabilization of α-tubulin.

These results suggest that the extra copy of TUB2 provided by duplication of chromosome VI is sufficient to explain the differences between pac10Δ plp1Δ and pac10Δ plp1Δ yap4Δ cells. It is possible that the deletion of YAP4 makes these cells susceptible to the consequences of duplication of the entire chromosome VI. Because the pac10Δ plp1Δ cells are aneuploid, and because we cannot rule out that cultures of these cells are heterogeneous, analyses of the consequence of low tubulin levels used the pac10Δ plp1Δ yap4 cells with or without the extra copy of TUB2. The results presented below establish that the phenotypes of the triple-mutant cells are partially suppressed by the extra copy of TUB2.

Cells with low tubulin heterodimer levels have mitotic phenotypes but can still segregate chromosomes accurately:

The data presented below demonstrate the phenotypic consequences of this difference in tubulin heterodimer levels. To summarize, we find that cells with 20% of wild-type heterodimer levels have significantly shorter spindles and a nuclear positioning defect. Nevertheless, they divide normally and segregate their chromosomes accurately.

Growth rate:

Despite the fact that they have such low tubulin heterodimer levels, pac10Δ plp1Δ yap4Δ cells do not have a detectable growth defect in logarithmic growth phase. In rich medium at 30°, wild-type, pac10Δ, pac10Δ plp1Δ yap4Δ, and pac10Δ plp1Δ yap4Δ + TUB2 cells all have doubling times of ∼95 min (data not shown); pac10Δ plp1Δ cells grow slightly slower (doubling time ≈ 110 min), as is normal for aneuploid cells. Thus, cells with 20% of the heterodimer levels of wild-type cells grow at a normal rate.

Synthetic interactions with mutations affecting the mitotic checkpoint:

We found that the spindle assembly checkpoint is required for normal growth of cells that have low tubulin heterodimer levels. When there are defects in mitotic spindle assembly resulting in unattached or improperly aligned chromosomes, the spindle checkpoint halts the cell cycle and thereby prevents chromosome missegregation (Amon 1999). We knocked out one of the major components of the spindle checkpoint, MAD2, in appropriate diploid strains and then sporulated them to recover various genotypes. Detailed analysis (see materials and methods) showed that three of these strains—pac10Δ, pac10Δ plp1Δ, and pac10Δ plp1Δ yap4Δ cells—require an intact spindle checkpoint for normal growth (data not shown) but cells that are wild type, plp1Δ, yap4Δ, or plp1Δ yap4Δ, do not require the spindle checkpoint for normal growth. Thus, cells with 65% of the amount of heterodimer such as plp1Δ do not require the spindle checkpoint, but strains with even lower levels of heterodimer do need the spindle checkpoint for normal growth. In the case of pac10Δ cells, we cannot distinguish between needing the checkpoint for low tubulin heterodimer levels or for normal growth with an excess of toxic β-tubulin.

A second surveillance mechanism in mitosis, the spindle position checkpoint, is not required in pac10Δ plp1Δ yap4Δ cells. This checkpoint monitors the position of the nucleus at anaphase and if the nucleus is not in line with the mother-bud axis, the checkpoint prevents mitotic exit (Lew and Burke 2003). Bub2p is required for this checkpoint. We disrupted BUB2 in appropriate diploid strains, sporulated, and found that Bub2p was not required in pac10Δ plp1Δ yap4Δ cells, in pac10Δ cells, or in any other single- or double-mutant combination. Thus, the spindle assembly checkpoint is required in these cells, but the spindle position checkpoint is not.

Cold sensitivity:

Although pac10Δ plp1Δ yap4Δ cells do not have a growth defect at 30°, they grow slower at reduced temperatures. Mutations that affect microtubule formation—for example, deletion of PAC10 (Geiser et al. 1997)—are often cold sensitive (Figure 4A). Each of these strains has comparable plating efficiencies at 30° and at 15° except for the pac10Δ plp1Δ yap4Δ triple mutant. The triple mutant has a reduced amount of viable colonies in the cold that can be rescued by additional β-tubulin (Figure 4B). In contrast, cells that are yap4Δ or plp1Δ are not cold sensitive and grow at the same rate as wild-type cells at all temperatures. There were no differences at a higher temperature of 34° (data not shown). Thus, tubulin heterodimer levels at 20% of wild type result in cells that are cold sensitive for growth, suggesting that increased tubulin levels may be needed to maintain normal timing of the cell cycle in the cold.

