Introduction

DNA polymerase β (Polβ) is the primary DNA polymerase responsible for gap-filling synthesis during the base excision repair (BER) pathway^1,2. This enzyme processes the central 1-nucleotide (nt) gap intermediate through two distinct activities: its polymerase domain catalyzes nucleotide insertion, while its N-terminal lyase (8-kDa) domain excises the 5’-deoxyribose phosphate (5’-dRP) moiety, enabling the short-patch (SP) BER sub-pathway. Polβ can also perform limited strand-displacement synthesis, a prerequisite for the alternative long-patch (LP) BER sub-pathway where 2-11 nucleotides are incorporated³. A defining biochemical feature of Polβ is its distributive mode of synthesis. Unlike processive replicative polymerases, Polβ does not remain tightly bound to its DNA template but instead dissociates after incorporating a short patch of 2-3 nucleotides⁴. However, this distributive kinetics also means that the overall efficiency of DNA synthesis is highly dependent on the enzyme’s rebinding rate. The resulting flap is cleaved by flap endonuclease 1 (FEN1), which stimulates Polβ activity in strand-displacement synthesis, in accordance with the “passing the baton” mechanism ^5–7. Notably, the fidelity of Polβ decreases substantially during strand-displacement synthesis compared to single-nucleotide gap filling, highlighting the critical importance of regulating its catalytic choice between these two modes⁸. Overexpression of Polβ, as observed in certain malignancies, is thought to promote a shift towards the long-patch BER (LP-BER) pathway and, under specific conditions, may even facilitate its aberrant recruitment into the nucleotide excision repair (NER) pathway. Such misregulation ultimately contributes to increased mutation rates and genomic instability^9,10.

The progression and efficiency of BER are tightly regulated by various factors, prominently including poly(ADP-ribose) polymerases 1 and 2 (PARP1 and PARP2). These enzymes function as molecular sensors of DNA strand breaks, binding to damaged DNA and catalyzing the synthesis of poly(ADP-ribose) (PAR) chains¹¹. PARylation serves as a signal for the recruitment of DNA repair factors but also modulates enzyme activity; for instance, automodification of PARP1 and PARP2 promotes their dissociation from DNA, thereby facilitating access for repair enzymes like Polβ^12–15. PARP1 and PARP2 exhibit distinct affinities for different BER intermediates: PARP1 binds effectively to undamaged DNA, AP-sites, gaps and nicks, whereas PARP2 shows a marked preference for gapped and nicked DNAs^12,13,16,17. Consequently, PARP1 influences all stages of BER^12,18, while PARP2 acts as a key regulator of final steps, including the handoff between Polβ and DNA ligase IIIα (LigIIIα)¹⁴.

A critical layer of DNA repair regulation is added by the packaging of genomic DNA into chromatin. The nucleosome core particle (NCP), the fundamental repeating unit of chromatin, presents a significant structural barrier to DNA repair machinery¹⁹. Extensive biochemical and recent structural studies, comprehensively detailed in an excellent recent review²⁰, have demonstrated that the activities of DNA glycosylases, apurinic/apyrimidinic endonuclease 1 (APE1), Polβ, and LigIIIα are substantially suppressed on nucleosomal DNA compared to naked DNA^14,21–24. The translational and rotational positioning of a DNA lesion within the nucleosome profoundly influences its accessibility to BER enzymes^25–28. For example, the activity of Polβ in gap-filling synthesis within the nucleosome follows a strong position-dependent gradient, being most efficient near the entry/exit sites^25,29 or linker DNA²⁶. Cryo-EM structures reveal that the reduced activity of Polβ in processing low-accessible lesions stems from the necessity for enzyme to engage in extensive “global DNA sculpting,” displacing ∼35 base pairs of DNA from the histone octamer to induce a ∼90° DNA bend²⁹. Similarly, ligation of the final nick by the XRCC1-LigIIIα complex is efficient only near the nucleosome periphery²³. This constrained repair capacity within chromatin is biologically significant, as it correlates with elevated DNA repair efficacy and mutation rates in nucleosome-dense genomic regions in vivo^30–32.

To overcome the chromatin barrier, cells employ active chromatin remodeling strategies^33,34. PARP1 and PARP2, in complex with the histone PARylation factor 1 (HPF1), play a pivotal role in this process by catalyzing PARylation of core histones, primarily on serine residues ^35,36. This HPF1-dependent histone PARylation triggers local chromatin relaxation facilitating the access of repair factors to lesions³⁷. While histone PARylation may directly relax nucleosome structure to facilitate Polβ activity on nucleosomal substrates in vitro³⁸, its principal function in vivo is likely the targeted recruitment of chromatin remodelers^39–42. ADP-ribosylation of histones in the vicinity of a nucleosomal DNA lesion, which is most efficiently catalyzed by PARP2^16,17,43, might serve as a signal to recruit ALC1. This remodeler would then catalyze nucleosome sliding away from the damage site, effectively repositioning the lesion into the more accessible linker DNA, where it becomes fully exposed to the repair machinery^42,44–46.

