The zinc finger of DNA ligase 3α binds to nucleosomes via an arginine anchor

Nagpal, Ashna; Schaich, Matthew A.; Weaver, Tyler M.; Roginskaya, Vera; Sallmyr, Annahita; Leary, Liam P.; Freudenthal, Bret D.; Tomkinson, Alan E.; Van Houten, Bennett

doi:10.1038/s41467-025-66320-8

Download PDF

Article
Open access
Published: 21 November 2025

The zinc finger of DNA ligase 3α binds to nucleosomes via an arginine anchor

Nature Communications volume 16, Article number: 11531 (2025) Cite this article

3721 Accesses
1 Altmetric
Metrics details

Subjects

Abstract

Ligation of DNA single strand breaks is critical for maintaining genome integrity during DNA replication and repair. DNA Ligase III (LIG3α) forms an important complex with X-ray cross complementing protein 1 (XRCC1) during single strand break and base excision repair. We utilize a real time single molecule approach to quantify DNA binding kinetics of HaloTag-LIG3α and XRCC1-YFP from nuclear extracts on long DNA substrates containing nicks, nucleosomes or nicks embedded in nucleosomes. LIG3α displays higher affinity for nicks than XRCC1 with the LIG3α catalytic core and N-terminal zinc finger (ZnF) competing for nick engagement. Surprisingly, compared to single strand breaks in naked DNA, LIG3α binds even more avidly to an undamaged nucleosome reconstituted on the 601-sequence, with binding dependent on two arginine residues in the N-terminal ZnF. These studies reveal insights into nick detection and identify an arginine anchor mechanism for LIG3α engagement with nucleosomes.

Activation of recombinases at specific DNA loci by zinc-finger domain insertions

Article Open access 31 January 2024

Rescuing the bacterial replisome at a nick requires recombinational repair and helicase reloading

Article Open access 26 November 2025

Controllable DNA hybridization by host–guest complexation-mediated ligand invasion

Article Open access 08 October 2022

Introduction

DNA ligases restore breaks in DNA phosphodiester bonds to maintain genome integrity. These breaks can be generated during DNA replication and repair as well as breaks resulting directly by DNA damaging agents. Human cells contain three forms of mammalian DNA ligases, I, III, and IV. Each of these enzymes have similar catalytic regions, including a DNA binding domain (DBD) and a catalytic core (CC) composed of nucleotidyl transferase (NTase) and oligonucleotide/oligosaccharide-fold binding domains (OBD)^{1,2,3,4,5,6,7,8}. While deletion of each of the LIG genes results in embryonic lethality in mice, it was not possible to establish cultures of embryonic fibroblasts from the LIG3 null embryos in contrast to LIG1 and LIG4 null embryos^9,10,11. Subsequently, it has been shown that the LIG3 gene is essential for cell viability under normal growh conditions because it encodes the only mitochondrial DNA ligase in addition to a nuclear version by alternative translation initiation^12,13,14.

The nuclear and mitochondrial versions of LIG3α both lack a nuclear localization signal (NLS). Nuclear LIG3α gains entry into the nucleus by forming a complex with the DNA repair scaffold protein, XRCC1 which has an NLS^{5,15,16,17,18,19} whereas a mitochondrial targeting sequence within the extra N-terminal region of mitochondrial LIG3α directs this protein to mitochondria where it functions to maintain the mitochondrial genome^13,20,21. Notably, disruption of the interaction between the C-terminal BRCT domains of LIG3α and XRCC1 markedly reduces the cellular LIG3α levels. This presumably reflects the role XRCC1 in stabilizing nuclear LIG3α, whereas LIG3α function in mitochondria is independent of XRCC1^13,20,21,22. Within the nucleus, the best characterized functions of LIG3α and XRCC1 are in the base excision repair (BER) and DNA single-stranded break (SSB) repair pathways^5,23. It is likely that these pathways play a critical role in ensuring the viability of differentiated cells since abortive ligation, defects in an interaction between LIG3α and fused in sarcoma (FUS), as well as inherited mutations in XRCC1 have been linked with neurodegeneration^24,25,26. LIG3α and XRCC1 also act in a backup repair mechanism when either of the major DNA double-strand break repair pathways are defective^{5,27,28,29,30}, and during the semi-conservative DNA replication via an alternative Okazaki-fragment ligation pathway^31,32,33,34.

In single strand break repair, the broken sugar-phosphate backbone is initially recognized by PARP1, which is activated to synthesize chains of poly(ADP-ribose) (PAR) that can be covalently linked to proteins, including PARP1 itself^35,36. PAR recruits XRCC1 and associated repair proteins, including LIG3α to the damage site^37,38,39. Interestingly, the central BRCT of XRCC1 has separate binding sites for PAR and DNA, both of which are involved in the recruitment and retention of XRCC1 at break sites⁴⁰. In BER, the combined action of DNA glycosylases and AP endonuclease generate a single strand break that is converted into a ligatable nick by Polβ through its lyase and polymerase activities. In the case of bi-functional DNA glycosylases that perform β-elimination, APE1 processes the conversion of 3’-phospho-α, β-unsaturated aldehyde (3’-PUA) into 3’-hydroxyl (3’-OH) and in the case of β, δ-elimination PNKP converts the 3’-P into 3’-OH. It has been suggested that repair intermediates are sequentially passed to BER enzymes with the XRCC1 scaffolding protein facilitating these hand-offs through interactions with the enzymes including LIG3α⁴¹. While PARP1 has been implicated in BER, its precise role in processing intermediates is not entirely clear. However, PARP1 clearly contributes to BER when the flux through the initial stages of the pathway exceeds the DNA synthesis and ligation steps^22,42.

LIG3α differs from the other ligases as it contains an N-terminal zinc finger (ZnF) motif that is highly related to the first two zinc finger motifs at the N-terminus of PARP1⁴³. The LIG3α ZnF appears to function as a single strand break sensor both in the presence and absence of activated PARP1⁴⁴. Biochemical and biophysical studies revealed that the LIG3α ZnF together with the DBD form a single strand break binding module that is insensitive to the break structure whereas the Ntase and OBD form a second single strand break binding module that preferentially binds to ligatable nicks⁴². This led to the proposal of a “jackknife” mechanism^6,45, in which the ZnF-DBD first senses the nick but is then displaced at a ligatable nick by the catalytic core (CC).

In contrast to biochemical and biophysical studies characterizing the interaction of LIG3α with naked DNA, there have been relatively few studies examining how LIG3α engages with intact and nicked DNA packaged into nucleosomes^46,47,48. Interestingly, it has been shown that LIG3α binds to condensed chromatin during metaphase⁴⁹, and that both LIG3α and XRCC1 associate with undamaged nucleosome core particle (NCPs)⁴. While it has been suggested that LIG3α/XRCC1 disrupts the nucleosome in order to ligate DNA nicks⁴⁷, more recent studies indicate that transient unwrapping of DNA, particularly in the regions where DNA enters and exits the nucleosome unwrapping, enables LIG3α/XRCC1 to join DNA nicks^46,50. Here we use single molecule kinetics to examine the binding of tagged versions of XRCC1 and LIG3α in nuclear extracts with intact and nicked DNA in the absence or presence of NCP. These experiments characterize the on and off rates of LIG3α and XRCC1, their order of assembly and disassembly, and how these proteins identify DNA nicks in dsDNA and chromatin. Importantly, using this single-molecule approach, we discover that the LIG3α ZnF domain utilizes two arginine anchor residues (R62/R64) for binding to undamaged NCP’s.

Results

Binding kinetics of LIG3α-XRCC1 on a λ DNA substrate containing eight observable 3’-OH- 5’-phosphate nicks using SMADNE

