A Novel Step in ß-Tubulin Folding Is Important for Heterodimer Formation in Saccharomyces cerevisiae

Corresponding author: Frank Solomon, Room 220, M.I.T., Cambridge, MA 02139., solomon{at}mit.edu (E-mail)

Communicating editor: F. WINSTON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Undimerized ß-tubulin is toxic in the yeast S. cerevisiae. It can arise if levels of ß-tubulin and -tubulin are unbalanced or if the tubulin heterodimer dissociates. We are using the toxicity of ß-tubulin to understand early steps in microtubule morphogenesis. We find that deletion of PLP1 suppresses toxic ß-tubulin formed by disparate levels of - and ß-tubulin. That suppression occurs either when -tubulin is modestly underexpressed relative to ß-tubulin or when ß-tubulin is inducibly and strongly overexpressed. Plp1p does not affect tubulin expression. Instead, a significant proportion of the undimerized ß-tubulin in plp1 cells is less toxic than that in wild-type cells. It is also less able to combine with -tubulin to form a heterodimer. As a result, plp1 cells have lower levels of heterodimer. Importantly, plp1 cells that also lack Pac10, a component of the GimC/PFD complex, are even less affected by free ß-tubulin. Our results suggest that Plp1p defines a novel early step in ß-tubulin folding.

MICROTUBULE function requires the participation of genes with a wide range of functions. Biochemical experiments first identified proteins that modulate the assembly and dynamics of the microtubule polymer. These proteins contribute to the formation and function of the many diverse microtubule structures (DESAI and MITCHISON 1997 ). Other experiments, both biochemical and genetic analyses, demonstrate activities that affect earlier steps in microtubule morphogenesis, in particular the folding of the tubulin polypeptides to allow formation of -ß-tubulin heterodimers, the subunits of the microtubule polymer.

The efficient formation of heterodimer is crucial for the cell. First, cell viability requires some minimal level of tubulin to support essential cell functions. Second, undimerized ß-tubulin, arising either because it is in excess with respect to -tubulin or because the heterodimer dissociates or does not form properly, is extremely toxic in yeast (BURKE et al. 1989 ; WEINSTEIN and SOLOMON 1990 ; JAVERZAT et al. 1996 ; VEGA et al. 1998 ). Even at low levels, undimerized ß-tubulin can disrupt microtubule assembly (SCHATZ et al. 1986 ); at higher levels, it is lethal. In contrast, a very large excess of -tubulin has only modest consequences for the cell (WEINSTEIN and SOLOMON 1990 ). Among cytoskeletal proteins, the problem of maintaining balance between major components—an issue encountered in other morphogenetic pathways (FLOOR 1970 ; STERNBERG 1976 )—affects only tubulin, but of course not actin, intermediate filament subunits, or even the prokaryotic predecessor of tubulin, FtsZ (ERICKSON 1995 ), all of which are monomeric proteins acting as primary subunits. Why the microtubule subunit contains two rather similar polypeptides, one of which is toxic on its own, is not understood.

Several gene products are involved in forming heterodimer. They include the cytosolic chaperonin (CCT), an essential structure that participates in the folding of many proteins (URSIC and CULBERTSON 1991 ; DUNN et al. 2001 ). In addition, the nonessential GimC or prefoldin complex (GimC/PFD; GEISSLER et al. 1998 ; VAINBERG et al. 1998 ) cooperates with CCT in folding a subset of those proteins, including actin and the tubulin family.

Third, five proteins act as cofactors in an in vitro assay for incorporation of tubulin polypeptides into heterodimer (TIAN et al. 1997 ). These cofactors are not essential for heterodimer formation in budding yeast (STEARNS et al. 1990 ; ARCHER et al. 1995 ; GEISER et al. 1997 ; HOYT et al. 1997 ) and may act in a pathway to reclaim dissociated heterodimers (FLEMING et al. 2000 ).

The specific role of GimC/PFD in the formation of functional cytoskeletal proteins is not known. It may interact with nascent polypeptides and transfer them to the CCT (VAINBERG et al. 1998 ; HANSEN et al. 1999 ), although the CCT itself may also interact with nascent chains bound to the ribosome (MCCALLUM et al. 2000 ). GimC/PFD may bind incompletely folded cytoskeletal proteins released from the CCT and help return them to the chaperonin for another round of sequestered folding (VAINBERG et al. 1998 ). Alternatively, GimC/PFD may prevent release of an incorrectly folded polypeptide from CCT (SIEGERS et al. 1999 ). The GimC/PFD complex interacts with both - and ß-tubulin (VAINBERG et al. 1998 ; SIEGERS et al. 1999 ; our unpublished results) and is important but not required for folding of both proteins.

A conspicuous phenotype of deleting PAC10, which encodes Pac10p/Gim2p, one of the yeast GimC/PFD components, is supersensitivity to microtubule depolymerizing drugs, a common microtubule phenotype (ALVAREZ et al. 1998 ; GEISSLER et al. 1998 ). In pac10 cells, this phenotype can be explained by changes in tubulin expression; the levels of both - and ß-tubulin are reduced to 55 and 85% of wild type, respectively (ALVAREZ et al. 1998 ). The resulting modest excess of ß-tubulin can account for the benomyl supersensitivity of pac10 cells, since overexpressed -tubulin or Rbl2p (rescues ß-tubulin lethality), a ß-tubulin-binding protein that protects cells against excess ß-tubulin (ARCHER et al. 1995 ), completely suppresses the drug phenotype (ALVAREZ et al. 1998 ; GEISSLER et al. 1998 ). In fact, the level of undimerized ß-tubulin in pac10 cells makes them inviable either in the absence of RBL2 or when the minor -tubulin gene TUB3 is deleted, producing an additional 15% undimerized ß-tubulin (ALVAREZ et al. 1998 ).

