Thymines opposite to bulky aristolactam-DNA adducts in duplex DNA are not targeted by human thymine-DNA glycosylase

Proteins

Phage T4 polynucleotide kinase was purchased from New England Biolabs (Evry, France). The E. coli BL21 Arctic Express (DE3) cells were purchased from Novagen-EMD Biosciences (Merck Chemicals, Nottingham, UK). The purified human uracil-DNA glycosylase (hUNGΔ84) was from laboratory stock (Morera et al., 2012).

The native full-length TDG^1–410 (TDG^FL) protein was produced as described previously (Talhaoui et al., 2013). In short, Arctic Express (DE3) cells were transfected with the expression vector pET28c-TDG^FL (for full-length TDG protein) and the transformants were grown on a shaker at 37 °C in LB, accompanied with 50 µg/ml of kanamycin, to OD_{600 nm} = 0.6–0.8. Thereafter, the temperature was decreased to 12 °C, the recombinant protein expression was induced by 0.2 mM isopropyl β-D-1-thiogalactopyranoside and the cultures were grown for another 15 h. The cells were collected by centrifugation and the bacterial pellets were lysed by a French press at 18,000 psi in buffer composed of 40 mM NaCl, 20 mM HEPES–KOH (pH 7.6) and 0.025% Nonidet P-40, complemented with Complete™ Protease Inhibitor Cocktail (Roche Diagnostics, Switzerland). The supernatants were collected after centrifugation of cell lysates at 40,000× g for 1 h at 4 °C, and adjusted to 20 mM imidazole and 500 mM NaCl before loading onto a HiTrap Chelating HP column (Amersham Biosciences, GE Healthcare, Chicago, IL, USA). All purification procedures were done at 4 °C. The column was rinsed with Buffer A1 (500 mM NaCl, 20 mM imidazole and 20 mM HEPES), and the remaining proteins were eluted with a continuous gradient of 20–500 mM imidazole in Buffer B1 (20 mM HEPES–KOH pH 7.6, 500 mM NaCl, 500 mM imidazole). The proteins were further purified by employing a Heparin HP affinity column. The fractions after HiTrap Chelating HP were collected and diluted tenfold in Buffer A2 (50 mM NaCl, 20 mM HEPES pH 7.6 and 0.0125% Nonidet P-40) and loaded on a 1-ml HiTrap Heparin column (Amersham Biosciences, Orsay, France). Proteins bound to the Heparin column were eluted by a 50–800 mM NaCl gradient in Buffer B2 (1,000 mM NaCl, 20 mM HEPES–KOH pH 7.6 and 0.0125% Nonidet P-40). Column fractions were analyzed by SDS-PAGE and the fractions with the pure His-tagged TDG^FL protein were kept at −80 °C in 50% glycerol. Concentrations of recombinant proteins in column fractions were measured by the Bradford assay. The activities of purified proteins were verified by using their classical DNA substrates immediately prior to use.

Oligonucleotides

The DNA oligonucleotide sequences used in this study are shown in Table 1. The oligonucleotides containing dA-ALI, dA-ALII and dG-ALII adducts were synthesized as described previously (Attaluri et al., 2010, 2014). All of the oligodeoxyribonucleotides composed of regular DNA base residues including their complementary counterparts and also oligonucleotides containing modified bases such as hypoxanthine (Hx) and uracil (U) were purchased from Eurogentec (Seraing, Belgium). The oligonucleotides were 5′-end labelled with γ-[³²P]-ATP (PerkinElmer, Waltham, MA, USA) and then annealed with corresponding complementary strands as described previously (Gelin et al., 2010). The ensuing duplex oligonucleotides are denoted to as X*•N, where X is a modified or regular nucleobase in the ³²P-labelled strand and N is a canonical DNA base opposite to X in the non-labelled complementary strand.

