Skip to main content
PMC home page
eLife logoLink to eLife
. 2016 Mar 25;5:e10288. doi: 10.7554/eLife.10288

An extensive program of periodic alternative splicing linked to cell cycle progression

Daniel Dominguez 1,2, Yi-Hsuan Tsai 1,3, Robert Weatheritt 4, Yang Wang 1,2, Benjamin J Blencowe 4,*, Zefeng Wang 1,5,*
Editor: Gene Yeo6
PMCID: PMC4884079  PMID: 27015110

Abstract

Progression through the mitotic cell cycle requires periodic regulation of gene function at the levels of transcription, translation, protein-protein interactions, post-translational modification and degradation. However, the role of alternative splicing (AS) in the temporal control of cell cycle is not well understood. By sequencing the human transcriptome through two continuous cell cycles, we identify ~1300 genes with cell cycle-dependent AS changes. These genes are significantly enriched in functions linked to cell cycle control, yet they do not significantly overlap genes subject to periodic changes in steady-state transcript levels. Many of the periodically spliced genes are controlled by the SR protein kinase CLK1, whose level undergoes cell cycle-dependent fluctuations via an auto-inhibitory circuit. Disruption of CLK1 causes pleiotropic cell cycle defects and loss of proliferation, whereas CLK1 over-expression is associated with various cancers. These results thus reveal a large program of CLK1-regulated periodic AS intimately associated with cell cycle control.

DOI: http://dx.doi.org/10.7554/eLife.10288.001

Research Organism: Human

eLife digest

Mitosis is a key step in the normal life cycle of a cell, during which one cell divides into two new cells. As a cell progresses through the cell cycle, it must carefully regulate its gene activity to switch particular genes on or off at specific moments. When a gene is activated its sequence is first copied into a temporary molecule called a transcript. These transcripts are then edited to form templates to build proteins. One way that a transcript can be edited is via a process called alternative splicing, in which different pieces of the transcript are cut and pasted together to form different versions of the final template. This allows different instructions to be obtained from a single gene, introducing an added layer of biological complexity. However, the role of alternative splicing in the timing of key events of the cell life cycle is not well understood.

Dominguez et al. have now looked for the genes that undergo alternative splicing during the cell cycle. The sequences of gene transcripts produced within human cells were collected while the cells went through two rounds of division. This approach revealed that around 1,300 genes are spliced in different ways at different stages of each cell cycle. Many of these genes were known to play roles in controlling the cell’s life cycle, but few of the genes showed large changes in the amount of total transcript that is generated over time.

Dominguez et al. also showed that an enzyme called CLK1 influences about half of the 1,300 periodically spliced genes during the cell cycle. The production of CLK1 is itself carefully controlled throughout the cell cycle, and the enzyme’s activity prevents its own overproduction. Further experiments showed that blocking CLK1’s activity while a cell is replicating its DNA halts the cell cycle, but blocking this enzyme’s activity after the cell had replicated its DNA did not. Given this pivotal role in the cell cycle, Dominguez et al. also examined the role of CLK1 in cancer cells and found that high levels of CLK1 in tumours were linked to lower survival rates. These findings indicate that CLK1 warrants further investigation, particularly in relation to its role in cancer.

DOI: http://dx.doi.org/10.7554/eLife.10288.002

Introduction

Alternative splicing (AS) is a critical step of gene regulation that greatly expands proteomic diversity. Nearly all (>90%) human genes undergo AS and a substantial fraction of the resulting isoforms are thought to have distinct functions (Pan et al., 2008; Wang et al., 2008). AS is tightly controlled, and its mis-regulation is a common cause of human diseases (Wang and Cooper, 2007). Generally, AS is regulated by cis-acting splicing regulatory elements that recruit trans-acting splicing factors to promote or inhibit splicing (Matera and Wang, 2014; Wang and Burge, 2008). Alterations in splicing factor expression have been observed in many cancers and are thought to activate cancer-specific splicing programs that control cell cycle progression, cellular proliferation and migration (David and Manley, 2010; Oltean and Bates, 2014). Consistent with these findings, several splicing factors function as oncogenes or tumor suppressors (Karni et al., 2007; Wang et al., 2014), and cancer-specific splicing alterations often affect genes that function in cell cycle control (Tsai et al., 2015).

Progression through the mitotic cell cycle requires periodic regulation of gene function that is primarily achieved through coordination of protein levels with specific cell cycle stages (Harashima et al., 2013; Vermeulen et al., 2003). This temporal coordination enables timely control of molecular events that ensure accurate chromatin duplication and daughter cell segregation. Periodic gene function is conventionally thought to be achieved through stage-dependent gene transcription (Bertoli et al., 2013), translation (Grabek et al., 2015), protein-protein interactions (Satyanarayana and Kaldis, 2009), post-translational protein modifications, and ubiquitin-dependent protein degradation (Mocciaro and Rape, 2012). Although AS is one of the most widespread mechanisms involved in gene regulation, the relationship between the global coordination of AS and the cell cycle has not been investigated.

Major families of splicing factors include the Serine-Arginine rich proteins (SR) proteins and the heterogeneous nuclear ribonucleoproteins (hnRNPs), whose levels and activities vary across cell types. SR proteins generally contain one or two RNA recognition motifs (RRMs) and a domain rich in alternating Arg and Ser residues (RS domain). Generally, RRM domains confer RNA binding specificity while the RS domain mediates protein-protein and protein-RNA interactions to affect splicing (Long and Caceres, 2009; Zhou and Fu, 2013). Post-translational modifications of SR proteins, most notably phosphorylation, modulate their splicing regulatory capacity by altering protein localization, stability or activity (Gui et al., 1994; Lai et al., 2003; Prasad et al., 1999; Shin and Manley, 2002). Dynamic changes in SR protein phosphorylation have been detected after DNA damage (Edmond et al., 2011; Leva et al., 2012) and during the cell cycle (Gui et al., 1994; Shin and Manley, 2002), suggesting that regulation of AS may have important roles in cell cycle control. However, the functional consequences of SR protein (de)phosphorylation during the cell cycle are largely unclear.

Through a global-scale analysis of the human transcriptome at single-nucleotide resolution through two continuous cell cycles, we have identified widespread periodic changes in AS that are coordinated with specific stages of the cell cycle. These periodic AS events belong to a set of genes that is largely separate from the set of genes periodically regulated during the cell cycle at the transcript level, yet the AS regulated set is significantly enriched in cell cycle- associated functions. We further demonstrate that a significant fraction of the periodic AS events is regulated by the SR protein kinase, CLK1, and that CLK itself is also subject to cell cycle-dependent regulation. Moreover, inhibition or depletion of CLK1 causes pleiotropic defects in mitosis that lead to cell death or G1/S arrest, suggesting that the temporal regulation of splicing by CLK1 is critical for cell cycle progression. The discovery of periodic AS thus reveals a widespread yet previously underappreciated mechanism for the regulation of gene function during the cell cycle.

Results

Alternative splicing is coordinated with different cell cycle phases

To systematically investigate the regulation of AS during the cell cycle, we performed an RNA-Seq analysis of synchronously dividing cells using a total of 2.3 billion reads generated across all stages (G1, S, G2 and M) of two complete rounds of the cell cycle (Figure 1—figure supplement 1A). To maximize the detection of regulated AS events, we used the complementary analysis pipelines, MISO and VAST-TOOLS (Katz et al., 2010; Irimia et al., 2014). These pipelines have different detection specificities and employ partially overlapping reference sets of annotated AS events, and therefore afford a more comprehensive analysis when employed together. Both pipelines were used to determine PSI (the percent of transcript with an exon spliced in) and PIR (the percent of transcripts with an intron retained). Alternative exons detected by both pipelines had highly correlated PSI values (Figure 1—figure supplement 1G; see below). Consistent with previous results (Bar-Joseph et al., 2008; Whitfield et al., 2002), transcripts from approximately 14.2% (1182) of expressed genes displayed periodic differences in steady-state levels between two or more cell cycle stages (see below). Remarkably, 15.6% (1293) of expressed genes also contained 1747 periodically-regulated AS events, among a total of ~40,000 detected splicing events (FDR < 2.5%; Figure 1A and Figure 1—figure supplement 1B,D).

Figure 1. Global detection of periodic cell cycle-dependent alternative splicing.

(A) Heat map representation of periodically spliced events. Row-normalized relative PSI values are shown. Diagram below indicates cell cycle phase. (B) Overlap between periodically spliced genes and periodically expressed genes detected by RNA-Seq. (C) Heat map representation of enriched Gene Ontology terms shown as log (p-value). Three gene sets were analyzed separately: all genes with periodic AS, genes with periodic AS only, and genes with both periodic AS and periodic expression. (D) Real-time quantitative PCR analysis of periodic retained introns and total mRNAs for three selected genes. Cells were synchronized by double thymidine block and samples were collected 0, 3, 6, 9, 12 and 15 hr post release. Errors bars represent standard deviation of the mean. Diagram below indicates cell cycle stage. (E) Schematic representation of AURKB AS pattern. Line graph showing the relationship between intron retention and mRNA levels for the AURKB gene across the cell cycle. Percent intron retention (solid red line) across cell cycle was used to determine the fraction of total mRNAs (solid blue line) not containing an intron, i.e. ‘corrected’ mRNA levels (dashed blue line).

DOI: http://dx.doi.org/10.7554/eLife.10288.003

Figure 1.

Figure 1—figure supplement 1. Identification of periodic AS by multiple analysis pipelines.

Figure 1—figure supplement 1.

(A) Number of sequencing reads per sample (top). RNA-Seq reads and periodic seeds used for the identification of all periodically expressed and spliced genes (bottom, see methods). (B,C) Dot plot of periodic score and false discovery rate (FDR) for each exon analyzed by the MISO and VAST-TOOLS analysis pipelines. Dashed lines show FDR and periodic score cutoff (see methods). (D) Heat map representation of periodically-spliced events identified by the VAST-TOOLS pipeline. Data are row-normalized. Diagram below indicates cell cycle stage. (E) Bar graphs showing the number of periodic AS events identified separated by event type and shown as a fraction of total events identified (SE: skipped exon, RI: retained intron, A3: alternative 3’splice site, A5: alternative 5’splice site). MISO analysis (left panel in blue) and VAST-TOOLS analysis (right panel in red). (F) Venn diagram representation of the overlap between periodic AS identified by VAST-TOOLS and periodically expressed mRNAs (top). Venn diagram representation of the overlap between periodic AS events as identified by both VAST-TOOLS and MISO (bottom, see methods). (G) Spearman’s rank correlation analysis of each cell cycle time point according to commonly detected alternative exons by MISO and VAST-TOOLS. Spearman’s rho values are shown in heat map.
DOI: http://dx.doi.org/10.7554/eLife.10288.004
Open in a new tab

Importantly, as has been observed previously for AS regulatory networks (Pan et al., 2004), the majority of genes with periodic AS events did not overlap those with periodic steady-state changes in mRNA expression. This indicates that genes with periodic changes in AS and transcript levels are largely independently regulated during the cell cycle (Figure 1B). Further supporting this conclusion, we did not observe a significant correlation (positive or negative) between exon PSI values and mRNA expression levels for genes with both periodic expression and periodic exon skipping (data not shown). A gene ontology (GO) analysis reveals that genes with periodic AS, like those with periodic transcript level changes (Bar-Joseph et al., 2008; Whitfield et al., 2002), are significantly enriched in cell cycle-related functional categories, including M-phase, nuclear division and DNA metabolic process (Figure 1C; adjusted p<0.05 for all listed categories, FDR<10%) (Supplementary file 1). Similar GO enrichment results were observed when removing the relatively small fraction (10%) of periodically spliced genes that also display significant mRNA expression changes across the cell cycle (Figure 1C). These results thus reveal that numerous genes not previously linked to the cell cycle, as well as previously defined cell cycle-associated genes thought to be constantly expressed across the cell cycle, are in fact subject to periodic regulation at the level of AS (Supplementary file 1 for a full list).

Among the different classes of AS analyzed (cassette exons, alternative 5'/3' splice sites and intron retention [IR]), periodically regulated IR events were over-represented (relative to the background frequency of annotated IR events) by ~2.2 fold whereas periodically regulated cassette exons, represent the next most frequent periodic class of AS (p=2.2×10-16, Fisher’s exact test, Figure 1—figure supplement 1E). Quantitative RT-PCR assays across different cell cycle stages validated periodic IR events detected by RNA-Seq (Figure 1D). Interestingly, one of these IR events is in transcripts encoding aurora kinase B (AURKB), a critical mitotic factor regulated at the levels of transcription, protein localization, phosphorylation and ubiquitination (Carmena et al., 2012; Lens et al., 2010). The AURKB retained intron is predicted to introduce a premature termination codon that elicits mRNA degradation through nonsense mediated decay, and is thus expected to result in reduced levels of AURKB protein. The splicing of the retained intron lags behind changes in the total AURKB mRNA expression (Figure 1E). We computationally corrected levels of fully spliced, protein coding AURKB mRNA by taking into account the fraction of intron-retaining (i.e. non-productive) transcripts across the cell cycle stages (Figure 1E). The expression curve for corrected AURKB mRNA levels is substantially different from total AURKB transcript levels, with a shifted peak coinciding with mitosis. Periodically-regulated IR events detected in other genes, including those with known cell cycle functions such as HMG20B and RAD52, are similarly expected to affect the cell cycle timing of mRNA expression (Figure 1A,D). Collectively, these results provide evidence that the temporal control of retained intron AS provides an important mechanism for establishing the timing of expression of AURKB mRNA and protein, as well as of the timing of expression of additional genes during the cell cycle.

