A novel non-canonical PIP-box mediates PARG interaction with PCNA

Figures

• 

  Figure 1.. PARG is acetylated at multiple sites in vivo. (A) Schematic presentation of lysine residues in the PARG regulatory and catalytic region that were found acetylated by mass spectrometry analysis. (B) FLAG co-immunoprecipitation of transiently co-transfected FLAG-PARG and myc-hMOF, myc-Tip60, myc-PCAF, myc-CBP and p300-HA acetyltransferases. 10 μM TSA and 20 mM NAM were added 6 h prior to cell lysis. Asterisk denotes a non-PARG band migrating at the same height in inputs blotted with anti-AcK.

• 

  Figure 2.. A non-canonical PIP-box mediates PARG interaction with PCNA in vivo. (A) K409 mutation abrogates PARG interaction with PCNA. Co-immunoprecipitation analysis of interaction between PCNA and FLAG-PARG K409A, K409R and K409Q mutants. 5% of input and 50% of eluate was loaded on the gels. (B) Fluorescent band intensity of K409 mutants relative to the wild-type PARG was quantified by normalizing eluted PCNA band intensity to the corresponding eluted FLAG-PARG band intensity from three independent experiments. (C) Alignment of non-canonical PIP-box motifs. (D) PARG non-canonical PIP-box motif is located within the disordered N-terminal PARG region. Disorder probability was calculated using IUpred software. (E) The circular dichroism (CD) spectrum of the N-terminal PARG region shows a negative minimum around 200 nm and a negative shoulder around 222 nm, suggesting high intrinsic disorder with a minor fraction of α-helix. (F) Schematic presentation of FRET resulting from the PARG–EGFP and mRFP–PCNA interaction. (G) Amplitude-averaged fluorescence lifetime map of PARG–EGFP in the nucleus of G1/G2 U2OS cells expressing PARG–EGFP + mRFP-empty (left) and PARG–EGFP + mRFP–PCNA (right). The…

• 

  Figure 3.. A non-canonical PIP-box mediates PARG interaction with PCNA in vitro. (A) ITC analysis of binding affinity between full-length his6-tagged PCNA (150 μM) and PARG wild-type, K409A, K409R and K409ac peptides (1.5 mM). (B) Dissociation constants (Kd) for wild-type, K409A, K409R and K409ac peptides. Peptide sequence is indicated under the graph; K409 is indicated in bold. Average from two independent experiments with two technical replicates are shown. Error bars represent sd.

• 

  Figure 4.. The putative N-terminal PIP-box is dispensable for PARG–PCNA interaction. (A) Location and alignment of PARG putative and non-canonical PIP-box. (B) Co-immunoprecipitation analysis of interaction between PCNA and FLAG-PARG truncation mutants and the putative PIP-box mutant Q76A. 5% of input and 50% of eluate was loaded on the gels. (C) Fluorescent band intensity of truncation mutants relative to the wild-type PARG was quantified by normalizing eluted PCNA band intensity to the corresponding eluted FLAG-PARG band intensity from three independent experiments. Error bars represent sd.

• 

  Figure 5.. Co-crystal structure of the PARG non-canonical PIP-box peptide bound to PCNA. (A) Three PARG PIP-box peptides (green) are bound to a PCNA trimer (red). (B) PARG PIP-box peptide binds the interdomain connector loop (IDCL) of PCNA. (C) PARG PIP-box peptide is nested in the hydrophobic region of PCNA surrounded by negatively charged residues. The color coded electrostatic surface potential of PCNA was drawn using the Adaptive Poisson–Boltzmann Solver package within PyMol. The electrostatic potential ranges from −5 (red) to +5 (blue) kT/e. (D) Electrostatic and hydrophobic interactions between PARG PIP-box peptide and PCNA. (E) Electrostatic and hydrophobic interactions of PARG K409 with PCNA S43 and V45. (F) ITC analysis of the changes in binding affinities due to mutations of PARG PIP-box residues. Two independent experiments with two technical replicates were performed. Error bars represent sd.

