Copyright 2001 by the Genetics Society of America

The Kar3-Interacting Protein Cik1p Plays a Critical Role in Passage ThroughMeiosis I in Saccharomyces cerevisiae

Robert M. Q. Shanks, Rebecca J. Kamieniecki and Dean S. Dawson

Department of Molecular Microbiology, Tufts University, Boston, Massachusetts 02111

Manuscript received March 21, 2001Accepted for publication August 16, 2001

ABSTRACTMeiosis I in Saccharomyces cerevisiae is dependent upon the motor protein Kar3. Absence of Kar3p in

meiosis results in an arrest in prophase I. Cik1p and Vik1p are kinesin-associated proteins known tomodulate the function of Kar3p in the microtubule-dependent processes of karyogamy and mitosis.Experiments were performed to determine whether Cik1p and Vik1p are also important for the functionof Kar3p during meiosis. The meiotic phenotypes of a cik1 mutant were found to be similar to those ofkar3 mutants. Cells without Cik1p exhibit a meiotic defect in homologous recombination and synaptonemalcomplex formation. Most cik1 mutant cells, like kar3 mutants, arrest in meiotic prophase; however, in cik1mutants this arrest is less severe. These data are consistent with the model that Cik1p is necessary forsome, but not all, of the roles of Kar3p in meiosis I. vik1 mutants sporulate at wild-type levels, but havereduced spore viability. This loss in viability is partially attributable to vegetative chromosome loss in vik1diploids. Cellular localization experiments reveal that Kar3p, Cik1p, and Vik1p are present throughoutmeiosis and are consistent with Cik1p and Vik1p having different meiotic roles.

MEIOSIS, mitosis, and karyogamy (nuclear fusion) the forces of two other kinesin-like motor proteins,depend upon microtubule function. In the bud- Kip1p and Cin8p (Saunders et al. 1997b). Certain muta-

ding yeast Saccharomyces cerevisiae, one kinesin-like motor tions in kar3 can partially suppress the inviable pheno-protein, Kar3p, has been reported to function in all of type of kip1, cin8-3 double mutants at restrictive condi-these processes. Kar3p is essential for karyogamy (Meluh tions (Saunders et al. 1997b). kar3 mutants areand Rose 1990). In newly formed zygotes, Kar3p is dependent on several of the mitotic spindle checkpointthought to function by cross-bridging microtubules be- components for viability (Roof et al. 1991; Hardwicktween fusing nuclei, drawing them together by minus- et al. 1999). Finally, Kar3p has been shown to have aend-directed forces (Meluh and Rose 1990). Kar3p is role in mitotic spindle positioning (Cottingham et al.important but not essential for mitosis. Vegetative kar3 1999).mutant cells display a number of phenotypes, including Kar3p is essential for meiosis (Bascom-Slack andslow growth, which is the consequence of �40% of cells Dawson 1997). kar3 cells arrest reversibly in prophaseexperiencing cell cycle arrest, temperature sensitivity, of meiosis I, with bushy prophase spindles, incompleteshort bipolar spindles, and longer, more numerous cyto- synaptonemal complex (SC) formation, modest levelsplasmic microtubules (Meluh and Rose 1990; Roof et of double-strand breaks, and highly reduced levels ofal. 1991; Saunders et al. 1997b). The temperature sensi- heteroallelic recombination (Bascom-Slack and Daw-tivity of vegetatively growing kar3 mutants is thought to son 1997).be attributable to the exaggerated cytoplasmic microtu- In mating and vegetative cells it has been demon-bules, as the temperature sensitivity is suppressible by strated that the activity of Kar3p is influenced by interac-the addition of microtubule-destabilizing drugs, such as tions with other kinesin-associated proteins (KAPs) thatbenomyl, or by mutations in certain genes required specify its function (Manning et al. 1999). Kar3p hasfor microtubule assembly and stability, such as BIM1 two known KAPs, Cik1p and Vik1p, which have been(Tirnauer et al. 1999) and TUB3 (Saunders et al. implicated in directing Kar3p function in different mi-1997b). These observations suggest that an important crotubule-dependent events in mitotic cells (Page andmitotic role of Kar3p is the mediation of microtubule Snyder 1992; Page et al. 1994). For example, the invia-depolymerization (Saunders et al. 1997b), an activity bility of cin8-3, kip1 double mutants is suppressible bythat has been demonstrated for the protein in vitro (En- a third mutation in either VIK1 or KAR3 mutants, butdow et al.1994). In mitosis, Kar3p is thought to oppose not in CIK1 (Manning et al. 1999). Conversely, cik1 and

kar3 mutants are temperature sensitive, whereas vik1mutants are not (Manning et al. 1999). While Vik1p

Corresponding author: Dean S. Dawson, Department of Molecular has no known role in karyogamy (Manning et al. 1999),Biology, Tufts University, 136 Harrison Ave., M & V 404, Boston, MA02111. E-mail: dean.dawson@tufts.edu Cik1p has been shown to be responsible for the cyto-

Genetics 159: 939–951 (November 2001)

940 R. M. Q. Shanks, R. J. Kamieniecki and D. S. Dawson

bp of KAR3 DNA (205 bp upstream to 2211 bp downstreamplasmic localization of Kar3p during this process (Pageof the start codon), flanked by EagI and SalI restriction enzymeet al. 1994). Cik1p and Vik1p also appear to affect thesites introduced by PCR.

distribution of Kar3p along the mitotic spindle (Man- pRS40 and pRS41 were made by ligating a VIK1-containingning et al. 1999). Another study showed that the Kar3p- NgoMIV-KpnI fragment of pRS316-VIK1, a gift from Mike Sny-

der, into the integrating plasmids pRS404 and pRS406 (Sikor-Cik1p complex, but not the Kar3p-Vik1p complex, isski and Heiter 1989), respectively. These plasmids were di-involved in spindle positioning; however, both com-gested with NruI to direct their integration into the chromosomeplexes are involved in spindle integrity (Cottinghamat the VIK1 locus.

et al. 1999). pD168 is composed of a pUN105 vector (Elledge andThe specific functions of Kar3p in meiosis are unclear. Davis 1988) that carries a KAR3-GFP fusion gene under the

control of the KAR3 promoter. The green fluorescent proteinOne approach to addressing this issue is to determine(GFP) sequence (a gift from Aaron Straight) has the S65T,the meiotic roles of proteins known to interact withV163A, and S175G mutations that result in a thermoresistantKar3p in other cell cycle stages. Here we characterizeGFP (Siemering et al. 1996).

the meiotic phenotypes of cik1 and vik1 mutants. The pRM2 was constructed by ligating a HpaI fragment con-meiotic phenotypes of cik1 mutants are consistent with taining the ARG4 open reading frame and promoter into the

HpaI site of ADE1 on pXW123 (a gift from Jim Haber).the model that Cik1p is critical for the function of Kar3pcik1 deletions were constructed by transforming a haploidin prophase of meiosis I. vik1 mutants have a spore

strain with a PCR-generated KanMX cassette (Guldener et al.viability defect that is at least partially attributable to1996) flanked by 40 bp of CIK1 sequence corresponding to

a mitotic chromosome maintenance defect, suggesting the beginning and end of the CIK1 open reading frame. G418that its role in meiosis is either not essential or can be [200 mg/liter in YPD, geneticin from GIBCO BRL (Gaithers-

burg, MD), product no. 11811]-resistant colonies were se-provided by other proteins. We localized Kar3p, Cik1p,lected and tested for temperature sensitivity. Temperature-and Vik1p in meiotic cells and observed differentialsensitive candidates were identified, and the correct insertionstaining between Cik1p and Vik1p, further suggestingwas confirmed by Southern blot analysis (Southern 1975;

that these proteins have different meiotic roles. data not shown). With regard to temperature sensitivity, inabil-ity to mate, and microtubule morphology, strains bearing thecik1::KanMX allele behave phenotypically as reported pre-

MATERIALS AND METHODS viously for cik1 null mutants (Page and Snyder 1992).vik1 deletions were constructed by transforming a haploidYeast strains, culture techniques, and cytology: Genotypes

strain with a PCR-generated KanMX cassette with 45 bp ofof strains used in this study are listed in Table 1. All of theseVIK1-flanking homology (Guldener et al. 1996). vik1 mutantsstrains are congenic derivatives of S288C (Nicolas et al. 1989)were confirmed by PCR.except where noted in text. Strain 2003 was a gift from Dan

