Biological validation that SF3b is a target of the antitumor macrolide pladienolide


A. Yokoi, Eisai Co., Ltd, Tokodai 5-1-3, Tsukuba, Ibaraki 300-2635, Japan
Fax: +81 29 847 2759
Tel: +81 29 847 5808


Pladienolide is a naturally occurring macrolide that binds to the SF3b complex to inhibit mRNA splicing. It has not been fully validated whether the splicing impairment is a relevant mechanism for the potent antitumor activity of pladienolide. We established pladienolide-resistant clones from WiDr and DLD1 colorectal cancer cells that were insensitive to the inhibitory action of pladienolide on cell proliferation and splicing. An mRNA-Seq differential analysis revealed that these two cell lines have an identical mutation at Arg1074 in the gene for SF3B1, which encodes a subunit of the SF3b complex. Reverse expression of the mutant protein transferred pladienolide resistance to WiDr cells. Furthermore, immunoprecipitation analysis using a radiolabeled probe showed that the mutation impaired the binding affinity of paldienolide to its target. These results clearly demonstrate that pladienolide exerts its potent activity by targeting SF3b and also suggest that inhibition of SF3b is a promising drug target for anticancer therapy.

Structured digital abstract






precursor messenger RNA


small nuclear ribonucleoprotein


Pladienolide is a 12-membered macrolide that was first isolated from Streptomyces platensis Mer-11107 during a cell-based assay that evaluated the suppression of hypoxia-induced gene expression controlled by the human vascular endothelial growth factor promoter [1,2]. Pladienolide B has been shown to arrest cell-cycle progression during the G1 phase and the G2/M transition of the cell cycle in vitro, and also to inhibit tumor growth in several human cancer xenograft models in mice [3]. COMPARE analysis with panel screening of 39 human cancer cell lines indicated that pladienolide B has a unique mode of antitumor action, unlike that of the anticancer drugs currently in clinical use [4]. Photoaffinity-labeling studies identified splicing factor SF3b as a major binding target molecule for pladienolide [5]. SF3b is a key component of the U2 small nuclear ribonucleoprotein (snRNP) complex that is responsible for the splicing of precursor messenger RNA (pre-mRNA) and the formation of mature mRNA. Recently, E7107, a synthetic derivative of pladienolide D, was shown to prevent the tight binding of U2 snRNP to pre-mRNA, resulting in the formation of defective spliceosomes. E7107 was also shown to impair an ATP-dependent remodeling event in U2 snRNP that exposes the branchpoint binding region [6].

SF3b is a 450 kDa complex that comprises seven subunits: SF3B1, SF3B2, SF3B3, SF3B4, SF3B5, SF3B14 and PHF5A [7–10]. At 155 kDa, SF3B1 is the largest subunit, and it cross-links to pre-mRNA both 5′ and 3′ of the branchpoint [11]. The N-terminal 450 amino acid region of SF3B1 functions as a scaffold to facilitate interaction with other splicing factors such as U2AF65 and SF3B14 [12]. The C-terminal region of SF3B1 contains 22 tandem repeats of the HEAT motif [13]. A single HEAT motif consists of approximately 40 amino acids forming two antiparallel α-helices, and tandem repeats of the motif are present in a variety of proteins, such as protein phosphatase 2A, importin-β and eukaryotic initiation factor 4G [14–16]. The HEAT repeats of SF3B1 meander around the SF3b complex, enclosing SF3B14 [17].

Although the binding affinities of pladienolide derivatives to SF3b correlate with their inhibition of cell proliferation [5], the precise antitumor mechanism of the compounds has not been fully elucidated. In the present study, we show that a R1074H mutation in the gene for SF3B1 confers resistance to the inhibitory action of pladienolide on cell proliferation and splicing by impairing the ability of pladienolide to bind to the SF3b complex. We also present evidence that the targeting of SF3b is the direct mechanism for the antitumor activity of pladienolide.


