Clustered DNA damage, influence on damage excision by XRS5 nuclear extracts and Escherichia coli Nth and Fpg proteins.

Ionizing radiation and radiomimetic anticancer agents induce clustered DNA damage, which are thought to reflect the biological severity. Escherichia coli Nth and Fpg and nuclear extracts from XRS5, a Chinese hamster ovary Ku-deficient cell line, have been used to study the influence on their substrate recognition by the presence of a neighboring damage or an abasic site on the opposite strand, as models of clustered DNA damage. These proteins were tested for their efficiency to induce a single-strand break on a (32)P-labeled oligonucleotide containing either an abasic (AP) site, dihydrothymine (DHT), 7,8-dihydro-8-oxo-2'deoxyguanine, or 7, 8-dihydro-8-oxo-2'deoxyadenine at positions 1, 3, or 5 base pairs 5' or 3' to either an AP site or DHT on the labeled strand. DHT excision is much more affected than cleavage of an AP site by the presence of other damage. The effect on DHT excision is greatest with a neighboring AP site, with the effect being asymmetric with Nth and Fpg. Therefore, this large inhibition of the excision of DHT by the presence of an opposite AP site may minimize the formation of double-strand breaks in the processing of DNA clustered damages.

Radiation and radiation mimetic anticancer agents cause DNA damage, and it is thought that clustered damage (in which at least two damages are produced within less than 10 base pairs) is implicated in the biological severity of radiation since it is less repairable (1,2). The complexity of radiationinduced clustered DNA damage increases on increasing the ionizing density of the radiation (LET). 1 From track structure simulations, ϳ20% of double-strand breaks are associated with other damages for low LET radiation but increased to Ͼ20% for double-strand breaks induced by high LET ␣-radiation (3). Indirect experimental evidence supporting the role of DNA damage complexity comes from the reduced repairability of double-strand breaks induced in cellular DNA by high LET radiation (4,5) and the increased complexity of single-strand breaks on increasing radiation quality as revealed using cell extracts (6 -9). If base damages within a cluster are on opposite strands and both excised, this gives rise to double-strand breaks. Therefore, it is of great significance to understand the way in which several damages in close proximity are recognized/processed by the cell.
Although the chemical nature of oxidative damage produced by oxidative stress and by ionizing radiation are similar, the unique feature of ionizing radiation and radiation mimetic agents is their ability to produce clustered damage. There are only few studies about the excision of a damage substrate in the vicinity of another damage. In particular, synthesized oligonucleotides containing damage at specific sites were used to focus on the efficiencies of endonucleases VIII (Nei) and III (Nth) to excise either thymine glycol or DHT when opposite a singlestrand gap (10) as well as the efficiency of Fpg to excise 8-oxo-G or AP site opposite a gap (11) or 8-oxo-G near a formylamine on the same strand (12). Chaudhry and Weinfeld (13) focused on the efficiency of Nth to process either clustered abasic site or two DHTs on opposed strands within a plasmid.
Since major base modifications produced by ionizing radiation are DHT (through reduction pathway), 8-oxoG, and 8-oxoA (through oxidative pathways) (24), specific DNA constructions have been synthesized with these lesions inserted at precisely known positions and varied the positions systematically relative to each other or to an AP site. This study represents the first investigation to address the way in which clustered DNA base damage is removed by nuclear extracts (from Ku-deficient CHO cells, XRS5; see Refs. [25][26][27] in comparison with damage removal by Nth-and Fpg-purified enzymes. The nuclear extracts from the Ku-deficient CHO-derived cell line XRS5 were used to gain insight into the role of a composite of enzyme present in the nucleus of mammalian cells.

