SON controls cell-cycle progression by coordinated regulation of RNA splicing

Abstract

It has been suspected that cell-cycle progression might be functionally coupled with RNA processing. However, little is known about the role of the precise splicing control in cell-cycle progression. Here, we report that SON, a large Ser/Arg (SR)-related protein, is a splicing cofactor contributing to efficient splicing of cell-cycle regulators. Downregulation of SON leads to severe impairment of spindle pole separation, microtubule dynamics, and genome integrity. These molecular defects result from inadequate RNA splicing of a specific set of cell-cycle-related genes that possess weak splice sites. Furthermore, we show that SON facilitates the interaction of SR proteins with RNA polymerase II and other key spliceosome components, suggesting its function in efficient cotranscriptional RNA processing. These results reveal a mechanism for controlling cell-cycle progression through SON-dependent constitutive splicing at suboptimal splice sites, with strong implications for its role in cancer and other human diseases.

Figure 1. SON knockdown causes multiple defects in mitotic chromosome arrangement, spindle pole separation and microtubule dynamics

(A) Mitotic chromosome misalignment in SON-depleted cells. Wright-Giemsa staining revealed misaligned chromosomes in SON siRNA-transfected K562 cells. Chromosomes that have shifted to an abnormal location are marked with red arrows. (B) Increase in pro-metaphase cells and incomplete separation of mitotic poles. HeLa cells were transfected with control siRNA or SON siRNA, and immunostained with an Aurora kinase A antibody and DAPI. Then, mitotic cells at different phases and with different mitotic pole status were counted under the fluorescence microscope. Bars represent mean ± SD from three independent experiments (n > 300 cells per experiment, *P < 0.005, **P < 0.02, t-test). (C) Incomplete mitotic spindle pole separation and failure in chromosome alignment caused by SON knockdown. HeLa cells were prepared and immunostained as described above (green, Aurora kinase A for mitotic spindle poles; blue, DAPI for DNA). (D) Abnormal spindle pole/centrosome amplification in SON siRNA-transfected K562 cells during mitosis and interphase. Cells were stained with an Aurora kinase A antibody (green) and DAPI (blue). (E) Defects in mitotic spindle formation in SON siRNA-transfected cells. HeLa cells transfected with control siRNA or SON siRNA were fixed (day 3), and stained with anti-α-tubulin for microtubules (green) and DAPI for DNA (blue). (F) Abnormal microtubule organization in interphase cells after SON depletion. Cells were prepared and stained as described in (E). (G) Microtubule re-growth assay in HeLa cells transfected with SON siRNA or control siRNA (day 2). Cells were cold-treated to depolymerize microtubules and fixed at different time points after incubation in warm media (0, 30 sec, 2 min, 5 min, and 20 min). Microtubules were stained with α-tubulin antibody and analyzed by fluorescence microscopy.

Figure 2. SON knockdown causes abnormal nuclear structures, aneuploidy/polyploidy and DNA breaks

(A) Abnormal nuclear structure caused by SON knockdown. HeLa cells were transfected with SON siRNA, and after 3 days, immunostained with SON antibody together with DAPI. SON knockdown causes nuclear buds/lobes and micronuclei (marked with arrows). (B) Multinuclear HeLa cells were observed after transient SON knockdown. HeLa cells were transfected with SON siRNA and shown is a photo of representative multinuclear cells observed 7 days after SON siRNA transfection. (C) Multinuclear cells were observed after SON siRNA transfection in U937 human leukemic monocyte lymphoma cell line (day 6). (D) SON knockdown causes aneuploidy. BJ human primary fibroblasts were transfected with control siRNA or SON siRNA, and DNA content was measure by propidium iodide staining and flow cytometric analysis (day 7), showing that SON knockdown increases cells with >2n and >4n DNA. (E) SON knockdown causes double-stranded DNA breaks. HeLa cells were transfected with SON siRNA, and stained with γH2A.X antibody. Cells with SON knockdown (indicated by arrows) showed γH2A.X foci.

Figure 3. SON knockdown causes down-regulation of genes involved in DNA maintenance/integrity and cell cycle progression

(A) Analysis of top functions of 659 genes that showed significant changes after SON knockdown by Ingenuity Pathway Analysis (IPA) software. Fischer’s exact test was used to calculate a p-value determining the probability that each biological function and/or disease assigned to that network is due to chance alone. (B) Analysis of top function of 472 genes that are significantly down-regulated after SON knockdown using IPA, as described in (A). (C) Representative genes that are down-regulated by SON knockdown. Down-regulated genes which belong to the functional groups for DNA replication/recombination/repair and cell cycle were determined by IPA, and representative genes were listed in the Venn diagram. (D) Decrease in protein levels of TUBG1, TUBGCP2 and AKT1 after SON knockdown. Whole cell lysates from control or SON siRNA-transfected HeLa cells were prepared 3 days after transfection, and immunoblotted for SON, TUBG1 (γ-tubulin), TUBGCP2 (γ-tubulin complex protein 2) and AKT1. TUBA1B (α-tubulin) and TUBGCP4 (γ-tubulin complex protein 4) were also blotted as unaffected controls. (E) Decrease in pericentrin (PCNT) level in the centrosome after SON knockdown detected by immunostaining (red for pericentrin, green for α-tubulin, blue for DNA).

