Impact of DNA ligase inhibition on the nick sealing of polβ nucleotide insertion products at the downstream steps of base excision repair pathway

Abstract

DNA ligase (LIG) I and IIIα finalize base excision repair (BER) by sealing a nick product after nucleotide insertion by DNA polymerase (pol) β at the downstream steps. We previously demonstrated that a functional interplay between polβ and BER ligases is critical for efficient repair, and polβ mismatch or oxidized nucleotide insertions confound the final ligation step. Yet, how targeting downstream enzymes with small molecule inhibitors could affect this coordination remains unknown. Here, we report that DNA ligase inhibitors, L67 and L82-G17, slightly enhance hypersensitivity to oxidative stress-inducing agent, KBrO₃, in polβ^+/+ cells more than polβ^-/- null cells. We showed less efficient ligation after polβ nucleotide insertions in the presence of the DNA ligase inhibitors. Furthermore, the mutations at the ligase inhibitor binding sites (G448, R451, A455) of LIG1 significantly affect nick DNA binding affinity and nick sealing efficiency. Finally, our results demonstrated that the BER ligases seal a gap repair intermediate by the effect of polβ inhibitor that diminishes gap filling activity. Overall, our results contribute to understand how the BER inhibitors against downstream enzymes, polβ, LIG1, and LIGIIIα, could impact the efficiency of gap filling and subsequent nick sealing at the final steps leading to the formation of deleterious repair intermediates.

Introduction

Base excision repair (BER) is the critical process for preventing the mutagenic and lethal consequences of DNA lesions generated by endogenous sources and environmental toxicants [1,2]. BER mainly repairs the oxidative base damage 8-oxo-2ʹ-deoxyguanine (8-oxodG), abasic sites, and single-strand breaks [3,4]. Such DNA lesions are highly abundant and can interfere with DNA replication and transcription leading to genomic instability, events that underlie human diseases such as neurological disorders and cancer [5,6]. The BER pathway includes a series of sequential enzymatic steps that require a tightly coordinated function of the core repair proteins: DNA glycosylase, AP-Endonuclease 1 (APE1), DNA polymerase (pol) β, and DNA ligase [7–10]. It has been shown that the BER proteins transfer the repair intermediates from one enzyme to the next in the sequential manner before releasing the repaired DNA so that the product of one reaction could serve as a substrate for the subsequent reaction in the repair pathway, which is known as the mechanism of substrate-product channeling [11]. Otherwise, persistent exposure of the DNA repair intermediates such as single-strand breaks (SSBs) could be cytotoxic and mutagenic that may initiate signaling pathways, leading to cell death and eventually genomic instability [12–14]. During the downstream steps of the BER pathway, polβ fills a gap repair intermediate by a single nucleotide incorporation, which generates a nick repair product with 3ʹ-OH and 5ʹ-PO₄ ends to be sealed subsequently by DNA ligase 1 (LIG1) or DNA ligase IIIα (LIGIIIα) to complete the BER process [15]. The exchangeability between the BER DNA ligases has been reported [16]. Our previous studies demonstrated that a functional interplay between polβ and the BER ligases, LIG1 and LIGIIIα, is critical for efficient repair [17–24]. For example, we reported that polβ mismatch, ribonucleotide, or oxidized nucleotide (i.e. 8oxodGTP) insertions can confound the next and final DNA ligation step and lead to aberrant BER [25–27].

The large proportion of therapeutic DNA damage such as platinum drugs, alkylating agents, ionizing radiation as well as external and internal radiotherapy produce DNA modifications including base damage and SSBs, which rely on BER for their repair and cytotoxicity in cancer cells [28–30]. It has been reported that the efficiency of BER, a front-line DNA repair mechanism that removes alkylation damage from DNA, impacts the final therapeutic effectiveness of alkylating anticancer drugs that are still commonly used in the treatment of several types of cancers [31–34]. In addition, because of a higher oxidative status in cancer cells than normal cells, BER, as a major repair pathway for the repair of oxidative DNA damage, can mediate chemotherapy resistance such as cisplatin cytotoxicity [35]. Furthermore, the BER proteins of the repair pathway have been considered as promising targets to develop anti-cancer compounds to increase sensitivity against several chemotherapy and/or radiotherapy agents in combination treatments [30–35]. This is particularly important for specific tumor types that contain BER gene and protein overexpression. As an alternative strategy to target BER, killing cancer cells specifically via a synthetic lethal partnership where a tumorigenic mutation becomes reliant on BER to survive has been widely used, and this includes the inhibition of Poly(ADP-ribose) polymerase 1 (PARP-1) in HR-defective cancers harboring BRCA1/2 mutations [36,37]. Several small molecule compounds have been also developed against the BER enzymes that function at the earlier steps of the repair pathway, such as 8-oxoguanine DNA glycosylase-1 (OGG1), Nei Like DNA Glycosylase 1 (NEIL1), and APE1 [38–43].

There are several small molecule inhibitor compounds have been developed against the downstream repair proteins, polβ and DNA ligases. For example, it has been shown that the polβ small molecular inhibitor NSC666715 [4-chloro-N-[5-4-chloroanilino)-1H-1,2,4-triazol-3-yl]-5-methyl-2-sulfanylbenzenesulfonamide] induces Temozolomide (TMZ)-induced cytotoxicity, causes accumulation of AP sites, replication fork double-strand breaks, and leads to senescence with an increase in p53 and p21 levels in HCT116 colon cancer cells [44–46]. In addition to this small molecule compound, polβ inhibitor pamoic acid has been found to be toxic to BRCA2 deficient cells leading to cell cycle arrest and increased apoptosis [47]. Moreover, Pro-13 is known as an irreversible inhibitor with a profound synergic effect in combination with MMS, and natamycin has been shown as an inhibitor of polβ strand displacement activity [48,49]. Although more than ~100 small molecule inhibitor compounds have been identified during the past two decades, most of them are either not potent to become a drug or not pharmaceutically specific because of the presence of similar other repair DNA polymerases such as polλ [50]. However, it has been widely accepted that polβ has a significant role in chemotherapeutic agent resistance because of its overexpression that reduces the efficiency of anticancer drug therapies and has been considered as a promising therapeutic target for cancer treatment [51].

