|Home | About | Journals | Submit | Contact Us | Français|
Histone H3 serine 10 phosphorylation is a hallmark of mitotic chromosomes but its full function remains to be elucidated. We report here that two SR protein splicing factors, SRp20 and ASF/SF2, associate with interphase chromatin, are released from hyperphosphorylated mitotic chromosomes but reassociate with chromatin late in M-phase. Inhibition of Aurora B kinase diminished histone H3 serine 10 phosphorylation and increased SRp20 and ASF/SF2 retention on mitotic chromosomes. Unexpectedly, we also found that HP1 proteins interact with ASF/SF2 in mitotic cells. Strikingly, siRNA-mediated knockdown of ASF/SF2 caused retention of HP1 proteins on mitotic chromatin. Finally, ASF/SF2-depleted cells released from a mitotic block displayed delayed G0/G1 entry, suggesting a functional consequence of these interactions. These findings underscore the evolving role of histone H3 phosphorylation and demonstrate a direct, functional and histone modification-regulated association of SRp20 and ASF/SF2 with chromatin.
Histones play a central role in modulating protein assembly on and release from the chromatin fiber, thereby regulating numerous chromatin-based processes; including transcription, replication, condensation, repair and segregation (Grewal and Moazed, 2003; Groth et al., 2007; Kouzarides, 2007; Li et al., 2007). The dynamic interaction of proteins on chromatin is regulated primarily by post-translational modifications of the N-terminal tails of each of the four core histones (Grunstein, 1998; Kouzarides, 2007; Luger and Richmond, 1998; Strahl and Allis, 2000). The diversity of histone modifications, including acetylation, methylation, phosphorylation, ADP-ribosylation and ubiquitination, has led to the “histone code” hypothesis and a chromatin signaling network model in which each modification, alone or in combination, serves a distinct function (Schreiber and Bernstein, 2002; Strahl and Allis, 2000). In essence, these modifications direct recruitment or release of proteins that alter chromatin structure and/or regulate DNA-based processes (Kutney et al., 2004). For example, the bromodomain of human p300/CBP- associated factor binds acetylated lysine residues within the tail of histone H3 and H4, and is associated with transcriptionally active genes (Dhalluin et al., 1999). In contrast, chromodomain protein heterochromatin protein 1 (HP1) binding to methylated lysine 9 of histone H3 contributes to the maintenance of large heterochromatic chromosomal domains and transcriptionally repressed genes (Bannister et al., 2001; Lachner et al., 2001; Maison and Almouzni, 2004).
Histone H3 serine 10 phosphorylation (H3S10P) plays a critical role in chromosome condensation and chromosome segregation during mitosis and meiosis (Mellone et al., 2003; Nowak and Corces, 2004; Wei et al., 1998; Wei et al., 1999). During M-phase, when H3S10P is high, most HP1 is released from chromatin despite persistence of histone H3 lysine 9 trimethylation. Thus, a methyl-phospho switch is believed to regulate HP1 dissociation from mitotic chromosomes (Fischle et al., 2005; Hirota et al., 2005). Results from another study suggest that phospho-acetylation (S10P-K14Ac) of histone H3 is involved in release of HP1 from chromatin (Mateescu et al., 2004). While these results point to a critical role for H3S10P in chromosome dynamics during mitosis, the full function of this particular histone H3 modification remains to be elucidated.
Members of the SR protein family are known to be important for splicing of mRNA precursors (Fu, 1993; Graveley, 2000; Manley and Tacke, 1996). These proteins are modular and comprised of one or more RNA recognition motifs (RRMs) and an arginine-serine rich (RS) domain. The RRM recognizes specific target sequences within the pre-mRNA, while the RS domain, which contains multiple phosphorylated serine residues, functions by modulating protein-protein, and perhaps protein-RNA interactions (Fu, 1993; Shen and Green, 2004). SR proteins also participate in other cellular processes (Huang and Steitz, 2005). For example, SR proteins, ASF/SF2, and SC35 have been shown to be critical for maintenance of genome stability and cell cycle progression (Li and Manley, 2005; Li et al., 2005; Xiao et al., 2007). Genetic depletion of ASF/SF2 led to formation of transcriptional R-loops and DNA rearrangements and caused G2/M cell cycle arrest, suggesting that SR proteins may have a role in chromatin dynamics and function. Similarly, in mouse embryonic fibroblasts loss of SC35 resulted in G2/M cell cycle arrest and genomic instability.
To gain more insight into the identity and function of chromatin-associated proteins and the emerging roles of histone H3S10P, we utilized a protein affinity purification scheme. Unexpectedly, we found that two SR proteins, SRp20 and ASF/SF2, bind to histone tails and their chromatin binding is regulated by histone H3S10P. We provide several lines of in vitro and in vivo evidence demonstrating that these SR proteins associate with interphase and late/post mitotic chromatin but are dissociated from mitotic chromatin. Strikingly, siRNA-mediated knockdown of ASF/SF2 led to retention of HP1 proteins on mitotic chromatin and caused a delay in G0/G1 entry of cells. We propose that these SR proteins have a novel “chromatin-sensor” activity, which together with HP1, may be necessary for proper cell-cycle progression.
