LncRNA-dependent nuclear stress bodies promote intron retention through SR protein phosphorylation

Abstract

A number of long noncoding RNAs (lncRNAs) are induced in response to specific stresses to construct membrane-less nuclear bodies; however, their function remains poorly understood. Here, we report the role of nuclear stress bodies (nSBs) formed on highly repetitive satellite III (HSATIII) lncRNAs derived from primate-specific satellite III repeats upon thermal stress exposure. A transcriptomic analysis revealed that depletion of HSATIII lncRNAs, resulting in elimination of nSBs, promoted splicing of 533 retained introns during thermal stress recovery. A HSATIII-Comprehensive identification of RNA-binding proteins by mass spectrometry (ChIRP-MS) analysis identified multiple splicing factors in nSBs, including serine and arginine-rich pre-mRNA splicing factors (SRSFs), the phosphorylation states of which affect splicing patterns. SRSFs are rapidly de-phosphorylated upon thermal stress exposure. During stress recovery, CDC like kinase 1 (CLK1) was recruited to nSBs and accelerated the re-phosphorylation of SRSF9, thereby promoting target intron retention. Our findings suggest that HSATIII-dependent nSBs serve as a conditional platform for phosphorylation of SRSFs by CLK1 to promote the rapid adaptation of gene expression through intron retention following thermal stress exposure.

Keywords: intron retention; noncoding RNA; nuclear stress bodies; phosphorylation; splicing factors.

Figure 1. HSATIII lncRNAs control intron retention of a specific set of genes

1. A

   Outline of the screening for HSATIII‐regulated genes during thermal stress recovery. HeLa cells were transfected with a HSATIII ASO (HSATIII KD) or HSATIII sense oligonucleotide (control), exposed to thermal stress. Nuclear polyA(+) RNAs were analyzed by next‐generation sequencing (NGS). NGS data have been deposited in the DDBJ Sequence Read Archive (DRA) (accession number: DRA007304).

2. B

   HSATIII ASO‐mediated depletion of nSBs. Thermal stress‐exposed HeLa cells (42°C for 2 h and recovery for 1 h at 37°C) were visualized by HSATIII‐FISH and immunofluorescence using an anti‐SAFB antibody or anti‐HNRNPM antibody. The nuclei were stained with DAPI. Scale bar: 10 μm.

3. C

   qRT–PCR validation of HSATIII knockdown. The graph shows the qRT–PCR level of HSATIII RNAs in control and HSATIII knockdown cells under three conditions: 37°C, 42°C for 2 h, and thermal stress followed by recovery at 37°C for 1 h (Recovery). Expression levels were calculated as ratios to GAPDH mRNA and were normalized to the levels in control cells under thermal stress conditions (42°C for 2 h). Data are shown as the mean ± SD (n = 3). HSATIII RNAs ratios (%) are indicated.

4. D

   MA plot (log₂ fold change over the average expression level) of all detected introns in HSATIII KD and control cells (n = 3). The introns with significant changes in their expression levels (adjusted P‐value < 0.01) are shown as green dots. Among these introns, up‐regulated (log₂ fold change > 1) and down‐regulated (log₂ fold change < −1) introns are shown with magenta and light blue dots, respectively. Genes and introns that were experimentally validated in subsequent experiments are indicated as gene symbol‐i (intron) #. Total numbers of introns and genes utilized were 144,769 and 32,623, respectively.

5. E

   The numbers of introns and exons that were affected by HSATIII knockdown (fold change > 2).

6. F, G

   Examples of RNA‐seq read maps of HSATIII target RNAs. Affected introns and exons are indicated by blue (down‐regulated upon HSATIII knockdown) and magenta (up‐regulated) boxes.

7. H

   Cumulative frequency curves of the lengths of adjacent exons of 399 HSATIII‐up‐regulated internal introns. The lengths of whole annotated internal introns are shown as a reference.

Source data are available online for this figure.

Figure EV1. Effect of HSATIII knockdown on other nuclear bodies and gene expression

1. A, B

   Normal assembly of nuclear speckles upon HSATIII knockdown. 16 h after HSATIII ASO transfection, HeLa cells were exposed to thermal stress (42°C for 2 h followed by recovery for 1 h at 37°C) and stained using anti‐SRSF2 antibody with the HSATIII‐FISH probe (A) or anti‐SRSF3 antibody (B). Nuclei were stained with DAPI. Scale bar: 10 μm.