Whereas the mechanisms of BER within the nucleosome core have been increasingly elucidated, the repair landscape in linker DNA regions – the stretches connecting nucleosome cores – remains less characterized. The nucleosome itself imposes a strong bias for the SP-BER pathway, effectively limiting DNA repair synthesis within the nucleosome core to single-nucleotide insertion; this phenomenon is less pronounced in linker DNA⁴⁷.

It was previously established that Polβ-catalyzed strand-displacement synthesis in linker DNA produced a repair patch whose size is spatially constrained by the proximity of the nucleosome⁴⁷. However, these investigations, performed in nuclear extracts, do not fully delineate the specific interplay between Polβ and its regulators like PARPs in the context of linker DNA. Furthermore, the entry/exit site of the nucleosome is associated with the linker histone H1, which promotes higher-order chromatin compaction and can further hinder the access of repair enzymes to DNA lesions^22,48. Recent findings indicate that the dissociation of H1 from chromatin facilitates DNA repair, suggesting its dynamics is a key regulatory point^22,49.

In this study, we investigate the activity of Polβ on linker DNA and its functional interplay with the linker histone H1 and the PARP1/PARP2 system, including the HPF1 cofactor. We specifically focus on how an adjacent nucleosome core particle and its associated proteins modulate Polβ’s choice between gap-filling and strand-displacement synthesis, a critical decision point in the BER sub-pathway selection.

Results

Stimulation of Polβ activity in Strand-Displacement Synthesis by an Adjacent Nucleosome Core Particle

Previous studies have consistently demonstrated that Polβ activity is substantially constrained within the nucleosome core, while its behavior on linker DNA adjacent to a nucleosome core particle (NCP) remains less explored. To investigate this aspect, we generated a nucleosome substrate with 50 bp and 30 bp linker arms, containing a 1-nt gap (gap-NCP227) located 26 bp from the 5’ end of the longer linker (Fig. 1a). The nucleosome was reconstituted using the Widom 603 positioning sequence, which ensures a well-defined translational and rotational setting of the DNA.

Fig. 1.

Unexpectedly, we found that the efficiency of Polβ-catalyzed strand-displacement synthesis was higher on the gap-NCP227 substrate compared to the respective naked DNA (gap-DNA227). In identical reaction conditions, Polβ generated longer extension products when acting on gap-NCP (Fig. 1a). Kinetic experiments show more efficient strand-displacement synthesis on gap-NCP substrate at each time point (Fig. 1b). The initial rate of the single-nucleotide gap-filling reaction was also higher for gap-NCP than for gap-DNA: 24 fmol/s vs 12 fmol/s (Fig. 1c).

Considering the distributive nature of Polβ strand-displacement synthesis (typically incorporating 3 nucleotides per binding event), we propose several potential explanations for this phenomenon. First, the presence of NCP could influence the binding affinity of Polβ, potentially facilitating more rapid enzyme rebinding to the DNA template. Indeed, we observed a stepwise accumulation of DNA extension products with increasing length by increments of three nucleotides (e.g., +3, +6). This pattern is consistent with a model where the NCP enhances processivity by promoting enzyme rebinding. However, the binding affinity of Polβ was very similar for both gap-DNA227 and gap-NCP227, as detailed below.

Another factor that may influence the efficiency of strand-displacement synthesis is the interaction between the displaced single-stranded DNA flap (5’-flap) and the NCP. It is plausible that the flexible N-terminal tails of core histones could interact with the displaced flap, thereby facilitating destabilization and unwinding of DNA helix and enhancing the synthesis. Notably, the simple addition of free core histones to gap-DNA227 did not stimulate Polβ activity (Supplementary Fig. 1). On the contrary, free histones present at excessive relative to Polβ concentrations suppressed strand-displacement synthesis, likely due to competition with Polβ for the substrate binding via non-specific electrostatic interaction with the DNA.

To investigate interplay between the DNA strand-displacement synthesis and DNA flap cleavage, we analyzed the Polβ polymerase activity in the presence of flap endonuclease 1 (FEN1). Consistent with previous reports ⁷, the addition of FEN1 stimulated overall DNA synthesis by Polβ on both naked gap-DNA227 and gap-NCP227 substrate, as evidenced by an increased yield of long extension products (Fig. 1d). This is attributed to FEN1 resolving inhibitory secondary structures in the displaced flap, allowing Polβ to perform more processive synthesis. Strikingly, a particular pattern of products emerged specifically on the nucleosome substrate. The electrophoretic profile revealed a pronounced attenuation of bands corresponding to the incorporation of every third nucleotide (+2, +5, +8), creating a periodic pattern (Fig. 1d, lane 4). In contrast, the product distribution detected for the naked DNA was more uniform, without the pronounced periodic intensity pattern (Fig. 1d, lane 2). This observation suggests that exactly in the nucleosome, the cleavage of the displaced flap by FEN1 is exceptionally efficient and tightly coupled to the synthetic cycle of Polβ. We hypothesize that upon the formation of a 3-nt flap – a preferred substrate for FEN1 in the presence of NCP – the endonuclease immediately cleaves it, effectively “resetting” the substrate and allowing Polβ to rapidly initiate the next synthetic cycle. This highly coordinated cycle results in the observed rhythmic extension pattern, illustrating a processive “passing the baton” mechanism uniquely facilitated by the nucleosome scaffold.