Regarding the small molecule inhibitors developed against DNA ligases, based on the first crystal structure of LIG1 in complex with nick DNA [52], the computer-aided drug design was used to identify potential DNA ligase inhibitors that were predicted to bind to the DNA binding domain (DBD) of LIG1 [53–56]. The putative DNA binding pocket that is in direct contact with nick, compromising the amino acid residues Gly(G)448, Arg(R)451, and Ala(A)455, has been identified by in silico database screening [57]. Through these studies, two DNA ligase inhibitors have been developed: L82-G17 (an inhibitor of LIG1) and L67 (an inhibitor of both LIG1 and LIGIIIα). These inhibitors have no effect on the formation of the covalent enzyme-adenylate (LIG-AMP) intermediate that occurs independently of the DNA binding at the initial step of the ligation reaction and is required to activate the ligase to bind to a nick DNA [58]. The effectives of the ligase inhibitors have been shown to be cytotoxic in cell culture assays, potentiate the cytotoxic effects of DNA-damaging agents, and induce cell death in cancer cells [59–62]. For example, DNA ligase inhibitors increase cytotoxicity in therapy-resistant breast cancer cells and tyrosine kinase inhibitor-resistant chronic myeloid leukemia with the combination of DNA-damaging agents such as DNA-alkylating agent methyl methanesulfonate (MMS) or PARP inhibitors by targeting alternative-non-homologous end joining (alt-NHEJ) with an elevated steady-state level of LIGIIIα [63].

In the present study, we used the small molecule inhibitor compounds, SKG-63-1 (an inhibitor of polβ), L82-G17 (an inhibitor of LIG1), and L67 (an inhibitor of both LIG1 and LIGIIIα), to target polβ and both BER ligases finalizing the repair pathway at the downstream steps. Our results demonstrated that the ligase inhibitors result in reduced ligation efficiency of nick substrates by LIG1 and LIGIIIα in vitro. We also showed a decrease in the amount of ligation products after polβ dGTP:C insertion in the presence of the ligase inhibitors. Furthermore, we observed that the BER ligases seal a gap repair intermediate, leading to gap ligation and formation of single nucleotide mutagenesis products, in case of an aberrant gap filling by the effect of polβ inhibitor SKG-63-1 in vitro.

In the current work, we also demonstrated that polβ^+/+ wild-type and polβ^-/- null Mouse Embryonic Fibroblast (MEF) cells exhibit similar sensitivity to DNA ligase inhibitors L67 and L82-G17. Our co-treatments of both MEF cells with oxidative-stress inducing agent potassium bromate (KBrO₃) and the ligase inhibitors demonstrated that L67 sensitizes polβ^+/+ cells to KBrO₃ more than polβ^-/- cells. Furthermore, we showed a slight increase in the formation of ligation failure products after polβ 8oxodGTP:A insertion in the presence of the ligase inhibitor L67 and LIGIIIα in vitro.

Finally, we generated the LIG1 mutants harboring amino acid substitutions at the ligase inhibitor binding sites (G448D, R451D, A455D) residing in the DNA binding domain of the ligase and observed reduced ligation efficiency of the nick DNA substrates with preinserted 3ʹ-dG:C, 3ʹ-8oxodG:A, and 3ʹ-8oxodG:C. Our real-time kinetics measurements with LIG1 mutants demonstrated ~10-fold difference in nick DNA binding affinity of wild-type protein. Overall, this study contributes to understanding how the inhibitors against the downstream BER enzymes, polβ, LIG1, and LIGIIIα, could lead to deviations in the proper interplay between polβ gap filling and subsequent nick sealing, which could result in the formation of aberrant repair intermediates.

Methods

Preparation of DNA ligase I mutant constructs

Plasmid DNA constructs of full-length (1-919) DNA ligase I (LIG1) for the mutants harboring G448D, R451D, and A455D mutations at the ligase inhibitor binding sites (G448, R451, A455) were cloned into the pET-24b expression vector using the wild-type LIG1 construct [17–27]. The coding sequences of all mutants were confirmed by DNA sequencing.

Protein purifications

Human DNA ligase I (LIG1) wild-type and mutants (G448D, R451D, A455D) were purified as described previously [17–27]. The recombinant proteins carrying 6x-his tag were overexpressed in Rosetta (DE3) pLysS E. coli cells and grown in Terrific Broth (TB) media with kanamycin (50 μgml⁻¹) and chloramphenicol (34 μgml⁻¹) at 37°C. Once the OD₆₀₀ reached 1.0, the cells were induced with 0.5 mM isopropyl β-D-thiogalactoside (IPTG), and the overexpression continued overnight at 28°C. Cells were harvested by centrifugation, and the cell pellet was lysed in the lysis buffer containing 50 mM Tris-HCl (pH 7.0), 500 mM NaCl, 20 mM imidazole, 2 mM β-mercaptoethanol, 5% glycerol, and 1 mM PMSF by sonication at 4°C. The lysate was pelleted at 31,000× g for 90 min at 4°C. The cell lysis solution was filter clarified and then loaded onto HisTrap HP column that was previously equilibrated with the binding buffer containing 50 mM Tris-HCl (pH 7.0), 500 mM NaCl, 20 mM imidazole, 2 mM β-mercaptoethanol, and 5% glycerol. The column was washed with the binding buffer and then by the washing buffer containing 50 mM Tris-HCl (pH 7.0), 500 mM NaCl, 35 mM imidazole, 2 mM β-mercaptoethanol, and 5% glycerol. LIG1 protein was finally eluted with an increasing imidazole gradient 0–500 mM at 4°C. The proteins were then subsequently loaded onto a HiTrap Heparin column that was equilibrated with the low salt binding buffer containing 20 mM Tris-HCl (pH 7.0), 50 mM NaCl, 2 mM β-mercaptoethanol, and 5% glycerol, and then eluted with a linear gradient of NaCl up to 1 M. The proteins were further purified by the size exclusion chromatography using Superdex 200 column in the buffer containing 20 mM Tris-HCl (pH 7.0), 200 mM NaCl, and 1 mM DTT. Human wild-type full-length (1-922 amino acids) DNA ligase IIIα (LIGIIIα) with 6x-his tag (pET-29a) was overexpressed in BL21(DE3) E. coli cells in Lysogeny Broth (LB) media at 37°C for 8 h and induced with 0.5 mM IPTG as described [17–27]. The protein overexpression was continued overnight at 28°C. The cells were harvested, lysed at 4°C, and then clarified as described above. LIGIIIα protein was purified by HisTrap HP column with an increasing imidazole gradient (20–300 mM) elution at 4°C. The collected fractions were then further purified by Superdex 200 increase 10/300 column in the buffer containing 50 mM Tris-HCl (pH 7.0), 500 mM NaCl, glycerol 5%, 1 mM DTT. Human full-length wild-type polβ with a GST-tag (pGEX-6p-1) were overexpressed in BL21(DE3) E. coli cells in LB media at 37°C for 8 h and induced with 0.5 mM IPTG when OD₆₀₀ reached 0.8–1.0. The cells were then grown overnight at 28°C, and the cell lysis was obtained at 4°C by sonication in the lysis buffer containing 1X PBS (pH 7.3), 200 mM NaCl, 1 mM DTT, and cOmplete protease inhibitor cocktail. The cell lysate was pelleted at 16,000× rpm for 1.5 h and then clarified by centrifugation and filtration. The supernatant was loaded onto a GSTrap HP column and purified with the elution buffer containing 50 mM Tris-HCl (pH 8.0) and 10 mM reduced glutathione. To cleave the GST-tag, the recombinant polβ protein was incubated with PreScission protease (1:1) for 16 h at 4°C in the buffer containing 1X PBS (pH 7.3), 200 mM NaCl, and 1 mM DTT. After the cleavage, the polβ protein was subsequently passed through GSTrap HP column and collected in the flowthrough. Polβ proteins were then further purified by Superdex 200 increase 10/300 column in the buffer containing 50 mM Tris-HCl (pH 7.5), and 400 mM NaCl. All proteins purified in this study were stored at −80°C in aliquots for single session use. Protein quality was evaluated on a 10% SDS-PAGE gel, and protein concentrations were measured using absorbance at 280 nm.

