SEARCH

SEARCH BY CITATION

Keywords:

  • Arabidopsis;
  • BY-2 cells;
  • DNA double-strand breaks (DSBs);
  • DNA repair;
  • E2F factor;
  • H2AX;
  • repair foci

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • Cellular responses to DNA double-strand breaks (DSBs) are linked in mammals and yeasts to the phosphorylated histones H2AX (γH2AX) repair foci which are multiproteic nuclear complexes responsible for DSB sensing and signalling. However, neither the components of these foci nor their role are yet known in plants.
  • In this paper, we describe the effects of γH2AX deficiency in Arabidopsis thaliana plants challenged with DSBs in terms of genotoxic sensitivity and E2F-mediated transcriptional responses.
  • We further establish the existence, restrictive to the G1/S transition, of specific DSB-induced foci containing tobacco E2F transcription factors, in both A. thaliana roots and BY-2 tobacco cells. These E2F foci partially colocalize with γH2AX foci while their formation is ataxia telangiectasia mutated (ATM)-dependent, requires the E2F transactivation domain with its retinoblastoma-binding site and is optimal in the presence of functional H2AXs.
  • Overall, our results unveil a new interplay between plant H2AX and E2F transcriptional activators during the DSB response.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Exposure to genotoxins (i.e. DNA-damaging agents) triggers a wide range of biological signalling pathways leading to DNA damage repair, cell cycle arrest or cell death in the presence of excessive DNA damage. Proper coordination of all these responses is sustained by the repair foci which are large proteic complexes forming in the vicinity of the DNA damage site. Descriptions of these repair foci according to the type of damage have been largely documented over the last years in yeasts and animals (Lisby et al., 2004; Niida & Nakanishi, 2006).

Since the seminal work by Rogakou et al., 1999 with insect, yeast and mammalian cells, the phosphorylated histones H2AX (γH2AX) foci have been considered as solid markers of DNA double-strand breaks (DSBs) although their precise function remains a matter of debate (Lobrich et al., 2010). The formation of γH2AX repair foci occurs rapidly after DSB induction and has been proved to be mainly dependent on the kinase ataxia telangiectasia mutated (ATM) (Burma et al., 2001), even if the ataxia telangiectasia and Rad-3-related (ATR) kinase might also contribute, to a lesser extent, to this process, especially in the context of replicative stress (Ward & Chen, 2001). The loss of H2AX in mammals compromises genomic stability and γH2AX-deficient mice are radiation-sensitive as well as growth-retarded (Celeste et al., 2003a; Franco et al., 2006). In human cells, the direct interaction between γH2AX and Mediator of DNA damage Checkpoint protein 1 (MDC1) is demonstrated as critical for DNA damage checkpoint activation (Stewart et al., 2003; Stucki et al., 2005), by promoting correct accumulation of other repair proteins such as 53BP1 and BRCA1 to sites of DSBs (Bassing & Alt, 2004; Kolas et al., 2007). In addition to phosphorylation, ubiquitination and acetylation of H2AX have important effects, as they facilitate the DNA damage responses (Vissers et al., 2008; Ikura et al., 2007). The spatiotemporal dynamics of γH2AX foci has also been discussed recently, considering that these structures might vary their composition and fulfil different functions according to the cell cycle progression (Iliakis, 2010; Nakamura et al., 2010).

In plants, several components of the DSB repair network are conserved. Orthologues of ATM and H2AX have notably been characterized in Arabidopsis (Garcia et al., 2003; Friesner et al., 2005). Besides its involvement in γH2AX foci formation (Friesner et al., 2005), ATM plays a pivotal role in the robust transcriptional response induced by DSB in Arabidopsis (Chen et al., 2003; Culligan et al., 2006; Ricaud et al., 2007). Recently, the Suppressor Of Gamma response 1 (SOG1) was described as a key mediator in this response (Yoshiyama et al., 2009). For our part, we have shown that the specific DSB induction of RAD51 and TSO2, encoding proteins involved, respectively, in homologous recombination repair and dNTP supply, is dependent on E2Fa – a member of the E2F transcription factors family (Roa et al., 2009).

So far, no functional role has been ascribed to the γH2AX foci in plants, and no other DSB-induced foci-forming proteins have been mentioned. Here, through genetic, molecular and cellular approaches, we established that γH2AX foci were instrumental both in efficient DSB repair and in transcriptional responses in Arabidopsis. We also demonstrated that E2Fa and H2AX contributed additively to the global DSB repair system. Additionally, we observed – in Arabidopsis and BY-2 cells – some unprecedented E2F foci which formed specifically at the G1/S transition in response to DSBs, as long as ATM sensing and γH2AX foci formation were not impaired. Besides, these E2F foci colocalized with γH2AX foci. Altogether, our results provide new data concerning the interplay between γH2AX and E2F foci in DNA DSB repair responses.

Materials and Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Plant material, growth and treatment conditions

Wildtype (WT) and transgenic Arabidopsis thaliana (L.) Heynh. plants are of the Columbia (Col-0) ecotype. They were sterilized and grown in vitro as described previously (Roa et al., 2009). The BY-2 tobacco cell suspension was maintained by weekly subculture as described previously (Chaboutéet al., 2002). Treatments were carried out either by sowing seeds on an MS-agar medium supplemented with genotoxins or by incubating the plantlets or cells in a liquid MS medium supplemented with the chemicals. Drug concentrations were 10−5 M (unless otherwise mentioned) for bleomycin (BLM; Laboratoire Thissen, Belgium), 5 nM for camptothecin (CPT; Sigma), 10 mM for hydroxyurea (HU; Sigma) and 5 mM for caffeine (Johnson Matthey Company, Karlsruhe, Germany).

