ABSTRACT
In addition to nucleotide excision repair (NER), the fission yeast
Schizosaccharomyces pombe
possesses a UV damage endonuclease (UVDE) for the excision of cyclobutane
pyrimidine dimers and 6-4 pyrimidine pyrimidones. We have previously described UVDE as part of an
alternative excision repair pathway, UVDR, for UV damage repair. The existence
of two excision repair processes has long been postulated to exist in
S.pombe
, as NER-deficient mutants are still proficient in the excision of UV
photoproducts. UVDE recognizes the phosphodiester bond immediately 5
'
of the UV photoproducts as the initiating event in this process. We show here
that UVDE activity is inducible at both the level of
uve1
+ mRNA and UVDE enzyme activity. Further, we show that UVDE activity is
regulated by the product of the
rad12
gene.
Multiple pathways exist for the repair of the major cytotoxic and carcinogenic
UV photoproducts, cyclobutane pyrimidine dimers (CPDs) and 6-4 pyrimidine pyrimidones (6-4 PPs). These include nucleotide excision repair (NER),
photoreactivation, recombination and UV damage repair (UVDR). NER is the
classical DNA excision repair system, which is the major repair pathway in
nearly all organisms responsible for the repair of not only UV photoproducts
but a variety of DNA damage types, including bulky lesions and cross-linked DNA (
1
-
3
). Photoreactivation, often referred to as light-dependent repair, had until recently thought to only be involved in
reactivating CPDs (
4
), however, recently a 6-4 PP photoreactivating enzyme was described (
5
). In photoreactivation, a photolyase recognizes the bonds joining adjacent
pyrimidines and in the presence of near UV light resolves the bases back to
monomers. Recombination is a less well-characterized mechanism of DNA repair, where DNA damage is removed by
recombination with its sister chromatid (
6
,
7
). UVDR is the most recently described DNA repair pathway. While the exact
mechanism of this process is not known, repair of UV photoproducts is initiated
by cleavage of the phosphodiester bond immediately 5' of either CPDs or 6-4 PPs. The enzyme which catalyzes this reaction, first described
in
Schizosaccharomyces pombe
, was named SPDE, for
S.pombe
DNA endonuclease (
8
), or UVDE, for UV damage endonuclease (
9
). A similar endonuclease has been identified in
Neurospora crassa
(
10
) and a homolog of the gene exists in
Bacillus subtilis
(
9
). To maintain continuity of nomenclature, we will refer to the
S.pombe
endonuclease as UVDE, for UV damage endonuclease, as defined by Takao
et al.
(
9
). Further, we will refer to the gene encoding UVDE as
uve1
+
and the UVDE-dependent repair pathway UVDR, for UV damage repair.
Our studies have previously shown that UVDR is distinct from NER, based on both
genetic and biochemical evidence (
11
). Yeast double mutant strains which carry mutations in genes involved in both
NER (
rad13-A
) and UVDR (
rad12-502
) are hypersensitive to UV light (
11
). Further, the site of 5' incision in UVDR is at the phosphodiester bond immediately adjacent to
the site of damage (
8
), while both 5' and 3' incisions in NER occur at a distance from the site of damage (
1
). Following 5' incision, DNA repair synthesis can be demonstrated
in vitro
(
11
). DNA repair synthesis is deficient in cell extracts prepared from
rad12-502
mutants and this deficiency can be complemented by the addition of partially
purified UVPE (
11
). This demonstrates that the reason UVDR is defective in
rad12
extracts is because of limiting UVDE activity. The existence of a second DNA
excision repair pathway for the removal of UV photoproducts was clearly
demonstrated genetically using antibodies directed against CPDs and 6-4 PPs (
12
). It has been proposed that UVDE may act to initiate a recombinational repair
process, which involves the
rad2
,
rad18
and
rph51
gene products (
13
). However, studies using our UVDR
in vitro
repair system (
14
) indicate that extracts prepared from both
rad2-44
and
rad18-X
cells have normal levels of repair synthesis (K.Sidik, unpublished data).