Interaction of Polβ and its Possible Competitors with Gapped DNA and Nucleosome Substrates

Next, we compared binding of Polβ with its nucleosome substrates using mass photometry (MP), that measures the mass of individual biomolecules, allowing for the quantitative label-free analysis of complex formation ⁵⁰. The theoretical mass of Polβ is 39 kDa, and of the nucleosome 248 kDa (108 kDa histone octamer + 140 kDa DNA). At a 3:1 ratio of Polβ and gap-NCP227 concentrations, we detected a complex with mass of 281±15 kDa (Fig 2a), indicating a 1:1 ratio of interaction (Polβ•gap-NCP227). Complexes of 325, 371 and 414 kDa formed at increased gap-NCP227:Polβ ratio contained two, three and four protein molecules. The calculated theoretical masses of these complexes are presented in Fig. 2b.

Fig. 2.

Additional EMSA experiments indicated formation of higher-order complexes of Polβ with gap-DNA227 and gap-NCP227, confirming the results of the MP assay (Supplementary Fig. 2a). Importantly, we observed no oligomeric complexes of Polβ with non-gapped DNA227 or NCP227 substrates in either EMSA or mass photometry experiments (Supplementary Fig. 2a, b)., indicating that the cooperative binding is characteristic for Polβ interaction with its specific gapped substrate.

Then, we compared binding affinities of Polβ and its possible competitors (linker histone H1, PARP1 and PARP2) for gap-DNA227, gap-NCP227, and nick-NCP227 (the product of 1-nt gap-filling synthesis), using fluorescence titration assay to evaluate EC₅₀ values (Table 1, Supplementary Fig. 2, c-f).

Table 1.

As shown in Table 1, the linker histone H1 exhibits the lowest affinity for DNA and nucleosomes, though its affinity for the NCP was 1.5-fold higher than for the naked DNA (p ≤ 0.05), consistent with its specific binding at the nucleosome entry/exit site. PARP1 and PARP2 displayed comparable affinities for gap-DNA227 and gap-NCP227. However, PARP2 affinity for nick-NCP227 was higher than that of either PARP1 or Polβ (p ≤ 0.01 and 0.05, respectively), aligning with previous reports ^13,14. As expected, Polβ exhibited higher affinity for the gapped substrates than for the single nucleotide insertion product (nick-NCP) (p ≤ 0.05).

Linker Histone H1 Restricts Polβ Strand-Displacement Synthesis at the Nucleosome Entry/Exit Site

Linker histone H1 binds to linker DNA and facilitates chromatin compaction, which could potentially inhibit activities of DNA-dependent enzymes near the NCP boundary. We therefore analyzed the influence of H1 on Polβ-catalyzed strand-displacement synthesis. In our gap-NCP substrate, the incorporation of 24 nucleotides corresponds to synthesis within the linker DNA, and the subsequent synthesis proceeds into the nucleosomal DNA at superhelical locations (SHLs). While previous studies established severe constraints for strand-displacement synthesis within the NCP itself, here we detected the DNA synthesis initiating in the linker region and progressing into the nucleosome (Fig. 3a, lanes 1 and 3).

Fig. 3.

Under near single-turnover experimental conditions (i.e. at equimolar concentrations of Polβ and DNA substrate), we revealed no impact of H1 on Polβ activity on the naked DNA (Fig. 3a, lanes 1, 2). This result can be explained by non-specific binding of H1 to the naked DNA away from the gap site, occupied by Polβ due to its higher affinity for the gapped DNA/NCP (Table 1). In contrast, H1 significantly restricted the length of strand-displacement products synthesized near the nucleosome entry/exit site (Fig. 3a, lanes 3, 4), leading to accumulation of shorter extension products. Our data demonstrate that Polβ strand-displacement activity can be regulated not only by the nucleosome core but by the linker histone H1 also.

