Son is essential for nuclear speckle organization and cell cycle progression

Abstract

Subnuclear organization and spatiotemporal regulation of pre-mRNA processing factors is essential for the production of mature protein-coding mRNAs. We have discovered that a large protein called Son has a novel role in maintaining proper nuclear organization of pre-mRNA processing factors in nuclear speckles. The primary sequence of Son contains a concentrated region of multiple unique tandem repeat motifs that may support a role for Son as a scaffolding protein for RNA processing factors in nuclear speckles. We used RNA interference (RNAi) approaches and high-resolution microscopy techniques to study the functions of Son in the context of intact cells. Although Son precisely colocalizes with pre-mRNA splicing factors in nuclear speckles, its depletion by RNAi leads to cell cycle arrest in metaphase and causes dramatic disorganization of small nuclear ribonuclear protein and serine-arginine rich protein splicing factors during interphase. Here, we propose that Son is essential for appropriate subnuclear organization of pre-mRNA splicing factors and for promoting normal cell cycle progression.

Figure 1.

Sequence motifs in Son. The amino-terminal half of Son contains multiple novel sequence motif repeats that are unique to Son. Repeats composed of unique sequences are shown here as sets of different colored bars, with each bar representing an individual repeat. The carboxy-terminal region of Son contains motifs such as a RS domain, a double-stranded RNA binding domain, and a G patch. Amino acids corresponding to C-terminal deletion constructs are indicated.

Figure 2.

Son is a component of nuclear speckles that is enriched in insoluble nuclear fractions. (A) Localization of YFP-Son in HeLa cells stably transfected with YFP-tagged full-length Son cDNA and processed for immunolocalization of splicing factor SF2/ASF. YFP-Son (a) colocalizes with SF2/ASF (b) at nuclear speckles (arrow). (B) Immunofluorescence using rabbit polyclonal anti-Son WU14 antibody shows that WU14 labeling (i, arrow) and WU15 labeling (q, arrow) precisely colocalize with labeling of splicing factor SF2/ASF (arrows in j and r) at nuclear speckles, whereas WU14 preimmune serum (e) and WU15 preimmune serum (m) do not label any specific nuclear domain. (C) Peptides used to generate Son antisera block immunoreactivity. The peptide corresponding to amino acid sequence in the amino terminus of Son that was used to generate the WU14 and WU15 rabbit polyclonal anti-Son antibodies blocks immunoreactivity of WU14 and WU15 by immunofluorescence. Blocking with 10 μM amino-terminal Son peptide before incubation of HeLa cells with WU14 antibody blocks WU14 immunolabeling (Ci) and WU15 immunolabeling (Ck) at nuclear speckles. Reactions without peptide (C, a and c) and with unspecific C-terminal Son peptide (C, e and g) do not block WU14 immunolabeling. (D) HeLa-S3 cells fractionated to yield nuclear insoluble proteins contain Son. An immunoblot with anti-Son antibody WU15 shows reactivity with a large protein that is enriched in a nuclear insoluble fraction of HeLa cells (lane 3) compared with whole cell extract (lane 1) or nuclear extract (lane 2). The bottom band could be an alternate splice form or breakdown product of Son. Treatment of HeLa cells with siRNA duplexes that deplete Son specifically reduced labeling of these bands (lane 4). DNA was stained with 4,6-diamidino-2-phenylindole (DAPI). Bar, 5 μm.

Figure 3.

siRNA-mediated depletion of Son alters the localization of U1-70K snRNP protein. HeLa cells were transfected with empty vehicle (mock), luciferase siRNA (control), Son siRNA 1, or Son siRNA 4. 24 h posttransfection, the cells were fixed and processed for dual immunolocalization of Son and U1-70K. Son signal is nearly absent after Son depletion (i and m). U1-70K localization in nuclear speckles was altered after depletion of Son with siRNA 1 (j, arrow) but not with mock (b, arrow) or control (f, arrow) siRNAs. siRNA-mediated depletion of Son by using Son siRNA 4 showed the same alteration in U1-70K nuclear speckle localization (n, arrow) compared with mock and control siRNAs, but did not alter the U1-70K distribution in Cajal bodies (n, arrowhead). DNA was stained with DAPI. Bar, 5 μm.

Figure 4.

siRNA-mediated depletion of Son alters the localization of SF2/ASF. (A) SF2/ASF localization is altered after siRNA-mediated Son depletion. HeLa cells were treated with empty vehicle (mock), luciferase siRNA (control), or Son siRNA 1. Twenty-four hours posttransfection, HeLa cells were fixed and processed for dual immunolocalization of Son and SF2/ASF. Son signal is nearly absent after Son depletion (i and m). SF2/ASF localization in nuclear speckles was altered after depletion of Son with siRNA 1 (j, arrow) but not with mock (b, arrow) or control (f, arrow) siRNAs. siRNA-mediated depletion of Son by using Son siRNA 4 showed the same alteration in SF2/ASF nuclear speckle localization (n, arrow) compared with mock and control siRNAs. DNA was stained with DAPI. Bar, 5 μm. (B) Son expression is reduced after RNAi. Semiquantitative RT-PCR showed a significant reduction in Son mRNA expression after siRNA-mediated Son depletion (lanes 4 and 8), whereas it was not reduced in cells treated with mock transfection (lanes 2 and 6) or control siRNAs (lanes 3 and 7). Note that GAPDH mRNA was not reduced after any of the treatments (lanes 2–4 and 6–8). Neither mRNA was detected when reverse transcriptase was omitted from the reaction (lanes 1 and 5).