ADE1 was disrupted by ARG4 by transforming yeast withBurke.pRM2 that had been digested with XhoI. Arg� prototrophsYeast cell culture and cytological techniques, including im-were screened for red colonies to confirm the ADE1 disrup-munofluorescence of fixed and spread cells, were as describedtion.previously (Bascom-Slack and Dawson 1997; Kamieniecki

The carboxy termini of Kar3p, Cik1p, and Vik1p wereet al. 2000). The sporulation regimen in this report was per-tagged with 13 � Myc, using the PCR-based gene modificationformed as described previously (Bascom-Slack and Dawsonsystem of Longtine et al. (1998). The KANMX cassette (from1997) with the exception that we added adenine to the sporula-pFA6a-13Myc-KanMX) was used for tagging KAR3, while thetion medium to a final concentration of 10 �g/ml. WeaklyHIS3MX cassette (from pFA6a-13Myc-His3MX6) was used forstaining cytoplasmic foci were observed in experiments usingtagging CIK1 and VIK1.anti-Myc primary antibody. These signals were also seen in

Genetic techniques: Commitment to heteroallelic interho-control experiments with a strain carrying no Myc-epitope-molog recombination, return to vegetative growth, and thetagged genes. Therefore, weak cytoplasmic staining was ig-quantitative viable haploid formation analyses were as de-nored in our experimental strains. Images were collected usingscribed previously (Bascom-Slack and Dawson 1997).an Olympus BX60 microscope with a 1.4 NA �100 objective

Chromosome I nondisjunction frequencies: Chromosome I non-and a Hamumatsu (Bridgewater, NJ) model C4742-95 cooleddisjunction frequencies were determined in DRS144 andCCD camera. Images were collected and manipulated usingDRS146. In these strains, one copy of chromosome I had beenOpenlab 2.2 software.modified such that ADE1 was replaced by ARG4. Both diploidsPlasmid construction, gene disruptions, and gene tagging:are also homozygous for the CAN1 gene, which confers sensi-CIK1 was cloned by complementing the cik1 temperature-tivity to canavanine. Cultures were sporulated in liquid me-sensitive phenotype as described by others (Page and Snyderdium for 48–72 hr and then were treated with zymolyase and1992). The complementing plasmid, dubbed pRS1, containsTriton X-100 to kill nonspores (Herman and Rine 1997).a 9606-bp fragment of chromosome XIII DNA (SaccharomycesSpores were then plated onto complete medium lacking ade-Genome Database, chromosome XIII coordinates 657470–nine and arginine. Among cells able to grow on this medium667076). This fragment contains the entire open readingare haploids disomic for chromosome I and unsporulatedframe of CIK1. We have since subcloned a SacI-KpnI fragmentdiploids that survived the zymolyase/detergent treatment.containing only the CIK1 open reading frame and sur-[Spores were also plated onto rich medium to determine therounding untranslated sequences into pRS415 (Sikorski andfrequency of colony-forming units (CFU) in the detergent-Heiter 1989) to create pRS12, which complements the sporu-treated cells.] To distinguish between these categories, colo-lation defect of cik1 mutants as well as pRS1.nies were tested to determine the ploidy of chromosome VpRS3 was made by inserting a KAR3-containing fragmentusing a previously described method (Zeng and Saunders(EagI-SalI; see below) into pCY204, which is a YCP50 vector2000). Fifty patches of Ade� Arg� colonies on a YPD platecontaining HO under the control of its own promoter (Rus-

sell et al. 1986). The KAR3-containing insert consists of 2416 were replica plated to medium containing canavanine and

941Meiosis in cik1 Mutants

TABLE 1

Strains used in this study

Strain Genotype

2003 MATa leu2 trp1 ura3-52 can1 hom3 cyh2 ade6 NDC10::GFP-TRP1 ade2 his3Dd513.3A MATa leu2-3,112 trp1-289 can1 ura3-52 his3�1 ade2 kar3::HIS3Dd513.5C MAT� PGK1::TRP1 leu2::LYS2 lys2�B trp1-289 ura3-52 cyh2-1 his3�1 kar3::HIS3DRS49.6B MATa PGK1::TRP1 leu2-3,112 lys2�B trp1-289 ura3-52 cyh2-1 his3�1 cik1::KANMX4DRS49.7B MATa leu2-3,112 LYS2 trp1-289 ura3-52 cyh2-1 his3�1 cik1::KANMX4DRS50.5D MAT� leu2-3,112 trp1-289 can1 ura3-52 his3�1 ade2 cik1::KANMX4DRS60.1D MATa leu2-3,112 trp1-289 ura3-52 arg4�57,RV ade2 cik1::KANMX4DRS61.11D MAT� trp1-289 can1 ura3-52 arg4�42 his3�1 ade2 cik1::KANMX4DRS137.8D MATa leu2-3,112 trp1-289 ura3-52 arg4�42 his3�1 vik1::KANMX4DRS137.14C MAT� leu2-3,112 trp1-289 ura3-52 arg4�42 his3�1 vik1::KANMX4DRS104.11C MATa lys2 ura3 arg4 ade2 his3 TUB4::GFP-HISMX6DRS144.1 MAT� leu-3,112 trp1-289 ura3-52 arg4�42 his3�1 VIK1 ade1::ARG4/ADE1 (disomic for chromosome I)DRS144.2 MAT� leu-3,112 trp1-289 ura3-52 arg4�42 his3�1 vik1::KANMX4 ade1::ARG4/ADE1 (disomic for chromosome I)DRSK22.2A MAT� leu2 lys2 trp1 ura3 his3 KAR3::13xMyc-KANMX6DRSK22.10A MATa leu2 lys2 trp1 ura3 his3 KAR3::13xMyc-KANMX6DRSK23.3A MAT� leu2 lys2 trp1 ura3 his3 CIK1::13xMyc-HIS3MX6DRSK23.3B MATa leu2 LYS2 trp1 ura3 his3 CIK1::13xMyc-HIS3MX6DRSK28.2A MAT� leu2 LYS2 trp1 ura3 his3 VIK1::13xMyc-HIS3MX6DRSK28.2B MATa leu2 lys2 trp1 ura3 his3 VIK1::13xMyc-HIS3MX6TRK73 MAT� leu2-3,112 trp1-289 ura3-52 his3�1 CIK1::13xMyc-HIS3MX6TRK80 MAT� leu2-3,112 trp1-289 ura3-52 his3�1 KAR3::13xMyc-KANMX6TRS89 as DRS137.8D but vik1::KAN::TRP1::VIK1TRS90 as DRS137.14C but vik1::KAN::URA3::VIK1TRS93 as DRS137.8D but vik1::KAN::TRP1::VIK1 ade1::ARG4TRS95 as DRS137.14C but ade1::ARG4 ura3-52::URA3DJ1 Dd513.5C � Dd513.3ADRS51 DRS49.6B � DRS50.5DDRS52 DRS49.7B � DRS50.5DDRS65 DRS60.1D � DRS61.11DDRS106 DRS104.11C � TRK73DRS142 DRS137.8D � DRS137.14CDRS143 TRS89 � TRS90DRS144 TRS90 � TRS93DRS146 DRS137.8D � TRS95DRS151 2003 � TRK80DRSK38 DRSK23.3A � DRSK23.3BDRSK40 DRSK22.2A � DRSK22.10ADRSK42 DRSK28.2A � DRSK 28.2B

were then subjected to 0 or 5000 �J of UV irradiation from cultures of both DRS144 and DRS146 were grown in liquidmedium lacking both adenine and arginine to maintain botha Stratagene (La Jolla, CA) Stratolinker model 2400 and al-

lowed to grow for 2 days at 30�. Cells with an intact copy of copies of chromosome I. When these cultures reached 1–2 �107 cells/ml they were plated onto YPD and allowed to growthe CAN1 gene (found on chromosome V ) die on medium

containing canavanine. Therefore, haploids are much more for 5 days at 30�. The frequency of red-sectored colonies wasdetermined (DRS144 had 0 red-sectored colonies/16,963 totallikely than diploids to survive on canavanine-containing me-