2. C

   Normal assembly of paraspeckles upon HSATIII knockdown. 16 h after HSATIII ASO transfection, HeLa cells were exposed to thermal stress and stained using NEAT1 ASOs and anti‐SFPQ antibody as paraspeckle markers. Scale bar: 10 μm.

3. D, E

   Box plot of total area of nuclear speckles (D) and paraspeckles (E) in each nucleus. Nuclear speckles, paraspeckles, and nuclei areas are defined by binarized images of SRSF2, NEAT1, and DAPI, respectively. Mean is indicated by X (n = 39 (SRSF2, control), 44 (SRSF2, HSATIII KD), 30 (NEAT1, control), 50 (NEAT1, HSATIII KD) nuclei). The first and third quartiles are the ends of the box, the median is indicated with the vertical line in the box, and the minimum and maximum are the ends of the whiskers. The outliers are indicated with open circles. P‐values (Mann–Whitney U‐test) are shown above the graphs.

4. F

   MA plot (log₂ fold change over average expression level) of all detected exons (n = 3). Significant change (green), significant and > 2‐fold increase (magenta), and decrease (light blue). Total numbers of exons and genes utilized are 162,989 and 32,623, respectively.

Source data are available online for this figure.

Figure 2. HSATIII lncRNAs control intron retention by regulating splicing

1. A

   Validation of HSATIII target introns by qRT–PCR. The graphs show the relative amounts of the intron‐retaining (IR) (upper) and spliced (lower) forms in control and HSATIII knockdown cells under three conditions: 37°C (normal), 42°C for 2 h (thermal stress), and thermal stress followed by recovery at 37°C for 1 h. Expression levels were calculated as the ratio of each RNA to GAPDH mRNA and were normalized to the levels in control cells under normal conditions (37°C). Data are shown as the mean ± SD (n = 3); *P < 0.05 (Sidak's multiple comparison test).

2. B

   Nuclear localization of the intron‐retaining RNAs. The relative amounts of intron‐retaining RNAs in the nuclear and cytoplasmic fractions were quantified by qRT–PCR and are represented as the ratio (% of the total). GAPDH mRNA and U1 snRNA were used as cytoplasmic and nuclear controls, respectively. Data are shown as the mean ± SD (n = 3).

3. C

   Overview of the qRT–PCR analysis of newly synthesized RNA within 1 h after thermal stress removal. CHX, cycloheximide; EU, 5‐ethynyl uridine.

4. D, E

   The levels of HSATIII target introns in newly synthesized RNAs within 1 h after thermal stress removal, as determined by qRT–PCR. The graphs show the changes in the expression levels of the intron‐retaining (IR) (D) and spliced (E) forms. Expression levels were calculated as the ratio of each RNA to GAPDH mRNA and were normalized to the level in the control cells. Data are shown as the mean ± SD (n = 3); *P < 0.05 (multiple t‐test modified by Holm–Sidak's method).

Figure 3. The HSATIII lncRNA is necessary and sufficient to promote nuclear intron retention

1. Splicing isoforms of the CLK1 pre‐mRNA. The retained introns are indicated by red lines. An asterisk indicates the position of the premature termination codon in the nonsense‐mediated mRNA decay‐targeted isoform.

2. Time course analysis of the splicing pattern of CLK1 pre‐mRNAs in control and HSATIII knockdown (HSATIII KD) cells by semi‐quantitative RT–PCR. Arrows indicate the positions of PCR primers. The GAPDH mRNA was used as an internal control.

3. Quantification of the data shown in (B). Data are shown as the mean ± SD (n = 3); *P < 0.05 (Sidak's multiple comparison test).

4. Thermal stress‐induced expression of HSATIII in CHO (His9) cells. The cells were visualized by HSATIII‐FISH (green), and the nuclei were stained with DAPI (blue). Scale bar: 10 μm.

