Modulation of Tubulin Polypeptide Ratios by the Yeast Protein Pac10p

Corresponding author: Frank Solomon, Department of Biology and Center for Cancer Research, Building E17-Room 220, M.I.T., Cambridge, MA 02139, solomon{at}mit.edu (E-mail).

Communicating editor: D. BOTSTEIN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Normal assembly and function of microtubules require maintenance of the proper levels of several proteins, including the tubulin polypeptides themselves. For example, in yeast a significant excess of ß-tubulin causes rapid microtubule disassembly and subsequent cell death. Even the modest excess of ß-tubulin produced by genetic alterations such as deletion of the minor -tubulin gene TUB3 affects cell growth and can confer microtubule phenotypes. We show here that the levels of the yeast protein Pac10p affect the relative levels of the tubulin polypeptides. Cells deleted for PAC10 have the same phenotypes as do cells that express reduced levels of -tubulin or Rbl2p, two proteins that bind ß-tubulin. Conversely, overexpression of Pac10p enhances the ability of -tubulin or Rbl2p to suppress the lethality associated with excess ß-tubulin. However, Pac10p is itself not a ß-tubulin binding protein. Pac10 null cells show a 30% decrease in the ratio of -tubulin to ß-tubulin. The results suggest that Pac10p modulates the level of -tubulin in the cell, and so influences microtubule morphogenesis and tubulin metabolism.

EARLY steps in the microtubule assembly pathway affect proper folding of the nascent tubulin chains and their incorporation into the heterodimer. Genetic and biochemical evidence demonstrates that folding of - and ß-tubulin is mediated by the Tcp-1p chaperone complex (GAO et al. 1992 ; YAFFE et al. 1992 ; STERNLICHT et al. 1993 ). In vitro, other factors are essential, either to finish the folding reaction or to stabilize the tubulin chains until they are dimerized (GAO et al. 1993 ; MELKI et al. 1996 ; TIAN et al. 1996 ). In the budding yeast Saccharomyces cerevisiae, mutations in presumptive chaperone complex components affect microtubule assembly and function (URSIC and CULBERTSON 1991 ; CHEN et al. 1994 ). However, these studies have not fully demonstrated how tubulin chains fold and assemble in cells.

The interactions of undimerized tubulin chains may have considerable physiological significance. In yeast, genetic configurations that produce higher than wild-type ratios of ß- to -tubulin are toxic (BURKE et al. 1989; KATZ et al. 1990 ; WEINSTEIN and SOLOMON 1990 ). Acute overexpression of ß-tubulin causes rapid, quantitative microtubule disassembly and subsequently a 10⁴-fold decrease in cell viability. In contrast, overexpression of -tubulin does not cause microtubule disassembly and is only slightly toxic (WEINSTEIN and SOLOMON 1990 ). These different properties of - and ß-tubulin imply functional differences between the two proteins. For example, perhaps ß-tubulin sequences are more important for interactions between the -ß tubulin heterodimer and factors essential for microtubule assembly than are -tubulin sequences. Consequently, excess free ß-tubulin could be an effective competitive inhibitor of assembly and thus toxic (WEINSTEIN and SOLOMON 1992 ). The molecular targets of ß-tubulin toxicity are as yet unidentified.

ß-Tubulin lethality is efficiently suppressed by concomitant overexpression of -tubulin, presumably by sequestering the excess ß-tubulin in heterodimer. A screen for other genes that when overexpressed would also rescue ß-tubulin lethality identified three RBL genes; one of them, RBL2, suppresses ß-tubulin lethality as well as overexpressed -tubulin (ARCHER et al. 1995 ). Like -tubulin, Rbl2p binds specifically to ß-tubulin. Cofactor A, a protein originally identified as part of an in vitro assay for ß-tubulin folding (CAMPO et al. 1994 ; GAO et al. 1994 ; TIAN et al. 1996 ), is structurally and functionally homologous to RBL2 (ARCHER et al. 1995 ). Cofactor A is thought to bind to a relatively unfolded form of ß-tubulin; however, in vivo Rbl2p can bind to ß-tubulin both before and after it has been incorporated into heterodimer (ARCHER et al. 1998 ). Thus, the available in vivo evidence does not clearly define a role for Rbl2p in ß-tubulin folding. Indeed, Rbl2p may participate in the formation of heterodimer or it may act as a buffer of free ß-tubulin.