dium after mild mutagenesis. Candidate Arg� Ade� patches colonies, while DRS146 had 38 red-sectored colonies/11,829total colonies).with high rates of papillation on canavanine medium were

considered haploid. All colonies classified as haploid on the Haploid vegetative chromosome I loss frequency: The rate of vege-tative chromosome I loss in haploids was determined in hap-basis of the chromosome V analysis behaved as maters in

mating-type tests. The frequency of chromosome I disomy in loid strains disomic for chromosome I. A number of stepswere performed to construct the strains used in this experi-haploid spore products of DRS144 and DRS146 was deter-

mined as follows: frequency of haploid chromosome I disomy � ment. First, DRS144 was sporulated and the cultures weretreated with zymolyase and detergent to kill nonspores (seeAde� Arg� disomes (CFU/ml of Ade� Arg� prototrophs �

frequency of chromosome V monosomes)/total haploids above). Ade� Arg� colonies were isolated and potential hap-loids were screened for, using a mating-type test. Ade� Arg�(CFU/ml). These experiments were done in parallel strains

and the results were averaged. maters (presumably either haploids disomic for chromosomeI or diploids expressing a single mating type) were subjected toDiploid vegetative chromosome I loss frequency: The rate of vegeta-

tive chromosome I loss in diploid cells was determined in the chromosome V canavanine test (see above) to determinechromosome V ploidy and thus distinguish disomic haploidsthe above-mentioned strains, DRS144 and DRS146. Triplicate

942 R. M. Q. Shanks, R. J. Kamieniecki and D. S. Dawson

from diploids. Chromosome XVI ploidy could also be deducedbecause the two VIK1 loci were differentially marked, one byTRP1 and the other by URA3 (see Table 1). Colonies thatwere disomic for chromosome I, monosomic for chromosomeV, exhibited a or � mating behavior, and carried the URA3(but not the TRP1) copy of chromosome XVI were used togenerate vik1::KANMX and VIK1 isogenic strains carrying thechromosome I disome. These cells were treated with 5-fluoro-orotic acid (5-FOA) to identify cells that had “looped out” theURA3 gene integrated at the VIK1 locus. Because the URA3gene was flanked on one side by the wild-type VIK1 and onthe other side by a vik1::KANMX allele, the FOAR derivativescould be either Vik1� or Vik1�. The VIK1 genotype of theFOAR derivatives was determined, and an isogenic VIK1 andvik1 pair was used to determine the loss frequency of chromo-some I. The assay was the same as in the chromosome lossassay used for diploids discussed above (vik1� had 38 red-sectored CFU/9667 total CFU, while VIK1 had 2 red-sectoredCFU/8145 total CFU).

Figure 1.—cik1 mutants are unable to efficiently form viableRESULTS haploid meiotic products. The frequency of haploidization

was monitored by determining the frequency of ade2, can1,cik1 mutants are defective in spore formation: To and cyh2 cells, following the introduction of isogenic cik1determine the sporulation phenotype conferred by a (DRS52) and CIK1 (DRS52 � pRS1) diploid strains (can1/cik1 null mutation, a diploid cik1 strain (DRS52) was CAN1, cyh1/CYH1, and ade2/ADE2) into sporulation medium.

Aliquots of cells harvested from the cultures at timed intervalsinduced to undergo meiosis. The same strain bearingwere plated on rich medium and complete medium lackinga CIK1 centromeric plasmid was evaluated in parallelarginine and supplemented with cyclohexamide and canavan-as an isogenic control. Cells that had been incubated ine. The frequency of haploidization was determined by com-

in sporulation medium for 96 hr were gathered and paring the fraction of Ade�, Canr, and Cyhr CFUs at each timeobserved with bright-field microscopy. While CIK1 cells point relative to the total CFUs on the nonselective medium.

For cik1 mutants at least 1.2 � 106 cells were plated at eachexhibited 25.4% ascus formation, cik1 strains formedtime point. The haploid production frequency was also deter-no asci at all (n 210). The complete absence of sporu-mined for an isogenic pair of strains, one kar3 (DJ1) and thelation observed here contrasts with a previous report other KAR3 (DJ1 � pRS3). At all time points kar3 mutants

that cik1� cells, also in an S288C-derived strain back- produced haploids below the level of detection (�10�6).ground, exhibited a considerably reduced sporulationefficiency and the formation of dyads rather than tetradsfor those cells that did form asci (Kurihara et al. 1996). and Cyhr colonies produced by the cik1 strains were

indeed haploids, we tested the mating types of five iso-A more sensitive genetic experiment was performedto test the severity of the meiotic defect in cik1 mutants lates. All five that were tested mated with either a MATa

or a MAT� strain, suggesting that they contained a singleby directly selecting for viable haploid meiotic productsformed by a diploid cik1 strain (DRS52). DRS52 is het- copy of chromosome III and supporting the conclusion

that cik1 mutants can complete meiosis, albeit very inef-erozygous for alleles that confer resistance to the drugscanavanine and cyclohexamide (CAN1S/can1r, CYH2S/ ficiently. In a parallel experiment we monitored the

ability of kar3 mutants to form viable haploid products.cyh2r). For both of these genes, the allele conferringsensitivity to the drug is dominant, such that the diploid Consistent with an earlier report (Bascom-Slack and

Dawson 1997), the kar3 strain yielded no haploid colo-(DRS52) will die on medium containing either of thesedrugs. If DRS52 were able to sporulate, 25% of its spores nies among 106 cells tested.

cik1 mutants arrest in prophase I of meiosis: To deter-would inherit both the can1r and cyh2r alleles, allowingthem to live in a medium containing both drugs. Addi- mine the meiotic stage at which cik1 mutants are

blocked, we monitored the fraction of cells that wouldtionally, DRS52 is heterozygous at the ADE2 locus(ADE2/ade2) so that it produces white colonies, while progress beyond meiotic prophase I. Cells categorized as

beyond prophase I included unbudded cells containingone-half of its haploid products would inherit the ade2allele and would produce red colonies. To determine two or more 4,6-diamidino-2-phenylindole (DAPI) di-

hydrochloride (a DNA staining dye)-stained chromatinthe frequency of haploid production, we scored produc-tion of Canr, Cyhr, red colonies in aliquots of cells har- masses and/or bipolar spindles. In CIK1 cells, 34.5% of

cells were postprophase at 46 hr. By 46 hr only 2.5% ofvested from synchronous meiotic cultures of cik1 (DRS52)and CIK1 (DRS52 � pRS1) strains (Figure 1). The cik1 cik1 mutants had proceeded beyond prophase I (Figure

2A). Of the cells that had gone beyond prophase I,strain produced Ade�, Canr, and Cyhr CFUs at 0.2%the efficiency of the isogenic strain harboring a CIK1 DAPI staining revealed equal numbers of binucleate

and tetranucleate cells, although no spores were ob-plasmid. To test further whether the rare Ade�, Canr,

943Meiosis in cik1 Mutants

Figure 2.—cik1 mutantcells arrest before meta-phase I. (A) The frequencyof cells exhibiting postpro-phase I morphologies wasdetermined for both thewild-type (DRS51 � pRS1)and cik1 (DRS51) strains at22 and 46 hr after meioticinduction. The meiotic stageto which cells had progressedwas determined by evaluat-ing both nuclear morphol-ogy (determined by DAPIstaining; B) and spindle mor-phology (indirect immuno-fluorescent staining of mi-crotubules; C). Results are

tabulated in A (n � 175). B and C show a field of cik1 mutants at 22 hr. Most cells arrest as mononucleate cells with a monopolarmicrotubule array, but some cells proceed to form bi- and multinucleate cells (top right in B). Bar, 2 �m.

