Dynamic Alterations of Histone H3 Phospho-acetylation Correlate With Radio Sensitivity of Mitotic Cells During DNA Damage

Background - Histone Post Translational Modications (PTMs) change in a cell cycle dependent manner and also orchestrate the DNA repair process for radiation induced DNA damage. Mitosis is the most radiosensitive phase of the cell cycle but the epigenetic events that regulate its radiosensitivity remain elusive. Results - This study explored the dynamics between histone marks H3S10/S28ph, H3K9ac and γH2AX during mitotic DNA damage response. The presence of a mononucleosome level association between γH2AX and H3S10ph was observed only during mitosis. This association was abrogated upon cell cycle progression and chromatin de-condensation, concomitant with chromatin recruitment of DNA repair proteins Ku70 and Rad51. Moreover, the levels of H3S10/28ph remained unchanged upon DNA damage during mitosis, but decreased in a cell cycle dependent manner upon mitotic exit. However, the population that arose after mitotic progression of damaged cells comprised of binucleated tetraploid cells. This population was epigenetically distinct from interphase cells, characterized by reduced H3S10/S28ph, increased H3K9ac and more open chromatin conguration. These epigenetic features correlated with decreased survival potential of this population. The low levels of H3S10/28ph were attributed to decreased protein translation and chromatin recruitment of histone kinase Mitogen and Stress-activated Kinase 1 (MSK1) along with persistent levels of Protein phosphatase1 catalytic subunit α (PP1α). Conclusions – This study suggests that a unique epigenetic landscape attained during and after mitotic DNA damage collectively contributed to mitotic radiosensitivity. The ndings of this study have potential clinical signicance in terms of tackling resistance against anti-mitotic chemotherapeutic agents.

Surprisingly, DDR activation during mitosis is associated with deleterious consequences like lagging chromosomes and anaphase bridge formation (14) , (22). The chromatin recruitment of 53BP1 during mitosis causes micronuclei generation, sister telomere fusion and confers a hypersensitive phenotype to mitotic cells (19). The rate of mitosis progression can also be impeded by activation of the spindle assembly checkpoint (23) , (24). However, mitotic cells can bypass the spindle assembly checkpoint by Cyclin B degradation, and undergo mitotic slippage. This results in formation of tetraploid binucleated cells that arise due to cytokinesis failure. Such binucleated cells that have neither completed mitosis nor entered the G 1 phase, are called 4N-intermediate cells (25)(26)(27)(28)(29)(30). Due to inter and intra-cellular heterogeneity, the fate of such cells can be either mitotic cell death or senescence. However, these cells can also re-enter the cell cycle, undergo one round of DNA replication but get arrested in the subsequent G 2 phase (31)(32)(33)(34). However, survival and proliferation of such cells can be detrimental due to their high tumorigenic potential and drug resistant nature (35)(36)(37)(38).
Histone Post Translational Modi cations (PTMs) change according to cell cycle phases and aid in the recruitment of chromatin remodelers and repair proteins. Therefore, histone PTM alterations that occur in response to DNA damage could be a cell cycle phase speci c event (39)(40)(41). Earlier studies from our group reported a dynamic kinetics of H3S10ph, observed only during the DDR of the relatively radioresistant G 0 /G 1 phase (42,43) Our recent report also indicates an association of altered histone phosphoacetylation with breast cancer radio-resistance (44). These reports suggest that a distinct interplay of H3 phospho-acetylation could regulate cell cycle phase speci c radio resistance.
Keeping in mind the above propositions, this study revealed histone phospho-acetylation to be an epigenetic determinant of mitotic radiosensitivity. A mitosis-speci c mononucleosome level correlation was observed between H3 phospho-acetylation and γH2AX, concomitant with negligible chromatin recruitment of DNA repair proteins. Additionally, radiated mitotic cells exhibited defects during mitotic progression and showed phenotypic variations like binucleated cells. The reduced cell survival of this population correlated with decreased H3S10/28ph, elevated H3K9ac and a de-condensed chromatin state. The H3 phosphorylation of such cells was regulated by enhanced stability of the phosphatase Protein Phosphatase 1 catalytic subunit α (PP1α) and reduced translation of Mitogen and Stressactivated Kinase 1 (MSK1) kinase. This study provides evidence of a distinct epigenetic milieu that contributes to radio-sensitivity of mitotic cells. The in-depth epigenetic analysis performed in this study can be utilized in clinics to design novel (or combinatorial) epigenetic therapies that can potentially target the surviving binucleated tetraploid cells, deemed to be tumorigenic and chemo-resistant in nature.