RBL2 is a nonessential gene. rbl2 strains do display modest microtubule phenotypes. We previously showed that some -tubulin mutants require Rbl2p for viability. Because such genetic interactions can identify functions that are redundant with or act in conjunction with those of Rbl2p, we screened for mutations in genes other than -tubulin that are synthetically lethal with rbl2. Here we describe one such gene, PAC10. This gene was previously identified in a screen for genes that are required in the absence of CIN8, which encodes a microtubule motor protein (GEISER et al. 1997 ). We find that pac10 null strains display phenotypes similar to those associated with deletion of RBL2 (ARCHER et al. 1995 ) or of the minor -tubulin gene TUB3 (SCHATZ et al. 1986B ). Unlike -tubulin and Rbl2p, however, Pac10p does not form a complex with ß-tubulin. Instead, our data suggest that PAC10 expression levels affect the ratio of -tubulin to ß-tubulin, probably by modulating the level of -tubulin. This effect explains the several microtubule-related phenotypes of altered Pac10p levels and suggests how Pac10p may function in early steps of microtubule morphogenesis.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains and microbiological techniques:
Genetic manipulations and growth media were standard methods (SHERMAN et al. 1986 ). The strains and plasmids used in this study are listed in Table 1.

Mutagenesis and mutant isolation:
We mutagenized JAY551 (rbl2) with ethylmethane sulfonate resulting in 40% viability. We plated 60,000 cells on media lacking uracil and replica plated to 5-fluoroorotic acid (5-FOA) to select for the ability to lose the plasmid pA21A bearing RBL2. After this first selection we isolated 105 candidates unable to grow on 5-FOA. The 5-FOA sensitivity of eight of those strains was rescued by transformation with plasmid pJA33 bearing RBL2 marked with HIS3. The strains bearing the synthetic lethal mutation were backcrossed to a wild-type strain (FSY183) and segregants were tested for the mutant allele. To test allelism between the synthetic lethal mutations and -tubulin, we crossed mutant strains to PAY60, a derivative of DBY2282 (provided by D. BOTSTEIN, Stanford University), in which the TUB1 locus (linked to the TUB3 locus) is marked with LEU2. If the synthetic lethal mutation is in one of the -tubulin genes, its phenotype should segregate away from the LEU marker.

Immunological techniques:
We followed standard procedures for immunoblots and immunofluorescence (SOLOMON et al. 1992 ), using anti--tubulin antibody #345 and anti-ß-tubulin antibody #206 at a dilution of 1/3500 for the immunoblots; and antibody #206 at 1/2000 for immunofluorescence (WEINSTEIN and SOLOMON 1990 ).

Cloning of RKS2/PAC10:
We used strain PAY3 (rks2-1) as a host to clone RKS2. We transformed these cells with a S. cerevisiae genomic DNA library on a centromeric plasmid marked with URA3 (provided by C. THOMPSON and R. YOUNG, M.I.T.). We tested the 40,000 transformants for recovered resistance to 30 µg/ml benomyl and so identified 30 candidates. We isolated the suppressing plasmids from each, and characterized the inserts by restriction mapping and so identified a region common to all the inserts. Partial DNA sequencing of that region demonstrated that it was identical to PAC10 [(GenBank accession no. U29137 (GEISER et al. 1997 )]. A plasmid, pPA36, carrying a 1.4-kb BamHI-KpnI fragment that includes the entire PAC10 gene was created by cutting pPA1 with BamHI and KpnI and religating this fragment into the backbone carrying PAC10.

Disruption of PAC10 and PAC2:
To disrupt the entire PAC10 open reading frame (ORF), we used PCR to flank the HIS3 gene with the 5'- and 3'-noncoding regions of PAC10 (815 bp upstream of the initiation codon and 718 bp downstream of the termination codon). The PCR primers for the 5'-noncoding region were 5'-TCAGAAGGCAATGCTGAATC-3' and 5'-AGATCTCCAAAGAAAAATAAAGGGCA-3'; and for the 3'-noncoding region, 5'-AGATCTATGTGCGTACAGTTTTC TGC-3' and 5'-GCAGTGGTGATGATGATTGG-3'. The two fragments were cloned into the pGEM-vector (Promega, Madison, WI), generating the plasmid pPA10. The primers create a BamHI site to permit cloning of a BamHI fragment carrying the HIS3 gene. This PAC10::HIS3 fragment was cut from the plasmid and transformed into wild-type diploids (FSY185). We checked the transformants for the correct integration of the disruption fragment at the PAC10 locus by PCR. A His⁺ haploid containing the desired integration was backcrossed against wild-type cells and renamed strain PAY169 (pac10).

A similar approach was used to disrupt the PAC2 gene. The oligonucleotides 5'-TTCTTCTGGTGCAGTCAACG-3' and 5'-GGATCCATCTCTGAAATTCGTTTTGC-3') were used to generate a 1050-bp domain of the 5' region; and GGATCCCCTTTTAGATTGTAAGCGGA-3' and 5'-CAAAGACGGTAAACTAAAACAGCA-3') were used to generate an 800-bp fragment of the 3' region. A pac2 haploid was renamed as strain PAY175. The cin1 strain (JFY206) was provided by J. FLEMING.