The SR protein kinase CLK1 fluctuates during the cell cycle

Alternative splicing is generally regulated by the concerted action of multiple cis-elements that recruit cognate splicing factors. Consistently, analysis of our RNA-seq data revealed 96 RNA binding proteins (RBPs) with periodic mRNA expression, including RS domain-containing factors like SRSF2, SRSF8, TRA2A and SRSF6 (Figure 2—figure supplement 1A). These 96 RBPs were significantly enriched in the GO term 'splicing regulation' (adjusted p=10-4, Figure 2—figure supplement 1B), indicating that periodic AS is likely controlled by multiple RBPs. Correlations between these RBPs and periodic splicing events were also identified (Figure 2—figure supplement 1C). For example, SRSF2 expression is significantly correlated with the splicing pattern of a retained intron in the SRSF2 transcript. Further supporting a role for these RBPs in controlling periodic splicing was the identification of RNA motifs bound by a subset of periodically expressed RBPs (Figure 2—figure supplement 1D). To further examine periodic RBP regulation during cell cycle, we measured the abundance of known splicing regulatory proteins at different stages of the cell cycle by immunoblotting (Figure 2A). Among the proteins analyzed, CDC-like kinase 1 (CLK1), an important regulator of the Ser/Arg (SR) repeat family of splicing regulators, displayed the strongest cyclic expression peaking at the G2/M phase (Figure 2A,B), consistent with the results of a recent mass-spectrometry-based screen for cycling proteins (Ly et al., 2014). CLK1 is one of four human CLK paralogs (CLK1-4) and is known to regulate AS via altering the phosphorylation status of multiple SR proteins (Duncan et al., 1997; Jiang et al., 2009; Ninomiya et al., 2011; Prasad et al., 1999). Notably, the levels of other detectable CLK paralogs, as well as members of another SR protein kinase, SRPK1, did not change significantly at the level of RNA and/or protein during the cell cycle (Figure 2B and Supplementary file 1).

Figure 2. Cell cycle-dependent regulation of CLK1.

(A) Immunoblot analysis of proteins involved in splicing regulation in synchronized HeLa cells after release from double thymidine block. (B) Immunoblot analysis of selected proteins in asynchronous HeLa cells or cells arrested at different cell cycle stages. Stably expressed exogenous CLK1 levels were also assessed during the cell cycle (bottom panel). (C) Immunoblot of endogenous CLK1 (top) and exogenously-expressed wild type (CLK1wt) or kinase catalytically inactive (CLK1KD) proteins (bottom) upon treatment with 10 µM TG003. (D) Co-expression of CLK1WT and CLK1KD at different ratios. (E) Immunoprecipitation of CLK1 proteins co-expressed with myc-ubiquitin. Cells were treated with 10 μM TG003 and 10 μM MG132 prior to sample collection. (F) Immunoblot analysis of lysates from cells synchronized upon early S phase (double thymidine) release with or without TG003 treatment.

DOI: http://dx.doi.org/10.7554/eLife.10288.005

Figure 2.

Figure 2—figure supplement 1. Periodic expression of RBPs.

Figure 2—figure supplement 1.

(A) Heat map representation of RNA-bound proteins (RBPs) with periodic expression. Row-normalized FPKM levels are shown. (B) GO analyses for the functional enrichment in the periodic RBPs. (C) Number of periodic AS events that significantly correlate (Spearman’s Rho > |.75|, p<.05) with the expression pattern of each RBP during cell cycle. Expression pattern of known two known splicing factors, SRSF2 and ESRP2, is shown in inset. (D) Average PSI values of periodic that peak at either G1 (red line) or M phase (blue line) (top panel). k-mer enrichment in periodic exons as judged by Z score and separated by cell cycle phase (y-axis = G1-S and x-axis = G2-M) (bottom panel).
DOI: http://dx.doi.org/10.7554/eLife.10288.006
Figure 2—figure supplement 2. Regulation of CLK1 proteins levels during the cell cycle is degradation-dependent.

Figure 2—figure supplement 2.

(A) CLK1 mRNAs as measured in synchronized cells by RNA-Seq (left) or in cells arrested at each cell cycle phase, followed by quantitative RT-PCR (right). (B) Diagram depicting alternative the splicing pattern for CLK1 pre-mRNA (left). The short form represents skipping of exon 4 that introduces a premature stop codon and is targeted by the non-sense mediated decay pathway, while long form represents the full-length active isoform. Right panels show levels of CLK1 variants, as measured by semi-quantitative RT-PCR with primers that simultaneously detect both forms, in cells upon early S phase release or in cells arrested at specific stages. (C) Protein stability of CLK1 is affected by its activity. Cyclohexamide chase experiment was used to measure the stability of CLK1 in cells expressing either CLK1wt or the catalytic mutant CLK1KD. MG132 was used to block proteasomal degradation. (D) Immunoprecipitation of Flag-CLK1 from cells after 3 hrs of MG132 treatment and subsequent detection of polyubiquitination by immonoblotting. The arrow indicates the expected position of unmodified CLK1. Bottom panel shows FLAG-CLK1 protein. (E) HeLa cells stably expressing Flag-CLK1 were synchronized by double thymidine block in the presence or absence of TG003. Phosphorylated histone 3B was detected as cell cycle marker.
DOI: http://dx.doi.org/10.7554/eLife.10288.007
Open in a new tab

Given that both CLK1 protein levels and known CLK1 substrates are periodically expressed, we decided to further investigate the role of CLK1 in the context of cell cycle. The levels of total CLK1 mRNA, as well as the levels of specific CLK1 splice variants, did not change significantly during the cell cycle (Figure 2—figure supplement 2A,B) indicating that periodic expression of CLK1 is controlled at the level of protein translation and/or turnover. Consistent with this, an exogenously expressed CLK1 protein displayed cell cycle-dependent fluctuations similar to those observed for endogenous CLK1 protein (Figure 2B). Moreover, CLK1 was rapidly degraded upon inhibition of translation by cycloheximide, and this effect was reversed by co-treatment with the proteasome inhibitor MG132 (Figure 2—figure supplement 2C). Additionally, polyubiquitination of Flag-tagged CLK1 was detected following immunoprecipitation with anti-Flag antibody from cells treated with MG132 (Figure 2—figure supplement 2D). These data suggest that the levels of CLK1 protein are controlled by ubiquitin-mediated degradation in a cell cycle-dependent manner.

Periodically regulated protein levels are often controlled through negative feedback circuits involving auto-regulatory loops. CLK1 has been reported to auto-phosphorylate on several residues (Ben-David et al., 1991). To investigate whether auto-phosphorylation of CLK1 affects its periodic regulation, we tested whether blocking its kinase activity affects its stability. Inhibition of CLK1 kinase activity using a selective inhibitor, TG003 (Muraki et al., 2004), markedly stabilizes both endogenous and exogenously expressed CLK1 proteins (Figure 2C). Moreover, activity-dependent destabilization of CLK1 was observed with a wild type (WT) protein, but not with a catalytically inactive (KD) mutant (Figure 2C, left panel). We further observe that WT CLK1 is rapidly degraded upon cycloheximide treatment, whereas the KD mutant is more stable (Figure 2—figure supplement 2C). We next tested whether CLK1 activity is sufficient to trigger its own degradation by co-expressing KD CLK1 with increasing amounts of WT CLK1. As expected, increasing amounts of WT CLK1 reduces levels of KD CLK1 (Figure 2D). Consistent with these results, WT CLK1 is more highly polyubiquitinated compared to the KD mutant (Figure 2E, compare lanes 2 to 4), and treatment with TG003 reduces polyubiquitination levels (Figure 2E, lane 2 vs. 3 and lane 4 vs. 5). Decreased polyubiquitination of WT CLK1 is more prevalent than is apparent upon TG003 treatment, as CLK1 is stabilized by TG003 inhibition and thus more total Flag-CLK1 is immunoprecipitated (Figure 2E). To further examine whether this auto-feedback loop is required for changes in CLK1 protein levels during the cell cycle, we treated synchronized cells with TG003 (or DMSO as a control) and measured CLK1 protein levels. We observed that CLK1 inhibition prevents its turnover after the G2/M phase for both endogenous and exogenously expressed kinases (Figure 2F and Figure 2—figure supplement 2E). Taken together, these results provide strong evidence that CLK1 protein levels are controlled by ubiquitin-mediated proteolysis in a cell cycle stage-specific manner, and that an activity-dependent negative feedback loop is required for this periodic regulation. These results further suggest that changes in the levels of CLK1 could account for many of the periodically regulated AS transitions we have detected during the cell cycle.

CLK1 regulates AS events in genes with critical roles in cell cycle control

RNA-Seq analysis of cells treated with TG003 revealed 892 AS events (in 665 genes) that significantly change after CLK1 inhibition (Figure 3A), including known CLK1-regulated splicing events (e.g. exon 4 of CLK1 [Duncan et al., 1997]). It is worth noting that TG003 can also inhibit CLK4 (although to a lesser extent than inhibition of CLK1). However, RNAi of CLK1 is sufficient to recapitulate the phenotype of CLK1/4 inhibition (see below and Figure 4) (Fedorov et al., 2011; Muraki et al., 2004). Intron retention and cassette exons are the most overrepresented types of AS affected by TG003 (Figure 3A). Most (70%) of the CLK1-regulated exons display increased skipping upon CLK1 inhibition, whereas 87% of CLK1-regulated introns show increased retention (Figure 3—figure supplement 1B and C), consistent with a recently reported role for CLK1 in the regulation of retained introns (Boutz et al., 2015). Of nine analyzed TG003-affected AS events detected by RNA-Seq analysis, all were validated by semi-quantitative RT-PCR assays (Figure 3B). These observations indicate that CLK1 inhibition mainly suppresses splicing, consistent with a general requirement for phosphorylation of SR proteins to promote splicing activity (Irimia et al., 2014; Prasad et al., 1999; Tsai et al., 2015). Importantly, there is a significant overlap between genes with cell cycle periodic AS events (Figure 1) and those with CLK1-regulated AS events, involving 156 genes (p=8.5×10-10, hyper-geometric test). In contrast, consistent with the results in Figure 1, we do not observe a significant overlap between genes containing CLK1-regulated AS events and periodically expressed genes. These results thus support a widespread and rapidly acting role for CLK1 in controlling cell cycle-regulated AS. Indeed, CLK1 inhibition induces rapid (within 3–6 hr) changes in AS among several analyzed cases (Figure 3—figure supplement 1D).

Figure 3. CLK1 regulates a network of genes that control cell cycle progression.

(A) Identification of endogenous CLK1 targets by RNA-Seq. Numbers of different AS types affected by treatment with the CLK1 inhibitor TG003 (left graph). SE, skipped exon; RI, retained intron; A3E, alternative 3’ exon; A5E, alternative 5’ exon. Fraction of total analyzed events that were affected by TG003 treatment (right graph). (B) Validation of TG003-responsive AS events by semi-quantitative RT-PCR. The bar graph shows the max-delta PSI for each AS event tested in a 24-hr time course of inhibition with 20μM TG003. (C) Representation of cell cycle control genes with CLK1-dependent AS events, organized by cell cycle phase and function. (D) Schematic representation of CHEK2 alternative splicing, showing that exon 9 encodes a region overlapping the kinase domain (upper panel). Semi-quantitative RT-PCR assessment of CHEK2 isoforms after treatment with TG003 or over-expression of the indicated factors (lower left panel). RNAi of SRSF1 in cells and subsequent analysis of CHEK2 splicing by semi-quantitative RT-PCR (lower right panel). PSI values are shown below gel. (E) Normalized CENPE total mRNA expression during an unperturbed cell cycle (triangle denotes mitosis, left panel) and diagram of CENPE splicing (right). TG003-treatment of HeLa cells released from G1/S arrest followed by semi-quantitative RT-PCR analysis of CENPE isoforms (bar graph). (F) Schematic of RNA-Seq analysis of CLK1 inhibition during cell cycle (left). AS events that were identified as being differentially regulated between G1 and G2 phase (top bar of bar graph) and number of events that were blocked by the indicated conditions (bottom 4 bars).

DOI: http://dx.doi.org/10.7554/eLife.10288.008

Figure 3.

Figure 3—figure supplement 1. CLK1 regulates a network of genes that control cell cycle progression.

Figure 3—figure supplement 1.

(A) Identification of endogenous CLK1 targets by RNA-Seq. Cells were treated with 10 μM TG003 for 18 hr and poly-A+ RNA was sequenced. Genes significantly up-regulated or down-regulated upon CLK1 inhibition (log(FPKM) are plotted, all plotted genes meet p<10-7). (B) Scatter plot representation of cassette exon inclusion after TG003 treatment by MISO analysis. Log Bayes factor values are shown on y-axis and delta PSI on x-axis. (C) Scatter plot representation of retained intron levels after TG003 treatment. Log Bayes factor values are shown on y-axis and delta PSI on x-axis. (C) Time course experiment after TG003 treatment. Cells were collected at the indicated time points and the absolute delta PSI is plotted. (E) Immonoblot analysis of reciprocal co-immunoprecipation experiments between CHEK2 isoforms as indicated.
DOI: http://dx.doi.org/10.7554/eLife.10288.009
Open in a new tab

Figure 4. CLK1 is required for cell cycle progression and proliferation.