EXPERIMENTAL PROCEDURES
Substrate Oligonucleotides-The oligonucleotides were purchased from Genosys or Glen Research. The sequences of the 40-mer oligonucleotides are presented in Table I. Strand 2 contains either DHT, 8-oxoG, 8-oxoA, or a uracil (or the corresponding undamaged base as controls) at a fixed position (position X). Strand 1 contains a uracil (that could be remove by uracil-DNA-glycosylase (Ung) to give an abasic site), a DHT, or thymine as a control at given but variable positions (position Y) opposite to the fixed damage on strand 2. The oligonucleotides were 32 P 5Ј-end-labeled using 10 units of T4 polynucleotide kinase (Life Technologies, Inc.) with 50 Ci of [␥-32 P]ATP (6,000 Ci/mmol, 10 mCi/ ml, NEN Life Science Products) in 25 l of the recommended buffer for 1 h at 37°C. Following purification on a 12% denaturing polyacrylamide gel, the labeled oligonucleotide was hybridized with 1.5-fold excess of the purified non-radiolabeled complementary strand. That the annealing was efficient was verified by migration on a native 10% polyacrylamide gel. To prepare the oligonucleotides containing an abasic site at given positions, the 32 P-labeled double-stranded oligonucleotides containing a uracil were treated with 1 unit of uracil-DNA-glycosylase (Life Technologies, Inc.) for 30 min at 37°C in 50 l of buffer (10 mM Tris-HCl, pH 7.5, 50 mM NaCl, 1 mM EDTA).
Purified Proteins-The purified Nth protein was a generous gift from Prof. Rick Wood (Imperial Cancer Research Fund) (28). The purified Fpg protein was extracted and purified as described by Boiteux et al. (29).
Preparation of Nuclear Extracts-The nuclear extracts were prepared from a Ku-deficient CHO-derived cell line, XRS5 (25)(26)(27), to avoid possible interference by Ku binding to linear DNA (30). The cells were grown in exponential phase in ␣-complemented minimum Eagle's medium (ICN Biomedical Inc.) supplemented with 10% fetal calf serum (Glasform). The cells were harvested by centrifugation at 1,000 ϫ g for 10 min at 4°C, and the pellet was washed twice in 30 ml of phosphatebuffered saline volume of phosphate-buffered saline. The pellet of cells (1 volume) was then resuspended in an equal volume of buffer (10 mM HEPES, pH 7.9, 100 mM KCl, 1.5 mM MgCl 2 , 0.5 mM DTT) and incubated on ice for 15 min. The cytoplasmic membranes were broken by drawing 10 times into a 0.5-m diameter needle. After a brief centrifugation at 12,000 ϫ g, the supernatant was removed, and the nucleus pellet was resuspended in 2/3 volume of high salt buffer (20 mM HEPES, pH 7.9, 420 mM NaCl, 25% glycerol, 1.5 mM MgCl 2 , 0.5 mM DTT, 0.5 mM PMSF) for 30 min with agitation on ice. After a 10-min centrifugation at 12,000 ϫ g, the supernatant was dialyzed 3 times over a period of 2 h against 1 liter of incubating buffer (20 mM HEPES, pH 7.9, 100 mM KCl, 0.2 mM EDTA, 20% glycerol containing 0.5 mM DTT, 0.5 mM PMSF). The protein concentration was estimated using the Bradford colorimetric technique, and the aliquots of nuclear extracts were stored at Ϫ80°C.
Cleavage Assays for Single-strand Break Analysis-The doublestranded oligonucleotides (10,000 cpm, 200 fmol) were incubated with increasing amounts of Nth, Fpg proteins, or XRS5 nuclear extracts, as specified in the legends of the figures, in 5 l of the incubation buffer (20 mM HEPES, pH 7.9, 100 mM KCl, 0.2 mM EDTA, 20% glycerol) for 30 min at 37°C. Subsequently, 5 l of the denaturing stop solution was added (98% formamide, 0.025% bromphenol blue, 0.025% xylene cyanol, 2 mM EDTA, pH 8.0) to the samples that were then subjected to electrophoresis on a 12% denaturing polyacrylamide gel containing 8 M urea in 1ϫ TBE (89 mM Tris-HCl, 89 mM boric acid, 2 mM EDTA, pH 8.3) for 90 min at 85 watts. The dried gel was then exposed to a Bio-Rad PhosphorImager screen to visualize cleaved and full-length DNA fragments using phosphorimagery (Bio-Rad, Molecular Imager® FX) and quantified using Quantity One software (Bio-Rad) to determine the excision efficiency of each enzyme for each of the DNA sequences used.
