To test whether DNA replication stress can induce soluble H2AX, we used the human osteosarcoma cell line U-2 OS that, unlike GIST cells, does not depend on a single oncogenically activated tyrosine kinase such as KIT and has bona fide normal DNA damage response pathways. Cells were transfected with constructs encoding GFP-tagged H2AX, mutant GFP-H2AX-S139A or empty GFP vector 20. Expression of overexpressed H2AX proteins was distinguishable from endogenous H2AX based on the increased molecular weight (Fig. 1A). After transient transfection (24 h), a considerable amount of GFP-tagged H2AX proteins were detected in the soluble nuclear compartment (Fig. 1B; lanes S3, untreated, middle and right panels). When these cells were treated with aphidicolin, a DNA polymerase α inhibitor that induces stalling of replication forks, a rapid increase of soluble proteins was detected in the S3 fractions within 1 h (Fig. 1B). The ribonucleotide reductase inhibitor hydroxyurea (HU), which induces replication stress through nucleotide depletion, had a similar effect (data not shown). Treatment with aphidicolin or HU also led to an increase of endogenous H2AX in the soluble fractions (Fig. 1B). These results highlight previous findings in yeast showing excess levels of free histones in cells in which DNA replication and histone deposition are uncoupled 171819. Transfection with mutant H2AX in which the phosphorylation site at S139 has been abolished (H2AX-S139A) had a similar effect as wild-type H2AX, suggesting that the increase of soluble H2AX is not dependent on this phosphorylation site. To test whether our results allow generalization, we transiently transfected the human lung cancer cell line H1299 with GFP-tagged H2AX, H2AX-S139A and empty vector. Treatment with aphidicolin for 1 h, also led to an increased fraction of soluble H2AX indicating that his response is not cell type specific (Additional File 1).

To further corroborate our results, U-2 OS cell populations stably expressing GFP-tagged H2AX, H2AX-S139A or empty vector were generated. In stable cell populations, H2AX was undetectable in S3 fractions from untreated cells (Fig. 1C). However, treatment with aphidicolin for 1 h rapidly resulted in an increase of soluble H2AX and H2AX-S139A (Fig. 1C; S3 fractions).

Taken together, these results show that transient stalling of DNA replication forks is associated with an increase of soluble histone H2AX.

We next tested whether an increase of soluble H2AX can lead to enhanced apoptosis. As shown in Fig. 1B, transient overexpression (24 h) of H2AX is associated with a proportion of H2AX remaining in the soluble nuclear fraction. At the same time, we detected a significant 3.3-fold increase of apoptotic cells (5.9%; p ≤ 0.001) in comparison to empty vector controls (1.8%; Fig. 2A,B). Apoptotic cells were identified based on the characteristic nuclear morphology including DNA condensation and fragmentation and only GFP-positive, transfected cells were counted (Fig. 2B). Likewise, overexpression of mutant H2AX-S139A also caused a 3.8-fold increase of apoptotic cells (6.9%; p ≤ 0.01; Fig. 2A,B). Hence, apoptosis induction does not critically depend on serine 139 phosphorylation, which is underscored by the finding that cells transfected with H2AX-S139A lacked immunostaining for γ-H2AX but still underwent apoptosis (Fig. 2B; bottom panels).

To determine whether replication stress sensitizes H2AX-overexpressing cells to undergo apoptosis, we used U-2 OS cell populations stably expressing either empty GFP vector, GFP-tagged H2AX or GFP-H2AX-S139A. Cells were treated with aphidicolin for 24 h and the proportion of apoptotic cells was assessed (Fig. 2C). A significant 2.2-fold increase of apoptotic U-2 OS cells stably expressing GFP-H2AX was detected from 4.5% in DMSO-treated controls to 9.7% in aphidicolin-treated populations (p ≤ 0.05). In addition, a significant 9.2-fold increase of apoptotic cells was detected in U-2 OS cells stably expressing GFP-H2AX-S139A from 0.5% in DMSO-treated controls to 4.6% in aphidicolin-treated cells (p ≤ 0.0001). No apoptotic cells were found in populations expressing empty GFP vector, regardless of the drug treatment (Fig. 2C).

We tested next whether overexpression of core histone H2A was able to stimulate apoptosis. Transiently transfected cell populations were stained with an anti-myc-tag antibody and assessed for apoptotic nuclear morphology. The proportion of transfected cells was similar to H2AX-transfected populations. No enhanced cell death was detected when myc-tagged H2A was overexpressed under transient conditions in comparison to empty myc-tag vector (Fig. 2D,E).

