Histone deacetylases 1 and 2 maintain S-phase chromatin and DNA replication fork progression
© Bhaskara et al.; licensee BioMed Central Ltd. 2013
Received: 25 April 2013
Accepted: 26 July 2013
Published: 15 August 2013
Skip to main content
© Bhaskara et al.; licensee BioMed Central Ltd. 2013
Received: 25 April 2013
Accepted: 26 July 2013
Published: 15 August 2013
Histone deacetylases (HDACs) play a critical role in the maintenance of genome stability. Class I HDACs, histone deacetylase 1 and 2 (Hdac1 and Hdac2) are recruited to the replication fork by virtue of their interactions with the replication machinery. However, functions for Hdac1 and Hdac2 (Hdacs1,2) in DNA replication are not fully understood.
Using genetic knockdown systems and novel Hdacs1,2-selective inhibitors, we found that loss of Hdacs1,2 leads to a reduction in the replication fork velocity, and an increase in replication stress response culminating in DNA damage. These observed defects are due to a direct role for Hdacs1,2 in DNA replication, as transcription of genes involved in replication was not affected in the absence of Hdacs1,2. We found that loss of Hdacs1,2 functions increases histone acetylation (ac) on chromatin in S-phase cells and affects nascent chromatin structure, as evidenced by the altered sensitivity of newly synthesized DNA to nuclease digestion. Specifically, H4K16ac, a histone modification involved in chromatin decompaction, is increased on nascent chromatin upon abolishing Hdacs1,2 activities. It was previously shown that H4K16ac interferes with the functions of SMARCA5, an ATP-dependent ISWI family chromatin remodeler. We found SMARCA5 also associates with nascent DNA and loss of SMARCA5 decreases replication fork velocity similar to the loss or inhibition of Hdacs1,2.
Our studies reveal important roles for Hdacs1,2 in nascent chromatin structure maintenance and regulation of SMARCA5 chromatin-remodeler function, which together are required for proper replication fork progression and genome stability in S-phase.
Histone deacetylase inhibitors (HDAC inhibitors/HDIs) are potent anticancer drugs. Several broad-spectrum inhibitors are in various stages of clinical trials for both solid tumors and hematopoietic malignancies . Two of these compounds (SAHA/Vorinostat and Depsipeptide/Romidepsin) have gained FDA approval for their use against T-cell cutaneous lymphomas. SAHA and Depsipeptide target class I HDACs (Hdacs 1, 2, 3 and 8) . Therefore, it is imperative to study and understand the specific functions of individual HDACs in order to ascertain drug specificity and design targeted therapeutics with increased potency and minimal side effects.
Hdac1 and Hdac2 are the core enzymes in three distinct protein complexes (Sin3, NURD and CoREST) that have diverse cellular functions . Targeted deletion of Hdac1 led to embryonic lethality . Hdac2-null pups die within a month due to cardiac defects and abnormalities in myocyte proliferation . Knockout of either Hdac1 or Hdac2 had minimal effect on hematopoiesis and on the cell cycle, likely due to compensation for one by the other, as they are highly similar proteins. However, deletion of both genes dramatically impaired proliferation in multiple cell types by blocking cells at the G1 to S phase transition [6, 7]. Additionally, double knockout of these enzymes caused mitotic catastrophe in fibrosarcoma cells . While a role for HDACs in transcription is well established [9–11], these enzymes also function in DNA replication. Hdac1 interacts with proliferating cell nuclear antigen (PCNA, a DNA replication processivity factor) . Also, Hdacs1,2 associate with newly replicated DNA . However, the precise functions for Hdacs1,2 in DNA replication are still not fully understood. Broad-spectrum (‘pan’) HDAC inhibitors inhibit cell cycle progression and kill cancer cells by triggering DNA damage during DNA replication/S-phase . Hence, it is important to understand how Hdacs1,2 function during replication and how these HDACs affect nascent chromatin.
In this study, we aimed to understand the mechanistic roles for Hdacs1,2 in DNA replication. Using novel selective inhibitors and genetic knockdown systems, we show that Hdacs1,2 functions are required for the proper progression of the DNA replication fork, and loss of Hdacs1,2 activities leads to the activation of replication stress and DNA damage response. This defect in replication is not simply caused by changes in transcription, as gene expression for factors involved in replication remain unchanged following Hdacs1,2-inhibitor treatment. Mechanistically, the defect in replication in the absence of Hdacs1,2 functions can be attributed to an altered chromatin structure, due to increased histone acetylation on S-phase chromatin, especially increased H4K12ac and H4K16ac levels. H4K12ac and H4K16ac antagonize substrate recognition and nucleosome remodeling activity of SMARCA5, an ISWI family chromatin remodeler [15, 16]. In this study, we further show that SMARCA5 is present on nascent chromatin, and loss of SMARCA5 also leads to a decrease in the replication fork velocity and activation of the replication stress response. This demonstrates an important role for SMARCA5 in replication fork progression in mammalian cells. Overall, we provide a model wherein Hdacs1,2 affect DNA replication fork progression by regulating histone acetylation on nascent chromatin and SMARCA5 activity.
