PARP enzyme de novo synthesis of protein-free poly(ADP-ribose)

Abstract

PARP enzymes transfer ADP-ribose from NAD⁺ onto proteins as a covalent modification that regulates multiple aspects of cell biology. Here, we identify an undiscovered catalytic activity for human PARP1: de novo generation of free PAR molecules that are not attached to proteins. Free PAR production arises when a molecule of NAD⁺ or ADP-ribose docks in the PARP1 acceptor site and attaches to an NAD⁺ molecule bound to the donor site, releasing nicotinamide and initiating ADP-ribose chains that emanate from NAD⁺/ADP-ribose rather than protein. Free PAR is also produced by human PARP2 and the PARP enzyme Tankyrase. We demonstrate that free PAR in cells is generated mostly by PARP1 de novo synthesis activity rather than by PAR-degrading enzymes PAR glycohydrolase (PARG), ARH3, and TARG1 releasing PAR from protein. The coincident production of free PAR and protein-linked modifications alters models for PAR signaling and broadens the scope of PARP enzyme signaling capacity.

Keywords: ADP-ribose; ARH3; HPF1; PARG; PARP1; PARP2; Parthanatos; TARG1; Tankyrase; poly(ADP-ribose).

Figure 1.. PARP1 catalysis of ADP-ribose modifications.

(A) PARP1 domains flexibly linked when not bound to DNA damage; HD blocks NAD⁺ access to ART and restricts catalysis. On DNA damage, PARP1 domains assemble on break and engage the HD, leading to HD conformation that permits NAD⁺ binding to ART donor site for initiation of ADP-ribose modifications. PARP1 attaches ADP-ribose to itself (automodification), with a preference for Glu/Ser located in a region that follows BRCT fold. An ADP-ribose modification docks in ART acceptor site for extension with an ADP-ribose unit that comes from NAD⁺ in the donor site. (B) PARP enzyme reactions were fractionated using TCA precipitation and treated as indicated.

Figure 2.. PARP1 ART domain produces free PAR.

(A) Urea-PAGE of TCA-precipitated PARP reaction that yields soluble (sol) and insoluble (ins) fractions. PAR produced in vitro by PARP1 ART (10 μM) with 20 mM NAD⁺ for 30 min analyzed on urea-PAGE stained with SYBR Gold. Where indicated, purified PAR samples were treated with PARG (2 μM). (B) Purified PAR from soluble and insoluble samples in panel A were HPLC analyzed. 280 nm absorbance profile from salt gradient elutions are shown. NAD⁺, ADPr, and AMP standards were also analyzed. (C) HPLC-MS. Soluble and insoluble fractions from TCA precipitation were resolved on urea-PAGE as in panel A. Gel sections that migrated near the 7-mer standard were excised and eluted for HPLC-MS. MS identified PAR species as labeled with associated negative charge (z value). # designates molecules that lost a water. See Table S1. (D) Chemical structure of protein-linked PAR chains showing typical terminus of PAR 1″ end, yielding indicated PAR chains. Schematic structure for NAD⁺-(ADPr)_n, in which 1″ end of PAR is a complete NAD⁺ molecule (nicotinamide + ADPr). PAR of this type cannot arise from protein-linked PAR, as nicotinamide is released from NAD⁺ during protein modification and chain extension.

Figure 3.. NAD⁺ and ADPr molecules bind the acceptor site and initiate free PAR synthesis.

(A) ART domain reactions. Green arrows labeled 1, 2, and 3 represent three different outcomes for ADP-ribose generated when NAD⁺ is utilized in the donor site: 1. ADPr released into solution (hydrolysis); 2. ADPr transferred to form protein-ADPr (initiation); 3. ADPr transferred to ADPr molecule occupying the acceptor site (extension). Nicotinamide is released in all scenarios. Canonically, acceptor site extends ADPr attached to protein (red arrow, right: protein-linked PAR). An individual ADPr (or NAD⁺) docks in acceptor site to be extended to create de novo free PAR (red arrow, left). (B) Urea-PAGE of PAR from TCA-fractionated reaction: PARP1 ART (10 μM) in presence of 100 and 500 μM NAD⁺ for 30 min. (C) Urea-PAGE of soluble samples from TCA-fractionated PARP reactions as in panel B, performed using 500 μM NAD⁺ and with/without carba-NAD⁺ at 500 μM. (D) HPLC-MS of PAR-containing soluble reactions from panel C (without/with carba-NAD⁺). PAR species in two m/z regions with associated negative charge (z value). (E) X-ray structure of PARP1 ART bound to BAD in donor site and carba-NAD⁺ in acceptor site (this study, PDB: 9BPY). BAD and carba-NAD⁺ are displayed as sticks with orange carbons. Weighted 2F_O-F_C electron density map (1σ contour) is overlayed surrounding BAD and carba-NAD⁺ molecules. (F) Focused view of ART residues in donor site that bind BAD and residues in acceptor site that bind carba-NAD⁺. Residues are numbered and displayed as sticks with green carbons. See Figure S1, Table S2, and Table S3.

