Copyright 0 1996 by the Society of.America

CDPl, a Novel Saccharomyces cermisiae Gene Required for Proper Nuclear Division and Segregation

Pamela K. Foreman and Ronald W. Davis Department of Biochemistry, Stanford University School of Medicine, Stanford, California 94305 Manuscript received May 27, 1996 Accepted for publication September 12, 1996

ABSTRACT To identify new gene products involvedin chromosome segregation,we isolated Saccharomyces cermisiae mutants that require centromere binding factor I (Cbflp) for viability. One Cbflpdependent mutant (denoted cdpl-1) was selected for further analysis. The CDPl gene encodes a novel 125-kD that is notably similar to previously identified mouse, human and Caenorhabditis elegans . CDPlA and cdpl-1 mutant cells were temperature sensitive for growth. At the permissive temperature, cdpl-1 and cdplA cells lost at a frequencies -20-fold and -110-fold higher than wild-type cells, respectively. These mutantsalso displayed unusually long andnumerous bundles of cytoplasmic microtu- bulesas revealed by immunofluorescent staining. In addition, we occasionally observed improperly oriented mitotic spindles, residing entirely withinone of the cells. Presumably as a result of undergoing nuclear division with improperly oriented spindles, a large percentage of cdpl cells had accumulated multiple nuclei. While cdpl mutant cells were hypersensitive to the microtubuledisrupting compound thiabendazole, they showed increased resistance to the closely related compound benomyl relative to wild-type cells. Takentogether, these results suggestthat Cdplp plays a role in governing tubulin dynamics within the cell and may interact directly with microtubules or tubulin.

HE faithful transmission of genetic information re- the nucleus in the bud neck and in orienting the spin- T quires that when cells divide, each daughter cell dle so that it projects from the mother cell into the bud receives one and only one copy of each chromosome. (JACOBS et al. 1988; PALMERet al. 1992; SULLIVANand During in eukaryotic cells, the duplicated chro- HUFFAKER1992). Microtubule-based mechanochemical mosomes bind to a microtubule-based spindle appara- motors are thought to generate the forces that account tus that guides them through a series of movements, for the movements of the nucleus and chromosomes culminating in the separationof chromosomes into two during mitosis and that separate the spindle polebodies complete complements.Each of the chromosome com- (HOWet al. 1992,1993; ROOF et al. 1992; SAUNDERSand plements is packaged into the nucleus of a daughter HOW 1992; ESHELet al. 1993; LI et al. 1993; MIDDLETON cell. and CARBON1994; SAUNDERSet al. 1995). Following The assembly, function and disassembly of the mi- mitosis, nuclear MTs become undetectable, whereas cy- totic spindle are complex processes that are only par- toplasmic MTs persist throughout the cell cycle. In S. tially understood. In Saccharomyces cerevisiae, the nuclear cerevisiae, the function of cytoplasmic MTs during in- envelope does not break down during mitosis and the terphase is unknown. The differential regulation of spindle is contained entirely within the nucleus. The polymerization and depolymerization of the various microtubule organizing center, or spindle pole body, classes of MTs, the precise coordination of MT-medi- is embedded in the nuclear envelope. Early in the cell ated events with progression through the cell cycle, and cycle, the spindle pole bodies are duplicated and mi- the diversity of functions performed by MTs suggests grate to positions defining the spindle poles (reviewed that asignificant number of as yetunidentified proteins in BYERS1981). Before the onset ofmitosis, nuclear are involved in MT dynamics and function. microtubules(MT) polymerize fromthe duplicated In most organisms, spindle MTs attach to chromo- spindle pole bodies. A subset of these MTs, called ki- somes at the , a multiprotein complex as- netochore MTs, form attachmentsto the chromosomes. sembled at the centromere. In S. cerevisiae, the centro- Another subset consists of interdigitating MTs emanat- mere is relatively small, consisting of three conserved ing from opposite poles (WINEYet al. 1995). M'lcrotu- DNA elements, CDEI,CDEII and CDEIII,which to- bules also project from the spindle pole bodies into getherspan - 125 bp (reviewed in CARBON1984; the cytoplasm (the equivalent of astral MTs) . During CLARKEand CMOS 1985; FITZCERALD1987; CARBON mitosis, these cytoplasmic MTs playa role in positioning and CLARKE1990; CLARKE1990; MURPHYand FITZGER- ALD 1990). Several kinetochore constituents have been Corresponding author: Pamela K. Foreman, Alza Corporation, 950 identified and cloned. The products of the CTF13, Page Mill Rd., Palo Alto, CA 94306. iVDClO/CBF2/CTFI4 and CEP3/CBF3b genes form a

