PARP1-dependent DNA-protein crosslink repair

Abstract

DNA-protein crosslinks (DPCs) are toxic lesions that inhibit DNA related processes. Post-translational modifications (PTMs), including SUMOylation and ubiquitylation, play a central role in DPC resolution, but whether other PTMs are also involved remains elusive. Here, we identify a DPC repair pathway orchestrated by poly-ADP-ribosylation (PARylation). Using Xenopus egg extracts, we show that DPCs on single-stranded DNA gaps can be targeted for degradation via a replication-independent mechanism. During this process, DPCs are initially PARylated by PARP1 and subsequently ubiquitylated and degraded by the proteasome. Notably, PARP1-mediated DPC resolution is required for resolving topoisomerase 1-DNA cleavage complexes (TOP1ccs) induced by camptothecin. Using the Flp-nick system, we further reveal that in the absence of PARP1 activity, the TOP1cc-like lesion persists and induces replisome disassembly when encountered by a DNA replication fork. In summary, our work uncovers a PARP1-mediated DPC repair pathway that may underlie the synergistic toxicity between TOP1 poisons and PARP inhibitors.

Fig. 1. DPCs on ssDNA gaps are ubiquitylated in a PARP1-dependent manner.

A Schematic of DPC pull-down assay. At given time points, the DPC plasmid is pulled down under stringent conditions, the DNA is digested by benzonase treatment, and M.HpaII is analyzed via immunoblotting. Note that although the M.HpaII antibody is generated against full-length M.HpaII, it is unlikely to recognize all degradation products. B pMH^ssDNA or pMH^dsDNA were incubated in high-speed supernatant extract (HSS, which is an extract that does not support DNA replication) and recovered by DPC pull-down at the indicated time points and immunoblotted for crosslinked M.HpaII. Molecular weight marker (kDa) is indicated on the left side of the blot and in all subsequent blots presented in this manuscript. C pMH^ssDNA was incubated in SPRTN- or SPRTN- and RFWD3-depleted HSS and retrieved at the indicated time points by DPC pull-down as in B. D pMH^ssDNA repair in SPRTN- and RFWD3-depleted HSS. Reactions were supplemented with untagged- or FLAG-tagged recombinant ubiquitin. For each condition, a sample was retrieved at 1 and 10 min, and was first recovered by DPC pull-down (DPC-PD), subsequently by FLAG pull-down (FLAG-PD), and immunoblotted for crosslinked M.HpaII. E pMH^ssDNA was incubated in SPRTN- and RFWD3-depleted HSS in the presence of the indicated inhibitors. DPCs were recovered via DPC pull-downs at the indicated time points and immunoblotted for crosslinked M.HpaII. Note that the small upshift of M.HpaII signal observed in the presence of Ub.E1i (lanes 5 and 6) is due to M.HpaII PARylation (see Fig. 2D). F pMH^ssDNA repair in SPRTN- and RFWD3-depleted HSS, which was further mock- or PARP1-depleted in the presence and absence of PARP inhibitor (PARPi). Samples were recovered by DPC pull-down and immunoblotted for crosslinked M.HpaII. The asterisk marks an unspecific band. G Scheme of PARP1. It consists of three main domains: an N-terminal DNA-binding domain (DBD) consisting of zinc-finger (ZF) motifs, a central BRCT domain-containing automodification domain, and a conserved C-terminal catalytic domain (CD). H SPRTN-RFWD3-depleted HSS was further mock- or PARP1-depleted and blotted with PARP1 antibody. PARP1-depleted extracts were supplemented with either buffer (+Buf.), recombinant hPARP1 (+WT), or recombinant catalytically impaired E988K mutant (+E988K). I Add-back rescue experiment using the extracts from H. Samples were recovered by DPC pull-down and immunoblotted against M.HpaII. Source data are provided as a Source Data file.

Fig. 2. PARP1-dependent DPC PARylation triggers DPC ubiquitylation and resolution.

A Overview table of PARP1 mutants. B SPRTN-RFWD3-depleted HSS was further mock- or PARP1-depleted and immunoblotted with PARP1 antibody. PARP1-depleted extracts were also supplemented with either buffer (+Buf.), DNA-binding deficient mutant (+ΔZF1-2), automodification-deficient mutant (+3SA), or catalytically impaired mutant PARP1 (+E988K). C Add-back rescue experiment using the extracts from B. DPCs were recovered by DPC pull-down and monitored via blotting against M.HpaII. The asterisk denotes an unspecific band. D pMH^ssDNA was incubated in RFWD3-SPRTN-depleted HSS in the presence of PARPi or E1 ubiquitin-activating enzyme inhibitor (Ub.E1i) alone or in combination. Samples were recovered via DPC pull-down and immunoblotted with either M.HpaII or Poly/Mono-ADP Ribose antibody (α-PAR). E pMH^ssDNA was incubated in SPRTN-RFWD3-depleted HSS, which was then either mock- or PARP1-depleted in the presence of Ub.E1i where indicated. DPC degradation and PARylation were monitored as in D. F pMH^ssDNA and pMH^dsDNA were incubated in SPRTN-RFWD3-depleted HSS supplemented with PARGi where indicated. DPC degradation and PARylation were monitored as in D. Source data are provided as a Source Data file.

Fig. 3. Plasmid Pull-down Mass Spectrometry (PP-MS) reveals proteasome and PAR-dependent E3 ubiquitin ligase recruitments.

