INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Microtubule functions in Saccharomyce cerevisiae are apparently simpler than those in higher eukaryotes (reviewed in Solomon, 1991). Previously identified microtubule-dependent functions in this yeast are chromosome separation during mitosis, as well as nuclear migration during mitosis and mating. However, within this relatively limited repertoire, microtubules perform complex functions, forming reversible associations with the cell cortex at precise locations and times during the cell cycle (Goode et al., 1999; reviewed in Heil-Chapdelaine et al., 1999; Miller and Rose, 1999; Adames and Cooper, 2000; Heil-Chapdelaine et al., 2000; Korinek et al., 2000; Lee et al., 2000; Yeh et al., 2000; Farkasovsky and Küntzel, 2001).

In S. cerevisiae, cytokinesis takes place at the mother-bud neck. Therefore, for equal distribution of chromosomes, the nucleus must first migrate to the bud neck. The mitotic spindle then aligns along the mother-bud axis. This preanaphase nuclear migration and alignment of the spindle require associations between astral microtubules and the cortex of the cell. As the cell enters anaphase, the spindle elongates and is translocated into the bud neck. This second nuclear migration process also requires astral microtubule-cortex interactions.

Several proteins have been identified that assist in both phases of nuclear migration during mitosis. Some proteins (i.e., Kar9p, Bim1p/Yeb1p, Num1p) localize to the tips of the mother and/or bud cortex and capture astral microtubule ends. This microtubule-cortex interaction is thought to move the nucleus to the bud neck. Other nuclear migration proteins (i.e., Dyn1p, Act5p/Arp1p, Nip100p) are required for spindle movement into the bud neck via sliding of astral microtubules along the surface of the cortex. Both types of astral microtubule-cortex interactions, involving either microtubule tips or microtubule sides, are reversible.

To further investigate microtubule functions during the cell cycle, we constructed a yeast strain such that the sole source of -tubulin protein was a mutant form in which cys-377 was changed to serine. We chose this residue for site-directed mutagenesis because it corresponds to the primary site of palmitoylation identified in porcine brain -tubulin (Ozols and Caron, 1997), and it is conserved in all known sequences of -tubulin. In addition, palmitoylation is a reversible posttranslational modification that is found in membrane-associated proteins but not cytosolic proteins.

We show here that site-directed mutagenesis of cys-377 resulted in astral microtubule defects during anaphase that were strikingly similar to a class of mutations in nuclear migration proteins that are involved in astral microtubule-cortex interactions (i.e., Kar9p, Bim1p/Yeb1p, dynein, dynactin complex proteins). In addition, cys-to-ser mutation of cys-377 reduced in vivo palmitoylation of -tubulin by >60%. These data suggest that cys-377 of -tubulin plays a role in astral microtubule-dependent functions during mitosis, perhaps through protein-protein interactions with known nuclear migration proteins, and possibly through lipid-protein interactions via palmitoylated -tubulin.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Reagents

Cerulenin (Sigma, St. Louis, MO) was stored at 20°C as a 3 mg/ml stock solution in dimethyl sulfoxide. For labeling, either [9,10-³H]palmitate (specific activity 52 Ci/mmol; PerkinElmer Life Science Products, Boston, MA) or [1-¹⁴C]palmitate (specific activity 55 mCi/mmol; PerkinElmer Life Science Products) was dried down in a Savant Speedvac evaporator and redissolved in ethanol to form a 100× stock solution. 4',6'-Diamidino-2-phenylindole (DAPI; Sigma) was stored at 20°C as a 1-mg/ml stock solution in H₂O. Benomyl (Sigma) was stored at 20°C as a 10-mg/ml solution in dimethyl sulfoxide. A 10× Calcofluor stock (1 mg/ml in H₂O; Sigma) was made immediately before use. Rhodamine-phalloidin (Molecular Probes, Eugene, OR) was stored at 20°C as a 200 U/ml solution in methanol.

Strains and Plasmids

Yeast strains and plasmids are listed in Table 1. We used standard methods for yeast manipulations (Rose et al., 1990; Solomon et al., 1992).

Construction of Haploid tub1 Mutant Strains

The TUB1:LEU2:CEN4:ARS1:ampR plasmid pRB539 was mutated with the use of a modification of the Transformer Site Directed Mutagenesis kit (CLONTECH, Palo Alto, CA). Due to the large size of the parental plasmid (~11 kb), we added two primers in addition to the mutation and selection primers, to synthesize the tub1 plasmids. The two additional primers were AGTCATCGAATTTGATTCTGTGCGATAGCG and AGCGCTTCGTTAATACAGATGTAGGTGTTCC and the selection primer was TGCGGTATTTCACACC-GCCCATGGTGCACTCTC. The two additional primers, the selection primer, and the mutagenic primer bound equidistantly around pRB539. Primers used to mutate TUB1 were designed to change amino acid 377 or 214 from cysteine to serine. To create pLJS377, we used the primer CACAATTGGCTACTGTGGATAGGGCCGTCTC-TATGTTGTCAAATACC, which changed codon 377 from TGC (Cys) to AGC (Ser) and introduced a silent mutation that removed an MscI site. To create pLJS214, we used the primer GCTATTTACGACATGAGCAAAAGAAACTTGGACATCCCAAGACC, which changed codon 214 from TGT (Cys) to TCT (Ser) and introduced a silent mutation that removed an EcoRV site. Plasmids that lost the appropriate restriction site after mutagenesis were sequenced to confirm that the correct tub1 mutations were introduced.

To create the double mutant plasmid (pJB2137), pLJS214 was cut with HindIII. A fragment, containing all of the vector and a majority of the -tubulin gene, was purified; this fragment was missing the 3' end of the gene, including the coding region for cys-377. pLJS377 was cut with HindIII, and the fragment containing the C377S mutation was purified. The C377S fragment was ligated into the C214S fragment. The resulting construct contained a deletion of some of the 3'-noncoding region of TUB1 in the vector. The plasmid was sequenced to confirm that the correct tub1 mutations were introduced.

