Copyright 0 1990 by the Genetics Society of America

Mitotic Transmission of Artificial Chromosomes in cdc Mutants of the Yeast, Saccharomyces cerevisiae

Robert E. EileenHogant andDouglas Koshlandt

*Department of Population Dynamics, Divisionof Reproductive Biology, The Johns Hopkins UniversitySchool of Hygiene and Public Health, Baltimore, Maryland 21205, and ?Departmentof Embryology, The Carnegie Institution of Washington, Baltimore, Maryland 21210 Manuscript received January1 1, 1990 Accepted for publicationMay 2, 1990

ABSTRACT In the yeast, Saccharomyces cerevisiae, cell division cycle (CDC) genes have been identified whose products are required for the executionof different steps in the cell cycle. In this study, the fidelity of transmission of a 14-kb circular minichromosome anda 155-kb linear chromosome fragmentwas examined in cell divisions where specific CDC products were temporarily inactivated with either inhibitors, or temperature sensitive mutations in the appropriate CDC gene. All of the cdc mutants previously shown to induce loss of endogenous linear chromosomes also induced loss of a circular minichromosome anda large linear chromosome fragmentin our study (either 1:0 or 2:O loss events). Therefore, the efficient transmission of these artificial chromosomes depends upon the same trans factors that are required for the efficient transmission of endogenous chromosomes. In a subset of cdc mutants (cdc6, cdc7 and cdcl6), the rate of minichromosome loss was significantly greater than the rate ofloss of the linear chromosome fragment, suggesting that a structural feature of the minichromosome (nucleotidecontent, length or topology) makesthe minichromosome hypersensitive to the level of function of these CDC gene products. In another subset of cdc mutants (cdc7 and cdcl7), the relativerate of 1:O events to2:O events differed for the minichromosome and chromosome fragment, suggestingthat the typeof chromosomeloss event observedin these mutantswas dependent upon chromosome structure. Finally, we show that 2:O events for the minichromosomecan occur by both a RAD52 dependent and RAD52 independent mechanism. These results are discussed in the context of the molecular functionsof the CDC products.

N the yeast, Saccharomyces cermisiae, the construc- tide content, size or topology) which ensures a high I tion and genetic analyses of small circular mini- fidelity of transmission during mitotic cell divisions. chromosomes have been instrumental in understand- In the absence of this structural feature, minichro- ing the structural features of chromosomes that are mosomes may be acted upon by the same replication essential for their properreplication and segregation. and segregation machinerythat acts upon endogenous For example, the analysis of minichromosomes has chromosomes but may bepoor substrates for this led to identification and characterizationof cis se- machinery atone or more steps in the cell cycle. quences required for theinitiation of DNA replication Alternatively, the absence of this structural featurein (HSIAOand CARBON1979; STINCHCOMB,STRUHL and minichromosomes may cause them to be transmitted DAVIS 1979)and for the function of centromeres by secondary replication or segregation processes that (CLARKE and CARBON, STINCHCOMB,1980; MANN and are inherently less faithful. DAVIS1982). Circular minichromosomes that contain An understanding of the molecular mechanism of an origin of replication (ARS)and a centromere (CEN) replication and segregation will also require the iden- are transmittedproperly in 99% of cell divisions tification and characterization of the trans acting fac- (CLARKEand CARBON, 1980; FITZGERALD-HAYESet torsthat comprise the replication and segregation al. 1982). However, their rate of loss per cell division machinery. In S. cerevisiae, temperature-sensitive mu- is still several orders of magnitude greater than en- tants have been identified which arrest at different dogenous (ESPOSITOet al. 1982;HARTWELL et al. stages of the cell cycle when they are shifted to the 1982) or large artificial linear chromosomes (MUR- nonpermissive temperature (PRINGLE and HARTWELL RAY, SCHULTES and SZOSTAK 1986; HIETER et al. 1981). A subset of these cell division cycle (cdc) mu- 1985) suggesting that minichromosomes lack a struc- tants identify genes whose products are required for tural feature(s) of endogenous chromosomes (nucleo- the execution of discrete steps in GI (CDC28, CDCI,

' To whom reprint requests should be addressed at the CarnegieInstitu- CDC7), S (CDC2,CDC6, CDC9, CDC17), G2-M tion. (CDCI3, CDCI6, CDC20, CDC23) and late M phase