Figure 4.. DNA damage-activated PARP1 produces free PAR with HPF1 regulation.

(A) Urea-PAGE of PAR from TCA-fractionated reaction with full-length PARP1 (10 μM) and DNA (10 μM) using 100 and 500 μM NAD⁺ for 30 minutes. (B) Quantification of relative intensities from reactions in panel A. Integrated density (ID) values were normalized to highest signal (insoluble, 500 μM NAD⁺). Bars represent averages of three independent experiments; error bars represent standard deviations. Points represent values obtained for individual experiments. (C) Reaction using full-length PARP1 as in panel A, using 500 μM NAD⁺ and incubated for time points indicated. (D) Quantification of panel C, as described in panel B. (E) Reaction described in panel A performed with PARP1 (5 μM) and DNA (5 μM) and 500 μM NAD⁺ for 30 min in absence/presence of HPF1 at ratio 1:10 or 1:1 (HPF1:PARP1). (F) Quantification of panel E, as described in panel B. Two-sample, two-sided t tests compared relative PAR production between with/without HPF1 (*p< 0.05, **p< 0.005). (G) Reaction with PARP1 (10 μM) and DNA (10 μM) using 100 and 500 μM NAD⁺ for 30 minutes in absence/presence of HPF1 at ratio 1:20 (HPF1:PARP1). (H) Quantification of panel G, as described in panel F. See Figure S2.

Figure 5.. PARP2 and TNKS1 produce free PAR.

(A) Urea-PAGE of PAR from TCA-fractionated reactions with PARP2 (10 μM), DNA (10 μM), and 500 μM NAD⁺ for 30 min, without or with HPF1 (0.5 or 10 μM; a 1:20 or 1:1 ratio of HPF1:PARP2). (B) Quantification of relative intensities from panel A. Integrated density (ID) values were normalized to highest signal (insoluble, 1:20 HPF1:PARP2 ratio). Bars represent averages of three independent experiments; error bars represent the associated standard deviations. Points represent values obtained in individual experiments. (*p< 0.05, ns=not significant). (C) Urea-PAGE of PAR from TCA-fractionated reactions with TNKS1 ART or TNKS1 ARC-SAM-ART (10 μM) and 20 mM NAD⁺ for indicated times. (D and E) Quantification of relative PAR intensities from panel C, with TNKS1 ART in panel D and TNKS1 ARC-SAM-ART in panel E. See also Figure S3 and Table S6.

Figure 6.. Isolation of free PAR from cultured cells.

(A) Urea-PAGE of PAR from TCA-fractionated HEK293 cells incubated with/without PARG inhibitor (20 μM) and with/without MMS (0.01%) treatment for 30 min. (B) SDS-PAGE of cell aliquot from panel A, collected prior to TCA fractionation, and resolved on SDS-PAGE. Quantified lane intensities estimated relative cell quantities that resulted from treatments in panel A. (C) Whole cell analysis as in panel B, except gel was transferred to membrane for Western blot, probed with pan anti-ADPr binding reagent. (D) PAR intensity quantification from indicated treatments in panel A (+PARGi, ±MMS, soluble versus insoluble). PAR intensities were corrected for cell amounts using quantified SDS-PAGE in panel B. Relative abundance of soluble versus insoluble fraction is a percentage of total PAR (soluble + insoluble). Bars represent averages of three independent experiments; error bars represent standard deviations. Points represent values obtained for individual experiments. See also Figure S4.

Figure 7.. PARP1 de novo synthesis of free PAR in cells.

(A) Urea-PAGE of PAR isolated from indicated HEK293 cells treated with PARGi (20 μM) and MMS (0.01%) for 30 min. Cells were TCA-fractionated. (B) SDS-PAGE of cell aliquot from panel A, collected prior to TCA fractionation and resolved on SDS-PAGE. Gel quantification estimated relative cell quantities from treatments in panel A. (C) PAR intensity quantification from cells in panel A, corrected for variations in cell amounts using quantified SDS-PAGE in panel B. Relative intensities are a ratio of sample intensity over intensity of soluble fraction of HEK293 WT cells. Bars represent averages of three independent experiments; error bars represent standard deviations. Points represent values from individual experiments. Two-sample, two-sided t tests compared relative PAR production between samples as shown (ns, not significant). (D) Western blots performed using anti-ARH3, anti-TARG1, or anti-HPF1 antibodies to confirm KO cell lines. (E) HPLC-MS of purified PAR samples from TCA-fractionated HEK293 WT cells. PAR species in two m/z regions with associated negative charge (z value). (F) HPLC-MS of purified PAR from TCA-fractionated HEK293 WT cells treated with/without carba-NR. PAR species in two m/z regions from soluble samples with associated negative charge (z value). See also Figure S5 and Table S7 and S8.