Nucleosome unwrapping and PARP1 allostery drive affinities for chromatin and DNA breaks

Abstract

Poly(ADP-ribose) polymerase 1 (PARP1) detects DNA strand breaks that occur in duplex DNA and chromatin. Here, correlative optical tweezers and fluorescence microscopy reveal how single molecules of PARP1 identify single-strand breaks (i.e., nicks), undamaged nucleosome core particles (NCP) and NCPs containing DNA nicks. Fluorescently-tagged PARP1 or PARP2 from nuclear extracts binds nicks with nanomolar affinity but does not engage undamaged dsDNA regions. In contrast, PARP1 avidly binds undamaged NCPs, and partial NCP unwrapping induced by DNA tension significantly increases PARP1 on rate and affinity. Catalytically dead PARP1 or EB-47 inhibition greatly increases PARP1 affinity to DNA nicks and undamaged NCP, implicating a mechanism where PARP1 reverse allostery regulates PARP1 retention to undamaged chromatin. We also monitor ADP-ribosylation in real time upon PARP1 binding undamaged or nicked NCPs. These results provide key mechanistic insights into domain allostery and how pharmacological intervention alters PARP1 binding dynamics for therapeutic impacts.

Fig. 1. Fluorescently-tagged PARP1 engages nicked DNA substrates.

a A structural model of HaloTag-PARP1 built from aligning the alphafold3 structure with PARP1 structure PDB code 2N8A and HaloTag modeled from PDB code 6U32. Domains are colored as in the diagram beneath. b A schematic for performing these single-molecule experiments, in which streptavidin-coated beads are captured in optical traps (I), DNA (biotin = blue dots) tethered between the beads (II), the substrate is washed in a buffer channel (III), and then lastly moved into a channel with nuclear extracts to visualize overexpressed PARP1 binding events (IV). c To generate DNA nicks, Nt.BspQI digestion creates eight observable nicks distributed through the length of lambda DNA as shown. d Representative kymograph data of HaloTag-PARP1 binding nick sites (white lines are stationary events), as well as an example of dwell/gap times utilized to determine on and off rates. e A cumulative residence time distribution (CRTD) of the dwell times observed, with fit shown in red. f A cumulative gap time distribution (CGTD) determines the on rate for binding, with fit shown in green. These two rates can then be combined to determine K_D values for interaction. DNA figures created in BioRender by Schaich, M. (2025). https://BioRender.com/ufjd01j.

Fig. 2. ZnF domains 1 and 2 regulate nick binding.

a A representative kymograph (white lines indicate PARP1 binding events) of PARP1 ZnF1-2 binding nicked lambda DNA. Resultant CRTD and CGTD plots are shown below, as well as the measured K_D. b PARP2 engages nick sites (white lines represent binding). Off rate and on rate measurements are shown below, as well as resultant K_D. Created in BioRender. Schaich, M. (2025). https://BioRender.com/ufjd01j.

Fig. 3. PARP1 avidly binds nucleosome particles in the absence of DNA damage.

a A schematic showing the strategy for ligating reconstituted 601 nucleosomes into the DNA handles kit (LUMICKS), as well as a resultant 2D scan demonstrating colocalization between the Cy3 signal of the nucleosome and the ATTO 647 N dye (represented by red A). b Representative kymograph of WT PARP1 engaging fully wrapped nucleosomes at low DNA tension, with the distal (weak) arm marked in orange (from PDB 4JJN). CRTD and CGTD plots shown on the right, as well as resultant affinity measurement. c At tensions > 5 pN, the distal arm (orange) unwraps, see model shown. A representative kymograph is shown, as is the CRTD, CGTD, and affinity measurement. Nucleosome models created in BioRender. Schaich, M. (2025) https://BioRender.com/du7zz94.

Fig. 4. DNA non-ligatable nicks within nucleosome core particles are detected by PARP1.

a Structural model of the nucleosomes generated from PDB 4JJN, with sites of each nick site marked along the structure (with SHL0 labeled red, SHL-2.5 labeled cyan, and SHL−4.5 labeled yellow). Underneath, a model was generated for the partial nucleosome unwrapping that occurs at DNA tensions greater than 5 pN. b Lifetimes and on rate values for the binding kinetics of NCP or control nick binding. Bar graphs are colored by nick position in the same way that the model is shown, and low-tension datasets are shown on the left with high tension datasets on the right (hatched bars). Values for each bar graph were determined by fitting an exponential decay to the dwell times/gap times from four biological replicates into CRTD/CGTD plots (two different batches of nuclear extracts and NCPS pooled together). See Table 1 for bar values and Supplementary Fig. 4 for all datapoints associated with these plots as well as their fits and associated errors on their CRTD/CGTD plots. Error bars represent error of the fit from the plotting.

Fig. 5. PARP1 exhibits catalytic activity at the single-molecule level on its substrates.

a In the presence of NAD and HaloTag-PARP1 (red), PBZ-mRuby2 (green) engages the nucleosome site after PARP1 first binds it. CRTD and CGTD for the PBZ is shown to the right. b Of the PARP1 and PBZ events, many of them interacted on the nucleosome at the same time (middle of Venn diagram) vs alone. These categories are further broken down by order of assembly and disassembly as shown on the right, with category 9 (bound PARP1 visited by PBZ before dissociating) dominating. c Nicked nucleosomes also activate PARP1 as indicated by green PBZ-mRuby2 binding following PARP1 binding. CRTD and CGTD plots are also shown to the right for PBZ interactions. d The colocalization and categories for this substrate are also displayed. Datasets were generated from multiple scans of three and five DNA tethers, respectively, and bar graphs throughout display mean values of these observations as well as standard errors of the mean. Models created in BioRender. Schaich, M. (2025). https://BioRender.com/du7zz94.

Fig. 6. The role of PARP1 allostery on DNA nick and nucleosome binding kinetics.

a A representative kymograph of catalytically-dead PARP1 binding nicked DNA, with several events lasting over 100 s (white lines). CRTD and CGTD plots are also shown. b With 100 µM EB-47 present (which saturates the protein to > 99% bound), WT-PARP1 events last much longer (white lines), and c with 100 µM Olaparib included. Determination of on and off rates are shown below, as well as the calculated K_D for the interaction. d Binding kinetics and resultant CRTD/CGTD plots for catalytically-dead PARP1 and e inhibited PARP1 on undamaged nucleosomes. With each CRTD plot, the fit for uninhibited WT PARP1 on nicked DNA is shown as a dotted red line for reference (τ = 2.3 s). Created in BioRender. Schaich, M. (2025). https://BioRender.com/oml86m7.

Fig. 7. Detection of damage and chromatin by PARP1.

(Top) Model for how PARP1 uses its ZnF domains to rapidly sample nicks before fully engaging to reside for longer lifetimes that can be extended by pro-retention inhibitors like EB-47 or catalytic inactivation. ZnF domains are shown in blue and green, BRCT domain in orange, WGR domain in purple, HD domain in beige, and ART domain in pink. (Bottom) PARP1 engages and ADP-ribosylates (orange) undamaged and damaged chromatin. The affinity of PARP for NCPs can be increased in cases where catalysis does not occur through a reverse allostery mechanism. Further, nicks are recognized within the nucleosomes by PARP1, but partial unwrapping eliminates that specificity window. Created in BioRender. Schaich, M. (2025). https://BioRender.com/e38jk20.