DNA double-strand breaks (DSBs) can efficiently kill cancer cells, but they can also produce unwanted chromosome rearrangements when DNA ends from different DSBs are erroneously joined. Movement of DSB-containing chromatin domains might facilitate these DSB interactions and promote the formation of chromosome rearrangements. Therefore, we analyzed the mobility of chromatin domains containing DSBs, marked by the fluorescently tagged DSB marker 53BP1, in living mammalian cells and compared it with the mobility of undamaged chromatin on a time-scale relevant for DSB repair. We found that chromatin domains containing DSBs are substantially more mobile than intact chromatin, and are capable of roaming a more than twofold larger area of the cell nucleus. Moreover, this increased DSB mobility, but not the mobility of undamaged chromatin, can be reduced by agents that affect higher-order chromatin organization.

Most anticancer therapies induce multiple DSBs to kill cancer cells, but concurrent induction of DSBs in non-tumor cells can result in chromosome rearrangements that might be a source of new therapy-related tumors in treated patients (Allan and Travis, 2005; Stephens et al., 2011). DSBs represent the most challenging type of DNA damage (Suzuki et al., 2003). A failure to rejoin DSBs leads to cell death, whereas joining of DNA ends originating from different DSBs results in structural chromosome rearrangements. The mechanisms that control the formation of chromosome rearrangements are a subject of ongoing debate. One of the favored theories postulates that there are interactions between separately generated DSBs followed by incorrect joining of DSB ends. According to this theory, DSB proximity and movement are factors that are crucial to the process.

The cellular response to DSBs starts with a complex signaling cascade leading to alterations in the organization and composition of large megabase chromatin domains surrounding the breaks (Murr et al., 2006; Rogakou et al., 1999; Ziv et al., 2006; van Attikum and Gasser, 2005). Such a large-scale reorganization could affect chromatin mobility. Experiments with yeast and mammalian cells all indicate that DSB-containing chromatin domains, referred to as ionizing-radiation-induced foci (IRIF), are mobile in the micrometer range (Aten et al., 2004; Kruhlak et al., 2006; Soutoglou et al., 2007).

An accurate description of the behavior of unrepaired DSBs in the cell nucleus, and of their mobility in particular, is of key importance in understanding how DSB interactions might be initiated. If DSB mobility does have a role in the formation of chromosome rearrangements, the ability to manipulate the movement of DSBs might reduce the dangerous side-effects of anticancer therapy. To obtain a detailed insight into the nature of DSB movement, we analyzed the mobility of IRIF in living mammalian cells and compared it with the mobility of undamaged chromatin domains, telomeres and centromeres. We also investigated the involvement of various chromatin- and repair-related processes in DSB movement with the aim of reducing DSB mobility.

Visualization of chromatin domains

Undamaged chromatin domains in U2OS cells, ~1 Mb in size, were visualized by incorporation of the fluorescent nucleotide analog Cy3–dUTP (Fig. 1A) (Pliss et al., 2009). The mobility of fluorescently labeled, capped telomeres was measured in U2OS cells that were transiently transfected with a TRF1–dsRED expression construct (Fig. 1B), and the mobility of centromeres in U2OS cells stably expressing the centromeric factor CENPB–GFP was also measured (Fig. 1C) (Shelby et al., 1996). To analyze the mobility of DSB-containing chromatin, we studied IRIF in γ-irradiated U2OS cells expressing the 53BP1–GFP fusion protein (Fig. 1D) (Bekker-Jensen et al., 2005). 53BP1–GFP forms IRIF that colocalize with γH2AX in nuclei of irradiated cells (Fig. 1E).

Fig. 1.

Visualization of intact and DSB-containing chromatin domains in U2OS cells. (A) Intact chromatin domains labeled with Cy3–dUTP (red). (B) Cells with telomeres labeled with TRF1–mCherry (red). (C) Cells with centromeres labeled with CENP-B–GFP (green). (D) Cells expressing 53BP1–GFP (green), exposed to 5 Gy of γ-radiation and fixed 30 minutes later. (E) Cells treated as in D, additionally stained for γH2AX (red). The inset shows the intensity profile of a single confocal scan measured along the white bar. Scale bar, 5 μm.

Fig. 1.

Visualization of intact and DSB-containing chromatin domains in U2OS cells. (A) Intact chromatin domains labeled with Cy3–dUTP (red). (B) Cells with telomeres labeled with TRF1–mCherry (red). (C) Cells with centromeres labeled with CENP-B–GFP (green). (D) Cells expressing 53BP1–GFP (green), exposed to 5 Gy of γ-radiation and fixed 30 minutes later. (E) Cells treated as in D, additionally stained for γH2AX (red). The inset shows the intensity profile of a single confocal scan measured along the white bar. Scale bar, 5 μm.