5. Ectopic HSATIII‐induced nSB assembly in CHO (His9) cells. The nSBs (yellow arrowheads) were visualized by HSATIII‐FISH and immunofluorescence using an anti‐SRSF1 antibody. The nuclei were stained with DAPI. Scale bar: 10 μm.

6. Time course analysis of the splicing pattern of Chinese hamster Clk1 pre‐mRNAs in control (CHO) and CHO (His9) cells by semi‐quantitative RT–PCR. Arrows indicate the positions of PCR primers. The GAPDH mRNA was used as an internal control.

7. Quantification of the data shown in (F). Data are shown as the mean ± SD (n = 3); *P < 0.05 (Sidak's multiple comparison test).

8. The effect of HSATIII knockdown on pooling of intron‐retaining RNAs in control (CHO) and CHO (His9) cells (42°C for 2 h and recovery for 2 h at 37°C). Arrows indicate the positions of PCR primers.

Source data are available online for this figure.

Figure EV2. Effect of HSATIII knockdown on intron retention of TAF1D and DNAJB9 pre‐mRNAs

1. Splicing isoforms of TAF1D and DNAJB9. The retained introns are indicated by red lines. An asterisk indicates the position of the premature termination codon (PTC) in the NMD‐targeted isoform.

2. Time course analysis of splicing pattern of TAF1D and DNAJB9 pre‐mRNAs in control and HSATIII knockdown (HSATIII KD) cells by semi‐quantitative RT–PCR. Arrows indicate the positions of PCR primers. GAPDH mRNA was used as internal control.

3. Specificity of the HSATIII‐dependent splicing control. Semi‐quantitative RT–PCR validation of the splicing pattern of newly synthesized RNAs within 1 h after thermal stress removal (see also Fig 2C). Arrows indicate the positions of PCR primers. CLK1 and GAPDH pre‐mRNAs were used as positive and internal controls, respectively.

Source data are available online for this figure.

Figure 4. Identification of HSATIII‐interacting proteins by ChIRP

1. A

   Overview of the ChIRP procedure used to identify HSATIII‐interacting proteins in thermal stress‐exposed cells. HSATIII‐interacting proteins were crosslinked and captured with a biotinylated HSATIII ASO. As a negative control, the same procedure was performed using control (37°C) cells and/or a control probe.

2. B

   RT–PCR validation of the specific precipitation of HSATIII by ChIRP. The NEAT1 ncRNA and GAPDH mRNA were used as negative controls. Input: 100%.

3. C

   Silver staining of the coprecipitated proteins. The predicted mobilities of five known nSB proteins are indicated. Input: 0.1%.

4. D

   Western blot analyses of proteins identified by the HSATIII‐ChIRP experiment. Known nSB proteins (SAFB, SRSF1, SRSF7, SRSF9, and HNRNPM) were used as positive controls, and ELAVL1, PSPC1, and α‐tubulin were used as negative controls. SRSF3, SRSF5, TRA2B, and THRAP3 are newly identified nSB components.

5. E–G

   Colocalization of novel nSB proteins with HSATIII. Endogenous TRA2B (E), THRAP3 (F), and FLAG‐tagged PPHLN1 (G) were stained with anti‐TRA2B, anti‐THRAP3, and anti‐FLAG antibodies, respectively. HSATIII was visualized by RNA‐FISH using a dig‐labeled HSATIII ASO, and the nuclei were stained with DAPI. Scale bar: 10 μm.

Source data are available online for this figure.

Figure 5. CLK1 is recruited to nSBs during thermal stress recovery

1. A

   Specific interaction of CLK1 with HSATIII during thermal stress recovery. HSATIII‐ChIRP at various time points during and after thermal stress exposure. Endogenous CLK1, SRSF7, and SRSF9 were detected using anti‐CLK1, anti‐SRSF7, and anti‐SRSF9 antibodies, respectively. HSATIII was detected by ChIRP‐RNA purification, followed by RT–PCR. Input: 0.1% for CLK1, 1% for SRSF7 and SRSF9, and 100% for HSATIII.

2. B

   The domain structures of wild‐type (WT) CLK1, the catalytically inactive mutant (CLK1 KR), and the N‐ and C‐terminal partial fragments (CLK1ΔC and CLK1ΔN).