served. The majority of cik1 cells arrested at prophase lute (Figure 3B). By 72 hr after meiotic induction, re-combinants were detectable at �1% of the level exhib-I with a single chromatin mass and monopolar tubulin

arrays (see Figure 2, B and C). ited by the CIK1 control. This result is similar to thatseen in kar3 mutants (Bascom-Slack and DawsonThe cik1 meiotic arrest is reversible: To test whether

the cik1 meiotic arrest is the consequence of irreversible 1997).Strains deficient for Cik1p form an incomplete synap-damage we determined whether cik1 cells that were ar-

rested in meiosis could return to mitotic growth when tonemal complex: SC, a tripartite proteinaceous matrixthat forms between synapsed homologs, is a definingplated on rich medium. The number of CFUs was deter-

mined by gathering cells throughout meiosis and plat- landmark of meiotic prophase. To determine if cik1mutants form SC we used indirect immunofluorescenceing dilutions on rich medium (YPD). The viability of

cik1 strains remains high for the first 24 hr (by which to observe Zip1p, a component of the central elementof the SC (Sym et al. 1993), in cik1 and CIK1 cells.time the cells have reached their prophase arrest), after

which there is a gradual drop in viability while an iso- Aliquots of synchronous meiotic cultures were harvestedat 12.5, 13.5, and 14.5 hr postmeiotic induction. Spreadgenic CIK1 strain retains high levels of viability for at

least 72 hr (Figure 3A). Ability of the cik1 mutants to nuclei (n 489 per strain) from these cultures revealthat, while approximately equal fractions of nuclei fromresume growth if returned to rich medium indicates that

meiotic arrest is not attributable to irreversible cellular wild-type and mutant strains had Zip1p staining (10%for cik1, 12% for CIK1), the cik1 strain was severelydamage. Instead this result suggests the block in meiotic

progress is a regulatory arrest rather than a lethal event. defective in assembling the Zip1p into the continuousworm-like deposits that are characteristic of mature SC.Consistent with this notion, rare cik1 cells are observed

to proceed beyond prophase I and form bi- or tetranu- Among the Zip1p-staining CIK1 cells, 49% showed mul-tiple worm-like structures suggestive of complete SCcleate cells with normal-looking spindles and DAPI

masses (Figure 2). formation. In contrast, in cik1 cells evidence of completesynapsis of all of the chromosome pairs was never seen.Heteroallelic interhomolog recombination is impaired

in cik1 mutants: A landmark of prophase of meiosis I is Of the Zip1p-staining cik1 cells, 17% exhibited someworm-like staining (as in Figure 4B), but we never ob-the commitment to meiotic levels of interhomolog re-

combination (Sherman and Roman 1963; Esposito served more than approximately eight (Figure 4B) con-tinuous complexes per spread. The majority of cik1and Esposito 1974). To assay for this landmark, iso-

genic cik1 (DRS65) and CIK1 (DRS65 � pRS1) strains strains had a discontinuous, punctate Zip1p-stainingpattern (Figure 4C). Many of the Zip1p-staining cik1were constructed that contained arg4 heteroallelic muta-

tions (Bascom-Slack and Dawson 1998). These strains cells (22%; 14/64) displayed a bright staining focus thatis likely a polycomplex (Figure 4, B and C, arrowheads);allowed us to assay the frequency of recombination be-

tween the two mutations by selecting for Arg� proto- this was rare (4%; 2/51) in the CIK1 spreads that hadZip1p staining. Polycomplexes are often seen in meiotictrophs. The frequency of prototroph formation was mea-

sured by returning aliquots of the meiotic culture to mutants that arrest in prophase of meiosis I without thesynapsis of homologous chromosomes (Klapholz et al.vegetative growth on medium with and without argi-

nine. The cik1 strain exhibited a strong reduction in 1985; Sym and Roeder 1995). Similarly, kar3 mutantsfail to produce mature SC and experience polycom-prototroph formation, although the defect was not abso-

944 R. M. Q. Shanks, R. J. Kamieniecki and D. S. Dawson

Figure 3.—cik1 mutants arrest reversibly in meiosis and are Figure 4.—Synaptonemal complex formation in cik1 mu-defective in the process of interhomolog recombination. (A) tants. Cultures of a wild-type strain (A, DRS51 � pRS1) andcik1 mutants arrest reversibly in meiosis. The total number of a cik1 strain (B and C, DRS51) were harvested 12.5–14.5 hrcolony-forming units was determined by inducing isogenic postmeiotic induction. Their nuclei were spread and stainedwild-type (DRS65 � pRS1) and cik1 (DRS65) strains to un- with the DNA-specific dye DAPI (column 1) and anti-Zip1pdergo meiosis, taking aliquots, plating dilutions on YPD, and antibodies (column 2). Wild-type cultures often displayed fullcounting the number of colonies after 4 days of growth at SC formation (A). cik1 mutant cells rarely show partial SC30�. (B) cik1 mutants are defective in interhomolog recombi- formation (B) and often display the more punctate patternnation. Isogenic cultures of CIK1 (DRS65 � pRS1) and cik1 of Zip1p staining shown in C. Polycomplexes (arrowheads)(DRS65) strains, heterozygous for a pair of arg4 heteroalleles were observed in 3.9% of wild-type and 21.9% of cik1 mutant(arg4RV and arg4�42), were induced to undergo meiosis. Ali- nuclei.quots were removed over the course of 3 days; these wereplated onto complete medium without arginine to determinethe number of Arg� recombinants. Aliquots were also plated

spores disomic for the nondisjoined chromosome andonto nonselective medium to determine the total number ofcolony-forming units. The cik1 mutants exhibited 1.2% of the two inviable spores bereft of the chromosome or by losswild-type levels of recombination at 72 hr. The same cultures of a chromosome in either vegetative growth or meiosiswere used for both assays (A and B). These were single experi- prior to anaphase I. The resulting 2n � 1 cell undergo-ments that were representative of reproducible phenotypes.

ing meiosis would yield two spores with, and two sporeswithout, a copy of the lost chromosome yielding 2:2tetrads.plexes in �25% of pachytene nuclei (Bascom-Slack

and Dawson 1997). vik1 mutants do not experience elevated levels ofmeiosis I nondisjunction: Consistent with the hypothesisvik1 mutants exhibit reduced spore viability: To deter-

mine if Kar3p’s other well-known KAP has a role in that Vik1p has a role in meiosis I, its transcription isinduced in meiosis (Chu et al. 1998), and the proteinmeiosis we deleted VIK1 and assayed for meiotic defects.

Although vik1 mutants formed spores at wild-type levels, is evident in prophase I cells (see below and Figure 8).To determine whether the increased level of two-sporethey had a reduction in spore viability from 91% for

wild-type to 78% for isogenic vik1 mutants (Table 2). viable tetrads in vik1 mutants was from a defect in chro-mosome disjunction at the meiosis I division, we de-Tetrad dissection revealed that the two-spore viable class

of tetrad viability increased nearly 20-fold in vik1 mu- signed a strain for which spore products resulting frommeiosis I nondisjunction events could be selected. Thetants, and the zero-spore viable class had also increased

(Table 2). Elevated levels of two-spore and zero-spore vik1 mutant and control strains were modified so thattheir two copies of chromosome I were differentiallytetrad viability classes can be explained by increased

levels of meiosis I nondisjunction leading to two viable marked with prototrophic genes such that spores diso-

945Meiosis in cik1 Mutants

TABLE 2 and the loss of one copy of chromosome I was assayedby the generation of red sectors (see materials andSpore viability is reduced in vik1 mutantsmethods). vik1 mutant haploids had a 16-fold increasein the loss of chromosome I (Table 3). This is slightlySpore Tetrad viability classes (%)

viability higher than the 3-fold increase in chromosome lossGenotype (%) 4:0 3:1 2:2 1:3 0:4 monitored using a reporter chromosome III fragment

(Manning et al. 1999). The frequency of chromosomevik1� 78.1 41.5 33.3 17.6 3.8 3.8I loss in the vik1 diploid strain was at least 50-fold higherVIK1 90.8 72.5 26.6 0.9 2.8 0than in the isogenic wild type (Table 3). If the loss

Spores of isogenic strains (DRS142 vik1� and DRS143 VIK1)frequency we measured for chromosome I in diploidswere dissected on rich media (YPD). After 3 days of growthis similar to that of the other chromosomes, then the(30�) viability was assayed. N � 159 for DRS142 and N � 109

for DRS143. Sporulation frequencies were 18.6% for DRS142 overall decrease in spore viability due to mitotic chromo-and 17.4% for DRS143 when assayed by light microscopy some loss would be 1.5% (frequency of chromosome I(n 500). loss � 16 chromosomes � two inviable spores per tet-

rad), a value that would contribute to the loss of sporeviability in vik1 strains.mic for chromosome I could be isolated (see materials