Genetics 144: 1387-1397 (December, 1996) 1388 P. K. Foreman and R. W'. Davis complex that binds to the CDEIII region of the centro- rupting drugs were dissolved in DMSO (benomyl) (DuPont) mere (LECHNERand CARBON1991; DOHENYet al. 1993; or dimethylformamide (thiabendazole) (Sigma) and added to molten media. GOH and KILMARTIN 1993; JIANG et al. 1993; LECHNER Plasmids: pMW29LYS2was provided byM. WAHLBEKG 1994; STRUNNIKOVet nl. 1995). Although the exact func- (University of Texas Southwestern Medical Center). It is a tions of these gene products are currently unknown, derivative of pMW29 (ZIELERet al. 1995), a pUC18-based plas- they are all essential for viability. The CBFl/CEPl/CPFl mid containing CEN4 ARSl ME3 LYS2 GAL1 promoter-poly- (hereafter referred to as CBFI) gene product binds to linker-GAL7 terminator. In YCpCBFAL, the CBFl gene is cloned into the polylinker region of pMW29LYS2 along with the CDEI region of the centromere (BAKERet al. 1989; 254 bp of the CBFl promoter. In this configuration the CBFI CM and DAWS1989, 1990;JIANG and PHILIPPSEN1989; gene is expressed from its own promoter and is not under BAKERand MASISON 1990; MELLOR et al. 1990). Cells the control of the GAL1 promoter. plC12A95-2922 contains lacking Cbflp are viable, but lose chromosomes at a an EroRI-ClaI (bps -196-95) fragment and a BglII-EcoRI (bps frequency - 10-fold higher than wild-type cells.In addi- 2922-3695) fragment from the CDPl genejoined at theEroRI tion, like manymutants defective in chromosomesegre- sites and subcloned into ClnI, BnmHI-cleaved pRS304 (which contains the TRPl marker) (SIKORSKIand HIETER 1989). gation, cbflA mutants are more sensitive than wild-type Screen for CBFldependent mutants: Yeast strains YNN534 cells to microtubule disrupting drugs (CN and DAL~IS and YNN535 were mutagenized with ethylmethane sulfonate 1990). Therefore, although notessential for growth un- or UV light to "30% viability (SHERMAN1991). UV-mutagen- der laboratory conditions, Cbflp plays a role main- ized cells were allowed to recover in the darkfor 48 hr. Follow- taining the high fidelity with which chromosomes are ing mutagenesis, -50,000 cells were plated on WD and the resulting colonies were examined for the presence of white segregated. sectors. Colonies that appeared to lack sectors were patched Cells lacking Cbflp are also unable to grow in the to WD plates and reexamined for sectoring using a dissecting absence of exogenously provided methionine (BAKER microscope. Three hundred thirty-seven candidates failed to and MASISON1990; CAI and DAVIS1990). This methio- sector in this test and were streaked to medium containing 0.2% a-aminoadipate (Sigma). One hundred thirty-one mu- nine auxotrophy is due, at least in part, to a require- tants did not grow on this medium and were transformed in ment for Cbflp in of the MET16 gene parallel with the plasmid pUN55 (ELL.EDGEand DAVIS1988) (THOMASrt al. 1992; O'CONNELLet al. 1995) We have or its derivative YCpCBFUS, which differs only in that it con- shown previously that single amino acid alterations in tains the CBFl gene. In 48 mutants, sectoring and growth on Cbflp can preferentially impair either its abilityto func- a-aminoadipate-containing medium were restored by YCpCB- FUS, but not by pUN55. tion in chromosome segregation or its ability to func- Cloning of CDPl: rdpl-1 cells were transformed, using the tion in transcription (FOREMANand DAVIS1993). These lithium acetate method (ITOet al. 1983),with a yeast genomic results demonstrate that the roles Cbflp plays in each library that carries a Urn? marker on a centromere-based of these capacities are mechanistically distinct. (low copy number) plasmid (RAMER et nl. 1992). Approxi- Because chromosome segregation is a fundamental mately 20,000 URA" colonies were replica-plated to synthetic aspect of cell division, it is important to identify and minimal media containing necessary supplements as well as uracil and 0.2% a-aminoadipate. After 2 days of growth, repli- define the roles of gene products involved in this pro- cas were once again made onto fresh a-aminoadipate-con- cess. We have performeda genetic screenfor novel taining medium. Surviving colonies were replicated to YPD gene products that play a role in chromosome segrega- plates, plates containing 0.1 % 5-fluoroorotic acid (5-FOA), tion. This screen makes use of the fact that chromo- and plates lacking methionine. Colonies that were white (con- some segregation is more efficient in cells that contain firming that they had lost pCBFAL) and failed to grow on both methionine-free medium (indicating that they were not Cbflp than in cells from which the CBFl gene has been CBFl-containing clones) and 5-FOA-containing medium were deleted. Mutations in other genes whose products are chosen for further analysis. involved in chromosome segregation might be ex- Two plasmids, containing overlapping DNA fragments of pected to exacerbatethe effects of a CBFl deletion, -3.4 and -3.7 kb were recovered that complemented the resulting in inviability. Based on this rationale, we have CBFl dependence, temperaturesensitivity and thiabendazole hypersensitivity of the rdpl mutant. The plasmids were iso- performed a screen for mutants thatcan be propagated lated and transformed into Eschrrichia coli by standard meth- in the presence of Cbflp, but that are unable to grow ods (GLTHRIEand FINK 1991). in the absence of Cbflp. These mutants have been des- Sequencing was performed by the Beckman Center PAN ignated cd$ (Cbflp-dependent). In this paper, we de- facility by thermocycling using custom oligonucleotide prim- scribe one of these mutants, cdpl, which appears to have ers, dye terminators and Taq DNA polymerase. Reaction prod- ucts were analyzed using an Applied Biosystems model 373 a defect in microtubule function,resulting in improper sequencing system. Both strands were sequenced. The se- chromosome segregation and nuclear division. quence was analyzed using the Intelligenetics Suite (Intelli- genetics, Inc., MountainView, CA), GCG (Genetics computer MATERIALSAND METHODS group, Madison, WI), and NCBI (Bethesda, MD) sequence analysis packages. Strains, mediaand microbiological techniques: Yeast To demonstratelinkage of the complementingDNA to the strains used in this study were congenic (W303 derivatives) mutant phenotypes, an 897-bp EcoRI fragment (bps 2798- and are listed in Table 1. Standard molecular cloning tech- 3695) containing the3'-end of the single open readingframe niques (SAMBROOKrt al. 1989) and yeast genetic techniques was subcloned into the integrating plasmid pRS304,which and media were used (SHERMAN1991). Microtubule dis- carries the TRPl marker. (SIKORSKIand HIETER1989). The Role of Cdplp in Nuclear Division 1389

