HSP90 regulates DNA repair via the interaction between XRCC1 and DNA polymerase β

Abstract

Cellular DNA repair processes are crucial to maintain genome stability and integrity. In DNA base excision repair, a tight heterodimer complex formed by DNA polymerase β (Polβ) and XRCC1 is thought to facilitate repair by recruiting Polβ to DNA damage sites. Here we show that disruption of the complex does not impact DNA damage response or DNA repair. Instead, the heterodimer formation is required to prevent ubiquitylation and degradation of Polβ. In contrast, the stability of the XRCC1 monomer is protected from CHIP-mediated ubiquitylation by interaction with the binding partner HSP90. In response to cellular proliferation and DNA damage, proteasome and HSP90-mediated regulation of Polβ and XRCC1 alters the DNA repair complex architecture. We propose that protein stability, mediated by DNA repair protein complex formation, functions as a regulatory mechanism for DNA repair pathway choice in the context of cell cycle progression and genome surveillance.

Figure 1. Complex formation between DNA polymerase β and XRCC1 is not essential for the cellular response to DNA damage

(A) Structure (pdb3lqc) depicting oxidized XRCC1 (residues 1–151) bound to the Polβ(residues 142–335). The image is a cartoon rendition of the palm and thumb domains of Polβ in orange with a mesh illustrating the surface of the structure and a space-filling rendition of the oxidized form of XRCC1 in grey with a solid illustrating the surface of the structure. Amino acids L301 (yellow), V303 (cyan) and V306 (magenta) are shown using a space-filling rendering. The images were generated using PyMOL (Molecular Graphics System, Version 1.2r3pre; Schrödinger, LLC). (B) Stable LN428 cell lines expressing Flag-Polβ(WT) or the V303 loop mutants were probed for Polβ/XRCC1 complex formation by IP of the lentiviral-expressed Flag-Polβ transgene via the N-terminal Flag epitope tag and probing for XRCC1 and Flag-Polβ by immunoblot (See also Supplementary Figure 9). (C) Stable LN428 cell lines expressing Flag-Polβ(WT) or Flag-Polβ(TM) were probed for Polβ/XRCC1 complex formation by IP of XRCC1 (XRCC1-Ab) and probing for XRCC1 and Polβ by immunoblot. (D) Cell viability of LN428 cells expressing Flag-Polβ(WT), Flag-Polβ(TM), Flag-Polβ(K72A) or EGFP (as indicated) after H₂O₂ treatment, as measured by the Long-term assay. Plots show the relative surviving fraction as compared to untreated (control) cells. Means are calculated from triplicate values in each experiment. Results indicate the mean ± SD of three independent experiments. (E) Cell viability of LN428/MPG/Polβ-KD cells expressing Flag-Polβ(WT), Flag-Polβ(TM), Flag-Polβ(K72A) or EGFP (as indicated) after MNNG treatment, as measured by the MTS assay 48 hours after exposure. Plots show the relative surviving fraction as compared to untreated (control) cells. Results indicate the mean ± SD of three independent experiments. (F) Immunoblot of PAR to determine activation of PARP after exposure to MNNG (5μM) for the time indicated for LN428/MPG/Polβ-KD cells expressing Flag-Polβ(K72A) (left panel) or EGFP (right panel). PARP1 and PCNA protein expression levels are also shown as loading controls. (G) Immunoblot of PAR to determine activation of PARP after exposure to MNNG (5μM) for the time indicated for LN428/MPG/Polβ-KD cells expressing Flag-Polβ(WT) (left panel) or Flag-Polβ(TM) (right panel). PARP1 and PCNA protein expression levels are also shown as loading controls.

Figure 2. DNA polymerase β is recruited to DNA damage sites via PARP1 activation independent of XRCC1 complex formation