A Overview of reaction conditions for PP-MS analysis. B Heatmap showing the mean of the z-scored log₂ label-free quantitation (LFQ) intensity (i.e., protein abundance) from four biochemical replicates of pCTRL and pMH^ssDNA incubated in SPRTN-RFWD3-depleted HSS in the presence of the indicated inhibitors. Dynamic proteins, responsive to PARGi and/or PARPi treatment, were selected. C pMH^ssDNA was incubated in SPRTN-RFWD3-depleted HSS, which was then additionally mock-depleted, PARP1-depleted, or PARP1-depleted and supplemented with recombinant PARP1 (Input, left WB). Plasmids were recovered via plasmid pull-down and protein recruitment to the plasmid was monitored with the indicated antibodies. D pMH^ssDNA was incubated in SPRTN-RFWD3-depleted HSS, which was further mock-depleted, PSA1-depleted, DDI2-depleted, or PSA1- and DDI2-depleted. Samples were recovered by DPC pull-down and immunoblotted against M.HpaII. Source data are provided as a Source Data file.

Fig. 4. PARP1 orchestrates TOP1cc and TOP1cc-like DPC repair.

A Schematic illustrating the current model of TOP1cc repair. B PARP1 deletion suppresses the visibility of camptothecin-induced SSBs. The indicated RPE-1 cells were incubated with 10 µM camptothecin for 1 h and then processed for alkaline comet assays to measure unrepaired SSBs. Data show the median Tail moment of 300 cells combined from three independent experiments (100 cells/experiment). Significant differences were determined by 2-way ANOVA with Sidak’s post hoc multiple comparisons test. CPT denotes camptothecin. **** denotes p-value ≤0.0001. C Proteasome inhibition and/or PARP1 deletion suppresses the visibility of camptothecin-induced SSBs in the alkaline comet assay pre-treated with proteasome inhibitor (50 µM MG132) 2 h prior to and during camptothecin treatment as above. D To generate pFLP, Flp-nick is crosslinked to a plasmid containing the FRT site. Products of Flp-nick incubation with a CTRL (pCTRL) or FRT-site containing plasmid (pFRT) were analyzed on a native agarose gel. E pFLP was incubated in non-replicating egg extracts (1:2 HSS:NPE ratio) in the absence and presence of Ub.E1i. Reaction samples were analyzed by native agarose gel electrophoresis. F Quantification of the experiment shown in E. Error bars represents the SD of the mean. n = 3 independent experiments. Significant differences were determined by a two-tailed unpaired t-test. ** denotes p-value ≤0.01 (p = 0.0061). G Analysis of protein recruitment to pFLP compared to pCTRL via PP-MS. Plasmids were recovered at 10 min after addition in non-replicating egg extracts (1:2 HSS:NPE ratio). The volcano plot shows the difference in the abundance of proteins between the two sample conditions (x-axis), plotted against the p-value resulting from two-tailed Student’s two-sample t-testing (y-axis). Proteins significantly down- or up-regulated (cutoff line represents permutation-based FDR < 5%) are represented in red or blue, respectively. n = 4. H pFLP was incubated in non-replicating egg extracts (1:2 HSS:NPE ratio) in the absence and presence of the indicated inhibitors. Reaction samples were analyzed as in (E). I Quantification of the experiment shown in (H) as in (F). Error bars represent the SD of the mean. n = 3 independent experiments. Significant differences were determined by a two-tailed unpaired t-test. ** denotes p-value ≤0.01 compared to the Mock control (p = 0.0052 for PARPi and p = 0.0061 for Ub.E1i). J Add-back rescue experiment with WT and E988K PARP1 in non-replicating egg extracts (1:2 HSS:NPE ratio). Reaction samples were analyzed as in (E). K Samples from (J) were quantified as in (F). Error bars represent the SD of the mean. n = 3 independent experiments. Significant differences were determined by a two-tailed unpaired t-test. *** and **** denote p-value ≤0.001 and p-value ≤ 0.0001 compared to the Mock control, respectively. ** denotes p-value ≤0.01 compared to the PARP1 depletion control (p = 0.0054). * denotes p-value ≤0.05 compared to Mock control (p = 0.02). L pFLP^PK was incubated in non-replicating egg extracts (1:2 HSS:NPE ratio) egg extracts in the absence and presence of Ub.E1i or in PARP1-depleted egg extracts. Reaction samples were as in E. Source data are provided as a Source Data file.

Fig. 5. Inhibition of pFLP repair leads to replisome disassembly during DNA replication.

A pCTRL and pFLP were replicated in the presence or absence of [α-³²P]dATP and PARPi (licensing in one volume of HSS for 60 min followed by the addition of two volumes of NPE). Samples were retrieved at indicated time points following NPE addition and analyzed by native agarose gel electrophoresis (top radiograph) or western blotting (bottom immunoblots). Samples were immunoblotted for p.CHK1 and loading control (ORC2). B Top, leftward, and rightward fork models of replisome disassembly when encountering the Flp-nick crosslink. Bottom, pFLP was replicated in the presence of [α-³²P]dATP and with or without PARPi. Samples were retrieved at indicated time points, phenol-chloroform extracted, digested with PstI and SapI, and resolved on a denaturing polyacrylamide gel. C Schematic of pFLP linearization by ScaI and the potential DNA species occurring on a 2D gel upon plasmid replication. D Indicated samples from (B) were linearized by ScaI and run in two dimensions. Source data are provided as a Source Data file.

Fig. 6. Mechanism of PARP1-mediated DPC resolution.

A Illustration of PARP1-mediated targeting of TOP1cc-like lesions. B In the absence of repair (e.g., PARPi) TOP1cc-like lesions encountered on the leading strand template induce CMG run-off. C Unrepaired TOP1cc-like lesions encountered on the lagging strand template induce CMG disassembly.