Volume 1, Issue 4, Pages 299-308 (April 2012) RBMX: A Regulator for Maintenance and Centromeric Protection of Sister Chromatid Cohesion  Sachihiro Matsunaga, Hideaki Takata, Akihiro Morimoto, Kayoko Hayashihara, Tsunehito Higashi, Kouhei Akatsuchi, Eri Mizusawa, Mariko Yamakawa, Mamoru Ashida, Tomoko M. Matsunaga, Takachika Azuma, Susumu Uchiyama, Kiichi Fukui  Cell Reports  Volume 1, Issue 4, Pages 299-308 (April 2012) DOI: 10.1016/j.celrep.2012.02.005 Copyright © 2012 The Authors Terms and Conditions

Cell Reports 2012 1, 299-308DOI: (10.1016/j.celrep.2012.02.005) Copyright © 2012 The Authors Terms and Conditions

Figure 1 Localization of RBMX and Mitotic Progression Defect (A) HeLa cells were stained with anti-RBMX antibody (green). DNA was stained with Hoechst 33342 (blue). Deconvolution processing was performed on all images. Scale bar, 5 μm. (B) RBMX was depleted in HeLa cells stably expressing GFP-H1.2. The number (top right) in each panel shows the time point (min) after nuclear envelope breakdown (NEBD). In RBMX-depleted cells, four types of aberrant chromosome dynamics (I–IV) were observed: I, chromosomes never formed the distinct metaphase plate and were scattered throughout the cytoplasm; II, chromosomes were aligned at the metaphase plate but began to scatter throughout the cytoplasm after a long arrest at metaphase; III, chromosomes were segregated without alignment at the metaphase plate; and IV, chromosomes were decondensed in early mitosis, and the cell became apoptotic. (C) Frequencies of the four types of aberration of chromosome dynamics in RBMX RNAi cells. Categories I–IV correspond to the numerals indicated in (B). (D) The time intervals between NEBD and the onset of anaphase were measured in live-cell imaging of control and RBMX RNAi cells. RBMX-depleted cells showed a significant delay in the onset of anaphase. (E) Changes in the localization of outer kinetochore proteins. After immunostaining of CENP-E, CENP-F, Hec1, and CENP-C, the fluorescence intensities of each protein were quantified. (n > 3; mean values and SD are shown). Error bars represent SD in triplicate experiments. See also Figure S1. Cell Reports 2012 1, 299-308DOI: (10.1016/j.celrep.2012.02.005) Copyright © 2012 The Authors Terms and Conditions

Figure 2 Domain Analyses of RBMX (A) Domain structure of RBMX. (B) Mitotic index was calculated after immunostaining with anti-GFP and anti-α-tubulin antibodies. HeLa cells were transfected with RNAi-refractory GFP-tagged RBMX deletion mutants and were subsequently transfected with RBMX siRNA. Error bars represent SD in triplicate experiments. (C) Localization of RBMX deletion mutants. HeLa cells transiently expressing GFP-tagged RBMX deletion mutants were fixed with 4% PFA and immunostained with anti-GFP antibody. Scale bar, 5 μm. (D) Fluorescence recovery of RBMX deletion mutants was monitored and quantified. Lines represent the best-fitting curves for the individual data sets. (n ≥ 10; mean values and SD are shown). See also Figure S2. Cell Reports 2012 1, 299-308DOI: (10.1016/j.celrep.2012.02.005) Copyright © 2012 The Authors Terms and Conditions