Functionally distinct chromatin domains have diverse mobilities

To illustrate movement of chromatin domains, we first plotted 2D trajectories of 100 randomly selected domains, which emerged from the same origin in the XY plane (Fig. 2A). Our analysis revealed that IRIF are substantially more mobile than Cy3-labeled chromatin domains (Cy3CDs) and centromeres (Fig. 2). At the end of the 60-minute observation period, the roaming range of the IRIF – expressed as mean square displacement (MSD) – had increased 1.7-fold relative to the Cy3CD MSD, and the nuclear area roamed by the DSB was 2.2-fold higher {the nuclear area is proportional to [(MSD)3/2]}. The average stepsize (ΔS) covered by the IRIF during the 2-minute intervals was ~20% larger than the ΔS of Cy3CDs (Fig. 2). IRIF mobility measured in normal human fibroblasts was also higher than the mobility of Cy3CDs, indicating that increased IRIF mobility is a general phenomenon, rather than a cell-line-specific effect (supplementary material Fig. S1C). We analyzed the mobility of foci marking DSBs induced by the topoisomerase II inhibitor etoposide (Osheroff, 1989). Etoposide, a frequently used anticancer drug, promotes the formation of chromosome rearrangements that can lead to specific types of leukemia in non-transformed cells. Etoposide-induced foci showed 2.2- and 3.2-fold increases in MSD and the nuclear area roamed, when compared with Cy3CDs (supplementary material Fig. S1). The etoposide-induced foci disappeared faster than IRIF signals and could be followed for only 40 minutes. A more detailed analysis showed that local damage did not globally influence chromatin mobility, because a dose of 10 Gy of γ-radiation did not enhance the mobility of Cy3CDs (supplementary material Fig. S2). Centromeres displayed mobility that was similar to that of Cy3CDs. Telomeres, however, were as mobile as IRIF. Previous observations have also indicated that telomeres have a relatively high mobility (Molenaar et al., 2003; de Vos et al., 2009), which might stem from the fact that ends of chromosomes are attached to the bulk of chromatin by one chromatin fiber only, in contrast to other chromatin domains. It should be noted that the mobility of telomeres in U2OS cells might be altered by activity of the alternative lengthening of telomeres (ALT) maintenance pathway, hallmarked by telomeric γ-H2AX or 53BP1 foci (Cesare et al., 2009). Although we cannot completely exclude the possibility, we did not observe the classical signs of telomere dysfunction in the cell line used for our study. Another strong indication that the telomere mobility measured in our study is not affected by their ALT-state comes from the study of de Vos and colleagues who reported the nearly identical mobility of telomeres in ECV-304, a non-ALT cell line (a derivative of the human bladder carcinoma) (de Vos et al., 2009).

Fig. 2.

Mobility of the DSB-containing and intact chromatin domains. (A) 100 randomly selected trajectories depicting movement of the indicated type of chromatin domains during 60 minutes are plotted as if they originated from the same point on the 2D plane. The color of the segments of the trajectories represents the time after the start of imaging. (B) MSD of the indicated types of chromatin domains. The bar graph shows the MSD of the respective domains averaged over the last three time points. (C) Average distance covered by the indicated type of chromatin domains per 2-minute interval. *P<0.05, assessed using the Student's t-test. Error bars represent the s.e.m. At least 30 cells and five foci per cell were analyzed per data series.

Fig. 2.

Mobility of the DSB-containing and intact chromatin domains. (A) 100 randomly selected trajectories depicting movement of the indicated type of chromatin domains during 60 minutes are plotted as if they originated from the same point on the 2D plane. The color of the segments of the trajectories represents the time after the start of imaging. (B) MSD of the indicated types of chromatin domains. The bar graph shows the MSD of the respective domains averaged over the last three time points. (C) Average distance covered by the indicated type of chromatin domains per 2-minute interval. *P<0.05, assessed using the Student's t-test. Error bars represent the s.e.m. At least 30 cells and five foci per cell were analyzed per data series.

Interestingly, of the chromatin domains that were analyzed, most MSD curves did not reach a plateau within 60 minutes, indicating that the full range of local movement is not reached within this time period. Alternatively, the fact that the MSD curves failed to reach a plateau might be explained by the long time-scale drift of large chromosome territories (Zink et al., 1998).

Some of the chromatin domains analyzed here differ in size – IRIF and centromeres are ~1 μm in diameter, and telomeres and Cy3CDs are roughly half that size – this might influence the results of a direct comparison in two ways. First, the domain size might influence the calculation of its position. However, to calculate domain positions we used the intensity-based center of gravity, a parameter that is not affected by object size. Second, the domain mobility is likely to be influenced by its size owing to physical constraints imposed by chromatin organization. However, our results indicate that, in the case of IRIF, other factors play a more substantial role. Accordingly, Fig. 2 shows that although IRIF display a mobility similar to telomeres, which are much smaller, the mobility of IRIF is greater than that of Cy3CDs.

The results of our analysis contrast with conclusions drawn by Kruhlak and colleagues (Kruhlak et al., 2006). In their experiments, the mobility of chromatin pre-sensitized by Hoechst and damaged by laser UV light was similar to that of intact chromatin when measured over a 20-minute period. This comparison was based on analysis of the dynamics of relatively large chromatin regions. Random movements of individual DSBs within these regions could have partly cancelled out, resulting in a lower overall mobility compared with the mobility of the individual IRIF analyzed in our experiments. An alternative explanation could be that the mobility of DSBs varies depending on the type of treatment used to induce DNA damage. Our observation that DSBs produced by the topoisomerase II inhibitor etoposide are substantially more mobile than DSBs induced by ionizing radiation demonstrates that this is indeed possible. Kruhlak and colleagues also published data on the mobility of 53BP1–GFP IRIF in U2OS cells (Kruhlak et al., 2006). The mobility of 53BP1–GFP IRIF reported in that study was higher than the mobility measured in our study (MSD of 0.9 μm2 after 50 minutes, compared with ~0.3 μm2, respectively), but absence of corresponding data on the mobility of undamaged chromatin in the study by Kruhlak and colleagues hampers a direct comparison with our results and conclusions.