We recently developed and validated the SMADNE approach (Single-Molecule Analysis of DNA binding Proteins from Nuclear Extracts)⁵¹, which utilizes fluorescently tagged DNA-binding proteins expressed in nuclear extracts to study real time DNA binding dynamics at the single molecule level using a LUMICKS C-Trap. In this experimental setup, we designed a structural model of heterodimeric complex of N-terminal tagged HaloTag-635 LIG3α and C-terminal YFP tagged XRCC1 bound to nicked DNA (PDB codes 3L2P and 3K77) using Alphafold 3 (Fig. 1a) and a domain structure of LIG3α (nuclear) bound to C-terminal BRCT2 domain of XRCC1 via BRCT domain, forming a heterodimeric complex (Fig. 1b). The optical tweezers in the LUMICKS C-Trap were used to trap streptavidin coated polystyrene beads. By flowing in biotinylated λ DNA substrate containing eight observable ligatable nicks, one strand of DNA is then strung up between the beads (Fig. 1c, left panel, see Methods). Nuclear extracts expressing fluorescently tagged LIG3α with HaloTag conjugated to Janelia-Fluor (JF) 635 dye on the N-terminus and XRCC1 fused to YFP at its C-terminus (Fig. 1a, b, Supplementary Figs. 1, 2. see “Methods”) were diluted 1:10 in binding buffer (20 mM HEPES pH 7.5, 150 mM NaCl, 1 mM DTT, 5% glycerol, 0.5 mg/ml BSA,1 mM trolox) and then flowed into the LUMICKS C-Trap flow cell containing the tethered λ DNA. Flow was stopped, and a 2D confocal image was generated to monitor binding of these proteins to the DNA (Fig. 1c, middle panel). We then collected kymographs, where the Y-axis represents the position on the DNA and the X-axis represents time, in 1-dimensional scanning (kymograph) mode at 10 frames per second using the nicked λ DNA (Fig. 1c, right panel). These kymographs revealed HaloTag-635-LIG3α binding events as red fluorescent time streaks, XRCC1-YFP binding events as blue fluorescent time streaks, and the simultaneous binding of LIG3α and XRCC1 as pink fluorescent time streaks. Using the SMADNE approach, we calculated the binding lifetimes (Tau) and k_off using CRTD (cumulative residence time distribution) analysis (Fig. 1d.i and e.i) of all the binding events. Binding event durations were best fit to a one phase exponential decay, yielding lifetimes for LIG3α, Tau = 5.1 s ( ± 0.02), and XRCC1, Tau = 6.3 s ( ± 0.03) (Fig. 1d.i for LIG3α & e.i for XRCC1). By measuring the gap times (the spaces between events at the same DNA position) and fitting to a CGTD (cumulative gap time distribution) curve, a k_{on app} for LIG3α and XRCC1 could be determined (Fig. 1d.ii for LIG3α & Fig. 1e.ii for XRCC1). Using the k_off, k_{on app}, and the protein concentration in the flow-cell (see “Methods” and Supplementary Fig. 1C), we also determined equilibrium dissociation constant (K_D) of LIG3α and XRCC1 for the nicked DNA substrate. This analysis revealed that LIG3α binds the nicked DNA substrate with a K_D = 0.12 ± 0.003 nM, whereas XRCC1 binds to the nicked DNA substrate with a K_D = 1.5 ± 0.04 nM (see Table 1). Together, our data indicates that both LIG3α and XRCC1 bind the nicked DNA substrate with high affinity, although XRCC1 binds with a ~ 15-fold weaker affinity than LIG3α (see Table 1).

**Fig. 1: SMADNE analysis of HaloTag635-LIG3α-XRCC1-YFP on eight observable nicks in λ DNA.**

Table 1 Binding affinities of HaloTag-635-LIG3α-XRCC1-YFP with nicked λ DNA substrates

Full size table

Surprisingly, while LIG3α and XRCC1 sometimes co-localize (17.1%), most of the events were single binding events of LIG3α (56.8%) or XRCC1 (26.1%) as shown in (Supplementary Fig. 1a). Thus, although it has been assumed that nuclear LIG3α functions as an obligate heterodimer with XRCC1, our studies indicate that, at the concentrations used in these experiments (0.1–2 nM) (see calibration curves in Supplementary Fig. 2c, d), a significant fraction of these proteins are acting independently. In addition to the two single binding events, there are nine possible overlapping binding events with unique orders of assembly and disassembly (Fig. 1f). Using a co-localization script (see “Material & Methods”) to classify these 11 different types of binding events, we found that the most common co-localization type was category 7, in which both LIG3α and XRCC1 arrive and dissociate together. In addition, binding events in which LIG3α and XRCC1 arrive together (categories 6-8) predominate, followed by binding events in which LIG3α arrives first (categories 8-11). Co-localization categories 3–5, in which XRCC1 arrives first, were very rare. Thus, in accord with their binding affinities, it is evident that LIG3α is driving the overlapping binding events, binding first or simultaneously with XRCC1 at the nick (Fig. 1g, see Supplementary Movie 1 for working model). However, when the two proteins co-localize, their average lifetime of both proteins together on the DNA was only 0.7 s, suggesting the BRCT interaction between the two proteins is weaker than the affinity of LIG3α for a nick (Supplementary Fig. 3).

Visualizing nick sealing over time

LIG3α seals a 3’-OH and a 5’-phosphate via a three-step catalytic process that harnesses the energy of ATP. During this process LIG3α catalyzes the transfer of AMP from ATP to the active site lysine 421 that forms an adenylated enzyme intermediate. This intermediate then transfers AMP group to the 5’-phosphate at the break, and 3’-OH attacks the adenylated 5’-phosphate to generate a new phosphodiester bond and release AMP thereby sealing the nick². While prior bulk biochemical studies have reported the binding kinetics of LIG3α in presence of ATP and Mg²⁺ ⁴⁸, we found that, under our assay conditions, the addition of Mg²⁺ resulted in much lower number of binding events which lasted shorter (Supplementary Fig. 4). In order to determine how much LIG3α is charged with AMP in the nuclear extracts, we investigated the role of ATP and LIG3α adenylation in two ways. First, we followed the number of LIG3α binding positions, which started with 7–8 binding sites but gradually decreased over time as nicks were ligated. In this way we measured the ligation rate based on lifetime of the last binding event prior to the loss of subsequent binding at that same site. (Fig. 2 ai, aii, b, c). Eventually over a period of 10-min of data collection at 10 pN tension we observed much less frequent nick binding over time compared to initial LIG3α binding rates (Fig. 2 ai, aii). By taking the meantime of all the last binding events, in which no further binding to those sites were observed, we could calculate an actual ligation rate that is a combination of k_{AMP transfer}, k_seal, and the rate of product release, which was 0.37 ± 0.05 s^–1 (Fig. 2c). This rate was very close to a previously reported rate constant for adenylyl transfer, k_transfer, (0.89 s^–1), which appears to be the rate-limiting step for ligation⁴⁸. In the second approach, using bulk biochemical analysis nuclear extracts incubated with nicked DNA duplex in the absence or presence of ATP and Mg²⁺ in our working buffer (see “Material and Methods”). Under these conditions, the majority of the ligation events (>95%) were carried out by the tagged LIG3α (Fig. 2 di, di, dii). As expected, there was more ligation under multiple turnover conditions with ATP and Mg²⁺ (Fig. 2diii). By measuring ligation in the absence of ATP and Mg²⁺, we estimate approximately ~33% of the LIG3α molecules were charged with AMP during extract preparation (Fig. 2diii, e).

**Fig. 2: Visualizing nick sealing in real time at 10 pN tension and analysis of LIG3α variants showing ligation.**

5’-OH moieties and catalytically dead HaloTag-635-LIG3α (K421A) alter nick binding kinetics

Pre-adenylated LIG3α transfers the AMP moiety covalently linked to K421 in the NTase domain⁴² to the 5’-P terminus of the nick generating a 5’-5’ adenylated DNA intermediate (Ap-DNA) complex⁵². To test the importance of 5’P in nick binding, we pre-incubated the nicked λ DNA with calf-intestinal alkaline phosphatase (CIP) which removes 5’ phosphate groups leaving non-ligatable 5’-OH nick termini (see Methods). Using our single-molecule approach we observed the binding events for HaloTag-635-LIG3α-XRCC1-YFP binding to CIP-treated nicked λ DNA generating 8 observable non-ligatable nick sites at 10 pN tension as shown in the corresponding kymograph (Fig. 3a). WT LIG3α had ~16-fold reduced binding affinity (K_D = 1.6 ± 0.3 nM, Fig. 3b. ii) for the CIP-treated nicked λ DNA (see Table 1). In addition, the binding of XRCC1 to the CIP-treated nicked λ DNA was also reduced, exhibiting an ~11-fold weaker affinity (K_D = 17 ± 0.9 nM) (Fig. 3c. ii, and Table 1). In accord with their reduced binding affinities and shorter binding lifetimes there was ~1.7 fold reduced co-localization complexes (9.7%), when compared to the WT (17.1%) and overall fewer co-localization complexes were formed (Supplementary Fig. 1a, b, Fig. 3d, e). This clearly indicates that if 5’-P is substituted with a 5’ -OH its blocks its 3-step catalytic transfer of the AMP from the active lysine, making it an abortive ligation intermediate to which LIG3α along with XRCC1 has poor binding kinetics. These studies are consistent with bulk biochemical studies that have previously shown the LIG3α had a twofold reduced binding affinity by removal of the 5’-phosphate^6,45.

To further probe the function of adenylated K421 in nick binding, we replaced K421 with an alanine residue (Fig. 3f) and measured the binding kinetics to ligatable nicks using SMADNE following coexpression with XRCC1-YFP (Fig. 3g). The binding lifetimes for LIG3α K421A-XRCC1 were threefold shorter when compared to similar studies with WT-HaloTag-635-LIG3α (Fig. 3h, i, see Table 1). The durations were then fit to a one phase exponential CRTD fit, yielding an average lifetime for HaloTag-635-K421A-LIG3α of 1.7 s ( ± 0.02), and 1.4 s ( ± 0.02), for XRCC1 (Fig. 3h, i). When compared to wild type LIG3α, the catalytically dead K421A mutant exhibits fourfold lower binding (K_D = 0.40 ± 0.01 nM), Table 1. The shorter lifetime binding of the K421A mutant resulted in a twofold reduction in co-localization with XRCC1 i.e., 7.6% when compared to the WT (Supplementary fig. 1c). These data indicate that lysine 421 and/or its adenylation contributes to nick binding.

