Molecular Cell, Vol. 6, 487–492, August, 2000, Copyright 2000 by Cell Press
The DNA Damage Checkpoint Signal
in Budding Yeast Is Nuclear Limited
Janos Demeter,* Sang Eun Lee,†
James E. Haber,† and Tim Stearns*‡
* Department of Biological Sciences and
Department of Genetics
Stanford University
Stanford, California 94305
† Rosenstiel Center and
Department of Biology
Brandeis University
Waltham, Massachusetts 02454
Summary
The nature of the DNA damage–induced checkpoint
signal that causes the arrest of cells prior to mitosis
is unknown. To determine if this signal is transmitted
through the cytoplasm or is confined to the nucleus,
we created binucleate heterokaryon yeast cells in
which one nucleus suffered an unrepairable doublestrand break, and the second nucleus was undamaged. In most of these binucleate cells, the damaged
nucleus arrested prior to spindle elongation, while the
undamaged nucleus completed mitosis, even when
the strength of the damage signal was increased. The
arrest of the damaged nucleus was dependent upon
the function of the RAD9 checkpoint gene. Thus, the
DNA damage checkpoint causing G2/M arrest is regulated by a signal that is nuclear limited.
Introduction
Cell cycle checkpoints are biochemical pathways that
make the execution of one cell cycle step conditional
on the completion of a previous step. One of the first
observations of a cell cycle checkpoint came from the
experiments of Rao and Johnson (1970). By fusing mammalian cells that were in different parts of the cell cycle,
they found that a G2 nucleus delayed mitosis until a G1
or S phase nucleus in the same cytoplasm finished DNA
replication, then the two nuclei entered mitosis synchronously. These results indicated that the nuclei could
communicate their replication state to the cytoplasm,
where an integration of the signals originating in the
different nuclei occurred, determining whether the cell
could enter mitosis.
The DNA damage checkpoint monitors the state of
the chromosomal DNA. When DNA is damaged, the cell
is prevented from entering mitosis until the damage is
repaired (Weinert, 1998). Much is known about the
mechanism of transduction of the DNA damage signal,
which involves a cascade of protein kinases (Weinert,
1998; Sanchez et al. 1999), but the nature of the damage
signal itself and the manner in which the checkpoint
targets the cell cycle machinery are still unclear.
In several organisms, activation of the DNA damage
checkpoint leads to an inhibitory phosphorylation of the
‡ To whom correspondence should be addressed (e-mail: stearns@
stanford.edu).
mitosis-activating kinase Cdk1, keeping its activity at
the interphase level (Rhind and Russell, 1998) and thus
preventing entry into mitosis. In addition, the localization
of several key components is regulated; for example, in
mammalian cells Cdk1 is prevented from entering the
nucleus in response to DNA damage by a 14-3-3 protein
(Jin et al., 1998; Chan et al., 1999). In fission yeast,
although Cdk1 seems to be predominantly nuclear at
all times, the localization of its activating phosphatase,
Cdc25, changes in response to DNA damage, becoming
cytoplasmic (Furnari et al., 1999). These observations
suggest that control of the entry into mitosis after DNA
damage might require transmission of the DNA damage
signal to the cytoplasmic compartment, although they
could also be explained by a nuclear-limited active export of a key mitotic component from the nucleus.
In the budding yeast Saccharomyces cerevisiae, inhibitory phosphorylation of Cdc28p, the homolog of
Cdk1, does not play a role in cell cycle arrest after DNA
damage (Amon et al., 1992; Sorger and Murray, 1992).
Instead, the DNA damage checkpoint appears to regulate the stability of the cyclin subunit of the mitotic kinase
(Sanchez et al., 1999; Tinker-Kulberg and Morgan, 1999).
Although the endpoint of the arrest in S. cerevisiae is
different, it is mediated by the same set of checkpoint
protein kinases as are found in both fission yeast and
mammalian cells, suggesting that the pathway from DNA
damage to cell cycle arrest is fundamentally the same
(Sanchez et al., 1999; Tinker-Kulberg and Morgan, 1999).
To gain insight into the nature of the DNA damage
checkpoint and how it regulates the cell cycle machinery, we adopted an approach similar to the original Rao
and Johnson (1970) experiments, but using a well-characterized and precise way of inducing DNA damage in
yeast cells. We show that the inhibitory checkpoint signal issued by a chromosome with a single unrepairable
double-strand break (DSB) is nuclear limited and not
transmitted to an undamaged nucleus in the same cytoplasm, i.e., that it is nondiffusible in yeast.