TABLE I Yeast strains

Strain Genotype Source YCRY2 a ade2-1 his3-11,15 tql-1 leu2-3,112 ura3-1 R. FULLER YCRY3 a/a ade2-1/ade2-1 his3-11,15/his3-l1,15 tql-l/tql-l lm2- R. FULLER 3, I 12/ ku2-3, I 12 ura3-1/ ura3-1 YNN531 a ade2-l his3-11,15 tql-1 leu2-3,112 ura3-I, cbf1A::LEUZ FOREMANand DAVIS(1993) YMW80 a ade2-1 ade3A22 his3-11,15 leu2-3,112 tql-I ura3-I 4s:: hisG M. WAHLBERG YNN532 a ad&-1 ade3A22 his3-11,IS leu2-3,112 trpl-I ura3-1 lys2: : hisG This study cbfl A : : LEU2 YNN533 a ade2-1 ade3A22 his3-11,15 leu2-3,112 tql-1 ura3-1 lys2::hisG This study cbflA : : LEU2 YNN534 a ade2-1 ade3A22 his3-11,15 leu2-3,112 trpl-1 ura3-I lys2::hisG This study cbflA : : LEU2 [YCpCBFAL] YNN535 a ade2-1 ade3A22 his3-11,15 leu2-3,112 tql-1 ura3-1 lys2::hisG This study cbflA : : LEU2 [YCpCBFAL] YNN536 a ade2-1 his3-l1,15 trpl-1 leu2-3,112 ura3-l cdpl-l This study YNN537 a ade2-1 ade3A22 his3-11,15 leu2-3,112 tql-l ura3-1 lys2::hisG This study cbflA :: LEU2 cdpl-l [YCpCBFAL] YNN538 a ade2-1 ade3A22 his3-11,15 leu2-3,112 tql-I ura3-1 lys2::hisG This study cVA: 1 LEU2 cdpl-1 [YCpCBFUS] YNN539 a ade2-1 ade3A22 his3-11,15 leu2-3,112 tql-1 ura3-1 lys2::hisG This study cbflA : : LEU2 CDPl/ cdpl : : TRPl YNN540 a/a ade2-1/a&2-1 his3-ll,15/his3-11,15 trpl-l/trpl-I leu2- This study 3,112/leu2-3,112 ura3-1/ura3-1 cdpl-l/cdpl-l YNN541 a/a a&2-1/ade2-1 his3-11,15/his3-11,15 trpl-I/trpl-I leu2- This study 3,112/1eu2-3,112 ura3-l/ura3-1 CFIII(D8B.d) URA3 SUPPlI YNN542 a/a ade2-l/adeZ-l his3-11,15/his3-11,15 trpl-l/trpl-l leu2- This study 3,112/leu2-3,112 ura3-l/ura3-1 cdpl-l/cdpI-I CFIII(D8B.d) uRA3 SUP11 YNN543 a add-1 his3-11,15 tql-1 leu2-3,112 ura3-I cdplA::TRPl This study YNN544 a/a adeZ-I/ade2-1 his3-ll,15/his3-l1,15 trpl-l/trpl-I leu2- This study 3,112l/eu2-3,112 ura3-l/ura3-l cdplA::TRPl/cdplA:: TRPl YNN545 a/a ndeZ-l/ade2-1 his3-Il,15/his3-l1,15 trpl-I/trpl-I leu2- This study 3,1121/eu2-3,112 ura3-l/ura3-l cdplA::TRPl/cdplA:: TRPl CFIII(D8B.d) URA3 SUP1 1 resulting plasmid, pRS304RI900, was linearized withBgnI, ized with EcoRI and used to transform YCRY3. TRPI+ trans- which cleaves within the subcloned fragment, and used to formants were analyzed by Southern blot to confirm that the transform strain YNN532. Integration of the plasmid at the plasmid had integrated at theCDPl locus. One transformant, CDPl locus was confirmed in strain YNN539 (a TRPI’ trans- heterozygous for a deletion of nucleotides 95-2922 of the formant) by Southern blotting. The integration resulted in a WPl gene, was sporulated and tetrads were dissected to ob partial duplication of CDPl without disruption of the wild- tain haploid TRPI+ cells lacking OPl. type gene. YNN539 was mated to YNN537 and the resulting diploid was sporulated. In all 23 tetrads analyzed, the cdpl temperature-sensitive phenotype and theTRPl marker segre- RESULTS gated away from one another,indicating that thecloned DNA Identification of yeast mutants that depend on CBFl is tightly linked to the CDPl gene. Chromosome fragmentloss assay: Strains YCRY3, YNN540, for viability: Understandard laboratory conditions, and YNN544 were transformed with the plasmid pYCF5/ Cbflp is a nonessential gene product thatincreases the CEN6 (HEGEMANNet al. 1988) linearized at the Not1 site. The fidelity with which chromosomes are segregated. Pre- resulting strains, YNN541, YNN542 and YNN545, carried an sumably, certain aspects of the chromosome segrega- -135-kb linear fragment of chromosome ZZZ bearing ZRA3 tion process are made more efficient by Cbflp. Toinves- and SUP1 1 markers. Quantitation of the loss frequency of the marked chromosome fragment was performed as described tigate which processes might be facilitated by Cbflp, by SHEROet al. (1991). we performed a screen for mutants in which the CBFl Cell fixationand staining: To countnuclei, cells were fixed gene is essential for viability. with ethanol andstained with 1 ~g/ml4’,6diamino-2-phenyl- The overall strategy used in this screen is similar to indole (DAPI) as described (PRINGLEet al. 1991). Microtu- previous studies (KRANZ and HOLM1990; BENDERand bules were stained with YOL-34rat anti-tubulin (Accurate Sci- entific Corporation) antibodies and fluorescein-labeled goat PRINGLE1991; ZIELER et al. 1995). The background anti-rat secondary antibodies (Cappel) (STEARNSet al. 1990). strains used (YNN532 and YNN533, which differ from Disruption of the CDPl gene: plC12A95-2922 was linear- one another onlywith respect to mating type)have 1390 P. K. Foreman and R. W. Davis deletions of the cbfl, add andlys2 genes. These strains TABLE 2 also contain an ade2-1 mutation, which, in the presence Capacity of cbfl separation of function alleles to support of an otherwise intact adenine biosynthetic pathway, growth in the cdpl mutant results in the accumulation of red pigment (JONES and FINK 1981). The plasmid YCpCBFAL, which contains ReleventDoubling genotype time (hr) the wild-type CBFl, ADE? and LYS2CBFl genes, was trans- CDPl 1.89 formedinto these strains to make YNN534 and CDPI cbfl-9 2.29 YNN535. Because Ade3p lies upstreamcbfl-16 of Ade2p in the CDPl 3.04 adenine biosynthetic pathway, cells containingCBFl YCpCB- cdpl-l 2.74 FAL (which confers wild-type Ade3p function)cbfl-9 are red, cdpl-I 6.01 whereas cells lacking the plasmid are white.cbfl-16 Since CBFl cdpl-l 34.20 is a nonessential gene, cells that lose YCpCBFAL are Strains YNN531 (CDPl cbflA) and YNN537 (cdpl-I cbflA) viable. As a consequence wild-type colonies are com- were transformed with plasmids containing CBFI, rbfl-9 or posed of red sectors (consisting of cells containing YCp- cbfl-16. YCpCBFAL was evicted from YNN537 transformants CBFAL) and white sectors (consisting of cells from by selection on a-aminoadipate. Doubling times were deter- mined at 30" by measuring the optical density of liquid cul- which this plasmid has been lost).However, in mutants tures at 600 nm. that depend on Cbflpfor viability, cellsthat lose YCpC- BFAL areunable togrow. CBFl-dependent mutants, therefore, form colonies thatare red with no white (CAI and DAVIS1990; STEARNS et al. 1990; BROWNet sectors. al. 1993). Twelveof the 48 CBFl-dependent mutants Strains YNN534 and YNN535 were mutagenized and displayed an increased sensitivity to thiabendazole rela- screened for red nonsectoringcolonies. Three hundred tive to wild-type cells. thirty-seven mutants that appeared to require YCpCB- Each of the 12 CBFl-dependent, thiabendazole-hy- FAL by this criterion were then subjected to a secondary persensitive mutants was backcrossed to wild-type cells. screen in which they were assessed for their ability to In four of these mutants, the CBFl-dependence and grow on media containing a-aminoadipate. a-aminoad- thiabendazole hypersensitivity cosegregated as defects ipate selects against cells with wild-type Lys2p function in single genes upon sporulation of the diploids. Each (encodedon the indicator plasmid, YCpCBFAI,). of the mutations was recessive and belonged to a sepa- Therefore only those cells from which YCpCBFAL has rate complementation group. The four complementa- been lost are able to grow on this medium. One hun- tion groups were designated mP1, CDP2, clup3 and dred thirty-one of the 337 nonsectoring mutants failed CDP4. The single allele of CLIP1 isolated, cdpl-I, failed to grow on a-aminoadipate-containing medium, indi- to grow at 3'7" and was analyzed in greater detail. cating that these mutantsrequired some sequence The cdpl mutant depends on CBFl for its function within YCpCBFAL for ongoing cell division. in chromosome segregation: To further explore thena- To determine whetherit was the CBFl gene in YCpC- ture of the dependence of the cdpl mutant on CBFl BFAL required by these mutants for viability, two plas- for ongoingcell division,two mutant alleles of cbfl were mids unrelated to YCpCBFAL were individually trans- examined for theirability to support growth of the cdpl- formedinto each candidate mutant. These two I mutant. cbfl-9 and cbfl-16 contain single amino acid plasmids, pYCBFUS and pUN55, differ only in that pYC- substitutions that preferentially impair the ability of BFUS carries CBFl, whereas pUN55 does not. In 48 of Cbflp to functionin transcriptional activation and the remaining 131 candidates, only pYCBFUS restored chromosomesegregation, respectively (FOREMANand both the ability to form white sectors and the ability DAVIS1993). Cells containing cbfl-9 are largely compe- to grow on medium containing a-aminoadipate. The tent for thechromosome segregation function of presence of a second CBFl gene in these mutants there- Cbflp, but are methionine auxotrophs, indicating that fore relieves selectivepressure for retainingYCpCBFAL, they have a severe defect in Cbflp-mediated transcrip- strongly suggesting that these mutants dependon tional activation. In contrast, cells containing cbfl-16 Cbflp expression for viability. are markedly compromised in chromosome segrega- Since Cbflp functions in both transcriptional activa- tion, but phenotypically wildtype for transcriptional tion and in chromosomesegregation, an additional activation. Each of these alleles was substituted for a analysis was performed to assist in identifying mutants wild-type copy of CBFl in wild-type and cdpl-1 mutant that depend on CBFl primarily for its function in chro- cells. Table 2 shows that the cbfl-9, cbfl-16 and cdpl-1 mosome segregation. The 48 CBFl-dependent mutants alleles each have modest effects on the rate of growth were tested for theirability to grow on media containing of cells on rich medium at 30". In cells bearing both the microtubule-disrupting compound thiabendazole. cbfl-9 and cdpl-1, the doubling time is increased to ap- Failure to grow in the presence of low concentrations proximately twice that of cells carrying the cdpl-1 muta- of related benzimidazole compounds is a feature shared tion alone and almost three times that of cells con- by many mutants defective in chromosome segregation taining cbfl-9 alone. Interestingly, in cells bearing both Role of Cdplp in Nuclear Division 1391 cbfl-16 and cdpl-1 the doubling time is more than 12 databases is a hypothetical 1174amino acid human pro- times slower than cdpl-1 cells and 11 times slower than tein (GenBank D63875) that is 27% identical over its cells containing cbfl-16 alone. The fact that cells con- entirelength. A recently reported mouse protein, taining cbfl-16 grow dramatically slower than cells con- p150TSP, displaysa similar level of identity and is 99% taining cbfl-9 in the cdpl-1 background (but only identical to the human sequence (MALEK et al. 1996). slightly slowerin a wild-type background) suggests that In addition, the predicted Cdplp sequence is 19% iden- the cdpl mutant depends onCBFl primarily for its func- tical to a C. ekgans ORF ( SULSTONet al. 1992) (PIR tion in chromosome segregation, and supports the no- S28279) that is 31% identical to pl5OTSp.Cdplp may, tion that the cdpl-1 mutant may be defective in some therefore, belong to an evolutionarily conserved family aspect of chromosome segregation. Furthermore, these of proteins. results formally demonstrate that the cdpl-1 mutant de- To demonstrate that the ORF recovered by comple- pends on expression of Cbflp, rather than merely the menting the cdpl-1 mutation represents the CDPl gene, presence of the CBFl gene, for viability. a TRPl marker was integrated at thelocus of the cloned Cloning and sequence analysis of the CDPl gene: The gene by homologous recombination. In the resulting CDPl gene was cloned by complementing the depen- strain (YNN539), the CDPl gene is partially duplicated dence of the cdpl-1 mutant on CBFl. A URA3marked, without disruption of the wild-type gene. YNN539 was low copy number yeast genomic plasmid library with an mated to YNN537 (cdpl-1) and tetrads were dissected. average insert size of 4 kb (RAMER et al. 1992) was trans In all 23 tetrads analyzed, the cdpl temperature-sensitive formed into YNN537. Approximately20,000 trans- phenotype and the TRPl marker segregated away from formants were replica plated to medium containing a- one another, indicating that the cloned DNA is tightly aminoadipate to select against cells containingthe linked to the CDPI gene (<4.5 cM). CBFI-bearingplasmid YCpCBFAL.Six transformants The cdpl gene is essential at 37" and is required for were obtained that grew robustly on this medium, indi- sporulation: To further explore the role of Cdplp in cating that they no longerrequired YCpCBFAL for cell division, the CDPl gene was deleted by one step growth. Three of these maintained methionine proto- gene replacement (ROTHSTEIN1983). Replacement of trophy following eviction ofYCpCBFAL,suggesting that amino acids 31-974 with the TRPl gene revealed that they had received a wild-type copy of CBFl from the cells lacking Cdplp grew very poorly at 30°, and like library. The remaining three were methionine auxo- cells containing cdpl-1,were inviable at 37". Bothof trophs. To assess whether growth on a-aminoadipate these defects were rescued by a CDPl-containing plas- required sequences the transformants had received mid. Figure 2A shows a time course of growth arrest from the library, theywere plated on medium con- for cdpl-1 and cdplA cells at 37". At the restrictive tem- taining 5-FOA,which selects against cells containing perature, the behavior of cells containing these alleles the URA3 marker. All six clones failed to grow on this is indistinguishable. Growth arrests slowly over the medium indicating that following eviction of YCpCB- course of two to three generations and the distribution FAL, they required the library plasmid for growth. Li- of cells through the cell cycle after 3 and 8 hr at the brary plasmids were isolated from two met- and one restrictive temperature is very similar to that observed met' transformant. Restriction mapping of the plasmid in wild-type cells (data not shown).As shown in Figure that conferred methionine prototrophyconfirmed that 2B, cdpl cells lose viability concomitant with the cessa- theinsert contained the CBFl gene. Transformation tion of division. In addition, diploid cells homozygous of YNN537 with each of the remaining two plasmids, for either cdpl-1 or cdplA fail to sporulate. derived from the methionine auxotrophs, rescued the To determine whether cells lacking Cdplp also de- CBF1-dependence, thiabendazole hypersensitivity and pendon Cbflp for viability,cdplA cells (strain temperature sensitivityof this strain. These plasmids YNN543) were crossed with cbflA cells (YNN531), in had inserts of -3.4 and -3.7 kb that overlapped over which the CBFl gene has been replaced with the LEU2 the length of the smaller clone. gene. The resulting LEUf, TRP+ diploids were sporu- The larger plasmid insert was sequenced entirely. It lated and tetrads were dissected. Sixteen tetrads were contained a 3413-bp open reading frame (OW) (Fig- analyzed and although colonies wereobserved that ure 1) spanning almost the entireinsert. A nearly identi- were either LEU+(cbflA) or TRP+(cdplA), no colonies cal sequence derived from the left arm of chromosome were observed that were both LEU+ and TRP'. There- XVwas recently identified by the yeast genome sequenc- fore, like cells containing the cdpl-1 allele, cdplA cells ingproject (CASS~YORet al. 1995).The twose- require Cbflp for growth. quences differ at amino acids 197, 461, 674, 786, 903, cdpl mutants are differentially sensitive to benzimid- 907,926,927,956,957,959 and987. These differences azole compounds: Likemany mutants defectivein may arise from sequence variations between the differ- chromosome segregation, the cdpl-1 mutant was unable ent strains used in these studies. The 1077-amino acid to grow on a media containing a low concentration protein encoded by this ORF has a predicted molecular of the benzimidazole compound thiabendazole. Both weight of 125 kD. The most similar sequence in the thiabendazole and the closely related benzimidazole 1392 P. K. Foreman and R. W. Davis