DNA repair activity assays

Unless otherwise specified, the standard reaction mixture (20 µl) contained 20 nM ³²P-labelled duplex oligonucleotide, 20 mM Tris–HCl (pH 8.0), 1 mM ethylenediamine tetraacetate (EDTA), 100 µg/ml bovine serum albumin (BSA), 1 mM dithiothreitol (DTT) and 200 nM TDG^FL. The reaction mixes were incubated at 37 °C for 5–60 min when assessing specific TDG-activities (G•T* and G•U* duplexes; the asterisk indicates the labelled strand having the defined base in a specific position) or 1, 6 and 18 h when measuring TDG-catalyzed aberrant or futile activity (T*•dA-ALI, T•dA-ALI*, T*•dA-ALII, T•dA-ALII*, T*•dG-ALII, T•dG-ALII*, T*•A, T•A* duplexes). The reaction mix for UNGΔ84 composed of 20 mM Tris–HCl (pH 8.0), 100 mM NaCl, 1 mM DTT, 100 µg/ml BSA, 1 mM EDTA was incubated at 37 °C for 30 min. All reactions were arrested with 10 µl of a stop solution consisting of 20 mM EDTA and 0.5% SDS. Unless otherwise specified, after incubation, the samples were incubated with 0.1 M NaOH for 3 min at 95 °C to cleave AP sites remained after DNA glycosylases catalyzed base excision, following the mixes were neutralized with HCl. Alternatively, the samples were treated with light piperidine (LP): 10% piperidine for 30 min at 37 °C and then neutralized with HCl. The reaction mixture for examining the combined AP endonuclease and 3′→5′ exonuclease activity of APE1 contained 200 nM TDG^FL in 20 mM Tris–HCl (pH 8.0), 1 mM DTT, 100 µg/ml BSA, 1 mM MgCl₂ and was incubated for 1 h at 37 °C. Subsequently increasing concentrations of APE1 (5–200 nM) were added to the sample and the incubation was continued for another 3 h. All samples were desalted with a Sephadex G-25 DNA grade SF column (GE Healthcare, Chicago, IL, USA) equilibrated in 7.5 M urea and run through denaturing 20% (w/v) polyacrylamide gel (7 M urea, 0.5×TBE, 42 °C) electrophoresis. The gels were scanned using a Fuji FLA-3000 Phosphor Screen (Fujifilm, Tokyo, Japan) and Typhoon FLA 9500 imager (GE Healthcare, Chicago, IL, USA) and quantified using Image Gauge v4.0 software (Fujifilm, Tokyo, Japan). At least three independent experiments were performed for all measurements. To generate size markers, 10 nM ³²P-labelled T*•G and T*•Hx duplexes were incubated with 200 nM TDG^FL for 5–60 min at 37 °C.

To calculate cleavage efficiency, we employed adjusted cleaved fraction (ACF) method which is useful when short DNA cleavage fragments can be generated from long DNA cleavage fragments in DNA substrates containing multiple cleavage site. We use following formula to calculate ACF:

$$\mathrm{A}\mathrm{C}\mathrm{F}=(\mathrm{T}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}\mathrm{I}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{n}\mathrm{s}\mathrm{i}\mathrm{t}\mathrm{y}\mathrm{o}\mathrm{f}\mathrm{C}\mathrm{l}\mathrm{e}\mathrm{a}\mathrm{v}\mathrm{e}\mathrm{d}\mathrm{B}\mathrm{a}\mathrm{n}\mathrm{d}\mathrm{s})/\mathrm{T}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}\mathrm{D}\mathrm{N}\mathrm{A}\mathrm{I}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{n}\mathrm{s}\mathrm{i}\mathrm{t}\mathrm{y}$$

Total intensity of cleaved bands includes the sum of the intensities of both long and short cleaved fragments. Total DNA Intensity is the sum of intensities of the non-cleaved full-length DNA band and all cleaved fragments (long and short). The statistical analysis of quantitative data presented in Figs. 3–7 can be found in Supplemental Information (Tables S1–S5).

Upon long incubation, human native TDG excises T opposite to A, and to a lesser extent T opposite to dA-ALI adduct in duplex DNA