Generation of transgenic cells and plants

The sequences of the NtE2F constructs were amplified via PCR with specific primers (3 and 4 for NtE2F; 5 and 6 for the truncated version of NtE2FMU; Supporting Information, Table S1) and total cDNA from tobacco as a template. Using the Gateway® technology (Invitrogen), the amplicons were then cloned into the pK7WGF2 vector (Karimi et al., 2002) to be fused to the EGFP (Enhanced Green Fluorescent Protein) under the control of the 35S promoter. The constructs were then introduced into Agrobacterium LBA4404 for transformation of BY-2 cells as previously described (Chaboutéet al., 2000) and into GV3101 to transform Arabidopsis, using the floral-dip method (Clough & Bent, 1998). NtE2F:GFP line was crossed with previously characterized atm−/− and e2fa−/− lines (Garcia et al., 2003; Roa et al., 2009). Subsequently, all transgenic plants were screened for the selection of homozygous lines.

The premiH2AX construct was amplified using the pre-miR171 cDNA as a template (Parizotto et al., 2004) and specific primers (1–2; Table S1). It was then cloned into the pBin61 vector at the Spe I/Xho I sites. After transformation, homozygous lines were selected. The premiH2AX construct was also introgressed into the e2fa as well as NtE2F:GFP lines, and homozygous lines were used in the experiments.

Immunolabelling

Plant fixation and immunostaining were performed as described previously (Friesner et al., 2005). The primary antibodies were rabbit anti-γH2AX (Friesner et al., 2005) and polyclonal chicken anti-GFP (Molecular Probes, Eugene, OR, USA), diluted at 1 : 500. The secondary antibodies for γH2AX detection were, depending on the experiments, Alexa 546 goat anti-rabbit conjugate (Molecular Probes) for red signals and Hilyte Fluor 488 goat anti-rabbit (AnaSpec, San Jose, CA, USA) for green signals, both applied at 1 : 500. For colocalization analyses, the secondary antibody for GFP detection was Alexa 488 goat anti-chicken (Molecular Probes) applied at 1 : 200.

Microscopy

Images of Arabidopsis root tip cells and BY-2 cells were captured using a Zeiss LSM510 laser scanning microscope with a C-Apochromat (63X; v1.2 W Korr) water objective lens. Excitation/emission wavelengths were 488 nm/505–545 nm and 543/long pass 560 nm according to the fluorophores. Images were processed using the Zeiss LSM510 version 2.8 and ImageJ v.1.43 (Rasband, W.S., NIH, Bethesda, MD, USA, http://imagej.nih.gov/ij/, 1997–2011).

Histone protein extraction

Plants were ground in nuclear isolation buffer (0.25 M sucrose, 60 mM KCl, 15 mM NaCl, 5 mM MgCl2, 1 mM CaCl2, 15 mM Pipes pH 6.8, 0.8% Triton X-100) plus 0.1% of a protease inhibitor cocktail (Roche Diagnostics) and 50 mM of phosphatase inhibitor sodium ortho-vanadate (P9599, Sigma). The plant extract was filtered twice through miracloth (Calbiochem, Darmstadt, Germany). After centrifugation at 10 000 g for 20 min at 4°C, the pellet was submitted to acid extraction through resuspension in 0.4 M sulphuric acid and incubation on ice for 1 h. After centrifugation at 15 000 g for 5 min at 4°C, the soluble proteins from the supernatant were precipitated overnight with acetone at –20°C, then spun down at 7000 g for 15 min at 4°C, and resuspended in 4 M urea.

Western blots

Protein extracts were analysed on 10 or 15% sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) and electro-transferred onto a PVDF membrane (Millipore). Immunoblotting for the detection of γH2AX was carried out as previously described (Friesner et al., 2005) with some modifications: PBS was used instead of TBS and the secondary anti-rabbit antibody was linked to the anti-rabbit alkaline phosphatase (Bio-Rad). The immunodetection of GFP:E2F proteins was performed using a polyclonal rabbit anti-GFP (kindly provided by J.L. Evrard, IBMP) diluted to 1 : 5000 and the secondary anti-rabbit antibody was linked to the alkaline phosphatase. Revelation of the blots was performed with the BCIP/NBT system (Roche Applied Science).

Reverse transcription-polymerase chain reaction (RT-PCR) analysis

RNA was extracted and processed for RT-PCR reactions as described (Roa et al., 2009). Amplicons were analysed by semiquantitative PCR, using agarose gel electrophoresis and DNA fragments were quantified with the Quantity One software program (Bio-Rad). 18S and actin were used as standards. The results are presented as relative mRNA levels compared with standards. Alternatively, quantitative Real-Time PCR was performed using Qr evaluation as previously described (Roa et al., 2009).

Neutral COMET assays

Assays were performed on 8-d-old plantlets according to (Roa et al., 2009). The quantification of the comet figures was carried out on nontreated plants and BLM-treated plants, and was related to an arbitrary scoring of the comet figures as described previously (Collins, 2004). In each assay, 200 comets were scored and the results represent the mean values ± SDs from three independent experiments.