Kar3p, Cik1p, and Vik1p are present during meiosisand methods for details). We found that the rate ofI and meiosis II: To determine the temporal and subcel-chromosome I disomy among vik1 spores was indistin-lular localization patterns of Kar3p, Cik1p, and Vik1p,guishable from that observed in wild-type strains (TableMyc-epitope-tagged versions of these genes were con-3). Similar results were found for chromosomes III andstructed (see materials and methods). The Myc-tagsV (data not shown). We conclude that the reduced sporeare C-terminal and the native promoters of the genes areviability of vik1 mutants is not because of a meiosis Iretained. The tagged genes were introduced throughnondisjunction defect.multiple backcrosses into the SK-1 background (KaneChromosome I is lost at high frequency during vegeta-and Roth 1974) to ensure synchronous and efficienttive growth in vik1 mutant diploids: To determine ifsporulation (Figures 5F, 6E, and 8E). Diploid cells, ho-lowered spore viability in vik1 mutants stems mainlymozygous for the tagged genes, efficiently formed tet-from a mitotic rather than a meiotic problem, we deter-rads with high viability (�92% viability for Cik1p-Mycmined the rates of chromosome I loss in vik1 mutantsand Vik1p-Myc and 86.5% viability for Kar3p-Myc; n �in the aforementioned diploid strains (DRS144 and22 tetrads). These strains, as well as a congenic strainDRS146) and in haploid strains disomic for chromo-with no Myc-tag, were sporulated, and samples weresome I (DRS144.1 and DRS144.2; see Table 1). Cellsharvested at timed intervals for microscopic analysis.were grown in medium that selected for both copies of

chromosome I. These were then plated on rich medium Cells were fixed and treated with either DAPI and antitu-

TABLE 3

vik1 chromosome behavior

Frequency of Frequency ofdisomic spore chromosome I

Genotype Ploidy formationa lossb

vik1� Diploid 2.3 � 10�3 � 1.7 � 10�3 4.6 � 10�4 � 1.1 � 10�4

VIK1 Diploid 2.1 � 10�3 � 1.5 � 10�3 8.4 � 10�6

vik1� 1n � 1 ND 5.6 � 10�4 � 6.7 � 10�5

VIK1 1n � 1 ND 3.4 � 10�5 � 5.1 � 10�5

ND, not determined.a Diploids heterozygous for centromeric markers on chromosome I (ADE1/ade1::ARG4) were sporulated

(DRS144 VIK1; DRS146 vik1�). The frequency of spore products disomic for chromosome I was determined(see materials and methods) in duplicate experiments. One standard deviation of error is shown with thefrequency of disomy.

b The frequency of mitotic chromosome I loss was determined in either diploids (DRS144 and DRS146) orhaploids disomic for chromosome I (DRS144.1 and DRS144.2) by assaying for red-sectored colony formationon YPD. We estimate from the observed relative frequency of one-half vs. smaller sectors (observed ratio 7:38;expected ratio 7:42) that chromosome loss in the first three cell generations of colony formation gave visiblesectors. Therefore, the loss frequency was determined to be the number of sectored colonies/total colonies �7 (the number of mitotic divisions in the first three generations of colony formation). Experiments were donein triplicate (diploids) or duplicate (haploids); the average with one standard deviation of error is shown. n 11,000 total colonies for diploids and n 8000 colonies for 1n � 1 haploids.

946 R. M. Q. Shanks, R. J. Kamieniecki and D. S. Dawson

bulin antibodies to monitor progress of the culturesthrough meiosis or with DAPI and anti-Myc antibodiesto evaluate localization of Kar3p, Cik1p, or Vik1p inmeiosis relative to the nucleus.

Kar3p: Kar3p-Myc staining was observed throughoutmeiosis I and meiosis II until after the breakdown ofthe spindles (Figure 5). In this experiment and in thosewith Cik1p and Vik1p, cells that had progressed beyondanaphase II exhibited a natural fluorescence that ob-scured the fluorescent signal used to detect Myc-stain-ing, making it difficult to draw conclusions about thepresence of Kar3p, Cik1p, or Vik1p in the late stagesof sporulation. Epitope-tagged versions of Kar3p havebeen shown previously to exhibit localization to thespindle pole bodies (SPBs) and to spindles of mitoticcells and, to a lesser extent, a pattern of general nuclearstaining or “nuclear patches” (Meluh and Rose 1990;Manning et al. 1999). All three types of localizationwere also seen in meiotic cells (Figure 5). In early mei-otic cells that are in prophase (t � 2 and 4 hr), Kar3ptypically localized to a single nuclear focus and brightpatches distributed throughout the nucleus (Figure5A). As cells progressed into anaphase I, Kar3p localizedin patches throughout the nucleus and to two foci (seeFigure 5B; t � 4 hr and beyond). These foci presumablycorrespond to the spindle poles, and in separate experi-ments (in S288c-background cells) we found that thesefoci colocalize with Tub4-GFP, a spindle pole body com-ponent (Marschall et al. 1996; data not shown). Thelocalization of Kar3p to both the poles and nuclearpatches continues through anaphase I and II (Figure5, C–E). Throughout meiosis, localization of Kar3p tothe region of the spindle between the poles was limited,observed in only 28% of Myc-staining nuclei (see thetop spindle in Figure 5D for an example).

Cik1p: Cik1p and tagged versions of Cik1p have beenshown to localize to the nucleus, spindle, and SPB ofvegetatively growing cells (Page et al. 1994; Manninget al. 1999). In mitotic and meiotic cells we observedthe bright nuclear patch staining pattern of Cik1p-Myclocalization similar to that of Kar3p-Myc (Figure 6, A–D). We also noted a SPB-like staining pattern, although

Figure 5.—Localization of Kar3p in meiosis. Diploid cellsthis was rare (at t � 6 hr, 6% of mononucleate cells(DRSK40) with a functional 13x-Myc-tagged version of Kar3pwith Myc staining exhibited a single focus while 8% had were induced to undergo a synchronous meiosis. Aliquots

two foci, compared to 54 and 10% for Kar3p-Myc at the were removed at timed intervals, and the cells were preparedsame time point n � 50). Throughout meiosis the SPB- for examination by fluorescent microscopy. Chromatin mor-

phology was evaluated by staining with DAPI (red). Kar3plike staining could be detected, although it was infre-was observed by indirect immunofluorescence (green). (A)quent and usually faint (indicated by arrowheads inProphase I. (B) Early anaphase I. (C) Late anaphase I. (D)Figure 6, A and B). We were able to confirm that these Early anaphase II. (E) Late anaphase II. (F) Progression of

Cik1p-Myc foci are at SPBs by demonstrating colocaliza- the culture through meiosis was monitored by quantifyingtion with Tub4p, a component of the SPB (Figure 7, mono-, bi-, and tetranucleate cells. The fraction of cells with

detectable Kar3p staining was determined for each time pointD–F). The two notable differences between Cik1p-Myc(open circles). Bars, 2 �m.and Kar3p-Myc staining occurred in the frequency and

intensity of SPB-like staining (see above) throughoutmeiosis and in binucleate cells that had just experiencedanaphase I, judging by the stretched, bilobed shape oftheir nuclei. In randomly chosen binucleate cells, 61%

947Meiosis in cik1 Mutants

Figure 7.—Kar3p localization in prophase nuclear spreadsand colocalization of Cik1p and the SPB component, Tub4p,in whole meiotic cells. In A, B, and C, meiotic nuclei werespread and prepared for fluorescent microscopy. (A and B)Nuclei (DJ1 � pD168) were stained with antisera for Zip1p(an SC component, shown in red) and Kar3p-GFP (shown ingreen). Prophase I cells representative of two classes of Kar3pstaining patterns are shown (A, a single focus 96%; B, multiplefoci 4%; n � 50). (C) Nuclei (DRS151) were stained withDAPI (chromatin stain, blue) and antisera against Ndc10p-GFP (a kinetochore protein, green) and Kar3p-Myc (red).D–F are a montage of whole meiotic cells from the same field.Meiotic cells (DRS106) were prepared for immunofluorescentmicroscopy. (D) Cik1p-Myc staining (green); (E) Tub4p-GFPfoci (red); and (F) overlay showing the colocalization of Cik1pand Tub4p in nuclei with bipolar spindles. Bar, 5 �m.