FIG. 1. Excision efficiency of several substrates by Nth, Fpg, and XRS5 nuclear extracts. A, representation of the excision of a base damage (X) on the labeled strand (*) of a doublestrand oligonucleotide by increasing amount of protein and its visualization after migration on a 12% denaturing polyacrylamide gel. B-D, comparison of the percentage of removal of several substrates by increasing amount of Nth, Fpg, and XRS5 nuclear extracts, respectively. The standard errors are represented, and the lanes are used to determine the quantity of protein that removes 50% of each base damage.

TABLE I
Oligonucleotide sequences X and Y represent AP, DHT, 8-oxoG, 8-oxoA or the normal corresponding base (T for AP and DHT, G for 8-oxoG or A for 8-oxoA). N represents normal base complementary to the X base (A opposite to AP or DHT, C opposite to 8-oxoG:C or T opposite to 8-oxoA). Ϫ5 up to Ϫ1, position of the X base 3Ј from the Y base, on the complementary strand. ϩ1 up to ϩ8, position of the X base 5Ј from the Y base, on the complementary strand.
Measure of Double-strand Breaks Induced by the Proteins-The experiments were done as described above for the analysis of singlestrand breaks, except the reactions were stopped by adding 5 l of a non-denaturing solution (40% sucrose, 0.025% bromphenol blue, 0.025% xylene cyanol, 5 mM EDTA, pH 8.0). The samples were run on a 10% native polyacrylamide gel in 1ϫ TBE for 3 h at 300 V, dried, and quantified using the Bio-Rad PhosphorImager as described above.

RESULTS
Excision Efficiency of Different Damages by Nth, Fpg, and XRS5 Nuclear Extracts-The efficiency of the various proteins to remove specific single substrates from DNA was first compared as controls using denaturing gels. With a fixed concentration of double-stranded oligonucleotides containing one specific damage and increasing amounts of either purified proteins or XRS5 nuclear extracts, the excision efficiency was determined from the amount of cleaved DNA compared with the total amount of DNA present ( Fig. 1A and Table I). As presented in Fig. 1B, Nth is 100-fold more efficient at removing an AP site than DHT, from the relative concentrations of Nth to FIG. 2. Effect of the presence of another opposite damage on the AP site excision efficiency by Nth. The oligonucleotide containing a AP site is labeled and hybridized to a complementary strand containing or not (control) another damage opposite the positions between 5 bases 5Ј (Ϫ5) and 8 bases 3Ј (ϩ8) from the AP site (see Table I Table I. The diagrams reflected the fold inhibition/activation by comparison to the control containing no other damage on the strand opposite to the AP site from 3 to 5 different experiments. The error bars represented the standard deviation of the values presented on the right of the diagrams. give 50% cleavage of the DNA. As shown in Fig. 1C with Fpg, the AP site is also the most efficiently removed damage, whereas 8-oxoG and DHT are excised less efficiently (7.5-and 75-fold, respectively) than an AP site. It is worth noting that Fpg does not excise 8-oxoA from DNA even when using a high concentration (200 ng) of protein.
In an approach to mimic DNA damage processing in cells, XRS5 nuclear extracts have been used to determine its efficiency to excise AP site, uracil, 8-oxoG or DHT from DNA. Fig.  1D shows that an AP site is cleaved most efficiently by nuclear extracts, whereas uracil and 8-oxoG are excised less efficiently (70-and 40-fold, respectively) by nuclear extract than an AP site. DHT is the least efficiently excised damage, requiring 180-fold greater quantity of nuclear extract to cleave 20% of the DNA compared with that for excision of the AP site (Fig. 1D).
The 3Ј end of the AP site containing oligonucleotides obtained after cleavage of the AP site Nth, Fpg proteins, or XRS5 nuclear extracts migrates at a different position in gel (data not shown). These bands correspond to a 3Ј-phosphoaldehyde obtained by a ␤ elimination using Nth, to a 3Ј phosphate processed via a ␤-␦ elimination by Fpg (31-33) and 3Ј OH termini using nuclear extracts in which exonucleases or 3Ј-phosphatases clean up the 3Ј ends to give the 3Ј OH termini required by DNA polymerases to repair the strand break.