Synchronization experiments

BY-2 cells were synchronized using aphidicolin as described previously (Chaboutéet al., 2002). The mitotic index was evaluated by counting the mitotic figures in DAPI stained nuclei. In parallel, S-phase was monitored through expression analyses using the tobacco S-phase marker H3 gene as described (Proust et al., 1999). The ribosomal gene 18S was used as a standard.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

AtH2AXa-b genes are constitutively expressed and AtH2AXa-b proteins form foci upon BLM exposure

In Arabidopsis, two H2AX orthologues, H2AXa and H2AXb (Friesner et al., 2005), have been identified. Both proteins are very similar to yeast and mammalian H2AXs (68–70% of identity) as well as to each other (98.6% of identity). In their C-terminal part, they share the same SQE motif which represents a consensus phosphorylation site for ATM, or alternatively other kinases (Friesner et al., 2005).

While in mammals several functional studies had already described the relevance of H2AX in the DSB responses (Bassing et al., 2002; Celeste et al., 2003b), such an analysis had not been performed in plants so far.

We first characterized the expression of H2AXa and H2AXb and noticed that the two genes were expressed similarly in different organs, albeit with a lower level in the cauline leaves (Fig. 1a). During plant development, we also observed that the abundance for both transcripts remained almost constant even if a slight increase in H2AXa expression could be noted 15 d after germination (Fig. 1a).

image

Figure 1. Characterization of Arabidopsis H2Axs. (a) Analysis of H2AXa and H2AXb expression by semiquantitative RT-PCR in different plant organs (R, roots; Fl, flowers; Cl, cauline leaves) and at different developmental stages (7, 12 and 15 d after germination). (b) Time-course analysis using western blot of the γH2AX levels in 8-d-old plantlets upon exposure to bleomycin (BLM). The antibody used was raised against the phosphorylated C-terminal tail common to H2AXa and H2AXb. As a loading control (LC), Coomassie blue staining of the immunoblot is presented. (c) Immunolocalization of γH2AX (green) in Arabidopsis root tip cells in nontreated plants (control) and after a 2 h BLM treatment, using the same antibody as in (b). Eight-day-old plantlets were used in the experiments. Images were captured through confocal microscopy. Bars, 5μm.

Download figure to PowerPoint

Furthermore, and consistently with results obtained upon gamma irradiation (Friesner et al., 2005), the amount of γH2AX proteins was shown to increase after treatment with the DSB-inducer drug BLM in a time-dependent manner (Fig. 1b). In parallel, γH2AX foci were also detected in Arabidopsis root tips upon exposure to BLM, using immunocytological approaches (Fig. 1c).

Characterization of the miH2AX line

To assess the effect of γH2AX deficiency in Arabidopsis, it was important to obtain a decreased expression for both H2AX genes. Since no KO or knocked-down T-DNA lines were available, an RNAi approach was used. We mutated the Arabidopsis pre-mir171 cDNA (Parizotto et al., 2004) using PCR elongation with primers harbouring mismatches, in order to generate the pre-miH2AX, which we placed under the control of a 35S promoter. After transitory expression in Nicotiana benthamiana leaves, where constitutive H2AX-GFP expression was down-regulated in the presence of the pre-miH2AX (data not shown), an Arabidopsis stably transformed homozygous line (named miH2AX in the following) was selected because of its decreased H2AXa and H2AXb expression levels (of 85 and 48%, respectively, in comparison with WT) (Fig. 2a). In addition, the level of γH2AX proteins in this line was below the threshold of detection in our experimental conditions (Fig. 2b), suggesting that, besides the degradation of the H2AXa and H2AXb mRNAs, the miH2AX construct might also cause the inhibition of their translation. Finally, the percentage of cells presenting γH2AX foci were also drastically reduced in miH2AX upon BLM exposure (Fig. 2c), confirming that the miH2AX line was affected in γH2AX functions.

image

Figure 2. Characterization of the Arabidopsis miH2AX line. Experiments were performed on 8-d-old plantlets. (a) Expression levels of H2AXa and H2AXb in the miH2AX line (closed bars) compared with the wildtype (WT; open bars). Semiquantitative reverse transcription-polymerase chain reaction (RT-PCR) experiments were repeated three times, SDs are indicated. Actin was used as a standard. (b) Detection through western blot of γH2AX levels in the WT and in the miH2AX line after a 2 h bleomycin (BLM) treatment. As a loading control (LC), Coomassie blue staining of the immunoblot is presented. (c) Comparison of the percentage of cells presenting γH2AX foci (CPγF) in WT and miH2AX plants in response to BLM after a 2 h BLM treatment. Two hundred cells from 20 different root tips were analysed randomly in two independent experiments. Error bars indicate SD. In the lower part, representative confocal images are presented; arrows indicate foci. Bar, 5μm.

