ABSTRACT
The
Escherichia coli
exonuclease III (AP endonuclease VI) is a DNA-repair enzyme that hydrolyzes the phosphodiester bond 5
'
to an abasic site in DNA. To study how the enzyme recognizes the abasic site,
we used oligonucleotides containing a synthetic abasic site at any desired
position in the sequence. We prepared oligonucleotides containing an abasic
residue such as 2
'
-deoxyribosylformamide, 2
'
-deoxyribose, 1
'
,2
'
- dideoxy ribofuranose or propanediol. Duplex oligonucleotides containing an abasic
residue used in this study were cleaved on the 5
'
side of the abasic site by exonuclease III in spite of the varieties of the
bases opposite and adjacent to the abasic site. In addition, we observed that
the enzyme cleaved single-stranded oligonucleotides containing an abasic site on the 5
'
side of the abasic site. These findings suggest that the enzyme may principally
recognize the DNA-pocket formed at an abasic site. The indole ring of the tryptophan 212
residue of the exonuclease III is probably intercalated to the abasic site. The
tryptophan in the vicinity of the catalytic site is conserved in the type II AP
endonuclease from various organisms.
Apurinic/apyrimidinic (AP) endonucleases play an important role in the repair of abasic sites in DNA (
1
,
2
). Unless lesions are repaired prior to DNA replication, the non-coding lesions promote misincorporation of nucleotides and mutagenesis.
Abasic sites are the common lesions that arise in cellular DNA. Abasic sites
are generated both spontaneously by the hydrolysis of glycosidic bonds (
3
,
4
) and enzymatically by
N
-glycosidases that remove misincorporated bases (
5
). Damaged bases due to active oxygen species formed by ionizing radiation (
6
) are fragmented into small remains. Generally, abasic residues involve the base alterations that are hydrolyzed after base modification.
The major AP endonuclease in
Escherichia coli
is the exonuclease III, which accounts for over 80% of the cellular AP endonuclease activity (
7
,
8
) and belongs to a class II AP endonuclease. Furthermore, this enzyme is a multifunctional enzyme which exhibits 3'-5' exonuclease (exonuclease III), 3'-phosphomonoesterase, 3'-repair diesterase, and
ribonuclease H activities (
9
-
12
). This class II AP endonuclease hydrolyzes the phosphodiester bond 5' to the AP sites. Cloning and base sequencing of the class II AP
endonucleases from various organisms such as human (
13
), bovine (
14
), mouse (
15
),
Drosophila melanogaster
(
16
),
Streptococcus pneumoniae
(
17
)
and
Bacillus subtilis
(
18
)
have been reported. A comparison of the amino acid sequences among these enzymes
reveals a remarkably high homology to exonuclease III of
E.coli.
This enzyme removes both the 2'-deoxyribosyl residue and [alpha],[beta]-unsaturated compound from the 3' terminus created by the class I AP
endonuclease (
19
-
21
). The class I AP endonucleases, such as endonuclease III in
E.coli
, hydrolyze the phosphodiester bond 3' to the AP site (
19
). There is no homology in amino acid sequences between the class I and class II
AP endonucleases.
The substrate specificity of exonuclease III has been investigated using DNAs damaged by reagents and physical treatments. Recently there has been marked development in the application of synthetic
oligonucleotides containing an abasic residue as substrates for AP endonucleases (
22
-
24
). Previously, we reported the chemical synthesis and properties of an
oligodeoxynucleotide containing a 2'-deoxyribosylformamide residue (
25
,
26
). The synthetic oligonucleotides containing a modified tetrahydrofuran moiety,
which is isosteric with 2'-deoxyribose and an analog of the acyclic sugar moiety, are cleaved on the 5' side of the abasic site by exonuclease III (
22
). Furthermore, exonuclease III cleaves the phosphodiester bond 5' to the
O-
alkylhydroxylamine
N-
glycosides (
27
) as well as that 5' to the urea
N-
glycosides (
28
) in DNA. However, thymine glycol
N-
glycoside, dihydrothymine
N-
glycosides and formamidopyrimidine
N-
glycoside do not serve as substrates (
27
). From these results, it has been suggested that a secondary amine at the
N-
glycosyl bond and lack of base pairing by the damaged base are required for
recognition and cleavage of the AP site by exonuclease III. Recently, crystal
structures of exonuclease III and a ternary complex made up of Mn
2+
, dCMP, and this enzyme have been reported (
29
). From analysis of the crystal structure of exonuclease III, it has been
assumed that an extra-helical base on the DNA strand opposite to the AP site plays an important
role in the recognition by exonuclease III.
