ATR Contributes More Than ATM in Intra-S-Phase Checkpoint Activation after IR, and DNA-PKcs Facilitates Recovery: Evidence for Modular Integration of ATM/ATR/DNA-PKcs Functions

The intra-S-phase checkpoint was among the first reported cell cycle checkpoints in mammalian cells. It transiently slows down the rate of DNA replication after DNA damage to facilitate repair and thus prevents genomic instability. The ionizing radiation (IR)-induced intra-S-phase checkpoint in mammalian cells is thought to be mainly dependent upon the kinase activity of ATM. Defects in the intra-S-phase checkpoint result in radio-resistant DNA synthesis (RDS), which promotes genomic instability. ATM belongs to the PI3K kinase family along with ATR and DNA-PKcs. ATR has been shown to be the key kinase for intra-S-phase checkpoint signaling in yeast and has also been implicated in this checkpoint in higher eukaryotes. Recently, contributions of DNA-PKcs to IR-induced G2-checkpoint could also be established. Whether and how ATR and DNA-PKcs are involved in the IR-induced intra-S-phase checkpoint in mammalian cells is incompletely characterized. Here, we investigated the contributions of ATM, ATR, and DNA-PKcs to intra-S-phase checkpoint activation after exposure to IR of human and hamster cells. The results suggest that the activities of both ATM and ATR are essential for efficient intra-S-phase checkpoint activation. Indeed, in a wild-type genetic background, ATR inhibition generates stronger checkpoint defects than ATM inhibition. Similar to G2 checkpoint, DNA-PKcs contributes to the recovery from the intra-S-phase checkpoint. DNA-PKcs–deficient cells show persistent, mainly ATR-dependent intra-S-phase checkpoints. A correlation between the degree of DSB end resection and the strength of the intra-S-phase checkpoint is observed, which again compares well to the G2 checkpoint response. We conclude that the organization of the intra-S-phase checkpoint has a similar mechanistic organization to that of the G2 checkpoint in cells irradiated in the G2 phase.


Introduction
Cell cycle checkpoints are the caretakers of genomic stability [1][2][3]. They stop cell cycle progression following DNA damage, transiently or permanently, and thus provide time for repair [2,4,5] before the cell progresses further into the cell cycle; alternatively, checkpoints drive cells into programmed cell death or senescence. Failure to activate checkpoints leads to genomic instability through associated failures in replication or repair [6]. Replicationdependent radiosensitization is among the most promising approaches towards improving and most importantly the interplay among ATM, ATR, and DNA-PKcs in the IR-induced intra-S-phase checkpoint.
We report that in contrast to the G2 checkpoint response of cells irradiated in the S phase, which is solely dependent on ATR, the intra-S-phase checkpoint, i.e., the checkpoint activated in S-phase cells that have also been irradiated in the S phase, is dependent on the activities of both ATM and ATR. This response is similar to the G 2 checkpoint-activated in cells irradiated in the G 2 phase. Similar to the IR-induced G 2 checkpoint, DNA-PKcs is crucial for the recovery from the intra-S-phase checkpoint, which is also in agreement with our previous findings [32]. Interestingly, ATR dominates over ATM in signaling persistent intra-S-phase checkpoint in DNA-PKcs deficient or inhibited cells. Similar to the IR-induced G 2 checkpoint, the magnitude of the intra-S-phase checkpoint correlates with the degree of resection. Inhibition of ATM or ATR individually, significantly reduces resection, and combined inhibition abolishes resection. Ablation, but not chemical inhibition, of DNA-PKcs exerts a significant increase in resection upon irradiation, and this feeds directly into the potentiation of the intra-S-phase checkpoint.

