The effect of DNA damage on the pattern of immune-detectable DNA methylation in mouse embryonic fibroblasts

Author’s Accepted Manuscript The effect of DNA damage on the pattern of immune-detectable DNA methylation in mouse embryonic fibroblasts Selcen Çelik, Yan Li, Chris O’Neill www.elsevier.com/locate/yexcr

PII: DOI: Reference:

S0014-4827(15)30079-3 http://dx.doi.org/10.1016/j.yexcr.2015.08.017 YEXCR10038

To appear in: Experimental Cell Research Received date: 26 March 2015 Revised date: 31 July 2015 Accepted date: 27 August 2015 Cite this article as: Selcen Çelik, Yan Li and Chris O’Neill, The effect of DNA damage on the pattern of immune-detectable DNA methylation in mouse embryonic fibroblasts, Experimental Cell Research, http://dx.doi.org/10.1016/j.yexcr.2015.08.017 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

The effect of DNA damage on the pattern of immuneimmune-detectable DNA methylation in mouse embryonic fibroblasts

Selcen Çelik1,2,3, Yan Li1 and Chris O’Neill1

1

Human Reproduction Unit, Kolling Institute for Medical Research, Sydney Medical School, University of Sydney, Sydney, 2065, Australia 2

Current address: Department of Molecular Biology and Genetics, Faculty of Science, Karadeniz Technical University, Trabzon, 61080, Turkey

3

Correspondence: [email protected], +90 462 377 42 72 (phone)

Short title: The patterns of DNA methylation after DNA damage

Key Words: Epigenetics, cytosine methylation, methyl-binding protein-1, heterochromatin, genotoxic stress, DNA damage, DNA repair, trypsin, antigen retrieval, immuneimmune-staining, mouse, fibroblast

Abbreviations:

5meC

(5-methyl

cytosine),

MBD1

(methyl-binding

protein-1),

HP1-β

(heterochromatin protein 1-β), MEFs (mouse embryonic fibroblasts)

1

Abstract The methylation of cytosine at CpG dinucleotides (5meC) is an important epigenetic mechanism that governs genome stability and gene expression. Important ontological and pathological transitions are associated with marked global changes in detectable levels of methylation. We have previously found wo pools of immuneimmune-detectable 5meC exist within cells, a pool that can bedetected after acid treatment of fixed cells to denature chromatin and another large but variable pool that requires a further tryptic digestion step for complete epitope retrieval.retrieval The trypsin-sensitive pool has been shown to be largely associated with the heterochromatic fraction (by a heterochromatin marker, HP1-β) of the genome, and the size of this pool varies with the growth disposition of cells. Since DNA damage imposes large changes on chromatin structure the present study analyzedzedhow such changes influences the faithful immunological detection of 5meC within mouse embryonic fibroblasts. DNA damage was induced by either UV-irradiation or doxorubicin treatment, each of which resulted in increased levels of immune-detectable 5meC at 24-48 h after treatment. There was a marked trypsin-sensitive pool of 5meC in these cells which was significantly increased after DNA damage. The increased levels of 5meC staining predominantly co-located with heterochromatic foci within nuclei, as assessed by HP1-β staining. The relative amount of masked 5meC after DNA damage was positively associated with increased levels of HP1-β. The methyl binding protein, MBD1, was a less reliable measure of changes in 5meC, with a significant fraction of 5meC not being marked by MBD1. The cyto-epigenetic approaches used here reveal dynamism in the levels and localizationz of eimmuno-detectable 5meC within the nuclei of fibroblasts in response to DNA damage.

2

Introduction Methylation of cytosinecat CpG dinucleotides (5meC) of thegenome has important roles in the maintenance of genome stability as well as influencing the patterns of gene expression. MethylationM can also influence chromatin structure within the nucleus as DNA in heterochromatic regions is commonly hypermethylated [1-4]. Sincemajor changes in chromatin conformation and structure are known to occur during ontogeny [5-7], throughout the cell-cycle [6,8,9] and in response to DNA damage [6,10], it is of interest to understand whether 5meC levels are associated with these changes. The use of immunolocalizationimmunolocalization with specific antibodies provides the significant advantage of ingdetection of changes in the localization of5meC within the architecture of the nucleus of individual cells [11,12]. It has the further very important advantage of being able to distinguish between the known range of covalent modifications to cytosine (5meC; 5hmC-5´hydoxymethylycytosine;

5fC,

ydoxymethylycytosine5´-formylcytosine;

and

5caC,

5´-

carboxycytosinerboxycytosine) in the same cell, an analysis not possible with modern forms of chemical measurement.measurement The power of immunolocalizationimmuno can only be fully realized, however, if the conditions used faithfully allow detection of antigens present. This requires that antibody binding isperformed under equilibrium binding conditions and that sample preparation allows the all of the epitope to be exposed to the aqueous phase so that antibody binding can be achieved. These pre-conditions are particularly challenging for antigens within chromatin because of the complex and variable conformationand structure of chromatin [2,3,7,13]. DNA is by its nature highly coiled and forms many higher order structures. This complexity is compounded by the many histone and non-histone proteins that bind to DNA and the range of covalent and other modifications of chromatin components that can occur. Conventionally, the detection of 5meC has relied upon denaturation of chromatin by brief acid treatment to cause solvent exposure of 5meC. We have recently shown, however, that this form of epitope retrieval leaves a large amount5meC masked from immunodetection [14-16]. This was most evident in the newly fertilized mouse embryo where the use of conventional immuno-staining methods showed a progressive loss of 5meC staining resulting in an appearance of almost complete demethylation by the time of the first cell division [17-19]. Further analysis showed that this loss of 5meC staining was not primarily due to a loss of5meC but was due to a progressive trypsin-sensitive masking of the antigen that accompanied changes to chromatin conformation and structure during the zygotic cell-

3

cycle. Conventional immune-localization of 5meC has typically involved a brief exposure of fixed cells to acid treatment and it has been generally reasoned that this provides sufficient denaturation of chromatin to expose the 5meC.. Yet a brief period of tryptic digestion of the fixed embryos allowed much more 5meC antigen to be retrieved. This T methodologyrevealed that there were only relatively minor changes in the levels of 5meC in the early embryo, and did not support claims of active global demethylation in the zygote. Analysis of mouse embryonic fibroblasts (MEFs) showed that these cells also had a degree oftrypsin-sensitive masking of 5meC [14]. While the near complete masking of 5meC seen in the zygote was not evidentevident in MEFs, there was still a large trypsinsensitive pool of 5meC. This pool was predominantly associated with the heterochromatic fraction of the genome. Most notably, it was found that 5meC associated with heterochromatin varied with the growth disposition of cells [14]. For instance, proliferative cells displayed a relatively greater accumulation of 5meC within the trypsin-sensitive pool than quiescent cells, and this was in large part associated with changes in the heterochromatin fraction between these two growth states (14). DNA damage and the resulting repair processes are associated with marked changes in the extent of chromatin compaction and heterochromatin expansion, and this is considered to allow repair factors to access damaged sites of DNA [6,10]. It is also noteworthy that molecular models for the catalytic demethylation of cytosine implicate the involvement of DNA repair processes, including the base excision repair pathway [20-26]. Given the changes in chromatin organisation caused by DNA damage and repair processes wewinvestigated if various forms of genotoxic stress influenced the levels of solvent exposure and trypsin-sensitive masking of 5meC in MEFs. We chose two treatments to induce DNA damage; UV irradiation can predominantly induce single strand DNA breaks (SSBs) [27,28], while doxorubicin causes double strand DNA breaks (DSBs) [29,30]. The study finds that both forms of DNA damage caused changed levels, localization and trypsin-sensitive masking of immuneimmune-detectable 5meC in MEFs. This was associated with the marked changes within the heterochromatin fraction of the genome. The findings show that 5meC within the fibroblast nucleus displayed significant dynamism that can be detected in a cost-effective and efficient manner by careful attention to the conditions required for full antigen retrieval during immuno-staining.