35 70 -L 35 MTNAMKVEGY PSMEWPTSLD IPLKASEELV GIDLETDLPD DPTDLKTLLV EENSEKEHWL TIALAYCNHG

105 140 KTNEGIKLIE MALDVE'QNSE RASLHTFLTW AHLNLAKGQS LSVETKEHEL TQAELNLKDA IGFDPTWIGN

175 210 MLATVELYYQ RGHYDKALET SDLFVKSIHA EDHRSGRQSK PNCLFLLLRA KLLYQKKNYM ASLKIFQELL

245 280 VINPVLQPDP RIGIGLCFWQ LKDSKMAIKS WQRALQLNPK NTSASILVLL GEFRESFTNS TNDKTFKEAF

315 350 TKALSDLNNI FSENQHNPVL LTLLQTYYYF KGDYQTVLDI YHHRILKMSP MIAKIVLSES SFWCGRAHYA

385 420 LGDYRKSFIM FQESLKKNED NLLAKLGLGQ TQIKNNLLEE SIITFENLYK TNESLQELNY ILGMLYAGKA

455 490 FDAKTAKNTS AKEQSNLNEK ALKYLERYLK LTLATKNQLV VSRAYLVISQ LYELQNQYKT SLDYLSKALE

525 560 EMEFIKKEIP LEVLNNLACY HFINGDFIKA DDLFKQAKAK VSDKDESVNI TLEYNIARTN EKNDCEKSES

595 630 IYSQVTSLHP AYIAARIRNL YLKFAQSKIE DSDMSTEMNK LLDLNKSDLE IRSFYGWYLK NSKERKNNEK

665 700 STTHNKETLV KYNSHDAYALISLANLWTI ARDGKKSRNP KEQEKSKHSY LKAIQLYQKV LQVDPFNIFA

735 770 AQGLAIIFAE SKRLGPALEI LRKVRDSLDN EDVQLNLAHC YLEMREYGKA IENYELVLKK FDNEKTRPHI

805 840 LNLLGRAWYA RAIKERSVNF YQKALENAKT ALDLFVKESS KSKFIHSVKF NIALLHFQIA ETLRRSNPKF

875 910 RTVQQIKDSL EGLKEGLELF RELNDLKEFN MIPKEELEQR IQLGETTMKS ALERSLNEQE EFEKEQSAKI

945 & 980 DEARKILEEN ELKEQERMKQ EEEARRLKLE KQAEEYRKLQ DEAQKLIQER EAMAISEHNV KDDSDLSDKD

1015 1050 NEYDEEKPRQ KRKRSTKTKN SGESKRRKAA KKTLSDSDEDDDDWKKPSH NKGKKSQLSN EFIEDSDEEE

1077 AQMSGSEQNK NDDNDENNDN DDNDGLF FIGURE1.-Predicted amino acid sequence of Cdplp. The endpointsof the region deleted in strains YNN543, YNN544 and YNN545 are indicated by vertical arrows. The full nucleotide sequence of WPl is available from GenBank under accession number U31217. compound benomyl (the active breakdown product of medium. Figure 3 shows that the cdpl mutant cells are which is carbendazim) bind to tubulin heterodimers unable to grow on medium containing 80 pg/ml thia- and inhibit microtubule polymerization in vivo and in bendazole, whereas the growth of wild-type cells is not vitro (HOEBEKEet al. 1976; DAVIDSEand FLACH1977, eliminated until the concentration reaches 120 pg/ml. 1978; FRIEDMANand PLATZER1978) reviewed in DA- In contrast, the growth of wild-type cells is significantly VIDSE (1986). To determine the relative effects of these inhibited at 10 pg/ml benomyl, whereas 15 pg/ml is compounds on thecdpl mutants, serial dilutions of wild- required to observe a significant effect on the growth type and mutant cells were spotted onto solid media of the cdpl mutant cells. containing various concentrations of benomyl or thia- cdpl mutants exhibit defects in nuclear divisionand bendazole. Unexpectedly, although cdpl-1 and cdplA chromosome segregation: The nuclear morphology of cellswere considerably more sensitive than wild-type cdpl mutants was examined. Wild-type, cdpl-l and cells to thiabendazole in the medium,they grew signifi- cdplA cells were grownlogarithmically at 30" and then cantly better thanwild-type cellson benomyl-containing shifted to 37" for 3 hr. Cells were fixed and the nuclei Role of Cdplp in Nuclear Division 1393 A B

