A PP4-phosphatase complex dephosphorylates gamma-H2AX generated during DNA replication

Abstract

The histone H2A variant H2AX is rapidly phosphorylated in response to DNA double-stranded breaks to produce gamma-H2AX. gamma-H2AX stabilizes cell-cycle checkpoint proteins and DNA repair factors at the break site. We previously found that the protein phosphatase PP2A is required to resolve gamma-H2AX foci and complete DNA repair after exogenous DNA damage. Here we describe a three-protein PP4 phosphatase complex in mammalian cells, containing PP4C, PP4R2, and PP4R3beta, that specifically dephosphorylates ATR-mediated gamma-H2AX generated during DNA replication. PP4 efficiently dephosphorylates gamma-H2AX within mononucleosomes in vitro and does not directly alter ATR or checkpoint kinase activity, suggesting that PP4 acts directly on gamma-H2AX in cells. When the PP4 complex is silenced, repair of DNA replication-mediated breaks is inefficient, and cells are hypersensitive to DNA replication inhibitors, but not radiomimetic drugs. Therefore, gamma-H2AX elimination at DNA damage foci is required for DNA damage repair, but accomplishing this task involves distinct phosphatases with potentially overlapping roles.

Fig. 1. PP4C dephosphorylates γ–H2AX in vitro, and silencing PP4C increases cellular γ–H2AX in the presence and absence of exogenous DNA damage

A) Schematic depiction of the PP4C catalytic domain and sequence comparison of the respective catalytic domains of human PP4C and PP2AC with S. cerevisiae Pph3. Alignment of Pph3 with either PP4C or PP2AC shows relatively stronger homology between Pph3 and PP4C. The bars and highlighted residues represent the amino acids which were identical between PP4C and Pph3 but different for PP2AC. The percentage of identity and similarity between yeast Pph3 and PP4C or PP2AC is represented in a table in the lower right. B) Silencing PP4C increases basal γ-H2AX. HeLa cells were transfected with control or PP4C siRNAs and harvested after 48 h. Two different siRNA duplexes to PP4C were used individually or in combination, and one siRNA complex was used for PP2AC. Immunoblots were probed for PP4C, PP2AC or β-actin. Numbers in parentheses indicate signal intensity as a percent of the control. Subsequent experiments were performed using PP4C siRNA duplex #1, which efficiently silenced PP4C. PP4C (upper panel), PP2AC (lower panel) or control siRNA-transfected HeLa cells, treated or not with CPT (2 µM, 1 h) were harvested at indicated times and immunoblots of whole cell extracts were probed for γ–H2AX or H2AX. Relative to control, PP4C-silenced cells have higher levels of γ-H2AX in untreated cells and after CPT treatment, whereas PP2AC-silenced cells show a relative increase in γ-H2AX only after CPT treatment. C) PP4 dephosphorylates γ–H2AX in vitro as efficiently as PP2A. HA-tagged PP4C and PP2AC were in vitro transcribed and translated and the respective proteins were immunopurified using anti-HA beads without cross contamination (right). Immunoprecipitated PP4C and PP2AC were serially diluted in the phosphatase reaction. PP4 dephosphorylates human γ–H2AX assembled in mononucleosomes as efficiently as PP2A (left).

Fig. 2. Knocking down the PP4C-R2-R3β complex alters basal γ-H2AX levels

A) Reciprocal immunoprecipitation/immunoblot of PP4 complex proteins in 293T cell lysates. The PP4C antibody worked for immunoblot, but failed to immunoprecipitate detectable amounts of PP4C. The lower band in the PP4C blot is a nonspecific cross-reacting band. The PP4 complex subunits R1, R2, R3α and R3β associated with PP4C and did not interact with the other phosphatases, PP2AC and PPM1D. B) Schematic representation of the distinct PP4 complexes. C) siRNAs targeting specific PP4 subunits reduce their expression without altering other PP4 subunits or tubulin. HeLa cells transfected with the indicated siRNAs were harvested after 40 h and analyzed by immunoblot. Experiments described henceforth were performed using these siRNAs. Numbers in parentheses indicate signal intensity as a percent of control. D) Knocking down certain PP4 subunits increases basal γ-H2AX levels. Control or PP4 subunit specific siRNA-transfected HeLa cells, treated or not with CPT (2 µM, 1 h), were harvested at indicated times and whole cell extracts were probed for γ–H2AX or H2AX. PP4R3β, PP4R2 and PP4C deficient cells have increased γ-H2AX both in untreated cells and after CPT treatment.

Fig. 3. Knocking down PP4C or R3β induces γ–H2AX foci in the absence of exogenous DNA damage

PP4C (A) or PP4R3β (B) or control siRNA-transfected HeLa cells, treated or not with CPT (2 µM, 1 h), were fixed and stained with anti-γH2AX and DAPI and visualized using laser-scanning confocal microscopy. γ-H2AX foci form in untreated cells deficient in PP4C or PP4R3β.

Fig. 4. Subcellular localization of PP4C, PP4R3β and PP4R2 does not substantially change in response to exogenous DNA damage

Untreated or CPT-treated (5 µM, 1 h, 37°C) HeLa cells were washed to remove CPT and incubated for 2 h, before fixing and staining for γ–H2AX and PP4C (A), PP4R3β (B) or PP4R2 (C). Images were visualized using laser-scanning confocal microscopy. PP4C is primarily perinuclear with some nuclear staining, but PP4R3β and PP4R2 are predominantly nuclear, suggesting that the PP4C-R2-R3β complex mostly resides in the nucleus. None of these components substantially changed their localization following CPT treatment. D) The PP4C-R2-R3β complex is associated with chromatin in fractionated nuclei from HeLa cells, treated or not with 2 mM HU for 24 h or with 10 uM CPT for 2 or 4 h. Although PP4R2 was also found in the nuclear soluble fraction, PP4C and PPR3β were exclusively detected in the chromatin fraction. MDC1 served as a control for the chromatin fraction.