Effect of Neighboring Damage on the Protein Recognition of an AP Site-The efficiency of Nth to excise an AP site was determined when further damage is positioned on the nonlabeled strand at defined positions opposite to the AP site. In Fig. 2A, this other damage, an AP site, was situated from 5 bases 3Ј up to 8 bases 5Ј from the fixed AP site of interest (position X) on the labeled strand of DNA (see Table I). In Fig.  2, B and C, 8-oxoG and 8-oxoA were positioned 5, 3, or 1 base(s) 5Ј or 3Ј from the AP site on the labeled strand. The largest effect on excision of the AP site by Nth is a 2.1-fold inhibition due to the presence of an AP site one base 5Ј opposite to the AP site of interest (position ϩ1 in Fig. 2A). An AP site situated at any of the other positions as well as 8-oxoG, 8-oxoA, or DHT (Fig. 2, B- Table I). These oligonucleotides are incubated with increasing amounts of Nth (A), Fpg (B), or XRS5 nuclear extracts (C). The diagrams reflected the fold inhibition/activation by comparison of the control containing no other damage to an AP site on the opposite strand. Similar results were obtained using XRS5 nuclear extracts where 8-oxoG, 8-oxoA, or DHT do not significantly influence the efficiency of cleavage of the AP site excision (Fig. 4, B and C and data not shown, respectively). However, the presence of another AP site results in an enhancement of the excision efficiency of the AP site when the other AP site is positioned between 1 and 5 bases 5Ј to the AP site of interest on the complementary strand (Fig. 4A), with maximum activation seen when the other AP site is 3 bases away in the 5Ј direction (position ϩ3, Fig. 4A). This enhancement contrasts with the inhibition seen for excision by Fpg and Nth. No effect was observed if the other AP site is 3Ј or more than 8 bases 5Ј from the AP site on the labeled strand.
Therefore, the major effect on excision on an AP site is due to the presence of another AP site at various positions opposite the AP site at the fixed X position on the labeled strand presented above. To assess whether the inhibition of the cutting efficiency for both Nth and Fpg and the stimulation of AP excision by XRS5 nuclear extracts is strand-or sequence-specific, we performed similar experiments but using oligonucleotides that were 5Ј-labeled on the strand containing an AP site at the variable Y positions. The results are presented in Fig. 5. The inhibitory effect on excision of an AP site at the variable positions Y by Nth (Fig. 5A) is comparable with that at the fixed position X ( Fig. 2A) and similarly shows an inhibitory effect of about 2 for an AP site 1 base 5Ј to the AP site of interest. The inhibitory pattern of an opposite AP site on the excision of an AP site by Fpg is similar when the AP site is in position Y (Fig.  5B) or X (Fig. 3A) of the labeled strand. The inhibitory effect is up to 3.9 and 3.3 when the AP site is positioned 1 base 5Ј opposite to the AP site at position Y (Fig. 5B) or X (Fig. 3A), respectively. Similar concordance of results is obtained for XRS5 nuclear extracts using oligonucleotide labeled on the strand that contains the AP site at position X (Fig. 4A) or Y (Fig. 5C). Some minor variations in the level of activation are seen and may reflect differences in the sequence context of the AP site, but activation of the excision of the AP site by the nuclear extracts is clearly correlated to the relative positions of the two opposite AP sites. The inhibitory/activation effects by an AP site on the excision of an opposite AP site does not depend significantly on sequence context.
Effect of Neighboring Damage on the Excision of DHT by Nth-We have tested the ability of Nth, Fpg, and XRS5 nuclear extract to excise DHT from DNA when another damage is positioned on the other nonlabeled strand. The largest inhibitory effect of DHT excision by Nth is seen in the presence of an AP site opposite to the DHT, and the effect is asymmetrical (Fig. 6). The inhibition occurs when the AP site is between Ϫ3 and ϩ8 bases from the DHT site, with a maximum of 52-fold inhibition when the AP site is positioned 1 base 5Ј from the DHT (Fig. 6A). Uracil-containing oligonucleotides show a minor effect on the efficiency of excision of DHT by Nth (Fig. 6B). A minor effect was also observed with DHT opposite to DHT at the various positions shown in Fig. 6C. In contrast, 8-oxoG and 8-oxoA positioned 1 base 5Ј from the DHT (positions ϩ1, Fig. 6, D and E) have an inhibitory effect on the excision of DHT by Nth. Therefore, the inhibitory effect of the second damage is always asymmetrical in double-stranded DNA.