PARP1 and PARP2 Distinctly Modulate Polβ Activity on Nucleosomal Substrates

We further explored PARP1/PARP2 induced modulation of strand-displacement synthesis across a range of Polβ concentrations. The both PARPs reduced the yield of strand-displacement products (Fig.4a, b). As expected, the extension by higher concentrations of Polβ was more effective in both the absence and presence of PARP1/PARP2. PARP1 appeared to suppress both 1-nt gap-filling and strand-displacement synthesis due to competition with Polβ for the gapped DNA/NCP substrate. PARP2 was similar to PARP1 in suppression of strand-displacement but less active in inhibiting the first nucleotide insertion, suggesting that PARP2 competes more strongly with Polβ for binding to the nicked DNA/NCP intermediate. Indeed, PARP2 has the highest affinity for nick-NCP227 compared to those of Polβ and PARP1 (Table 1). The difference between PARP1 and PARP2 in their inhibitory effects was further observed in experiments with fixed Polβ and increasing PARP concentrations. Even at a 10-fold excess over Polβ, PARP2 displayed no suppression of nicked intermediate production, in contrast to PARP1, which exhibited strong inhibition (Fig. 4c, d; Supplementary Fig. 3a). PARP2-induced inhibition of the single-nucleotide gap-filling reaction (performed in the presence of dTTP only) could be detected at a lower Polβ concentration and a shorter reaction time, when the extent of substrate conversion was below 50% (Supplementary Fig. 3b). Thus, competition between PARP2 and Polβ for interaction with the 1-nt DNA gap strongly depends on the experimental conditions.

Fig. 4.

Interestingly, the inhibition effects of both PARP1 and PARP2 were more pronounced on the nucleosome substrate than on the naked DNA. The maximum length of product synthesized in the presence of each PARP was always greater on the DNA than on the NCP at identical Polβ concentrations (Fig. 4a, b). Similarly, lower PARP1 and PARP2 concentrations were required to suppress synthesis on the NCP. Although no substantial difference in the binding affinity of Polβ, PARP1, or PARP2 for gap-DNA versus gap-NCP was detected (Table 1), it is evident that the proximal nucleosome modulates the functional interplay between PARP enzymes and Polβ at the DNA lesion.

PARylation Antagonizes the Inhibitory Action of PARPs and H1 on Strand-Displacement Synthesis

In the presence of NAD⁺, PARP1 and PARP2 catalyze poly(ADP-ribose) (PAR) synthesis and transfer to target proteins, including automodification. AutoPARylation promotes the dissociation of PARPs from damaged DNA. Furthermore, these enzymes can PARylate core and linker histones in the presence of HPF1. We analyzed how addition of NAD⁺ modulates the effects of PARP1 and PARP2 on Polβ-catalyzed DNA synthesis on the nucleosome, including in the presence of H1 and HPF1. As expected, the inhibitory action of PARP1 and PARP2 was nearly completely abolished in the presence of NAD⁺; the efficiency of strand-displacement synthesis was restored under PARylation conditions (Fig. 5a, b; compare lanes 1 and 6). PARP1 suppressed the DNA synthesis most efficiently, compared to PARP2 and especially to H1, and this effect predominated when PARP1 and H1 were present together (Fig. 5a, compare lanes 3, 4 and 5). However, effects produced by PARP2 and histone H1 present together were different from those detected for each protein alone (Fig. 5b, compare lane 5 with lanes 3 and 4), indicating combined action of the both proteins. Most intriguingly, the restrictive effect on the length of extension products produced by H1 was not observed in PARylation conditions (Fig. 5a, b; compare lane 4 with lane 7). This fact was also observed in the presence of HPF1 (Supplementary Fig. 4). Additional experiments show that the inhibitory action of PARP1/PARP2 and H1 in PARylation conditions depends on the NAD⁺ concentration known to control the efficiency of PAR elongation (Supplementary Fig. 5). These results suggest that the activity of PARP1 and PARP2 is sufficient to displace H1 from the nucleosome entry/exit site. This displacement could occur either through direct PARylation of H1 or via its non-covalent interaction with PAR⁵¹. To clarify the mechanism, we further analyzed the PARP1/PARP2 catalyzed modification of H1.

Fig. 5.

Linker histone H1 is modified by PARP1 and PARP2 in a HPF1-dependent manner

The PARylation-dependent action of histone H1 on Polβ activity prompted us to explore the direct PARylation of H1. The PARylation reaction was performed using gap-NCP for PARP1/PARP2 activation in the presence and absence of HPF1 (Fig. 6). PARP1 was more active in the automodification than in histone modification, whereas PARP2 catalyzed more efficiently ADP-ribosylation of histones (Fig. 6a, c), consistent with our previous studies ^16,17,43. The presence of H1 reduced the level of HPF1-dependent modification of core histones by the both enzymes due to redistribution of PARylation targets between the linker and core histones (Fig. 6a, c; compare samples 4, 8 with samples 2, 7). In the absence of HPF1, modification of H1 was detected only at its increased concentration as evidenced from analysis of ADP-ribosylation products after treatment of reaction mixtures with poly(ADP-ribose) glycohydrolase (PARG) (Fig. 6b, d, e and Supplementary Fig. 6). Thus, the very low level of HPF1-independent PARylation of H1 by PARP1/PARP2 is most likely insufficient to induce the linker histone dissociation. Additional fluorescence anisotropy experiments showed that while H1 slowed down the dissociation kinetics of PARP1/PARP2 from the complex with gap-NCP, it did not affect the final level of anisotropy (Supplementary Fig. 7)., suggesting that H1 dissociates from the complex together with PARPs under PARylation conditions. Therefore, the eviction of H1 is mediated rather by its high-affinity interaction with PAR (demonstrated previously ^51,52) than by direct PARylation (which is minimal without HPF1). We propose that automodification of PARP1 and PARP2 leads to the recruitment of H1 away from the nucleosome via attraction to PAR, ultimately permitting strand-displacement synthesis to proceed through the nucleosome entry/exit region.