Four plasmids (TUB1 and the three tub1 plasmids) were transformed separately into the haploid strain DBY2384 that was deleted for both TUB1 and TUB3, the two -tubulin genes in S. cerevisiae. Before transformation, strain DBY2384 was kept alive by a 2 µ:TUB3:URA3 plasmid to provide -tubulin function. Plasmid shuffle (Boeke et al., 1987) was used to replace the TUB3 plasmid with either TUB1 or tub1 plasmids. Transformants were streaked onto 5-fluroorotic acid (5-FOA) plates to select for loss of the TUB3 plasmid (Boeke et al., 1984). Resulting strains contained as their sole source of -tubulin either TUB1 (strain LTY430), C214S tub1 (strain LTY496), C377S tub1 (strain LTY429), or C214S/C377S tub1 (strain JBY212).

Construction of Diploid C377S tub1 Overexpressor Strain

To construct the C377S tub1 overexpressor plasmid (pJFS377), pQX3, which contains GAL1-10-TUB1, was cut with BstEII and BsrGI. A fragment, containing all of the vector and a majority of the TUB1 gene, was purified. This fragment was missing the 3' end of the TUB1 gene, including the coding region for cys-377. pLJS377 was cut with BstEII and BsrGI, and the fragment containing the C377S tub1 mutation was purified. The C377S tub1 fragment was ligated into the GAL1-10-TUB1 fragment. The resulting construct was sequenced to confirm that the correct tub1 mutation was introduced.

The diploid strain FSY279 was streaked onto 5-FOA plates to select for loss of the GAL1-10-TUB1 plasmid (pQX3). Cells were then transformed with three different GAL1-10-C377S tub1 plasmids. Transformed strains were maintained in synthetic complete (SC) medium (U, L). To assess galactose induction of Tub1p and C377S tub1p synthesis, strain FSY279 (containing the GAL1-10-TUB1 plasmid) and strains containing the GAL1-10-C377S tub1 plasmids were grown for up to 4 h in SC (U, L) and 2% galactose. At several time points, cells were lysed and extracts were assayed for relative levels of - and -tubulin by SDS-PAGE and immunoblot analysis. Strain JFY2705 produced similar levels of -tubulin as strain FSY279 at all time points. After 4 h in the presence of galactose, -tubulin levels increased by approximately fivefold and -tubulin increased by approximately twofold in both strains.

Microscopy

Microtubules were visualized with the use of the rat monoclonal anti-tubulin antibody YOL1/34 and fluorescein isothiocyanate-labeled goat anti-rat antibody (Rose et al., 1990; Solomon et al., 1992). Nuclei were visualized by staining with DAPI. Bud scars were visualized by staining with Calcofluor (Rose et al., 1990). Actin filaments and cortical actin patches were visualized by staining with rhodamine-phalloidin (Miller et al., 1999). Cells were viewed at the Center for Biomedical Imaging Technology, University of Connecticut Health Center, Farmington, CT, with a Zeiss Axiovert 100 microscope equipped with a 63×/1.25 or 63×/1.40 objective lens. Images were recorded with a cooled charged-coupled device camera (Photometrics, Phoenix, AZ). Differential interference and phase contrast microscopy was performed with the same system.

Immunoblot Analysis

Immunoblot analysis of - and -tubulin levels was performed with rabbit polyclonal anti--tubulin (#345) and anti--tubulin (#206) antibodies (Weinstein and Solomon, 1990), with the use of enhanced chemiluminescence detection (Amersham Pharmacia Biotech, Arlington Heights, IL).

Benomyl Sensitivity

To examine benomyl sensitivity, serial dilutions of cells were placed onto YPD plates containing benomyl (0-40 µg/ml) at 10 µg/ml increments. Plates were incubated at various temperatures until control plates, containing no benomyl, showed confluent growth.

In Vivo Palmitoylation of Yeast Proteins and Immunoprecipitation of Tubulin

Strains FSY279 and JFY2705 were grown overnight at 30°C in SC medium containing 2% raffinose. At the zero time point, cells (2 × 10⁸) were inoculated into 1 ml of YP medium containing 2% galactose. The pH of the media was adjusted to 6.8; this has been found to increase uptake of exogenous palmitate (Powers et al., 1986). To reduce endogenous levels of palmitate, cerulenin was added to a final concentration of 3 µg/ml (Funabashi et al., 1989). Incubation of FSY279 cells with cerulenin for 4 h did not affect the percentage of large-budded cells, nor did it alter the extent or pattern of microtubules (Caron, unpublished data). After 3 h at 30°C, [³H]palmitate (1 mCi/ml) was added. In one set of experiments, [¹⁴C]palmitate (0.1 mCi/ml) replaced [³H]palmitate because carbon 14 is a stronger beta-emitter than tritium. Cells were incubated for an additional hour and then washed once in ice-cold lysis buffer (phosphate-buffered saline, 1% Triton X-100, 0.5% SDS, 1 mM -mercaptoethanol). Cell pellets were frozen in dry ice/ethanol, and resuspended in 300 µl of lysis buffer containing protease inhibitors (Caron, 1997). An equal volume of glass beads was added, and cells were broken by vortexing six times in 1-min bursts followed by chilling on ice. Insoluble material was removed by centrifugation at 14,000 × g for 10 min at 4°C. Protein concentrations of resulting supernatants were determined with the DC Protein Assay kit (Bio-Rad, Richmond, CA).

To examine [³H] or [¹⁴C]palmitoylated proteins in whole cell extracts, yeast proteins were precipitated with chloroform/methanol (Wessel and Flügge, 1984), resuspended in sample buffer (Laemmli, 1970), and analyzed by one-dimensional (1D)-PAGE and fluorography.

To immunoprecipitate either - or -tubulin, cell lysates (1 mg) were diluted fivefold with immunoprecipitation buffer (phosphate-buffered saline, 1% Triton X-100, 1 mM EGTA, and protease inhibitors) and incubated with an excess of the mouse monoclonal antibody A1BG7 for -tubulin or B1BE2 for -tubulin (Vega et al., 1998). Under these lysis and immunoprecipitation conditions, - and -tubulin did not coimmunoprecipitate (Caron et al., 1985; Caron, unpublished data). Immunoprecipitated proteins were subjected to 1D-PAGE (Laemmli, 1970) with the use of dithiothreitol (10 mM) as the reducing agent. Gels were stained with Coomassie blue, photographed with an -Innotech IS-1000 Digital Imaging System, and fluorographed with Kodak X-OMAT AR x-ray film at 70°C. Approximately 10% of the immunoprecipitated protein was subjected to immunoblot analysis to determine - and -tubulin levels.