Here, we were interested in studying the phenomenon of aberrant repair catalyzed by TDG on the DNA duplexes containing dA-AL adducts. To examine whether TDG could excise T opposite to dA-AL adducts in duplex DNA, we incubated the TDG^FL protein with 5′-³²P-labelled duplex oligonucleotides containing G, Hx and the dA-ALI adduct in two different sequence contexts: CpXpG, “favorable” for the mutagenesis by AA, referred to as 14-03 and “unfavorable” ApXpT referred to as 14-04 (were X is for G, Hx and dA-ALI). Other than this difference, the oligonucleotides were identical. In addition, to avoid non-specific degradation of the DNA substrate by possible contaminating nuclease activities, we assembled a dumbbell-shaped duplex (dmbDNA) consisting of a short 27-mer oligonucleotide containing either G or Hx or dA-ALI adduct at position 16 in 14-03 or 14-04 contexts (Fig. 3 and Table 1) and a long 55-mer strand forming hairpins at both ends and single-strand gap complementary to the 27-mer oligonucleotide (referred as “c1403-55” and “c1404-55” Table 1). The same dmbDNA construct was used in our previous study to measure activities of human major AP endonuclease 1, APE1 (Bazlekowa-Karaban et al., 2019). The 27-mer 14-03 and 14-04 oligonucleotides, containing G, Hx or dA-ALI residues were annealed to the 5′-³²P labelled 55-mers c1403-55 and c1404-55, respectively, so to place a regular T opposite to the position 16. This hybridization resulted in the formation of 41-mer dmbDNA substrates in which the T26 residue is located opposite to G, Hx or dA-ALI in either 14-03 or 14-04 sequence context (Figs. 3B, 3D). Next, the dmbDNA substrates G•T*, Hx•T* and dA-ALI•T* (asterisk indicates the labelled DNA strand with the defined residue) were incubated in the presence of 20 or 200 nM TDG^FL in a reaction buffer, with or without 100 mM NaCl either for 1 h or for 18 h at 37 °C, followed by the treatment with a hot alkaline to cleave AP sites produced by DNA glycosylase action. As expected, when using the DNA substrate (14-03) containing mismatched T in a favored TpG/CpX context, 20 nM TDG^FL excised T opposite to G and Hx in 41-mer G•T* and Hx•T* 14-03 dmbDNA, respectively, and generated a 25-mer cleavage fragment with good efficiency, namely 6–11% of T in 1 h (Fig. 3A, lanes 2–5). TDG^FL excised T opposite to G or Hx with similar efficiency and the presence of NaCl slightly inhibited the DNA glycosylase activity. In agreement with previous observations, when presented with a mismatch in an unfavorable TpT/ApX (14-04) sequence context, TDG^FL excised T opposite to G or Hx with 2- or 5-fold lower efficiency, respectively, as compared to the favorable TpG/CpX (14-03) context (Fig. 3C, lanes 2–5). Interestingly, increased ionic strength markedly inhibited TDG-catalyzed aberrant excision of T opposite to Hx in an unsuitable sequence context (Fig. 3C, lane 5 vs Fig. 3A, lane 5).

In contrast, incubation of 200 nM TDG^FL with 41-mer dA-ALI•T* 14-03 dmbDNA for 1, 6 or 18 h at 37 °C resulted in the generation of multiple cleavage products, the most prominent being the 35-, 29-, 25-, 17- and 15-mer cleavage fragments of variable intensities (Fig. 3A, lanes 7–9 and 10–12), indicating excision of regular T residues opposite to A at positions 36, 30, 18 and 16 in the 55-mer hairpin oligonucleotide (Fig. 3B). The TDG^FL-catalyzed futile excision of T opposite to A increased in a time-dependent manner (Figs. 3A, 3C, lanes 7–12). Notably, TDG^FL preferentially excised T opposite to A in position 30 of the 55-mer, cleaving 30.4% and 13.2% of T after 18 h of incubation (without or with 100 mM NaCl, respectively; Fig. 3A, lanes 9 and 12), whereas T opposite to the bulky dA-ALI adduct in the favorable TpG/CpX (14-03) context was excised at only 3.2% (no NaCl) and 1.5% (100 mM NaCl), even worse than the excision of T opposite to regular A in positions 36, 18 and 16 (lanes 9 and 12). Excision of T opposite to the bulky dA-ALI adduct in an unfavorable TpT/ApX (14-04) context was further decreased two-fold as compared to the favorable 14-03 context (Fig. 3C, lanes 9–11 vs Fig. 2A, lanes 9–11). In summary, these results suggest that the presence of dA-ALI adducts in DNA duplex: (i) does not stimulate aberrant removal of T opposite to the bulky lesion; (ii) does not interfere with futile excision of neighboring pyrimidines in the sequence contexts used.