Analysis of suppression of ß-tubulin lethality:
We transformed JAY47 with several combinations of TUB1, RBL2, and PAC10 plasmids. To determine the extent of suppression of ß-tubulin lethality, we plated the transformants to galactose (inducing) and glucose (noninducing) media. The extent of suppression is expressed as the percentage of cells growing on galactose vs. glucose media. We made an overexpression version of the PAC10 gene by amplifying the coding region with PCR and cloning the fragment into the SalI, NotI sites of a pGAL-URA3-CEN vector (LIU et al. 1992 ) to create pPA23 (pGal-PAC10). The primers for the PAC10 gene were RKS2/SalI: 5'-GTCGACTATGGACACACTGTTCAACTCCA-3' and RKS2/NotI 5'-GCGGCCGCACAGACACATTATATCTTGAG-3' creating pPA23 (pGAL-RKS2). The construction pPA23 was checked for its ability to rescue the benomyl supersensitivity of pac10 (PAY169) in a galactose-dependent manner. The other plasmids used in the experiment were pDK44 (TUB1-LYS2-CEN) and pJA33 (RBL2-HIS3-CEN).

Sensitivity to ß-tubulin lethality:
We made the diploid strain PAY224 by crossing PAY169 (rks2) and FSY626 (TUB2-LEU2-GalTUB2), then sporulated to generate strains PAY231 (rks2, TUB2-LEU2-GalTUB2) and PAY232 (TUB2-LEU2-GalTUB2). To test the effect of ß-tubulin overexpression, PAY231 and PAY232 cells were grown overnight in raffinose media at 30°. At 0 hr, galactose was added to 2%, and at different time points samples were obtained. To test for viability, we counted the cells in each sample and plated them to glucose plates.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Identification of genes synthetic lethal with rbl2:
Cells lacking Rbl2p have conditional microtubule phenotypes, and rbl2 is synthetically lethal with specific mutant alleles of -tubulin (ARCHER et al. 1995 ). The vertebrate homolog of Rbl2p, cofactor A, may be involved in folding nascent ß-tubulin chains in vitro (GAO et al. 1994 ). To learn more about the cellular functions in which Rbl2p participates, we screened for new mutations that make this gene essential. We mutagenized a rbl2 strain bearing a CEN plasmid (pA21A) expressing genomic RBL2 and the URA3 marker (JAY551; see MATERIALS AND METHODS for details). This screen identified eight independent strains that require wild-type RBL2 for growth. Each of the eight strains is benomyl supersensitive (Ben^ss), two are cold sensitive (at 15°) and one is thermosensitive (at 37°). Backcrossing the mutagenized strains to wild-type cells (FSY183) demonstrates that these conditional phenotypes are recessive. Because rbl2 is known to be synthetically lethal with specific -tubulin mutations, we tested to determine if the double mutants could be rescued by excess -tubulin. In all of the mutant strains, the presence of excess -tubulin provided by genomic TUB1 on a low-copy plasmid (pRB539) relieves the need for RBL2. In seven of the eight strains, the extra copy of TUB1 also fully suppresses the Ben^ss phenotype, as expected if the mutation were in either of the -tubulin genes. However, in one of the strains—PAY1—we noticed that excess -tubulin does not completely restore wild-type growth on benomyl, especially at higher concentrations (30–40 µg/ml) of the drug. In a direct test for allelism with -tubulin, sporulation of the diploid resulting from crossing PAY1 with a strain bearing a LEU2 marker integrated next to TUB1 (PAY60) demonstrates that the benomyl supersensitivity segregates independently of the LEU2 marker. Therefore, the new mutation is unlikely to reside in either -tubulin gene TUB1 or TUB3, which are themselves linked. We provisionally named the mutated locus rks2-1 (RBL2 Knockout Synthetic lethal).

Cloning of RKS2:
To identify the wild-type RKS2 sequence, we transformed PAY3 (rks2-1, rbl2, pJA33) with an S. cerevisiae genomic library marked with URA3 and tested transformants for suppression of the benomyl supersensitivity. About 0.1% of the 4 x 10⁴ transformants were able to grow on 30 µg/ml benomyl. Characterization of several of the suppressing plasmids demonstrated they contained three genomic fragments that shared a single domain. The overlapping region corresponds to a 600-bp ORF that predicts a 199aa protein of 23.1 kD. A mutation in this same sequence previously arose from a screen for genes synthetically lethal with deletion of the nonessential mitotic motor CIN8 (GEISER et al. 1997 ). That report named the sequence PAC10 and described two mutant alleles. Therefore, we renamed the rks2 mutation from our screen pac10-3.