At present, little is known about the binding mechanism of AP endonuclease to
the abasic site. In this paper, we report the substrate specificity of exonuclease III using synthetic oligonucleotides
containing an abasic residue such as 2'-deoxyribosylformamide (F), 2'-deoxyribose (D), 1'2'-dideoxyribofuranose (H), and
propanediol (P) at the desired position. We have found that not only double-stranded DNA but also single-stranded DNA containing an abasic site is a good substrate for the AP endonuclease activity
of exonuclease III. Finally, we propose that the conserved tryptophan residue
spatially near the catalytic site in the class II AP endonuclease might be attributable to recognition of an AP site.
Oligonucleotides were synthesized using an Applied Biosystems model 391S DNA
synthesizer on a 0.2 [mu]mol scale. Protected DNA phosphoramidite monomers were purchased from
Applied Biosystems and Glen Research (Sterling, VA). Phosphoramidite monomers
of abasic residues (F, H, and P shown in Fig.
1
) were prepared and used as a 0.1 M acetonitrile solution (
22
,
30
). Fully protected oligonucleotides containing F, H, and P were deblocked and
purified using the same procedure for the purification of natural oligonucleotides. An alkali labile oligonucleotide containing a 2'-deoxyribose (D) was prepared by enzymatic treatment of an oligonucleotide containing 2'-deoxyuridine with uracil
N-
glycosidase (
31
) (Perkin Elmer Cetus, Norwalk, CT). The duplex DNA was prepared by annealing an
oligonucleotide containing an AP site and an excess complementary
oligonucleotide. The 5' end of the oligonucleotide was labeled with [[gamma]-
32
P]ATP (Amersham) by T4 polynucleotide kinase (Nippon Gene, Toyama). The nucleotide sequences of the oligonucleotides used in this study are shown in Figure
1
.
Escherichia coli
exonuclease III was purchased from Takara Shuzo (Kyoto, Japan). Using sodium dodecyl sulfate-polyacrylamide gel electrophoresis followed by silver-staining, no impurity in this commercial product was detected.
Approximately 4.4 pmol of double-stranded or single-stranded DNA (0.44 [mu]M) was dissolved in 10 [mu]l of 77 mM NaCl, 5 mM MgCl
2
, 10 mM DTT, 66 mM Tris-HCl, pH 8.0, and then exonuclease III (0.4 nM, 4 nM, 40 nM) was added to the solution. The incubation mixture was maintained at 23oC for 5 min. The reaction was quenched by the addition of 5 [mu]l of the stop solution (7 M urea, 10 mM EDTA, 0.1% bromophenol
blue, and 0.1% xylene cyanol), and then the solution was boiled for 5 min. The
reaction products were analyzed by electrophoresis through a 20% denaturing (7
M urea) polyacrylamide gel (19:1 ratio of acrylamide to bisacrylamide) using
0.4* TBE (90 mM Tris-borate, pH 8.3, 2 mM EDTA). DNA fragments were visualized by
autoradiography. Relative efficiencies of the cleavage reactions with the
different substrates were determined by densitometric autoradiography scanning
with a densitometer (Shimadzu CS-9000).
The reaction mixtures (10 [mu]l) containing exonuclease III (1.2 [mu]M) and DNAs (0.12 [mu]M) in buffer (77 mM NaCl, 10 mM EDTA, 10 mM DTT, and 66 mM Tris-HCl, pH 8.0) were incubated at room temperature for 30 min.
To each sample, 3 [mu]l of saturated sucrose solution was added. The mixtures were analyzed by electrophoresis through a 15% native gel (29:1 ratio of acrylamide to bisacrylamide) in 0.4* TBE buffer at room temperature. The DNA bands in the gel were
visualized by autoradiography.
Oligonucleotides containing a synthetic abasic site and structures of synthetic abasic sites are shown in Figure
1
. The duplexes have 4mer overhangs at both 3' terminals to prevent 3'-5' exonucleolytic cleavage by exonuclease III. The single-stranded oligonucleotides containing an abasic site were not cleaved by
the 3'-exonucleolytic activity of exonuclease III (data not shown). Here we examined the
substrate specificity of exonuclease III using the synthetic oligonucleotides.