To further examine if histone acetylation marks increase upon inhibition of Hdacs1,2 activities, we chose to selectively inactivate these two enzymes using novel, benzymilic class small molecule inhibitors (RGFP898 and RGFP233, henceforth referred to as 898 and 233, respectively). We first determined the selectivity of these two molecules towards Hdacs1,2. The IC50 values obtained using in vitro HDAC assays showed 233 and 898 inhibit Hdacs1,2 activities at a low concentration (Additional file 3: Figure S3A). Unlike SAHA, the inhibitory activity of RGFP106 (another benzamide-type inhibitor similar to 898 or 233) was previously shown to remain unchanged even after 100-fold dilution of the inhibitor-enzyme mixture and histone acetylation did not return to basal levels even after washing away the inhibitor . Therefore, these benzamide-type Hdacs1,2 inhibitors are slow and tight-binding compounds. We next examined the efficacy of 898 and 233 to inhibit Hdacs1,2 in NIH3T3 cells. An increase in histone acetylation was observed following treatment of NIH3T3 cells with 2 to 10 μM 898 (Additional file 3: Figure S3B). We then determined the minimum concentration range required to inhibit Hdac1,2 activities and to increase histone acetylation in NIH3T3 cells. A robust inhibition of only Hdacs1,2 activities was observed at lower concentrations of 898 or 233 (3.0 to 3.75 μM) (Figure 1D, 1E). To ensure the reduced enzyme activity is not due to differences in the enzyme concentrations used in the assay, we checked and confirmed that, indeed, equal amount of Hdac1, Hdac2 and Hdac3 were present in the immunoprecipitates (Additional file 4: Figure S4). Collectively, these characterization studies confirmed the efficacy of 898 and 233 as Hdac1,2-selective inhibitors, and provided us the minimal, effective concentration range for these two inhibitors to be used in our studies (3 to 3.75 μM).
Similar to the knockdown of Hdacs1,2 (Figure 1C), inhibition of Hdacs1,2 in vivo using the selective inhibitors (898 or 233) also resulted in an increase in H4K5ac, H4K12ac and H3K9,K14ac levels when compared to cells treated with vehicle alone (DMSO) (Figure 1F-G). Given their high sequence homology [22, 23], we sought to further confirm the specificity of 233 and 898 towards only Hdacs1,2 and not Hdac3. To this end, we used fibrosarcoma cells containing floxed alleles of either Hdac1 and Hdac2 (Hdac1Fl/FlHdac2Fl/Fl) or Hdac3 (Hdac3Fl/Fl) to obtain conditional knockout of these enzymes upon expressing Cre recombinase . Efficient depletion of Hdacs1,2 and Hdac3 were observed in these cells following infection with an adenovirus-containing Cre recombinase (Ad-Cre) (Figure 1H and 1I). Conditional deletion of Hdacs1,2 in fibrosarcoma cells led to a significant increase in H4K5ac (Figure 1H), whereas deletion of Hdac3 led to a subtle increase in H4K5ac (Figure 1I). Treatment of Hdac1,2 knockout cells with 233 or 898 did not result in any further increase in H4K5ac (Figure 1H, Additional file 5: Figure S5A and S5B), confirming that these two inhibitors are selective for Hdacs1,2. Addition of 233 or 898 to Hdac3 knockout cells resulted in a significant increase in H4K5ac (Figure 1I). This increase in H4K5ac is an additive effect obtained due to the inhibition of Hdacs1,2 activities by these two molecules combined with the loss of Hdac3 activity (Figure 1I and Additional file 5: Figures S5C and S5D). Taken together, our studies using genetic systems and selective inhibitors reveal a role for Hdacs1,2 in the removal of histone deposition marks.
To examine if Hdacs1,2 activities are required for DNA replication, we released serum-starved cells into S-phase and treated them for 12 h, 18 h or 24 h with either the vehicle (DMSO) or the Hdac1,2-selective inhibitor (898 or 233). Nascent DNA was labeled with BrdU prior to harvesting. Changes in BrdU incorporation were assessed by slot blot analysis using equal amounts of genomic DNA isolated from DMSO-, 898- or 233-treated cells. We observed a two-fold reduction in BrdU incorporation in cells treated with the 898 or 233 when compared to the untreated control cells (Figure 2B and Additional file 9: Figure S9). Importantly, this finding suggests that Hdacs1,2 activities are required for efficient synthesis of nascent DNA during DNA replication.