Taken together, these results indicate that overexpression of histone H2AX, but not histone H2A, can sensitize cells to undergo apoptosis.

In order to determine the mechanisms that underlie H2AX induced cell death, we performed a morphological analysis of U-2 OS cells transiently transfected with empty GFP vector, GFP-H2AX or GFP-H2AX-S139A (Fig. 3A). Excessive amounts of histones have previously been shown to result in non-specific binding to DNA and/or chromatin leading to structural alterations 1621. In transient transfection experiments, we detected an increase of cells with aberrant chromatin aggregation from 4% in empty vector controls to 17.2% in H2AX-transfected cells (4.3-fold; p ≤ 0.001) and 14.8% H2AX-S139A-transfected cells (3.7-fold; p ≤ 0.001; Fig. 3A,B). U-2 OS cells transiently transfected with core histone H2A showed a moderate increase of chromatin aggregation from 5.5% in empty vector controls to 10.4% in H2A-transfected cells that was not statistically significant (Fig. 3C). Chromatin aggregation induced by H2AX was not equivalent to apoptosis since cells with chromatin aggregation did not show TUNEL positivity as exemplified in Fig. 3D (bottom panels).

We lastly asked whether chromatin aggregation induced by overexpression of histone H2AX can lead to a downregulation of active gene transcription 16. The majority of cellular mRNA transcription is carried out by RNA pol II. The C-terminal domain (CTD) of RNA pol II consists of tandem heptad repeats with the consensus sequence Y₁S₂P₃T₄S₅P₆S₇ that become hyperphosphorylated and hypophosphorylated during the transcription cycle. Phosphorylation at serine 5 is associated with transcription initiation whereas serine 2 phosphorylation is predominantly found during transcription elongation 22. Active gene transcription can be visualized by immunofluorescence microscopy against RNA pol II S2 and typically results in a fine granular nuclear staining pattern (Fig. 4A; top panels) 2324. The transcriptional co-activators CREB-binding protein (CBP) and p300 regulate RNA pol II-mediated gene transcription by connecting sequence-specific transcription factors to the basal transcription machinery 25. These proteins display a similar staining pattern during ongoing transcription as RNA pol II. Transcriptional inhibition has been shown to induce a redistribution of the S2 phosphorylated elongating form of RNA pol II (pRNA pol II S2) into speckled nuclear domains and to lead to RNA pol II degradation 2627. Likewise, CBP and p300 are redistributed into nuclear speckles upon transcriptional inhibition 27.

We detected a significant increase of cells with redistribution of pRNA pol II S2 into nuclear speckles or loss of pRNA pol II S2 expression below detection level by immunofluorescence microscopy in cells ectopically expressing H2AX (2.9-fold; 36.5%; p ≤ 0.05) or H2AX-S139A (2.7-fold; 34.2%; p ≤ 0.005) when compared to controls (12.5%; Fig. 4A,B). In the vast majority of these cells, nuclear redistribution of RNA pol II S2 coincided with chromatin aggregation (Fig. 4A; middle panels). No significant effect on pRNA pol II S2 distribution was detected when H2A-transfected cells were analyzed (Fig. 4A,B). In addition to pRNA pol II S2, we found the transcriptional co-activator CBP to be displaced from chromatin in cells transfected either with H2AX (3.2-fold increase, 36%; p ≤ 0.05) or H2AX-S139A (2.1-fold increase, 24%; p ≤ 0.01) in comparison to empty vector controls (11.4%; Fig. 4D,E). Similar results were found when cells were analyzed for expression of p300 (Fig. 4F). Furthermore, pRNA pol II S2 was displaced from chromatin after treatment with aphidicolin as shown by immunoblotting further corroborating the results obtained by immunofluorescence staining (Fig. 4G). Since transcription is required for cell viability, our results suggest that the ability of H2AX to abrogate ongoing transcription may sensitize cells to undergo apoptosis.