Genetics 125: 763-774 (August, 1990) 764 R. E. Palmer, E. Hogan and D. Koshland

(CDC14, CDCl5)of the cell cycle(PRINGLE and HART- effect of cdc mutations on endogenous chromosome WELL 1981). Given their cell cycle phenotype, these transmission (HARTWELLand SMITH1985). CDC genes are likely candidates for encoding trans factors required for proper chromosometransmission. MATERIALS AND METHODS In fact, when the cdc mutants were incubated at a semi-permissive temperature, conditions that partially Strains: All of the cdc mutations used in this study were inactivate the CDC gene product, many of these mu- obtained by back crossing at least five times the original cdc mutations isolated in A364A to strains congenic with the tants lose endogenous chromosome at elevated rates original A364A parent. All other markers were obtained by demonstrating that these CDC products are involved back crossing the mutation from anotherSaccharomyces cere- in ensuring proper chromosome transmission (HART- uisiae strain into theA364A congenic background by at least WELL and SMITH1985). ten back crosses. The haploid strains usedin this study The molecular functions of some of the CDC gene (Table 1) were constructed by standard genetic methods (SHERMAN,FINK and HICKS 1986) and thentransformed products have been identified by cytological and bio- (ITO et al. 1983)with the minichromosome, pDK243 (KOSH- chemical analysis of arrested cells, while the activity LAND, KENT and HARTWELL1985). The homozygous cdc of other CDC products have been inferred from amino diploid strains containing the chromosome fragment (Table acid similarities with other better characterized pro- 1) were constructed as follows. First, the chromosome frag- teins. The CDC28 and CDC7 genes encode protein ment,CF352 (Figure l), was constructed in the haploid strain, 45 13-12 1, by the chromosome fragmentation pro- kinases (LORINCZand REED 1984; PATTERSONet al. tocol (GERRING,CONNELLY and HIETER1989; VOLLRATHet 1986); the former shows extensive similarity to the al. 1988). This strain was then crossed with a set of congenic kinase subunit of maturationpromoting factor, an haploid cdc strains (Table l), anda set of congenic haploid essential regulator of the G2 to M transition in many cdc strains were obtained that contained the chromosome organisms (ARIONet al. 1988; BEACH,DURKACZ and fragment. These strains were then mated with the haploid cdc strains listed in Table 1 to generate homozygous cdc NURSE1982; GAUTIERet al. 1988). The CDC4 prod- diploids with 1 copy of the chromosome fragment. uct shares homology tothe @-adrenergicreceptor Growth conditions of exponentially growing cells and (YOCHEMand BYERS1987). CDC2 and CDCl7 encode temporarily arrested cells: The cdc mutant strains (Table DNA polymerases (SITNEY,BUDD and CAMPBELL 1) were grown at 23"to saturation in YPD (SHERMAN, FINK and HICKS1986) supplemented with adenine. These cul- 1989; BOULETet al. 1989; CARSON1987), while CDC9 tures were used to inoculate YPD. When cells reached mid- encodes DNA ligase (BARKERand JOHNSTON1983). log phase (5 X lo6 cells/ml), an aliquot was diluted with However, the function of many of the CDC gene water approximately 1000-fold and spread onto YPD plates. products remains to be elucidated. The remaining cells were temporarily arrested by shifting them tothe nonpermissive condition for 3 hr. The cdc In this paper, we examined the effect of temporarily mutants were temporarily arrested by incubating them at reducing the functionof different CDC gene products 36" for 3 hr. Wild-type cells were temporarily arrested by on the mitotic transmission of two artificial chromo- incubating them in the presence of nocodazole or hydrox- somes, a 14-kb circular minichromosome and a 155- yurea (a final concentration of 0.1 mM and 0.1 M, respec- kb linear chromosome fragment. Both an increase or tively) for3 hr at 23" and 36". The time spentat the nonpermissive condition was sufficient to inactivate most of decrease in the ploidy of these test chromosomes could the CDC gene product in the majority of cells as evidenced be detected in individual cell divisions through the by the fact that greater than 90% of the cells were arrested use of a visual assay for chromosome ploidy (KOSH- at the appropriate stage of the cell cycle. After 3 hr at the LAND, KENT and HARTWELL1985; KOSHLAND and nonpermissive condition, arrested cells were diluted with water and plated on YPD plates at 23" where they reentered HIETER1987). In this study, we identified mutations the cell cycleand formeda colony. The dilution and plating in cdc genes that reduce the fidelity of minichromo- were done rapidly to ensure thatarrested cells did not some transmission as evidenced by an increase in the complete division before they were plated. rate of aberrant celldivisions where ploidy of the Half-sector analysis of minichromosome and chromo- some fragment ploidy: The ploidy of the minichromosome minichromosome changed. For each cdc mutant, we and chromosome fragment were monitored in the wild-type determined the types of changes in minichromosome and cdc strains by a visual assay for chromosome ploidy in ploidy that were observed and their rate of occur- which cells that contained 0, 1 or 2 copies of these artificial rence. Finally, we compared the fidelity of transmis- chromosomes were white, pink and red,respectively (KOSH- sion of the circular minichromosome and large linear LAND, KENT and HARTWELL1985). The change in the ploidy of the chromosome fragment was monitored in dip- chromosomefragment in a subset of cdc mutants. loids instead of haploid strains because the distinction be- These experiments provide insights into the role of tween the red andpink color of colonies was easier to make CDC products in the transmission of minichromo- in diploids strains. Five to seven days after exponentially somes, the cause of the high basal instability of mini- growing cells and temporarily arrested cells were plated, the chromosomes, and the ability of different CDC prod- colony size and color were sufficiently developed to deter- mine the ploidy of the artificial chromosomes in the colony. ucts to transmit chromosomes with different struc- Each colony was scored as white, pink, red, oras pink:white, ture. These insights are discussed in the context of red:white, or pink:red half-sectored. The different type of the molecular functions of the CDC products and the half-sectored colonies reflected 1:0, 2:0,and 2:l events Transmission of Yeast Chromosomes 765