To exclude the possibility that the observed effects are caused by fluorescence imaging conditions, we tested the toxicity of the illumination regimes. Neither fluorescence imaging nor irradiation induced substantial cytotoxicity during the observation period, as confirmed by the prolonged time-lapse phase-contrast observation of illuminated and/or irradiated cells (see Materials and Methods).

IRIF mobility depends on chromatin organization and varies with cell cycle phase

A wealth of data indicates that the presence of DSBs leads to local chromatin relaxation, possibly to enable repair factors to access the damaged DNA (e.g. Falk et al., 2007; Ziv et al., 2006). It is tempting to speculate that low-density chromatin would be more mobile than condensed chromatin. Chromatin relaxation might then provide a straightforward elegant explanation for the increased mobility of IRIF. In our experiments, decondensed euchromatic Cy3CDs in desynchronized cells displaying early-S-phase labeling patterns (Fig. 3B) were more mobile than the condensed heterochromatic Cy3CDs in cells displaying mid- or late-S-phase patterns (Fig. 3A,C). This is in agreement with previous results (Pliss et al., 2009). The affirmation that relaxation increases the mobility of undamaged chromatin motivated us to investigate the effect of chromatin condensation on IRIF movement.

In order to test the influence of chromatin organization on the mobility of DSBs, we focused on a small fraction of persistent IRIF that could be traced up to 4 hours after irradiation and on IRIF persisting 24 hours after irradiation in cells incubated with ATM inhibitor Ku55933, as these lasting DSBs are frequently associated with heterochromatic regions (Goodarzi et al., 2008). Our results show that heterochromatin-associated persistent IRIF were, indeed, significantly less mobile (~40%) compared with randomly distributed IRIF imaged early after irradiation (Fig. 3D), in agreement with the hypothesis that chromatin relaxation is one of the factors leading to enhanced IRIF mobility.

In addition to the epigenetically determined differences between euchromatin and heterochromatin, nuclear chromatin undergoes dramatic changes during progression of the cell cycle (Chuang and Belmont, 2007). Therefore, we examined the cell cycle dependence of IRIF mobility. In these experiments, we incubated cells with the thymidine analog BrdU immediately after imaging. This allowed us to discriminate between BrdU-negative G1 or G2 nuclei (Fig. 3F) and BrdU-positive S-phase nuclei (Fig. 3E). Our results show that IRIF move significantly (~35%) less in S-phase nuclei (Fig. 3G), which might be surprising in view of the replication-related unwinding and decondensation of chromatin.

The outcome of the latter experiment indicates that chromatin relaxation or decondensation is not the only factor affecting IRIF mobility. We, therefore, conclude that other types of changes in the organization or composition of chromatin might also influence IRIF mobility. Of special interest to us were changes in chromatin organization or composition that can be induced on demand.

Fig. 3.

Mobility of euchromatin versus heterochromatin domains and effects of labeling on the cell cycle. (A,B) Cy3CDs in cells labeled during mid or late (A) or early (B) S-phase and fixed 48 hours later. (C) MSD of Cy3CDs in cells labeled during mid or late, or early S-phase. Cells were analyzed 48 hours post-labeling. (D) MSD of IRIF imaged immediately after irradiation (green), of persistent IRIF imaged 4 hours after exposing cells to γ-radiation (red) and of IRIF 24 hours after irradiation (IR) in cells incubated in the presence of the ATM inhibitor Ku55933 (purple). The bar graph shows the MSD averaged over the last three time points. *P<0.05, assessed using the Student's t-test. Error bars represent the s.e.m. At least 30 cells were analyzed per data series. (E–G) Cells expressing 53BP1–GFP were irradiated, imaged for 60 minutes and then incubated for 5 minutes in the presence of BrdU to label cells undergoing DNA replication. (E) BrdU pattern of cells in S phase undergoing DNA replication. (F) No BrdU incorporation in G1, G0 or G2 cells. (G) MSD of IRIF in BrdU-positive (replicating) cells (red) and BrdU-negative (non-replicating) cells (green). The bar graph shows the MSD averaged over the last three time points. *P<0.05, assessed using the Student's t-test. Error bars represent the s.e.m. At least 30 cells and at least five foci per cell were analyzed per data series.

Fig. 3.

Mobility of euchromatin versus heterochromatin domains and effects of labeling on the cell cycle. (A,B) Cy3CDs in cells labeled during mid or late (A) or early (B) S-phase and fixed 48 hours later. (C) MSD of Cy3CDs in cells labeled during mid or late, or early S-phase. Cells were analyzed 48 hours post-labeling. (D) MSD of IRIF imaged immediately after irradiation (green), of persistent IRIF imaged 4 hours after exposing cells to γ-radiation (red) and of IRIF 24 hours after irradiation (IR) in cells incubated in the presence of the ATM inhibitor Ku55933 (purple). The bar graph shows the MSD averaged over the last three time points. *P<0.05, assessed using the Student's t-test. Error bars represent the s.e.m. At least 30 cells were analyzed per data series. (E–G) Cells expressing 53BP1–GFP were irradiated, imaged for 60 minutes and then incubated for 5 minutes in the presence of BrdU to label cells undergoing DNA replication. (E) BrdU pattern of cells in S phase undergoing DNA replication. (F) No BrdU incorporation in G1, G0 or G2 cells. (G) MSD of IRIF in BrdU-positive (replicating) cells (red) and BrdU-negative (non-replicating) cells (green). The bar graph shows the MSD averaged over the last three time points. *P<0.05, assessed using the Student's t-test. Error bars represent the s.e.m. At least 30 cells and at least five foci per cell were analyzed per data series.