Bulky dA-ALI adduct can be removed by APE1 3′→5′ exonuclease after excision of thymine 3′-next to the lesion by TDG

Next, we examined whether TDG^FL acts on pyrimidines in the top DNA strand containing dA-ALI adduct. For this, we used dmbDNA with the 5′-³²P-labelled 27-mer strand containing either G or dA-ALI in position 16 annealed to the non-labelled 55-mer hairpin oligonucleotide. As shown in Fig. 4A, TDG^FL acts very weakly on the 27-mer 14-03 strand excising 1.9% of C at position 11 in CpA context after 18 h of incubation (lane 8). Strikingly, TDG^FL acted more efficiently towards the 27-mer 14-04 strand, generating a 16-mer cleavage product via excision of 23.6% of T at position 17 immediately 3′ to the dA-ALI adduct (Fig. 4C, lane 16). Excision of T by TDG would generate an AP site, which then should be cleaved by APE1 to generate a single-strand break 3′ to the damaged adenine residue. This break might become persistent, since the presence of the bulky adduct at the 3′ terminus would make it a poor substrate for DNA polymerase β in the downstream steps of the BER pathway (Fig. 1). Previously, we showed that APE1 can remove bulky 8,5′-cyclopurine DNA adducts at 3′-termini of DNA strand breaks or gaps (Mazouzi et al., 2013). Therefore, we examined whether after the futile excision of T 3′-next to dA-ALI by TDG, APE1 could cleave the resulting AP site and remove the bulky adduct by its 3′→5′ exonuclease activity. For this, we incubated TDG^FL with A*•T or dA-ALI*•T 14-04 dmbDNA in a buffer containing MgCl₂ for 1 h and either treated with light piperidine or incubated in the presence of varying concentrations of APE1. As shown in Fig. 5A, TDG^FL excised T next to A and dA-ALI with low but detectable efficiency and generated cleavage fragments after light piperidine (LP) treatment (lanes 4 and 10). When the products of reaction were incubated in the presence of 5 nM APE1, we observed a 16-mer fragment, indicating cleavage of the AP site by this enzyme (Fig. 5A, lanes 5 and 11 and Fig. 5C). Note that APE1-generated cleavage products that contain hydroxyl group at 3′-termini (lanes 5–7 and 11–13) migrated faster as compared to LP-generated fragments (lanes 4 and 10) that contain 3′-α,β-unsaturated aldehyde. Upon increasing the concentrations of APE1 up to 20 and 50 nM, we observed the appearance of the products of 3′-exonuclease degradation in both A*•T and dA-ALI*•T dmbDNA, indicative of the removal of 3′-terminal nucleotides dA and dA-ALI (Fig. 5A, lanes 6–7 and 11–13 and Fig. 5C). With increasing concentrations of APE1 up to 200 nM, the 16-mer cleavage products completely disappeared, suggesting complete degradation of the fragment by the 3′→5′ exonuclease activity of APE1 (Fig. 5A, lanes 8 and 14 and Fig. 5C). Curiously, excision of T next to dA-ALI was slightly higher compared to dA, leading to the possibility that the presence of a neighboring dA-ALI adduct provides easier access for TDG^FL. Taken together, these results suggest that APE1 can remove bulky dA-ALI adducts at 3′-termini of DNA strand breaks with good efficiency. However, the TDG-catalyzed activity towards 3′ neighboring T next to dA-ALI is still very low to have physiological relevance even when combined with APE1 action.

Influence of bulky dA-ALI and II adducts on TDG-catalyzed futile excision of pyrimidines

