Characterisation of
Schizosaccharomyces pombe rad31
, a UBA-related gene required for DNA damage tolerance
Characterisation of Schizosaccharomyces pombe rad31 , a UBA-related gene required for DNA damage tolerance
Maryam
Shayeghi
,
Claudette L.
Doe
,
Manoochehr
Tavassoli
and
Felicity Z.
Watts*
Biochemistry, School of Biological Sciences, University of Sussex, Falmer,
Brighton
BN1 9QG,
UK
Received December 16, 1996;
Revised and Accepted February 3, 1997
DDBJ/EMBL/GenBank accession no. Y08805
ABSTRACT
The fission yeast
rad31-1
mutant is sensitive to both UV and ionising radiation and exhibits a growth
defect at 35
o
C. In addition, the mutant displays defects in cell morphology and nuclear
division at 26
o
C which are exaggerated at 35
o
C. We have cloned the
rad31
gene and have shown that it is not essential for viability, although cells
containing a disrupted
rad31
gene grow slowly. The null allele has similar cell and nuclear morphologies to
the original allele and displays an extremely high frequency of loss of
minichromosomes.
rad31
is not required for either the S/M or G2/M checkpoint, however double mutant
analysis indicates that
rad31
acts in a process which is defective in the checkpoint
rad
mutants and which involves
hus5
. Sequence analysis indicates that
rad31
encodes a protein which is related to ubiquitin activating proteins and more
particularly to an ORF in
Saccharomyces cerevisiae
and to the
Arabidopsis thaliana AXR1
and human APP-BP1 genes. We have isolated the
S.cerevisiae
sequence, which we have named
RHC31
(
r
ad31
homologue in
S.
c
erevisiae
),
since we show that it can complement the slow growth phenotype and radiation
sensitivity of
S.pombe rad31
.
INTRODUCTION
Co-ordination of cell cycle events is essential if cells are to undergo
successful mitoses and pass on faithful copies of their genetic material to
daughter cells. Thus, controls exist which ensure that mitosis is dependent on
the completion of DNA replication and repair of DNA damage (
1
-
3
). This dependency of mitosis on DNA structure has been investigated through the
analysis of checkpoint mutants in both
Saccharomyces cerevisiae
and
Schizosaccharomyces pombe
(
1
-
4
). Inhibition of DNA synthesis brought about by inhibitors such as hydroxyurea
(HU) or cell cycle (
cdc
) mutants evokes the S/M checkpoint which, in
S.pombe
, arrests the cell cycle through tyrosine phosphorylation of the p34
cdc2
protein (
5
,
6
). The presence of DNA damage also results in cell cycle arrest (
2
,
3
).
In
S.pombe
, the DNA damage checkpoint requires the action of
chk1
(
4
,
7
),
rad24
and
rad25
(
8
). It also requires the checkpoint
rad
genes which include
rad1
,
rad3
,
rad9
,
rad17
,
rad26
and
hus1
. These six genes are also needed for the replication checkpoint (
2
-
4
,
9
). Exposure of the corresponding checkpoint mutants to either DNA synthesis
inhibitors or DNA damaging agents results in mitotic catastrophe, as observed
by a `
cut'
phenotype (
10
). In addition to their requirement for the S/M and damage checkpoints, the
checkpoint
rad
genes have been proposed to have a further function which is required for the
recovery from irradiation or S phase arrest (
4
,
9
). This is based on the observation that the checkpoint
rad
mutants are substantially more sensitive to radiation than mutants defective
solely in the checkpoint function, such as
chk1
(
4
,
7
). In addition to the checkpoint
rad
genes, this recovery process or damage tolerance mechanism has been shown to
require the
hus5
gene (
11
).
hus5
mutants are moderately sensitive to radiation and HU, but are not defective in
either the replication or damage checkpoints. The
hus5
gene encodes a ubiquitin conjugating enzyme, deletion of which results in
microcolonies which display multiple mitotic abnormalities.