Download figure to PowerPoint

Phenotypically, in the absence of genotoxins, root growth was similar between miH2AX and WT plants (Fig. 3a), but appeared to be mildly delayed in miH2AX in the presence of the DSB inducers – CPT and BLM (Fig. 3b–c).

image

Figure 3. Phenotypic analyses of the Arabidopsis miH2AX line in response to double-strand breaks (DSBs). (a, b) Measurement of root growth in wildtype (WT, closed bars) and miH2AX (open bars) plants within 21 d after germination on medium without genotoxins (a) or in the presence of camptothecin (CPT, b). (c) Measurement of root growth within 10 d after germination on medium containing 10−6 bleomycin (BLM). (d) Evaluation of DSBs in COMET assays in the WT and without genotoxins (no BLM), after a single 6 h BLM treatment (BLM) and after a 6 h BLM treatment followed by a 30 min recovery period (BLM + recovery). WT, closed bars; miH2AX, open bars. Eight-day-old plantlets were used in the experiments. Each comet figure was given an arbitrary score as presented in the upper inset. The results were obtained from three independent experiments. Error bars indicate SD. (e) Evaluation of TSO2 expression by quantitative reverse transcription-polymerase chain reaction (RT-PCR) in WT and miH2AX plantlets. Relative mRNA levels correspond to standardization with 18S. Analyses were performed on total RNA extract from 8-d-old plantlets either treated with BLM (open bars, 2 h) or untreated (closed bars). The results were obtained from three independent experiments. Error bars indicate SD.

Download figure to PowerPoint

To relate this sensitivity to genotoxins with possible DSB repair defects, we evaluated DSB repair efficiency in miH2AX plants using neutral COMET assays (Fig. 3d). We first noticed that in the absence of genotoxins the basal amounts of DSBs in the miH2AX line and in the WT were similar, which a priori excluded the possibility that deficiency in H2AXs could result in higher genomic instability. Surprisingly, we also observed that the amounts of DSBs after a 6h BLM treatment were not significantly different between the miH2AX and the WT lines, possibly because of the length of the treatment or/and the high BLM concentration used in the experiment (10−5M), which may lead to saturating effects. However, when plants were submitted to a 30 min recovery after BLM treatment, the amount of remaining DSBs was twice as high in miH2AX plants as in the WT for which the level of DNA DSBs dropped significantly (Fig. 3d). Altogether, these findings strongly suggested that miH2AX plants were not more susceptible than WT to generate DSBs in response to genotoxins, but were rather impaired in their ability to repair newly emerging DSBs.

Since previous results showed that deficiency in DSB repair may be correlated to defects in transcriptional response, we further analysed the response of the DSB transcriptional marker TSO2 (Roa et al., 2009). Without genotoxins, the TSO2 mRNA level in miH2AX plantlets was not affected compared with the WT. However, when miH2AX plants faced a 2h BLM exposure, the TSO2 mRNA level was reduced by more than five times (Fig. 3e).

Overall, these results showed that the miH2AX plants were impaired not only in repair efficiency but also in the TSO2 transcriptional response induced by BLM.

Interplay between E2Fa and γH2AXs in the DSB repair response

As the DSB-induced accumulation of TSO2 mRNAs was AtE2Fa-dependent (Roa et al., 2009), we explored the genetic interaction between AtE2Fa and AtH2AXa-b in response to DSBs. To do so, the miH2AX construct was introgressed into e2fa, which is completely deficient in the DSB-induced TSO2 up-regulation (Roa et al., 2009). The resulting e2fa × miH2AX line exhibited a similar decrease in H2AXa and H2AXb expression as observed in the miH2AX line (data not shown). Singularly, the e2fa x miH2AX line was affected about twice as strongly as the single mutants miH2AX and e2fa, regarding sensitivity to CPT (measured by root growth) as well as DNA repair efficiency (measured in COMET assays) (Fig. 4a–b). Consequently, we inferred that AtE2Fa and AtH2AXa-b acted in an additive way in the DSB response.

image

Figure 4. Phenotypic comparisons between Arabidopsis miH2AX, e2fa and miH2AX × e2fa lines. (a) Measurement of root growth within 21 d after germination on medium containing camptothecin. (b) Evaluation of double-strand breaks (DSBs) in COMET assays after a 6 h bleomycin treatment followed by a 30 min recovery period. Eight-day-old plantlets were used in the experiments. Maximum damage was normalized as 100% at t = 0. Three independent experiments were performed. Error bars indicate SD.

Download figure to PowerPoint

DSB-induced cellular response of a tobacco E2F factor includes foci formation and requires its C-terminal part

To further unravel the interplay between plant E2Fs and H2AXs, we considered their subcellular localization upon exposure to genotoxins. One crucial aspect could be the possible relocalization of some E2F factors after generation of DSBs.

The Arabidopsis E2F family includes six members split into two classes. AtEFa, b and c belong to the first one and AtDEL1, 2 and 3 to the second one. Unlike AtDEL1-3, AtEFa-c possesses a transactivation (TA) domain containing a conserved retinoblastoma (Rb)-binding site which is critical for the G1/S checkpoint regulation (Mariconti et al., 2002). In addition, AtE2Fa and b are transcriptional activators, while AtE2Fc and all AtDEL factors are transcriptional repressors. In the animal E2F family, as already shown, the misexpression of one member could induce deregulation of the other members (Kong et al., 2007). To avoid such a complication in Arabidopsis with the overexpression of one member, we chose to focus on NtE2F which is the only characterized E2F factor from tobacco (Sekine et al., 1999). This factor is induced in response to high doses of UV-C in BY-2 cells (Lincker et al., 2004), possesses an operative NLS (Fig. S1a–d) and harbours structural similarities with AtE2Fa-b, including a TA domain and an Rb-binding site in its C-terminal part (Lincker et al., 2008). Moreover, the GFP-NtE2F fusion, driven by a 35S promoter, is able to transactivate the E2F-regulated RNR gene in BY-2 cells (Fig. S2a). In the Arabidopsis e2fa mutant (Roa et al., 2009), it can also restore the specific TSO2 induction in response to DSBs (Fig. S2b). We therefore concluded that the GFP:NtE2F construct was functional and that NtE2F could be considered as a typical representative of the plant E2F transcriptional activators.