The catalytic core (CC) of LIG3α competes with the ZnF domain for nick engagement

While previous bulk biochemical studies with purified proteins have shown that the LIG3α N-terminal zinc finger domain plays an important role in nick binding and ligation under physiological salt concentrations and intermolecular duplex ligation, it is not required for catalysis but instead it appears to function as an initial nick sensor^45,49,53,54. These studies identified that deletion of the ZnF in LIG3α did not significantly change the DNA binding affinity of LIG3α to an intact duplex⁴⁵. To further understand the role of the LIG3α N-terminal zinc finger ZnF in nick recognition, we expressed versions of LIG3α that either lack the ZnF domain (ΔZnF) (see Fig. 4a, b) or lack the entire catalytic core (LIG3α ZnF-BRCT) (see Fig. 4f, g) and observed their bindings to ligatable nicks. LIG3α ΔZnF binds to ligatable nicked DNA almost as well as the full-length WT LIG3α construct (K_D = 0.93 ± 0.007 nM versus K_D = 0.12 ± 0.003 nM; Fig. 4c & Fig. 1dii, Table 1). In contrast the ZnF-BRCT variant binds with ~32-fold weaker affinity than WT LIG3α (K_D = 3.2 ± 0.2 nM) (Fig. 4 h.i, Fig. 1dii). In the presence of ZnF-BRCT variant, XRCC1 also shows a similar drop in affinity (K_D = 4.0 ± 0.8 nM) (Fig. 4i). The lower on rates (k_on) for ZnF-BRCT (2.4 ± 0.12 × 10⁷ s^–1 M^-1) when compared to WT (1.7 ± 0.04 × 10⁹ s^–1 M^-1), as well as the rapid on rate of ΔZnF (8.6 ± 0.6 × 10⁸ s^–1 M^–1) suggests that nick sensing occurs through a concerted effort of both the ZnF and the catalytic core (CC). Interestingly, there is a twofold decrease in colocalization of XRCC1 with the ZnF-BRCT variant when compared to the ΔZnF construct, which may indicate that XRCC1 interacts with the CC in addition to the BRCT domain (7.5% versus 16.3%) (Fig. 4e, j, Supplementary Fig. 1d, e). Together these data suggests that the N-terminal zinc finger domain of LIG3α in the ΔZnF & ZnF-BRCT variant has a ~ 2.5-fold longer lifetimes when compared to the WT LIG3α. These data suggest that there is a coordinated rapid handoff between the ZnF to the CC during nick sensing by WT LIG3α whereas, in the absence of either ZnF or the CC, nick sensing is increased.

Robust engagement of LIG3α with undamaged 601 NCPs at single molecule level

In the nucleus, DNA repair occurs in the context of chromatin, in particular the nucleosomal core particle⁵⁵. Previous biochemical studies have examined the association of repair proteins with undamaged NCPs and the activity of BER enzymes on DNA damage within a NCP^4,47,56 A drawback with these studies is that the DNA fragments containing NCP have open free DNA ends, which are also bound by LIG3α, as well as other repair proteins and could interfere with LIG3α binding kinetics to nicks^4,57. To characterize the interactions of LIG3α and XRCC1 with nucleosomes having no free DNA ends, Cy3-labeled human histone nucleosomes were reconstituted onto a 191 bp DNA duplexes containing a central Widom 601 sequence⁵⁸. The reconstituted nucleosomes were then ligated into 6 kb biotinylated DNA handles (DNA tethering kit from LUMICKS), and the ligated substrate was tethered between optically trapped streptavidin beads on the C-Trap (Fig. 5a). Using this approach, we quantified the binding kinetics of LIG3α and XRCC1 to nucleosomes in a substrate that more closely mimics native chromatin. Since DNA tension greater than 5 pN results in unwrapping the distal arm of nucleosomes reconstituted on the 601 sequence, all the experiments with the NCP substrate were performed at 4–5 pN tension to ensure full DNA wrapping⁵⁹.

**Fig. 5: Robust engagement of LIG3α with an undamaged 601 NCP at the single molecule level.**

Surprisingly, single-molecule imaging revealed that WT-LIG3α avidly binds to an undamaged NCPs with a lifetime of 24 ± 0.6 s (Fig. 5b) and has a ~ 3-fold tighter binding affinity (K_D = 0.04 ± 0.002 nM) compared to a ligatable nick (K_D = 0.12 ± 0.03 nM). However, compared to LIG3α, XRCC1 binds 20-fold less tightly to an undamaged NCP as compared to LIG3α (K_D = 0.60 ± 0.05 nM; Fig. 5c, Table 2). Order of assembly analysis identified that LIG3α not only avidly binds to the undamaged NCP but also often binds independent of XRCC1 (Supplementary Fig. 5a). When LIG3α and XRCC1 binding to an NCP overlap, LIG3α almost always binds first prior to XRCC1 engagement (Fig. 5d, see Supplementary Movie 2). To determine the region of LIG3α that mediates NCP binding, we performed single molecule imaging with LIG3α variants either lacking the ZnF or the CC (Fig. 5e, f). This domain analysis found that the ZnF-BRCT avidly binds to an undamaged NCP with high affinity (K_D = 1.1 ± 0.3 nM). This is threefold tighter than this variant’s binding affinity to a ligatable nick in naked DNA. Notably, deletion of the ZnF binding domain (ΔZnF-LIG3α variant) of LIG3α abrogated the interaction of LIG3α with an undamaged NCP (Supplementary Fig. 6). After identifying the ZnF domain as critical for the NCP interaction, we hypothesized that the R62 & R64 residues within the ZnF domain of LIG3α may act as arginine anchors that bind the acidic patch of the nucleosome (Supplementary Fig. 6)⁶⁰. Indeed, substitution of Arg62 and Arg64 to alanine caused LIG3α to no longer bind to the undamaged NCP (Supplementary Fig. 6). Since the putative Arg anchors (R62 and R64) are located in the ZnF domain of LIG3α, we presume that the anchor binds to the acidic patch of the NCP.

Table 2 Binding affinities of HaloTag-635-LIG3α-XRCC1-YFP with 601-NCP substrates

Full size table

Non-ligatable nick position in the NCP alters LIG3α and XRCC1 binding kinetics

Chromatin is subjected to various environmental and enzymological factors that can generate DNA SSBs^57,61. Having found that LIG3α binds with high affinity to undamaged NCPs, we examined the binding of LIG3α and XRCC1 to nicks within nucleosomes. Using our 601 NCP substrate, we introduced a non-ligatable nick containing a 3’-OH and 5’-OH at three different positions within the NCP (one super helical location, SHL, at a time: 0, –2.5, or –4.5) and performed single molecule imaging (Fig. 6a). Non-ligatable nicks within the NCP were necessary to avoid sealing of the nicks during long substrate preparation. LIG3α binding lifetimes for SHL -4.5 were longer (Tau = 17 ± 3 s) compared when compared to SHL -2.5 (Tau = 12 ± 4 s) and SHL 0 (Tau = 10 s ± 2) but were shorter than binding to the undamaged NCP (Tau = 24 ± 0.6 s) (Fig. 6b, c, d). Also, examination of the frequency of co-localization for LIG3α and XRCC1 indicated the highest colocalization of the two proteins occurred when the nicks was a SHL -4.5, followed by SHL -2.5, and finally SHL 0 (Supplementary Fig. 5 c). The CRTD plots fitted a two-phase exponential curve composed of some long binding events (Tau slow), and shorter events (Tau fast) (see Table 2). We used a two-phase exponential fit assuming that the long binding events (Tau slow), are the ZnF engagements with the nucleosome through the arginine anchor, whereas shorter events (Tau fast) are from the LIG3α ZnF domain disengaging from the nucleosome and showing shorter nick binding. These data suggest a process in which that the ZnF domain of LIG3α can switch from relatively stable binding to the NCP (Tau slow) to rapidly sensing the nicks (Tau fast) embedded within the NCP. While the three SHL nick positions exhibited relatively similar binding kinetics for LIG3α and XRCC1, the highest affinity substrate was SHL –4.5 (see Table 2). We also examined LIG3α binding kinetics to a single non-ligatable nick in the absence of an NCP, and LIG3α exhibited a decreased binding affinity (K_D = 2.5 ± 0.1 nM), similar to the CIP-treated nicked λ DNA substrate (K_D = 1.6 ± 0.3 nM) (Supplementary Fig. 7). These data indicate that the nick sensing role of ZnF causes release from binding to the nucleosome, causing a decrease in the overall dwell time of LIG3α on the nick containing nucleosomes, with the nick at SHL 0 being the strongest competitor for nucleosome binding as it is closest to the acidic patch (Fig. 7).