In Vivo Palmitoylation of Yeast Proteins and DEAE Chromatography of Tubulin

Strains FSY279 and JFY2705 were grown overnight at 30°C in SC medium containing 2% raffinose. At the zero time point, cells (2 × 10⁸) were inoculated into 20 ml of YP medium (pH 6.8) containing 2% galactose. Cerulenin was added to a final concentration of 3 µg/ml. After 4 h at 30°C, [³H]palmitate (0.1 mCi/ml) was added. Cells were incubated for an additional hour and then washed twice with ice-cold distilled water, frozen in dry ice/ethanol, and resuspended in 300 µl of lysis buffer (0.1 M piperazine-N,N'-bis(2-ethanesulfonic acid) [PIPES], pH 6.8, 10 mM MgSO₄, 2 mM EGTA, 0.1 mM GTP, 5 mM dithiothreitol, 1% Triton X-100) with protease inhibitors (1 µg/ml antipain, 1 µg/ml aprotinin, 2 µg/ml leupeptin, 2 µg/ml pepstatin A, 2 µg/ml chymostatin, 1 mM N--p-tosyl-L-arginine methyl ester). An equal volume of glass beads was added, and cells were broken by vortexing six times in 1-min bursts followed by chilling on ice. Supernatants were transferred to fresh tubes. Beads were washed with 200 µl of lysis buffer, and the second set of supernatants was combined with the first. Additional protease inhibitors (2 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 5 mM phenanthroline) were added. Supernatants were centrifuged first at 1,086 × g for 10 min followed by 100,000 × g for 60 min at 4°C.

DEAE chromatography was performed as described by Barnes et al. (1992) with some modifications. All procedures were performed at 4°C. Glycerol and NaCl were added to supernatants to final concentrations of 10% and 0.1 M, respectively. Supernatants (~2.2 mg of protein per strain) were loaded onto 1-ml columns of DE-52 resin (Whatman, Clifton, NJ) equilibrated with column buffer (0.1 M PIPES, pH 6.8, 10 mM MgSO₄, 2 mM EGTA, 0.1 mM GTP, 0.1 M NaCl, 10% glycerol, 1% Triton X-100, and protease inhibitors). The column was washed with 5 volumes of column buffer. Tubulin was eluted with 2 volumes of bump buffer (0.1 M PIPES, pH 6.8, 10 MgSO₄, 2 mM EGTA, 0.1 mM GTP, 0.6 M NaCl, 10% glycerol, 1% Triton X-100, and protease inhibitors). Fractions (100 µl) were collected.

Duplicate samples of each fraction were subjected to SDS-PAGE followed by either Coomassie blue staining or immunoblot analysis of -tubulin levels. Fractions containing tubulin were pooled and proteins were concentrated by chloroform/methanol precipitation (Wessel and Flügge, 1984). Precipitated protein was solubilized in sample buffer (Laemmli, 1970), and duplicate samples were subjected to SDS-PAGE followed by either Coomassie blue staining and fluorography or immunoblot analysis of -tubulin levels.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Generation of TUB1, C377S tub1, C214S tub1, and C214S/C377S tub1 Haploid Strains

We used site-directed mutagenesis to construct plasmids in which either cys-377 or cys-214 or both of TUB1 was changed to serine. The haploid yeast strain DBY2384 was transformed with either the TUB1, C377S tub1, C214S tub1, or C214S/C377S tub1 plasmids. Strain DBY2384 has deleted copies of chromosomal TUB1 and TUB3, the two -tubulin genes in S. cerevisiae, but contains a plasmid with TUB3 for -tubulin function, and URA3 as a selective marker. Transformants were obtained for all four plasmids indicating that coexpression of C377S tub1, C214S tub1, or C214S/C377S tub1 with TUB3 was not deleterious to the cells. When transformants were grown on 5-FOA to select for the ability to lose the TUB3 plasmid (Boeke et al., 1984), colonies were obtained with either TUB1, C377S tub1, C214S tub1, or C214S/C377S tub1 as their sole source of -tubulin. These results demonstrate that neither cys-377 nor cys-214 is essential.

Morphology of TUB1 and tub1 Mutant Haploid Strains

We examined whether the cys-to-ser -tubulin mutations affected microtubules or position of nuclei in cells grown at 30°C. Cells were processed for immunofluorescence microscopy of microtubules and DAPI staining of DNA as described in MATERIALS AND METHODS. In G1 of the cell cycle, TUB1 cells contained a set of short astral microtubules emanating from a single site on the nuclear membrane into the cytoplasm. All three tub1 mutants displayed the TUB1 microtubule configuration while in G1 (Figure 1A).

View larger version (49K): [in this window] [in a new window]   Figure 1.   Microtubule structures in C214S tub1 and C377S tub1 cells grown at 30°C. Cells were grown to log phase at 30°C, harvested, and processed for microscopy as described in MATERIALS AND METHODS. Microtubules were visualized by immunofluorescence staining of tubulin. Chromatin was visualized by DAPI staining. (A) C214S tub1 cell in G1 phase of the cell cycle. (B) C214S tub1 cell in M phase. (C-H) C377S tub1 cells in M phase. At least 500 cells were counted for each strain. A summary of results for mitotic cells is shown in the accompanying table.

As cells entered mitosis, spindles in all four strains oriented along the mother-bud axis. In anaphase, astral microtubules in TUB1 cells emanated from both ends of the spindle until reaching the cortex of both the mother and bud. As spindles continued to elongate, astral microtubules shortened and chromosomes moved to the cortex of both the mother and bud.