Polβ nucleotide insertion assays

The insertion assays were performed to measure nucleotide insertion by polβ using one nucleotide gap DNA substrate (Supplementary Table 1) in the reaction mixture containing 50 mM Tris-HCl (pH 7.5), 100 mM KCl, 10 mM MgCl₂, 1 mM ATP, 1 mM DTT, 100 µgml^-1 BSA, 1% glycerol, dNTP (100 µM), and DNA substrate (500 nM) in a final volume of 10 µl. The reaction was initiated by the addition of the polβ (10 nM), and the reaction mixtures were incubated at 37°C for the time points as indicated in the figure legends. The reaction products were then mixed with an equal amount of gel loading buffer containing 95% formamide, 20 mM EDTA, 0.02% bromophenol blue, and 0.02% xylene cyanol and separated by electrophoresis on an 18% polyacrylamide gel. The gels were finally scanned with a Typhoon PhosphorImager RGB, and the data were analyzed using ImageQuant software. Polβ nucleotide insertion assays were performed similarly in the absence and presence of polβ (SKG-63-1) and DNA ligase (L67 and L82-G17) inhibitors.

Polβ nucleotide insertion coupled to ligation assays

The coupled assays were performed to measure nucleotide insertion by polβ and subsequent nick sealing by DNA ligase in the same reaction mixture using one nucleotide gap DNA substrate (Supplementary Table 1). The reaction was initiated by the addition of pre-incubated enzyme mixture of polβ (10 nM) and DNA ligase (100 nM) to the reaction mixture containing 50 mM Tris-HCl (pH 7.5), 100 mM KCl, 10 mM MgCl₂, 1 mM ATP, 1 mM DTT, 100 μg ml^-1 BSA, 1% glycerol, DNA substrate (500 nM), and dNTP (100 µM) in a final volume of 10 μl. The reaction mixtures were then incubated at 37°C for the reaction times as indicated in the figure legends, quenched by mixing with an equal volume of loading dye, and analyzed as described above. The coupled assays were performed similarly in the absence and presence of polβ (SKG-63-1) and DNA ligase (L67 and L82-G17) inhibitors.

DNA ligation assays

Ligation assays were performed using nick DNA substrates with preinserted 3ʹ-dG:C, 3ʹ-8oxodG:A, and 3ʹ-8oxodG:C (Supplementary Table 1) in the reaction mixture containing 50 mM Tris-HCl (pH 7.5), 100 mM KCl, 10 mM MgCl₂, 1 mM ATP, 1 mM DTT, 100 µgml⁻¹ BSA, 1% glycerol, and nick DNA substrate (500 nM) in a final volume of 10 µl. The reaction was initiated by the addition of DNA ligase (100 nM), incubated at 37°C, and stopped at the time points indicated in the figure legends by mixing with an equal volume of loading dye. The reaction products were analyzed as described above. Ligation assays were performed similarly in the absence and presence of polβ (SKG-63-1) and DNA ligase (L67 and L82-G17) inhibitors.

Nick DNA binding measurements

Nick DNA binding kinetics of LIG1 (wild-type and mutants) were measured by Biolayer Interferometry (BLI) assays in real time using the Octet QKe system (Fortebio). BLI experiments were performed at 20°C in 96-well microplates with agitation set to 1000 rpm in the absence and presence of nick DNA with 3ʹ-biotin label. Streptavidin (SA) biosensors were used to attach the biotin-labeled DNA. The SA biosensors were hydrated in the buffer containing 50 mM Tris-HCl (pH 7.5), 100 mM KCl, and 1 mM DTT at 20°C for 20 min. The sensors were then immersed in nick DNA (40 nM) in the buffer for 300 sec. After recording an initial baseline in the buffer (60 s), the sensors with DNA were exposed to LIG1 at the concentration range as indicated in the figure legends for 240 s association and then in the buffer for 240 s dissociation. In all measurements, the affinity constants (K_D), the association (k_on), and dissociation (k_off) rates were calculated using the ForteBio Data Analysis software with 1:1 binding model. The association rate = k_on [ligand][analyte] and the dissociation rate = k_off [ligand-analyte]. At equilibrium, forward and reverse rates are equal. All images were drawn using Graph Pad Prism 9. BLI assays were performed similarly in the presence of DNA ligase inhibitors L67 and L82-G17.

Thermal stability assays

Fluorescence-based thermal shift assays were carried out with 96 well plates in the CFX96 RT-PCR detection system (Bio-Rad). The assays were performed with DNA ligase (5 µM) in the absence and presence of nick DNA (50 µM). After incubation at 4°C for 1 h, 20× Sypro Orange protein dye was added, and the plate was then sealed with an optical seal and centrifuged. Thermal scan ranged from 10 to 95°C with a temperature ramp rate of 0.20°C/min. The fluorescence intensity upon binding of Sypro Orange was measured with excitation/emission of 533/580 nm. Data analysis and report generation were performed using the CFX Maestro software (Bio-Rad). Melting temperature (T_m) values were calculated manually from the negative derivative plot at the point of inflection of the curve (the midpoint for protein unfolding). All images were drawn using Graph Pad Prism 9. Thermal shift assays were performed similarly in the absence and presence of nick DNA.

Structure modeling of LIG1 mutants

Structure modeling was performed to predict DNA ligase inhibitor mutants G448D, R451D, and A455D based on previously solved crystal structure of human LIG1 (PDB:6P0C) bound to adenylated DNA containing C:G base-pair [64]. All structural images were drawn using PyMOL (Schrödinger).

Cell culture and cytotoxicity assays

Polβ^+/+ wild-type (36.3) and polβ^-/- null (38D4) mouse embryonic fibroblasts (MEF) cell lines were previously described [12–14,25]. The cells were routinely grown in a 10% CO₂ humidified chamber at 34°C in Dulbecco’s modified Eagle’s medium supplemented with GlutaMAX (GIBCO), 10% fetal bovine serum (Millipore Sigma) 1% anti-anti 100× (Penicillin/Streptomycin) (GIBCO) and Hygromycin B (80 µg/ml) (Invitrogen). Using polβ^+/+ wild-type and polβ^-/- null MEF cells, cell viability was measured using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay [65]. Both MEF cells were diluted in the cell culture medium as described above without hygromycin and seeded at a density of 2000 viable cells/well in 96 well flat bottom plates. The cells were then incubated at 34°C in a 10% CO₂ incubator for 24 h allowing them to adhere to the plate. All incubations were carried out under the same growth conditions. The cells were then exposed to a concentration range of oxidative stress-inducing agent potassium bromate (KBrO₃) alone and in combination treatments with the DNA ligase inhibitors L67 and L82-G17 as indicated in the figure legends. The plates were incubated for 72 h after which the MTT solution (5 mg/ml) was added to each well, and incubation continued for 4 h more. A solubilization agent of an acidic 10% SDS was finally added to the plates which were then incubated at RT overnight in the dark on a shaker. Absorbance at 560 nm was measured using a GloMax multi detection system plate reader (Promega). Cell viability was expressed as a percentage of the value obtained with control cells (cell survival % control). All experiments were carried out in triplicate. All images were drawn using Graph Pad Prism 9.