Results
To test whether the DNA damage checkpoint signal is
diffusible from one nucleus to another in yeast cells, we
wished to create binucleate cells, activate the damage
checkpoint in one of the nuclei, and then determine the
fate of the two nuclei (Figure 1). There are three possible
outcomes from such an experiment (Figure 1). If the
DNA damage signal is diffusible, then activation of the
damage checkpoint should prevent both nuclei from
dividing. If the DNA damage signal is not diffusible, then
activation of the damage checkpoint might act only locally by preventing the damaged nucleus from dividing,
but allowing the undamaged one to divide, resulting in
what we will call mononuclear division. The third possibility is that the undamaged nucleus provides a dominant signal that overrides the activated checkpoint in
the damaged nucleus, resulting in a binuclear division.
To carry out this experiment, three conditions had to
be met. First, we needed to create a binucleate cell,
then to induce defined DNA damage in one of the two
nuclei of that cell, and finally to distinguish the damaged
Molecular Cell
488
Figure 1. Experimental Setup and Potential Outcomes
Cells of the indicated relevant genotypes were mated (cell fusion)
and treated with hydroxyurea (HU arrest) to synchronize all cells
at S phase. HO endonuclease was expressed from a galactoseinducible promoter (PGAL-HO). After release from the hydroxyurea
block, cells were fixed and stained for immunofluorescence microscopy by anti-␣-tubulin antibody to visualize microtubules (red), antiGFP antibody to detect the damaged nucleus (green), and DAPI to
observe nuclei (blue). The three potential phenotypes are shown in
the lower panel: if the DNA damage checkpoint is diffusible, neither
nucleus would divide; if the normal nucleus provides a dominant
signal, both would divide; and if the DNA damage signal is nondiffusible, only the undamaged nucleus would divide.
and undamaged nuclei from each other in the ensuing
cell cycle (Figure 1). Binucleate cells were created by
mating two haploid yeast strains, one of which carried
the kar1-⌬15 mutation. This mutation causes a block
in nuclear fusion, resulting in zygotes that have two
separate nuclei (Vallen et al., 1992) but are otherwise
normal. Defined DNA damage was created by expressing the yeast HO endonuclease in the binucleate zygote
from a galactose-inducible promoter. The HO endonuclease normally cleaves a single site in the genome at
MAT, the expressed mating-type locus (Kostriken et al.,
1983). The HO cleavage site is also present at the transcriptionally silent HML and HMR mating type loci, but
these sites are not normally accessible for cutting. A
single persistent DSB generated by HO is sufficient to
cause cells to arrest prior to anaphase (Sandell and
Zakian, 1993; Toczyski et al., 1997; Lee et al., 1998).
To arrange for the DSB to occur in only one nucleus
in the binucleate zygote, the Kar1⫹ parent contained a
deletion of the HO cleavage site at the MATa locus,
rendering it immune to the action of HO. Mixing of the
cytoplasms after mating would allow access of the HO
endonuclease to the kar1-⌬15 nucleus, which contains
a cleavable MAT␣ locus; thus, this nucleus would suffer
a DSB (Lee et al., 1998). Although this break can be
repaired by recombination with the HML and HMR loci,
in the continuous presence of high levels of the HO
endonuclease the site is rapidly recleaved (Connolly et
al., 1988). To eliminate the possibility that this repair
was altering the response to the DNA damage, we also
performed the experiments with strains deleted for the
HML and HMR loci, where no repair by recombination
is possible.
To distinguish the two nuclei within the zygote after
mating, we marked one of the chromosomes in the kar1⌬15 nucleus with a GFP tag. We used a system developed to visualize chromosomes by binding of a GFPlacI fusion protein to tandem repeats of Lac operator
sites (lacO) integrated into chromosome III (Straight et
al., 1996). In this case, the Kar1⫹ strain expressed the
GFP-lacI fusion protein and the kar1-⌬15 mating partner
contained the lacO sites on chromosome III. Thus, the
zygote will contain one nucleus with a single green fluorescent dot, which will become two dots when mitosis
segregates the sister chromatids. Entry into anaphase
of mitosis by individual nuclei was assayed by antitubulin immunofluorescence to visualize mitotic spindle
elongation. To maximize the number of cells in mitosis
at a given time point, zygotes were synchronized in S
phase by treatment with hydroxyurea for 4.5 hr then
washed to allow the cells to resume cell cycle progression. Typically, no zygotes contained nuclei undergoing
mitosis at the time of release from the hydroxyurea
block, and only a low percentage (⬍10%) had entered
mitosis 1 hr after release. Most of the zygotes had begun
nuclear division at the 2 hr time point, with some residual
division at 3 hr.