Results
Mitotic cells exposed to ionizing radiation (IR) exhibit G 2 /M arrest, high levels of H3K9ac and reduced H3S10/28ph histone PTMs.
A time dependent analysis provided information about histone PTM alterations after mitotic DNA damage (Fig. 1A-B). During mitosis, increased levels of γH2AX were observed upon IR treatment while H3S10/28ph remained unchanged ( Fig. 1A and B, C-E γH2AX channel). After 2 hours of nocodazole release, more than 50% of the non-radiated cells had exited mitosis and entered the G 0 /G 1 phase.
However, only 30% of the radiated cells had entered the G 0 /G 1 phase 4 hours after nocodazole release ( Fig. 1A). This indicated that radiation exposure could lead to a delay in mitotic progression. Mitotic exit and the G 0 /G 1 phase entry were marked by reduced H3S10/28ph and increased H3K9ac levels, irrespective of radiation treatment ( Fig.1 A-E, Additional File 1 A-C). This was concomitant with conversion of the condensed mitotic chromatin to a de-condensed state, reminiscent of an interphase nucleus (Fig 1 C-E; DAPI channel and Additional File 1 A-C DAPI Channel).
After radiation and nocodazole release, decreased levels of γH2AX suggested ongoing repair ( Fig 1B). However, there was an incremental increase in the percentage G 2 /M phase cells (Fig 1A). Notably, the levels of H3S10/S28ph were neither comparable to the non-radiated cells nor increased upon G 2 /M phase enrichment (

Cellular morphology of mitotic cells after DNA damage and cell cycle progression
Contrary to the expected G 0 /G 1 phase arrest, the G 2 /M enrichment after mitotic exit was perplexing.
Therefore, an immuno-uorescence based analysis was performed to analyze the cellular morphology of these cells. The nuclear shape after IR and mitotic progression was distorted and fragmented, compared to well-rounded nuclei of non-radiated cells (Fig 2A, marked by red arrows). Cells exposed to IR also showed presence of micronuclei (Fig 2A, depicted by yellow arrows), chromatin-bridge ( Fig. 2B marked by white arrows) and "grape phenotype" of the nuclei (Fig. 2C, represented by white arrows). Interestingly, coincident with the previously observed G 2 /M enrichment (Fig. 1A), bi-nucleated cells were detectable after IR and nocodazole release (Fig 2A, white arrows). This provided strong evidence that the previously observed G 2 /M state comprised of bi-nucleated tetraploid cells, that were detected by ow cytometer to have same ploidy as the G 2 /M phase.

Cell division defects in mitotic cells subjected to radiation
Live-cell microscopy was performed to analyze radiation-associated defects that could lead to the formation of bi-nucleated tetraploid cells. A signi cant delay in initiation of cell division was seen after IR exposure ( Fig 3A white arrows in IR+ve, White arrows in 3A IR-ve also point towards cell division, quanti ed in 3B.). Fusion of daughter cells after cell division was also observed (Fig 3A depicted by red arrows, zoomed out gures and 3C marked by red arrow; 1h 11 min. depicts cell division and 1h 42 min. depicts cell fusion). In addition to daughter cell fusion, three distinct events were also observed when radiated mitotic cells resumed cell cycle progression. Firstly, radiated mitotic cells were able to complete the cell division and divide into two daughter cells (Fig. 3C marked by black arrows; 0 min. is starting of time lapse and 40 min. is cell division). Secondly, cells did not initiate cell division even two hours after nocodazole release (Fig. 3C represented by blue arrows; 0 min. is starting of time lapse and 2h 50 min. represents no cell division). Thirdly, radiated cells also gave rise to asymmetric sized daughter cells ( Fig.  3C depicted by green arrows, 0 min. is starting of time lapse). Therefore, radiation-associated defects in cell cycle progression mitotic cells led to the formation of unique phenotypes.
Analysis of cell cycle DNA repair and chromatin alterations in mitotic cells subjected to radiation In context of DDR, there was no recruitment of either NHEJ speci c Ku70 and HR speci c Rad51 repair proteins on chromatin during mitosis (Fig. 4A). The recruitment of these proteins was concomitant with reduced γH2AX levels, observed 4 hours after nocodazole release (Fig. 4A). A comparison between the population formed after radiation and nocodazole release and an actual G 1 phase population showed cyclin B levels to be reduced but detectable upto 4 hours after IR and nocodazole release, followed by complete absence 24 hours after radiation. Additionally, IR treated population had reduced levels of cyclin D upon cell cycle progression, in comparison to non-radiated and G 0 /G 1 phase cells (Fig. 4B).