Western blot analysis

Western-blot analysis was performed to evaluate the protein levels of polβ, LIG1, and LIGIIIα in polβ^+/+ wild-type and polβ^-/- null MEF cells. Both cells were lysed with RIPA buffer containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% Triton, 0.1% SDS, 0.5% Sodium deoxycholate, 2 mM EDTA, and a cocktail of protease inhibitors. Cell lysate protein concentration was measured, an equal amount of total protein (40 µg) was separated on SDS-PAGE and subsequently transferred to a nitrocellulose membrane (Thermo Fisher Scientific). The membranes were blocked with 5% nonfat dry milk followed by probing with primary antibodies overnight with the indicated dilutions below. After washing with Tris-buffered saline containing 0.05% Tween 20 (TBST), a secondary antibody was conjugated to horseradish peroxidase (Invitrogen cat: 31480 at 1:15,000 dilution in TBST), and then was added for 1 hr. Finally, the blots were washed and developed using enhanced chemiluminescence (Pierce™ ECL Western Blotting Substrate Thermo Fisher Scientific cat: PI32106) according to the manufacturer’s protocol and then were imaged (Amersham Imager 680). Antibodies and dilutions were as follows: LIG1 1:200 (Novus Bio cat: NB100-119), LIG IIIα 1:500 (Novus Bio Catalog cat: NBP1-41190), polβ 1:1000, Beta Actin 1:1000 (Cell Signaling cat: 3700).

Results

Impact of DNA ligase inhibitors on the efficiency of nick sealing and binding by LIG1 and LIGIIIα

We first examined the ligation of nick DNA substrates by LIG1 and LIGIIIα in the presence of ligase inhibitors at a concentration range of 0.5–5 µM of L67 and 2.5–20 µM of L82-G17 (Fig. 1A). For LIG1, we showed a decrease in the ligation products along with an increase in the formation of DNA-AMP intermediate in the presence of L82-G17 (Fig. 1B), while there was a greatly reduced nick sealing efficiency by LIG1 in the presence of L67 (Fig. 1C). For LIGIIIα and L67, our results demonstrated a completely diminished nick sealing accompanied by an accumulation of DNA-AMP intermediate products (Fig. 1D). For both inhibitors, we obtained a concentration-dependent decrease in the amount of ligation products by LIG1 and LIGIIIα (Fig. 1E–G).

Figure 1.

Impact of DNA ligase inhibitors on nick sealing by LIG1 and LIGIIIα. (A) Scheme shows nick DNA substrate, and the products of ligation and DNA-AMP intermediate observed in the ligation assays in the absence and presence of the ligase inhibitor. (B–C) Line 1 is the negative enzyme control of the nick DNA substrate with 3ʹ-dG:C, and line 2 is the positive control of ligation reaction by LIG1 in the absence of the ligase inhibitor. Lanes 3–7 are the ligation products by LIG1 in the presence of the ligase inhibitors L82-G17 and L67, respectively. (D) Line 1 is the negative enzyme control of the nick DNA substrate with 3ʹ-dG:C, and line 2 is the positive control of ligation reaction by LIGIIIα in the absence of the ligase inhibitor. Lanes 3–8 are the ligation products by LIGIIIα in the presence of L67. (E-G) Graphs show ligase inhibitor concentration-dependent change in the amount of ligation products and the data represent the average of three independent experiments ± SD.

We then measured the nick DNA binding kinetics of both DNA ligases in the presence of L67 and L82-G17 (Fig. 2). In comparison with nick binding affinity of LIG1 (K_D: 2 nM) in the absence of the ligase inhibitor (Supplementary Fig. 1A), we observed no significant difference in the K_D at 3 and 12 µM concentrations of L67 and there was a completely diminished nick binding at 25 µM (Fig. 2A–C). However, in the presence of L82-G17 at all concentrations tested (12–50 µM), we did not observe any difference in the affinity of nick DNA binding by LIG1 (Fig. 2D–F). Furthermore, in comparison with the nick DNA binding affinity of LIGIIIα (K_D: 6 nM) in the absence of ligase inhibitor (Supplementary Fig. 1B), there was a slight difference at 3 and 12 µM concentrations and no binding at the highest concentrations of L67 (Fig. 2G–I).

Figure 2.

Impact of DNA ligase inhibitors on nick DNA binding by LIG1 and LIGIIIα. (A–F) Real-time binding kinetics of nick DNA are shown for LIG1 in the presence of ligase inhibitors L67 (A-C) and L82-G17 (D–F). (G–I) Real-time binding kinetics of nick DNA are shown for LIGIIIα in the presence of ligase inhibitor L67. The sensorgrams are shown for the concentrations range of DNA ligase (0–640 nM), where nick DNA with a biotin label is immobilized on the streptavidin biosensors.

Efficiency of nick sealing after polβ nucleotide insertion in the presence of DNA ligase inhibitors

We further investigated the impact of the ligase inhibitors on nick sealing by LIG1 and LIGIIIα after polβ correct nucleotide insertion in the presence of L67 and L82-G17. In the control experiments in the absence of the ligase inhibitors, we observed polβ dGTP:C insertion products (after an incorporation of dGTP opposite C in a gap repair intermediate by polβ) and their conversion into ligation products (sealing of resulting nick repair product by DNA ligase) as a function of time in the reaction. For LIG1 and L82-G17 (Fig. 3A), our results demonstrated less efficient conversion of polβ insertion products to a complete ligation product by the effect of ligase inhibition (Fig. 3B, lanes 2–8) in comparison with the control experiment without the inhibitor (Fig. 3B, lanes 9–15). We observed a time-dependent increase in nick sealing products and obtained ~2-fold difference in the amount of ligation products in the absence versus the presence of L82-G17 (Fig. 3C). For LIGIIIα and L67 (Fig. 4A), our results showed a time-dependent increase in nick sealing of polβ insertion products in the control reaction (Fig. 4B, lanes 2–7), while the conversion of these insertions to ligation products by LIGIIIα is greatly reduced in the presence of L67 (Fig. 4B, lanes 8–13). For the initial time points of the reaction, we observed ~4 to 6-fold difference in the amount of ligation products (Fig. 4C). The comparisons of the ligation products in the presence of LIG1/L82-G17 versus LIGIIIα/L67 demonstrated the impact of ligase inhibition on the interplay between polβ and BER ligases for the nick sealing efficiency after polβ dGTP:C insertion (Supplementary Fig. 2). In addition, we tested the impact of L67 and L82-G17 on polβ gap filling activity in the insertion assays and our results showed a slight decrease in the amount of polβ dGTP:C insertion products at the initial time points (10 and 30 s) of the reaction in the presence of both ligase inhibitors L67 and L82-G17, while the later time points demonstrate no significant difference (Supplementary Fig. 3).