The Activation of the Intra-S-Phase Checkpoint Is Strongly Dependent on the Functions of Both ATM and ATR
The impact of ATM and ATR on intra-S-phase checkpoint was first measured in wildtype Chinese hamster ovarian (CHO) cells. CHO cells offer an attractive model for DNA damage response (DDR) studies owing to the availability of relevant DSB repair mutants, which are employed in this study and are described later. Exponentially growing CHO cells were exposed to 10 Gy of IR in the presence or absence of ATM inhibitor (ATMi), and/or ATR inhibitor (ATRi), and DNA synthesis was measured at the indicated times after IR, using tritiated thymidine ( 3 H) incorporation into DNA. The reduction in 3 H incorporation reflects inhibition of DNA synthesis owing to the activation of the intra-S-phase checkpoint.
CHO cells show, at 1 h post-IR, about 40% reduction in DNA synthesis. DNA synthesis recovers eventually and reaches~80% of controls within 3 h post-IR ( Figure 1A). ATM inhibition partially abrogates the checkpoint that now causes only a 20% reduction in DNA synthesis, which remains relatively unchanged until 3 h ( Figure 1B). ATR inhibition strongly abrogates the checkpoint, with only~10% transient reduction in DNA synthesis at 0.5 h and a quick recovery to control levels at 1 h.
Indeed, an overshooting in DNA synthesis is noted at 3 h post-IR ( Figure 1B). The combined inhibition of ATM and ATR results in nearly complete abrogation of the intra-S-phase checkpoint at all three levels: initiation, maintenance, and recovery ( Figure 1C). Since ATR and its downstream target Chk1 have been shown to affect origin firing during unperturbed DNA replication [33][34][35][36][37], we additionally checked the impact of ATRi on DNA synthesis in unirradiated cells. A slight increase in DNA synthesis is observed at 1 and 3 h after ATRi treatment, when compared to 0.5 h (data not shown). This increase might explain the overshooting observed in DNA synthesis at 3 h after IR.
In addition to CHO cells, we investigated the impact of ATM and ATR on the intra-S-phase checkpoint in the human lung carcinoma cell line, A549. A549 cells show upon irradiation a robust checkpoint (~70% reduction in DNA synthesis) at 30 min post-IR ( Figure 2A). The checkpoint remains strong until 1 h, with clear evidence of recovery at 3 h post-IR. Here again, ATM inhibition partially suppresses the checkpoint (Figure 2A), yielding only about a 40% reduction in DNA synthesis. ATR inhibition again exerts a stronger checkpoint abrogation ( Figure 2B), resulting in only~20% reduction in DNA synthesis. The combined inhibition of ATM and ATR abrogates the checkpoint to a similar extent ( Figure 2C) as the inhibition of ATR alone. The above set of data, collectively, suggests that both ATM and ATR kinases contribute to the activation of the IR-induced intra-S-phase checkpoint in rodents as well as in human cells, with ATR exerting a slightly stronger effect than ATM (see below for a model of the mechanism).  ATRi. Error bars represent the standard deviation from 3 independent experiments. The significance of differences between individual measurements is indicated by * symbol: * p < 0.05, ** p < 0.01, *** p < 0.001; n.s. non-significant.
In addition to CHO cells, we investigated the impact of ATM and ATR on the intra-S-phase checkpoint in the human lung carcinoma cell line, A549. A549 cells show upon irradiation a robust checkpoint (~70% reduction in DNA synthesis) at 30 min post-IR (Figure 2A). The checkpoint remains strong until 1 h, with clear evidence of recovery at 3 h post-IR. Here again, ATM inhibition partially suppresses the checkpoint (Figure 2A), yielding only about a 40% reduction in DNA synthesis. ATR inhibition again exerts a stronger checkpoint abrogation ( Figure 2B), resulting in only ~20% reduction in DNA synthesis. The combined inhibition of ATM and ATR abrogates the checkpoint to a similar extent ( Figure 2C) as the inhibition of ATR alone. The above set of data, collectively, suggests that both ATM and ATR kinases contribute to the activation of the IR-induced intra-S-phase checkpoint in rodents as well as in human cells, with ATR exerting a slightly stronger effect than ATM (see below for a model of the mechanism). The significance of differences between individual measurements is indicated by * symbol: * p < 0.05, ** p < 0.01, *** p < 0.001.

DNA-PKcs Contributes to the Recovery from IR-Induced Intra-S-Phase Checkpoint
DNA-PKcs has been recently implicated in the recovery from IR-induced G2 checkpoint in cells irradiated either in the S or G2 phase of the cell cycle [26,27]. Earlier work provided evidence that DNA-PK is important for the recovery from IR-induced intra-Sphase checkpoint [32], and we sought to follow up on this, taking advantage of the avail- The significance of differences between individual measurements is indicated by * symbol: * p < 0.05, ** p < 0.01, *** p < 0.001.