Additionally, MNase digestion revealed that radiated mitotic cells adopted a more de-condensed chromatin state, compared to non-radiated cells ( Fig. 4E and F, points of difference marked by black arrows in 4F). Since the population generated after cell cycle progression of radiated mitotic cells was epigenetically distinct from a G 0 /G 1 phase population, these cells were also assessed for their survival potential. Thus, ploidy-based Fluorescence Activated Cell Sorting (FACS) was performed to separate the G 0 /G 1, S and G 2 /M phase cell population 48 hours after radiation and nocodazole release (Fig. 4C).
Subsequent exposure of these cell populations to IR revealed a drastic reduction in the cell survival potential, even without radiation exposure. However, it was interesting that despite a reduced survival capacity, there were some cells that were able to proliferate and produce colonies (Fig. 4D).
Histone PTMs γH2AX, H3S10/28ph and H3K9ac co-localization during mitotic DNA damage Previous observations suggested that upon IR exposure, γH2AX and H3S10ph/S28ph co-occurred during mitosis, but not after mitotic progression (Fig 1B). This suggested an interphase speci c inverse correlation between γH2AX and H3S10ph. Thus, alterations of mitotic marks H3S10/28ph were assessed during mitotic DDR. Immuno-uorescence analysis revealed partial co-localization of γH2AX with H3S10/28ph during mitosis (seen as yellow regions). However, levels of H3K9ac in mitotic cells were too low for detection and co-localization analysis (Fig 5A-C, 0 hour time point). Interestingly, there were two striking features of cells generated after mitotic progression of radiated cells. Firstly, no co-localization was observed between γH2AX and H3S10ph, contrary to the scenario during mitosis (Fig 5A, 4 hours time point and zoomed out image). Both the histone marks were observed to form very distinct spatial foci despite being present in the same nucleus ( Fig. 5A zoomed out image). Secondly, it was also apparent that cells having more intense staining of γH2AX had a dramatically reduced intensity of H3S10ph ( Fig.   5A zoomed out image). H3K9ac mark was also not observed to co-localize with γH2AX ( Fig. 5C zoomed out image) and H3S28ph levels were too low to comment (Fig. 5B zoomed out image).
A reason for the observed co-localization between H3S10ph and γH2AX during mitosis could be the highly condensed chromatin state. The close proximity of the red and green uorophores (representing H3S10ph and γH2AX, respectively) could lead to yellow co-localizing regions. However, upon mitotic progression and chromatin de-condensation, the uorophores could get spatially apart, thereby causing abrogation of co-localization. Thus to ascertain whether H3S10ph and γH2AX actually co-localized during mitosis, a mono-nucleosomal co-immunoprecipitation was performed (Fig 5D and Additional File  3 A-C). Notably, both γH2AX and H3S10ph marks co-occurred on the same nucleosome during mitosis ( Fig. 5D, 0 hour time point). Nucleosomes having γH2AX and H3S10ph marks also harbored H3S28ph and H3K9ac during mitosis. However, upon chromatin de-condensation, H3S10/S28ph and H3K9ac marks were absent from γH2AX-containing nucleosomes, corroborating the immuno uorescence-based observation (Fig. 5D, 4 hour time point). Interestingly, upon mitotic progression, the nucleosomes containing H3S10ph harbored H3S28ph and H3K9ac but not γH2AX. Hence, a dynamic rearrangement of chromatin upon mitotic progression led to mutual exclusion of γH2AX and H3S10ph in interphase, contrary to their co-occurrence during mitosis, corroborating our previous ndings of similar nature during G 0 /G 1 phase speci c DDR (43).

