ABSTRACT
HAP1 is a divalent cation-dependent endonuclease from human cells with specificity for
apurinic/apyrimidinic (AP) sites in DNA. Extraction of the essential metal ion
from purified HAP1 stabilized its binding to an oligonucleotide containing a single AP site, permitting AP site binding studies to be undertaken using gel retardation assays. Binding
of HAP1 to such an oligonucleotide was dependent upon the presence of an AP
site. Previous structural and modelling studies have suggested a role for
Asn212 (Asn153 in exonuclease III, the bacterial homologue of HAP1) in substrate recognition. Substitution of alanine for Asn212 abolished the AP endonuclease activity of purified recombinant HAP1 protein.
More conservative substitutions of aspartate or glutamine for Asn212 still led to a reduction in specific activity of at least 300-fold. Moreover, none of the three Asn212 substitution mutants of HAP1 possessed detectable AP site binding activity
in vitro
. This study indicates that chelation of the active site metal ion in HAP1
stabilizes the complex of the protein with AP sites and identifies an active
site asparagine residue as an important component of AP site recognition by the
HAP1 protein.
Apurinic/apyrimidinic (AP) sites are biologically important lesions in DNA for three reasons. First, they are amongst the most common forms of
damage to arise in cellular DNA and are generated both by spontaneous
hydrolysis of the
N
-glycosyl bond linking the base to the sugar-phosphate backbone and by exposure of cells to DNA damaging agents.
Secondly, they are non-instructional to DNA polymerases and are, therefore, potentially pro-mutagenic. Thirdly, they can inhibit DNA synthesis and, as a consequence, induce cell death (for reviews see
1
-
6
). All organisms studied thus far express dedicated DNA repair enzymes, termed
AP endonucleases, that initiate repair of AP sites (
2
-
4
,
6
). The major class of these enzymes cleave the phosphodiester backbone 5' of the AP site via a hydrolytic mechanism to generate a DNA single-strand break with a 3'-hydroxyl and a 5'-deoxyribose phosphate terminus. AP site repair is then completed by
phosphodiesterases, DNA polymerases and DNA ligases (
2
-
4
,
6
).
HAP1 (also called APE, APEX and Ref-1) is the major AP endonuclease expressed by human cells (
7
-
10
). This 37 kDa protein is a structural and functional homologue of the
Escherichia coli
exonuclease III protein (
7
,
8
,
11
,
12
). Recent studies have gone some way towards improving our understanding of the
precise catalytic mechanism of action of both exonuclease III and HAP1. The
high resolution crystal structure of exonuclease III revealed that three amino
acid residues (Glu34, Asp229 and His259) are necessary either for binding of a
divalent cation or for activation of a water molecule to a nucleophile that
cleaves the scissile phosphate group of the DNA (
11
). Mutagenesis studies on the HAP1 cDNA have confirmed that the equivalent
residues in the HAP1 protein (Glu96, Asp283 and His309) are important for
catalysis (
12
,
13
). The role of the essential metal ion is probably not to directly activate a
water molecule, but rather to facilitate cleavage of the phosphodiester
backbone by correctly orientating and/or polarizing the P-O3' bond (
11
,
12
).
Little is known about the precise mechanism by which HAP1 or its homologues
recognize their DNA substrate, since the crystal structure of an AP
endonuclease complexed with DNA has not been reported. Mol
et al
. (
11
) have presented the structure of the ternary complex of exonuclease III with a
metal ion and dCMP. Based upon that structure, it was suggested that at least
two amino acids are important for interactions with the dCMP moeity. The side
chain of one of these residues, Asn153, was proposed to hydrogen bond both with
the nucleotide O3' position and with the 5'-phosphate group of dCMP (
11
). The residue equivalent to Asn153 is conserved amongst all members of the
exonuclease III family of AP endonucleases and presumably, therefore, performs an important role in the action of these enzymes (for a review of AP endonuclease sequence similarities see
6
).