To identify novel modification-sensitive histone interacting proteins, we incubated HeLa cell nuclear extracts with differentially modified biotinylated histone peptides and analyzed bound proteins by SDS-PAGE and silver stain followed by mass spectrometry of excised bands. The biotinylated histone tail peptides used in the initial experiments were histone H3 tail peptides; methylated at lysine 9 (H3-K9Me) and phosphorylated at serine 10 (H3-S10P). Mass spectrometric analysis of selected isolated bands that bound to H3-K9Me but not H3-S10P were identified as HP1α, -β, -γ and SRp20 (Figure 1A). The role of H3-K9Me and H3-S10P in binding and release of HP1 proteins from chromatin has been reported before (Bannister et al., 2001; Fischle et al., 2005; Hirota et al., 2005; Lachner et al., 2001; Maison and Almouzni, 2004). The unexpected binding of the SR protein splicing factor to this histone H3 tail prompted us to further characterize the SRp20-histone H3 interaction in vitro. HeLa cell nuclear extracts were incubated with selected biotinylated histone H3 tail peptides and bound proteins were analyzed by SDS PAGE followed by immunoblot with anti-SRp20 antibody. Endogenous SRp20 bound to unmodified histone H3, H3-K9Ac, H3-K14Ac and H3-K9Me peptides. However, phosphorylation of serine 10 significantly reduced SRp20 association with histone H3 peptide tails (Figure 1B).
We next investigated whether another SR family member, ASF/SF2, also binds histones in a modification-selective manner. To this end, HeLa cell nuclear extracts were incubated with biotinylated histone H3 tail peptides. The histone H3 peptide tails were pre-bound to streptavidin agarose beads and the complexes were analyzed by immunoblotting with anti-ASF/SF2 antibody. Endogenous ASF/SF2 indeed bound strongly to the unmodified, methylated and acetylated histone H3 tail-peptides, while no detectable binding between ASF/SF2 and either H3-S10P or H3-S10P/K14Ac peptides was observed (Figure 1C). As a control, neither SRp20 (data not shown) nor ASF/SF2 bound to histone H2B tails, further demonstrating the specificity of the histone H3 interaction (Figure 1C).
Since SRp20 bound to the histone H3 tail in vitro, we wanted to identify the domain of SRp20 necessary for this interaction. SRp20 consists of one RRM and a carboxyl-terminal RS domain (Bourgeois et al., 2004). We expressed and purified these domains as GST fusion-proteins and analyzed their interaction with acid-extracted histones isolated from asynchronously growing HeLa cells. Under conditions where approximately equal amounts of fusion proteins were used, the GST-RS domain but not the GST-RRM domain associated with histone H3 (Figure 1D), suggesting that the RS domain of SRp20 is sufficient for the histone H3 interaction.
To confirm the specificity of SRp20 binding to histone H3, GST pull-down assays were performed with recombinant SRp20 and acid-extracted histones isolated from asynchronously growing HeLa cells. Bound proteins were analyzed by immunoblot using anti-histone H3 and anti-phosphoserine 10 histone H3 antibodies. Unlike the endogenous protein, bacterially expressed GST-SRp20 bound both unmodified histone H3 and histone H3 phosphorylated at serine 10 (Figure 2A). SR proteins are known to be phosphorylated at multiple sites in vivo (Bourgeois et al., 2004; Ngo et al., 2005). Therefore, we investigated whether the phosphorylation status of SRp20 accounted for the discrepancy in histone binding between endogenous and recombinant SRp20. To test this hypothesis, we treated HeLa cell nuclear extracts or recombinant GST-SRp20 with calf intestinal phosphatase (CIP). Immunoblot analysis revealed a change in the SDS-PAGE mobility of cellular SRp20, but not of GST-SRp20, suggesting that only endogenous SRp20 was phosphorylated (Figure 2B). We then asked whether in vitro phosphorylated GST-SRp20 fails to associate with histone H3-S10P tails. GST-SRp20 was phosphorylated by pre-incubation with a HeLa cell S100 fraction (a major source of the SR protein kinase 1, SRPK1, activity) and ATP (Figure 2C) (Gui et al., 1994b). Following incubation, the beads were washed to remove additional proteins and ATP prior to addition of acid-extracted histones. In vitro phosphorylated GST-SRp20 failed to associate with serine 10 phosphorylated histone H3 (Figure 2D, lane 4). In contrast, unphosphorylated GST-SRp20 bound to serine 10 phosphorylated histone H3 (Figure 2D, lane 3).
The above data suggest that the phosphorylation state of ASF/SF2 and SRp20 is an important determinant in their histone H3 binding capabilities. To test this hypothesis, we mock- or CIP-treated endogenous SRp20 and ASF/SF2 and determined their ability to associate with a biotinylated histone H3-S10P peptide tail. A mobility shift was observed for SRp20 and ASF/SF2 treated with CIP, suggesting that endogenous SRp20 and ASF/SF2 were, as expected, phosphorylated (Figure 2E, input). Equal protein amounts were added in both −CIP and +CIP inputs as determined by protein quantification, despite the discrepancy in immunoblot signal (Figure 2E, input). Consistent with the above data with purified recombinant proteins, CIP-treated endogenous SRp20 and ASF/SF2 associated with histone H3-S10P peptide, whereas their phosphorylated counterparts did not despite the fact that 1 mg of nuclear extract was used in the binding experiment (Figure 2E; see figure legend for details). These results show that phosphorylated SRp20 and ASF/SF2 fail to bind to histone H3-S10P tails significantly, and that association requires dephosphorylation of the histone H3 tail. It remains to be determined whether a single or multiple site phosphorylation event influences the binding properties of SRp20 and ASF/SF2.