(A) Immunoblot analysis of CLK1 proteins after stable shRNA knockdown in HeLa cells. Bottom, DNA content as measured by propidium iodide staining following flow cytometry. (B) Immunofluorescence microscopy of A549 cells depleted of CLK1 by shRNA (top row), cells treated with 10 µM TG003 for 12 hr (middle row), and a control treatment with DMSO (bottom row); green: tubulin, red: emerin (nuclear envelope), and blue: DAPI. Scale bar 10 µm. Right bar graph shows the quantification of multinucleated cells. p values determined using Student’s t-test. (C) Static frames from a live-cell high-content imaging movie of HeLa cells expressing Histone H2B-GFP and treated with TG003 (top panel). Time after start of the experiment is indicated; EP, end point (~960 min). TG003 treated cells with apparent cell division defects (indicated by arrowheads in the bottom field) are shown in two independent fields. (D) Synchronized HeLa cells were treated with 20 µM TG003 at the indicated time points (0, 5, and 10 hr) and analyzed by propidium iodide staining and flow cytometry to measure DNA content. Percent of 2N (lower bar graph) and 4N (upper bar graph) cells were quantified at each time point as indicated in the treatment scheme (top). (E) Colony formation assay of HeLa cells depleted of CLK1 by shRNA, or continuously treated with TG003 or KHCB-19 at the indicated concentrations. (F) Box plot representation of CLK1 mRNA expression levels in paired normal and tumorous kidney tissue. 72 cases were analyzed. (G) Kaplan-Meier plot showing survival differences between patients with kidney tumors with high CLK1 (red, upper quartile) or reduced CLK1 (blue, lower three quartiles) expression. (H) Number of cancer-associated AS events that are also regulated by CLK1 in different tumor types. BRCA, Breast invasive carcinoma; COAD, Colorectal adenocarcinoma; KIRC, Kidney renal clear cell carcinoma; LUAD, Lung adenocarcinoma; LUSC, Lung squamous cell carcinoma; LIHC, liver hepatocellular carcinoma.

DOI: http://dx.doi.org/10.7554/eLife.10288.010

Figure 4.

Figure 4—figure supplement 1. Loss of CLK1 results in cell cycle defects in multiple cell types.

Figure 4—figure supplement 1.

(A) Cell cycle composition as measured by propidium iodide staining of DNA and flow cytometry analysis of cells that have been depleted of CLK1 by the indicated shRNAs. (B) Representative histograms of propidium iodide stained H157cells to determine cell cycle defect after RNAi of CLK1. (C) HeLa cells treated with TG003 also have defective cell division. Immunofluorescence microscopy was used to detect multi-nucleated cells, and a representative field is shown (green:tubulin, red:emerin, blue:DAPI). (D) Representative histograms of DNA content in synchronous HeLa cells treated with TG003. The time after early S phase release (when TG003 was added) is indicated, These data are associated with Figure 4 of the main text. (E) Representative image from anchorage-independent growth assays (soft agar assay) of HeLa cells after depletion of CLK1. (F) Relative mRNA expression levels of CENPE and HMMR during cell cycle. Data obtained from RNA-Seq analysis. Dashed line represents 6 hr after release from G1/S.
DOI: http://dx.doi.org/10.7554/eLife.10288.011
Figure 4—figure supplement 2. CLK1 mis-regulation in human cancer.

Figure 4—figure supplement 2.

(A) Overlap of AS events that were altered in kidney cancers (as compared to normal kidney samples) and AS that was altered in asynchronous cells treated with TG003 (left panel). Pie chart denoting if AS in cancer occurred in the expected direction, that is, normal kidney resembled TG003 treatment while tumor kidney resembled untreated cells (see methods). (B) Boxplot representation of CLK1, CLK2, CLK3 and CLK4 mRNA levels in 72 paired normal vs. cancer kidney cancer (cRCC) samples. Kolmogorov-Smirnov test significance for each factors is as follows CLK1: p=3 × 10-5, CLK2: p=1.2 × 10-7, CLK3: p=2.8 × 10-12, CLK4: p=2.2 × 10-16. (C) Kaplan-Meier plot of kidney (cRCC) patients with tumors expressing high CLK4 (red, upper quartile) vs. normal CLK4 (blue, 1–3 quartile). (D) PARD3 exon (chr10:34661426–34661464) PSI levels in five cancer cases are shown as an example of TG003-sensitive AS which is altered between normal and cancer tissues.
DOI: http://dx.doi.org/10.7554/eLife.10288.012
Open in a new tab

Supporting an important role for CLK1 in cell cycle progression, genes whose AS levels are affected by CLK1 inhibition are significantly enriched in the GO terms cell cycle phase, M-phase, DNA metabolic processes, nuclear division, DNA damage response, and cytokinesis (adjusted p<0.05 and FDR <20% for all listed GO terms; full list in Supplementary file 2). The affected genes function at various stages of cell cycle including the G1/S transition (Figure 3C). Mitotic processes were, however, associated with the largest number of CLK1-target genes with AS changes and included examples that function in centriole duplication (CEP70, CEP120, CEP290, CEP68, CDK5RAP2), metaphase and anaphase (e.g. CENPK, CENPE, CENPN), and cytokinesis (e.g. SEPT2, SEPT10, ANLN) (additional examples in Figure 3C).

To further investigate the functional consequence of CLK1-dependent AS, we selected two examples in genes that have important roles in the cell cycle: checkpoint kinase 2 (CHEK2), a tumor suppressor that controls the cellular response to DNA damage and cell cycle entry (Paronetto et al., 2011; Staalesen et al., 2004), and centromere-associated protein E (CENPE), a kinetochore-associated motor protein that functions in chromosome alignment and segregation during mitosis (Kim et al., 2008). We detected a TG003 dose-dependent increase in CHEK2 exon 9 inclusion, whereas overexpression of WT CLK1 induced exon 9 skipping, an event that removes the CHEK2 kinase domain (Figure 3D). Expression of WT CLK1 in the presence of TG003, or a catalytically inactive CLK1, had little to no effect on the splicing of this exon (Figure 3D, bottom panels), indicating that the catalytic activity of CLK1 is essential for regulating CHEK2 AS. Over-expression of the SR protein splicing regulator SRSF1, a known target of CLK1 (Prasad et al., 1999), had a similar effect as over-expression of CLK1, resulting in CHEK2 exon 9 skipping, whereas knockdown of SRSF1 had the opposite effect (Figure 3D, right panel). Furthermore, the activation of CHEK2 requires homodimerization (Shen et al., 2004), and we observe that the CHEK2 isoform lacking exon 9 still interacts with full-length protein (Figure 3—figure supplement 1E), suggesting that this CLK1-regulated isoform may function in a dominant-negative manner to attenuate CHEK2 activity.

CENPE is known to be tightly controlled at multiple levels (including transcription, localization, phosphorylation and degradation), and disruption of its regulation leads to pronounced mitotic defects. CENPE AS generates long and short isoforms (Supplementary file 2), with the predominant variant being the short isoform that lacks amino acids 1972–2068. Inhibition of CLK1 rapidly shifts CENPE splicing to produce predominantly the long isoform (Figure 3E), and this is accompanied by a reduction in CENPE protein levels during G2/M phase, presumably due to the instability of the long isoform (Figure 3E). These data thus show that CLK1 controls the AS of major cell cycle regulators, and therefore suggest that inhibition of CLK1 may alter cell cycle progression.

To investigate this, we next performed an RNA-Seq analysis of synchronized cells at G1 and early G2 (when CLK1 accumulates), following inhibition of CLK1 with TG003. As a specificity control, we performed a parallel RNA-Seq analysis using a structurally distinct CLK1 inhibitor, KHCB-19 (Fedorov et al., 2011). Strikingly, of 1498 AS events that change between G1 (t=0) and G2 (t=6) ~94% display reduced changes following treatment with the two drugs (Figure 3F), with 65% commonly affected by both drugs, thus supporting an important role for CLK1 in controlling cell-cycle dependent splicing.

CLK1 is required for normal mitosis and cell proliferation

Given that CLK1 regulates the AS of many cell cycle factors (Figure 3C), we next examined whether it is necessary for cell cycle progression. Knockdown of CLK1 using shRNAs led to an accumulation of cells with 4N DNA content in multiple cell types, specifically HeLa, H157, and A549 (Figure 4A, Figure 4—figure supplement 1A,B), and the extent of this accumulation correlated with the degree of knockdown (Figure 4—figure supplement 1A). We also observed a significant increase in multi-nucleation, a common consequence of defective chromosome segregation or cytokinesis, following shRNA-knockdown or TG003 inhibition of CLK1 in the treated cells (Figure 4B and Figure 4—figure supplement 1C). To visualize the effect of CLK1 inhibition on mitosis at a single cell level, we performed time-lapse high-content microscopy on live cells stably expressing a GFP-histone 2B fusion protein, to track changes in chromatin. TG003-treated cells entered mitosis normally, as measured by nuclear envelope breakdown, but displayed delayed or aberrant cytokinesis, typically resulting in multi-polar divisions, increased time in metaphase, failure to undergo chromatin de-condensation and eventual cell death (Figure 4C and Videos 1–5). To further determine at what cell cycle stage CLK1 activity is required, we inhibited CLK1 using TG003 at different time points after early S phase release. Consistent with the imaging data, both control and TG003-treated cells entered mitosis normally, as measured by 4N DNA content. However, inhibition of CLK1 before late S-phase impaired progression through mitosis, whereas cells treated 5 hr after early S phase release underwent a round of normal mitotic division, although failed to enter the next cell cycle (Figure 4D and Figure 4—figure supplement 1D). These results suggest that the primary defects caused by CLK1 inhibition occur in late S-phase and G2 phase, which is when CLK1 levels normally begin to rise (Figure 2A). This conclusion is further supported by the observation that CLK1-dependent AS targets, such as those detected in HMMR and CENPE, are periodically expressed during cell cycle and peak during G2 and M phase (Figure 4—figure supplement 1E). Taken together with the earlier results, these data support an important and multifaceted role for CLK1 in the control of cell cycle progression through its function in the global regulation of periodic AS.

Video 1. Live-cell imaging of control HeLa cells stably expressing a GFP-H2B.

Download video file (918.2KB, mp4)
DOI: 10.7554/eLife.10288.013
Open in a new tab

Cells were synchronized by single-thymidine block and released and imaged at 10X magnification (every ~15 min) for 960 min.

DOI: http://dx.doi.org/10.7554/eLife.10288.013

Video 2. Live-cell imaging of control HeLa cells stably expressing GFP-H2B.

Download video file (905.6KB, mp4)
DOI: 10.7554/eLife.10288.014
Open in a new tab

Cells were synchronized by single-thymidine block and released into 1 μM of TG003 and imaged at 10X magnification (every ~15 min) for 960 min.

DOI: http://dx.doi.org/10.7554/eLife.10288.014

Video 3. Zoom of live-cell imaging of control HeLa cells.

Download video file (17.6KB, mp4)
DOI: 10.7554/eLife.10288.015
Open in a new tab

Data associated with Figure 4D.

DOI: http://dx.doi.org/10.7554/eLife.10288.015

Video 4. Zoom of live-cell imaging of TG003-treated HeLa cells.

Download video file (191.6KB, mp4)
DOI: 10.7554/eLife.10288.016
Open in a new tab

Data associated with Figure 4.

DOI: http://dx.doi.org/10.7554/eLife.10288.016

Video 5. Zoom of live-cell imaging of TG003-treated HeLa cells.

Download video file (85.4KB, mp4)
DOI: 10.7554/eLife.10288.017
Open in a new tab

Data associated with Figure 4.

DOI: http://dx.doi.org/10.7554/eLife.10288.017

CLK1 expression and CLK1-regulated AS is altered in kidney cancer

The importance of CLK1 for faithful progression through the cell cycle suggests it may play a role the control of cell proliferation in cancer. Supporting this, shRNA knockdown or chemical inhibition of CLK1 with TG003 or KHCB-19 in HeLa cells results in a near complete block in cell proliferation, as measured by anchorage-dependent and -independent colony formation assays (Figure 4E and Figure 4—figure supplement 1F). As mentioned above, CLK1 is likely the primary target of inhibition in these experiments since RNAi of CLK1 recapitulates the phenotype seen with these chemical inhibitors.

Using RNA-Seq data from the Cancer Genome Atlas (TCGA) (Cancer Genome Atlas Network, 2013) we observe that CLK1 displays significantly higher expression in 72 kidney tumors compared to matched normal tissue samples (Figure 4F, p=10-5, Kolmogorov-Smirnov test). Consistently, most CLK1-controlled AS events that are altered in tumors have expected splicing changes (Figure 4—figure supplement 2A). Furthermore, patients with tumors that have elevated CLK1 expression (i.e. the upper quartile of all samples) have significantly lower survival rates relative to other patients in the comparison group (Figure 4G, p=0.007). While there was also an increase in CLK2, CLK3 and CLK4 mRNA expression in these tumors, CLK1 displayed the highest relative mRNA expression levels compared to CLK2 and CLK3 (Figure 4—figure supplement 2B). Levels of the CLK4, which is ~80% identical to CLK1, did not correlate with survival differences despite its increased levels in tumors (Figure 4—figure supplement 2C). Consistent with this, CLK1-regulated AS events, as defined by the RNA-Seq analysis in Figure 3, were also altered across multiple tumor types, including breast, colon, lung and liver (Figure 4H and Figure 4—figure supplement 2D). These data are further consistent with a multi-faceted role for CLK1 in regulating cell cycle progression, and also suggest that CLK1 contributes to increased cell proliferation in cancer, at least in part through its role in controlling periodic AS.

Discussion

Previous studies have shown that splicing and the cell cycle are intimately connected processes. Indeed, cell cycle division (CDC) loci originally defined in S. cerevisiae, namely cdc5 and cdc40, were subsequently shown to encode spliceosomal components (Ben-Yehuda et al., 2000; McDonald et al., 1999). Moreover, genome-wide RNAi screens for new AS regulators of apoptosis genes in human cells revealed that factors involved in cell-cycle control, in addition to RNA processing components, were among the most significantly enriched hits (Moore et al., 2010; Tejedor et al., 2015). An RNAi screen performed in Drosophila cells for genes required for cell-cycle progression identified numerous splicing components (Björklund et al., 2006) as well as a Drosophila ortholog of CLK kinases, Darkener of apricot Doa (Bettencourt-Dias et al., 2004). In other studies, negative control of splicing during M phase was shown to be dependent on dephosphorylation of the SR family protein, SRSF10 (Shin and Manley, 2002), and the mitotic regulator aurora kinase A (AURKA) was shown to control the AS regulatory activity of SRSF1 (Moore et al., 2010). Our study shows for the first time that AS patterns are subject to extensive periodic regulation, in part via a global control mechanism involving cell cycle fluctuations of the SR protein kinase CLK1. At least one likely function of this periodic AS regulation is to control the timing of activation of AURKB (Figure 1), as well as of numerous other key cell cycle factors shown here to be subject to periodic AS. The definition of an extensive, periodically-regulated AS program in the present study thus opens the door to understanding the functions of an additional layer of regulation associated with cell cycle control and cancer.