10

p

Ec

0 5 10 15 20 25 0 5 10 15 20 25 Hours at 37" Hours at 37" FI(;L'KE2.--C;rowth CIIWC and \kIbility of dj11 mutant cells at 37". YCRY3 (wild-type, O), YNN.540 (cdpl-f,0) and YNN.544 (u//IlA, ) crlls were grown logarithmically at 30" for sevrral hours anrl shifted to 37". (A) Optical density of cultures at various times following the temperature shift. Idrntical growth cunes are obtained if cells are counted rather than measuring their number by optical density. (R) Viability 01' cells was determined by removing aliqrmts of cells at the indicated times, counting on a haemocytometer and plating onto YPD plates. The fraction of cells that formed colonies on the platcs is indicated. were observed by DAPI staining. Nuclear staining pat- The DAPI stainingpattern of cdpl mutant cells was terns were scored as either anucleate,having two nuclei more diffuse than in wild-type cells. The diffuse DAPI per cell, or having three or morenuclei per cell. Figure staining was particularly pronounced at37". As a result, 4 shows the results of this analysis.Less than 1% of it was difficult to establish with certainty in some cases unbuckled or budded wild-type cells were binucleate how many nuclei were present and whether somecells and no cells were observed that had more than hvo truly lacked a nucleus. Only cells in which the nuclear nuclei. In contrast, cdpl mutant cells displayed abnor- configuration was unambiguous (>go% ) were scored. mal nuclear configurations at both the permissive and To determine whether the cdjll mutant? exhibited a restrictive temperatures. At 37", -35% of cdplA cells chromosome segregation defect, the chromosome loss contained hvo ormore nuclei, and -9% were an- frequency was determined in wild-type, cdj)l-l and ucleate.Multinucleate andanucleate cells were oh cclplA cells atthe permissive temperature. Diploid served in cells containing the cdpl-I allele at frequen- strains were constructed that containeda marked -1%- cies considerably lower than cells containing the null kb linear fragment of chromosome 111. Table 3 shows allele at 30". At 37", the frequency of cdpl-f cells dis- the loss frequency for this marked chromosome frag- playing abnormal nuclear configurationsincreased dra- ment in wild-type and cdjll mutant cells. At SO", the cdjll- matically, and was between one-third and one-half of I mutant lost the Chromosome fragment ata frequency the frequencyof rtIplA cells with similar configurations. -10- to %-fold higher than wild-type cells. In cd/jlA

FIGIIRE:3.-Differential benomyl sensitivity of mdfif mutant none (carbendazim) thiabendazole cells to microtubule-dis- nlpting compounds. Ap 5 10 15 40 80" 120 proximately 5000 logarith- WT WT mically grown YCRY2 (wild- type), YNN.536 (rdpf-f) and YNN.543 (+fA) cells Cdpl-1 wcrc spotted onto plates containingthe indicated compounds. The numbers cdp IA abo\re each colrlmn reflect the concentration of the compound in pg/ml. Cells were allowed to grow at "26" on YPD for 3 days, on thiabendazolefor 8 clays anrl on benomyl for 4 days. 1394 P. K. Foreman and R. W. Davis

TABLE 3 tained entirely within one cell (presumably the Chromosome loss in cdpl mutants mother). As a result, both daughter nuclei are con- tained within a single cell (Figure 5C). This phenome- Strain CDPl genotype Loss frequencyCFIII non is likely the mechanism by which binucleate and mutinucleate cells arise. YNN541 CLIP1 1.0 t 0.4 X 10-~ YNN542 cdpl-l 1.8 t 0.4 X lo-' YNN545 cdpl A 1.1 t 0.3 X 10" DISCUSSION This paper describes a novel yeast gene, CDPl, that cells, the loss rate was -60- to 230-fold higher than appears to play a general role in governing MT dynam- in wild-type cells. Both the cdpl-1 and cdplA mutants ics. Cellsdefective in Cdplp function exhibit abnormal therefore exhibit moresevere chromosome loss defects microtubule morphology typified by unusually long and than the cbflA mutant (BAKERand MASISON 1990; CAI numerous cytoplasmic MT bundles. Cytoplasmic MTs and DAVIS1990; FOREMANand DAVIS1993). function in migration of the nucleus into the budneck cdpl mutants exhibit abnormal microtubule morphol- before chromosome segregation and in orienting the ogy: The differential sensitivity of the cdpl mutants to mitotic spindle (JACOBS et al. 1988; PALMERet al. 1992; microtubule-disrupting drugs suggested the possibility SULLIVANand HUFFAKER1992). The frequent occur- that Cdplp might play a role in tubulin function. To rence of cells withimproperly oriented spindles as well further explore this possibility, microtubules were oh as binucleate and multinucleate cells suggeststhat these served in intact fixed cells by indirect immunofluores- processes are defective in cdpl mutants. In addition, cence staining using anti-tubulin antibodies. Figure 5A the fact that these mutants lose chromosomes at a sub- shows the pattern of staining observed in wild-type cells stantially elevated frequency suggests that nuclear MTs grown at 37". The top panels show a cytoplasmic micro- may also be altered in the cdpl mutant. tubule array characteristic of unbudded cells and the The differential sensitivity of cdpl mutant cells to the bottom panels show cellsin successive stages of mitosis. related microtubule-disrupting compounds thiabenda- Figure 5B shows the microtubule-staining pattern of zole and benomyl suggests the possibility that Cdplp cdpl-1 cellsgrown logarithmically for several genera- closely interacts with MTs or tubulin. Thesebenzimidaz- tions and shifted to 37" for 3 hr. At all stages of the cell ole compounds bind to tubulin heterodimers and in- cycle, >80% of these cells were larger and displayed hibit microtubule polymerization in vivo and in vitro dramatically longer arrays of cytoplasmic microtubules (HOEBEKEet al. 1976; DAVIDSEand FLACH1977, 1978; than were observed in wild-typecells. Inaddition, FRIEDMANand PLATZER1978) (reviewed in DAVIDSE whereas wild-type cells rarely show more than three cy- 1986).A single amino acid substitution in the 0-tubulin toplasmic MT bundles large enough to see by this gene of Aspergzllus nidulans causes differential sensitivity method, cdpl cells typically had at least four bundles to MT disrupting drugs UUNG and OAKLEY1990). How- emanating from each spindle pole. The relative abun- ever, mutations in othergenes have not previously been dance of MT bundles was particularly apparent in un- shown to have this effect. Studies of the binding of budded cells (compare top panels Figure 5, A and B) . these compounds to tubulins derived from a variety of At 30°, cdpl-1 cells were similar in appearance to wild- species have shown that the bindingaffinity apparently typecells, with 2-3% of cells containinglong cyto- solely determines whether or not a benzimidazole has plasmic MT arrays. cdplA cells displayed longer and antifungal activity (DAVIDSE1986; DAVIDSEand FLACH more numerouscytoplasmic MT bundles thanwild-type 1977). Therefore, one implication of our data might cells at both 30" and 37" (data not shown). be that in cdpl mutant cells the tubulin bindingaffinity In wild-type cells, the spindle projects through the of thiabendazole is increased and the binding affinity bud neck with one pole in the mother cell and the of benomyl is decreased. Such an effect might be oh other in the bud. In cdpl cells grown at the 37", we served if Cdplp normally interacts with tubulin directly occasionally (<5%) observed spindles that were con- or indirectly in a manner that alters the structure or