To confirm the in vitro peptide binding data, we next examined whether SRp20 and ASF/SF2 interact with nucleosomes in intact cells. Previous findings indicate that in mitotic cells, global levels of histone H3 acetylation and methylation do not change, whereas H3 phosphorylation significantly increases (Fischle et al., 2005). To increase histone H3S10P, cells were treated with 50 ng/mL nocodazole for 12h. Chromatin fractions were prepared by lysing cells in hypotonic buffer. Nuclei from cells grown under normal conditions or treated with nocodazole were isolated by low speed centrifugation and subjected to micrococcal nuclease (MNase) digestion to prepare nucleosomes for immunoprecipitation with anti-SRp20 or anti-ASF/SF2 antibodies. Nocodazole treatment did not significantly alter SRp20 and ASF/SF2 levels and both could be specifically immunoprecipitated with their respective antibodies from control and nocodazole treated cells (Figure 3A). Nucleosomes with acetylated, methylated and phosphorylated histone H3 are present in nocodazole-treated cells (Figures 3B, input). Endogenous SRp20 and ASF/SF2 co-immunoprecipitated with nucleosomes containing unmodified (Figure 3B, first panel), acetylated at lysine 9 (Figure 3B, third panel) and methylated at lysine 9 (Figure 3B, fourth panel) histone H3. Importantly, we failed to detect any measurable interaction of SRp20 or ASF/SF2 with nucleosomes containing histone H3S10P (Figure 3B, second panel). Additionally, SRp20 co-immunoprecipitated with unmodified histone H3 in control cells (Figure 3B, first panel, left).
To support our immunoprecipitation results suggesting that SRp20 and ASF/SF2 associate with endogenous histones, we performed colocalization studies of SRp20 and ASF/SF2 with histone H3 in asynchronously growing HeLa cells using immunocytochemistry. As expected, the level of H3S10P was low to undetectable in interphase cells (data not shown). While the SR proteins stained diffusely throughout the nucleus with characteristic nuclear speckled patterns (Caceres et al., 1997; Lamond and Spector, 2003), we found that a significant amount of SRp20 and ASF/SF2 colocalized with histone H3 in interphase cells (Figure 3C top panel, d-d′ and bottom panel, i-i′ respectively). Similarly, SR protein/DNA merged images show that a significant amount of endogenous SRp20 and ASF/SF2 associated with chromosomes in interphase cells (Figure 3C, e-e′ and j-j′). Together, these results strongly suggest association of SRp20 and ASF/SF2 with interphase chromatin and H3S10P-sensitive interaction of SRp20 and ASF/SF2 with nucleosomes/chromatin.
In light of our observations that SRp20 and ASF/SF2 associated with interphase chromatin but failed to interact with histone H3 phosphorylated at serine 10 in vitro and with hyperphosphorylated histones/nucleosomes isolated from cultured cells, we examined association of the SR proteins with chromatin as a consequence of histone phosphorylation as cells progress through mitosis. HeLa cells were synchronized by a double thymidine block and mitotically arrested with a short nocodazole treatment 9h after release from the second thymidine block to enhance the number of cells in mitosis. Mitotic cells were then harvested, fixed with paraformaldehyde, permeabilized with methanol and probed with anti-ASF/SF2 and anti-histone H3-S10P antibodies followed by Hoechst staining for DNA (Figure 4B). Consistent with previous reports mitotic cells showed extensive histone H3 hyperphosphorylation (Figure 4B, a) (Hendzel et al., 1997; McManus and Hendzel, 2006). In contrast to its extensive association with interphase chromatin (Figure 3C), a significant reduction in ASF/SF2 association with mitotic chromosomes (Figure 4B, e-e′) became evident. Similar results were also obtained with nocodazole-arrested mitotic cells (data not shown) and with SRp20 (Figure 5B, e-e′). This finding suggests that H3S10P is an important determinant for release of SR proteins from mitotic chromosomes. Residual ASF/SF2 and SRp20 association with mitotic chromatin (Figure 4B, e′ and Figure 5B, e′, respectively) might represent their interaction with unmodified, acetylated or methylated nucleosomal domains (Figure 3B).