We are using Saccharomyces cerevisiae to study tubulin heterodimer formation as an early step in microtubule morphogenesis, using the toxicity of free ß-tubulin as a probe of its state in vivo. We screened for loss-of-function mutations that would rescue a strain that contains three nonessential mutations affecting tubulin expression and that is inviable in the absence of a low-copy plasmid expressing -tubulin. This screen identifies deletion of the PLP1 gene as a suppressor of this lethality. We show that plp1 rescues cells from excess ß-tubulin in several circumstances, including high-level overexpression. Suppression by plp1 does not occur through differential expression of Rbl2p or tubulin proteins. Instead, it affects the properties of undimerized ß-tubulin. The data suggest that Plp1p may affect the state of ß-tubulin in the cell by facilitating the efficient transfer of nascent ß-tubulin polypeptides to the folding apparatus.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains and media:
All yeast strains are derivatives of FSY183 (WEINSTEIN and SOLOMON 1990 ) and are listed in Table 1. We used standard yeast manipulation methods and media (SHERMAN et al. 1986 ; GUTHRIE and FINK 1991 ; SOLOMON et al. 1992 ) and we used the pNKY51 vector containing hisG-URA3-hisG sequences to disrupt the entire PLP1 and GRR1 open reading frames (ALANI et al. 1987 ). STE4 and STE18 were deleted using a PCR-based method (LONGTINE et al. 1998 ). Deletion of PAC10, TUB3, and RBL2 was previously described (ARCHER et al. 1995 ; ALVAREZ et al. 1998 ; ABRUZZI et al. 2002 ). We hemagglutinin (HA)-tagged the chromosomal PLP1 gene using the pFA6a-3HA-kanMX6 module (LONGTINE et al. 1998 ).

Mutagenesis:
We mutagenized SSY14 (pac10::HIS3 grr1::hisG tub3::hisG + pAIA510b/TUB1) with the mTn-lacZ/LEU2 insertion library (BURNS et al. 1994 ) according to the Yale Genome Analysis Center protocol (http://ygac.med.yale.edu/). We plated library transformants to synthetic complete media lacking leucine, uracil, and histidine. Colonies grew 4 days and then were replica plated to 5-fluoroorotic acid (5-FOA). We used the vectorette PCR rescue method described by the Botstein lab (http://genome-www.stanford.edu/group/botlab/protocols/vectorette.html) to identify the position of the inserts.

DNA sequencing:
Sequencing of the transposon-inserted allele of PLP1 was performed by the MIT Biopolymers Facility.

Immunoblotting:
We followed standard procedures for immunoblotting (SOLOMON et al. 1992 ), using anti--tubulin antibody 345 (SCHATZ et al. 1987 ) and anti-ß-tubulin antibody 206 (BOND et al. 1986 ). Tubulin protein levels were normalized to carboxypeptidase Y (CPY). Immunoblot detection was performed as previously described (ABRUZZI et al. 2002 ).

GroEL trap experiments:
Strains with integrated TUB2-LEU2-GAL-TUB2 were transformed with a plasmid containing the gene encoding GroEL D87K trap under a copper inducible promotor (generous gift of U. Hartl). The plasmid (pSal4-T-GroEL) was described in SIEGERS et al. 1999 and was made by subcloning the Escherichia coli GroEL D87K mutant (FARR et al. 1997 ) under the CUP1 promoter. The strains were grown in raffinose to log phase. At 0 hr, CuSO₄ to 100 µM and galactose to 2% were added, and cells were plated to noninducing media at different time points.

Gel filtration chromatography:
To determine the state of tubulin polypeptides in each strain, we used gel filtration chromatography as described previously (ABRUZZI et al. 2002 ).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

A deletion of PLP1 rescues the synthetic lethality of pac10 grr1 tub3 triple mutants:
To find genes involved in the regulation of tubulin expression or toxicity, we screened for the rescue of pac10 grr1 tub3 triple mutants using a transposon-based insertion library. As noted above, both pac10 and tub3 mutants contain undimerized ß-tubulin. Each of those mutations is synthetically lethal with grr1 (A. SMITH, M. MAGENDANTZ and F. SOLOMON, unpublished results). The screen is based on the observation that this triple mutant can survive with a plasmid containing TUB1, the major -tubulin gene. Therefore, the inviability of the triple mutant is most likely due to the presence of toxic, undimerized ß-tubulin. Rescue of the triple mutant must suppress at least two of the three mutations since each of the pairwise combinations is synthetically lethal, but each combination is suppressible by either excess -tubulin or excess Rbl2p (ALVAREZ et al. 1998 ; S. LACEFIELD and F. SOLOMON, unpublished results). However, we used the triple mutant strain for our suppressor screen because it gave a 10-fold lower background of false positives relative to any of the double-mutant strains. Thus, mutations that allow these cells to live could be involved in regulating expression levels of -tubulin, ß-tubulin, or Rbl2p or in regulating the toxicity of undimerized ß-tubulin.