In the present study we have analysed a new radiation and HU sensitive mutant of
S.pombe
,
rad31
, which displays a `
cut'
phenotype in a small percentage of cells, even in the absence of radiation, in
a manner reminiscent of
hus5.62
(
11
). To further investigate the role of the
rad31
gene, we cloned it by complementation of the UV sensitive and slow growth
phenotypes. The gene product has homology to the
Arabidopsis AXR1
protein (
12
), the human APP-BP1 protein (
13
), an
S.cerevisiae
open reading frame (ORF), as well as the N-terminal region of ubiquitin activating proteins (UBAs) (e.g.
14
). We also demonstrate that the
S.cerevisiae
ORF complements the slow growth and radiation sensitive phenotypes of the
rad31.d
null allele. The Rad31 protein appears to be a novel member of a conserved
family of UBA-related proteins which do not contain the conserved cysteine required for
ubiquitin activation. Disruption of the gene is not lethal, however the cells
grow slowly with a high proportion of the cells containing condensed
chromosomes. The disruptants have a very high loss of minichromosomes.
Epistasis analysis indicates that
rad31
functions in a process which is defective in the checkpoint
rad
mutants, and more specifically, in a process requiring
hus5
. The potential role of the Rad31 protein in damage tolerance/recovery from
radiation is discussed.
MATERIALS AND METHODS
Strains and growth media
Schizosaccharomyces pombe
strains used
are listed in
Table
1
.
Escherichia coli
strains, procedures and growth media used for the routine culture and
maintenance of
S.pombe
strains were as described previously (
15
).
rad31::ura4, Ch16(ade6-m216)
,
, ade6-m210, ura4-D18, h
-
sp.X34
rad31::ura4, rad8::ura4, ade6-704, leu1-32 h
-
sp.X35
rad31::ura4, hhp2::leu1, ade6-704, h
+
sp.X36
rad31::ura4, rad17::ura4, ade6-704, leu1-32 h
-
Cloning of the
rad31
gene and cDNA from
S.pombe
and the
RHC31
gene from
S.cerevisiae
The pURSP1 and pURSP2
S.pombe
genomic libraries (
16
) were used to transform a
rad31-1 ura4-D18
strain (sp.134) to uracil prototrophy. Two selection procedures were used to
isolate the
rad31
gene. In one, the transformation plates were incubated for 12 h at 26oC and then shifted to 35oC for 5 days. In the other, 20 000
ura
+
transformants, obtained following incubation of the plates at 26oC for 5 days, were subjected to three rounds of selection following
exposure to increased doses of UV irradiation as described in our previous work
(
15
). Both procedures resulted in the isolation of an identical plasmid, pMS1. The
rad31
cDNA was amplified by PCR using an
S.pombe
cDNA library (
17
) with the primers MS14 (GAAGAGGCCATGGGGAATCACA) and MS15
(TAGAAGGATCCTAAATAACAGAATACC). The amplified product was cloned into a T
vector.
The
RHC31
gene was amplified using
S.cerevisiae
(strain Y153) genomic DNA by PCR using primers Sc31a (ATCAGTAACCATGGATATG) and
Sc31b (AGGTATGTAGATATCTAAA) and cloned into pREP1N (a modified version of pREP1
(
18
) in which the
Nde
I site has been replaced with an
Nco
I site).
DNA sequence analysis
The 3.5 kb
Pst
I-
Sph
I fragment containing the
rad31
gene was cloned into M13 in both orientations and one clone was subjected to
exonuclease III deletions as described by Henikoff (
19
). Overlapping deletions were sequenced and aligned by computer analysis. The
opposite strand was sequenced using synthetic oligonucleotide primers. The four
introns were confirmed by sequence analysis, using synthetic oligonucleotides,
of the
rad31
cDNA. The sequence has been submitted to the EMBL database and has the
accession number Y08805.
Percentage of aberrant morphologies in cultures of
rad31
and
hus5
Strain
Ennucleates
Stretched/fragmented
Condensed
cdc
cut
chromatin
chromosomes
rad31-1
25oC
2
6
7.5
18
2.5
35oC 3 h
2.5
7.5
15
20
4
rad31.d
25oC
1
3.5
10
24
2.5
35oC 3 h
3.5
7
12.5
28
5.5
hus5-62
25oC
0.5
1.5
2.5
11
0.5
35oC 3 h
1.5
5.5
12
20
3
% cells with aberrant morphologies was estimated as described by Al-Khodairy
et al
. (11).