In contrast to TUB1 cells, abnormal microtubule phenotypes were pronounced in mitotic cells expressing C377S tub1p: ~45% of mitotic C377S tub1 cells displayed at least one of the abnormal astral microtubule structures shown in Figure 1, C-H. These cells often appeared to contain an excess number of astral microtubules, some of which were abnormally long and were bent around the periphery of the cell (Figure 1, C-H). Some astral microtubules in the C377S tub1 mutant were also incorrectly oriented: as shown in Figure 1, E and F, astral microtubules emerging from the spindle pole body nearest the bud grew into the mother cell instead of the bud. Abnormally long mitotic spindles (Figure 1C) were also observed in the C377S tub1 cells.

The majority of mitotic C214S tub1 cells were indistinguishable from TUB1 cells (Figure 1B and accompanying table). However, a small number of mitotic C214S tub1 cells (12%) displayed long astral microtubules that curved around the cortex of the bud. Abnormally long spindles were not found in either TUB1 or C214S tub1 cells.

The double mutant C214S/C377S tub1 was indistinguishable from the C377S tub1 mutant, displaying similar patterns of aberrant astral microtubule arrays and abnormally long spindles during anaphase. However, the abnormal microtubule phenotype of the double mutant appeared no more pronounced than that of the single mutant C377S tub1: ~39% of the C214S/C377S tub1 cells had abnormal astral microtubules, 6% had long spindles, and 55% resembled TUB1 cells.

When cells were grown at 15°C, microtubules in TUB1, C377S tub1, and C214S tub1 cells were again indistinguishable during G1 of the cell cycle. During mitosis, microtubule patterns in C214S tub1 cells were most often (94%) indistinguishable from those in TUB1 cells (Figure 2A). In contrast, ~31% of mitotic C377S tub1 cells displayed abnormally long mitotic spindles that bent around the periphery of the mother or the bud or both. In some cases, elongated spindles formed a "figure 8" pattern (Figure 2B), whereas other spindles were shaped like a question mark with the longer, more bent region of the spindle located in the mother cell (Figure 2C). Finally, in 29% of mitotic C377S tub1 cells, abnormally long astral microtubules were observed. In contrast, no abnormally long astral microtubules were found in either TUB1 or C214S tub1 cells.

View larger version (36K): [in this window] [in a new window]   Figure 2.   Microtubule structures in mitotic C214S tub1 and C377S tub1 cells grown at 15°C. Cells were grown to log phase at 15°C, harvested, and processed for microscopy as described in MATERIALS AND METHODS. Microtubules were visualized by immunofluorescence staining of tubulin. Chromatin was visualized by staining with DAPI. (A) C214S tub1 cell. (B and C) C377S tub1 cells. At least 500 cells were counted for each strain. A summary of results for mitotic cells is shown in the accompanying table.

In wild-type cells, when a spindle elongates to approximately half the length of the mother cell, it translocates as a unit from the mother into the bud neck, and spindle elongation continues (Kahana et al., 1995; Yeh et al., 1995). In some C377S tub1 cells, long spindles were positioned primarily in the mother (Figure 1, F and H), suggesting that this mutation may affect spindle translocation. To test this possibility, we selected mitotic cells with clearly separated chromosomes and determined the percentage of the spindle that had traversed the bud neck (Figure 3). In the majority of TUB1 cells, chromosomes in the mother and bud were approximately equidistant from the bud neck. No TUB1 cells were found with <30% of the spindle in the bud. In contrast, the majority of C377S tub1 cells had spindles located primarily in the mother cell; values for C214S tub1 cells were intermediate between TUB1 and C377S tub1 cells. Despite an apparent translocation defect, nuclei in C377S tub1 cells were always found in both the mother and the bud suggesting that chromatin destined for the daughter cell was positioned far enough through the bud neck to prevent generation of anucleate cells. This conclusion is supported by the fact that no anucleate or multinucleate cells were observed in the C377Stub1 strain.

View larger version (23K): [in this window] [in a new window]   Figure 3.   Percentage of anaphase spindle that traversed the bud neck. Cells were grown in YPD medium at 30°C to log phase, harvested, and fixed in ethanol (McMillan and Tatchell, 1994). DNA and bud scars were stained with DAPI and Calcofluor, respectively, and cells were examined by phase and fluorescence microscopy as described in MATERIALS AND METHODS. Random fields of cells were photographed. Cells in anaphase B (defined as mitotic cells with clearly separated DNA) were analyzed by measuring the distance between DNA in the mother and bud, and the distance between the bud neck (identified by phase microscopy and Calcofluor staining) and DNA in the bud. The mother and bud were differentiated by size, with the use of phase microscopy, and/or by Calcofluor staining. The percentage of the spindle that had traversed the bud neck was then determined. TUB1 (n = 38), C377S tub1 (n = 105), and C214S tub1 (n = 73).

Spindles with separated chromosomes were not always laterally centered in C377S tub1 mutants (Figure 1G). As shown in Table 2, 44% of late anaphase C377S tub1 cells had spindles that were not centered. This contrasts with TUB1 cells in which 97% of the spindles are maintained in a more central position. A small percentage of C214S tub1 cells (13%) displayed spindles that were not laterally centered.

Staining of actin filaments and cortical actin patches with rhodamine-phalloidin revealed no apparent differences between TUB1, C377S tub1, and C214S tub1 cells. These data indicate that neither the C377S tub1 nor C214S tub1 mutation caused global damage to either the actin cytoskeleton or the actin-dependent polarity system in yeast.

Growth, Benomyl Sensitivity, and -Tubulin Levels of Strains Expressing TUB1, C377S tub1, C241S tub1, or C214S/C377S tub1

Cold or temperature sensitivity is often used to select for -tubulin mutants (Schatz et al., 1988). To determine whether our tub1 mutants displayed cold- and/or temperature-sensitive phenotypes, growth at 15, 30, and 37°C was examined. On YPD plates, the three tub1 mutant strains grew as well as TUB1 cells at all three temperatures, demonstrating that the mutations had no major effect on growth. To measure growth rates, cells were incubated in YPD medium at 15, 30, and 37°C, and at various time points, cells were counted. Again, no differences were found between the tub1 mutants and TUB1 cells. Doubling times for all four strains at 15, 30, and 37°C were 8.0, 1.6, and 2.0 h, respectively. These studies suggest that the tub1 mutations did not affect the ability of -tubulin to assemble into microtubules, a property of the protein that is exquisitely sensitive to conformational changes in the molecule (Sackett et al., 1994).