CDPl 0 ax@ 0oI@ CeJlS FIGURE 4.-Nuclear DNA in cdpl mutants. shah BdyPe Tern %m*tc %*with 2 % Celt with23 hunted DAF'Istained cells were scored for abnormal nu- & nuclei nu& clear morphology. Cells growing logarithmically at 30" or shifted to 37" for 3 hr were fixed and stained YiXY2 CDPl 30' 1.9 0.4 0 403 withwere CellsDAPI. scored for nuclear morphol- YiXY2 CDPl 31' 2.0 0.6 0 325 ogy and aberrant cells containing zero, two, or YNN.536 ~4pI-I 30' 4.4 2.2 0.4 265 more orthree nuclei per mother cell are tabu- 438 lated. Nuclei contained in presumptive daughter YNN536 C@I-1 31' 5.7 6.8 5 154 cells (always zero or one) are not included. YNN543 cdpld 30' 10.2 14.2 9.6 YNN543 cdpld 31' 9.2 20.1 14.1 129 Role of Cdplp in Nuclear1395 Division

tion of function alleles of ctfl, that cdpl mutants depend on Cbflpprimarily for its function in chromo- some segregation. It is possible that both Cdplp and Cbflp interactdirectly with MTs at the centromere. Alternatively, the cumulative effects of impaired MT function (either nuclearor cytoplasmic), due to a muta- tion in cdpl and impaired kinetochore function due to a cifl mutation, might result in an inability to sustain ongoing cell division. For example, one possibility is that in the absence of Cdplp, MT turnover might be sufficiently altered so that the number of encounters between nascent MTs and the kinetochore is lowered. In this case failure to capture a MT in a few encounters (perhaps as a result of defective cbflp function) or failure to maintainattachment could causechromo- some loss. If Cbflp were involved in improving theeffi- ciency of capturing or maintaining attachment ato na- scent kinetochoreMT (perhaps by altering centromere or kinetochore function), thecumulative effects of the two mutationsmight prove lethal. In this regard, it should be noted that neither the cdpl-I nor the cbjlA mutation causes morethan a modesteffect on cell growth in rich media making it unlikely that the syn- thetic lethality caused by the combined mutations is nonspecific. Numerous other models could also explain the relationship between CBFI and CDP1 and further efforts to characterize the genetic interaction are cur- rently underway. Comparison of the predicted sequenceof Cdplpwith other proteins in the databases revealed a mouse pro- tein, p150T”, that is notably similar insequence. p150Tsrhas been shown to localize to the nucleus in mouse cells and binds to peptides containing SH2 de mains (MALEK et nl. 1996).Two other anonymous ORFs from human and C. ekgnns also display considerable similarity to bothCdplp andpl5OTSr. p150Tsp and these ORFs contain putative tetratricopeptide repeat (TPR) FIGLIRE5.-.\licrotuldc staining pattcrn ol’ wild-type and rdjll mutant cells. The microtul,rtle morphology of wild-type sequences arranged in tandem(MAL.EK et al. 1996). The and cdpl mutant cells \vas determined bv indirect immuno- TPR motif is a %-amino acid motif that may mediate fluorescence (leftpanels). Nuclei werevisualized by DAPI interactions with other TPR-containingproteins (SI- staining (center panels). The DIC images of the correspond- KORSKI pt nl. 1990, 1991, 1993). For example,the ing fields areshown in the right panels. (A) YCRY3 (wild- Cdc27p TPR repeats are important in the interaction type) cells at progressive stages of the cell cycle. (R) YNN540 (rdpl-Z) cells at approximately equivalent stages of the cell betweenCdc27p and Cdc23p, which form part of a cvcle. (C) A YNN540 (cd/)l-l)cell in whichthe spindle is complex that promotes anaphase in yeast (LAMB pt nl. contained entirely within the mother cell. Bar, 10 pm. 1994). In the mouse protein, p150T”, the domain con- tainingthe TPRmediates self-association. Although accessibility of the benzimidazole binding site. Further p15Ors” and Cdplp sequences display a fairly high de- strldies will be required to test this notion as well as to gree of identity within the putative TPR repeats, the determinewhether Cdplp interactsdirectly or indi- predicted Cdplp sequence contains few of the signa- rectly with MTs or tubulin. ture amino acids found in TPR repeats. Furtherstudies In cells lacking Cdplp, the CBFl geneproduct is will be required to determine if these regions mediate essential forviability. The basis for thesynthetic lethality interactions with other TPR-containingproteins in between rffl and cdj~lmutations is unclear. Cbflp binds yeast. to an 8-bp sequence located both within centromeres Several other yeast proteins that may also play roles (CDEI) and upstream of several genes involved in me- in MT function have been identified by various ap thionine biosynthesis, where it plays a role in transcrip proaches, [r.g., Cdc20p (HARTWELLpt nl. 1973; HART- tional activation. M’e have demonstrated, using separa- WELL and SMITH1985; SETHI Pt nl. 1991); Cinlp,Cin2p, 1396 P. K. Foreman and R. W. Davis

