Chlamydia trachomatis Inhibits Homologous Recombination Repair of DNA Breaks by Interfering with PP2A Signaling

Chlamydia trachomatis induces DNA double-strand breaks in host cells but simultaneously inhibits proper DNA damage response and repair mechanisms. This may render host cells prone to loss of genetic integrity and transformation. Here we show that C. trachomatis prevents activation of the key DNA damage response mediator ATM by preventing the release from PP2A, leading to a complete absence of homologous recombination repair in host cells.

I n recent years, several epidemiological studies have implicated infections with the Gram-negative bacterial pathogen Chlamydia trachomatis in the development of cervical (1,2) and ovarian (3) carcinomas. These cancers together show eight distinct validated mutational signatures (http://cancer.sanger.ac.uk/cosmic/signatures), one of which is attributed to defective homologous recombination (HR) repair. How these mutational processes are initiated during the course of carcinogenesis is largely unknown. It is thus intriguing that C. trachomatis, an obligate intracellular bacterium causing chronic, asymptomatic infections, promotes DNA double-strand breaks (DSBs) and modulates a range of host cellular functions and signal transduction pathways, including those involved in preserving cellular and genomic integrity (4)(5)(6)(7)(8)(9)(10) and immune activation and apoptosis induction (11).
DSBs represent the most dangerous form of damage, as they cannot always be correctly repaired and thus pose the risk of genomic instability and chromosomal rearrangements (12). The DNA damage response (DDR) in mammalian cells prevents the accumulation of mutations and genomic instability (12,13). Ataxia-telangiectasia mutated (ATM), an apical activator induced by DSBs, plays a critical role in relaying a strong, widespread signal to numerous downstream effectors involved in multiple processes (14), including repair, cell cycle, and cell death (15). Activation of ATM can occur by direct interaction with single-stranded DNA or oligonucleotides at DSBs, as well as by a direct interaction with the MRE11-RAD50-NBS1 (MRN) complex, whose members are themselves ATM substrates (16). Thus, a positive-feedback loop between ATM and its substrate proteins seems to be required for its correct positioning at break sites.
In our previous study, we showed that Chlamydia induces DSBs but simultaneously suppresses the activation and recruitment of ATM and MRE11 to the damage sites (8,17). However, the molecular mechanisms by which Chlamydia suppresses activation of ATM signaling in the face of extensive DSBs and its consequences for the function of the error-free HR repair pathway remain unknown. Phosphorylation and dephosphorylation of proteins appear to be crucial for activating the DDR within minutes of DNA damage (18), suggesting a prime role for protein phosphatases in regulating the DDR (16). Protein phosphatase 2A (PP2A), a serine/threonine phosphatase, has been implicated in regulation of ATM activity in response to radiation-induced DSBs (19). PP2A holoenzymes are heterotrimers consisting of a core dimer scaffold (A) and a catalytic (C) subunit that is associated with one of the regulatory (B) subunits. Posttranslational modification in the C-terminal part of the catalytic subunit regulates the phosphatase activity of PP2A. Phosphorylation of tyrosine residue 307 (Y307) on the C subunit results in decreased PP2A enzyme activity (20,21). Upon the formation of irradiation-induced DSBs, the PP2AC-B55␣ regulatory subunit of PP2A, which normally facilitates association with ATM, rapidly dissociates, leading to ATM autophosphorylation and activation (19,22).
Here, we addressed the involvement of PP2A in the failure to mount an adequate response to DSBs in Chlamydia-infected cells. Interestingly, PP2A activity remained high and exhibited a persistent interaction with ATM, keeping it in an inactive state despite the DSBs. This is in line with data indicating a remarkable loss of HR repair function in C. trachomatis-infected cells. Treatment with okadaic acid (OA), an inhibitor of PP2A, led to increased phosphorylation of Y307 on the PP2A catalytic subunit and released ATM from PP2A, resulting in ATM phosphorylation. The activated ATM was then recruited to chromatin and initiated downstream signaling, leading to checkpoint kinase 2 (Chk2)mediated G 2 /M cell cycle checkpoint activation and restoration of HR DNA repair. Taking the data together, this study revealed an intriguing mechanism by which Chlamydia modulates host signaling to support its intracellular development. By inhibiting ATM signaling, this pathogen inactivates an essential high-fidelity HR pathway and predisposes infected cells to mutagenesis.