SSY14 cells, pac10 grr1 tub3 covered with TUB1 on a URA3 marked CEN plasmid, grow on synthetic complete medium but not on medium containing 5-FOA, which allows growth of only cells that can survive the loss of the covering plasmid. Using the mTn-lacZ/LEU2 insertion library (BURNS et al. 1994 ), we recovered three independent insertions that could survive in the absence of the plasmid containing TUB1. All three insertions were within the gene PLP1 (FLANARY et al. 2000 ). However, we screened with only 1x coverage and do not believe the screen was saturated. To verify this rescue, we recreated the quadruple mutant by deleting the entire open reading frame of PLP1 in SSY14 cells and showed that it can grow, although slowly, without the extra -tubulin expressed from the covering plasmid (Fig 1A). The long doubling times of the triple mutant suppressed by either excess TUB1 (9 hr) or plp1 (12 hr) are largely determined by the grr1 mutation, which substantially slows the growth of wild-type cells (1.6 vs. 6 hr). The plp1 itself does not affect growth rate.

View larger version (102K): In this window In a new window Download PPT slide   Figure 1. pac10 grr1 tub3 plp1 cells are viable and express very low tubulin levels. (A) pac10 grr1 tub3 plp1 cells can survive in the absence of a plasmid containing TUB1. Serial dilutions (1:5) of saturated yeast cultures were grown on standard media (YPD) and photographed after 4 days of growth. pac10 grr1 tub3 plp1 mutants grow more slowly than pac10 and grr1 mutants. pac10 grr1 tub3 mutants cannot survive on 5-FOA plates in the absence of the plasmid containing TUB1. (B) -Tubulin and ß-tubulin levels of wild-type and pac10 grr1 tub3 plp1 cell extracts were determined by immunoblotting. Values reported (from an average of three experiments) are the levels of tubulin polypeptides normalized to CPY; the wild-type values are defined as 1.0.

The pac10 grr1 tub3 plp1 quadruple mutant survives with very low tubulin levels:
We used immunoblotting to measure the levels of Tub1p and Tub2p in the pac10 grr1 tub3 plp1 quadruple mutant. A deletion of either PAC10 or GRR1 causes a decrease of 20–45% in the levels of both Tub1p and Tub2p (ALVAREZ et al. 1998 ; A. SMITH, M. MAGENDANTZ and F. SOLOMON, unpublished results). In addition, deletion of TUB3 decreases the total amount of -tubulin by 15% (SCHATZ et al. 1986 ). Normalizing to a control protein (carboxypeptidase Y), we found that tubulin levels are dramatically reduced in the quadruple mutant: -tubulin levels to 27 ± 5% and ß-tubulin levels to 32 ± 6% of wild type (Fig 1B). The quadruple mutant cells grow even more slowly than grr1 cells, which have an elongated G₁ phase (BARRAL et al. 1995 ). The quadruple mutant has the lowest level of tubulin shown to be sufficient for growth in yeast. Cells with 50% of the wild-type tubulin complement grow at normal rates (KATZ et al. 1990 ).

plp1 suppresses the toxicity of free ß-tubulin:
Deletion of PLP1 partially rescues the benomyl supersensitivity of pac10 and tub3 cells (Fig 2). This suppression is apparent at relatively low benomyl concentrations—10 µg/ml for pac10 plp1 cells and 4µg/ml for tub3 plp1 cells. At higher drug concentrations, the suppression is not detectable. Thus, plp1 suppresses the phenotypes of undimerized ß-tubulin arising from altered expression levels. The plp1 mutation in an otherwise wild-type background does not confer either benomyl resistance or sensitivity at any concentration of drug.

View larger version (53K): In this window In a new window Download PPT slide   Figure 2. plp1 suppresses the benomyl phenotypes of pac10 and tub3 at low benomyl concentrations. Wild type, plp1, pac10, pac10 plp1, tub3, and tub3 plp1 cells were grown overnight in rich media. Serial dilutions (1:5) of a saturated culture were spotted onto rich plates (YPD) or plates containing benomyl (Ben 10 µg/ml) as indicated and photographed after 3 days of growth.

Plp1p's microtubule function is independent of phosducin-like homology:
The gene encoding the yeast Plp1p (for phosducin-like protein) was identified and named on the basis of its sequence similarity to mammalian phosducin (FLANARY et al. 2000 ). In retinal cells, phosducin binds the ß-subunits of G proteins and so inhibits their rebinding to the -subunit to reconstitute the trimeric protein. In yeast, the only known Gß protein function is in the mating response pathway. Mating pheromone binding is signaled through the G protein ß-subunits Ste4p and Ste18p. Consistent with the sequence homology, Plp1p can bind Ste4p and Ste18p released from the mating pheromone receptor when -factor is present (FLANARY et al. 2000 ). Overexpression of Plp1 does significantly suppress pheromone-induced gene expression, but neither deletion nor overexpression of PLP1 has a significant effect on mating response (FLANARY et al. 2000 ). The related yeast protein, Plp2, is essential, but its function relative to Plp1 is unclear, since overexpression of Plp1 does not suppress plp2 (H. DOLHMAN, personal communication).

We determined that neither Ste4p nor Ste18p is required for the rescue of microtubule mutants by plp1. First, plp1 rescue of pac10 grr1 cells is unaffected by deletion of STE4 and STE18 (data not shown). Also, plp1 still reduces the benomyl sensitivity of pac10 ste4 ste18. Thus, the role of Plp1p in microtubule regulation is independent of any known functional homology to phosducin.