Figure 3 RBMX Functional Analyses in Chromosome Construction (A) Chromosome morphology in control and in RBMX-depleted cells. Cells were synchronized at mitosis using 0.1 μg/ml colcemid for 3 hr, and spread on a glass slide using Cytospin. Chromosomes were visualized with DAPI. Scale bar, 5 μm. (B) Chromosome morphologies were categorized into four groups: closed arm, chromosomes maintained sister chromatid cohesion over their entire length; open arm, sister chromatids cohered at the centromeric region, but not along the chromosome arm; mild separation, sister chromatids were partially separated, and their cohesion was maintained at the centromere or on the chromosome arm, and primary constriction was not observed at the centromere; and complete separation, sister chromatids were completely separated. DNA was stained with DAPI (blue), and kinetochores were visualized with CENP-A (green). Values were derived from more than 57 cells in 3 independent experiments. (C) RBMX knockdown cells were immunostained with anti-Sgo1 antibody (green) and anti-CREST serum (red). Enlargement shows an example of the centromeric region. Scale bar, 5 μm. (D) RBMX knockdown cells were immunostained with anti-B56α antibody (green) and anti-CREST serum (red). Enlargement shows an example of the centromeric region. Scale bar, 5 μm. (E) RBMX knockdown cells were immunostained with anti-Scc1 (green) antibody. Scale bar, 5 μm. (F) Plk1 knockdown, and RBMX and Plk1 double-knockdown cells were immunostained with anti-Sgo1 antibody (green) and anti-CREST serum (red). Scale bar, 5 μm. See also Figure S3. Cell Reports 2012 1, 299-308DOI: (10.1016/j.celrep.2012.02.005) Copyright © 2012 The Authors Terms and Conditions

Figure 4 RBMX Is Required for Cohesion Maintenance (A) The distance between sister chromatids was measured by FISH experiments. HeLa cells were synchronized by double-thymidine treatment. Cells were harvested 5 hr after release from the second arrest. FISH analysis using a probe specific for the trisomic tff1 locus on chromosome 21 successfully detected two pairs of FISH signals (green) in the control and RBMX RNAi cells. DNA was counterstained with DAPI (blue). The paired signals are magnified (lower left in each image). The distance between paired FISH signals was measured. Scale bar, 5 μm. (B) Depletion of RBMX at G2 phase activated the G2/M checkpoint. HeLa cells were synchronized by double-thymidine treatment. RBMX and histone H3 were inactivated at G2 phase by chromophore-assisted light inactivation with KillerRed fused with RBMX (KR-RBMX) and histone H3 (KR-H3). The controls are HeLa cells expressing GFP-H1.2. The control, KR-RBMX, and KR-H3 cells were irradiated with green light 5.5 hr after release from the second arrest (control light, KR-RBMX light, and KR-H3). The time of mitotic entrance was determined by detecting the nuclear envelope breakdown using live imaging based on GFP-H1.2 fluorescence. (C) Immunoblotting was performed using immunoprecipitates from the nuclear fraction. Immunoprecipitates (IP) were obtained with Scc1 antibody (Scc1) or a control antibody and analyzed by immunoblotting with Scc1, Smc3, SA1, and RBMX antibodies. Of the IP protein amount, 1.25% was loaded for the input. (D) Immunoblotting was performed using immunoprecipitates from the nuclear fraction of HeLa cells transiently expressing GFP or stably expressing GFP-RBMX. Immunoprecipitates (IP) were obtained with GFP antibody and analyzed by immunoblotting with RBMX, Scc1, Smc3, and Wapl antibodies. Of the IP protein amount, 1.25% was loaded for the input. (E) Immunoblotting was performed using immunoprecipitates from the nuclear fraction at S and G2 phases, the chromatin (M-ch) and cytoplasm fractions (M-cy) at M phase of synchronized cells stably expressing GFP-RBMX. Immunoprecipitates (IP) were obtained with GFP antibody and analyzed by immunoblotting with Scc1, Smc3, and GFP antibodies. Of the IP protein, 1.25% was loaded for the input. (F) Immunoblotting was performed using immunoprecipitates from the nuclear fraction of HeLa cells transiently expressing GFP-TRR-del. Immunoprecipitates (IP) were obtained with GFP antibody and analyzed by immunoblotting with Scc1 and Smc3 antibodies. Of the IP protein, 1.25% was loaded for the input. (G) Chromosome morphologies in knockdown cells were categorized into three groups: closed arm, chromosomes maintained sister chromatid cohesion over their entire length; open arm, sister chromatids cohered at the centromeric region, but not along the chromosome arm; and separation, sister chromatids were separated. See also Figure S4. Cell Reports 2012 1, 299-308DOI: (10.1016/j.celrep.2012.02.005) Copyright © 2012 The Authors Terms and Conditions

