Volume 33, Issue 4, Pages 450-461 (February 2009) Chromatin Binding of SRp20 and ASF/SF2 and Dissociation from Mitotic Chromosomes Is Modulated by Histone H3 Serine 10 Phosphorylation  Rebecca J. Loomis, Yoshinori Naoe, J. Brandon Parker, Velibor Savic, Matthew R. Bozovsky, Todd Macfarlan, James L. Manley, Debabrata Chakravarti  Molecular Cell  Volume 33, Issue 4, Pages 450-461 (February 2009) DOI: 10.1016/j.molcel.2009.02.003 Copyright © 2009 Elsevier Inc. Terms and Conditions

Figure 1 SRp20 and ASF/SF2 Bind Histone H3 in a Modification-Selective Manner (A) Silver stain of a gel showing the proteins that differentially bound to indicated histone H3 peptide tails. The indicated bands were identified by protein microsequencing. (B and C) Endogenous SRp20 (B) or ASF/SF2 (C) does not bind histone H3 phosphorylated at serine 10. HeLa cell nuclear extracts were incubated with indicated biotinylated histone H3 tail peptides. Bound proteins were subject to immunoblot analysis with anti-SRp20 or ASF/SF2 antibody. (D) Domain structure of SRp20 (top panel). SRp20 binds histones through the RS domain (lower panel). Acid-extracted histones isolated from asynchronously growing HeLa cells were separately incubated with equal amounts of indicated GST-fusion proteins and pulled down with glutathione-sepharose beads. Bound proteins were subject to immunoblot analysis with anti-histone H3 antibody. Molecular Cell 2009 33, 450-461DOI: (10.1016/j.molcel.2009.02.003) Copyright © 2009 Elsevier Inc. Terms and Conditions

Figure 2 Phosphorylation of SRp20 and ASF/SF2 Determines Their Modification-Selective Binding to Histone H3 (A) Bacterially expressed SRp20 bound both unmodified histone H3 and histone H3 phosphorylated at serine 10. GST-SRp20 or GST alone was incubated with acid-extracted HeLa cell histones. Bound proteins were subject to immunoblot analysis with anti-histone H3 or anti-phosphoserine 10 histone H3 antibodies. (B) Endogenous SRp20 is phosphorylated. HeLa cell nuclear extracts or GST-SRp20 were treated with or without CIP and run on SDS-PAGE for immunoblot analysis. (C) GST-SRp20 can be phosphorylated in vitro. Recombinant GST-SRp20 was either mock incubated, incubated with buffer or incubated with ATP and HeLa cell S100 fraction, and analyzed by immunoblot with anti-SRp20 and anti-phosphoserine antibodies. (D) Phosphorylated GST-SRp20 no longer binds to histone H3 when phosphorylated at serine 10. GST-SRp20 was preincubated with ATP and HeLa cell S100 fraction prior to GST pull-down assay with acid-extracted histones. (E) Endogenous hyperphosphorylated SR proteins do not bind to histone H3-S10P peptide tail, whereas in vitro hypophosphorylated SR proteins do. Endogenous SRp20 and ASF/SF2 were mock-incubated or incubated with CIP, followed by incubation with phosphoserine 10 histone H3-biotinylated peptide and analyzed by immunoblot with anti-SRp20 and anti-ASF/SF2 antibodies. Input represents 50 μg or 5% of total soluble protein fraction used in the immunoprecipitation reactions. Unbound samples represent 10% of the total volume recovered following immunoprecipitation reactions (approximately equivalent to 100 μg total protein used in the immunoprecipitation). Bound fractions represent all the proteins from the immunoprecipitation reactions that bound to the histone H3S10P peptide. Molecular Cell 2009 33, 450-461DOI: (10.1016/j.molcel.2009.02.003) Copyright © 2009 Elsevier Inc. Terms and Conditions