Defective DNA replication in the absence of Hdacs1,2 activities might be a cause for the reduced BrdU incorporation in 898- or 233-treated S-phase cells (Figure 2B and Additional file 9: Figure S9). To test this possibility, we used the molecular combing assay to examine if loss or inhibition of Hdacs1,2 affects replication fork progression. In the combing assay, cells are sequentially pulse-labeled with two different halogenated thymidine analogs (IdU, iodo-deoxyuridine and CldU, choloro-deoxyuridine) to independently mark initiation/early elongation events and subsequent progression of the fork at replicating origins. After spreading the DNA fibers on a slide (combing), newly replicated regions are detected using fluorescent dye-labeled antibodies that specifically recognize the two different incorporated thymidine analogs. Length of the fluorescence signal and the labeling time are then used to calculate the replication fork velocity (Kb/min). In this assay, defects in replication fork progression or elongation can be further exacerbated via stalling the fork using a dose of hydroxyurea (HU, a ribonucleotide reductase inhibitor) that does not cause fork collapse .
To measure changes, if any, in the replication fork velocity upon abrogation of Hdacs1,2 activities, we released serum-starved cells into S-phase and treated them with DMSO (vehicle) or with the Hdacs1,2-selective inhibitor (898 or 233) followed by pulse labeling of nascent DNA with IdU (Figure 2C). After removing any free IdU, we performed the second pulse labeling with CldU in the presence of HU. We then measured the length of the two pulse-labels to obtain replication fork velocities, which were further classified into categories based on the distance travelled by the fork. We performed box plot analysis to measure the average fork velocity (Figure 2C, 2D). We also classified fibers into bins of increasing velocities to examine if loss of Hdacs1,2 activities affects fibers of a particular velocity range (Additional file 10: Figure S10A). With IdU labeling, replication fork velocities were reduced in the presence of Hdacs1,2-selective inhibitor (898 or 233) compared to the control (Figure 2C and Additional file 10: Figure S10A, see 0.95 to 1.6 Kb/min). Defects in replication fork progression due to inhibition of Hdacs1,2 were also evident for CldU labeling in the presence of hydroxyurea (HU), which slows/stalls fork progression (Figure 2C and Additional file 10: Figure S10A). In addition, we observed a severe reduction in the fork velocity upon treatment of NIH3T3 cells with SAHA that inhibits Hdacs1, 2 and 3 (Additional file 11: Figure S11). We further measured the replication fork velocities in NIH3T3 cells following knockdown of Hdacs1,2. Loss of Hdacs1,2 caused a decrease in fork velocity (Figure 2D, IdU label and Additional file 10: Figure S10B), which was further affected in the presence of hydroxyurea (Figure 2D, CldU label). This decrease in fork velocities found upon loss of Hdacs1,2 correlates well with that observed upon preventing their activities using selective inhibitors (Figure 2C). Collectively, our findings demonstrate that Hdac1,2 functions are required for maintaining normal replication fork rates.
We next sought to explore the molecular mechanism(s) by which Hdacs1,2 might promote replication fork progression. Reduced fork velocity might be due to a shortage of cellular dNTP pool, as seen upon hydroxyurea treatment , or alternatively, due to a decrease in transcription of genes involved in nucleotide biosynthesis. Treatment of cells with trichostatin A (TSA), a pan-HDAC inhibitor, reduced fork velocity, due to its effect on pyrimidine biosynthesis. TSA treatment decreased the expression of CTP synthetase 1 and thymidylate synthetase genes, which in turn reduced pyrimidine biosynthesis . To examine whether reduced fork velocity in 898-treated cells is linked to defects in transcription, we used RNA-seq to determine differential gene expression in three independent DMSO or 898-treated S-phase cells. We observed differential expression of 70 genes, including upregulation of cytochrome Cyb561 gene (Figure 3E, Additional file 16: Figure S16A, Additional file 17: Table S1). However, transcript levels for CTP synthetase 1 gene were not affected in all three independently treated samples (Figure 3E). Also, expression of genes coding for factors involved in DNA replication or DNA damage response remained unchanged (Additional file 16: Figure S16B). Therefore, our gene-expression analysis suggests that the reduced fork velocity and DNA damage observed upon inhibition of Hdac1,2 activities are not likely due to altered gene transcription.