We expected that two undamaged nuclei in the common cytoplasm of a zygote would undergo a synchronous, binuclear division in response to activation of mitosis. This was confirmed in control experiments in
which the Kar1⫹ ⫻ kar1-⌬15 zygote did not express the
inducible HO endonuclease. The zygotes resulting from
the cross of TSY1260 with TSY1261 (Table 1) showed
binuclear division 84.7% of the time (Figures 2, top, and
3a). In those zygotes in which only one nucleus divided,
Table 1. Strains Used in This Study
Strain
Genotype
Source
TSY1260
TSY1261
TSY1262
MAT␣ kar1-⌬15 his3-200 leu2:lacO256:LEU2 ura3-52
MATa his3-⌬200 leu2-3,112 lys2-801 ura3:PCUP1-GFP-lacI:URA3
mata-⌬117 ⌬ho ⌬hml::ADE1 ⌬hmr::ADE1 ade1-100 leu2-3,112 lys5 trp1::hisG ura3:PCUP1-GFP-lacI:URA3
ade3::PGAL10-HO
mata-⌬117 ⌬ho ⌬hml::ADE1 ⌬hmr::ADE1 ade1-100 leu2-3,112 lys5 trp1::hisG ura3:PCUP1-GFP-lacI:URA3
ade3::PGAL10-HO rad9::kanMX6
MAT␣ kar1-⌬15 his3-200 leu2:lacO256:LEU2 ura3-52 rad9::kanMX6
MATa his3-⌬200 leu2-3,112 lys2-801 ura3:PCUP1-GFP-lacI:URA3 yku70::kanMX6
mata-⌬117 ⌬ho ⌬hml::ADE1 ⌬hmr::ADE1 ade1-100 leu2,3-112 lys5 trp1::hisG ura3:PCUP1-GFP-lacI:URA3
ade3::PGAL10-HO yku70::kanMX6
MAT␣ ⌬ho ⌬hml::ADE1 ⌬hmr::ADE1 ade1-100 leu2:lacO256:LEU2 lys5 trp1::hisG ura3-52 kar1-⌬15
MAT␣ ⌬ho ⌬hml::ADE1 ⌬hmr::ADE1 ade1-100 leu2:lacO256:LEU2 lys5 trp1::hisG ura3-52 kar1-⌬15
yku70::kanMX6
MAT␣ ⌬ho ⌬hml::ADE1 ⌬hmr::ADE1 ade1-100 leu2:lacO256:LEU2 lys5 trp1::hisG ura3-52::HO cut
site:URA3 kar1-⌬15
This study
This study
This study
TSY1263
TSY1264
TSY1265
TSY1266
TSY1267
TSY1268
TSY1272
This study
This study
This study
This study
This study
This study
This study
A Nuclear-Limited DNA Damage Checkpoint Signal
489
Figure 3. Nuclear Divisions in Heterokaryon Zygotes
Figure 2. The DNA Damage Signal Is Nuclear Limited
Zygotes without (WT) or with DNA damage (PGAL-HO) that are undergoing nuclear division are shown. Zygotes are stained for ␣-tubulin
to visualize the microtubules (red), GFP to visualize the lacO-marked
chromosome in the MAT␣ nucleus (green), and DAPI to visualize
DNA (blue). A Nomarski image of the zygotes is also included in the
lower right of each set. In the absence of DNA damage (WT), in the
majority of the zygotes both nuclei divided, indicated by the crossed
long spindles. When high levels of HO endonuclease were present
(PGAL-HO), the undamaged nucleus in most zygotes entered mitosis
whereas the damaged nucleus, marked with GFP, did not.
the GFP-marked and the unmarked nucleus were
equally likely to divide.
