The mitotic kinase Plk1 contributes to the DNA damage response (DDR) by targeting multiple factors downstream of the core responder kinase ATM/ATR. In this study, we show that Polo-like kinase 1 (Plk1) also phosphorylates key factors upstream of ATM/ATR and regulates their DDR-related functions. Plk1 phosphorylated Mre11, a component of the Mre11/Rad50/Nbs1 (MRN) complex, at serine 649 (S649) during DDR. Phosphorylation of Mre11-S649 by Plk1 primed subsequent CK2-mediated phosphorylation at Mre11-serine 688 (S688). Phosphorylation of Mre11 at S649/S688 inhibited loading of the MRN complex to damaged DNA, leading to both premature DNA damage checkpoint termination and inhibition of DNA repair. Tumors expressing phosphomimetic Mre11 were more sensitive to the PARP inhibitor olaparib, compared with those expressing unphosphorylatable Mre11, suggesting that patients with elevated Plk1 expression might benefit from olaparib treatment. Cancer Res; 77(12); 3169–80. ©2017 AACR.
Although it has been well documented that Polo-like kinase 1 (Plk1) has critical roles in many mitotic events (1, 2), mounting evidence suggests that Plk1 is also involved in many nonmitotic events, including DNA damage response (DDR; ref. 3) and G2 DNA damage checkpoint recovery (4). Throughout the life of an organism, the integrity of genome is continuously challenged by DNA lesions due to exposure to both exogenous (genotoxic chemicals or radiation) or endogenous (normal metabolic byproducts or DNA replication) agents. The DDR pathway is essential for the maintenance of genomic stability. The DDR cascade can be broadly divided into two phases: an early cascade comprising recruitment of the Mre11-Rad50-Nbs1 (MRN) complex, autoactivation of ATM/ATR, and generation of γH2AX, followed by a later cascade, induced by RNF8 that mediates the retention of RNF168, BRCA1, and 53BP1 (5). Upon activation, ATM/ATR phosphorylate a large number of substrates to regulate various downstream pathways. Cells have evolved multiple pathways in response to DNA damage that span from arrest of the cell-cycle progression (DNA damage checkpoint), repair of the DNA lesions (DNA repair), to the reestablishment of cellular homeostasis (checkpoint recovery). The DNA lesions are first sensed by the multifunctional MRN complex, which in turn activates ATM/ATR and triggers DNA damage response. There are two major competing pathways for repair of DNA double-strand breaks (DSB): homologous recombination (HR) and nonhomologous end-joining (NHEJ). Mre11 has both endo- and exonuclease activities against a variety of single-stranded DNA (ssDNA) and double-stranded DNA (dsDNA) substrates (6, 7) and these nuclease activities contribute to both HR (8) and NHEJ (9). Of note, it was reported that Plk1 also directly phosphorylates the Rad51 recombinase and facilitates HR in response to DSBs (10).
To have enough time to repair DNA lesions, cells will activate the checkpoint pathway to stop cell cycle. Activated ATM and ATR phosphorylate checkpoint effector kinases Chk2/1, which subsequently phosphorylate substrates including Cdc25, resulting in inactivation of Cdk1. ATM and ATR also directly phosphorylate Bora, which is recognized and degraded by E3 ligase SCFβTRCP (11). Once Bora is destructed, mitotic entry is restricted by Cdk1/cyclin B inhibition partially due to Plk1 inactivation (12). DNA damage–induced cell-cycle arrest will allow cells to repair damaged DNA. Upon completion of DNA repair, the checkpoint has to be silenced to restart cell-cycle progression. Compared with activation of the DNA damage checkpoint and repair, the exact mechanism of checkpoint recovery is poorly understood. Previous work showed that Plk1 plays a crucial role during recovery from G2 DNA damage checkpoint (4). Furthermore, the activity of Plk1 is required for inactivation of the ATR/Chk1 pathway through phosphorylation-induced degradation of ATR-mediator protein Claspin (13). Recently, it has been shown that Plk1 inactivates the ATM–Chk2 pathway through phosphorylation of 53BP1 and Chk2 (14). In response to G2 DNA damage, p53 pathway is also activated to lead to cell-cycle arrest. Plk1-associated activity inactivates p53 by phosphorylating GTSE1, a negative regulator of p53 (15). Collectively, the studies summarized above revealed an essential role of Plk1 in checkpoint recovery by targeting multiple factors downstream of ATM and ATR. Whether Plk1 regulates functions of proteins upstream of ATM and ATR is not known.