Changes in IRIF mobility after treatments that affect chromatin organization and composition

Recently published data indicate that 53BP1, a protein involved in local modification of chromatin in response to DSBs, promotes interactions between DSBs during variable, diverse and joining [V(D)J] recombination (Difilippantonio et al., 2008). Moreover, the mobility of uncapped telomeres, which physiologically resemble one-ended DSBs and recruit DSB repair proteins, is decreased after 53BP1 knockdown (Dimitrova et al., 2008). Taken together, these results suggest that 53BP1 also influences the mobility of DSBs. To examine this possibility, we measured the mobility of IRIF in MDC1–GFP-expressing U2OS cells transfected with siRNA targeted against TP53BP1. Even though TP53BP1 was effectively downregulated, we observed no reduction of IRIF mobility in these cells when compared with cells transfected with scrambled siRNA (Fig. 4A), suggesting that the mobility of IRIF is regulated by mechanisms that are independent of 53BP1. Furthermore, we did not observe a change in IRIF mobility when we inhibited Tip60 (also known as KAT5), a protein also implicated in chromatin remodeling at DSB sites (Sun et al., 2006) (Fig. 4A). Even though neither an siRNA approach nor chemical inhibition can fully suppress protein activity, and we cannot completely exclude the involvement of 53BP1 or Tip60 in increased mobility of IRIF, our results did not provide any indication that inhibition of these early-repair-related-factors can reduce the mobility of IRIF.

As chromatin remodeling requires metabolic energy, we examined the effect of ATP depletion on IRIF mobility. We found that a combination of the ATP synthesis inhibitors 2-deoxyglucose and sodium azide significantly (34%) decreased the MSD of IRIF (Fig. 4B). Subsequently, when we explored the impact of transcription, a process strongly associated with chromatin remodeling, we detected only a moderate effect on the movement of IRIF. Treatment with the transcription inhibitor 5,6-dichloro-1β-D-ribofuranosylbenzimidazole (DRB) did reduce IRIF mobility but the change was not statistically significant (Fig. 4C). Several studies indicate that ATP supply and transcription influence movement of undamaged chromatin (Dundr et al., 2007; Heun et al., 2001; Levi et al., 2005; Marshall et al., 1997; Mearini and Fackelmayer, 2006). Likewise, our results indicate that energy-dependent chromatin remodeling processes can contribute to IRIF movement.

Fig. 4.

Influence of DSB-related and global chromatin modifications on IRIF mobility. (A) Left panel: the MSD of MDC1–GFP IRIF in U2OS cells transfected with scrambled siRNA or siRNA targeting TP53BP1, or incubated for 24 hours with Tip60 inhibitor anacardic acid. The bar graph shows MSD of IRIF averaged over the last three time points. Right panel: detection of 53BP1 by western blotting in cells transfected with scrambled siRNA (left lane) or siRNA targeting 53BP1 (right lane). (B) MSD of IRIF in cells exposed to 5 Gy of γ-radiation, incubated for 30 minutes in the presence or the absence of ATP synthesis inhibitors 2-deoxyglucose and sodium azide, and imaged for 60 minutes without refreshing the medium. The bar graph shows MSD of IRIF averaged over the last three time points. (C) MSD of IRIF in cells incubated for 4 hours in the presence or the absence of transcription inhibitor DRB, irradiated as in A and imaged for 60 minutes. The bar graph shows MSD of IRIF averaged over the last three time points. (D) MSD of IRIF in cells irradiated and imaged after incubation for 24 hours with 150 nM histone deacetylase inhibitor TSA, 48 hours with 0.75 μM non-methylable cytidine analog 5-azacytidine, or 1 hour with 1 μM histone acetyltransferase inhibitor curcumin. The bar graph shows the MSD of IRIF under respective conditions, averaged over the last three time points. *P<0.05, assessed using the Student's t-test. Error bars represent the s.e.m. At least 30 cells and at least five foci per cell were analyzed per data series.

Fig. 4.

Influence of DSB-related and global chromatin modifications on IRIF mobility. (A) Left panel: the MSD of MDC1–GFP IRIF in U2OS cells transfected with scrambled siRNA or siRNA targeting TP53BP1, or incubated for 24 hours with Tip60 inhibitor anacardic acid. The bar graph shows MSD of IRIF averaged over the last three time points. Right panel: detection of 53BP1 by western blotting in cells transfected with scrambled siRNA (left lane) or siRNA targeting 53BP1 (right lane). (B) MSD of IRIF in cells exposed to 5 Gy of γ-radiation, incubated for 30 minutes in the presence or the absence of ATP synthesis inhibitors 2-deoxyglucose and sodium azide, and imaged for 60 minutes without refreshing the medium. The bar graph shows MSD of IRIF averaged over the last three time points. (C) MSD of IRIF in cells incubated for 4 hours in the presence or the absence of transcription inhibitor DRB, irradiated as in A and imaged for 60 minutes. The bar graph shows MSD of IRIF averaged over the last three time points. (D) MSD of IRIF in cells irradiated and imaged after incubation for 24 hours with 150 nM histone deacetylase inhibitor TSA, 48 hours with 0.75 μM non-methylable cytidine analog 5-azacytidine, or 1 hour with 1 μM histone acetyltransferase inhibitor curcumin. The bar graph shows the MSD of IRIF under respective conditions, averaged over the last three time points. *P<0.05, assessed using the Student's t-test. Error bars represent the s.e.m. At least 30 cells and at least five foci per cell were analyzed per data series.