Radiation survival and minichromosome loss assay
Gamma irradiation was carried out using a
137
Cs gamma source with a dose rate of 12 Gy/min. UV irradiation was carried out
directly on freshly plated cells using a Stratagene `Stratalinker'. The
minichromosome loss assay of Niwa
et al
. (
20
) was used as described in our previous work (
21
).
RESULTS
Characterisation of the
rad31-1
mutant phenotype
The
rad31-1
mutant was isolated during a search for new ionising radiation sensitive
mutants undertaken with a view to identifying novel genes involved in
recombination (
22
). The radiation sensitive phenotypes of the
rad31-1
mutant were investigated following exposure to both UV and gamma irradiation
(Fig.
1
a and b). At high doses of both UV and gamma radiation
rad31-1
displays similar sensitivities to known recombination mutants, such as
rad32.d
(
22
) and is much less sensitive to both UV and ionising radiation than the
checkpoint
rad
mutant
rad9
(
23
). It is also much less sensitive to UV radiation than a nucleotide excision
repair mutant, such as
rad15
(
15
). However, at low doses of radiation (<600 Gy or 100 J/m
2
)
rad31-1
is relatively insensitive to both UV and ionising radiation, with sensitivities
close to wild-type levels.
During routine handling, it was noted that cultures of
rad31-1
mutant cells
grew slowly (with 50% greater doubling time at 25oC and 66% greater at 35oC than wild-type cells) and stationary phase cultures contained a high
proportion of inviable cells (5-6-fold that observed with wild-type cultures). Cell and nuclear morphologies of
rad31-1
were therefore compared to wild-type. DAPI staining and microscopic analysis indicated that
rad31-1
cells display an aberrant cell morphology and an apparent defect in chromosome
segregation (Fig.
2
). Cells were generally elongated and irregularly shaped, displaying a `
cut'
phenotype in 2-3% cells at 25oC (Table
2
), with the defects being more severe at 35oC. This aberrant morphology, seen even in the absence of DNA damage,
suggests that the
rad31
gene is likely to be required for a process in addition to, or other than, DNA
repair. This phenotype is reminiscent of the `constant cut' phenotype of
hus5.17
observed by Enoch
et al
. (
9
).
Cloning and sequencing of
rad31
Disruption of the
rad31
gene results in slow growth and mitotic defects
To determine whether the
rad31
gene is essential for viability two gene disruption constructs were created as
described in Materials and Methods (Fig.
3
A). We decided to create two constructs, since the first results in removal of 2
kb of DNA, but includes 1.2 kb of sequence downstream of the gene and hence may
affect an adjacent gene, while the other involves disruption of the ORF at amino acid 77. Cells containing either of the disruptions are viable with the
same mutant phenotype and the radiation and aberrant cell morphology phenotypes
of both alleles can be complemented by the
rad31
cDNA cloned in pREP41 (data not shown), indicating that all detectable
phenotypes result from loss of a single gene. Cultures derived from
rad31.d
ura
+
spores grew slowly compared to wild-type strains and accumulated cells with aberrant cell and nuclear
morphologies. DAPI staining of DNA indicated a number of differences between
rad31.d
and wild-type cells (Fig.
2
, Table
2
). Firstly, the DNA in many
rad31.d
cells appeared to be in a condensed state (10% at 25oC), unlike the DNA in wild-type cells. In a small proportion of cells (~1% at 25oC) three discrete DAPI stained entities were detected,
presumably representing the three
S.pombe
chromosomes. Secondly, 2.5% of cells at 25oC and 5.5% of the cells at 35oC displayed a `
cut'
phenotype with septation occurring prior to completion of nuclear division and
in some cases (1% at 25oC) there was evidence of stretched chromatin. Thirdly, at 35oC, 7% of the cells displayed spotty staining suggesting that the
chromatin was being fragmented. Extensive culturing of the
rad31.d
strain resulted in cells with extremely low cell viability, presumably due to
loss or fragmentation of the chromosomes (Fig.