Establishment of pladienolide-resistant cells

To elucidate the antitumor mechanism of pladienolide, we established pladienolide-resistant (-R) cells using WiDr and DLD1 human colorectal cancer cell lines. WiDr-R and DLD1-R clones were obtained from their parental cells by stepwise selection with increasing concentrations of pladienolide B and E7107, respectively, followed by limiting dilution cloning. Pladienolide B showed potent antiproliferative action against the parental WiDr and DLD1 cells, with IC50 values of 0.5 and 8.5 nm, respectively; however, WiDr-R and DLD1-R cells continued to proliferate even when treated with 100 nm pladienolide B (Fig. 1B). We also confirmed that WiDr-R and DLD1-R cells exerted cross-resistance against E7107 (Fig. S1). To further characterize the resistant cell lines, we assessed the pladienolide-induced inhibition of splicing. After 4 h of exposure to pladienolide B, the amounts of unspliced mRNA of DNAJB1, CDKN1B, RIOK3 and BRD2 were measured using quantitative PCR analysis. Pladienolide B significantly increased the unspliced forms of the mRNA in WiDr and DLD1 cells in a dose-dependent manner. However, accumulation of the unspliced mRNAs was not observed in WiDr-R and DLD1-R cells treated with pladienolide B (Fig. 2). These results demonstrate that the WiDr-R and DLD1-R cells are resistant to both the cell growth suppression and splicing inhibition of pladienolide.

Figure 1.

 Establishment of pladienolide-resistant cell lines. (A) Structures of pladienolides B and D, 3H probe and E7107. (B) Antiproliferative action of pladienolide B against WiDr/WiDr-R and DLD1/DLD1-R cells. Cell viability was assessed with a colorimetric assay after 72 h of treatment with pladienolide B. Each data point represents the mean ± SD from three separate experiments, each performed in triplicate.

Figure 2.

 Splicing inhibition by pladienolide B. (A) WiDr/WiDr-R cells and (B) DLD1/DLD1-R cells were treated with 1, 10 or 100 nm pladienolide B for 4 h or left untreated. Unspliced mRNA was quantified using quantitative RT-PCR using primers specific for each intron. Each data point represents the mean ± SD in triplicate.

Identification of the SF3B1 mutation in pladienolide-resistant cells

To examine the mechanism of resistance in WiDr-R and DLD1-R cells, we performed mRNA-Seq differential analysis using the Illumina Genetic Analyzer (Illumina, San Diego, CA, USA). Deep sequencing was used to compare mRNA sequences in paired parental and resistant WiDr and DLD1 cells. The two sets of analyses generated 98 and 83 Mb sequences, respectively, which each aligned with the human reference sequence, and identified four and 87 mutations specific to the resistant lines in WiDr/WiDr-R and DLD1/DLD1-R pairs, respectively (Fig. 3A). One mutation in the gene for SF3B1 that results in the replacement of Arg1074 with a His was common between the two resistant clones. The R1074H mutation was confirmed using the Sanger sequencing method, which demonstrated that the mutant mRNA was heterozygously expressed in WiDr-R cells but was homozygously expressed in DLD1-R cells (Fig. 3B).

Figure 3.

 R1074H mutation detected in the gene for SF3B1. (A) Venn diagram describing the number of mutations specific to each resistant line from the mRNA-Seq differential analysis, in which mRNA sequences were compared between parental and resistant pairs of WiDr or DLD1 cells. (B) Illustration of the sequencing results. Conversion of CGT into CAT in codon 1074, which results in the replacement of an Arg with a His, was confirmed using capillary sequencing.