Figure 3.

Impact of ligase inhibitors on the nick sealing of polβ nucleotide insertion products by LIG1. (A) Scheme shows gap DNA substrate and products of polβ dGTP:C insertion and ligation by LIG1 observed in the coupled assays in the absence and presence of the ligase inhibitor L82-G17. (B) Line 1 is the negative enzyme control of gap DNA substrate, lanes 2–8 and 9–15 are the ligation products by LIG1 after polβ dGTP:C insertion in the presence and absence of the ligase inhibitor L82-G17 (12 µM), respectively, and correspond to 10, 30, 60, 90, 120, 180, and 210 s. (C) The graph shows time-dependent changes in the amount of ligation products, and the data represent the average of three independent experiments ± SD.

Figure 4.

Impact of ligase inhibitor on the nick sealing of polβ nucleotide insertion products by LIGIIIα. (A) The scheme shows gap DNA substrate and products of polβ dGTP:C insertion and ligation by LIGIIIα observed in the coupled assays in the absence and presence of the ligase inhibitor L67 (3 µM). (B) Line 1 is the negative enzyme control of gap DNA substrate, lanes 2–7 and 8–13 are the ligation products by LIGIIIα after polβ dGTP:C insertion in the absence and presence of the ligase inhibitor L67, respectively, and correspond to 30, 60, 90, 120, 180, and 210 s. (C) The graph shows time-dependent changes in the amount of ligation products, and the data represent the average of three independent experiments ± SD.

Impact of DNA ligase inhibitors on cytotoxicity to oxidative stress in BER cells

To understand the impact of the ligase inhibition on the viability of the model BER cells, we investigated the cell survival in polβ^+/+ wild-type and polβ^-/- null MEF cells in the presence of the DNA ligase inhibitors. We first validated that the protein expression levels of LIG1 and LIGIIIα that were found to be similar in both cells based on the western-blot analysis (Supplementary Fig. 4).

Our results demonstrated that polβ^+/+ wild-type and polβ^-/- null cells exhibit similar sensitivity to the ligase inhibitors L67 and L82-G17 (Fig. 5A–B and Supplementary Fig. 5). We previously reported that polβ^+/+ wild-type cells exhibit more sensitivity to oxidative stress-inducing agent KBrO₃ than polβ^-/- null cells [25]. Accordingly, we observed more cell killing in polβ^+/+ wild-type cells than polβ^-/- null cells as a function of 0–1 mM concentration range of KBrO₃ (Fig. 5C).

Figure 5.

Impact of DNA ligase inhibition on cellular response in BER cells. (A-B) Plots show cell survival of polβ^+/+ wild-type and polβ^-/- null cells in the presence of the DNA ligase inhibitors L67 (A) and L82-G17 (B). The data represent mean values ± SD from three independent experiments. (C) The plot shows cell survival of polβ^+/+ wild-type and polβ^-/- null cells as a function of KBrO₃ concentration. The data represent mean values ± SD from three independent experiments.

We then further investigated the impact of the ligase inhibition by the effect of L67 or L82-G17 on the cell survival in both BER cells after their co-treatment with KBrO₃. Our results showed that the ligase inhibitors L67 and L82-G17 slightly enhance hypersensitivity to this genomic stress in both polβ^+/+ and polβ^-/- cells (Fig. 6A–B). When we compared the IC₅₀ levels (Fig. 6C), our results demonstrated no significant difference between MEF cells in comparison with the alone treatments (Supplementary Fig. 6).

Figure 6.

Impact of DNA ligase inhibition on cytotoxicity to oxidative stress in BER cells. (A–B) DNA ligase inhibitors enhance oxidative stress in polβ^+/+ wild-type and polβ^-/- null cells co-treated with KBrO₃ and L67 (A) or L82-G17 (B). The data represent mean values ± SD from three independent experiments. (C) The table shows IC₅₀ levels of polβ^+/+ and polβ^-/- cells for co-treatment of L82-G17 (inhibitor of LIG1) or L67 (inhibitor of both LIG1 and LIGIIIα) and KBrO₃.

Impact of DNA ligase inhibitors on mutagenic nick sealing of 8oxoG-containing repair intermediates

We then investigated the ligation efficiency after polβ oxidized nucleotide (8-oxodGTP) insertion by LIG1 and LIGIIIα in the presence of the ligase inhibitors in vitro (Figs 7A and 8A). For LIG1, in the absence of the ligase inhibitor, in line with our previous report [25], we obtained mutagenic ligation of polβ 8-oxodGTP:A insertion product appeared along with the ligase failure products with 5ʹ-AMP (Fig. 7B, lanes 2–5). In the presence of the ligase inhibitors L82-G17 and L67 (Fig. 7B, lanes 6–9 and 10–13, respectively), our results showed a decrease in nick sealing efficiency after polβ 8-oxodGTP:A insertion. The quantification of the reaction products showed no significant difference in the amount of ligation products and time-dependent increase in the formation of ligation failure products by the effect of both ligase inhibitors (Fig. 7C–D). For LIGIIIα, in the absence of the ligase inhibitor, we obtained the products of mutagenic ligation and ligase failure after polβ 8-oxodGTP:A insertion (Fig. 8B, lanes 2–5). In the presence of the ligase inhibitor L67, our results showed a diminished nick sealing (Fig. 8B, lanes 6–9). There was ~4-fold decrease in the amount of mutagenic nick-sealing products and time-dependent increase in the formation of ligation failure products by LIGIIIα in the presence of L67 (Fig. 8C–D).

Figure 7.

Impact of the ligase inhibitors on the nick sealing of polβ oxidized nucleotide insertion products by LIG1. (A) Scheme shows gap DNA substrate, and products of polβ 8-oxodGTP insertion and ligation observed in the coupled assays in the absence and presence of the ligase inhibitor. (B) Line 1 is the negative enzyme control of gap DNA substrate, and lanes 2–5 are the reaction products in the absence of the ligase inhibitor. Lanes 6–9 and 10–13 are the reaction products by LIG1 after polβ 8-oxodGTP:A insertion in the presence of L82-G17 and L67, respectively, and correspond to 0.5, 1, 3, and 5 min. (C–D) Graphs show time-dependent changes in the amount of ligation and failure products, and the data represent the average of three independent experiments ± SD.