Plp1p is also distinct from the recently described rat protein PhLP (MCLAUGHLIN et al. 2002 ). PhLP is related to phosducin but is only 15% identical to yeast Plp1p and Plp2p. Overexpression of PhLP severely affects protein folding, while overexpressed Plp1p has no phenotype (data not shown). Thus, there appears to be no significant functional relationship between these proteins.

plp1 does not suppress undimerized ß-tubulin by modifying tubulin expression levels:
The benomyl sensitivity of pac10 cells can be attributed at least in part to excess ß-tubulin, since it is rescued by overexpression of Rbl2p (ALVAREZ et al. 1998 ; GEISSLER et al. 1998 ). Therefore, the suppression by plp1 could be achieved by increasing the amount of -tubulin or by decreasing the amount of ß-tubulin in the cell. However, the ratio of ß- to -tubulin in pac10 plp1 is similar to that in pac10 (Fig 3A). In agreement with previous reports (ALVAREZ et al. 1998 ; GEISSLER et al. 1998 ), in pac10 cells - and ß-tubulin levels are reduced to 54 ± 10% and 85 ± 10% of wild type, respectively. The ratio of ß- to -tubulin is 1.6, showing that there is an excess of ß-tubulin compared to -tubulin. In a pac10 plp1 double mutant, -tubulin (48 ± 5% of wild type) and ß-tubulin (92 ± 3% of wild type) levels are similar to those in pac10 cells within experimental error. The ratio of ß- to -tubulin is 1.9. There is no effect of plp1 on tubulin levels in an otherwise wild-type background: -tubulin levels are 95 ± 10% and ß-tubulin levels are 96 ± 10%. Therefore, despite the fact that the pac10 plp1 double mutant actually has levels of tubulin polypeptides similar to those of the pac10 mutant, it has lower sensitivity to benomyl.

View larger version (137K): In this window In a new window Download PPT slide   Figure 3. plp1 does not rescue by decreasing ß-tubulin levels, increasing -tubulin levels, or through Rbl2p expression. (A) plp1 does not affect tubulin levels in wild-type or pac10 cells. Extracts from wild type, plp1, pac10, and pac10 plp1 were analyzed for levels of -tubulin, ß-tubulin, and CPY by immunoblotting, as for Fig 1B. Values reported are the average of five experiments. (B) A deletion of PLP1 rescues rbl2 tub3 and rbl2 pac10. Cells deleted for both RBL2 and TUB3 carrying a low-copy plasmid expressing TUB3 and URA3 grow normally on standard media (glucose) but do not grow on media containing 5-FOA, which selects for cells that have lost the plasmid expressing URA3 (5-FOA). In contrast, if the rbl2 tub3 and rbl2 pac10 are also deleted for PLP1, they grow normally in both the absence and the presence of the covering plasmid. Saturated cultures were spotted onto both media and photographed after 3 days of growth.

plp1 does not suppress undimerized ß-tubulin through RBL2:
The toxicity of overexpressing ß-tubulin can be suppressed by overexpression of either -tubulin or Rbl2p (WEINSTEIN and SOLOMON 1990 ; ARCHER et al. 1995 ). Therefore, it is possible that the plp1 mutation upregulates Rbl2p expression to suppress strains carrying excess ß-tubulin. However, plp1 rescues the synthetic lethality of rbl2 tub3 double mutants and rbl2 pac10 (Fig 3B). Therefore, suppression by plp1 is not mediated by Rbl2 function.

plp1 protects cells against high-level overexpression of ß-tubulin:
The toxicity of overexpressed ß-tubulin can be demonstrated in two ways. Both assays demonstrate that plp1 suppresses the effects of even high levels of undimerized ß-tubulin. First, haploid cells carrying an additional chromosomal copy of ß-tubulin under the control of the galactose-inducible promoter show dramatically reduced plating efficiency on galactose plates; only 0.06% of cells form colonies on medium containing galactose compared to on medium containing glucose (Fig 4A). Deletion of PAC10 makes the cells even more sensitive: 0.03% live on galactose. In both of these backgrounds, the plp1 mutation increases plating efficiency 4-fold to 0.27% in PAC10 cells and 300-fold to 18% in pac10 cells.

View larger version (74K): In this window In a new window Download PPT slide   Figure 4. pac10 plp1 rescues ß-tubulin lethality. (A) Strains containing an integrated galactose-inducible ß-tubulin-expressing gene were grown overnight in raffinose and then plated on glucose and galactose media. The percentage survival was determined by comparing the number of colonies on the galactose and glucose media. (B) Strains containing an integrated galactose-inducible ß-tubulin-expressing gene were grown in raffinose to early log phase. At time zero, galactose was added to induce overexpression of TUB2. At the given time points, the cells were plated to glucose (to repress the galactose promoter) and the plating efficiency was determined. (C) Cultures of the four strains shown in B were collected after growth in raffinose and after 5 hr of galactose induction, and immunoblots were used to determine the amount of ß-tubulin and control protein CPY in each strain. All four strains showed comparable fold increases in ß-tubulin.

Second, the plp1 mutation affects the kinetics of ß-tubulin toxicity. Cells were grown to log phase in raffinose, and then, at zero time, galactose was added to induce overexpression of ß-tubulin. At various times, cells were plated to glucose to repress the galactose-inducible promoter and to enable counts of viable cells. Compared to otherwise wild-type cells, pac10 cells lose viability more rapidly (ALVAREZ et al. 1998 ). In both backgrounds, plp1 substantially slowed the rate of cell death (Fig 4B). In cells lacking the prefoldin component Pac10, plp1 has an even more dramatic phenotype.

Both the plating efficiency and kinetic experiments show that deletion of PLP1 reduces the toxicity of excess ß-tubulin to a greater extent in pac10 than in PAC10 cells (Fig 4), even though the pac10 mutation makes cells substantially more sensitive to overexpressed ß-tubulin. Indeed, the pac10 plp1 double mutants are much more resistant to excess ß-tubulin than are wild-type cells. Control experiments demonstrate that the extent of ß-tubulin overexpression was comparable in each of these strains (Fig 4C).