Using site-directed mutagenesis to introduce point mutations into the HAP1 cDNA, we
have analysed whether Asn212, the residue equivalent to Asn153 in exonuclease
III, is vital for the enzymatic activity of HAP1. More specifically, we have addressed whether mutants lacking this residue have an altered ability to interact with
DNA containing an AP site. In order to do this, we have developed a gel
retardation assay for AP site recognition by the HAP1 protein which exploits
the observation that AP site binding is stabilized by extraction of the
essential metal ion from the active site of the HAP1 protein. We show that
Asn212 is necessary for the AP endonuclease activity of the HAP1 protein and
that mutants carrying amino acid substitutions at position 212 fail to bind efficiently to AP site-containing oligonucleotide substrates.
The HAP1 protein coding region was amplified using the polymerase chain reaction (PCR). The following primers were used to incorporate 5'
Xho
I sites (shown in bold):
5'-primer
AGAGAG
CTCGAG
CCGAAGCGTGGGAAAAAGGGA
3'-primer
AGAGAG
CTCGAG
GGTGTCACAGTGCTAGGTATAG
PCR was performed using 1 [mu]g HAP1 cDNA template with 10 rounds of amplification consisting of: 92oC, 30 s; 50oC, 30 s; 72oC, 1 min. The PCR product was electrophoresed on a 1% agarose
gel and the band was excised and purified using a Geneclean kit (BIO 101). The
product was then digested with
Xho
I and subcloned into the
Xho
I site of pET-14b (Novagen). This vector encodes a leader peptide incorporating an N-terminal hexahistidine tag that allows purification of the
recombinant fusion protein through a Ni
2+
charged His[middot]Bind column. The construct was verified by nucleotide sequencing of the
HAP1 cDNA using the dideoxy chain termination method with Sequenase enzyme (US Biochemical Corp).
Mutagenesis was performed by the PCR-based method of Landt
et al.
(
14
), as previously described by Barzilay
et al
. (
12
,
13
). The primers used to create each mutant form are detailed below (altered
codons shown in bold):
HAP1-N212A primer
TTCATGTGCCAC
AGC
GAGGTCTCCACA
HAP1-N212D primer
TTCATGTGCCAC
ATC
GAGGTCTCCACA
HAP1-N212Q primer
TTCATGTGCCAC
TTG
GAGGTCTCCACA
The mutant forms of HAP1 cDNA were cloned into the
Xho
I site of pET-14b for high level expression and purification, as described below.
The wild-type and mutant pET-14b/HAP1 constructs were transformed into BL21.pLysS cells, an
E.coli
strain incorporating a chromosomally integrated T7 polymerase gene under control
of the
lac
promoter. The cells were grown at 30oC in a shaking incubator until the culture reached an OD
600
of 0.5. Expression of HAP1 from the T7 promoter was induced for 2 h by addition
of 0.4 mM IPTG (final concentration). The cells were then harvested, lysed by
sonication and the cell debris was pelleted by ultracentrifugation (39 000
r.p.m., 4oC, 20 min in a Beckman 45 Ti rotor). The supernatant was then loaded onto a phosphocellulose P11 column and the bound proteins were eluted with a 100-1250 mM NaCl gradient. The fractions containing the HAP1 protein (eluting
at ~600 mM NaCl) were pooled and extensively dialysed against binding buffer (5
mM imidazole, 500 mM NaCl, 20 mM Tris-HCl, pH 7.9).
The dialysed protein was loaded onto a Ni
2+
charged His[middot]Bind resin column (Novagen) and the column washed with 10 vol binding
buffer, followed by 6 vol wash buffer (60 mM imidazole, 500 mM NaCl, 20 mM Tris-HCl, pH 7.9). Bound proteins were eluted with 2 vol elution buffer (1 M
imidazole, 500 mM NaCl, 20 mM Tris-HCl, pH 7.9). The purified proteins were stored at -70oC in storage buffer (50% glycerol, 125 mM NaCl, 20 mM Tris-HCl, pH 7.9). All proteins were >95% pure as judged by
electrophoresis on 12% SDS-polyacrylamide gels (
15
) and staining with Coomassie blue.