It is conceivable that Hdacs1,2 might target a non-histone protein(s) with an important role in DNA replication. Smc3, a subunit of the cohesin complex, regulates replication and is acetylated by Eco1 acetyl transferase . To test whether loss or inhibition of Hdacs1,2 affects Smc3 acetylation, we used an antibody that specifically recognizes the acetylated form of Smc3. While Smc3ac levels on chromatin increased in S-phase cells, we did not observe any change in the levels of Smc3 or its acetylated form following treatment with 898 (Additional file 18: Figure S17A) or following siRNA-mediated knockdown of Hdacs1,2 (Additional file 18: Figure S17B). These results suggest that Hdacs1,2 are not involved in regulating Smc3 or its acetylation, and this agrees well with the recent finding that Hdac8 is involved in Smc3 deacetylation . It is possible that Hdacs1,2 target some other non-histone protein(s) involved in DNA replication.
Heterochromatin protein 1 (HP1) has been linked to assembly and maintenance of heterochromatin and to DNA replication . HP1 binds methylated H3K9 . Since H3K9,K14ac is increased in the absence of Hdacs1,2 functions (Figure 1C, Additional file 3: Figure S3B), we tested whether Hdacs1,2 play a role in the chromatin binding of replication-associated forms of HP1 (that is, HP1α and HP1γ)  indirectly via their regulation of histone modifications. We found chromatin-bound levels of HP1α and HP1γ in 898-treated S-phase cells were not affected compared to the untreated control cells (Figure 3F). Similar results were obtained following knockdown of Hdacs1,2 (Figure 3G). These findings suggest that global heterochromatin is not affected upon transient inhibition of Hdacs1,2 activities, and rule out reduced HP1 binding to chromatin as a reason for the DNA replication defects observed upon abrogating Hdacs1,2 functions.
Modulation of chromatin structure around a replication fork is achieved by the concerted action of histone variants, histone modifying enzymes, chromatin remodelers, histone chaperones and numerous chromatin-binding factors. It is conceivable that histone acetylation might be required to maintain a permissive chromatin conformation for the replication fork to progress. During S-phase, newly synthesized histone H4 is acetylated at K5 and K12 residues and deposited onto nascent chromatin by CAF-1, a histone chaperone . Also, H4K16ac is enriched at initiation zones and at early replication regions . Removal of H4K5ac and H4K12ac by HDACs following their deposition onto nascent chromatin was proposed to be an event in chromatin maturation during DNA replication . Using selective inhibitors, we now show that Hdacs1,2 target histone deposition marks within S-phase cells and on nascent DNA (Figure 5A). Since H4K16ac prevents chromatin compaction, we propose a model, wherein Hdacs1,2 remove H4K16ac to allow chromatin restoration/reassembly step following DNA replication (Figure 10B). Indeed, we find that in the absence of Hdacs1,2 activities the nascent chromatin and nucleosomes at candidate replication loci are more susceptible to nuclease digestion (Figure 9), indicative of a more ‘permissive’ or ‘loose’ chromatin conformation. Hence, it is tempting to speculate that this atypical chromatin structure signals the replication machinery to stall, collapse, and trigger the DNA damage response. The resulting double-strand break and S-phase lesions could cause severe chromosome segregation defects. We find DNA damage and stress response activation upon inhibiting Hdacs1,2 (Figures 3A-B, Additional files 12, 13, 14: Figures S12, S13, S14), which corroborates previous findings that loss of Hdacs1,2 causes severe mitotic catastrophe .
Chromatin maturation not only involves histone deacetylation, but also nucleosome remodeling. Human cells contain two isoforms of ISWI: SNF2H and SNF2L . The SNF2H/SMARCA5 complex remodels nucleosomes to allow smooth movement of the fork, especially through the heterochromatin . SMARCA5 interacts with PCNA and with Hdacs1,2 [43, 48]. In yeast, loss of Iswi2 and Ino80 causes defects in replication fork progression . Mammalian ACF1-SNF2H is required for replication through heterochromatin, and depletion of SMARCA5 decreases BrdU incorporation in HeLa cells . However, function for SMARCA5 in fork elongation in mammalian cells was not known. In this study, we show that SMARCA5 associates with nascent DNA (Figure 6D), and its occupancy on chromatin and at candidate replication origins increases in S-phase (Figure 6A and Figure 7). Importantly, we show that SMARCA5 has a direct role in controlling fork elongation, as its deletion reduces replication fork rates (Figure 10A).