The shared pattern of Kar3p and Cik1p staining priorto anaphase, followed by the detection of Cik1p but not

Figure 6.—Localization of Cik1p in meiosis. Diploid cells Kar3p at the midzone, is consistent with a dissociation(DRSK38) with a functional 13x-Myc-tagged version of Cik1p of some or all of the Kar3p-Cik1p complexes as cellswere induced to undergo a synchronous meiosis. Aliquots

enter anaphase I. Alternatively, it may be that Kar3p-were removed at timed intervals, and the cells were preparedMyc, but not Cik1p-Myc, is difficult to detect at thefor examination by fluorescent microscopy. Chromatin mor-

phology was evaluated by staining with DAPI (red). Cik1p midzone for any number of technical reasons.was observed by indirect immunofluorescence (green). (A) The nuclear patching phenotype exhibited by Kar3pProphase I, arrowhead points to SPB-like focus. (B) Early and Cik1p is consistent with the association of theseanaphase I, arrowheads point to SPB-like foci. (C) Late ana-

proteins with chromosomes. To test this hypothesis wephase I, the arrow indicates apparent midzone localization.looked for association of Kar3p with isolated chromo-(D) Anaphase II. (E) Progression of the culture through meio-

sis was monitored by quantifying mono-, bi-, and tetranucleate somes in S288c-background cells. Nuclear spreads werecells. The fraction of cells with detectable Cik1p staining was performed on cells in meiotic prophase to determinedetermined for each time point (open circles). Bars, 5 �m. whether Kar3p foci were found in association with con-

densed pachytene chromosomes. The chromosomespreads were stained with anti-Zip1p and anti-GFP anti-(20/33) had clear Cik1p-Myc staining remaining be-bodies to detect the synaptonemal complex and thetween the two nuclei (Figure 6C, arrow), compared toGFP-tagged versions of Kar3p expressed in the cells. Inonly 8% (3/36) for Kar3p-Myc staining. This midzonethe majority of spread nuclei we saw a single focus (92%,staining pattern has been seen for a number of proteinsn � 46) at the edge of the chromosome mass; thisfrom diverse organisms that are involved in chromo-focus sometimes looked like two closely associated focisome and mitotic spindle stability or function, for exam-(Figure 7A), as would be expected for duplicated butple, cut7 of Schizosaccharomyces pombe (Hagan and Yanag-unseparated SPBs. A fraction (8%, n � 4) of the spreadida 1992), Ase1p and Slk19p of S. cerevisiae (Pellman etnuclei with Kar3p-GFP staining displayed a single brightal. 1995; Zeng et al. 1999), Mast of Drosophila (Lemos et al.staining focus and several (3–20) smaller foci (Figure2000), and Plk1 of humans (Arnaud et al. 1998; Wianny et7B). Similar staining patterns were seen for Cik1p-Mycal. 1998). Some of these proteins localize to the microtu-in spread prophase I cells (data not shown). The factbule-organizing centers before they localize to the spin-

dle midzone during anaphase of mitosis and/or meiosis. that Kar3p and Cik1p nuclear patches could be observed

948 R. M. Q. Shanks, R. J. Kamieniecki and D. S. Dawson

in the majority of whole prophase cells, while they wererarely observed associated with spread chromosomes,suggests that either the nuclear patches are not associ-ated with chromatin in prophase I of meiosis or thatthe methods used for chromosome spreading disruptthe association of Kar3p and Cik1p with chromatin butnot with the SPBs. To determine if the rare Kar3p (orCik1p) foci that were found associated with spread chro-mosomes were localized to the kinetochores, we com-pared the localization of Kar3p with Ndc10p, a knownkinetochore protein (Goh and Kilmartin 1993; Jianget al. 1993) in nuclear spreads (DRS151). We found thatneither Kar3p-Myc nor Cik1p-Myc (not shown) coloca-lized with Ndc10p-GFP during prophase I (Figure 7C;n 70).

Vik1p: Vik1p-Myc had a staining pattern that was dis-tinct from that of Kar3p and Cik1p in that the nuclearpatching phenotype was rarely seen (at t � 6 hr, 6% ofcells had nuclear patches of Vik1p, compared to 36 and85% for Kar3p and Cik1p, respectively; n � 50); thenuclear patching phenotype was not observed in a no-epitope control strain. Instead, depending on the stageof meiotic progression, one to four foci were seen inmeiotic cells (Figure 8, A–D). Throughout meiosis, Vik1pappears to be primarily associated with the SPBs, as hasbeen shown in vegetative cells (Manning et al. 1999).

cik1, vik1, and kar3 mutants differ in the length oftheir meiotic monopolar microtubule arrays: Cik1p hasbeen reported to affect the length and number of micro-tubules in mitotic cells (Page and Snyder 1992). cik1mutants display pronounced nuclear microtubule arraysin vegetative and �-factor arrested cells, but shorter bi-polar spindles (Page and Snyder 1992; Manning et al.1999). Similar phenotypes have been reported for kar3mutants (Meluh and Rose 1990; Saunders et al. 1997a);however, vik1 mutants were reported to have microtu-bule arrays and bipolar spindles indistinguishable fromthose of wild-type strains (Manning et al. 1999). In meio- Figure 8.—Localization of Vik1p in meiosis. Diploid cellssis, kar3 mutants have been shown to arrest with promi- (DRSK42) with a functional 13x-Myc-tagged version of Vik1p

were induced to undergo a synchronous meiosis. Aliquotsnent nuclear microtubule arrays (Bascom-Slack andwere removed at timed intervals, and the cells were preparedDawson 1997). To see if the meiotic phenotypes of cik1for examination by fluorescent microscopy. Chromatin mor-and vik1 were similar to the reported mitotic pheno-phology was evaluated by staining with DAPI (red). Vik1p was

types, we assayed the length of nuclear monopolar tu- observed by indirect immunofluorescence (green). (A) Twobulin arrays in meiotic time courses. To assay the differ- cells: The one at the bottom left is in prophase I and the one

at the top right is in early anaphase I. (B) Late anaphase I.ences among strains, we measured the length of the(C) Early anaphase II. (D) Late anaphase II. (E) Progressionlongest nuclear microtubule bundle in monopolar tu-of the culture through meiosis was monitored by quantifyingbulin arrays of randomly chosen wild-type, cik1, vik1,mono-, bi-, and tetranucleate cells. The fraction of cells with

and kar3 meiotic diploids. detectable Vik1p staining was determined for each time pointThe monopolar arrays in wild-type cells showed a (open circles). Bars, 2 �m.

distribution of lengths at t � 0 (when cells were trans-ferred into sporulation medium) with an average lengthof 1.21 �m. As meiosis progressed, the average length meiosis. vik1 mutants have vegetative and meiotic mono-

polar arrays with lengths that are indistinguishable fromof the arrays increased (Figure 9). Note that most wild-type cells that will complete meiosis have formed bipolar wild type (Figure 9). cik1 mutant microtubule arrays are

similar in length to wild-type arrays in vegetative cellsspindles before the 26-hr time point. Thus, the cellsthat remain with monopolar arrays at 26 hr do not grown in YPAcetate (t � 0); however, in early (t � 6

hr) and middle (t � 12 hr) meiosis cik1 mutants haverepresent cells undergoing normal passage through

949Meiosis in cik1 Mutants

defects in the levels of meiotic recombination and theformation of mature synaptonemal complexes. We alsoshow that Cik1p and Kar3p have similar but distinctlocalization patterns in meiotic cells.

