INTRODUCTION
Ultraviolet light (UV) induces two major types of DNA damage:
cyclobutane type pyrimidine dimers (CPDs) and (6-4)
photoproducts [(6-4)PPs]. To ensure survival in a UV-rich
environment, various DNA repair mechanisms exist to remove
UV-induced DNA damage. One of the simplest and most
effective ways to repair UV-induced damage is photoreactivation,
in which a single enzyme named photolyase eliminates CPDs by
transfer of absorbed visible light energy to CPDs as electrons to
split CPDs by b-elimination (1). In contrast and in addition to this
single enzyme process, another important and efficient repair
mechanism for UV-induced DNA damage is nucleotide excision
repair (NER) which requires a large protein complex, recently
termed the `repairosome'. It contains more than ten independent
gene products which recognize DNA damage and introduce
incisions into the neighbourhood of the damaged sites on DNA
(2,3).
Recently, we isolated a novel gene encoding a protein with an
incision activity acting against both CPDs and (6-4)PPs from the
filamentous fungus Neurospora crassa (4). Since N.crassa does
not possess NER, this repair activity seems to play a substitutional
role for NER in this organism. We designate this gene product,
which is defective in the UV-sensitive mutant mus-18, as a
N.crassa UV-damage endonuclease (NC-UVDE).
A similar enzymatic activity has been reported in an extract of
the fission yeast Schizosaccharomyces pombe (5). To identify the
gene responsible for this activity, we applied the same method as
that used for isolation of NC-UVDE, namely, the complementation
of repair-deficient E.coli cells after introduction of a foreign cDNA
library and selection by UV irradiation of the transformants. We
isolated a S.pombe homolog of the NC-UVDE gene. Herein we
report the characterization of the gene and its product as well as
a homolog of UVDE from the Gram positive bacterium Bacillus
subtilis.
Cloning of S.pombe cDNA
Schizosaccharomyces pombe cDNA library (Clontech) in a
vector pGADGH was introduced in E.coli SY2 (JM107
Dphr::Cmr DuvrA::Kmr DrecA::Tetr). Selection of UV-resistant
clones was performed as previously reported (4). All UV-resistant
transformants contained the same plasmid designated as
pSpUVDE. The sequence of the insert was determined from both
orientations by auto sequencer model DSQ-1000 (Simadzu).
Expression of the cloned gene
A cDNA fragment was amplified with primers having
appropriate restriction sites and inserted into the expression
vector pFLAG2 (Kodak), generating pFDE. E.coli XL-1 Blue
cells transformed with the expression construct pFDE were
treated with 1 mM IPTG at OD = 0.6 for 5 h. Cells were collected,
suspended in a buffer containing 50 mM Tris-HCl (pH 7.5), 100
mM NaCl, 1 mM EDTA, 1 mM DTT and 1 mM PMSF, and
sonicated. Supernatant of the cell lysate was loaded on a
heparin-Sepharose column with a linear gradient of 0.1-1 M
NaCl in buffer A (50 mM Tris-HCl, 1 mM EDTA and 1 mM
DTT, pH 7.5). Activity of UV-dimer endonuclease, which
converts closed circular UV-irradiated plasmid into an open
circular form, was assayed as detailed below. Fractions around
0.6 M NaCl were collected, diluted to 0.3 M NaCl with buffer A,
and subsequently applied on a Hi-Trap Blue column (Pharmacia).
Elution was made with a gradient of 0.3-2.0 M NaCl in buffer A
containing 10% glycerol. Fractions at 1.5 M NaCl with activity
peaks were collected and concentrated by centricon 30 (Amicon).
Purity was around 35% as judged by SDS-PAGE. This fraction
(BS-fraction) was used for in vivo plasmid nicking and incision
assays.
Plasmid nicking assay
Close circular plasmid was exposed to UV (1 kJ/m2) and incubated
with BS-fraction or column fractions at 37_C for 30 min. The
reaction was stopped by heating at 65_C for 10 min and analyzed
on an agarose gel. Optimal conditions of the plasmid nicking
activity for the BS-fraction were determined as 50 mM Tris-HCl
(pH 7.9), 50 mM KCl, 15 mM MgCl2 and 1 mM DTT, by varying
the concentration of the respective ingredients and changing
buffer pH.