Phenotypes of pac10 mutant cells:
To characterize this presumptive pac10 mutation further, and to establish that the synthetic lethal mutation is indeed allelic to PAC10, we removed the entire PAC10 ORF by integrative transformation in the wild-type diploid strain FSY185. We used PCR to confirm the presence of one wild-type and one disrupted copy of PAC10 in the resulting diploid (see MATERIALS AND METHODS). Sporulation of these heterozygotes produced tetrads containing primarily four viable spores, and the marker identifying the pac10 disruption segregated 2:2. Thus pac10, like pac10-3, is viable. We created a diploid strain, PAY-223, designed to be heterozygous at the PAC10 locus (pac10/pac10-3), homozygous for rbl2, and carrying wild-type RBL2 on a low-copy plasmid marked with the URA3 gene. Sporulation of this strain demonstrated that all segregants require the RBL2 plasmid for viability. These results provide further evidence that pac10-3 is indeed a mutant allele of PAC10. They also show that the null allele of pac10, like pac10-3, is synthetically lethal with rbl2.

Analysis of pac10 cells demonstrates that they display the conditional phenotypes (Ben^ss, moderate Cs^- at 15°) of pac10-3. These phenotypes are similar to those displayed by cells containing a moderate excess of ß-tubulin due to deletion of the minor -tubulin gene, TUB3 (SCHATZ et al. 1986B ). Neither pac10 nor pac10-3 has abnormal microtubules at either 30° or at 15°, as assessed by immunofluorescence.

Suppression of the pac10, rbl2 synthetic lethality by mutant -tubulins:
The phenotypes associated with the pac10 mutants—synthetic lethality with rbl2 and benomyl supersensitivity at modest (20 µg/ml) concentrations of benomyl—are largely suppressed by a low-copy plasmid bearing the major -tubulin gene, TUB1 (Figure 1 and Figure 2). However, analysis of several cold-sensitive mutant -tubulins shows that they vary in their ability to rescue these phenotypes. We transformed PAY189 cells (rbl2, pac10, pCEN-RBL2-URA3) with plasmids bearing tub1 mutant alleles (SCHATZ et al. 1988 ). We assayed for the ability of these mutant genes to support growth in the absence of plasmid-borne wild-type RBL2. The tub1 mutants we tested included representatives from each of the three classes originally described: those arresting with no microtubules (class 1), with too many microtubules (class 2), or with disorganized microtubules (class 3). We found that all of the mutant -tubulins assayed can suppress the lethal phenotype at 30°, which is their permissive temperature. However, at their restrictive temperature (15°), a subset of the mutant -tubulins do not support growth without wild-type RBL2 (Table 2). Interestingly, the particular mutants that fail to suppress (tub1-724, -728, -738, and -759) have two other properties in common. First, all are of class I and arrest with no microtubules. Second, each of these specific -tubulin mutations is synthetically lethal with rbl2 (ARCHER et al. 1995 ).

View larger version (44K): In this window In a new window Download PPT slide   Figure 1. —The rbl2, pac10 synthetic lethal interaction is rescued by overexpression of -tubulin. Haploid cells bearing rbl2 and a CEN plasmid encoding RBL2 and marked with the URA3 gene (top row) can grow normally on YPD or on medium containing 5-FOA. Deletion in the same strain of the PAC10 gene (middle row) causes these same cells to die on 5-FOA. This lethality is efficiently suppressed by the presence of a second plasmid encoding -tubulin (third row).

View larger version (115K): In this window In a new window Download PPT slide   Figure 2. —The benomyl supersensitivity of pac10 cells is suppressed by overexpression of -tubulin. pac10 cells (top row) fail to grow on solid medium containing 20 µg/ml benomyl. The presence of either PAC10 (middle row) or TUB1 (bottom row) on low-copy plasmids restores wild-type growth.

View this table: In this window In a new window   Table 2. Allele-specific suppression of the rbl2, pac10 synthetic lethal interaction by overexpression of -tubulin mutants at restrictive temperature

Overexpression of PAC10 does not have an RBL phenotype:
The requirement for either Pac10p or Rbl2p for vegetative growth could be explained if these two proteins independently carried out similar functions. To address this question, we tested if overexpressed Pac10p, like excess Rbl2p, could rescue cells from the lethality associated with excess ß-tubulin. JAY47 diploid cells carry a third copy of the ß-tubulin gene under the control of the inducible GAL promoter and integrated at the normal TUB2 locus. These cells die rapidly in medium containing galactose (ARCHER et al. 1995 ). High levels of either -tubulin or Rbl2p rescue these cells nearly completely, and even a single extra copy of either gene provides significant rescue (10²-fold relative to unsuppressed strains). However, PAC10 under control of its own promoter or of the galactose promoter has no detectable effect on ß-tubulin lethality.