• 

  Figure 6.. PARG non-canonical PIP-box binds PCNA in a canonical mode. (A) Overlay of crystal structures of the PARG non-canonical PIP-box peptide and non-canonical PIP-box peptides from Pol κ (2ZVL) and Pol η (2ZVK). (B) Overlay of crystal structures of the PARG non-canonical PIP-box peptide and canonical PIP-box peptides from p21 (1AXC), polymerase delta p66 subunit (1U76) and FEN1 (1U7B).

• 

  Figure 7.. The effect of PARG mutations spanning the PARG non-canonical PIP-box on PCNA binding in vivo. (A,C) Co-immunoprecipitation analysis of interaction between PCNA and FLAG-PARG constructs. 5% of input and 50% of eluate was loaded on the gels. (B,D) Fluorescent band intensity of PARG constructs relative to the wild-type PARG was quantified by normalizing eluted PCNA band intensity to the corresponding eluted FLAG-PARG band intensity from three independent experiments. Error bars represent sd. P-values were calculated using Student's t-test.

• 

  Figure 8.. Dynamic interaction of the PARG peptide with PCNA. (A) Representative MD conformations of the PARG–PCNA complex obtained from the cluster analysis of concatenated MD trajectories (see Supplementary Figure S7 for details) and aligned in the reference frame of PCNA. PCNA is shown in cartoon (left panel) and surface representations (right panel), whereby the latter is coloured according to the molecular hydrophobicity potential (MHP). A PARG MD conformation closest to X-ray structure (backbone RMSD of 1.5 Å) is shown in cartoon representation. ‘Core’ interacting PARG residues are shown in line representation, while K409 is shown in stick representation. (B) Average frequency of contacts between PARG residues and PCNA, calculated from the MD simulations. (C) Detailed statistics for contact pairs of ‘core’ interacting PARG residues (bold characters). The inset shows intra-molecular H-bonds formed by D412.

• 

  Figure 9.. K409 mutation delocalizes PARG from PCNA-containing replication foci. (A) Immunofluorescence analysis of the localization of transiently transfected PARG-YFP wild-type and mutants throughout interphase. PCNA and Aurora B were used as markers of S phase and G2 phase, respectively. All PCNA and Aurora B images were adjusted with the same brightness/contrast settings. PARG-YFP images were differently adjusted to account for variable transfection levels. (B) The percentage of PARG mutants displaying punctate S-phase pattern. (C) 3D-SIM images of PARG–PCNA co-localization in S-phase U2OS cells. U2OS cells were transfected with PARG–EGFP wild-type, K409A, K409R and Q76A mutants and stained with anti-GFP and anti-PCNA. Representative images of co-localized pixels are shown. (D) Analysis of PARG–PCNA colocalization using Pearson colocalisation coefficient (Rcoloc). For Tetra spec beads Rcoloc = 0.9095 was calculated, while negative control without anti-PCNA staining gave Rcoloc = 0.005. Scale bars indicate 5 μm. Error bars represent sd. P-values were calculated using Student's t-test.

• 

  Figure 10.. K409 mutations impair PARG recruitment to the sites of laser-induced DNA damage in U2OS cells. DNA damage was induced with a 355 nm pulsed laser as a line across the nucleus followed by live imaging analysis of the recruitment of PARG-YFP wild-type and mutants to the laser line. (A) Maximum recruitment live images of PARG-YFP wild-type and point mutants in different stages of the cell cycle. Cells were fixed at the point of maximum recruitment and stained with anti-PCNA and anti-Aurora B as markers of S-phase and G2 cells, respectively. Imperfect alignment of composite pictures stems from the difference in acquisition (live vs fixed). Scale bars indicate 5 μm. (B and C) Analysis of the recruitment kinetics of (B) wild-type PARG-YFP, Q76A and K409 mutants in G1/G2 cells by spinning disc microscopy and (C) wild-type PARG and K409A in the presence of PARP1/2 inhibitors olaparib (10 μM) and talazoparib (100 nM) applied for 1 h prior to imaging. (D) Simultaneous imaging of mRFP–PCNA and PARG–EGFP WT and K409L recruitment and dissociation from laser-induced DNA damage sites by split-view widefield microscopy. Relative fluorescence intensity was rescaled according to maximum recruitment…