Since many RBPs are found to be periodically expressed (Figure 4A), it is likely that other aspects of mRNA metabolism are also coordinated with cell cycle stages. For example, differential degradation could potentially contribute to the observed periodic fluctuation of splice isoforms during cell cycle. The degradation of mRNA is closely linked with alternative polyadenylation, which has emerged as a critical mechanism that controls mRNA translation and stability. Generally, shortened 3’ UTRs are found in rapidly dividing cells and more aggressive cancers (Mayr and Bartel, 2009). This study identified 94 cases periodic alternative poly-A site usage (data not shown) in genes known to regulate cell cycle and/or proliferation, including SON, CENPF and EPCAM (Ahn et al., 2011; Bomont et al., 2005; Chaves-Perez et al., 2013). In addition, many mRNAs have recently been found to be translated in a cell cycle dependent fashion (Aviner et al., 2015; Maslon et al., 2014; Stumpf et al., 2013). Interestingly, a fraction of periodically translated genes are also periodically spliced, including key regulators of cell cycle (e.g., AURKA, AURKB, TTBK1 and DICER1). This observation is consistent with recent findings that the regulation of AS and translation may be coupled (Sterne-Weiler et al., 2013). In summary, we have demonstrated that AS is subject to extensive temporal regulation during the cell cycle in a manner that appears to be highly integrated with orthogonal layers of cell cycle control. These results thus provide a new perspective on cell cycle regulation that should be taken into consideration when studying this fundamental biological process, both in the context of normal physiology and diseases including cancers.

Materials and methods

Cell culture and synchronizations

HeLa (a kind gift from J. Trejo), HEK 293T (from ATCC CRL-3216) and A549 (kind gift from W. Kim) cells were maintained in DMEM (Gibco) medium supplemented with 10% FBS (Gibco). All cells were cultured in humidified incubators with 5% CO2. Cell cycle synchronization was adapted from the protocol of Whitfield et al. (Whitfield et al., 2002); ~750,000 log phase HeLa cells were plated in 15 cm dishes in complete media and allowed to attach for 16 hr, reaching <30% confluence. Cells were subsequently treated with 2 mM thymidine (Sigma-Aldrich, St. Louis, MO) for a total of 18 hr, washed 2 times with 1xPBS, and supplemented with fresh complete media for 10 hr. 2 mM thymidine was subsequently added for a second block of 18 hr and washed as described previously. Mitotic block was performed by double thymidine arrest (as above) and release in fresh media for 3.5 hr followed by addition of nocodazole 100 μM (Sigma) for 10 hr. G1 block was performed by serum starvation for 72 hr in DMEM containing 0.05% FBS. For RNA-Seq, cells (both adherent and detached) were harvested every 1.5 hr for 30 hr and frozen immediately for purification of total RNA. To block the activity of CLK1, cells were treated with TG003 (Sigma), KHCB-19 (Tocris, Bristol, UK). To block activity of the proteasome cells were treated with MG132 (Sigma). Drugs were-suspended in DMSO and added to growing cultures at the indicated concentrations and times.

Flow cytometry and cell cycle analysis

Cells were harvested with trypsin treatment, washed 2 times in cold 1xPBS and subsequently fixed in 80% ice cold ethanol for at least 4 hr. Cells were then washed twice with 1xPBS and suspended in propidium iodide/RNase staining buffer (BD Pharmingen, cat # 550825). Cells were analyzed by flow cytometry to count 10,000 cells that satisfied gating criteria. Data collected were analyzed using ModFit software to discern 2N (G1), S-phase, and 4N (G2 and M) composition.

Mapping and filtering of RNA-Seq data

RNA-Seq reads were mapped to the human genome (build hg19) using the MapSplice informatics tool with default parameters (Wang et al., 2010). The mapped reads were further analyzed with Cufflinks to calculate the level of gene expression with FPKM (Fragments Per Kilobase per Million mapped reads) (Trapnell et al., 2010). The levels of alternatively spliced isoforms were quantified with MISO (Mixture-of-Isoforms) probabilistic framework (Katz et al., 2010) using the annotated AS events for human hg19version 2. The levels of alternatively spliced isoforms were also quantified with VAST-TOOLS using the event annotation as previously described (Irimia et al., 2014). Each AS event was assigned a PSI or PIR value to represent the percent of transcripts with the exon spliced in, or the intron retained, respectively.

Identification of periodic AS

For identification of periodic AS raw PSI/PIR values were normalized as:

normalized(Φns)=Φns−ΦminsΦmaxs

where s = 1 to 32,109 for all splicing events; n = 1 to 14 for the 14 samples; Φmin is the minimum and Φmax is the maximum PSI value among the 14 samples.

To identify periodic AS events, normalized gene expression values (normalized FPKM values as en) for the well-known periodic gene, CCNB1, CCNA2, CCNB2, and CENPE, were used as a starting point to subsequently add curves with broader or sharper peaks as well as shifted to left or right, resulting in 7 periodic expression curves that cover all the phases of cell cycle (Figure 1—figure supplement 1A). We term these 'ideal seed curves', which capture intermittent peak times and phase shifts that were not well represented within the initial known periodic genes. To identify genes with similar splicing patterns across the cell cycle, we computed the EuclideanDistanceED of each AS event s to the model seed curves m as follows:

EDm,s=∑n=114|normalized(enm)−normalized(Φns)|

where m = 1 to 7 for all model seed curves, s = 1 to 32,109 for all AS events.

Based on the ranking of distance, a similar cutoff of ED≤2.75 was set as a minimum requirement for periodic AS. Lastly, we calculated a false discovery rate (FDR) by shuffling PSI values across the 14 time points 10,000 times and calculating how often a random shuffle had a better periodic score than the true periodic score for that event. A maximum FDR of 2.5% was required for a splicing event to be periodic.

Heat maps, correlations, GO-term analysis, overlap analysis and statistics

Heat maps, hierarchical clustering, and Pearson correlations were generated using GENE-E (www.broadinstitute.org/cancer/software/GENE-E/). All heat maps shown are row-normalized for presentation purposes. Spearman’s rank correlation with average linkage was used for clustering. DAVID (http://david.abcc.ncifcrf.gov/gene2gene.jsp) was used for all gene ontology enrichment; terms shown are for biological process (GOTERM_BP_FAT). To test for significance in overlap analysis, overlapping genes in two data sets (i.e. TG003-treatment and periodic AS) and a background set of only co-detected events was used (i.e. genes detected in both experiments). Significance of overlapping gene sets was assessed using the hyper-geometric test. For overlap and correlation analysis between VAST-TOOLS and MISO we used two MISO AS event annotations (HG18 and HG19), due to differences in the input annotation files for these two pipelines. To identify the 4,343 overlapping exons in the heatmap, we used MISO hg19v1 annotations and MISO hg18 annotations, and the VAST-TOOLS annotations. Student’s t-test was used to measure significant in cell cycle defects (multi-nucleation and flow-cytometry) as well as semi-quantitative RT-PCR assessment of splice variants. To identify AS events blocked by TG003 or KHCB19, a Student’s t-test statistic was used. If a change between G1/S (t=0) and G2 (t=6) was significant, but not significant in the presence of inhibitors we consider that event to be blocked. For over-representation of periodic introns, we performed Fisher’s exact test in a 2x2 contingency table as compared to skipped exons. For differences in expression of CLK1 mRNAs in kidney cancers a Kolmogorov-Smirnov test was performed. For survival differences, the survdiff function in the R survival package was used (as discussed in methods below).

Plasmid construction, transfections and RNAi

The expression constructs were generated by cloning the cDNA of CLK1 into pCDNA3 (for transient expression) or pCDH (for stable transfection) backbones with different epitope tags (HA or Flag) at the N- or C-terminus. The Myc-His-Ubiquitin expression vector is a gift from Dr. Gary Johnson’s lab, and the Histone H2B-GFP expression vector is gift from Dr. Angelique Whitehurst’s lab. Plasmid transfections were performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s protocol. The lentiviral vectors of shRNAs were obtained from Addgene in pLKO.1 TRC cloning plasmid through UNC core facility as part of mammalian gene knockdown consortium. Lentiviral infections were performed according to the manufacturer's instruction from System Biosciences (SBI).

Semi-quantitative RT-PCR assays for monitoring splicing

Cells transfected with shRNA constructs or treated with TG003 were lysed and total RNA was extracted using the Trizol method (Life Technologies). Purified RNA was treated with 1U of RNAse-free DNAase (Promega) for 1 hr at 37°C and reverse transcribed using random hexamer cDNA preparation kit (Applied Biosystem). One-tenth of the RT product was used as the template for PCR amplification (25 cycles of amplification, with a trace amount of Cy5-dCTP in addition to non-fluorescent dNTPs) using gene specific primers listed in Supplementary file 4. The resulting gels were scanned with a Typhoon 8600 Imager (GE Healthcare), and analyzed with ImageQuant 5.2 software (Molecular Dynamics/GE Healthcare). Real time PCR was carried out using the SYBR Green kit (Invitrogen) and GAPDH as an internal control.

Immunoblotting and immunoprecipitation

Proteins were extracted in lysis buffer (CHAPS 1% w/v, 150 mM NaCl, 50 mM MgCl2 with protease inhibitor), resolved by SDS-PAGE and transferred onto PVDF membrane. For immunoprecipitaion experiments to detect ubiquitination, cells were co-transfected with Flag-CLK1 and myc-ubiquitin constructs as above. 36 hr later, TG003 (20 μM) was added for 18 hr. 4 hr prior to harvesting, 10 μM of MG132 was added to the media. Proteins were extracted in lysis buffer as above with the addition of NEM. Incubation with EZ-View FLAG Beads (Sigma) was performed for 2 hr at 4°C. Samples were extensively washed according to the manufacturer’s protocol and subjected to immunoblotting. Antibodies and dilutions are listed in Supplementary file 4.

Immunofluorescence and high-content live-cell imaging

For immunofluorescence microscopy, cells were plated on glass coverslips coated with poly-L-Lysine. Cells were then washed twice with 1xPBS, fixed with 4% formaldehyde (Sigma), permeabilized with 0.05% Triton X-100 (Promega) and blocked with 3% BSA (Fisher); all dilutions were made in 1XPBS. For live cell imaging, HeLa cells transduced with Histone H2B-GFP was stably selected as described previously (Cappell et al., 2010). Cells were plated in a 6-well format and treated with 2 mM thymidine for 24 hr, subsequently washed and released in fresh complete medium with or without TG003 (20 µM). Cells were imaged using the BD Pathway Microscope with a 10X objective.

Colony formation assays

HeLa cells stably producing shRNAs targeting CLK1 or control shRNAs were plated at low density (1000 cells/6 cm2) in standard culture medium and allowed to proliferate for 9 days. Cells were then fixed and stained with crystal violet at room temperature. The dried plates were used for estimations of colony diameter and number.

Kidney cancer analysis of CLK1 and alternative splicing

RNA-Seq data from the The Cancer Genome Atlas (Ciriello et al., 2013) was processed as previously described (Tsai et al., 2015). Briefly, for mRNA expression, RSEM expression values for the indicated genes were analyzed in 79 paired KIRC (tumor and normal) samples and the paired ks-test was used to test significance. For alternative splicing analysis data from BRCA: Breast invasive carcinoma, COAD: Colorectal adenocarcinoma, KIRC: Kidney renal clear cell carcinoma, LUAD: Lung adenocarcinoma, LUSC: Lung squamous cell carcinoma, LIHC: liver hepatocellular carcinoma were analyzed through the MISO pipeline as described above (Tsai et al., 2015). Relapse-free survival was analyzed using Kaplan Meier plots. All plots and statistical analyses (survdiff) were generated using the R package version 3.1.1 survival function.

Acknowledgements

We thank Dr. Jan Prins and Darshan Singh for helpful suggestions in analyzing RNA-Seq data, and Drs. Angelique Whitehurst, William Marzluff and Jean Cook for sharing reagents. We thank Drs. Jun Zhu and Ting Ni for advice and assistance in generating strand-specific RNA library. Dr. Rebecca Sinnott provided assistance in live cell imaging experiments. We thank Drs. William Marzluff, Chris Burge, Jean Cook, Michael Emanuele and Mauro Calabrese for helpful comments on the manuscript. This work is supported by an NIH grant (R01CA158283) and Jefferson-Pilot fellowship to ZW, and by grants from the Canadian Institutes of Health Research to BJB. BJB holds the Banbury Chair of Medical Research at the University of Toronto.

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Funding Information

This paper was supported by the following grant:

  • National Cancer Institute R01CA158283 to Zefeng Wang.

Additional information

Competing interests

BJB: Reviewing editor, eLife.

The other authors declare that no competing interests exist.

Author contributions

DD, Conceived the project design, Performed the experiments, Analyzed data, Wrote the paper.

Y-HT, Analyzed data, Acquisition of data.

RW, Analyzed data, Drafting or revising the article.

YW, Analyzed data, Acquisition of data.

BJB, Analyzed data, Wrote the paper.