The ability of Rbl2p to bind ß-tubulin is likely to reflect some aspect of its function in vivo. However, we are unable to detect any physical interaction between Pac10p and ß-tubulin or -tubulin, even when both proteins are overexpressed. We searched for such complexes in extracts from cells expressing either the His₆- or HA-tagged versions of Pac10p. We analyzed those extracts using Ni-NTA beads to bind the His₆-Pac10p, or by immunoprecipitation with antibodies against the HA epitope, -tubulin or ß-tubulin. In each case, we failed to find specific association between either tubulin polypeptide and Pac10p. Both of the modified versions of Pac10p complement the pac10 null phenotype, and therefore are functional. Under similar conditions, we can isolate Rbl2p-ß-tubulin complexes (ARCHER et al. 1998 ) as well as the -/ß-tubulin heterodimer. These results suggest that Pac10p does not form a stable complex with ß-tubulin.

Levels of Pac10p and sensitivity to ß-tubulin:
Several genetic and physiological experiments show that cells are sensitive to perturbations in the balance between - and ß-tubulins. Parallel analyses suggest that Pac10p levels affect that balance. First, overexpression of ß-tubulin kills pac10 cells much more rapidly than wild-type cells (Figure 3). Four hours after induction of ß-tubulin overexpression, viability of pac10 cells is 10-fold lower than that of wild-type cells. This supersensitivity to excess ß-tubulin is comparable to that conferred by deletion of RBL2 (ARCHER et al. 1995 ).

View larger version (14K): In this window In a new window Download PPT slide   Figure 3. —pac10 cells are more supersensitive to over-expression of ß-tubulin. Haploid cells, either "wild type" (PAY232) or "pac10" (PAY231) and containing an extra copy of the ß-tubulin gene under the inducible GAL promoter were grown overnight in selective raffinose media at 30°. At time 0 hr, galactose (final concentration, 2%) was added. At various times, aliquots of both cultures were counted for cell number, and appropriate fractions plated on glucose-containing medium. "Viability" represents the fraction of cells counted that gave rise to colonies.

Second, increased levels of Pac10p enhance the ability of both -tubulin and Rbl2p to rescue cells from ß-tubulin overexpression (Figure 4). Typically, cells containing GAL-TUB2 form colonies on galactose with 0.01% of the efficiency of cells plated on glucose. The presence of an extra copy of either TUB1 or RBL2 under control of their own promoters increases that ratio to about 2%, whereas overexpression of PAC10 itself has no effect on survival on galactose (Figure 4). However, concomitant overexpression of PAC10 enhances the ability of an extra copy of either RBL2 or TUB1 to promote growth in the presence of excess ß-tubulin. When PAC10 is present on a low-copy plasmid and under control of its own promoter, the percentage of viable colonies on galactose increases by about twofold when co-overexpressed with TUB1 or RBL2. Co-overexpression of even higher levels of Pac10p, achieved using the galactose-inducible promoter, increases the viability by about eightfold. These results suggest that, although Pac10p cannot itself suppress ß-tubulin lethality (Figure 4), it can enhance the ability of other genes to do so.

View larger version (14K): In this window In a new window Download PPT slide   Figure 4. —Pac10p enhances suppression of ß-tubulin lethality by Rbl2p and Tub1p. JAY47 cells (diploid cells containing a third intergrated copy of TUB2 under control of the GAL promoter) were transformed with several combinations of the plasmids encoding TUB1 (pRB539), RBL2 (pA21A), PAC10 (pPA36), and GAL-PAC10 (pPA23) (Table 1). Aliquots were withdrawn from exponential phase cultures in glucose media, and the extent of suppression was calculated as a percentage of cells on galactose (inducing) vs. glucose (noninducing) plates that could form colonies. CONTROL is JAY47 transformed with YCpGAL. The control, and JAY47 cells expressing GAL-PAC10, gave ~0.01% colonies on galactose vs. glucose.

PAC10 influence on tubulin levels:
Many of the consequences of altered Pac10p levels described above are consistent with the idea that Pac10p affects the activity of the tubulin chains in vivo. Accordingly, we measured - and ß-tubulin in pac10 and wild-type cells using immunoblots. The results of a typical experiment are shown in Table 3. We find that the levels of - and ß-tubulin are reduced in pac10 cells compared to wild type. However, the decrease in -tubulin is greater, so the resulting ratio of -tubulin to ß-tubulin is ~30% lower in the mutants. The ratio is restored to its wild-type value when the mutant is transformed with a low-copy plasmid carrying the PAC10 gene. This relationship between the ratio of tubulin chains and the presence of PAC10 was found in four independent experiments.

View this table: In this window In a new window   Table 3. Absence of Pac10p decreases the -tubulin/ß-tubulin ratio in the cell

Genetic interactions of PAC10:
The data described in Table 3 provide a rationale for the synthetic lethality between pac10 and rbl2. The absence of the ß-tubulin binding activity of Rbl2p would be expected to enhance the cells' sensitivity to the imbalance in -tubulin to ß-tubulin produced by the absence of Pac10p. These relationships also rationalize the suppression of the pac10, rbl2 synthetic lethality by TUB1 (Figure 1), because an extra copy of that gene should provide more -tubulin.