RESULTS
Chlamydia trachomatis-induced ATM inactivation is mediated by protein phosphatase 2A. Irradiation-induced DSBs elicit PP2A-mediated ATM activation (18). However, C. trachomatis infection suppresses the phosphorylation-mediated activation of ATM despite induction of extensive DSBs ( Fig. 1A and B). Here we investigated the role of PP2A in regulating ATM suppression after the formation of C. trachomatis-induced DSBs. Treatment of C. trachomatis-infected cells with the protein phosphatase inhibitor okadaic acid (OA) at different concentrations (19) for the last 2 h or for different periods of time during infection led to increased phosphorylation of ATM at Ser1981 (pATM) in a concentration-and exposure time-dependent manner compared to untreated infected cells ( Fig. 1C and D). While OA treatment itself did not induce DSBs, OA treatment of C. trachomatis-infected cells led to enhanced induction of DSBs, as demonstrated by neutral comet assay (Fig. 1E). However, no increase in the levels of phospho-H2AX (S139) (␥H2AX), a hallmark of DSBs, was observed as demonstrated by immunoblot analysis (Fig. 1F). Further, to analyze the specific role of PP2A in the ATM inhibition observed in C. trachomatis-infected cells, small interfering RNA (siRNA)mediated knockdown (KD) of the PP2A catalytic alpha subunit PPP2CA was performed. PPP2CA KD led to ATM activation in C. trachomatis-infected cells ( Fig. 1G and H) in contrast to C. trachomatis-infected cells transfected with a neutral control (siLuci). Thus,   PP2A plays a key role in the suppression of ATM activation after C. trachomatis induction of DSBs. Dynamic interaction of ATM and PP2A persists despite the presence of Chlamydia-induced DNA double-strand breaks. Phosphorylation of PP2A catalytic subunit C at Y307 (PP2A-C pY307) inhibits binding of the regulatory subunit B, an essential component for interaction with target proteins. To evaluate whether functionally active PP2A is directly responsible for inhibiting ATM activation, we analyzed the level of PP2A-C pY307 in C. trachomatis-infected cells with and without OA treatment. The levels of pY307 remained unaffected in response to C. trachomatis infection or treatment with FTY720, an immunomodulator that activates PP2A, which served as a positive control ( Fig. 2A). Interestingly, a dramatic increase in PP2A-C pY307 levels in C. trachomatis-infected cells treated with OA was observed, indicating that PP2A was present in its active form in the C. trachomatis-infected cells. Further, we performed an in situ proximity ligation assay (PLA) using specific antibodies against PP2A and ATM to visualize protein-protein interactions. We observed a stronger interaction between PP2A and ATM in C. trachomatis-infected cells than in uninfected cells. However, additional treatment of C. trachomatis-infected cells with OA led to Hsp60 β-Actin reduced interactions of PP2A and ATM, similarly to what is observed after etoposideinduced DSBs ( Fig. 2B and C). Taking the data together, PP2A remains active in C. trachomatis-infected cells and continues to interact with ATM, thus maintaining it in a dephosphorylated state despite the presence of DSBs. The G 2 /M cell cycle checkpoint is activated in C. trachomatis-infected cells upon PP2A inhibition. ATM acts as a dual-function protein in DNA damage repair as well as in transduction of signals eliciting cell cycle checkpoints to provide enough time for DNA repair and/or apoptosis induction, depending on the extent of DNA damage (8,15,23). Here we assessed whether activation of ATM following PP2A inhibition restores the functional cell cycle checkpoint in C. trachomatis-infected cells. To this end, cells infected for 45 h with or without additional treatment of OA during the last 6, 12, 20, or 24 h were subjected to immunoblot analysis for phosphorylation of the checkpoint protein Chk2, an effector of active ATM kinase. Levels of phosphorylated Chk2 increased in infected cells treated with OA in contrast to untreated infected cells, depending on the duration of treatment (Fig. 3A). Further, we analyzed whether activation of Chk2 protein would culminate in cell cycle arrest. Since C. trachomatis is an intracellular pathogen, cell cycle analysis based on DNA content using a fluorescence-activated cell sorting (FACS) approach remains suboptimal, as C. trachomatis DNA interferes with the quantification. For this reason, we used the novel and powerful FUCCI (fluorescence ubiquitin cell cycle indicator) cell system, which utilizes fluorescent proteins in combination with two components of the DNA replication control system of higher eukaryotes: the licensing factor Cdt1 and its inhibitor geminin. The levels of abundance of Cdt1 and geminin show inverse patterns during the cell cycle, with opposing effects on DNA replication. Cdt1 