3. C

   Specific interaction of HSATIII with CLK1 proteins during stress recovery. HSATIII‐ChIRP/Western blotting was performed using HeLa cells expressing FLAG‐CLK1 WT or the KR mutant. FLAG‐tagged CLK1 and SRSF9 were detected by Western blotting using an anti‐FLAG and an anti‐SRSF9 antibody, respectively. HSATIII was detected by ChIRP followed by RT–PCR. Input: 1% for Western blotting, 100% for RT–PCR.

4. D

   Re‐translocation of FLAG‐CLK1 to nSBs after thermal stress removal. HeLa cells were transfected with a FLAG‐CLK1 expression vector, cultured for 16 h, and then exposed to thermal stress (42°C for 2 h) with or without recovery (37°C for 1 h). FLAG‐CLK1 was visualized with an anti‐FLAG antibody. Nuclear stress bodies and nuclei were visualized by HSATIII‐FISH and DAPI staining, respectively. Scale bar: 10 μm.

5. E

   Box plot of colocalization correlation coefficients between HSATIII and FLAG‐CLK1 WT. The first and third quartiles are the ends of the box, the median is indicated by the vertical line in the box, and the minimum and maximum are the ends of the whiskers (n = 20 nuclei for each condition). *P < 0.05 (Mann–Whitney U‐test).

6. F

   Recovery phase‐specific interaction of HSATIII with the CLK1 N‐terminal region. HSATIII‐ChIRP/Western blotting was performed on HeLa cells expressing FLAG‐CLK1ΔC or FLAG‐CLK1ΔN using the same procedure as in (C). Asterisk: an unidentified partial fragment of FLAG‐CLK1ΔN.

7. G, H

   Dependency of SRSF9 on the interaction of CLK1 (G) and CLK1ΔC (H) with HSATIII. HSATIII‐ChIRP was performed using control and SRSF9‐depleted cells expressing FLAG‐CLK1 proteins (42°C for 2 h and recovery for 1 h at 37°C). Input: 1%. Graphs represent coprecipitation ratios of FLAG‐CLK1 proteins to HSATIII in the control and SRSF9‐depleted cells. Data are shown as the mean ± SD (n = 3); *P < 0.05 (Welch's t‐test).

Source data are available online for this figure.

Figure EV3. Recruitment of N‐terminal IDRs of CLK family proteins to nSBs and nuclear speckles

1. A–H

   HeLa cells were transfected with FLAG‐CLK1 expression plasmid, cultured for 16 h, and exposed to thermal stress (42°C for 2 h) followed by recovery (37°C for 1 h). FLAG‐tagged CLK1 or its mutant (see also Fig 5B) was visualized with anti‐FLAG antibody. nSBs (A, E) or nuclear speckles (B, F) and nuclei were visualized by HSATIII‐FISH, IF using the anti‐SRSF2 antibody, and with DAPI, respectively. (Scale bar: 10 μm). Box plots (C, D, G, H) of correlation coefficients between localization of FLAG‐tagged CLK1 proteins and HSATIII (C, G) or SRSF2 (D, H). The first and third quartiles are the ends of the box, the median is indicated with the vertical line in the box, and the minimum and maximum are the ends of the whiskers. The outliers are indicated with open circles. *P < 0.05 (Mann–Whitney U‐test (C, G), Kruskal–Wallis test, followed by Dunn's multiple comparison test (D, H)) (n = 20 nuclei (C, G, and H), n = 15 (37°C), n = 14 (42°C 2 h), and n = 16 nuclei (42°C 2 h followed by recovery for 1 h at 37°C (D)).

2. I

   Western blot of nuclear and cytoplasmic fractions of HeLa cells expressing FLAG‐CLK1 truncated protein. FLAG‐CLK1 proteins were detected using anti‐FLAG antibody. SRSF9 and α‐tubulin were used as nuclear and cytoplasmic fraction controls, respectively. Asterisk; an unidentified partial fragment of FLAG‐CLK1ΔN.