A direct test of this model is to analyze the effects of pac10 in other genetic backgrounds expected to alter the ratio of - to ß-tubulin. We showed previously that tub3 strains are viable, but supersensitive to benomyl (KATZ et al. 1990 ). Tub3p contributes ~15% of the cells' -tubulin (SCHATZ et al. 1986B ), so the properties of the tub3 strain are explicable in terms of excess ß-tubulin. We transformed the tub3 strain PAY290 with the PAC10::HIS3 fragment of pPA10 and selected for strains that had stably integrated the HIS3 marker. The majority of those isolates were unable to lose the pPA46 plasmid (pTUB3-CEN-URA3). We confirmed that in those strains the chromosomal copy of PAC10 had been disrupted using PCR Southerns. In the His⁺ isolates that could lose the plasmid, the disruption fragment integrated elsewhere in the genome. These results demonstrate that the double mutant pac10, tub3 is not viable, probably because of the presence of excess ß-tubulin in these cells.

We also have determined whether other genes thought to participate in tubulin polypeptide metabolism interact with PAC10. In particular, vertebrate homologs of Cin1p and Pac2p are essential for the chaperone-mediated incorporation of denatured ß-tubulin into -/ß-tubulin heterodimers in vitro (TIAN et al. 1996 , TIAN et al. 1997 ). Disruptions of both of these genes are lethal in strains lacking the Cin8p mitotic motor (GEISER et al. 1997 ), as is pac10. Cin1p also has been implicated in ß-tubulin folding in vivo (HOYT et al. 1997 ). We created diploid strains by crossing pac10 haploids (PAY170) with strains bearing deletions of either cin1 (JFY209) or pac2 (PAY175). The resulting diploids also contained pPA36 (pPAC10-CEN-URA3). Spores containing the double deletions are viable as long as the plasmid is maintained, but they are unable to grow on medium containing 5-FOA. This requirement for the PAC10 plasmid demonstrates that both cin1 and pac2 are synthetically lethal with pac10. A low-copy plasmid containing the TUB1 gene rescues pac10, pac2 and pac10, cin1 cells (data not shown), suggesting that these synthetic lethal interactions depend at least in part on the consequences of excess ß-tubulin. The results suggest that these gene products all impinge upon the same essential function, and that their proper stoichiometry is important for cell growth.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

PAC10 originally was identified as a gene required for viability in the absence of the mitotic motor protein Cin8p (GEISER et al. 1997 ). However, like some of the other genes so identified, Pac10p does not appear to have motor functions, so its absence could act indirectly to exacerbate the sublethal consequences of a CIN8 deletion. Our independent identification of PAC10 as important for cellular functions involving the tubulin chains supports this view. The results presented above also give some insight into those functions and how PAC10 may participate in them. Specifically, we find that cellular -tubulin levels, and consequently the -/ß-tubulin ratio, are affected by levels of Pac10p. Previous studies demonstrate that a depressed ratio of -/ß-tubulin affects microtubule function adversely. We hypothesize that this defect in combination with either the absence of the Cin8p motor protein or of the Rbl2p ß-tubulin binding protein may severely disrupt essential microtubule functions.

Levels of Pac10p are important for in vivo microtubule functions:
In the absence of Pac10p, cells become supersensitive to the microtubule depolymerizing drug benomyl, also a property of cells that have a small deficit in -tubulin (KATZ et al. 1990 ). Similarly, pac10 nulls are dependent upon the presence of the ß-tubulin-binding protein Rbl2p for growth. Both of those phenotypes are substantially suppressed by excess -tubulin. Those data are explicable if a consequence of the absence of Pac10p is a decrease in -tubulin levels, and thus a reduced capacity to bind ß-tubulin and so suppress its toxic effects. That deletion of PAC10, like deletion of RBL2, renders cells supersensitive to ß-tubulin overexpression supports that interpretation.

However, Pac10p does not form a stable complex with ß-tubulin. We cannot detect a physical association between Pac10p and either tubulin chain under conditions where we can readily isolate both Rbl2p-ß-tubulin complexes and the -ß tubulin heterodimer itself. An in vivo test of ß-tubulin binding also fails for Pac10p: even GAL-induced expression of PAC10 does not increase the ability of cells to survive induced overexpression of ß-tubulin under conditions in which both -tubulin and Rbl2p act as strong suppressors. Therefore, the effect of deletion of PAC10 on sensitivity to ß-tubulin is likely exercised indirectly.