Many tubulin mutations increase the amount of time cells spend in mitosis (Schatz et al., 1988). The percentage of cells in mitosis is indicated by the percentage of large-budded cells within a population. With the use of phase contrast microscopy, we examined cells growing logarithmically in YPD medium, and counted those as large-budded if the bud was at least half the size of the mother. We found no significant differences in the percentage of large-budded cells between TUB1 cells and the three tub1 mutants at 15, 30, or 37°C. Examination of the overall size of both budded and unbudded cells revealed no differences between wild type and the mutants.

Increased sensitivity of cells to the anti-tubulin drug benomyl is a common phenotype of tubulin mutants (Schatz et al., 1986, 1988, Richards et al., 2000). We, therefore, compared sensitivity of TUB1 cells and the three tub1 mutant strains to different concentrations of the drug. As shown in Figure 4, the C377S tub1 mutant displayed increased sensitivity to benomyl relative to TUB1 cells, whereas the C214S tub1 mutant was more resistant; the double mutant was intermediate between the two single mutants. The same relative sensitivities to benomyl were found when TUB1 and tub1 mutant cells were grown at 15°C. The level of -tubulin in the three mutant strains was compared with that in TUB1 cells by immunoblot analysis of whole cell extracts (Figure 5). Similar levels of -tubulin were found in the TUB1 and C377S tub1 strains. However, both C214S tub1 and C214S/C377S tub1 cells contained ~30% more -tubulin than TUB1 cells. This increased level of -tubulin protein in the C214S tub1 strain may account for the increased benomyl resistance found in these cells relative to the TUB1 strain; to our knowledge, benomyl resistance is the only phenotype associated with excess -tubulin relative to -tubulin (Schatz et al., 1986). Likewise, increased -tubulin expression in the double mutant may result in suppression of benomyl sensitivity that is found in the C377S tub1 mutant.

View larger version (18K): [in this window] [in a new window]   Figure 4.   Benomyl sensitivity of TUB1 and tub1 strains. Serial dilutions of cells were spotted onto YPD plates containing 0-40 µg/ml benomyl. Cells grown on 10, 20, or 30 µg/ml benomyl at 25°C are shown.

View larger version (79K): [in this window] [in a new window]   Figure 5.   -Tubulin levels in TUB1 and tub1 haploid strains. Cells were grown to log phase at 30°C, harvested, lysed, and processed for immunoblot analysis with anti--tubulin antibody as described in MATERIALS AND METHODS. Relative levels of -tubulin were determined by scanning blots with an -Innotech IS1000 Digital Imaging System and setting -tubulin levels in the TUB1 extract at 100%. Differences in protein loading onto SDS-PAGE gels were adjusted after scanning the corresponding Coomassie blue-stained gel. Molecular weight standards (kDa) are shown in the left lane.

In Vivo Palmitoylation of Wild-Type -Tubulin in Yeast

Our previous studies suggest that cys-377 in Tub1p of S. cerevisiae may be posttranslationally modified by palmitoylation (Caron, 1997; Ozols and Caron, 1997). To determine whether wild-type -tubulin in yeast is indeed palmitoylated in vivo, the following experiments were performed. Initially, we examined in vivo [³H]palmitoylation of tubulin in haploids and diploids that produce wild-type levels of tubulin. However, we could not detect [³H]palmitoylated tubulin within reasonable exposure times. This is most likely because S. cerevisiae contains relatively low levels of tubulin. For example, in platelets, where we were able to detect [³H]palmitoylation of tubulin (Caron, 1997), tubulin comprises ~3% of the total protein (Steiner and Ikeda, 1979), whereas in yeast only 0.05% of the protein is tubulin (Kilmartin, 1981). To circumvent this problem, we chose a diploid strain, FSY279, for biochemical analyses. This strain contains high copy galactose-inducible TUB1 (-tubulin) and TUB2 (-tubulin) overexpression plasmids (Weinstein and Solomon, 1990). After a 4-h induction with galactose, - and -tubulin levels increased by approximately five- and twofold, respectively, and a more extensive and dense array of microtubules formed. To decrease the level of endogenous palmitate, cells were preincubated with cerulenin (3 µg/ml) before addition of [³H]palmitate (Mitchell et al., 1994; Song and Dohlman, 1996). A time course study with strain FSY279 revealed maximal [³H]palmitoylation of proteins after incubating cells with cerulenin for 4 h and adding [³H]palmitate for the last hour.

To determine whether tubulin in strain FSY279 is [³H]palmitoylated in vivo, cells were incubated in labeling medium containing 2% galactose to induce expression of tubulin (see MATERIALS AND METHODS). Analysis of whole cell extracts by 1D-PAGE revealed that several proteins were [³H]palmitoylated and that the pattern of labeled proteins was distinct from that of Coomassie blue-stained proteins (Figure 6A). When - and -tubulin were immunoprecipitated from whole cell extracts, [³H]palmitoylated -tubulin was detected; a much weaker signal from [³H]palmitoylated -tubulin was also found (Figure 6B). Immunoblot analysis of immunoprecipitated proteins demonstrated that there was no cross-contamination of - or -tubulin in the immunoprecipitants.

View larger version (38K): [in this window] [in a new window]   Figure 6.   [3H]Palmitoylation of tubulin in vivo. Overexpressor strain FSY279 (TUB1) was incubated with 2% galactose and [3H]palmitate as described in MATERIALS AND METHODS. Cells were lysed and processed for immunoprecipitation of - and -tubulin. (A) Autoradiograph and corresponding Coomassie blue-stained gel of proteins from whole cell lysate after 1D-PAGE. Molecular weight markers (kDa) are shown on the right. (B) Autoradiograph of [3H]palmitoylated protein immunoprecipitated with either anti--tubulin (left lane) or anti--tubulin (right lane) antibody.