4

Material and Methods Cell Culture Mouse embryonic fibroblasts (MEFs) from Day 13.5 embryos were collected using standard methods and then cultured in Dulbecco’s modified eagle medium (DMEM) (Thermo Fisher Scientific Inc, Utah, USA; Cat. No. SH30243.FS). Media were supplemented with 10% (v/v) fetal bovine serum (FBS) (Invitrogen Life Technologies, Carlsbad, USA; Cat. No. 10099-141), 1% (v/v) 1 x MEM-Non-Essential Amino Acids solution (Invitrogen, Cat. No. 11140-050), 50U/ml penicillin (Sigma-Aldrich Co.; St. Louis, MO, USA; Cat. No P3032) and 50 µg/ml streptomycin (Sigma, Cat. No. S6501) at 37°C with 5% CO2 in air. Cells were grown as proliferative (1 day-culture) or quiescent (24h serum-starved confluent culture) cultures. Genotoxic stress Quiescent MEFs were exposed to UV-irradiation (253.7nm at a dose of 1.95J/cm2)for (i) 15 sec followed by post-incubation either for 1h or (ii) 12 min followed by culture for 48h.Cells were in HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) buffered media (pH 7.4) (Invitrogen, Cat. No. 12430) at 37°C, during exposure. HEPES-buffered media included bovine serum albumin (BSA) (3mg/ml). Control cells were untreated at 37°C with 5%CO2 in air. Proliferative MEFs were treated with doxorubicin (Sigma, Cat. No. D1515) at doses of 0 or 50 nM in complete media for 24h. DNA damage – Immunofluorescence for γ-H2A.X (phosphoS139) staining After treatment, MEFs were washed with 1xPhosphate-buffered saline PBS) and fixed with 2% (w/v) paraformaldehyde (PFA) (Sigma, Cat. No. P6148) for 30 min at RT (room temperature). Cells were permeabilized with 2% (w/v) PFA containing 0.3% (v/v) tween-20 (Sigma, Cat. No. P7949) and 0.2% (v/v) triton-x (Bio-Rad Laboratories Inc., CA, USA, Cat. No.161-0407) for 30 min at RT. Cells were washed in 1xPBS (w/v) (Sigma, Cat. No. D5773) for 30 min and blocked in 30% (v/v) goat serum (Sigma, Cat. No. G9023) in 2 mg/ml BSA in PBS with 0.05% triton-x (w/v) (BSA, from Sigma, Cat. No. A1470) for 3h at RT. Cells were washed in 1xPBS for 30 min at RT. Cells were incubated in primary antibodies (1:250): rabbit polyclonal anti-γ-H2A.X (phosphoS139) (Abcam, UK; Cat. No. ab2893) to detectDNA breaks; or a non-immune rabbit IgG (Sigma, Cat. No.I5006) in 2 mg/ml BSA in PBS with 0.05% triton-x (w/v) at 4°C overnight. Cells were washed in 1xPBS for 30 min, and then were incubated with FITC (fluorescein isothiocyanate) (1:200) (Sigma, Cat No.F1262) in 2 mg/ml BSA in

5

PBS with 0.05% triton-x (w/v) for 1h at RT in the dark. Cells were washed in 1xPBS for 30 min, and then mounted with PBS. Immunofluorescence for DNA methylation (double staining for 5meC and MBD1) After treatment, MEFs were washed with 1xPBS (w/v) and fixed with 4% (w/v) PFA for 30 min at RT. Cells were then washed with 1xPBT (w/v) (PBT, Phosphate-buffered saline with tween-20) for 30 min. Cells were permeabilized with 1xPBS including 0.5% (v/v) triton-x and 0.5% (v/v) tween-20 for 40 min at RT. Cells were treated with 4N HCl containing 0.1% (v/v) triton-x for 10 min at RT. Acid was removed by extensive washing with 1xPBT. Cells were treated with trypsin (0.25%) either for 0 or 1 min at 37°C. Trypsin was inactivated by washing with an equal volume of pre-warmed media including 90% (v/v) DMEM, 3 mg/ml BSA and 10% (v/v) sheep serum, for 1 min at 37°C and then washed with 1xPBT (w/v) followed by blocking in 10% (v/v) goat and 30% (v/v) sheep serum in 1xPBT at 4°C overnight. Cells were incubated in primary antibodies: rabbit anti-MBD1 (1:50) or nonimmune rabbit IgG (1:50) in 2 mg/ml BSA in 1xPBT overnight at 4°C. Cells were washed with 1xPBT for 10 min at RT followed by incubation with primary antibodies: mouse anti-5meC (1:60) or nonimmune mouse IgG (1:60) in 2 mg/ml BSA in 1xPBT for 1h at RT. Cells were washed with 1xPBT (w/v) for 30 min. Cells then were incubated with secondary antibody mix including anti-rabbit-Texas Red (1:200) andanti-mouse-FITC (1:200) in 1xPBT with 2 mg BSA/ml for 1h in the dark at RT. Cells were washed with 1xPBT (w/v) for 30 min at RT. Cells were mounted with 1xPBS including bisBenzimide Hoechst 33342 (Sigma, Cat. No. B2261) (4 µg/ml). Immunofluorescence for double staining of MBD1 and heterochromatin (HP1-β) After treatment, MEFs were washed with 1xPBS (w/v) and fixed with 4% (w/v) PFA for 30 min at RT. Cells were washed with 1xPBT (w/v) for 30 min. Cells were permeabilized with 1xPBS including 0.5% (v/v) triton-x and 0.5% (v/v) tween-20 for 40 min at RT. Cells were treated with 4N HCl containing 0.1% (v/v) triton-x for 10 min at RT. Cells were washed with 1xPBT (w/v) for 30 min at RT followed by two-step blocking: (i) in 30% (v/v) goat serum in 1xPBT at 4°C overnight, and then (ii) in 30% (v/v) sheep serum in 1xPBS for 2h at RT. Cells were incubated in primary antibody mixes including either rabbit anti-MBD1 (1:50) (Abcam, Cat. No. ab3753) and mouse anti-HP1-β (1:200) (Abcam, Cat. No. ab101425); or non-immune rabbit IgG (1:50) (Sigma, Cat. No. I5006) and mouse IgG (1:200) (Sigma; Cat. No. M7894) in 2 mg/ml BSA in 1xPBT (w/v) at 4°C overnight. Cells were washed with 1xPBT (w/v) for 30 min at RT. Cells then were incubated with secondary antibody mix including anti rabbitTexas Red (1:200) (Abcam, Cat. No. ab6719) and anti mouse-FITC (1:200) (Sigma, Cat. No. 6257) in