Figure 5.

siRNA-mediated depletion of Son alters the localization of core EJC proteins. (A) YFP-MLN51 localization is altered after siRNA-mediated Son depletion. HeLa cells were treated with empty vehicle (mock), luciferase siRNA (control), or Son siRNA 4. Twenty-four hours posttransfection, HeLa cells were fixed and processed for immunolocalization of Son. Son signal is nearly absent after Son depletion (i). YFP-MLN51 localization in nuclear speckles was altered after depletion of Son with siRNA 4 (j, arrow) but not with mock (b, arrow) or control (f, arrow) siRNAs. (B) YFP-Magoh localization is altered after siRNA-mediated Son depletion. HeLa cells were treated with empty vehicle (mock), luciferase siRNA (control) or Son siRNA 4. Twenty-four hours posttransfection, HeLa cells were fixed and processed for immunolocalization of Son. Son signal is nearly absent after Son depletion (u). YFP-Magoh localization in nuclear speckles was altered after depletion of Son with siRNA 4 (v, arrow) but not with mock (n, arrow) or control (r, arrow) siRNAs. DNA was stained with DAPI. Bar, 5 μm.

Figure 6.

siRNA-mediated depletion of Son alters the localization of nuclear speckle polyA⁺ RNA. HeLa cells were treated with luciferase siRNA (control), or Son siRNA 4. Twenty-four hours posttransfection, HeLa cells were fixed and processed for immunolocalization of SF2/ASF to monitor nuclear speckle phenotype and RNA FISH with oligo(dT) probe to detect polyA⁺ RNA. Oligo(dT) probe localization in nuclear speckles was altered after depletion of Son with siRNA 4 (e, arrow) in the same way as SF2/ASF (f, arrow), whereas both showed normal nuclear speckle localization after treatment with control siRNA (a–d). DNA was stained with DAPI. Bar, 5 μm.

Figure 7.

Nuclear speckle reorganization after Son depletion is not a result of reduced global transcription, and transcription inhibition causes a different phenotype. (A) Global labeling of transcription sites in situ. HeLa cells were treated with control siRNA (a–d) or Son siRNAs (e–l) for 24 h and then permeabilized for uptake of Br-UTP to label nascent transcripts. Bromo-UTP incorporation was equally detectable in nuclei of Son-depleted cells (f and j) and cells treated with contol siRNA (b). Bromo-UTP labeling was absent in cells treated with α-amanitin (n). (B) Son reorganization after RNA pol II transcription inhibition. HeLa cells were treated with alpha-amanitin for 6 h and processed for immunolocalization of Son and U1-70K. U1-70K and Son redistributed to enlarged, rounded speckles upon inhibiting RNA polymerase II with α-amanitin (u, arrow) compared with untreated cells (q, arrow). DNA was stained with DAPI. Bar, 5 μm.

Figure 8.

Pre-mRNA splicing factor protein expression level is not altered by Son depletion. (Top) HeLa cells treated with siRNAs against Son were extracted for immunoblot analysis of splicing factors. Expression of splicing factors U1-70K and SF2/ASF were not significantly reduced in extracts from Son depleted cells (lanes 3 and 4) compared with cell treated with mock (lane 1) and control (lane 2) siRNAs. The actin immunoblot is shown to demonstrate equal loading of total protein in each lane. (Bottom) Son depletion was confirmed by semiquantitative RT-PCR. A duplicate siRNA depletion experiment was performed in parallel to confirm complete reduction in Son mRNA expression after 48 h of siRNA-mediated Son depletion. Son mRNA was nearly abolished after treatment of cells with Son siRNA 1 (lane 8) and Son siRNA 4 (lane 9), whereas it was not reduced in cells treated with mock (lane 6) or control (lane 7) siRNAs. GAPDH mRNA was not reduced after any of the treatments (lanes 6–9). Neither mRNA was detected when reverse transcriptase is omitted from the reaction (lane 5).

Figure 9.

Nuclear speckles reorganize directly to a doughnut-like morphology after Son depletion. Time-lapse microscopy was performed on HeLa cells that stably express YFP-SF2/ASF and were transfected with Son siRNA 4. The inset in a indicates the region of interest enlarged in the bottom panels. Live imaging was initiated 13 h after siRNA transfection. The morphology of nuclear speckles changed directly from an irregular compacted organization (a, arrows), to an open doughnut-shaped organization (f, arrows). Time after siRNA transfection is indicated in hours. Video of this data set shows the transition from speckles to doughnuts (see Supplemental Movie). Bar, 5 μm.