ZW, Conceived the project design, Analyzed data, Wrote the paper.

Additional files

Supplementary file 1. Periodic splicing events identified during the HeLa cell cycle and associated GO terms.

Data associated with Figure 1.

DOI: http://dx.doi.org/10.7554/eLife.10288.018

elife-10288-supp1.xlsx (9.8MB, xlsx)
DOI: 10.7554/eLife.10288.018
Supplementary file 2. AS events altered by CLK1 inhibition (TG003 10 μM for 16 hr) in 293T cells and associated GO terms.

Data associated with Figure 3.

DOI: http://dx.doi.org/10.7554/eLife.10288.019

elife-10288-supp2.xlsx (201.4KB, xlsx)
DOI: 10.7554/eLife.10288.019
Supplementary file 3. AS events altered by CLK1 inhibition (TG003 10 μM, KHCB19 10 μM) in synchronized HeLa cells.

Data associated with Figure 3.

DOI: http://dx.doi.org/10.7554/eLife.10288.020

elife-10288-supp3.xlsx (2.1MB, xlsx)
DOI: 10.7554/eLife.10288.020
Supplementary file 4. List of PCR primers, shRNA sequences and antibodies used in this study.

DOI: http://dx.doi.org/10.7554/eLife.10288.021

elife-10288-supp4.xlsx (12.4KB, xlsx)
DOI: 10.7554/eLife.10288.021

Major datasets

The following dataset was generated:

Dominguez D, Tsai Y-H, Wang Z,2016,An extensive program of periodic alternative splicing linked to cell cycle progression,http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE81485,Publicly available at the Gene Expression Omnibus NCBI website (accession no: GSE81485)

References

  1. Ahn EY, DeKelver RC, Lo MC, Nguyen TA, Matsuura S, Boyapati A, Pandit S, Fu XD, Zhang DE. SON controls cell-cycle progression by coordinated regulation of RNA splicing. Molecular Cell. 2011;42:185–198. doi: 10.1016/j.molcel.2011.03.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Aviner R, Shenoy A, Elroy-Stein O, Geiger T. Uncovering hidden layers of cell cycle regulation through integrative multi-omic analysis. PLoS Genetics. 2015;11:e10288. doi: 10.1371/journal.pgen.1005554. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bar-Joseph Z, Siegfried Z, Brandeis M, Brors B, Lu Y, Eils R, Dynlacht BD, Simon I. Genome-wide transcriptional analysis of the human cell cycle identifies genes differentially regulated in normal and cancer cells. Proceedings of the National Academy of Sciences of the United States of America. 2008;105:955–960. doi: 10.1073/pnas.0704723105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Ben-David Y, Letwin K, Tannock L, Bernstein A, Pawson T. A mammalian protein kinase with potential for serine/threonine and tyrosine phosphorylation is related to cell cycle regulators. The EMBO Journal. 1991;10:317–325. doi: 10.1002/j.1460-2075.1991.tb07952.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Ben-Yehuda S, Dix I, Russell CS, McGarvey M, Beggs JD, Kupiec M. Genetic and physical interactions between factors involved in both cell cycle progression and pre-mRNA splicing in Saccharomyces cerevisiae. Genetics. 2000;156:1503–1517. doi: 10.1093/genetics/156.4.1503. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bertoli C, Skotheim JM, de Bruin RA. Control of cell cycle transcription during G1 and S phases. Nature Reviews. Molecular Cell Biology. 2013;14:518–528. doi: 10.1038/nrm3629. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bettencourt-Dias M, Giet R, Sinka R, Mazumdar A, Lock WG, Balloux F, Zafiropoulos PJ, Yamaguchi S, Winter S, Carthew RW, Cooper M, Jones D, Frenz L, Glover DM. Genome-wide survey of protein kinases required for cell cycle progression. Nature. 2004;432:980–987. doi: 10.1038/nature03160. [DOI] [PubMed] [Google Scholar]
  8. Björklund M, Taipale M, Varjosalo M, Saharinen J, Lahdenperä J, Taipale J. Identification of pathways regulating cell size and cell-cycle progression by RNAi. Nature. 2006;439:1009–1013. doi: 10.1038/nature04469. [DOI] [PubMed] [Google Scholar]
  9. Bomont P, Maddox P, Shah JV, Desai AB, Cleveland DW. Unstable microtubule capture at kinetochores depleted of the centromere-associated protein CENP-F. The EMBO Journal. 2005;24:3927–3939. doi: 10.1038/sj.emboj.7600848. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Boutz PL, Bhutkar A, Sharp PA. Detained introns are a novel, widespread class of post-transcriptionally spliced introns. Genes & Development. 2015;29:63–80. doi: 10.1101/gad.247361.114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Cancer Genome Atlas Research Network Comprehensive molecular characterization of clear cell renal cell carcinoma. Nature. 2013;499:43–49. doi: 10.1038/nature12222. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Cappell KM, Larson B, Sciaky N, Whitehurst AW. Symplekin specifies mitotic fidelity by supporting microtubule dynamics. Molecular and Cellular Biology. 2010;30:5135–5144. doi: 10.1128/MCB.00758-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Carmena M, Wheelock M, Funabiki H, Earnshaw WC. The chromosomal passenger complex (CPC): from easy rider to the godfather of mitosis. Nature Reviews. Molecular Cell Biology. 2012;13:789–803. doi: 10.1038/nrm3474. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Chaves-Pérez A, Mack B, Maetzel D, Kremling H, Eggert C, Harréus U, Gires O. EpCAM regulates cell cycle progression via control of cyclin D1 expression. Oncogene. 2013;32:641–650. doi: 10.1038/onc.2012.75. [DOI] [PubMed] [Google Scholar]
  15. Ciriello G, Miller ML, Aksoy BA, Senbabaoglu Y, Schultz N, Sander C. Emerging landscape of oncogenic signatures across human cancers. Nature Genetics. 2013;45:1127–1133. doi: 10.1038/ng.2762. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. David CJ, Manley JL. Alternative pre-mRNA splicing regulation in cancer: pathways and programs unhinged. Genes & Development. 2010;24:2343–2364. doi: 10.1101/gad.1973010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Duncan PI, Stojdl DF, Marius RM, Bell JC. In vivo regulation of alternative pre-mRNA splicing by the Clk1 protein kinase. Molecular and Cellular Biology. 1997;17:5996–6001. doi: 10.1128/mcb.17.10.5996. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Edmond V, Moysan E, Khochbin S, Matthias P, Brambilla C, Brambilla E, Gazzeri S, Eymin B. Acetylation and phosphorylation of SRSF2 control cell fate decision in response to cisplatin. The EMBO Journal. 2011;30:510–523. doi: 10.1038/emboj.2010.333. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Fedorov O, Huber K, Eisenreich A, Filippakopoulos P, King O, Bullock AN, Szklarczyk D, Jensen LJ, Fabbro D, Trappe J, Rauch U, Bracher F, Knapp S. Specific CLK inhibitors from a novel chemotype for regulation of alternative splicing. Chemistry & Biology. 2011;18:67–76. doi: 10.1016/j.chembiol.2010.11.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Grabek KR, Diniz Behn C, Barsh GS, Hesselberth JR, Martin SL. Enhanced stability and polyadenylation of select mRNAs support rapid thermogenesis in the brown fat of a hibernator. eLife. 2015;4 doi: 10.7554/eLife.04517. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Gui JF, Lane WS, Fu XD. A serine kinase regulates intracellular localization of splicing factors in the cell cycle. Nature. 1994;369:678–682. doi: 10.1038/369678a0. [DOI] [PubMed] [Google Scholar]
  22. Harashima H, Dissmeyer N, Schnittger A. Cell cycle control across the eukaryotic kingdom. Trends in Cell Biology. 2013;23:345–356. doi: 10.1016/j.tcb.2013.03.002. [DOI] [PubMed] [Google Scholar]
  23. Irimia M, Weatheritt RJ, Ellis JD, Parikshak NN, Gonatopoulos-Pournatzis T, Babor M, Quesnel-Vallières M, Tapial J, Raj B, O'Hanlon D, Barrios-Rodiles M, Sternberg MJ, Cordes SP, Roth FP, Wrana JL, Geschwind DH, Blencowe BJ. A highly conserved program of neuronal microexons is misregulated in autistic brains. Cell. 2014;159:1511–1523. doi: 10.1016/j.cell.2014.11.035. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Jiang K, Patel NA, Watson JE, Apostolatos H, Kleiman E, Hanson O, Hagiwara M, Cooper DR. Akt2 regulation of Cdc2-like kinases (Clk/Sty), serine/arginine-rich (SR) protein phosphorylation, and insulin-induced alternative splicing of PKCbetaII messenger ribonucleic acid. Endocrinology. 2009;150:2087–2097. doi: 10.1210/en.2008-0818. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Karni R, de Stanchina E, Lowe SW, Sinha R, Mu D, Krainer AR. The gene encoding the splicing factor SF2/ASF is a proto-oncogene. Nature Structural & Molecular Biology. 2007;14:185–193. doi: 10.1038/nsmb1209. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Katz Y, Wang ET, Airoldi EM, Burge CB. Analysis and design of RNA sequencing experiments for identifying isoform regulation. Nature Methods. 2010;7:1009–1015. doi: 10.1038/nmeth.1528. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kim Y, Heuser JE, Waterman CM, Cleveland DW. CENP-E combines a slow, processive motor and a flexible coiled coil to produce an essential motile kinetochore tether. The Journal of Cell Biology. 2008;181:411–419. doi: 10.1083/jcb.200802189. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Lai MC, Lin RI, Tarn WY. Differential effects of hyperphosphorylation on splicing factor SRp55. The Biochemical Journal. 2003;371:937–945. doi: 10.1042/BJ20021827. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Lens SM, Voest EE, Medema RH. Shared and separate functions of polo-like kinases and aurora kinases in cancer. Nature Reviews. Cancer. 2010;10:825–841. doi: 10.1038/nrc2964. [DOI] [PubMed] [Google Scholar]
  30. Leva V, Giuliano S, Bardoni A, Camerini S, Crescenzi M, Lisa A, Biamonti G, Montecucco A. Phosphorylation of SRSF1 is modulated by replicational stress. Nucleic Acids Research. 2012;40:1106–1117. doi: 10.1093/nar/gkr837. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Long JC, Caceres JF. The SR protein family of splicing factors: master regulators of gene expression. The Biochemical Journal. 2009;417:15–27. doi: 10.1042/BJ20081501. [DOI] [PubMed] [Google Scholar]
  32. Ly T, Ahmad Y, Shlien A, Soroka D, Mills A, Emanuele MJ, Stratton MR, Lamond AI. A proteomic chronology of gene expression through the cell cycle in human myeloid leukemia cells. eLife. 2014;3:e10288. doi: 10.7554/elife.01630. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Maslon MM, Heras SR, Bellora N, Eyras E, Cáceres JF. The translational landscape of the splicing factor SRSF1 and its role in mitosis. eLife. 2014:e10288. doi: 10.7554/eLife.02028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Matera AG, Wang Z. A day in the life of the spliceosome. Nature Reviews. Molecular Cell Biology. 2014;15:108–121. doi: 10.1038/nrm3742. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Mayr C, Bartel DP. Widespread shortening of 3'UTRs by alternative cleavage and polyadenylation activates oncogenes in cancer cells. Cell. 2009;138:673–684. doi: 10.1016/j.cell.2009.06.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. McDonald WH, Ohi R, Smelkova N, Frendewey D, Gould KL. Myb-related fission yeast cdc5p is a component of a 40S snRNP-containing complex and is essential for pre-mRNA splicing. Molecular and Cellular Biology. 1999;19:5352–5362. doi: 10.1128/mcb.19.8.5352. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Mocciaro A, Rape M. Emerging regulatory mechanisms in ubiquitin-dependent cell cycle control. Journal of Cell Science. 2012;125:255–263. doi: 10.1242/jcs.091199. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Moore MJ, Wang Q, Kennedy CJ, Silver PA. An alternative splicing network links cell-cycle control to apoptosis. Cell. 2010;142:625–636. doi: 10.1016/j.cell.2010.07.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Muraki M, Ohkawara B, Hosoya T, Onogi H, Koizumi J, Koizumi T, Sumi K, Yomoda J, Murray MV, Kimura H, Furuichi K, Shibuya H, Krainer AR, Suzuki M, Hagiwara M. Manipulation of alternative splicing by a newly developed inhibitor of Clks. The Journal of Biological Chemistry. 2004;279:24246–24254. doi: 10.1074/jbc.M314298200. [DOI] [PubMed] [Google Scholar]
  40. Ninomiya K, Kataoka N, Hagiwara M. Stress-responsive maturation of Clk1/4 pre-mRNAs promotes phosphorylation of SR splicing factor. The Journal of Cell Biology. 2011;195:27–40. doi: 10.1083/jcb.201107093. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Oltean S, Bates DO. Hallmarks of alternative splicing in cancer. Oncogene. 2014;33:5311–5318. doi: 10.1038/onc.2013.533. [DOI] [PubMed] [Google Scholar]
  42. Pan Q, Shai O, Misquitta C, Zhang W, Saltzman AL, Mohammad N, Babak T, Siu H, Hughes TR, Morris QD, Frey BJ, Blencowe BJ. Revealing global regulatory features of mammalian alternative splicing using a quantitative microarray platform. Molecular Cell. 2004;16:929–941. doi: 10.1016/j.molcel.2004.12.004. [DOI] [PubMed] [Google Scholar]
  43. Pan Q, Shai O, Lee LJ, Frey BJ, Blencowe BJ. Deep surveying of alternative splicing complexity in the human transcriptome by high-throughput sequencing. Nature Genetics. 2008;40:1413–1415. doi: 10.1038/ng.259. [DOI] [PubMed] [Google Scholar]
  44. Paronetto MP, Miñana B, Valcárcel J. The Ewing sarcoma protein regulates DNA damage-induced alternative splicing. Molecular Cell. 2011;43:353–368. doi: 10.1016/j.molcel.2011.05.035. [DOI] [PubMed] [Google Scholar]
  45. Prasad J, Colwill K, Pawson T, Manley JL. The protein kinase Clk/Sty directly modulates SR protein activity: both hyper- and hypophosphorylation inhibit splicing. Molecular and Cellular Biology. 1999;19:6991–7000. doi: 10.1128/mcb.19.10.6991. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Satyanarayana A, Kaldis P. Mammalian cell-cycle regulation: several Cdks, numerous cyclins and diverse compensatory mechanisms. Oncogene. 2009;28:2925–2939. doi: 10.1038/onc.2009.170. [DOI] [PubMed] [Google Scholar]
  47. Shen H, Kan JL, Green MR. Arginine-serine-rich domains bound at splicing enhancers contact the branchpoint to promote prespliceosome assembly. Molecular Cell. 2004;13:367–376. doi: 10.1016/s1097-2765(04)00025-5. [DOI] [PubMed] [Google Scholar]
  48. Shin C, Manley JL. The SR protein SRp38 represses splicing in M phase cells. Cell. 2002;111:407–417. doi: 10.1016/s0092-8674(02)01038-3. [DOI] [PubMed] [Google Scholar]
  49. Staalesen V, Falck J, Geisler S, Bartkova J, Børresen-Dale AL, Lukas J, Lillehaug JR, Bartek J, Lønning PE. Alternative splicing and mutation status of CHEK2 in stage III breast cancer. Oncogene. 2004;23:8535–8544. doi: 10.1038/sj.onc.1207928. [DOI] [PubMed] [Google Scholar]
  50. Sterne-Weiler T, Martinez-Nunez RT, Howard JM, Cvitovik I, Katzman S, Tariq MA, Pourmand N, Sanford JR. Frac-seq reveals isoform-specific recruitment to polyribosomes. Genome Research. 2013;23:1615–1623. doi: 10.1101/gr.148585.112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Stumpf CR, Moreno MV, Olshen AB, Taylor BS, Ruggero D. The translational landscape of the mammalian cell cycle. Molecular Cell. 2013;52:574–582. doi: 10.1016/j.molcel.2013.09.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Tejedor JR, Papasaikas P, Valcárcel J. Genome-wide identification of Fas/CD95 alternative splicing regulators reveals links with iron homeostasis. Molecular Cell. 2015;57:23–38. doi: 10.1016/j.molcel.2014.10.029. [DOI] [PubMed] [Google Scholar]
  53. Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van Baren MJ, Salzberg SL, Wold BJ, Pachter L. Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nature Biotechnology. 2010;28:511–515. doi: 10.1038/nbt.1621. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Tsai YS, Dominguez D, Gomez SM, Wang Z. Transcriptome-wide identification and study of cancer-specific splicing events across multiple tumors. Oncotarget. 2015;6:6825–6839. doi: 10.18632/oncotarget.3145. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Vermeulen K, Van Bockstaele DR, Berneman ZN. The cell cycle: a review of regulation, deregulation and therapeutic targets in cancer. Cell Proliferation. 2003;36:131–149. doi: 10.1046/j.1365-2184.2003.00266.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Wang GS, Cooper TA. Splicing in disease: disruption of the splicing code and the decoding machinery. Nature Reviews. Genetics. 2007;8:749–761. doi: 10.1038/nrg2164. [DOI] [PubMed] [Google Scholar]
  57. Wang ET, Sandberg R, Luo S, Khrebtukova I, Zhang L, Mayr C, Kingsmore SF, Schroth GP, Burge CB. Alternative isoform regulation in human tissue transcriptomes. Nature. 2008;456:470–476. doi: 10.1038/nature07509. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Wang Z, Burge CB. Splicing regulation: from a parts list of regulatory elements to an integrated splicing code. RNA. 2008;14:802–813. doi: 10.1261/rna.876308. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Wang K, Singh D, Zeng Z, Coleman SJ, Huang Y, Savich GL, He X, Mieczkowski P, Grimm SA, Perou CM, MacLeod JN, Chiang DY, Prins JF, Liu J. MapSplice: accurate mapping of RNA-seq reads for splice junction discovery. Nucleic Acids Research. 2010;38:e10288. doi: 10.1093/nar/gkq622. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Wang Y, Chen D, Qian H, Tsai YS, Shao S, Liu Q, Dominguez D, Wang Z. The splicing factor RBM4 controls apoptosis, proliferation, and migration to suppress tumor progression. Cancer Cell. 2014;26:374–389. doi: 10.1016/j.ccr.2014.07.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Whitfield ML, Sherlock G, Saldanha AJ, Murray JI, Ball CA, Alexander KE, Matese JC, Perou CM, Hurt MM, Brown PO, Botstein D. Identification of genes periodically expressed in the human cell cycle and their expression in tumors. Molecular Biology of the Cell. 2002;13:1977–2000. doi: 10.1091/mbc.02-02-0030.. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Zhou Z, Fu XD. Regulation of splicing by SR proteins and SR protein-specific kinases. Chromosoma. 2013;122:191–207. doi: 10.1007/s00412-013-0407-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
eLife. 2016 Mar 25;5:e10288. doi: 10.7554/eLife.10288.029