Overexpression of Plk1, observed in many cancers, has often been correlated with disease stage, histologic grade, poor prognosis, metastatic potential, and survival (1). Consequently, several Plk1 inhibitors have been developed, with some agents showing encouraging results in various cancer cell lines and xenograft models of human cancer (16, 17). Of these, volasertib (BI6727) has been most extensively studied clinically. In contrast to marked antitumor activity in multiple cancer models, overall antitumor activity of volasertib in patients with solid tumor has been modest in trials (16). To fully understand the role of Plk1 in the context of ionizing radiation (IR)-induced carcinogenesis, we need dissect how Plk1-associated kinase activity regulates DDR and its subsequent repair in more details.
Materials and Methods
Cell culture, transfections, and RNAi
HeLa, U2OS, HCT116, and HEK293T cells were cultured in DMEM supplemented with 10% FBS, 100 U/mL penicillin, and 100 U/mL streptomycin at 37°C in 8% CO2. Cell lines were obtained from ATCC in 2006. Additional cell line authentication and mycoplasma testing have not been tested after its purchase. The numbers of passages of cells used in the study were below 20. Plasmid transfections were performed with Liopfectamine (Invitrogen), whereas siRNA was transfected with Transmessenger transfection reagent (Qiagen). Cells were examined at 30 hours after transfection for immunoblotting. The pKD-CK2 construct used to deplete CK2 was obtained from Upstate Biotechnology. The following target sequence was used to design siRNA against human Plk1: AAGGGCGGCTTTGCCAAGTGCTT. To deplete Mre11, oligonucleotides with the following sense and antisense sequences were used for the cloning of short hairpin RNA (shRNA)-encoding sequences in lentiviral vector: 5′-GATCCCCGGCACTGAGAAACAT GCAATTCAAGAGAT TGCATGTTTCTCAGTGCCTTTTTGGAAA-3′ (sense) and 5′-AGCTTTTCCAAAAAGGCAC T GAGAAACATGCAATCTCTTGAATTGCATGTTTCTCAGTGCCGGG-3′ (antisense).
The phospho-specific antibodies against Mre11-S649 and Mre11-S688 were generated by Proteintech. In brief, two peptides containing phospho-S649 and phospho-S688 were synthesized and used to immunize rabbits. After the antibodies were affinity purified, a series of control experiments were performed to confirm its specificity. The antibodies against ATM (NB100-104) and phospho-ATM (p-S1981; NB100-81803) are from Novus Biologicals. The antibodies against Nbs1 (GTX70224), and Rad50 (GTX70228) were purchased from GeneTex. The antibodies against phospho-Chk2 (p-T68) (2661), Chk2 (2662), Mre11 (4847), and CK2 (2656) were obtained from Cell Signaling Technology. The antibodies against γ-H2AX (ab22551) were from Abcam, whereas the antibodies against Plk1 (05-844) were from Millipore.
Synchronization and checkpoint assay
Cells were incubated with thymidine (2.5 mmol/L) for 16 hours to arrest at the G1–S transition. Thymidine was washed away thoroughly to release cells into fresh medium. At 8 hours postrelease, the second round of thymidine were added to arrest cells at the G1–S boundary. Where indicated, the G2–M DNA damage checkpoint was activated by treating cells with 0.5 μmol/L doxorubicin for 1 hour at 7 hours postrelease from the double thymidine block (DTB). To inactivate DNA damage signaling and allow mitotic entry, caffeine (5 mmol/L) was added to inhibit ATM and ATR checkpoint kinases.
Recombinant protein purification
Various GST-tagged human Mre11 constructs were subcloned into pGEX-KG, expressed in E. coli, and purified. Point mutations were generated using the QuickChange Site-Directed Mutagenesis Kit (Stratagene). To purify Plk1 kinase, Hi5 insect cells were infected with baculovirus encoding GST-Plk1 and harvested 36 hours after infection. Recombinant GST-fusion proteins were affinity-purified by incubation with glutathione-agarose beads, followed by extensive washes with STE buffer (10 mmol/L Tris-HCl, pH 8.0, 1 mmol/L EDTA, 150 mmol/L NaCl) and elution with glutathione elution buffer (40 mmol/L glutathione, 50 mmol/L Tris-HCl, 10 mmol/L DTT, 200 mmol/L NaCl, pH 8.0).
In vitro kinase assay
In vitro kinase assays were performed in TBMD buffer [50 mmol/L Tris (pH 7.5), 10 mmol/L MgCl2, 5 mmol/L dithiothreitol, 2 mmol/L EGTA, 0.5 mmol/L sodium vanadate, 20 mmol/L p-nitrophenyl phosphate] supplemented with 125 μmol/L ATP and 10 μCi of [γ- 32P] ATP at 30°C for 30 minutes in the presence of GST-Mre11 proteins. After the reaction mixtures were resolved by SDS-PAGE, the gels were stained with Coomassie brilliant blue, dried, and subjected to autoradiography.