Incision assay using synthetic oligonucleotides
Incision of UV-irradiated oligomer was assayed in essentially the
same way as previously described (5). An oligonucleotide containing
two dipyrimidine sequences (underlined),
54-GTATACACACACGTATGCATCATGTTATACGCACACAC
AGTGCATACA-CATATAGC-34, was either 54-labelled with
[g-32P]ATP by T4 polynucleotide kinase (Takara) or 34-labelled with
[a-32P]ddATP by terminal deoxynucleotidyl transferase
(Boehringer), annealed to its complementary oligomer and exposed
to UV (3 kJ/m2). The DNA (10 fmol) was incubated with
BS-fraction, NC-UVDE or T4 PD-DNA glycosylase at 37_C for 60
min. The reaction was stopped by adding an equal volume of 90%
formamide loading solution, heated at 95_C for 2 min, and analyzed
on a 15% polyacrylamide denaturation gel. For the size marker,
end-labelled, unirradiated DNA digested with NlaIII or NsiI was
mixed and run on a lane. For preparing DNA containing CPD or
(6-4)PP, photoreactivation of the 54-labelled, UV-irradiated DNA
with Drosophila phr or Analycistis phr was performed as described
previously (4).
Construction of truncated UVDE gene
The cDNA was truncated from the N-terminus by digesting it at
the ScaI or ClaI site and inserted into pUC18 or pGEM7Zf+
vectors (which generated a small portion of lacZ fusion), resulting
in pDN101 or pDN349, respectively. To generate additional
N-terminal deletion constructs, PCR was made with a 54 primer
starting at the 233th or 274th amino acid position and inserted in
a modified pFLAG2 vector in which the FLAG peptide is fused to
the C-terminus of the insert (pDN232 or pDN273, respectively).
An antibody to the FLAG sequence (Kodak) was used to detect
peptide production in transformants of pDN232 and pDN273.
C-terminal deletion construct pDC564 was made by digesting
pFDE with EcoO109I and BglII (in MCS of the vector),
blunt-ended by the Klenow fragment of DNA polymerase and
religated.
Cloning of Bacillus subtilis homolog of eukaryotic UVDE
A homology search of the database with the conserved amino acid
sequence between NC-UVDE and SP-UVDE was performed by
TBLAST (6), predicting a homologous sequence in B.subtilis
genomic sequence (Z49782). The bacterial genomic fragment
showing UVDE homology was amplified with primers
(54-CTATGATTTTCAGATTCGGGTTCGTTT-34 and
54-CATTTATGACTTCCATTGCAGCGCACC-34 for the 54- and
34-end of the ORF, respectively) and cloned into pT7-Blue
(Novagen). The plasmid was introduced into SY2 cells and
examined for UV resistance.
RESULTS
Cloning of a cDNA complementing UV sensitivity of
repair deficient E.coli host cells
We introduced a cDNA library made from S.pombe into the E.coli
strain SY2, which lacks DNA repair activity in all the three
pathways essential for repair of UV-induced DNA damages,
namely, recombination, NER and photoreactivation. UV-resistant
transformants were isolated after three rounds of UV-selection
procedures (Fig. 1) and a plasmid harbouring a 2 kb insert with
the longest open reading frame (ORF) consisting of 598 amino
acid (aa) was isolated. The presence of an in-frame stop codon
upstream of the first start codon for translation yielding the
longest ORF, suggests that it is the start codon for translation of
this gene. The deduced aa sequence was compared with that of the
previously isolated NC-UVDE gene (Fig. 2) and was found to
have 36.6% identity. The similarity between the two sequences
was found in the C-terminal two-thirds region (around 54%),
while only 16% identity was found in the N-terminal one-third of
the proteins. The nucleotide sequence data reported in this paper
will appear in the DDBJ, EMBL and GenBank nucleotide
sequence databases with the accession number
D78571.
Characterization of a recombinant protein obtained
from E.coli host cells
We expressed the cloned cDNA from the putative start codon using
the tac promoter in E.coli and analyzed whether the S.pombe
protein has an activity similar to that of NC-UVDE to cleave
UV-irradiated DNA. Since cell lysate of a SY2 transformant was
found to convert UV-irradiated supercoiled plasmid DNA to a
relaxed form, the purification process was monitored by the
nicking activity of UV-irradiated supercoiled DNA. Applying
cell lysate on a heparin-Sepharose and subsequently to a
blue-Sepharose column, ~35% homogeneity of the protein
without bacterial nuclease activity was obtained (not shown).
Since further purification led to an immediate inactivation of the
enzymatic cleavage activity, we used the blue-Sepharose fraction
(BS-fraction) in all in vitro assays reported herein.
First, the requirements of S.pombe UVDE for plasmid nicking
activity were examined. The BS-fraction was mixed with
UV-irradiated and non-irradiated supercoiled plasmid DNA of
two different sizes. Only the UV-irradiated DNA was incised
when appropriate concentrations of magnesium and potassium
were included in the reaction (Fig. 3A). The effect of ATP on the
plasmid nicking activity was also examined using the amount of
the BS-fraction which converted 60% of UV-irradiated plasmid
to a relaxed form in the complete reaction. As shown in Figure
3B, the addition of ATP did not enhance the activity. The required
presence of magnesium and salt as well as ATP independence of
SP-UVDE for its activity are the same as found for
NC--UVDE.