Further studies will be necessary to resolve the steps following UVDE incision
in this repair process.
Some enzymes involved in various DNA repair processes are induced in response to
DNA damage. In
Escherichia coli
transcription activation is controlled by the RecA protein (
15
). In
Saccharomyces cerevisiae
there are a number of genes which are activated in response to DNA damage. Some
of these gene products are involved directly in DNA repair, such as RAD2, RAD7,
RAD18, RAD51, RAD54 and SNM1, while others are involved in DNA metabolism, such
as RNR1, RNR2 and RNR3 (
16
-
22
). In
S.pombe
subtraction cloning was used to clone four cDNAs whose mRNAs were induced in
response to UV damage (
23
). Recently, it was reported that the
S.pombe
recombination repair gene
rhp51
is transcriptionally induced in response to DNA damage (
24
). Analysis of the
rhp51
promoter region revealed that it contained damage-responsive elements (DRE) homologous to sequences identified in
S.cerevisiae
as being involved in regulating induction by DNA damage. Interestingly,
rhp51
encodes a protein with amino acid similarity to the
E.coli
RecA protein. In this paper we show that UVDE is inducible at the level of
transcription and that UVDR, as measured by
in vitro
excision repair, is similarly induced. Further, we show that the product of the
rad12
+
gene does not code for UVDE, but rather is a regulator of UVDE activity.
Schizosaccharomyces pombe
was cultured by standard techniques (
25
). Complete genotypes of the strains used in this study are summarized in Table
1
. Sp272,
h
-S
rad12-502
, was constructed by outcrossing Sp264 (
h
+N
rad12-502
) twice with 972 (
h
-S
).
Table 1
Whole cell extracts were prepared from 10
10
-10
11
S.pombe
cells. These cells were grown to late log phase in 1.5* YEA (7.5 g/l yeast extract, 45 g/l dextrose, 100 mg/l adenine). Cells
were collected by centrifugation, washed in water, resuspended in an equal
volume of extraction buffer [20 mM Tris-HCl, pH 7.9, 10% glycerol, 1 mM EDTA, 10 mM MgCl
2
, 0.3 M (NH
4
)
2
SO
4
, 1 mM PMSF, 1 mM DTT] and frozen at -80oC. Frozen cells were thawed and lysis was performed using a 50 ml
bead beater (BioSpec Products Inc). After separating the beads, the cellular
debris was removed by centrifugation for 1 h at 100 000
g
. The supernatant was dialyzed for 5 h to overnight against 100 vol dialysis
buffer (20 mM HEPES-KOH, pH 7.6, 10 mM MgSO
4
, 10 mM EGTA, 5 mM DTT, 20% v/v glycerol, 1 mM PMSF). Protein concentrations of
the extracts, determined by Bradford assay (BioRad), were between 30 and 60
mg/ml.
Cells used to produce UV-induced extracts were grown in YEA to late log phase, collected by
centrifugation, washed with water and resuspended in 1 vol water. Cells were
placed in a large Petri dish or a glass tray and irradiated with constant
mixing. Inductions were performed with 254 nm light, using a total dose of 200
or 400 J/m
2
(as indicated) and a dose rate of 2.68 J/m
2
/s. The cells were transferred to fresh YEA and incubated with shaking at 30oC for the appropriate times. Cells were then collected by centrifugation
and mixed with an equal volume of extraction buffer prior to freezing at -70oC. Extracts were prepared as described above. Viability experiments
on cells irradiated by this protocol yielded 90-75% viability, indicating the actual dose absorbed by the yeast was 50-100 J/m
2
for the 200 or 400 J/m
2
total doses given respectively.
The 6-4 PP 49mer and CPD 49mer were synthesized as described (
27
). 3'-End-labeling was carried out by incubating 1 pmol of either
oligonucleotide with [[alpha]-
32
P]dGTP (50 [mu]Ci, 3000 Ci/mmol), 0.2 mM dATP and 5 U T4 DNA polymerase for 40 min at 14oC. This created the 3'-end-labeled 6-4 PP or CPD 51mer. UV-damaged plasmid DNA was prepared by
spotting 10 [mu]l droplets of supercoiled pUC18 DNA in TE at 0.1 [mu]g/[mu]l onto a sheet of parafilm. The DNA was exposed to 100 J/m
2
254 nm light.