The SF3B1 mutation confers pladienolide resistance

To investigate whether the R1074H mutation in the gene for SF3B1 directly confers pladienolide resistance to cells, we performed a reverse-expression analysis of the mutant protein using parental WiDr cells. Accordingly, two types of viral vectors were prepared: (a) an SF3B1 expression vector, which expressed wild-type or R1074H-type hemagglutinin (HA)-tagged SF3B1, and (b) an SF3B1 knockdown vector, which targeted the 3′-UTR sequences of SF3B1 to decrease the endogenous expression of the gene (Fig. 4A). After transduction of both expression and knockdown viral vectors into WiDr cells, we confirmed the expression of the HA-tagged SF3B1 protein by western blot analysis using anti-HA serum (Fig. 4B). Knockdown of the endogenous SF3B1 expression was also evaluated by quantitative PCR analysis using primers designed for the SF3B1 3′-UTR (Fig. 4C), and the results obtained showed that most of the endogenous SF3B1 protein was replaced with the exogenous HA-tagged SF3B1. The effect of pladienolide B on the cell growth of these transduced WiDr cells was investigated. Pladienolide B showed potent action with IC50 values of 0.8 nm in wild-type SF3B1 expressed/endogenous SF3B1-knockdown cells (Fig. 5A and Fig. S2A). This value was almost identical to that in nontransduced WiDr cells (Fig. 1B), demonstrating that exogenous expression of SF3B1 did not affect the sensitivity to pladienolide B. By contrast, the R1074H SF3B1 expressed/endogenous SF3B1-knockdown cells showed complete resistance to cell growth suppression by pladienolide B. We further assessed the inhibition of splicing by pladienolide B in these transduced cells. Accumulation of unspliced pre-mRNA of DNAJB1, CDKN1B and RIOK3 was not observed in R1074H-expressing cells (Fig. 5B and Fig. S2B). These results demonstrate that reverse expression of SF3B1 (R1074H) transferred pladienolide resistance to the cells.

Figure 4.

 Establishment of WiDr cells expressing R1074H SF3B1. WiDr cells were transduced by the SF3B1 expression viral vector pCLXIH-SF3B1 [wild-type (Wt) or R1074H] and the endogenous SF3B1 knockdown viral vector pLenti/shSF3B1 (3989 or 4145). (A) Diagram of the SF3B1 (Wt/R1074H) expression viral vectors and the endogenous SF3B1 knockdown viral vectors. HA-SF3B1 indicates a HA-tagged SF3B1 cDNA. shSF3B1 indicates a shRNA template, which targets the 3′-UTR of SF3B1. Targeting sequences of shSF3B1 (3989) and shSF3B1 (4145) correspond to nucleotides 3989–4009 and 4145–4165 in the gene for SF3B1, respectively. (B) Expression of SF3B1 was detected using anti-HA and anti-SF3B1 sera. pLenti/shLuc was used as a negative control for the knockdown viral vector. (C) Decreased expression of endogenous SF3B1 was revealed using quantitative RT-PCR using primers designed for the 3′-UTR region of the gene for SF3B1.

Figure 5.

 Reverse expression of SF3B1 (R1074H) transferred the pladienolide resistance to WiDr cells. (A) Antiproliferative activity of pladienolide B against WiDr cells transduced by pCLXIH-SF3B1 [wild-type (Wt) or R1074H] and pLenti/shSF3B1 (3989). Cell viability was assessed after 72 h of treatment with pladienolide B using a colorimetric assay. Each data point represents the mean ± SD from three separate experiments, each performed in triplicate. (B) Splicing inhibition by pladienolide B in WiDr cells transduced by pCLXIH-SF3B1 (Wt or R1074H) and pLenti/shSF3B1 (3989). Cells were treated with 100 nm of pladienolide B for 4 h. Unspliced DNAJB1 mRNA, CDKN1B mRNA and RIOK3 mRNA were quantified using quantitative RT-PCR using primers specific for each intron. Each data point represents the mean ± SD in triplicate.

The SF3B1 mutation impairs the binding of pladienolide to SF3b

To further understand the mechanism of resistance in pladienolide-resistant cells, we studied the interaction of pladienolide with the SF3b complex containing mutant SF3B1 using a radiolabeled probe. First, we immunoprecipitated the SF3b complex from the nuclear fraction of DLD1 and DLD1-R cells with anti-SF3B1 serum. Western blot analysis showed that the SF3b subunits SF3B1, SF3B2 and SF3B3 were coprecipitated equally from DLD1 and DLD1-R cells (Fig. 6A). Next, we treated the immunoprecipitates with a 3H probe and evaluated the 3H signal using scintillation counting. The probe is a radiolabeled derivative of pladienolide B, which was shown to compete with a series of pladienolide for binding to the SF3b complex [5]. Comparison of the radioactivity between the DLD1 and DLD1-R cells showed that the immunoprecipitate from the DLD1 cells retained 27 times the amount of 3H probe retained in DLD1-R cells (Fig. 6B). This demonstrated that the mutant SF3B1 protein in DLD1-R reduced the ability of the probe to bind to the SF3b complex.