TABLE 1

Strains and genotypes used in this study

Strains Genotype BP5063Mata leu21Mataleu2ade21ade2 ade3lade3 HIS3lhis3 his7/His7 cdc6-llcdch-1 canllcanl sap3lsap3 [cf352 chrlll (D8Bleft) CEN6 LEU2 ade3-2p] BP.55Mata 19 leu21Mata leu2ade2/ade2 ade3lade3 his71his7 URAllural cdc7-llcdc7-1 canllcanlsap3/sap3 [cf352 chrlll (D8Bleft) CEN6 LEU2 ade3-2p] EH520.5 Mataleu21Mata leu2ade21ade2 ade31ade3 his71His7 cdcl4-l/cdcl4-1canllcanl sap3/sap3 [cf352 chrll (D8Bleft) CEN6 LEU2 ade3-2p] BP50Mata10 leu2lMata leu2 ade21ade2 ade3/ade3 HIS3/his3 his71HIS7 cdcl6-l/cdcl6-1canllcanl sap3/sap3 [cf352 chrlll (D8Bleft) CEN6 LEU2 ade3-2p] BPS524Mataleu21Mata leu2ade21ade2 ade3/ade3 his7/his7 URAIluraIcdcl7-l/cdcl7-1 canllcanl sap31 sap3 [cf352 chrlll (D8Bleft)CEN6 LEU2 ade3-2p] BP5061 MATa leu2IMATaleu2ade21ade2 ade31ade3 his3/his3 his7/HIS7cdc20-ll cdc20-canl/canl sap3lsap3 [cf352 chrlll (D8Bleft) CEN6 LEU2add-2p] 4521-001/pDK243Mata leu2 ade2 ade? canl sap3 ural his71LEU2 ade3-2p 4522-281IpDK243 MATa cdc28-1 leu2ade2 ade3 canl sap? his71LEU2 ade3-2p 4523-041/pDK243Mata cdc4-1 leu2ade2 ade3 canl sap3 his71LEU2 ade3-2p 4524-1-31pDK243Mata cdc7-1 leu2ade2 ade3 canl sap3 ural his71LEU2 ade3-2p 4525-061/pDK243Mata cdcb-1 leu2ade2 ade3 canl sap3 his7/LEU2 ade3-2p 4527-0211pDK243Mata cdc2-1 leu2ade2 ade3 canl sap? his71LEU2 ade3-2p 4528-091/pDK243Mata cdc9-1 leu2ade2 ade3 can1 sap? ural his71LEU2 ade3-2p 4529-1311pDK243Mata cdcl3-1 leu2 ade2 ade3 canl sap? his71LEU2 ade3-2p 4530-161/pDK243Mata cdclb-1 leu2 ade2 ade3 canl sap3 his7/LEU2 ade3-2p 4531-4-2IpDK243Mata cdc20-1 leu2ade2 ade3 canl sap3 ural his7/LEU2 ade3-2p 4532-171lpDK243Mata cdcl7-1 leu2 ade2 ade3 canl sap3 ural his71LEU2 ade?-2p 4535-1411pDK243Mata cdcl4-1 leu2 ade2 ade3 canl sap3 his71LEU2 ade3-2p 4536-151/pDK243Mata cdcl5-1 leu2 ade2 ade3 canl sap3 ural his71LEU2 ade3-2p 4541-8-lIpDK243Mata TOPIIleu2 ade2 ade3 canl sap3 ural his7/LEU2 ade3-2p 4563-9-21pDK243 MATa cdcl3-1 his7 ade2ade3 rad52 canl sap31LEU2ade3-2p 4564-9-lIpDK243 MATa cdcl6-1 his7 leu2ade2 ade3 rad52 canl sap31LEU2 ade3-2p 4567-11-41pDK243 MATa cdcl7-1 his7 leu2ade2 ade? rad52canl sap3/LEU2ade3-2p 4571-8-21pDK243 MATa cdc7-1 his7 leu2ade2 ade3 rad52 canl sapSILEU2ade3-2p 4565-1l-lIpDK243 MATa his7 leu2ade2 ade3rad52 canl sap3ILEUZade3-2p Each strain contained a leu2 marker used in transformation and crosses, and ade2 and ade? markers which were necessary for monitoring the ploidy of chromosomes harboring theADE3-2p marker by the visual assay for chromosome ploidy.All markers after theslash are markers coniributed by the mini chromosome^ pDK243 (see Figure 1). which occurred in the first division after plating (see Figure the ploidy of the chromosome in one or both of the 1). The rate of these events per cell division for cells con- progeny cells differed from its ploidy in the parental taining 1 copy of the artificial chromosome was calculated asthe number of half-sectored colonies divided by the cell. These aberrant events led to half-sectored colo- number of pink colonies plus the number of half-sectored nies when the two primary progeny inherited differ- colonies (the total number of cells with one copy prior to ent numbers of the chromosome(Figure 1B). The plating). The scoring of red-white (2:O)and pink-white (1:O) ploidy of the test chromosome in these half-sectored half-sectored colonies was usually unambiguous. However, the difference between the red and pink color did vary in colonies were identified by a visual assay for chromo- the different cdc mutants. some ploidy (KOSHLAND, KENT and HARTWELL1985). The frequency of half-sectored colonies among the RESULTS total colonies was adirect measure of therate of Rationale for experimental design: The amount aberrant events in the first cell division after cells were offunctional CDC geneproduct incells couldbe plated. When wild-type cells or cdc cells were plated reduced during a single cell division because these prior to their shift to the nonpermissive conditions, products could be temporarily inactivated either by the frequency of half-sectored colonies reflected the briefly exposing wild-type cells tothe appropriate rate of aberrant events per cell division of exponen- inhibitors or by briefly exposing cdc mutants to the tially growing wild-type or cdc mutant cells. When nonpermissive temperature (see MATERIALS AND wild-type or mutant cells were plated just after being METHODS). Wild-type and cdc cells contained a non- released from arrest, then the first division of these essential test chromosome, either a14-kb circular cells on the plate was the completion of a division in minichromosome, pDK243, or a 155 kb linear chro- which cells had been previously arrested. Therefore, mosome fragment, CF352 (FigureIA). Aberrant the frequency of half-sectored colonies produced by transmission of these chromosomes occurred when these cells reflected the rate of aberrant events in cell 766 R. E. Palmer, E. Hogan and D. Koshland