**Fig. 6: Positioning of non-ligatable nicks within the 601NCP alters LIG3α and XRCC1 binding kinetics.**

**Fig. 7: Detection of damage and chromatin by LIG3α.**

Discussion

DNA nick sealing is an essential step in the SSB and BER pathways orchestrated by XRCC1 and LIG3α^42,62,63. This single molecule study provides mechanistic insights into the kinetics of LIG3α and XRCC1 interactions with substrates containing non-ligatable and ligatable SSBs in naked DNA, nucleosomes without DNA damage and nucleosomes with non-ligatable nicks. Since the DNA substrates were embedded in long DNA tethers suspended between two beads, this single molecule approach overcomes limitations of other biochemical and biophysical approaches in which these nicked DNA substrates have free DNA ends. We discovered that LIG3α has a 12.5-fold higher affinity for ligatable nicks (K_D = 0.12 nM) than XRCC1 (K_D = 1.5 nM), and that the two proteins are not obligate heterodimers, with co-localizing events at nicks constituting only 17.1% of the total events. Since the steady state levels of LIG3α are reduced in cells expressing mutant versions XRCC1 that are defective in complex formation with LIG3α, it has been assumed that XRCC1 is required for the stability of LIG3α and that these proteins functions as a heterodimer in nuclear DNA repair. While the low protein concentrations of the nuclear extract used for the single molecule studies may result in the dissociation of the LIG3α-XRCC1 complex, it is possible that the function of XRCC1 is to interact with LIG3α in the cytoplasm and direct its transport into the nucleus where it is protected from proteolysis and can function independently from XRCC1. Notably, XRCC1 is not present in mitochondria and the levels and function of mitochondrial LIG3α, which has an N-terminal mitochondrial targeting sequence generated by alternative translation initiation, are not impacted in xrcc1 cells with reduced total cellular levels of LIG3α²⁰.

Previous studies using bulk fluorescence polarization assays reported a K_D for LIG3α binding to nicks of 300 nM, three orders of magnitude less tight than what we determined⁴⁵. One confounder in those studies was that the nick was in a short duplex and, since LIG3α has affinity for the ends of duplex DNA, the authors were forced to work at 250 mM NaCl to reveal specific binding to nicks. By embedding nicks in a long DNA substrate and measuring on and off rates we have overcome this previous limitation. Recent bulk biochemistry SPR experiment indicate XRCC1 has affinity (54 nM) for nicks which is considerably weaker affinity than what we measured in our studies (K_D = 0.6 nM)⁶⁴. Our real-time tracking of order assembly (Fig. 1) indicates that when the proteins co-localize at nicks either LIG3α and XRCC1 bind together at the site of the nick or LIG3α binds first recruiting XRCC1 through the BRCT domain interaction. We observed ligation in real time at the single molecule level, with the rate of apparent rate of nick ligation for LIG3α was 0.37 ± 0.05 s^–1 (Fig. 2). Previous bulk biochemical single turnover experiments indicated that while the rate of sealing is rather fast, 19 s^–1, and the AMP transfer step from LIG3α to the 5’phosphate was rate limiting at 0.89 s^–1 at 37 ^oC⁴⁸. While our data were generated at 25 ^oC and in a different buffer, these two values agree very well. To understand the effect of adenylation in LIG3α, we tested a K421A variant which is catalytically dead due to the inability to be charged with AMP. This variant had a four-fold higher K_D for nicks due to its shorter dwell time. We determined that ~33% of overexpressed LIG3α is fully charged with AMP (Supplementary Fig. 4), thus some of the shorter binding events we are observing with WT protein, could also be due to uncharged LIG3α. We sought to examine the effect of ATP and Mg²⁺ in the extract and reaction buffer, but this was unsuccessful since Mg²⁺ seemed to greatly decrease binding to nicks in our system (Supplementary Fig. 4).

Domain mapping studies provided mechanistic insights into how LIG3α detects and processed ligatable nicked DNA. Removal of the ZnF domain, caused 2.5-fold longer dwell times (see Table 1). The on rates were five-fold more for ΔZnF LIG3α construct (8.6 ± 0.6 ×10⁸ s^–1 M^–1) when compared to full length construct (1.7 ± 0.04 ×10⁹ s^-1 M^-1) and a K_D within error of the full-length protein (see Table 1). Surprisingly the ZnF-BRCT construct of LIG3α also exhibited a twofold longer dwell time compared to full length (see Table 1) but had a 69.2-fold lower on rate (2.4 ± 0.1 × 10⁷ s^–1 M^–1) when compared to the full-length protein (1.7 ± 0.04 × 10⁹ s^–1 M^–1). These data clearly suggest that the ZnF and the CC domain effectively compete for nick binding in the full-length protein in partial support of the jack-knife model first proposed by Ellenberger and coworkers⁴⁵. Rapid shuttling between the ZnF and the CC increases the binding affinity of full length LIG3α for nicks. These data (Fig. 4) are consistent with previous biochemical studies showing that the ZnF domain is required for efficient ligation of SSB^45,65 and also blunt DNA ends⁶. However, since the ZnF-BRCT construct bound 27-fold less tight than the full length, the ZnF of LIG3α provides another important function to LIG3α, as it is the only mammalian DNA ligase containing a ZnF.

This work presented in Fig. 5 unequivocally established that LIG3α binds to undamaged NCPs with high affinity (K_D = 0.04 nM). Subsequent domain analysis indicated that two critical Arg in the ZnF domain are essential for the high affinity binding to a nucleosome. Loss of the ZnF or substituting both Arg62 and Arg64 for Ala in LIG3α completely abrogated binding to nucleosomes, whereas the ligase activity of these variants was higher than WT (Supplementary Fig. 6h, i). It remains unknown why the ZnF motif evolved in some species, but our discovery that the arginine anchor resides within the ZnF and that the RARA sequence in the arginine anchor is highly conserved in all organisms containing a ZnF on LIG3α (see Supplementary Fig. 6) argues that this motif has a dual role (Arg anchor binds to an NCP and the ZnF binds to a nick) in chromatin architecture and DNA repair within nucleosomes compared to other ligases that lack the ZnF domain or arginine anchor¹. In accord with our results, previous mass spectrometry screening experiments of proteins that bound to different regions of nucleosomes found that LIG3α bound to nucleosome but did not identify a precise binding position within the nucleosome, probably due to the confounding binding to the free double-strand ends in this study⁴. Notably, the structurally conserved ZnF1 in PARP1 does not contain these arginine anchors, suggesting that PARP1 potentially binds nucleosomes through an alternate interaction that does not compete with LIG3α. Arginine anchors have been discovered in a number of proteins that bind directly to nucleosomes, but never to our knowledge in a ZnF domain of a protein⁶⁰. We also found that placing a non-ligatable nick in a nucleosome apparently weakens the overall affinity of LIG3α to a nucleosome (Table 2), mostly through a decrease in binding lifetimes. These data suggest that the binding affinity of the ZnF to the acidic patch is altered by nearby nicks, implying that in a biological setting LIG3α may be “pre-positioned” on chromatin for immediate response once nicks are generated. Future studies are aimed at further exploring the role of histone variants in the acidic patch of H2A.Z. It should be noted that modeling of LIG3α CC domain with nicks within the nucleosomes at SHL 0, –2.5, and –4.5 would indicate it cannot engage nicks fully without significant nucleosome unwrapping and the transient binding is probably due only to the ZnF engagement with the nick (Supplementary fig. 8). This is consistent with the finding that nicks in nucleosomes are ligated more slowly⁵⁶.

The relatively high affinity for NCPs without DNA breaks points to roles of LIG3α outside of DNA repair and may be more relevant to its role in chromatin structure and remodeling during transcription and DNA replication. LIG3α binding to histones within a nucleosome may regulate chromatin state and play an important role in the chromatin architecture. Early evidence has suggested that LIG3α has an affinity for condensed chromatin in metaphase chromosomes⁴⁹. Other factors such as post-translational modifications (PTMs) may alter interactions of LIG3α with chromatin. To this end it is interesting to note that LIG3α is phosphorylated on Ser123 during early S phase by CDK2⁶⁶ and that oxidative stress induces loss of Ser123 phosphorylation in an ATM dependent manner probably through activation of a phosphatase. Also, a recent study suggests that LIG3α can be recruited into condensates of PARylated nucleosomes⁶⁷. Thus, histone modifications and PARP1 activity may play an important role in LIG3α binding to chromatin⁶⁷. Future single molecule studies using nucleosome arrays and other protein factors such as PARP1 will provide insights into how LIG3α interacts with chromatin and is shuttled towards sites of damage during DNA replication, repair, and transcription.

These studies have several potential limitations. All experiments in this study were performed with nuclear extracts overexpressed with LIG3α and XRCC1 coexpressed together in U2OS cells because LIG3α lacks a nuclear localization sequence and is dependent upon its association with XRCC1 for nuclear localization. This increases the efficiency of data collection but does not preserve the stoichiometry to other factors. Further, the presence of endogenous LIG3α and XRCC1 along with other dark proteins in the extracts may compete with the overexpressed protein, although these are at a much lower concentration (20–50-fold) when seen on an SDS-gel (Supplementary Fig. 2).

NCPs in this study were reconstituted on the strong Widom 601 sequence to ensure consistency and reproducibility. Future work with NCPs on alternate sequences and substrates with multiple positioned nucleosomes, which may more accurately recapitulate biological nucleosomes, will need to be investigated in the future. For studying the role of LIG3α to seal nicks, we specifically used ligatable nicks to study its binding kinetics on duplex DNA. For nucleosomes, we had to use non-ligatable nicks by necessity, because a ligatable nick would have been sealed by T4 ligase during the ligation of the LUMICKS handles onto the reconstituted nucleosomes. This would affect studying the LIG3α as the binding kinetics were altered with the presence of a 5’-OH on the break as seen in (Supplementary Fig. 7). The localization precision of binding events ( ~ 50-100 bp with JF635 labeled HaloTag) prevents the unambiguous determination of the position of LIG3α-XRCC1 binding events on an NCP, and in the case of a nicked NCP, it cannot be determined if LIG3α is binding the nick or a site on the histones.