DNA-PKcs Contributes to the Recovery from IR-Induced Intra-S-Phase Checkpoint
DNA-PKcs has been recently implicated in the recovery from IR-induced G 2 checkpoint in cells irradiated either in the S or G 2 phase of the cell cycle [26,27]. Earlier work provided evidence that DNA-PK is important for the recovery from IR-induced intra-S-phase checkpoint [32], and we sought to follow up on this, taking advantage of the availability of highly specific inhibitors of ATM, ATR, and DNA-PKcs. We first measured the effect of DNA-PKcs on intra-S-phase checkpoint using a specific small-molecule inhibitor (DNA-PKi) and DNA-PKcs mutants of human and CHO origin.
Inhibition of DNA-PK results in slightly stronger inhibition of DNA synthesis, as compared to untreated samples of CHO and A549 cells. Notably, DNA-PK-inhibited cells show persistent checkpoint, indicating suppression of checkpoint recovery in both rodent and human cells ( Figure 3A,B). Similar responses could also be measured in DNA-PKcs mutants of human M059J, Chinese hamster (V3) origin, and HCT116 DNA-PK −/− (Figures 3C,D and S1 respectively). These results implicate DNA-PKcs in intra-S-phase checkpoint recovery, as previously reported for the G 2 checkpoint [26,27].
We also tested, in similar experiments, HCT116 WT and their LIG4-deficient counterpart HCT116 LIG4 −/− , and compared the results to those obtained with DNA-PK −/− cells generated on the same genetic background ( Figure S1). In this set of cell lines, activation of the intra-S-phase checkpoint is similar, but recovery deviates. While wild-type cells begin recovering DNA synthesis 2 h after IR, HCT116 LIG4 −/− cells begin recovering only after 3 h but do reach control levels at 6 h. On the other hand, their DNA-PKcs-deficient counterparts show sustained suppression of DNA replication at 6 h with no evidence of recovery.
XR1 cells, a CHO mutant in the LIG4 cofactor XRCC4, show activation of the intra-Sphase checkpoint, with kinetics similar to wild-type CHO cells and their DNA-PKcs-deficient counterparts, V3. However, while XR1 cells show delayed recovery, DNA-PKcsdeficient cells show sustained suppression of DNA replication ( Figure 3D,F). We conclude that DNA-PK is essential for efficient recovery from the intra-S-phase checkpoint and that this effect is specific for DNA-PKcs and may not be entirely attributed to the associated inhibition of c-NHEJ that occurs to a greater extent in LIG4 mutants. It is evident that in some but not all cell lines, c-NHEJ suppression by defects in factors other than DNA-PKcs delays checkpoint recovery. The variability among cell lines of this effect and the underpinning parameters will require further study in future work. The question as to which kinase (ATM and/or ATR) sustains the persistent intra-S-phase checkpoint in a DNA-PKcs-deficient background is examined next.  To investigate whether suppression of c-NHEJ-mediated DSB repair is solely responsible for the checkpoint potentiation observed in DNA-PK inhibited or mutant cells, we measured intra-S-phase checkpoint in the immortalized normal human fibroblast cells, 82-6 hTert, and compared the results to those obtained in two LIG4 mutants, 180BR and 411BR. We selected the LIG4 mutation for these experiments, as it is known to generate the strongest defect on DSB repair by c-NHEJ. The results summarized in Figure 3E show that 82-6 hTert, 180BR, and 411BR cells activate and recover from the checkpoint with similar kinetics, despite the expected longer persistence of DSBs in 180BR and 411BR cells.
We also tested, in similar experiments, HCT116 WT and their LIG4-deficient counterpart HCT116 LIG4 −/− , and compared the results to those obtained with DNA-PK −/− cells generated on the same genetic background ( Figure S1). In this set of cell lines, activation of the intra-S-phase checkpoint is similar, but recovery deviates. While wild-type cells begin recovering DNA synthesis 2 h after IR, HCT116 LIG4 −/− cells begin recovering only after 3 h but do reach control levels at 6 h. On the other hand, their DNA-PKcs-deficient counterparts show sustained suppression of DNA replication at 6 h with no evidence of recovery.
XR1 cells, a CHO mutant in the LIG4 cofactor XRCC4, show activation of the intra-S-phase checkpoint, with kinetics similar to wild-type CHO cells and their DNA-PKcsdeficient counterparts, V3. However, while XR1 cells show delayed recovery, DNA-PKcsdeficient cells show sustained suppression of DNA replication ( Figure 3D,F). We conclude that DNA-PK is essential for efficient recovery from the intra-S-phase checkpoint and that this effect is specific for DNA-PKcs and may not be entirely attributed to the associated inhibition of c-NHEJ that occurs to a greater extent in LIG4 mutants. It is evident that in some but not all cell lines, c-NHEJ suppression by defects in factors other than DNA-PKcs delays checkpoint recovery. The variability among cell lines of this effect and the underpinning parameters will require further study in future work. The question as to which kinase (ATM and/or ATR) sustains the persistent intra-S-phase checkpoint in a DNA-PKcs-deficient background is examined next.