pDK243 CEN3

CF352 "_ 145 kb (I55 kb) D ___ Left arm chr /I/ D8B EN6 LEU2 ade3 -2p

A chromosome "yeast cell

shift to non permissive conditions to reduce cdc function

shift to permissive conditions to restorecdc function and allow cells to complete cell division

primary progeny aa a0 transmission pattern 1:l 1 :o of chromosome I I transmission pattern in clonal descendants B

FIGURE1 .-Structure of artificial chromosomes and experimental design to follow changes in their ploidy in cdc mutants. A. Diagram (not drawn to scale) of the two artificial test chromosomes, the 155-kb linear chromosome fragment, CF352, and the 14-kb circular minichro- mosome, pDK243, whose ploidy is monitored in this study. The centromere DNA of the chromosome fragment (CEN6) is derived from chromosome VI. The right arm of the chromosome fragment contains a telomere (solid black), pBR322 sequences, the ADE3-2p gene which was used as a color marker to follow the ploidy of the chromosome (see MATERIALS AND METHODS) and the LEU2 gene which was used as a selectable marker in crosses and transformation. The left arm of the chromosome fragment contains 145 kb of sequences derived from the left arm of chromosome 111, starting at the chromosome Ill sequence, D8B (10 kb from the centromere), and extending to the telomere at the left end. The minichromosome, pDK243 (KOSHLAND, KENT and HARTWELL1985). contains the origin of replication, ARSI, the centromere DNA sequences from chromosome 111, CEN3, the LEU2 gene for selection during crosses and transformation, and theADE3-2p for following changes in its ploidy. B, The function of different CDC gene products was reduced temporarily as described in the text. Since cells that were temporarily arrested were plated before they have a chance to complete cell division, the primary progeny of these cells were physically juxtapose next to each other on the plate. In cells where the reduction in CDC function had no affect on the transmission of the test chromosome (indicated by a black dot), a 1:l segregation pattern was produced in the primary progeny. Since mostof the clonal descendants of the primary progeny contained 1 copy, a colony was formed in which most of the cells contained 1 copy of the minichromosome. Alternatively, the reduction in a given CDC function led to aberrant segregation patterns in the primary progeny (1:O. 2:O and 2:1 events). Since all divisions after the first division on the plate occurred at the pernlissive conditions, the change of ploidy, generated when cells were tenlporarily arrested, was stably perpetuated in the clonal descendants, giving rise to half-sectored colonies. Because the artificial chromosomes were marked with the ADE3-2p gene and the cdc strains were also ade2 ade3, these half-sectored colonies could be detected by the visual assay for chromosome ploidy (see MATERIALS AND METHODS). Transmission of Yeast Chromosomes 767 divisions where CDC function was temporarily re- that no statistically significant conclusions could be duced. Any difference in the absolute rateof 1:O and drawn about their induction. 2:O events, or in the fold induction, that was %fold Basal rate of 1:0 and 2:O events in exponentially greater than wild-type values was considered signifi- dividing wild-type cells and cdc mutants: The rates cant (see legendof Figure 2 andFigure 3). The of 1:0 and 2:O events for the minichromosome per occurrence of 2: 1 events was so rare in exponentially wild-type cell division were 0.01 and 0.005, respec- growing cells and in cells with reduced CDC function tively (Figure 2, A and B). These values are similar to

Temporarily arrested A Exponential

0.12 - c,cv) FIGURE2.-Rate of 1:0 and 2:O 5 .2 events forthe minichromosome, > .E 0.10- pDK243, in cdc mutants. The rates of 1:0 and 2:O events for the mini- chromosome, pDK243 (Figure 1A) e== 0.08- were measured in wild-type cells and aa cdc mutants (see Table 1) growing -0 exponentially at 23" (solid bar) and 0 in wild-type and cdc mutants tempo- L 0.06- rarily arrested (striped bar) by incu- bation for 3 hr at the nonpermissive CR K condition (see text and MATERIALS 0.04 -f AND METHODS). The rates in expo- nentially growing cells and temporar- i ily arrested cells were the average of 0.02 2-4 independent trials in which 500- 1000 colonies were scored. The ab- solute rates of 1:0 and 2:O events 0.00 measured in a given trial differed by wt 3 2817 7 4 2 9 6 Hl 1316 N top2 1514 less than 50% (data not shown) from the mean. When wild-type cells were subjected to conditions used to arrest the cdc mutants (3 hr at 36"), the rates of 1:0 and 2:O events were iden- 0.14 - G1 S LA tical to those rates observed in cells G2-M incubated only at 23". Similarly, the M lack of a striped area for a few cdc 0.12 - mutants indicates that the rates after temporary arrest andthe rates in mu- tant cells growing exponentially at 23" were the same. The cdc mutants 0.10 - are grouped according to the stage in the cellcycle where they arrest. The cdc3 mutant blocks cytokinesis 0.08 - but does not block the nuclear cycle. Though top2 does not give a uniform arrest, its function is needed prior to 0.06 - anaphase (HOLMet al. 1985); there- fore, it was placed in the G2-M class. N, nocodazole arrested wild-type 0.04 - cells. HU, hydroxyurea arrested wild-type cells.