UVDE assays were carried out essentially as described (
11
). Whole cell extract (100 [mu]g) was incubated with 0.02 pmol 3'-end-labeled 6-4 PP 51mer at 37oC for 5-15 min in 45 mM HEPES-KOH, pH 7.8, 70 mM KCl and 7 mM
MgCl
2
in a 20 [mu]l reaction. The samples were treated with proteinase K, extracted with
phenol/chloroform and the DNA analyzed on denaturing 15% polyacrylamide-urea gels. The gel was dried, exposed to X-ray film and the results were quantified on an Image Analysis
System (Fuji).
Extract preparation and reaction conditions were as described (
14
). Following incubation at 30oC for 2 h plasmid DNA was repurified (
14
). The DNA was separated on a 0.8% agarose gel. The gel was dried, exposed to X-ray film and the results were quantified on an Image Analysis System
(Fuji).
Details of the RIA have been published (
28
,
29
). Briefly,
S.pombe
was grown to late log phase in YEA, collected, resuspended in an equal volume
of dH
2
O and irradiated with 200 J/m
2
254 nm UV light. The cells were returned to YEA and grown for the indicated
times prior to harvest of total DNA. Next, poly(dA)[middot]poly(dT) (Boehringer-Mannheim) was nick translated with [[alpha]-
32
P]dTTP (Amersham) to a specific activity of 5-10 * 10
8
c.p.m./[mu]g. The labeled DNA was irradiated in water at a fluence rate of 14 J/m
2
/s, measured at 254 nm, for a total dose of 30 kJ. About 5-10 pg UV-irradiated, radiolabeled ligand competed with 7.5 [mu]g heat-denatured sample DNA for binding to antiserum. Rabbit
polyclonal antisera that bind 6-4 PPs or CPDs were added to TES (10 mM Tris, pH 7.8, 150 mM NaCl, 1 mM
EDTA) containing 0.15% gelatin (Type III; Sigma) at a concentration that
yielded 30-50% binding. The absolute specificities of these antisera for the 6-4 PP and
cis
,
syn
cyclobutane dimer have been demonstrated using mobility shift immunoassays of
damage-specific oligonucleotides (
30
). After overnight incubation at 4oC, goat anti-rabbit IgG (Calbiochem) and carrier [gamma]-globulin (Calbiochem) were added and incubated for 2-3 days at 4oC to form a precipitable immune complex. The
immune pellet was collected by centrifugation, dissolved in tissue solubilizer
(Amersham) and counted in a Packard liquid scintillation counter. For DNA
repair curves, percentage inhibition of sample DNA harvested at increasing
times post-irradiation was extrapolated through a linear regression of the unrepaired
sample harvested immediately after irradiation, to give the percentage
remaining photoproduct.
Schizosaccharomyces pombe
972 cells were grown in YEA to a density of 2 * 10
7
cells/ml at 32oC. Cells were collected and resuspended in 2 vol dH
2
O. The cells were irradiated with 400 J/m
2
254 nm UV light. Cells were then transferred to fresh YEA prewarmed to 32oC and grown with shaking for the indicated times. Cells were collected by
centrifugation at 3000
g
for 2 min and rapidly frozen. Because of the rapid induction times unirradiated
cells were collected and frozen and used to measure basal transcription. In
addition, aliquots of irradiated and unirradiated cells were plated and counted
for survival.