6

1xPBT 2 mg BSA /ml for 1h in the dark at RT. Cells were washed with 1xPBT (w/v) for 30 min at RT, and 1xPBS (w/v) wash once. Cells were mounted with 1xPBS including Hoechst (4 µg/ml). Microscopy and Image Analysis Cells were assessed by epifluorescence microscopy, Eclipse 80i (Nikon Instruments Inc., USA). Images were captured using CoolSnap cf camera (Photometrics, AZ., USA), and analyzed by Image-Pro Plus version 5.0 (Media Cybernetics, Inc., MD., USA). Slides were assessed using the same setting of UV power, and images were taken using the same exposure time (4 sec) for all treatments of replicates in each experiment. Original images without any further adjustments (i.e. contrast) were analyzed. Objects (nuclei) in images were randomly selected using drawing tool and analyzed. The total staining (sum value) of specific antigens in each cell (nucleus) was measured, and the level of staining is optical density (A.UAU., arbitrary units). The average staining of antigen in cells was shown using bar graphs. In some experiments, nuclei were subjectively grouped as having a predominant focal or diffuse staining, and the number of each group was compared. Undetectable and so unmeasurable non-immune IgG staining was defined as a negative control compared to the specific staining. Experiments were performed for at least 3 independent replicates. Statistics Statistical analyses were performed by SPSS program, version 19. Graphs were made using SPSS or Sigma Plot (Version 11.0). The total amount of 5meC staining and the numbers of 5meC foci within the nuclei were analyzed using UNIANOVA. The percentages of γ-H2AX-stained cells were arcsinetransformed, and compared using UNIANOVA. Differences between groups were determined by post-hoc test. The proportions of nuclei stained with antigens were compared using binary logistic regression analysis. The relationship between 5meC and MBD1 in a nucleus was analyzed by linear regression.

7

Results Three experimental models of DNA damage to MEFs were established (Table 1).). DNA damage was induced either by UV-irradiation or doxorubicin treatment, and the extent of damage was assessed by the staining of a DNA damage marker, phosphorylated histone 2A2A.X (γ-H2A.X). Model 1 involvedinvolved UV treatment of quiescent cells for 12 min followed by 1h further incubation. This resulted in an increase both in the total staining of γ-H2A.X (p<0.0001) (Fig 1A) and in the number of nuclei staining for γ-H2A.X (p<0.0001) (Fig 1B, C). Model 2. had a shorter period of exposure UV treatment (15 sec) of quiescent cells followed by 48h further cultureculture in complete media. This treatment resulted a significant increase in the detectable amount of γ-H2A.X (p<0.0001) (Fig 1D, E) andalso in the proportion of nuclei with with an accumulation of γ-H2A.X after 48h(p<0.0001) (Fig 1F). In model 3mdoxorubicin treatment (50nM) of proliferative cells for 24h caused increased γH2A.X staining (p<0.0001) (Fig 1G, H), and increased number of nuclei staining for γ-H2A.X (p<0.0001) (Fig 1I). Untreated proliferative cells (in model 3) showed a higher level of γ-H2A.X staining than quiescent cells (untreated controls in models 1 and 2), as is consistent with γ-H2A.X also marking mitotic DNA of normal cells [31-33]. Each of the treatments waswas validated to have caused significant levels of DNA damage based upon the accumulation of γ-H2A.X staining above their respective untreated controls.controls. In model 1, UV treatment caused an increase in the acid-sensitive pool of immune-detectable 5meC (p<0.0001). Epitope retrieval by an additional tryptic digestion step increased the total levels of 5meC staining in both untreated and UV-treated cells (p<0.0001) but the relative increase caused by tryptic digestion was less in UV-treated cells than untreated controls (Fig 2A). This result indicates that an acute response of cells to UV exposure was an overall small increase in the immune-detectableimmunodetectable pool of 5meC but a decrease in the relative size of the trypsin-sensitive pool. The binding of MBD1 to DNA may serve as aproxy measure of 5meC. UV-treatment in model 1 had no effect on the acid-sensitive or trypsin-sensitive levels of MBD1 staining (p>0.05), but the overall level of MBD1 detected was higher after trypsin treatment (p<0.0001) (Fig 2B). Co-staining of cells for 5meC and MBD1 showed at there was a similar relationship between the levels both antigens after either acid or acid plus trypsin treatment in untreated control cells (Fig 2C), but in UV-treated cells tryptic digestionincreased the amount of MBD1 staining relative to the amount of 5meC detected (Fig 2D). Observations of the pattern of staining in individual nuclei showed that much of the MBD1 staining foci co-stained with 5meC, both in untreated and UV-treated cells (represented in orange boxes, Fig 2E), but there were some MBD1-staining foci that did not co-stained with 5meC in

8

each treatment. This analysis also showed that the overall pattern of 5meC staining was similar in both UV-treated and untreated cells with a low level of diffuse 5meC staining across the nucleoplasm and a number of 5meC-intense staining foci (that corresponded to Hoechst-enriched staining foci) (Fig 2E). In UV-treated cells more of these 5meC-intense foci were detected without the necessity for tryptic digestion (Fig 2E) and this seemed to largely account for the increase in the acid-sensitive 5meC staining level observed in UV-treated cells. This analysis indicates that a response to the acute DNA damage induced by brief UV treatment was ana acute decrease in the amount of trypsin-sensitive masking of the 5meC epitope located within Hoechst-staining intense foci. There was also an increase in the proportion of these foci that co-stained with MBD1. The results showshow that MBD1 is only a partialpartial measure ofof 5meC levels, and this was improved by brief tryptic digestion of cells. The longer term effects of UV-irradiation were assessed ini model 2 and this showed that at 48 h after treatment for 15 sec there were increasedlevels of both 5meC and MBD1 detected after epitope retrieval by acid alone and this increase was greater after epitope retrieval by acid plus trypsin treatment (Fig 3A). The increase in immune-detectableimmunodetectable 5meC after tryptic digestion was evident in both untreated and UV-irradiated cells, but was greater after UV treatment (Fig 3A, B). The amount of 5meC and MBD1MBD1 antigen detected was positively correlated for both untreated and UV-treated cells (p<0.0001) (Fig 3C, D). The relationship between 5meC and MBD1 staining was stronger in UV-treated cells than untreated cells after trypsin treatment (Fig 3C, D). This analysis also showed that after UV treatment a number of the MBD1 foci did not co-localize either with 5meC or Hoechst stain-intense chromocentres in acid-treated cells (represented in white boxes). This was less evident after tryptic digestion with a higher level of 5meC and MBD1 colocalization at these chromocentres chromocentres (represented in orange boxes) (Fig 3E). InI the quiescent cells little extra trypsin-sensitive staining of 5meC was detected and much of the 5meC staining was primarily restricted to the chromocentres. The distribution of Hoechst-stained chromocentres was similar for untreated and UV-treated cells and was unaffected by brief tryptic digestion. In untreated cells, many of the chromocentres were only lightly co-stained for 5meC or MBD1 (Fig 3E). In UV-treated cells, thee level of trypsin-sensitive masking of the 5meC was increased (Fig 3B,E). Thus, 48h after UV exposure quiescent MEFs had a marked increase in the immunedetectableimmunodetectable levels of 5meC compared to untreated cells and much of this increase was within the trypsin-sensitive 5meC pool and was predominantly associated with the Hoechst staining-intense chromocentreschromocentres.