We next wished to determine if H3S10P is indeed a critical regulator of SR protein-chromatin interactions and sufficient to cause dissociation of the SR proteins from mitotic chromatin. To investigate this, HeLa cells were synchronized by a double thymidine block and released for 9h, followed by a short nocodazole treatment to enhance the number of cells in mitosis. We then performed a time course analysis of the disappearance of H3S10P and reassociation of SRp20 and ASF/SF2 following release from the M-phase block. Cells were harvested every 30 min after release, fixed with paraformaldehyde, permeabilized with methanol and probed with anti-ASF/SF2 (Figure 4A) or anti-SRp20 (Figure 5A) and anti-histone H3-S10P antibodies to analyze SR protein association/dissociation from chromatin in relation to the presence of H3S10P. Cells in prometaphase, metaphase and anaphase displayed high levels of H3S10P (Figure 4A, a, g, m and Figure 5A, a, g, m) and largely showed ASF/SF2 and SRp20 to be dissociated (excluded) from mitotic chromatin (Figure 4A, f, l, r and Figure 5A, f, l, r respectively). As cells progressed through telophase, a late stage of mitosis, the levels of H3S10P decreased (Figure 4A, s and Figure 5A, s) and was accompanied by a dramatic reassociation of both ASF/SF2 and SRp20 with chromatin (Figure 4A, x and Figure 5A, x respectively), even before chromatin decondensation was complete. These data suggest that SR protein dissociation from chromatin during M-phase and their reassociation during/following telophase is dependent on histone H3S10P.
To confirm that histone H3S10P is indeed required for release of SRp20 and ASF/SF2 from mitotic chromosomes, we next analyzed association of the SR proteins with mitotic chromatin when H3S10P was inhibited. It has been established that H3 serine 10 is phosphorylated by Aurora B kinase in mitotic cells (Giet and Glover, 2001; Hsu et al., 2000). Mitotic cells were obtained following a double thymidine block, a 10h release and 1h treatment with an Aurora B kinase inhibitor, ZM447439 (Ditchfield et al., 2003; Ditchfield et al., 2005). As expected, the Aurora B kinase inhibitor significantly reduced H3S10P on condensed mitotic chromatin (Figure 4B, f; f;4C4C middle panel and Figure 5B, f; f;5C5C middle panel). Importantly, we observed a concomitant increase in the retention of ASF/SF2 on the condensed mitotic chromatin when H3S10P was blocked (Figure 4B, compare e′ with j′; 4C top panel). As with ASF/SF2, we found that SRp20 associated with interphase chromatin (Figure 3C, upper panel) and was largely dissociated from mitotic chromosomes (Figure 5A). In the presence of the Aurora B kinase inhibitor, however, an increase in retention of SRp20 on condensed mitotic chromatin was observed (Figure 5B, compare e′ with j′; 5C top panel).
We next wished to corroborate our immunocytochemistry results (Figures 4B and and5B)5B) by biochemical fractionation and analysis. To this end, we collected mitotic cells following double thymidine block, a 9h release and 2h treatment with nocodazole and ZM447439. These cells were either fractionated to obtain a chromatin-enriched fraction as previously described (Fischle et al., 2005; Mendez and Stillman, 2000)(Figures 4C and and5C)5C) or lysed to obtain whole cell lysates (Figures 4D and and5D).5D). In both cases, lysates were analyzed for ASF/SF2 (Figure 4C–D, top panels), SRp20 (Figure 5C–D, top panels), H3S10P (Figures 4C–D, 5C–D, second panels) and loading control, either histone H1 (Figures 4C and and5C,5C, bottom panels) or actin (Figures 4D and and5D,5D, bottom panels) using the appropriate antibodies. Total levels of ASF/SF2 (Figure 4D, top panel) or SRp20 (Figure 5D, top panel) in whole cell extracts were unchanged following inhibition of histone H3S10P by Aurora B kinase inhibitor treatment. As expected, cells treated with the Aurora B kinase inhibitor significantly decreased histone H3S10P (Figures 4C–D and 5C–D, second panels). Importantly, consistent with our immunocytochemistry data (Figures 4B and and5B),5B), increased retention of both ASF/SF2 (Figure 4C, top panel) and SRp20 (Figure 5C, top panel) on mitotic chromatin in the absence of H3S10P was evident. As a positive control, and in agreement with the results of Fischle et al. (2005), we also found that HP1β was retained on mitotic chromatin in the absence of H3S10P (Fischle et al., 2005) (Figure 4C, third panel). Together with the results of Figure 3, these immunocytochemical and biochemical analyses demonstrate association of SRp20 and ASF/SF2 with interphase chromatin, their dissociation from mitotic chromatin and reassociation with late/post mitotic chromatin in a H3S10P-dependent manner.
Our data so far have shown that SRp20 and ASF/SF2 associate with interphase chromatin; are largely dissociated from mitotic chromatin in a H3S10P-dependent manner; and subsequently reassociate with chromatin late in M-phase when H3S10P diminishes. These observations raise the interesting question as to the function of SRp20 and ASF/SF2 dissociation from chromatin during M-phase and reassociation with post-mitotic chromatin. To address this issue, we considered the possibility that ASF/SF2 may be necessary to remove other chromatin-associated protein(s) to facilitate M-phase progression. HP1 proteins have been found to be recruited to discrete regions of the genome by histone H3 lysine 9 trimethylation where they are involved in controlling chromatin structure and dynamics – predominantly regulation of gene expression and heterochromatin formation (Fischle et al., 2005; Hirota et al., 2005). HP1 dissociation from chromatin is dependent on H3S10P, and treatment with the Aurora B kinase inhibitor leads to retention of HP1 proteins on mitotic chromatin (Fischle et al., 2005), similar to what we observed with SRp20 and ASF/SF2.
