- Email: [email protected]
Epigenetic aspects of HP1 exchange kinetics in apoptotic chromatin
Biochimie 95 (2013) 167e179
Contents lists available at SciVerse ScienceDirect
Biochimie journal homepage: www.elsevier.com/locate/biochi
Research paper
Epigenetic aspects of HP1 exchange kinetics in apoptotic chromatin a Legartová a, Al ta Jugová a, Lenka Stixová a, Stanislav Kozubek a, Miloslava Fojtová a, Son zbe b k Zdráhal , Gabriela Lochmanová b, Eva Bártová a, * Zbyne a b
Institute of Biophysics, Academy of Sciences of the Czech Republic, v.v.i., Královopolská 135, 612 65 Brno, Czech Republic Core FacilityeProteomics, Central European Institute of Technology, Masaryk University, Kamenice 753/5, Brno, Czech Republic
a r t i c l e i n f o
a b s t r a c t
Article history: Received 14 June 2012 Accepted 10 September 2012 Available online 27 September 2012
Apoptotic bodies are the most condensed form of chromatin. In general, chromatin structure and function are mostly dictated by histone post-translational modifications. Thus, we have analyzed the histone signature in apoptotic cells, characterized by pronounced chromatin condensation. Here, H2B mono-acetylation, and H3K9 and H4 acetylation was significantly decreased in apoptotic cells, which maintained a high level of H3K9 methylation. This phenotype was independent of p53 function and distinct levels of anti-apoptotic Bcl2 protein. Interestingly, after etoposide treatment of leukemia and multiple myeloma cells, H3K9 and H4 hypoacetylation was accompanied by increased H3K9me2, but not H3K9me1 or H3K9me3. In adherent mouse fibroblasts, a high level of H3K9me3 and histone deacetylation in apoptotic bodies was likely responsible for the pronounced (w40%) recovery of GFP-HP1a and GFP-HP1b after photobleaching. HP1 mobility in apoptotic cells appeared to be unique because limited exchange after photobleaching was observed for other epigenetically important proteins, including GFPJMJD2b histone demethylase (w10% fluorescence recovery) or Polycomb group-related GFP-BMI1 protein (w20% fluorescence recovery). These findings imply a novel fact that only certain subset of proteins in apoptotic bodies is dynamic. Ó 2012 Elsevier Masson SAS. All rights reserved.
Keywords: Apoptosis Histones HP1 Epigenetics Chromatin
1. Introduction Despite the fact that apoptotic chromatin is fragmented into large or oligonucleosomal fragments, apoptotic bodies represent the most condensed form of chromatin [1e3]. Typical apoptotic DNA fragmentation is mediated by caspase-activated DNase (CAD) and possibly other endonucleases [4]. It is well known that chromatin condensation occurs during cell differentiation, mitosis, and apoptosis. However, one question that remains is whether identical histone post-translational modifications occur on chromatin during all cellular processes that require chromatin compaction. In lymphoma cells, it has been observed that apoptosis is accompanied by the deubiquitination of histone H2A. Phosphorylation of histones H1, H2A, and H2B also occurs in apoptotic cells [5e7]. Interestingly, Wu et al. [8] found that histones are released from apoptotic nucleosomes. After specific pro-apoptotic stimuli, histones H2A, H2B, H3, and H4 are released from apoptotic DNA in Jurkat and HeLa cells, but this event was not observed in apoptotic NIH3T3 cells. Wu et al. [8] explained that CAD function is
* Corresponding author. Tel.: þ420 5 41517141; fax: þ420 5 41240498. E-mail address: [email protected] (E. Bártová). 0300-9084/$ e see front matter Ó 2012 Elsevier Masson SAS. All rights reserved. http://dx.doi.org/10.1016/j.biochi.2012.09.027
responsible for histones being released from the DNA of apoptotic cells and that the separation of histones from DNA is not simply a result of the DNA fragmentation process. The Wu et al. [8] study raised the question of whether apoptotic bodies are formed only from DNA or also contain some residual histones that maintain their post-translational modifications. From this view, it was observed that apoptotic nucleosomal fragmentation appears in parallel with phosphorylation of H2B. In addition, apoptotic chromatin condensation was accompanied by decreased H3 phosphorylation and the sites where phosphorylation occurs do not match the phosphorylation pattern of histones that accompanies mitotic chromosome condensation. Hurd et al. [9] have reported increased phosphorylation of threonine 45 of histone H3 (H3T45) in apoptotic cells. These authors suggest that the nucleosomal position of H3T45 is responsible for the structural changes within the nucleosomes that lead to the formation of DNA nicks and/or DNA fragmentation. This event is mediated by protein kinase C-delta that governs H3T45ph, in vitro and in vivo [9]. Despite this fact, it is not completely clear whether specific post-translational modifications of histones regulate apoptotic events, such as chromatin margination, or if complex histone signature is responsible for maintenance of apoptotic bodies as shown by Hurd et al. [9] for H3T45 phosphorylation. In the present study, we endeavored to characterize the
168
S. Legartová et al. / Biochimie 95 (2013) 167e179
histone signature of apoptotic bodies. Because heterochromatin protein 1 a (HP1a) partially maintains its diffusion ability in apoptotic cells [10], we also addressed HP1 kinetics and the HP1related histone signature in apoptotic cells. We also tested the levels of histone acetylation, H3K4 methylation, H3K9 methylation and H3K27 trimethylation in apoptotic cells. This was done in parallel with fluorescence recovery after photobleaching (FRAP) analysis of HP1a, HP1b proteins, H3K9-demethylase JMJD2b, and Polycomb group-related protein BMI1. Our results reveal the apoptosis-specific epigenetic features that are likely responsible for the maintenance of highly condensed chromatin of apoptotic bodies. 2. Materials and methods 2.1. Cell cultivation Human leukemia cells K-562, HL-60, and U-937 (obtained from European Collection of Cell Cultures, UK) and plasmocytoma ARH-77 cells were cultivated in RPMI medium supplemented with 10% fetal bovine serum (FBS); (PAN, Germany). Human multiple myeloma U266 cells (#ACC 9) (purchased from DSMZ, Braunschweig, Germany) were seeded at a density of 2 105 cells/ml and cultured in RPMI1640 medium supplemented with 10% FBS (PAN, Germany), 100 IU/ml penicillin, and 0.1 mg/ml streptomycin at 37 C in a humidified atmosphere of 5% CO2/95% air. Multiple myeloma MOLP-8 cells (DSMZ, Braunschweig, Germany) were grown in RPMI1640 medium supplemented with 20% FBS, 100 U/ml penicillin, and 0.1 mg/ml streptomycin at 37 C in a humidified atmosphere containing 5% CO2/95% air. Cells were treated for 24 h with an inhibitor of histone deacetylases, Trichostatin A (100 nM TSA, dissolved in ethanol), and with pro-apoptotic etoposide (4 h of 15 mM for MOLP8; 4 h of 75 mM for U-937; 8 h of 100 mM for ARH-77; 24 h of 200 mM for K-562, 4 h of 5 mM for HL-60; 4 h of 100 mM for U-266; and 24 h of 10 mM or 50 mM for MEFs). Concentrations of the pro-apoptotic agent etoposide were chosen according to 50% cell viability and induction of apoptosis, verified by lamin B fragmentation or according to morphology A-type lamins. Protein levels were analyzed by western blots and normalized to the total protein levels. 2.2. Western blotting After washing in PBS, the cells were lysed in sodium dodecyl sulfate (SDS) lysis buffer (50 mM TriseHCl, pH 7.5; 1% SDS; 10% glycerol) and further processed as in Jugová et al. [11]. Protein concentration was measured with a mQuant spectrophotometer. The following antibodies were used for western blots: antiH3K4me1 (#ab8895, Abcam); anti-H3K4me2 (#07-030, Upstate); anti-H3K4me3 (#ab8580, Abcam); anti-H3K9me1 (#ab9045, Abcam); anti-H3K9me2 (#07-212, Upstate); anti-H3K9me3 (#05-1250, Upstate); anti-H3K9 acetylation (#06-942, Upstate); anti-H4 acetylation (#382160, Calbiochem, Merck-Millipore); antiHP1a (#05-689, Upstate); anti-HP1b (MAB3448, clone 1MOD-1A9, Merck-Millipore); anti-H3K27me3 (#07-449, Upstate); antiSuv39h (#ab38637, Abcam); anti-p53 (#NCL-p53 CM1, Novacastra); anti-pRb1 (#sc-50, Santa Cruz Biotechnology); anti-lamin B (#sc-6217, Santa Cruz Biotechnology); anti-lamin A/C (#sc-7293, Santa Cruz Biotechnology); anti-BMI1 (#05-637, Upstate, USA); and antibody against JMJD2b (#ab-91549, Abcam); antibody against caspase 3 (#sc-7148, Santa Cruz Biotechnology); anti-caspase 8 (#9746, Cell Signaling Technology, Inc.); anti-caspase 9 (#9508, Cell Signaling Technology, Inc.); anti-Bcl2 (#sc-509, Santa Cruz Biotechnology). Dilutions were 1:1000e1:2500. The secondary antibodies were anti-rabbit IgG (A-4914, SigmaeAldrich, Czech Republic; dilution 1:2000), anti-mouse IgG (A-9044, Sigmae Aldrich, Czech Republic; dilution 1:2000), anti-goat IgG (A-4174,
Sigma-Aldrich, Czech Republic, dilution 1:2000), and anti-mouse IgG1 (sc-2060, Santa Cruz Biotechnology; dilution 1:500). 2.3. Immunofluorescence After fixation in 4% formaldehyde, for 10 min at room temperature (RT), cells were permeabilized with 0.2% Triton X-100 for 10 min, followed by 0.1% saponin (SigmaeAldrich, Czech Republic) for 12 min, and washed twice in PBS for 15 min. BSA (1%) dissolved in PBS was used as a blocking solution. Slides with cells were incubated in BSA for 1 h, washed for 15 min in PBS, and then incubated with primary antibodies. Appropriate secondary antibodies were used [12]. Image acquisition was performed by TSC SP5-X confocal microscopy (Leica Microsystems, Mannheim, Germany). For fluorochrome excitation, we used the white light laser (WLL, 470e670 nm in 1 nm increments) connected to the Leica confocal system. 2.4. Cell transfection Mouse embryonic fibroblasts (MEFs) were transfected by plasmids encoding GFP-HP1a, GFP-HP1b, GFP-BMI1, and GFP-JMJD2b (for detailed information, see Refs. [13,14]). A-type lamins were visualized by cell transfection by plasmid DNA encoding mCherrylamin A (pBABE-puro-GFP-wt-lamin A was purchased from Addgene [#17662] and it was re-cloned by dr. Oskar Laur to pBABE-puromCherry vector). Plasmid DNA was transformed into competent Escherichia coli DH5a and isolated following Stixová et al. [13]. Plasmid DNA (1e5 mg) was then used for transfection using the METAFECTENEÔPRO system (Biontex Laboratories GmbH, Germany). 2.5. Fluorescence recovery after photobleaching (FRAP) The FRAP mode of the Leica TSC SP5-X (Leica Microsystems, Mannheim, Germany) was used for these experiments. We used the argon laser (488 nm) for photobleaching and subsequent visualization of the cell nuclei. For analysis, we used immersion objective, 63, numeric aperture (N.A.) ¼ 1.4. Cell cultures were cultivated in microscopic Petri dishes (MatTek Corporation, P50G-0-30-F). For FRAP, the dishes with the cells were placed in a cultivation hood (EMBL Heidelberg, Germany) to guarantee optimal cultivation conditions (i.e., 37 C, 5% CO2 and optimal humidity). Pre-bleaching was performed for 5 intervals, bleaching 2 intervals, and postbleaching fluorescence intensity was measured for 30 intervals. The interval value was 0.67 s (determined by software, according to selected resolution 512 512 pixels). GFP fluorescence was bleached to the background level (the background value was subtracted from experimental data) and recovery of fluorescence was monitored using the LEICA LAS AF software tool for FRAP. 2.6. Histone extraction Histones were isolated as described earlier by Bonenfant et al. [15] with minor modifications. Briefly, the cells were washed twice with ice-cold PBS, resuspended in extraction buffer (80 mM NaCl, 20 mM EDTA, 1% Triton X-100, 45 mM sodium butyrate, and 100 mM PMSF), incubated for 20 min on ice, and centrifuged at 2000 g for 8 min. Pellets were resuspended in 900 ml ice-cold 0.2 M H2SO4 and incubated for 2 h with shaking at 4 C. After centrifugation at 16,000 g, supernatants were precipitated in trichloroacetic acid at a final concentration of 25% and incubated for 30 min on ice. After centrifugation at 5000 g for 30 min at 4 C, the pellet was washed with 50 mM HCl in acetone and then with acetone, and subsequently dissolved in water. The protein concentration was determined by the Bradford protein assay (Bio-Rad).