The results described above demonstrate that undimerized ß-tubulin produced by differential levels of expression relative to -tubulin shows substantially lower-than-expected toxicity in plp1 cells. This conclusion applies to both pac10 cells, which contain a modest excess of ß-tubulin, and cells overexpressing ß-tubulin under the control of the strong GAL promoter.

Additional -tubulin does not affect the benomyl phenotype of pac10 plp1:
Our findings raise the possibility that plp1 affects the properties of the excess ß-tubulin. To assay the state of ß-tubulin in the absence of Plp1p, we first asked whether the 40% excess undimerized ß-tubulin in pac10 plp1 cells could interact with -tubulin. The supersensitivity of pac10 cells to benomyl can be rescued by -tubulin on a low-copy plasmid (ALVAREZ et al. 1998 ; GEISSLER et al. 1998 ), because the additional -tubulin binds the excess ß-tubulin to form more heterodimer. Significantly, however, the residual benomyl sensitivity of pac10 plp1 is not rescued by additional -tubulin, even with high-level overexpression from a galactose-inducible promoter (Fig 5). This result suggests that, in the double mutant, the excess ß-tubulin is in a form that is unable to heterodimerize with -tubulin.

View larger version (72K): In this window In a new window Download PPT slide   Figure 5. The excess ß-tubulin in pac10 plp1 cells is in a nontoxic and nonfunctional form. Excess -tubulin does not suppress the benomyl phenotypes of pac10 plp1. Wild-type, pac10, and pac10 plp1 cells containing a plasmid carrying a galactose-inducible TUB1 gene or an empty plasmid were grown in selective media overnight. Serial (1:5) dilutions of a saturated culture were spotted onto rich plates (YPD) to ensure equivalent growth (not shown) and plates containing galactose (Gal) and 10 µg/ml of benomyl. Plates were photographed after 3 days of growth.

Cells deleted for PLP1 have normal levels of tubulin expression but reduced levels of heterodimer:
We next tested whether the loss of PLP1 affects the form of ß-tubulin in a wild-type background. As shown in Fig 3, the levels of - and ß-tubulin in plp1 cells are equivalent to those in wild-type cells. We used gel-filtration chromatography (ABRUZZI et al. 2002 ) to show that, in plp1 cells, about half of both the - and ß-tubulin proteins are present in a heterodimer peak as well as in a much larger form that elutes in the void volume of the column (Fig 6). In wild-type cells, >90% of the - and ß-tubulin is found in a single heterodimer peak. The tubulin in the void volume of the column found in plp1 is most likely aggregated (ABRUZZI et al. 2002 ). This could be a form of nontoxic ß-tubulin present in the form of an aggregate in plp1 cells.

View larger version (12K): In this window In a new window Download PPT slide   Figure 6. plp1 cells contain aggregated tubulin. (Left) Wild-type and (right) plp1 cell extracts were resolved on a Sephacryl 300S column, and fractions were analyzed for - and ß-tubulin by immunoblot. The graphs represent the percentage of the total amount of tubulin in each fraction. -tubulin, solid line; ß-tubulin, dashed line.

The GroEL trap for unfolded proteins suppresses ß-tubulin toxicity:
The results described above suggest that PLP1 function helps convert ß-tubulin to a form that can become either heterodimer or, in the absence of sufficient -tubulin, toxic. To probe the nature of toxic ß-tubulin, we introduced a mutant version of the gene encoding the bacterial GroEL subunit under control of the inducible yeast copper promoter. This GroEL mutant, D87K, binds unfolded polypeptides but is defective in ATP hydrolysis and so does not release them; it has been used previously as a trap for unfolded proteins in yeast (WEISSMAN et al. 1994 ; FARR et al. 1997 ; SIEGERS et al. 1999 ). We asked whether ectopic expression of GroEL D87K affected loss of viability in cells overexpressing ß-tubulin in wild-type and plp1 cells. If overexpressed ß-tubulin has the same potential to become toxic in wild-type and plp1 cells, the trap should rescue to the same extent in the two strains.

The genes encoding ß-tubulin and the GroEL trap were induced for 5 hr, and the cells then were plated to medium permissive for growth but that repressed inducible promoters. Under these conditions, overexpression of ß-tubulin kills 80% of the cells, but when the GroEL trap is co-overexpressed only 35% of the cells are killed (Fig 7). Thus, the presence of the GroEL trap is sufficient to substantially suppress ß-tubulin lethality. These results suggest that the exogenous chaperonin can bind a form of ß-tubulin with the potential to be toxic. However, the presence of the GroEL trap only modestly increases the survival of plp1 cells overexpressing ß-tubulin from 45 to 65% (Fig 7). This difference is not due to a difference in the amount of ß-tubulin bound by the GroEL trap, which is the same in both wild-type and plp1 backgrounds, as determined by co-immunoprecipitation experiments with anti-GroEL (data not shown). Taken together, these results suggest that the ß-tubulin present in a wild-type cell is more likely to become toxic than that in a plp1 cell. The GroEL trap reduces the pool of ß-tubulin in both strains, but since much of the ß-tubulin in plp1 is not or does not become toxic, the decrease in levels does not greatly enhance rescue.