The AP endonuclease activities of the native and mutant HAP1 proteins were
assayed using a double-stranded oligonucleotide labeled at the 5'-end with
32
P and containing a single AP site, as described by Barzilay
et al
. (
12
,
13
). Reactions were preformed as described previously (
12
,
13
) and the reaction products were electrophoresed on a 12% denaturing
polyacrylamide gel. The gels were dried onto Whatman 3MM paper and exposed to a
32
P-sensitive phosphor screen (Molecular Dynamics). The screens were scanned on a PhosphorImager 425 (Molecular Dynamics) and analysed
using the ImageQuant software. The amount of cleavage was quantified by
reference both to the uncleaved strand (to equalize for gel loading) and to a
positive control for AP site cleavage (incubation with 500 mM piperidine, 98oC, 30 min).
This assay utilizes the same
32
P-labeled, double-stranded oligonucleotide that was used for the AP endonuclease assay
(
12
,
13
). The oligonucleotide contained a single internal uracil, shown underlined and
in bold below.
Oligo
A5'-GATCTGATTCCCCATCTCCTCAGTTTCACT
CGCATG
Oligo
B5'-CGGTGCAGAAGTGAAACTGAGGAGATGGGGAATCA
An internal AP site was generated by excision of the uracil using uracil-DNA
glycosylase. This reaction was performed in a 10 [mu]l volume containing 0.5 ng oligonucleotide, 0.002 U human uracil-DNA
glycosylase (kindly supplied by Dr G. Slupphaug) and 1* reaction buffer (60 mM Tris-HCl, pH 8.0, 1 mM DTT, 100 [mu]g/ml BSA) per sample. The mixture was incubated at 37oC for 15 min.
AP site binding was assayed with 0.5 ng AP site-containing oligonucleotide and various concentrations of HAP1 proteins in
a 20 [mu]l reaction containing 1 [mu]g poly(dI[middot]dC) and 1* reaction buffer (4 mM EDTA, 4% glycerol, 10 mM HEPES-KOH, pH 7.9, 50 mM NaCl, 100 [mu]g/ml BSA). The samples were incubated on ice
for 20 min, before loading onto a 5% polyacrylamide gel. Samples were
electrophoresed (without the addition of loading dye) at 50 V in cold 0.5* TBE buffer. A control lane containing only loading dye was run on each
gel to monitor the progress of the electrophoresis. The samples were then dried
onto Whatman 3MM paper and exposed to X-ray film.
Non-specific DNA binding was assayed as described above for AP site binding,
but reactions contained 0.5 ng oligonucleotide that had not been treated with
uracil glycosylase, together with various HAP1 protein concentrations. These
reactions did not contain poly(dI[middot]dC).
We have reported previously that purified HAP1 protein does not form a stable
complex
in vitro
with DNA containing AP sites (
13
). This is presumably because the enzyme interacts only in a transient manner with its substrate during the catalysis of phosphodiester bond cleavage. We investigated whether by inactivating the catalytic
activity of the enzyme via chelation of the essential active site divalent
cation we could detect a more stable complex of the HAP1 protein with an oligonucleotide containing an AP site. Figure
1
shows that a specific complex of HAP1 with an oligonucleotide containing a
single AP site could be detected using gel retardation assays when the enzyme
was treated with 4 mM EDTA. This specific binding occurred even in the presence
of 1 [mu]g non-specific competitor, poly(dI[middot]dC).
Mol
et al
. (
11
) have presented the structure of the ternary complex of
E.coli
exonuclease III with Mn
2+
and dCMP. In this structure the dCMP moeity was shown to be hydrogen bonded to
an asparagine residue (Asn153) located in the active site of the enzyme.
Inspection of the amino acid sequence of the members of the exonuclease III
family of AP endonucleases indicated that the residue equivalent to Asn153 in
exonuclease III was completely conserved (Fig.
2
). To investigate the role of this asparagine residue (Asn212 in HAP1) in AP
site recognition/binding, we mutated the HAP1 cDNA to replace Asn212 in the
HAP1 protein with three alternative amino acids. The Asn212 residue was
converted to glutamine (HAP1-N212Q), aspartic acid (HAP1-N212D) or alanine (HAP1-N212A) and the mutant proteins containing these single amino
acid substitutions were purified in each case to near homogeneity in
E.coli
(Fig.