Total RNA was isolated by lysing the cells with glass beads. The frozen cell
pellets (0.4 ml) were resuspended in 4 ml Trizol (Gibco BRL) in 50 ml conical
tubes and enough glass beads (0.5 mm) added so that no liquid remained. The
cells were vortexed for 2 * 40 s. Then, 4 ml Trizol were added and mixed followed by the addition of
1.6 ml CHCl
3
. The samples were again mixed and the aqueous layer was separated by
centrifugation. The aqueous layer was extracted with an equal volume of
phenol/chloroform and isopropanol precipitated. Total RNA was collected by
centrifugation and the pellets briefly air dried. The RNA was suspended in 400 [mu]l DEPC-treated dH
2
O. Based on absorbance at 260 nm, between 1 and 1.5 mg total RNA were recovered.
Poly(A)
+
mRNA was isolated on Oligotex (Qiagen). Between 20 and 35 [mu]g poly(A)
+
mRNA were recovered from each sample.
Northern blot analysis (
31
) was done by resuspending 5 [mu]g poly(A)
+
mRNA from each sample in 10 [mu]l loading buffer (50% formamide, 20% formaldehyde, 20 mM MOPS, pH 7.0, 1 mM
sodium acetate, 1 mM Na
2
EDTA and 400 [mu]g ethidium bromide). The samples were heated to 65oC for 10 min and loaded onto a 1.2% agarose gel containing 2.2%
formaldehyde, 20 mM MOPS, pH 7.0, 1 mM sodium acetate and 1 mM Na
2
EDTA. The RNAs were separated by electrophoresis for 3 h at 80 V. The gel was
washed for 5 min in dH
2
O and blotted onto Zetablot (BioRad). Blots were probed with
32
P-labeled PCR product produced from UVDE DNA using
Pfu
I polymerase (Stratagene). Quantitation of mRNA levels from the
leu1
gene were carried out by probing with a
32
P-labeled
leu1
PCR product. Following washing the blots were either autoradiographed or
analyzed on a Molecular Dynamics phosphorimager.
A wild-type genomic library, made by a partial
Hin
dIII digest cloned into pWH5, was screened for UVDE sequences. The probe was
made by PCR amplifying a region of the UVDE cDNA using [[alpha]-
32
P]dCTP and [[alpha]-
32
P]dGTP in the reaction. The clone (pUVDE12) contained several
Hin
dIII fragments. One of these fragments of ~5.4 kb was shown to contain the entire sequence for UVDE. This fragment was
subcloned into pUC18 and named pgUV2. The sequencing of
uve1
+
and its promoter region was accomplished by both conventional dideoxy
sequencing using Sequenase (US Biochemicals) and by automated sequencing (ABI).
Previous data had shown that extracts prepared from
rad12-502
cells were deficient in UVPE activity based on an
in vitro
excision repair system (
11
). Furthermore, when
rad12-502
cells were crossed with the NER-deficient mutant strain
rad13-A
the resulting double mutant was hypersensitive to UV light. Based on these data
the
rad12-502 rad13-A
double mutant cells were tested for excision of UV photoproducts in an
in vivo
assay. In this assay cells were grown to late log phase, exposed to UV light
and their DNA isolated at various times following irradiation. UV adducts
remaining in the DNA were measured by an immunoassay using antibodies specific
to CPDs or 6-4 PPs. Our results demonstrated that this double mutant was still
proficient in the excision of both CPDs and 6-4 PPs (Fig.
1
). These data contrast with similar experiments carried out in
S.cerevisiae
, where elimination of NER function alone is sufficient to prevent excision of
UV photoproducts
in vivo
(
12
).
To show that induction of UVDE activity was DNA damage dependent and not due to
different levels at specific points in the cell cycle, UVDE activity was
measured as cells exited from a
cdc25-22
-dependent G
2
arrest. The
cdc25-22
mutant arrests the cell cycle at a point concomitant with the radiation-induced G
2
checkpoint (
32
,
33
). Cells were synchronized at the G
2
/M transition by culturing at 36oC, the restrictive temperature for
cdc25-22
. Following release at 25oC, the permissive temperature for
cdc25-22
, there were no significant changes in UVDE activity as cells progressed through
the cell cycle (Fig.
4
). The time course presented includes two mitoses, with H1 kinase activity peaks
at 20 and 140 min. These results indicate that the UVDE activity increase in
response to UV light depends on UV-induced damage and not cell cycle phase changes after UV light exposure.