Decision letter

Editor: Gene Yeo1

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your work entitled "An Extensive Program of Periodic Alternative Splicing Linked to Cell Cycle Progression" for peer review at eLife. Your submission has been favorably evaluated by Aviv Regev (Senior editor), a Reviewing editor (Gene Yeo), and three reviewers.

The reviewers have discussed the reviews with one another and the Reviewing editor has drafted this decision to help you prepare a revised submission.

Summary:

The authors identify widespread periodic changes in alternative splicing (AS) (>1000 AS events) that are coordinated with stages of the cell cycle, noticing that periodically regulated intron retention is prominent. The authors discover that SR protein kinase CLK1 is subject to cell cycle-dependent changes and appears to play a central role in the control of AS during cell cycle.

Essential revisions:

1) There were major concerns with regards to the poor overlap of the two pipelines (VAST and MISO) in the identification of AS events. While the reviewers and myself appreciate that the point of the manuscript is not a systematic comparison of both software, the authors are expected to explain the rationale for decisions taken which likely impact the interpretation of the results. Greater detail in the bioinformatics analysis is expected in the revision. Also, why are both tools used? And given the relatively poor overlap, is the use of either or both of these tools justified?

2) A more complete treatment of the contribution of other splicing factors to cell cycle AS is necessary. I do not believe it has to be exhaustive, but in its present form, the manuscript led to concerns by the reviewers that CLK1 is not necessarily a clear outlier (and therefore candidate) in terms of its regulation, compared to other factors. CLK4 modification and levels should be certainly verified by Western blot analysis. Based on these new results (in the revision), the authors are encouraged to alter/broaden their title (as suggested by Reviewer 3 as well).

3) The authors should address the major concerns of Reviewers 1 and 2 with regards to why is the CLK1 inhibitor used (rather than depletion) and address clearly their interpretation of specificity since the inhibitors also target CLK4.

4) The authors should address queries with regards to the PTC-induced degradation of AURKB (Reviewer 1) and known IR events (Reviewer 3).

Reviewer #1:

In this interesting paper, the Blencowe and Wang labs investigate the role of Alternative splicing (AS) during cell cycle progression.

First, Dominguez and co-authors carried out a sequence analysis of the human transcriptome during two continuous cell cycles. This analysis resulted in the identification of widespread periodic changes in AS that are coordinated with specific stages of the cell cycle. In particular, they identified 1,747 AS cell cycle-dependent AS events in 1,293 genes. They found that periodically regulated intron retention is the predominant event.

The authors next focused on the SR protein kinase, CLK1 and found that its expression also fluctuates during the cell cycle. These and other results suggested that CLK1 might have a specific role in regulating cell cycle-dependent AS. Following this, the authors went on to identify endogenous CLK1 AS targets using an RNA-Seq approach. Altogether, these data suggests that temporal regulation of splicing by CLK1 is critical for cell cycle progression. Finally, the authors conclusively show that CLK1 is indeed required for normal mitosis and cell proliferation. The authors put forward a model whereby CLK1 has an essential role in controlling cell cycle-dependent AS.

Defining a central role for CLK1 in the regulation of AS during cell cycle is of importance. The authors also nicely show that the levels of CLK1 are indeed self-regulated and depend on its own catalytic activity. The main targets of CLK1 are SR proteins, yet only in the case of CHEK2 pre-mRNA splicing, it is shown that the effect of CLK1 operates via SRSF1. It would be important to show some evidence that links the role of CLK1 in cell cycle regulated splicing with a particular SR protein/s.

In summary, this is a very good study that represents and important contribution to our understanding of Alternative splicing and cell cycle progression.

Specific comments:

1) In the Abstract and again in paragraph two of the Introduction, the authors refer to periodic regulation of gene function at the levels of transcription, protein modification and degradation, etc.; however, they ignore the role of mRNA translation. There are several papers describing the regulation of mRNA translation during mitosis, which would enrich the Discussion section. Some of these include:

Stumpf et al. (2013) The translational landscape of the mammalian cell cycle. Mol Cell. 2013 Nov 21;52(4):574-82. PubMed PMID: 24120665.

Maslon et al. (2014) The translational landscape of the splicing factor SRSF1 and its role in mitosis. eLife. 2014 May 6:e02028. doi: 10.7554/eLife.02028. PubMed PMID: 24842991.

2) Is there any overlap in those genes found here to be regulated at the AS level and those previously reported to be regulated by mRNA translation?

3) There seems to be independent regulation of transcriptionally and AS-regulated genes during the cell cycle. Are those genes that show an overlap (133) more likely to be regulated by the kinetic control mode of AS regulation? In other words, are those genes preferentially regulated by the processivity of RNA pol II?

4) On Figure 1E, has the PTC-induced degradation of AURKB been experimentally determined?

5) On Figure 3, why is a CLK1 inhibitor preferred to a depletion of CLK1 (either via CRISPR or siRNA approach)?

6) On Figure 3, is the CENPE AS also regulated by SRSF1, as is the case with the CHEK2 pre-mRNA?

Reviewer #2:

This manuscript by Dominguez et al., describes a thorough and important analysis of transcriptome changes during the cell cycle, with particular emphasis on alternative splicing. This analysis then motivates the authors to investigate a potential role of the kinase CLK1 in regulating cell cycle splicing and progression, and its possible link to cancer. If fully substantiated, this work would add tremendously to the growing body of knowledge regarding how different cellular growth conditions impact splicing. However, some additional analysis and important controls are required for this manuscript to achieve its full impact.

Firstly, much of the analysis of the RNA-Seq data is poorly defined. Why were VAST-TOOLs and MISO both used and what is the advantage of this? What is "Periodic score" in Figure 1—figure supplement 1B and C? The authors may feel these are readily apparent, but readers should not have to dig deeply through the Methods to understand the first figure. More importantly, Figure 1A should either be replaced with a version that gives raw PSI, or a version with raw PSI should be included in the supplement. Use of the "row normalized" method raises concerns that many of the changes observed in AS events are modest and potentially not of biologic significance.

Secondly, the justification for focusing on CLK1 is not clear. While Figure 3 and Figure 4 certainly support a role of the CLK family in cell cycle splicing, Figure 2 is not convincing that CLK1 is a clear outlier in terms of its regulation. A quick glance through the GE values in supplemental table 1 indicates that genes encoding several SR proteins vary throughout the cell cycle. Have SR proteins other than SRSF1 been analyzed by Western? Moreover, CLK4 should be certainly tested by Western – due to it being a target of TG003 and KHCB-19 (see next point).

Finally, there is significant concern that both the inhibitors used target CLK4 as well as (or more potently than) CLK1. The authors should either conclusively rule out a role for CLK4, or broaden their conclusions to acknowledge a potential role for this family member.

Reviewer #3:

In this work the authors create a genome wide map of oscillating alternative splicing events, and identify CLK1 as a major regulator of a subset of these oscillating events. Roughly, two thirds the paper are focused on exploring the regulatory role of CLK1 through a series of elaborate and thoughtful experiments after establishing it plays a major role in regulating oscillating AS. We found relatively few and mostly minor concerns regarding these later sections. Most critical concerns regard the first, genome wide, analysis.

1) The two pipelines used, based on VAST and MISO, are basically in severe disagreement with respect to which AS events are oscillating. Figure 1—figure supplement 1F shows that not only is the number of events detected very different (244 vs 513) but the overlap of those is a mere 27% or 12%(!). This is coming from two pipelines that use the same downstream procedure to check for oscillation and differ only in their approach to quantify the raw PSI values fed into the pipeline. This poor overlap puts the entire genome wide mapping presented here into questioning. Figure 1—figure supplement 1G may have been added to address this concern (it's not explained) but actually does little to alleviate this concern: Figure 1—figure supplement 1G basically shows that the correlation between VAST and MISO based PSI values are much better for the same condition (diagonal) than for different conditions when looking across the >4K events quantified. That is actually to be expected: most events are either highly included or highly excluded. Such highly included/excluded events would (a) make the extreme values dominating the correlation coefficient and (b) would make the majority of cases that are generally not changing and thus both methods agree on. In fact, a closer look at Figure 1—figure supplement 1G shows the average PSI correlation between VAST and MISO for the same condition (diagonal) is ~0.55 which is again troubling (btw, setting the dark red color for 0.55 is misleading). Some suggestions that may help alleviate the discrepancies between the two pipelines are given below.

Related to this: why does Figure 1—figure supplement 1D contain only 513 events but Figure 1A contains 1747 events? If Figure 1—figure supplement 1D is VAST's list why does Figure 1—figure supplement 1F show 513 for MISO?