When constitutively expressed in tobacco BY-2 cells, GFP-NtE2Fs were mainly nuclear in the absence of BLM (Fig. 5a). Remarkably, upon exposure to BLM, part of the GFP-NtE2Fs relocalized into discrete subnuclear bright foci (Fig. 5b). As a control, in a line constitutively expressing the single GFP, the pattern of fluorescent protein was not affected by the addition of genotoxins (Fig. 5a–b, GFP).

image

Figure 5. Characterization of the GFP:NtE2F foci. All pictures were captured using confocal microscopy. Bars, 5 μm. Eight-day-old Arabidopsis plantlets and mid-log phase BY-2 cells were used in the experiments (a, b). Localization of single GFP and of GFP:NtE2F expressed in BY-2 cells, without genotoxins (a) and after a 2 h bleomycin (BLM) treatment (b). (c) Localization of GFP:NtE2F in Arabidopsis thaliana Col-0 background, without genotoxins and after a 2 h BLM treatment. Root tip cells were analysed. (d) Western blot showing levels of GFP:NtE2F expressed in Col-0 and in e2fa backgrounds. As a loading control (LC), Coomassie blue staining of the immunoblot is presented. (e) Localization of GFP:NtE2F in Arabidopsis e2fa background, without genotoxins and after a 2 h BLM treatment. Root tip cells were analysed. (f, g) A truncated version of the NtE2F gene lacking the C-term domain was fused to the GFP and expressed in Col-0 plants. (f) The percentage of cells presenting foci (CPF) was evaluated in the GFP:NtE2FMU line vs the GFP:NtE2F line for both control (black bar) and BLM (grey bar) treatments. A total of 200 cells from 10 different root tips were randomly screened in two independent experiments. Error bars indicate SD. (g) Western blot showing levels of GFP:NtE2F and GFP: NtE2FMU. As a LC, Coomassie blue staining of the immunoblot is presented.

Download figure to PowerPoint

When constitutively expressed in an Arabidopsis Col-0 (GFP:NtE2F) background, GFP:NtE2F proteins were mainly detected in the highly dividing cells of the root tips; but they were considerably less expressed in differentiated root cells, suggesting a negative control of GFP:NtE2F in these non-proliferative cells. Moreover, in the absence of BLM, the subcellular localization of the GFP:NtE2F proteins was homogeneously nuclear, while in the presence of BLM, part of them relocalized into foci (Fig. 5c). For a 2h long treatment with BLM, we recorded about 2.3 foci per cell in the GFP:NtE2F line. This result was congruent with our independent quantification of γH2AX foci in the same genotoxic conditions, suggesting that the two processes might be related. In a second experiment, the GFP:NtE2F construct was introgressed into the e2fa background (e2fa x GFP:NtE2F) and a line was selected, in which the GFP:NtE2F proteins were present at a level similar to that in Col-0 (Fig. 5d). In this line, as in Col-0, we observed that the GFP:NtE2F proteins were mainly nuclear in the absence of BLM but relocalized into foci upon BLM treatment (Fig. 5e).

Altogether, using both homologous and heterologous systems, these findings tended to confirm that the formation of E2F foci was an aspect of the DSB cellular response conserved in plants.

Interestingly, a truncated version of GFP:NtE2F (called GFP:NtE2FMU), lacking the TA domain of NtE2F, exhibited no foci at all in response to BLM (Fig. 5f), although this construct was expressed at a level similar to GFP:NtE2F (Fig. 5g). We therefore concluded that the sequence which encompasses the TA domain as well as the regulating Rb-binding site was critical for GFP:NtE2F accumulation within foci.

Cellular response of E2F is cell cycle-regulated

Curiously, in both Arabidopsis and tobacco cells, the GFP-NtE2F foci were mainly observed in dividing cells (mid-log phase cells for BY-2 and meristematic tissue of the root tip for Arabidopsis). More precisely, only 7% of the observed Arabidopsis root meristematic cells and only 15% of the observed BY-2 mid-log phase cells presented GFP-NtE2F foci in response to BLM. Moreover, in Arabidopsis root meristem, the percentage of cells presenting foci (CPF) increased with the time of exposure to BLM (twice as many cells when the BLM treatment rises from 2 to 6 h) (Fig. 6a). Likewise, the foci number per cell (FNC) increased by c. 34% between 2 and 6 h (Fig. 6b). By contrast, when a 4 h period of recovery was allowed after a 2 h BLM treatment, the percentage of CPF as well as the FNC decreased significantly (Fig. 6a,b). As a whole, these findings were interesting because they did not only substantiate the fact that the temporal dimension of the GFP:NtE2F foci formation matched the typical dynamics of repair foci, but they also implied that the formation of these foci might be cell cycle-regulated.