(A) Diagram describing a U2OS cell line (U2OS-TRE) with an integrated single-copy tandem array of tetracycline-response elements (TREs) and the light-induced localization of a Killer Red-tetR fusion protein to the TRE tandem array, mediated by the specific interaction between the TRE and the tetR protein, depicted by a red spot. (B) Fluorescent images depicting DNA damage-induced foci of Killer Red-tetR and copGFP-Polβ(WT) or copGFP-Polβ(TM) expressed in U2OS-TRE cells after light exposure (10 min), as indicated in the figure. Arrows point to the foci induced by tetR-KR and copGFP-Polβ(WT) or copGFP-Polβ(TM) and after pre-incubation with the PARP inhibitor PJ-34. Scale bar in the image indicates 2 μm. (C) The relative intensity of copGFP-Polβ(WT) or copGFP-Polβ(TM) foci induced by Killer Red-mediated reactive oxygen species^,, was quantified as shown. Results indicate the mean ± SD of the analysis of ten independent cells. (D) Kinetics of copGFP-Polβ(WT) or copGFP-Polβ(TM) recruitment to DNA damage sites. Cells were treated with a 405nm laser (500ms) and the images were obtained at the indicated times. The relative intensity of foci was then quantified. Results indicate mean ± SD of three independent experiments. (E) The relative intensity of foci in cells treated with 10ms or 500ms 405nm laser at 5 min was quantified. Results indicate mean ± SD of three independent experiments. The representative images were shown in Supplementary Figure 2B. (F) & (G) Suppression of (F) copGFP-Polβ(WT) or (G) copGFP-Polβ(TM) recruitment to DNA damage-induced foci by the PARP inhibitors ABT-888, PJ34 and BMN-673. Stable LN428 cells expressing copGFP-Polβ(WT) or and copGFP-Polβ(TM) were pre-treated with PARP inhibitors (PJ34, 4μM; or ABT-888, 10μM; or BMN-673, 5μM) for 1hr or without PARP inhibitor treatment, then cells were exposed to the 405nm laser (50ms or 1000ms). Images of cells were obtained after 2 min laser treatment. Results indicate mean ± SD of five to eight independent cells. A one tailed t-test was used for the statistical analysis and the p-value was determined comparing cells treated with the PARP inhibitors as compared to cells without PARP inhibitor treatment.

Figure 3. DNA polymerase β stability depends on complex formation with XRCC1

(A) Immunoblot of nuclear lysates from stable LN428 cell lines expressing Flag-Polβ(WT) or the V303 loop mutants (as indicated), probing for the steady-state levels of Flag-Polβ, XRCC1 and PCNA, as indicated. A representative immunoblot image is shown. (B) The relative level of Flag-Polβ in whole cell lysates and nuclear lysates from stable LN428 cell lines expressing Flag-Polβ(WT) or Flag-Polβ(TM) and from stable LN428/XRCC1-KD cells expressing Flag-Polβ(WT), as determined by immunoblot analysis, as in panel (A) and Supplementary Figure 3D. The result indicates mean ± SD of two independent experiments. (C) Immunoblot of whole-cell lysates from stable LN428 cell lines expressing Flag-Polβ(WT) or Flag-Polβ(TM) after exposure to cycloheximide (Cyclo) for the times indicated. The immunoblot image depicts the levels of Flag-Polβ, XRCC1 and PCNA, before and after treatment with Cyclo, showing the stability of Flag-Polβ(WT) and the rapid degradation of Flag-Polβ(TM). A representative immunoblot image is shown. (D) The relative level of Flag-Polβ in whole-cell lysates from stable LN428 cell lines expressing Flag-Polβ(WT) or Flag-Polβ(TM) and from stable LN428/XRCC1-KD cells expressing Flag-Polβ(WT). Cells were treated with Cycloor Cyclo+MG132 (6 hrs) and protein levels determined by immunoblot analysis, as in panel (E). The result indicates mean ± SD of two independent experiments. (E) Immunoblot of whole-cell lysates from stable LN428 cell lines expressing Flag-Polβ(WT) or Flag-Polβ(TM) after exposure to Cyclo + the proteasome inhibitor MG132 for the times indicated. The immunoblot image depicts the levels of Flag-Polβ, XRCC1 and PCNA, before and after treatment with Cyclo+MG132, showing the stability of Flag-Polβ(WT) and the restored stability of Flag-Polβ(TM) due to proteasome inhibition. A representative immunoblot image is shown.