9

TheT treatment of proliferative cells with doxorubicin (model 3) caused a significant increase in both the acid and trypsin-sensitive immune-detectable levels of 5meC and MBD1(Fig 4A, B). The level of MBD1 and 5meC staining in individual cells was strongly correlated in both untreated and drugtreated cells (Fig 4C, D). Doxorubicin-treated treated cells had a high level of co-localization of 5meC and MBD1 after tryptic digestion, and both antigens showed marked accumulation at Hoechst staining chromocentreschromocentres(Fig 4E). After acid treatment alone, MBD1 staining both in untreated and doxorubicin treated cells tended to be under-represented within the chromocentres and over-represented in the rest of the nucleoplasm, while 5meC staining was predominantly associated with the chromocentres (Fig 4E). Doxorubicin treatment increased the amount of solvent exposed 5meC detected after acid treatment, including an increase associated with the chromocentres. Trypsin treatment further increased the overall level of 5meC and MBD1 detected. This revealed extensive co-localization of MBD1 and 5meC within many of the chromocentres (represented in yellow boxes). This co-localization was not commonly seen after acid treatment alone (represented in white boxes) (Fig 4E). This result shows that doxorubicin-induced DNA damage in proliferative MEFs caused an overall increase in immune-detectableimmunodetectable 5meC, and a significant amount of this 5meC was trypsin-sensitive. The different levels of 5meC in the untreated control cells in models reflectreflects the difference in the experimental design in each model (Table 1).. In model 1, cells were confluent and had serum removed from media, in model 2 these conditions were extended for a further 48 h after less UV exposure resulting in cells becoming entirely withdrawn from the cell-cycle. In model 3, cells were cultured under more sparse conditions in the presence of serum ensuring that most cellscell will be undergoing cell-cycle progression. We have previously shown [14] that the levels of methylation are the lowest in proliferative cells, increasedincrease in confluent cells and are the highest in quiescent cells and this pattern is confirmed in the untreated cells in the current study. Each of the three models of DNA damage caused an increased association of 5meC and MBD1 with Hoechst-intense nuclear chromocentres.chromocentres. The late response (24-48h) to DNA damage was associated with an increased level of trypsin-sensitive masking of DNA methylation but in the acute response (1 h) to DNA damage the level of masking was reduced. In each model much of the 5meC was associated with Hoechst-staining chromocentreschromocentres and these are considered to consist of heterochromatic regions of the nucleus [7,34,35]. We measured HP1-β to assess the effect of each model of DNA damage on the pattern of heterochromatin. HP1-β staining was subjectively classified as being predominantly dispersed throughout the nucleoplasm (diffuse staining) or showing a number of distinct staining foci (focal staining) (Fig 5A). In model 1, UV

10

treatment caused an acute reduction in the proportion of cells with focal HP1-β staining (p<0.0001) (Fig 5B) but there was increased focal HP1-β staining 48h after UV exposure (p<0.0001) (Fig 5C). Doxorubicin caused a small but significant increase in the proportion of nuclei with predominantly focallocalization of HP1-β (p<0.05) (Fig 5D). The total level of HP1-β staining showed an acute decrease in response to UV-induced DNA damage (p<0.0001) in model 1 (Fig 5E), increasedincreased by almost two-fold 48h after UV irradiation (p<0.0001) in model 2 (Fig 5F), and a small increase after doxorubicin treatment (p<0.0001) in model 3 (Fig 5G). The two treatments (models 2 and 3) that caused increased HP1-β staining were associated with an increase in the average nuclear size (Fig 5H-J).

We next assessed whether the Hoechst-intense chromocentres were heterochromatic as assessed by HP1-β accumulation at these sites. The co-localization of the methyl binding protein MBD1 at these chromocentres was also assessed. In. In untreated cells of each model, most of the Hoechst intense foci co-stained for HP1-β (with varying intensity) confirming the heterochromatic status of these

regions.

The

acute

response

to

UV

exposure

(Model

1)

was

for

chromocentreschromocentres to have reduced intensity of HP1-β staining (Fig 6 A, B) but there was an increased proportion of the chromocentreschromocentres that stained for both HP1-β and MBD1 compared to untreated cells (Fig 6 C). 48h after UV treatment the

chromocentreschromocentres became heavily stained for HP1-β but had little acid-sensitive MBD1 (Fig 7A-C). A broadly similar pattern of change was observed in doxorubicin-treated cells (Fig 8A-C). These experiments show that the Hoechst-intense chromocentreschromocentres present in MEFs are predominantly associated with heterochromatin. It also shows that the nature of this heterochromatin changed after DNA damage. It can be concluded that the trypsin-sensitive masking of 5meC was predominantly associated with the heterochromatic regions and that the increased masking that occurs as a late response to DNA damage was associated with increased accumulation of HP1-β at these heterochromatic foci while the reduced trypsin-sensitive masking of 5meC 1h after UV treatment was associated with a reduced intensity of HP1-β staining.

11

Discussion This study shows that immune-detectable immunodetectable levels of 5meC increased 24-48h after the induction of DNA damage by doxorubicin or UV-irradiation, respectively. These changes were evident after epitope retrieval by the conventional method of acid-induced denaturation of DNA; however,

this

method

generally

underestimated

the

total

levels

of

immune-

detectableimmunodetectable methylation. More More 5meC was detected after a further tryptic digestion step was added to the epitope retrieval process. A late effect (24-48 h) of DNA damage waswas a relative increase in the amount of 5meC that was masked in a trypsin-sensitive manner. Much of the increased 5meC detected after DNA damage was associated with the heterochromatic fraction of the genome and DNA damage resultedresulted in increased binding of HP1-β and MBD1 and localization of 5meC at many heterochromatic foci. By contrast, only a small change in the total level of 5meC was detected 1h after UV-induced DNA damage, but there was a reduction in the relative level of trypsin-sensitive masking of 5meC detected at this time and this was associated with reduced levels of HP1-β binding at most heterochromatic focifoci. The results show that as cells respond to DNA damage changes in the conformation of chromatin results and this is associated with changes in the total and trypsin-sensitive pool of immune-detectableimmunodetectable 5meC within the genome. These findings indicate a considerableconsiderable dynamism in the global levels and localization of this epigenetic mark occurs in response to DNA damage.. Immunolocalization of 5meC provides the enormous advantage that it allows analysis of methylation changes within individual cells. It also allows assessment of changes in the intranuclear localization of 5meC. ItIt is increasingly recognised that the three dimensional organisation of the genome is as important in regulating gene function as the two-dimensional sequence information [36].. When immunolocalization of 5meC is combined with detection of other cytosine derivatives or with genomic structures, such a satellite sequences, powerful insights into the roles of cytosine modifications into nuclear organisation and structure will be possible. The interpretation of this data, however, requires confidence that the immunolocalization faithfully reflects the true 5meC . Carefullocalization. Careful design and validation of immunodetection of DNA methylation is required to optimizeoptimise epitope retrieval in each system under analysis. Conditions for testing in each system include technical conditions, such as the antibody incubation time and temperature required to achieve equilibrium binding conditions [37], the form and extent of antigen retrieval procedures [14-16,38,39], and the nature of the blocking protocol used to minimize non-specific binding and maximize, the form and extent of antigen retrieval procedures , and the nature of the