Xenopus egg extracts
The use and care of Xenopus laevis were approved by the Institutional Animal Care and Use Committee of the University of North Carolina at Charlotte (Charlotte, NC). Low-speed supernatant egg extract was prepared according to our previously published approach (18).
Electrophoretic mobility shift assay
dsDNA end binding of proteins was assessed by EMSA as described previously (19). Oligonucleotides were chemically synthesized and purified by HPLC (IDT). The sequences of oligonucleotides were: TP423: 5′- CTG CAG GGT TTT TGT TCC AGT CTG TAG CAC TGT GTA AGA CAG GCC AGA TC-3′; and TP424: 5′-CAC AGT GCT ACA GAC TGG AAC AAA AAC CCT GCA GTA CTC TAC TCA TCT C- 3′. After the 5′ end of TP423 was labeled with [γ-32P]ATP and polynucleotide kinase (New England Biolabs), TP423 and TP424 oligonucleotides were annealed to generate the 3′ overhanging DNA duplexes. After the 3′ overhanging DNA duplexes (10 pmol/L) were subsequently incubated with purified Mre11 (10 nmol/L) proteins in the binding buffer [25 mmol/L MOPS, pH 7.0, 60 nmol/L KCl, 0.2% Tween, 2 mmol/L DTT, 1 mmol/L Mg(CH3COO)2] at 73°C for 15 minutes, gel loading buffer [250 mmol/L Tris-HCl (pH 7.5), 0.2% bromphenol blue and 40% glycerol] was added to the reaction. The reaction products were separated on a 5% nondenaturing polyacrylamide gel at 4°C in Tris borate/EDTA buffer for 2 hours. After the gels were dried and exposed to the phosphorimaging screen, images were acquired using Typoon FLA 9500 and subsequently by autoradiography.
After irradiation, the 0.5% agarose microgel with embedded cells was placed into the lysis buffer containing 2.5 mol/L NaCl, 0.1 mol/L EDTA, 10 mmol/L Tris base, and 1% Triton X-100 (pH 12.5), at 4°C for 1 hour. Thereafter, slides were equilibrated with fresh electrophoresis buffer (0.3 mol/L NaOH, 1 mmol/L EDTA) for 40 minutes to unwind DNA, and then electrophoresed for 30 minutes at 25 V. The nucleoids were subsequently washed in neutralizing buffer (0.4 mol/L Tris-HCl, pH 7.5), before staining with DAPI (5 μg/mL; ref. 20).
All procedures involving mice were guided by Purdue University Animal Care and Use Committee (protocol no: 1111000133E001). Mice, housed in the animal facility with free access to standard rodent chow and water, were under pathogen-free conditions and maintained in a 12 hours light/12 hours dark cycle. To generate tumors, U2OS cells (3 × 106 cells per mouse) were mixed with an equal volume of Matrigel (Collaborative Biomedical Products) and inoculated into the right flank of nude mice (Harlan Laboratories). Two weeks later, animals were randomized into treatment and control groups with 6 mice each. Olaparib (50 mg/kg) was injected into the tail vein twice weekly. Tumor volumes were measured by formula: V = L × W2/2 [V is volume mm3;), L is length (mm), W is width (mm)]. Tumors were fixed in 10% neutral buffered formalin, paraffin embedded, sectioned to 5 μmol/L, and stained using conventional hematoxylin and eosin staining.
All data are presented as means ± SD. Statistical calculations were performed with Microsoft Excel analysis tools. While a two-tailed, unpaired Student t test was used to assess the difference between the effects of treatment in cell lines, one-way ANOVA was used to determine statistically significant differences from the means in the animal study. P values of <0.05 were considered statistically significant (*, P < 0.05; **, P < 0.01).
Plk1 suppresses DDR by inactivation of the ATM/Chk2 pathway
To dissect how Plk1-associated kinase activity affects DDR, we first assessed the status of H2AX phosphorylation at S139 (γH2AX) following IR treatment in cells treated with BI2536, a Plk1 inhibitor (21). As shown in Fig. 1A and Supplementary Fig. S1A, BI2536 pretreatment significantly increased IR-induced γH2AX formation in U2OS cells. Next, U2OS cells were transfected with different forms of Plk1 (constitutively active T210D or kinase dead K82M mutant) and subjected to IR. Cells expressing Plk1-KM showed much high levels of γH2AX staining than cells expressing Plk1-TD mutant (Fig. 1B). Phosphorylation of Sgt1, a previously identified Plk1 substrate (22), was used as a control to indicate the kinase activity of Plk1. Furthermore, to understand how Plk1 regulates γH2AX formation due to IR, we sought to determine whether Plk1 is involved in the events upstream of H2AX phosphorylation. It has been established that the formation and expansion of γH2AX at DSB sites depends on activation of ATM kinase (23). Therefore, we asked whether there is a functional link between Plk1 and ATM during DDR by modulation of Plk1 activity. Pretreatment of cells with BI2536 significantly increased IR-induced phosphorylation, thus activation of ATM and its downstream target Chk2 (Fig. 1C; Supplementary Fig. S1B–S1D). Similarly, Plk1 knockdown by siRNA also markedly increased IR-induced phosphorylation and activation of ATM and Chk2 (Fig. 1D; Supplementary Fig. S1E). To further confirm the critical role of Plk1 in IR-induced activation of the ATM/Chk2 pathway, we also compared DDR in cells expressing different forms of Plk1. Upon IR, cells expressing Plk1-T210D showed reduced levels of phospho-ATM and phospho-Chk2 than cells expressing Plk1-K82M (Fig. 1E; Supplementary Fig. S1F and S1G). Collectively, these data support the notion that Plk1-associated kinase activity antagonizes IR-induced activation of ATM and its downstream targets Chk2 and H2AX.