0.02 -

0.00 wt 3 28 4 7 17 2 9 6 Hl 1316 N top2 1415 cdc 768 R. E. Palmer, E. Hogan and D. Koshland

20 I G1 18 -

Q) c. 16 - La 14 -

12 -

10 -

8-

6- P FIGURE 3.-The fold increase in 4- I, 1:0 and 2:O events for the minichro- mosonle in cells temporarily arrested 2- by different cdc mutations. The rate -a of 1:0 and 2:O events was measured in cdc mutants that were temporarily 0- I, .. 3 28 4 172 9 6W13 16 N top2 14 15 arrested by incubation for 3 hr at the nonpermissive temperature (see Fig- ure 2 and text). The fold increase in cdc 1:0 and 2:O events relative to wild- type cells was calculated by dividing the rate of these events in cells tem- porarily arrested by therate ob- G1 S LATE served in wild-type cells subjected to mock arrest conditions(see Figure 2). M Points are the mean of two to four independent trials; the standard de- viation for each value is indicated by I G2" the error bars. P

P P P P III 1f 3 28 4 7 13 16 N top; 14 15 cdc those previously reported for this plasmid in other 4 X events per cell division inwild-type cells. yeast strains (KOSHLAND, KENT and HARTWELL1985). These rates are in agreement with the published rate The rates of 1:0 and 2:O events in mutants grown at of loss for a nearly identical chromosome fragment in the permissive temperature were similar to the rates other yeast strains (Hegemann et al. 1988). Further- observed in exponentially growing wild-type cells (Fig- more, they are 20-fold lower than those rates observed ure 1). Therefore, when these cdc mutants were for pDK243 (Tables 2and 3), which is consistent with grown at their permissive temperature, sufficient CDC previous observations that large linear chromosomes function existed to ensure a fidelity of minichromo- are more stable than small circular chromosomes (Su- some transmission similar to wild-type cells. ROSKY, NEWLONand TYE1986; MURRAY, SCHULTES The rates of 1:O events and 2:O events for the 155 and SZOSTAK1986; HIETERet al. 1985). In exponen- kb linear chromosome fragment occurredat a rate of tially growing cdc7, cdcl7, cdcl6 and cdcl4 cells, the Transmission of Yeast Chromosomes 769 rates of 1:O and 2:O events for the linear fragment 0.12 I 5 were similar to the rates of these events in exponen- 0 .C .C 0.10 tially growing wild-type cells. Therefore, sufficient cdc > function apparently existed in these mutants at their s = 0.08 permissive temperature to ensure a fidelity of chro- W 0 mosome transmission similar to wild type. However, \ u) the rate of 1:0 events for the chromosome fragment 2 0.06 al W in the cdcb mutant strain was 10-fold higher than that W observed in the wild-type strain apparently because z 0.04 CDC6 function may be already partially reduced in c 0.02 this strain even at its permissive temperature. A simi- W d W lar increase in the absolute rate of 1:0 events for the K minichromosome may have occurred in cdc6 mutants 0.00 +- +- +- +- +- RAD52 growing at the permissive temperature, but this in- wt 7 13 16 17 cdc crease was obscured by the high basal rate of 1:0 FIGURE4.-Rad52 dependent 2:0 events. Cells harboring a cdc events observed forthe minichromosome in allstrains. mutation, the rad52 mutation, and the minichromosome, pDK243, Rates of 1:O and 2:O events for the minichromo- were constructed (see Table1). The rate of 2:0 events was measured some in cells reduced for cdc function: The fidelity using the half-sector colony assay (see Figure 1) in exponentially growing cellsand cells temporarily arrested as described in the text. of minichromosome transmission was reduced in The rate of 2:O events in cells grown at the permissive temperature many different cdc mutants when they were tempo- (solid bar) and after temporary arrest (striped bar) are shown. The rarily arrested as evidenced by an increase in the rate rate of 2:O events in the appropriate RAD52 cdc mutant (Figure 2) of 1:O events, 2:0 events, or both (Figures 2 and 3). were regraphed here for ease of comparison. Several observations support the conclusion that the induction of 1:O events was not observed in any of the increases in the rate of 1:0 and 2:O events resulted GImutants, and the induction of 2:O events was not from the temporary reduction in the function of spe- observed in late M mutants. Additional mutants and cific CDC products. Exposure of yeast cells to 36" was additional alleles of these genes must be analyzed to not sufficient by itself to increase the rate of 1:0 or confirm this preliminary pattern. Finally, mutants that 2:O events since wild-type cells shifted to 36" for 3 arrest in the same phase of the cell cyclegave different hours showed the same rate of 1:0 and 2:O events as responses. Among the S phase mutants, cdcb showed wild-type cells grown at 23" (Figure 2). Furthermore, an increase in mostly 1:0 events, while the cdc2, cdc9, different cdc mutants caused different changes in the and cdcl7 showed a dramatic increase in both 1:O and rates of 1:0 and 2:O events(Figures 2 and 3). No 2:O events. Similarly, among the two late M phase increase in the absolute ratesof 1:O or 2:O events were mutants, the cdcl4 mutant showed a dramaticincrease observed in cells temporarily arrested by cdc3 (cyto- in the 1:0 events while the cdcl5 mutant did not. kinesis), hydroxyurea (DNA synthesis), or nocodazole RAD52 dependent 2:O events: The cdc2,cdc9, (mitosis). The rate of 1:0 and 2:O events were induced cdcl7 and cdcl3 mutations had previously been shown to approximately equal values in cdcl6, cdcl7, cdc2 to increase the rate of recombination between homo- and cdc9 mutants. The rates of 1:0 events were signif- logs as well as the rate of loss of endogenous chro- icantly greater than the rateof 2:O events in cdc6 and mosomes; the appearance of recombinant progeny in cdcl4 mutants while therates of 2:O eventswere thesemutants was blocked by a rad52 mutation significantly greater than 1:0 events in cdc7, cdc28and (HARTWELLand SMITH1985). Therefore, it was pos- cdc4 mutants. sible that the RAD52 dependent hyperrecombination The induction of the 1:0 and 2:O events by cdc in these cdc mutants mightcause the observedincrease mutants were analyzed as a function of the stage of in their rate of 2:O events for the circular minichro- the cell cyclewhere the mutantsarrest. The induction mosome. To address this hypothesis the rate of 2:O of 1:0 and 2:O events was not confined to mutants events for the minichromosome was measured in dif- that act at a specific stage of the cell cycle (Figures 2 ferent cdc strains (Table 1) which also carried a rad52 and 3). For example an induction of 1:O events was mutation. observed in all S mutants (cdc2, cdc6, cdc9and cdcl7), The presence of the rad52 mutation had no effect in both G2-M mutants (cdcl3 and cdclb) as well as in on the basal rate of 2:O events in wild-type cells or one late M phase mutant (cdcl4). Similarly, an in- mutant cells grown at theirpermissive condition (Fig- crease in the rate of 2:O events was observed in all the ure 4). While rad52 mutations are known to cause GI mutants (weakly for cdc4 and cdc28 but strongly loss of endogenous chromosomes (MORTIMER,CON- for cdc7), in all S phase mutants except cdc6, and for TOPOULOU and SCHILD1981), the effect of this rad52 both G:, mutants. However, 1:0 and 2:O events may mutation on the transmission of the minichromosome not be inducible at all stages of the cell cycle since the was apparentlyobscured by the high basal rate of 770 R. E. Palmer. E.and Hogan D. Koshland minichromosome loss. The rad52 mutation did signif- TABLE 2 icantly reduce the rate of 2:O events in cells limited Rate of 1:0 and 2:O events per cell division for CDCI3 and CDCl7 function but notin cellslimited for CDC7 and CDC16 function, indicating that the MinichromosomeChromosomefragment induction of 2:O events in some cdc mutants was Strain l:o 2:o 1:0 2:o RAD52 dependent, andin others RAD52 independent. wt 0.010 0.006 0.0004 0.0004 Furthermore, the RAD52 dependent 2:O events were cdc7 0.015 (1.5) 0.050 (8) 0.0030 (7) 0.0010 (3) observed in the two cdc mutants (cdcl3 and cdcl7) cdc6 0.140 (14) 0.020 (3) 0.0200 (50) 0.0030 (9) which exhibit elevated levels of RAD52 dependent cdcZ7 0.050 (5) 0.070 (1 1) 0.0700 (175) 0.0040 (5) recombination (HARTWELLand SMITH1985) while cdcZ6 0.140 (14) 0.100 (16) 0.0050 (12) 0.0040 (10) cdcl4 0.120 (12) 0.010 (1.6) 0.0400 (120) 0.0005 (1) the RAD52 independent 2:O events were observed in Comparison of the absolute rate of 1:0 and 2:O events and the the two cdc mutants (cdc7 and cdcl6) which exhibit fold increase in 1:0 and 2:O events for the minichromosome and no increase in recombination (HARTWELLand SMITH chromosome fragment in cdc mutants. The rate of 1:0 and 2:O 1985). Theseresults are consistent with the hypothesis events for the chromosome fragment were measured in homozy- gous cdc diploids (Table 1) that were growing exponentially or that elevated levels of RAD52 dependent recombina- temporarily arrested. Absolute rates of 1:0 and 2:O events are the tion were responsible for the induction of 2:O events mean of 2 to 4 independent trials, approximately 5000 colonies in cdcl3 and cdcl7 cells and probably in the cdc2 and were scored. The data for the absolute rate of 1:0 and 2:O events forthe minichromosome are taken from Figures 2 and 3 and cdc9 mutants as well since these mutants also show represented here to facilitate a