To demonstrate substrate specificity and position of the nicks,
a synthetic oligomer was prepared, which contained one TC and
one TT dipyrimidine sequence, major potential sites for
UV-induced (6-4)PP and CPD, respectively (Fig. 4A).
Restriction sites for NsiI and NlaIII which cleave 54 of the first
thymines of TC and TT, respectively, were included so that the
restriction digests work as size markers. The oligomers were
either 54- or 34-labelled, annealed to its complementary strand
oligomer and UV irradiated. This DNA was mixed with the
BS-fraction and the reaction products were analyzed by
polyacrylamide gel electrophoresis (PAGE) as shown in Figure
4B and C. The reaction produced two fragments at the TC and TT
sites (Fig. 4B and C, lane 1) with the same migration as the DNA
fragments resulting from the reaction with NC-UVDE (Fig. 4B
and C, lane 2) and those of marker fragments (i.e. mixture of NsiI
and NlaIII digests). In contrast, the reaction of the DNA with T4
phage-derived CPD-DNA glycosylase showed a major band near
the TT site created by its AP-lyase action (Fig. 4B and C, lane 3).
Thus, the sites cleaved by SP-UVDE are different from those of
T4 CPD-DNA glycosylase and identical to those of NC-UVDE.
Judging from the positions of the nicked DNA fragments in
Figure 4, SP-UVDE introduced a single nick immediately 54 to
the damage leaving 34-OH and 54-phosphate termini at the nicked
site as is the case for NC-UVDE
(4).
In order to further confirm the substrate specificity, 54-labelled,
UV-irradiated DNA was photoreactivated with either photolyase
from Anacystis nidulans or Drosophila melanogaster prior to the
incision reaction. As shown in Figure 4B, when oligomer was
photoreactivated with Drosophila photolyase, which reverts
(6-4)PP lesions, the band intensity at the TC site was significantly
reduced for the SP-UVDE reaction (lane 4) and the NC-UVDE
reaction (lane 5). Photoreactivation of the DNA with Anacystis
photolyase, which reverts CPD lesions, diminished the band
intensity at the TT site of the SP-UVDE reaction (lane 7) and the
NC-UVDE reaction (lane 8). Reasonably, only photoreactivation
with Anacystis photolyase eliminated the substrates for T4
CPD-DNA glycosylase (lane 6 and 9). These results indicate that
the incision occurs immediately 54 to both CPDs and (6-4)PP
lesions as previously found for NC-UVDE (4). Furthermore, the
manner of the incision of the recombinant SP-UVDE is consistent
with the activity reported for the extract of S.pombe cells (5).
Minimum size of the cDNA indispensable for enzymatic
activity
The N-terminal one-third of the SP-UVDE is different in aa
sequence from that of the Neurospora homolog. We next
investigated, whether this part of the protein is necessary for the
complementing activity of UV sensitive E.coli host cells. Several
truncated SP-UVDE genes were constructed and introduced into
E.coli SY2 host cells. No change in UV resistance of the E.coli
transformants were observed until the N-terminal sequence was
deleted up to the 232th residue (Fig. 5A). Deletions from the
N-terminus to the 273th as well as the 349th residue led to a
phenotype as UV-sensitive as that of cells harbouring vector
plasmid (Fig. 5B). In order to confirm protein production in the
transformants, whole cell lysates were analyzed by Western
blotting using antibody against the tag fused at the C-terminus of
the recombinant protein. The transformants harbouring pDN232
and pDN273 produced truncated soluble proteins of reasonable
sizes (Fig. 5C), suggesting that the deletion up to 273th residue
of the gene first inactivates the enzymatic activity of the gene
product. On the other hand, even a small deletion of 35 residues
at the C-terminal sequence of the protein influenced the enzymatic
activity (Fig. 5A and B). From these results, it is concluded that the
C-terminal two-thirds of the protein, where aa sequences are well
conserved between two UVDEs, almost coincides with the region
indispensable for the enzymatic activity.
A bacterial homolog of UVDE
In a recently published database of the genomic nucleotide
sequence from the Gram positive bacterium Bacillus subtilis, an
ORF was found which showed a remarkable similarity in its
deduced aa sequence to the NC- and SP-UVDE (Fig. 2). We
obtained the genomic fragment of the ORF by polymerase chain
reaction and introduced it into E.coli SY2 cells. Figure 6 depicts
the complementing activity of UV sensitivity by introduction of
the bacterial gene. Thus, UVDE is distributed not only in
eukaryotes but also in eubacteria.