Persistent Intra-S-Phase Checkpoint in DNA-PKcs Mutants Is Independent of ATM Activity
As shown above, ATM inhibition partly abrogates the intra-S-phase checkpoint in wt Chinese hamster and human cells ( Figures 1A and 2A, respectively). We tested whether ATM is responsible for the persistent intra-S-phase checkpoint in DNA-PKcs deficient cells. Human glioma wild-type (M059K) and DNA-PK mutant (M059J) cells, as well as a CHO mutant of DNA-PKcs, V3, were treated with ATMi, and the percentage of DNA synthesis was measured. Wt M059K cells show significant abrogation of the checkpoint ( Figure 4A), which is similar to that observed in CHO and A549 cells ( Figures 1A and 2A, respectively). Notably, ATM inhibition has a much smaller effect on V3 ( Figure 4C) and particularly on M059J ( Figure 4B) cells, suggesting that ATM's contribution to the intra-S-phase checkpoint diminishes in DNA-PKcs-deficient cells. Similar results are obtained with XR1 cells ( Figure S2A).

ATR Is the Key Activator of the Persistent Intra-S-Phase Checkpoint in DNA-PKcs Mutants
Using an experimental set up similar to that described above for ATMi, we analyzed the impact of ATR inhibition on an intra-S-phase checkpoint in DNA-PKcs mutant cells. Cells were treated with ATRi, and the percentage of DNA synthesis was measured in wt and DNA-PKcs-deficient mutants. Wild-type M059K cells ( Figure 5A) show clear abrogation of the checkpoint, similar in magnitude to that observed in CHO or A549 cells ( Figures 1B and 2B, respectively). Notably, ATR inhibition also resulted in nearly complete abrogation of the intra-S-phase checkpoint in M059J and V3 cells ( Figure 5B,C), which demonstrates that the strong and persistent intra-S-phase checkpoint in DNA-PKcs mutants is mainly sustained by ATR. ATRi also abrogates the checkpoint observed in XR1 cells ( Figure S2B). CHO mutant of DNA-PKcs, V3, were treated with ATMi, and the percentage of DNA synthesis was measured. Wt M059K cells show significant abrogation of the checkpoint ( Figure 4A), which is similar to that observed in CHO and A549 cells ( Figures 1A and 2A, respectively). Notably, ATM inhibition has a much smaller effect on V3 ( Figure 4C) and particularly on M059J ( Figure 4B) cells, suggesting that ATM's contribution to the intra-Sphase checkpoint diminishes in DNA-PKcs-deficient cells. Similar results are obtained with XR1 cells ( Figure S2A). Error bars represent the standard deviation from 3 independent experiments. The significance of differences between individual measurements is indicated by * symbol: * p < 0.05, ** p < 0.01, *** p < 0.001; n.s. non-significant.