Methods

Cell lines

U2OS cells were cultured in 5% oxygen in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 4.5 g/l glucose, 10% fetal bovine serum (Gibco), and 100 µg/mL penicillin/100 µg/mL streptamycin (Life Technologies). To obtain transient overexpression of the fluorescent-tagged proteins of interest, 4 μg of plasmid per 4 million cells was used to transfect using the lipofectamine 2000 reagent and protocol, including 4–6 h of lipofectamine treatment before changing media and letting the plasmids express overnight (Thermo Fisher Cat# L3000008). Cells with overexpressed HaloTag fusions were treated with 100 nM ( ~ 10-100-fold molar excess) of fluorescent HaloTag ligand for 30 min at 37 °C (Janelia Fluor® 635 or 552 HaloTag® Ligand from Dr. Luke Lavis Laboratory, Janelia Research Campus).

Nuclear extraction

Nuclear extraction was performed the day after transient transfection using a nuclear extraction kit from Abcam (ab113474) as in the previously reported single-molecule method⁵¹. Resultant nuclear extracts were aliquoted and flash-frozen in liquid nitrogen prior to storage at –80 C in single-use aliquots. Upon use for single-molecule experiments, nuclear extracts were immediately diluted after thawing in buffer for experiments at typically a ratio of 1:10. Concentrations of labeled proteins were determined by comparing the background signal to standard curves of known concentrations, and the efficiency of the HaloTag labeling reaction as well as presence of free HaloTag dye was monitored vis SDS- PAGE (Supplementary Fig. 4).

Western blots of overexpressed proteins from nuclear extracts

Various amounts of extracts and purified proteins were loaded onto 4-20% tris-glycine polyacrylamide gels (Invitrogen; XP04202BOX) (Supplementary Fig. 1). Transfer was performed using polyvinylidene difluoride membrane followed by blocking in 20% nonfat dry milk (diluted in PBST: phosphate-buffered saline containing 0.1% Tween 20) for 1 h at room temperature. Membranes were incubated with primary antibodies for 2 h at room temperature or overnight at 4 °C, washed 3 × 10 min in PSBT, and incubated with peroxidase conjugated secondary antibodies for 1 h at room temperature. Membranes were washed again before developing using Super Signal West Femto Maximum Sensitivity Substrate (Thermo Fisher Scientific; #34095). Primary antibodies used: LIG3α (1:1000; GTX #GTX103172), Secondary antibodies used: anti-rabbit IgG (1:50,000). Primary antibodies used: XRCC1 (1:1000; abcam #ab134056), Secondary antibodies used: anti-rabbit IgG (1:50,000)

Nucleosome reconstitution

Histones were reconstituted onto the 601-sequence using a previously described salt dialysis protocol^68,69. Briefly, human histone H3 C96S C110A, H2A K119C, H2B, and H4 (Histone Source at Colorado State University) were incubated with either H2A/H2B or H3/H4 in 2 mg/mL guanidinium buffer at room temperature for 2 h, dialyzed against high salt refolding buffer a total of 3 times, and at least 8 h for each exchange at 4 °C. H2A K119C/H2B dimers and H3 C96S C110A/H4 tetramers were purified using a Superdex 200 column. To perform the maleimide labeling, twofold molar excess Cy3- malemidie dye was added H2A K119C/H2B dimer in 0.7 mM TCEP. Cy3-maleimide dye was added to the H2A K119C in a 2:1 molar ratio and incubated at room temperature for 1 h while rocking. Reactions were quenched with 10 mM DTT, the dimer was purified over a Superdex S200 column, and frozen in 50% glycerol. After confirming the stoichiometry with SDS gel, add equal volume of 100% glycerol to store the H2A K119C/H2B dimer and the H3 C96S C110A/H4 tetramer at –20 °C. To reconstitute an undamaged nucleosome core particle on DNA, the following ultramer sequences were ordered from IDT:

Top strand:5′-Phos-CAAC TGA GAC CAT GTA CCC AGT TCG AAT CGG ATG TAT ATA TCT GAC ACG TGC CTG GAG ACT AGG GAG TAA TCC CCT TGG CGG TTA AAA CGC GGG GGA CAG CGC GTA CGT GCG TTT AAG CGG TGC TAG AGC TGT CTA CGA CCA ATT GAG CGG CCT CGG CAC CGG GAT TCT CGA TAA CTC AGC AAT AGT GGG TCT CA- 3′

Bottom strand: 5′- Phos-ACCA TGA GAC CCA CTA TTG CTG AGT TAT CGA GAA TCC CGG TGC CGA GGC CGC TCA ATT GGT CGT AGA CAG CTC TAG CAC CGC TTA AAC GCA CGT ACG CGC TGT CCC CCG CGT TTT AAC CGC CAA GGG GAT TAC TCC CTA GTC TCC AGG CAC GTG TCA GAT ATA TAC ATC CGA TTC GAA CTG GGT ACA TGG TCT CA-3′

To generate NCPs with single-strand breaks, the “Top Strand” oligonucleotide was annealed to sets of two complementary oligonucleotides to generate non-ligatable nicks in defined locations as follows:

For SHL0:

Nick_SHL0_Istrand1: 5′-Phos-AC CAT GAG ACC CAC TAT TGC TGA GTT ATC GAG AAT CCC GGT GCC GAG GCC GCT CAA TTG GTC GTA GAC AGC TCT AGC ACC GCT TAA ACG CAC GTA CGC G-3′

Nick_SHL0_Istrand2: 5′-OH-CTG TCC CCC GCG TTT TAA CCG CCA AGG GGA TTA CTC CCT AGT CTC CAG GCA CGT GTC AGA TAT ATA CAT CCG ATT CGA ACT GGG TAC ATG GTC TCA-3′

For SHL-2.5:

Nick_2.5_Istrand1 (IDT Ultramer):5′-Phos-AC CAT GAG ACC CAC TAT TGC TGA GTT ATC GAG AAT CCC GGT GCC GAG GCC GCT CAA TTG GTC GTA GAC AGC TCT AGC ACC GCT TAA ACG CAC GTA CGC GCT GTC CCC CGC GTT TTA ACC GC-3′

Nick_2.5_Istrand2: 5′-OH-CAA GGG GAT TAC TCC CTA GTC TCC AGG CAC GTG TCA GAT ATA TAC ATC CGA TTC GAA CTG GGT ACA TGG TCT CA-3′

For SHL-4.5:

Nick_4.5_Istrand1 (IDT Ultramer): 5′-Phos-AC CAT GAG ACC CAC TAT TGC TGA GTT ATC GAG AAT CCC GGT GCC GAG GCC GCT CAA TTG GTC GTA GAC AGC TCT AGC ACC GCT TAA ACG CAC GTA CGC GCT GTC CCC CGC GTT TTA ACC GCC AAG GGG ATT ACT CCC TAG TCT-3′

Nick_4.5_Istrand2: 5′-OH-CCA GGC ACG TGT CAG ATA TAT ACA TCC GAT TCG AAC TGG GTA CAT GGT CTC A-3′

The annealed DNA, H2A K119C/H2B dimer, and the H3 C96S C110A/H4 tetramer are mixed in a 1:2:1 molar ratio and equilibrate in dialysis tubing against high salt buffer for 30 min. The high salt was then removed via dialysis, the reconstituted nucleosome concentrated, and heat shocked at 55 °C for 30 min. Then, sub-nucleosomes and free DNA were removed using a 10–40% sucrose gradient for 40 h at 125,000 x g at 4 °C. Fractions containing reconstituted nucleosomes are combined, buffered exchanged into TE buffer, and concentrated to ~1 µM and stored at 4 °C. Native PAGE analysis (Supplementary Fig. 7) revealed that these octamers were stable in these conditions for at least 3 months.

DNA substrate generation

Lambda DNA was biotinylated and isolated for C-Trap experiments as previously described, with aliquots stored at 20 ng/μL at –20 °C⁵¹. After thawing aliquots, they were stored at 4 °C for up to 2 weeks and then discarded. DNA with single-stranded breaks (nicked DNA) was generated by digesting 1 μg of DNA with the nickase Nt.BspQI (NEB) to generate 10 nicks (8 observable), cutting on the 3′ side of its recognition sequence. Of note, this nickase cuts outside of its recognition sequence, so each nick is flanked by unique DNA sequences, and binding events on this DNA represent an average of all the sequences. Two of these substrates are not observable because one is too close to the streptavidin beads, and two others are so close that they cannot be optically differentiated. After nicking, substrates were aliquoted and stored at –80 °C for up to one year⁵¹.

To tether defined substrates like the nucleosomes or substrates with a single ligatable nick, the “DNA tethering kit” from Lumicks was utilized. The protocol was followed as per the manufacturer’s instructions: briefly 50 pmol of the DNA fragment was incubated with the two biotinylated 6 kb handles, ligase, and ligase buffer, and placed in the dark to ligate overnight at room temperature. Following ligation, samples were diluted 1:300, and Cy3 fluorescence was utilized to confirm tethered samples had labeled NCPs before data collection.