The above experiments do not reveal whether chromatin organization alone has the capacity to affect the mobility of IRIF. To examine this question we focused on treatments that modify the organization of chromatin, without directly interfering with DNA repair or transcription. In these studies we used agents that, among other activities, affect the organization of chromatin at a global level by inhibiting DNA methylation, or histone acetylation and/or de-acetylation. When we applied 5-azacytidine, an inhibitor of DNA methylation, or trichostatin A (TSA), an inhibitor of histone deacetylation (supplementary material Fig. S3) and curcumin, a histone acetyltransferase inhibitor (Yoshida et al., 1995), we found that all these chromatin-modifying treatments induced a significant reduction in IRIF mobility (Fig. 4D). The curcumin treatment, in particular, reduced the IRIF MSD to a level that was only 20% higher than the MSD of undamaged Cy3CDs. This effect appeared to be limited to damaged chromatin, as curcumin treatment did not change the mobility of Cy3CDs (supplementary material Fig. S2). Importantly, treatment of cells with TSA, curcumin or 5-azacytidine did not disturb the progression of the cell cycle, as measured by BrdU incorporation assay. Furthermore, we did not detect substantial changes in repair, according to the kinetics of γ-H2AX IRIF disassembly after irradiation (supplementary material Fig. S3). Together, these experiments demonstrate that the modes of IRIF- and Cy3CD-movement are different and, moreover, that IRIF mobility can be reduced on demand.

An important consequence of chromatin mobility is the fusion of IRIF (Aten et al., 2004), which brings unrepaired DNA ends from initially distant DSBs into close proximity and might thereby increase the chance of induction of chromosome rearrangement. We indeed observed occasional fusions of multiple IRIF (Fig. 5), although it cannot be confirmed whether the actual rejoining of open DNA ends takes place within fused IRIF. In order to investigate whether changes in IRIF mobility affect IRIF fusion-rate, we measured the frequencies of IRIF fusions in cells treated with curcumin, which modifies IRIF mobility. Importantly, IRIF fusion frequencies were indeed decreased in cells treated with curcumin, probably owing to decreased IRIF mobility.

Fig. 5.

Fusion of multiple IRIF. (A) Multiple IRIF occasionally fuse. The image gallery shows the fusion of two IRIF initially spaced by ~2 μm. Scale bar, 2 μm apart. (B) IRIF fusion frequency is decreased in U2OS cells incubated in the presence of curcumin. Time-lapse movies of 53BP1–GFP IRIF in control cells and in cells incubated with curcumin, exposed to 5 Gy of γ-radiation were scored ‘blindly’ by two observers. The graph represents the average number of observed IRIF fusions per cell. Error bars represent the range of frequencies obtained by the two observers. At least 100 cells were scored per data point.

Fig. 5.

Fusion of multiple IRIF. (A) Multiple IRIF occasionally fuse. The image gallery shows the fusion of two IRIF initially spaced by ~2 μm. Scale bar, 2 μm apart. (B) IRIF fusion frequency is decreased in U2OS cells incubated in the presence of curcumin. Time-lapse movies of 53BP1–GFP IRIF in control cells and in cells incubated with curcumin, exposed to 5 Gy of γ-radiation were scored ‘blindly’ by two observers. The graph represents the average number of observed IRIF fusions per cell. Error bars represent the range of frequencies obtained by the two observers. At least 100 cells were scored per data point.

Conclusions

Our results demonstrate that the presence of DSBs locally increases the mobility of chromatin, resulting in a 2.2- to 3.2-fold larger nuclear area roamed by the DSBs compared with that of intact chromatin, and that the increase depends, to some extent, on the agent used to induce the DSBs. Moreover, IRIF mobility can be reduced by exogenous agents that affect higher-order chromatin organization. Factors that reduce IRIF mobility might not affect the mobility of undamaged chromatin or uncapped telomeres, or vice-versa, indicating that IRIF movement involves additional and/or different mechanisms.

The movement of DSBs might appear to be of little consequence as the range of their movement is less than a micrometer. However, their non-directional walk can cover an area in the cell nucleus of about 1 μm3, which is a highly relevant size in the context of cancer treatment. Each treatment in a fractionated therapy produces ~100 DSBs per cell, even in healthy tissues directly surrounding the malignancy. Considering that the volume of a typical mammalian cell nucleus is about 250 μm3, a single treatment can, thus, result in multiple accidental DSB interactions leading to chromosome rearrangements. Fusions between IRIF have been reported in cells exposed to alpha particles (Aten et al., 2004) and X-rays (Falk et al., 2007). Thus, the 2.2- and 3.2-fold increases in nuclear area roamed by DSBs induced by ionizing radiation and etoposide, when compared with that of undamaged chromatin, would be expected to increase the probability of DSB interactions. Furthermore, it is feasible that the reverse process, a reduction in DSB mobility, might decrease IRIF fusion frequency and reduce DSB interactions.