3. J

   Specific interaction of CLK family proteins with HSATIII during stress recovery. HSATIII‐ChIRP was performed in thermal stress‐exposed HeLa cells (42°C for 2 h followed by recovery for 1 h at 37°C) expressing FLAG‐CLK1ΔC, CLK2ΔC, and CLK4ΔC. FLAG‐tagged proteins and endogenous SRSF9, used as a ChIRP control, were detected by Western blotting using anti‐FLAG and anti‐SRSF9 antibody, respectively. Input (1%).

4. K

   Colocalization of FLAG‐CLK2 ΔC and FLAG‐CLK4 ΔC with HSATIII in thermal stress‐exposed HeLa cells (42°C for 2 h followed by recovery for 1 h at 37°C). Scale bar: 10 μm.

Source data are available online for this figure.

Figure 6. The HSATIII lncRNA accelerates CLK1‐dependent re‐phosphorylation of SRSF9

1. Localization of SRSF9 with HSATIII during thermal stress and recovery. Scale bar: 10 μm.

2. Recovery phase‐specific colocalization of CLK1 and SRSF9. HeLa cells were transfected with a FLAG‐CLK1 expression vector, cultured for 16 h, and exposed to thermal stress (42°C for 2 h) followed by recovery (37°C for 1 h). FLAG‐CLK1 and SRSF9 were co‐stained using anti‐FLAG and anti‐SRSF9 antibodies, respectively, and the nuclei were stained with DAPI. Scale bar: 10 μm.

3. Box plot of colocalization correlation coefficients between SRSF9 and FLAG‐CLK1 WT. The first and third quartiles are the ends of the box, the median is indicated with a vertical line in the box, and the minimum and maximum are the ends of the whiskers (n = 20 nuclei for each condition). *P < 0.05 (Mann–Whitney U‐test).

4. Scheme of sampling references for analysis of the phosphorylation states of SRSF9 in (F). The CLK1 inhibitor KH‐CB19 (10 μM) or DMSO as a control was administrated during the recovery period. The numbers correspond to the lane numbers in (F).

5. Conditions tested in the time course analysis of the phosphorylation states of SR proteins. HeLa cells were transfected with a HSATIII ASO or control SO, cultured for 16 h, exposed to thermal stress (42°C for 2 h), and cultured at 37°C for the indicated period.

6. The effect of HSATIII knockdown on time course‐dependent changes in SRSF9 phosphorylation. The phosphorylation states were estimated by Phos‐tag SDS–PAGE, followed by Western blotting using an anti‐SRSF9 antibody. SRSF9 detected by Western blotting of a normal SDS–PAGE gel is shown in the bottom panel. The arrows and arrowheads indicate the normal phosphorylated forms and thermal stress‐induced hypo‐ and de‐phosphorylated forms, respectively.

7. Quantification of the data shown in (F). Graphs show the time course changes in four representative SRSF9 phosphorylated forms in (F). Relative levels of band intensity were calculated as the ratio of each band to α‐tubulin and were normalized to those of control at 37°C (bands (a) and (b)) and 42°C (bands (c) and (d)). Since band (d) partially overlapped with the other band (asterisk in F) at 37°C and recovery 4 h, band (d) at those time points was not measured. Data are shown as the mean ± SD (n = 3); *P < 0.05 (Sidak's multiple comparison test).

8. The re‐phosphorylation status of SRSF9 within nSBs. Also, see Fig EV4G. Phosphorylated (arrows) and hypo‐ and de‐phosphorylated (arrowheads) SRSF9 in the nSB fractions (lanes 4, 6, and 8) were analyzed by Western blotting on Phos‐tag gels. KH‐CB19 (10 μM) or DMSO as a control was administrated during the recovery period. The lower panel shows a shorter exposure image of the phosphorylated forms. The asterisks indicate bands different from band (a) or (d). Input: 5%.

9. Quantification of the data shown in (H). Relative band intensities were calculated as the ratios of each band to total SRSF9 detected by Western blotting of a standard SDS–PAGE and were normalized to those at 42°C. Data are shown as the mean ± SD (n = 4) (Dunnett's multiple comparison test).

Source data are available online for this figure.