Plk1 phosphorylates Mre11 at S649
To dissect how Plk1 is involved in DDR, we first used the Xenopus egg extract system, in which DDR signaling can be reconstituted in vitro (24). Extracts prepared from Xenopus egg were treated with different forms of Plk1 proteins. Consistent with the findings in intact cells, pretreatment with constitutive active Plk1 markedly inhibited ATM autophosphorylation induced by dsDNA without altering ATM levels (Fig. 2A). Because the MRN complex is recruited to DNA DSBs before ATM activation by autophosphorylation, we tested the hypothesis that Plk1 prevents ATM activation via its direct phosphorylation towards the MRN complex. Accordingly, DNA-bound proteins were isolated from the egg extract, followed by in vitro kinase assay with purified Plk1 ± BI2536. As indicated in Fig. 2B, a protein with the size of 85 kDa, the molecular weight of Mre11, was phosphorylated by Plk1, suggesting that Mre11 might be a direct target of Plk1. In a separate search for novel Plk1-interacting proteins using mass spectrometry, we also identified Mre11 as a potential Plk1 substrate (25). To directly test whether Mre11 is a substrate of Plk1, a series of in vitro kinase assays were performed with purified Plk1 and various recombinant glutathione S-transferase (GST)-fused Mre11 regions in the presence of [γ-32P]ATP. As indicated, Plk1 directly phosphorylates the very C-terminal region of Mre11-aa 623-709 (Fig. 2C). Furthermore, site-directed mutagenesis was used to identify S649 and S688 as two major phosphorylation sites in vitro (Fig. 2D). To confirm the kinase/substrate relationship, two phospho-specific antibodies (pS649 and pS688) were generated. A series of control experiments were performed to characterize the specificity of the two antibodies. Purified GST-Mre1-aa 623-709 (WT, S649A, S688A) was incubated with Plk1 in the presence of unlabeled ATP, followed by anti-pS649 and anti-pS688 IB. As indicated, S649A and S688A mutations completely abolished the respective anti-pS649 and anti-pS688 Western signals, indicating that the antibodies recognize only phosphorylated forms of Mre11 at those two sites (Fig. 2E). Next, HEK293T cells expressing GFP-Mre11 (WT, S649A, S688A) were analyzed by immunoblotting (IB) with the two antibodies. Both pS649 and pS688 signals were detected in cells expressing Mre11-WT, but not in cells expressing Mre11-S649A or -S688A, respectively (Fig. 2F), indicating that both sites are phosphorylated in vivo. To test whether Plk1 is the kinase responsible for the two phosphorylation sites in vivo, Plk1 inhibitor BI2536 was used to inhibit Plk1 kinase activity. The pS649, but not the pS688, epitope of endogenous Mre11 in 293T cells was substantially reduced when cells had been pretreated with BI2536 (Fig. 2G). Conversely, phosphorylation of S649 and S688 of Mre11 significantly increased as cells enter mitosis in a well-synchronized culture, matching the expression level of Plk1 (Fig. 2H). Next, we asked whether other mitotic kinases are also responsible for these phosphorylation events. Accordingly, cells were arrested at mitosis by nocodazole and harvested after incubation with different protein kinase inhibitors (Supplementary Fig. S2A). Treatment with three Plk1 inhibitors, BI2536, BI6727, or GSK461364, all impaired the phosphorylation of Mre11-S649. In contrast, the phosphorylation level of Mre11-S649 was not affected by treatment with Aurora A inhibitor I or RO-3306, a Cdk1/2 inhibitor, suggesting that endogenous Mre11 is phosphorylated at S649 in a Plk1-dependent manner.