Fig. 6.

Fig. 7.

Interestingly, the yields of PARylated core histones and their lengths were quite different in the absence and presence of H1, but the yields of MARylated products were practically independent on the presence of H1 (compare lanes 2, 4, 7 and 8 in Fig. 6a with respective lanes in Fig. 6b). These data suggest that H1 inhibits primarily modification of core histones at the elongation stage. This could be related to different stabilities of complexes formed by the linker histone and core histones with DNA or speaks in favor of H1 interaction with PAR, which may influence the PAR length⁵³.

Discussion

The canonical view of chromatin as a passive barrier to DNA repair has been increasingly refined to encompass its role as an active regulator of repair pathway choice^19,28,47. While it is well established that the nucleosome core particle (NCP) profoundly suppresses the activity of BER enzymes, most studies have focused on lesions buried within the nucleosomal DNA and the regulatory landscape of the linker DNA has remained less explored⁴⁷. While a lesion within the core DNA is subject to severe steric hindrance^{14,21–24,38}, a lesion in the linker is physically accessible yet its processing is influenced by the nearby NCP. Here, we investigate DNA repair synthesis in the linker DNA and reveal how the adjacent NCP, in conjunction with associated proteins, creates a unique regulatory environment that modulates the activity of Polβ and its decision between short-patch (SP) and long-patch (LP) sub-pathways of BER.

Our most striking finding is that NCP can stimulate Polβ-catalyzed single-nucleotide gap-filling and strand-displacement synthesis on adjacent linker DNA, instead of expected suppression. We propose that the NCP may act as a “steric anchor” promoting the rapid rebinding of Polβ to its DNA substrate, thereby effectively increasing its local concentration and processivity. Considering the distributive nature of DNA synthesis catalyzed by Polβ, such facilitated rebinding would lead to an increased rate of DNA synthesis and a greater number of nucleotides incorporated, which is precisely what we observe in the presence of the nucleosome. This stimulatory effect extends beyond Polβ alone. Our data reveal that the presence of an adjacent nucleosome not only influences Polβ activity but also orchestrates its functional interplay with downstream enzyme of the LP BER pathway, FEN1. The rhythmic pattern of extension products, where bands corresponding to every third nucleotide are faint, detected specifically on the nucleosome substrate, is highly indicative of effective substrate channeling, probably via “passing the baton” mechanism ⁵. Consistent with this model, the interaction of PARP1, Polβ and FEN-1 at the branch-point BER intermediate has been directly demonstrated by their co-precipitation following photocrosslinking to this DNA intermediate in cell extracts⁵⁴. Importantly, this coordination on the nucleosome contrasts with the more stochastic and asynchronous product distribution on naked DNA, indicating that the nucleosome context actively facilitates efficient handoff between repair enzymes. In addition, the inhibitory effects of both PARP1 and PARP2 (discussed below) were more pronounced on the nucleosome substrate than on the naked DNA, despite similar binding affinities. It has been shown previously that the protein assembly for the specific repair process can be coordinated by protein-protein interactions which were determined for the core enzymes of BER^54–57. Here we propose that the nucleosome itself functions as an allosteric platform that modulates the assembly, activity and regulation of the BER machinery and fosters a coordinated substrate channeling. This hypothesis is supported by recent structural work revealed specific mechanisms of nucleosome binding by DNA repair enzymes^58,59.

Furthermore, we demonstrate that Polβ forms high-order complexes on gapped DNA and nucleosome substrates via multimerization. This propensity for self-association has been previously noted for many BER players, including PARP1 and PARP2^17,55,60 and for Polβ on gapped and template-primer structures⁶¹. This observation raises the possibility that multiple cycles of Polβ binding and flap endonuclease-catalyzed cleavage could be coordinated through such multimeric assemblies.