Figure 8.

Impact of the ligase inhibitors on the nick sealing of polβ oxidized nucleotide insertion products by LIGIIIα. (A) Scheme shows gap DNA substrate and products of polβ 8-oxodGTP insertion and ligation observed in the coupled assays in the absence and presence of the ligase inhibitor. (B) Line 1 is the negative enzyme control of gap DNA substrate, and lanes 2–5 are the reaction products in the absence of L67. Lanes 6-9 are the reaction products by LIGIIIα after polβ 8-oxodGTP:A insertion in the presence of L67 and correspond to 0.5, 1, 3, and 5 min. (C-D) Graphs show time-dependent changes in the amount of ligation and failure products, and the data represent the average of three independent experiments ± SD.

We then tested the impact of the ligase inhibitors on the ligation efficiency of nick DNA containing preinserted 3ʹ-8oxodG:A that mimics polβ 8oxodGTP:A insertion products (Supplementary Fig. 7A). In comparison with the mutagenic nick sealing by LIG1 in the absence of the ligase inhibitor, we showed that both L67 and L82-G17 lead to a decrease in the ligation efficiency and result in the formation of DNA-AMP intermediates (Supplementary Fig. 7B). In the presence of LIGIIIα and L67, we observed a diminished nick sealing of 3ʹ-8oxodG:A along with an accumulation of DNA-AMP intermediates (Supplementary Fig. 7C).

Impact of DNA ligase inhibitor binding mutants on nick binding and sealing efficiency of LIG1

DNA ligase inhibitors have been developed using computer-aided drug design and were predicted to bind to the putative DNA binding pocket of LIG1 (53-56). These are referred as the ligase inhibitor binding sites Gly(G)448, Arg(R)451, and Ala(A)455 residing in the DBD within C-terminal catalytic core of LIG1 protein (Fig. 9). To further examine the impact of ligase inhibition, we generated the amino acid substitutions at the inhibitor binding sites; G448D, R451D, and A455D; and characterized all three LIG1 mutants for their nick DNA binding affinity, protein stability, and the ligation efficiency of nick DNA substrates.

Figure 9.

Structure modeling of DNA ligase inhibitor binding mutants. (A–C) Ribbon diagrams showing LIG1 encircling nick DNA. The domains of the protein are depicted as the N-terminal domain (amino acids 1–262), and the catalytic core (amino acids 262–919) consisting of the DNA-binding domain (DBD), Adenylation domain (AdD), and OB-fold domain (OBD). LIG1 residues G448, R451, and A455 are putative DNA ligase inhibitor binding sites residing in DNA binding pocket are represented by sticks. (D) Schematic view showing the domain composition of LIG1 and the positions of the ligase inhibitor binding residues G448, R451, and A455.

We first investigated nick DNA binding affinity of the LIG1 mutants in real-time and questioned whether the mutations affect protein stability and folding by thermal shift denaturation assays (Fig. 10). Our real-time measurements demonstrated no significant difference in nick DNA binding between LIG1 wild-type (Supplementary Fig. 1A) and G448D mutant (Fig. 10A). However, we showed that R451D and A455D mutations at the ligase inhibitor binding site of LIG1 significantly affected the nick DNA binding (Fig. 10B–C) with ~10-fold difference in the K_D of wild-type protein (Fig. 10D). Furthermore, thermal stability assays in the absence and presence of nick DNA demonstrated minor differences in the melting temperature (T_m) between wild-type (~44°C) and the ligase inhibitor binding mutants of LIG1 which demonstrates that mutations in the inhibitor binding site do not disrupt global protein structure and folding (Fig. 10E and Supplementary Fig. 8).

Figure 10.

Nick DNA binding affinity and thermal stability of LIG1 mutants. (A–D) Real-time binding kinetics of nick DNA are shown for LIG1 mutants G448D (A), R451D (B), A455D (C). Table shows the comparison of the equilibrium binding constants (K_D) between LIG1 wild-type and mutants (D). The sensorgrams are shown for the concentrations range of LIG1 (0–640 nM), where nick DNA with a biotin label is immobilized on the streptavidin biosensors. (E) The table shows the average T_m values ± SD for LIG1 wild-type and mutants in the absence and presence of nick DNA. T_m values were calculated manually using the inverse derivative plot of RFU values using the average of three independent repeats.

To understand the impact of mutations at the ligase inhibitor binding sites on the end joining ability of LIG1 in vitro, we next performed the ligation assays using the nick DNA substrates with preinserted canonical 3ʹ-dG:C, as well as damaged 3ʹ-8oxodG:A and 3ʹ-8oxodG:C ends in vitro (Fig. 11A). Our results showed that LIG1 mutants G448D, R451D, and A455D can ligate the nick DNA containing 3ʹ-dG:C (Fig. 11B). However, in comparison with the ligation efficiency by wild-type LIG1 (Supplementary Fig. 9B, lanes 2–7), we obtained significant difference in the amount of ligation products with LIG1 mutants (~90-fold at 30 s) particularly at earlier time points of the reaction (Fig. 11C). For the nick DNA substrate containing 3ʹ-8oxodG:A, in comparison with LIG1 wild-type (Supplementary Fig. 9B, lanes 9–14), we observed both mutagenic nick sealing and ligation failure products for all three LIG1 mutants (Fig. 12A). The efficiency of ligation by LIG1 G448D mutant was similar to wild-type enzyme, while we observed less efficient mutagenic nick-sealing products for LIG1 R451D and A455D mutants (Fig. 12B). In the presence of nick DNA containing 3ʹ-8oxodG:C, in comparison with LIG1 wild-type (Supplementary Fig. 9B, lanes 16–21), LIG1 G448D mutant showed similar ligation efficiency (Fig. 12C). However, we did not observe any nick sealing of 3ʹ-8oxodG:C by LIG1 R451D and A455D mutants (Fig. 12D). In addition, our results showed different amount of the ligation failure products (5ʹ-AMP) for the LIG1 inhibitor binding mutants depending on the type of the mutation (Supplementary Fig. 10).

Figure 11.

Ligation of the nick DNA substrate with canonical end by LIG1 mutants. (A) Scheme shows nick DNA substrate and products of ligation and DNA-AMP intermediate observed in the ligation assays. (B) Line 1 is the negative enzyme control of the nick DNA substrate with 3ʹ-dG:C. Lanes 2–7, 8–13, and 14–19 are the ligation products by LIG1 mutants G448D, R451D, A455D, respectively, and correspond to 0.5, 1, 3, 5, 8, and 10 min. (C) The graph shows time-dependent changes in the amount of ligation products, and the data represent the average of three independent experiments ± SD.

Figure 12.