Figure 4.Figure 4.
pac10Δ plp1Δ yap4Δ cells grow normally but are cold sensitive. (A) Serial dilutions of saturated yeast cultures were plated onto rich plates and grown at 30° for 3 days and 15° for 6 days. (B) Equivalent cell numbers ...

Chromosome loss:

We found that cells with low tubulin heterodimer levels do not show enhanced chromosome missegregation. We analyzed wild-type, yap4Δ, plp1Δ, pac10Δ, pac10Δ plp1Δ, and pac10Δ plp1Δ yap4Δ strains for loss of a linear minichromosome (Spencer et al. 1990). We monitored these strains by half-sector analysis (see materials and methods). At 30°, no significant change in minichromosome loss was seen in any of the mutant backgrounds (Table 3). However, at 18° and 15°, >30% of pac10Δ cells had lost the minichromosome, compared to <0.5% of wild-type cells. This loss rate was rescued by deleting PLP1; pac10Δ plp1Δ and pac10Δ plp1Δ yap4Δ cells lost chromosomes at a frequency comparable to wild type. Thus, the increased rate of chromosome loss in the cold in pac10Δ was likely due to the excess of toxic β-tubulin and when this toxicity was suppressed, chromosome loss was also suppressed. The data also suggest that at 30°, 18°, and 15°, the cells that are viable with low amounts of tubulin heterodimer segregate chromosomes faithfully.

Minichromosome loss rates are normal in pac10Δ plp1Δ yap4Δ cells

Progress through mitosis:

Cells that are pac10Δ plp1Δ yap4Δ have a longer G2/M phase of the cell cycle compared to wild-type cells. Budding yeast have various bud sizes depending upon the stage of the cell cycle. An asynchronous population of cells grown to early log phase was scored for percentage of cells in the population with no buds or small, medium, or large buds (Table 4). The large-budded fraction represents the cells in G2/M of the cell cycle. Wild type, pac10Δ, and plp1Δ cells have similar percentages of large-budded cells, but the number of large-budded cells among pac10Δ plp1Δ yap4Δ cells almost doubles. Addition of one copy of TUB2 in pac10Δ plp1Δ yap4Δ cells fully suppresses that phenotype.

In an asynchronous culture, pac10Δ plp1Δ yap4Δ mutants have a prolonged mitosis

Nuclear position:

Since microtubules are also required for nuclear positioning during mitosis, we asked if cells with low tubulin heterodimer levels have nuclear positioning defects. Nuclear position was analyzed in the population of large-budded cells (Table 5). Sixty-nine percent of wild-type large-budded cells have two nuclei that are clearly separated into mother and daughter cells and separated from the neck. In 15%, the nuclei have their DNA straddling the neck between mother and daughter cells. Only 15% have a single nucleus located in one cell.

pac10Δ plp1Δ yap4Δ large-budded cells have improperly positioned nuclei

In pac10Δ plp1Δ yap4Δ cells, the nuclear positions were significantly different. Of the large-budded cells, 54% have two nuclei, one in the mother and one in the daughter. Of those that have two separated nuclei, most are very close to the neck instead of well separated (both of the nuclei are in the longitudinal quarters of the cell closer to the neck). This position is rarely seen in wild-type cells (2/200 nuclei counted). Again in contrast to normal cells, 32% of pac10Δ plp1Δ yap4Δ large-budded cells have a single nucleus in one of the cells, and 10% have a single nucleus straddling the neck. Interestingly, 4% of cells had two nuclei localized to one of the cells, indicative of a modest defect in nuclear segregation (Table 5). The nuclear positioning of plp1Δ, pac10Δ, and pac10Δ plp1Δ yap4Δ + TUB2 cells is similar to that of wild-type cells. In summary, pac10Δ plp1Δ yap4Δ cells have a nuclear positioning defect characterized by an increase in the percentage of large-budded cells with either a single nucleus in the mother cell or two separated nuclei that are very close together. Both of these nuclear positioning defects are rescued with the addition of excess β-tubulin.