S. Legartová et al. / Biochimie 95 (2013) 167e179
169
Fig. 1. Morphology of non-apoptotic and apoptotic cells. Apoptosis was induced by etoposide and TSA in A. multiple myeloma MOLP-8 cells and B. leukemia U-937 cells. Transmission light mode from confocal microscope Leica TCS SP5-X was used for image acquisition. Selected cells are magnified in individual figures and scale bars are shown for each panel. C. The number of apoptotic cells was counted on cell morphology in non-treated and etoposide- or TSA-treated MOLP-8 and U-937 cells.
2.7. Analysis of histone acetylation status by MALDI-TOF MS Separation of histone extracts and matrix-assisted laser desorption/ionisationetime of flight mass spectrometry (MALDIe TOF MS) analysis of histone fractions were performed as described previously [16] with minor modifications. Briefly, separation of
histones was performed on a Dionex Ultimate 3000 highperformance liquid chromatography (HPLC) system using an Agilent mRP-C18 High-Recovery Protein Column (0.5 100 mm, macroporous C18-bonded, 5-mm silica particles). Histone extracts (3 mg) were injected onto the column in solvent A and separated using a multi-step acetonitrile gradient at 70 C and a constant flow
170
S. Legartová et al. / Biochimie 95 (2013) 167e179
Fig. 2. Western blot data for epigenetic patterns in non-apoptotic and apoptotic cells. A. Lamin B, H3K9ac, H3K9me1/me2/me3, H3K4me1/me2/me3, H3K27me3, Suv39h, p53, and pRb1 levels were studied in plasma cell leukemia line ARH-77, leukemia HL-60, K-562 and U-937 cell lines, multiple myeloma U-266 and MOLP-8 cells. B. H4 acetylation was studied in selected cell lines (MOLP-8, ARH-77, K-562, U-937 and MEFs). Leukemia and multiple myeloma cell lines were treated by etoposide as written in Materials and methods and MEFs were treated by 50 mM etoposide. Protein levels were normalized to the total protein levels. Label “C” shows non-treated control cells; “ETOP” is abbreviation for etoposide treatment and “TSA” is abbreviation for Trichostatin A treatment.
rate of 9 ml/min. The concentration of solvent B was increased in the following manner: 0% over 8 min, from 36 to 60% over 30 min, 80% over 9 min [Solvent A: 2% acetonitrile (v/v), 0.1% trifluoroacetic acid (v/v) in water; solvent B: 80% acetonitrile (v/v), 0.1% trifluoroacetic acid (v/v)]. Particular histone fractions were collected by a Probot LC Packings Micro Fraction Collector, dried in a SpeedVac SPD 111V (Savant), dissolved in 15 ml 100 mM ammonium acetate at pH 4, and digested by endoproteinase Glu-C (Sequencing grade, Roche Diagnostics) at an enzyme ratio of 1:25 at 25 C for 90 min. Glu-C digests were analyzed with MALDIeTOF MS on an Ultraflex III mass spectrometer (Bruker Daltonik, Bremen, Germany). Mass spectra were acquired in linear positive mode (25 kV acceleration voltage) with
500 laser shots. External calibration of the mass spectra was performed using standard peaks of Escherichia coli DH5a. A sample in a volume of 0.6 ml was mixed with 2.4 ml matrix solution [20 mg/ml DHB in an acetonitrile:water:TFA (80:19:1, v/v/v) mixture]. Flex Analysis 3.0 software (Bruker Daltonik) was used for data processing. The relative abundance of the individual modified form was determined as the ratio of its single-ion signal intensity to the sum of the intensities of all relevant variants (0Ac, 1Ac, and 2Ac; see Supplement S1; Ac indicates histone acetylation) within the spectrum of the N-terminal fragment. A total of 15 spectra accumulations per sample were acquired; the Grubbs test for outliers was performed and the mean of relative abundance and 95 confidence
S. Legartová et al. / Biochimie 95 (2013) 167e179
171
Fig. 3. Immunofluorescence detection of H3K9 acetylation (ac) and H3K9 dimethylation (me2) in non-apoptotic and apoptotic U-937 cells. A. H3K9ac in control (non-treated), etoposide, and TSA treated U-937 cells. B. H3K9me2 in control (non-treated), etoposide, and TSA treated U-937 cells. Intact control cell and apoptotic cells, characterized by apoptotic bodies, are indicated by frames. Scale bars in microns are shown for each panel.
172
S. Legartová et al. / Biochimie 95 (2013) 167e179
intervals of the ratios were calculated. The significance of differences in acetylation status between controls and etoposide-treated samples was statistically evaluated by paired t-tests (P < 0.05). 3. Results 3.1. Analysis of the induction of apoptosis and the specific histone signature in apoptotic cells We induced apoptosis with etoposide and an inhibitor of histone deacetylases (HDACs), TSA. Etoposide and TSA treatment led to distinct morphologies in the entire cell population of selected cell lines, MOLP-8 (Fig. 1A) and U-937 (Fig. 1B), but both agents function as pro-apoptotic inducers (see Figs. 1C and 2A, lamin B fragmentation). In addition, we found that etoposide induced apoptosis in approximately 50% of the cell population, whereas TSA was substantially less efficient in the induction of morphologically
recognized apoptosis (Fig. 1C). Thus, for subsequent studies we used etoposide as a powerful inducer of programmed cell death. We first characterized apoptosis on the population level using western blots to detect changes in the histone signature induced by the pro-apoptotic agents. We studied H3K9 and H4 acetylation, H3K9me1/me2/me3, H3K4me1/me2/me3, K3K27me3, and the level of SUV39h histone methyltransferases (Fig. 2A). We also analyzed the levels of tumor suppressors p53 and pRb1. The effects of etoposide and TSA were studied in several leukemia cell lines (U-937, HL-60, K-562) and multiple myeloma cells (U-266, MOLP-8 and ARH-77). In all cell types tested, we observed that etoposide treatment reduced H3K9 acetylation, which in U-937, U-266, and MOLP-8 cells was accompanied by increased H3K9me2 levels (Fig. 2A). However, H3K9me1 and H3K9me3 levels were not changed in comparison with relevant non-treated controls following either treatment (Fig. 2A). Conversely, TSA, which also induced apoptosis as characterized by lamin B fragmentation,
Fig. 4. Immunofluorescence detection of H3K9 acetylation (ac) and H3K9 dimethylation (me2) in non-apoptotic and apoptotic MOLP-8 cells. A. H3K9ac in control, etoposide, and TSA treated MOLP-8 cells. B. H3K9me2 in control, etoposide, and TSA treated MOLP-8 cells. Apoptotic cells, characterized by apoptotic bodies, are indicated by frames. C. Magnified images from panels A and B, showing H3K9ac and H3K9me2 in apoptotic bodies. Possible release of histones (greeneyellow signals) is shown as (1) and residual H3K9ac histones in one apoptotic body are shown as (2). High levels of H3K9me2 (3) are shown in several apoptotic bodies of MOLP-8 cells. Scale bars in microns are shown for individual panels.
S. Legartová et al. / Biochimie 95 (2013) 167e179
173
caused H3K9 hyperacetylation (Fig. 2A). However, the hyperacetylation was not a characteristic of apoptotic cells, which lacked H3K9 acetylation even during TSA-induced cell death (see morphology of apoptotic cells in Fig. 3A). As shown by immunofluorescence in U-937 and MOLP-8 cells treated by TSA, apoptotic bodies were characterized by a low level of H3K9 acetylation and relatively high levels of H3K9me2 (Fig. 3A and B or 4AeC). Therefore, a high level of H3K9 acetylation, determined by western blots in entire cell population, is preferentially induced in non-apoptotic cells after TSA treatment (see Figs. 3A and 4A, intact cells with strong green fluorescence signal after TSA treatment). Moreover, our analysis shows that while some histones are likely released during the apoptotic process (Fig. 4C1, green spots in panel 1), some remain components of apoptotic bodies (Fig. 4C, panel 2 and panel 3; e.g. greeneyellow signals inside the apoptotic bodies). Additional western blots detected identical levels of H3K4me1/ me2/me3, SUV39h, and H3K27me3 in control, and etoposide- and TSA-treated cells (Fig. 2A). None of these events were influenced by p53 status, because similar trends in histone signature were observed in p53 mutant leukemia (HL-60, U-937 cells) and p53 positive myeloma cells (Fig. 2A). The level of H4 acetylation was also studied by western blots and we observed decreased levels of this histone marker after etoposide treatment in MOLP-8, K-562, U-937 cells and mouse embryonic fibroblasts (MEFs). An exception was ARH-77 multiple myeloma cells, which were characterized by an identical level of H4 acetylation in non-apoptotic and apoptotic etoposide-treated cells (Fig. 2B). 3.2. Apoptotic pathways studied in etoposide and TSA treated cells We next looked for distinctions in the apoptotic pathways induced by etoposide and TSA in p53 negative (U-937) and p53 positive (MOLP-8) cells. Western blots showed that both cell lines had increased levels of the 17 kDa fragment of caspase 3 after etoposide or TSA treatment. In comparison with control, the 43 kDa and 18 kDa fragments of caspase 8 were more robust after TSA or etoposide treatment in U-937 cells. Similarly, high level of 18 kDa fragment was found in MOLP-8 cells after TSA treatment (Fig. 5). Only subtle distinctions were found in the level of caspase 9 between etoposide and TSA treated cells. However, p53 mutant U-937 cells were characterized by a high level of Bcl2 protein while p53 positive MOLP-8 cells had subtle Bcl2 fragments (see Fig. 5). Both etoposide and TSA induced strong gH2AX positivity (Fig. 5), which is a characteristic marker of DNA damage that can be induced by pro-apoptotic agents. The appearance of gH2AX was likely associated with apoptotic DNA fragmentation. 3.3. Epigenetic features and fluorescence recovery after photobleaching of HP1a, HP1b, JMJD2b, and BMI1 proteins in apoptotic MEFs We studied the kinetics of HP1a, HP1b, JMJD2b, and BMI1 proteins in non-apoptotic mouse embryonic fibroblasts (MEFs) (as a control) and during apoptosis (induced by etoposide) using the FRAP technique. Instead of cell growing in suspension, we chose MEFs for these experiments because adherent cells are more suitable for FRAP analysis. We asked whether epigenetic events, including H3K9 deacetylation, H3K9 methylation and additionally H3K27me3 in apoptotic bodies, are accompanied by specific behaviors of proteins that recognize these histone modifications. We first confirmed the hypoacetylation and H3K9 hypermethylation in apoptotic MEFs (Figs. 6A and 7AeE). We also determined the patterns of H3K9ac, H3K9me2, H3K9me3, and H3K27me3 in pre-apoptotic and apoptotic MEFs (Figs. 6 and 7). The
Fig. 5. Proteins of apoptotic pathways analyzed in U-937 and MOLP-8 cells. Western blots of caspase 3, caspase 8, caspase 9, Bcl2 protein, and phosphorylated histone H2AX (gH2AX) in control, etoposide- and TSA-treated U-937 and MOLP-8 cells.