qualitative comparison of the effect elevated levels of RAD52 dependent recombination. of reducing different CDC functions on the transmission of the Rates of 1:0 and 2:O events for a linear chromo- minichromosome and chromosome fragment.The fold increase in the rates of 1:0 and 2:O events observed in cdc mutants relative to some fragment in cells reduced for cdc function: To the rates observed in wild-type cells is shown in parentheses. It was address whether small circular minichromosomes and calculated by dividing the rate of these events in cells temporarily large linear chromosomes responded similarly to the arrested by the rate observed in wild-type (wt) cells subjected to mock arrest conditions. reduction in functional CDC product, we examined the effect of temporarily reducing CDC function on events and 2:O events was independent of the struc- the transmission of the 155-kblinear chromosome ture (the topology, length or nucleotide content) of fragment (Figure IA). The transmission of this chro- the test chromosome. In contrast, thecdcl7 mutation mosome fragment was analyzed in a subset of cdc caused an equal induction of 1:0 and 2:O events for mutations. These mutations were chosen either be- the minichromosome but a much greater increase in cause they acted at different stages of the cell cycleor 1:0 events relative to 2:O events for the chromosome because their effects on minichromosome transmis- fragment. For the minichromosome, the cdc7 muta- sion were different. Fortechnical reasons, the analysis tion caused an increase in the rate of 2:O events only of the transmission of the chromosome fragment was while forthe chromosomefragment this mutation performed in homozygous cdc diploids rather than in induced more 1:0 events than 2:O events. Therefore, haploid strains that were used to analyze the mini- in both the cdc7 and cdcl7 mutants the fold induction chromosome (see MATERIALS AND METHODS). How- of 1:0 events relative to 2:O events was dependent on ever, the rates of 1:0 and 2:O events observed for the the structure of the test chromosome. minichromosome in wild-type and cdcl6 diploid The absolute rate of loss (rate of 1:O events plus 2:O strains(both exponential and temporarilyarrested) events) of the minichromosome was compared with were identical to those observed in haploid cells (data the absolute rate of loss of the chromosome fragment not shown). Therefore, since the overall ploidy of the cell apparently had no affect on the transmission of in the different cdc mutants. The absolute rate of loss the minichromosome, it was valid to compare the rates of the circular minichromosome and linear chromo- of 1:0 and 2:O events forthe minichromosome in some fragment were within 3 fold of each other for haploid cells with those observed for the chromosome cells limited for CDCl7, CDCI4and CDC20 functions fragment in diploid cells. (Table 3). Therefore, the absolute rate of chromo- The pattern of 1:0 events and 2:O events for the some loss induced by thesemutants appears to be minichromosome was compared with the pattern of independent of chromosome size or structure. In con- 1:0 and 2:O events for the chromosome fragment in trast, the absolute rate of minichromosome loss ob- the different cdc strains (Table 2). The following cdc served in cells limited for CDC6, CDC7 and CDC16 mutations induced the same patterns of loss for both functions were respectively 8, 20 and 26 fold greater chromosomes: the cdcl4 mutation induced only 1:0 than the rates observed for the linear chromosome events; cdcl6 induced equally 1:0 and 2:O events; and fragment (Table 3). Therefore, the absolute rate of finally, cdc6 caused a small induction of 2:O events chromosome loss in cells limiting CDC6, CDC7 and and a very large induction of 1:0 events. Therefore, CDC16 function appeared to be sensitive to chromo- in cdc6, cdcl4 and cdcl6 mutants the pattern of 1:0 some size and orstructure. Transmission of Yeast Chromosomes 77 1 TABLE 3 quence of the hypersensitivity of the minichromosome Quantitative comparison of the rates of minichromosome loss to the level of activity of these gene products in wild- and chromosome fragment loss in cdc mutants type cells. This hypersensitivity of the circular mini- chromosome could result from its small size, circular Rate of chromosomeloss per cell topology, or nucleotidecontent. Understanding division for: Ratio of mini- chromosome to which of these factors is responsible for thehypersen- Chromosome chromosome Strain Minichromosomefragmentfragment Strain sitivityof minichromosome to CDC6,CDC7 and CDC16 products may help to elucidate the functions wt 0.0 16 0.0008 20 cdc7 0.065 0.0040 16 of these products as well as provide insight into the cdc6 0.160 0.0230 8 role of chromosome structure in chromosome trans- cdcl7 0.1201.7 0.0740 mission. cdcl6 0.240 0.0090 26 The cdc28 mutant caused minichromosome loss to cdcl4 0.130 0.0400 3 cdc20 0.0351.3 0.0280 increase to an absolute rate of 3% per cell division, while in the experiments of HARTWELLand SMITH The rates of 1:0 and 2:O events for the minichromosome and the chromosome fragment were measured in wild-type (wt) cells (1985), this mutant exhibited no detectable increase and in temporarily arrestedcdc mutants (Table2) and were summed inloss of alinear endogenous chromosome. The to give the total rate of chromosome loss. failure of linear chromosome loss to increase to an DISCUSSION absolute rate of 3% per cell division may reflect some hypersensitivity of the minichromosome to reductions A comparison of the rate of loss of circular min- in CDC28 function like that observed for CDCB, CDC7 ichromosomesand large linear chromosomes in and CDClG. However, theapparent difference be- cells withreduced CDC function: All of the cdc tween minichromosome loss andendogenous chro- mutants that induced loss of endogenous linear chro- mosome loss may simply reflect that CDC28 function mosomes in the study of HARTWELL andSMITH (1 985) was more significantly reduced in this study than in also induced loss of circular minichromosomes in our the experiments of HARTWELLand SMITH,perhaps study (either 1:0 or 2:O events). The similar response because different alleles of CDC28 and differentmeth- of minichromosomes and endogenous chromosomes ods for reducing CDC function were used. to the many different cdc mutants suggests that all of The induction of 1:0 and 2:O events in different these CDC gene products are important in the trans- cdc mutants: The induction of 1:0 and events in mission of both minichromosomes and endogenous 2:O the different cdc mutants provided new insights into chromosomes. Since the cell division cycle mutants appear to identify gene products that participate in the function of their gene products. In the following the major pathway of replication and segregation of discussion our interpretations of 1:0 and 2:O events endogenous chromosomes in yeast, the minichromo- are based upon thefollowing assumptions. A 1 :O event some appearsto be transmittedby this major pathway. for a test chromosome is diagnostic of the failure of Therefore, the high basal instability of minichromo- that chromosome to replicate or theloss of one of the somes apparently is not a consequence of minichro- two sister chromatids subsequent to replication. A 2:O mosomes being transmitted by some secondary path- event is diagnostic of a failure in chromosome segre- way, but rather because they may be poor substrates gation since the two sister chromatids end up in the forone or more steps in themajor pathway (see same progenycell. However, it should be keptin mind below). Furthermore, since all of cdc mutants tested thatimproper chromosome segregation might be induce loss of both endogenous chromosomes and caused by defects in replication machinery as well as circular minichromosomes, the CDC products must in the segregation machinery. For example, a failure act directly or indirectly on structural features(nucle- to complete DNA replication presumably would pro- otides or proteins) shared by minichromosomes and hibit sister chromatid separation. The partially repli- endogenous chromosomes. Therefore, none of these cated molecule might be pulled to one pole, and then CDC gene products appear to interact with proteins its replication might be completed thereafteror within or sequences that are present exclusively on endoge- the next cell cycle giving rise to the 2:O event. nous linear chromosomes such as telomere proteins GI mutants: The cdc28 and cdc4 mutations induce or telomere sequences. only 2:O events forthe circular minichromosomes. The absolute rate of minichromosome loss was sig- Given the observation that thesemutants seem to nificantly greater than the rate of loss of the linear arrestthe cellcycle before DNA synthesis is even chromosome fragment for cdc6, cdc7 and cdcl6 mu- initiated (PRINGLEand HARTWELL198 l), it may seem tants suggesting that the minichromosome is hyper- surprising that these mutants shouldcause a defect in sensitive to the level of function of these CDC gene chromosomesegregation. However, the earliest products. In fact, the high basallevel of minichro- events of the cell cycle in budding yeast involve mor- mosome loss observed in all cellsmay bea conse- phogenesis of the spindle pole bodies, the microtubule 772 R. E. Palmer. E.and Hogan D. Koshland