The ISWI family chromatin remodelers have the catalytic ATPase domain and a SANT domain that binds the H4 tail. H4K16ac and H4K12ac were shown to interfere with the ability of ISWI to interact with the H4 tail and they inhibit the ATPase activity of the ISWI complex [15, 16]. We show Hdacs1,2 target H4K16ac on newly synthesized DNA (Additional file 20: Figure S19). Also, we show that H4K12ac and H4K16ac are targeted by Hdacs1,2 in S-phase cells (Figure 5A, Figure 8D-F). Interestingly, we find an inverse correlation for the occupancy of H4K16ac and SMARCA5 at candidate replication origins (α-globin, β-globin and pancreatic amylase) in S-phase cells (Figure 7, Figure 8D-F). While H4K16ac levels are high at these loci in the non-replicating G0/G1 phase, they are reduced by Hdacs1,2 as cells enter and progress through the S-phase (Figure 8D-F). On the other hand, SMARCA5 levels are low at these loci in G0/G1 phase and increase during the S-phase. Therefore, these findings support a model (Figure 10B), wherein Hdacs1,2 might regulate SMARCA5 activity at the replication forks via removal of H4K12ac and H4K16ac. Increase in H4K12ac and H4K16ac around the fork upon loss or inhibition of Hdacs1,2 functions might inhibit SMARCA5-mediated chromatin remodeling, which is necessary for fork progression (Figure 10A), resulting in fork stalling and collapse. Collectively, we favor a model (Figure 10B), wherein Hdacs1,2 control nascent chromatin structure in two modes: one, affecting nucleosome structure and chromatin packaging by directly regulating histone acetylation and two, by regulating nucleosome remodeling via modulation of chromatin remodeler activity. Since abrogation of Hdacs1,2 functions alone is sufficient to impair DNA replication and compromise chromatin and genome stability during S-phase, selective inhibition of Hdacs1,2 might be an efficient therapeutic strategy to minimize the side effects of pan-HDIs that are currently used in cancer treatment.
In this study, we report functions for Hdacs1,2 in maintaining normal replication fork progression and in nascent chromatin maintenance using novel Hdacs1,2-selective inhibitors and siRNA-mediated genetic knockdown systems. SAHA, a pan-inhibitor that targets all class I HDACs, was shown to affect replication fork velocity in cancer cells without affecting transcription . Here, we show that Hdacs1,2 (a subset of SAHA targets) play a direct role in DNA replication without disrupting transcription of genes involved in DNA replication, repair or nucleotide biosynthesis. We further show that inhibiting Hdacs1,2 alters nascent chromatin architecture (histone acetylation and compaction), reduces replication fork velocity and triggers DNA damage response. These findings highlight the important role for Hdacs1,2 in genome stability maintenance. In addition, we show that SMARCA5, an ISWI-family chromatin remodeler, is present on nascent chromatin and is required for proper progression of DNA replication. Therefore, in this study, we have connected the functions of a chromatin remodeler (SMARCA5), histone modifications (H4K12ac and H4K16ac) and histone deacetylases (Hdacs1,2 that target H4K12ac and H4K16ac) to the progression of replication fork.
HeLa and HEK 293 cells were cultured in DMEM containing 10% fetal bovine serum (Hyclone, Logan, UT, USA), 1% penicillin-streptomycin and 1% glutamine. NIH3T3 cells were serum starved for 72 h in 0.5% serum containing media. NIH3T3 cells were cultured in DMEM (Cellgro™, Tewksbury, MA, USA) containing 10% fetal calf serum, 1% penicillin-streptomycin and 1% glutamine. NIH3T3 cells were serum starved for 72 h in 0.5% serum containing media. Fibrosarcoma cells for conditional knockout of Hdac1,2 or Hdac3 were cultured as described previously .
Cells were transfected with siGenome SMART pool for mouse Hdac1, or siGenome SMART pool for mouse Hdac2, siGenome SMART pool for human SMARCA5, or with non-specific control pool (siRNA negative control) as described previously . All siRNAs were purchased from Dharmacon (Lafayette, CO, USA).
NIH3T3 cells transfected with non-targeting (NT) or Hdac1 & Hdac2 siRNAs were labeled with 20 μM IdU (iodo-deoxyuridine) for 15 min following 72 h post-transfection. Cells were washed with PBS and labeled with 100 μM CldU (chloro-deoxyuridine) in the presence of 250 μM hydroxyurea for 20 min. Cells were lysed with spreading buffer (0.5% SDS in 200 mM Tris–HCl, pH 7.4 and 50 mM EDTA) and DNA fibers were spread on silane-coated slides. Following fixation and DNA denaturation, immunofluorescence (IF) was performed using anti-IdU and anti-CldU antibodies and mouse anti-BrdU-conjugated to Alexa 488 and rat anti-CldU conjugated to Alexa 594 (secondary antibodies). Fiber images were captured using an Axioscope microscope. The lengths of approximately 100 fiber tracks were measured using the ImageJ software. The fiber length (μm) was converted into Kb DNA length after taking the stretching factor (1 μm = 2 Kb DNA) into consideration. The resulting value was then divided by the incubation time to obtain the fork velocity.
