PARP2 mediates branched poly ADP-ribosylation in response to DNA damage

Abstract

Poly(ADP-ribosyl)ation (PARylation) is a posttranslational modification involved in multiple biological processes, including DNA damage repair. This modification is catalyzed by poly(ADP-ribose) polymerase (PARP) family of enzymes. PARylation is composed of both linear and branched polymers of poly(ADP-ribose) (PAR). However, the biochemical mechanism of polymerization and biological functions of branched PAR chains are elusive. Here we show that PARP2 is preferentially activated by PAR and subsequently catalyzes branched PAR chain synthesis. Notably, the direct binding to PAR by the N-terminus of PARP2 promotes the enzymatic activity of PARP2 toward the branched PAR chain synthesis. Moreover, the PBZ domain of APLF recognizes the branched PAR chain and regulates chromatin remodeling to DNA damage response. This unique feature of PAR-dependent PARP2 activation and subsequent PARylation mediates the participation of PARP2 in DNA damage repair. Thus, our results reveal an important molecular mechanism of branched PAR synthesis and a key biological function of branched PARylation.

Fig. 1

Loss-of-PARP2 affects the branched PAR chain formation. a Diagram depicting the procedure of sample preparation for dot blot, UV spectroscopy or mass spectrometry. Wide type or Parp2^−/− MEFs were treated with 500 μM H₂O₂ for 10 min to induce DNA damage. PAR was extracted and followed by digestion with pyrophosphatase (PPase) and alkaline phosphatase (AP) prior to LC–MS/MS. b Dot blot assays were performed with anti-PAR antibody. 5X means five-fold loading samples. Actin was used as the control of cell lysates (monoclonal antibody, Sigma A2228). c Depletion of PARP2-induced minor reduction of total PAR in response to DNA damage, whereas lacking PARP1 remarkably suppressed total PAR level. The levels of PAR were examined at 259 nm using UV spectroscopy. d Mass spectrometry detection of adenosine (Ado), ribosyladenosine (R-Ado) and diribosyladenosine (R₂-Ado). LC–MS/MS quantitative analysis showed that Ado (terminal PAR unit) and R₂-Ado (branched PAR unit) were remarkably reduced in PARP2-deficient cells, but not R-Ado (linear PAR unit). e Only full length PARP2, but not the E545A mutant restores the branched PAR chain formation. PARP2 was deleted in U2OS cells and reconstituted with either full length PARP2 (ΔPARP2 + FL) or the E545A mutant (ΔPARP2 + E545A, left panel). The levels of R₂-Ado were measured by LC–MS/MS (middle and right panels). Data are represented as mean ± s.d. as indicated from three independent experiments. Significance of differences was evaluated by Student’s t-test. NS: non significant; *statistically significant (p < 0.05); ***statistically significant (p < 0.001)

Fig. 2

PARP2 mediates the branched PAR chain formation. a A diagram of PARP2-dependent branched PAR chain formation. b PARP2 catalyzes branched PAR chain polymerization. The activator PAR was examined by mass spectrometry following the in vitro PARP2-dependent PARylation. c PARP1 cannot PARylate the existed PAR chains in the same tested condition as b. d The PARP2E545A mutant fails to catalyze additional PAR chains on the activator PAR. Data are represented as mean ± s.d. as indicated from three independent experiments. Significance of differences was evaluated by Student’s t-test. NS: non significant; ***statistically significant (p < 0.001)

Fig. 3

N-terminus of PARP2 mediates the branched PAR chain formation. a The NTR domain of PARP2 recognizes PAR. The recombinant GST fusion proteins were incubated with ³²P labeled PAR. Protein-PAR complex was pulled down by glutathione agarose beads followed by autoradiography (top panel). Recombinant GST and GST-RNF146 WWE domain were used as the negative and positive controls (NC and PC), respectively. The GST fusion proteins were also examined by the SDS-PAGE followed with Coomassie blue staining (bottom panel). b The NTR region of PARP2 is required for the interaction with PAR. Full length PARP2, Δ NTR or Δ WGR were incubated with ³²P labeled PAR. Associated PAR was examined (top panel). Recombinant PARP2 was stained with Coomassie blue (bottom panel). c The NTR domain is required for the activation of PARP2. Biotin-NAD⁺ was used in the in vitro PARP2-dependent PARylation assays. PARP2-dependent PARylation was examined by dot blot using streptavidin-HRP. The proteins were also examined by the SDS-PAGE followed with Coomassie blue staining (bottom panel). d The levels of PARP2 in U2OS and PARP2-null cells were examined by western blot (upper panel). The PARP2-null cells were reconstituted with full length PARP2 or the Δ NTR mutant. And their expression was confirmed by western blot (lower panel). e The NTR domain is required for the branched PAR chain synthesis in response to DNA damage. Cells were treated with 500 μM H₂O₂ for 10 min. PAR was extracted and processed by PPase and AP. R₂-Ado was quantified by LC–MS/MS examination. Data are represented as mean ± s.d. as indicated from three independent experiments. Significance of differences was evaluated by Student’s t-test. NS: non significant; ***statistically significant (p < 0.001)

