Human TDP1, APE1 and TREX1 repair 3'-DNA-peptide/protein cross-links arising from abasic sites in vitro

Abstract

Histones and many other proteins react with abundant endogenous DNA lesions, apurinic/apyrimidinic (abasic, AP) sites and/or 3'-phospho-α,β-unsaturated aldehyde (3'-PUA), to form unstable but long-lived Schiff base DNA-protein cross-links at 3'-DNA termini (3'-PUA-protein DPCs). Poly (ADP-ribose) polymerase 1 (PARP1) cross-links to the AP site in a similar manner but the Schiff base is reduced by PARP1's intrinsic redox capacity, yielding a stable 3'-PUA-PARP1 DPC. Eradicating these DPCs is critical for maintaining the genome integrity because 3'-hydroxyl is required for DNA synthesis and ligation. But how they are repaired is not well understood. Herein, we chemically synthesized 3'-PUA-aminooxylysine-peptide adducts that closely resemble the proteolytic 3'-PUA-protein DPCs, and found that they can be repaired by human tyrosyl-DNA phosphodiesterase 1 (TDP1), AP endonuclease 1 (APE1) and three-prime repair exonuclease 1 (TREX1). We characterized these novel repair pathways by measuring the kinetic constants and determining the effect of cross-linked peptide length, flanking DNA structure, and the opposite nucleobase. We further found that these nucleases can directly repair 3'-PUA-histone DPCs, but not 3'-PUA-PARP1 DPCs unless proteolysis occurs initially. Collectively, we demonstrated that in vitro 3'-PUA-protein DPCs can be repaired by TDP1, APE1, and TREX1 following proteolysis, but the proteolysis is not absolutely required for smaller DPCs.

Scheme 1.

Structure of AP, 3′-PUA, C7-NH₂ and 3′-dRP.

Scheme 2.

Mechanistic formation of 3′-PUA–protein DPCs.

Scheme 3.

Synthesis of 3′-PUA-OxyLys by oxime ligation.

Figure 1.

Removal of 3′-PUA-OxyLys within nicked DNA by TDP1. (A) The structure of nicked DNA containing 3′-PUA-OxyLys and 5′-phosphate. The red dashed arrow illustrates the excision site of TDP1. (B) A plot showing the efficiency of TDP1 (10 or 20 nM) to remove 3′-PUA-OxyLys within nicked DNA (50 nM) at 37 °C as a function of time. The data is from three independent experiments. (C) A representative 20% urea–PAGE gel showing the removal of 3′-PUA-OxyLys within the nicked DNA (80 nM) by wild type (WT) but not the catalytically inactive (H263A) TDP1. The reactions were carried out at 37°C for 30 min. (D) A representative 20% urea–PAGE gel showing the removal of 3′-PUA-OxyLys within the nicked DNA by TDP1 with or without an additional treatment by T4 polynucleotide kinase (T4 PNK). Step 1, 3′-PUA-OxyLys (80 nM) was treated by TDP1 (50 nM) at 37°C for 30 min. Step 2, to the mixture in Step 1, T4-PNK (0.5 unit/μl) was added and incubated at 37°C for 30 min. The urea–PAGE gels in (C) and (D) were scanned using the fluorescence of 6-FAM.

Figure 2.

Removal of 3′-PUA-OxyLys within nicked DNA by APE1. (A) The structure of nicked DNA containing 3′-PUA-OxyLys and 5′-phosphate. The red dashed arrow illustrates the excision site of APE1. (B) A plot showing the removal of 3′-PUA-OxyLys within nicked DNA (50 nM) by APE1 (0.2 or 1 nM) at 37°C as a function of time. The data is from three independent experiments. (C) A representative 20% urea–PAGE gel showing the removal of 3′-PUA-OxyLys within the nicked DNA (80 nM) by APE1 (20 nM) at 37°C for 30 min. The gel was scanned using the fluorescence of 6-FAM.

Figure 3.

Removal of 3′-PUA-OxyLys within nicked DNA by TREX1. (A) The structure of nicked DNA containing 3′-PUA-OxyLys and 5′-phosphate. The free 3′-termini were blocked by phosphorylation to prevent the undesired degradation by TREX1. The red dashed arrow illustrates the processing of 3′-PUA-OxyLys by TREX1. (B) A representative 20% urea–PAGE gel showing the removal of 3′-PUA-OxyLys within nicked DNA (20 nM) by indicated concentrations of TREX1_1–242 at 37°C for 30 min. (C) A plot showing the removal efficiency of 3′-PUA-OxyLys within nicked DNA (50 nM) by TREX1 (5 nM) at 37°C as a function of time. The data is from three independent experiments. (D) A representative 20% urea–PAGE gel showing the removal of 3′-PUA-OxyLys within the nicked DNA by increasing concentrations of wild type (WT) or catalytically inactive (D18N) TREX1_1–242. The reactions were carried out by incubating the substrate (20 nM) and indicated concentration of enzyme at 37°C for 30 min. The urea–PAGE gels in (B) and (D) were scanned using the fluorescence of 6-FAM.