Figure 3 Analysis of Endogenous SRp20 and ASF/SF2-Histone Interactions (A and B) Endogenous SRp20 and ASF/SF2 bound to unmodified, acetylated, and methylated nucleosomes, but neither bound to phosphorylated nucleosomes. Nucleosomes were prepared from control and nocodazole-treated cells as described in the Experimental Procedures and immunoprecipitated with anti-SRp20 or anti-ASF/SF2 antibodies. Bound proteins were subject to immunoblot analysis with (A) anti-SRp20 (left panels) or ASF/SF2 (right panels), (B) anti-histone H3 (first panel), anti-phosphoserine 10 histone H3 (second panel), anti-acetyl lysine 9 histone H3 (third panel), or anti-dimethyl lysine 9 histone H3 (fourth panel) antibodies. Inputs represent 5% of the amount used for immunoprecipitation. (C) Both SRp20 and ASF/SF2 colocalized with total histone H3 and interphase chromatin. HeLa cells were fixed and immunostained for SRp20 (Cb) or ASF/SF2 (Cg) and total histone H3 (Ca and Cf). Hoechst staining showed nuclear staining of DNA (Cc and Ch). Merged images of SRp20/histone H3 (Cd), SRp20/DNA (Ce), ASF/SF2/histone H3 (Ci), and ASF/SF2/DNA (Cj) are shown. A section from each field was enlarged to better visualize colocalization (Cd′, Ce′, Ci′, and Cj′). Data are representative of three independent experiments for each ASF/SF2 and SRp20 with histone H3 where approximately 30 interphase cells were analyzed. Molecular Cell 2009 33, 450-461DOI: (10.1016/j.molcel.2009.02.003) Copyright © 2009 Elsevier Inc. Terms and Conditions

Figure 4 ASF/SF2 Dissociation from Mitotic Chromatin Is Dependent on Serine 10 Phosphorylation ASF/SF2 is dissociated from hyperphosphorylated mitotic chromosomes. (A) ASF/SF2 is primarily excluded from chromatin during prometaphase (Aa–Af), metaphase (Ag–Al), and anaphase (Am–Ar) and reassociates with histone H3 as serine 10 phosphorylation levels decrease during telophase (As–Ax). HeLa cells were synchronized using a double thymidine block, released and blocked in M-phase using a short treatment with nocodazole (2 hr). Cells were released from the M-phase block and collected for microscopy every 30 min for analysis of ASF/SF2-chromatin association in relation to histone H3S10P levels. Data are representative of three independent experiments where 100 mitotic cells were analyzed (approximately 25/stage). (B) ASF/SF2 dissociation from mitotic chromatin is dependent on H3S10P. Mitotically growing HeLa cells were treated without (Ba–Be′) or with (Bf–Bj′) Aurora B kinase inhibitor ZM447439. Cells were fixed and immunostained with ASF/SF2 (Bb and Bg) and histone H3-S10P (Ba and Bf) antibodies. Hoechst staining identified condensed chromatin (Bc and Bh). A section from each field was enlarged to better visualize colocalization or a lack thereof of ASF/SF2 and phosphoserine 10-histone H3 (Bd and Bd′; Bi and Bi′), ASF/SF2 and chromosome (Be and Be′; Bj and Bj′). Data are representative of three independent experiments, where approximately 40 mitotic cells were analyzed. All stages (prometaphase, metaphase, anaphase, and telophase) demonstrated similar colocalization patterns—dissociation of ASF/SF2 from mitotic chromatin when H3S10P was high and retention when H3S10P was inhibited with the Aurora B kinase inhibitor. (C and D) Inhibition of histone H3S10P increased retention of ASF/SF2 on mitotic chromatin without affecting ASF/SF2 protein levels. Mitotically growing HeLa cells were treated without (−) or with (+) Aurora B kinase inhibitor ZM447439. Chromatin-enriched fraction (C) or whole-cell extract (D) was subject to immunoblot analysis with indicated antibodies. Molecular Cell 2009 33, 450-461DOI: (10.1016/j.molcel.2009.02.003) Copyright © 2009 Elsevier Inc. Terms and Conditions