organizing centers, that are key to the formation and (KOSHLANDet al. 1987). The subsequent resolution of function of the spindle. In fact, early steps in spindle the dimer intotwo monomers before the next mitosis pole body morphogenesis are blocked in cdc28 and would give rise to a 2:O event.A rad52 mutation cdc4 arrested cells (BYERSand GOETSCH1974; BYERS would block the induction of 2:O events by blocking 1981). Therefore, it is not surprising that temporarily sister chromatidexchange between the sister mini- reducing the functionof these two gene productsmay chromatids. perturb the normal assembly of the spindle pole bod- If sister chromatid exchange in cdcl7 mutants can ies. When cdc28 and cdc4 cells reenter the cell cycle, induce 2:0 events for the circular minichromosome, aberrant spindles may be formed and subsequently then other cdc mutations which have elevated levels malfunction during mitosis leading to chromosome of sister chromatid exchange should also induce 2:O nondisjunction. eventsfor the minichromosome. Limiting cells for The cdc7 mutants have extremely complex cell cycle CDC2, CDC9 or CDC13 function stimulates recombi- phenotypes. The initial characterization of this mu- nation between homologs (HARTWELLand SMITH tant indicatedthat it arrest cellswith botha G1 1985). Since mutations in these genes arrestcells after phenotype (failure to initiate DNA synthesis) (HART- the initiation of DNA synthesis but prior to chromo- WELL 1973) and a G2-M, phenotype (formation of a some separation (PRINGLEand HARTWELL198 l), that bipolar spindle) (BYERSand GOETSCH1974). The is within the window of the cellcycle when sister function of this protein became more unclear when chromatidexchange can occur, it is reasonable to subsequentexperiments demonstrated that a small suggest that these mutations may induce sister chro- amountnon-random DNA synthesis doesoccur in matid exchange as well as homolog exchange. Inter- mitotically arrested cdc7 cells and that cdc7 mutations estingly, mutations in CDC2,CDC9 and CDC13 all block meiosis after thecompletion of premeiotic DNA induced 2:0 events. While mutations in CDC14 and synthesis (SCHILDand BYERS1978; SIMCHEN1974). CDC6 also stimulate homolog exchange, they appar- In this study we document another perplexing phe- ently arrest cells at stages of the cell cycle where sister notype of this mutant. The transmission of the mini- chromatid exchangemay not be possible because sister chromosome and large linear chromosome fragment chromatids have alreadybeen separated (cdcl4) differfrom each otherboth qualitatively (the fold (BYERSand GOETSCH1947) or because sister chro- induction of 1:0 relative to 2:O events) and quantita- matids have not yet been formed (cdc6) (HARTWELL tively (the absolute rate of loss). Recently, the CDC7 1976). Interestingly, thesetwo mutants induce1:0 but product has been shown to be a protein kinase (PAT- not 2:O events for the minichromosome. TERSON et al. 1986). Perhaps the identification of its G2-M:The cdclb mutant induces 1:0 and 2:O events substrates will help to elucidate the relationship be- for both the minichromosome and the chromosome tween these unusual phenotypes. fragment. Unlike the S phase mutants the 2:0 events S phasemutants and RAD52 dependent 2:O were not dependent upon the RAD52 gene product events: The product of the cdcl7 genes is required or on the topology of the chromosome. In addition, for progression through S phase because it encodes a the cdclb mutant does not induce recombination be- DNA polymerase essential for DNA replication (CAR- tween homologs (HARTWELLand SMITH 1985). SON 1987). Impairing its function apparently leaves Therefore, cdclb mutants do not appear to induce lesions in the DNA which stimulate RAD52 dependent 2:O events as a consequence of aberrant DNA metab- recombination between homologs (HARTWELLand olism. These phenotypes coupled with its arrest at GP- SMITH1985). In this study, we show that the cdcl7 M boundary suggest that theCDCl6 product is a good mutation also increase the rate of 1:0 and 2:O events candidate for a protein directly required for proper for the minichromosome. This increase in the rate of initiation or completion of chromosome segregation. 2:0 events was observed onlyin the RAD52 back- The increase in 1:0 events in cdclb mutants is not ground and only with the circular minichromosome, inconsistent with its putative role in chromosome seg- but not the linear chromosome fragment. regation since improperchromosome segregation To explain these observations, we propose that the could occasionally lead to destruction of one of the cdcl7 mutation stimulates RAD52 dependent recom- two sister chromatids. bination between sister chromatids as well as homo- Impairing topoisomerase 11 function or microtu- logs (HARTWELLand SMITH1985). Sister chromatid bule assembly might be expected to increase the rate exchange between linear sister chromatids would sim- of 2:O events. Topoisomerase I1 functions at mitosis ply cause the exchange of sequences distal to thecross to allow chromosome segregationby removing tangles over, while exchange between circularchromatids between sister chromatids caused by catenation of the would cause the formation of a dicentric dimercircle. sister DNA molecules (HOLMet al. 1985). Indeed, in Dicentric minichromosomes often survive mitosis in this study and in a study of linear chromosome frag- yeast and segregate to one of the two progeny cells ments (HOLM,STEARNS and BOTSTEIN1989), mutants Transmission of Yeast Chromosomes 773 in topoisomerase I1 did show and increase in the rate thank PHIL HIETER, GARY KARPEN, KAREN LAMARCO and FORREST of 2:O events. Similarly, microtubules are essential to SPENCERfor their comments on the manuscript. CHRISTINENOR- MAN assisted with the preparation of the manuscript. R. E. P. was the movement of chromosomes during all phases of supported by NationalInstitutes of Health Training Grant HD mitosis. However, impairing microtubuleassembly for 07276 awardedto The Johns Hopkins School of Hygiene and 3 hr did not show an increase in 2:O events for the PublicHealth, Department of Population Dynamics, Division of minichromosome. Given the role of microtubules in Reproductive Biology, and a Howard Hughes predoctoral grant chromosomesegregation this result is surprising. awardedto the Carnegie Institution of Washington. D. K.is a Lucille P. Markey Scholar and this work was supported in part by However, the ability of cells to transmit linear endog- a grant from theLucille P. Markey Charitable Trust. enous chromosomes also seemed to be unaffected by microtubule depolymerizing reagents unless cells are LITERATURECITED incubated with these drugs for periods as long as 24 hr (WOOD1982). ARION, D., L. MEIJER, L. BRIZUELAand D. BEACH,1988 cdc 2 is a component of the M phase-specific histone HI kinase: evi- Late M: Cells arrested by cdcl4 and cdcl5 mutations dence for identity with MPF. Cell 55: 371-378. block in late M phase (PRINGLE andHARTWELL 198 1; BARKER,D. G., and L. H. JOHNSTON,1983 Saccharomycescereuisiae BYERS1981). Therefore, the fact that the cdcl4 mu- cdc9, a structural gene foryeast DNA ligase which complements tant induced a much greater number of 1:0 events Schizosaccharomyces pombe cdcl7. Eur. J. Biochem. 134: 315- than 2:O events suggests that adefect in terminal 319. BEACH,D. H., B. DURKACZ andP. M. NURSE,1982 Functionally events in mitosis can lead to loss of one of the repli- homologous cell cycle controlgenes in budding and fission cated sister chromatids from one of the two segre- yeast. Nature 300 706-709. gated genomes. Since the cdcl5 mutant fails to induce BOULET,A,, M. SIMON,G. FAYE, G. A. BAUER, P. M. BURGERS, 1:O events, arresting cells at this terminal stage of the 1989 Structureand function of the Saccharomyces cereuisiae cellcycle apparently is not sufficient to induce 1:0 CDC2 gene encoding the large subunit of DNA polymerase 111. EMBO J. 8: 1849-1854. events. The CDC14 geneproduct may representa BYERS,B. 1981 Cytology of the Yeast Life Cycle, pp. 59-96 in classof proteinsthat are requiredto ensure that Molecular Biology of the Yeast Saccharomyces: Life Cycle and Inher- segregated chromosomes are packaged into the two itance, edited by J. N.STRATHERN, E. W. JONES and J. R. daughter nuclei when the nucleus divides at the end BROACH. Cold Spring Harbor Laboratory, Cold Spring Harbor, of mitosis. A defect in these proteins may cause chro- N.Y. BYERS,B., and L. GOETSCH,1974 Duplication of spindle plaques mosomes to leech into the cytoplasm and be lost. and integration of the yeast cell cycle. Cold Spring Harbor Alternatively, DNA metabolism may occur in this late Symp. Quan. Biol. 38: 123- 13 1. stage of mitosis, which is important for replication or CARSON,M. J., 1987 CDCZ7, the structural gene for DNA polym- segregation in thenext cell division. Afailure to eraseI of yeast: mitotic hyperrecombination and effects on properly perform this metabolism may lead to chro- telomere metabolism. Ph.D. thesis, University of Washington, mosome damageand subsequentdestruction. The Seattle. CLARKE, L., andJ. CARBON,1980 Isolation of a yeast centromere fact that the cdcl4 mutant does induce homolog re- andconstruction of functional small circularchromosomes. combination is consistent with its product playing Nature 287: 504-509. some rolein DNA metabolism (HARTWELL andSMITH ESPOSITO, M. S., D. T. MALEAS,K. A. BJORNSTADand C. V. 1985). BRUSCHI,1982 Simultaneousdetection of changes in chro- Conclusion: In this study we have analyzed the mosome number, gene conversion and intergenic recombina- tion during mitosis ofSaccharomyces cereuisiae spontaneous and transmission of artificial chromosomes in cdc mutants. ultraviolet light induced events. Curr. Genet.6 5-1 1. An understanding of the chromosome loss events (1 :O FITZGERALD-HAYES, M.,M. J. BUHLER,T. G. COOPERand J. CAR- or 2:O events) that are induced by mutations affecting BON, 1982 Isolation andsubcloning analysis of functional defined stages of the cell cycle provides a framework centromere DNA (CEN1I) from Saccharomyces cerevisiae chro- for analyzing changes in chromosome ploidy that oc- mosome XI. Mol. Cell. Biol. 2: 82-87. GAUTIER,J., C. NORBURY,M. LOHKA, P. NURSEand J. MALLER, cur in mutants of other less well-characterized genes. 1988 Purifiedmaturation-promoting factor contains the This study also provides a framework for thinking product of a Xenopus homolog of the fission yeast cell cycle about themode of action of environmental toxins that control gene cdc2+. Cell 54: 433-439. reduce the fidelity of chromosome transmission. In GERRING,S. L., C. CONNELLYand P. HIETER, 1989 Positional this study, we showed that the reduction in activity of mapping of genes by chromosome blotting and chromosome fragmentation. Methods Enzymol. (in press). most CDC gene products during a single cell cycle is HARTWELL,L. H., 1973 Three additionalgenes required for sufficient to inducechromosome loss. Therefore, deoxyribonucleic acid synthesis in Saccharomyces cereuisiae. J. chromosome loss in yeast potentially could be stimu- Bacteriol. 115: 966-974. lated by a very brief (one cell cycle) exposure to any HARTWELL,L. H., 1976 Sequentialfunction of gene products environmental toxin that reduces the function of one relative to DNA synthesis in the yeast cell cycle. J. Mol. Biol. 104: 803-817. of the CDC gene products. HARTWELL, L. H., and D. SMITH,1985 Altered fidelity of mitotic chromosome transmission in cell cycle mutants of S. cereuisiae. This workbenefited greatly from the comments of LELAND Genetics 110: 381-395. HARTWELL. We thank GLENBLUSTEIN for technical assistance, and HARTWELL,L. H., S. K. DUTCHER,J. S. WOOD and B. GARVIK, 774 R. E. Palmer, E. Hogan and D. Koshland