DISCUSSION
The presence of a second excision repair pathway for UV-induced
DNA damage was first suggested in the fission yeast
Schizosaccharomyces pombe (7) and recently, it was shown that
this pathway is distinct from NER (8). This alternative excision
repair pathway involves an enzyme similar to the UV
endonuclease, which we identified in N.crassa (9). We showed
here that S.pombe has a homolog of the N.crassa gene, which
encodes an endonuclease for UV-induced CPDs as well as
(6-4)PPs. Since the enzymatic activity of the recombinant protein
is identical to the reported enzymatic activity in the extract of
S.pombe cells (5), we concluded that the reported UV
endonuclease activity in S.pombe is derived from the cloned gene.
Thus, we showed the existence of a second excision repair
pathway, besides the well-known NER, for UV-induced DNA
damage in this widely-studied yeast
species.
We previously found a number of consecutive hydrophilic aa
sequences in the NC-UVDE, including a sequence
4Xlys-2Xgly-2X(LysArg) found near the C-terminal region with
similarity to a part of protamine (4). Alignment of the aa sequences
of NC- and SP-UVDEs (Fig. 2) indicates that this sequence is not
well conserved, although several deletions in this sequence
inactivates the complementing activity of the NC-UVDE (4). From
the experiments using N- or C-terminal deletion constructs, only the
carboxyl two-thirds of the SP-UVDE turned out to be essential for
the complementing activity. We do not know the function of the little
conserved N-terminal region of the UVDE proteins. This region,
which is abundant in Ser and Glu residues for both proteins, may
have functions to interact with other proteins.
The aa sequence alignments between the eukaryotic UVDEs and
the possible bacterial homolog found in Bacillus subtilis show a
27% identity. Although this is a rather low similarity for the protein
with specific functions, judging from the complementing activity
of UV sensitivity in E.coli host cells, the bacterial protein is
probably structurally and functionally similar to the eukaryotic
homologs.
Lac promoter expression of the Bacillus gene in forward
orientation is toxic for the host, especially for the cells in the
stationary phase, as shown by our finding that almost all cells from
the overnight culture were not viable. This phenomenon was
previously observed in an E.coli culture harbouring the
Neurospora UVDE gene (unpublished data). Reproducible
survival curves of transformants with the Bacillus gene were
obtained when the coding sequence was introduced in the reverse
orientation to the promoter in the plasmid, suggesting that high
expression of the gene product is disadvantageous for the E.coli
host cells.
The presence of the NC-UVDE gene homolog in yeast and
bacteria increases the significance of this alternative DNA repair
pathway. We previously pointed out the possibility that
NC-UVDE may play a substitutional role for NER in N.crassa
which does not possess efficient NER activity (4). Although
S.pombe has NER, photolyase activity for CPD has not been
found (10). Similarly, photolyase gene seems to be missing in
Bacillus subtilis (11,12). Thus, SP-UVDE in S.pombe and its
homolog in the B.subtilis could substitute for the
photoreactivation of CPDs in these organisms.
The function and interaction of the S.pombe UVDE with other
repair systems in the well characterized yeast system are more easily
analyzed than in Neurospora. It is now of interest to identify the
processing of the nicked UV damage. We previously introduced the
NC-UVDE gene into various repair deficient mutants of the budding
yeast Saccharomyces cerevisiae and found that the repair pathway
initiated with the nicking by UVDE can be completed in the rad2
mutant which is deficient in the 34 nicking activity of NER. There are
most probably other proteins involved in damage processing in this
alternative DNA repair pathway.
ACKNOWLEDGEMENTS
We would like to thank Drs Takeshi Todo and Andries P.M. Eker
for providing us with Drosophila and Anacystis photolyases,
respectively. The presence of nucleotide sequence in the Bacillus
database homologous to a part of the UVDE gene of N.crassa was
first informed by Dr Paul W. Doetsch, for which we are very
grateful. Technical assistance by Ms Izumi Chiba and Ms Junko
Kikuchi is acknowledged. This work was supported by a
Grant-in-Aid for Scientific Research on Priority Area No.
07270101 from the Ministry of Education, Science, Sports and
Culture of Japan to A. Yasui.
1 Sancar,A. (1994) Biochemistry, 33, 2-9.
3 Hoeijmakers,J.H.J. (1993) Trends Genet., 9, 173-177.
6 Altschul,S.F., Gish,W., Miller,W., Myers,E.W. and Lipman,D.J. (1990) J.
Mol. Biol. 215, 403-410.
7 Birnbolm,H.C. and Nasim,A. (1975) Mol. Gen. Genet. 136, 1-8.
8 McCready,S., Car,A.M. and Lehmann,A.R. (1993) Mol. Microbiol. 10,
885-890.
10 Yasui,A., Eker,A.P.M. and Koken,M. (1 989) Mutat. Res., 217, 3-10.
11 Quirk,P.G., Hicks,D.B. and Krulwich,T.A. (1993) J. Biol. Chem., 268,
678-685.
* To whom correspondence should be addressed