Binucleate zygotes with a DSB in one of the two nuclei
were made by crossing TSY1260 with TSY1262, which
expresses the HO endonuclease. Of all such zygotes
dividing by 3 hr after release from the hydroxyurea arrest, 68.4% displayed mononuclear division in which
the undamaged nucleus divided while the GFP-marked,
damaged nucleus remained undivided (Figure 2, bottom, and 3a). This result indicates that the inhibitory
(a) Percentage of mitotic zygotes with mononuclear division. The
relevant genotypes are indicated under the graph: WT, no DNA
damage (TSY1260 ⫻ TSY1261); HO, DNA damage (TSY1260 ⫻
TSY1262); HO rad9⌬, DNA damage, no Rad9p (TSY1263 ⫻
TSY1264); HO hml⌬ hmr⌬, DNA damage, no silent mating type information in the damaged nucleus (TSY1267 ⫻ TSY1262); HO hml⌬
hmr⌬ yku70⌬, DNA damage, no silent mating type information and
no Yku70p (TSY1268 ⫻ TSY1266); HO hml⌬ hmr⌬ 2⫻ cut site, DNA
damage, no silent mating type information, and a second HO cut
site introduced at the URA3 locus (TSY1262 ⫻ TSY1272).
(b) Increased DNA damage signal does not increase delay in the
division of the undamaged nucleus. The graph shows the percentage of zygotes undergoing nuclear division 2 hr after hydroxyurea
release. The respective strains are the same as in (a). hml⌬ hmr⌬
yku70⌬ shows the result from the cross between TSY1268 ⫻
TSY1265; hml⌬ hmr⌬ 2⫻ cut site: TSY1261 ⫻ TSY1272. In the absence of DNA damage (WT), mostly binuclear divisions were observed, whereas in the presence of DNA damage, mostly mononuclear divisions, as shown in (a).
signal that is elicited by DNA damage in one of the nuclei
acts only locally and is not transmitted to the other
nucleus.
If the mononuclear division we observed in the zygotes with one damaged nucleus was due to the action
of the DNA damage checkpoint acting on the damaged
nucleus, then it should be possible to abrogate it by
deletion of RAD9, which abolishes the HO-induced DSB
checkpoint (Lee et al., 1998; Weinert, 1998). When the
above experiment was repeated with rad9 deletion mutations in both parent strains, in the majority of zygotes
(85.6%) both nuclei divided (Figure 3a), similar to the
results in the absence of DNA damage. This indicates
Molecular Cell
490
that the arrest of the damaged nucleus is wholly dependent on the DNA checkpoint and not due to structural
defects that prevent the execution of mitosis.
In a substantial number of zygotes (31.6%) in which
damage was induced, the damaged nucleus divided
together with the undamaged nucleus (Figure 3a). One
possible mechanism for the escape of the damaged
nucleus from the cell cycle arrest is recombinational
repair of the DSB damage using the silent HML and HMR
loci present in the damaged nucleus as gene conversion
donors. The repaired MAT locus could then be cut again
by the HO endonuclease, but the temporarily intact DNA
might allow some cells to progress through mitosis. We
tested this hypothesis by inducing the HO endonuclease
in MAT␣ haploid cells bearing either wild-type HML␣
and HMRa donor loci, or hml⌬ and hmr⌬ mutations. In
the presence of recombination donors, ⵑ30% of the
cells were able to resume growth within 5 hr, whereas
in the absence of recombination donors, virtually all
cells remained arrested (Lee, et al., 1998; and data not
shown). Thus, we repeated the heterokaryon experiment
using a MAT␣ hml⌬ hmr⌬ strain (TSY1267) as a parent,
eliminating the possibility of recombinational repair of
the DSB. Under these conditions, 93.3% of the zygotes
displayed mononuclear division with the damaged nucleus remaining undivided (Figure 3a), supporting the
hypothesis that the originally observed fraction of binuclear divisions was due to repair of the damage at the
MAT locus.
In the above experiments, we used hydroxyurea to
synchronize the zygotes prior to mitosis. Hydroxyurea
blocks DNA replication and activates the replication
checkpoint. To exclude the possibility of interference
between the hydroxyurea-generated signal and the DNA
damage signal, we performed the heterokaryon experiment in the absence of hydroxyurea treatment. Although
fewer cells were observed to be in division due to the
lack of synchronization, there was no quantitative difference in the behavior of the zygotes. Without DNA damage, 89.2% of zygotes displayed binuclear nuclear division, while in the presence of DNA damage in one
nucleus, 94.4% of zygotes displayed mononuclear division in which the damaged nucleus remained arrested.
These experiments demonstrate that a single unrepaired DSB in one nucleus creates a checkpoint-mediated arrest of that nucleus only, without affecting the
division of a second, undamaged nucleus in the same
cytoplasm, suggesting that the DNA damage signal is
not diffusible.