Single-molecule experiments

DNA tether formation and positioning

Single-molecule experiments were performed on a LUMICKS C-Trap instrument⁷⁰. For experiments with lambda DNA (~48 kb), channels one, two, and three were filled with 4.0–4.9 μm polystyrene streptavidin beads (LUMICKS), biotinylated DNA, and buffer of interest, respectively, with trapping lasers set to 100%, 30% overall power, and 50% Trap 1 split. For the 12 kb substrates, trapping laser power was reduced to 15% overall power, and 1.5–1.7 μm streptavidin beads were utilized. All three channels were flowed at a pressure of 0.3 bar to maintain laminar flow while catching beads in each trap. Then the traps were moved to channel two and distance varied between the traps while looking for a force response to tether a DNA substrate between the two beads as expected by the extensible Worm-like Chain Model⁷¹. Channel three and four (containing the nuclear extracts of interest) were flowed at 0.3 bar for at least 10 s to introduce nuclear extracts into the flow cell while keeping the DNA substrate in the buffer alone. DNA tension was then defined in the absence of laminar flow (either 4–5 pN for wrapped NCPs or 10 pN for dsDNA or unwrapped DNA) and then kymographs were collected when the DNA moved to the nuclear extract.

Confocal imaging

YFP was excited with a 488 nm laser and emission collected in a 500–550 nm band pass filter, and HaloTag-JF-635 was excited with a 638 nm laser and emission collected in a 650–750 nm band pass filter. All data was collected with a 1.2 NA 60X water emersion objective and fluorescence measured with single-photon avalanche photodiode detectors. Each laser was set to 5% power and scanned at a rate of 10 frames per second. This framerate allowed for a pulsed excitation approach and a ~threefold increase in fluorophore lifetime before photobleaching compared to continuous scanning⁵¹.

Data analysis

Single molecule data was exported with Bluelake and analyzed using custom software from LUMICKS (Pylake). The utility C-Trap.h5 Visualization GUI was used for figure generation (Watters, 2020). KymoWidgetGreedy widget from LUMICKS was utilized for line tracking of single-molecule events, performing a gaussian fit over the fluorescent intensity (Mangeol et al., 2016). After tracking the lines, the position and time data for each line was used to determine each line’s duration and the duration of gaps between lines for on rate calculation. As previously reported, YFP blinking was observed to last up to 2 s⁷², thus events that occurred at the same position less than two seconds apart were connected and considered one binding approach as blinking⁵¹. Python scripts used to calculate photobleaching decay constants and co-localization analysis have been deposited at https://harbor.lumicks.com/scripts.

Photobleaching analysis

Photobleaching decay constants were determined as previously described for each fluorophore by collecting kymographs with continuous exposure on fluorophores nonspecifically adsorbed at the bottom of the slide. Total intensities were binned by one second intervals and fit to an exponential decay function. To examine the impact of photobleaching on the measured off rates, both the raw values and corrected lifetimes for each dataset are shown in Supplementary information (see Table 1). As the correction caused only slightly altered most values, the raw values are reported in the text.

Note: HaloTag-635 used for LIG3α has a 93% labeling efficiency and as high as 25% of the YFP tagged XRCC1 may not have reached full maturation.

Quantification and statistical analysis

For each experiment, the number of observations analyzed has been included in the figure and/or in the figure legends. The types of errors displayed are also indicated in the figure legends and tables. Each dataset represents at least two replicates with two batches of nuclear extracts, and the datasets with nucleosomes combine two batches of reconstitution for each type (undamaged and each nick position).

Biochemical assays

Oligonucleotide substrates

Synthetic DNA oligonucleotides (Integrated DNA Technologies) were designed to generate a Cy3-labeled nicked duplex substrate. The upper left strand contained a 5′ Cy3 fluorophore (5′-Cy3-CGA CGG CCA GTG CAG GGT TTC-3′), the upper right strand was 5′ phosphorylated (5′-Phos-GTA AAG TCA CGA CCG TCA TGC-3′), and the complementary lower strand carried a 5′ biotin modification (5′-biotin-GCA TGA CGG TCG TGA CTT TAC GAA ACC CTG CAC TGG CCG TCG-3′).

Equimolar concentrations of all three oligonucleotides were mixed, heated to 94 °C for 2 min, and annealed by gradual cooling at a rate of 1.5 °C min⁻¹ to 25 °C using a thermocycler. The annealed nicked duplex was stored at −20 °C until use.

Ligation assay

Ligation efficiency was measured using Cy3-labeled nicked duplex DNA and U2OS nuclear extracts. Each 10 µL reaction contained 30 mM Tris-HCl (pH 7.8), 10 mM DTT, 10 mM MgCl₂, 1 mM ATP, and 3 µL of nuclear extract. Reactions were incubated at room temperature for 10 min.

Negative controls were prepared in parallel using buffer lacking MgCl₂ and ATP (30 mM Tris-HCl, pH 7.8, and 10 mM DTT). Following incubation, reactions were mixed with 30 µL of pre-washed streptavidin magnetic beads (Thermo Fisher Scientific) equilibrated in 30 mM Tris-HCl (pH 7.5) and incubated for 10 min at room temperature with gentle agitation. Beads were captured magnetically, and unbound oligonucleotides were removed by three washes with Tris buffer.

Bound oligonucleotides were eluted by adding 20 µL of stop buffer (formamide, 30 mM EDTA, and 1.5% SDS) and heating at 100 °C for 5 min. The denatured samples were separated from the beads by magnetic pull-down, and the supernatant was analyzed by electrophoresis.

Gel electrophoresis and detection

Samples were resolved on a 20% denaturing urea polyacrylamide gel and electrophoresed at 250 V in 1× TBE buffer until adequate separation was achieved. Cy3 fluorescence was visualized using a Typhoon.

Western blotting

For protein detection, the indicated amount of soluble fraction and purified Lig3 + XRCC1 were loaded onto a 7.5% SDS-PAGE gel. The gel was then transferred onto PVDF membranes (Millipore) which ran at 100 volts for 2 h.

Membranes were blocked overnight at 4 °C with 5% bovine serum albumin (BSA) in 1x Tris buffer (10 mM, pH 7.5). The membrane was then incubated with a primary antibody against LIG3 for 1 h at room temperature (GeneTex, Cat# 103172, 1:2000, 1 h, RT). The membrane was washed 3 times for 5 min each and incubated with an HRP-conjugated anti-rabbit secondary antibody for 45 min at room temperature (Bio-Rad, Cat# 1706515, 1:20000, 45 min, RT). Finally, the membrane was washed three times for ten minutes before detection. Blots were developed using ECL reagent (Cytiva) and imaged on a ChemiDoc imaging system (Bio-Rad).

Data availability

All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Large files containing kymograph data (~1–2 gigabytes each) are available from the corresponding author upon request. Source data are provided with this paper.

Code availability

Python scripts used to calculate lifetime and MSD analysis have been deposited at https://github.com/schaichm. Python scripts used to calculate photobleaching decay constants and co-localization analysis have been deposited at https://harbor.lumicks.com/scripts.