Alterations in histone modifying enzyme levels upon mitotic DDR
Since the population generated after radiated mitotic cells that resumed cell cycle progression had reduced levels of H3S10ph, the protein levels of histone modifying kinases and phosphatases were assessed. The phosphatases Protein phosphatase1 catalytic subunit α (PP1α) and MAP Kinase Phosphatase-1 (MKP-1) showed persistent levels up to 48 and 24 hours nocodazole release, respectively. This pattern was followed irrespective of radiation treatment but with higher levels in the radiated cells  Fig.1A for cell cycle). The non-radiated cells showed an increase in MSK1 levels 2 hours after nocodazole release, while no such increase was observed even up to 4 hours after radiation. This was a very crucial and distinguishing event between radiated and non-radiated cells. Interestingly, these time points signi ed the G 0 /G 1 phase entry for these populations ( Fig. 6A and Fig.1A for cell cycle). Apart from no induction of MSK1 protein, there was also a rapid decline in its levels. Additionally the phosphatase PP1α, but not MKP-1 showed chromatin recruitment upon mitotic progression (irrespective of radiation exposure). On the other contrary, MKP-1 was substantially enriched in the nucleo-cytoplasmic fraction (NCF) (Fig. 6D). The kinases AURKB and MSK1 also showed reduced chromatin recruitment after DNA damage and mitotic progression. The increase in H3K9ac after mitotic progression was concurrent with recruitment of histone acetyl transferases (HATs) GCN5 and PCAF on chromatin, irrespective of radiation exposure (Fig. 6D). Interestingly, reduced levels of H3S10ph and increase of H3K9ac were concomitant with chromatin recruitment of (a) HATs, (b) PP1α, and (c) reduced chromatin recruitment of HDAC1 ( Fig. 6 C and D).
A previous in silico study from our group revealed MSK1 kinase to have a reduced a nity towards H3 peptides acetylated at positions H3K9 and K14 (45). Since histone PTMs act in a combinatorial manner, the presence of a distinct set of histone PTMs could lead to an increased chromatin recruitment of a protein. Thus it was hypothesized that the acetylation on residues H3K9/K14 could lead to enhanced recruitment of phosphatase PP1α on chromatin. To investigate this, molecular modelling was performed using Swiss model software for protein structures available for PP1α and MKP-1 with a set of differentially modi ed histone H3 peptides (Fig 6E). The modi cations on the histone peptides were H3S10ph, H3K9ac and H3K14ac in combinations of unmodi ed, dual and triply modi ed histone Nterminal tails. The calculated haddock score predicted the extent of a nity for a speci c combination of phospho-acetylated H3 tail. It was observed that both MKP-1 and PP1α had comparable a nity for unmodi ed H3 peptides. Similar observation was seen in case of peptides having combination of only H3S10ph along with only one acetyl mark (either H3K9ac or H3K14ac). PP1α showed increased a nity for peptides having one or both H3K9ac/H3K14ac marks with absence of H3S10ph. Remarkably, this similar H3 phospho-acetyl PTM milieu was mimicked by radiated and mitosis progressed cells, with increased H3K9ac but negligible levels of H3S10ph (Fig. 6C). Therefore, these data indicated that in silico, the chromatin recruitment of H3 phosphatases MKP-1 and PP1α could be in uenced by PTM(s) present or absent on nearby residue(s), apart from the H3S10 position.

Regulation of histone modifying kinases and phosphatases upon mitotic DNA damage and cell cycle progression
To understand the regulation of H3S10ph modifying enzymes, transcript-level alterations were analyzed in response to mitotic DDR (Fig. 7A). The transcript levels of MKP-1 and MSK1 were signi cantly increased 2 hours after radiation and mitotic progression while those of PP1α and AURKB were unchanged, except decreased AURKB expression 24 hours after radiation (Fig. 7A). The information provided by transcript level analysis of histone modifying kinases and phosphatases was insu cient in explaining the alterations at protein levels. This indicated that these proteins could be regulated by protein translation or degradation. Treatment of mitotic cells with protein translation inhibitor cycloheximide (CHX) did not affect mitotic progression cells, as indicated by cyclin B levels (Fig. 7B). No increase in the levels of p53 upon CHX treatment and DNA damage induction con rmed the activity of CHX. The level of MSK1 protein increased 6 hours after nocodazole release in the non-radiated cells. Such an induction was not observed upon CHX treatment, irrespective of radiation exposure. This suggested MSK1 protein levels were regulated by translation of its mRNA upon mitotic progression.
Hence despite transcriptional up-regulation, there was no increase in MSK1 protein levels due to radiation induced translation-related defects. The phosphatase MKP-1 followed a cyclical pattern of increase of protein levels at 2 hours, followed by a reduction at 6 hours after radiation. These alterations at the protein levels were concomitant with the transcript level alterations Upon CHX treatment, the MKP-1 protein levels diminished, thereby suggesting MKP-1 was regulated by both transcription and translation ( Fig. 7 A and B). Interestingly, the protein levels of PP1α were increased even upon CHX treatment, with unchanged transcript levels. These data indicated that protein stabilization could play an important role in regulating the level of PP1α.