2) Basic "normalized PSI" (subheading “Identification of periodic AS”) means we are only looking at relative changes of PSI. This may introduce a lot of variability/noise to the analysis and may contribute to the substantial differences between the results from the two analysis pipelines (see above). The authors may be better off screening for events for which (a) min(max(PSI) > VAL1) and (b) max(PSI) – min(PSI) > VAL2. Similarly, increasing the threshold on number of reads per event to be included could help. For example, using VAL1=VAL2 = 20% could help avoid fluctuations that may just appear as periodic or changing dramatically in a relatively small/insignificant range.

Related to this: the authors do not explain how events are screened for coverage across the experiments.

3) The method for detecting the oscillating AS events and its effect on their results is not discussed or evaluated. Specifically, the authors use 7 previously characterized profiles from GE analysis. It's not clear whether the same kind of profiles is the best choice for analyzing periodic AS. For example, profiles 3,4 are similar in phase but one has a wider shape – why would that be the best fit for AS profiles? It seems reasonable to at least compare the results from this approach to an unbiased approach where the entire set of AS events are clustered with no specific periodic profiles and then the most prominent periodic profiles are extracted or executing a more directed search against theoretical periodic profiles.

4) More convincing/exhaustive search of RBPs that may contribute to cell cycle AS:

While there's significant overlap between CLK1 inhibitor AS events and periodic cell cycle events, this only explains a fraction of all the periodic events they find. It would be good to acknowledge at least that there are likely other factors out there that contribute to this program of AS. The authors should discuss how much of the oscillation signal is explained by CLK1. The WB data from Figure 2A is not exhaustive, but it is convincing for those factors tested. It should be straightforward to supplement this analysis by applying their pipelines and RNA-seq data to report known/suspected splicing factors and/or RBPs that change at the level of GE or AS. Any changes found in this analysis could potentially contribute to the large AS program they observe beyond CLK1 regulation.

5) Analysis on IR/AS-PTC introducing events:

Figure 1D/E suggests that periodic IR can affect transcript expression levels (probably by the introduction of PTCs) in the case of AURKB and some other genes they examined by qRT-PCR that are important for cell cycle progression. However Figure 1B argues little overlap between splicing regulation and mRNA expression levels, save for 133 overlapping events. Are the events in the overlap of periodic AS and periodic gene expression overrepresented for IR like in their example for AURKB and/or other PTC-introducing AS? The same could be done with CLK1 regulated AS and gene expression that is mentioned in the text later (paragraph one, subheading “CLK1 regulates AS events in genes with critical roles in cell cycle control”).

6) Blot for phospho-SR proteins: Figure 2A shows that CLK1 seems to be unique among splicing factors tested in periodic protein expression level changes. An obvious mechanism that likely contributes to the observed periodic splicing changes is altered SR protein phosphorylation, particularly since SRSF1 doesn't change protein expression (Figure 2A). Can the authors use a phospho-RS-specific antibody (1H4?) to see if the phosphorylation state of any SR proteins is altered through the cell cycle and, if so, if any of these phosphorylation changes are deadened upon CLK1 inhibition with TG003?

7) Experimentally test known IR event identified in their analysis in CLK1: Figure 2—figure supplements 1A and B show neither CLK1 mRNA levels nor inclusion of exon 4 changes significantly through the cell cycle. Boutz et al. (2014) and others cited within have described regulated intron retention (or "detained introns") of the upstream and downstream introns flanking CLK1 exon 4 that affects the transcript's localization and stability. Since periodic IR is suggested to be a common point of AS regulation through the cell cycle in this paper, showing that these introns are or are not retained across the cell cycle using additional primer sets, as was done for exon 4 skipping in Figure 2—figure supplements 1B, could further rule out splicing regulation or add another interesting layer of regulation controlling CLK1 in the context of the cell cycle. Indeed, the MISO analysis in Supplementary file 1 calls IR in this region as one of the 1747 periodic splicing events (event: chr2:201726189-201725961:-@chr2:201724469-201724403:-) so this may very well be an additional layer of regulation upstream of the kinase-dependent, ubiquitin-mediated turnover that the authors convincingly demonstrated in Figure 2—figure supplements 1C and Figure 2C–E. Similarly, the authors could test and report results for variations of CLK1's 3'UTR.

eLife. 2016 Mar 25;5:e10288. doi: 10.7554/eLife.10288.030

Author response

Essential revisions:

1) There were major concerns with regards to the poor overlap of the two pipelines (VAST and MISO) in the identification of AS events. While the reviewers and myself appreciate that the point of the manuscript is not a systematic comparison of both software, the authors are expected to explain the rationale for decisions taken which likely impact the interpretation of the results. Greater detail in the bioinformatics analysis is expected in the revision. Also, why are both tools used? And given the relatively poor overlap, is the use of either or both of these tools justified?

The MISO and VAST-TOOLS pipelines utilize different statistical frameworks for the assessment of AS events, which result in different sensitivities for detection of AS events. More importantly, different libraries of annotated AS events are also utilized by these pipelines, such that they capture overlapping but also distinct events. For example, VAST-TOOLS includes annotations for microexons and a larger set of retained introns than MISO. Accordingly, employing the two pipelines provides a more comprehensive survey of periodic AS, whereas a complete overlap in the detected AS events is not expected. Importantly, we observe a strong correlation between PSI values for periodic AS events detected by both pipelines (Spearman’s Rho > 0.8, p value < 10-16). We have clarified these points in paragraph one, subheading “Alternative splicing is coordinated with different cell cycle phases” and Figure 1—figure supplement 1G in the manuscript of the revised manuscript and below in response to Reviewer 2 and Author response image 3.

Author response image 3. (A) Scatter plot representation of raw PSI values called by VAST- TOOLS and MISO pipelines.

Author response image 3.

Open in a new tab

(B) Heat map representation of possible pair-wise correlation between commonly detected exons by VAST-TOOLS and MISO. Note scale ranges from Spearman’s Rho 0.75-1.0. (C) Scatter plot representation of periodic scores of AS events detected by MISO and VAST- TOOLS pipelines. Spearman’s Rho is shown above with a P value of > 2.2e-16. (D) Boxplot representation of periodic scores measured by VAST-TOOLS or MISO for events detected for events detected as periodic by the different pipelines (statistical significance was measured by Kolmogorov–Smirnov test, *** = 1e-16).

DOI: http://dx.doi.org/10.7554/eLife.10288.024

It is also worth noting that raw read data were mapped, filtered and analyzed for AS and gene expression independently using MISO and VAST-TOOLS (see Methods). The fact that we obtain very similar results with both pipelines indicates that our overall results are robust and can be recapitulated using independent pipelines. We therefore respectfully disagree with one of the reviewers that the “poor overlap puts the entire genome wide mapping presented here into questioning.”

2) A more complete treatment of the contribution of other splicing factors to cell cycle AS is necessary. I do not believe it has to be exhaustive, but in its present form, the manuscript led to concerns by the reviewers that CLK1 is not necessarily a clear outlier (and therefore candidate) in terms of its regulation, compared to other factors.

Our study focuses in CLK1 for the following reasons: (1) It shows a periodic expression pattern; (2) It affects a large number of AS events by modifying multiple SR proteins that are key splicing factors; and (3) Its levels are controlled by a self-inhibitory feedback circuit that likely accounts for its periodic regulation. While we fully agree that CLK1 is not the only key regulator of cell cycle dependent splicing, an in depth, comprehensive analyses of the role of other splicing regulators in periodic regulation of AS during the cell cycle is beyond the scope of this study.

Nevertheless, to address the reviewers’ comments we have included a new analysis in our manuscript that systematically examines all annotated RNA-binding proteins for periodic expression patterns that may relate to cell cycle dependent splicing regulation. The results from these new analyses are described on page 7 and 8, summarized in Figure 2—figure supplement 1 and Table 1 in the manuscript, and additional information is provided below for the reviewers’ consideration.

Analysis of our RNA-seq data revealed that 96 RNA binding proteins (RBPs) with periodic mRNA expression, and include RS domain-containing factors like SRSF2, SRSF8, TRA2A and SRSF6 (Author response image 1A). These 96 RBPs were significantly enriched in the GO term “splicing regulation” (adjusted P = 10-4, Author response image 1B), indicating that periodic AS is likely controlled by multiple RBPs. Correlations between these RBPs and periodic splicing events were also identified (Author response image 1C), SRSF2 expression for example, significantly correlated with the splicing of the pattern of a retained intron in the SRSF2 transcript. Further supporting a role for these RBPs in controlling periodic splicing was the identification of RNA motifs bound by a subset of periodically expressed RBPs (Author response image 1D).

Author response image 1. (A) Heat map representation of RNA-bound proteins (RBPs) with periodic expression.

Author response image 1.

Open in a new tab

Row-normalized FPKM levels are shown. (B) GO analyses for the functional enrichment in the periodic RBPs. (C) Number of periodic AS events that significantly correlate (Spearman’s Rho > [.75], P <..05) with the expression pattern of each RBP during cell cycle. Expression pattern of two known splicing factors, SRSF2 and ESRP2, is shown in inset. (D) Average PSI values of periodic that peak at either G1 (red line) or M phase (blue line) (top panel). k- mer enrichment in periodic exons as judged by Z score and separated by cell cycle phase (y-axis = G1-S and x-axis = G2-M) (bottom panel).

DOI: http://dx.doi.org/10.7554/eLife.10288.022

CLK4 modification and levels should be certainly verified by Western blot analysis. Based on these new results (in the revision), the authors are encouraged to alter/broaden their title (as suggested by Reviewer 3 as well).

We have repeatedly attempted to measure the levels of CLK4 by immunoblot. However a reliable antibody to CLK4 is not available.

The reviewers are correct in pointing out that the small molecules utilized in our study likely also inhibit the activity of CLK4. However, shRNA mediated depletion of CLK4 did not reveal significant alteration in cell cycle-associated phenotypes (whereas knockdown of CLK1 did; see below), which is another reason why we focused our efforts on CLK1. Furthermore, in our subsequent kidney tumor analysis, CLK4 expression did not correlate with patient outcomes, whereas CLK1 expression did.

3) The authors should address the major concerns of Reviewers 1 and 2 with regards to why is the CLK1 inhibitor used (rather than depletion) and address clearly their interpretation of specificity since the inhibitors also target CLK4.

The main advantage of employing CLK inhibitors vs. RNAi or CRISPR depletion is that they have rapid effects, which is critical for studying temporal control of splicing as it causes rapid changes in splicing (Figure 3—figure supplement 1D). Moreover, due to the self- regulatory feedback of CLK1, its efficient depletion in the context of synchronization experiments was not possible with siRNA pools. As such, we found that the only viable way to assess the activity of the kinase in cell cycle control was to use available chemical inhibitors. Nevertheless, as mentioned above, we show in the manuscript that knockdown of CLK1 using a shRNA phenocopies cell cycle defects observed with both inhibitors, providing evidence that the effects of these inhibitors are due to CLK1 inhibition.

4) The authors should address queries with regards to the PTC-induced degradation of AURKB (Reviewer 1) and known IR events (Reviewer 3).

We carried out additional experiments to address this point. As expected, AURKB intron retention isoforms significantly increase upon treatment with cyclohexamide or an inhibitor of the NMD pathway.

Reviewer #1:

Specific comments:

1) In the Abstract and again in paragraph two of the Introduction, the authors refer to periodic regulation of gene function at the levels of transcription, protein modification and degradation, etc.; however, they ignore the role of mRNA translation. There are several papers describing the regulation of mRNA translation during mitosis, which would enrich the Discussion section. Some of these include:

Stumpf. et al. (2013) The translational landscape of the mammalian cell cycle. Mol Cell. 2013 Nov 21;52(4):574-82. PubMed PMID: 24120665.

Maslon et al. (2014) The translational landscape of the splicing factor SRSF1 and its role in mitosis. eLife. 2014 May 6:e02028. doi: 10.7554/eLife.02028. PubMed PMID: 24842991.

We apologize for this oversight, and have now included this point and cited the references in the new Discussion (paragraph two, Discussion).

2) Is there any overlap in those genes found here to be regulated at the AS level and those previously reported to be regulated by mRNA translation?

See our response above.

3) There seems to be independent regulation of transcriptionally and AS-regulated genes during the cell cycle. Are those genes that show an overlap (133) more likely to be regulated by the kinetic control mode of AS regulation? In other words, are those genes preferentially regulated by the processivity of RNA pol II?

This is a very insightful comment, which raises the possibility that the 133 overlapping genes may be periodically spliced as a result of slower or faster polymerase movement across the gene during transcription. According to the conventional kinetic model, certain genes may be transcribed more rapidly in a specific cell cycle stage, which could potentially modulate AS. Unfortunately, we currently lack measurements for pol II elongation rates in the corresponding genes making it difficult to address the reviewer’s comment at the present time. Related to this question are additional analyses regarding the correlation between a gene’s expression and alternative splicing during cell cycle (please see our response to Reviewer 3, point 3).

4) On Figure 1E, has the PTC-induced degradation of AURKB been experimentally determined?

We have addressed this point and included a supplementary figure on this. Please see the essential revision point #4 (and Author response image 2) for details.

Author response image 2. AURKB splicing isoform with retained intron that contains a pre-mature stop codon is indeed a NMD target.

Author response image 2.

Open in a new tab

The inhibition of NMD causes accumulation of the intron-containing isoform.

DOI: http://dx.doi.org/10.7554/eLife.10288.023

5) On Figure 3, why is a CLK1 inhibitor preferred to a depletion of CLK1 (either via CRISPR or siRNA approach)?

Please see response above.

6) On Figure 3, is the CENPE AS also regulated by SRSF1, as is the case with the CHEK2 pre-mRNA?