Figure 4. SON is required for efficient intron removal at constitutive splice site on a selective group of genes and binds to the RNA of those genes

(A) Detection of the unspliced form of RNAs in SON-depleted cells by PCR analysis. RNAs were prepared from control or SON siRNA transfected HeLa cells, and RT-PCR analyses were done using primers targeting two neighboring constitutive exons for indicated genes (E, exon; F, forward primer; R, reverse primer). PCRs for TUBA1B and AURKA were done as controls that were not down-regulated by SON siRNA. Arrows with solid lines in black indicate detected bands and dotted lines in gray indicate the expected size of undetected bands. (B) A schematic diagram of exons (E1 – E11) and introns (I1 – I10) of the TUBG1 gene and the primer sets designed for PCR shown in (C). (C) PCR analysis using primer sets shown in (B) to compare splicing efficiencies of TUBG1 splicing junctions in control (ct) and SON siRNA-transfected (s) cells. (D) Interaction of SON with RNA. UV-crosslinking and immunoprecipitation (CLIP) was performed with control IgG or SON antibody and associated RNAs were analyzed by RT-PCR. (E) SON depletion causes impaired splicing of TUBG1 minigene. An intron and the flanking exons were cloned from the TUBG1 gene (exon 7 – 8 region and exon 8 – 9 region) and the TUBA1B gene (exon 2 – 3 region) to downstream of the CMV promoter to make minigene constructs. HeLa cells pre-treated with control or SON siRNA were transfected with the minigenes, and RT-PCR was performed to detect unspliced and spliced RNA.

Figure 5. The C-terminal domain containing the RS domain and the G-patch is necessary for SON’s activity in splicing

(A) Domains and unique amino acid repeats in the SON protein. Full length SON (known as isoform f) is composed of 2,426 amino acids. It contains an RS domain and two RNA-binding motifs at the C-terminus and a putative DNA-binding domain. In addition, SON contains unique amino acid repeats that span most of the N-terminal region. The features of each repeat are presented. (B) Various SON fragments generated for the splicing rescue experiment. siRNA-resistant SON (siRR-SON) was generated by mutagenesis (marked by *), and various deletion mutants were generated from siRR-SON. (C) Splicing rescue by different SON fragments. HeLa cells were transfected with control or SON siRNA, and then transfected with TUBG1 minigene (exon 7 - 8) together with various SON cDNA fragments listed in (B). RT-PCR was performed to detect unspliced and spliced RNA and the photo is the representative result. Relative amount of spliced form was calculated by measuring the density of spliced form (the amount of spliced form in the siRR-SON lane was set as 1). Bars represent mean ± SD from 5 independent experiments (*P < 0.005, **P < 0.03, t-test, when compared to SON siRNA only, the 2nd lane).

Figure 6. SON is required for processing of weak constitutive splice sites and facilitates the interaction of SR proteins with RNA polymerase II

(A) Splice site sequences in primary transcripts of wild type and mutant minigenes. Consensus sequences in splice sites are presented. The wild type and modified 5′ and 3′ splice site sequences between exon 7 and exon 8 of TUBG1, as well as 5′ and 3′ splice site sequences between exon 2 and exon 3 of TUBA1B, are presented. Upper cases indicate exon sequences, lower cases indicate intron sequences. Mutated nucleotides are marked in red with an aster. (B) Various mutant minigene constructs generated for splicing assay. Asters indicate nucleotide changes shown in (A). Intron 7 of TUBG1 was replaced by intron 2 of TUBA1B to generate α2-Intron. (C) SON is required for processing of weak splice sites. Minigenes shown in (B) were transfected into HeLa cells pre-treated with control or SON siRNA. RT-PCR was performed to detect unspliced and spliced RNA. (D) SON knockdown altered SC35 localization, resulting in completely round shaped SC35 dots, and moderately affects snRNP localization, but does not affect localization of the CTD of RNAP II. Cells were immunostained for SC35, 2,2,7-tri-methylguanosine (TMG for snRNA) and the CTD of RNAP II (detected by 8WG16). (E) SON facilitates SC35 interaction with the CTD of RNAP II, U1 snRNP and U2AF₆₅. HEK293 cells expressing V5-tagged SC35 were transfected with control or SON siRNA, and V5-immunoprecipitation was performed to pull down SC35 complex. HEK293 cells without V5-SC35 expression were included as a control. Cell lysates and V5-immunoprecipites were immunoblotted with antibodies indicated. Due to the weak affinity of SON antibody in immunoblot, more concentrated lysates were used to detect SON from the lysates (panel marked with *).

Figure 7. Proposed models for the role of SON as a splicing co-factor

(A) If splice sites are strong (which can be due to optimal sequences in the intron), interactions between pre-mRNA and spliceosome components are strong and stable. Therefore, spliceosome complex is formed for splicing regardless presence or absence of SON. When splice sites are weak (due to suboptimal sequences), spliceosome components form weak and unstable interactions (represented by dotted lines) with pre-mRNA in the absence of SON, while stable interactions can be assured by presence of SON. The process may involve interaction(s) between SON and other critical splicing factors, including SR proteins. (B) A model for the role of SON as a co-factor in efficient transcription-splicing coupling. During transcription in wild type cells (left), SON interacts with DNA, RNA, SR proteins and other early spliceosome components through its DNA-binding domain (DB), RNA-binding motifs (RB) and RS domain (RS), thereby facilitating recruitment of early spliceosome components to the CTD of RNAP II. The long and unique amino acid repeats in SON may help this protein stretch and make contact with multiple components. Such organization and connection by SON may help efficient and immediate spliceosome assembly on the nascent pre-mRNA with a weak splice site, resulting in efficient splicing. In the absence of SON (right), SR proteins and other early spliceosome components are not efficiently recruited to the CTD of RNAP II, and DNA, RNA and proteins are not closely associated with each other. Therefore, co-transcriptional spliceosome assembly is not efficient, and splice site recognition/selection on a weak splice site is not accurate, resulting in delayed or impaired splicing.