CK2 is responsible for Mre11-S688 phosphorylation in vivo
Although Plk1 phosphorylates Mre11-S688 in vitro (Fig. 2D), inhibition of Plk1 did not significantly affect Mre11-S688 phosphorylation in vivo (Fig. 2G). It has been established that both Plk1 and CK2 prefer a target sequence containing negatively charged residues nearby and that two kinases share the same sites in some substrates (26, 27). Therefore, we examined the ability of CK2 to phosphorylate Mre11-S688. In vitro kinase assays revealed efficient phosphorylation of GST-Mre11-WT, but not GST-Mre11-S688A, by CK2 (Fig. 2I). Notably, inhibition of CK2 activity with its inhibitor TBCA greatly reduced the pS688 epitope in 293T cells (Fig. 2J), indicating that CK2 is the major kinase responsible for Mre11-S688 phosphorylation. To further confirm these results, RNAi-mediated Plk1 or CK2 depletion was performed. As indicated, depletion of Plk1 abolished the pS649 epitope, but not the pS688 epitope. In contrast, the pS688, but not the pS649 epitope, was substantially reduced when CK2 was depleted (Fig. 2K). Thus, we concluded that Plk1 phosphorylates Mre11-S649 and that CK2 targets Mre11-S688 in vivo.
Phosphorylation of Mre11 by Plk1 and CK2 is required for G2 DNA damage checkpoint recovery
Next, to examine whether the effect of Plk1-associated kinase activity on IR-induced activation of ATM and its downstream targets Chk2 and H2AX is dependent on Mre11-S649 phosphorylation, we reconstituted Mre11-depleted cells with Mre11-S649A mutant. We found that complementation with Mre11-S649A mutant in Mre11-depleted cells blocked BI2536-induced IR sensitivity (Supplementary Fig. S2B and S2C). Because Mre11 phosphorylation at S649/S688 is clearly cell cycle regulated (Fig. 2H), we asked whether Mre11 is involved in normal mitotic progression by Mre11 knockdown using siRNA targeting the 3′ UTR. As a control, we found that Plk1-depleted cells accumulated with a 4N DNA content, indicating a G2–M arrest of these cells. In contrast, no difference in FACS profile or cyclin B level was observed upon Mre11 depletion (Supplementary Fig. S3A), indicating that loss of Mre11 does not interfere with normal mitotic progression. Because Plk1 plays a central role in G2 DNA damage checkpoint recovery and because Mre11 is an initial DNA damage response protein, we next tested whether Mre11-S649/S688 phosphorylation affects G2 DNA damage checkpoint recovery. Toward this end, we synchronized U2OS cells at the G1–S boundary using the DTB and then allowed cells to progress through the cell cycle. At 7 hours after release, a time at which the majority of cells had completed S-phase (Supplementary Fig. S3B), cells were treated with IR to induce DNA damage during G2 phase (Fig. 3A). Strikingly, we found that treatment of cells with IR had reduced Mre11 phosphorylation at S649 and S688 without affecting the level of Mre11 (Fig. 3B). Of note, phosphorylation of Plk1-T210 and Sgt1 was inhibited by IR, indicating that Plk1 is inactivated upon DNA damage (Fig. 3B). Subsequently, we mimicked checkpoint silencing by addition of ATM/ATR checkpoint kinase inhibitor caffeine to override the DNA damage checkpoint thus to follow checkpoint recovery. As shown in Fig. 3C, DNA damage induced by doxorubicin blocked the cells into mitosis and treatment with caffeine drove the cells to reenter mitosis. In contrast, Plk1 inhibition with BI2536 prevented caffeine-induced mitotic reentry, suggesting that Plk1 activity is required for G2 DNA damage checkpoint recovery (Fig. 3C; Supplementary Fig. S3C). Similarly, we found that levels of pS649 and pS688 were considerably reduced in the presence of doxorubicin compared with untreated samples (Fig. 3D). In contrast, enhanced phosphorylation at S649/S688 was detected during the caffeine-induced checkpoint recovery, but remained to be either unphosphorylated at S649 or weakly phosphorylated at S688 in the presence of BI2536 (Fig. 3D), suggesting that the phosphorylation of S649/S688 during the G2 checkpoint recovery is Plk1 dependent. This result is consistent with the outcome described above in unperturbed cell, where the level of pS688 was slightly reduced upon Plk1 depletion (Fig. 2K), indicating that Plk1 is also involved in phosphorylation of Mre11 at S688. We also noticed that the pS688 epitope was apparently reduced in Mre11-S649A mutant background in vitro (Fig. 2E) and in vivo (Fig. 2F), and that the intensity of the pS688 epitope in cells expressing Mre11-S649D obviously increased (Supplementary Fig. S3D). Of note, several studies previously indicated that Plk1 and CK2 can act in concert to regulate one substrate (10, 26). Therefore, we evaluated whether Plk1 phosphorylation at S649 might modify the efficiency of CK2-mediated phosphorylation at S688 by using of sequential kinase assays. We first incubated GST-Mre11-aa 623-709 immobilized on glutathione beads with Plk1 or CK2, and after washing off the kinase, a second kinase reaction was carried out. Strikingly, prephosphorylation of Mre11 with Plk1 substantially facilitated subsequent phosphorylation by CK2, indicated by a significantly increased level of the pS688 epitope. In contrast, CK2 prephosphorylation of Mre11 did not alter the efficiency of Plk1-dependent phosphorylation of Mre11 at S688 (Fig. 3E). These results support the notion that during DNA damage checkpoint recovery Mre11 is sequentially modified by Plk1-dependent priming phosphorylation at S649, followed by CK2-mediated phosphorylation at S688.