In contrast to the other chromatin modifying enzymes, the protein level of AURKB declined irrespective of radiation and CHX treatment. This pointed towards the role of protein degradation in regulation of AURKB levels. Mitotic cells treated with velcade, an inhibitor of the proteasome machinery, showed sustained levels of cyclin B and AURKB (Fig. 7C). Conversely, untreated mitotic cells progressed to G 0 /G 1 phase showed reduced protein levels of AURKB and cyclin B. This indicated that AURKB protein levels were regulated by proteasome-mediated degradation upon mitotic exit and interphase entry. Additionally, inhibition of protein degradation was unable to rescue the levels of MSK1 kinase, thereby indicating protein translation to be the regulator of MSK1 protein levels upon mitotic progression. In context of the phosphatases, minor accumulation of MKP-1 protein was observed upon velcade treatment and radiation exposure, thereby implying the turnover of MKP-1 to be regulated by its degradation also. Strangely, PP1α levels remained unchanged upon both protein translation and degradation inhibition. This strongly pointed towards involvement of PTM-associated protein stabilization (of PP1α or its associated regulatory subunit). Thus, the non-recovery of H3S10ph after mitotic DNA damage was a result of reduced translation of the kinase MSK1 or persistent presence of phosphatases MKP-1 and PP1α.

Discussion
Interphase, comprising of G 1 , S and G 2 phases, is the longest part of the cell cycle and the DDR events associated with it are well elucidated. The emphasis of this study was to understand regulation of histone H3 phospho-acetylation during mitotic DDR. We report a mitosis speci c mononucleosome level association of H3S10ph and γH2AX that is abrogated upon interphase entry and unique cellular phenotypes that arise after mitotic DNA damage as a consequence of radiation exposure. Such cells have decreased survival potential and a distinct epigenetic pro le. The reduced H3S10/28ph level in these cells was contributed by decreased protein translation and chromatin recruitment of Msk1 kinase and enhanced stability of phosphatase PP1α.
The interest in H3S10ph and its nearby acetyl mark(s) originates from our previous reports about association of H3 phospho-acetylation with G 0 /G 1 phase speci c DDR and radio-resistance (42)(43)(44).
However, the role of H3S10ph during mitotic DDR is not well explored. Our data for cell-cycle dependent reduction of H3S10/28ph upon mitotic progression and co-occurrence of these marks with γH2AX are in complete agreement with a previous study (46). Also in corroboration with a previous report, we observed reduced levels of H3K9ac during mitosis that increased upon mitotic exit (47). This was attributed to chromatin recruitment of Histone Acetyl Transferases (HATs) Gcn5 and PCAF. However, our data does not suggest DNA damage associated decrease of H3K9ac during mitotic DDR and nocodazole release (46). Several factors such as duration of nocodazole treatment, speci c time point analysis after interphase entry, cell line and extent of DNA damage often lead to such contrasting observations, that re ects the dynamic nature of DDR associated epigenetic alterations (48).
Perhaps the most confounding observation of the study was the G 2 /M arrest of IR exposed mitotic cells.
This was actually a technical limitation of ow cytometry based cell cycle analysis, which detected bi-nucleated tetraploid cells as G 2 /M phase cells. Absence of cyclin B and reduced cyclin D levels has been reported for cells that aborted mitosis but were unable to enter the G 1 phase(33). Our live cell analysis suggests that such a "4N-intermediate" / bi-nucleated tetraploid population could also arise due to daughter cell fusion. However, further studies are required to understand radiation-associated defects in the cytokinesis process that leads to formation of bi-nucleated tetraploids.
Notably, a mixed population, comprising of G 0 /G 1 , S and G 2 /M phase cells was generated after IR exposure to mitotic cells. Such a population could arise due to (a) normal cell cycle progression after DNA repair, (b) activation of interphase DNA damage checkpoints or (c) re-entry of binucleated cells into cell cycle. This phenotypic variability is reported to arise due to a competition between independent networks of cell survival and cell death, leading to intra-cell variation (31). Most remarkably, radiated mitotic cells adopted a more de-condensed chromatin con guration after cell cycle progression, which correlated with their reduced cell survival. This observation is in complete concordance with our recent report that suggests increased heterochromatinization upon acquirement of radio-resistance (44). Hence, these data strongly indicate that global chromatin con guration plays an essential role in determining cell fate towards radiation induced DNA damage. However, our interpretation is based on a collective analysis of diverse cell types formed after mitotic progression. Hence studies utilizing a single-cell based approach to elucidate the "cause or consequence" of a speci c phenotype are essential. Ongoing studies are also aimed at understanding how H3S10/28ph and H3K9ac could affect the transcriptome of binucleated tetraploid cells that tips the scales in favor cell death or survival.