Figure 4. The C-terminal domain of DNA polymerase β is targeted for proteasome-mediated degradation by ubiquitylation on K206/K244

(A) Immunoblot of whole cell lysates from stable LN428 cell lines expressing Flag-Polβ(TM) after expression of HA-ubiquitin. Flag-Polβ(TM) was isolated by IP and probed for the HA-ubiquitin modification by immunoblot with the HA-Ab. The level of Flag-Polβ(TM) in the WCL and the input level of Flag-Polβ(TM) is detected using the M2 Flag-Ab. (B) The relative level of Flag-Polβ in whole-cell lysates from stable LN428 cell lines expressing Flag-Polβ(WT), Flag-Polβ(TM) and the indicated K206A, K244A or K206A/K244A mutants. The predicted ubiquitylation sites K206 and K244 reverse the instability of Flag-Polβ(TM), as determined by immunoblot analysis, as in Supplementary Figure 4D. The result indicates mean ± SD of two independent experiments. (C) Immunoblot of whole-cell lysates from stable LN428 cell lines expressing Flag-Polβ(TM) or Flag-Polβ(TM/K206A/K244A) after exposure to Cyclo for the times indicated. The immunoblot image depicts the levels of Flag-Polβ, XRCC1 and PCNA, before and after treatment with Cyclo, showing the rapid degradation of Flag-Polβ(TM) and the stability of Flag-Polβ(TM/K206A/K244A). A representative immunoblot image is shown. For quantified results, see Supplementary Figure 4E. (D) Immunoblot of whole cell lysates from stable LN428 cell lines expressing Flag-Polβ(TM) or Flag-Polβ(TM/K206A/K244A) after expression of HA-ubiquitin. Flag-Polβ(TM) or Flag-Polβ(TM/K206A/K244A) was isolated by IP and probed for the HA-ubiquitin modification by immunoblot with the HA-Ab, showing the ubiquitylation of Flag-Polβ(TM) but not Flag-Polβ(TM/K206A/K244A). The level of Flag-Polβ(TM) or Flag-Polβ(TM/K206A/K244A) in the whole cell lysates and the input level of Flag-Polβ(TM) or Flag-Polβ(TM/K206A/K244A) is detected using the M2 Flag-Ab.

Figure 5. Unbound XRCC1 is an HSP90 client protein

(A) HSP90 interacts with XRCC1 but not other BER related proteins. Whole cell lysates (WCL) were prepared from proliferating T98G/MPG/Polβ-KD cells expressing either EGFP, Flag-Polβ(WT) or Flag-Polβ(TM) (left panel). XRCC1 and other BER related proteins were immuno-precipitated with an HSP90 antibody and examined by immunoblot as shown (right panel). (B) HA-HSP90 interacts with XRCC1 in Polβ-KO MEFs, not WT MEFs. WT (92TAg) and Polβ-KO (88TAg) MEFs, after expression of HA-HSP90, were probed for HSP90 interacting proteins by IP of the expressed HA-HSP90 via the N-terminal HA epitope tag and probing for XRCC1 by immunoblot. Input HA-HSP90 is also shown. (C) HSP90 inhibitor17-AAG treatment induces the degradation of XRCC1 in cells lacking Polβ (LN428/Polβ-KD cells and MEFs), as indicated: Immunoblot of WCLs from stable LN428 cell lines expressing Flag-Polβ(WT) or Flag-Polβ(TM) and from stable LN428/XRCC1-KD cells expressing Flag-Polβ(WT) (Shown in Top Panel)and Immunoblot of WCLs from WT (92TAg) and Polβ-KO (88TAg) MEFs (Shown in Bottom Panel)after exposure to 17-AAG at the concentrations indicated. The immunoblot image depicts the levels of XRCC1, Flag-Polβ and PCNA, before and after treatment with 17-AAG. A representative immunoblot image is shown. For quantified results, see Supplementary Figure 5D, E. (D) Cycloheximide-enhanced degradation of XRCC1 in LN428/Flag-Polβ(TM) cells treated with 17-AAG is protected by MG132: (Top Panel) Immunoblot of WCLs from stable LN428 cell lines expressing Flag-Polβ(TM) after exposure to 17-AAG (0 or 10μM, as indicated) and cycloheximide (Cyclo) for the times indicated. A representative immunoblot image is shown. (Bottom Panel) Immunoblot of WCLs from stable T98G/Polβ-KD cell lines expressing EGFP after exposure to 17-AAG (10μM) and Cyclo or Cyclo+MG132 for the times indicated. A representative image is shown. For quantified results, see Supplementary Figure 5F. (E) HSP90 knockdown induces the degradation of XRCC1 in LN428/Flag-Polβ(TM) and LN428/MPG/Polβ-KD/Flag-Polβ(TM) cells: Immunoblot of WCLs from LN428 cell lines expressing Flag-Polβ(TM)or LN428/MPG/Polβ-KD cells expressing Flag-Polβ(TM) after lentiviral-mediated expression of GFP or HSP90-specific shRNA, as indicated. A representative immunoblot image is shown.