Figure EV4. HSATIII‐dependent localization and phosphorylation of nSB‐localizing SRSFs

1. A

   The effect of HSATIII knockdown on SRSF9 localization. HeLa cells were transfected with HSATIII ASO (HSATIII KD) or control ASO (control), cultured for 16 h, and exposed to thermal stress (42°C for 2 h and 37°C for 1 h). The cells were stained by HSATIII‐FISH, IF using SRSF9 antibody, and DAPI (Scale bar: 10 μm).

2. B, C

   The subnuclear distribution of nSBs and nuclear speckles. HeLa cells (42°C for 2 h and 37°C for 1 h) were stained with an anti‐SRSF9 antibody, an anti‐SRSF2 antibody, and a HSATIII‐FISH probe. The nuclei were stained with DAPI. Scale bar: 10 μm.

3. D

   Box plot of colocalization correlation coefficient among SRSF2, SRSF9, and HSATIII in HeLa nuclei (42°C for 2 h and 37°C for 1 h). The first and third quartiles are the ends of the box, the median is indicated with the vertical line in the box, and the minimum and maximum are the ends of the whiskers. The outliers are indicated with open circles. (n = 20 nuclei for each). *P < 0.05 (Kruskal–Wallis test followed by Dunn's multiple comparison test).

4. E

   The effect of HSATIII knockdown on time course‐dependent changes in SRSF1 phosphorylation. The phosphorylation states were estimated by Phos‐tag SDS–PAGE, followed by Western blotting with an anti‐SRSF1 antibody. SRSF1 detected by Western blotting of a normal SDS–PAGE gel is shown in the bottom panel. The arrow and bracket indicate the normal phosphorylated forms and thermal stress‐induced hypo‐ and de‐phosphorylated forms, respectively.

5. F

   Phosphorylation states of SRSFs. SRSFs with longer RS domains were detected by standard SDS–PAGE Western blotting with pan‐phospho SR antibody (1H4) in HSATIII KD and control cells under the different temperature conditions shown above the panel. α‐tubulin is a loading control.

6. G

   Scheme of Phos‐tag Western blotting coupled with HSATIII‐ChIRP used to validate the phosphorylation states of SRSF9 within nSBs.

7. H

   Standard SDS–PAGE of the HSATIII‐ChIRP samples, followed by Western blotting.

8. I

   Phos‐tag SDS–PAGE of the HSATIII‐ChIRP samples, followed by Western blotting. SRSF1 was detected with anti‐SRSF1 antibody. The bottom panel shows the emphasized images. Input (5%). The arrow and bracket indicate the normal phosphorylated forms and thermal stress‐induced hypo‐ and de‐phosphorylated forms, respectively.

9. J

   Phos‐tag Western blotting of SRSF1 and SRSF9 upon phosphatase treatment. HeLa cell proteins were treated with Lambda protein phosphatase for the indicated periods. An arrow, an arrowhead, and asterisks with bracket indicate the original position, completely de‐phosphorylated, and transiently increased partially de‐phosphorylated bands, respectively.

Source data are available online for this figure.

Figure EV5. Intron retention controlled by SRSF9

1. A

   Western blot validation of SRSF knockdown. HeLa cells were transfected with each Stealth siRNA, cultured for 48 h, and analyzed by Western blot using anti‐SRSF1, anti‐SRSF7, and anti‐SRSF9 antibodies. α‐tubulin was detected as a loading control.

2. B

   Normal assembly of nSBs in SRSF7 or SRSF9 knockdown cells. 48 h after siRNA transfection, HeLa cells were exposed to thermal stress and stained by HSATIII‐FISH. Nuclei were stained with DAPI (Scale bar: 10 μm).

3. C

   Quantification of (B). Box plot represents relative HSATIII intensities in nuclei. Mean is indicated by X. The first and third quartiles are the ends of the box, the median is indicated with the vertical line in the box, and the minimum and maximum are the ends of the whiskers. The outliers are indicated with open circles (n = 40 nuclei for each) *P < 0.05 (Kruskal–Wallis test followed by Dunn's multiple comparison test).

4. D

   SRSFs eCLIP tag on the retaining introns and adjacent exons of CLK1 and TAF1D.

5. E

   The amino acid sequence of human SRSF9. In the SRSF9‐de‐phosphorylated mutant, serines in SR and SP dipeptides (indicated in magenta) were replaced by alanines.