apoptotic bodies or their peripheries were completely stained by antibody against H3K9me3; however, H3K9me2 was absent from apoptotic bodies (compare Figs. 6B and 7AeE). H3K27me3, a BMI1 binding partner [17], was also positioned at the periphery of apoptotic bodies (compare Fig. 7F control intact cell nuclei with panel 7G showing H3K27me3 on the periphery of apoptotic bodies) or removed from apoptotic chromatin (Fig. 7G, bottom panels). The epigenetic events observed on the single cell level were accompanied by decreased levels of BMI1 protein in the apoptotic cell population, as revealed by western blots (Fig. 8A). This was again accompanied by reduced levels of H3K9 acetylation in apoptotic cells (Fig. 6A or 8A, second line in the western blot panel). Additionally, H4 acetylation was reduced in MEFs after etoposide treatment (Fig. 2B) that caused A-type lamin degradation, which is specific for apoptotic events (Fig. 8B,C). However, western blots showed that the levels of other epigenetic markers (H3K9me2, H3K9me3, H3K27me3, HP1a, and HP1b) were not changed by etoposide treatment. One exception was the increased levels of H3K36me3 in MEFs exposed to a high concentration of etoposide (Fig. 8A). We next performed FRAP analysis for JMJD2b, BMI1, HP1a, and HP1b, proteins in non-apoptotic and apoptotic cells (Fig. 9A; see also pictorial illustration of the FRAP procedure in Supplement S2 and S3). We observed significantly reduced fluorescence recovery for JMJD2b (Fig. 9Ba; w10% recovery) and BMI1 proteins (Fig. 9Bb; w20% recovery) in etoposide-induced apoptotic cells (Fig. 9B; red diamonds). Conversely, HP1a and HP1b recovered more rapidly, to 40% of pre-bleaching levels, in apoptotic cells (Fig. 9Bc and Bd; red squares). However, the HP1 recovery time after photobleaching was more variable in comparison with JMJD2b (compare standard errors in panel Fig. 9Ba with data in panels Fig. 9Bc and Bd). This correlates well with the pattern of H3K9me3
174
S. Legartová et al. / Biochimie 95 (2013) 167e179
Fig. 6. Immunofluorescence detection of H3K9 acetylation (ac) and H3K9 dimethylation (me2) in non-apoptotic and apoptotic MEFs. A. H3K9ac in non-apoptotic and apoptotic MEFs. Cell nuclei, characterized by apoptotic chromatin margination and apoptotic bodies are shown in frames. B. H3K9me2 in non-apoptotic and apoptotic MEFs. In white frames, see 3D-visualization of H3K9ac and H3K9me2 in intact cells. In red frame, panel B, see H3K9me2 in apoptotic bodies. Apoptosis was induced by 10 mM etoposide. Scale bars in microns are shown for individual panels.
S. Legartová et al. / Biochimie 95 (2013) 167e179
175
in apoptotic bodies (Fig. 7E, frames). For explanation, the pronounced variation in fluorescence recovery (shown as mean substantial standard errors; S.E.) for HP1a and HP1b is consistent with the variation between two possibilities: that a high level of H3K9me3 is present in apoptotic bodies or that H3K9me3 is located at the periphery of apoptotic bodies (Fig. 7E). These results imply that HP1 sub-types and H3K9 methylation are functional in apoptotic cells, which appears to be important for compactness of apoptotic bodies. Variability in fluorescence recovery was also observed for GFPtagged BMI1 protein, a component of the Polycomb Group Complex 1 (PRC1) that binds H3K27me3. In this case, H3K27me3 was either absent or positioned at the periphery of apoptotic bodies (Fig. 7G, frames) and GFP-BMI1 protein did not recover significantly after photobleaching in apoptotic cells (Fig. 9Bb). 3.4. Analysis of histone signature in apoptotic cells by mass spectrometry The effect of etoposide on the acetylation level of histone H4 in HL-60, U-937, and MEF cells was examined (Fig. 10A). Histone extracts were separated by RP-HPLC (for representative chromatograms see Supplement S1A) and fractions corresponding to histone H4 were collected. Changes in acetylation state were examined within the N-terminal Glu-C fragment of H4 (1e53 AA) containing the most frequent forms of modification. The relative abundance of each histone H4 variant (non-, mono- and diacetylated) in etoposide-treated samples was compared with controls. The signal intensity of the di-acetylated fragment in the spectra of etoposide-treated HL-60 and MEFs cells significantly decreased together with an increase in the signal intensity of the non-acetylated fragment, especially in HL-60 cells (Fig. 10A). However, a statistically insignificant decrease in the di-acetylated fragment associated with slight increase in the non-acetylated fragment was observed in the spectrum of U-937 cells (Fig. 10A). In addition, a notable decrease in the relative abundance of the H4 mono-acetylated variant was observed in the spectrum of HL-60 cells (Fig. 10A, summary Fig. 10B). Using a similar approach changes in acetylation levels within the N-terminal Glu-C fragment of histones H2B and H2A (1e35 AA for H2B and 1e56 AA for H2A, respectively) after etoposide treatment were also determined. Notable changes in the abundance of acetylated forms, especially a decrease in several acetylation profiles, were observed for all cell lines examined (see summary in Fig. 10B). Etoposide treatment also influenced the modification state of histone H3, but the complexity of particular forms undergoing post-translational modifications did not allow reliable interpretation of the etoposide effect (not shown). 4. Discussion
Fig. 7. Patterns of H3K9me3 and H3K27me3 in non-apoptotic and apoptotic MEFs. A. H3K9me3 (red) in interphase nuclei (blue) of the entire population of MEFs. B. 3D-visualization of H3K9me3 in non-apoptotic cells. C. 3D visualization of H3K9me3 in pre-apoptotic cells, characterized by chromatin margination. D. H3K9me3 occupied periphery of pre-apoptotic interphase nuclei of MEFs. Quantification was performed by Leica LAS AF software, according to selected regions of interest (ROIs) (see enclosed graph; FI means fluorescence intensity). E. Level of H3K9me3 in apoptotic bodies (frames). F. H3K27me3 in non-apoptotic cells compared with panel G. showing the pattern of H3K27me3 in apoptotic bodies (frames). Apoptosis was induced by 10 mM etoposide. Scale bars in microns are shown for individual panels.
In this study, we have shown that the chromatin of apoptotic cells has a specific histone signature, which is likely important for the maintenance of the highly condensed structure of apoptotic bodies. We found that the pronounced histone hypoacetylation and the high level of H3K9 methylation in apoptotic bodies is likely essential for the dynamic exchange of HP1 protein in apoptotic chromatin (Figs. 2, 6, 7, and 9B). On the other hand diffusion properties were limited for JMJD2b and BMI1 proteins in apoptotic cells (Fig. 9Ba, Bb). The proposed unique function of histone acetylation, H3K9 methylation and HP1 in apoptotic cells is supported by our additional results. Although, we detected similar levels of H3K9me3 in non-apoptotic and apoptotic MEFs, increased level of H3K36me3 accompanied apoptotic events (Fig. 8A; ETOP 50 mM). This could be linked to the limited function of JMJD2b in apoptotic
176
S. Legartová et al. / Biochimie 95 (2013) 167e179
Fig. 8. Western blot analysis of selected protein levels in MEFs and morphology of apoptotic cells. A. Levels of BMI1, H3K9 acetylation (ac), H3K9me2, H3K9me3, H3K27me3, H3K36me3, HP1a, HP1b, JMJD2b proteins were analyzed by western blots in MEFs. Label “C” represents the western blot data in non-treated cells, by “NC” are labeled MEFs, exposed to 0.001% DMSO (solvent for etoposide), and by ETOP 10 mM and ETOP 50 mM are labeled etoposide treated MEFs. B. Analysis of apoptotic markers, mCherry-tagged A-type lamins, showed distinct morphology of non-apoptotic and apoptotic living cells. Lamin A/C apoptotic fragmentation in MEFs was analyzed by western blotting. C. Percentage of nonapoptotic and apoptotic cells, distinguished according to morphology of nuclear lamina visualized by mCherry tagged A-type lamins, in non-apoptotic (control) and apoptotic (etoposide-treated) cells.
bodies, in which the fluorescence recovery of JMJD2b was significantly reduced to w10% (Fig. 9Ba). As a better explanation, JMJD2b antagonizes H3K9me3 in centromeric heterochromatin and H3K36me3 in the entire genome of intact cells [18,19]. Thus, the abrogated function of JMJD2b in apoptotic cells should rather lead to increased or stable levels of the relevant histone marks and it was confirmed in our experimental model (Fig. 8A). Our data for epigenome specificity in apoptotic cells are in a good agreement with Hurd et al. [9], which showed increased H3T45 phosphorylation in apoptotic bodies. Moreover, Ajiro [5] pointed out H2B phosphorylation as a universal marker of apoptotic DNA fragmentation that, on the chromosomal level, appears as the disorganization of chromosomal territories detected as random chromosomal segments of varying sizes [20]. BoixChornet et al. [21] showed H4K5, H4K8, H4K12, and H4K16 deacetylation in apoptotic chromatin of HL-60 cells, but surprisingly H3K9 acetylation and H3K9me2 levels remained stable during apoptotic process. In any case, these authors as the first showed specific histone signature of apoptotic cells, characterized by hypoacetylated histones H4. Cheng et al. [22] additionally found
dense H3K27me1 signal in staurosporine-induced apoptotic cells. Here, using three experimental approaches (western blotting, immunofluorescence, and mass spectrometry) we observed a general decrease in the acetylation state of all histones, although certain cell-type specific differences were observed (Figs. 2A, B, and 10). The results of our experiments particularize the general view on histone signatures in apoptotic cells. We provide information on changes in the balance between non-acetylated, mono-acetylated, and di-acetylated histones that can be induced by pro-apoptotic stimuli (Fig. 10). Interestingly, mass spectrometry showed a significant decrease in the di-acetylated form of histone H4 and monoacetylated form of H2B in apoptotic cell populations. This was logically balanced by an increased level of non-acetylated histones (summary in Fig. 10B). We also confirmed a prior report on GFP-HP1a and GFP-HP1b exchange kinetics in apoptotic cells [10]. As published by Cheutin et al. [10], HP1 isoforms recover rapidly after photobleaching in apoptotic bodies. Fast recovery of HP1 after FRAP was found in euchromatin and heterochromatin, but in different extent [10,23].