Since ASF/SF2 and HP1 proteins are all significantly released from hyperphosphorylated mitotic chromatin, we reasoned that they may interact with each other and that such an interaction may be necessary for their release. To determine if ASF/SF2 and HP1β interact, we analyzed the association between ASF/SF2 and HP1β in control siRNA treated mitotic cells by immunocytochemistry. Although not extensive, we nonetheless observed colocalization of ASF/SF2 and HP1β in control siRNA treated mitotic cells in which both proteins were primarily excluded from chromatin (Figure 6A, e-e′ and Figure 6D where the same cell is shown with different overlapping images). To provide direct support for an interaction, we made mitotic soluble protein extracts, immunoprecipitated with either control IgG or anti-ASF/SF2 antibody and analyzed bound proteins by immunoblotting for HP1β. In agreement with the results of Figure 6A, we found that a fraction of ASF/SF2 and HP1β (Figure 6B) associated in mitotic cell extracts. In the future, it will be important to determine whether binding of SRp20 and ASF/SF2 to histone H3K9Me plays any functional role in the above mentioned process.
To address the possibility that the ASF/SF2-HP1β association in mitotic cells is important for HP1 release from mitotic chromatin, we determined the effect of ASF/SF2 knockdown on HP1 protein association with mitotic chromatin. We transfected HeLa cells with either control or ASF/SF2 siRNAs, fixed cells with paraformaldehyde, permeabilized with methanol and analyzed HP1β association by immunocytochemistry (Figure 6D). ASF/SF2 protein levels were significantly diminished after 48h of RNAi treatment (Figure 6C). Consistent with previous findings, we observed HP1β remained nuclear but was predominantly excluded from chromatin in control siRNA treated mitotic cells (Figure 6D, a, e-e′ and (Fischle et al., 2005)). Importantly, more HP1β (Figure 6D, compare e′ with j′) was retained on mitotic chromatin in ASF/SF2-knockdown cells despite the presence of H3S10P (Supplemental Figure 1).
We next wished to confirm the immunocytochemistry results with biochemical fractionation and analysis. To this end, we collected mitotic cells after ASF/SF2 knockdown and nocodazole treatment, and they were either fractionated to obtain a chromatin-enriched fraction as previously described (Fischle et al., 2005; Mendez and Stillman, 2000) (Figure 6E) or lysed to obtain whole cell lysates (Figure 6F). Lysates were analyzed for HP1β, ASF/SF2 and a loading control, either histone H1 or actin using appropriate antibodies. Our findings demonstrate that increased amounts of HP1β were indeed retained on mitotic chromatin upon ASF/SF2 knockdown (Figure 6E top panel) despite no significant change in the levels of HP1β (Figure 6F top panel). Additional immunoprecipitation, immunocytochemical and biochemical fractionation analyses were conducted for HP1α and HPγ in the ASF/SF2 knockdown cells and similar results were obtained (Supplemental Figure 2). Collectively, our findings strongly suggest that along with histone H3 phosphorylation, ASF/SF2 also plays a potentially important role in mediating HP1 protein dissociation from mitotic chromatin.
Since SRp20 and ASF/SF2 associated with interphase and post-mitotic, but not mitotic, chromatin, we next analyzed whether ASF/SF2 plays a role in affecting cell cycle progression. We tested this possibility by utilizing DT40-ASF cells, which are a derivative of chicken DT40 cells where the only copy of the ASF/SF2 gene is a human cDNA under the control of a tetracycline (tet)-repressible promoter (Wang et al., 1996). DT40-ASF cells were treated with 1 μg/mL doxycycline for 16h before adding 100 ng/mL nocodazole for 8h to prevent cell cycle progression past prometaphase (Figure 7A). ASF/SF2 protein expression levels were significantly diminished after 24h of doxycycline treatment, which represents hour 0 of the schematic shown in Figure 7A (Figure 7C inset). After the nocodazole treatment, we released the cells and analyzed the effect of ASF/SF2 depletion on cell cycle progression via flow cytometry using propidium iodide to stain the DNA. Cell cycle analysis indicated that ASF/SF2 depletion inhibited G2/M progression (Figure 7B) and subsequently delayed entry of cells into G0/G1 (Figure 7C). Subsequently, as expected, and consistent with our overall hypothesis inhibition of histone H3S10P by treatment with the Aurora B kinase inhibitor essentially prevented all cells from entry into G0/G1 in both ASF/SF2 depleted and normal cells (Supplemental Figure 3). Although our experimental design does not directly demonstrate a causal relationship between ASF/SF2 chromatin reassociation at the end of telophase and cell entry into G0/G1 phase, our findings together demonstrate that depletion of endogenous ASF/SF2 alters retention of HP1 proteins on mitotic chromatin and prevents G0/G1 entry after release from an M-phase block.
This study has provided the first demonstration that two members of the human SR protein family, SRp20 and ASF/SF2, associate with interphase chromatin, dissociate from mitotic chromosomes following histone H3S10P, and reassociate with post-mitotic chromatin. In the future it will be important to determine whether this is a common property of all members of the SR family of proteins. It appears that the release of SR proteins from chromatin is determined by SR protein phosphorylation by SRPKs and histone H3 serine 10 phosphorylation by Aurora B kinase. Consistent with this view, SRPK1 has been shown to translocate to the nucleus at the onset of M-phase, and some SR proteins have been shown to be hyperphosphorylated in M-phase (Ding et al., 2006; Gui et al., 1994a). We propose that a dual phosphorylation event involving SRPK-mediated phosphorylation of SR proteins and Aurora B kinase mediated H3 phosphorylation probably regulates the dynamics of SR protein interactions with chromatin. Although primarily nuclear, shuttling of SRp20 and ASF/SF2 to the cytoplasm when transcription is blocked or Clk/Sty kinase is exogenously expressed, has been observed (Caceres et al., 1998). It remains to be seen whether these events are linked and thus establish a regulatory circuit.