The stimulatory effect of nucleosome is balanced by the suppressive action of linker histone H1, which in turn is controlled by PARP-catalyzed PAR synthesis. The linker histone H1 contributing to chromatin compaction^62–64 limits Polβ-catalyzed DNA repair synthesis at the entry/exit site of nucleosome. Critically, this restriction is reversed by PARP1/PARP2-catalyzed PARylation. While eviction of histone H1 in vivo and its function as a high-affinity PAR-reader in vitro have been established^{49,51,52,65,66}, the dominant mechanism of H1 PARylation-dependent function remains unclear⁶⁷. Our data show that H1 displacement from DNA occurs even when its direct PARylation (detected in the absence of HPF1) is minimal, favoring a model where H1 is sequestered via PAR-mediated interaction with automodified PARP (Fig. 7). This provides a dynamic, PARylation-controlled mechanism for primary chromatin relaxation required for DNA repair.

Our findings refine the distinct roles of PARP1 and PARP2 in BER. The both PARPs are well established to suppress strand-displacement synthesis via competition with BER enzymes for DNA intermediates^{12,13,38,54,68}. The distinct binding preferences of PARP1 and PARP2 for the BER DNA intermediates^13,14,17, coupled with their stage-specific inhibitory effects, suggest a functional specialization: PARP1 appears to modulate predominantly the initial stages of BER, while PARP2 exerts a major impact on the following steps (DNA synthesis and ligation). In our experiments, both enzymes compete with Polβ for substrate binding, but their modes of inhibition are distinct. PARP1 suppresses 1-nt gap-filling and strand-displacement synthesis, consistent with its high affinity for the initial gapped intermediate. In contrast, PARP2, with its pronounced preference for the nicked DNA, specifically and potently inhibits the strand-displacement and is less effective at the stage of the first nucleotide insertion. Thus, the DNA synthesis step itself demarcates the functional boundary between the regulatory spheres of PARP1 and PARP2 within the BER pathway. This positions PARP2 as a dedicated «game changer» of the SP/LP BER decision point. By “capping” the nick after gap-filling, PARP2 prevents strand-displacement, biasing repair towards the short-patch pathway – a critical function in linker DNA, which lacks the inherent physical constraints of the nucleosome core. Importantly, this role of PARP2 in binding and protecting nicked DNA intermediates provides a mechanistic basis for its reported involvement in PARP inhibitor cytotoxicity. It has been proposed that PARP2, trapped on nicks associated with Okazaki fragments during DNA replication, prevents their ligation, leading to cell death^69–71.

Thus, our study elucidates a multi-layered system to regulate Polβ activity at the crucial interface between the nucleosome core and the linker DNA. The different affinities of PARP1 and PARP2 for distinct BER intermediates create a regulatory cascade to choose between short- and long-patch repair. The nucleosome may act as an active participant that influences the organization of repair factors, their kinetics, and the processing of DNA damage.

Materials and methods

Recombinant wild-type human PARP1 and murine PARP2 were purified as described previously⁷². Recombinant wild-type human HPF1 and H1.0 were purified as described previously ⁴³. Human APE1, FEN1 and rat Polβ were purified as described previously^73–76. All proteins were tested for DNA polymerase activity (Supplementary Fig. 8a). E. coli uracil-DNA glycosylase (UDG) and T. aquticus DNA polymerase were from Biosan (Novosibirsk, Russia). Core histones were isolated from Gallus gallus erythrocytes and purified as described previously ⁷⁷. pGEM-3z/603 was a gift from J. Widom (Addgene plasmid #26658; http://n2t.net/addgene:26658; RRID:Addgene_26658). Synthetic oligodeoxyribonucleotides were purchased from Genterra (Moscow, Russia). The ³²P-labelled NAD⁺ was synthesized enzymatically as described previously⁷⁸, using [α-³²P]ATP (with specific activity of 3000 Ci/mmol, synthesized in the Laboratory of Biotechnology, ICBFM, Novosibirsk, Russia). NAD⁺ and reagents for electrophoresis and basic components of buffers were purchased from Sigma–Aldrich (USA). Double-distilled water was used to prepare all buffer solutions and reaction samples. Sequences of the oligonucleotides are presented in Table 2.

Table 2.

Methods

Preparation of DNA and nucleosomes

Model 227 bp DNA containing the 147-bp core of the 603 nucleosome positioning sequence ⁷⁹ was generated by PCR from the pGEM-3z/603 plasmid using the oligonucleotide primers (Table 2). Nucleosomes were assembled by mixing DNA and core histones in a high-salted buffer containing 2 M NaCl, dialyzed against the NaCl concentration gradient from 2 M to 65 mM during 6 h at 4 °C and then against the buffer with 10 mM NaCl overnight at 4 °C with gentle stirring. The optimal ratios of DNA and histones were determined in preliminary quick-time experiment as described in the published protocol⁷⁷. The homogeneity of the DNA and NCP samples was analyzed by the EMSA on a 4% nondenaturing PAG (Supplementary Fig. 8b).