1982 The fidelity of mitotic chromosome reproduction in S. mosome length controls mitotic chromosome segregation in cerevisiae. Rec. Adv. Yeast Mol. Biol. 1: 28-38. yeast. Cell 45: 529-536. HEGEMANN,J. H., J. H. SHERO,G. COTTAREL,P. PHILIPPSENand PATTERSON,M., R. S. SCLAFANI,W. L. FANGMAN andJ. ROSAMOND, P. HIETER, 1988 Mutational analysis of the centromere DNA 1986 Molecular characterization of cell cycle gene CDC7 fromchromosome VI of Saccharomyces cerevisiae. Mol. Cell. from Saccharomyces cereuisiae. Mol. Cell. Biol. 6: 1590-1598. PRINGLE,J. R., and L. H. HARTWELL,1981 The Saccharomyces Biol. 8: 2523-2528. cerevisiae cell cycle, pp. 97-142 in The Molecular Biology ofthe HIETER, P.,C. MANN,M. SNYDER and R.W. DAVIS,1985 Mitotic Yeast Saccharomyces: Life Cycle and Inheritance, edited by J. N. stability of yeast chromosomes: A colony color assay that meas- STRATHERN,E. W. JONES andJ. R. BROACH.Cold Spring ures nondisjunction and chromosome loss. Cell 40: 381-392. Harbor Laboratory, Cold Spring Harbor, N.Y. HOLM, C., T. GOTO,J. C. WANGand D. BOTSTEIN,1985 DNA SCHILD,D., and B. BYERS,1978 Meiotic effects of DNA-defective topoisomerase I1 is required at the timeof mitosis in yeast. Cell cell division cycle mutations of Saccharomyces cerevisiae. Chro- 41: 553-563. mosoma 70: 109-130. HOLM,C., T. STEARNSand D. BOTSTEIN,1989 DNAtopoisom- SHERMAN,F., G. R. FINK,and J. B. HICKS,1986 Methods in Yeast erase I1 mustact at mitosis toprevent nondisjunction and Genetics. Cold Spring Harbor Laboratory,Cold Spring Harbor, chromosome breakage. Mol. Cell. Biol. 9: 159-168. N.Y. HSIAO,C.-L., and J. CARBON,1979 High-frequency transforma- SIMCHEN,G., 1974Are mitotic functionsrequired in meiosis? Genetics 76 745-753. tion of yeast by plasmids containing thecloned yeast Argl gene. SITNEY,K. C., M. E. BUDD and J. L.CAMPBELL, 1989 DNA Proc. Natl. Acad. Sci. 76: 3829-3833. polymerase 111, a second essential DNA polymerase, is encoded ITO, H., Y. FUKUDA,K. MURATA~~~A. KIMURA, 1983 Transforma- by the S. cerevisiae CDCZ gene. Cell 56: 599-605. tion of intact yeast cells treated with alkali cations. J. Bacteriol. STINCHCOMB,D. T., C. MANN and R.W. DAVIS,1982 Centromeric 153: 163-168. DNA from Saccharomyces cerevisiae. J. Mol. Biol. 158: 157-179. KOSHLAND, D.,and P. HIETER, 1987 Visual assay for chromosome STINCHCOMB,D. T., K. STRUHLand R. W. DAVIS,1979 Isolation ploidy. Methods Enzymol. 155: 351-372. and characterizationof a yeast chromosomal replicator. Nature KOSHLAND,D., J. C. KENTand L. H. HARTWELL, 1985 Genetic 282: 39-43. analysis of the mitotic transmission of minichromosomes. Cell SUROSKY, R.T., C. S. NEWLONand B-K. TYE,1986 The mitotic 40: 393-403. stability of deletion derivativesof chromosome 111 in yeast. Proc. Natl. Acad. Sci. 83: 414-418. KOSHLAND,D., L.RUTLEDGE, M. FITZGERALD-HAYESand L. H. VOLLRATH,D., R. W. DAVIS, C.CONNELLY and P. HIETER, HARTWELL,1987 A genetic analysis of dicentric minichro- 1988 Physical mapping of large DNA by chromosome frag- mosonles in Saccharomyces cerevisiae. Cell 48: 801-81 2. mentation. Proc. Natl. Acad. Sci. 85: 6027-6031. LORINCZ,A. T., and S. I. REED, 1984 Primary structural homol- WOOD,J. S., 1982Genetic effects of methyl benzimidazole-2-yl ogy between the product of yeast cell division control gene carbamate onSaccharomyces cereuisiae. Mol. Cell. Biol. 2: 1064- CDC28 and vertebrate oncogenes. Nature307: 183-185. 1079. MORTIMER,R. K., R. CONTOPOULOUand D. SCHILD, 1981 Mitotic YOCHEM,J., and B. BYERS, 1987Structural comparison ofthe chromosome lossin aradiation-sensitive strain of the yeast yeast cell division cycle gene CDC4 and a related pseudogene. Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. 78: 5778-5782. J. Mol. Biol. 195: 233-245. MURRAY,A. W., N. P. SCHULTESand J. W. SZOSTAK,1986 Chro- Communicating editor: M. CARLSON