Another interesting feature of the SR protein interaction with chromatin is that while a dual modification of histone H3 appears necessary for HP1 release (Fischle et al., 2005; Hirota et al., 2005; Mateescu et al., 2004), H3S10P alone is sufficient for release of SRp20 and ASF/SF2 from mitotic chromatin. While the interaction of SRp20 and ASF/SF2 with chromatin might appear unexpected, previous work established that inactivation of ASF/SF2 causes transcription-related and cell-cycle progression defects, suggesting a role for ASF/SF2 in modulation of chromosome dynamics. Li et al. (2005) suggested that a mechanism other than an alteration in mRNA splicing, specifically the generation of DNA double strand breaks resulting from cotranscriptional R-loop formation (Li and Manley, 2005; Li et al., 2005), could be responsible for the G2/M transition defects in cells depleted of ASF/SF2. Similarly, roles for another splicing factor, SC35, beyond its function in RNA processing have recently been reported in chromatin based processes: SC35 regulates transcription elongation of specific genes and also plays a critical role in regulating genomic stability and cell cycle progression (Lin et al., 2008; Xiao et al., 2007). Based on these and our current findings, we propose that chromatin association/dissociation properties of the SR proteins also contributes to proper cell cycle progression, and that they may contribute to the G2/M defect, as well as explain the delayed G0/G1 entry we observed upon ASF/SF2 depletion.
SRp20 and ASF/SF2’s chromatin association/dissociation properties provide insight into how these proteins may work to regulate proper chromatin function. We hypothesize that the release of SR proteins from hyperphosphorylated chromatin and the subsequent reassociation of SR proteins with chromatin once histone H3S10P has diminished are both important events for proper chromatin function. siRNA-mediated knockdown of ASF/SF2 caused retention of HP1 proteins on histone H3 despite serine 10 phosphorylation, suggesting that dissociation of ASF/SF2 from hyperphosphorylated histone H3 also influences HP1 dissociation from mitotic chromatin. Indeed, Fischle et al. (2005) suggested that in addition to H3S10P, other mechanisms may be involved in mitotic release of HP1 from chromatin, such as further modification of HP1 proteins and/or their interaction partners (Fischle et al., 2005). Our findings implicate ASF/SF2 as one such interacting partner responsible for HP1 release from mitotic chromatin. HP1 and ASF/SF2 are both associated with interphase chromatin, are dissociated from chromatin upon H3S10P and are associated with each other in mitotic cells. Due to the inherent limitations of mammalian cells for mutational analysis, neither this study nor previous ones (Fischle et al., 2005; Hirota et al., 2005) provided a direct causal relationship linking the chromatin association/dissociation properties of SRp20, ASF/SF2 and HP1 proteins with M-phase progression and chromosome segregation. However, M-phase progression and chromosome segregation defects have been noted in S. pombe and Tetrahymena mutants defective in H3S10P (Mellone et al., 2003; Wei et al., 1999). Additionally in S. pombe, the HP1 homolog, Swi6, is required for proper chromatin segregation (Pidoux and Allshire, 2004).
Our studies in conjunction with the above results provide a strong correlation between SR and HP1 proteins and their release from chromatin with proper chromosome segregation and M-phase progression. It is, therefore, reasonable to suggest that the coordinated removal of HP1 and ASF/SF2 during M-phase is probably necessary to allow access by factors required for mediating proper chromatin condensation and faithful chromosome segregation. In this context we note that although 14-3-3 proteins were shown to bind H3S10P, they did not significantly associate with hyperphosphorylated condensed chromosomes in mitotic cells. Since H3S10P is also implicated in transcriptional activation, it appears that 14-3-3 and H3S10P association may be restricted to transcription of inducible genes (Macdonald et al., 2005).
We also found that removal of an M-phase block in cells depleted of ASF/SF2 significantly delayed both exit from G2/M and entry into G0/G1. There are several possibilities for why ASF/SF2 depleted cells, blocked in M-phase, would delay G0/G1 entry. The first is that ASF/SF2 depletion blocks the cells at G2, as suggested by Li et al. (Li et al., 2005). These authors demonstrated that DT40-ASF cells depleted of ASF/SF2 induced a G2-block and an increase in apoptotic cell death by 72h post-tet treatment. However, it is important to note that our experimental time frame was significantly shorter (28.5h) than theirs and we did not see increased apoptotic cell death as indicated by the absence of a sub-G0/G1 peak in all time points analyzed within the experimental time frame (28.5h). This, by no means, eliminates the possibility that part of the G2/M increase we see in our cell cycle analysis is due to a G2 block and not an M-phase block. The second, more intriguing possibility for the delay of G0/G1 entry upon ASF/SF2 depletion and release from an M-phase block is because ASF/SF2 is not present to reassociate with chromatin in telophase, once H3S10P has diminished. ASF/SF2 reassociation with chromatin at the end of mitosis may be a trigger for M-phase completion and it may allow additional factors, such as HP1 proteins, to reassociate with chromatin as the cell enters G0/G1. We propose that inhibition of histone H3S10P in ZM-treated cells, ASF/SF2 and SRp20, as well as HP1 proteins, are prevented from dissociating from mitotic chromatin and thus exacerbating the effect we see with ASF/SF2 depletion alone.