Ligation of the nick DNA substrate with damaged end by LIG1 mutants. (A,C) Line 1 is the negative enzyme control of the nick DNA substrates with 3ʹ-8oxodG:A (A) and 3ʹ-8oxodG:C (C). Lanes 2-7, 8-13, and 14-19 are the ligation products by LIG1 mutants G448D, R451D, A455D, respectively, and corresponds to 0.5, 1, 3, 5, 8, and 10 min, in the presence of nick DNA substrates with 3ʹ-8oxodG:A (A) and 3ʹ-8oxodG:C (C). (B, D) Graphs show time-dependent changes in the amount of ligation products and the data represent the average of three independent experiments ± SD.

Structure and function analyses of DNA ligase inhibitor binding mutants

To gain insight into the possible structure–function mechanism underlying the defects in the ligase inhibitor binding mutants, we performed homology modeling to determine the impact of LIG1 mutations using the crystal structure of human LIG1/nick DNA complex [52]. Homology modeling revealed that the mutations at the ligase inhibitor binding sites provide small perturbations in the structure of the DNA binding pocket as shown in the wild-type protein (Fig. 13A). From these models, G448D is predicted to have the smallest effect on DNA binding since the negatively charged carboxyl moiety of D448 is predicted to be farther from the DNA backbone than the other single mutants (Fig. 13B). R451D is predicted to have a more detrimental effect on DNA binding as the R451D mutation flips the charge at that residue from positive to negative, leading to charge-charge repulsion (Fig. 13C). A455D is predicted to be the mutation that places the negatively charged carboxyl moiety of D455 closest to the negatively charged DNA backbone, resulting in a more severe charge-charge repulsion (Fig. 13D).

Figure 13.

Homology modeling of DNA ligase inhibitor binding mutants. (A-D) Homology models are shown for LIG1 wild-type (A), G448D (B), R451D (C), A455D (D). LIG1 residues G448, R451, and A455 are putative DNA ligase inhibitor binding sites residing in DNA binding pocket are represented by sticks.

Impact of polβ inhibitor on the ligation of polβ nucleotide insertion products

In addition to the DNA ligase inhibitors, in the present study, we investigated the efficiency of nick sealing after polβ dGTP:C insertion in the presence of polβ inhibitor SKG-63-1 (44–46). We first tested the effect of SKG-63-1 on polβ dGTP:C insertion activity at 0-10 µM concentration range (Supplementary Fig. 11A). Our results demonstrated the inhibitor concentration-dependent decrease in the insertion products in comparison with a time-dependent increase in polβ insertion in the control reaction (Supplementary Fig. 11B–C). There was ~4-fold decrease in the amount of polβ dGTP:C insertion products at 10 µM of polβ inhibitor SKG-63-1 (Supplementary Fig. 11D). Furthermore, we wanted to test if the polβ inhibitor has any effect on the nick-sealing activity of LIG1 and LIGIIIα. Our results demonstrated no significant effect on the ligation efficiency of both BER ligases for nick DNA substrate with preinserted 3ʹ-dG:C in the presence of SKG-63-1 (Supplementary Fig. 12).

We then investigated the ligation of polβ dGTP:C insertion products by LIG1 and LIGIIIα in the presence of the polβ inhibitor (Fig. 14A). For both DNA ligases, our results demonstrated no difference in the efficiency of nick sealing at 5 µM of SKG-63-1 in comparison with efficient ligation of polβ dGTP:C insertion products in the control reaction without polβ or ligase inhibitor (Fig. 14B). This could be due a relatively slight effect on the gap filling activity of polβ itself by the polβ inhibitor SKG-63-1 at this low concentration. However, at the highest concentration (10 µM) of the polβ inhibitor SKG-63-1, LIG1 and LIGIIIα attempt sealing a gap DNA itself resulting in the formation of gap ligation product (Fig. 14C). Gap ligation versus a complete ligation of nick polβ dGTP:C insertion product by LIG1 and LIGIIIα was revealed by the difference in the size of these products (Fig. 14C, lane 2 versus lanes 3–8 and 9–14). Our results demonstrated ~2-fold difference in the amount of the insertion products between 5 and 10 µM concentrations of the inhibitor. However, for 10 µM inhibitor concentration, the products of coupled reaction were mainly gap ligation, where both DNA ligases attempt to ligate gap DNA itself without polβ insertion. The reason of the difference in terms of polβ insertion products that were observed in the insertion assays, but not in the coupled assays at 10 µM inhibitor concentration, could be due to the intrinsic ability of both ligases to bind to gap DNA. LIG1 and LIG3α can bind to one nucleotide gap in the case of slower or diminished gap filling by polβ, which generates a gap ligation product as we previously reported [26,27].

Figure 14.

Impact of polβ inhibition on nick sealing efficiency of insertion products by LIG1 and LIGIIIα. (A) The scheme shows gap DNA substrate, and the products of polβ dGTP:C insertion and ligation observed in the coupled assays in the absence and presence of polβ inhibitor SKG-63-1. (B) Line 1 is the negative enzyme control of gap DNA substrate, lanes 2–7 and 8–13 are the ligation products by LIG1 and LIGIIIα, respectively, after polβ dGTP:C insertion in the presence of 5 µM polβ inhibitor SKG-63-1, and correspond to 10, 30, 60, 90, 120, and 180 s. (C) Line 1 is the negative enzyme control of gap DNA substrate, and line 2 is the positive control of ligation reaction after polβ dGTP:C insertion in the absence of polβ inhibitor. Lanes 3–8 and 9–14 are the ligation products by LIG1 and LIGIIIα, respectively, after polβ dGTP:C insertion in the presence of 10 µM polβ inhibitor SKG-63-1, and correspond to 10, 30, 60, 90, 120, and 180 s.

Discussion

The BER repairs genotoxic and oxidative lesions as well as DNA base modification products that are important for both the etiology and treatment of cancer [4]. The BER pathway is a multi-protein process that relies on the assembly, function, and coordination of a large repair protein complexes and requires a tight coordination that is mediated by protein-protein interactions between repair proteins to hand off the potentially toxic and mutagenic DNA intermediates from one step to the next in a sequential manner to repair a base damage efficiently [7,11]. The downstream steps of the BER pathway involve a coordination between gap filling by polβ and subsequent nick sealing by LIG1 or LIGIIIα [15]. As it has been considered that the protein-protein interactions could be a promising strategy for the development of new drugs [66], we suggest that the functional interplay between polβ and BER ligases at the final steps could be one of these directions to develop small molecules against the interactome of downstream enzymes. However, the concept of redundancy and ‘backup’ repair pathways could be challenging. Indeed, the BER ligases, LIG1 and LIGIIIα, finalize almost all DNA excision repair pathways including nucleotide excision repair and mismatch repair with the last nick-sealing step [67]. LIG1 also plays a critical role in the maturation of Okazaki fragments in DNA replication [68]. In addition, the intermediates of the BER pathway, such as AP-sites and strand breaks, are themselves genotoxic [12–14].