S. Legartová et al. / Biochimie 95 (2013) 167e179
177
Fig. 9. Fluorescence recovery after photobleaching (FRAP) for GFP-JMJD2b, GFP-BMI1, GFP-HP1a, GFP-HP1b proteins. A. Morphology of non-apoptotic and apoptotic cells transfected by full length GFP-JMJD2b, GFP-BMI1, GFP-HP1a, and GFP-HP1b. B. FRAP analysis of JMJD2b (a), BMI1 (b), HP1a (c), and HP1b (d) proteins in non-apoptotic (control) MEFs; etoposide treated non-apoptotic cells (nuclei with no apoptotic morphology), and etoposide-induced apoptotic cells. GFP-HP1a and GFP-HP1b fluorescence recovery after photobleaching was performed for HP1 located away from heterochromatin foci and accumulated into nuclear foci. Average relative fluorescence is shown with standard errors (S.E.). The cell nuclei after photobleaching were monitored during 23 s.
178
S. Legartová et al. / Biochimie 95 (2013) 167e179
Fig. 10. MALDIeTOF MS analysis of histone N-termini after etoposide treatment. Histone extracts from HL-60, U-937, and MEF cells were separated by RP-HPLC, digested by Glu-C, and analyzed by MALDIeTOF MS. A. Differences in the relative abundance of H4 N-terminal variants (1e53 AA; non-, mono-, and di-acetylated forms) of etoposide-treated samples and controls (data are shown as means and 95% confidence limits; n ¼ 15). For representative MALDIeTOF MS spectra see Supplement S1B. B. Overview of the effect of etoposide on histones H2A, H2B, and H4 acetylation status in particular cell lines.
Fluorescence recovery of HP1 isoforms after photobleaching can be changed by the inhibition of transcriptional elongation through actinomycin D [10,13]. Interestingly, identical HP1 exchanged kinetics, as after actinomycin D treatment, was observed in heterochromatic apoptotic cells, characterized by w30e40% recovery of HP1 after photobleaching [10]; (Fig. 9Bc and Bd). This observation may be tightly linked to the transcriptional silencing that occurs after actinomycin D treatment and/or in apoptotic cells. Altered dynamics of HP1 isoforms, JMJD2b and BMI-1 proteins may be attributed to the changes in posttranslational histone modifications. Especially de-acetylation, that occur together with changes in chromatin compaction, accompanied by apoptotic DNA fragmentation. 5. Conclusion The findings of this study contribute to our knowledge of apoptosis-related epigenetics. In several cell types we showed the existence of global histone hypoacetylation, especially H3K9 hypoacetylation and a decrease in H4 di-acetylation or H2B monoacetylation. These events were accompanied by an increased level of H3K9 methylation (me2 or me3) in apoptotic cells. The H3K9 acetylation pattern was cell-type specific (Figs. 2A, B and 8A), similar to the change in acetylation state studied by mass spectrometry (Fig. 10B). Levels of other histone markers (H3K4me1/me2/me3; H3K9me1; H3K27me3) were identical in non-apoptotic and apoptotic cell populations. In addition to the results on nonacetylated, mono- and di-acetylated histones in apoptotic cells, another novel observation is that not all components of highly compacted apoptotic heterochromatin have identical kinetic properties. As described above, GFP-tagged HP1 protein recovered more rapidly in apoptotic cells, while fluorescence of GFP-JMJD2b has never recovered up to more then 10%. Similarly, fluorescence recovery of GFP-BMI1 protein in apoptotic cells was w20% of prebleaching intensity (Fig. 9A, B). It is in contrast to intact cells, characterized by the BMI1 fluorescence recovery to w60e70% of prebleaching intensity in heterochromatin and 80e90% recovery of BMI1 in chromatin away from PcG bodies [24]. These data unambiguously document specificity in protein diffusion in apoptotic chromatin, generally characterized by histone deacetylation.
Taken together, it seems evident that apoptotic bodies represent a functionally specialized form of heterochromatin, characterized by a unique histone signature and kinetics of chromatin-related proteins, including HP1 sub-types, BMI1, and JMJD2b. Here, we pointed out that HP1 protein, H3K9 methylation and histone deacetylation, unlike JMJD2b, BMI1, H3K27me3 and H3K4 methylation, seem to be the most important epigenetic factors, responsible for the maintenance of apoptotic bodies, which represent the most condensed form of chromatin. Conflict of interest We declare no conflict of interest. Acknowledgments Our work was supported by following agencies: Grant Agency of Czech Republic, grant Nos.: P302/10/1022; P302/12/G157 and by Ministry of Education Youth and Sports of the Czech Republic, grant No.: LD11020. EB is also a member of the EU COST action TD09/05 and principal investigator of EU Marie Curie project PIRSES-GA2010-269156-LCS and project of LS supported by ASCR, v.v.i., No: M200041271. Work of ZZ and GL was supported by the project “CEITEC - Central European Institute of Technology” (CZ.1.05/ 1.1.00/02.0068) from European Regional Development Fund. For re-cloning assistance we thank Dr. Oskar Laur, Emory University Scholl of Medicine, Atlanta, USA. Appendix A. Supplementary material Supplementary material associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biochi. 2012.09.027. References [1] A.H. Wyllie, J.F. Kerr, A.R. Curie, Cell death: the significance of apoptosis, Int. Rev. Cytol. 68 (1980) 251e306. [2] G.M. Cohen, X.M. Sun, H. Fearnhead, M. MacFarlane, D.G. Brown, R.T. Snowden, D. Dinsdale, Formation of large molecular weight fragments of
S. Legartová et al. / Biochimie 95 (2013) 167e179
[3] [4] [5] [6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
DNA is a key committed step of apoptosis in thymocytes, J. Immunol. 153 (1994) 507e516. S.J. Martin, D.R. Green, T.G. Cotter, Dicing with death: dissecting the components of the apoptosis machinery, Trends Biochem. Sci. 19 (1994) 26e30. M.O. Hengartner, Apoptosis. DNA destroyers, Nature 412 (2001) 27e29. K. Ajiro, Histone H2B phosphorylation in mammalian apoptotic cells. An association with DNA fragmentation, J. Biol. Chem. 275 (2000) 439e443. E.P. Rogakou, W. Nieves-Neira, C. Boon, Y. Pommier, W.M. Bonner, Initiation of DNA fragmentation during apoptosis induces phosphorylation of H2AX histone at serine 139, J. Biol. Chem. 275 (2000) 9390e9395. P. Waring, T. Khan, A. Sjaarda, Apoptosis induced by gliotoxin is preceded by phosphorylation of histone H3 and enhanced sensitivity of chromatin to nuclease digestion, J. Biol. Chem. 272 (1997) 17929e17936. M. Wu, L.G. Xu, X. Li, Z. Zhai, H.B. Shu, AMID, an apoptosis-inducing factorhomologous mitochondrion-associated protein, induces caspase-independent apoptosis, J. Biol. Chem. 277 (2002) 25617e25623. P.J. Hurd, A.J. Bannister, K. Halls, M.A. Dawson, M. Vermeulen, J.V. Olsen, H. Ismail, J. Somers, M. Mann, T. Owen-Hughes, I. Gout, T. Kouzarides, Phosphorylation of histone H3 Thr-45 is linked to apoptosis, J. Biol. Chem. 284 (2009) 16575e16583. T. Cheutin, A.J. McNairn, T. Jenuwein, D.M. Gilbert, P.B. Singh, T. Misteli, Maintenance of stable heterochromatin domains by dynamic HP1 binding, Science 299 (2003) 721e725. A. Jugová, G. Sustá cková, S. Legartová, L. Stixová, S. Kozubek, E. Bártová, Effects of epigenetic-based anti-cancer drugs in leukaemia and multiple myeloma cells, Cell Biol. Int. 35 (2011) 1195e1203. E. Bártová, J. Pacherník, A. Harni carová, A. Kovarík, M. Kovaríková, J. Hofmanová, M. Skalníková, M. Kozubek, S. Kozubek, Nuclear levels and patterns of histone H3 modification and HP1 proteins after inhibition of histone deacetylases, J. Cell Sci. 118 (2005) 5035e5046. k, S. Legartová, S. Kozubek, HeteroL. Stixová, E. Bártová, P. Matula, O. Dane geneity in the kinetics of nuclear proteins and trajectories of substructures associated with heterochromatin, Epigenet. Chromatin 4 (2011) 5. E. Bártová, L. Stixová, G. Galiová, A. Harni carová Horáková, S. Legartová, S. Kozubek, Mutant genetic background affects the functional rearrangement
[15]
[16]
[17] [18]
[19]
[20] [21]
[22]
[23]
[24]
179
and kinetic properties of JMJD2b histone demethylase, J. Mol. Biol. 405 (2011) 679e695. D. Bonenfant, H. Towbin, M. Coulot, P. Schindler, D.R. Mueller, J. van Oostrum, Analysis of dynamic changes in post-translational modifications of human histones during cell cycle by mass spectrometry, Mol. Cell. Proteomics 6 (2007) 1917e1932. L. Cin cárová, G. Lochmanová, K. Nováková, P. Sultesová, H. Kone cná, L. Fajkusová, J. Fajkus, Z. Zdráhal, A combined approach for the study of histone deacetylase inhibitors, Mol Biosyst. 8 (2012) 2937e2945. R. Paro, H. Strutt, G. Cavalli, Heritable chromatin states induced by the Polycomb and trithorax group genes, Novartis Found. Symp. 214 (1998) 51e61. B.D. Fodor, S. Kubicek, M. Yonezawa, R.J. O’Sullivan, R. Sengupta, L. PerezBurgos, S. Opravil, K. Mechtler, G. Schotta, T. Jenuwein, Jmjd2b antagonizes H3K9 trimethylation at pericentric heterochromatin in mammalian cells, Genes Dev. 20 (2006) 1557e1562. A. Harni carová Horáková, G. Galiová, S. Legartová, S. Kozubek, P. Matula, E. Bártová, Chromocentre integrity and epigenetic marks, J. Struct. Biol. 169 (2010) 124e133. E. Bártová, P. Jirsová, M. Fojtová, K. Sou cek, S. Kozubek, Chromosomal territory segmentation in apoptotic cells, Cell. Mol. Life Sci. 60 (2003) 979e990. M. Boix-Chornet, M.F. Fraga, A. Villar-Garea, R. Caballero, J. Espada, A. Nunez, J. Casado, C. Largo, J.I. Casal, J.C. Cigudosa, L. Franco, M. Esteller, E. Ballestar, Release of hypoacetylated and trimethylated histone H4 is an epigenetic marker of early apoptosis, J. Biol. Chem. 281 (2006) 13540e13547. M.F. Cheng, C.H. Lee, K.T. Hsia, G.S. Huang, H.S. Lee, Methylation of histone H3 lysine 27 associated with apoptosis in osteosarcoma cells induced by staurosporine, Histol. Histopathol. 24 (2009) 1105e1111. R. Festenstein, S.N. Pagakis, K. Hiragami, D. Lyon, A. Verreault, B. Sekkali, D. Kioussis, Modulation of heterochromatin protein 1 dynamics in primary mammalian cells, Science 299 (2003) 719e721. G. Sustá cková, S. Kozubek, L. Stixová, S. Legartová, P. Matula, D. Orlova, E. Bártová, Acetylation-dependent nuclear arrangement and recruitment of BMI1 protein to UV-damaged chromatin, J. Cell. Physiol. 227 (2012) 1838e1850.