Figure 10.

Expression of siRNA-resistant Son rescues nuclear speckle organization in cells depleted of endogenous Son. HeLa cells transfected with plasmid encoding YFP-tagged siRNA-resistant Son (YFP-siR-Son) were plated for siRNA treatment. The YFP tag allowed visualization of exogenous Son in transiently transfected cells, as well as detection of exogenous Son mRNA by RT-PCR oligos that target the YFP-encoding region. Endogenous Son was detected by RT-PCR oligos that target the 3′-untranslated region that is absent in YFP-Son and YFP-siR-Son. Duplicate experiments were performed to simultaneously analyze nuclear speckle phenotype (A) and endogenous and siRNA-resistant Son expression (B). (A) YFP-siR-Son (a–h) and YFP-siR-Son (1-1493; i–p) colocalized with SF2/ASF in nuclear speckles. Treatment of cells with control siRNAs (a–d and i–l) did not affect speckle morphology. Treatment with Son siRNA 4 showed the expected nuclear speckle reorganization (i–l, arrow), but neighboring cells protected by YFP-siR-Son FL or YFP-siR-Son (1-1493) expression showed a normal nuclear speckle phenotype (YFP-positive cells in e–f and m–n, respectively). (B) YFP-siR-Son resists depletion by Son siRNA 4. Semiquantitative RT-PCR was performed to target the YFP-encoding region that amplifies both YFP-Son and YFP-siR-Son (YFP, top row), the 3′-untranslated region of endogenous Son mRNA (Son, middle row), or GAPDH (bottom row). HeLa cells were transfected with either YFP-Son (YFP-Son; left set of panels), siRNA resistant YFP-Son (YFP-siR-Son FL, middle set of panels), or siRNA-resistant Son repeats (YFP-siR-Son 1-1493, right set of panels) before siRNA treatment. YFP-Son was used as a control for evaluating siRNA depletion of a non-siRNA resistant YFP-tagged Son mRNA by RT-PCR only. Individual RT-PCR reactions were performed on total RNA from each sample. Endogenous Son mRNA is significantly reduced after siRNA treatment with Son-specific siRNA 4 (middle row, lanes 4–5, 9–10, and 14–15) but not with mock transfection or control siRNAs (middle row, lanes 2–3, 7–8, and 12–13). While the nonresistant YFP-Son was reduced by treatment with both Son siRNAs (top row, left panel, lanes 4–5), the siRNA resistant YFP-siR-Son constructs were not reduced by treatment with siRNA 4 (top row, lanes 10 and 15) but were depleted by siRNA 1 (top row, lanes 9 and 14). GAPDH mRNA was not reduced after any of the treatments (bottom row, lanes 2–5, 7–10, and 12–15). No mRNA was detected when reverse transcriptase was omitted from the reaction (lanes 1, 6, and 11).

Figure 11.

Son depletion causes cell growth defects and cell cycle arrest. (A) Cell growth curves. Sixteen sets of siRNA experiments (set = mock, control, si1, and si4) were performed and cell counts were obtained on a ViCell after a time course. Growth curves indicate that cell growth stalled in Son-depleted cells at 48 h, whereas cells continued to grow beyond 48 h after mock and control siRNA treatments. The graph depicts average numbers of cells from three separate groups of experiments. (B) Cell cycle is altered in Son-depleted cells. The siRNA experiments from the 48-h time point in part A were also analyzed by flow cytometry; data shown is representative of one of the four experimental groups. Untreated cells were used to set the gating on the flow cytometer and showed expected peaks at G1 and G2/M. Graphs of DNA content are shown for mock transfected, control siRNA treated cells and cells treated with Son siRNAs 1 and 4. A corresponding table of the percent of cells in each cell cycle stage is shown. (C) Immunoblot analysis of Son-depleted cells shows an increase in histone H3 modification (serine 10 phosphorylation). Cells from the 48-h time point in A were extracted in Laemmli buffer for immunoblot analysis. There was an increase in phosphorylation of histone H3Ser(10) in extracts from Son-depleted cells (lanes 3 and 4) as compared with mock (lane 1) and control extracts (lane 2). Extract from nocodazole arrested HeLa cells (N, lane 5) was used as a positive control for elevated H3 Ser(10)P in mitosis. Actin immunoblotting was used to confirm equal loading of total protein in each lane.

Figure 12.

Son depleted cells arrest in metaphase. Son siRNA mixture or Son siRNA 4 were transfected to HeLa cells stably expressing histone H2B-GFP at 10 and 100 nM, respectively. (A) After transfection, GFP fluorescence was observed every 12 h. An increase in mitotic cells was seen after 36 h of Son depletion (i–j and n–o). (B) Number of cells in mitosis scored in A plotted over time. (C) Mitotic index and number of cells in each mitotic phase were scored for >1000 cells. Son depletion caused arrest of mitosis in metaphase as determined by the dramatic doubling of cells in prometaphase/metaphase and almost complete absence of cells in anaphase and telophase.