Figure 4.

Removal of 3′-PUA–OxyLys–peptide cross-links within nicked DNA by TDP1, APE1 and TREX1. (A) Illustration of nicked DNA containing 3′-PUA–OxyLys–peptides and 5′-phosphate. (B–D) The plots showing the efficiency of adduct removal as a function of cross-linked peptide length by TDP1 (B), APE1 (C), and TREX1 (D). In (B) and (C), the reactions were carried out by incubating the substrate (80 nM) with TDP1 (30 nM) or APE1 (10 nM) at 37°C for 30 min. In (D), the reactions were performed by incubating the substrate (20 nM) with TREX1 (10 nM) at 37°C for 20 min. The data is from two independent experiments with each one in triplicate. The P-values (*P < 0.05; **P < 0.01) were determined by two-tailed unpaired t test. n.s., not significant. The P-values are 0.009 (B, 1 versus 5-mer), 0.118 (B, 1 versus 10-mer), 0.002 (B, 5 versus 10-mer), 0.003 (C, 1 versus 5-mer), 0.005 (C, 1 versus 10-mer), 0.026 (C, 5 versus 10-mer), 0.118 (D, 1 versus 5-mer), 0.002 (D, 1 versus 10-mer), and 0.010 (D, 5 versus 10-mer).

Figure 5.

Flanking DNA structure effect on 3′-PUA–OxyLys–peptide_10mer removal by TDP1, APE1 and TREX1. (A) Substrates containing 3′-PUA–OxyLys–peptide_10mer with different flanking structure. (B–D) The plots showing the adduct removal by TDP1 (B), APE1 (C) and TREX1 (D) as a function of flanking DNA structure. The free 3′-termini used for TREX1 experiments were blocked by phosphorylation. The reactions were carried out by incubating the substrates (20 nM) at 37°C with TDP1 (50 nM, 30 min), APE1 (10 nM, 1 h) or TREX1 (2 nM, 30 min). The data is from two independent experiments with each one in triplicate.

Figure 6.

Effect of the opposite nucleobase on removal of 3′-PUA–OxyLys–peptide_10mer by TDP1, APE1 and TREX1. (A) Substrates containing 3′-PUA–OxyLys–peptide_10mer with different opposite nucleobases. (B–D) The plots showing the adduct removal by TDP1 (B), APE1 (C) and TREX1 (D). The reactions were carried out by incubating the substrate (20 nM) with TDP1 (50 nM, 30 min), APE1 (10 nM, 30 min) or TREX1 (2 nM, 15 min) at 37°C. The data is from three independent experiments. The P-values (*P < 0.05; **P < 0.01) were determined by two-tailed unpaired t test. For clarity, the difference between A and G, or T and C in panel D, which is not significant (n.s.), is not annotated. The P-values in panel D (TREX1) are 0.003 (A versus T), 0.398 (A versus G), 0.003 (A versus C), 0.003 (T versus G), 0.065 (T versus C) and 0.002 (G versus C).

Figure 7.

Removal of 3′-PUA–histone DPCs by TDP1, APE1 and TREX1. (A) Structure of 3′-PUA–histone DPCs prepared by reductive amination. (B–D) Plots showing the efficiency of 3′-PUA–histone DPC removal by TDP1 (B), APE1 (C) and TREX1 (D). The reactions were carried out by incubating the DPC substrates (60 nM) and indicated concentration of enzyme at 37°C for 60 min. The data is from three independent experiments.

Figure 8.

Repair of proteinase K-digested 3′-PUA–PARP1 DPCs by TDP1, APE1 and TREX1. (A) Illustration of preparing and repairing the 3′-PUA–PARP1 DPC and its proteolyzed products. (B–D) Representative 20% urea–PAGE gels and plots showing the repair of 3′-PUA–PARP1 peptides by TDP1 (B), APE1 (C) and TREX1 (D). The reactions were carried out by incubating the mixture of unreacted AP sites and 3′-PUA-peptide adducts (100 nM) with indicated concentrations of enzyme at 37°C for 60 min. The data is from two independent experiments. The urea–PAGE gels in B–D were scanned using the fluorescence of 6-FAM.