Fig. 4

The recruitment of PARP2 to DNA lesions is mediated by PARP1-dependent PARylation. a The recruitment kinetics of PARP2 to DNA lesions. GFP-PARP1 or PARP2 were expressed in U2OS cells. Following laser micro-irradiation treatment, the recruitment of PARP1 or PARP2 was examined with live-cell imaging at the indicated time points (left panel). The relocation kinetics is shown in the right panel. Data represent mean ± s.d. from three biologically independent experiments (right panel). At least 20 cells were included in each experiment. b The recruitment of endogenous PARP2 was examined following laser micro-irradiation by anti-PARP2 antibodies. c The NTR domain alone is sufficient to be recruited to the DNA damage sites. The recruitment kinetics of GFP-NTR at different time points was measured. d PARP1 is required for the recruitment of PARP2. GFP-PARP2 or the NTR domain was expressed in the Parp1^+/+ or Parp1^−/− MEFs. The recruitment kinetics was examined at indicated time points. The percentage of recruitment under different PARP1 conditions is shown in the right panel e. The catalytic activity of PARP1 is required for the recruitment of PARP2. PARP1-deficient U2OS cells were stably expressing the full length PARP1 or the E988A mutant (catalytic inactive mutant). GFP-PARP2 was expressed in the cells, and the recruitment kinetics was measured. Scale bar represents 5 μm

Fig. 5

APLF recognizes the branched PAR. a Diagrammatic representation of the sample preparation. PAR was incubated with APLF PBZ or the WWE domain of RNF146, and then partially digested by PARG. The protein-associated residues were further treated with PPase and AP, and then subjected to LC–MS/MS. APLF PBZ (b) or the WWE domain of RNF146 (c) bound residues were examined. d The recruitment of APLF to DNA lesions is dependent on PARP2. GFP-APLF was expressed in U2OS or PARP2-null cells. The recruitment kinetics of APLF was examined following DNA damage at various time intervals. Scale bar represents 5 μm. Data are represented as mean ± s.d. as indicated from three independent experiments. Significance of differences was evaluated by Student’s t-test. NS: non significant; ***statistically significant (p < 0.001)

Fig. 6

PARP2-dependent branched PAR chain is important for histone H3 removal in response to DNA damage. a Doxycycline (DOX)-inducible Cas9 (iCas9) was expressed in the U2OS or U2OS PARP2-null cells. The expression of iCas9 was examined by western blot using anti-FLAG antibody (monoclonal antibody, Sigma F1804, the inset). A gRNA facilitates a single DSB as the AAVS1 locus. P1 represents a pair of primers flanking both sides of the DSB for the analysis of DSB repair. P2 is a pair of primers specific to a region 2 kb upstream to the DSB and serves as a negative control of DSB repair. P3 and P4 are two sets of primers located at 0.3 kb upstream and downstream respectively from the DSB. These two primers were used for the detection of histone H3 removal in response to DNA damage. b γH₂AX (polyclonal antibody, Abcam ab11174) was examined as a surrogate maker of the solo DSB at AAVS1 locus using immunofluorescence analysis. c Histone H3 removal is severely compromised in the absence of PARP2. Histone H3 removal was examined by ChIP assays with antibody (polyclonal antibody, Millipore 06-755) using P3 and P4 primers. d Depletion of APLF also suppressed histone H3 removal at the DSB. Depletion of APLF was verified by western blot using anti-APLF antibody (polyclonal antibody, Thermo Fisher Scientific PA5-39776, left panel). e DSB repair was monitored by q-PCR using P1 and P2. f Loss of either PARP2 or APLF significantly suppresses DSB repair. DSB repair in cells lacking PARP2 or APLF was examined by q-PCR. Scale bar represents 5 μm