ChIP assays were performed as described previously .
Immunofluorescence was performed as described previously .
BrdU-PI analysis was performed as described previously .
PCR primers for ChIP analysis at replication origins were described previously .
NIH3T3 cells were labeled with 20 μM BrdU, washed with ice-cold phosphate-buffered saline (PBS). Cell lysis buffer (10 mM Tris–HCl, pH 7.4, 300 mM sucrose, 3 mM CaCl2, 2 mM Mg(CH3COO)2, 0.5% NP-40, 5 mM dithiothreitol (DTT) and 1X Roche protease inhibitor cocktail) was added to the plate and left on ice for 5 min. Cells were scraped following lysis, Dounce homogenized fifty times, spun at 1000 rpm for 5 min at 4°C. Nuclei were stored in nuclei storage buffer (50 mM Tris- HCl, pH 8.3, 40% glycerol, 0.1 mM EDTA, 5 mM Mg(CH3COO)2, 5 mM DTT and 1X Roche protease inhibitor cocktail) and stored at −80°C until use.
Nuclear extract was prepared in RIPA buffer supplemented with protease inhibitors (Roche) and precleared with Protein A-agarose beads (Millipore, Billerica, MA, USA) for 20 min at 4°C with constant rotation. The precleared lysate was then incubated with anti-PCNA antibody for 12 h at 4°C with constant rotation. Protein A-agarose beads equilibrated in RIPA buffer was then added to the samples, and pull-down was done at 4°C for 1 h with constant rotation. The beads were then washed with RIPA buffer for three times and resuspended in 1X SDS sample buffer before Western analysis.
NIH3T3 cells were washed with ice-cold PBS. Cells were scraped and spun at 3000 rpm at 4°C for 5 min. The cell pellet was resuspended in buffer A (10 mM HEPES, pH 7.9, 10 mM KCl, 1.5 mM MgCl2, 0.34 M sucrose, 10% glycerol, 1 mM dithiothreitol (DTT), and protease inhibitor cocktail). Triton X-100 (0.1% final concentration) was added to the extract, and incubated on ice for 8 min. Nuclei (fraction P1) were collected by centrifugation (5 min, 8000 rpm, 4°C). The nuclear pellet was resuspended in RIPA buffer containing protease inhibitor cocktail and sonicated to solubilize the chromatin. To prepare chromatin extract, the isolated nuclei (fraction P1) were resuspended in hypotonic buffer (3 mM EDTA, 0.2 mM EGTA, 1 mM DTT) supplemented with protease inhibitor cocktail (Roche, Penzberg, Upper Bavaria, Germany) and incubated on ice for 30 min. Following centrifugation, the pellet was resuspended in RIPA buffer containing protease inhibitors and sonicated briefly to solubilize the chromatin. Protein concentrations were measured using Bio-Rad Protein Assay or Bio-Rad DC™ protein assay kits (Bio-Rad, Hercules, CA, USA)’.
NIH3T3 cells were treated with DMSO (control) or the selective HDAC inhibitors for 24 h. Nuclei were isolated as described above and resuspended in HERR buffer (20 mM HEPES pH 7.9, 150 mM KCl, 0.1% NP40, 10% glycerol, 2 mM EDTA). Extracts were sonicated and precleared with Protein A-agarose beads (Millipore) for 20 min at 4°C with constant rotation. The precleared lysate was then incubated with anti-Hdac1 or -Hdac2 or -Hdac3 antibodies for 4 hr at 4°C with constant rotation. Protein A-agarose beads were then added to samples and pull down was done at 4°C for 1 hr with constant rotation. The beads were then washed with HERR buffer for three times and HDAC assay was performed as per the recommended protocol provided with the Fluor-de-Lys™ HDAC fluorometric activity assay kit (Enzo Life Sciences, Farmingdale, NY, USA). The activity was measured using the 2104 EnVision™ Multilabel Reader (PerkinElmer; excitation at 340 nm and emission at 495 and 520 nm). In vitro HDAC assays were performed using recombinant HDAC enzymes and the Fluor-de-Lys™ HDAC fluorometric activity assay kit.