In Vivo Palmitoylation of Tubulin from TUB1 and C377S tub1 Overexpressor Strains: Analysis of -Tubulin Protein by Immunoprecipitation

Because cys-377 in TUB1 may be a palmitoylation site (Ozols and Caron, 1997), we next sought to determine whether cys-to-ser mutation of this residue affected palmitoylation of the Tub1p in vivo. To this end, we constructed a diploid strain (JFY2705) that overexpresses C377S tub1p by approximately fivefold when cells are grown in the presence of galactose for 4 h (see MATERIALS AND METHODS). The rate and level of induction of -tubulin in strain JFY2705 was similar to that in strain FSY279 (Figure 7).

View larger version (62K): [in this window] [in a new window]   Figure 7.   Induction of -tubulin levels in TUB1 and C377S tub1 overexpressor strains. Strains FSY279 (TUB1) and JFY2705 (C377S tub1) were incubated with 2% galactose for up to 4 h. At several time points, cells were harvested, lysed, and processed for immunoblot analysis with anti--tubulin antibody. The corresponding Coomassie blue-stained gel is shown with molecular weight standards (kDa) in the left lane. (a) TUB1. (b) C377S tub1.

To determine whether the C377S mutation affected the palmitoylation of -tubulin, the TUB1 (FSY279) and C377S tub1 (JFY2705) overexpressor strains were incubated with galactose and [¹⁴C]palmitate as described in MATERIALS AND METHODS. Cells were lysed and -tubulin was immunoprecipitated. Duplicate samples of the immunoprecipitants were subjected to 1D-PAGE and immunoblot analysis with anti--tubulin antibody. As shown in Figure 8, in vivo palmitoylation of -tubulin from the C377S mutant strain was reduced by ~50% compared with -tubulin from the TUB1 strain.

View larger version (21K): [in this window] [in a new window]   Figure 8.   Effect of the C377S tub1 mutation on in vivo palmitoylation of -tubulin: use of immunoprecipitation. Overexpressor strains, FSY279 (TUB1) and JFY2705 (C377S tub1), were incubated with 2% galactose for 4 h; [14C]palmitate for the last hour. Cells were lysed and extracts were processed for immunoprecipitation with anti--tubulin antibody as described in MATERIALS AND METHODS. Immunoprecipitants were subjected to 1D-PAGE followed by Coomassie blue staining and fluorography. A fraction (10%) of the immunoprecipitants was subjected to 1D-PAGE and processed for immunoblot analysis with anti--tubulin antibody. The autoradiograph shows [14C]palmitoylated -tubulin in TUB1 (a) and C377S tub1 (b) immunoprecipitants. The corresponding immunoblot shows -tubulin levels in TUB1 (a) and C377S tub1 (b) immunoprecipitants.

In Vivo Palmitoylation of Tubulin from TUB1 and C377S tub1 Overexpressor Strains: Analysis of -Tubulin Protein by DEAE Chromatography

We next developed a second approach for analysis of palmitoylated tubulin that did not require immunoprecipitation of the protein. After incubation with galactose and [³H]palmitate, tubulin was partially purified from both the TUB1 (FSY279) and C377S tub1 (JFY2705) overexpressor strains by DEAE chromatography as described in MATERIALS AND METHODS. DEAE fractions containing -tubulin were identified by immunoblot analysis, pooled, and proteins were concentrated. Extracts were then processed for SDS-PAGE and either Coomassie blue staining and fluorography, or immunoblot analysis with anti--tubulin antibody.

As shown in Figure 9, similar patterns of Coomassie blue-stained proteins were found in DEAE extracts from TUB1 and C377S tub1 overexpressor strains. Tubulin in these extracts represented ~5% of the protein and was easily seen after SDS-PAGE and Coomassie blue staining. Because of this substantial enrichment of tubulin, we were able to analyze these fractions without immunoprecipitation.

View larger version (83K): [in this window] [in a new window]   Figure 9.   Effect of the C377S tub1 mutation on in vivo palmitoylation of tubulin: use of DEAE chromatography. Overexpressor strains, FSY279 (TUB1) and JFY2705 (C377S tub1) were incubated with 2% galactose and [3H]palmitate before cell lysis and DEAE chromatography (see MATERIALS AND METHODS). DEAE extracts containing -tubulin were subjected to SDS-PAGE followed by Coomassie blue staining and fluorography. Duplicate samples were subjected to immunoblot analysis with anti--tubulin antibody. Molecular weight markers (kDa) are in the left lane of the Coomassie blue-stained gel. The position of -tubulin is shown with an arrow. (a) TUB1 extract. (b) C377S tub1 extract.

Fluorography of the Coomassie blue-stained gels revealed similar patterns of [³H]palmitoylated proteins between extracts from the two strains, but some differences in intensity were found (Figure 9). Examination of the tubulin band revealed [³H]palmitoylated tubulin from the control strain (TUB1). However, the level of [³H]palmitoylated tubulin in the C377S tub1 strain was reduced by >60% of that in the control; immunoblot analysis of -tubulin demonstrated that the TUB1 and C377S tub1 extracts contained similar levels of the protein (Figure 9).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Astral Microtubules and Nuclear Positioning Are Defective in Anaphase C377S tub1 Cells

Replacement of cys-377 with serine did not have global effects on microtubule-dependent functions, but instead resulted in an apparent anaphase-specific defect in astral microtubule functions.

In S. cerevisiae, microtubules are involved in nuclear migration during mitosis and mating, and chromosome separation during cell division (Solomon, 1991). Recent studies, however, show that the role of microtubules in these processes is rather complex, being required at several specific times and for different reasons. Figure 10, A-G, is a compilation derived from those studies that have broadened our knowledge of the role of microtubules in S. cerevisiae (Jacobs et al., 1988; Snyder et al., 1991; Palmer et al., 1992; Sullivan and Huffaker, 1992; McMillan and Tatchell, 1994; Farkasovsky and Küntzel, 1995; Kahana et al., 1995; Yeh et al., 1995; Doyle and Botstein, 1996; Waddle et al., 1996; Shaw et al., 1997; Carminati and Stearns, 1997; Miller and Rose, 1998; Goode et al., 1999; Heil-Chapdelaine et al., 1999; Tirnauer et al., 1999; Adames and Cooper, 2000; Heil-Chapdelaine et al., 2000; Yeh et al., 2000; Farkasovsky and Küntzel, 2001).