Contents lists available at SciVerse ScienceDirect
Biochimie journal homepage: www.elsevier.com/locate/biochi
Research paper
Epigenetic aspects of HP1 exchange kinetics in apoptotic chromatin a Legartová a, Al ta Jugová a, Lenka Stixová a, Stanislav Kozubek a, Miloslava Fojtová a, Son zbe b k Zdráhal , Gabriela Lochmanová b, Eva Bártová a, * Zbyne a b
Institute of Biophysics, Academy of Sciences of the Czech Republic, v.v.i., Královopolská 135, 612 65 Brno, Czech Republic Core FacilityeProteomics, Central European Institute of Technology, Masaryk University, Kamenice 753/5, Brno, Czech Republic
a r t i c l e i n f o
a b s t r a c t
Article history: Received 14 June 2012 Accepted 10 September 2012 Available online 27 September 2012
Apoptotic bodies are the most condensed form of chromatin. In general, chromatin structure and function are mostly dictated by histone post-translational modifications. Thus, we have analyzed the histone signature in apoptotic cells, characterized by pronounced chromatin condensation. Here, H2B mono-acetylation, and H3K9 and H4 acetylation was significantly decreased in apoptotic cells, which maintained a high level of H3K9 methylation. This phenotype was independent of p53 function and distinct levels of anti-apoptotic Bcl2 protein. Interestingly, after etoposide treatment of leukemia and multiple myeloma cells, H3K9 and H4 hypoacetylation was accompanied by increased H3K9me2, but not H3K9me1 or H3K9me3. In adherent mouse fibroblasts, a high level of H3K9me3 and histone deacetylation in apoptotic bodies was likely responsible for the pronounced (w40%) recovery of GFP-HP1a and GFP-HP1b after photobleaching. HP1 mobility in apoptotic cells appeared to be unique because limited exchange after photobleaching was observed for other epigenetically important proteins, including GFPJMJD2b histone demethylase (w10% fluorescence recovery) or Polycomb group-related GFP-BMI1 protein (w20% fluorescence recovery). These findings imply a novel fact that only certain subset of proteins in apoptotic bodies is dynamic. Ó 2012 Elsevier Masson SAS. All rights reserved.
Keywords: Apoptosis Histones HP1 Epigenetics Chromatin
1. Introduction Despite the fact that apoptotic chromatin is fragmented into large or oligonucleosomal fragments, apoptotic bodies represent the most condensed form of chromatin [1e3]. Typical apoptotic DNA fragmentation is mediated by caspase-activated DNase (CAD) and possibly other endonucleases [4]. It is well known that chromatin condensation occurs during cell differentiation, mitosis, and apoptosis. However, one question that remains is whether identical histone post-translational modifications occur on chromatin during all cellular processes that require chromatin compaction. In lymphoma cells, it has been observed that apoptosis is accompanied by the deubiquitination of histone H2A. Phosphorylation of histones H1, H2A, and H2B also occurs in apoptotic cells [5e7]. Interestingly, Wu et al. [8] found that histones are released from apoptotic nucleosomes. After specific pro-apoptotic stimuli, histones H2A, H2B, H3, and H4 are released from apoptotic DNA in Jurkat and HeLa cells, but this event was not observed in apoptotic NIH3T3 cells. Wu et al. [8] explained that CAD function is
* Corresponding author. Tel.: þ420 5 41517141; fax: þ420 5 41240498. E-mail address: [email protected] (E. Bártová). 0300-9084/$ e see front matter Ó 2012 Elsevier Masson SAS. All rights reserved. http://dx.doi.org/10.1016/j.biochi.2012.09.027
responsible for histones being released from the DNA of apoptotic cells and that the separation of histones from DNA is not simply a result of the DNA fragmentation process. The Wu et al. [8] study raised the question of whether apoptotic bodies are formed only from DNA or also contain some residual histones that maintain their post-translational modifications. From this view, it was observed that apoptotic nucleosomal fragmentation appears in parallel with phosphorylation of H2B. In addition, apoptotic chromatin condensation was accompanied by decreased H3 phosphorylation and the sites where phosphorylation occurs do not match the phosphorylation pattern of histones that accompanies mitotic chromosome condensation. Hurd et al. [9] have reported increased phosphorylation of threonine 45 of histone H3 (H3T45) in apoptotic cells. These authors suggest that the nucleosomal position of H3T45 is responsible for the structural changes within the nucleosomes that lead to the formation of DNA nicks and/or DNA fragmentation. This event is mediated by protein kinase C-delta that governs H3T45ph, in vitro and in vivo [9]. Despite this fact, it is not completely clear whether specific post-translational modifications of histones regulate apoptotic events, such as chromatin margination, or if complex histone signature is responsible for maintenance of apoptotic bodies as shown by Hurd et al. [9] for H3T45 phosphorylation. In the present study, we endeavored to characterize the
168
S. Legartová et al. / Biochimie 95 (2013) 167e179
histone signature of apoptotic bodies. Because heterochromatin protein 1 a (HP1a) partially maintains its diffusion ability in apoptotic cells [10], we also addressed HP1 kinetics and the HP1related histone signature in apoptotic cells. We also tested the levels of histone acetylation, H3K4 methylation, H3K9 methylation and H3K27 trimethylation in apoptotic cells. This was done in parallel with fluorescence recovery after photobleaching (FRAP) analysis of HP1a, HP1b proteins, H3K9-demethylase JMJD2b, and Polycomb group-related protein BMI1. Our results reveal the apoptosis-specific epigenetic features that are likely responsible for the maintenance of highly condensed chromatin of apoptotic bodies. 2. Materials and methods 2.1. Cell cultivation Human leukemia cells K-562, HL-60, and U-937 (obtained from European Collection of Cell Cultures, UK) and plasmocytoma ARH-77 cells were cultivated in RPMI medium supplemented with 10% fetal bovine serum (FBS); (PAN, Germany). Human multiple myeloma U266 cells (#ACC 9) (purchased from DSMZ, Braunschweig, Germany) were seeded at a density of 2 105 cells/ml and cultured in RPMI1640 medium supplemented with 10% FBS (PAN, Germany), 100 IU/ml penicillin, and 0.1 mg/ml streptomycin at 37 C in a humidified atmosphere of 5% CO2/95% air. Multiple myeloma MOLP-8 cells (DSMZ, Braunschweig, Germany) were grown in RPMI1640 medium supplemented with 20% FBS, 100 U/ml penicillin, and 0.1 mg/ml streptomycin at 37 C in a humidified atmosphere containing 5% CO2/95% air. Cells were treated for 24 h with an inhibitor of histone deacetylases, Trichostatin A (100 nM TSA, dissolved in ethanol), and with pro-apoptotic etoposide (4 h of 15 mM for MOLP8; 4 h of 75 mM for U-937; 8 h of 100 mM for ARH-77; 24 h of 200 mM for K-562, 4 h of 5 mM for HL-60; 4 h of 100 mM for U-266; and 24 h of 10 mM or 50 mM for MEFs). Concentrations of the pro-apoptotic agent etoposide were chosen according to 50% cell viability and induction of apoptosis, verified by lamin B fragmentation or according to morphology A-type lamins. Protein levels were analyzed by western blots and normalized to the total protein levels. 2.2. Western blotting After washing in PBS, the cells were lysed in sodium dodecyl sulfate (SDS) lysis buffer (50 mM TriseHCl, pH 7.5; 1% SDS; 10% glycerol) and further processed as in Jugová et al. [11]. Protein concentration was measured with a mQuant spectrophotometer. The following antibodies were used for western blots: antiH3K4me1 (#ab8895, Abcam); anti-H3K4me2 (#07-030, Upstate); anti-H3K4me3 (#ab8580, Abcam); anti-H3K9me1 (#ab9045, Abcam); anti-H3K9me2 (#07-212, Upstate); anti-H3K9me3 (#05-1250, Upstate); anti-H3K9 acetylation (#06-942, Upstate); anti-H4 acetylation (#382160, Calbiochem, Merck-Millipore); antiHP1a (#05-689, Upstate); anti-HP1b (MAB3448, clone 1MOD-1A9, Merck-Millipore); anti-H3K27me3 (#07-449, Upstate); antiSuv39h (#ab38637, Abcam); anti-p53 (#NCL-p53 CM1, Novacastra); anti-pRb1 (#sc-50, Santa Cruz Biotechnology); anti-lamin B (#sc-6217, Santa Cruz Biotechnology); anti-lamin A/C (#sc-7293, Santa Cruz Biotechnology); anti-BMI1 (#05-637, Upstate, USA); and antibody against JMJD2b (#ab-91549, Abcam); antibody against caspase 3 (#sc-7148, Santa Cruz Biotechnology); anti-caspase 8 (#9746, Cell Signaling Technology, Inc.); anti-caspase 9 (#9508, Cell Signaling Technology, Inc.); anti-Bcl2 (#sc-509, Santa Cruz Biotechnology). Dilutions were 1:1000e1:2500. The secondary antibodies were anti-rabbit IgG (A-4914, SigmaeAldrich, Czech Republic; dilution 1:2000), anti-mouse IgG (A-9044, Sigmae Aldrich, Czech Republic; dilution 1:2000), anti-goat IgG (A-4174,
Sigma-Aldrich, Czech Republic, dilution 1:2000), and anti-mouse IgG1 (sc-2060, Santa Cruz Biotechnology; dilution 1:500). 2.3. Immunofluorescence After fixation in 4% formaldehyde, for 10 min at room temperature (RT), cells were permeabilized with 0.2% Triton X-100 for 10 min, followed by 0.1% saponin (SigmaeAldrich, Czech Republic) for 12 min, and washed twice in PBS for 15 min. BSA (1%) dissolved in PBS was used as a blocking solution. Slides with cells were incubated in BSA for 1 h, washed for 15 min in PBS, and then incubated with primary antibodies. Appropriate secondary antibodies were used [12]. Image acquisition was performed by TSC SP5-X confocal microscopy (Leica Microsystems, Mannheim, Germany). For fluorochrome excitation, we used the white light laser (WLL, 470e670 nm in 1 nm increments) connected to the Leica confocal system. 2.4. Cell transfection Mouse embryonic fibroblasts (MEFs) were transfected by plasmids encoding GFP-HP1a, GFP-HP1b, GFP-BMI1, and GFP-JMJD2b (for detailed information, see Refs. [13,14]). A-type lamins were visualized by cell transfection by plasmid DNA encoding mCherrylamin A (pBABE-puro-GFP-wt-lamin A was purchased from Addgene [#17662] and it was re-cloned by dr. Oskar Laur to pBABE-puromCherry vector). Plasmid DNA was transformed into competent Escherichia coli DH5a and isolated following Stixová et al. [13]. Plasmid DNA (1e5 mg) was then used for transfection using the METAFECTENEÔPRO system (Biontex Laboratories GmbH, Germany). 2.5. Fluorescence recovery after photobleaching (FRAP) The FRAP mode of the Leica TSC SP5-X (Leica Microsystems, Mannheim, Germany) was used for these experiments. We used the argon laser (488 nm) for photobleaching and subsequent visualization of the cell nuclei. For analysis, we used immersion objective, 63, numeric aperture (N.A.) ¼ 1.4. Cell cultures were cultivated in microscopic Petri dishes (MatTek Corporation, P50G-0-30-F). For FRAP, the dishes with the cells were placed in a cultivation hood (EMBL Heidelberg, Germany) to guarantee optimal cultivation conditions (i.e., 37 C, 5% CO2 and optimal humidity). Pre-bleaching was performed for 5 intervals, bleaching 2 intervals, and postbleaching fluorescence intensity was measured for 30 intervals. The interval value was 0.67 s (determined by software, according to selected resolution 512 512 pixels). GFP fluorescence was bleached to the background level (the background value was subtracted from experimental data) and recovery of fluorescence was monitored using the LEICA LAS AF software tool for FRAP. 2.6. Histone extraction Histones were isolated as described earlier by Bonenfant et al. [15] with minor modifications. Briefly, the cells were washed twice with ice-cold PBS, resuspended in extraction buffer (80 mM NaCl, 20 mM EDTA, 1% Triton X-100, 45 mM sodium butyrate, and 100 mM PMSF), incubated for 20 min on ice, and centrifuged at 2000 g for 8 min. Pellets were resuspended in 900 ml ice-cold 0.2 M H2SO4 and incubated for 2 h with shaking at 4 C. After centrifugation at 16,000 g, supernatants were precipitated in trichloroacetic acid at a final concentration of 25% and incubated for 30 min on ice. After centrifugation at 5000 g for 30 min at 4 C, the pellet was washed with 50 mM HCl in acetone and then with acetone, and subsequently dissolved in water. The protein concentration was determined by the Bradford protein assay (Bio-Rad).