Nuclei were isolated from cells treated with DMSO (control) or Hdacs1,2-selective inhibitor (898), or from cells transfected with non-targeting siRNA or Hdacs1,2-specific siRNA followed by labeling with 20 μM BrdU for 1 hr. MNase digestion of nuclei was performed essentially as described previously . Briefly, equal numbers of nuclei were digested with 0, 0.5, 2 or 8 units of MNase (Worthington Biochemical Co., Lakewood, NJ, USA) for 5 min at 37°C. Nuclease digestion was terminated following addition of equal volume of 2X stop buffer (20 mM Tris. Cl, pH 7.5, 200 mM NaCl, 2% SDS and 20 mM EDTA). Following nuclease digestion, genomic DNA was isolated using the Easy-DNA™ Kit (Invitrogen) as per the manufacturer’s instructions. DNA concentration was measured using NanoDrop 1000 spectrophotometer (ThermoScientific). Undigested DNA (0.5 μg) or MNase-digested DNA (1 μg) were resolved in a 1.5% agarose gel and stained with ethidium bromide. The DNA was then transferred onto a charged membrane (Hybond™-N+, GE Healthcare, Pittsburgh, PA, USA) using capillary transfer following the standard Southern blotting procedure. The membrane was subjected to UV crosslinking to immobilize DNA. Membrane-bound BrdU-labeled nascent DNA was detected by Western blotting using an anti-BrdU antibody (used at 1:2000 dilution).
NIH3T3 cells were serum starved for 72 h and released into serum-rich medium to progress into S-phase in the presence of DMSO or 3 μM Hdacs1,2-selective inhibitor (898) and treated for 20 h prior to crosslinking with 0.1% formaldehyde for 10 min. Serum-starved NIH3T3 cells prior to release into S-phase were also crosslinked similarly with formaldehyde. Crosslinking was quenched by adding glycine (125 mM final concentration) and incubating for 5 min at room temperature. Nuclei were isolated as described above and counted. Equal number of nuclei from different treatments (0.5 × 106) were washed once in Buffer D (50 mM Tris- HCl, pH 8.0, 25% glycerol, 0.1 mM EDTA, 5 mM Mg(CH3COO)2, 5 mM DTT) and then resuspended in 0.2 ml of Buffer MN (15 mM Tris, pH 7.4, 60 mM KCl, 15 mM NaCl, 0.5 mM DTT, 0.25 M sucrose and 1 mM CaCl2). Nuclei (0.1 ml) were digested with micrococcal nuclease (10 units, Worthington Biochemical Co.) by incubation for 7 min at 37°C and 400 μg RNase A (Qiagen, Vinlo, Limberg, Germany) was added and digestion was continued for an additional 3 min. Digestion was stopped by adding an equal volume of 2X stop buffer (20 mM Tris. Cl, pH 7.5, 200 mM NaCl, 2% SDS and 20 mM EDTA). To the remaining nuclei (0.1 ml, set aside for undigested DNA control) an equal volume of 2X stop buffer was added. DNA from undigested and MNase-digested nuclei was isolated using Qiagen DNeasy™ Kit. MNase-digested DNA was resolved in a 1.8% agarose gel. Mononucleosomal DNA (approximately 146 bp) was excised and extracted from the gel using Qiagen Gel Extraction Kit. The yield of undigested DNA and purified mononucleosomal DNA was measured using Quant-iT™ PicoGreen™ dsDNA Assay Kit (Invitrogen, Carlsbad, CA, USA) in a Qubit™ 2.0 Fluorometer. DNA amounts obtained from various samples were normalized to make them equal. Equal amount of DNA was then used as template in qPCR in real time using the primers for α-globin, β-globin and pancreatic amylase loci. For each primer, percentage protection was calculated by comparing Ct values of undigested DNA versus mononucleosomal DNA for DMSO or 898 treatments and further normalized to Ct values of undigested DNA versus mononucleosomal DNA obtained from the serum starved cells.
For measuring BrdU incorporation, NIH3T3 cells in S-phase were treated with DMSO or 3 μM Hdacs1,2-selective inhibitor (898) for 12 h, 18 h or 24 h followed by labeling of cells with 20 μM BrdU for 1 hr. Genomic DNA was isolated using Qiagen DNeasy™ Kit and extensively digested with RNaseA. DNA was quantified in a Qubit™ 2.0 Fluorometer using Quant-iT™ PicoGreen™ dsDNA Assay Kit (Invitrogen). Equal amount DNA (500 ng) from the different samples was resuspended in 40 μl water and denatured by adding 10 volumes of 0.4 N NaOH and incubation at room temperature for 30 min. Equal volume of 1 M Tris. Cl, pH 6.8 was added to neutralize and samples were placed on ice. Aliquots were made to obtain varying amounts of DNA (50, 25 and 12.5 ng) for each sample in a total volume of 100 μl. DNA was then transferred on to a Zeta-Probe GT Membrane (Bio-Rad, Hercules, CA, USA) using a Slot Blot apparatus (Schleicher and Schuell Minifold II). DNA was immobilized to the membrane using a UV crosslinker and western blotting was done with anti-BrdU antibody. The linear range for detecting BrdU-labeled DNA was initially determined using DNA isolated from the DMSO control. Using a two-fold serial dilution of DNA, we determined that 50 ng to 6.25 ng to be in the linear range of detection in western blotting using the anti-BrdU antibody (1:500 dilution) and ECL2 Western Blotting Substrate (Thermo Scientific Pierce, Waltham, MA, USA).