Since the presence of an AP site greatly inhibits the excision of DHT by Nth, we examined whether this inhibition of singlestrand break formation correlated with the formation of a double-strand break. Therefore, the same experiment was performed as in Fig. 6A, but the samples were divided as follows: one-half ran on a 12% denaturing polyacrylamide gel, and the other half ran on a 10% native polyacrylamide gel. The percentage of single-or double-strand breaks induced by Nth in the DNA-specific oligonucleotides is represented in Fig. 7, A  and B, respectively. A high yield of single-strand breaks but no double-strand breaks are produced in the control that contains only one damage. Fig. 7, A and B, shows similar dependence on the influence of the other damage for excision of the damages to give single-and double-strand breaks. Only oligonucleotide Ϫ5 has fewer double-strand breaks in comparison with the level of single-strand break. This is probably due to the fact that the 5 base pairs, between the two single-strands breaks, are not properly dehybridized in the sample and during the migration.
Effect of a Neighboring Damage on the Excision of DHT by Fpg-As shown in Fig. 8A, the largest inhibitory effect for Fpg excision of DHT occurs when an AP site is positioned opposite to the DHT of interest. This inhibition is asymmetrical and is greatest when the AP site is positioned 3 or 1 bases 3Ј to the DHT (positions Ϫ3 and Ϫ1) with 50-and 43-fold inhibition, respectively. In contrast to the results with Nth, the removal of DHT by Fpg is also inhibited 8.4 -12.2-fold by the presence of uracil located between 5 bases 3Ј to 5 bases 5Ј of the DHT on the opposite strand, as presented in Fig. 8B. The presence of an 8-oxoG opposite to the DHT also has an inhibitory effect and is maximum when 8-oxoG in positioned at 3 bases 3Ј or 5Ј of the DHT (positions Ϫ3 and ϩ3 in Fig. 8D, respectively). In contrast, the presence of another DHT or an 8-oxoA, 1 base 5Ј (position ϩ1, Fig. 8C) and 1 base 3Ј to the DHT (position Ϫ1, Fig. 8E), respectively, on the labeled strand gives 2.1-and 3.6-fold activation. No effect was seen when DHT or 8-oxoA is present at any other site opposite to the DHT of interest.
By using double labeling probes containing an AP site opposite to DHT, the same level of DHT or AP site was excised by Nth or Fpg as that observed with the singly labeling probes described above (Figs. 2D and 6A and Figs. 3D and 8A). In the same way, identical results were obtained using 5Ј-32 P-labeled probes on the DHT-containing strand that was hybridized to the complementary oligonucleotides containing an AP site with either a 5Ј-OH or a 5Ј-P termini. These results suggest that the presence of 5Ј OH or 5Ј-P termini does not change the inhibitory effect of the presence of an AP site on the excision of DHT by Nth or Fpg.
Influence of Another Damage on the Excision of DHT by XRS5 Nuclear Extract-As with the purified Nth and Fpg proteins, the presence of an AP site opposite to a DHT causes a large decrease of the efficiency of the excision of DHT by the XRS5 nuclear extract (Fig. 9A). In contrast to the findings with Nth and Fpg (Figs. 6A and 8A, respectively), the effect is symmetric. The largest inhibition is seen when the AP site is 5 bases 3Ј or 5Ј of the DHT (positions Ϫ5 and ϩ5 in Fig. 9A), and the inhibition is less when the AP site is at positions Ϫ1 and ϩ1 to DHT (Fig. 9A). The effect of uracil on DHT excision (Fig. 9B) gives the same inhibitory profile as that of an AP site, but the level of inhibition is significantly less. The presence of either another DHT or 8-oxoG does not have any significant effect on the action of the nuclear extracts (Fig. 9, C and D). However, it is worth noting that, as for the incubation with Fpg (Fig. 8D), the profile of the effect in the presence of 8-oxoG is the same but less efficient when incubated with nuclear extract (Fig. 9D).

FIG. 7. Single-strand and doublestrand break formation by Nth.
Increasing amounts of Nth (250 pg up to 50 ng) from an oligonucleotide containing a DHT on the labeled strand and an AP site at various positions on the opposite strand. The reaction is performed as for Fig. 3A, and the samples are separated on a 12% denaturing polyacrylamide gel to measure the percentage of single-strand breaks (A) or on a 10% native polyacrylamide gel to determine the level of double-strand breaks induced by Nth (B). The lanes Ϫ5 up to ϩ5 and represent the different positions of the AP site from 5 bases 5Ј up to 5 bases 3Ј from the DHT as described in Table I; lane C represents the control DNA containing only one DHT in A but is not presented in B since it does not give any double-strand break.