Deletion of PAC10 changes the stoichiometry of the tubulin chains. The levels of both tubulin chains are decreased in the mutant cells, but the decline in -tubulin is greater, so that the /ß ratio decreases by 30%. The balance of tubulin polypeptides is tightly regulated at the level of protein. For example, cells carrying a single extra copy of TUB1 do display a proportional increase in the amount of -tubulin mRNA, but the -tubulin polypeptide level is very nearly the same as in wild-type cells (KATZ et al. 1990 ). Presumably, the -tubulin synthesized in these cells that is in excess of the ß-tubulin complement is unstable and degraded. Perhaps in the case of pac10 cells, then, the decreased levels of -tubulin result in undimerized ß-tubulin. The diminished /ß tubulin ratio could explain the several phenotypes of pac10 nulls and the ability of -tubulin overexpression to suppress those phenotypes.

ß-tubulin lethality and its suppression:
Excess ß-tubulin is much more toxic than either excess -tubulin or excess heterodimer (BURKE et al. 1989; WEINSTEIN and SOLOMON 1990 ). Presumably, undimerized ß-tubulin, but not undimerized -tubulin, competes with the heterodimer for binding to factors essential for microtubule assembly and cell growth (WEINSTEIN and SOLOMON 1992 ). This model is formally analogous to the balance of components hypothesis, which illuminated the consequences of altered stoichiometries of components in phage morphogenesis (FLOOR 1970 ; STERNBERG 1976 ). We do not yet know the identity of the targets of free ß-tubulin.

The suppression of ß-tubulin lethality by -tubulin or Rbl2p is likely to be based on their ability to bind the free ß-tubulin. Although the details of these interactions are not understood, we do know that the suppression has at least one striking feature. Cells containing ß-tubulin, and a low-copy plasmid bearing either RBL2 or TUB1 under control of their own promoters, show about 1% suppression of ß-tubulin lethality, 100-fold greater than the 0.01% of the cells without the plasmid. Surprisingly, those suppressed cells form colonies the same size as those formed by wild-type cells. Clearly, then, nearly all of the products of each mitosis must be viable. If the proportion of viable mitotic products was lower—for example, 1%—the colonies would be much smaller. This behavior may mean that once cells pass over a threshold event, they can survive excess ß-tubulin. According to this model, the presence of the suppressor increases the probability that they will pass over such a threshold.

This model also provides us with a way of thinking about the effects of increased PAC10 expression on survival of excess ß-tubulin. We note that overexpression of PAC10 enhances the suppression of ß-tubulin lethality by modest increases in the levels of Rbl2p and -tubulin, increasing survival by two- to eightfold. Although overexpression of PAC10 is itself not sufficient to increase survival, it may provide sufficient -tubulin to act cooperatively with excess Rbl2p or -tubulin.

The molecular role of PAC10:
Pac10p and Rbl2p do not appear to have redundant functions. We do not yet know in what way Pac10p acts to affect levels of -tubulin protein. It shows no structural relationship to transcription factors, and thus is unlikely to affect -tubulin mRNA synthesis. As noted by its original identifiers, Pac10p does have homologs in other organisms, including humans where it is believed to bind to a tumor suppressor gene product that itself has no obvious homolog in S. cerevisiae (GEISER et al. 1997 ). It may interact with -tubulin mRNA or protein to stabilize them, although we have been unable to identify a stable complex with the latter. It may also be involved in folding of -tubulin, although no homolog among the proteins essential in the in vitro assay for such activities is known (TIAN et al. 1996 ). Experiments to distinguish among those possibilities are in progress.

	ACKNOWLEDGMENTS

We thank S. SANDERS (M.I.T.) and the members of our laboratory for valuable discussions. P.A. was supported by a postdoctoral training fellowship from Fundacion Ramon Areces (Spain). A.S. and J.F. were supported in part by a predoctoral training grant from the National Institutes of Health to the M.I.T. Department of Biology. Work in our laboratory is supported by a grant from the National Institute of Medical Sciences.

Manuscript received September 15, 1997; Accepted for publication February 26, 1998.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ARCHER, J. E., 1996 Analysis of microtubule morphogenesis in vivo. Ph.D. Thesis, Massachusetts Institute of Technology, Cambridge, MA.

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].

ARCHER, J. E., M. MAGENDANTZ, L. VEGA, and F. SOLOMON, 1998  Formation and function of the Rbl2p-ß-tubulin complex. Mol. Cell. Biol. in press.

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

CAMPO, R., A. FONTALBA, L. SANCHEZ, and J. ZABALA, 1994  A 14kDa release factor is involved in GTP-dependent beta-tubulin folding. FEBS Lett. 353:162-166[Medline].

CHEN, X., D. SULLIVAN, and T. HUFFAKER, 1994  Two yeast genes with similarity to TCP-1 are required for microtubule and actin function in vivo.. Proc. Natl. Acad. Sci. USA 91:9111-9115[Abstract/Free Full Text].

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

GAO, Y., J. THOMAS, R. CHOW, G.-H. LEE, and N. COWAN, 1992  A cytoplasmic chaperonin that catalyzed beta-actin folding. Cell 69:1043-1050[Medline].