Figure 6.

 The R1074H mutation in SF3B1 impairs the binding of pladienolide to SF3b. (A) Immunoprecipitation of the SF3b components. Anti-SF3B1 serum was used to immunoprecipitate (IP) DLD1 or DLD1-R cells. Immunoblots were performed using anti-SF3B1, anti-SF3B2 and anti-SF3B3 sera. Asterisks indicate proteins detected nonspecifically. The molecular weight is indicated to the left. (B) Radioactivity in the immunoprecipitated nuclear fraction. Samples were immunoprecipitated with anti-SF3B1 serum and treated with 100 nm3H probe for 1 h. The 3H signals of the immunoprecipitated samples were measured using a liquid scintillation analyzer.


In the present study, we have shown that the targeting of SF3b is the direct mechanism of the antitumor activity of pladienolide. We first established pladienolide-resistant lines (i.e. WiDr-R and DLD1-R) that were resistant to the cell growth suppression and splicing inhibition actions of pladienolide. The mRNA-Seq differential analysis showed that the WiDr-R and DLD1-R cells had a common mutation in the gene for SF3B1 that resulted in the replacement of Arg1074 with a His. We next confirmed that reverse expression of the mutant transferred the pladienolide resistance to WiDr cells. Furthermore, immunoprecipitation analysis with a radiolabeled probe showed that the binding affinity of pladienolide to SF3b was impaired in DLD1-R cells in which the R1074H SF3B1 was exclusively expressed. These results clearly demonstrate that pladienolide exerts its antitumor activity only by targeting SF3b.

Although both WiDr-R and DLD1-R cells exhibited resistance to pladienolide, the cell lines had different characteristics. In WiDr-R cells, a high concentration of pladienolide B suppressed cell proliferation by 30%, whereas, in DLD1-R cells, no effect was noted at any of the doses examined (Fig. 1B). The expression of both wild-type and mutant-type of SF3B1 mRNA was confirmed in WiDr-R cells, although only the mutant form was expressed in DLD1-R cells (Fig. 3B), suggesting that the heterozygous expression of the gene might confer semi-resistance in WiDr-R cells. Indeed, the transduced WiDr cells, in which R1074H SF3B1 was expressed and endogenous SF3B1 was knocked down, exhibited more resistance to the antiproliferative action of pladienolide B than WiDr-R cells (Fig. 5A and Fig. S2A).

It has been reported that WiDr cells are replication error-negative and that DLD1 cells are replication error-positive [18]. In DLD1 cells, both alleles of the gene for MSH6, which is required for the initial recognition of mismatched nucleotides, contain frameshift mutations that result in a truncated product [19]. The DNA-repair contexts of the cells appear to directly affect the number of acquired mutations in both WiDr-R and DLD1-R cells: four mutations were identified in WiDr-R cells and 87 mutations were identified in DLD1-R cells.

It is noteworthy that the two independent clones of WiDr-R and DLD1-R have an identical R1074H alteration that still permits splicing to occur. There are no reports concerning single nucleotide polymorphisms and somatic mutations at Arg1074 in the gene [20,21], suggesting that the alteration is an acquired mutation by the continuous exposure of pladienolide. The Arg1074 is conserved from Saccharomyces cerevisiae to human and located at the consensus residue 21 in the HEAT repeat of SF3B1 [13]. The function of the consensus Arg is well described in the structural study of the PP2A PR65/A subunit in that the residue is located adjacent to the inter loop of two α-helices and stabilizes both the secondary and tertiary HEAT structures by interaction with an aspartic acid located in front of six residues of the Arg [14]. The corresponding aspartic acid does not exist around Arg1074 in SF3B1. Furthermore, four Args remain at this position and three Asp-Arg pairs exist among the 22 HEAT repeats of the protein. Despite its irregular HEAT motifs with the limited number of Asp-Arg pairs, the amino acid sequence of the SF3B1 C-terminal is highly conserved among species: the identities between human and homolog SF3B1 C-terminals are: mouse, 99.8%; Xenopus, 99.5%; and Caenorhabditis elegans, 82.1% [22]. This suggests that there is little allowance for SF3B1 to change amino acids and still keep its splicing function. The replacement of Arg1074 with a His might represent a rare exception in terms of not interfering with the HEAT structure of SF3B1, although there is no His at position 21 in any other repeats of the protein.