If the observed mitotic phosphorylation of Mre11 is critical for the G2 DNA damage checkpoint recovery, we would expect to observe altered kinetics of G2–M transition when the phosphorylation site mutants of Mre11 are expressed under genotoxic condition. To directly examine whether Mre11-S649/S688 phosphorylation is involved in checkpoint recovery, endogenous Mre11 was replaced with RNAi-resistant Mre11 constructs (WT, S649A or S688A), and subjected to DNA damage and recovery process as above. Cells expressing Mre11-WT were able to recovery from G2 arrest (Fig. 3F, left). In contrast, cells expressing Mre11-S649A or -S688A showed recovery defects (Fig. 3F, middle and right). Furthermore, the levels of pS649 and pS688 were considerably reduced in the presence of doxorubicin compared with untreated samples. In contrast, significantly enhanced phosphorylation at S649/S688 was detected during the caffeine-induced checkpoint recovery (Fig. 3F, left). Interestingly, Mre11-S649A mutation apparently blocked caffeine-induced phosphorylation of S688 (Fig. 3F, middle), further confirming that the priming phosphorylation of S649 by Plk1 is required for the subsequent phosphorylation of S688 by CK2 during the checkpoint recovery. Although Mre11-S688A mutation did not affect caffeine-induced phosphorylation of S649 by Plk1, cells expressing Mre11-S688A failed to enter mitosis, arguing that CK2-mediated phosphorylation at S688 is also critical for checkpoint recovery (Fig. 3F, right). Consistent with this observation, inhibition of CK2 activity partially prevented caffeine-induced mitotic reentry after DNA damage (Supplementary Fig. S3E). Because the polo-box domain (PBD) has been responsible for interacting with most of known Plk1 substrates, we also performed co-immunoprecipitation (co-IP) analysis for GFP-Mre11 with different domains of Flag-Plk1. As indicated, the PBD, but not the kinase domain of Plk1, interacts with Mre11 (Supplementary Fig. S3F).
Mre11 phosphorylation at S649/S688 inhibits its binding to dsDNA and antagonizes the ATM signaling
Mre11 has two potential DNA-binding-domain (DBD) with S649/S688 located within the second DBD (Supplementary Fig. S4A; ref. 28). Therefore, we speculated that Mre11-S649/S688 phosphorylation might modulate its ability to load onto DNA. To explore this possibility, we analyzed DNA binding of different forms of recombinant Mre11 (WT, SA, SD) following incubation of Xenopus egg extracts with biotinylated linear DNA. As shown in Fig. 4A, substitution of S649/S688 with alanine resulted in a significant increase in Mre11 binding to DNA. In contrast, the DNA-binding ability of Mre11 was almost completely abolished upon introduction of aspartic acid mutations in S649 and S688 sites (Fig. 4A). Furthermore, pretreating recombinant WT GST-Mre11 with Plk1 and CK2 reduced the amounts of DNA-bound Mre11, but this effect was not observed when the S649A/S688A mutant was used (Fig. 4B). To further confirm that Mre11-S649/S688 phosphorylation regulates the Mre11/DNA binding, we performed electrophoretic mobility shift assay (EMSA) in a cell-free system. As shown in Fig. 4C and Supplementary Fig. S4B, the binding of Mre11 to the 3′ overhang dsDNA was inhibited in Mre11-S649D, Mre11-S688D, and Mre11-S649D/S688D mutants, suggesting that the phosphorylation of Mre11 at S649 and S688 directly disrupts the Mre11/DNA complex. To validate these findings in intact cells, we examined the effect of Mre11 phosphorylation on the MRN complex recruitment to the DSBs. Immunofluorescence analysis showed that reconstitution of Mre11-depleted cells with Mre11-WT, but not the -S649D/S688D mutant, restores MRN complex recruitment to DSBs (Fig. 4D–F). Therefore, Mre11 phosphorylation at S649/S688 prevents the recruitment of the MRN complex to DSBs, likely due to reduced DNA-binding ability of phosphorylated Mre11 (Fig. 4A–C).