We found that, although Vik1p is present in meiosis,vik1 mutants have no dramatic meiotic defects. We dem-onstrated that vik1 mutants exhibit reduced spore viabil-ity, which is due to neither meiosis I nor meiosis IInondisjunction (data not shown), but is caused, at leastin part, by the loss of chromosomes during vegetativecell divisions. It is also possible that the elevated levelsof two-spore viable tetrads that we observed in the vik1mutants are due to loss of chromosomes in meiosis prior

Figure 9.—Monopolar tubulin array length in kar3, cik1,to anaphase I, and our assays would not detect this typeand vik1 mutants. The length of the longest microtubule bun-of error. The loss of chromosomes could occur by adle in monopolar tubulin arrays was measured microscopi-

cally. Wild-type (DRS144), vik1 (DRS146), cik1 (DRS52), and simple failure of chromosomes to attach to the spindle.kar3 (DJ1) strains were induced to undergo meiosis. Aliquots The Ncd motor protein of Drosophila is homologouswere removed at the indicated times and processed for tubulin to Kar3p and is required for organizing meiotic spindlesimmunofluorescence. Mononucleate cells were randomly cho-

in Drosophila oocytes. In ncd mutants chromosomessen and their microtubule bundles were measured. The histo-can be lost because of their failure to be captured bygrams display the average monopolar array length at each

time point. All strains had a range of microtubule length, and the disorganized microtubules (Hatsumi and Endowbars showing 1 SD reflect this distribution. (n � 25 arrays for 1992; Endow and Komma 1996). Although the assemblyT � 0; n � 30 for meiotic samples.) of meiotic spindles in yeast and Drosophila oocytes is

drastically different, the Vik1p-Kar3p complex may havean organizational function that parallels that of Ncd,

arrays that are 20% longer than wild type and this differ- and vik1 defects may result in mild spindle disorganiza-ence increases to 40% at t � 26 hr (Figure 9). kar3 tion and meiotic chromosome loss that yields elevatedmutants show highly variable and elongated prophase levels of two-spore viable tetrads.arrays. Vegetative kar3 mutant nuclear microtubule What are Cik1p and Kar3p doing in meiosis I? Thearrays are, on average, twice as long as those of wild- major events of prophase I include formation of double-type cells and the length of their long microtubule arrays strand breaks, interhomolog recombination, synapsis ofis maintained throughout meiosis (Figure 9). The exag- homologous chromosomes, and the establishment of agerated length of kar3 mutant arrays is not an artifact meiosis I spindle. Cik1p and Kar3p may be directly orof prolonged arrest in prophase I, because wild-type indirectly necessary for these events. Many organisms,cells at 26 hr with prophase I microtubule arrays have including S. cerevisiae, reorganize the positions of thepresumably arrested in meiosis I as well and do not chromosomes within the cell during prophase I (Dern-exhibit the greatly elongated microtubule phenotype of berg et al. 1995; Zickler and Kleckner 1998). In thiskar3 mutants. Over time, the length of the cik1 monopo- process, telomeres become clustered with the chromo-lar arrays approaches that of the kar3 mutants, but even some arms radiating outward in an organization re-at 26 hr the kar3 arrays are longer. Unlike the other ferred to as the bouquet. Bouquet formation may facili-strains, kar3 mutants frequently exhibited unusual spin- tate homologous recombination, as agents that disruptdle morphologies, as has been reported previously (Bas- bouquet formation have been demonstrated to disruptcom-Slack and Dawson 1997). These include bushy chiasma formation (Loidl 1990). In budding yeast,nuclear monopolar arrays, tandem (side-by-side) nu- both genetic and cell biological methods have beenclear monopolar arrays, cytoplasmic microtubules that used to demonstrate clustering of telomeres in prophasespan the diameter of the cell, and cytoplasmic accumula- I (Goldman and Lichten 1996; Hayashi et al. 1998;tions of microtubules into bright-staining foci adjacent Trelles-Sticken et al. 1999). Failure in this prealign-to the SPB. ment is one way to explain how defects in a motor

protein might result in a defect in interhomolog recom-bination. Alternatively, the recombination defects of

DISCUSSIONthese mutants might be an indirect consequence of anarrest triggered by defects in spindle assembly or func-We have characterized the meiotic phenotypes of cik1

mutants and found them consistent with the model that tion.The experiments presented here are consistent withCik1p is critical for the function of Kar3p in prophase of

meiosis I. This conclusion is based upon several meiotic the model that in meiosis, as in mitosis, Kar3p performssome roles independently of Cik1p. kar3 mutants havephenotypes shared by both cik1 and kar3 mutants. Cells

mutated for these genes arrest in prophase I with similar more highly elongated nuclear microtubule arrays in

950 R. M. Q. Shanks, R. J. Kamieniecki and D. S. Dawson

Endow, S. A., S. J. Kang, L. L. Satterwhite, M. D. Rose, V. P. Skeenmeiosis than do cik1 mutants. Cik1p is deposited at theet al., 1994 Yeast Kar3 is a minus-end microtubule motor protein

spindle midzone at anaphase I while this seems not to that destabilizes microtubules preferentially at the minus ends.EMBO J. 13: 2708–2713.be true for the majority of Kar3p. kar3 mutants have a

Esposito, M. S., and R. E. Esposito, 1974 Genetic recombinationmore severe prophase I arrest than do cik1 mutants.and commitment to meiosis in Saccharomyces. Proc. Natl. Acad.

cik1 mutants produce rare viable haploids and display Sci. USA 71: 3172–3176.Goh, P. Y., and J. V. Kilmartin, 1993 NDC10 : a gene involved inpostprophase spindle morphologies at detectable levels,

chromosome segregation in Saccharomyces cerevisiae. J. Cell Biol.while kar3 mutants were observed to do neither. In an-121: 503–512.

other study (Kurihara et al. 1996), kar3 mutants pro- Goldman, A. S. H., and M. Lichten, 1996 The efficiency of meioticrecombination between dispersed sequences in Saccharomyces cere-duced no spores, while cik1 mutants produced low levelsvisiae depends on their chromosomal location. Genetics 144:of asci that contained two spores. We attribute the differ-43–55.

ent cik1 sporulation phenotypes described between Guldener, U., S. Heck, T. Fielder, J. Beinhauer and J. H. Hege-mann, 1996 A new efficient gene disruption cassette for re-these two studies to either the well-documented differ-peated use in budding yeast. Nucleic Acids Res. 25: 2519–2524.ences in meiotic mutant phenotypes in different strains

Hagan, I., and M. Yanagida, 1992 Kinesin-related cut7 protein asso-or to differences in sporulation regimens. In both strains ciates with mitotic and meiotic spindles in fission yeast. Nature

356: 74–76.the kar3 phenotype is more severe than the cik1 pheno-Hardwick, K. G., R. Li, C. Mistrot, R. H. Chen, P. Dann et al.,type.

1999 Lesions in many different spindle components activate theIn summary, Vik1p appears to play a minor or redun- spindle checkpoint in the budding yeast Saccharomyces cerevisiae.

Genetics 152: 509–518.dant role in meiosis. The loss in spore viability of vik1Hatsumi, M., and S. A. Endow, 1992 Mutants of the microtubulemutants can partially be attributed to chromosome loss

motor protein, nonclaret disjunctional, affect spindle structurein mitotic cell divisions that precede meiosis; meiotic and chromosome movement in meiosis and mitosis. J. Cell Sci.

101: 547–559.chromosome loss may also contribute to this phenotype.Hayashi, A., H. Ogawa, K. Kohno, S. M. Gasser and Y. Hiraoka,The severe defects in recombination and synapsis exhib-

1998 Meiotic behaviours of chromosomes and microtubules inited by both cik1 and kar3 mutants are consistent with budding yeast: relocalization of centromeres and telomeres dur-

ing meiotic prophase. Genes Cells 3: 587–601.the hypothesis that Cik1p acts as a KAP to Kar3p in itsHerman, P. K., and J. Rine, 1997 Yeast spore germination: a require-essential role in prophase I of meiosis. The overall slightly

ment for Ras protein activity during re-entry into the cell cycle.milder meiotic phenotypes of cik1 mutants suggest that EMBO J. 16: 6171–6181.

Jiang, W., J. Lechner and J. Carbon, 1993 Isolation and character-in meiosis, as in mitosis, Kar3p can perform some func-ization of a gene (CBF2) specifying a protein component of thetions independently of Cik1p.budding yeast kinetochore. J. Cell Biol. 121: 513–519.