View larger version (20K): In this window In a new window Download PPT slide   Figure 7. GroEL trap suppresses ß-tubulin toxicity and enhances suppression by plp1. Wild-type and plp1 cells containing an integrated galactose-inducible ß-tubulin gene, with and without a plasmid containing a copper-inducible GroEL trap, were grown in raffinose media to saturation. Five hours after galactose and copper were added to induce expression of ß-tubulin and GroEL trap, cells were plated onto glucose. The graph represents the percentage of viable cells.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The experiments described above use ß-tubulin toxicity as a probe to understand the properties and regulation of this essential protein. Previously discovered suppressors of excess ß-tubulin lethality—-tubulin and Rbl2p—act by binding directly to the toxic protein. The screen used here for loss-of-function mutations as suppressors of excess ß-tubulin led to the identification of a protein, Plp1p, which is important in forming toxic ß-tubulin. In cells lacking Plp1p, the consequences of excess ß-tubulin are substantially diminished, especially in the absence of the GimC/PFD complex.

Plp1p mediates most but not all ß-tubulin toxicity:
The suppression of undimerized ß-tubulin toxicity by plp1 is manifest in those strains in which the excess is caused by higher levels of ß-tubulin than -tubulin. That condition occurs in strains deleted for the minor -tubulin gene TUB3, in strains deleted for the prefoldin component PAC10, and in strains that inducibly overexpress the ß-tubulin gene TUB2.

The ability of plp1 to rescue phenotypes caused by excess ß-tubulin could be explained if Plp1p had a role in tubulin expression, so that its deletion reduced expression of ß-tubulin or increased expression of -tubulin. However, plp1 in otherwise wild-type cells has no effect on expression of either tubulin. The levels of ß-tubulin are the same in tub3 and tub3 plp1 cells and essentially identical to wild-type levels (data not shown).

A role for Plp1p in microtubule morphogenesis:
The data suggest that PLP1 functions not in controlling tubulin expression levels but rather in folding ß-tubulin. Our results suggest that Plp1p function is relatively specific for ß-tubulin. If Plp1p were equally important for -tubulin folding, the suppression of modest levels of undimerized ß-tubulin would not occur. Instead, plp1 has no apparent effect on -tubulin folding. The aggregated -tubulin in plp1 cells likely arises from the absence of sufficient folded ß-tubulin with which it can form heterodimers. In plp1 cells, a substantial fraction of the total tubulin is aggregated compared to that in functional heterodimer. Such aggregates can be understood as arising due to inefficient folding of nascent ß-tubulin. Consequently, a proportion of both tubulins is undimerized and so the proteins tend to form aggregates (ABRUZZI et al. 2002 ). That Plp1p affects the proportion of tubulin in functional heterodimer provides a rationale for an activity that, under certain circumstances, is deleterious to the cell.

Where along the ß-tubulin folding pathway does Plp1p act? Two steps in that pathway in yeast have been defined by the role of the CCT chaperonin and the prefoldin complex. The function of Plp1p can be clearly distinguished from the proteins participating in each of those steps. First, unlike plp1, defects in the chaperonin components have deleterious effects on microtubules. Second, deletion of the prefoldin component Pac10 makes cells supersensitive to microtubule depolymerizing drugs and is not viable without the minor -tubulin gene TUB3; plp1 rescues both these phenotypes of pac10. A third possible point of action is in the formation of heterodimers. The properties of Plp1p are also distinct from those of the yeast homologs of proteins that mediate heterodimer formation in vitro. Those proteins, although not essential in S. cerevisiae, may participate in a salvage pathway to rescue dissociated heterodimers (FLEMING et al. 2000 ).

The data suggest that Plp1p functions at an early step in folding. One possibility is that it facilitates the efficient transfer of nascent ß-tubulin polypeptides from the ribosome to the cytosolic chaperonin (Fig 8). As a consequence, in plp1 cells a substantial fraction of tubulin polypeptides is released into the cytoplasm. There, the tubulin polypeptides can still interact with the GimC/PFD complex or the cytosolic chaperonin to be folded, although with lower efficiency. This defect can explain how plp1 rescues the phenotypes of tub3 and pac10 cells, both their benomyl supersensitivity and their synthetic lethal interaction with rbl2: because plp1 cells bring less ß-tubulin to the chaperonin, less ß-tubulin is properly folded and therefore is toxic. The same explanation applies to the protection by plp1 against the large excess of ß-tubulin produced by a strong inducible promoter.

View larger version (19K): In this window In a new window Download PPT slide   Figure 8. Model depicting how Plp1p increases the efficiency of folding of newly synthesized ß-tubulin polypeptides at a step distinct from GimC/prefoldin and TriC.

pac10 plp1 double mutants are much more resistant to ß-tubulin overexpression than are wild-type cells (Fig 4), despite the fact that pac10 alone is much more sensitive. This apparent contradiction can be explained in part in terms of a role for the GimC/prefoldin complex in recruiting free ß-tubulin to the CCT for folding (VAINBERG et al. 1998 ). In addition, it is important to note that the GimC/PFD complex seems to act asymmetrically on - and ß-tubulin: -tubulin levels are decreased more than ß-tubulin levels. Thus, in the absence of the GimC/PFD complex, there is less folded -tubulin than ß-tubulin, leading to an excess of ß-tubulin that is folded, functional, and toxic. Plp1p, however, acts specifically on ß-tubulin, with the consequence that pac10 plp1 mutants, although they accumulate undimerized -tubulin, have an even larger pool of ß-tubulin that is nonfunctional. Insight into the nature of this excess ß-tubulin is provided by the fact that increasing the levels of -tubulin by either low-copy plasmid or high-level expression driven by a strong inducible promoter does not further rescue the benomyl phenotypes of pac10 plp1. In contrast, the toxic ß-tubulin in pac10 cells is completely suppressed by overexpressed -tubulin. This result, and the aggregation of tubulin in plp1 cells, demonstrates that excess ß-tubulin can be in different conformations, depending in part upon whether or not Plp1p is present.