protein peaks in G 1 phase just before the onset of DNA replication and declines abruptly after S-phase initiation. In contrast, geminin levels are high during S and G 2 phase but are low during late mitosis and G 1 phase (24). These HeLa FUCCI cells (24), which exhibit green fluorescence during S/G 2 /M phase and red fluorescence during G 1 phase, in combination with additional immunostaining for S phase using bromodeoxyuridine (BrdU) antibody, enable the identification of cells that are in G 1 , S, or G 2 /M phase. Using automated image acquisition and analysis of FUCCI HeLa cells additionally labeled with BrdU and Hoechst, we found that cells infected with C. trachomatis proliferated without cell cycle arrest ( Fig. 3B and C). However, activation of ATM via OA treatment in infected cells led to increased accumulation of cells in G 2 /M phase. Representative images of C. trachomatisinfected HeLa FUCCI cells with or without OA (Fig. 3B) show increased levels of green fluorescent protein (GFP)-positive cells that occurred in an OA concentrationdependent manner (Fig. 3B). Further, quantification of GFP-positive and BrdU-negative cells indicated G 2 /M cell cycle arrest in response to PP2A inhibition (Fig. 3C). Cells treated with etoposide served as a positive control for DSB-induced G 2 /M cell cycle arrest. Further, we analyzed if the observed cell cycle checkpoint activation and enhanced G 2 /M arrest are dependent on ATM activation. To this end, C. trachomatisinfected cells were treated with OA alone or in combination with the ATM kinase inhibitor KU-55933 (ATMi) and analyzed for pChk2 by immunoblotting and for cell cycle profile using HeLa FUCCI cells. Increases in pChk2 levels, as well as increased proportions of cells in G 2 /M phase in infected cells treated with OA, are prevented when ATM activation is inhibited by ATMi (Fig. 3D and E). Thus, inhibition of PP2A rescues the C. trachomatis-imposed suppression of ATM activation-mediated cell cycle arrest.

PP2A impairs DNA repair by homologous recombination in Chlamydia-infected cells.
To investigate the efficiency of HR repair in the face of C. trachomatis-induced DSBs, we used the HEK293 DR-GFP reporter cell line. This cell line has an I-SceI site integrated into a full-length GFP gene (SceGFP), which allows the introduction of a single DSB via the rare-cutting endonuclease I-SceI, thus disrupting the GFP gene (25). Repair of this DSB by HR, using a downstream internal truncated GFP fragment as the template, results in a functional GFP gene. Using the HEK293 DR-GFP cell line, we further developed an automated microscopic assay using the ScanR system (Olympus) for simultaneous visualization and scoring of GFP-positive infected and uninfected cells from the same microscopic image. HEK293 DR-GFP cells were infected with C. trachomatis, and the numbers of GFP-positive cells were compared to the levels measured in uninfected cells. Interestingly, the numbers of GFP-positive cells in infected and uninfected cultures remained the same. However, even in the C. trachomatis-infected cultures, a GFP signal was observed only in cells that had remained noninfected, indicating an almost complete abrogation of HR in infected cells (Fig. 4). PP2A inhibi- tion led to recovery of HR in infected cells in a concentration-dependent manner ( Fig. 4A and 5B). Further, the restored functional HR is associated with the increased recruitment of activated ATM to the DSBs after PP2A inhibition in C. trachomatis-infected cells. Immunofluorescence analysis of pATM, ␥H2AX, and DNA in cells infected for 36 h with or without addition of OA during the last 20 h, or treated with etoposide for 2 h as a positive control, showed that, in contrast to etoposide-treated cells, C. trachomatisinfected cells lacked pATM foci at DSB sites (Fig. 5A). OA-treated cells showed increased localization of pATM to the damaged DNA after infection (Fig. 5A to C).
Proper HR repair depends on ATM-mediated phosphorylation of MRE11 S676 and S678, which is required for end resection by exonuclease 1, as well as S343 on MRN complex member NBS1 (16,26,27). Therefore, we used immunoblotting to analyze phosphorylation of MRE11 at S676 (pMRE11) and NBS1 at S343 (pNBS1) in C. trachomatis-infected cells with or without OA treatment. No increase in pMRE11 and pNBS1 levels could be observed compared to uninfected control cells. In contrast, C. trachomatis-infected cells treated with OA showed enhanced levels of pMRE11 and pNBS1 (Fig. 5D), in congruence with the increase in pATM levels described above (Fig. 1C and D and 5A). Taking the data together, our results show that C. trachomatis infection suppresses the HR repair response to DSBs as a consequence of PP2Amediated inactivation of ATM signaling, thereby leading to a predisposition to host cell mutagenesis.