12

blocking protocol used to minimise non-specific binding and maximise the signal to noise in the fluorescence channel. The improved methods of immunolocalization used here show that different interpretation of methylation patterns became evident compared to conventional methodology. It is recognised that DNA damage results in the recruitment of a wide range of proteins to the damage site. TheseThese are involved in the sensing and recognition of the damage and the coordinated organisation of the repair ofof lesions and thethe quality assurance of the repair process. Key sensors of DNA damage include γ-H2A.X [40] and we show here that the treatments imposed caused recruitment of γ-H2A.XX to multiple foci through chromatin.

DNA

methyltransferases (DNMT)1 and 3A and two alsoof repair and can catalyze selective methylation of the promoter-distal segment of the damaged DNA 41]. DNMT1 accumulation occurs in UV-induced lesions 42]43. It has been observed that after repair there is remodeling of the methylation pattern in a transcription-dependent manner and this is stabilized over a 10-15 day period [41]. Our observations that the level of methylation varied with time after irradiation and the nature of the irradiation used is consistent with these earlier findingsDNA methyltransferases (DNMT)1 and 3a and two proteins that regulate methylation activity, UHRF1 and GADD45A, are also recruited to sites of repair and can catalyse selective methylation of the promoter-distal segment of the damaged DNA [34]. DNMT1 accumulation occurs in UV-induced lesions [35] and was detected in micronuclei in

doxorubicin-treated cells [36]. It has been observed that after repair there is remodelling of the methylation pattern in a transcription dependent manner and this is stabilised over a 10-15 day period [34]. Our observations that the level of methylation varied with time after irradiation and the nature of the irradiation used is consistent with these earlier findings. Methylation of cytosine creates a site for the recruitment of a range of other proteins. It acts as a docking site for a range of proteins that have selective 5meC recognition and binding domains. We have chosen one example of this family of proteins, MBD1, for assessment. High but incomplete association between 5meC and MBD1 binding in the genome was detected, and this was greater after retrieval of both epitopes by trypsin digestion. Methyl binding proteins can serve as a docking site for a further set of proteins. For example, MBD1 [44] and MeCP2 [45] can bind the histone methylase SUV39H1 which acts to di- and tri-methylate histone 3 lysine 9. This modification is a canonical mark of heterochromatin. MBD1 also recruits and binds HP1 proteins and histone deacetylases, both of which are characteristic of compacted heterochromatic regions [44]. These interactions of the methyl binding proteins serve as self-reinforcing mechanism by which cytosine modification can profoundly change the conformation of chromatin. HP1 family heterochromatin

13

proteins are also recruited to a range of DNA damage sites [10,46,47] and the initial steps of DNA damage repair occur in compact chromatin domains [10,46,48]. Depletion of HP1 in cells subjected to genotoxic stress results in a delay in the repair of DSBs and elevated levels of apoptosis after irradiation [49], implicating the recruitment of HP1-β is as an important component of the repair process. The results of this study indicate that aspects of these chromatin changes also cause the masking of 5meC from immune-detectionimmunodetection without further processing by tryptic digestion. The critical components of this masking are yet to be defined. There was not, however, strict co-localization of 5meC and MBD1 at all sites across the nucleoplasm. This was particularly the case in cells were the epitope was recovered with acid treatment only. After a further brief tryptic digestion step much improved co-localization between the 5meC and MBD1 was observed. The molecular basis for this heterogeneity requires further investigation. This finding provides a caution to the use of MBD1 as a proxy measure of 5meC. It suggests that immunoprecipitation of MBD1 may be expected to pull down some regions of DNA that are not strongly methylated and conversely would fail to precipitate other heavily methylated regions of DNA. This important technical consideration indicates care and extensive validation of these strategies are required to allow valid interpretation of methylation studies. It is noteworthy that the tryptic digestion process used here to improve the localization of MBD1 in MEFs caused its almost complete degradation in other cell types (the one and two-cell embryo [15]). The observation that both 5meC and MBD1 co-stain in MEFs may suggest that steric competition between MBD1 and antibodies for the available 5meC is not a primary cause of trypsin-sensitive masking of 5meC. JustJust as the conditions required for optimal epitope retrieval vary depending upon the cell type and growth status, the same also seems to apply to other nuclear proteins such as MBD1, pointing to the requirement for careful validation of immunolocalization of the range of nuclear proteins. This study was not designed to assess whether these changes in chromatin structure were associated with specific sites of DNA damage, nor does it allow for a strict assessment of whether the changes in 5meC and heterochromatin were associated with particular stages of the DNA damage response. The staining methods validated here will allow further analysis of these questions and would be most suited to more selective models of damage such as Laser-induced linear DNA damage. This model, coupled with co-staining of markers of each stage of the damage sensing and repair processes will shed light on these aspects of the question. The marked increase in immune-detectable 5meC levels in cells with DNA damage was consistent with expected outcomes. The finding that holding MEFs in a mitotically quiescent state[12]. The

14

finding that hoingMEFs for a prolonged period was associated with a marked loss of 5meC staining was consistent with observations that proliferative MEFs had higher levels of total 5meC than quiescent cells [14]. While the heterogeneity in the 5meC levels in heterochromatic foci could conceivably reflect differences in the cell-cycle stage it is expected that all the cells in the quiescent culture will have exited the cell cycle. It has long been thought that 5meC is a relatively stable epigenetic modification within a cell lineage [50-54] but thist finding of a marked change in immunedetectableimmunodetectable 5meC levels as cells exit the cell-cycle and further changes as quiescence persists,together with the marked polymorphic nature of 5meC localization with heterochromatin adds to a growing body of evidence that dynamic changes in methylation plays important roles in governing cellular homeostasis within lineages [55-57]. This study confirms that a large but variable proportion of the nuclear 5meC is present within chromatin in a conformation that cannot be detected by conventional methods of immunolocalization. It further shows that both the overall levels and the trypsin-sensitive pools of 5meC change with time after exposure of cells to genotoxic stress. Much of the increased 5meC present after DNA damage is within the trypsin-sensitive pool and of this much is associated with the heterochromatic fraction of the cell. The study shows that the heterochromatin is polymorphic for the manner in which 5meC presented and the levels and nature of this polymorphism changes with the type of DNA damage and the time after damage is induced. The results illustrate the power of immunolocalization to detected marked changes in the three-dimensional allocation of 5meC within the genome and further highlight the critical need for thorough validation of epitope retrieval methods in the detection of nuclear antigens.