S. Legartová et al. / Biochimie 95 (2013) 167e179
169
Fig. 1. Morphology of non-apoptotic and apoptotic cells. Apoptosis was induced by etoposide and TSA in A. multiple myeloma MOLP-8 cells and B. leukemia U-937 cells. Transmission light mode from confocal microscope Leica TCS SP5-X was used for image acquisition. Selected cells are magnified in individual figures and scale bars are shown for each panel. C. The number of apoptotic cells was counted on cell morphology in non-treated and etoposide- or TSA-treated MOLP-8 and U-937 cells.
2.7. Analysis of histone acetylation status by MALDI-TOF MS Separation of histone extracts and matrix-assisted laser desorption/ionisationetime of flight mass spectrometry (MALDIe TOF MS) analysis of histone fractions were performed as described previously [16] with minor modifications. Briefly, separation of
histones was performed on a Dionex Ultimate 3000 highperformance liquid chromatography (HPLC) system using an Agilent mRP-C18 High-Recovery Protein Column (0.5 100 mm, macroporous C18-bonded, 5-mm silica particles). Histone extracts (3 mg) were injected onto the column in solvent A and separated using a multi-step acetonitrile gradient at 70 C and a constant flow
170
S. Legartová et al. / Biochimie 95 (2013) 167e179
Fig. 2. Western blot data for epigenetic patterns in non-apoptotic and apoptotic cells. A. Lamin B, H3K9ac, H3K9me1/me2/me3, H3K4me1/me2/me3, H3K27me3, Suv39h, p53, and pRb1 levels were studied in plasma cell leukemia line ARH-77, leukemia HL-60, K-562 and U-937 cell lines, multiple myeloma U-266 and MOLP-8 cells. B. H4 acetylation was studied in selected cell lines (MOLP-8, ARH-77, K-562, U-937 and MEFs). Leukemia and multiple myeloma cell lines were treated by etoposide as written in Materials and methods and MEFs were treated by 50 mM etoposide. Protein levels were normalized to the total protein levels. Label “C” shows non-treated control cells; “ETOP” is abbreviation for etoposide treatment and “TSA” is abbreviation for Trichostatin A treatment.
rate of 9 ml/min. The concentration of solvent B was increased in the following manner: 0% over 8 min, from 36 to 60% over 30 min, 80% over 9 min [Solvent A: 2% acetonitrile (v/v), 0.1% trifluoroacetic acid (v/v) in water; solvent B: 80% acetonitrile (v/v), 0.1% trifluoroacetic acid (v/v)]. Particular histone fractions were collected by a Probot LC Packings Micro Fraction Collector, dried in a SpeedVac SPD 111V (Savant), dissolved in 15 ml 100 mM ammonium acetate at pH 4, and digested by endoproteinase Glu-C (Sequencing grade, Roche Diagnostics) at an enzyme ratio of 1:25 at 25 C for 90 min. Glu-C digests were analyzed with MALDIeTOF MS on an Ultraflex III mass spectrometer (Bruker Daltonik, Bremen, Germany). Mass spectra were acquired in linear positive mode (25 kV acceleration voltage) with
500 laser shots. External calibration of the mass spectra was performed using standard peaks of Escherichia coli DH5a. A sample in a volume of 0.6 ml was mixed with 2.4 ml matrix solution [20 mg/ml DHB in an acetonitrile:water:TFA (80:19:1, v/v/v) mixture]. Flex Analysis 3.0 software (Bruker Daltonik) was used for data processing. The relative abundance of the individual modified form was determined as the ratio of its single-ion signal intensity to the sum of the intensities of all relevant variants (0Ac, 1Ac, and 2Ac; see Supplement S1; Ac indicates histone acetylation) within the spectrum of the N-terminal fragment. A total of 15 spectra accumulations per sample were acquired; the Grubbs test for outliers was performed and the mean of relative abundance and 95 confidence
S. Legartová et al. / Biochimie 95 (2013) 167e179
171
Fig. 3. Immunofluorescence detection of H3K9 acetylation (ac) and H3K9 dimethylation (me2) in non-apoptotic and apoptotic U-937 cells. A. H3K9ac in control (non-treated), etoposide, and TSA treated U-937 cells. B. H3K9me2 in control (non-treated), etoposide, and TSA treated U-937 cells. Intact control cell and apoptotic cells, characterized by apoptotic bodies, are indicated by frames. Scale bars in microns are shown for each panel.
172
S. Legartová et al. / Biochimie 95 (2013) 167e179
intervals of the ratios were calculated. The significance of differences in acetylation status between controls and etoposide-treated samples was statistically evaluated by paired t-tests (P < 0.05). 3. Results 3.1. Analysis of the induction of apoptosis and the specific histone signature in apoptotic cells We induced apoptosis with etoposide and an inhibitor of histone deacetylases (HDACs), TSA. Etoposide and TSA treatment led to distinct morphologies in the entire cell population of selected cell lines, MOLP-8 (Fig. 1A) and U-937 (Fig. 1B), but both agents function as pro-apoptotic inducers (see Figs. 1C and 2A, lamin B fragmentation). In addition, we found that etoposide induced apoptosis in approximately 50% of the cell population, whereas TSA was substantially less efficient in the induction of morphologically
recognized apoptosis (Fig. 1C). Thus, for subsequent studies we used etoposide as a powerful inducer of programmed cell death. We first characterized apoptosis on the population level using western blots to detect changes in the histone signature induced by the pro-apoptotic agents. We studied H3K9 and H4 acetylation, H3K9me1/me2/me3, H3K4me1/me2/me3, K3K27me3, and the level of SUV39h histone methyltransferases (Fig. 2A). We also analyzed the levels of tumor suppressors p53 and pRb1. The effects of etoposide and TSA were studied in several leukemia cell lines (U-937, HL-60, K-562) and multiple myeloma cells (U-266, MOLP-8 and ARH-77). In all cell types tested, we observed that etoposide treatment reduced H3K9 acetylation, which in U-937, U-266, and MOLP-8 cells was accompanied by increased H3K9me2 levels (Fig. 2A). However, H3K9me1 and H3K9me3 levels were not changed in comparison with relevant non-treated controls following either treatment (Fig. 2A). Conversely, TSA, which also induced apoptosis as characterized by lamin B fragmentation,
Fig. 4. Immunofluorescence detection of H3K9 acetylation (ac) and H3K9 dimethylation (me2) in non-apoptotic and apoptotic MOLP-8 cells. A. H3K9ac in control, etoposide, and TSA treated MOLP-8 cells. B. H3K9me2 in control, etoposide, and TSA treated MOLP-8 cells. Apoptotic cells, characterized by apoptotic bodies, are indicated by frames. C. Magnified images from panels A and B, showing H3K9ac and H3K9me2 in apoptotic bodies. Possible release of histones (greeneyellow signals) is shown as (1) and residual H3K9ac histones in one apoptotic body are shown as (2). High levels of H3K9me2 (3) are shown in several apoptotic bodies of MOLP-8 cells. Scale bars in microns are shown for individual panels.
S. Legartová et al. / Biochimie 95 (2013) 167e179
173
caused H3K9 hyperacetylation (Fig. 2A). However, the hyperacetylation was not a characteristic of apoptotic cells, which lacked H3K9 acetylation even during TSA-induced cell death (see morphology of apoptotic cells in Fig. 3A). As shown by immunofluorescence in U-937 and MOLP-8 cells treated by TSA, apoptotic bodies were characterized by a low level of H3K9 acetylation and relatively high levels of H3K9me2 (Fig. 3A and B or 4AeC). Therefore, a high level of H3K9 acetylation, determined by western blots in entire cell population, is preferentially induced in non-apoptotic cells after TSA treatment (see Figs. 3A and 4A, intact cells with strong green fluorescence signal after TSA treatment). Moreover, our analysis shows that while some histones are likely released during the apoptotic process (Fig. 4C1, green spots in panel 1), some remain components of apoptotic bodies (Fig. 4C, panel 2 and panel 3; e.g. greeneyellow signals inside the apoptotic bodies). Additional western blots detected identical levels of H3K4me1/ me2/me3, SUV39h, and H3K27me3 in control, and etoposide- and TSA-treated cells (Fig. 2A). None of these events were influenced by p53 status, because similar trends in histone signature were observed in p53 mutant leukemia (HL-60, U-937 cells) and p53 positive myeloma cells (Fig. 2A). The level of H4 acetylation was also studied by western blots and we observed decreased levels of this histone marker after etoposide treatment in MOLP-8, K-562, U-937 cells and mouse embryonic fibroblasts (MEFs). An exception was ARH-77 multiple myeloma cells, which were characterized by an identical level of H4 acetylation in non-apoptotic and apoptotic etoposide-treated cells (Fig. 2B). 3.2. Apoptotic pathways studied in etoposide and TSA treated cells We next looked for distinctions in the apoptotic pathways induced by etoposide and TSA in p53 negative (U-937) and p53 positive (MOLP-8) cells. Western blots showed that both cell lines had increased levels of the 17 kDa fragment of caspase 3 after etoposide or TSA treatment. In comparison with control, the 43 kDa and 18 kDa fragments of caspase 8 were more robust after TSA or etoposide treatment in U-937 cells. Similarly, high level of 18 kDa fragment was found in MOLP-8 cells after TSA treatment (Fig. 5). Only subtle distinctions were found in the level of caspase 9 between etoposide and TSA treated cells. However, p53 mutant U-937 cells were characterized by a high level of Bcl2 protein while p53 positive MOLP-8 cells had subtle Bcl2 fragments (see Fig. 5). Both etoposide and TSA induced strong gH2AX positivity (Fig. 5), which is a characteristic marker of DNA damage that can be induced by pro-apoptotic agents. The appearance of gH2AX was likely associated with apoptotic DNA fragmentation. 3.3. Epigenetic features and fluorescence recovery after photobleaching of HP1a, HP1b, JMJD2b, and BMI1 proteins in apoptotic MEFs We studied the kinetics of HP1a, HP1b, JMJD2b, and BMI1 proteins in non-apoptotic mouse embryonic fibroblasts (MEFs) (as a control) and during apoptosis (induced by etoposide) using the FRAP technique. Instead of cell growing in suspension, we chose MEFs for these experiments because adherent cells are more suitable for FRAP analysis. We asked whether epigenetic events, including H3K9 deacetylation, H3K9 methylation and additionally H3K27me3 in apoptotic bodies, are accompanied by specific behaviors of proteins that recognize these histone modifications. We first confirmed the hypoacetylation and H3K9 hypermethylation in apoptotic MEFs (Figs. 6A and 7AeE). We also determined the patterns of H3K9ac, H3K9me2, H3K9me3, and H3K27me3 in pre-apoptotic and apoptotic MEFs (Figs. 6 and 7). The
Fig. 5. Proteins of apoptotic pathways analyzed in U-937 and MOLP-8 cells. Western blots of caspase 3, caspase 8, caspase 9, Bcl2 protein, and phosphorylated histone H2AX (gH2AX) in control, etoposide- and TSA-treated U-937 and MOLP-8 cells.