Figure S1 RNAi of RBMX Using Two Different siRNA Sequences, Related to Figure 1 (A) Synchronized HeLa cells were harvested at the indicated time points [h] and total extracts were analyzed by immunoblotting using RBMX, α-tubulin, cyclin-B1 and histone H3 antibodies. (B) RNAi of RBMX using siRNAs with different sequences. For control siRNA, we used a random RNA sequence. For RBMX siRNA, two RBMX specific sequences (RBMX siRNA-1 and RBMX siRNA-2) were designed. Both RBMX siRNAs caused approximately 70% reduction in the expression of RBMX. Numbers at the bottom of the panels show the relative amounts of protein loaded for immunoblotting. (C) After 48 hr transfection with control siRNA, siRNA-1 or siRNA-2, HeLa cells were fixed with 4% PFA and stained with DAPI to analyze chromosome morphology. The depletion of RBMX by both siRNAs caused similar chromosome aberrations. (D) The percentage of chromosome aberration in mitotic cells. Chromosomes showing aberrations were categorized as either non-aligned or misaligned chromosomes. “RBMX rescue” indicates RBMX-depleted cells (using siRNA-1) into which an RNAi-refractory RBMX vector was transfected. The percentage of chromosome aberration was almost same for siRNA-1 and siRNA-2. n = 5; > 1000 cells were counted in each experiment. (E) Control metaphase cells and RBMX-depleted cells showing aberrant chromosomes were stained with anti-cyclin B1 (green) after 4% PFA fixation. DNA was stained with Hoechst33342 (blue). Bar, 5 μm. (F) Effect of RBMX depletion with RBMX siRNA-1 and siRNA-2 on protein expression of Securin, sororin, and tubulin. Immunoblotting was performed using HeLa cell extracts. (G) Metaphase arrest in RBMX-depleted cells. 45 hr after siRNA transfection, HeLa cells were incubated with 20 μM MG132 for 3 hr to inhibit the anaphase transition. n = 3; > 1000 cells were counted in each experiment. Error bars represent SD in triplicate experiments. (H–J) Immunoblotting showing protein levels in RNAi cells. The levels of Scc1, Sgo1, and methylation level of tri-methylated K9 of histone H3 (H3K9Me3) (H), the levels of Bub1 and BubR1 (I), and the levels of Plk1, HP1α, and Mad2 (J), were not changed after depletion of RBMX by RNAi. Numbers at the bottom of the panels show the relative amounts of protein loaded for immunoblotting. (K–N) Localization of outer kinetochore proteins in RBMX-depleted cells. CENP-E (K), CENP-F (L), Hec1 (M) and Mis13 (N) were visualized by immunostaining using their antibodies in control and RBMX RNAi cells. Centromeres and DNA were stained with CREST (red) and DAPI (blue). Enlargements show examples of the centromeric regions. Bars, 5 μm. (O) Localization of inner kinetochore proteins in RBMX-depleted cells. CENP-A and CENP-C were visualized by immunostaining using their antibodies in control and RBMX RNAi cells. DNA was counter stained with DAPI (blue). There were no differences in the localizations of CENP-A and CENP-C between the control and RBMX RNAi cells. Bar, 5 μm. Cell Reports 2012 1, 299-308DOI: (10.1016/j.celrep.2012.02.005) Copyright © 2012 The Authors Terms and Conditions