Acknowledgements This work was supported by grants from the Australian National Health and Medical Research Council to CO and a Turkish Government Postgraduate Scholarship to SC. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

15

References

1. Akhmanova A, Verkerk T, Langeveld A, Grosveld F, Galjart N (2000) Characterisation of transcriptionally active and inactive chromatin domains in neurons. Journal of Cell Science 113: 4463-4474. 2. Hajkova P, Ancelin K, Waldmann T, Lacoste N, Lange U, et al. (2008) Chromatin Dynamics during Epigenetic Reprogramming in the Mouse Germ Line. Nature 452 (7189):: 877-881. 3. Wongtawan T, Taylor JE, Lawson KA, Wilmut I, Pennings S (2011) Histone H4K20me3 and HP1 alpha are late heterochromatin markers in development, but present in undifferentiated embryonic stem cells. Journal of Cell Science 124: 1878-1890. 4. Kang YK, Koo DB, Park JS, Choi YH, Lee KK, et al. (2001) Differential inheritance modes of DNA methylation between euchromatic and heterochromatic DNA sequences in ageing fetal bovine fibroblasts. FEBS Letters 498: 1-5. 5. Kreiling JA, Tamamori-Adachi M, Sexton AN, Jeyapalan JC, Munoz-Nagar U, et al. (2011) Ageassociated increase in heterochromatic marks in murine and primate tissues. Aging Cell 10: 292-304. 6. Terzoudi GI, Hatzi VI, Donta-Bakoyianni C, Pantelias GE (2011) Chromatin dynamics during cell cycle mediate conversion of DNA damage into chromatid breaks and affect formation of chromosomal aberrations: Biological and clinical significance. Mutation ResearchFundamental and Molecular Mechanisms of Mutagenesis 711: 174-186. 7. Grigoryev SA, Nikitina T, Pehrson JR, Singh PB, Wodcock CL (2004) Dynamic relocation of epigenetic chromatin markers reveals an active role of constitutive heterochromatin in the transition from proliferation and quiescence. Journal of Cell Science 117: 6153-6162. 8. Woodcock CL, Ghosh RP (2010) Chromatin Higher-order Structure and Dynamics. Cold Spring Harbor Perspectives in Biology 2. 9. Weisman-Shomer P, Kaftory A, Fry M (1979) Replicative Activity of Isolated Chromatin from Proliferating and Quiescent Early Passage and Aging Cultured Mouse Cells. Journal of Cell Physiology 101: 219-228. 10. Baldeyron C, Soria G, Roche D, Cook AJL, Almouzni G (2011) HP1 alpha recruitment to DNA damage by p150CAF-1 promotes homologous recombination repair. Journal of Cell Biology 193: 81-95. 11. Brown SE, Weaver ICG, Meaney MJ, Szyf M (2008) Regional-specific global cytosine methylation and DNA methyltransferase expression in the adult rat hippocampus. Neuroscience Letters 440: 49-53. 12. Miller OJ, Schnedl W, Allen J, Erlanger BF (1974) 5-Methylcytosine Localized in Mammalian Constitutive Heterochromatin. Nature 251: 636-637. 13. Bridger JM, Boyle S, Kill IR, Bickmore WA (2000) Re-modelling of nuclear architecture in quiescent and senescent human fibroblasts. Current Biology 10: 149-152. 14. Çelik S, Li Y, O'Neill C (2014) The exit of mouse embryonic fibroblasts from the cell-cycle changes the nature of solvent exposure of the 5´-methylcytosine epitope within chromatin. PloS ONE 9 (4): e92523. 15. Li Y, O'Neill C (2012) Persistance of Cytosine Methylation of DNA following Fertilisation in the Mouse. PloS ONE 7 (1): e30687.16. Li Y, O'Neill C (2013) 5'-methylcytosine and 5'hydroxymethylcytosine Each Provide Epigenetic Information to the Mouse Zygote. PloS ONE 8: e63689. 17. Iqbal K, Jin SG, Pfeifer GP, Szabo PE (2011) Reprogramming of the paternal genome upon fertilization involves genome-wide oxidation of 5-methylcytosine. Proceedings of the National Academy of Sciences of the United States of America 108: 3642-3647.

16

18. McLay D, Clarke H (2003) Remodelling the Paternal Chromatin at Fertilization in Mammals. Reproduction 125: 625-633. 19. Smallwood SA, Tomizawa S, Krueger F, Ruf N, Carli N, et al. (2011) Dynamic CpG island methylation landscape in oocytes and preimplantation embryos. Nature Genetics 43: 811U126. 20. Cortazar D, Kunz C, Selfridge J, Lettieri T, Saito Y, et al. (2011) Embryonic lethal phenotype reveals a function of TDG in maintaining epigenetic stability. Nature 470: 419-U210. 21. Cortellino S, Xu JF, Sannai M, Moore R, Caretti E, et al. (2011) Thymine DNA Glycosylase Is Essential for Active DNA Demethylation by Linked Deamination-Base Excision Repair. Cell 146: 67-79. 22. Guo JU, Su YJ, Zhong C, Ming GL, Song HJ (2011) Hydroxylation of 5-Methylcytosine by TET1 Promotes Active DNA Demethylation in the Adult Brain. Cell 145: 423-434. 23. Tahiliani M, Koh KP, Shen YH, Pastor WA, Bandukwala H, et al. (2009) Conversion of 5Methylcytosine to 5-Hydroxymethylcytosine in Mammalian DNA by MLL Partner TET1. Science 324: 930-935. 24. Ito S, Shen L, Dai Q, Wu SC, Collins LB, et al. (2011) Tet Proteins Can Convert 5-Methylcytosine to 5-Formylcytosine and 5-Carboxylcytosine. Science 333: 1300-1303. 25. He YF, Li BZ, Li Z, Liu P, Wang Y, et al. (2011) Tet-Mediated Formation of 5-Carboxylcytosine and Its Excision by TDG in Mammalian DNA. Science 333: 1303-1307. 26. Maiti A, Drohat AC (2011) Thymine DNA Glycosylase Can Rapidly Excise 5-Formylcytosine and 5Carboxylcytosine Potential Implications for Active Demethylation of CpG sites. Journal of Biological Chemistry 286: 35334-35338. 27. Mone MJ, Volker M, Nikaido O, Mullenders LHF, van Zeeland AA, et al. (2001) Local UV induced DNA damage in cell nuclei results in local transcription inhibition. EMBO Reports 2: 10131017. 28. Loignon M, Drobetsky EA (2002) The initiation of UV-induced G(1) arrest in human cells is independent of the p53/p21/pRb pathway but can be attenuated through expression of the HPV E7 oncoprotein. Carcinogenesis 23: 35-45. 29. Kothapalli N, Sarath G, Zempleni J (2005) Biotinylation of K12 in Histone H4 Decreases in Response to DNA Double-Starnd Breaks in Human JAr Chorioarcinoma Cells The Journal of Nutrition 135: 2337-2342. 30. L'Ecuyer T, Sanjeev S, Thomas R, Novak R, Das L, et al. (2006) DNA damage is an early event in doxorubicin-induced cardiac myocyte death. American Journal of Physiology-Heart and Circulatory Physiology 291: H1273-H1280. 31. Ichijima Y, Sakasai R, Okita N, Asahina K, Mizutani S, et al. (2005) Phosphorylation of histone H2AX at M phase in human cells without DNA damage response. Biochemical and Biophysical Research Communications 336: 807-812. 32. McManus KJ, Hendzel MJ (2005) ATM-dependent DNA damage-independent mitotic phosphorylation of H2AX in normally growing mammalian cells. Molecular Biology of the Cell 16: 5013-5025. 33. Tu WZ, Li B, Huang B, Wang Y, Liu XD, et al. (2013) gamma H2AX foci formation in the absence of DNA damage: Mitotic H2AX phosphorylation is mediated by the DNA-PKcs/CHK2 pathway. Febs Letters 587: 3437-3443. 34. Festenstein R, Pagakis SN, Hiragami K, Lyon D, Verreault A, et al. (2003) Modulation of heterochromatin protein 1 dynamics in primary mammalian cells. Science 299: 719-721. 35. Lomberk G, Wallrath LL, Urrutia R (2006) The Heterochromatin Protein 1 family. Genome Biology 7:228 36. Lavelle C (2014) Pack, unpack, bend, twist, pull, push: the physical side of gene expression. Current Opinion in Genetics & Development 25: 74-84.