DISCUSSION
This report reveals that Chlamydia, despite inducing DSBs, modulates a host regulatory inhibition mechanism to alter the usual DDR and homologous recombination repair factors (18). Phosphorylation of ATM on Ser1981 following DSB induction leads to its activation as an apical activator and mobilizer of DDR proteins (15,23). Failure of ATM activation can have severe effects on genomic stability. We have previously reported that C. trachomatis induces enhanced ROS production, which is essential for its development but as a consequence induces DSBs in the host DNA that fail to activate either ATM or MRE11 (8). C. trachomatis is also known to modulate mitogenactivated protein kinase (MAPK) signaling, contributing to aberrant cell proliferation despite the presence of DSBs (4,8).
Several studies have proposed alternative mechanisms for ATM activation, including direct interaction with single-stranded DNA or oligonucleotides at DSBs (16). However, these mechanisms clearly fail to elicit an ATM response in C. trachomatis-infected cells despite the presence of DSBs. While the role of kinases in the DDR has been well established, the complex roles of protein dephosphorylation by phosphatases are only just emerging. Dephosphorylation events have been implicated in keeping DDR factors inactive during normal cell growth (28,29) and in the inactivation of checkpoint arrest following DNA repair. Further, spatial and temporal regulation of dephosphorylation events mediated by phosphatases is crucial for the DDR (28).
Activation of ATM in response to irradiation-induced DSBs has been shown to require disruption of its constitutive association with PP2A (19). In contrast, our results show that C. trachomatis interferes with the normal PP2A response to DSBs, thus inhibiting the ATM-mediated DDR. ATM activity is further regulated by protein phosphatases PP1 and PP5, with PP1 inhibiting ATM activity (19,30), while PP5 activates it by an as-yet-unknown mechanism (31). However, we demonstrated that PP2A is specifically involved in suppressing ATM activation in C. trachomatis-infected cells, since siRNA-mediated knockdown of the catalytic alpha subunit of PP2A led to activation of ATM. Nevertheless, continued PP2A activity in C. trachomatis-infected cells might have additional effects on host cells, as it is known to play important roles in regulating multiple signaling pathways such as the phosphatidylinositol 3-kinase (PI3K), Wnt, and RAS pathways (32).
We investigated in more depth the mechanism by which C. trachomatis infection regulates PP2A. PP2A activity is regulated via posttranslational modifications of its catalytic subunit. Methylation of the catalytic subunit enhances activity by facilitating binding of the regulatory B subunit, an essential event for interaction with target proteins (20). In contrast, phosphorylation of the catalytic subunit PP2A-C at pY307 inhibits binding of the B subunit and interaction with the target protein. The phosphatase inhibitor OA enhances phosphorylation at PP2A-C pY307, potently inactivating PP2A (33). Here, we found that while C. trachomatis-infected cells show minimal phosphorylation of PP2A-C Y307, comparably to uninfected cells, OA treatment leads to inactivation of PP2A by increasing PP2A-C pY307 and dissociation of ATM, resulting in ATM phosphorylation.
Phosphorylated ATM modulates the activity of several downstream effector protein kinases, leading to the coordinated activation of the DNA repair machinery and cell cycle checkpoints (34). Checkpoint kinase 2 (Chk2) is a well-known effector of pATM (35). ATM activation following irradiation-induced DNA damage leads to Chk2 phosphorylation and G 2 /M cell cycle arrest (35,36). In C. trachomatis-infected cells, PP2Amediated suppression of ATM phosphorylation prevents Chk2 phosphorylation and cell cycle arrest. Since HR occurs only during the S and G 2 phases, when sister chromatids are available as templates, lack of ATM and the consequent checkpoint activation suppress HR and shift DSB repair to the error-prone non-homologous end joining (NHEJ) pathway, which repairs breaks during all phases of the cell cycle as it does not require sequence homology (14,37,38).
Once ATM is phosphorylated, it undergoes spatial relocalization to the nucleus, where it initiates repair by phosphorylating proteins that are recruited to DSBs. These form discrete foci of phosphorylated ATM substrates (such as histone H2AX, NBS1, and MRE11) (15). This recruitment of pATM is suppressed by C. trachomatis infection, but our findings show that it can be partially rescued by inactivating PP2A. As a conse-quence, ATM-mediated phosphorylation of MRE11 and NBS1 is also restored, allowing HR repair of DSBs to proceed. We have shown previously that the ATM substrate H2AX is phosphorylated by the NHEJ pathway effector DNA-dependent protein kinase catalytic subunit (DNA-PKcs) in response to C. trachomatis-induced DSBs in the absence of ATM activation (8). This suggests that the NHEJ repair pathway in C. trachomatisinfected cells is predominantly active. In congruence, infected cells exhibit defective repair of DSBs, leading to loss of nucleotides and mismatch incorporation of nucleotides (8). Furthermore, Chlamydia infection causes cytokinesis defects, leading to multinucleated host cells, increased supernumerary centrosomes, abnormal spindle poles, and chromosomal segregation defects (39,40) that could culminate in genomic instability. In the present study, we directly confirmed-using a reporter cell line-that HR does not take place in C. trachomatis-infected cells, as ATM activation is blocked by PP2A.