A ne balance between the levels, activity and spatio-temporal regulation of chromatin modifying enzymes maintains the equilibrium of histone phospho-acetylation. According to our in silico and biochemical analysis, we propose that increased histone acetylation positively in uences PP1α chromatin recruitment. This was found to be in agreement with a report that suggested enhanced chromatin recruitment of PP1δ upon histone acetylation induction during mitosis (49). Further, the plausible causes of reduced H3S10ph after radiation and mitotic exit were decreased MSK1 kinase levels and persistent presence of PP1α phosphatase. A hypothesis that emerges from these data suggests that histone hyperacetylation-associated chromatin recruitment of PP1α could regulate activity of the kinase(s). While hypo-acetylated histone tails were found to be preferred substrates for AURKB mediated mitotic H3S10ph(50), it was also reported that presence of PP1 negatively regulates the kinase activity of AURKB in Xenopus interphase chromatin extracts (51). Most interestingly, our data suggests PP1α chromatin recruitment to occur irrespective of radiation exposure. So how does PP1α regulate kinase activity in interphase cells that arise after mitotic radiation? There can be two plausible explanations that shed light on this proposition. Firstly, the kinase activity of MSK1 is regulated by auto-phosphorylation and ERK 1/2 or p38 mediated phosphorylation (52). Therefore it could be possible that PP1α regulates either an upstream activator of MSK1 or directly the kinase. Secondly, the substrate speci city of PP1 catalytic subunit depends on its association with different regulatory subunits. Association of PP1γ with Repo-Man has been reported to regulate histone de-phosphorylation that occurs upon mitosis to G 1 transition(53,54). Formation of similar complex by PP1α could also regulate MSK1 kinase activity during interphase, in response to DNA damage. Possibly PP1 could associate with a unique regulatory subunit upon radiation treatment, which directs its speci city towards interphase kinases. These are interesting possibilities that can bring to light the complexity of kinase-phosphatase regulation during mitotic DDR.
In context of mitotic DDR, our data suggests co-occurance of γH2AX and H3S10ph on the same mononucleosome during mitosis but not interphase. This was concomitant with an independent observation of Ku70 and Rad51 chromatin localization only upon mitotic exit. Our previous report also suggests G 0 /G 1 phase speci c H3S10ph decrease to be crucial for DNA repair(43) , (42). This suggests that the mono-nucleosomal co-existence of H3S10ph and γH2AX might not provide a permissive environment for initiation of DNA repair during mitosis. There are four probable possibilities that led us to this conclusion. Firstly, H3S10ph is associated with DNA-RNA hybrid R-loop formation that is conserved across yeast, nematode and human cells (55). Hence, dynamic alterations of H3S10ph during mitotic DDR could be detrimental to genome stability by formation of such structures during cell division process. Secondly, it is elusive whether γH2AX-H3S10ph form homotypic or heterotypic nucleosomes.
This could lead to nucleosome architecture alterations, which might hinder chromatin recruitment of remodeling factors required during DNA repair. This proposition seems plausible since structural alteration of chromatin by remodeling complexes, especially when the cell is preoccupied with the division process, might lead to genomic instability. Additionally, to further explore this possibility, it needs to be ascertained whether H3S10ph-γH2AX containing nucleosomes are present at the DNA damage site/DSB or in its vicinity. Thirdly, the activity, levels, spatio-temporal presence and accessibility of the H3S10ph and γH2AX modifying kinases and phosphatases could also govern their co-occurance. Finally, is possible that apart from H3S10ph and γH2AX, presence or absence of other PTMs is required for mitotic DNA-damage response activation, and such a milieu is attained upon chromatin de-condensation in G 1 phase. This hypothesis is based on a report by Clouaire et. al. that uses arti cial double strand break (DSB) induction for elucidation of the histone PTM milieu around the break site (56). A similar analysis of the histone PTMs adjoining the DSBs during mitotic DDR could provide strong evidence of a "code" required to halt DDR in mitosis but initiate it in interphase.
Our study strongly suggests that key epigenetic differences between mitotic and interphase cells could lead to differences in their DDR. The report also suggests that the radio-sensitivity of mitotic cells is contributed by the epigenetic state of the population that arises after damaged mitotic cells resume cell cycle (Fig. 8). It is interesting to note tetraploid cells are tumorigenic in nature, so their survival and proliferation could lead to tumor relapse even after successful radiation regime. Therefore, an understanding of the epigenetic features of such cells could open up avenues for utilization of nextgeneration mitotic inhibitor therapy or combinatorial therapies with HDAC inhibitors to target the residual detrimental population.