To generate a single-nucleotide gap in DNA substrates, reaction mixtures containing 1 μM DNA or NCP, uracil-DNA glycosylase (UDG, 1 activity unit per 0.6 pmol of DNA/NCP), and 0.1 µM AP endonuclease 1 (APE1) in reaction buffer (50 mM Tris-HCl (pH 8.0), 50 mM NaCl, 5 mM MgCl₂, 1 mM DTT) were incubated for 30 min at 37 °C. To obtain nicked NCP (nick-NCP) substrate, reaction mixtures containing 166 nM gap-NCP, 16.6 nM Polβ and 10 μM dTTP in the reaction buffer were incubated for 10 min at 37 °C. Then the mixture was diluted with buffer (50 mM Tris-HCl (pH 8.0), 50 mM NaCl, 10 mM EDTA, 1 mM DTT) to 6 nM concentration of NCP and incubated for 10 min at 45°C. The homogeneity of nick-NCP was analyzed by EMSA on a 4% nondenaturing PAG (Supplementary Fig. 8b).

Fluorescence studies of Polβ/H1/PARP1/PARP2 interaction with DNAs/NCPs

The affinity of proteins for DNA and NCP substrates was evaluated by determining the half-maximal effective concentration (EC₅₀) of protein complexes with FAM-labeled DNA or NCP using fluorescence anisotropy measurements ⁸⁰. Briefly, reaction mixtures containing 0–400 nM PARP1/PARP2/H1/Polβ and 3 nM of gap-DNA (227 bp), gap-NCP, or nick-NCP in binding buffer (50 mM Tris-HCl pH 8.0, 50 mM NaCl, 10 mM EDTA, 1 mM DTT) were prepared on ice in a 384-well plate and incubated for 5 min at room temperature. Fluorescence anisotropy measurements were performed using a CLARIOstar microplate reader (BMG Labtech). Fluorescent probes were excited at 495 nm, and emission was detected at 520 nm. Each measurement consisted of 50 flashes per well, and fluorescence intensity values were averaged automatically. Measurements for each well were performed three times at 1-min intervals. The averaged values were used for final plotting, and EC₅₀ values were calculated using the SMART Control Data Analysis software (BMG LABTECH). The data were plotted (F vs C) and fitted to four-parameter logistic equation: F = F₀ + (F_∞ - F₀)/[1 + (EC₅₀/C)ⁿ], where F is the measured fluorescence anisotropy of a solution containing the labeled DNA at a given concentration (C) of proteins, F₀ is the fluorescence anisotropy of solution of the labeled DNA alone, F_∞ is the fluorescence anisotropy of the labeled DNA saturated with the protein, EC₅₀ is the concentration of the protein at which F - F₀ = (F_∞ - F₀)/2, and n is the Hill coefficient, which denotes the steepness (slope) of the nonlinear curve.

Mass Photometry measurements of Polβ interaction with gap-NCP

The stoichiometry of Polβ complexes with gap-NCP was assessed by mass photometry using a TwoMP mass photometer (Refeyn, UK). Briefly, 10 μl buffer (50 mM Tris-HCl (pH 8,0), 50 mM NaCl, 5 mM MgCl₂, 1 mM DTT) was loaded to the sample chamber and the objective was focused by autofocus function. Then 0.5 μl sample containing 60 nM gap-NCP alone or together with 180 nM or 200 nM Polβ was added to the chamber and mixed by pipetting. The buffer was preliminary ultrafiltered by using Sartorius vivaspin columns with 1000 kDa pore PES membrane filter. The reaction mixture was prepared on ice and incubated for 5 min at room temperature. Data were collected for 1 min using the AcquireMP software (Refeyn, UK). All samples were measured using the expanded detection area. The mass photometry signals were calibrated using BSA (69 kDa) (Sigma-Aldrich, USA), recombinant PARP1 protein (113 kDa) and recombinant Cas9 protein (158 kDa). Mass photometry data were analyzed using the DiscoverMP software (Refeyn, UK) to calculate relative molecular populations from the areas of the Gaussian peaks (representing free NCP and its complexes with one or several Polβ molecules).

Testing of Polβ-catalyzed DNA synthesis

DNA synthesis catalyzed by Polβ was performed in 5 µL reaction mixtures containing 50 nM gap-DNA (227 bp) or gap-NCP, 20 nM Polβ (unless stated otherwise), and either 100 µM of each of four dNTPs (for strand-displacement synthesis) or 500 nM dTTP (for gap-filling synthesis) in reaction buffer (50 mM Tris-HCl pH 8.0, 50 mM NaCl, 5 mM MgCl₂, 1 mM DTT). Polβ concentration was varied from 3 nM to 75 nM depending on the experiment. To analyze the influence of FEN1, H1, PARP1 or PARP2 on the Polβ DNA polymerase activity mixtures were supplemented with 25 nM FEN1, 170 nM H1, 50–200 nM PAPR1 or PARP2 respectively. To analyze the influence of H1 and PARylation on Polβ activity, mixtures (the final volume was 10 μl) of 50 nM gap-DNA or gap-NCP, 100 nM PAPR1 or PARP2, 75 nM H1, and 10 μM NAD⁺ in reaction buffer (50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 5 mM MgCl₂, 1 mM DTT) were incubated at 37°C for 30 min for the PARylation reaction. After this, Polβ and dNTPs were added to the mixtures to final concentrations of 20 nM Polβ and 100 μM of each dNTP. After adding of Polβ and dNTPs, the mixtures were incubated at 37 °C for 10 min.