Thus, to promote cell cycle progression and survival for its propagation, C. trachomatis interferes with the usual function of PP2A, thereby suppressing the ATM-HR axis responsible for high-fidelity repair of DSBs. Further, the functions of PP2A and ATM are known not only to regulate the cellular response to DSBs but also to function in various branches of metabolism and cell signaling (41), indicating the magnitude of the effect on diverse signaling cascades in C. trachomatis-infected cells-thus leading to a predisposition to mutagenesis and loss of cellular homeostasis.
Chlamydia infections. C. trachomatis L2 (ATCC VR-902B) stocks were prepared as described earlier (4). Unless otherwise stated, Chlamydia infection experiments were performed at a multiplicity of infection (MOI) of 5 in infection medium (DMEM supplemented with 5% FCS, 2 mM glutamine, and 1 mM sodium pyruvate). The medium was exchanged at 2 h postinfection (p.i.), and cells were grown at 35°C in 5% CO 2 .
SDS-PAGE and Western blotting. Cells grown in six-well plates and treated per experimental requirements were washed with phosphate-buffered saline (PBS) and lysed with 300 l of SDS sample buffer (3% 2-mercaptoethanol, 20% glycerin, 0.05% bromophenol blue, 3% SDS). Cell lysates were collected and incubated at 95°C for 10 min. Samples were stored at Ϫ20°C until required. SDS-PAGE and Western blotting were performed as described earlier (4).
Indirect immunofluorescence microscopy. Indirect immunofluorescence microscopy was performed as described in reference 8. Briefly, cells incubated in CSK buffer {10 mM PIPES [piperazine-N,N=bis(2-ethanesulfonic acid)] (pH 6.8), 100 mM NaCl, 300 mM sucrose, 3 mM MgCl 2 , 1 mM EGTA, 0.5% TritonX-100} on ice for 10 min and then fixed with 2% paraformaldehyde (PFA) at room temperature for 30 min. Cells were washed with PBST (PBS-0.1% Tween 20) and then incubated in blocking buffer (3% bovine serum albumin [BSA]-PBST) for 30 min followed by 1 h of incubation at room temperature with primary antibodies. Cells were washed and incubated for 1 h at room temperature with the appropriate fluorochrome-conjugated secondary antibodies. Cells were also stained for DNA with Draq5-PBS or Hoechst stain-PBS for 5 min before being mounted with Mowiol. The fluorochromes were visualized with Cy2, Cy3, and Cy5 filters. A series of images with z-stacks were acquired with a laser scanning confocal microscope (Leica) and further processed with Image J and Photoshop CS3 (Adobe Systems). Integrated intensities were calculated using ImageJ software.
Proximity ligation assay. Proximity ligation assays were performed as previously described (4). Briefly, End1E67E7 cells grown on coverslips in 24-well plates were infected with C. trachomatis. Cells were washed twice with PBS and then fixed with ice-cold methanol overnight at 4°C. Incubation with antibodies against ATM and PP2A-A was performed with a Duolink In Situ Orange goat/rabbit starter kit (Sigma) according to the manufacturer's instructions. A series of images with z-stacks were acquired with a laser scanning confocal microscope (Leica), analyzed with ImageJ software, and further processed by Photoshop CS3 (Adobe Systems).
Cell cycle analysis using FUCCI HeLa cells. Fucci HeLa cells were seeded in 96-well plates (5,000 cells/well) and then infected with C. trachomatis L2 on the following day. Cells were then treated with OA with or without 10 M ATM inhibitor (KU-55933). Cells were treated with BrdU at a 1:1,000 dilution in 5% FCS-DMEM for 1 h before PFA fixation was performed followed by incubation with 2 M HCL for 20 min and then 0.1 M sodium borate (Na 2 B 4 O 7 ) (pH 8.5) for 2 min. The cells were permeabilized and stained with anti-BrdU (1:1,000) primary antibody followed by fluorochrome-conjugated secondary antibody and Hoechst stain. Finally, images were acquired by the use of an automated microscope and Cy2, Cy3, Cy5, and DAPI (4=,6-diamidino-2-phenylindole) filters, and data were analyzed by the use of ScanR software and a custom-developed image analysis assay.