To determine if the induced levels of UVDE were due to increased transcription
of the
uve1
gene, wild-type cells were grown to late log phase and irradiated with 400 J/m
2
254 nm UV light. Total RNA was isolated and poly(A)
+
mRNA selected. Preliminary studies had shown that the
uve1
+
mRNA was not detected in total RNA by Northern blot analysis. Based on
absorbance at 260 nm, very similar recoveries of mRNA for each sample were
obtained. This was borne out by the results of probing with a
leu1
probe, which showed very similar levels of
leu1
mRNA in each sample. Because these cells were irradiated in suspension we
wanted to check survival in order to determine how this dose related to that of
cells irradiated on plates. Cells before and after UV irradiation were plated
and counted for survival. In two separate experiments this dose of UV light
gave 74 and 75% survival, which is an equivalent dose of plated cells of 80 J/m
2
.
Poly(A)
+
mRNA (5 [mu]g) was separated on 1.2% agarose-formaldehyde gels, blotted and hybridized to a fragment of
uve1
+
generated by PCR. The blot was visualized and quantitated using a
phosphorimager (Fig.
5
). The UVDE mRNA band migrates at ~2.3 kb. A second slower migrating band is visible which we have determined
to be cross-hybridization with rRNA. The data indicate that the mRNA levels elevate
very quickly, increasing to 2.5-fold higher than unirradiated within 10 min of irradiation, then rapidly
returning to normal levels. Induction was so rapid that cells collected by
centrifugation and quick frozen immediately following irradiation showed
significant induction (data not shown). For this reason induction was compared
with unirradiated cells.
Since induction of
uve1
following DNA damage is transcriptionally regulated we compared its promoter
sequence with that of the
S.pombe
rhp51
, gene whose induction following DNA damage is also transcriptionally regulated.
The promoter region of
uve1
(Fig.
6
) appears to contain two DREs which share homology with regulatory sequences in
S.cerevisiae.
The two sequences, labeled DRE1 (CATGGCCTTC) and DRE2 (CTGGGAATGA), share
reasonable homology with the DRE sequences of
rph51
and those of
S.cerevisiae
(C[T/G][T/A]GG[T/A]NT[T/C][A/C]). In addition, a search of the sequence shows an exact 9 nt match with the c-Jun binding site (TGACGTAAC) at position -220.
Genetic data and
in vivo
studies had previously shown that
S.pombe
possesses an excision repair pathway independent of NER for the removal of UV
photoproducts (
12
). Recent studies by our laboratories have shown that this second DNA excision
repair pathway, which we have named UVDR, for UV damage repair, relies on the
enzyme UVDE for the removal of both CPDs and 6-4 PPs (
11
). In that study we showed that extracts prepared from
rad12-502
cells were deficient in UVDE activity and that repair activity could be
restored by adding back UVDE. This data demonstrated that UVDE is required in
this reaction. The fact that partially purified UVDE and purified
mus-18 protein, the
Neurospora crassa
homolog of UVDE, can recognize and cleave at CPDs and 6-4 PPs (
8
,
10
) suggests that this endonuclease alone is the damage recognition and repair
initiating event in this process. In this study we have provided data
demonstrating that UVDR is inducible and that the induction includes increased
UVDE activity. While other unidentified proteins involved in this reaction may
also be elevated in response to damage, our earlier studies indicated that UVDE
was limiting in this reaction (
11
). Interestingly, while the
rad12
gene product is required for maintaining normal basal levels of UVDE,
rad12-502
cells induce normally, suggesting that induction of UVDE is independent of
rad12
. The mechanism of
rad12
regulation is currently under study. It is also unclear from these results why
rad12-502 rad13-A
double mutants are hypersensitive to UV damage, which we previously reported,
while they have normal induction of UVDE activity. There are two possible
explanations: the lack of basal levels of UVDE make the cells more sensitive or
that the
rad12
gene product plays a broader regulatory role beyond UVDE regulation.