Summary:
The data reported here suggest that Plp1p acts in the ß-tubulin folding pathway to help create heterodimerizable ß-tubulin. In the absence of PLP1, sufficient ß-tubulin is made and folded correctly to support cell viability; however, some is in an aggregated form. It is the reduced efficiency of folding ß-tubulin that promotes aggregate formation in plp1 cells and so rescues them from excess ß-tubulin. The precise role and interactions of Plp1p, and the nature of the toxic form of ß-tubulin, are among the interesting questions that remain.

	ACKNOWLEDGMENTS

We thank P. Phillippsen for pFA vectors, H. Ploegh for anti-CPY, K. Siegers and U. Hartl for the GroEL trap construct, M. Snyder for the transposon library, and N. Kleckner for pNKY51; A. Amon, G. Fink, and R.T. Sauer for critical evaluation of the manuscript; S. Bell and P. Sharp for discussion of experiments; and the members of our laboratory for their valuable contributions. S.L. was supported by a predoctoral fellowship from the Ludwig Cancer Fund and by a predoctoral training grant from 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.

Manuscript received April 5, 2003; Accepted for publication June 18, 2003.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ABRUZZI, K. C., A. SMITH, W. CHEN, and F. SOLOMON, 2002  Protection from free ß-tubulin by the ß-tubulin binding protein Rbl2p. Mol. Cell. Biol. 22:138-147.[Abstract/Free Full Text]

ALANI, E., L. CAO, and N. KLECKNER, 1987  A method for gene disruption that allows repeated use of URA3 selection in the construction of multiply disrupted yeast strains. Genetics 116:541-545.[Abstract/Free Full Text]

ALVAREZ, P., A. SMITH, J. FLEMING, and F. SOLOMON, 1998  Modulation of tubulin polypeptide ratios by the yeast protein Pac10p. Genetics 149:857-864.[Abstract/Free Full Text]

ARCHER, J. E., L. R. VEGA, and F. SOLOMON, 1995  Rbl2p, a yeast protein that binds to ß-tubulin and participates in microtubule function in vivo.. Cell 82:425-434.[Medline]

BARRAL, Y., S. JENTSCH, and C. MANN, 1995  G₁ cyclin turnover and nutrient uptake are controlled by a common pathway in yeast. Genes Dev. 9:399-409.[Abstract/Free Full Text]

BOND, J. F., J. L. FRIDOVICH-KEIL, L. PILLUS, R. C. MULLIGAN, and F. SOLOMON, 1986  A chicken-yeast chimeric ß-tubulin protein is incorporated into mouse microtubules in vivo. Cell 44:461-468.[Medline]

BURKE, D., P. GASDASKA, and L. HARTWELL, 1989  Dominant effects of tubulin overexpression in Saccharomyces cerevisiae.. Mol. Cell. Biol. 9:1049-1059.[Abstract/Free Full Text]

BURNS, N., B. GRIMWADE, P. B. ROSS-MACDONALD, E. Y. CHOI, and K. FINBERG et al., 1994  Large-scale analysis of gene expression, protein localization, and gene disruption in Saccharomyces cerevisiae.. Genes Dev. 8:1087-1105.[Abstract/Free Full Text]

DESAI, A. and T. J. MITCHISON, 1997  Microtubule polymerization dynamics. Annu. Rev. Cell Dev. Biol. 13:83-117.[Medline]

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.[Medline]

ERICKSON, H. P., 1995  FtsZ, a prokaryotic homolog of tubulin? Cell 80:367-370.[Medline]

FARR, G. W., E. C. SCHARL, R. J. SCHUMACHER, S. SONDEK, and A. L. HORWICH, 1997  Chaperonin-mediated folding in the eukaryotic cytosol proceeds through rounds of release of native and nonnative forms. Cell 89:927-937.[Medline]

FLANARY, P. L., P. R. DIBELLO, P. ESTRADA, and H. G. DOHLMAN, 2000  Functional analysis of Plp1 and Plp2, two homologues of phosducin in yeast. J. Biol. Chem. 275:18462-18469.[Abstract/Free Full Text]

FLEMING, J., L. VEGA, and F. SOLOMON, 2000  Function of tubulin binding proteins in vivo.. Genetics 156:69-80.[Abstract/Free Full Text]

FLOOR, E., 1970  Interaction of morphogenetic genes of bacteriophage T4. J. Mol. Biol. 47:293-306.[Medline]

GEISER, J. R., E. J. SCHOTT, T. J. KINGSBURY, N. B. COLE, and 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.[Abstract]

GEISSLER, S., K. SIEGERS, and E. SCHIEBEL, 1998  A novel protein complex promoting formation of functional alpha- and gamma-tubulin. EMBO J. 17:952-966.[Medline]

GUTHRIE, C., and G. FINK, 1991 Guide to Yeast Genetics and Molecular Biology. Academic Press, New York.

HANSEN, W. J., N. J. COWAN, and W. J. WELCH, 1999  Prefoldin-nascent chain complexes in the folding of cytoskeletal proteins. J. Cell Biol. 145:265-277.[Abstract/Free Full Text]

HOYT, M. A., J. P. MACKE, B. T. ROBERTS, and J. R. GEISER, 1997  Saccharomyces cerevisiae PAC2 functions with CIN1, 2 and 4 in a pathway leading to normal microtubule stability. Genetics 146:849-857.[Abstract]

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.[Abstract/Free Full Text]

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.