3
). Each of the mutant proteins was fully soluble in
E.coli
and had chromatographic properties similar to that of the wild-type recombinant HAP1 protein. These data, taken together with those
indicating that the three mutant proteins were still capable of binding non-specifically to undamaged DNA (see below), indicate that the overall
tertiary structure of the mutant proteins was not significantly perturbed.
To quantify the specific activity of the mutant proteins, we performed AP
endonuclease assays using an oligonucleotide substrate containing a single AP
site (as reported previously;
13
). Figure
4
shows that each of the mutant proteins had a dramatically reduced specific
activity as an AP endonuclease. The HAP1- N212A mutant lacked any detectable AP endonuclease activity, while the
HAP1-N212Q and HAP1-N212D mutants retained a low level of activity (300- and 900-fold reductions respectively, compared with wild-type HAP1).
In order to investigate whether the deficiency in AP endonuclease activity
exhibited by the Asn212 substitution mutants was a reflection of an inability
of the mutant proteins to interact either non-specifically with DNA or specifically with an AP site, we performed gel
retardation assays using the HAP1-N212Q mutant. We have shown previously (
13
) that HAP1 can bind with low affinity to `undamaged' DNA in reactions lacking
non-specific competitor nucleic acid [poly(dI[middot]dC)]. In that study it was shown that at concentrations of HAP1 >5
[mu]g/ml, the oligonucleotide substrate was retarded to the extent that it
remained within the wells of an acrylamide gel (
13
). The HAP1-N212Q mutant, when present at concentrations >5 [mu]g/ml, retained the ability to retard an undamaged oligonucleotide,
indicating that loss of Asn212 did not adversely influence the ability of HAP1 to bind non-specifically to DNA (data not shown). In contrast, even at concentrations as high as
50 [mu]g/ml, the HAP1-N212Q mutant protein failed to form a stable complex with an
oligonucleotide containing an AP site. In the same experiment, the wild-type HAP1 protein was shown to bind to this substrate (Fig.
5
). This AP site binding assay was performed under conditions that were identical
to those described in Figure
1
, using reaction mixtures containing both 4 mM EDTA and poly(dI[middot]dC) competitor. Very similar data to those generated with the HAP1-N212Q protein were obtained with the HAP1-N212D and HAP1-N212A mutants (data not shown).
We have investigated the binding of the HAP1 DNA repair enzyme to its principal DNA substrate, the AP site. We have shown that catalytically
inactivated HAP1 protein, lacking a metal ion in the active site pocket, can
form a complex with an AP site-containing oligonucleotide that is sufficiently stable to permit detection using gel retardation assays. Furthermore, we have identified an amino acid
residue, Asn212, that performs an important function in AP site
recognition/binding by the HAP1 enzyme.
There is compelling evidence that metal ions play an essential role in the
catalytic action of AP endonucleases and of many other hydrolytic nucleases (
11
,
12
,
16
-
21
). Structural analyses on
E.coli
exonuclease III revealed the presence of a single metal ion within the proposed catalytic active site bound to the carboxylate side chain of a glutamate residue (Glu34, equivalent to Glu96 in HAP1; see
11
). This glutamate residue, as well as the four flanking amino acids, are
conserved in all members of the exonuclease III family of AP endonucleases (
6
) and previous site-directed mutagenesis studies have confirmed that Glu96 plays a vital role
in metal ion binding in HAP1 and is essential for high level AP endonuclease
activity (
12
). Moreover, Gu
et al
. (
22
) have shown that the equivalent residue in Rrp1, the
Drosophila
homologue of HAP1, is important for enzymatic activity and for the ability of
Rrp1 to functionally substitute for exonuclease III in
E.coli
. The simplest interpretation of the data presented here is that extraction of
the metal ion from the active site of HAP1 has little or no effect on substrate
recognition, but nevertheless prevents phosphodiester bond cleavage. This would
be consistent with the proposed role of the metal ion in the catalytic action of exonuclease III and HAP1, in which the positive charge of the cation
leads to polarization of the P-O3' bond, facilitating nucleophilic attack on this bond by an
activated water molecule. It is possible that the metal ion plays a dual role
by also acting to accurately position the scissile phosphate in the enzyme
active site. We would suggest that in the presence of EDTA the rate of
apoenzyme dissociation from an AP site is reduced significantly, since the process is presumably triggered by DNA
cleavage under normal circumstances. However, this cleavage reaction is prevented in the absence of an active site cation.