Figure 5 SRp20 Dissociation from Mitotic Chromatin Is Dependent on Serine 10 Phosphorylation SRp20 is dissociated from hyperphosphorylated mitotic chromosomes. (A) SRp20 is primarily excluded from chromatin during prometaphase (Aa–Af), metaphase (Ag–Al), and anaphase (Am–Ar) and reassociates with histone H3 as serine 10 phosphorylation levels decrease during telophase (As–Ax). Experimental procedures and data analysis are similar to those described under Figure 4. (B) SRp20 dissociation from mitotic chromatin is dependent on H3S10P. Mitotically growing HeLa cells were treated without (Ba–Be′) or with (Bf–Bj′) Aurora B kinase inhibitor ZM447439. Cells were fixed and immunostained with SRp20 (Bb and Bg) and histone H3-S10P (Ba and Bf) antibodies. Hoechst staining identified condensed chromatin (Bc and Bh). A section from each field was enlarged to better visualize colocalization or a lack thereof of SRp20 and phosphoserine 10-histone H3 (Bd and Bd′; Bi and Bi′) and SRp20 and chromosome (Be and Be′; Bj and Bj′). Data are representative of three independent experiments where approximately 40 mitotic cells were analyzed. (C and D) Inhibition of histone H3S10P increased retention of SRp20 on mitotic chromatin without affecting SRp20 protein levels. Mitotically growing HeLa cells were treated without (−) or with (+) Aurora B kinase inhibitor ZM447439. Chromatin-enriched fraction (C) or whole-cell extract (D) was subject to immunoblot analysis with indicated antibodies. Molecular Cell 2009 33, 450-461DOI: (10.1016/j.molcel.2009.02.003) Copyright © 2009 Elsevier Inc. Terms and Conditions

Figure 6 ASF/SF2 Associates with HP1β in Mitotic Cells and ASF/SF2 Knockdown Leads to Retention of HP1β on Mitotic Chromatin (A) ASF/SF2 and HP1β colocalize in mitotic cells (Aa–Ae′). (B) ASF/SF2 and HP1β associate in HeLa mitotic cells. HeLa cells were mitotically arrested using 100 ng/ml nocodazole. The soluble protein fraction was used for immunoprecipitation with either control IgG or anti-ASF/SF2 antibody. Bound proteins were subject to immunoblot analysis with anti-HP1β antibody. Input represents 10 μg of total lysate used in the experiment. (C) RNAi treatment of HeLa cells significantly reduces ASF/SF2 expression. HeLa cells were treated with control or ASF/SF2 siRNA two times, 24 hr apart, using Lipofectamine 2000. Cells were collected for immunoblot analysis 48 hr after the first transfection and probed with anti-ASF/SF2 and anti-actin antibodies. (D–F) ASF/SF2 knockdown led to an increased retention of HP1β on mitotic chromatin despite histone H3S10P and without affecting HP1β protein levels. (D) Top panel in (D) is the same cell as that shown in (A) except with different overlapping images. Cells were fixed with paraformaldehyde, permeabilized with methanol, and stained with anti-HP1β and anti-ASF/SF2 antibodies at 48 hr post-first transfection. ASF/SF2 knockdown led to increased association of HP1β with chromatin (compare De and De′ with Dj and Dj′) despite the presence of H3S10P in metaphase cells. HP1β data are representative of 3 independent experiments and approximately 50 mitotic cells were analyzed. (E and F) HeLa cells were treated with control or ASF/SF2 siRNA two times, 24 hr apart, and then with nocodazole for 12 hr before harvesting to enrich for mitotic cells. Chromatin-enriched fraction (E) or whole-cell extract (F) was subject to immunoblot analysis with indicated antibodies. Molecular Cell 2009 33, 450-461DOI: (10.1016/j.molcel.2009.02.003) Copyright © 2009 Elsevier Inc. Terms and Conditions

Figure 7 ASF/SF2 Depletion Delays G0/G1 Entry Depletion of ASF/SF2, followed by an M-phase block and release, caused cells to be delayed entering G0/G1. DT40-ASF cells were treated with doxycycline for 16 hr, followed by cotreatment with doxycycline and nocodazole for 8 hr, and then released and analyzed for cell-cycle progression by propidium iodide staining and subsequent flow cytometry analysis. (A) Schematic of experimental design shown in (B) and (C). (B) DT40 ASF cells depleted of ASF/SF2 demonstrated a significant increase in the percentage of G2/M cells compared to ASF/SF2-expressing cells. (C) ASF/SF2 depletion in DT40-ASF cells demonstrated a significant decrease in the percentage of cells entering G0/G1 compared to cells expressing ASF/SF2. ([C], inset) DT40-ASF chicken cells, treated with 1 μg/ml of doxycycline for 24 hr, showed depletion of ASF/SF2. Cells were collected for immunoblot analysis and probed with anti-ASF/SF2 and anti-actin antibodies. Data shown in (B) and (C) are representative of four independent experiments. Molecular Cell 2009 33, 450-461DOI: (10.1016/j.molcel.2009.02.003) Copyright © 2009 Elsevier Inc. Terms and Conditions