To measure the mobility of the various labeled chromatin domains, we captured 3D time-lapse movies of cells at 2-minute intervals for 60 minutes and corrected the individual images for shift and rotation of cell nuclei during the imaging. Movements of chromatin domains were interpreted as a restrained random walk. We determined the average distance (ΔS) covered by the domains during the 2-minute intervals between images and the mean squared displacement (MSD) in the 3D images, as a function of time (Krawczyk et al., 2008). To compare the range of movement of chromatin domains, we used the average of the last three MSD values. Cells that displayed extensive morphing of their nuclei during the imaging period were excluded from the analysis.

Cell culture and treatments

U2OS cells were cultured in DMEM (Gibco) supplemented with 10% FCS, 200 mM L-glutamine (Gibco) and penicillin-streptomycin (Gibco) in a moist atmosphere containing 10% CO2. For Cy3–dUTP labeling, confluent cell cultures grown on coverslips were scratch-loaded in the presence of 50 μM Cy3–dUTP (Sigma-Aldrich) in PBS (Gibco). The scratching, performed using a sterile needle, temporarily damages cellular membranes, allowing nuclear penetration of Cy3–dUTP that is then incorporated into the DNA of replicating cells. For BrdU labeling, positions of all imaged cells were recorded using coverslips with etched numbered grids. Cells were incubated in the presence of 10 μM BrdU (Sigma-Aldrich) for 5 minutes immediately after live-cell imaging. Subsequently, cells were fixed, stained for BrdU and imaged using the same microscope, and recorded positions were analyzed. It was, therefore, possible to determine BrdU content in those exact cells whose live images were captured earlier. The following inhibitor concentrations were used (unless stated otherwise): TSA (Sigma-Aldrich), 150 nM, 24 hours before irradiation; 5-azacytidine (5-AzaC) (Sigma-Aldrich), 0.75 μM, 48 hours before irradiation; 5,6-dichloro-1β-D-ribofuranosylbenzimidazole (DRB) (Sigma-Aldrich), 100 μM, 4 hours before irradiation; anacardic acid (AA) (Sigma-Aldrich), 100 nM, 1 hour before irradiation; curcumin (Sigma-Aldrich), 1 μM, 60 minutes before irradiation. Cells were irradiated with a 137Cs source at a dose rate of 0.6 Gy/minute for a total dose of 5 or 8 Gy, or treated with 15 μg/ml etoposide (Sigma-Aldrich) for 15 minutes.

siRNA and western blotting

Cells were transfected with scrambled siRNA or siRNA targeting TP53BP1 (Dharmacon) using Lipofectamine 2000 (Invitrogen) and standard protocols. Western blot analysis and imaging of siRNA-transfected cells was performed 48 hours post-transfection. Cells were lysed in a buffer containing 2% SDS, 10% glycerol, 60 mM Tris-HCl pH 6.8, 50 mM dithiothreitol (DTT), protease and phosphatase inhibitors, and 0.02% Bromophenol Blue. After fractionation by SDS-PAGE, proteins were transferred on to a nitrocellulose membrane and probed with the relevant antibodies.

Time-lapse microscopy and image processing

Cells were imaged using a Leica IRBE (Leica Microsystems) inverted wide-field microscope, 63X oil objective, at 37°C in an atmosphere containing 10% CO2. Z-stacks of five images, 300 nm apart along the z-axis, were captured at 2-minute intervals. The 3D images were then reconstructed using Huygens Professional (Scientific Volume Imaging). Semi-automated 3D-image registration and object tracking was performed as described previously (Krawczyk et al., 2008). In brief, translational movement and rotation during the imaging was eliminated using a data-alignment approach based on the iterative closest point (ICP) algorithm. This method requires extraction of positions of the IRIF from the pictures using image thresholding, followed by calculation of the individual centers of gravity (the most stable parameter describing their position). To correct for the movement of the nucleus between time-point zero and a given time point, the coordinates of all centers of gravity at a given time-point are aligned with the coordinates at time-point zero using the ICP. The aligned coordinate sets are then checked for IRIF, which, within a single time-interval, moved over a distance exceeding a preset value. This apparent movement can be a consequence of optical merging or the disappearance of objects. Such events are removed from the coordinate sets and excluded from the analysis. Following alignment, the spatio-temporal properties of the IRIF in the cell nucleus can be analyzed.