In summary, this work has established a novel association of two SR proteins, SRp20 and ASF/SF2, with interphase chromosomes; demonstrates their release from hyperphosphorylated mitotic chromosomes and reassociation with post-mitotic chromatin; and provides insight into our evolving understanding of the function of H3S10P. The ability of the SR proteins to associate with and dissociate from chromosomes in a histone H3 modification-selective manner may account for G2/M cell cycle arrest and defects in G0/G1 accumulation. Significantly, the SR proteins have similar association/dissociation characteristics as HP1 proteins. ASF/SF2 and HP1 proteins associate in mitotic cells and knockdown of ASF/SF2 led to retention of HP1 proteins on mitotic chromatin, suggesting a mechanistic link between the release of ASF/SF2 and HP1 proteins from mitotic chromatin. Future studies will investigate the underlying mechanisms of SR protein association/dissociation from chromatin and the potential role of SR proteins in directly regulating chromatin function and cell cycle progression.
Expression constructs were comprised of PCR-amplified fragments of SRp20 and its derivatives cloned into pGEX-4T1 (CLONTECH) vectors.
HeLa cells were grown in monolayers in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal calf serum, penicillin/streptomycin and L-glutamine and incubated in a 5% CO2/95% humidified atmosphere at 37°C. To induce M-phase arrest, 50 ng/mL nocodazole was used for 12h, where indicated. Synchronization of HeLa cells was performed using a double thymidine block. Cells were released from the second block by washing and incubating in fresh culture media. After approximately 10 h, as the cells neared mitosis, they were treated with 2 μM ZM447439 (an Aurora B kinase inhibitor, Tocris Bioscience, (Ditchfield et al., 2003; Ditchfield et al., 2005)) or DMSO control. After 1h treatment, mitotic cells were shaken off the plates, washed with 1xPBS and spun onto coated cytospin slides using Cytospin 2 (Shandon) at 1000 rpm for 4 min. For analysis of cell cycle progression after mitotic block, HeLa cells were synchronized using double thymidine block, 8h after the second release, 50 ng/mL nocodazole was added for 2.5h before cells were released from the mitotic block and analyzed by immunocytostaining. Immunocytostaining was performed using the method as described under Immunocytochemistry.
DT40-ASF chicken cells are cultured at 1x105 cells/mL in RPMI medium supplemented with 10% fetal calf serum, 1% chicken serum, penicillin/streptomycin with anti-mycotics and L-glutamine and incubated in a 5% CO2/95% humidified atmosphere at 37°C (Li et al., 2005; Wang et al., 1996). ASF depletion was induced using 1 μg/mL doxycycline for indicated times. To block cells in mitosis, they were cultured in 100 ng/mL of nocodazole for 8h.
See details in Supplemental Data
See details in Supplemental Data
See details in Supplemental Data
HeLa cells are treated with 50 ng/mL nocodazole for 12h to induce M-phase arrest. Mitotic cells were collected and lysed in lysis buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.5% NP-40, 50 mM NaF, 1 mM EDTA, 1mM DTT and protease inhibitor cocktail). Cell lysates (1 mg) were treated with 100 units of calf intestinal phosphatase (CIP) (NEB) or mock-treated at 37°C for 1h. After CIP inactivation by addition of 10 mM sodium vanadate and 10 mM EDTA, lysate was incubated with histone H3S10P-biotinylated peptide (AbCam) or no peptide for 45 min at room temperature. The peptide and associated proteins were pulled-down using streptavidin-agarose beads (Roche) at 4°C for 1h. The beads and associated proteins were washed and bound proteins were separated by SDS-PAGE and analyzed by immunoblot.
HeLa cells were suspended in nuclei isolation buffer (NIB; 15 mM Tris-HCl [pH 7.5], 60 mM KCl, 15 mM NaCl, 5 mM MgCl2, 1 mM CaCl2, 2 mM sodium vanadate, 10 mM NaF, 250 mM sucrose, protease inhibitor cocktail and 200 μg/mL RNase A). An equal volume of NIB containing 0.6% NP-40 was added to the cells, the suspension was gently mixed and incubated on ice for 5 min. Nuclei were pelleted by centrifugation at 2000xg for 5 min at 4°C and digested at 37°C for 5 min with 15 U of MNase (GE Healthcare) in 300 μL of NIB buffer. Following centrifugation for 10 min at 10,000xg at 4°C, the supernatant was removed by aspiration and the pellet was resuspended by pipetting in 300 μL of ice-cold 2 mM EDTA (pH 8.0). The nucleosome preparation was incubated with protein A-beads (Invitrogen) and either 10 μL of anti-SRp20 antibody (Invitrogen) or anti-ASF/SF2 antibody (Invitrogen) for 16h at 4°C. The immunoprecipitates were washed 5 times with 0.1% NP-40 containing wash buffer. Bound proteins were separated by SDS-PAGE and analyzed by immunoblot. For HP1 protein-ASF/SF2 association experiments, 1 mg (for HP1β) or 2.5 mg (for HP1γ) of soluble protein isolated from mitotically arrested HeLa cells was used for the immunoprecipitations with control IgG or mouse anti-ASF/SF2 antibody (Invitrogen). Bound proteins were separated by SDS-PAGE and analyzed by immunoblot.