An Increased Damage Signal Does Not Affect Division
of the Undamaged Nucleus
An alternative interpretation of the results presented
above is that the signal is diffusible, but a single DSB
simply does not generate a large enough signal to prevent the undamaged nucleus from dividing. To test this,
we increased the double strand break–induced damage
signal by either incorporating a second HO cut site in
the genome, or by adding a mutation that alters the
processing of the double strand break. Previous studies
have shown that haploid yeast cells are able to adapt to
the presence of an unrepaired single DSB and progress
through the cell cycle after a delay of several hours
(Sandell and Zakian, 1993; Toczyski et al., 1997; Lee et
al., 1998), but this adaptation can be eliminated either
by introducing a second DSB or by deleting YKU70 (Lee
et al., 1998), suggesting that both alterations function to
measurably increase the strength of the damage signal.
YKU70 encodes one of two Ku protein subunits that bind
and protect the exposed DNA ends from exonuclease
degradation. The yku70⌬ mutation results in an increased rate of 5⬘ to 3⬘ DNA degradation at the HO cut
site (Lee et al., 1998). When the heterokaryon experiment
was repeated with strains bearing either of these alterations, we found that neither a second HO cut site in
the susceptible nucleus nor presence of the yku70 deletion in both parents affected the outcome. In both cases,
greater than 90% of zygotes exhibited mononuclear division, indicating that the increased DNA damage signal
was still not able to arrest the undamaged nucleus (Figure 3a). These data are most consistent with the interpretation that the signal that prevents entry into anaphase
is nondiffusible.
We noted in the heterokaryon experiments that although the presence of the damaged nucleus did not
prevent the undamaged nucleus from dividing, it did
delay the onset of nuclear division. We compared the
percentage of zygotes undergoing nuclear division at 2
hr following release from the hydroxyurea arrest, in the
presence or absence of DNA damage in one nucleus
(Figure 3b). Without induced damage, 85.1% of largebudded zygotes were undergoing nuclear division,
whereas with induced damage only 65.1% of large budded zygotes were in nuclear division at the same time
point. This delay was eliminated when both parents bore
rad9⌬ mutations (Figure 3b), indicating that the delay is
DNA damage dependent. If this delay were a manifestation of a diffusible DNA damage checkpoint signal, then
we might expect that the delay would be greater if the
DNA damage signal was increased. Interestingly, neither
the addition of a second HO cut site nor the deletion of
YKU70 changed the relative kinetics of division with and
without DNA damage (Figure 3b).
Discussion
By inducing a single DSB into one nucleus of a binucleate heterokaryon, we have shown that unrepaired chromosomal damage in one nucleus does not prevent the
division of the undamaged nucleus present in the same
cytoplasm. The arrest of mitosis in the nucleus suffering
the HO-induced DSB arrest depends on the RAD9-mediated DNA damage checkpoint. We conclude from these
results that the DNA damage response in budding yeast
is nuclear limited.
Although several proteins involved in the signaling of
DNA damage have been identified, many aspects of the
process remain undefined. First, it is not known which
proteins directly detect DNA damage, or even if the
primary signal is the break itself, the single-stranded
DNA produced by resecting the ends by a 5⬘ to 3⬘ exonuclease, or even the oligonucleotide products liberated
by this resection. Second, it is not yet clear if all of the
important downstream targets of the signal have been
identified. Our results appear to rule out the possibility
that the primary signal is diffusible, or that the principal
means of signaling G2/M arrest would be through the
induction of a damage-induced mRNA, whose protein
product would not be nuclear limited.
We imagine several possible explanations of the nuclear-limited nature of the DNA damage response. In
the first, the proteins responsible for the cell cycle arrest
A Nuclear-Limited DNA Damage Checkpoint Signal
491
are themselves nuclear limited. Pds1p, which is an inhibitor of the anaphase promoting complex (Cohen-Fix and
Koshland, 1999), is localized to the nucleus (Cohen-Fix
et al., 1996) and it is possible that neither Pds1p nor
the protein kinases involved in its phosphorylation in
response to DNA damage (Cohen-Fix and Koshland,
1997) exit the nucleus once they have originally been
imported from the cytoplasm. Thus, the mitosis inhibiting activity of Pds1p would be limited to the damaged
nucleus. In a second possibility, regulated protein localization plays in the arrest after DNA damage. This could
involve the active export of a key cell cycle regulator
from the nucleus, such as a cyclin. In animal cells, cyclin
B1 localization to the nucleus is controlled by a balance
between nuclear import and nuclear export (Yang et al.,
1998), and regulation of nuclear translocation of cyclin
B1 is important in the DNA damage checkpoint (Toyoshima et al., 1998). In fission yeast, a key step in cell
cycle control involves the relocalization of the Cdc25
phosphatase from the cytoplasm to the nucleus, an
event that appears to be prevented after DNA damage,
in response to checkpoint protein phosphorylation
(Furnari et al., 1999). Exclusion of a mitotic activator from
the damaged nucleus would not affect the undamaged
nucleus. A third possibility is that the strength of the
DNA damage signal in our experiment was insufficient
to arrest the undamaged nucleus. We believe we have
ruled out this possibility by carrying out the same experiment with either yku70⌬ strains or a strain with a second
HO cut site, without a change in the outcome.