Spindle morphology:

The unusually small distance between two separated nuclei and the low tubulin heterodimer levels in the pac10Δ plp1Δ yap4Δ triple mutants led to the analysis of the spindle length in these cells. Consistent with their nuclear positioning defect, the spindle lengths are greatly reduced in pac10Δ plp1Δ yap4Δ cells. Spindle lengths were measured for large-budded cells with two distinct nuclei representing late anaphase/telophase spindles (Figure 5). Wild-type cells have an average spindle length of 10.02 ± 1.29 μm (Table 6). The spindles are, on average, 91 ± 6% of the length of the cell (Table 6). Similarly, yap4Δ, plp1Δ, and pac10Δ single-mutant cells have spindles comparable to wild type (data not shown). However, pac10Δ plp1Δ yap4Δ cells have significantly shorter spindles. The average length is 6.4 ± 1.13 μm with a range from 4.61 to 8.35 μm. This length is sufficient to span the distance through the neck to segregate chromosomes, but is much shorter than normal late anaphase/telophase spindles. In addition, since pac10Δ plp1Δ yap4Δ cells are the same length as wild-type cells, these short spindles only span 49 ± 6% of the cell length. As a consequence, the astral microtubules do not contact the distal ends of the cell, but instead contact the cell cortex at sites closest to the end of the spindle (Figure 5B). There was not a defect in the length of the cytoplasmic microtubules (data not shown). Finally, the spindles in pac10Δ plp1Δ yap4Δ + TUB2 cells are indistinguishable from those in wild-type cells (Figure 5C and Table 6). These observations on spindle morphology show that: (1) cells with 20% the amount of tubulin heterodimer have significantly shorter spindles than wild-type cells, and (2) normal spindle length can be obtained in these cells by increasing heterodimer levels.

Figure 5.
Spindle lengths are shorter in pac10Δ plp1Δ yap4Δ compared to wild type. Deconvolution microscopy of (A) wild-type, (B) pac10Δ plp1Δ yap4Δ, and (C) pac10Δ plp1Δ yap4Δ + TUB2 ...
Average spindle lengths and ranges in wild-type, pac10Δ plp1Δ yap4Δ and pac10Δ plp1Δ yap4Δ + TUB2 cells


The results presented above demonstrate that tubulin heterodimer levels can be decreased by mutations in the folding pathway and that the cell can maintain many normal microtubule functions with low heterodimer levels. Previous work identified both Pac10p and Plp1p as modulators of tubulin folding (Geissler et al. 1998; Lacefield and Solomon 2003). We show that the absence of these proteins can affect tubulin heterodimer levels. Cells lacking Pac10p, a member of the GimC/prefoldin complex that helps fold tubulin, have lower heterodimer levels (Tables 1 and 7). Deletion of the PLP1 gene significantly compromises β-tubulin folding, also resulting in a decrease in tubulin heterodimer levels (Table 1 and Lacefield and Solomon 2003). However, when both PAC10 and PLP1 are deleted along with the deletion of YAP4, a gene encoding a putative transcription factor (Fernandes et al. 1997), tubulin heterodimer levels are substantially decreased.

Cells with low tubulin heterodimer levels have some microtubule phenotypes but can still segregate chromosomes accurately:

An earlier attempt to study the consequence of low tubulin heterodimer levels characterized a diploid strain heterozygous for β-tubulin; those cells had only 50% of wild-type tubulin levels and were essentially normal with respect to microtubule function (Katz et al. 1990). From the work presented here, we know that cells with just 35% of wild-type heterodimer levels also have essentially normal microtubules. It is only when the cells have 20% heterodimer levels that we see these microtubule phenotypes: prolonged mitosis, shortened spindle, and abnormal nuclear position. These cells can still undergo faithful chromosome segregation and grow at rates equivalent to wild type at 30° in rich conditions. A shorter spindle can therefore be tolerated and function normally under these conditions. The cell likely maintains the same division time by decreasing the length of G1 since the percentage of cells in G1 is much smaller than in wild-type cells (Table 4).