At the beginning, we reexamined the result that the phosphodiester bond 5' to the abasic site of a duplex DNA is hydrolyzed by exonuclease III in
spite of the abasic residue structure. The 5'-end labeled oligonucleotides containing a regular abasic residue
(D), an alkaline-stable analog (H) of the regular 2'-deoxyribose residue, an anucleosidic analog (P), or a 2'-deoxyribosyl formamide (F) were annealed with the complementary
oligonucleotide having dA opposite to the abasic site. These duplexes (abbreviations: D:A, H:A, P:A, and F:A, respectively) were incubated with exonuclease III and the products were analyzed by 7 M urea polyacrylamide gel electrophoresis (Fig.
2
a). At a low enzyme concentration, these oligonucleotides containing an abasic
site were cleaved to give the predictable 15mer oligonucleotide (lane 1). This result means that the enzyme cleaves on the 5'-side of AP sites to produce 3'-OH groups. At a high enzyme concentration, the 3'-exonucleolytic products from the 15mer
were observed as a ladder (lanes 2 and 3).
Based on the three-dimensional structure of exonuclease III, it has been presumed that an
unpaired base on the DNA strand opposite the abasic site might be recognized by
the enzyme (
29
). If the unpaired base plays an important role in the recognition of the AP
site by exonuclease III, the type of unpaired base must influence the degree of
binding and cleavage of exonuclease III. We prepared a series of duplexes constructed from the oligonucleotide containing an abasic site (H, P) and the complementary oligonucleotide (Fig.
1
). The duplexes are shown as H:N and P:N (N; A, T, G, or C), respectively. These
duplexes were incubated with exonuclease III and the products were analyzed by 7 M urea PAGE (Fig.
3
a). At a low enzyme concentration (lanes 1), the product (15mer) cleaved at the
phosphodiester bond 5' to the abasic site was observed. At a high enzyme concentration (lanes
2), the exonucleolytic products from the 15mer were observed. The binding of the enzyme to the
duplexes was examined. After the mixing of the enzyme in the duplex solutions,
the samples were analyzed using native PAGE (Fig.
3
b). The mobilities of the duplexes were reduced by the addition of exonuclease
III (>95%). These results indicated that the extent of endonucleolytic cleavage
of the duplexes and bindings to the duplexes were scarcely influenced by the
unpaired base opposite the abasic site. The slight difference in the
endonucleolytic cleavage (Fig.
4
b) might be attributable to differences in the structures in the vicinity of the
abasic site.
Since sequences in the vicinity of the abasic site and opposite the abasic site
were not absolutely required for the AP endonuclease activity of exonuclease
III, we speculated that a single-stranded DNA containing an abasic site might be cleaved by exonuclease
III. When the single-stranded oligonucleotides H and P were incubated with exonuclease III (40
nM), they were resolved at the 5' side of the abasic site with less efficiency than a duplex containing an abasic site (Fig.
4
a and b). As previously seen in the duplexes containing an abasic site, the cleavage sites of the single-stranded oligomers containing an abasic site were the same. In order to
confirm that a single-stranded oligonucleotide containing an abasic site is cleaved by
exonuclease III at the abasic site, we prepared another oligomer, d-GCGATGACTAACG
H
TACTAGGCTTCCGAGCC, as a substrate. When the single-stranded oligomer was incubated with exonuclease III (5 nM, 50 nM), it was also cleaved at the 5' side of the abasic site (data not shown). The binding of the
enzyme to the single-stranded oligomer containing an abasic site was then examined. After mixing of the single-stranded oligomer and the enzyme, the samples were analyzed using native PAGE (Fig.
4
c). Lanes H
-
and P
-
show that these oligonucleotides containing an abasic residue do not form a
self-aggregated duplex. No complex formation between the single-stranded oligomer and the enzyme was detected (lanes H
+
and P
+
). These results show that exonuclease III cleaves the 5' side of the abasic site even though the binding affinity of exonuclease
III to the abasic site on the single-stranded oligomer is weak. The presence of the hole induced by deletion of
a base residue from the DNA strand might be the key determinant for the
exonuclease III-mediated cleavage of DNA containing an abasic residue.
Our present study has provided significant clues for the AP site recognition
mechanism of exonuclease III. We have shown that not only double-stranded DNA but also single-stranded DNA containing an AP site was a good substrate for
exonuclease III despite the nucleotide sequence in the vicinity of the AP site.
