The EMBO Journal Vol.17 No.19 pp.5525–5528, 1998
EMBO MEMBER’S REVIEW
DNA damage checkpoint in budding yeast
Maria Pia Longhese, Marco Foiani,
Marco Muzi-Falconi, Giovanna Lucchini and
Paolo Plevani
1
Dipartimento di Genetica e Biologia dei Microrganismi,
Via Celoria 26, 20133 Milano, Italy
1
Corresponding author
e-mail: plevanip@imiucca.csi.unimi.it
Eukaryotic cells have evolved a network of control
mechanisms, known as checkpoints, which coordinate
cell-cycle progression in response to internal and
external cues. The yeast Saccharomyces cerevisiae has
been invaluable in dissecting genetically the DNA
damage checkpoint pathway. Recent results on post-
translational modifications and protein–protein inter-
actions of some key factors provide new insights into
the architecture of checkpoint protein complexes and
their order of function.
Keywords: budding yeast/cell cycle/checkpoints/DNA
damage/DNA replication
Note: Throughout this review, the Schizosaccharomyces
pombe genes are indicated with a superscript ‘Sp’.
Introduction
Checkpoints are genetically controlled surveillance
mechanisms that ensure the interdependency of cell-cycle
events (for reviews see Hartwell and Weinert, 1989;
Murray, 1992; Elledge, 1996; Paulovich et al., 1997;
Weinert, 1998). Both intrinsic and extrinsic checkpoints
can be envisaged: intrinsic mechanisms act in each cell
cycle under unperturbed conditions to ensure the proper
temporal order of events, while extrinsic mechanisms are
activated only when alterations are detected. The DNA
damage checkpoint represents the subset of extrinsic
surveillance mechanisms that are triggered in response to
DNA insults. The activation of this pathway leads to the
induction of a set of genes required for the resolution of
the damage (Aboussekhra et al., 1996; Kiser and Weinert,
1996) and to a temporary inhibition of cell-cycle progres-
sion, in order to prevent replication and segregation of
damaged DNA. Failure to respond properly to DNA
alterations can lead to increased genomic instability, which
is one of the most prominent hallmarks of cancer cells
(Hartwell and Kastan, 1994).
The cell cycle is transiently arrested at different stages
depending on the phase at which DNA alterations occur
(G
1
, S and G
2
). Three responses have been characterized
in budding yeast, which are known as the G
1
/S, intra-S
and G
2
/M DNA damage checkpoints. It has become clear
© Oxford University Press 5525
that even though the players might be different, the general
mechanism underlaying the DNA damage checkpoint
response is the same in the different phases of the cell cycle.
The recognition of DNA damage
The first step in the DNA damage and replication check-
point pathways is the recognition of particular DNA
structures or alterations. One of the crucial questions in
this field is to define the signals that activate the checkpoint.
In fact, genotoxic agents cause many types of primary
lesions that can be converted to secondary lesions during
replication of a damaged template. DNA replication across
a single-strand nick is likely to cause replication fork
collapse and production of a double-strand break, while
single-strand gaps can be generated if replication is arrested
in front of a covalently modified base, and DNA synthesis
resumes downstream of the damage.
It is thus evident that eukaryotic cells must have evolved
a complex network of systems that allow them to respond
to this variety of DNA perturbations. All of these various
lesions could be directly recognized by a number of
checkpoint proteins, either alone or in specialized subcom-
plexes, or they could be processed to a common inter-
mediate that triggers the checkpoint activation. Since the
checkpoint response might also be influenced by the stage
of the cell cycle at which the damage occurs, multiple
sensors probably recognize the signals in specific phases
of the cell cycle.
Most of the key players in the checkpoint response in
Saccharomyces cerevisiae have been identified and have
structural and functional equivalents in Schizosaccharo-
myces pombe and in human cells, thus providing an
important contribution to the understanding of checkpoint
controls in all eukaryotes (Table I).
The emphasis within the field is currently on defining the
biochemical properties, and the functional and structural
interactions among checkpoint proteins. In budding yeast,
the RAD9, RAD17, RAD24, MEC3 and DDC1 gene
products are specifically required for a proper DNA
damage response, and are proposed to act at an early step
of damage recognition at any stage of the cell cycle.
Conversely, the DNA replication proteins Pol ε and Rfc5
appear to sense both replication blocks and DNA damage
during DNA synthesis (Navas et al., 1996; Sugimoto
et al., 1997). Pol ε is a replicative DNA polymerase,
while Rfc5 is a subunit of the Replication Factor C (RF-C)
complex which, after binding to template-primer junctions,
loads the proliferating cell nuclear antigen (PCNA) clamp
onto DNA, thereby recruiting DNA polymerases to the
site of DNA replication. The RF-C subunits are structurally
related to each other and to the RAD24 gene product.
Genetic and biochemical studies indicate functional and
physical interactions between RF-C and Rad24 (Lydall

M.P.Longhese et al.
Table I. DNA damage checkpoint genes are evolutionarily conserved
Function S.cerevisiae S.pombe Human
gene gene gene
Protein kinase (?) MEC1 rad3 ATM, ATR
RAD9 rhp9/crb2
MEC3
DDC1 rad9 HRAD9
RAD17 rad1 HRAD1
RAD24 rad17 HRAD17
Protein kinase RAD53 cds1
DNA polymerase ε POL2 cdc20 POLε
DPB11 cut5
Replication Factor C RFC5 HRFC38
Replication Factor A RFA1 rad11 HRPA1
DNA primase PRI1 PRIM1
PDS1 cut2
and Weinert, 1997; Shimomura et al., 1998), suggesting
that the function of RF-C in the checkpoints may not be
restricted to S phase. Based on the role of RF-C in DNA
replication, it is tempting to speculate that the RF-C–
Rad24 complex may play a role in loading checkpoint or
repair proteins onto damaged DNA.