The phenotypes associated with low tubulin heterodimer levels in the pac10Δ plp1Δ yap4Δ triple-mutant strain resemble those of certain tubulin mutations that alter microtubule dynamics. Microtubules undergo dynamic instability during polymerization, which refers to periods of slow growth and periods of rapid shrinkage (Howard and Hyman 2003). Changes in growth and shrinkage phases allow microtubules to move chromosomes and nuclei to different locations in the cell. If microtubule dynamics occur faster or slower than normal, the cell may be defective in nuclear localization or chromosome segregation. For example, there are two alleles of β-tubulin, tub2-150 and tub2-T143G, that result in altered microtubule dynamics and have defects in either nuclear localization or cell cycle timing (Machin et al. 1995; Dougherty et al. 2001; Dorn et al. 2005). Both mutants have short anaphase spindles like those we see in the pac10Δ plp1Δ yap4Δ triple mutant.

The tub2-150 mutant was isolated as a conditional allele that can grow only in the presence of the microtubule depolymerizing drug benomyl (Thomas et al. 1985). The benomyl suppresses the increased growth and shrinkage rates of microtubules in this strain (Dorn et al. 2005). In addition to the short anaphase spindles, this strain also has a nuclear localization defect similar to that of the pac10Δ plp1Δ yap4Δ triple mutant (Machin et al. 1995). However, in contrast to the triple mutant, the spindles are also mispositioned. These cells rapidly lose viability due to nuclear missegregation, but the triple-mutant cells segregate nuclei correctly 96% of the time. Furthermore, the triple mutant does not require the spindle position checkpoint, suggesting that the nuclear positioning defect can be resolved.

The tub2-T143G mutant has a similar short spindle defect with spindles that just span the length of the bud neck, as well as an increase in large-budded cells with mispositioned nuclei (Dougherty et al. 2001). This mutation in β-tubulin is close to the site where a guanine nucleotide is bound. It alters microtubule dynamics, causing slower growing and more stable microtubules. Unlike the pac10Δ plp1Δ yap4Δ triple mutant, the cells have an increased doubling time.

The low level of heterodimer in pac10Δ plp1Δ yap4Δ triple-mutant cells could affect microtubule dynamics. Most likely, the lower heterodimer concentrations will lead to slower growth of microtubules in the cell, but an intriguing future direction will be to measure the rate of microtubule growth and shrinkage in the pac10Δ plp1Δ yap4Δ triple-mutant strain.

Functional β-tubulin is limiting for heterodimer formation in pac10Δ plp1Δ yap4Δ cells:

The pac10Δ plp1Δ yap4Δ triple mutant has a tubulin heterodimer level 20% that of wild-type cells. This low level of tubulin heterodimer is due to a limited amount of functional β-tubulin. Addition of a single copy of the β-tubulin gene increases both α- and β-tubulin protein levels and also increases heterodimer levels from 20 to 50% (Figure 1 and Table 1). Excess α-tubulin without a partner of functional β-tubulin is likely degraded (Katz et al. 1990). An increase of β-tubulin, even inefficiently folded as in pac10Δ plp1Δ, appears to stabilize some of the α-tubulin, increasing the amount of heterodimer in the cell.

Surprisingly, pac10Δ plp1Δ cells increase β-tubulin expression by duplicating the β-tubulin chromosome to raise heterodimer levels. This duplication does not occur in pac10Δ plp1Δ yap4Δ cells. We are unsure of the role that Yap4p plays in the tubulin pathway other than allowing the duplication when present and prohibiting the duplication when absent. Perhaps, when YAP4 is deleted, the excess of another gene on chromosome VI in addition to TUB2 cannot be tolerated.

Among the interesting features of this duplication is the fact that, ordinarily, an increase of β-tubulin expression is toxic. That it is not in the pac10Δ plp1Δ background is probably due to the presence of excess α-tubulin available to bind the β-tubulin and also to the fact that plp1Δ suppresses the toxicity of excess β-tubulin (Lacefield and Solomon 2003). These aneuploid cells grow slower but must have some sort of advantage to maintain this aneuploid state. Indeed, when we introduce an additional copy of TUB2 to pac10Δ plp1Δ yap4Δ cells, we observe a rescue of the microtubule phenotypes. Thus, even though cells can survive and segregate chromosomes normally with a short spindle, this is not the preferred state. For this reason, we believe that 20% is the lowest level of tubulin heterodimer possible to maintain normal growth. Yeast cells with low tubulin heterodimer levels growing in the wild may have a strong disadvantage due to the slow growth in the cold and the mitotic defects.