Treatment toxicity analysis

Live-cell fluorescence imaging is inherently toxic and might induce artifacts. We, thus, carefully examined the effects of imaging conditions on the cell cycle and induction of cell death. We exposed cells to the fluorescence imaging conditions used throughout the study. Subsequently, we monitored the cells by acquiring phase-contrast images for up to 48 hours. We did not detect a substantial influence of the imaging procedure on the duration of the cell cycle or on the induction of cell death in the generation in which the cells were illuminated, as directly assessed from the time-lapse movies. Cell cycle duration was 24.5 hours (s.d. = 3.5 hours) and 24.4 hours (s.d. = 2.5 hours) for control cells and illuminated cells, respectively. No cell death was observed within 24 hours after illumination. Furthermore, there was very limited influence of fluorescence imaging on cells additionally exposed to 5 Gy of γ-radiation. No cell death was observed within the first 10 hours of imaging and ~3% of cells died within 24 hours. Further addition of DRB, anacardic acid, sodium azide and 2-deoxyglucose, 5-azacitidine and TSA induced 6, 7, 7, 11 and 12% cell death within 24 hours after the end of the fluorescence imaging, respectively. No signs of toxicity induced by these treatments were observed within the first 4 hours after the end of fluorescence imaging. Treatment with 1 μM curcumin for 1 hour before imaging resulted in an increase in the number and size of vacuoles ~1–3 hours after the end of imaging and clear signs of apoptotic cell death within 3–5 hours after the end of imaging. Importantly, this treatment did not induce any discernible effects on cell behavior or morphology within 1 hour after fluorescence imaging. Cell cycle analysis using BrdU incorporation revealed no changes in cell cycle distribution upon treatment of cells with either curcumin, TSA or 5-azacitidine (supplementary material Fig. S3).