6. F

   Replacement of endogenous SRSF9 with FLAG‐SRSF9‐WT and the mutant (FLAG‐SRSF9‐mut). HeLa cells were transfected with control or SRSF9 siRNA (siControl or siSRSF9) in combination with an expression plasmid of siRNA‐resistant FLAG‐SRSF9‐WT or FLAG‐SRSF9‐mut, cultured for 48 h and analyzed by Western blotting. The empty plasmid was used as negative control (empty). Arrowheads and an arrow indicate exogenous and endogenous SRSF9, respectively.

7. G

   Frequency of the SRSF9 eCLIP tag on HSATIII‐up‐regulated introns and adjacent exons. High and low coverage introns in the transcriptome samples (n = 3) in Fig 1B are used as matched controls.

Source data are available online for this figure.

Figure 7. The role of SRSF9 re‐phosphorylation in controlling HSATIII‐dependent intron retention

1. A, B

   Validation of the effect of SRSF9 on retention/splicing of HSATIII target introns by qRT–PCR. The relative expression levels of intron‐retaining (IR) forms (A) and spliced forms (B) of newly transcribed RNAs during the recovery period are shown. See also the scheme in Fig 2C. Expression levels were calculated as the ratio of each RNA to GAPDH mRNA and were normalized to the control. Data are shown as the mean ± SD (n = 3); *P < 0.05 (multiple t‐test modified by Holm–Sidak's method).

2. C, D

   Time course analysis of the splicing patterns of CLK1 and TAF1D intron‐retaining RNAs in control and HSATIII knockdown cells by semi‐quantitative RT–PCR. Arrows indicate the positions of primers for RT–PCR. GAPDH mRNA was used as an internal control. The experiment was performed in triplicate. Quantification of the data shown in (C). Data are shown as the mean ± SD (n = 3); *P < 0.05 (Sidak's multiple comparison test).

3. E

   Scheme for analyzing the effect of CLK1 inhibition on intron retention by newly transcribed RNAs after stress removal. CHX, cycloheximide; EU, 5‐ethynyl uridine.

4. F

   Validation of the requirement of CLK1 catalytic activity for promotion of intron retention by qRT–PCR. Data are shown as the mean ± SD (n = 3). P‐values (multiple t‐test modified by Holm–Sidak's method) are shown.

5. G

   The diagrams show the domain structures of wild‐type (WT) and mutant (mut) SRSF9, in which all SR/SP dipeptides were replaced by AR/AP. FLAG‐SRSF9‐WT and FLAG‐SRSF9‐mut were extracted from transfected HeLa cells under normal (37°C) and thermal stress (42°C for 2 h) conditions, separated in a Phos‐tag gel, and detected by Western blot using an anti‐FLAG antibody. The arrowhead indicates the thermal stress‐induced de‐phosphorylated form of FLAG‐SRSF9‐WT.

6. H, I

   The role of phosphorylation of SRSF9 on splicing of the TAF1D (H) and DNAJB9 (I) reporters. Each splicing reporter plasmid was transfected into HeLa cells along with a siRNA (control or SRSF9‐specific) and a siRNA‐resistant SRSF9‐WT or SRSF9‐mut expression plasmid (or pEGFP‐C1 vector as the empty vehicle (−)) 48 h before the assay. Semi‐quantitative RT–PCR analyses revealed the splicing patterns of the reporter RNAs purified from the thermal stress‐exposed cells (42°C for 2 h and 37°C for 1 h). The positions of the PCR primers are indicated by arrows. Graphs represent the relative changes in the spliced/unspliced ratios, normalized to the control sample with control siRNA and empty plasmid (−). Data are shown as the mean ± SD (n = 3); *P < 0.05 (Dunnett's multiple comparison test).

7. J

   A model for the function of nSBs in splicing control. Thermal stress‐induced HSATIII integrates subsets of de‐phosphorylated SRSFs and SR‐related proteins. After stress removal, CLK1 relocates to nSBs to promote rapid re‐phosphorylation of the sequestrated SRSFs, especially SRSF9. Re‐phosphorylated SRSF9 suppresses splicing of target introns to accumulate intron‐retaining RNAs in the nucleus.

Source data are available online for this figure.