Materials And Methods
Cell culture, synchronization: MCF7 cell line was a kind gift by Dr. Santosh Kumar Sandur, BARC, India.
U87 cell line was provided by Dr. Neelam Shirsat, ACTREC and AGS cell line was purchased from ATCC. MCF7 and U87 cells were cultured in DMEM (Invitrogen, USA) media while RPMI (Invitrogen, USA) was used for AGS cell line. The medium was supplemented with 2mM glutamine (Sigma, USA), 10% fetal bovine serum (Gibco, USA). Cell were maintained at 37°C and 5% CO 2 . Synchronization in G 0 /G 1 phase of the cell cycle was done by serum starvation using 0.02% serum for 72 hours. Serum starvation was stopped by adding 10% FBS containing medium 6 hours before radiation to allow cells to enter G 1 phase.
Synchronization in pro-metaphase was performed by incubating cells with 200ng/ml nocodazole for 18 hours.
Nocodazole release of mitotic cells and inhibitor/drug treatment-Nocodazole release of mitotic cells was performed by shake off method (57,58). Mitotic cells were subjected to radiation while the medium still contained nocodazole. Immediately after irradiation, mitotic cells were collected by gently tapping the plates to dislodge mitotic cells. The mitotic cells collected in the medium were pelleted by centrifugation at 3000 rpm for 2 minutes at 4°C. Nocodazole (Sigma, USA) containing medium was discarded and cells washed once with 1X PBS to remove residual nocodazole and centrifuged as described above. The mitotic cells were re-suspended in nocodazole free medium and seeded into fresh 100mm 2 culture plates to allow mitotic progression. Cycloheximide (Sigma, USA) (50μg/ml) and proteasome inhibitor velcade (1μM/ml) were added 1 hour before radiation treatment and also after nocodazole release in the necessary controls. Drug treatment was given as Cisplatin (Sigma, USA) (2μg/ml for 4 hours), Ultra Violet (10J/m 2 for 15 minutes) rays, Adriamycin (Sigma, USA) (10μg/ml for 4 hours) and H 2 O 2 (Sigma, USA) (500μM for 15 minutes).
Cell irradiation: A Co-60 radioactive source present in Bhabhatron-II (Panacea Medical Technologies Ltd. and Bhabha Atomic Research Centre (BARC), India) installed at the Department of Radiation Oncology, ACTREC was used to subject the cells to ionizing radiation. Field size was 25cm x 25cm, Source-to-skin distance (SSD) as 80cm and gantry was angled at 180° to the specimen.
Histone Isolation and total cell lysate preparation: Histone isolation was performed as described earlier (43). The chromatin bound histones were extracted using acid extraction method. The histones obtained in the nal pellet was resuspended in 0.1% β-mercaptoethanol (Sigma, USA) and stored in -20°C. Total cell lysates were prepared by sonication of the cell pellet in the buffer described above and cleared of debris by centrifugation at 17000rpm for 30 mins at 4°C. The supernatant obtained was used as total cell lysate.
Western blotting: Histones and total cell lysates were resolved on 18% and 10% SDS-Poly-Acrylamide Gel Electrophoresis (SDS-PAGE) respectively, transferred on PVDF membrane (Immobillion) and subjected to western blotting. List of antibodies used in the study is provided in Table 1.
Immuno uorescence Microscopy-Immuno uorescence was performed as described previously (43). Imaging was done using Zeiss 510 Meta confocal microscope. Image analysis was performed using FIJI software.
Live cell imaging and analysis -MCF7 cells were synchronized in pro-metaphase by nocodazole arrest in a glass-bottomed plate. Immediately after radiation, nocodazole-containing medium was discarded and cells were carefully washed twice with 1X PBS. Nocodazole-free medium was added to the plate and imaging was initiated. Cells were maintained at 95% relative humidity and 5% CO 2 for 3 hours during imaging process. Imaging was performed on Leica SP8 confocal microscope at 63X magni cation. Images were processed using the Leica LASX software.
Flow cytometry based cell cycle analysis and sorting: Cell cycle analysis was carried out using propidium iodide (Sigma, USA) based DNA content analysis as described previously (43). DNA content analysis was carried out using Fluorescence-Activated Cell Sorting (FACS) Calibur ow cytometer (Becton Dickinson) and analysis done using MODFIT software by Verity house. Cell sorting was performed using VYBRANT green dye (Invitrogen, USA), as per manufacturer's instructions.