The above results showed that after a long incubation time at 37 °C TDG^FL is capable of excising T and C residues opposite to A and G in TpG and CpA contexts, respectively, in regular non-damaged DNA duplexes to a significant extent (Fig. 3A, lanes 7–9 and Figs. 4A, 4C, lanes 8 and 16). In addition, TDG^FL can excise T opposite to dA-ALI adduct in an aberrant manner, but with much lower efficiency as compared to the futile excision of T in T•A base pairs at different positions of the same DNA duplex substrate (Fig. 3A). Since the results shown above and our recent published study (Manapkyzy et al., 2024) clearly demonstrated the absence of nuclease contamination in our enzyme preparations, we decided to use short blunt-end DNA duplexes for further characterization of TDG-catalysed repair. For this, 24 mer 14-07 and 10-05 duplex oligonucleotides with the same sequence, but containing either dA-ALI or dA-ALII adduct at position 14, respectively (Table 1), were incubated with TDG^FL. It should be noted that the use of short DNA duplex substrates allowed us to reduce the number of cleavage sites as compared to long 55 mer oligonucleotide utilized to construct dmbDNA, which in turn simplified analysis and interpretation of the results. As shown in Fig. 6A, TDG^FL exhibited efficient futile excision of thymine residues in the bottom and top strands of non-damaged regular A•T* and A*•T 10-05 duplex oligonucleotide after 1 or 18 h incubation (lanes 4–5 and 7–8). Intriguingly, the presence of dA-ALII adduct in dA-ALII•T* and dA-ALII*•T 10-05 duplexes dramatically reduced and changed the pattern of the futile activity (lanes 12–13 and 15–16). When acting upon regular A*•T 10-05 duplex, TDG^FL excised in a futile manner 15% and 54% of T opposite to A at position 11 in the ³²P-labelled top strand after 1 or 18 h, respectively, (lanes 7–8). In the 10-05 dA-ALII*•T duplex, excision of T11 was reduced approximately tenfold (lane 16), but the excision pattern remained the same, with T at positions 11 and 9 as preferred sites (lanes 7–8 vs 15–16). In a regular 10-05 A•T* duplex with the labelled bottom strand, TDG also excises T at position 11, but with ~5-fold less efficiency (Fig. 6A, lanes 4–5), as compared to the excision of T11 in the top strand of the same duplex (lanes 7–8). In addition, we observed a faint band migrating at position of 14 mer, suggesting that TDG^FL excises C15 opposite to G, but with a very low efficiency (0.4%) (lane 5). Strikingly, the presence of dA-ALII adduct changed completely the pattern of futile activity on the bottom strand of 10-05 dA-ALII•T* duplex: TDG^FL exhibits a 20-fold stimulation of excision of C15 (8.2%) opposite to G and a 7-fold inhibition of excision of T11 (1.4%) opposite to the dA-ALII adduct (lane 13) as compared to the regular A•T* 10-05 duplex (lane 5). It should be noted that we used 10-05 G•U15* duplex incubated with TDG to generate 14 mer size marker (lane 10), which migrates at the same position as the faint cleavage product in (lane 5) and major product in (lane 13). Next, we examined the pattern of TDG^FL-catalyzed excisions of pyrimidines in 14-07 A•T and dA-ALI•T duplexes. As shown in Fig. 6C, the presence of dA-ALI adduct influences the futile removal of T and C by TDG^FL in a manner very similar to that of dA-ALII (lanes 4–5, 7–8 and 10–11). Overall, these results suggest that: (i) TDG excises thymine opposite to dA-AL adducts poorly compared with its canonical G•T substrate; (ii) the presence of dA-AL adducts in duplex DNA rather inhibits futile excision of both opposite thymine in complementary strand and neighboring 5′ upstream thymine in the same adducted strand; (iii) depending on sequence contexts the presence of dA-AL adducts may also stimulate excision of neighboring regular cytosine in complementary strand.

To examine further the effect of sequence context and the presence of bulky purine adducts on TDG activities, we measured the DNA glycosylase activities using a 19-mer 14-06 duplex containing dA-ALI adduct at position 13 and a 24-mer 10-13 duplex with the same sequence as 10-05 but containing a dG-LII adduct at position 10 (Table 1). As shown in Fig. 7A, after 18 h of incubation TDG^FL exhibited very weak futile activity towards short regular A•T* and A*•T 14-06 duplexes, since we observed excision of 3.1% of T11 opposite to A in the bottom strand and 2.8% of C8 opposite to G in the top strand of the A•T* 14-06 substrate (lanes 5 and 8). The presence of dA-ALI in 14-06 duplex did not change the pattern of futile activity, but stimulated excision of T11 opposite to A in the bottom strand of dA-ALI•T* threefold, while reducing the excision of C8 opposite to G in the top strand of dA-ALI*•T duplex also threefold (lanes 11 and 14).

As shown in Fig. 7C, similar to 10-05 duplex, TDG^FL exhibited robust futile activities towards 24-mer regular A•T* and A*•T 10-13 duplexes, excising 22.5% of T11 in the bottom strand and 17.7% of T11 in the top strand (lanes 5 and 8). Replacement of G with the bulky dG-ALII adduct stimulated the excision of T11 in the bottom strand about threefold, whereas the excision of T15 opposite modified guanine remained very weak (lane 11). The excision of T11 immediately 3′ to dG-ALII in the top strand was also strongly inhibited (lane 14). No other changes in the pattern of futile activity were observed as compared with the regular 10-13 duplex. Altogether, the results suggest that (i) TDG^FL inefficiently recognizes thymine opposite to either a bulky adenine or a guanine adduct compared with its canonical G•T substrate; (ii) the presence of a dG-ALII adduct depending on the sequence context, can strongly inhibit futile excision of immediately neighboring pyrimidines.