Cin4p (HOYI et al. 1990; STEARNSet al. 1990); Biklp binding protein that is able to distinguish single base-pair muta- tions in its recognition site. Mol. Cell. Biol. 9: 2544-2550. (BERLINet al. 1990);help (PELLMANet al. 1995); Kar9p CAI, M., and R. W. DAVIS,1990 Yeast centromere binding protein (KURIHARAet al. 1994)l. In some instances, these pro- CBFl, of the helix-loop-helix protein family, is required forchro- teins have been shown to localizeto the spindle appara- mosome stability and methionineprototrophy. Cell 61: 437-446. CARBON,J., 1984 Yeast centromeres:structure and function. Cell tus (BERLINet al. 1990; BARNESet al. 1992). Like the 37: 351-353. CDPI gene, the CINl, CIN2, CIN4, and BIKl genes are CARBON,J., and L. CI~KE,1990 Centromere structure andfunction nonessential (BERLIN et al, 1990; HOYI et al. 1990; in budding and fission yeasts. New Biol. 2: 10-19. CASSAMAYOK,A,, M. ALDEA, C. CASAS, E. HERRERO,F.-J. GAMOel ul., STEARNSet al. 1990). However, in contrast to cdpl mu- 1995 DNA sequence analysis of a 13 kbp fragment of the left tant cells,which have unusually long and numerous arm of yeastchromosome XV containing seven new open reading bundles of MTs, cells individually lacking these gene frames. Yeast 11: 1281-1288. CIARKE,L., 1990 Centromeres of budding and fission yeasts. Trends products display diminished cytoplasmic MT arrays Genet. 6: 150-154. when grown at restrictive temperatures (BERLINet al. CIARKE,I,., and J.CAKBON, 1985 The structure and functionof yeast 1990; HOY~et al. 1990). In addition, overexpression of centromeres. Annu. Rev. Genet. 19: 29-56. DAVIDSF.,L. C., 1986 Benzimidazole fungicides: mechanism of ac- Biklp results in a transient increase in the length of tion and biological impact. Ann. Rev. Phytopathol. 24 43-65. cytoplasmic MTs (BERLINet at. 1990). Itwill be of inter- DAVIDSE,L. C., and W. FIACH,1977 Differential binding of methyl est to investigate the relationship between these gene benzimidazol-2yl carbamate to fungal tubulinas a mechanism of products and Cdplp. resistance to this antimitotic agent in mutant strains of Asperpllus nidulans. J. Cell Bid. 72: 174-193. Mutations in threeother genes (CBP2, mP3, and DAVIDSE,L. C., and W. FIACH, 1978 Interaction of thiabendazole CDP4) were identified in this study that cause depen- with fungal tubulin. Biochim. Biophys. Acta 543: 82-90. dence on CBFl for viability and hypersensitivity to thia- DOHENY,K. F., P. K. SORGER,A. A. HYMAN,S. TUGENDREICH,F. SPEN- (:PRet al., 1993 Identification of essential components of the bendazole. The products of these genes may constitute S. cerevisiae kinetochore. Cell 73: 761-774. additional proteins involved in MT dynamics. Since sin- EILEDGE,S. J., and R. W. DAVIS,1988 A family of versatile centrtr gle alleles of these genes were identified, the genetic rneric vectors designed foruse in the sectoring-shuffle mutagene- sis assay in Saccharomyces cerevisiae. Gene 70: 303-312. screen employed to identify the CDP genes is unlikely ESHEI.,D., L. A. URRESTARAZU,S. VISSERS, J. C. JAUNIAUX,C. J. VAN to be saturated. Therefore, additional new genes may VI.IET-REEDIJKrt al., 1993 Cytoplasmic dynein is required for be identified by this method. Analysis of these genes is normal nuclear segregation in yeast. Proc. Natl. Acad. Sci. USA 11172-11176. likely to provide new insights into the functionof MTs, FITZGERALD,H. M., 1987 Yeast centromeres. Yeast 3: 187-200. their structuralcomposition and the regulation of their FOREMAN,P. R,and R. W. DAVIS,1993 Point mutations that sepa- assembly. rate the role of Saccharomyces cereuisinecentromere binding factor 1 in chromosome segregationfrom its role in transcriptional The authors thankMARK WAHI.BERG and HEIGE ZIEIXRfor invalu- activation. Genetics 135: 287-296. able assistance with the genetic screen used in this study. We are FRIEDMAN,P. A., and E. G. PIATZER,1978 lnteraction of anthelmin- indebted to TIMSTEARNS and GEORIANA BARNESfor helpful discus- tic benzimidazoles and benzimidazole derivatives with bovine brain tubulin. Biochim. Biophys. Acta 544 605-614. sions aswell as LAURA "ALL, ANNEVII.I.ENEWE and ANNA Golf, P.Y., and J. V. KII.MARTIN,1993 NDC10: a gene involved in ASTROMOFF for comments on the manuscript. This work was sup- chromosome segregation in Sacrharomyces cerpuzsiar. J. Cell Biol. ported by National Institutes of Health grant 3 R37 HG009198. 121: 503-512. HARTMTI.~.,I.. H., and D. SMITH,1985 Altered fidelity of mitotic chromosome transmission in cell cycle mutants of Saccharo- LITERATURE CITED myces ccrevisiae. Genetics 110: 381 -395. HAR~T.I.I.,I.. H., R. K. MORTIMER,J. CIXOTTI and M. CUI.OTTI,1973 BAKER,R. E., and D. C. MMISON, 1990 Isolation of the gene encod- Genetic control of the cell division cycle in yeast. V. Genetic ing the Saccharomyces cerevisiae centromere-binding protein 74 CP1. Mol. Cell Bid. 10: 2458-2467. analysis of cdc mutants. Genetics 267-286. HEGEMANN, J. H.,J. H. SHERO,G. COTI'AKEI.,P. PHII.IPPSEN and BAKER,R. E., H. M. FITZGERAIDand T. C. O'BRIEN,1989 Purifica- P. tion of the yeast centromere binding protein CP1 and a muta- HIETER,1988 Mutational analysis of centromere DNA from tional analysisof its binding site. J. Bid. Chem. 264 10843- chromosome VI of Saccharomyces cerevisiae. Mol. Cell. Biol. 8: 10850. 2523-2535. BARNES,G., K. A. LOUIE andD. BOI'STEIN,1992 Yeast proteins associ- HOEBEKE, J.,G. VANNIJEN and M. DEBRABANUER,1976 Interaction of oncodazole (R17934), a new anti-tumoral drug with rat brain ated with microtubules in vitro and in vivo. Mol. Biol. Cell 3: 29- 47. tubulin. Biochem. Biophys. Res. Comm. 69: 319-324. Hour, M. A,, T. S'I'EARNYand D. BOTSTEIN,1990 Chromosome insta- BENDER,A., and J. R. PRINGIX,3991 Use of a screen for synthetic lethal and multicopy suppresseemutant3 to identify two new bility mutants of Saccharomyces cerevisiae that are defective in genes involved in morphogenesis in Saccharomyces cerpuisiae. Mol. microtubllle-mediated processes. Mol. Cell. Bid. 10: 223-234. Cell. Biol. 11: 1295-1305. Hour, M. A,, I,. HE, K R 1,oo and W. S. SAUNUERS,1992 Two Sac- BERLIN,V., C. A. STYT.ESand G. R. FINK,1990 BIK1, a protein re- charomyces cerevisiae kinesin-related gene products required for quired for microtubule function during mating and mitosis in mitotic spindle assembly. J. Cell Biol. 118: 109-120. Saccharomyces cerevisiae, colocalizes with tubulin. J. Cell Bid. Hour, M. A., L. HE, L. TOTISand W. S. SAUNDERS,1993 Loss of 111: 2573-2586. function of Saccharomyces cerevisiae kinesin-related ClN8 and BROWN,M. T., L. GOE'IXH and L. H. HARMI.L, 1993 MFZ is re- KIP1 is suppressed by KAR3 motor domain mutations. Genetics quired for mitotic spindle integrity duringanaphase spindle 135: 35-44. elongation in Succharomyres cerwisiae. J. Cell Biol. 123 387-403. ITO, H., K. FL~KUDA andA. KIMURA, 1983 Transformation of intact BYERS,B., 1981 Cytology of the yeast life cycle, pp. 59-71 in The yeast cells treated with alkali cations. J. Bacterid. 153: 163-168. MobcularBiology qthe Yeast Saccharomyces, edited by J. N. STRATH- JACOBS,C. W., A. E. M. ADAMS, P. J. SZANISZIL)andJ. R. PRINGI.E,1988 ERN,E. W. JONES and J. R. BROACH.Cold Spring HarborLabora- Functions of microtubules in the Saccharomyces cPrevisiuecel1 CyCk. tory, Cold Spring Harbor, NY. J. Cell Bid. 107: 1409-1426. CAI, M.,T.,and R. W. DAVIS,1989 Purification of ayeast centromere- JIAN(;, W. D., and P. PHII.IPPSEN,1989 Purification of a protein bind- Role of Cdplp in Nuclear Division 1397