Fig. 5. ATR is responsible for γ-H2AX formation in PP4-deficient cells and ATR activity is not regulated by PP4

A) ATM inhibition does not affect basal γ-H2AX levels in PP4-silenced cells. PP4C, R3β or control siRNA-transfected HeLa cells were incubated with DMSO or 5 µM KU55933 (ATM inhibitor) and whole cell extracts were analyzed by immunoblot probed for γ–H2AX or H2AX (upper panel). To verify ATM inhibition, cells were irradiated (10 Gy) and immunoblotted for phospho1981-ATM or ATM. B) Inhibition of ATR, but not DNA-PK or ATM, decreases γ-H2AX in PP4-silenced cells. PP4C, R3β or control siRNA-transfected HeLa cells were incubated with 20 µM or 100 µM Wortmannin or with 4 mM caffeine and whole cell extracts analyzed by immunoblot for γ–H2AX or H2AX. C) Knocking down PP4 in ATR-deficient cells has no effect on basal γ-H2AX levels. PP4C and PP4R3β were knocked down in Seckel fibroblasts (right), and γ-H2AX levels analyzed by immunoblot in untreated cells or after irradiation (2 Gy). There was no effect of PP4-knock down on γ-H2AX in these cells (left). D) PP4 knockdown does not affect ATR activity. ATR was immunopurified from PP4 subunit-specific or control siRNA-transfected HeLa cells. The kinase activity of the ATR immunocomplex from PP4-deficient cells was assayed relative to the activity in control cells. The lower panel shows comparable amounts of immunoprecipitated ATR and ATRIP. siRNAs used were black, control; light dots, PP4C; dark dots, PP4R2; white, PP4R3β. Mean ±SD are shown. E) PP4 deficiency does not activate the checkpoint proteins or p53. PP4C, PP4R3β or control siRNA-transfected HeLa cells were incubated with DMSO or 5 mM HU 48 h after transfection and immunoblots were performed on whole cell extracts with the indicated antibodies (upper panel). HeLa cells transfected with PP4C, PP4R3β or control siRNAs were harvested at indicated times after transfection and immunoblots of whole cell extracts probed with the indicated antibodies (lower panel).

Fig. 6. γ-H2AX formation in PP4-knocked down cells is dependent on cell division

A) Knocking down PP4C in non-dividing cells (macrophages) has no effect on γ-H2AX. Human macrophages were transfected with control siRNAs or siRNAs targeting PP2AC or PP4. Immunoblots showed significant knockdown of PP2AC and PP4C (upper panel). Macrophages in which PP4C (middle panel) or PP2AC (lower panel) were knocked down were irradiated (5 Gy) and immunoblots of whole cell extracts at indicated times probed for γ–H2AX or H2AX. Macrophages with lower amounts of PP4C showed no change in γ-H2AX relative to control cells, but γ-H2AX persisted longer in cells with reduced PP2AC. B) γ-H2AX foci formation in PP4-deficient cells is induced by DNA replication. Serum starved 293T cells were transfected with siRNAs targeting PP4C, PP4R2 or PP4R3β. Transfected cells were released from the G0/G1 block by adding serum and BrdU was added to visualize cells that had undergone DNA replication. After 12 h, cells were washed before fixing and staining for BrdU, γ–H2AX and DAPI. Images were visualized using laser-scanning confocal microscopy. After partial depletion of PP4 complex proteins, only cells that were BrdU positive (had undergone DNA replication) showed γ-H2AX foci.

Fig. 7. PP4 is required to repair DNA replication-mediated DNA damage

A) Knocking down the PP4 complex impairs DNA repair in dividing cells. PP4C, PP4R3β or control siRNA-transfected 293T cells were serum starved for 2 d. Cells were either mainta ined in low serum or released from G0/G1 block by adding serum and analyzed by single cell gel electrophoresis (comet assay) 12 h later. Representative images are shown above. The comet tail moment of 75 cells for each condition was quantified using NIH Image software. The comet moment was normalized to that of control cells (black bar) and expressed as a percentage. DNA damage in the PP4C (white bar, p<0.02) and R3β (dotted bar, p<0.02) deficient cells was significantly increased only after DNA replication resumed. B) Measurement of HR-mediated repair of an I-SceI-induced DSB. U2OS cells carrying the recombination substrate (DR-GFP) were transfected with PP4C, PP4R3β or control siRNAs. I-SceI expression plasmid was transfected after 24 h and GFP+ cells measured 48 h later. HR repair was significantly impaired in cells transfected with PP4C (p<0.01) or R3β (p< 0.01) siRNA. C) PP4-silenced cells are hypersensitive to DNA replication inhibitors. Cell viability was analyzed by MTT assay relative to untreated cells. Curves were generated from 3 independent experiments. PP4C or PP4R3β deficient cells were hypersensitive to CPT (p<0.001, left) and HU (p<0.002, right), but not to bleomycin (lower) relative to control cells. In panels (A–C), the mean ±S.D. of representative experiments are shown.