apoptotic bodies or their peripheries were completely stained by antibody against H3K9me3; however, H3K9me2 was absent from apoptotic bodies (compare Figs. 6B and 7AeE). H3K27me3, a BMI1 binding partner [17], was also positioned at the periphery of apoptotic bodies (compare Fig. 7F control intact cell nuclei with panel 7G showing H3K27me3 on the periphery of apoptotic bodies) or removed from apoptotic chromatin (Fig. 7G, bottom panels). The epigenetic events observed on the single cell level were accompanied by decreased levels of BMI1 protein in the apoptotic cell population, as revealed by western blots (Fig. 8A). This was again accompanied by reduced levels of H3K9 acetylation in apoptotic cells (Fig. 6A or 8A, second line in the western blot panel). Additionally, H4 acetylation was reduced in MEFs after etoposide treatment (Fig. 2B) that caused A-type lamin degradation, which is specific for apoptotic events (Fig. 8B,C). However, western blots showed that the levels of other epigenetic markers (H3K9me2, H3K9me3, H3K27me3, HP1a, and HP1b) were not changed by etoposide treatment. One exception was the increased levels of H3K36me3 in MEFs exposed to a high concentration of etoposide (Fig. 8A). We next performed FRAP analysis for JMJD2b, BMI1, HP1a, and HP1b, proteins in non-apoptotic and apoptotic cells (Fig. 9A; see also pictorial illustration of the FRAP procedure in Supplement S2 and S3). We observed significantly reduced fluorescence recovery for JMJD2b (Fig. 9Ba; w10% recovery) and BMI1 proteins (Fig. 9Bb; w20% recovery) in etoposide-induced apoptotic cells (Fig. 9B; red diamonds). Conversely, HP1a and HP1b recovered more rapidly, to 40% of pre-bleaching levels, in apoptotic cells (Fig. 9Bc and Bd; red squares). However, the HP1 recovery time after photobleaching was more variable in comparison with JMJD2b (compare standard errors in panel Fig. 9Ba with data in panels Fig. 9Bc and Bd). This correlates well with the pattern of H3K9me3
174
S. Legartová et al. / Biochimie 95 (2013) 167e179
Fig. 6. Immunofluorescence detection of H3K9 acetylation (ac) and H3K9 dimethylation (me2) in non-apoptotic and apoptotic MEFs. A. H3K9ac in non-apoptotic and apoptotic MEFs. Cell nuclei, characterized by apoptotic chromatin margination and apoptotic bodies are shown in frames. B. H3K9me2 in non-apoptotic and apoptotic MEFs. In white frames, see 3D-visualization of H3K9ac and H3K9me2 in intact cells. In red frame, panel B, see H3K9me2 in apoptotic bodies. Apoptosis was induced by 10 mM etoposide. Scale bars in microns are shown for individual panels.
S. Legartová et al. / Biochimie 95 (2013) 167e179
175
in apoptotic bodies (Fig. 7E, frames). For explanation, the pronounced variation in fluorescence recovery (shown as mean substantial standard errors; S.E.) for HP1a and HP1b is consistent with the variation between two possibilities: that a high level of H3K9me3 is present in apoptotic bodies or that H3K9me3 is located at the periphery of apoptotic bodies (Fig. 7E). These results imply that HP1 sub-types and H3K9 methylation are functional in apoptotic cells, which appears to be important for compactness of apoptotic bodies. Variability in fluorescence recovery was also observed for GFPtagged BMI1 protein, a component of the Polycomb Group Complex 1 (PRC1) that binds H3K27me3. In this case, H3K27me3 was either absent or positioned at the periphery of apoptotic bodies (Fig. 7G, frames) and GFP-BMI1 protein did not recover significantly after photobleaching in apoptotic cells (Fig. 9Bb). 3.4. Analysis of histone signature in apoptotic cells by mass spectrometry The effect of etoposide on the acetylation level of histone H4 in HL-60, U-937, and MEF cells was examined (Fig. 10A). Histone extracts were separated by RP-HPLC (for representative chromatograms see Supplement S1A) and fractions corresponding to histone H4 were collected. Changes in acetylation state were examined within the N-terminal Glu-C fragment of H4 (1e53 AA) containing the most frequent forms of modification. The relative abundance of each histone H4 variant (non-, mono- and diacetylated) in etoposide-treated samples was compared with controls. The signal intensity of the di-acetylated fragment in the spectra of etoposide-treated HL-60 and MEFs cells significantly decreased together with an increase in the signal intensity of the non-acetylated fragment, especially in HL-60 cells (Fig. 10A). However, a statistically insignificant decrease in the di-acetylated fragment associated with slight increase in the non-acetylated fragment was observed in the spectrum of U-937 cells (Fig. 10A). In addition, a notable decrease in the relative abundance of the H4 mono-acetylated variant was observed in the spectrum of HL-60 cells (Fig. 10A, summary Fig. 10B). Using a similar approach changes in acetylation levels within the N-terminal Glu-C fragment of histones H2B and H2A (1e35 AA for H2B and 1e56 AA for H2A, respectively) after etoposide treatment were also determined. Notable changes in the abundance of acetylated forms, especially a decrease in several acetylation profiles, were observed for all cell lines examined (see summary in Fig. 10B). Etoposide treatment also influenced the modification state of histone H3, but the complexity of particular forms undergoing post-translational modifications did not allow reliable interpretation of the etoposide effect (not shown). 4. Discussion
Fig. 7. Patterns of H3K9me3 and H3K27me3 in non-apoptotic and apoptotic MEFs. A. H3K9me3 (red) in interphase nuclei (blue) of the entire population of MEFs. B. 3D-visualization of H3K9me3 in non-apoptotic cells. C. 3D visualization of H3K9me3 in pre-apoptotic cells, characterized by chromatin margination. D. H3K9me3 occupied periphery of pre-apoptotic interphase nuclei of MEFs. Quantification was performed by Leica LAS AF software, according to selected regions of interest (ROIs) (see enclosed graph; FI means fluorescence intensity). E. Level of H3K9me3 in apoptotic bodies (frames). F. H3K27me3 in non-apoptotic cells compared with panel G. showing the pattern of H3K27me3 in apoptotic bodies (frames). Apoptosis was induced by 10 mM etoposide. Scale bars in microns are shown for individual panels.
In this study, we have shown that the chromatin of apoptotic cells has a specific histone signature, which is likely important for the maintenance of the highly condensed structure of apoptotic bodies. We found that the pronounced histone hypoacetylation and the high level of H3K9 methylation in apoptotic bodies is likely essential for the dynamic exchange of HP1 protein in apoptotic chromatin (Figs. 2, 6, 7, and 9B). On the other hand diffusion properties were limited for JMJD2b and BMI1 proteins in apoptotic cells (Fig. 9Ba, Bb). The proposed unique function of histone acetylation, H3K9 methylation and HP1 in apoptotic cells is supported by our additional results. Although, we detected similar levels of H3K9me3 in non-apoptotic and apoptotic MEFs, increased level of H3K36me3 accompanied apoptotic events (Fig. 8A; ETOP 50 mM). This could be linked to the limited function of JMJD2b in apoptotic
176
S. Legartová et al. / Biochimie 95 (2013) 167e179
Fig. 8. Western blot analysis of selected protein levels in MEFs and morphology of apoptotic cells. A. Levels of BMI1, H3K9 acetylation (ac), H3K9me2, H3K9me3, H3K27me3, H3K36me3, HP1a, HP1b, JMJD2b proteins were analyzed by western blots in MEFs. Label “C” represents the western blot data in non-treated cells, by “NC” are labeled MEFs, exposed to 0.001% DMSO (solvent for etoposide), and by ETOP 10 mM and ETOP 50 mM are labeled etoposide treated MEFs. B. Analysis of apoptotic markers, mCherry-tagged A-type lamins, showed distinct morphology of non-apoptotic and apoptotic living cells. Lamin A/C apoptotic fragmentation in MEFs was analyzed by western blotting. C. Percentage of nonapoptotic and apoptotic cells, distinguished according to morphology of nuclear lamina visualized by mCherry tagged A-type lamins, in non-apoptotic (control) and apoptotic (etoposide-treated) cells.
bodies, in which the fluorescence recovery of JMJD2b was significantly reduced to w10% (Fig. 9Ba). As a better explanation, JMJD2b antagonizes H3K9me3 in centromeric heterochromatin and H3K36me3 in the entire genome of intact cells [18,19]. Thus, the abrogated function of JMJD2b in apoptotic cells should rather lead to increased or stable levels of the relevant histone marks and it was confirmed in our experimental model (Fig. 8A). Our data for epigenome specificity in apoptotic cells are in a good agreement with Hurd et al. [9], which showed increased H3T45 phosphorylation in apoptotic bodies. Moreover, Ajiro [5] pointed out H2B phosphorylation as a universal marker of apoptotic DNA fragmentation that, on the chromosomal level, appears as the disorganization of chromosomal territories detected as random chromosomal segments of varying sizes [20]. BoixChornet et al. [21] showed H4K5, H4K8, H4K12, and H4K16 deacetylation in apoptotic chromatin of HL-60 cells, but surprisingly H3K9 acetylation and H3K9me2 levels remained stable during apoptotic process. In any case, these authors as the first showed specific histone signature of apoptotic cells, characterized by hypoacetylated histones H4. Cheng et al. [22] additionally found
dense H3K27me1 signal in staurosporine-induced apoptotic cells. Here, using three experimental approaches (western blotting, immunofluorescence, and mass spectrometry) we observed a general decrease in the acetylation state of all histones, although certain cell-type specific differences were observed (Figs. 2A, B, and 10). The results of our experiments particularize the general view on histone signatures in apoptotic cells. We provide information on changes in the balance between non-acetylated, mono-acetylated, and di-acetylated histones that can be induced by pro-apoptotic stimuli (Fig. 10). Interestingly, mass spectrometry showed a significant decrease in the di-acetylated form of histone H4 and monoacetylated form of H2B in apoptotic cell populations. This was logically balanced by an increased level of non-acetylated histones (summary in Fig. 10B). We also confirmed a prior report on GFP-HP1a and GFP-HP1b exchange kinetics in apoptotic cells [10]. As published by Cheutin et al. [10], HP1 isoforms recover rapidly after photobleaching in apoptotic bodies. Fast recovery of HP1 after FRAP was found in euchromatin and heterochromatin, but in different extent [10,23].