References

Tomkinson, A. E., Vijayakumar, S., Pascal, J. M. & Ellenberger, T. DNA ligases: Structure, reaction mechanism, and function. Chem. Rev. 106, 687–699 (2006).
Article PubMed Google Scholar
Ellenberger, T. & Tomkinson, A. E. Eukaryotic DNA ligases: Structural and functional insights. Annu. Rev. Biochem. 77, 313–338 (2008).
Article PubMed PubMed Central Google Scholar
Martin, I. V. & MacNeill, S. A. ATP-dependent DNA ligases. Genome Biol. 3, 3005. https://doi.org/10.1186/gb-2002-3-4-reviews3005 (2002).
Skrajna, A. et al. Comprehensive nucleosome interactome screen establishes fundamental principles of nucleosome binding. Nucleic Acids Res. 48, 9415–9432 (2020).
Article PubMed PubMed Central Google Scholar
Sallmyr, A., Bhandari, S. K., Naila, T. & Tomkinson, A. E. Mammalian DNA ligases; roles in maintaining genome integrity. J. Mol. Biol. 436, 168276 (2023).
Article PubMed PubMed Central Google Scholar
Cotner-Gohara, E. et al. Human DNA ligase III recognizes DNA ends by dynamic switching between two DNA-bound states. Biochemistry 49, 6165–6176 (2010).
Article PubMed Google Scholar
Kaminski, A. M. et al. Structures of DNA-bound human ligase IV catalytic core reveal insights into substrate binding and catalysis. Nat. Commun. 9, 2642 (2018).
Article ADS PubMed PubMed Central Google Scholar
Pascal, J. M., O’Brien, P. J., Tomkinson, A. E. & Ellenberger, T. Human DNA ligase I completely encircles and partially unwinds nicked DNA. Nature 432, 473–478 (2004).
Article ADS PubMed Google Scholar
Cinnella, G. et al. Effects of assisted ventilation on the work of breathing: volume-controlled versus pressure-controlled ventilation. Am. J. Respir. Crit. Care Med. 153, 1025–1033 (1996).
Article PubMed Google Scholar
Bentley, S. D. et al. Complete genome sequence of the model actinomycete Streptomyces coelicolor A3(2). Nature 417, 141–147 (2002).
Article ADS PubMed Google Scholar
Barnes, D. E., Stamp, G., Rosewell, I., Denzel, A. & Lindahl, T. Targeted disruption of the gene encoding DNA ligase IV leads to lethality in embryonic mice. Curr. Biol. 8, 1395–1398 (1998).
Article PubMed Google Scholar
Gao, Y. et al. DNA ligase III is critical for mtDNA integrity but not Xrcc1-mediated nuclear DNA repair. Nature 471, 240–244 (2011).
Article ADS PubMed PubMed Central Google Scholar
Lakshmipathy, U. & Campbell, C. The human DNA ligase III gene encodes nuclear and mitochondrial proteins. Mol. Cell Biol. 19, 3869–3876 (1999).
Article PubMed PubMed Central Google Scholar
Simsek, D. et al. Crucial role for DNA ligase III in mitochondria but not in Xrcc1-dependent repair. Nature 471, 245–248 (2011).
Article ADS PubMed PubMed Central Google Scholar
Dulic, A. et al. BRCT domain interactions in the heterodimeric DNA repair protein XRCC1-DNA ligase III. Biochemistry 40, 5906–5913 (2001).
Article PubMed Google Scholar
Nash, R. A., Caldecott, K. W., Barnes, D. E. & Lindahl, T. XRCC1 protein interacts with one of two distinct forms of DNA ligase III. Biochemistry 36, 5207–5211 (1997).
Article PubMed Google Scholar
Cuneo, M. J., Gabel, S. A., Krahn, J. M., Ricker, M. A. & London, R. E. The structural basis for partitioning of the XRCC1/DNA ligase III-{alpha} BRCT-mediated dimer complexes. Nucleic Acids Res. 39, 7816–7827 (2011).
Article PubMed PubMed Central Google Scholar
Hammel, M. et al. An atypical BRCT-BRCT interaction with the XRCC1 scaffold protein compacts human DNA Ligase IIIalpha within a flexible DNA repair complex. Nucleic Acids Res. 49, 306–321 (2021).
Article PubMed Google Scholar
Kirby, T. W. et al. Nuclear Localization of the DNA Repair Scaffold XRCC1: Uncovering the Functional Role of a Bipartite NLS. Sci. Rep. 5, 13405 (2015).
Article ADS PubMed PubMed Central Google Scholar
Lakshmipathy, U. & Campbell, C. Mitochondrial DNA ligase III function is independent of Xrcc1. Nucleic Acids Res. 28, 3880–3886 (2000).
Article PubMed PubMed Central Google Scholar
Lakshmipathy, U. & Campbell, C. Antisense-mediated decrease in DNA ligase III expression results in reduced mitochondrial DNA integrity. Nucleic Acids Res 29, 668–676 (2001).
Article PubMed PubMed Central Google Scholar
Demin, A. A. et al. XRCC1 prevents toxic PARP1 trapping during DNA base excision repair. Mol. Cell 81, 3018–3030.e3015 (2021).
Article PubMed PubMed Central Google Scholar
Fan, J. & Wilson, D. M. 3rd Protein-protein interactions and posttranslational modifications in mammalian base excision repair. Free Radic. Biol. Med. 38, 1121–1138 (2005).
Article PubMed Google Scholar
Ahel, I. et al. The neurodegenerative disease protein aprataxin resolves abortive DNA ligation intermediates. Nature 443, 713–716 (2006).
Article ADS PubMed Google Scholar
Wang, H. et al. Mutant FUS causes DNA ligation defects to inhibit oxidative damage repair in Amyotrophic Lateral Sclerosis. Nat. Commun. 9, 3683 (2018).
Article ADS PubMed PubMed Central Google Scholar
Hoch, N. C. et al. XRCC1 mutation is associated with PARP1 hyperactivation and cerebellar ataxia. Nature 541, 87–91 (2017).
Article ADS PubMed Google Scholar
Wang, H. C. et al. DNA ligase III as a candidate component of backup pathways of nonhomologous end joining. Cancer Res. 65, 4020–4030 (2005).
Article PubMed Google Scholar
Sallmyr, A. & Tomkinson, A. E. Repair of DNA double-strand breaks by mammalian alternative end-joining pathways. J. Biol. Chem. 293, 10536–10546 (2018).
Article PubMed PubMed Central Google Scholar
Simsek, D. et al. DNA ligase III promotes alternative nonhomologous end-joining during chromosomal translocation formation. PLoS Genet. 7, e1002080 (2011).
Article PubMed PubMed Central Google Scholar
Simsek, D. & Jasin, M. Alternative end-joining is suppressed by the canonical NHEJ component Xrcc4-ligase IV during chromosomal translocation formation. Nat. Struct. Mol. Biol. 17, 410–416 (2010).
Article PubMed PubMed Central Google Scholar
Arakawa, H. & Iliakis, G. Alternative Okazaki fragment ligation pathway by DNA ligase III. Genes (Basel) 6, 385–398 (2015).
Article PubMed Google Scholar
Han, L., Masani, S., Hsieh, C. L. & Yu, K. DNA ligase I is not essential for Mammalian cell viability. Cell Rep. 7, 316–320 (2014).
Article PubMed PubMed Central Google Scholar
Le Chalony, C. et al. Partial complementation of a DNA ligase I deficiency by DNA ligase III and its impact on cell survival and telomere stability in mammalian cells. Cell. Mol. Life Sci. 69, 2933–2949 (2012).
Article PubMed PubMed Central Google Scholar
Bhandari, S. K. et al. Unchanged PCNA and DNMT1 dynamics during replication in DNA ligase I-deficient cells but abnormal chromatin levels of non-replicative histone H1. Sci. Rep. 13, 4363 (2023).
Article ADS PubMed PubMed Central Google Scholar
Pascal, J. M. The comings and goings of PARP-1 in response to DNA damage. DNA Repair (Amst.) 71, 177–182 (2018).
Article PubMed Google Scholar
Langelier, M. F. et al. PARP enzyme de novo synthesis of protein-free poly(ADP-ribose). Mol. Cell. https://doi.org/10.1016/j.molcel.2024.10.024 (2024).
Okano, S., Lan, L., Caldecott, K. W., Mori, T. & Yasui, A. Spatial and temporal cellular responses to single-strand breaks in human cells. Mol. Cell Biol. 23, 3974–3981 (2003).
Article PubMed PubMed Central Google Scholar
Okano, S., Lan, L., Tomkinson, A. E. & Yasui, A. Translocation of XRCC1 and DNA ligase IIIalpha from centrosomes to chromosomes in response to DNA damage in mitotic human cells. Nucleic Acids Res. 33, 422–429 (2005).
Article PubMed PubMed Central Google Scholar
Caldecott, K. W. XRCC1 protein; Form and function. DNA Repair (Amst.) 81, 102664 (2019).
Article PubMed Google Scholar
Polo, L. M. et al. Efficient single-strand break repair requires binding to both poly(ADP-ribose) and DNA by the central BRCT domain of XRCC1. Cell Rep. 26, 573–581.e575 (2019).
Article PubMed PubMed Central Google Scholar
Wilson, S. H. & Kunkel, T. A. Passing the baton in base excision repair. Nat. Struct. Biol. 7, 176–178 (2000).
Article PubMed Google Scholar
Sallmyr, A., Bhandari, S. K., Naila, T. & Tomkinson, A. E. Mammalian DNA ligases; roles in maintaining genome integrity. J. Mol. Biol. 436, 168276 (2024).
Article PubMed Google Scholar
Mackey, Z. B. et al. DNA ligase III is recruited to DNA strand breaks by a zinc finger motif homologous to that of poly(ADP-ribose) polymerase. Identification of two functionally distinct DNA binding regions within DNA ligase III. J. Biol. Chem. 274, 21679–21687 (1999).
Article PubMed Google Scholar
Leppard, J. B., Dong, Z., Mackey, Z. B. & Tomkinson, A. E. Physical and functional interaction between DNA ligase IIIalpha and poly(ADP-Ribose) polymerase 1 in DNA single-strand break repair. Mol. Cell Biol. 23, 5919–5927 (2003).
Article PubMed PubMed Central Google Scholar
Cotner-Gohara, E., Kim, I. K., Tomkinson, A. E. & Ellenberger, T. Two DNA-binding and nick recognition modules in human DNA ligase III. J. Biol. Chem. 283, 10764–10772 (2008).
Article PubMed PubMed Central Google Scholar
Cannan, W. J., Rashid, I., Tomkinson, A. E., Wallace, S. S. & Pederson, D. S. The human ligase IIIα-XRCC1 protein complex performs DNA nick repair after transient unwrapping of nucleosomal DNA. J. Biol. Chem. 292, 5227–5238 (2017).
Article PubMed PubMed Central Google Scholar
Odell, I. D. et al. Nucleosome disruption by DNA ligase III-XRCC1 promotes efficient base excision repair. Mol. Cell. Biol. 31, 4623–4632 (2011).
Article PubMed PubMed Central Google Scholar
McNally, J. R. & O’Brien, P. J. Kinetic analyses of single-stranded break repair by human DNA ligase III isoforms reveal biochemical differences from DNA ligase I. J. Biol. Chem. 292, 15870–15879 (2017).
Article PubMed PubMed Central Google Scholar
Mackey, Z. B. et al. An alternative splicing event which occurs in mouse pachytene spermatocytes generates a form of DNA ligase III with distinct biochemical properties that may function in meiotic recombination. Mol. Cell Biol. 17, 989–998 (1997).
Article PubMed PubMed Central Google Scholar
Sutton, T. B. et al. Global screening of base excision repair in nucleosome core particles. DNA Repair (Amst.) 144, 103777 (2024).
Article PubMed Google Scholar
Schaich, M. A. et al. Single-molecule analysis of DNA-binding proteins from nuclear extracts (SMADNE). Nucleic Acids Res. 51, e39 (2023).
Article PubMed PubMed Central Google Scholar
Crut, A., Nair, P. A., Koster, D. A., Shuman, S. & Dekker, N. H. Dynamics of phosphodiester synthesis by DNA ligase. Proc. Natl. Acad. Sci. USA 105, 6894–6899 (2008).
Article ADS PubMed PubMed Central Google Scholar
Abdou, I., Poirier, G. G., Hendzel, M. J. & Weinfeld, M. DNA ligase III acts as a DNA strand break sensor in the cellular orchestration of DNA strand break repair. Nucleic Acids Res. 43, 875–892 (2015).
Article PubMed Google Scholar
Taylor, R. M., Whitehouse, C. J. & Caldecott, K. W. The DNA ligase III zinc finger stimulates binding to DNA secondary structure and promotes end joining. Nucleic Acids Res. 28, 3558–3563 (2000).
Article PubMed PubMed Central Google Scholar
Luger, K., Mäder, A. W., Richmond, R. K., Sargent, D. F. & Richmond, T. J. Crystal structure of the nucleosome core particle at 2.8 Å resolution. Nature 389, 251–260 (1997).
Article ADS PubMed Google Scholar
Cannan, W. J., Rashid, I., Tomkinson, A. E., Wallace, S. S. & Pederson, D. S. The human ligase IIIalpha-XRCC1 protein complex performs DNA nick repair after transient unwrapping of nucleosomal DNA. J. Biol. Chem. 292, 5227–5238 (2017).
Article PubMed PubMed Central Google Scholar
David, S. S., O’Shea, V. L. & Kundu, S. Base-excision repair of oxidative DNA damage. Nature 447, 941–950 (2007).
Article ADS PubMed PubMed Central Google Scholar
Lowary, P. T. & Widom, J. New DNA sequence rules for high affinity binding to histone octamer and sequence-directed nucleosome positioning. J. Mol. Biol. 276, 19–42 (1998).
Article PubMed Google Scholar
Díaz-Celis, C. et al. Assignment of structural transitions during mechanical unwrapping of nucleosomes and their disassembly products. Proc. Natl. Acad. Sci. USA 119, e2206513119 (2022).
Article PubMed PubMed Central Google Scholar
McGinty, R. K. & Tan, S. Principles of nucleosome recognition by chromatin factors and enzymes. Curr. Opin. Struct. Biol. 71, 16–26 (2021).
Article PubMed PubMed Central Google Scholar
Duclos, S., Doublié, S. & Wallace, S. S. In The Cellular Response to the Genotoxic Insult Issues in Toxicology (eds Greim, H. & Albertini, R.) 115–159 (The Royal Society of Chemistry, 2012).
Sallmyr, A., Rashid, I., Bhandari, S. K., Naila, T. & Tomkinson, A. E. Human DNA ligases in replication and repair. DNA Repair (Amst.) 93, 102908 (2020).
Article PubMed Google Scholar
Tomkinson, A. E. & Sallmyr, A. Structure and function of the DNA ligases encoded by the mammalian LIG3 gene. Gene 531, 150–157 (2013).
Article PubMed Google Scholar
Tang, Q. & Caglayan, M. The scaffold protein XRCC1 stabilizes the formation of polbeta/gap DNA and ligase IIIalpha/nick DNA complexes in base excision repair. J. Biol. Chem. 297, 101025 (2021).
Article PubMed PubMed Central Google Scholar
McNally, J. R. Kinetic Analysis of Human DNA Ligase III Doctor of Philosophy thesis, University of Michigan. (2019).
Dong, Z. & Tomkinson, A. E. ATM mediates oxidative stress-induced dephosphorylation of DNA ligase IIIalpha. Nucleic Acids Res. 34, 5721–5279 (2006).
Article PubMed PubMed Central Google Scholar
Nosella, M. L. et al. Poly(ADP-ribosyl)ation enhances nucleosome dynamics and organizes DNA damage repair components within biomolecular condensates. Mol. Cell 84, 429–446 e417 (2024).
Article PubMed Google Scholar
Ryan, B. J., Weaver, T. M., Spencer, J. J. & Freudenthal, B. D. In Base Excision Repair Pathway: Methods and Protocols (eds Kishor, K. B. & Tapas, K. H.) 55–76 (Springer US, 2023).
Schnable, B. L. et al. Thymine DNA glycosylase combines sliding, hopping, and nucleosome interactions to efficiently search for 5-formylcytosine. bioRxiv. https://doi.org/10.1101/2023.10.04.560925 (2024).
Hashemi Shabestari, M., Meijering, A. E. C., Roos, W. H., Wuite, G. J. L. & Peterman, E. J. G. In Methods in Enzymology Vol. 582 (eds Maria, S. & Chemla, Y. R.) 85–119 (Academic Press, 2017).
Wang, M. D., Yin, H., Landick, R., Gelles, J. & Block, S. M. Stretching DNA with optical tweezers. Biophys. J. 72, 1335–1346 (1997).
Article ADS PubMed PubMed Central Google Scholar
Dickson, R. M., Cubitt, A. B., Tsien, R. Y. & Moerner, W. E. On/off blinking and switching behaviour of single molecules of green fluorescent protein. Nature 388, 355–358 (1997).
Article ADS PubMed Google Scholar