ing to the CDEI subregion of Saccharomyces cerevisiae centro- proteins required forassembly of the mitotic spindle. J. Cell Biol. mere DNA. Mol. Cell. Biol. 9: 5585-5593. 118: 95-108. JIANG, W., J. LECHNERand J. CARBON, 1993 Isolation and character- ROTHSTEIN,R. J., 1983 One-step gene disruption in yeast. Methods ization of a gene (CBF2) specifjmg a protein component of the Enzymol. 101: 202-210. budding yeast kinetochore. J. Cell Biol. 121: 513-519. SAMBROOK,J., E. F. FRITSCHand T. WNIATIS,1989 Molecular Clon- JONES,E. W., and G. R. FINK.1981 Regulation of amino acid and ing, A Laboratoly Manual. Cold Spring Harbor Laboratory Press, nucleotide biosynthesis in yeast. pp. 181-299 in The Molecular Cold Spring Harbor, NY. Biology of the Yeast Saccharomyces: Metabolism and Gene Expression, SAUNDERS,W. S., and M. A. Hour, 1992 Kinesin-related proteins re- edited by J. R. BROACHand E. W. JONES.Cold Spring Harbor quired for structural integrity of the mitotic spindle. Cell 70: Laboratory, Cold Spring Harbor, NY. 451-458. JUNC, M. K., and B. R. OAKLEY,1990 Identification of an amino acid SAUNDERS,W. S., D. KOSHLAND,D. ESHEI.,I. R. GIBBONSand M.A. substitution in the benA, P-tubulin gene of Aspergillus nidulans Hour, 1995 Saccharomyces cerevisiae kinesin- and dynein-re- that confers thiabendazole resistance and benomyl supersensitiv- lated proteins required for anaphase chromosome segregation. ity. Cell Motil. Cytoskel. 17: 87-94. J. Cell Biol. 128: 617-624. KRANZ, J., and C. HOLM,1990 Cloning by function: an alternative SETHI, N., M. C. MONTEAGUDO,D. KOSHLAND,E. HOGANand D. J. approach for identifying yeast homologs of genes from other BURKE,1991 The CDC2O gene product of Saccharomyces cere- organisms. Proc. Natl. Acad. Sci. 87: 6629-6633. visiae, a beta-transducin homolog, is required for a subset of KURIHARA,L. J., C. T. BEH, M. LATTERICH, R. SCHEKMANand M. D. microtubuledependent cellular processes. Mol. Cell. Biol. 11: Rost, 1994 Nuclearcongression andmembrane fusion: two 5592-5602. distinct events in the yeast karyogamy pathway. J. Cell Biol. 126: SHERMAN,F., 1991 Getting started with yeast, pp. 3-20 in Guide to 91 1-923. YeastGenetics and , edited by C. GUTHRIEand WR,J. R., W. A. MICHAUD,R. S. SIKORSKIand P. A. HIETER,1994 G. R. FINK.Academic Press, San Diego. CdclGp, Cdc23p and Cdc27p form a complexessential for mito- SHERO,J. H., M. KOVAL,F. SPENCER,R. E. PALMER,P. HIETERet al., sis. EMBO J. 13: 4321-4328. 1991 Analysis of chromosome segregation in Saccharomyces LECHN~R,J., 1994 A zinc finger protein, essential for chromosome cerevisiae. Methods Enzymol 194: 749-773. segregation, constitutes a putative DNA binding subunit of the SIKORSKI,R. S., and P. HIETER,1989 A system of shuttle vectors and Saccharomyces cerevisiae kinetochore complex, Cbf'3. EMBO J. yeast host strains designed for efficient manipulation of DNA in 13: 5203-5211. Saccharomyces cmvisiae. Genetics 122: 19-27. LECHNF.R,J., and J. CARBON, 1991 A 240kd multisubunit protein SIKORSKI,R. S., M. S. BOGUSKI,M. GOEBL.and P. HIETER,1990 A complex, CBF3, is a major component of the budding yeast cen- repeating aminoacid motif in CDC23 defines afamily of proteins tromere. Cell 64: 717-725. and a new relationship among genes required for mitosis and LI, Y. Y., E. YEH, T. HAYSand K. BLOOM,1993 Disruption of mitotic RNA synthesis. Cell 60: 307-317. spindle orientation in a yeast dynein mutant. Proc. Natl. Acad. SIKORSKI,R. S., W. A. MICHAUD,J. C. WOOTTON,M. S. BOGUSKI,C. Sci. USA 90: 10096-10100. CONNELLYet al., 1991 TPR proteins as essential components of MALEK, S. N., C. H. YANG,W. C. EARNSHAW,C. A. KOZAKand S. DESI- the yeast cell cycle. Cold Spring Harbor Symp. Quant. Biol. 56 DERIO,1996 pl50TSP, a conserved nuclearphosphoprotein 663-673. that contains multiple tetratricopepetide repeats and binds spe- SIKORSKI,R. S., W. A. MICHAUDand P. HIETER,1993 p62cdc23 of cifically to SH2 domains. J. Biol. Chem. 271: 6952-6962. Saccharomyces cerevisiae: a nuclear tetratricopeptide repeat pro- MELLOR,J., W. JIANG,M. FUNK,J. RATHJEN, C. A. BARNESet al., 1990 tein with two mutable domains. Mol. Cell. Biol. 13 1212-1221. CPF1, a yeast protein which functions in centromeres and pro- STEARNS,T., M. A. HOW and D. BOTSTEIN,1990 Yeast mutants sensi- moters. EMBO J. 9: 4017-4026. tive to antimicrotubule drugs define three genes that affect mi- MIDDLETON,R, and J. CARBON, 1994KARSencoded kinesin is a crotubule function. Genetics 124 251-262. minusenddirected motor thatfunctions with centromere bind- STRATHERN,J. N. and D.R. HIGGINS,1991 Recoveryof plasmids ing proteins (CBF3) on an in vitro yeast kinetochore. Proc. Natl. from yeast into Escherichia coli: shuttle vectors, pp. 319-329 in Acad. Sci. USA 91: 7212-7216. Guide to Yeast Genetics and Molecular Biology, edited by C. GUTHRIE MURPHY,M., and H. M. FITZGERALD,1990 Cis- and trans-acting fac- and G. R. FINK.Academic Press, San Diego, CA. tors involved in centromere function in Saccharomyces cerevis- STRUNNIKOV,A. V., J. KINGSBURY and D. KOSHLAND,1995 CEP3 en- iae. Mol. Microbiol. 4 329-336. codes a centromere protein of Saccharomyces cerevisiae. J. Cell O'CONNEI.I.,K. F., Y. SURDIN-KERJANand R. E. BAKER,1995 Role of Biol. 128: 749-760. the Saccharomyces cerevisiae general regulatory factor CP1 in SULLIVAN,D. S., and T. C. HUFFAKER,1992 Astral microtubules are methionine biosynthetic gene transcription. Mol. Cell. Biol. 15: not required for anaphaseB in Saccharomyces cerevisiae. J. Cell 1879-1888. Biol. 119: 379-388. PALMER,R. E., D. S. SULLIVAN,T. HUFFAKERand D. KOSHLAND,1992 SULSTON,J., Z. Du, K. THOMAS,R. WIISON,L. HILLIERet al., 1992 Role of astral microtubules and actin in spindle orientation and Nature 356: 37-41. migration in the budding yeast, Saccharomycescerevisiae. J. Cell THOMAS,D., I. JACQUEMIN and K. Y. SURDIN,1992 MET4, a leucine Biol. 119 583-593. zipper protein, and centromere-binding factor 1 are both re- quired for transcriptional activation of sulfur metabolism in Sac- PELLMAN,D., M. BAGGET,H. TU andG. R. FINK,1995 Two microtu- charomyces cerevisiae. Mol. Cell. Biol. 12: 1719-1727. bule-associated proteins required for anaphase spindle move- WINEY,M., C. L. WY,E. T. O'TOOLE,D. N. WTRONARDE,T. H.J. ment in Saccharomyces cerevisiae. J. Cell Biol. 130 1373-1385. GIDDINGSet al., 1995 Threedimensional ultrastructural analysis PRINGLE,J. R., A. E. ADAMS, D. G. DRUBINand B. K. HAARER,1991 of the Saccharomyces cerevisiae mitotic spindle. J.Cell Biol. 129: Immunofluorescence methods foryeast. Methods Enzymol. 194: 1601-1615. 565-602. ZIELER,H. A., M. WALBERGand P. BERG,1995 Suppression of muta- RAMER,S. W., S. J. ELLEDCEand R. W. DAVIS,1992 Dominant genet- tions in two Saccharomyces cerevisiae genes by the adenovirus ics using a yeast genomic library under the control of a strong E1A protein. Mol. Cell. Biol. 15: 3227-3237. inducible promoter.Proc. Natl. Acad. Sci. USA 89: 11589- 11593. ROOF, D.M., P. B. MEI.UH and M.D. ROSE, 1992 Kinesin-related Communicating editor: F. WINSTON