Figure S2 Chromosome Morphology Rescued by Domain-Deletion Plasmids, Related to Figure 2 (A) Localization of the GFP-tagged TRR of RBMX. HeLa cells transiently expressing GFP-tagged RBMX deletion mutants were fixed with 4% PFA and immunostained with anti-GFP antibody. Bar, 5 μm. (B) Chromosome morphology in RBMX-depleted cells with domain-deletion rescue plasmids. Cells were synchronized at mitosis using 0.1 μg/ml colcemid for 3 hr, and spread on a glass slide using Cytospin. Chromosomes were visualized with DAPI. Bar, 5 μm. (C) Chromosome morphologies were categorized into four groups as shown in Figure 3B. Cell Reports 2012 1, 299-308DOI: (10.1016/j.celrep.2012.02.005) Copyright © 2012 The Authors Terms and Conditions

Figure S3 Localization of Kinetochore Proteins in RBMX-Depleted Cells, Related to Figure 3 (A) Expression level of Scc1 in Scc1-depleted cells. Immunoblotting was performed using the HeLa cell extract. (B) Localization of outer kinetochore proteins in Scc1-depleted cells. CENP-E and Hec1 were visualized by immunostaining using their antibodies in Scc1 RNAi cells. Bar, 5 μm. (C) Quantification of centromeric localization of outer kinetochore proteins in Scc1-depleted cells. After immunostaining of CENP-E and Hec1, the fluorescence intensities of each protein were quantified using the signal intensity of CREST serum. (n > 3; mean values and SD are shown). Error bars represent SD in triplicate experiments. Cell Reports 2012 1, 299-308DOI: (10.1016/j.celrep.2012.02.005) Copyright © 2012 The Authors Terms and Conditions

Figure S4 Association of RBMX with Chromatin and Cohesin, Related to Figure 4 (A) Association of RBMX with chromatin. The amount of RBMX was analyzed using cytoplasmic (Cy), nuclear (Nu) and chromatin (Ch) fractions prepared from HeLa cells synchronized by double thymidine treatment. The Nu and Ch fractions were prepared using different concentrations (0 and 500 mM) of NaCl. These fractions were subjected to immunoblotting with anti-Smc3. Scc1, α-tubulin and RBMX antibodies. CBB shows histones stained with Coomassie brilliant blue. (B and C) Cohesin-independent association of RBMX with chromatin. RBMX-GFP expressing cells were transfected with Scc1 siRNA and synchronized in the G2 phase. Cells were fixed with 4% PFA (fixed cells) (B) or were pre-extracted with 0.1% Triton X-100 before fixation (pre-extracted cells) (C). Cells were immunostained by anti-GFP antibody (green) and DNA was stained with Hoechst33342 (blue). Quantification of fluorescence levels is shown in the bar charts (n > 3; mean values and SD are shown). Error bars represent SD in triplicate experiments. (D–F) Chromophore-assisted light inactivation of RBMX at G2 phase. HeLa cells expressing histone H1.2 fused with GFP (H1-GFP), and RBMX or histone H3 fused with KillerRed were synchronized with thymidine. Live imaging was performed using differential interference contrast (DIC) and GFP fluorescence. The number on the top of each panel shows the time point (min) after the release. Without the irradiation, the cell expressing histone H1.2 fused with GFP and RBMX fused with KillerRed started to divide 372 min after the release from thymidine arrest. Two daughter cells were formed at 417 min (D). At 330 min after the release from thymidine arrest, the cells expressing histone H1.2 fused with GFP and RBMX or histone H3 fused with KillerRed were irradiated using 550 nm green light. The irradiated cells co-fused with KillerRed-RBMX did not enter mitosis even at 720 min (E) while those co-fused with KillerRed-H3 entered apoptosis at 330 min (F). Bars, 5 μm. (G) Native interaction of RBMX with Scc1. Immunoblotting was performed using immunoprecipitates from the nuclear fraction of HeLa cells. Immunoprecipitates (IP) were obtained with RBMX antibody and analyzed by immunoblotting with RBMX and Scc1 antibodies. 1.25% of the IP protein amount was loaded for the input. Cell Reports 2012 1, 299-308DOI: (10.1016/j.celrep.2012.02.005) Copyright © 2012 The Authors Terms and Conditions