Mitotic HeLa cells were harvested and whole cell extracts were prepared by lysing cells in lysis buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.5% NP-40, 50 mM NaF, 1 mM EDTA, 1 mM DTT and protease inhibitor cocktail) for 1h at 4°C. For chromatin fractionation, we followed the protocol previously described (Fischle et al., 2005; Mendez and Stillman, 2000). Briefly, mitotic HeLa cells were resuspended in 10 mM HEPES (pH 7.9), 10 mM KCl, 1.5 mM MgCl2, 0.34 M sucrose, 10% glycerol, 1 mM DTT, 1 mM PMSF, 0.1% Triton X-100 and protease inhibitor cocktail and incubated on ice for 5 min followed by centrifugation at 1300xg for 5 min at 4°C. The supernatant represents the soluble protein fraction. The remaining pellet is washed two times with HEPES buffer described above and resuspended in 3 mM EDTA, 0.2 mM EGTA, 1 mM DTT and protease inhibitor cocktail and incubated on ice for 30 min, followed by centrifugation at 1700xg for 5 min at 4°C. The pellet is then resuspended in the original HEPES buffer, sonicated to release chromatin associated proteins and centrifuged to remove residual insoluble proteins and nuclear matrix. The resulting supernatant is designated as the chromatin-enriched fraction and was probed with anti-ASF/SF2 (Invitrogen), anti-SRp20 (Invitrogen), anti-HP1α (Upstate), anti-HP1β (Upstate), anti-HP1γ (AbCam), anti-H3S10P (Cell Signaling) and/or anti-histone H1 (Upstate) antibodies.
HeLa cells were plated on coverslips in a 12-well plate at 2.5x104 cells/well the day before the first transfection. Thirty nM control RNAi or ASF/SF2 RNAi (Invitrogen) were transfected using Lipofectamine 2000 (Invitrogen). The RNAi transfection was repeated 24h later. Cells were immunocytostained according to method described under Immunocytochemistry 48h after the first transfection. RNAi treated HeLa cells used for biochemical fractionation analysis or whole cell lysate analysis by SDS-PAGE were treated with 100 ng/mL nocodazole for 12h before harvesting to enrich for mitotic cells.
HeLa cells grown in a monolayer on coverslips were fixed in 4% paraformaldehyde for 15 min at 4°C and permeabilized for 30s with ice-cold methanol. Coverslips were washed 3 times with 1% milk/150 mM sodium acetate in 1xPBS and 3 times with 1% milk in 1xPBS and incubated with anti-SRp20 (Invitrogen) or anti-ASF/SF2 (Invitrogen) antibodies with either anti-H3 S10P (Upstate), anti-H3 K9Me2 (Upstate), anti-H3 acetylated lysine (AbCam), anti-H3 (Upstate), anti-HP1β (Upstate), anti-HP1α (Upstate) or anti-HP1γ (Upstate) antibodies overnight at 4°C. After washing 3 times in 1% milk in 1xPBS, coverslips were subsequently incubated with both anti-mouse-Alexa 488 and anti-rabbit-Alexa 563 (Invitrogen) for 1h at room temperature, followed by Hoechst 33342 (Invitrogen) staining to visualize DNA. Cells were visualized under 100X magnification on a LSM 510 Meta microscope (Carl Zeiss) and the midplane of each cell was imaged.
DT40-ASF chicken cells were harvested at desired times, washed twice in cold 1 mL 1xPBS, resuspended to single cell suspension in 50 μL 1xPBS prior to fixation with ethanol (950 μL ice-cold 70% ethanol). Cells were stored at −20°C in fixative for at least 2h. Cells were centrifuged for 5 min at 200xg and ethanol was aspirated. The cell pellet was resuspended in 1 mL 1x PBS, centrifuged for 5 min at 200xg, decanted and wash was repeated two more times. The cell pellet was resuspended in 500 μL of PI staining solution (50 μg/mL PI, 0.1% Triton X-100, 2 mg RNase A in 1xPBS) and incubated at 37°C for 20 min. Flow cytometry analysis was performed at the Robert H. Lurie Comprehensive Cancer Center Flow Cytometry Core Facility.
We wish to thank Drs. I. Schulman, S. Huang, H. Kiyokawa, T. Volpe, G. Ghosh and anonymous reviewers for suggestions and comments on the manuscript. We also thank Dr. Teng-Leong Chew and Northwestern University Robert H. Lurie Comprehensive Cancer Center Flow Cytometry and Microscope Core Facility for help with FACS analysis and confocal imaging. This work was supported by grant from the NIH (RO1-DK65148) to D.C. and (R01-GM48259) to J.L.M. MRB received support from NIH-T32 DK 007169. The authors declare no conflicts of interest.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.