Reactive oxygen species (ROS) are highly reactive molecules produced by endogenous and exogenous sources such as mitochondria, ionizing radiation, and environmental toxicants [69]. Both genomic DNA and the nucleotide pool (deoxyribonucleotides or dNTPs) are significant targets of ROS, and Guanine is particularly susceptible to oxidation due to its lowest redox potential [70–72]. During DNA replication and repair, oxidized dNTPs such as 8-oxo-2ʹ-deoxyguanosine-5ʹ-triphosphate (8-oxodGTP) can be incorporated by DNA polymerases, which could generate a nick product with damaged end for DNA ligase [73]. In our previous studies, we reported that DNA ligation step is compromised after polβ insertion of 8-oxodGTP, leading to DNA ligase failure by LIG1 or LIGIIIα, and the formation of abortive ligation products containing 5ʹ-adenylate (AMP) in vitro [25]. We have also shown that polβ^-/- mouse embryonic fibroblast (MEF) cells display an increased resistance to the oxidative stress-inducing agent potassium bromate (KBrO₃) in comparison to polβ^+/+ cells and MTH1 gene deletion enhances this cytotoxicity to KBrO₃ in polβ^+/+ cells, suggesting the role of oxidized nucleotide incorporation by polβ on the cellular sensitivity phenotype to oxidative stress [25]. However, the role of DNA ligases for this polβ-mediated oxidized nucleotide incorporation-induced cell death in BER cells under oxidative stress conditions remains unknown. Small molecule inhibitors developed against polβ and DNA ligases have been considered as promising candidates for the development of anticancer agents. DNA ligase inhibitors have been characterized using computer-aided rational drug design strategy based on the DNA binding pocket of LIG1 by Dr. Alan Tomkinson’s laboratory to identify potential competitive and uncompetitive compounds that prevent DNA binding of human LIG1, LIGIIIα, and LIGIV [53–59].

In the present study, we used the ligase inhibitor compounds L67 (an inhibitor of both LIG1 and LIGIIIα) and L82-G17 (an inhibitor of LIG1) to understand the impact of ligase inhibition on cell survival in model BER cells. Our cell experiments showed more sensitivity to KBrO₃ in polβ^+/+ wild-type cells than polβ^-/- null cells, which is consistent with our previous report [25]. Yet, the polβ^+/+ and polβ^-/- cells exhibit similar sensitivity to the ligase inhibition when treated with L67 or L82-G17. Furthermore, we observed a slightly increased cytotoxicity to KBrO₃ upon the inhibition of the BER ligases, which was more effective at lower concentrations of L67. In polβ^-/- null cells, we demonstrated that ablation of polβ results in a modest rescue in cell survival, which could be due to the other polymerases such as polλ that can complement polβ deficiency during the BER of oxidative DNA damage as reported previously in MEF cells [74]. It is also important to note the potential incorporation of 8-oxodGTP during DNA replication by replicative polymerases, especially in mitochondria, which also leads to the formation of genomic 8oxodG, resulting in further production of cytotoxic strand breaks [75]. In addition, it has been shown that LIGIIIα plays a back-up role in DNA replication while LIG1 is considered as a predominant ligase that is responsible for joining Okazaki fragments during replicative DNA synthesis [76]. Furthermore, the impact of LIG1 and/or LIG3 gene depletion in polβ^+/+ and polβ^-/- cells should be explored to compare the impact of ligase inhibition on the polβ/BER ligase interplay at the downstream steps.

In our cell survival studies, we selected KBrO₃ as oxidative stress-inducing agent that has been shown to cause the formation of reactive metabolites that preferentially oxidize Guanine residues in the nucleotide pool [77]. It has been previously reported that polβ^-/- cells show more hypersensitivity to another oxidizing agent H₂O₂ than polβ^+/+ cells [74]. We believe the difference is likely due to the nature of the oxidizing agent-mediated DNA damage in cells as H₂O₂ has been widely utilized to induce single-strand breaks and various forms of base damage in mammalian cells [78–80]. Further cell experiments in both BER cells with H₂O₂ and the ligase inhibitors should be investigated to compare the impact with that of KBrO₃. We suggest that KBrO₃-induced oxidized nucleotide accumulation is consistent with an adverse effect of polβ-mediated oxidized nucleotide insertion leading to ligase failure in the presence of the ligase inhibitors and resulting accumulation of stalled BER intermediates that can trigger more cytotoxicity in polβ^+/+ cells.

Our results with purified proteins and the ligase inhibitors in reconstituted in vitro assays showed that the ligase inhibition results in a decrease in the efficiency of ligation for nick repair intermediates as well as in nick sealing products of polβ nucleotide insertion by LIG1 and LIGIIIα. Accordingly, we showed an increase in the ligase failure products following polβ oxidized nucleotide 8-oxodGTP:A insertion in vitro. In addition to the ligase inhibitors, we also tested the impact of polβ inhibition on the ligation of nick repair product and observed that the BER ligases ligate the gap DNA itself resulting in the formation of aberrant repair intermediate, i.e. gap ligation product, in case of no gap filling by polβ by the effect of the inhibitor. These findings could contribute to the understanding of the importance of functional interplay between polβ and DNA ligase at the downstream steps of the BER pathway. These studies could be expanded to better understand the impact of scaffolding factors and other repair proteins, such as X-ray repair cross-complementing protein 1 (XRCC1) and Poly(ADP-ribose) polymerase 1 (PARP1), on the repair pathway coordination involving polβ and LIG1/LIGIIIα [4].

In the present study, we also investigated how the mutations at the ligase inhibitor binding sites of LIG1 could affect the nick binding and sealing. Through the biochemical interrogation of the LIG1 mutants, our results demonstrated that the end-joining activity of LIG1 is compromised by G448D, R451D, and A455D mutations. Our previous structures of LIG1 revealed that the active site engages with base substitution errors that mimic polβ mismatch insertion products and ribonucleotides depending on the base pairing feature of 3ʹ-terminus/template base at the nick site [81,82]. Further structure/function studies with LIG1 mutants should be conducted to better understand the effect of mutations at the ligase inhibitor binding sites on the fidelity of LIG1 at atomic resolution. Furthermore, X-ray crystallography studies of LIG1 bound to small molecule inhibitor compounds will be aimed to uncover atomic insight into the mechanism of ligase inhibition.

Overall findings demonstrate that the deviations in the interplay between polβ and LIG1/LIGIIIα by the effect of small molecule inhibitors that target gap filling and nick sealing at the downstream steps of the BER pathway could affect repair outcomes and could be the source of toxic and mutagenic repair intermediates. This coordination at the final steps might be considered as a promising target to develop a new generation of small molecule inhibitors to increase the selectivity and efficiency of current DNA ligase and/or polβ inhibitors.

Supplementary data

Supplementary data is available at Mutagenesis Online.

Conflict of interest

None declared.

Funding

This work was supported by a grant 1R35GM147111-01 from the National Institute of General Medical Sciences (NIGMS).

Data availability

All data are contained within the manuscript. Further information and requests for materials used in this research are available from the authors upon reasonable request and should be directed to Dr Melike Çağlayan (caglayanm@ufl.edu).