LONGTINE, M. S., A. MCKENZIE, D. J. DEMARINI, N. G. SHAH, and A. WACH et al., 1998  Additional modules for versatile and economical PCR-based gene deletion and modification in Saccharomyces cerevisiae.. Yeast 14:953-961.[Medline]

MCCALLUM, C. D., H. DO, A. E. JOHNSON, and J. FRYDMAN, 2000  The interaction of the chaperonin tailless complex polypeptide 1 (TCP1) ring complex (TRiC) with ribosome-bound nascent chains examined using photo-cross-linking. J. Cell Biol. 149:591-601.[Abstract/Free Full Text]

MCLAUGHLIN, J. N., C. D. THULIN, S. J. HART, K. A. RESING, and N. G. AHN et al., 2002  Regulatory interaction of phosducin-like protein with the cytosolic chaperonin complex. Proc. Natl. Acad. Sci. USA 99:117-121.

SCHATZ, P. J., F. SOLOMON, and D. BOTSTEIN, 1986  Genetically essential and nonessential -tubulin genes specify functionally interchangeable proteins. Mol. Cell. Biol. 6:3722-3733.[Abstract/Free Full Text]

SCHATZ, P. J., G. E. GEORGES, F. SOLOMON, and D. BOTSTEIN, 1987  Insertions of up to 17 amino acids into a region of -tubulin do not disrupt function in vivo.. Mol. Cell. Biol. 7:3799-3805.[Abstract/Free Full Text]

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.

SIEGERS, K., T. WALDMANN, M. R. LEROUX, K. GREIN, and A. SHEVCHENKO et al., 1999  Compartmentation of protein foldling in vivo: sequestration of non-native polypeptides by the chaperonin-GimC system. EMBO J. 18:75-84.[Medline]

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.

STEARNS, T., M. HOYT, and D. BOTSTEIN, 1990  Yeast mutants sensitive to antimicrotubule drugs define the genes that affect microtubule function. Genetics 124:251-262.[Abstract]

STERNBERG, N., 1976  A genetic analysis of bacteriophage l head assembly. Virology 71:568-582.[Medline]

TIAN, G., S. LEWIS, B. FEIERBACH, T. STEARNS, and H. ROMMELAERE et al., 1997  Tubulin subunits exist in an activated conformational state generated and maintained by protein cofactors. J. Cell Biol. 138:821-832.[Abstract/Free Full Text]

URSIC, D. and M. CULBERTSON, 1991  The yeast homolog to mouse Tcp-1 affects microtubule-mediated processes. Mol. Cell. Biol. 11:2629-2640.[Abstract/Free Full Text]

VAINBERG, I. E., S. A. LEWIS, H. ROMMELAERE, C. AMPE, and J. VANDEKERCKHOVE et al., 1998  Prefoldin, a chaperone that delivers unfolded proteins to cytosolic chaperonin. Cell 93:863-873.[Medline]

VEGA, L., J. FLEMING, and F. SOLOMON, 1998  An -tubulin mutant destabilizes the heterodimer: phenotypic consequences and interactions with tubulin-binding proteins. Mol. Biol. Cell 9:2349-2360.[Abstract/Free Full Text]

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.[Abstract/Free Full Text]

WEISSMAN, J., Y. KASHI, W. FENTON, and A. HORWICH, 1994  GroEL-mediated protein folding proceeds by multiple rounds of binding and release of nonnative forms. Cell 78:693-702.[Medline]

This article has been cited by other articles:

				M. M. Castellano and R. Sablowski Phosducin-Like Protein 3 Is Required for Microtubule-Dependent Steps of Cell Division but Not for Meristem Growth in Arabidopsis PLANT CELL, April 1, 2008; 20(4): 969 - 981. [Abstract] [Full Text] [PDF]

				P. C. Stirling, M. Srayko, K. S. Takhar, A. Pozniakovsky, A. A. Hyman, and M. R. Leroux Functional Interaction between Phosducin-like Protein 2 and Cytosolic Chaperonin Is Essential for Cytoskeletal Protein Function and Cell Cycle Progression Mol. Biol. Cell, June 1, 2007; 18(6): 2336 - 2345. [Abstract] [Full Text] [PDF]

				S. Lacefield, M. Magendantz, and F. Solomon Consequences of Defective Tubulin Folding on Heterodimer Levels, Mitosis and Spindle Morphology in Saccharomyces cerevisiae Genetics, June 1, 2006; 173(2): 635 - 646. [Abstract] [Full Text] [PDF]

				P. C. Stirling, J. Cuellar, G. A. Alfaro, F. El Khadali, C. T. Beh, J. M. Valpuesta, R. Melki, and M. R. Leroux PhLP3 Modulates CCT-mediated Actin and Tubulin Folding via Ternary Complexes with Substrates J. Biol. Chem., March 17, 2006; 281(11): 7012 - 7021. [Abstract] [Full Text] [PDF]

				J. C. Knol, R. Engel, M. Blaauw, A. J. W. G. Visser, and P. J. M. van Haastert The Phosducin-Like Protein PhLP1 Is Essential for G{beta}{gamma} Dimer Formation in Dictyostelium discoideum Mol. Cell. Biol., September 15, 2005; 25(18): 8393 - 8400. [Abstract] [Full Text] [PDF]

				J. Martin-Benito, S. Bertrand, T. Hu, P. J. Ludtke, J. N. McLaughlin, B. M. Willardson, J. L. Carrascosa, and J. M. Valpuesta Structure of the complex between the cytosolic chaperonin CCT and phosducin-like protein PNAS, December 14, 2004; 101(50): 17410 - 17415. [Abstract] [Full Text] [PDF]