In the present study, we have shown that the R1074H mutation in the gene for SF3B1 impaired the ability of pladienolide to bind to its target. Previously, using a photoaffinity-labeled probe experiment, we showed that pladienolide binds to SF3B3, comprising a component of the U2 snRNP that is tightly bound to SF3B1 [5]. These results indicate that pladienolide might be located in the vicinity of both SF3B1 and SF3B3. In our model, pladienolide fails to bind to SF3B3 because of physical interference by the R1074H mutation of SF3B1. In the previous study, the signal of the coprecipitated 3H probe correlated with the quantity of the SF3b complex but not with the quantity of the SF3B3 protein [5]. This suggests that the binding affinity of pladienolide might be maintained in the context of the SF3b complex including SF3B1, SF3B2 and SF3B3 [5]. Furthermore, another splicing inhibitor, spliceostatin, which also binds to SF3b and inhibits splicing, bound to SF3B1 more efficiently than to SF3B3 under strict washing conditions [23]. It is possible that these types of compounds inhibit splicing by fitting into a space between SF3B3 and SF3B1 (Fig. 7).

Figure 7.

 Model of the binding site of pladienolide to SF3b. (A) Pladienolide inhibits splicing by fitting into a space between SF3B3 and SF3B1. (B) The R1074H mutation in SF3B1 impairs the binding of pladienolide to SF3b.

The present study also indicates the binding mode of SF3B1 and SF3B3. The fact that the R1074H mutation in the SF3B1 protein impairs the binding of pladienolide to SF3B3 implies that Arg1074 of SF3B1, which is in the sixteenth HEAT motif repeat, might be located near an area of access to the SF3B3 protein. Studies aiming to characterize the HEAT repeat domain in SF3B1 are needed to further understand the structure of the SF3b complex. This will also provide insight into the mode of the interaction between the complex and pladienolide.

In conclusion, our genetic and chemical probe approaches demonstrate that a R1074H mutation in the gene for SF3B1 confers resistance to the pladienolide activity by impairing the ability of pladienolide to bind to SF3b. This finding reveals that pladienolide exerts antitumor effects by targeting SF3b and suggests that the splicing machinery is a novel type of antitumor drug target.

Materials and methods

Cell lines and pladienolide

WiDr and DLD1 human colorectal cancer cell lines were obtained from Dainippon Sumitomo Pharma (Osaka, Japan) and the American Tissue Culture Collection (Manassas, VA, USA), respectively. Cells were cultured in RPMI-1640 medium (Wako, Osaka, Japan) with 10% fetal bovine serum and 1% penicillin and streptomycin. The cells were incubated at 37 °C in a humidified atmosphere of 5% CO2. Pladienolide B, E7107 and an 3H probe were prepared as previously described [1,2,5]. To obtain pladienolide-resistant cells, WiDr and DLD1 cells were treated with increasing concentrations of pladienolide B (3–100 nm) and E7107 (30–100 nM) in the presence of 10 μm verapamil (EMD Biosciences, San Diego, CA, USA), respectively. After stepwise selection, WiDr-R and DLD1-R cell lines were isolated from the each resistant population by limiting dilution cloning. The Uniprot accession numbers associated with the proteins used in the present study are: SF3B1, O75533; SF3B2, Q13435; and SF3B3, Q15393.

Cell viability assay

Cells were plated at a density of 3 × 103 cells per well in a 96-well microtiter plate and incubated overnight. Serial dilutions of pladienolide were added into each well. After 72 h, 10 μL of WST-8 reagent (Dojindo, Kumamoto, Japan) was added to each well. A450 was monitored and compared with a reference measurement at A660 using a rainbow microplate reader (SLT Lab Instruments, Salzburg, Austria).