In gel retardation assays that were performed in the presence of the non-specific competitor poly(dI[middot]dC) we have shown that formation of a stable complex between HAP1 and an oligonuceotide is only observed when the DNA contains an AP site. Nevertheless, it is clear from the current and previous work (
13
) that HAP1 can bind with low affinity to undamaged DNA. However, we have shown that at molar ratios of undamaged DNA to AP site-containing DNA of up to 500-fold, the binding of HAP1 to an AP site cannot be competed off
(unpublished data). Since all available evidence indicates that HAP1 and other AP endonucleases perform the dual role of damage recognition and phosphodiester backbone cleavage without the requirement for auxilliary proteins, the success of these enzymes in rapidly searching complex genomes for AP sites
must presumably be dependent upon an ability to efficiently discriminate between damaged and undamaged DNA.
We have shown that substitution of alanine for Asn212 reduces both AP endonuclease activity and AP site binding to a level that is below the lower limit of detection in our assays. This is consistent with structural data indicating that the equivalent residue (Asn153) is located within the active site of
E.coli
exonuclease III (
11
). Although the precise role of Asn212 is yet to be defined, the combined
results of Mol
et al
. (
11
) and those presented here strongly suggest that Asn212 and its equivalent
residues are intimately involved in substrate recognition by this family of AP
endonucleases. Although data from structural studies on exonuclease III (or
HAP1) in the presence of an AP site-containing oligonucleotide are not yet available, the structure of the
ternary complex of exonuclease III with Mn
2+
and dCMP revealed that Asn153 makes hydrogen bonding contacts with the dCMP
moeity. This suggests that the ability of this asparagine residue to hydrogen
bond to components of the DNA substrate are important for AP site binding. The
amino acid residues used to substitute for Asn212 were chosen in one case to
remove hydrogen bonding potential via elimination of a polar side chain (Asn -> Ala), in the second case to retain hydrogen bonding potential and the
amide-containing polar side chain, whilst altering side chain length (Asn -> Gln), and in the third case to convert the amide terminal group to a
carboxylate (Asn -> Asp). Whilst it was perhaps predictable that the Asp -> Ala substitution eliminated activity, it was more surprising to find that the two
apparently conservative substitutions had such a drastic effect on both the
specific activity and the AP site binding properties of the mutant enzymes.
These data indicate that the HAP1 protein will not tolerate even minor
modifications within the region of the active site pocket that contains the
Asn212 side chain.
In summary, we have developed a gel retardation binding system for the study of
AP site recognition by the HAP1 protein and have shown that Asn212 is an
essential component both for AP site binding and for the AP endonuclease
activity of HAP1. We are now in a position to further delineate the molecular
interactions that are required for recognition of an AP site by this DNA repair
enzyme and to investigate whether the sequence flanking the AP site or, more
specifically, the residue opposite the AP site influences the efficiency of AP site recognition or phosphodiester bond cleavage by the HAP1 protein.
We thank Drs P. Gallinari and J. Jiricny for alerting our attention to the
effects of EDTA on HAP1 DNA binding and to Dr J. Tainer for information
regarding the structure of exonuclease III. We also thank members of the
Molecular Oncology Laboratory for useful discussions, Drs G. Barzilay and C. Norbury for critical reading of the
manuscript and E. Clemson for typing the manuscript. This work was supported by the Imperial Cancer Research Fund.
*To whom correspondence should be addressed. Tel: +44 1865 222 417; Fax: +44
1865 222 431; Email: hickson@icrf.icnet.uk
REFERENCES
Return