GAO, Y., I. VAINBERG, R. CHOW, and N. COWAN, 1993  Two cofactors and cytoplasmic chaperonin are required for the folding of - and ß-tubulin. Mol. Cell. Biol. 13:2478-2485[Abstract/Free Full Text].

GAO, Y., R. MELKI, P. WALDEN, S. LEWIS, and C. AMPE et al., 1994  A novel cochaperonin that modulates the ATPase activity of cytoplasmic chaperonin. Cell 125:989-996.

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].

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].

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.

KIRKPATRICK, D. and F. SOLOMON, 1994  Overexpression of yeast homologs of the mammalian checkpoint gene RCC1 suppresses the class of -tubulin mutations that arrest with excess microtubules. Genetics 137:381-392[Abstract].

LIU, H., J. KRIZEK, and A. BRETSCHER, 1992  Construction of a GAL1-regulated yeast cDNA expression library and its application to the identification of genes whose overexpression causes lethality in yeast. Genetics 132:665-673[Abstract].

MELKI, R., H. ROMMELAERE, R. LEGUY, J. VANDEKERCKHOVE, and C. AMPE, 1996  Cofactor A is a molecular chaperone required for beta-tubulin-folding: functional and structural characterization. Biochemistry 35:10422-10435[Medline].

SCHATZ, P., L. PILLUS, F. GRISAFI, F. SOLOMON, and D. BOTSTEIN, 1986a  Two functional alpha tubulin genes of the yeast Saccharomyces cerevisiae encode divergent proteins. Mol. Cell. Biol. 6:3711-3721[Abstract/Free Full Text].

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

SCHATZ, P., F. SOLOMON, and D. BOTSTEIN, 1988  Isolation and characterization of conditional-lethal mutation in the TUB1 -tubulin gene of the yeast Saccharomyces cerevisiae. Genetics 120:681-695[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.

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, New York.

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

STERNLICHT, H., G. FARR, M. STERNLICHT, J. DRISCOLL, and K. WILLISON et al., 1993  The t-complex polypeptide 1 complex is a chaperonin for tubulin and actin in vivo.. Proc. Natl. Acad. Sci. USA 90:9422-9426[Abstract/Free Full Text].

TIAN, G., Y. HUANG, H. ROMMELAERE, J. VANDEKERCKHOVE, and C. AMPE et al., 1996  Pathway leading to correctly folded ß-tubulin. Cell 86:287-296[Medline].

TIAN, G., S. LEWIS, B. FEIERBACH, R. 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].

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].

WEINSTEIN, B. and F. SOLOMON, 1992  Microtubule assembly and phage morphogenesis: new results and classical paradigms. Mol. Microbiol. 6:677-681[Medline].

YAFFE, M., G. FARR, D. MIKLOS, A. HORWICH, and M. STERNLICHT et al., 1992  TCP-1 complex is a molecular chaperone in tubulin biogenesis. Nature 358:245-248[Medline].

This article has been cited by other articles:

				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]

				S. Lacefield and F. Solomon A Novel Step in {beta}-Tubulin Folding Is Important for Heterodimer Formation in Saccharomyces cerevisiae Genetics, October 1, 2003; 165(2): 531 - 541. [Abstract] [Full Text] [PDF]

				C. Her, X. Wu, M. D. Griswold, and F. Zhou Human MutS Homologue MSH4 Physically Interacts with von Hippel-Lindau Tumor Suppressor-binding Protein 1 Cancer Res., February 15, 2003; 63(4): 865 - 872. [Abstract] [Full Text] [PDF]

				K. C. Abruzzi, A. Smith, W. Chen, and F. Solomon Protection from Free {beta}-Tubulin by the {beta}-Tubulin Binding Protein Rbl2p Mol. Cell. Biol., January 1, 2002; 22(1): 138 - 147. [Abstract] [Full Text] [PDF]

				J. A. Fleming, L. R. Vega, and F. Solomon Function of Tubulin Binding Proteins in Vivo Genetics, September 1, 2000; 156(1): 69 - 80. [Abstract] [Full Text]

				T. M. Sandrock, S. M. Brower, K. A. Toenjes, and A. E. M. Adams Suppressor Analysis of Fimbrin (Sac6p) Overexpression in Yeast Genetics, April 1, 1999; 151(4): 1287 - 1297. [Abstract] [Full Text]

				E. Grishchuk and J. McIntosh Sto1p, a fission yeast protein similar to tubulin folding cofactor E, plays an essential role in mitotic microtubule assembly J. Cell Sci., January 6, 1999; 112(12): 1979 - 1988. [Abstract] [PDF]

				L. R. Vega, J. Fleming, and F. Solomon An alpha -Tubulin Mutant Destabilizes the Heterodimer: Phenotypic Consequences and Interactions with Tubulin-binding Proteins Mol. Biol. Cell, September 1, 1998; 9(9): 2349 - 2360. [Abstract] [Full Text]