Quantitative PCR analysis

Total RNA was extracted using an RNeasy Mini Kit (Qiagen, Hilden, Germany) and converted to cDNA using a High-Capacity cDNA Archive Kit (Applied Biosystems Foster City, CA, USA). Gene expression levels were quantified using an ABI7900HT Sequence Detection System (Applied Biosystems). The amounts of unspliced mRNA of DNAJB1, CDKN1B, RIOK3 and BRD2 were determined using SYBR Green (Power SYBR Green PCR Master Mix; Applied Biosystems) real-time quantitative RT-PCR using specific primers designed for the intron of each gene (Table 1). The amount of endogenous SF3B1 mRNA was determined using primers designed for the 3′-UTR of the gene. PCR using 0.2 μm of each primer was performed on 10 ng of the obtained cDNA. PCR conditions were: 2 min at 50 °C, 10 min at 95 °C, 40 cycles of 20 s at 94 °C and 20 s at 55 °C, and 30 s at 72 °C. 18S rRNA was used as the control for normalization.

Table 1.   Sequences of primers used in the present study.
PrimerSequence (5′–3′)

mRNA-Seq differential analysis

Total RNA was extracted from WiDr and DLD1 cells using an RNeasy Mini Kit (Qiagen). The preparation of the mRNA-Seq library was carried out in accordance with the manufacturer’s instructions (Illumina). PolyA-RNA was purified from 2 μg of the total RNA using Sera-Mag Magnetic Oligo(dT) Beads (Illumina). The PolyA-RNA was fragmented by metal hydrolysis in fragmentation buffer (Illumina) for 5 min at 94 °C. The reaction was stopped by the addition of fragmentation stop solution (Illumina). The fragmented RNA was converted to first-stranded cDNA using SuperScript II Reverse Transcriptase (Invitrogen, Carlsbad, CA, USA) and random primers (Illumina). This was then followed by second-strand synthesis to create double-stranded cDNA fragments. The fragments were incubated with T4 DNA polymerase (Illumina) and Klenow DNA polymerase (Illumina) for 30 min at 20 °C. Adenosine overhangs were added to the blunt ends using a Klenow fragment (Illumina) for 30 min at 37 °C followed by ligation to PE Adapter Oligo Mix (Illumina) using T4 DNA Ligase (Illumina) for 15 min at room temperature. Adapter-ligated products of the desired size range were purified using a 2% Tris-acetate-EDTA agarose gel. Nine-tenths of each product was enriched by 15 cycles of PCR with Phusion DNA polymerase (Finnzymes Oy, Espoo, Finland), PCR Primer PE 1.0 and PCR Primer PE 2.0 (Illumina). The thermocycling parameters were 30 s at 98 °C; 15 cycles of 10 s at 98 °C, 30 s at 65 °C and 30 s at 72 °C; and 5 min at 72 °C. The reaction products were purified using a QIAquick PCR Purification Kit (Qiagen). DNA quality was assessed and quantified using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) and a Nanodrop 7500 spectrophotometer (Nanodrop, Wilmington, DE, USA). The DNA was diluted to 10 nm and mRNA-Seq libraries were sequenced using an Illumina Genetic Analyzer (Illumina) in accordance with the manufacturer’s instructions.

In the bioinformatics data analysis aiming to extract mutations associated with drug resistance, we applied four filters against all locations on the genome with pile-up base frequencies of A, T, G and C. The first filter was a chi-squared test, which can distinguish differences in base frequencies between two samples; in this case, the pladienolide-resistant and parental cells. The cut-off value was 0.05. The second filter was a depth filter. We chose a depth of sequence coverage of five or more for both samples. In addition to the statistical filters, we also applied a filter to detect differences in the base frequencies using the cosine coefficient of the frequencies. We set a cut-off value of 0.895, which means one-third or less overlap of frequencies between the two samples. The final filter that we applied was a nonsynonymous single nucleotide polymorphism filter to focus on the functional changes caused by mutations.

Sanger sequence validation

The SF3B1 fragment was amplified from cDNA using PCR primers flanking the R1074H mutation. Priming sites for M13 Forward −21 and M13 reverse were attached to the 5′ ends of the primers to allow direct Sanger sequencing of amplified products. After purification of the amplicons, cycle sequencing was performed using a BigDye Terminator kit, version 3.1 (Applied Biosystems). The products of the sequencing reactions were purified using a Performa DTR Ultra 96-Well Plate Kit (Edge BioSystems, Gaithersburg, MD, USA) and sequenced in a 16-capillary ABI PRISM 3130xl Genetic Analyzer (Applied Biosystems).