The only significant effect of 8-oxoA is when positioned 1 base 3Ј (position Ϫ1, Fig. 9E) opposite to the DHT, giving a 2.1-fold inhibition.

DISCUSSION
The purpose of the present study was to determine the effect of neighboring DNA damage on the excision of a specific DNA damage to gain an insight into the effects of clustered DNA damage on DNA damage repair. Purified proteins involved in the DNA damage recognition and excision and nuclear extracts, as a composite of the cellular proteins, were used to assess how such clustered damage may compromise their efficiency to be repaired in cells even though the individual damages are readily processed (Fig. 1). This study is the first to assess the recognition and processing of complex damage by cell extracts, as a more complex model of repair, and compare the effects with base excision repair enzymes, used as benchmarks. For instance, the AP lyase activity of Nth and Fpg is about 1 order of magnitude greater than their glycosylase activities when comparing the cleavage of an AP site with that for excision of a base damage. The cleavage of an AP site by the extract was also more efficient than cleavage as a result of excision of the base damages. These effects are similar to those using Fpg, but quantitatively, the differences are greater with the extracts. This difference may reflect that only a few proteins are able to excise 8-oxoG such as Fpg or OGG1 proteins in the cellular environment (34) in comparison with the larger number of proteins that remove AP sites from DNA. The differences between the AP site and the excision of uracil (70-fold) is probably due to the presence of uracil-DNA-glycosylase in the nuclear extracts that convert uracil into an AP site that is subsequently cleaved by proteins.
With the purified proteins, the end group termini (data not shown) are consistent with the AP lyase activity of Nth involving a ␤-elimination process as described by Bailly and Verly (35). The 3Ј OH end termini produced by cleavage of AP sites by the nuclear extract are consistent with the clean up of the end termini.
The biggest effect on the efficiency of cleavage of an AP site, reflecting the AP lyase activity, by Nth, Fpg, and XRS5 nuclear extracts is the presence of a neighboring AP site on the complementary strand. The large inhibition (Ͻ4-fold) of the AP lyase activity occurs with Fpg when another AP site is situated from Ϫ5 to ϩ8 of the AP site monitored (Figs. 3A and 5B). The cleavage of an AP site by Nth is only affected by an immediately neighboring AP site. The cleavage of an AP site by Nth, Fpg, and extract is not greatly affected by the presence of a vicinal base damage, DHT, 8-oxoG, and 8-oxoA. For nuclear extracts, the efficiency of cleavage of an AP site is enhanced when another AP site is positioned opposite at ϩ1, ϩ3, and ϩ5 and is not significantly dependent on the sequence context as demonstrated in Fig. 5.
The major inhibition (Ͼ50-fold), however, is seen on conver- sion of DHT into a single-strand break by Nth or Fpg (see below for DHT excision by Fpg) when an AP site is on the opposite strand and the influence extends over several base pairs of separation (Figs. 6A and 8A). Since the inhibitory effect of a neighboring AP site on excision of DHT is greater than on cleavage of an AP site, it is suggested that the AP site on the nonlabeled strand is rapidly converted into a single-strand break since both Nth and Fpg are more efficient at cleavage of an AP site than excision of DHT (Fig. 1, B and C). From comparison of the yields of single-strand and double-strand breaks (Fig. 7) following Nth excision of DHT, it is apparent that inhibition of double-strand break formation is of the same magnitude and trend as that for formation of a single-strand break. Therefore, it is suggested that a single-strand break on the non-labeled complementary strand at the AP site position is already formed prior to removal of the DHT. This finding is comparable with that for the removal of thymine glycol opposite a gap by Nth (10). However, the inhibitory effect of an AP site on the excision of thymine glycol is symmetrical in contrast to the asymmetry seen for excision of DHT opposite to an AP site by Nth. In contrast to the large inhibition by an AP site on DHT excision, the majority of the inhibition of an opposing base damage (DHT, 8-oxoA, 8-oxoG) or uracil by Nth occurs when the second damage is at ϩ1 site of the DHT of interest. The lack of effect of 8-oxoG on the excision of DHT on the opposite strand, consistent with previous studies (10), is when 8-oxoG is more than 1 base 5Ј or at every position 3Ј to the DHT on the complementary strand (Fig. 6D). The asymmetrical effects observed may be correlated to the asymmetrical binding of Nth to the strand opposite to the recognized damage. Indeed, DNase I footprinting experiments (36) show that Nth protects an AP site from 4 bases 3Ј to 4 bases 5Ј on the AP-containing strand and, on the complementary strand, from 1 base 3Ј to 9 bases 5Ј from the AP site.