ATR Is the Key Activator of the Persistent Intra-S-Phase Checkpoint in DNA-PKcs Mutants
Using an experimental set up similar to that described above for ATMi, we analyzed the impact of ATR inhibition on an intra-S-phase checkpoint in DNA-PKcs mutant cells. Cells were treated with ATRi, and the percentage of DNA synthesis was measured in wt and DNA-PKcs-deficient mutants. Wild-type M059K cells ( Figure 5A) show clear abrogation of the checkpoint, similar in magnitude to that observed in CHO or A549 cells (Figures 1B and 2B, respectively). Notably, ATR inhibition also resulted in nearly complete abrogation of the intra-S-phase checkpoint in M059J and V3 cells ( Figure 5B,C), which demonstrates that the strong and persistent intra-S-phase checkpoint in DNA-PKcs mutants is mainly sustained by ATR. ATRi also abrogates the checkpoint observed in XR1 cells ( Figure S2B). Our recent studies on IR-induced checkpoint signaling have shown a correlation between the degree of resection and the activation of G1 or G2 phase checkpoints. Indeed,

The Degree of DNA End-Resection at DSBs in the S-Phase Correlates with the Strength of the Intra-S-Phase Checkpoint
Our recent studies on IR-induced checkpoint signaling have shown a correlation between the degree of resection and the activation of G 1 or G 2 phase checkpoints. Indeed, ATM-, ATR-, and DNA-PKcs-mediated modulation of resection is a key factor in these responses [4,26,27,29]. Here, we examined whether the degree of resection in S-phase cells correlates with the strength of IR-induced intra-S-phase checkpoint. We measured resection in CHO cells after exposure to 20 Gy of IR, in the presence or absence of the above-discussed inhibitors of these kinases. RPA70 signal and EdU signal were measured by FACS at 3 h post-IR to enable cell-cycle specific analysis-here analysis in S-phase cells.
RPA70 intensity in the S phase is measured by gating S-phase cells as shown in Figure 6A. Irradiated CHO cells show an increase in RPA70 signal when compared to 0 Gy controls, indicating IR-induced DNA end resection ( Figure 6B upper panel). DNA-PKcs inhibition ( Figure 6B second panel from top) has no effect on resection, as reported earlier [26,27,38]. Inhibition of ATM ( Figure 6B third panel from top) or ATR individually ( Figure 6B fourth panel from top) partly suppresses resection. Notably, combined inhibition of ATM and ATR results in nearly complete inhibition of resection ( Figure 6B bottom panel). Suppression of resection through ATMi or ATRi individually, but particularly in combination, reflects well the observed impact of these kinases on the intra-S-phase checkpoint. In addition, the DNA-PKcs mutant, V3, shows clearly elevated resection in S phase when compared to wild-type CHO cells ( Figure 6C), which is in line with similar results obtained in G 2 -phase irradiated cells [26,27,38].

ATM and ATR Contribute to the Intra-S-Phase Checkpoint
Multiple studies have implicated ATM as the main kinase for the activation of IRinduced cell cycle checkpoints in mammalian cells [39,40], including the intra-S-phase