In addition to the nuclear-limited arrest of mitosis, we
did observe a delay in the time at which the undamaged
nucleus entered mitosis (Figure 3b). This could be explained by the induction of a secondary signal in the
damaged nucleus that is diffusible through the cytoplasm, such as a DNA damage–induced mRNA. Another
possibility is that a second checkpoint signal, not directly related to DNA damage, is created in our experiment. One such signal could be the checkpoint associated with monitoring the completion of S phase.
Induction of DSB damage during zygote formation is
expected to happen in G1, soon after the cells have
mated. In the MAT␣ haploid parent cells, induction of a
DSB in G1 leads to a delay in S phase entry of about 1
hr (J.D., unpublished observation). In the heterokaryon
experiments reported here, the temporary delay in progression of the damaged nucleus through S phase, and
the transient activation of the replication checkpoint,
could explain the delay we observed in the division of
the undamaged nucleus. An alternative explanation
could be based on the recent observation that the DNA
damage checkpoint bifurcates downstream of Mec1p
and that activation of only Rad53p or only Chk1p leads
to a delay in cell cycle progression (Gardner et al., 1999;
Sanchez et al., 1999). If only one of these pathways or
its targets is transmittable to the undamaged nucleus,
it could cause a delay in the division of the undamaged
nucleus. Further experiments are needed to address
this question. We also do not exclude the possibility
that an arrest of the undamaged nucleus could be briefly
established but cannot be maintained by damage occurring in another nucleus.
We have shown that DNA damage in one nucleus
prevents the division of only that nucleus in a yeast
heterokaryon. This result is the opposite of a what was
observed in recent experiments in mammalian cells (Rieder and Cole, 1998), in which one of two nuclei in a
binucleate cell was damaged by laser irradiation, leading
to a permanent arrest of both nuclei in interphase. However, the interpretation of this experiment is complicated
because it was not possible to demonstrate either that
DNA damage was the cause of the observed arrest,
or that the arrest was mediated by the DNA damage
checkpoint, both of which we dealt with directly in the
yeast experiments.
Aside from the experimental differences, there are
other differences in the way mammalian cells and budding yeast carry out mitosis, which could explain the
differences in the way heterokaryons respond to DNA
damage. For example, yeast cells have a closed mitosis,
in which chromosome segregation takes place within
an intact nuclear envelope, whereas animal cells have
an open mitosis in which the nuclear envelope breaks
down. Although this would seem to be a significant difference bearing on the experiments presented here, we
note that the damage and arrest in mammalian cells take
place at a time in the cycle when the nuclear envelope is
cytologically intact. Another difference between mammalian cells and budding yeast is the potentially different
target points at which the damage checkpoint impinges
on the cell cycle machinery. In mammalian cells and
fission yeast, activation of the checkpoint leads to an
inhibitory phosphorylation of Cdk1 and maintenance of
cytoplasmic localization of cyclin B1, arresting the cells
with low mitotic kinase activity prior to mitosis. In budding yeast, phosphorylation of this inhibitory site does
not play the central role in the checkpoint response, and
cells arrest with high mitotic kinase activity, although a
regulated G2/M transition cannot be excluded (Gardner
et al., 1999; Sanchez et al., 1999).
The idea that a checkpoint signal could be limited to
the nucleus is supported by experiments of Dillin and
Rine (1998) on the checkpoint responses associated
with failure of DNA replication. They found that a nucleus
in which the DNA did not replicate due to a mutation in
a subunit of the origin recognition complex failed to
prevent division of another nucleus in the same cytoplasm. Although their assay lacked the resolution to
establish this point, the results suggest that the DNA
replication checkpoint might also be nuclear limited. It
will be of great interest to determine whether the DNA
replication signals and the DNA damage arrest signals
that we have analyzed here both depend on the same
nuclear-limited mechanisms.