image

Figure 6. GFP:NtE2F foci formation is cell cycle-regulated. (a) Percentage of cells presenting foci (CPF) in the root meristem of 8-d-old GFP-NtE2F Arabidopsis plantlets, after a 2 h bleomycin (BLM) treatment (2H), a 6 h BLM treatment (6H) and a 2 h BLM treatment followed by a 4 h recovery period (4HR). A total of 200 cells from 10 different root tips were randomly screened in two independent experiments. Error bars indicate SD. (b) Evaluation of foci number per cell (FNC) in the root meristem of GFP-NtE2F, in the same conditions as (a). (c) Evolution of the percentage of CPF after a 2 h BLM treatment in different fractions of a synchronized transgenic BY-2 cell culture (lower panel). As a control, the value of CPF for a nontreated fraction is presented. Synchronization was monitored with the percentage of cells in mitosis (upper panel) and with H3 expression levels (middle panel). AP, aphidicholin.

Download figure to PowerPoint

To test the latter hypothesis, we analysed the GFP:NtE2F foci formation during the cell cycle in synchronized BY-2 cells expressing GFP:NtE2F. Cell cycle progression was monitored through mitotic index evaluation as well as the analysis of the S-phase H3 gene expression. For each cell cycle stage, a sample of the synchronized cells was treated with BLM for 2 h, then the number of GFP:NtE2F foci was determined. Strikingly, these foci formed preferentially in the G1 phase or at the G1/S transition. They were also present at a significantly lower level in the S phase but were practically absent at the G2/M transition (Fig. 6c). This clearly showed that the E2F foci were part of the DSB response essentially in the cellular context of the G1/S transition.

NtE2F foci are ATM-dependent and colocalize with γH2AX foci

Ataxia telangiectasia mutated and ATR are known to be the key sensors of the DSB signalling. Using caffeine as an inhibitor of both ATM and ATR kinases (Sarkaria et al., 1999), we could no longer detect GFP:NtE2F foci in Arabidopsis root tips upon BLM exposure (Fig. 7a-caffeine). Interestingly, with another genotoxin – hydroxyurea (HU), triggering an ATR-dependent DNA damage signalling (Culligan et al., 2004) as well as in the crossed line atm× GFP:NtE2F challenged with BLM – we could not observe any GFP:NtE2F foci (Fig. 7a-HU and 7a-atm). Altogether, these results strongly suggested that GFP:NtE2F foci formation was mainly ATM-dependent.

image

Figure 7. GFP-NtE2F foci are ataxia telangiectasia mutated (ATM)-dependent and colocalize with γH2AX foci. All pictures were captured using confocal microscopy. Bars, 5 μm. Eight-day-old Arabidopsis plantlets were used in the experiments (a) Localization of GFP-NtE2F, after a 2 h bleomycin (BLM) treatment (GFP-NtE2F BLM), after a 2 h BLM treatment in the presence of caffeine (GFP-NtE2F BLM + caffeine), after a 2 h hydroxyurea treatment (GFP-NtE2F HU) and after a 2 h BLM treatment in the GFP-NtE2F × atm line (GFP-NtE2F × atm BLM). (b) Immunodetection of GFP:NtE2F foci (green), γH2AX foci (red) and colocalization (merge). Left and right panels are two examples from two independent experiments.

Download figure to PowerPoint

In other organisms, several ATM-dependent mediators and transducers of the DSB signalling were identified through their ability to colocalize and/or interact with γH2AX foci, and so we investigated a possible colocalization of γH2AX foci with NtE2F foci, using immunolocalization experiments on Arabidopsis root tips exposed to BLM.

We first noticed that only a fraction of the cells displaying γH2AX foci also presented GFP:NtE2F foci. This was actually in agreement with the results establishing that NtE2F foci occurred exclusively in the G1 phase and at the G1/S transition, while γH2AX foci were detectable in all phases of the cell cycle. Then, in this fraction of cells presenting GFP:NtE2F and γH2AX foci, we showed that both colocalized in response to BLM (Fig. 7b). However, this colocalization was only partial (80% of the foci), suggesting that the formations of NtE2F and γH2AX foci were not completely synchronous.

To check if such colocalization might be related to a direct interaction between GFP:NtE2F and AtH2AXs, we performed yeast-two hybrid (Y2H) assays as well as immunoprecipitations using the NBS1/Mre11 interaction as a positive control (Waterworth et al., 2007). However, no direct interaction between GFP:NtE2F and AtH2AXs could be detected (data not shown). Since in mammals NBS1 had been described as directly interacting with γH2AX (Kobayashi et al., 2002), we also tested possible interactions between AtH2AXs and the AtNBS1 candidate (Waterworth et al., 2007). But again, no interaction was detected (data not shown).

Overall these results suggested either that the interactions may require intermediate factors or post-translational modifications unavailable in our experimental conditions (for instance, we could not determine if AtH2AX proteins were proficiently phosphorylated upon genotoxic exposure in yeasts) or, alternatively, that the mechanisms of interactions within repair foci in plants were distinguishable from other eukaryotes.