We thank A. Amon, I. Goldhaber-Gordon, A. Grossman, G. Mandel, T. Orr-Weaver, M. Rosbash, A. Seshan, P. Sharp, L. Vega, and members of our laboratory for valuable comments and advice on this work; P. Hieter, A. Murray, P. Phillipsen, and J. Yang for strains and reagents; and D. Rines and P. Sorger for help with microscopy. S.L. was supported by a predoctoral fellowship from the Ludwig Cancer Fund and by a predoctoral training grant from the National Institute of General Medical Sciences to M.I.T. This work was supported by a grant from National Institute of General Medical Sciences to F.S.


  • Abruzzi, K., A. Smith, W. Chen and F. Solomon, 2002. Protection from free β-tubulin by the β-tubulin binding protein Rbl2p. Mol. Cell. Biol. 22: 138–147. [PMC free article] [PubMed]
  • Alvarez, P., A. Smith, J. Fleming and F. Solomon, 1998. Modulation of tubulin polypeptide ratios by the yeast protein Pac10p. Genetics 149: 857–864. [PMC free article] [PubMed]
  • Amon, A., 1999. The spindle checkpoint. Curr. Opin. Genet. Dev. 9: 69–75. [PubMed]
  • Archer, J., M. Magendantz, L. Vega and F. Solomon, 1998. Formation and function of the Rbl2p-b-tubulin complex. Mol. Cell. Biol. 18: 1757–1762. [PMC free article] [PubMed]
  • Ausubel, F. N., R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman et al., 1989. Current Protocols in Molecular Biology. John Wiley & Sons, New York.
  • Bond, J. F., J. L. Fridovich-Keil, L. Pillus, R. C. Mulligan and F. Solomon, 1986. A chicken-yeast chimeric β-tubulin protein is incorporated into mouse microtubules in vivo. Cell 44: 461–468. [PubMed]
  • Burke, D., P. Gasdaska and L. Hartwell, 1989. Dominant effects of tubulin overexpression in Saccharomyces cerevisiae. Mol. Cell. Biol. 9: 1049–1059. [PMC free article] [PubMed]
  • Dorn, J. F., K. Jaqaman, D. R. Rines, G. S. Jelson, P. K. Sorger et al., 2005. Yeast kinetochore microtubule dynamics analyzed by high-resolution three-dimensional microscopy. Biophys. J. 89: 2835–2854. [PMC free article] [PubMed]
  • Dougherty, C. A., C. R. Sage, A. Davis and K. W. Farrell, 2001. Mutation in the β-tubulin signature motif suppresses microtubule GTPase activity and dynamics, and slows mitosis. Biochemistry 40: 15725–15732. [PubMed]
  • Dunn, A. Y., M. W. Melville and J. Frydman, 2001. Review: cellular substrates of the eukaryotic chaperonin TRiC/CCT. J. Struct. Biol. 135: 176–184. [PubMed]
  • Fernandes, L., C. Rodrigues-Pousada and K. Struhl, 1997. Yap, a novel family of eight bZIP proteins in Saccharomyces cerevisiae with distinct biological functions. Mol. Cell. Biol. 17: 6982–6993. [PMC free article] [PubMed]
  • Geiser, J. R., E. J. Schott, T. J. Kingsbury, N. B. Cole, L. J. Totis et al., 1997. Saccharomyces cerevisiae genes required in the absence of the CIN8-encoded spindle motor act in functionally diverse mitotic pathways. Mol. Biol. Cell 8: 1035–1050. [PMC free article] [PubMed]
  • Geissler, S., K. Siegers and E. Schiebel, 1998. A novel protein complex promoting formation of functional α- and γ-tubulin. EMBO J. 17: 952–966. [PMC free article] [PubMed]
  • Guthrie, C., and G. Fink, 1991. Guide to Yeast Genetics and Molecular Biology. Academic Press, New York.
  • Howard, J., and A. A. Hyman, 2003. Dynamics and mechanics of the microtubule plus end. Nature 422: 753–758. [PubMed]
  • Javerzat, J., G. Cranston and R. C. Allshire, 1996. Fission yeast genes which disrupt mitotic chromosome segregation when overexpressed. Nucleic Acids Res. 24: 4676–4683. [PMC free article] [PubMed]