For the analysis of BrdU-labeled DNA obtained from ChIP assays, input and ChIP DNA were eluted in 50 μl water. The input DNA yield was measured using NanoDrop 1000 spectrophotometer (ThermoScientific) to ensure that the DNA amount to be used in Slot Blot assay remains in the linear range of detection (<50 ng). The input and immunoprecipitated DNA (50 μl) were denatured by adding 2.5 volumes of 0.4 N NaOH and incubating at room temperature for 30 min. Equal volume of 1 M Tris–HCl, pH 6.8 (175 μl) was then added to neutralize and samples were placed on ice. A serial dilution of DNA from input and immunoprecipitated DNA for the Slot Blot assay were prepared as follows: For input DNA, 100 μl denatured DNA; 50 μl denatured DNA + 50 μl H2O; 25 μl denatured DNA + 75 μl H2O. For ChIP DNA, 200 μl denatured DNA; 100 μl denatured DNA + 100 μl H2O; 50 μl denatured DNA + 150 μl H2O. The DNA were then transferred on to a Zeta-Probe GT Membrane (Bio-Rad, Hercules, CA, USA) using a Slot Blot apparatus (Schleicher and Schuell Minifold II) and processed as described above. For the modified ChIP assays with Brdu pulse-chase, NIH3T3 or HeLa cells were labeled with 20 μM BrdU for 30 min. Following BrdU-labeling, cells were washed twice with PBS to remove unincorporated BrdU and grown in fresh medium. For the chase, cells were fixed with formaldehyde for ChIP analyses at 15 min, 30 min and 60 min time points.
Total RNA was isolated from NIH3T3 cells that were released into S-phase from serum starvation for 20 h either in the presence of DMSO or 3 μM 898 using the Versagene RNA isolation kit (5 Prime). RNA was prepared from three different sets of DMSO- and 898-treated cells and sequenced using the Illumina Hiseq2000 sequencer. Standard gene analysis was performed using the open source USeq/DESeq analysis packages. In brief, this involves aligning each replica dataset to a genome index that has the standard mm10 chromosomes plus an artificial chromosome containing all known and all theoretical splice junctions. After alignment splice junction coordinates are converted to genomic coordinates, counts for each gene was collected, and a multi-replica treatment versus control comparison was performed using DESeq (http://genomebiology.com/2010/11/10/R106). Genes passing two thresholds, an FDR of <10% and absolute log2 ratio of 1, were considered differentially expressed and used in subsequent analysis. A window scanning, no known annotation analysis was also performed using the USeq MultipleReplicaScanSeqs application. This analysis generated the genome wide log2 ratio and FDR window summary tracks.
For preparation of whole cell extracts, cell pellets were washed with PBS and sonicated in RIPA buffer with protease inhibitors (Roche protease inhibitor cocktail) prior to western analyses. Antibodies used in this study are listed in the supplementary table (Additional file 23: Table S2).
Ad-Cre infection of fibrosarcoma cells was performed as described previously .
Adenovirus-containing Cre recombinase
Arbitrary fluorescence units
Chromatin assembly factor-1
Fluorescence-activated cell sorting
2: Histone deacetylase 1 and 2
Heterochromatin protein 1
Proliferating cell nuclear antigen
Polymerase chain reaction
Replication protein A
Suberoylanilide hydroxamic acid
SWI/SNF-related matrix associated actin-dependent regulator of chromatin subfamily A member 5
TATA binding protein
We thank David Jones and Maria McDowell for their comments on the manuscript. We thank Hannah Wellman, Jean Ansolabehere, Brandt Jones and Saravanan Ramakrishnan for their technical assistance. We thank Cedric Clapier and Takuya Abe for their suggestions. We thank David Nix and Brian Dalley at the HCI genomics core for assistance with the RNA-seq analysis. We are grateful to Kim Boucher for the statistical analyses. We thank Jun Qin and Beom-Jun Kim for the Smc3 cell lines and Katsuhiko Shirahige for the Smc3ac antibody. We specially thank Scott Hiebert for his kind encouragement and support of this work, and comments on this manuscript. This work was supported by the Huntsman Cancer Institute and Radiation Oncology Department Development Funds provided to SB.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.