ATM and ATR Contribute to the Intra-S-Phase Checkpoint
Multiple studies have implicated ATM as the main kinase for the activation of IRinduced cell cycle checkpoints in mammalian cells [39,40], including the intra-S-phase checkpoint investigated here [14,41]. ATR, owing to the requirement for RPA-coated ssDNA for its activation [20], is considered to be mainly involved in the UV-induced intra-S-phase checkpoint and the regulation of DNA replication stress [14,16,42]. As a consequence, the contributions of ATR to IR-induced checkpoints are typically estimated as rather secondary and partly ATM-dependent [26,27].
However, recent results from our laboratory implicate ATR in IR-induced checkpoint responses much stronger than hitherto thought and suggest that ATM, ATR, and DNA-PKcs operate as a module to regulate IR-induced checkpoints. Notably, the putative ATM/ATR/DNA-PKcs module displays striking mechanistic adaptations that accommodate not only the progression of the irradiated cell through the cell cycle but also increases in IR dose, i.e., increases in the load to DSBs present in the genome [26,27,29]. In the current paper, we examined for the first time the crosstalk among ATM, ATR, and DNA-PKcs in the regulation of IR-induced intra-S-phase checkpoint. This study, as well as our previous studies [26,27], significantly benefited from the relatively recent availability of highly specific inhibitors of ATM, ATR, and DNA-PKcs, which dramatically ease experimentation, complement genetic approaches, and help to overcome problems associated with the lethality of genetic ATR ablation.
Our results show that, in line with previous studies [28], both ATM as well as ATR contribute to the activation of the intra-S-phase checkpoint and that indeed the contribution of ATR is stronger (Figure 7). The stronger effect of ATR inhibition in the suppression of the intra-S-phase checkpoint is notable because we previously showed that ATM remains functional in ATR inhibited cells [26]. Because combined application of ATMi and ATRi generates an effect only slightly stronger than that of ATRi alone, we surmise that in the absence of ATR, ATM is unable to significantly contribute to the activation of the intra-Sphase checkpoint. This raises ATR to the key regulator of the intra-S-phase checkpoint. highly specific inhibitors of ATM, ATR, and DNA-PKcs, which dramatically ease experimentation, complement genetic approaches, and help to overcome problems associated with the lethality of genetic ATR ablation.
Our results show that, in line with previous studies [28], both ATM as well as ATR contribute to the activation of the intra-S-phase checkpoint and that indeed the contribution of ATR is stronger (Figure 7). The stronger effect of ATR inhibition in the suppression of the intra-S-phase checkpoint is notable because we previously showed that ATM remains functional in ATR inhibited cells [26]. Because combined application of ATMi and ATRi generates an effect only slightly stronger than that of ATRi alone, we surmise that in the absence of ATR, ATM is unable to significantly contribute to the activation of the intra-S-phase checkpoint. This raises ATR to the key regulator of the intra-S-phase checkpoint.
These results are fully in line with those recently published for G2-checkpoint activation in cells also irradiated in G2 phase [26]. Indeed, in this case, and at low doses of IR, ATM and ATR epistatically regulate the G2 checkpoint, and inhibition of either kinase abrogates the checkpoint almost completely. On the other hand, at high IR doses that approach the range of doses used in the present study, ATM and ATR act independently, and inhibition of both is required to fully suppress the checkpoint. This is similar to what we observe for the intra-S-phase checkpoint in some cell lines.
Strikingly, when cells are irradiated in S phase, as was the case in all experiments presented here, the G2 checkpoint activated when cells reach the G2 phase uses ATM and ATR in an entirely different manner. Indeed, under these conditions, checkpoint activation relies entirely on ATR, and ATM contributes to its recovery as does also DNA-PKcs (see below) [27]. A comparison of these results to those presented here uncovers highly interesting and potentially relevant adaptations in the crosstalk between ATM and ATR, as S-phase irradiated cells progress through the cell cycle and arrest in the following G2 phase. Figure 7. A model depicting the crosstalk between ATM, ATR, and DNA-PKcs in the regulation of checkpoint activation and recovery for S-phase cells exposed to high IR doses. The checkpoint is measured either in S phase (intra-S-phase checkpoint), as described in the present paper, or after progression of cells into G2 phase, as described earlier [27]. See discussion for further details.
Why do the roles of ATM and ATR change among the two types of checkpoints experienced by cells irradiated in S phase? One reason might be that in the period of the cell cycle where the intra-S-phase checkpoint is activated, the genome is present in both unreplicated G1-like and replicated G2-like forms. However, when S-phase cells enter the G2 phase, DNA is exclusively in G2-like form. Consequently, G1-like DNA in S phase requires both ATM and ATR for checkpoint activation, as reported recently for the G1 checkpoint at high IR doses [29]. However, the G2-like DNA activates the checkpoint using mainly ATR. It has been reported that the G1 and the intra-S-phase checkpoints involve a similar Figure 7. A model depicting the crosstalk between ATM, ATR, and DNA-PKcs in the regulation of checkpoint activation and recovery for S-phase cells exposed to high IR doses. The checkpoint is measured either in S phase (intra-S-phase checkpoint), as described in the present paper, or after progression of cells into G 2 phase, as described earlier [27]. See discussion for further details.
These results are fully in line with those recently published for G 2 -checkpoint activation in cells also irradiated in G 2 phase [26]. Indeed, in this case, and at low doses of IR, ATM and ATR epistatically regulate the G2 checkpoint, and inhibition of either kinase abrogates the checkpoint almost completely. On the other hand, at high IR doses that approach the range of doses used in the present study, ATM and ATR act independently, and inhibition of both is required to fully suppress the checkpoint. This is similar to what we observe for the intra-S-phase checkpoint in some cell lines.
Strikingly, when cells are irradiated in S phase, as was the case in all experiments presented here, the G 2 checkpoint activated when cells reach the G 2 phase uses ATM and ATR in an entirely different manner. Indeed, under these conditions, checkpoint activation relies entirely on ATR, and ATM contributes to its recovery as does also DNA-PKcs (see below) [27]. A comparison of these results to those presented here uncovers highly interesting and potentially relevant adaptations in the crosstalk between ATM and ATR, as S-phase irradiated cells progress through the cell cycle and arrest in the following G 2 phase.
Why do the roles of ATM and ATR change among the two types of checkpoints experienced by cells irradiated in S phase? One reason might be that in the period of the cell cycle where the intra-S-phase checkpoint is activated, the genome is present in both unreplicated G 1 -like and replicated G 2 -like forms. However, when S-phase cells enter the G 2 phase, DNA is exclusively in G 2 -like form. Consequently, G 1 -like DNA in S phase requires both ATM and ATR for checkpoint activation, as reported recently for the G 1 checkpoint at high IR doses [29]. However, the G 2 -like DNA activates the checkpoint using mainly ATR. It has been reported that the G 1 and the intra-S-phase checkpoints involve a similar signaling cascade [43][44][45][46][47], and in accordance, the combined action of ATM and ATR at high DSB loads.
The intra-S-phase checkpoint is transient in nature, lasts only a few minutes, and may not provide enough time for the completion of DNA repair. Thus, cells with unrepaired DNA damage may enter the G 2 phase and arrest at the G 2 /M border [48,49]. This checkpoint relies solely on ATR [27] and connects to HR but not c-NHEJ [4,50]. Indeed, S-phase-irradiated cells enter the G 2 phase up to 8 h after IR, which leaves enough time for c-NHEJ to complete. It is likely, therefore, that the coordination of the G 2 checkpoint with c-NHEJ is lost, leaving HR and thus ATR as the sole contributor. In other words, the contribution of ATM to checkpoint activation is an early event and can only be observed in the phase of the cell cycle, where cells are irradiated. When cells move to the subsequent phase of the cell cycle, ATM mainly contributes to checkpoint recovery. Whether the intra-S-phase checkpoint facilitates DSB repair and how, and which repair pathways are favored during its activation requires further investigations.
The observations summarized here on the activation in S-phase-irradiated cells of the intra-S-phase and the G 2 -checkpoint and the roles of ATM and ATR in these responses are schematically shown in Figure 7.