  • Katz, W., B. Weinstein and F. Solomon, 1990. Regulation of tubulin levels and microtubule assembly in Saccharomyces cerevisiae: consequences of altered tubulin gene copy number in yeast. Mol. Cell. Biol. 10: 2730–2736. [PMC free article] [PubMed]
  • Lacefield, S., and F. Solomon, 2003. A novel step in β-tubulin folding is important for heterodimer formation in Saccharomyces cerevisiae. Genetics 165: 531–541. [PMC free article] [PubMed]
  • Leroux, M., M. Fandrich, D. Klunker, K. Siegers, A. Lupas et al., 1999. MtGimC, a novel archeal chaperone related to the eukaryotic chaperonin GimC/prefoldin. EMBO J. 18: 6730–6743. [PMC free article] [PubMed]
  • Lew, D. J., and D. J. Burke, 2003. The spindle assembly and spindle position checkpoints. Annu. Rev. Genet. 37: 251–282. [PubMed]
  • Longtine, M. S., A. McKenzie, D. J. Demarini, N. G. Shah, A. Wach et al., 1998. Additional modules for versatile and economical PCR-based gene deletion and modification in Saccharomyces cerevisiae. Yeast 14: 953–961. [PubMed]
  • Machin, N. A., J. M. Lee and G. Barnes, 1995. Microtubule stability in budding yeast: characterization and dosage suppression of a benomyl-dependent tubulin mutant. Mol. Biol. Cell 6: 1241–1259. [PMC free article] [PubMed]
  • Sambrook, J., E. F. Fritsch and T. Maniatis, 1989. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
  • Schatz, P., L. Pillus, F. Grisafi, F. Solomon and D. Botstein, 1986. Two functional α-tubulin genes of the yeast Saccharomyces cerevisiae encode divergent proteins. Mol. Cell. Biol. 6: 3711–3721. [PMC free article] [PubMed]
  • Schatz, P., F. Solomon and D. Botstein, 1988. Isolation and characterization of conditional-lethal mutations in the TUB1 a-tubulin gene of the yeast Saccharomyces cerevisiae. Genetics 120: 681–695. [PMC free article] [PubMed]
  • Schatz, P. J., G. E. Georges, F. Solomon and D. Botstein, 1987. Insertions of up to 17 amino acids into a region of a-tubulin do not disrupt function in vivo. Mol. Cell. Biol. 7: 3799–3805. [PMC free article] [PubMed]
  • Sherman, F., G. Fink and J. Hicks, 1986. Laboratory Course Manual for Methods in Yeast Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
  • Solomon, F., L. Connell, D. Kirkpatrick, V. Praitis and B. Weinstein, 1992. Methods for studying the yeast cytoskeleton, pp. 197–222 in The Cytoskeleton, edited by K. Carraway and C. Carraway. Oxford University Press, Oxford.
  • Spencer, F., C. Gerring, C. Connelly and P. Hieter, 1990. Mitotic chromosome transmission fidelity mutants in Saccharomyces cerevisiae. Genetics 124: 237–249. [PMC free article] [PubMed]
  • Thomas, J. H., N. F. Neff and D. Botstein, 1985. Isolation and characterization of mutations in the b-tubulin gene of Saccharomyces cerevisiae. Genetics 112: 715–734. [PMC free article] [PubMed]
  • Ursic, D., and M. Culbertson, 1991. The yeast homolog to mouse Tcp-1 affects microtubule-mediated processes. Mol. Cell. Biol. 11: 2629–2640. [PMC free article] [PubMed]
  • Vainberg, I., S. Lewis, H. Rommelaere, C. Ampe, J. Vandekerckhove et al., 1998. Prefoldin, a chaperone that delivers unfolded proteins to the cytosolic chaperonin. Cell 93: 863–873. [PubMed]
  • Weinstein, B., and F. Solomon, 1990. Phenotypic consequences of tubulin overproduction in Saccharomyces cerevisiae: differences between alpha-tubulin and beta-tubulin. Mol. Cell. Biol. 10: 5295–5304. [PMC free article] [PubMed]

Articles from Genetics are provided here courtesy of Genetics Society of America
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...