Plasmid construction and viral production

We constructed a synthetic humanized SF3B1 ORF (mxSF3B1) by ligating the N-terminal half of the mouse SF3B1 ORF and the C-terminal half of the Xenopus SF3B1 ORF, followed by replacement of three amino acid residues in the Xenopus ORF with three human-type residues. This was carried out because it was reported previously that cloning of human SF3B1 cDNA is impossible [13] and also because we have experienced similar difficulties with mouse SF3B1 cDNA. The mouse SF3B1 ORF was amplified from mouse brain Quick-Clone cDNA (Clontech, Mountain View, CA, USA) using FastStart HiFi DNA polymerase (Roche Diagnostics, Mannheim, Germany) and a primer pair of T013-FB and T021-RH. A mouse SF3B1 ORF of approximately 4 kb was successfully amplified, although it was difficult to clone into vectors as a result of its toxicity in Escherichia coli. Therefore, the N-terminal half of mouse SF3B1, a fragment of approximately 1.5 kb obtained by BamHI-XhoI digestion, was cloned into a pBluescript II SK(+) vector (Stratagene, La Jolla, CA, USA). The C-terminal half of the gene for Xenopus SF3B1 was cloned by PCR using the Xenopus laevis embryo LAMBDA cDNA library (Stratagene) and a primer pair of xenoSAP155-F1/Bm and xenoSAP155-R1 (Table 1). The PCR product was cloned into a pBluescript II KS(+) vector (Stratagene) using BamHI and HindIII restriction sites. The plasmid, designated pCMVHA-mxSF3B1, was constructed by insertion of DNA fragments into the pcDNA3.1(−) vector (Invitrogen): the HA-Nhe-Bam-adaptor fragment (5′-GCTAGCACCATGAGAGGTTCGAACTACCCCTACGACGTGCCAGACTACGCTTCCCTGGGATCACTCGAGGAATTCGTCGACGGTACCGGGCCCGGATCC-3′) containing an initiation codon followed by the HA-epitope (YPYDVPDYA); the N-half of mouse SF3B1 (NheI-ScaI) fragment; and the C-half of the Xenopus SF3B1 (ScaI-HindIII) fragment. The resulting plasmid was used as a template for the site-directed mutagenesis to convert the Xenopus-unique amino acids. Three amino acids in the Xenopus SF3B1 (i.e. Val1247, Glu1291 and Thr1303) were changed to the human-type Leu, Asp and Ile, respectively, using the PrimeSTAR Mutagenesis Basal Kit (Takara Bio, Shiga, Japan). The primers used were: V1247L forward, V1247L reverse, E1291D forward, E1291D reverse, T1303I forward and T1303I reverse (Table 1). To generate the mutant allele, R1074H (single amino acid replacement of Arg1074 with a His), we performed PCR-based mutagenesis using the primers xSAP155-mutant-F and xSAP155-mutant-R (Table 1). Finally, the HA-mxSF3B1 ORFs (wild-type and R1074H) were inserted into a modified pCLXSN retroviral vector (IMGENEX, San Diego, CA, USA), in which the SV40 promoter-neo(r) sequence had been replaced with the IRES-hyg (r) from the pIREShyg vector (Clontech), resulting in pCLXIH-SF3B1(Wt) and pCLXIH-SF3B1(R1074H).