Download references

Acknowledgements

We appreciate the helpful discussions with Dr. Neil Kad while this work was in progress. We would like to thank Marco Simonetta (Lumicks) for generating DNA handles for tethering experiments and Dr. Janet Iwasa for creating the animation videos. This work was supported by NIH R35ES031638 (BVH), S10OD032158 (BVH) and 2P30CA047904 to the UPMC Hillman Cancer Center, R35GM128652 (BDF), F32ES034982 (MAS), F32GM140718 (TMW), R01ES012512 (AET) P01CA092584 (AET, PI Tainer), P30CA118100 to UNMCCC. The manuscript’s contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIEHS or NIH.

Author information

Authors and Affiliations

Department of Pharmacology and Chemical Biology, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA
Ashna Nagpal, Matthew A. Schaich & Bennett Van Houten
UPMC-Hillman Cancer Center, Pittsburgh, PA, USA
Ashna Nagpal, Matthew A. Schaich, Vera Roginskaya, Liam P. Leary & Bennett Van Houten
Department of Biochemistry and Molecular Genetics, University of Virginia School of Medicine and University of Virginia Comprehensive Cancer Center, Charlottesville, VA, USA
Tyler M. Weaver
University of New Mexico Comprehensive Cancer Center and the Departments of Internal Medicine, and Molecular Genetics & Microbiology, University of New Mexico Health Sciences Center, Albuquerque, NM, USA
Annahita Sallmyr & Alan E. Tomkinson
Department of Biochemistry and Molecular Biology, University of Kansas Medical Center, Kansas City, KS, USA
Bret D. Freudenthal

Authors

Ashna Nagpal
View author publications
Search author on:PubMed Google Scholar
Matthew A. Schaich
View author publications
Search author on:PubMed Google Scholar
Tyler M. Weaver
View author publications
Search author on:PubMed Google Scholar
Vera Roginskaya
View author publications
Search author on:PubMed Google Scholar
Annahita Sallmyr
View author publications
Search author on:PubMed Google Scholar
Liam P. Leary
View author publications
Search author on:PubMed Google Scholar
Bret D. Freudenthal
View author publications
Search author on:PubMed Google Scholar
Alan E. Tomkinson
View author publications
Search author on:PubMed Google Scholar
Bennett Van Houten
View author publications
Search author on:PubMed Google Scholar

Contributions

A.N.: Formal analysis, Conceptualization, Investigation, Visualization, Writing—original draft, Writing—review & editing. M.S.: Formal analysis, Investigation, Methodology, T.W.: Formal analysis, Investigation, Methodology, V.R.: Formal analysis, Investigation, Validation. L.L.: Formal analysis, Investigation. B.F.: Conceptualization, Project administration, Resources, Supervision, Writing—review & editing, A.S.: Formal analysis, Investigation, Validation, Writing—review & editing, A.T.: Conceptualization, Investigation, Writing—review & editing, B.V.H.: Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Writing—original draft, Writing—review & editing.

Corresponding author

Correspondence to Bennett Van Houten.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Gregory Lohman and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Description of Additional Supplementary Files

Supplementary Movie 1

Supplementary Movie 2

Transparent Peer Review file

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Nagpal, A., Schaich, M.A., Weaver, T.M. et al. The zinc finger of DNA ligase 3α binds to nucleosomes via an arginine anchor. Nat Commun 16, 11531 (2025). https://doi.org/10.1038/s41467-025-66320-8

Download citation

Received: 14 February 2025
Accepted: 04 November 2025
Published: 21 November 2025
Version of record: 29 December 2025
DOI: https://doi.org/10.1038/s41467-025-66320-8