4
a).
a
Data from Tavassoli
et al
. (22).
b
Data from Murray
et al
. (21).
rad31-deleted cells retain the DNA damage and replication checkpoints
To determine whether the `
cut'
phenotype observed in the absence of radiation was due to loss of the
replication checkpoint, exponentially growing cells were exposed to 20 mM HU
and sampled hourly. Unlike the checkpoint
rad
mutants,
rad31.d
cells are only slightly sensitive to HU (Fig.
4
b). Microscopic analysis of cell morphology after exposure to HU indicates that
there is no increase in the frequency of the `
cut'
phenotype in the first 6 h, and that cells are generally more elongated than
untreated
rad31.d
cells (data not shown), indicating that the replication checkpoint is still
intact. Exposure to 250 Gy ionising radiation also resulted in cell cycle
arrest and elongated cells (data not shown) indicating that the radiation
checkpoint is also functioning in
rad31.d
cells.
rad31-deleted strains show a very high incidence of loss of minichromosomes
The aberrant nuclear morphologies suggest that
rad31
is defective in chromosome segregation. To investigate this, the minichromosome-loss assay of Niwa
et al.
(
20
) was used. The
rad31.d
and
rad
+
strains containing the Ch16 minichromosome, sp.X31 and sp.1235 respectively,
were grown at 25 and 35oC for 10 generations and then plated at 25oC and scored for loss of Ch16 (Table
4
). The
rad31.d
strain showed a very high incidence of loss of the minichromosome, 9.9 and
13.4% loss per generation at 25 and 35.5oC respectively, compared to
rad
+
cells (0.019 and 0.029%).
a)
b)
Genetic interactions
A related gene exists in
S.cerevisiae
which can functionally complement the rad31D phenotype
Analysis of the sequence databases indicates that, apart from the UBA family of
proteins, there are three other sequences related to Rad31, namely the
AXR1
protein (
12
), the human APP-BP1 protein (
13
) and an ORF in
S.cerevisiae
identified through the genome sequencing project. The
S.cerevisiae
sequence [renamed here as
rad
homologue in
S.cerevisiae
(
RHC31
)] was isolated by PCR as described in Materials and Methods, and cloned into
the
S.pombe
expression plasmid pREP1N. The
RHC31
sequence was able to complement the slow growth and the radiation sensitive
phenotypes of
rad31.d
(Fig.
7
) indicating that
rad31
and
RHC31
are likely to be functional homologues.
DISCUSSION
In
S.pombe
the DNA damage and replication checkpoints have been shown to require the
functions of the checkpoint
rad
genes,
rad1
,
rad3
,
rad9
,
rad17
,
rad26
and
hus1
(
2
-
4
,
9
). The gene products are thought to act as a guardian complex and to signal to
the mitotic machinery to arrest mitosis. The proteins have also been proposed
to be required for a damage tolerance or recovery activity following the
blocking of replication in the presence of DNA damage (
9
). The basis of this damage tolerance is unknown, but the fact that it is
independent of the checkpoint functions has been demonstrated through analysis
of the
rad26.T12
allele which retains the damage checkpoint activity, but not the damage
tolerance function (
4
). In contrast, the
chk1
gene is required solely for the damage checkpoint function but not the damage
tolerance activity (
4
,
7
).
Although the nature of the damage tolerance process is unknown, it has been
shown to require a ubiquitin conjugating enzyme, Hus5 (
11
). Several recent reports on the human homologue to
hus5
suggest that it interacts with a range of proteins including Rad51 (
32
), three subunits of the centromere DNA-binding core complex, CBF3 (
33
) and the human papilloma type 16 E1 replication protein (
34
). This range of interactions may explain the pleiotropic nature of the
hus5
mutation which results in slow growth and defects in chromosome segregation and
mitosis.
We have characterised a new
S.pombe
mutant which displays a `constant cut' phenotype reminiscent of
hus5
, with low viability in stationary phase, indicating the
rad31
gene product is required even in the absence of radiation and in non-cycling cells. The mutant also has a very high rate of loss of
minichromosomes.