View larger version (22K): [in this window] [in a new window]   Figure 10.   Schematic of microtubules in TUB1 (wild-type) and C377S tub1 cells during the yeast cell cycle and a model for the function of palmitoylated -tubulin.

Comparison of TUB1 and C377S tub1 haploid strains presented here suggests that astral microtubules do not require cys-377 during G1 phase of the cell cycle (Figure 10A), for positioning nuclei at emerging bud sites (Figure 10B) or for formation of mitotic spindles or alignment of spindles along the mother-bud axis (Figure 10C). However, striking differences were found between TUB1 and C377S tub1 cells during anaphase. First, in C377S tub1 cells, correctly aligned spindles were often found primarily in the mother cell (Figures 1 and 3), suggesting a role for cys-377 in the spindle translocation step (Figure 10E). Second, spindles and astral microtubules in late anaphase mutant cells were often found along the sides of cells (Table 2), suggesting that cys-377 plays a role in establishing and/or maintaining microtubule structures in a laterally centered position. Third, some astral microtubules in the mutant were misoriented, suggesting that cys-377 is involved in positioning or maintaining the position of astral microtubules. Fourth, upon reaching the cortex, astral microtubules in the mutant continued to grow resulting in abnormally long microtubules that curved around the periphery of the cell. This suggests that cys-377 plays a role in regulating astral microtubule length. Finally, the abnormally long mitotic spindles suggest that cys-377 plays a role in the termination of spindle elongation (Figure 10F).

Similarity of C377S tub1 Mutant with Nuclear Migration Mutations

Mutation of several nontubulin proteins that are involved in mitotic nuclear migration (Dhc1, Arp1p/Act5p, Nip100p, Jnm1p, Num1p, Kar9p, Bim1p/Yeb1p, Crn1p) results in remarkably similar phenotypes to those described for the C377S -tubulin mutant. As with the C377S tub1 mutant, mutation of these nuclear migration proteins has little or no effect on cell growth at different temperatures (Kormanec et al., 1991; Eshel et al., 1993; Li et al., 1993; McMillan and Tatchell, 1994; Muhua et al., 1994; Kahana et al., 1998; Goode et al., 1999). However, phenotypes resulting from these mutations are mitosis specific. Like the C377S tub1 mutant, nuclear migration protein mutations show incomplete penetrance of phenotypes: the percentage of cells affected by the different nuclear migration mutations varies from 4 to 50%, depending on the phenotype and the temperature of growing cultures. A comparison of phenotypes in nuclear migration mutants and in the C377S tub1 mutant is shown in Table 3. In wild-type cells, distal ends of astral microtubules appear to intersect with Bim1p/Yeb1p, Kar9p and Jnm1p that concentrate at the tip of the bud cortex in large-budded cells (McMillan and Tatchell, 1994; Miller and Rose, 1998; Miller et al., 1999; Tirnauer et al., 1999; Adames and Cooper, 2000; Korinek et al., 2000; Lee et al., 2000). It was suggested that these proteins assist in the tethering of astral microtubules to this region of the bud cortex during anaphase. Num1p localizes to the cortex of the mother and the bud where it is also thought to tether astral microtubules (Farkasovsky and Küntzel, 1995; Heil-Chapdelaine et al., 2000; Farkasovsky and Küntzel, 2001). Arp1p/Act5p (Clark and Meyer, 1994; Muhua et al., 1994; Adames and Cooper, 2000), an S. cerevisiae-specific actin-related protein, and Nip100p (Kahana et al., 1998), the yeast homolog of p150^glued, are components of dynactin complex, which is required for dynein-mediated vesicle transport (Gill et al., 1991; Schroer and Sheetz, 1991). Dhc1p is the dynein heavy chain in S. cerevisiae. In a review of the function of dynactin and dynein, Schroer (1994) suggests that dynactin complex associates with the cell cortex and binds to dynein which, in turn, links astral microtubules to the cortex. Although Nip100p has so far only been localized to bud-directed spindle pole bodies (Kahana et al., 1998), Arp1p/Act5p and dynein have been localized to the cortex of mitotic cells (Adames and Cooper, 2000). Furthermore, dynein/dynactin complex is required for the lateral associations of astral microtubules with the cortex. A model is presented by Adames and Cooper (2000) in which dynein/dynactin complex, anchored to the cortex, binds the sides of astral microtubules and walks in the minus (or spindle pole body) end direction, thereby generating the large force needed to pull the nucleus into the bud neck during anaphase.

Finally, Crn1p is a highly conserved actin binding protein whose activity may intersect with that of cys-377 of -tubulin. As an abundant component of cortical actin patches, Crn1p may link the actin and microtubule cytoskeletons in yeast (Heil-Chapdelaine et al., 1998; Goode et al., 1999). crn1 strains show no defects in growth rates, and normal actin cables are maintained. However, cortical actin patches are absent. Interestingly, like C377S tub1 cells, crn1 mutants display excessively long and misoriented astral microtubules during anaphase (5% of the population), but no misorientation of spindles. This suggests the possibility that Crn1p in cortical patches (which localize to the bud tip during mitosis) may help to guide astral microtubules to the cortex of the bud tip. A similar actin-mediated guidance mechanism for microtubule-plasma membrane interactions is described for vertebrate cells (Kaverina et al., 1998).