The target sequences for SF3B1 knockdown were shSF3B1 (3989) 5′-GCACAGCTACTTCACACCTTA-3′ and shSF3B1 (4145) 5′-GCCAGTAGTGACCAAGAACAC-3′. These sequences were inserted into the pENTR/U6 vector (Invitrogen) in accordance with the manufacturer’s instructions. The resulting vectors, designated pENTR/shSF3B1 (3989) and pENTR/shSF3B1 (4145), had SF3B1 shRNA sequences under the control of the U6 promoter. The knockdown platform vector was constructed by ligating the Gateway Vector Conversion System Reading Frame Cassette C.1 (Invitrogen) and the cytomegalovirus promoter-EGFP fragment from pEGFP-N2 vector (Clontech) into the BamHI-KpnI restriction sites of the pLenti6/V5-GW/lacZ vector (Invitrogen). pLenti/shSF3B1 (3989) and pLenti/shSF3B1 (4145) vectors were constructed by recombination of the attL cassettes from the pENTR/shSF3B1 vectors into the knockdown platform vector using the LR Clonase II enzyme mix (Invitrogen).

For viral production, 293-EBNA cells (Invitrogen) were transfected with the packaging mix and the above viral vector using the TransIT-LT1 reagent (Mirus, Madison, WI, USA). Medium was exchanged 8 h after transfection, and the culture supernatant was collected after culturing for another 2 days. The collected culture supernatant was filtered through a 0.45 μm syringe filter (Millipore, Billerica, MA, USA) and used as the virus solution. The virus solution for SF3B1 expression was concentrated by ultracentrifugation at 50 000 g for 100 min at 4 °C. WiDr cells were transduced with the virus solution in the presence of 6 μg·mL−1 Polybrene (Sigma-Aldrich, St Louis, MO, USA).

Western blot analysis

Cells were lysed with SDS lysis buffer (10 mm Tris/HCl, pH 7.5, 1% SDS) containing Complete EDTA-free protease inhibitor cocktail (Roche Diagnostics) and 1 mm orthovanadate (Sigma-Aldrich). The protein content of the lysates was determined using a BCA Protein Assay Kit (Thermo Scientific, Rockford, IL, USA). Western blot analysis was performed using anti-SF3B1 serum (D221-3; MBL, Aichi, Japan), anti-SF3B2 serum (A-20; Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-SF3B3 serum (Abcam, Cambridge, UK), anti-HA serum (16B12; Covance, Princeton, NJ, USA) and anti-β-actin serum (AC-15; Sigma-Aldrich). AP-conjugated anti-mouse immunoglobulin (Ig)G (Chemicon, Temecula, CA, USA), AP-conjugated anti-goat IgG (Jackson ImmunoResearch, West Grove, PA, USA) and horseradish peroxidase-conjugated anti-mouse IgG (GE Healthcare, Little Chalfont, UK) were used as secondary antibodies. Protein bands were visualized using NBT/BCIP substrates (Pierce, Rockford, IL, USA) or Immobilon Western Chemiluminescent HRP substrate (Millipore).

Immunoprecipitation of the nuclear fraction from 3H probe-treated cells

Immunoprecipitation using anti-SF3B1 serum was performed as described previously [5]. Briefly, cells were homogenized and centrifuged at 2000 g for 10 min at 4 °C. To extract the nuclear proteins, the pellet was suspended in Tris-buffered 0.5 m NaCl solution and incubated on ice for 15 min. The suspension was centrifuged at 3000 g for 10 min at 4 °C to obtain the soluble nuclear fraction. Anti-SF3B1 serum was added to the nuclear fraction with buffer A and mixed gently on a rotator for 1 h. Protein A/G agarose (Santa Cruz Biotechnology) was added to the samples and mixed for an additional 2 h. The agarose beads were precipitated, suspended in wash buffer C, and mixed gently for 1 h at 4 °C. To the mixture, 100 nm of 3H probe was added and mixed for an additional 1 h. The immunoprecipitates were collected, and the radioactivity of the coprecipitated 3H probe in the precipitates was measured using a liquid scintillation analyzer (Tri-Carb 2700TR; Packard, Downers Grove, IL, USA).


We thank Akiyoshi Fukamizu (Life Science Center, Tsukuba Advanced Research Alliance and Graduate School of Life and Environmental Sciences, University of Tsukuba) for his helpful advice and suggestions regarding the present study and Toshio Tsuchida (Mercian Corporation Bioresource Laboratories) for supplying pladienolide B. We are grateful to Toshimitsu Uenaka, Takashi Owa, Yoshiya Oda and Kappei Tsukahara for coordinating the research team.

Conflict of interest

The authors are employees of Eisai Co., Ltd.