All reactions were terminated by adding 10 µL of 8 M urea with 20 mM EDTA, followed by heating at 80 °C for 10 min and 97 °C for 1 min. Reaction products were separated by electrophoresis on 20% denaturing polyacrylamide gels. FAM-labeled oligonucleotides of 26 nt and 50 nt length (Table 2) were used as gel-mobility controls corresponding to the length of substrate and linker. Gels were visualized using a Typhoon FLA 9500 scanner (GE Healthcare Life Sciences, USA). The relative intensities of bands corresponding to the substrate and products were quantified using Quantity One 4.6.6 software (Bio-Rad, USA).

Kinetic measurements of Polβ activity

The reaction of DNA synthesis via strand displacement catalyzed by Polβ was carried out in a 70 μl reaction mixture containing 50 mM Tris-HCl, pH 8.0, 50 mM NaCl, 5 mM MgCl₂, 50 nM gap-DNA (227 bp) or gap-NCP, 20 nM Polβ and 100 µM each of four dNTPs. Reactions were initiated by addition of dNTPs and stopped after 30 s, 2 min, 5 min and 10 min of incubation by adding 5 µl aliquot of the reaction mixture to 5 µl of 8 M urea with 20 mM EDTA. For kinetic analysis of gap-filling synthesis, the reaction mixture contained 50 mM Tris-HCl, pH 8.0, 50 mM NaCl, 5 mM MgCl₂, 200 nM gap-DNA/gap-NCP, 4 nM Polβ and 500 nM dTTP. Reactions were initiated by addition of dTTP and stopped after 20 s, 1 min, 3 min and 5 min of incubation. Reaction products were separated and analyzed as described in subsection above.

Protein PARylation reaction

Reaction mixtures containing 125 nM gap-NCP, 250 nM PARP1 or PARP2, 500 nM HPF1, 150 nM H1, and 1 µM [³²P]NAD⁺ in buffer (50 mM Tris-HCl pH 8.0, 50 mM NaCl, 5 mM MgCl₂, 1 mM DTT) were incubated for 30 min at 37 °C. Reactions were stopped by adding 1.2 µL of inhibitor solution (50 µM olaparib, 100 mM EDTA, 25 mM Tris-HCl pH 8.0, 25 mM NaCl, 500 µM DTT). When indicated, 1 µM PARG was added to the mixtures, followed by incubation for 1 h at 37 °C to digest PAR. Subsequently, 1.5 µL of SDS-PAGE sample buffer was added, and samples were heated at 97 °C for 5 min. Reaction products were separated by 20% SDS-PAGE, visualized by phosphorimaging, and quantified using a Typhoon imaging system (GE Healthcare Life Sciences) and Quantity One Basic software (Bio-Rad).

Data availability

Figure 6 contains the numerical data used to generate the figure.

Acknowledgements

We would like to thank the entire laboratory of bioorganic chemistry of enzymes for feedback. We acknowledge Svetlana N. Khodyreva for preparation of recombinant FEN1, Anton V. Endutkin and Dmitry O. Zharkov for supporting Mass Photometry experiments. The reported study was funded by the Russian state-funded project for ICBFM SB RAS № 125012300658-9 (use of shared equipment for experimental work, MP experiments) and by the Russian Science Foundation № 25-74-10025 (preparation of proteins and nucleosomes) and 25-74-30006 (biochemical experiments).

Additional information

Author contributions

D.M.S. performed the experiments, T.A.K. performed the experiments and created the figures. M.M.K., K.N.N., A.A.U. and N.A.M. contributed to the study with protein purification and nucleosome assembling. T.A.K., N.A.M. and O.I.L. analyzed the data and wrote the manuscript. All authors reviewed the results and approved the final version of the manuscript.

Funding

Russian Academy of Sciences (RAS) (125012300658-9)

• Konstantin N Naumenko

• Danil M Shtanov

• Tatyana A Kurgina

• Alexander A Ukraintsev

• Mikhail Kutuzov

• Nina A Moor

• Olga I Lavrik

Russian Science Foundation (RSF) (25-74-10025)

• Alexander A Ukraintsev

• Mikhail Kutuzov

• Konstantin N Naumenko

• Danil M Shtanov

• Nina A Moor

• Tatyana A Kurgina

• Olga I Lavrik

Russian Science Foundation (RSF) (25-74-30006)

• Alexander A Ukraintsev

• Tatyana A Kurgina

• Nina A Moor

• Konstantin N Naumenko

• Danil M Shtanov

• Mikhail Kutuzov

• Olga I Lavrik

Additional files

Supplementary material