Clonogenic assay: MCF7 cells sorted after VYBRANT dye treatment were counted and seeded in 6 well plates. 1000 cells were seeded per well, per cell cycle phase. Cells were irradiated and maintained for 14 days. Colonies were xed with 4% paraformaldehyde for 20 mins followed by washing with Phosphate Buffered Saline (PBS). Staining was performed using 0.5% crystal violet (59).
Cellular fractionation: Cellular fractionation into cytoplasmic, nuclear and chromatin fractions was performed as described previously (43). The nuclear and cytoplasmic fractions were pooled together to form nucleo-cytoplasmic fraction. It should be noted that western blots depicted in Fig. 4A and 7D are a part of the same experiment but depicted as separate gures. They are derived from the same experimental nucleo-cytoplasmic and chromatin lysates; hence have same histone H3 and GAPDH loading control blots.

Mono-nucleosomal Immunoprecipitation
Isolation of mono-nucleosomes was performed as previously described (43). Mono-nucleosome isolation from mitotic nuclei was done in MNase digestion buffer containing 10mM CaCl 2 . Mono-nucleosomes were prepared by incubating 1mg chromatin with 200 units of MNase (UBS, USA) for 30 minutes at 37°C. 200μg of chromatin was incubated with 2μg of anti-γH2AX, anti-H3S10ph and IgG antibodies. 20μl of magnetic DYNA beads (Thermo Fischer, USA) were added to chromatin-antibody mixture and incubated for 4 hours on a rotating platform. Separation of antibody-nucleosome complex and washing of bound complex was done using magnetic rack. 2X SDS loading dye was added to the bead bound complex. The samples were boiled, chilled and loaded on 18% SDS-PAGE gel, followed by western blotting with respective antibodies.
Molecular homology modelling -In silico modelling of MKP1 and PP1α with differentially modi ed H3 peptides was performed as previously described (45).
Quantitative PCR: RNA extraction from was done by Trizol method, followed by DNaseI treatment (Fermentas, USA) and cDNA synthesis using random hexamer primers (Revert-Aid cDNA Synthesis Kit, Thermo Scienti c, USA), strictly as per manufacturer's instructions. Real-time PCR was performed using gene speci c primers using ampli cation conditions of 30 seconds at 94°C, 1 minute at 60°C and 1 minute at 72°C for 30 cycles followed by 10 minute nal extension. List of primers used in the study is provided in Table 2. The expression of change upon radiation was plotted as fold change normalized to non-radiated cells of the indicated time points.
Statistical analysis -All numerical data were expressed as average of values obtained ± standard deviation (SD). Statistical signi cance was determined by conducting unpaired students t-test.

Declarations
Ethics approval and consent to participate Not applicable

Consent for publication
The paper was critically read by all authors and approved for publication.

Availability of data and materials
The data in manuscript is available    arrows indicate daughter cell fusion. Mins-Minutes. Statistical analysis was performed using unpaired ttest. **** p<0.0001, *** p<0.001, ** p<0.01, * p<0.05 and p>0.05 is considered non-signi cant. and 7D were a part of the same experiment but are depicted as separate gures. They are derived from the same experimental nucleo-cytoplasmic and chromatin lysates; hence have same histone H3 and GAPDH loading control blots. Co-localization analysis of γH2AX with H3S10/28ph and H3K9ac during mitotic DNA damage response.
Representative z-stack images for co-immuno uorescence based co-localization analysis of γH2AX with (A) H3S10ph, (B) H3S28ph and (C) H3K9ac at speci ed time points after radiation and nocodazole release. Time points 0 hour depicts time when cells were still in mitosis and 4 hours denotes time elapsed after nocodazole release and radiation. Inset shows zoom-out image at 4 hour time point. Yellow color denotes co-localization of indicated histone marks. DAPI acts as nuclear marker for all images (D) Mononucleosomal immuno-precipitation (IP) performed with anti-γH2AX and anti-H3S10ph antibody at the indicated time points after radiation and nocodazole release. Immuno-blotting (IB) performed with anti-γH2AX, anti-H3S10ph, anti-H3S28ph and anti-H3K9ac antibodies. Input is 10% amount of mononucleosomes used for immuno-precipitation. * and ** denote antibody heavy chain (55kDa) and light chain (25kDa) of the antibodies, respectively. Hrs. -Hours and Gy-Gray. Scale bar for all microscope-based images is 10μm. Radiation dose is 8Gy for all samples.