U-2 OS osteosarcoma cells and H1299 lung cancer cells were obtained from ATCC (Manassas, VA) and maintained in Dulbecco's modified Eagle medium (DMEM; Cambrex, Baltimore, MD) supplemented with 10% fetal bovine serum (FBS; Mediatech, Herndon, VA), 50 U/ml penicillin and 50 μg/ml streptomycin (both BioWhittaker, Baltimore, MD). Cells were transfected with an empty GFP vector or N-terminal GFP fusion plasmids encoding wild-type H2AX and H2AX in which serine 139 has been substituted by alanine (H2AX-S139A)20 using Fugene 6 (Invitrogen, Carlsbad, CA). In addition, empty myc-tag vector or myc-H2A was used for transient overexpression experiments. Stable cell populations were generated by selecting cells in G418-containing media for 48 h.

Whole cell lysates were prepared by scraping cells into RIPA buffer (1% NP-40, 50 mM Tris-HCl pH 8.0, 100 mM sodium fluoride, 30 mM sodium pyrophosphate, 2 mM sodium molybdate, 5 mM EDTA, 2 mM sodium orthovanadate) containing protease inhibitors (10 μg/ml aprotinin, 10 μg/ml leupeptin, 1 μM phenylmethylsulfonyl fluoride). Suspensions were incubated for one hour at 4°C and cleared by centrifugation for 30 min at 14,000 rpm at 4°C. Protein concentrations were determined by the Bradford assay (Biorad, Hercules, CA). 30 μg of protein were loaded on a 4–12% Bis-Tris gel (Invitrogen, Carlsbad, CA) and blotted onto a nitrocellulose membrane. Primary antibodies used were directed against GFP (Lab Vision, Fremont, CA), H2AX (Bethyl Laboratories, Montgomery, TX), ORC2 (BD Biosciences Pharmingen), phosphorylated RNA pol II S2 (Covance, Berkeley, CA) and actin (Sigma, St. Louis, MS).

For immunofluorescence micoscopic analyses, cells were grown on coverslips, washed in PBS, fixed in 4% paraformaldehyde/PBS (15 min), permeabilized with 1% Triton-X 100/PBS (15 min) and blocked in 10% normal donkey serum (Jackson Immunoresearch, West Grove, PA) for 15 min at room temperature. After incubation with primary antibodies overnight at 4°C in a humid chamber, cells were incubated with FITC-conjugated or Rhodamine Red-conjugated secondary antibodies (Jackson Immunoresearch, 2 hours at 37°C) and counterstained with DAPI (Vector Laboratories, Burlingame, CA). For double-immunofluorescence analysis, cells were stained for the first antigen followed by a second primary antibody and an AMCA-labeled secondary antibody (Jackson Immunoresearch). Primary antibodies used for immunofluorescence stainings were directed against γ-H2AX (Trevigen, Gaithersburg, MD), myc-tag (Cell Signaling, Danvers, MA), phosphorylated RNA pol II S2 (Covance, Berkeley, CA), CBP and p300 (both Santa Cruz, Santa Cruz, CA). Cells were analyzed using an Olympus AX70 epifluorescence microscope equipped with a SpotRT digital camera.

Subcellular fractionation was performed according to Mendez and Stillman 39. Cells were trypsinized, collected by centrifugation, washed with PBS, resuspended in Buffer A (10 mM HEPES, pH 7.9, 10 mM KCl, 1.5 mM MgCl₂, 10 mM NaF, 340 mM sucrose, 10% glycerol, 1 mM DTT, 0.1 mM PMSF, 1 mM sodium orthovanadate, 0.1% Triton-X 100, 5 μg/ml leupeptin 5 μg/ml aprotinin) and incubated on ice for 5 minutes. The nuclei (P1) were collected by low speed centrifugation at 1,300 g and lysed in Buffer B (3 mM EDTA, 0.2 mM EGTA, 1 mM DTT, 0.1 mM PMSF, 1 mM sodium orthovanadate, 5 μg/ml leupeptin, 5 μg/ml aprotinin) for 10 minutes on ice. The insoluble chromatin fraction (P3) was collected by centrifugation at 1,700 g. The supernatant contained the soluble nuclear proteins (S3). The chromatin pellet (P3) was resuspended in SDS sample buffer and the DNA was sheared by sonication in a 550 Sonic Dismembrator (Fisher) at 15% amplitude.

Apoptotic cells were visualized using the In Situ Cell Death Detection Kit (Roche Applied Sciences, Indianapolis, IN) according to manufacturer's instructions.

Statistical significance was assessed using Student's t test for independent samples. P values ≤ 0.05 were considered significant.