17

37. Çelik S (2015) Understanding the complexity of antigen retrieval of DNA methylation for immunofluorescence-based measurement and an approach to challenge. J Immunol Methods 416: 1-16. 38. Gustafsson J, M. O, S. M, Hoffman P (2010) Citric acid antigen retrieval (CAAR) for tryptic peptide imaging directly on archived formalin-fixed paraffin-embedded tissue. J Proteome Res 9: 4315-4328. 39. Yamashita S (2007) Heat-induced antigen retrieval: mechanisms and application to histochemistry. Progress in Histochemistry and Cytochemistry 41: 141-200. 40. Chiolo I, Tang J, Georgescu W, Costes SV (2013) Nuclear dynamics of radiation-induced foci in euchromatin and heterochromatin. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis 750: 56-66. 41. Morano A, Angrisano T, Russo G, Landi R, Pezone A, et al. (2014) Targeted DNA methylation by homology-directed repair in mammalian cells. Transcription reshapes methylation on the repaired gene. Nucleic Acids Research 42: 804-821. 42. Mortusewicz O, Schermelleh L, Walter J, Cardoso C, Leonhardt H (2005) Recruitment of DNA methyltransferase I to DNA repair sites. Proceedings of the National Academy of Sciences of the United States of America 102: 8905-8909. 43. Tan HH, Porter AG (2009) DNA methyltransferase 1 is a mediator of doxorubicin-induced genotoxicity in human cancer cells. Biochemical and Biophysical Research Communicatons 382: 462-467. 44. Fujita N, Watanabe S, Ichimura T, Tsuruzoe S, Shinkai Y, et al. (2003) Methyl-CpG Binding Domain 1 (MBD1) Interacts with the Suv39h1-HP1 Heterochromatic Complex for DNA Methylationbased Transcriptional Repression. Journal of Biological Chemistry 278: 24132-24138. 45. Fuks F, Hurd PJ, Deplus R, Kouzarides T (2003) The DNA methyltransferases associate with HP1 and the SUV39H1 histone methyltransferase. Nucleic Acids Research 31: 2305-2312. 46. Soria G, Polo S, Almouzni G (2012) The DNA damage response in chromatin: a focus on histone variants and heterochromatin proteins. Mol Cell 46: 722-734. 47. Dinant C, Luijsterburg M (2009) The Emerging Role of HP1 in the DNA Damage Response. Molecular and Cellular Biology 29: 6335-6340. 48. Luijsterburg MS, Dinant C, Lans H, Stap J, Wiernasz E, et al. (2009) Heterochromatin protein 1 is recruited to various types of DNA damage. Journal of Cell Biology 185: 577-586. 49. Lee Y-H, Kuo C-Y, Stark JM, Shih H-M, Ann DK (2013) HP1 promotes tumor suppressor BRCA1 functions during the DNA damage response. Nucleic Acids Research 41: 5784-5798. 50. Dhawan S, Georgia S, Tschen SI, Fan GP, Bhushan A (2011) Pancreatic beta Cell Identity Is Maintained by DNA Methylation-Mediated Repression of Arx. Developmental Cell 20: 419429. 51. Zachary DS, Meissner A (2013) DNA methylation: roles in mammalian development. Nature Reviews 14: 204-220. 52. Bar-Nur O, Russ HA, Efrat S, Benvenisty N (2011) Epigenetic Memory and Preferential LineageSpecific Differentiation in Induced Pluripotent Stem Cells Derived from Human Pancreatic Islet Beta Cells. Cell Stem Cell 9: 17-23. 53. Bird A (2002) DNA methylation patterns and epigenetic memory. Genes & Development 16: 6-21. 54. Wolf J (2013) Evolution of genomic imprinting as a coordinator of coadapted gene expression. Proceedings of the National Academy of Sciences of the United States of America. 110 (13): 5085-90 55. Lim ASP, Srivastava GP, Yu L, Chibnik LB, Xu J, et al. (2014) 24-Hour Rhythms of DNA Methylation and Their Relation with Rhythms of RNA Expression in the Human Dorsolateral Prefrontal Cortex. PLoS Genetics 10: e1004792. 56. Yang H, Lin H, Xu H, Zhang L, Cheng L, et al. (2014) TET-catalyzed 5-methylcytosine hydroxylation is dynamically regulated by metabolites. Cell Research 24: 1017-1020.

18

57. Kaas GA, Zhong C, Eason D, Ross D, Vachhani RV, et al. (2013) TET1 Controls CNS 5methylcytosine Hydroxylation, Active DNA Demethylation, Gene Transcription and Memory Formation. Neuron 79: (6):1086-93.

19

Table 1. Experimental Models Experimental Model

Cells used

Treatments

Model 1

Confluent MEFs (2-day culture in complete media) were serum starved for 24h (Quiescent)

12 min UV

1h in serum-free media

Model 2

Confluent MEFs (2-day culture in complete media) were serum starved for 24h (Quiescent)

15 sec UV

*

48h in complete media

Model 3

Proliferating cells (1-day culture in complete media)

Doxorubicin * (50nM)

24h in complete media

*

*

Incubation after (or with) treatment

Incubation after Cells used

Treatments (or with) treatment

Experimental Model Confluent MEFs (2-day culture Model 1

in complete media) were serum starved for 24h (Quiescent) Confluent MEFs (2-day culture

Model 2

- 12 min UV 1h in serum-free media -

- 15 sec UV

in complete media) were serum starved for 24h (Quiescent)

48h in complete media - Doxorubicin

Model 3

Proliferating cells (1-day culture

(50nM)

in complete media)

24h in complete media

-

For each model untreated control cells were run in parallel.