Figure 6. CHIP-mediated degradation of XRCC1 is regulated by HSP90

Over-expression of the E3 ligase CHIP enhances the 17-AAG mediated degradation of XRCC1 in (A) LN428/MPG/Polβ-KD/Flag-Polβ(TM) cells and (B) T98G/MPG/Polβ-KD/Flag-Polβ(TM) cells: Immunoblot of whole-cell lysates from stable LN428/MPG/Polβ-KD cell lines (Panel A) or T98G/MPG/Polβ-KD/Flag-Polβ(TM) cell lines (Panel B) expressing Flag-Polβ(WT), Flag-Polβ(TM) or EGFP after exposure to 17-AAG at the concentrations indicated and after over-expression of CHIP, as indicated. The immunoblot images depict the levels of XRCC1, CHIP and PCNA, before and after treatment with 17-AAG, showing the stability of XRCC1 in the cells expressing Flag-Polβ(WT) and the rapid degradation of XRCC1 in the cells expressing Flag-Polβ(TM) or EGFP. Representative immunoblot images are shown. (C) CHIP knockdown efficiently prevents 17-AAG mediated degradation of XRCC1 in LN428 cells: Immunoblot of whole cell lysates from stable LN428 cells expressing CHIP-specific shRNA (shCHIP.2 or shCHIP.1) or GFP (control) after exposure to 17-AAG at the concentration indicated. The immunoblot image shows the level of XRCC1, CHIP and PCNA (loading control), before and after treatment with 17-AAG.

Figure 7. Dynamic interaction of DNA polymerase β and HSP90 with XRCC1 regulates base excision repair sub-pathway choice

(A) Cell viability in response to MNNG, H₂O₂ or ionizing radiation. Left Panel: LN428/MPG or LN428/MPG/XRCC1-KD cells expressing Flag-Polβ(WT), Flag-Polβ(TM) or EGFP after MNNG treatment, as measured by the MTS assay 48 hours after exposure. Results indicate the mean ± SD of three independent experiments. Middle Panel: LN428 or LN428/XRCC1-KD cells expressing Flag-Polβ(WT), Flag-Polβ(TM) or EGFP after H₂O₂ treatment, as measured by the CyQuant assay 10 days after exposure. Results indicate the mean ± SD of three independent experiments. Right Panel: Clonogenic survival assay of LN428/Polβ-KD, LN428/Puro and LN428/XRCC1-KD cells after exposure to ionizing radiation. Data points represent means and SD of at least three experiments, each performed in triplicate. The arrows indicate the extent of the requirement for Polβ (red) or XRCC1 (blue). Plots show the relative surviving fraction as compared to untreated cells. (B) Polβ/XRCC1 and XRCC1/HSP90 heterodimer ratios in proliferating (P) and confluent (C) MEFs (92TAg) were probed for heterodimer formation by IP of XRCC1 (XRCC1-Ab) and probing for XRCC1, HSP90 and Polβ by immunoblot. A representative immunoblot is shown. *¼ input IP product for XRCC1; S, short exposure time; L, longer exposure time. Bar graphs are plotted with mean ± SD of two independent experiments. (C) The relative ratio of Polβ/XRCC1 and HSP90/XRCC1 was quantified following immunoprecipitation of XRCC1 and analysis for Polβ or HSP90 (Supplementary Figures 7C, E) from proliferating and confluent cells (92TAg MEFs) treated with different DNA damaging agents. Bar graphs are plotted with mean ± SD of two independent experiments. (D) Proposed model for the dynamic regulation of the stability and degradation of Polβ and XRCC1. Varied cellular conditions, such as HSP90-phosphorylation, alterations in expression or function related to the cell cycle or DNA damage response and cell-type specificity promote the formation of one of the two heterodimers. Conditions that increase Polβ/XRCC1 levels would favour a preference for Polβ-dependent BER whereas conditions that increase XRCC1/HSP90 levels would favour a preference for XRCC1-dependent or Polβ-independent BER. The degradation of XRCC1 is regulated by HSP90 and CHIP.