Increased levels of UVDE activity could be accounted for in a number of ways. A
trivial explanation for increased UVDE activity could be that
uve1
+
mRNA is synthesized only during the G
2
part of the cell cycle and that increased levels of UVDE following DNA damage
is simply due to the fact that the cells are arrested in G
2
. To test this possibility we studied UVDE levels through the cell cycle using a
synchronized cell population. Cells containing a temperature-sensitive
cdc25
mutation were blocked in G
2
by incubation at the non-permissive temperature for 3 h followed by resumption of growth by
shifting to the permissive temperature. UVDE activity was measured and shown to
be at essentially constant levels throughout the cell cycle (Fig.
4
). We next measured levels of the
uve1
+
transcript following UV irradiation. Northern blot analysis demonstrated that
UVDE mRNA levels increased (within 5 min) following exposure to UV light and
that these elevated levels rapidly return to normal (within 20 min). It is
interesting that while UVDE mRNA levels rise rapidly, peaking at 10 min, the
levels of UVDE activity peak at 60 min. Whether this signals another level of
regulation of UVDE activity or represents normal expression time is not clear
from these studies. Future studies are planned to study post-transcriptional regulation of UVDE.
Cell-free extracts prepared from
S.pombe
cells following UV irradiation were previously reported to have elevated base
excision repair activity (
34
). However, we believe that in fact they were measuring UVDR activity. Induction
of DNA repair genes by transcriptional activation has been reported previously
in
S.pombe
. Four transcripts were shown to be elevated following exposure to 254 nm UV
light, named
uvi15
+
,
uvi18
+
,
uvi22
+
and
uvi31
+
(
23
). Two of these genes,
uvi18
+
and
uvi31
+
, were induced only by UV light. It will be interesting to see if either of
these genes are involved in UVDR. This same group has recently reported that
rhp51
+
, the fission yeast homolog of the
E.coli
recA
and
S.cerevisiae
RAD51
genes, is transcriptionally regulated (
24
). They further showed that the promoter region of
rhp51
contains sequences sharing homology with regulatory elements described in
S.cerevisiae
. Deletion analysis showed that a region containing two DREs, DRE1 and DRE2, was
necessary for both basal levels of transcription and the inducible response.
The
S.pombe
DRE sequences, which act as positive regulators of
rhp51
, share sequence homology with upstream repressor sequences, negative regulatory
elements identified in
S.cerevisiae.
Promoter sequences in
rhp51
homologous with upstream activating sequences appear to be required for
maintaining basal levels of transcription. Finally, deletion of a region very
near the promoter led to loss of repression of
rhp51
expression, as transcription in these mutants is at the induced level. All of
this suggests a relatively complex mechanism of regulation of this gene.
Analysis of the sequences in the UVDE promoter (Fig.
6
) shows the presence of DRE-like sequences. In addition there is homology with a number of other
promoter elements, most notably the TGACGTAAC c-Jun binding sequence. The presence of a c-Jun binding site is interesting in the light of the fact that c-Jun, which is part of the AP-1 transcription complex, is activated in response to UV
damage in mammalian cells and that AP-1 binding has been implicated in the regulation DNA damage-induced genes (
35
,
36
). As with
rhp51
,
uve1
transcriptional control would appear to be complex, with regulation of both its
basal transcription levels and its induction. Ultimately, a detailed analysis
of the
uve1
+
promoter region will be required to determine all the elements involved in
uve1
+
regulation.
S.D. is a Career Scientist of the Ontario Cancer Treatment and Research
Foundation and this work was supported in part by OCTRF Start-up funds to S.D. G.A.F. was supported by NIH grant CA72647. D.L.M. was
supported by NIH grant ES05914.
Strain
Genotype
Reference
972
h
-S
25
Sp18
h
-S
cdc25-22
26
Sp264
h
+N
rad12-502
11
Sp269
h
-S
rad12-502 rad13-A
11
Sp272
h
-S
rad12-502
This study
Sp273
h
-S
leu1-32 rad13-A
11
REFERENCES
Return