The observation that RAD17, RAD24, MEC3 and DDC1
belong to the same epistasis group, while RAD9 is in a
group on its own (Lydall and Weinert, 1995; Longhese
et al., 1997), suggests that Rad9 and the Rad24 group of
proteins act in different branches of the checkpoint path-
way. Indeed, RAD9 and RAD24 have different effects on
the accumulation of single-strand DNA at telomeres in
cdc13 mutants, indicating that correct balance of their
activities is required for proper processing of at least some
peculiar type of lesion (Lydall and Weinert, 1996). It has
been shown recently that Mec3 and Ddc1 physically
interact in vivo, and that Rad17 is needed for this inter-
action (Paciotti et al., 1998). A similar complex also exists
in S.pombe, where Rad1
sp
(homologous to Rad17) interacts
with Hus1
sp
, and Rad9
sp
(homologous to Ddc1) is required
for complex formation (Kostrub et al., 1998). This indi-
cates that the structural organization of at least some
checkpoint protein complexes has been conserved during
evolution.
Ddc1 is phosphorylated in response to DNA damage,
and its modification correlates with DNA damage check-
point activation. Damage-induced Ddc1 phosphorylation
is totally dependent on a functional MEC1 gene and also
partially requires the Rad24 group of proteins (Longhese
et al., 1997; Paciotti et al., 1998). Mec1 is an essential
checkpoint factor which has been evolutionarily conserved
in eukaryotes. It belongs to the PI-3 kinase family, which
includes, among others, Tel1, Rad3
sp
, mammalian ATM
and ATR, and known protein kinases, such as the DNA-
dependent protein kinase (DNA-PK) (reviewed in Elledge
1996; Weinert, 1998). Rad3
sp
has an associated protein
kinase activity (Bentley et al., 1996) and, although a direct
biochemical demonstration is still lacking, Mec1 is also
likely to act as a protein kinase.
The finding that Ddc1 phosphorylation depends on
Mec1 and on the proteins encoded by genes of the
RAD24 epistasis group suggests that Mec1 may participate,
together with Rad24, Rad17, Mec3 and Ddc1, at an early
step of the DNA damage recognition process (Figure 1).
The observation that Rad9 is not required for Ddc1
5526
phosphorylation supports the notion that Rad9 might act
in a different branch of the DNA damage response
pathway. However, it has been recently found that Rad9
is phosphorylated after DNA damage and that this modi-
fication depends on Mec1, Tel1 and the Rad24 group of
proteins (A.Emili, personal communication; N.Lowndes,
personal communication; Sun et al., 1998). Rad9 phos-
phorylation appears to be physiologically relevant since it
correlates with checkpoint activation, and phosphorylated
Rad9 preferentially interacts with Rad53 (Emili, personal
communication; Sun et al., 1998). Together with the above
reported observations, the fact that Rad9 and Ddc1 are
required with Mec1 to phosphorylate Rad53 in response
to DNA damage (Sanchez et al., 1996; Sun et al., 1996;
Paciotti et al., 1998), and that Mec1 is necessary for Rad9
and Ddc1 phosphorylation, suggests that the checkpoint
response may be more complex than a simple linear
pathway.
Similar to what has been proposed for Rad3
sp
(Bentley
et al., 1996; Carr, 1997), we suggest that Mec1 might be
able to recognize specific DNA or protein–DNA structures.
This function could be influenced by interaction with
checkpoint proteins like Rad9 and the Rad24 group, which
could also confer a target specificity to Mec1 (Figure 1).
The last hypothesis is supported by the finding that Mec1
and Rad9, but not the Rad24 group of proteins, are
required for DNA damage-dependent phosphorylation of
Pds1, an inhibitor of the metaphase to anaphase transition
which is also involved in G
2
/M DNA damage checkpoint
(Cohen-Fix and Koshland 1997; Paciotti et al., 1998).
Intrinsic DNA damage checkpoint
Another interesting aspect is the possible existence of an
intrinsic DNA damage signal in a normal cell cycle, in
the absence of external cues. Indeed, the replication
process by itself can be genotoxic. Replication errors
occur stochastically during nucleotide incorporation, and
structural intermediates normally arising during unper-
turbed DNA replication, such as unwound DNA and
single-stranded regions, are more fragile than double-
stranded DNA organized in a chromatin structure. In
addition, single-strand and double-strand breaks are
generated by the nicking–closing activity of DNA topo-
isomerases, which are required to remove torsional stress
ahead of the replication forks.
Since the DNA replication process generates DNA
structures that may be similar to some of those produced
by DNA damage, it is important to define whether the
checkpoint can be activated by DNA synthesis per se,or
whether it becomes activated only when DNA is damaged
or replication is altered. It has been found recently that
Ddc1 is periodically phosphorylated during unperturbed
cell cycles concomitantly with entry and progression
through S phase, and that this modification is still depend-
ent on Mec1 and Mec3 (Longhese et al., 1997; Paciotti
et al., 1998). Furthermore, Ddc1 and Mec3 are also
physically associated in the absence of DNA damage,
which leads us to the hypothesis that a ‘guardian complex’
may constantly monitor the integrity of the genome.
From these data we propose the existence of an intrinsic
checkpoint signal during unperturbed DNA replication. In
S phase the checkpoint response may be in a pre-activated