Experimental Procedures
Standard yeast media and culture conditions were used (Adams et
al., 1997). Strains were derived from JKM179 (Lee et al., 1998) or
DBY4974. Deletions were done using the one-step replacement
method (Wach et al., 1994) and were verified using PCR. To introduce the second HO cut site at the URA3 locus on chromosome V,
an NdeI-NheI fragment of plasmid pNSU134 containing a 117 bp
HO cleavage site in a URA3 fragment was used for integration (Lee et
al., 1998). Mating experiments were performed by mixing overnight
YPGal cultures of the respective strains on a YPGal plate for 3 hr
at room temperature, then scraping the cells from the plate and
spreading them on YPGal ⫹ 100 mM hydroxyurea at room temperature. Zygotes were arrested on hydroxyurea plates for 4.5 hr, then
washed, and inoculated into liquid YPGal. Aliquots were fixed hourly
by addition of formaldehyde for up to 3 hr after release from the
hydroxyurea arrest. For the experiment without hydroxyurea, cells
were mated on YPGal plates for 7.5 hr and were fixed at this time
point and at 8.5 hr as above.
Standard immunofluorescence techniques (Adams et al., 1997)
Molecular Cell
492
were used to localize microtubules and GFP-lacI using YOL1/34
(Sera-Tec) and anti-GFP antibodies (gift of J. Kahana and P. Silver).
Zygotes were counted and classified by phenotypic category. Despite the presence of the kar1-⌬15 mutation, the nuclei did undergo
karyogamy in some of the zygotes (⬍5%–10%). These zygotes were
not included in the data shown on Figure 3. In each experiment,
between 100 and 300 zygotes were counted, and each experiment
was performed at least twice.
The percentage of asynchronous divisions in Figure 3a was determined by adding the number of zygotes with only one nucleus undergoing mitosis at all time points and dividing this number by the total
number of mitotic zygotes at all time points, averaged among several
independent experiments. To represent the kinetics of division in
Figure 3b, we calculated the percentage of zygotes undergoing
nuclear division by dividing the number of zygotes in which nuclear
division was occurring (mononuclear and binuclear) at the 2 hr time
point with the total number of zygotes at that time point.
Rao, P.N., and Johnson, R.T. (1970). Mammalian cell fusion: studies
on the regulation of DNA synthesis and mitosis. Nature 225, 159–164.
Acknowledgments
Straight, A.F., Belmont, A.S., Robinett, C.C., and Murray, A.W. (1996).
GFP tagging of budding yeast chromosomes reveals that proteinprotein interactions can mediate sister chromatid cohesion. Curr.
Biol. 6, 1599–1608.
We thank Mark Rose, Sue Biggins, and Andrew Murray for providing
constructs. J. D. is supported by NIH NRSA fellowship CA0930222. S. E. L. is a postdoctoral fellow of the Leukemia Society of
America. Support for this work was from Department of Energy
grant DOE 99ER62729 to J. E. H. J. E. H. is a member of the Keck
Institute for Cellular Visualization.
Received February 2, 2000; revised June 9, 2000.
References
Adams, A., Gottschling, D.E., Kaiser, C.A., and Stearns, T. (1997).
Methods in Yeast Genetics, 1997 Edition (Plainview, NY: Cold Spring
Harbor Laboratory Press).
Amon, A., Surana, U., Muroff, I., and Nasmyth, K. (1992). Regulation
of p34CDC28 tyrosine phosphorylation is not required for entry into
mitosis in S. cerevisiae. Nature 355, 368–371.
Chan, T.A., Hermeking, H., Lengauer, C., Kinzler, K.W., and Vogelstein, B. (1999). 14-3-3Sigma is required to prevent mitotic catastrophe after DNA damage. Nature 401, 616–620.
Cohen-Fix, O., and Koshland, D. (1997). The anaphase inhibitor of
Saccharomyces cerevisiae Pds1p is a target of the DNA damage
checkpoint pathway. Proc. Natl. Acad. Sci. USA 94, 14361–14366.
Cohen-Fix, O., and Koshland, D. (1999). Pds1p of budding yeast
has dual roles: inhibition of anaphase initiation and regulation of
mitotic exit. Genes Dev. 13, 1950–1959.