The NtE2F × miH2AX line displays fewer E2F foci but rescues the TSO2 transcriptional dampening

As NtE2F foci colocalized with γH2AX foci, we wondered whether H2AX was important for the recruitment of NtE2F in the DSB-induced foci. So we introgressed the miH2AX construct into Arabidopsis GFP:NtE2F and selected a homozygous line (named NtE2F × miH2AX in the following) for its H2AXa-b down-regulation, comparable to the one in miH2AX (data not shown). Even if the level of GFP:NtE2F was not affected in NtE2F × miH2AX (Fig. 8a), we observed an 80% decrease in the number of cells presenting NtE2F foci after a 2 h BLM treatment (Fig. 8b). It was thus clear that AtH2AXa-b were necessary for proper GFP:NtE2F foci formation.

image

Figure 8. Analysis of the GFP:NtE2F × miH2AX line. Eight-day-old plantlets were used in the experiments. (a) Western blot showing GFP:NtE2F levels in Arabidopsis GFP:NtE2F and NtE2F × miH2AX lines. As a loading control (LC), Coomassie blue staining of the immunoblot is presented. (b) Comparison of the percentage of cells presenting GFP:NtE2F foci (CPF) after a 2 h bleomycin (BLM) treatment in the GFP:NtE2F and the GFP:NtE2F × miH2AX lines. A total of 200 cells from 10 different root tips were randomly screened in two independent experiments. (c) TSO2 expression after a 2 h BLM treatment in the GFP:NtE2F × miH2AX line compared with the WT, GFP:NtE2F and miH2AX lines. Relative mRNA levels correspond to fold changes in gene expression in treated plants compared with nontreated plants. Results were obtained from three independent semiquantitative reverse transcription-polymerase chain reaction (RT-PCR) experiments. Error bars indicate SD. 18s was used as a standard.

Download figure to PowerPoint

Intriguingly the overexpression of GFP-NtE2F in the Col-0 background did not impact the TSO2 response to BLM in comparison to WT. However, in the NtE2F × miH2AX line, the BLM-induced TSO2 expression, which was dampened in the miH2AX line, was partially restored (Fig. 8c). These results suggested that the overexpression of the NtE2F transcriptional activator was sufficient to overcome the TSO2 transcriptional decrease in miH2AX and that this did not need NtE2F foci formation.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Relevance of H2AX in the plant DSB responses

The first objective of this study was to investigate the effects of H2AX’s deficiency in Arabidopsis. AtH2AX genes exist in two highly redundant copies and are more or less constitutively expressed in all plant organs (Fig. 1a). Even if this might be interpreted as a clue to their importance, we clearly established that the down-regulation of their expression levels (85% for H2AXa, 48% for H2AXb) had only mild impacts on the plant’s ability to cope with DNA DSBs. Indeed, without stress, miH2AX plants grew like the WT and, in the presence of DSB-inducer drugs such as BLM or CPT, they met moderate growth delay (Fig. 3a–c). Further COMET assays also showed that impaired recovery from DSBs, rather than higher genomic instability, might account for this phenotype (Fig. 3d). These results do not help to distinguish individual AtH2AXa from AtH2AXb, and the study of KO lines may be necessary for a complete understanding of the AtH2AXs functions; yet, they are consistent with reports stating that in mammals, H2AX contributes to DSB repair in a moderate way only (Pinto & Flaus, 2010).

Moreover, we demonstrated that the typical DSB induction of TSO2 disappeared in miH2AX (Fig. 3e). Given that our previous data showed the dependency of TSO2 up-regulation on AtE2Fa (Roa et al., 2009), this fact was of special interest because it entailed a possible epistatic pathway between AtE2Fa and AtH2AXs for the TSO2 phenotype. Strikingly, the study of an e2fa × miH2AX line revealed that, in addition to their common action for TSO2 regulation, AtE2Fa and AtH2AXs also contributed additively to the global DSB responses (Fig. 4).

On the other hand, it is known that proper activity of TSO2 is required for correct checkpoint activation and DNA repair after exposure to UV-C (Wang & Liu, 2006). The depletion of TSO2 transcripts might thus explain the phenotypes observed in miH2AX.

Overall, our results are evidence for a direct link between Arabidopsis H2AX and E2Fa proteins, with repercussions on the transactivation of an E2F-regulated gene. However, a more thorough survey of the transcriptional responses in the miH2AX line, targeting genes other than TSO2, should provide valuable supplementary information about the role of plant H2AXs in this domain, especially regarding the SOG1 functions (Yoshiyama et al., 2009).

GFP-NtE2F relocalization revealed the existence of plant E2F foci which are part of the plant DSB signalling

For a more extensive analysis of the interplay between plant H2AXs and E2Fs, we showed that a tobacco E2F factor fused to GFP relocalized upon BLM exposure to nuclear foci in A. thaliana root tip cells as well as in BY-2 cells (Fig. 5a–e). The number of these foci actually matched the number of γH2AX foci we could independently observe in the same experimental conditions. In comparison to previous results, the number of these foci corresponded to that obtained after a 2.5 Gy dose of gamma irradiation (Friesner et al., 2005), although the latter results were acquired on mitotic cells only, while ours were based on interphasic cells.

We further demonstrated that in Arabidopsis, GFP-NtE2F foci formation was time-dependent (Fig. 6a,b) and mainly controlled by ATM (Fig. 7a). In a previous report, Friesner et al., 2005 had shown that, in mitotic cells, ATR was responsible for 10% of the γH2AX foci, but our results here, with the atm mutant and the HU (Fig. 7a), made very unlikely any role of ATR in the NtE2F foci formation, even if we cannot totally exclude a possible implication of ATR for the few NtE2F foci observed in the S phase (Fig. 6c).