20

Legend Figure 1. Experimental models of DNA damage to MEFs. A-B-C Model 1. A shows the level of total γ-H2A.X staining in untreated and UV-treated (12 min) cells at 1h. B shows representative images of quiescent cells 1h after UV exposure or untreated controls. C shows the proportion (%) Model 1. The proportion (%)of cells with γ-H2A.X staining levels above background staining. Total number of nuclei analyzed are 630 and 1493 forB Two representative images of quiescent cells 1h after UV exposure or untreated and UV-treated cells, respectivelycontrols. D-E-F Model 2. D shows the level of total γ-H2A.X staining in untreated and UVtreated (15 sec) cells at 48h. E shows representative images of γ-H2A.X staining. F The proportion of cells (%) with γ-H2A.X staining above background. Total number of nuclei analyzed are 700 and 609 for untreated and UV-treated cells, respectively. G-H-I Model 3. G shows totalF-G Model 3. F Total level of γ-H2A.X staining 24h after doxorubicin treatment of cells and untreated controls. H shows representative images of γ-H2A.X staining in group. I shows the proportion (%) of cells with γ-H2A.X staining levels above background staining. Total numbers of nuclei analyzed are 426 and 447 for untreated and doxorubicin-treated cells, respectively. The results are from three independent replicates. Scale bar = 10 µm. AU - Arbitrary units of optical densitymicron. Error bars are the mean +/- s.e.m. **** p<0.001 Figure 2. The pattern of 5meC and MBD1 staining 15 min after UV (Model 1). A and B show the total amount of 5meC and MBD1, respectively. AU - Arbitrary units of optical density. C and D show the correlation between the amount of 5meC and MBD1 in untreated and UVtreated cells, respectively. r = Pearson correlation coefficient. E shows representative nuclei of untreated and UV-treated MEFs with (+) or without (-) tryptic digestion. White boxes represent MBD1 foci that do not coincide with Hoechst and 5meC foci, orange boxes represent the colocalization of MBD1 foci with 5meC and Hoechst foci. The results are from at least three independent replicates. Scale bar 5 µm. Numbers of nuclei (untreated) analyzed are 320 and 337, after acid alone and with trypsin, respectively. Numbers of nuclei (UV-treated) analyzed are 320 and 352, after acid alone and with trypsin, respectively. **** Scale bar 5 micron.**** p<0.0001 Figure 3. The pattern of 5meC and MBD1 48h after UV (Model 2). A and B show the total level of 5meC and MBD1, respectively. AU - Arbitrary units of optical density. C and D show the correlation between the amount of 5meC and MBD1 in untreated and UV-treated cells, respectively. r = Pearson correlation coefficient. E shows representative nuclei of untreated and UV-treated MEFs with (+) or without (-) tryptic digestion. White boxes represent MBD1 foci that

21

do not coincide with Hoechst and 5meC foci, orange boxes represent the co-localization of MBD1 foci with 5meC and Hoechst foci. The results are from at least three independent replicates. Scale bar 5 µm. Numbers of nuclei (untreated) analyzed are 311 and 334, after acid alone and with trypsin, respectively. Numbers of nuclei (UV-treated) analyzed are 271 and 241, after acid alone and with trypsin, respectively. **** Scale bar 5 micron.**** p<0.0001 Figure 4. The pattern of 5meC and MBD1 after doxorubicin (Model 3). A and B show the total level of 5meC and MBD1, respectively. AU - Arbitrary units of optical density. C and D show the correlation between the amount of 5meC and MBD1 in untreated and UV-treated cells, respectively. r = Pearson correlation coefficient. E shows representative nuclei of untreated and UV-treated MEFs with (+) or without (-) tryptic digestion. White boxes represent MBD1 foci that do not coincide with Hoechst and 5meC foci, orange boxes represent the co-localization of MBD1 foci with 5meC and Hoechst foci. The results are from at least three independent replicates. Scale bar 5 µm. Numbers of nuclei (untreated) analyzed are 377 and 328, after acid alone and with trypsin, respectively, and numbers of nuclei (UV-treated) analyzed are 299 and 309, after acid alone and with trypsin, respectively. ****micron.**** p<0.0001 Figure 5. The effect of DNA damage on the pattern of heterochromatin (Models 1-3). A shows three representative examples of diffuse and focal patterns of staining of the heterochromatin marker, HP1-β, and non-immune control staining. Scale bar = 10 Bar = ??µm. B, C and D show the proportion, of nuclei (%) with predominantly focal HP1-β staining (%) in models 1, 2 and 3, respectively. E, F and G show the level of HP1-β in models 1, 2 and 3, respectively. AU Arbitrary units of optical density. H, I and J show the average size of nuclei (pixels) after DNA damage of experimental models 1, 2 and 3, respectively. No difference was seen after UV in model 1 compared to untreated cells. But nuclei were bigger after DNA damage in models 2 and 3. The results are from at least three independent replicates. B,E,H: Model 1, C,F,I: Model 2, D,G,J: Model 3. * p<0.05, **** p<0.0001 Figure 6. The effect of DNA damage on the nuclear localization pattern of MBD1 and HP1-β (Model 1). Cells were double-stained for MBD1 and HP1-β and counterstained with Hoechst. A and B show two representative images for untreated and UV-treated cells, respectively. C shows the number of Hoechst-intense staining foci co-stained with HP1-β and/or MBD1 in each cell. The presence or absence of MBD1 and HP1-β staining was shown as + and -. Yellow, red and blue boxes indicate

22

representative staining of Hoechst-foci (+/-), (+/+) and (-/-), respectively. Scale bar = 10 µm.* p<0.05, ** p<0.01, **** p<0.0001 Figure 7. The effect of DNA damage on nuclear localization pattern of MBD1 and HP1-β (Model 2). Cells were double-stained for MBD1 and HP1-β and counterstained with Hoechst. A and B show two representative images for untreated and UV-treated cells, respectively. C shows the number of Hoechst-intense staining foci co-stained with HP1-β and/or MBD1 in each cell. The presence or absence of MBD1 and HP1-β staining was shown as + and -. Yellow, red and blue boxes indicate representative staining of Hoechst-foci (+/-), (+/+) and (-/-), respectively. Scale bar = 10 µm.* p<0.05, ** p<0.01, **** p<0.0001 Figure 8. The effect of DNA damage on nuclear localization pattern of MBD1 and HP1-β (Model 3). Cells were double-stained for MBD1 and HP1-β and counterstained with Hoechst. A and B show two representative images for untreated and UV-treated cells, respectively. C shows the number of Hoechst-intense staining foci co-stained with HP1-β and/or MBD1 in each cell. The presence or absence of MBD1 and HP1-β staining was shown as + and -. Yellow, red and blue boxes indicate representative staining of Hoechst-foci (+/-), (+/+) and (-/-), respectively. Scale bar = 10 µm. * p<0.05.

23

Highlights    

Immune-detectable methylcytosine exists within two pools: one retrieved after acid denaturation and the other required tryptic digestion. The size of the trypsin-sensitive pool varied in response to DNA damage, with an initial reduction in the relative size of this pool followed by a marked increased. This was associated with changes in the level of HP1-β within heterochromatic foci. This method of epitope retrieval of methylcytosine allowed detection of a remarkable dynamism in the level and location of this important epigenetic feature.

85

24

25

26

27

28

29

30

31

32