Allan
J. M.
,
Travis
L. B.
(
2005
).
Mechanisms of therapy-related carcinogenesis
.
Nat. Rev. Cancer
5
,
943
-
955
.
Aten
J.
,
Stap
J.
,
Krawczyk
P.
,
van Oven
C.
,
Hoebe
R.
,
Essers
J.
,
Kanaar
R.
(
2004
).
Dynamics of DNA double-strand breaks revealed by clustering of damaged chromosome domains
.
Science
303
,
92
-
95
.
Bekker-Jensen
S.
,
Lukas
C.
,
Melander
F.
,
Bartek
J.
,
Lukas
J.
(
2005
).
Dynamic assembly and sustained retention of 53BP1 at the sites of DNA damage are controlled by Mdc1/NFBD1
.
J. Cell Biol.
170
,
201
-
211
.
Cesare
A. J.
,
Kaul
Z.
,
Cohen
S. B.
,
Napier
C. E.
,
Pickett
H. A.
,
Neumann
A. A.
,
Reddel
R. R.
(
2009
).
Spontaneous occurrence of telomeric DNA damage response in the absence of chromosome fusions
.
Nat. Struct. Mol. Biol.
16
,
1244
-
1251
.
Chuang
C.
,
Belmont
A. S.
(
2007
).
Moving chromatin within the interphase nucleus-controlled transitions?
Semin. Cell Dev. Biol.
18
,
698
-
706
.
de Vos
W. H.
,
Hoebe
R. A.
,
Joss
G. H.
,
Haffmans
W.
,
Baatout
S.
,
Van Oostveldt
P.
,
Manders
E. M. M.
(
2009
).
Controlled light exposure microscopy reveals dynamic telomere microterritories throughout the cell cycle
.
Cytometry A
75
,
428
-
439
.
Difilippantonio
S.
,
Gapud
E.
,
Wong
N.
,
Huang
C.
,
Mahowald
G.
,
Chen
H. T.
,
Kruhlak
M. J.
,
Callen
E.
,
Livak
F.
,
Nussenzweig
M. C.
, et al. 
. (
2008
).
53BP1 facilitates long-range DNA end-joining during V(D)J recombination
.
Nature
456
,
529
-
533
.
Dimitrova
N.
,
Chen
Y. M.
,
Spector
D. L.
,
de Lange
T.
(
2008
).
53BP1 promotes non-homologous end joining of telomeres by increasing chromatin mobility
.
Nature
456
,
524
-
528
.
Dundr
M.
,
Ospina
J. K.
,
Sung
M.
,
John
S.
,
Upender
M.
,
Ried
T.
,
Hager
G. L.
,
Matera
A. G.
(
2007
).
Actin-dependent intranuclear repositioning of an active gene locus in vivo
.
J. Cell Biol.
179
,
1095
-
1103
.
Falk
M.
,
Lukasova
E.
,
Gabrielova
B.
,
Ondrej
V.
,
Kozubek
S.
(
2007
).
Chromatin dynamics during DSB repair
.
Biochim. Biophys. Acta
1773
,
1534
-
1545
.
Goodarzi
A. A.
,
Noon
A. T.
,
Deckbar
D.
,
Ziv
Y.
,
Shiloh
Y.
,
Löbrich
M.
,
Jeggo
P. A.
(
2008
).
ATM signaling facilitates repair of DNA double-strand breaks associated with heterochromatin
.
Mol. Cell
31
,
167
-
177
.
Heun
P.
,
Laroche
T.
,
Shimada
K.
,
Furrer
P.
,
Gasser
S. M.
(
2001
).
Chromosome dynamics in the yeast interphase nucleus
.
Science
294
,
2181
-
2186
.
Krawczyk
P. M.
,
Stap
J.
,
Hoebe
R. A.
,
van Oven
C. H.
,
Kanaar
R.
,
Aten
J. A.
(
2008
).
Analysis of the mobility of DNA double-strand break-containing chromosome domains in living Mammalian cells
.
Meth. Mol. Biol.
463
,
309
-
320
.
Kruhlak
M.
,
Celeste
A.
,
Dellaire
G.
,
Fernandez-Capetillo
O.
,
Muller
W.
,
McNally
J.
,
Bazett-Jones
D.
,
Nussenzweig
A.
(
2006
).
Changes in chromatin structure and mobility in living cells at sites of DNA double-strand breaks
.
J. Cell Biol.
172
,
823
-
834
.
Levi
V.
,
Ruan
Q.
,
Plutz
M.
,
Belmont
A.
,
Gratton
E.
(
2005
).
Chromatin dynamics in interphase cells revealed by tracking in a two-photon excitation microscope
.
Biophys. J.
89
,
4275
-
4285
.
Marshall
W. F.
,
Straight
A.
,
Marko
J. F.
,
Swedlow
J.
,
Dernburg
A.
,
Belmont
A.
,
Murray
A. W.
,
Agard
D. A.
,
Sedat
J. W.
(
1997
).
Interphase chromosomes undergo constrained diffusional motion in living cells
.
Curr. Biol.
7
,
930
-
939
.
Mearini
G.
,
Fackelmayer
F. O.
(
2006
).
Local chromatin mobility is independent of transcriptional activity
.
Cell Cycle
5
,
1989
-
1995
.
Molenaar
C.
,
Wiesmeijer
K.
,
Verwoerd
N. P.
,
Khazen
S.
,
Eils
R.
,
Tanke
H. J.
,
Dirks
R. W.
(
2003
).
Visualizing telomere dynamics in living mammalian cells using PNA probes
.
EMBO J.
22
,
6631
-
6641
.
Murr
R.
,
Loizou
J. I.
,
Yang
Y. G.
,
Cuenin
C.
,
Li
H.
,
Wang
Z. Q.
,
Herceg
Z.
(
2006
).
Histone acetylation by Trrap-Tip60 modulates loading of repair proteins and repair of DNA double-strand breaks
.
Nat. Cell Biol.
8
,
91
-
99
.
Osheroff
N.
(
1989
).
Effect of antineoplastic agents on the DNA cleavage/religation reaction of eukaryotic topoisomerase II: inhibition of DNA religation by etoposide
.
Biochemistry
28
,
6157
-
6160
.
Pliss
A.
,
Malyavantham
K.
,
Bhattacharya
S.
,
Zeitz
M.
,
Berezney
R.
(
2009
).
Chromatin dynamics is correlated with replication timing
.
Chromosoma
118
,
459
-
470
.
Rogakou
E.
,
Boon
C.
,
Redon
C.
,
Bonner
W.
(
1999
).
Megabase chromatin domains involved in DNA double-strand breaks in vivo
.
J. Cell Biol.
146
,
905
-
916
.
Shelby
R. D.
,
Hahn
K. M.
,
Sullivan
K. F.
(
1996
).
Dynamic elastic behavior of alpha-satellite DNA domains visualized in situ in living human cells
.
J. Cell Biol.
135
,
545
-
557
.
Soutoglou
E.
,
Dorn
J. F.
,
Sengupta
K.
,
Jasin
M.
,
Nussenzweig
A.
,
Ried
T.
,
Danuser
G.
,
Misteli
T.
(
2007
).
Positional stability of single double-strand breaks in mammalian cells
.
Nat. Cell Biol.
9
,
675
-
682
.
Stephens
P. J.
,
Greenman
C. D.
,
Fu
B.
,
Yang
F.
,
Bignell
G. R.
,
Mudie
L. J.
,
Pleasance
E. D.
,
Lau
K. W.
,
Beare
D.
,
Stebbings
L. A.
, et al. 
. (
2011
).
Massive genomic rearrangement acquired in a single catastrophic event during cancer development
.
Cell
144
,
27
-
40
.
Sun
Y.
,
Jiang
X.
,
Chen
S.
,
Price
B. D.
(
2006
).
Inhibition of histone acetyltransferase activity by anacardic acid sensitizes tumor cells to ionizing radiation
.
FEBS Lett.
580
,
4353
-
4356
.
Suzuki
K.
,
Ojima
M.
,
Kodama
S.
,
Watanabe
M.
(
2003
).
Radiation-induced DNA damage and delayed induced genomic instability
.
Oncogene
22
,
6988
-
6993
.
van Attikum
H.
,
Gasser
S. M.
(
2005
).
The histone code at DNA breaks: a guide to repair?
.
Nat. Rev. Mol. Cell Biol.
6
,
757
-
765
.
Yoshida
M.
,
Horinouchi
S.
,
Beppu
T.
(
1995
).
Trichostatin A and trapoxin: novel chemical probes for the role of histone acetylation in chromatin structure and function
.
BioEssays
17
,
423
-
430
.
Zink
D.
,
Cremer
T.
,
Saffrich
R.
,
Fischer
R.
,
Trendelenburg
M. F.
,
Ansorge
W.
,
Stelzer
E. H.
(
1998
).
Structure and dynamics of human interphase chromosome territories in vivo
.
Human Genet.
102
,
241
-
251
.
Ziv
Y.
,
Bielopolski
D.
,
Galanty
Y.
,
Lukas
C.
,
Taya
Y.
,
Schultz
D. C.
,
Lukas
J.
,
Bekker-Jensen
S.
,
Bartek
J.
,
Shiloh
Y.
(
2006
).
Chromatin relaxation in response to DNA double-strand breaks is modulated by a novel ATM-and KAP-1 dependent pathway
.
Nat. Cell Biol.
8
,
870
-
876
.

Supplementary information