Our results also revealed that functional γH2AX foci (Fig. 8b) as well as the C-terminal part of NtE2F were required for foci formation (Fig. 5f). These findings provide new insight into the E2F’s properties which are important for foci formation. In mammals, for instance, it has not been demonstrated that the DNA repair-induced E2F1 foci are ATM- and H2AX-dependent, or that the TA domain of E2F1 is important for its accumulation within foci. Instead, it has been suggested that E2F1 foci formation is dependent on interaction with TopBP1 via its N-terminal part (Stevens & La Thangue, 2004). Here, by contrast, we present results which strongly suggest that the ability of plant E2F to form foci in response to DSBs is directly correlated to its C-terminal part, which includes the TA domain and the Rb-binding site. In addition, we can also note that NtE2F possesses a canonical SQD/E motif in its TA domain, suggesting that an ATM-dependent phosphorylation of the E2F factor might be a prerequisite for foci formation. As supporting information, two-dimensional gel showed that indeed, upon BLM exposure, the GFP-NtE2Fs underwent a post-translational modification, which, however, could not be specified (Fig. S3, Methods S1).

Intriguingly, the NtE2F foci were apparently not involved in the regulation of TSO2 expression, as in NtE2F × miH2AX the TSO2 up-regulation was restored with no foci formation (Fig. 8b,c). This finding established, a priori, an independence between the E2F functions in transcriptional responses and within the foci. However, we could not exclude the possibility that the recruitment of other E2F factors to the repair foci might affect their transcriptional activities. It had been shown in mammals that the E2F1 foci formation at stalled replication forks correlated with the repression of E2F1 functions, including S-phase entry, transactivation and apoptosis (Liu et al., 2003). The same process might occur in plants and, in this context, the major G1/S occurrence of the NtE2F foci in our experiments (Fig. 6c) could imply that the recruitment of E2F factors into repair foci coincides with the inhibition of the proliferative functions and is therefore instrumental in the implementation of the G1/S checkpoint. Consistent with this hypothesis was the diminution, in miH2AX and upon BLM exposure, of the TSO2 messengers, which might be indicative of a leaky G1/S checkpoint in response to DSBs. Besides, S-phase γH2AX foci were recently observed in a relationship with an ATR-dependent DNA replication checkpoint (Amiard et al., 2010).

Additionally, we observed that the formation of NtE2F foci required γH2AX foci (Fig. 8b) and that both significantly overlapped in Arabidopsis root tip cells (Fig. 7b). Similar colocalization was already observed in mammals with E2F1, which promoted the recruitment to the DSBs of other repair factors such as NBS1 (Chen et al., 2011) known to directly interact with γH2AX (Kobayashi et al., 2002). However, in our experimental conditions, we could not bring to light any interactions between NtE2F, AtNBS1 and AtH2AXs, suggesting some differences between plants and mammals in the setting of the DSB-induced repair foci machinery. It is also noteworthy that in mammals, the interaction between MDC1 and γH2AX is considered a major process of the DSB response (Kinner et al., 2008), but that no homologue of MDC1 has yet been identified in plants.

Interestingly, Shechter et al. (2009) had already put forward the suggestion that the absence of GKK residues in the vicinity of the AtH2AXs SQE motif argued in favour of some differences between plant cells and other eukaryotes regarding the mechanisms of interactions within DSB repair foci.

In conclusion, with the miH2AX and GFP:NtE2F lines, this study has offered interesting tools to help untangle the complex interplay between plant H2AXs and plant E2F factors encompassing a TA domain. It has emphasized the functions of Arabidopsis H2AXs in the context of genotoxin stress and demonstrated the existence of DSB-induced E2F foci in plants. In future, the identification of partners recruited to γH2AX and E2F foci, as well as the elucidation of their post-translational modifications, should allow a wider understanding of the DSB cellular response in plants and of its difference with other eukaryotes.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

This research was supported in part by an Action Concertée (Biologie Moléculaire Cellulaire et Structurale), a grant from the Ministère de l’Education Nationale et de la Recherche. We thank Dr D. Inzé for providing GFP gateway vectors; Dr E. Herzog for the PCK-GFP vector; Dr J. D. Friesner and Prof. A. B. Britt for the anti-γH2AX antibody; Drs O. Voinnet and E. Parizotto for their advice on the miRNA strategy; and Dr C. E. West for providing NBS1 clone. We are grateful to Dr. B. Winsor and L. Blech for their critical reading of the manuscript. The inter-institute confocal microscopy platform was co-financed by the Centre National de la Recherche Scientifique, the University of Strasbourg, the Région Alsace, the Association de la Recherche sur le Cancer and the Ligue Nationale contre le Cancer.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Fig. S1 Nuclear localization of NtE2F is assumed by its N terminal NLS.

Fig. S2 The GFP:NtE2F fusion is functional.

Fig. S3 Post-translational modification of GFP:NtE2F in BY-2 cells treated with BLM.

Table S1 Pairs of primers used in RT-PCR or cloning

Methods S1 Preparation of nuclear extracts and two-dimensional electrophoresis.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office.

FilenameFormatSizeDescription
NPH_4062_sm_FigS1-S3.pptx1140KSupporting info item
NPH_4062_sm_TableS1_MethodsS1.docx17KSupporting info item