During the course of this study, it became apparent that DHT is a substrate for the Fpg protein (Fig. 1C, see Ref. 37) that usually recognizes purine damages, although some pyrimidine damages such as 5-hydroxycytosine and 5Јhydroxyuracil have previously been shown to be substrates for Fpg (20). Similar to the finding with Nth, the inhibitory effect on DHT excision of a neighboring base damage by Fpg is significantly less than that for the neighboring AP site. With Fpg, the largest effect on inhibition of the excision of DHT occurs when an AP site is positioned 3Ј to the DHT on the opposite strand with up to 50-fold inhibition. This effect is asymmetric, but the asymmetry is not related to the binding motif of Fpg that protects an AP site from DNase I footprinting by binding 2 bases 3Ј and 5Ј of the AP site but only one base opposite the AP site on the complementary strand (38,39). The influence of other damages is much more variable than that obtained with Nth. In particular, the presence of a uracil on the opposite strand has a 8.4 -12.2-fold inhibitory effect on DHT excision by FIG. 9. Effect of the presence of a neighboring damage on the DHT excision by XRS5 nuclear extracts. Double-stranded oligonucleotides as described previously (Fig. 6) are incubated with increasing amounts of XRS5 nuclear extracts (0.5 to 30 g). The diagrams reflected the fold inhibition/activation by comparison to the control containing no other damage on the strand opposite to the DHT. The white bars, gray bars, and black bars represent the opposite damage on the nonlabeled strand at positions 1, 3, or 5 nucleotides, respectively, in 5Ј or 3Ј from the DHT on the labeled strand. Triplicates reactions were performed, and the standard deviation are presented. Fpg (Fig. 8B) but no effect on DHT excision by Nth (Fig. 6B). This inhibition could allow the action of the Ung protein to form an abasic site that will be quickly converted into a singlestrand break. After repair of this break, the Fpg protein can then act more efficiently on the DHT opposite a normal strand.
With XRS5 nuclear extracts, the major effect on excision of a DHT is the presence of an AP site (Fig. 9A) or uracil (Fig. 9B), in contrast to the small effects seen with either a neighboring DHT, 8-oxoG, or 8-oxoA. In contrast to the purified enzymes, the largest inhibitory effects of a vicinal damage are seen on base damage excision and not on cleavage of an AP site. The effect of uracil on the excision of DHT by XRS5 nuclear extracts gives a similar pattern to that for the influence of an AP site on DHT removal, which does not show an asymmetrical response as observed using both Nth and Fpg. It is suggested that the effect of nuclear extract is much more complicated and results from a combination of effects of the enzymes that remove AP sites from DNA. The effect of 8-oxoG on cleavage of an AP site by the nuclear extract gives the same qualitative dependence as that with Fpg. This suggests that the effect with nuclear extracts might be due to the combined effects of an Fpg-like protein and Nth to give the lower inhibition. It is of interest that certain clustered damages could give an increased efficiency of excision of a base damage within the cluster. This is particularly apparent for a DHT 1 base 5Ј or an 8-oxoA 1 base 3Ј to DHT on the complementary strand with Fpg and an AP site 1 base up to 5 bases from another AP site with XRS5 nuclear extracts.
It is suggested that the big inhibitory effect of an opposite AP site on DHT recognition by all the proteins tested may minimize the probability of formation of a double-strand break so that the repair enzymes have time to rejoin the gap, generated at the AP site, prior to excision of DHT from the opposite strand. Therefore, it will be on interest to test further these combinations of clustered damages to determine if they increase the probability to give a double-strand break and how they affect the repairability of the damaged base, particularly with nuclear extracts.