Next, we asked whether the phosphorylation events affect the ATM signaling in response to DNA damage. As expected, DNA damage-induced activation of the ATM–Chk2 pathway was completely abolished upon Mre11 depletion (Fig. 4G, lanes 4–6). Reconstitution of Mre11-depleted cells with RNAi-resistant Mre11-WT completely rescued the phenotype (Fig. 4G, lanes 7–9). Of significance, cells expressing Mre11-S649D, Mre11-S688D, or Mre11-S649D/S688D mutants in the absence of endogenous Mre11 failed to activate the ATM/Chk2 pathway after DNA damage (Fig. 4G, lanes 10–12; Fig. 4H, lanes 2–4). To ensure that the finding is not IR-specific, we repeated the experiment with doxorubicin. Again, cells expressing Mre11-WT and various S649/S688 to alanine mutants, but not S649/S688 to aspartic acid mutants, responded to doxorubicin to activate ATM and its targets Chk2 and H2AX (Fig. 4H).
To rule out the possibility that the reduced DNA binding of Mre11-SD mutants could result from an unstable MRN complex, we examined integrity of the MRN complex in cells expressing different forms of Mre11 (WT, SA, SD). 293T cells were transiently transfected with siRNA to deplete Mre11. Notably, silencing Mre11 also resulted in a significant reduction in protein levels of Rad50 and Nbs1 (Supplementary Fig. S4C, compare lanes 1 and 2), which is consistent with a reduced expression of Nbs1 and Rad50 in the Mre11-deficient ATLD cells (29). This is because of a critical role of Mre11 to stabilize the MRN complex (30). In addition to the reduction of protein levels of Rad50 and Nbs1, there was also a reduction in Chk2 phosphorylation following IR when Mre11 was silenced (Supplementary Fig. S4C, compare lanes 3 and 4). Then, we reintroduced RNAi-resistant GFP-Mre11 constructs (WT, -2A, -2D) and monitored MRN complex stability. As indicated in Supplementary Fig. S4D, expression of all three forms of GFP-Mre11 (WT, -2A, -2D) rescued the normal levels of Rad50 and Nbs1, ruling out the possibility that the MRN complex might be unstable in cells expressing Mre11-SD mutants. Furthermore, we examined the interactions between different Mre11 variants and Rad50/Nbs1 and found that Mre11 mutants (2A, 2D) still coimmunoprecipitate with Rad50 and Nbs1 (Supplementary Fig. S4E and S4F). To further investigate the potential role of Mre11-SD mutations in the MRN complex formation, we carried out IF studies. The HCT116 cells, which were Mre11 deficient, were transiently transfected with different forms of GFP-Mre11 mutants. IF analysis indicated that different Mre11 mutants can restore Nbs1 protein level (Supplementary Fig. S4G). Collectively, we concluded that phosphorylation of Mre11 by Plk1 and CK2 does not affect the formation of the MRN complex.
Mre11-S649/S688 phosphorylation inhibits DNA repair
DSBs in eukaryotes are repaired by either HR or NHEJ, in which Mre11 is involved both. We examined the impact of Mre11-S649/S688 phosphorylation on the HR and NHEJ by utilizing two well-characterized GFP-based reporter systems: DR-GFP and EJ5-GFP. In brief, DR-GFP-HeLa or EJ5-GFP-HeLa cells were cotransfected with I-Scel–expressing vector and different Mre11 constructs (WT or S649D/S688D), and harvested for quantification of the GFP-positive population. Consistent with the previously reports, we found that Mre11 knockdown significantly decreased the fraction of GFP-positive cells (Supplementary Fig. S4H). Interestingly, much higher DNA repair events were detected in cells expressing Mre11-WT compared with cells expressing Mre11-S649D/S688D (Fig. 5A; Supplementary Fig. S4H), suggesting that phosphorylation of Mre11-S649/S688 inhibits DNA repair. Consistent with this observation, cells expressing Mre11-2D showed a markedly increased level of DNA damage than cells expressing Mre11-WT after 20 hours of repair incubation (Fig. 5B). Furthermore, cells expressing Mre11-2D also showed enhanced sensitivity to IR, indicated by both reduced cell survival (Fig. 5C) and colony-forming ability (Fig. 5D).
Considering that Mre11-defective cells had profound sensitivity to inhibitors of PARP (31), we also investigated whether Mre11-S649/S688 phosphorylation affects the cellular response to olaparib, a PARP inhibitor. Remarkably, cells expressing Mre11-WT and Mre11-2A variant exhibited significantly higher resistance to olaparib compared with cells expressing Mre11-2D (Fig. 5E). Next, we compared the efficacy of olaparib on xenograft tumors derived from U2OS cells expressing different forms of Mre11 (WT, 2A, 2D). Consistent with the in vitro data (Fig. 5E), tumors derived from Mre11-2D–expressing cells were much more sensitive to olaparib than tumors derived from cells expressing Mre11-WT or -2A mutant (Fig. 5F–H).