These results show that this enzyme probably recognizes the space produced by
deletion of the base. Kow has examined the AP endonuclease activity of exonuclease III on PM2 DNA containing a number of different damaged base residues (
27
). Within Kow's double-stranded DNA substrates, residues such as thymine glycol, dihydrothymine, [beta]-ureidoisobutylic acid, and formamidopyrimidine, which are almost the same size as those of the
normal bases, were not recognized as substrates. Furthermore, he has shown that
the smaller the
O
-alkylhydroxylamine residue at the AP site, the higher the AP endonucleolytic activity of exonuclease III. It has been demonstrated that abasic residues such as deoxyribose,
deoxyribosylurea and tetrahydrofuran, and acyclic sugar residues such as deoxyribitol and propanediol corresponding to the carbon atoms of the phosphodiester backbone were cleaved by exonuclease
III (
22
-
24
). Using double-stranded DNAs having an original base sequence, we have reevaluated the
fact that the double-stranded DNA containing an abasic or an anucleosidic residue is a good
substrate of exonuclease III. Furthermore, the small formamide adduct to the
sugar moiety, which is produced by ionizing irradiation from the thymine
residue (
32
) or by hydroxyl radical degradation from a guanine residue (
33
), does not influence the AP endonuclease activity (Fig.
2
a). In conclusion, these results and previous information suggest that the abasic residue itself
is not recognized by exonuclease III. Weiss has proposed that the exonuclease
III recognizes a space created by the unwinding of a terminal base pair at the
ends of the duplexes and by removal of a base from the duplexes (
34
). It was then predicted that the enzyme might act as an endonuclease on the
unpaired or mispaired regions of the duplexes. In this study, it was shown that
the DNAs containing an AC mismatched base pair, an AA mismatched base pair, and
a bulge structure were not cleaved at all. These noncleavable DNAs did not form
stable complexes with exonuclease III (Fig.
2
b). It is not likely that the exonuclease recognizes a merely local structural
strain on the DNA.
The three dimensional structure of exonuclease III has been revealed at 2.6 Å resolution by X-ray crystallography (
29
). The catalytic domain, made up of Asp229, His259, and Glu34, is within a
pocket surrounded by Gln112, Asn153, Try109, Asn7, and Trp212. From this crystallographic data, it has been speculated that recognition of the AP site includes interaction with an unpaired base on
the DNA strand opposite the AP site and Gln112 or Asp153 through hydrogen
bonding or Trp212 through base stacking. However, we have elucidated that the
binding affinity and endonucleolytic activity of exonuclease III were scarcely
influenced by the type of base opposite the AP site (Fig.
3
). In addition, we have found that single-stranded DNA containing an AP site is a cleavable substrate by exonuclease
III (Fig.
4
). Using single-stranded DNA containing an AP site, we have obtained unambiguous evidence which reveals that the unpaired base opposite to the AP site is not essential for the recognition of the AP
site by exonuclease III.
Oligopeptides containing aromatic amino acids can preferentially form stacked complexes with single-stranded nucleic acids (
35
,
36
). The tripeptide lysyl-tryptophyl-[alpha]-lysine (Lys-Trp-Lys) efficiently binds to apurinic DNA by stacking of the
indole ring with nucleic acid bases in damaged regions (
37
). It has been shown that the tripeptide binds to DNA through efficient stacking
of the tryptophyl residue to the apurinic site followed by cleavage of DNA at the apurinic site (
38
). In the three dimensional structure of exonuclease III, it is noteworthy that
there is a tryptophan residue (Trp212) near the catalytic site (
29
). This tryptophan residue of the type II AP endonuclease from
E.coli
is conserved in the other type II AP endonucleases from various organisms such as
B.subtilis
(
18
),
S.pneumoniae
(
17
),
Drosophila
(
16
), mouse (
15
), bovine (
14
), and human (
13
) (Fig.
5
). However, there is not a conserved tryptophan residue in the type I AP
endonuclease. This tryptophan residue lies in the protruding loop preceding the
[beta]VI sheet. The amino acid sequences adjacent to the tryptophan residue are
also appreciably conserved.
We thank Mr Tadashi Mineo and Mr Tomoyoshi Ogawa for preparing oligonucleotides
containing an abasic residue.
*To whom correspondence should be addressed. Tel: +81 268 221215; Fax: +81 268
240921; Email: shida@pterus.shinshu-u.ac.jp
REFERENCES
Return