DNA-PKcs Defects Cause Mainly ATR-Dependent Potentiation of Intra-S-Phase Checkpoint
Following successful repair or following adaptation to DNA damage, cells eventually re-enter the cell cycle through a process known as checkpoint recovery. Efficient recovery from a checkpoint is essential for the completion of checkpoint responses. DNA-PKcs has been implicated in the recovery from IR-induced G 2 and intra-S-phase checkpoint [26,27,32,[51][52][53], suggesting c-NHEJ-independent functions for the kinase.
In the present study, we confirmed that DNA-PKcs is crucial for the recovery from IR-induced intra-S-phase checkpoint, as reported earlier by Guan J. et al. [32]. Inhibition of DNA-PKcs causes a persistent S-phase checkpoint, which mainly depends on ATR activity. This response is similar to that reported earlier for the recovery from the G 2 checkpoint, both for cells irradiated in S, as well as in the G 2 phase of the cell cycle [26,27]. In addition, DNA-PKcs mutant cells show increased resection in irradiated S-phase cells, which likely causes stronger ATR activation and potentiation of the intra-S-phase checkpoint. It has also been reported that ATR regulates origin firing upon replication stress and that DNA-PK along with Chk1 serve as a backup pathway in this regulation when ATR is inhibited, highlighting additional interactions between ATR and DNA-PKcs [54].
The possibility that the delayed intra-S-phase checkpoint recovery in DNA-PKcsdeficient cells is due to compromised DSB repair through a defect in c-NHEJ is un-likely, as XRCC4 and LIG4 mutants, despite the potentiated checkpoint, exhibited checkpoint recovery when compared to their corresponding DNA-PK deficient counterparts ( Figures 3E,F and S1). In addition, recent studies from our group show that only DNA-PKcs mutation causes hyperresection upon irradiation in G 2 -phase cells, while cells with defects in other c-NHEJ factors show normal resection levels, comparable to control cells [26,38]. It has been shown that DNA-PKcs may crosstalk with ATM and ATR [53][54][55][56][57], suggesting a non-canonical, c-NHEJ-independent, function. Our results also suggest a non-canonical DDR function of DNA-PKcs in the IR-induced intra-S-phase checkpoint (Figure 7).
The results presented in this paper uncover the potential for coordinated functions of ATM, ATR, and DNA-PKcs in the intra-S-phase checkpoint (Figure 7), as has been reported earlier for IR-induced G 1 and G 2 checkpoint responses at high DSB loads [4,26,27,29]. One limitation of the current study is the absence of low dose results, owing to the reduced sensitivity of the assay employed. As a consequence, we could not reliably measure the intra-S-phase checkpoint at doses below 10 Gy. Thus, the interplay among ATM, ATR, and DNA-PKcs kinases in the regulation of the intra-S-phase checkpoint at low DSB loads remains elusive. Which DSB repair pathway/s engage during the intra-S-phase checkpoint will also need further detailed investigations.