Cohen-Fix, O., Peters, J.M., Kirschner, M.W., and Koshland, D.
(1996). Anaphase initiation in Saccharomyces cerevisiae is controlled by the APC-dependent degradation of the anaphase inhibitor
Pds1p. Genes Dev. 10, 3081–3093.
Connolly, B., White, C.I., and Haber, J.E. (1988). Physical monitoring
of mating type switching in Saccharomyces cerevisiae. Mol. Cell.
Biol. 8, 2342–2349.
Dillin, A., and Rine, J. (1998). Roles for ORC in M phase and S phase.
Science 279, 1733–1737.
Furnari, B., Blasina, A., Boddy, M.N., McGowan, C.H., and Russell,
P. (1999). Cdc25 inhibited in vivo and in vitro by checkpoint kinases
Cds1 and Chk1. Mol. Biol. Cell 10, 833–845.
Gardner, R., Putnam, C.W., and Weinert, T. (1999). RAD53, DUN1
and PDS1 define two parallel G2/M checkpoint pathways in budding
yeast. EMBO J. 18, 3173–3185.
Jin, P., Hardy, S., and Morgan, D.O. (1998). Nuclear localization of
cyclin B1 controls mitotic entry after DNA damage. J. Cell Biol. 141,
875–885.
Kostriken, R., Strathern, J.N., Klar, A.J., Hicks, J.B., and Heffron,
F. (1983). A site-specific endonuclease essential for mating-type
switching in Saccharomyces cerevisiae. Cell 35, 167–174.
Lee, S.E., Moore, J.K., Holmes, A., Umezu, K., Kolodner, R.D., and
Haber, J.E. (1998). Saccharomyces Ku70, Mre11/Rad50 and RPA
proteins regulate adaptation to G2/M arrest after DNA damage. Cell
94, 399–409.
Rhind, N., and Russell, P. (1998). Mitotic DNA damage and replication checkpoints in yeast. Curr. Opin. Cell Biol. 10, 749–758.
Rieder, C.L., and Cole, R.W. (1998). Entry into mitosis in vertebrate
somatic cells is guarded by a chromosome damage checkpoint that
reverses the cell cycle when triggered during early but not late
prophase. J. Cell Biol. 142, 1013–1022.
Sanchez, Y., Bachant, J., Wang, H., Hu, F., Liu, D., Tetzlaff, M., and
Elledge, S.J. (1999). Control of the DNA damage checkpoint by Chk1
and Rad53 protein kinases through distinct mechanisms. Science
286, 1166–1171.
Sandell, L.L., and Zakian, V.A. (1993). Loss of a yeast telomere:
arrest, recovery, and chromosome loss. Cell 75, 729–739.
Sorger, P.K., and Murray, A.W. (1992). S-phase feedback control in
budding yeast independent of tyrosine phosphorylation of
p34cdc28. Nature 355, 365–368.
Tinker-Kulberg, R.L., and Morgan, D.O. (1999). Pds1 and Esp1 control both anaphase and mitotic exit in normal cells and after DNA
damage. Genes Dev. 13, 1936–1949.
Toczyski, D.P., Galgoczy, D.J., and Hartwell, L.H. (1997). CDC5 and
CKII control adaptation to the yeast DNA damage checkpoint. Cell
90, 1097–1106.
Toyoshima, F., Moriguchi, T., Wada, A., Fukuda, M., and Nishida,
E. (1998). Nuclear export of cyclin B1 and its possible role in the
DNA damage-induced G2 checkpoint. EMBO J. 17, 2728–2735.
Vallen, E.A., Hiller, M.A., Scherson, T.Y., and Rose, M.D. (1992).
Separate domains of KAR1 mediate distinct functions in mitosis and
nuclear fusion. J. Cell Biol. 117, 1277–1287.
Wach, A., Brachat, A., Pohlmann, R., and Philippsen, P. (1994). New
heterologous modules for classical or PCR-based gene disruptions
in Saccharomyces cerevisiae. Yeast 10, 1793–1808.
Weinert, T. (1998). DNA damage checkpoints update: getting molecular. Curr. Opin. Genet. Dev. 8, 185–193.
Yang, J., Bardes, E. S., Moore, J. D., Brennan, J., Powers, M.A.,
and Kornbluth, S. (1998). Control of cyclin B1 localization through
regulated binding of the nuclear export factor CRM1. Genes Dev.
12, 2131–2143.