In Vivo Palmitoylation of Tubulin in Yeast

In addition to its effect on mitotic astral microtubules, cys-to-ser mutagenesis of cys-377 resulted in a substantial decrease (50-62%) in the in vivo palmitoylation of yeast -tubulin (Figures 8 and 9), suggesting that this cysteine residue is a site of palmitoylation. This level of reduction is found in many palmitoylated proteins with single amino acid mutations (Druey et al., 1999). The residual label in yeast tubulin may be due to palmitoylation of an unknown secondary site in -tubulin. Alternatively, the presumptive secondary site identified in porcine brain -tubulin (Caron and Ozols, 1997) may be palmitoylated in yeast -tubulin. However, the microtubule and nuclear migration phenotype of the double mutant (C214S/C377S tub1) was equivalent to the single mutant (C377S tub1), making unclear the functional significance of cys-214 as a secondary palmitoylation site, if it is one. A third possibility for residual label is palmitoylation of Tub3p; unlike the haploid strains used for phenotypic analysis, TUB3 was not deleted in the TUB1 and C377S tub1 diploid overexpressor strains used for biochemical studies. However, detection of palmitoylated Tub3p seems unlikely because this protein was expressed at 20-fold lower levels compared with C377S tub1p; in support of this conclusion, we note again that palmitoylation of Tub1p was not detected in the diploid strain without the galactose-induced fivefold increase in expression of the protein. Finally, under our immunoprecipitation conditions, -tubulin is not immunoprecipitated with anti--tubulin antibody (Caron et al., 1985; Caron, unpublished data). Therefore, residual label found after immunoprecipitation of -tubulin from C377S tub1 extracts (Figure 8) is not due to palmitoylation of Tub2p.

In support of cys-377 as a primary site of palmitoylation of Tub1p, tubulin partially purified from strain FSY279 (TUB1) was [³H]palmitoylated in our cell-free system (Caron, 1997), whereas tubulin from strain JFY2705 (C377S tub1) was not (Caron, unpublished data). This cell-free system has been shown to palmitoylate the same sites in nontubulin proteins as found in vivo (Druey et al., 1999).

Model for Function of Palmitoylated -Tubulin

In eukayotic cells, including S. cerevisiae, the long-chain fatty acid palmitate is covalently linked to cysteine residues via thioester bonds (reviewed in Bizzozero et al., 1994; Casey, 1995; Mumby, 1997). The reversibility of this modification suggests that palmitoylation regulates membrane activities. For example, palmitoylation results in the sequestration of some proteins, such as nitric-oxide synthase (Garcia-Cardena et al., 1996; Shaul et al., 1996) and proteins involved in T-cell receptor signaling (Zhang et al., 1998), within membrane microdomains. Palmitoylation also determines protein-protein interactions within the membrane. For example, palmitoylation of the -adrenergic receptor prevents its phosphorylation by cAMP-dependent protein kinase C, which in turn reduces its ability to interact with Gs protein (Moffett et al., 1996).

- and -Tubulin are palmitoylated in resting human platelets, and the level of palmitoylated tubulin decreases when platelets are activated with thrombin (Caron, 1997). This depalmitoylation of tubulin coincides with thrombin-induced disassembly of the peripheral band of microtubules juxtaposed to the cell cortex.

The phenotype of the C377S tub1 mutant is consistent with the hypothesis that palmitoylation of cys-377 plays a role in static microtubule-cortex interactions. First, the astral microtubule defects associated with this mutation occur during the same period of the cell cycle (anaphase B) in which static astral microtubule-cortex interactions occur (Carminati and Stearns, 1997; Shaw et al., 1997; Adames and Cooper, 2000). Other microtubule-cortex interactions occur earlier in mitosis, but these are not static (Shaw et al., 1997). As suggested by Shaw et al. (1997), these early, transient microtubule-cortex interactions may correct errors in bud site selection and spindle orientation. Second, astral microtubules in the C377S tub1 mutant do not interact with the bud tip of the cortex the same way that they do in TUB1 cells: instead of stopping at the bud tip, astral microtubules in the mutant continue to grow. Such a phenotype is consistent with a mutation that alters the affinity of -tubulin for membranes. Third, phenotypes associated with the C377S tub1 mutant are strikingly similar to those from a class of nuclear migration proteins whose functions involve microtubule-cortex interactions.

A simple "lock and key" model (Figure 10) for the function of palmitoylated -tubulin accommodates microtubule phenotypes observed in the C377S tub1 mutant as well as data on nuclear migration proteins and known functions of astral microtubules. As wild-type cells enter anaphase and astral microtubules lengthen, these microtubules associate laterally with the cortex of both the mother and bud via interactions with cortical-anchored dynein/dynactin complex (Adames and Cooper, 2000). Through these interactions, astral microtubules slide along the cortex, pulling the spindle into the bud neck (Figure 10D). On reaching the cortical tips, astral microtubules probe the surface for a specific region that contains nuclear migration proteins (i.e., Crn1p, Bim1p/Yeb1p, Kar9p, Jnm1p, Num1p) and palmitoylating enzymes. Through interactions with these proteins, -tubulin at or near the plus (or cortex) end of these microtubules is palmitoylated, which in turn locks the microtubule end onto the cortex. This tethering of the microtubule inhibits further microtubule assembly at its plus end and leads to the more static and rigid structure that is found during this stage of mitosis (Palmer et al., 1992; Carminati and Stearns, 1997; Shaw et al., 1997). Because the tethered microtubules (and attached spindle) are "locked" into a specific position, they assist in translocating the spindle through the bud neck and maintaining the spindle in a laterally centered position (Figure 10E). Once the spindle has elongated to the length of the cell, the shortened and tethered astral microtubules prevent further elongation (Figure 10F). Finally, -tubulin is depalmitoylated, the spindle disassembles, nuclei become centered in the mother and bud, and cytokinesis occurs (Figure 10G).

Microtubule-cortex interactions are also found in higher eukaryotes. As in S. cerevisiae, dynein/dynactin complex has been localized to cortical-microtubule attachment sites in Caenorhabditis elegans (Hyman and White, 1987; Hyman, 1989; Skop and White, 1998). Here, these interactions are required for asymmetric cell division during embryogenesis, which in turn leads to the polarization of cellular components and the generation of cell diversity (Gönczy and Hyman, 1996; reviewed in White and Strome, 1996). Microtubule-cortex interactions during cell division may be more important for the survival of higher eukaryotes than S. cerevisiae: unlike S. cerevisiae, the site of cytokinesis in higher eukaryotes is determined by the position of the mitotic spindle. It is possible that palmitoylation of -tubulin plays a role in establishing or maintaining static microtubule-cortex interactions required for the correct positioning of nuclei during cell division of higher eukaryotes.