Kamieniecki, R. J., R. M. Shanks and D. S. Dawson, 2000 Slk19pThe authors are grateful to Carol Bascom-Slack, Benedict J. Kemp,is necessary to prevent separation of sister chromatids in meiosisand Stu Powers for insight, discussions of the work, and critical readingI. Curr. Biol. 10: 1182–1190.of the manuscript. We thank Becky Boumil for Zip1p antisera and

Kane, S. M., and R. Roth, 1974 Carbohydrate metabolism duringTim Stearns and Dan Burke for yeast strains. This work was supportedascospore development in yeast. J. Bacteriol. 118: 8–14.by research grant no. 1-Fy99-579 from the March of Dimes Birth

Klapholz, S., C. S. Waddell and R. E. Esposito, 1985 The roleDefect Foundation to D.S.D. and a grant from the Department of of the SPO11 gene in meiotic recombination in yeast. GeneticsDefense (Breast Cancer Research grant no. BC-980 652) to R.M.Q.S. 110: 187–216.

Kurihara, L. J., B. G. Stewart, A. E. Gammie and M. D. Rose, 1996Kar4p, a karyogamy-specific component of the yeast pheromoneresponse pathway. Mol. Cell. Biol. 16: 3990–4002.

Lemos, C. L., P. Sampaio, H. Maiato, M. Costa, L. V. Omel’yanchukLITERATURE CITEDet al., 2000 Mast, a conserved microtubule-associated protein

Arnaud, L., J. Pines and E. A. Nigg, 1998 GFP tagging reveals required for bipolar mitotic spindle organization. EMBO J. 19:human Polo-like kinase 1 at the kinetochore/centromere region 3668–3682.of mitotic chromosomes. Chromosoma 107: 424–429. Loidl, J., 1990 The initiation of meiotic chromosome pairing: the

Bascom-Slack, C. A., and D. S. Dawson, 1997 The motor protein, cytological view. Genome 33: 759–778.Kar3, is essential for meiosis I. J. Cell Biol. 139: 459–467. Longtine, M. S., A. McKenzie, D. J. Demarini, N. G. Shah, A. Wach

Bascom-Slack, C. A., and D. Dawson, 1998 A physical assay for et al., 1998 Additional modules for versatile and economicaldetection of early meiotic recombination intermediates in Sac- PCR-based gene deletion and modification in Saccharomyces cerevis-charomyces cerevisiae. Mol. Gen. Genet. 258: 512–520. iae. Yeast 14: 953–961.

Chu, S., J. DeRisi, M. Eisen, J. Mulholland, D. Botstein et al., 1998 Manning, B. D., J. G. Barrett, J. A. Wallace, H. Granok andThe transcriptional program of sporulation in budding yeast. M. Snyder, 1999 Differential regulation of the Kar3p kinesin-Science 282: 699–705. related protein by two associated proteins, Cik1p and Vik1p. J.

Cottingham, F. R., L. Gheber, D. L. Miller and M. A. Hoyt, 1999 Cell Biol. 144: 1219–1233.Novel roles for Saccharomyces cerevisiae mitotic spindle motors. J. Marschall, L. G., R. L. Jeng, J. Mulholland and T. Stearns, 1996Cell Biol. 147: 335–350. Analysis of Tub4p, a yeast gamma-tubulin-like protein: implica-

Dernberg, A. F., J. W. Sedat, W. Z. Cande and H. W. Bass, 1995 tions for microtubule-organizing center function. J. Cell Biol.Cytology of telomeres, pp. 295–337 in Telomeres, edited by E. H. 134: 443–454.Blackburn and C. W. Grieder. Cold Spring Harbor Laboratory Meluh, P. B., and M. D. Rose, 1990 KAR3, a kinesin-related genePress, Cold Spring Harbor, NY. required for yeast nuclear fusion. Cell 60: 1029–1041.

Elledge, S. J., and R. W. Davis, 1988 A family of versatile centro- Nicolas, A., D. Treco, N. P. Schultes and J. W. Szostak, 1989meric vectors designed for use in the sectoring-shuffle mutagene- An initiation site for meiotic gene conversion in the yeast Saccharo-sis assay in Saccharomyces cerevisiae. Gene 70: 303–312. myces cerevisiae. Nature 338: 35–39.

Endow, S. A., and D. J. Komma, 1996 Centrosome and spindle Page, B. D., and M. Snyder, 1992 CIK1: a developmentally regulatedfunction of the Drosophila Ncd microtubule motor visualized in spindle pole body-associated protein important for microtubulelive embryos using Ncd-GFP fusion proteins. J. Cell Sci. 109: functions in Saccharomyces cerevisiae. Genes Dev. 6: 1414–1429.

Page, B. D., L. L. Satterwhite, M. D. Rose and M. Snyder, 19942429–2442.

951Meiosis in cik1 Mutants

Localization of the Kar3 kinesin heavy chain-related protein re- Southern, E. M., 1975 Detection of specific sequences among DNAfragments separated by gel electrophoresis. J. Mol. Biol. 98: 503–quires the Cik1 interacting protein. J. Cell Biol. 124: 507–519.517.Pellman, D., M. Bagget, Y. H. Tu, G. R. Fink and H. Tu, 1995 Two

Sym, M., and G. S. Roeder, 1995 Zip1-induced changes in synaptone-microtubule-associated proteins required for anaphase spindlemal complex structure and polycomplex assembly. J. Cell Biol.movement in Saccharomyces cerevisiae. J. Cell Biol. 130: 1373–1385.128: 455–466.Roof, D. M., P. B. Meluh and M. D. Rose, 1991 Multiple kinesin-

Sym, M., J. Engebrecht and G. S. Roeder, 1993 ZIP1 is a synaptone-related proteins in yeast mitosis. Cold Spring Harbor Symp.mal complex protein required for meiotic chromosome synapsis.Quant. Biol. 56: 693–703.Cell 72: 365–378.Russell, D. W., R. Jensen, M. J. Zoller, J. Burke, B. Errede et

Tirnauer, J. S., E. O’Toole, L. Berrueta, B. E. Bierer and D.al., 1986 Structure of the Saccharomyces cerevisiae HO gene andPellman, 1999 Yeast Bim1p promotes the G1-specific dynamicsanalysis of its upstream regulatory region. Mol. Cell. Biol. 6:of microtubules. J. Cell Biol. 145: 993–1007.4281–4294.

Trelles-Sticken, E., J. Loidl and H. Scherthan, 1999 BouquetSaunders, W., D. Hornack, V. Lengyel and C. Deng, 1997a Theformation in budding yeast: initiation of recombination is notSaccharomyces cerevisiae kinesin-related motor Kar3p acts at preana-required for meiotic telomere clustering. J. Cell Sci. 112: 651–658.phase spindle poles to limit the number and length of cytoplasmic

Wianny, F., A. Tavares, M. J. Evans, D. M. Glover and M. Zernicka-microtubules. J. Cell Biol. 137: 417–431.Goetz, 1998 Mouse polo-like kinase 1 associates with the acen-Saunders, W., V. Lengyel and M. A. Hoyt, 1997b Mitotic spindletriolar spindle poles, meiotic chromosomes and spindle midzonefunction in Saccharomyces cerevisiae requires a balance betweenduring oocyte maturation. Chromosoma 107: 430–439.different types of kinesin-related motors. Mol. Biol. Cell 8: 1025– Zeng, X., and W. S. Saunders, 2000 The Saccharomyces cerevisiae1033. centromere protein Slk19p is required for two successive divisions

Sherman, F., and H. Roman, 1963 Evidence for two types of allelic during meiosis. Genetics 155: 577–587.recombination in yeast. Genetics 48: 255–261. Zeng, X., J. A. Kahana, P. A. Silver, M. K. Morphew, J. R. McIntosh

Siemering, K. R., R. Golbik, R. Sever and J. Haseloff, 1996 Muta- et al., 1999 Slk19p is a centromere protein that functions totions that suppress the thermosensitivity of green fluorescent stabilize mitotic spindles. J. Cell Biol. 146: 415–425.protein. Curr. Biol. 6: 1653–1663. Zickler, D., and N. Kleckner, 1998 The leptotene-zygotene transi-

Sikorski, R. S., and P. Heiter, 1989 A system of shuttle and yeast tion of meiosis. Annu. Rev. Genet. 32: 619–697.host strains designed for efficient manipulation of DNA in Saccha-romyces cerevisiae. Genetics 122: 19–27. Communicating editor: M. D. Rose