A single plasmid species (pMS1) was obtained independently by complementation of
the radiation sensitivity and slow growth phenotype at 35oC. Despite extensive culturing, it was not possible to integrate pMS1 into
the genome to formally prove that the gene is identical to the one where the
original mutation lies, so the possibility exists that pMS1 encodes a
suppressor. However, we believe that this is unlikely since the same plasmid
was isolated by complementation of two different phenotypes, and that the
phenotype of the genomic disruptions is similar to that of the
rad31-1
mutant.
Sequence analysis indicates that the gene encodes a novel member of a ubiquitin
activating protein-related family. Ubiquitin activating proteins (UBAs) are required for the
initial step in the selective targeting of proteins by the covalent attachment
of ubiquitin (reviewed in
35
). The UBAs are involved in activating ubiquitin through the formation of a high
energy thiol-ester linkage between a cysteine residue in the UBA and the C-terminal glycine residue of ubiquitin. The significance of the
homology of the Rad31 protein to UBAs is unknown, particularly since the Rad31
protein does not contain the conserved cysteine required for ubiquitin
activation in UBAs (
24
). It is interesting to note that the
S.pombe
and
S.cerevisiae
members are similar in length to one another and appear to be functional
homologues. The AXR1 and APP-BP1 proteins both contain, in addition to the sequences conserved in the
yeast proteins, an extra internal domain. In APP-BP1 this domain has been shown to be involved in the binding of the
Alzheimer's amyloid precursor protein (
13
), suggesting that the domain is required for specific targeting of the protein.
None of the four proteins contains the equivalent of the conserved cysteine
residue required for activation of ubiquitin. While the involvement of AXR1 and
APP-BP1 in ubiquitin-related processes has been speculated about (
12
,
13
), our data showing that
rad31
and
hus5
[a UBC, (
11
)] are required for a common process strongly suggest that the
rad31
member of this family is required for a ubiquitin-dependent process.
Ubiquitin-related processes have been shown to be associated with several cell cycle
events. For example, there are several examples of ubiquitin-dependent modification and/or proteolysis of cyclins (e.g.
36
-
40
). Other cell cycle-related events involve the requirement for a ubiquitin-conjugating protein (Cdc34) for passage from G1 to S (
41
) and in
S.pombe
the
mts2
gene, which is required for chromosome segregation, has been shown to encode
the 26S protease required for ubiquitin-dependent proteolysis (
42
). The
S.cerevisiae
homologue to
hus5
(
UBC9
) has been implicated in cyclin B destabilisation (
39
). However, the equivalent proteins in
Xenopus
and clam have not been demonstrated to have the ability to ubiquitinate cyclins
(
37
). Unlike the case in
S.cerevisiae
, deletion of
hus5
is not lethal (
11
) suggesting that in
S.pombe
hus5
may also not be involved with cyclin destruction.
Cyclosomes or APCs (anaphase promoting complexes) have recently been identified
as cell cycle-regulated ubiquitin-protein ligases (
34
). They are required for entry into anaphase and have been demonstrated to be
involved in cyclin B destruction. In addition, they have also been shown to be
required for the ubiquitination of other proteins, e.g. proteins which function
to inhibit sister chromatid separation, such as Cut2 (
43
). The phenotypes displayed by
hus5
and
rad31
mutants lead us to speculate that the gene products may be involved in
directing ubiquitination of proteins other than cyclins which are targeted by
the cyclosome.
Our results presented here provide the first indication that a member of this
UBA-related family of proteins is actually associated with a ubiquitin-related process, however, the precise role of
rad31
and the nature of the events for which it is required remain to be elucidated.
We are now in a position to investigate whether there is Hus5-dependent ubiquitination of, for example, certain replication proteins and
whether Rad31 has a role in this process.
ACKNOWLEDGEMENTS
We thank A.M. Carr and J.M. Murray for helpful discussions. The work was
supported in part by CRC grants SP2212/0101 and SP2212/0102. CLD was supported
by a BBSRC studentship. FZW thanks the Royal Society and the Wellcome Trust for
travel grants.
REFERENCES
1 Weinert,T.A. and Hartwell,L.H. (1988) Science,241, 317-322.MEDLINE Abstract
2 Al-Khodairy,F. and Carr,A.M. (1992) EMBO J., 11, 1343-1350.