This is an interesting possibility but we feel that it is beyond the scope of the present study.

Reviewer #2:

This manuscript by Dominguez et al., describes a thorough and important analysis of transcriptome changes during the cell cycle, with particular emphasis on alternative splicing. This analysis then motivates the authors to investigate a potential role of the kinase CLK1 in regulating cell cycle splicing and progression, and its possible link to cancer. If fully substantiated, this work would add tremendously to the growing body of knowledge regarding how different cellular growth conditions impact splicing. However, some additional analysis and important controls are required for this manuscript to achieve its full impact.

Firstly, much of the analysis of the RNA-Seq data is poorly defined. Why were VAST-TOOLs and MISO both used and what is the advantage of this? What is "Periodic score" in Figure 1—figure supplement 1B and C? The authors may feel these are readily apparent, but readers should not have to dig deeply through the Methods to understand the first figure. More importantly, Figure 1A should either be replaced with a version that gives raw PSI, or a version with raw PSI should be included in the supplement. Use of the "row normalized" method raises concerns that many of the changes observed in AS events are modest and potentially not of biologic significance.

Please see our response to Dr. Yeo’s essential revision #1. We have included a figure with raw PSI values below as well as added data regarding the methodology of periodic score.

Secondly, the justification for focusing on CLK1 is not clear. While Figure 3 and Figure 4 certainly support a role of the CLK family in cell cycle splicing, Figure 2 is not convincing that CLK1 is a clear outlier in terms of its regulation. A quick glance through the GE values in supplemental table 1 indicates that genes encoding several SR proteins vary throughout the cell cycle. Have SR proteins other than SRSF1 been analyzed by Western? Moreover, CLK4 should be certainly tested by Western – due to it being a target of TG003 and KHCB-19 (see next point).

Finally, there is significant concern that both the inhibitors used target CLK4 as well as (or more potently than) CLK1. The authors should either conclusively rule out a role for CLK4, or broaden their conclusions to acknowledge a potential role for this family member.

Please see our response to the essential revision point #3.

Reviewer #3:

In this work the authors create a genome wide map of oscillating alternative splicing events, and identify CLK1 as a major regulator of a subset of these oscillating events. Roughly, two thirds the paper are focused on exploring the regulatory role of CLK1 through a series of elaborate and thoughtful experiments after establishing it plays a major role in regulating oscillating AS. We found relatively few and mostly minor concerns regarding these later sections. Most critical concerns regard the first, genome wide, analysis.

1) The two pipelines used, based on VAST and MISO, are basically in severe disagreement with respect to which AS events are oscillating. Figure 1—figure supplement 1F shows that not only is the number of events detected very different (244 vs 513) but the overlap of those is a mere 27% or 12%(!). This is coming from two pipelines that use the same downstream procedure to check for oscillation and differ only in their approach to quantify the raw PSI values fed into the pipeline. This poor overlap puts the entire genome wide mapping presented here into questioning. Figure 1—figure supplement 1G may have been added to address this concern (it's not explained) but actually does little to alleviate this concern: Figure 1—figure supplement 1G basically shows that the correlation between VAST and MISO based PSI values are much better for the same condition (diagonal) than for different conditions when looking across the >4K events quantified. That is actually to be expected: most events are either highly included or highly excluded. Such highly included/excluded events would (a) make the extreme values dominating the correlation coefficient and (b) would make the majority of cases that are generally not changing and thus both methods agree on. In fact, a closer look at Figure 1—figure supplement 1G shows the average PSI correlation between VAST and MISO for the same condition (diagonal) is ~0.55 which is again troubling (btw, setting the dark red color for 0.55 is misleading). Some suggestions that may help alleviate the discrepancies between the two pipelines are given below.

Related to this: why does Figure 1—figure supplement 1D contain only 513 events but Figure 1A contains 1747 events? If Figure 1—figure supplement 1D is VAST's list why does Figure 1—figure supplement 1F show 513 for MISO?

We have adjusted the color of the heat map scale in Figure 1—figure supplement 1G.

The Figure 1—figure supplement 1G heat map representation of the sample correlations is actually comparing normalized PSI (normalized across cell cycle), which is why higher positive correlations between samples in the cell cycle stage are observed. This was meant to demonstrate that even though the overlap between the two pipelines is not perfect, there is still a significant correlation between these two pipelines. Furthermore, it demonstrates that even when using 4,000 AS events periodicity can still be observed in aggregate.

For clarity we have now included the correlation by raw PSI value (which is much higher >0.8). When low and high PSI values are excluded (0.2 >PSI<0.8) the correlation between VAST-TOOLS and MISO are still positive and very significant (~0.5-0.6 Spearman’s Rho), showing that the positive and significant correlation between our AS measurements are not solely driven by high and low PSI values.

Furthermore, we found a positive correlation between periodic scores for events detected by both pipelines. In addition, AS events called periodic by one pipeline but missed by the other have significantly lower periodic scores (as determined by either pipeline) than events not identified as periodic in either analysis, suggesting differences in sensitivity (Figure 1D).

For more information regarding our rationale for the use of both VAST-TOOLS and MISO see the response to the editor above.

Please see our response to the essential point #1 for the related questions.

2) Basic "normalized PSI" (subheading “Identification of periodic AS”) means we are only looking at relative changes of PSI. This may introduce a lot of variability/noise to the analysis and may contribute to the substantial differences between the results from the two analysis pipelines (see above). The authors may be better off screening for events for which (a) min(max(PSI) > VAL1) and (b) max(PSI) –min(PSI) > VAL2. Similarly, increasing the threshold on number of reads per event to be included could help. For example, using VAL1=VAL2 = 20% could help avoid fluctuations that may just appear as periodic or changing dramatically in a relatively small/insignificant range.

In the main figures we used relative PSI values for presentation purposes, as it more clearly shows the data and patterns of periodicity. Below we show a heat map representation of raw delta PSI. That is, we are showing (PSI valuei – min(PSI val1-14)) for each event. This (a) represents the raw data and (b) shows the magnitude of change in raw PSI value that the reviewer has requested.

Author response image 4. Different version of the heat map representation of periodically spliced events.

Author response image 4.

Open in a new tab

Raw PSI values are shown. Diagram below indicates cell cycle phase. The types of AS events were also indicated.

DOI: http://dx.doi.org/10.7554/eLife.10288.025

Related to this: the authors do not explain how events are screened for coverage across the experiments.

We are showing only co-detected events. As stated above different mapping and annotation libraries were used for MISO and VAST-TOOLS.

3) The method for detecting the oscillating AS events and its effect on their results is not discussed or evaluated. Specifically, the authors use 7 previously characterized profiles from GE analysis. It's not clear whether the same kind of profiles is the best choice for analyzing periodic AS. For example, profiles 3,4 are similar in phase but one has a wider shape –

why would that be the best fit for AS profiles? It seems reasonable to at least compare the results from this approach to an unbiased approach where the entire set of AS events are clustered with no specific periodic profiles and then the most prominent periodic profiles are extracted or executing a more directed search against theoretical periodic profiles.

The difference in width of these curves came from analysis on periodic mRNA expression patterns (please see associated manuscript) where known and clearly periodic mRNAs were missed in our first pass analysis due to wider or narrower shapes expression curves. We dealt with this by including wider and narrower periodic seeds.

We also attempted an unbiased clustering approach for both AS and gene expression (this was the easiest first-pass analysis of the data), this analysis revealed very interesting results. However, we have detected ~30,000 AS events making it very difficult to de-convolute the clusters. We also found it hard to assign a specific “score” to AS events belonging to a cluster. Furthermore, we employed a Fourier transform in some of our preliminary analysis to identify periodic expression patterns as was previously reported by Whitfield et al. (2002), and found similar results. We did feel that the Fourier transform method likely works best when more sampling times are taken along the time course, which was a practical and financial limitation to our sequencing based approach. Below is a heat map representation of hierarchical clustering on all detected events. The bar to the right represents events identified as periodic (in red) by our current analysis.

Author response image 5. Unbiased hierarchical cluster of the PSI values of all AS events during cell cycle.

Author response image 5.

Open in a new tab

DOI: http://dx.doi.org/10.7554/eLife.10288.026

We appreciate the reviewer’s suggestions and would like to reiterate that we have attempted a variety of approaches to analyze these data and chose a concise presentation style to highlight our main point. We feel that an exhaustive comparison of methods to determine periodic AS/gene expression patterns is beyond the scope of this work.

4) More convincing/exhaustive search of RBPs that may contribute to cell cycle AS:

While there's significant overlap between CLK1 inhibitor AS events and periodic cell cycle events, this only explains a fraction of all the periodic events they find. It would be good to acknowledge at least that there are likely other factors out there that contribute to this program of AS. The authors should discuss how much of the oscillation signal is explained by CLK1. The WB data from Figure 2A is not exhaustive, but it is convincing for those factors tested. It should be straightforward to supplement this analysis by applying their pipelines and RNA-seq data to report known/suspected splicing factors and/or RBPs that change at the level of GE or AS. Any changes found in this analysis could potentially contribute to the large AS program they observe beyond CLK1 regulation.

This comment was addressed by additional data and new Figure 2—figure supplement 1 (Author response image 2). Please see the response to essential revision point #2.

5) Analysis on IR/AS-PTC introducing events:

Figure 1D/E suggests that periodic IR can affect transcript expression levels (probably by the introduction of PTCs) in the case of AURKB and some other genes they examined by qRT-PCR that are important for cell cycle progression. However Figure 1B argues little overlap between splicing regulation and mRNA expression levels, save for 133 overlapping events. Are the events in the overlap of periodic AS and periodic gene expression overrepresented for IR like in their example for AURKB and/or other PTC-introducing AS? The same could be done with CLK1 regulated AS and gene expression that is mentioned in the text later (paragraph one, subheading “CLK1 regulates AS events in genes with critical roles in cell cycle control”).

We have verified that AURKB is indeed an NMD target (essential revision #4). Although the overlap is small, we addressed this point and show a table with the overlapping event types (see our response to the other comment #3 of this same reviewer). As you can see here, there is no specific enrichment for any of the four AS types shown here. As suggested by the reviewer the same overlap analysis was carried out for CLK1-regulated events.

6) Blot for phospho-SR proteins: Figure 2A shows that CLK1 seems to be unique among splicing factors tested in periodic protein expression level changes. An obvious mechanism that likely contributes to the observed periodic splicing changes is altered SR protein phosphorylation, particularly since SRSF1 doesn't change protein expression (Figure 2A). Can the authors use a phospho-RS-specific antibody (1H4?) to see if the phosphorylation state of any SR proteins is altered through the cell cycle and, if so, if any of these phosphorylation changes are deadened upon CLK1 inhibition with TG003?

We have attempted this. Unfortunately the results were not consistently reproducible in cell synchrony experiments. Furthermore, we obtained different results when we used mab104 and mab1H4. In addition, conflicting results in terms of changes in the pattern of SR protein phosphorylation during cell cycle were reported previously by Fu and Manley’s labs using the same antibody and cell line.

7) Experimentally test known IR event identified in their analysis in CLK1: Figure 2—figure supplements 1A and B show neither CLK1 mRNA levels nor inclusion of exon 4 changes significantly through the cell cycle. Boutz et al. (2014) and others cited within have described regulated intron retention (or "detained introns") of the upstream and downstream introns flanking CLK1 exon 4 that affects the transcript's localization and stability. Since periodic IR is suggested to be a common point of AS regulation through the cell cycle in this paper, showing that these introns are or are not retained across the cell cycle using additional primer sets, as was done for exon 4 skipping in Figure 2—figure supplements 1B, could further rule out splicing regulation or add another interesting layer of regulation controlling CLK1 in the context of the cell cycle. Indeed, the MISO analysis in Supplementary file 1 calls IR in this region as one of the 1747 periodic splicing events (event: chr2:201726189-201725961:-@chr2:201724469-201724403:-) so this may very well be an additional layer of regulation upstream of the kinase-dependent, ubiquitin-mediated turnover that the authors convincingly demonstrated in Figure 2—figure supplements 1C and Figure 2C–E. Similarly, the authors could test and report results for variations of CLK1's 3'UTR.

Our data support a model in which CLK1 levels and activity change during the cell cycle. We agree with this reviewer that intron retention of CLK1 may very well be another form of regulation during cell cycle, but feel that investigating this aspect of CLK1 regulation is beyond the scope of what is already a result-rich manuscript.

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    Supplementary file 1. Periodic splicing events identified during the HeLa cell cycle and associated GO terms.

    Data associated with Figure 1.

    DOI: http://dx.doi.org/10.7554/eLife.10288.018

    elife-10288-supp1.xlsx (9.8MB, xlsx)
    DOI: 10.7554/eLife.10288.018
    Supplementary file 2. AS events altered by CLK1 inhibition (TG003 10 μM for 16 hr) in 293T cells and associated GO terms.

    Data associated with Figure 3.

    DOI: http://dx.doi.org/10.7554/eLife.10288.019

    elife-10288-supp2.xlsx (201.4KB, xlsx)
    DOI: 10.7554/eLife.10288.019
    Supplementary file 3. AS events altered by CLK1 inhibition (TG003 10 μM, KHCB19 10 μM) in synchronized HeLa cells.

    Data associated with Figure 3.

    DOI: http://dx.doi.org/10.7554/eLife.10288.020

    elife-10288-supp3.xlsx (2.1MB, xlsx)
    DOI: 10.7554/eLife.10288.020
    Supplementary file 4. List of PCR primers, shRNA sequences and antibodies used in this study.

    DOI: http://dx.doi.org/10.7554/eLife.10288.021

    elife-10288-supp4.xlsx (12.4KB, xlsx)
    DOI: 10.7554/eLife.10288.021

    Articles from eLife are provided here courtesy of eLife Sciences Publications, Ltd

    RESOURCES