Irradiation
Irradiations were carried out with an X-ray machine (GE-Healthcare, Chicago, IL, USA) operated at 320 kV, 10 mA with a 1.65 mm Al filter (effective photon energy approximately 90 kV), at a distance of 50 cm and a dose rate of approximately 1.3 Gy/min. The dosimetry was performed with a PTW and/or a chemical dosimeter, which was used to calibrate an in-field ionization monitor. Even exposure to radiation was ensured by rotating the radiation table. Cells were returned to the incubator immediately after IR.

Measurement of Radio-Resistant DNA Synthesis
DNA synthesis was measured by tritiated thymidine ( 3 H) incorporation in exponentially growing cells. 3 H was added for 15 min prior to collection of the indicated time point after IR exposure. Cells were trypsinized, counted, and loaded on glass microfiber filters (Whatman GF/A). The filters were washed 3X with a 10% solution of trichloroacetic acid (TCA), washed (3X) with deionized water, and incubated for 12-14 h in 0.5 mL of 0.5 N NaOH at 60 • C. The samples were subsequently neutralized with 0.5 mL of 0.5 N HCl, supplied with 10 mL scintillation fluid, and counted for 3 H-activity in a liquid scintillation counter (Perkin Elmer Tri-Carb ® 2910TR, Waltham, MA, USA). The rate of DNA synthesis was calculated from these measurements and is presented as the percentage of values measured in sham-irradiated controls.

Flow Cytometry-Based DSB end Resection Assay
DNA end resection was measured using a RPA70 staining-based flow cytometric method [4,26,27,58]. Briefly, exponentially growing cells were pulse-labeled for 30 min with 5 µM 5-ethynyl-2 -deoxyuridin (EdU). EdU-containing media were removed, and cells were rinsed once with pre-warmed PBS, supplemented with fresh pre-warmed growth medium prior to exposure to X-rays. Cells were collected at 3 h after IR by trypsinization. Non-irradiated cells were collected at 1 h after the EdU pulse. Unbound RPA was extracted by treating cells with ice-cold phosphate-buffered saline (PBS) containing 0.2% Triton X-100 for 2 min on ice. Cells were spun-down and fixed for 15 min in PFA solution at room temperature (RT). Cells were then blocked with PBG buffer overnight at 4 • C until preparation for immunostaining. Cells were incubated with a monoclonal antibody raised against RPA70, which was kindly provided by Dr. Hurwitz [59] for 1.5 h at RT. Cells were washed twice with PBS and incubated for 1 h at RT with a secondary antibody conjugated with AlexaFluor 488. Subsequently, the EdU signal was developed using an EdU staining kit according to the manufacturer's instructions. Finally, cells were stained with 40 µg/mL PI. Cell cycle specific analysis was carried out using flow cytometry combined with quantification by Kaluza 1.3, as described earlier [26,27].

Statistical Analysis
p values were calculated using t-test in Sigma Plot 11.0 and 14.0. The significance of differences between individual measurements is indicated by the * symbol representing p values: * p < 0.05, ** p < 0.01, *** p < 0.001; n.s. non-significant.