In this study, we show that (i) Plk1 phosphorylates Mre11 at S649 during G2 DNA damage recovery; (ii) Plk1-mediated phosphorylation stimulates subsequent phosphorylation of Mre11 at S689 by CK2; and (iii) Mre11 phosphorylation at S649/S689 drives premature checkpoint termination and reduced DNA repair, thus likely contributing to IR-associated carcinogenesis.
It has been established that Plk1 is a key player of checkpoint recovery by phosphorylating several checkpoint factors, leading to their inactivation (14). The activity of Plk1 is required for inactivation of both ATM/Chk2 and ATR/Chk1 pathways. While Plk1 phosphorylation of claspin leads to its degradation via SCFβTRCP, thus inactivating the ATR/Chk1 pathway (13), the ATM–Chk2 pathway is also inactivated by Plk1 due to its phosphorylation of 53BP1 and Chk2 (14). More importantly, Plk1 is a predictive marker and therapeutic target for radiotherapy in rectal cancer (32). Despite these progress, whether Plk1 affects molecules upstream of ATM/ATR remains unknown. Herein, we show that Mre11 is a direct target of Plk1 and CK2 during the checkpoint recovery process. On the basis of our data, we propose the following working model. Upon DSB induction, the MRN complex is efficiently recruited to the site of damage and initiates damage response. Processed DNA lesions lead to the recruitment and activation of ATM, which in turn phosphorylates its substrates, among them Bora. Phosphorylated Bora is degraded in a SCFβTrCP-dependent manner, resulting in Plk1 inactivation and cell-cycle arrest at G2. During the subsequent recovery process, Plk1 phosphorylation of Mre11-S649 facilitates subsequent phosphorylation of Mre11-S688 by CK2, leading to MRN complex release from DNA and inactivation of both ATM–Chk2 and ATR–Chk1 pathways. In Plk1-overexpressing cells, Plk1 signaling impairs DDR by inhibiting the ATM pathway, likely due to reduced recruitment of the MRN complex to DNA damage. Considering that radiation is one major carcinogenetic factor, our data clearly demonstrate that the level/activity of Plk1 plays a critical role in IR-induced carcinogenesis for two reasons. First, elevation/activation of Plk1 causes premature termination of DNA damage checkpoint, as we showed that Plk1 phosphorylation of Mre11 abolishes its binding ability to chromatin in response to DNA damage. Second, elevation/activation of Plk1 contributes to reduced DNA repair due to its phosphorylation toward Mre11. Altogether, a combination of premature termination of DNA damage checkpoint and reduced DNA repair will accelerate IR-associated carcinogenesis, suggesting that inhibition of Plk1 might be useful in prevention of IR-induced carcinogenesis.
It was previously described that Mre11 was phosphorylated at S649 both in vitro and in vivo, and in vitro kinase assay showed that Mre11-S649 was phosphorylated by CK2 (33). The unanswered questions were whether Mre11 S649 can be phosphorylated by CK2 in vivo and whether Mre11-S649 phosphorylation affects its function. It was established that both Plk1 and CK2 share the same sites in some substrates in vitro (26, 27). Here we demonstrated that Mre11 was phosphorylated at S649 by Plk1 both in vitro and in vivo, and that depletion of Plk1, but not CK2, impaired S649 phosphorylation, supporting the notion that Plk1 is the major kinase that phosphorylates Mre11 at S649 in vivo. Finally, it was shown that phosphorylation of the C terminus of Mre11 was generally inhibitory to its functions (34). Consistent with this notion, we demonstrated that phosphorylation of Mre11-S649 impaires its binding to damaged DNA and suppresses DNA damage response pathway. Therefore, we identified a novel molecular mechanism that links Plk1, the MRN complex, and DDR.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: Z. Li, N. Ahmad, X. Liu
Development of methodology: Z. Li, S. Yan
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Z. Li, J. Li, X. Liu
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Z. Li, J. Li, Y. Kong, N. Ahmad, X. Liu
Writing, review, and/or revision of the manuscript: Z. Li, N. Ahmad, X. Liu
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Z. Li, Y. Kong, S. Yan, X. Liu
Study supervision: Z. Li, X. Liu
This work was supported by NIH grants R01 CA157429 to X. Liu, R01 CA192894 to X. Liu, R01 CA196835 to X. Liu, R01 CA196634 to X. Liu, R01 AR059130 to N. Ahmad, R01 CA176748 to N. Ahmad, and R15 GM114713 to S. Yan. The work was also partially supported by Purdue University Center for Cancer Research (P30 CA023168).