S. Legartová et al. / Biochimie 95 (2013) 167e179
177
Fig. 9. Fluorescence recovery after photobleaching (FRAP) for GFP-JMJD2b, GFP-BMI1, GFP-HP1a, GFP-HP1b proteins. A. Morphology of non-apoptotic and apoptotic cells transfected by full length GFP-JMJD2b, GFP-BMI1, GFP-HP1a, and GFP-HP1b. B. FRAP analysis of JMJD2b (a), BMI1 (b), HP1a (c), and HP1b (d) proteins in non-apoptotic (control) MEFs; etoposide treated non-apoptotic cells (nuclei with no apoptotic morphology), and etoposide-induced apoptotic cells. GFP-HP1a and GFP-HP1b fluorescence recovery after photobleaching was performed for HP1 located away from heterochromatin foci and accumulated into nuclear foci. Average relative fluorescence is shown with standard errors (S.E.). The cell nuclei after photobleaching were monitored during 23 s.
178
S. Legartová et al. / Biochimie 95 (2013) 167e179
Fig. 10. MALDIeTOF MS analysis of histone N-termini after etoposide treatment. Histone extracts from HL-60, U-937, and MEF cells were separated by RP-HPLC, digested by Glu-C, and analyzed by MALDIeTOF MS. A. Differences in the relative abundance of H4 N-terminal variants (1e53 AA; non-, mono-, and di-acetylated forms) of etoposide-treated samples and controls (data are shown as means and 95% confidence limits; n ¼ 15). For representative MALDIeTOF MS spectra see Supplement S1B. B. Overview of the effect of etoposide on histones H2A, H2B, and H4 acetylation status in particular cell lines.
Fluorescence recovery of HP1 isoforms after photobleaching can be changed by the inhibition of transcriptional elongation through actinomycin D [10,13]. Interestingly, identical HP1 exchanged kinetics, as after actinomycin D treatment, was observed in heterochromatic apoptotic cells, characterized by w30e40% recovery of HP1 after photobleaching [10]; (Fig. 9Bc and Bd). This observation may be tightly linked to the transcriptional silencing that occurs after actinomycin D treatment and/or in apoptotic cells. Altered dynamics of HP1 isoforms, JMJD2b and BMI-1 proteins may be attributed to the changes in posttranslational histone modifications. Especially de-acetylation, that occur together with changes in chromatin compaction, accompanied by apoptotic DNA fragmentation. 5. Conclusion The findings of this study contribute to our knowledge of apoptosis-related epigenetics. In several cell types we showed the existence of global histone hypoacetylation, especially H3K9 hypoacetylation and a decrease in H4 di-acetylation or H2B monoacetylation. These events were accompanied by an increased level of H3K9 methylation (me2 or me3) in apoptotic cells. The H3K9 acetylation pattern was cell-type specific (Figs. 2A, B and 8A), similar to the change in acetylation state studied by mass spectrometry (Fig. 10B). Levels of other histone markers (H3K4me1/me2/me3; H3K9me1; H3K27me3) were identical in non-apoptotic and apoptotic cell populations. In addition to the results on nonacetylated, mono- and di-acetylated histones in apoptotic cells, another novel observation is that not all components of highly compacted apoptotic heterochromatin have identical kinetic properties. As described above, GFP-tagged HP1 protein recovered more rapidly in apoptotic cells, while fluorescence of GFP-JMJD2b has never recovered up to more then 10%. Similarly, fluorescence recovery of GFP-BMI1 protein in apoptotic cells was w20% of prebleaching intensity (Fig. 9A, B). It is in contrast to intact cells, characterized by the BMI1 fluorescence recovery to w60e70% of prebleaching intensity in heterochromatin and 80e90% recovery of BMI1 in chromatin away from PcG bodies [24]. These data unambiguously document specificity in protein diffusion in apoptotic chromatin, generally characterized by histone deacetylation.
Taken together, it seems evident that apoptotic bodies represent a functionally specialized form of heterochromatin, characterized by a unique histone signature and kinetics of chromatin-related proteins, including HP1 sub-types, BMI1, and JMJD2b. Here, we pointed out that HP1 protein, H3K9 methylation and histone deacetylation, unlike JMJD2b, BMI1, H3K27me3 and H3K4 methylation, seem to be the most important epigenetic factors, responsible for the maintenance of apoptotic bodies, which represent the most condensed form of chromatin. Conflict of interest We declare no conflict of interest. Acknowledgments Our work was supported by following agencies: Grant Agency of Czech Republic, grant Nos.: P302/10/1022; P302/12/G157 and by Ministry of Education Youth and Sports of the Czech Republic, grant No.: LD11020. EB is also a member of the EU COST action TD09/05 and principal investigator of EU Marie Curie project PIRSES-GA2010-269156-LCS and project of LS supported by ASCR, v.v.i., No: M200041271. Work of ZZ and GL was supported by the project “CEITEC - Central European Institute of Technology” (CZ.1.05/ 1.1.00/02.0068) from European Regional Development Fund. For re-cloning assistance we thank Dr. Oskar Laur, Emory University Scholl of Medicine, Atlanta, USA. Appendix A. Supplementary material Supplementary material associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biochi. 2012.09.027. References [1] A.H. Wyllie, J.F. Kerr, A.R. Curie, Cell death: the significance of apoptosis, Int. Rev. Cytol. 68 (1980) 251e306. [2] G.M. Cohen, X.M. Sun, H. Fearnhead, M. MacFarlane, D.G. Brown, R.T. Snowden, D. Dinsdale, Formation of large molecular weight fragments of
S. Legartová et al. / Biochimie 95 (2013) 167e179
[3] [4] [5] [6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
DNA is a key committed step of apoptosis in thymocytes, J. Immunol. 153 (1994) 507e516. S.J. Martin, D.R. Green, T.G. Cotter, Dicing with death: dissecting the components of the apoptosis machinery, Trends Biochem. Sci. 19 (1994) 26e30. M.O. Hengartner, Apoptosis. DNA destroyers, Nature 412 (2001) 27e29. K. Ajiro, Histone H2B phosphorylation in mammalian apoptotic cells. An association with DNA fragmentation, J. Biol. Chem. 275 (2000) 439e443. E.P. Rogakou, W. Nieves-Neira, C. Boon, Y. Pommier, W.M. Bonner, Initiation of DNA fragmentation during apoptosis induces phosphorylation of H2AX histone at serine 139, J. Biol. Chem. 275 (2000) 9390e9395. P. Waring, T. Khan, A. Sjaarda, Apoptosis induced by gliotoxin is preceded by phosphorylation of histone H3 and enhanced sensitivity of chromatin to nuclease digestion, J. Biol. Chem. 272 (1997) 17929e17936. M. Wu, L.G. Xu, X. Li, Z. Zhai, H.B. Shu, AMID, an apoptosis-inducing factorhomologous mitochondrion-associated protein, induces caspase-independent apoptosis, J. Biol. Chem. 277 (2002) 25617e25623. P.J. Hurd, A.J. Bannister, K. Halls, M.A. Dawson, M. Vermeulen, J.V. Olsen, H. Ismail, J. Somers, M. Mann, T. Owen-Hughes, I. Gout, T. Kouzarides, Phosphorylation of histone H3 Thr-45 is linked to apoptosis, J. Biol. Chem. 284 (2009) 16575e16583. T. Cheutin, A.J. McNairn, T. Jenuwein, D.M. Gilbert, P.B. Singh, T. Misteli, Maintenance of stable heterochromatin domains by dynamic HP1 binding, Science 299 (2003) 721e725. A. Jugová, G. Sustá cková, S. Legartová, L. Stixová, S. Kozubek, E. Bártová, Effects of epigenetic-based anti-cancer drugs in leukaemia and multiple myeloma cells, Cell Biol. Int. 35 (2011) 1195e1203. E. Bártová, J. Pacherník, A. Harni carová, A. Kovarík, M. Kovaríková, J. Hofmanová, M. Skalníková, M. Kozubek, S. Kozubek, Nuclear levels and patterns of histone H3 modification and HP1 proteins after inhibition of histone deacetylases, J. Cell Sci. 118 (2005) 5035e5046. k, S. Legartová, S. Kozubek, HeteroL. Stixová, E. Bártová, P. Matula, O. Dane geneity in the kinetics of nuclear proteins and trajectories of substructures associated with heterochromatin, Epigenet. Chromatin 4 (2011) 5. E. Bártová, L. Stixová, G. Galiová, A. Harni carová Horáková, S. Legartová, S. Kozubek, Mutant genetic background affects the functional rearrangement
[15]
[16]
[17] [18]
[19]
[20] [21]
[22]
[23]
[24]
179
and kinetic properties of JMJD2b histone demethylase, J. Mol. Biol. 405 (2011) 679e695. D. Bonenfant, H. Towbin, M. Coulot, P. Schindler, D.R. Mueller, J. van Oostrum, Analysis of dynamic changes in post-translational modifications of human histones during cell cycle by mass spectrometry, Mol. Cell. Proteomics 6 (2007) 1917e1932. L. Cin cárová, G. Lochmanová, K. Nováková, P. Sultesová, H. Kone cná, L. Fajkusová, J. Fajkus, Z. Zdráhal, A combined approach for the study of histone deacetylase inhibitors, Mol Biosyst. 8 (2012) 2937e2945. R. Paro, H. Strutt, G. Cavalli, Heritable chromatin states induced by the Polycomb and trithorax group genes, Novartis Found. Symp. 214 (1998) 51e61. B.D. Fodor, S. Kubicek, M. Yonezawa, R.J. O’Sullivan, R. Sengupta, L. PerezBurgos, S. Opravil, K. Mechtler, G. Schotta, T. Jenuwein, Jmjd2b antagonizes H3K9 trimethylation at pericentric heterochromatin in mammalian cells, Genes Dev. 20 (2006) 1557e1562. A. Harni carová Horáková, G. Galiová, S. Legartová, S. Kozubek, P. Matula, E. Bártová, Chromocentre integrity and epigenetic marks, J. Struct. Biol. 169 (2010) 124e133. E. Bártová, P. Jirsová, M. Fojtová, K. Sou cek, S. Kozubek, Chromosomal territory segmentation in apoptotic cells, Cell. Mol. Life Sci. 60 (2003) 979e990. M. Boix-Chornet, M.F. Fraga, A. Villar-Garea, R. Caballero, J. Espada, A. Nunez, J. Casado, C. Largo, J.I. Casal, J.C. Cigudosa, L. Franco, M. Esteller, E. Ballestar, Release of hypoacetylated and trimethylated histone H4 is an epigenetic marker of early apoptosis, J. Biol. Chem. 281 (2006) 13540e13547. M.F. Cheng, C.H. Lee, K.T. Hsia, G.S. Huang, H.S. Lee, Methylation of histone H3 lysine 27 associated with apoptosis in osteosarcoma cells induced by staurosporine, Histol. Histopathol. 24 (2009) 1105e1111. R. Festenstein, S.N. Pagakis, K. Hiragami, D. Lyon, A. Verreault, B. Sekkali, D. Kioussis, Modulation of heterochromatin protein 1 dynamics in primary mammalian cells, Science 299 (2003) 719e721. G. Sustá cková, S. Kozubek, L. Stixová, S. Legartová, P. Matula, D. Orlova, E. Bártová, Acetylation-dependent nuclear arrangement and recruitment of BMI1 protein to UV-damaged chromatin, J. Cell. Physiol. 227 (2012) 1838e1850.