Serine/arginine-rich splicing factor 3 (SRSF3) is a member of the SR protein family and plays wide-ranging roles in gene expression. The human SRSF3 gene generates two alternative splice transcripts, a major mRNA isoform (SRSF3-FL) encoding functional full-length protein and a premature termination codon (PTC)-containing isoform (SRSF3-PTC). The latter is degraded through nonsense-mediated mRNA decay (NMD). Treatment of a human colon cancer cell line (HCT116) with 100 μM sodium arsenite increased SRSF3-PTC mRNA levels without changing SRSF3-FL mRNA levels. A chemiluminescence-based NMD reporter assay system demonstrated that arsenite treatment inhibited NMD activity and increased SRSF3-PTC mRNA levels in the cytoplasm, facilitating translation of a truncated SRSF3 protein (SRSF3-TR) from SRSF3-PTC mRNA. SRSF3-TR lacked two-thirds of the Arg/Ser-rich (RS) domain whose phosphorylation state is known to be crucial for subcellular distribution. SRSF3-FL was localized in the nucleus, while overexpressed SRSF3-TR was diffusely distributed in the cytoplasm and the nucleus. A part of SRSF3-TR was also associated with stress granules in the cytoplasm. Interestingly, treatment of HCT116 cells with a small interference RNA specifically targeting SRSF3-PTC mRNA significantly attenuated arsenite-stimulated induction of c-JUN protein, its binding activity to the AP-1 binding site (−126 to 120 bp) in the interleukin (IL)-8 gene promoter, and AP-1 promoter activity, resulting in significant reduction of arsenite-stimulated IL-8 production. Our results suggest that SRSF3-TR may function as a positive regulator of oxidative stress-initiated inflammatory responses in colon cancer cells.
- SRSF3 gene
- oxidative stress
- alternative splicing
- nonsense-mediated mRNA decay
serine/arginine-rich splicing factor 3 (SRSF3) (also known as SRp20) is a member of the SR protein family that comprises at least 12 evolutionarily conserved and structurally related RNA-binding proteins (SRSF1 to SRSF12) (24). Like other SR proteins, SRSF3 binds cis-acting RNA elements through an NH2-terminal RNA recognition motif (RRM), and a COOH-terminal region enriched in Arg-Ser dipeptides (RS domain) mediates interactions with other proteins for regulation of both constitutive and alternative splicing of pre-mRNA (7). Recently, it has been shown that the SR protein family, including SRSF3, participates in many cellular processes including transcriptional elongation (22), mRNA transport (32), translation (31), and cell cycle regulation (16). In addition, a short hairpin RNA-based screen of splicing regulators has uncovered a novel function of SRSF3 as a negative regulator of interleukin (IL)-1β (28). Moreover, SRSF3 is overexpressed in human ovarian cancer, and its knockdown results in apoptosis of ovarian cancer cells (10). These findings suggest that SRSF3 contributes to biological networks for cell fate.
The human SRSF3 gene is composed of seven exons and eight introns and generates two mRNA isoforms through alternative splicing. SRSF3 exon 4 contains multiple premature stop codons (PTCs); therefore, a major SRSF3 mRNA isoform, which encodes a functional full-length SRSF3 protein (SRSF3-FL), excludes this exon 4. Alternative splicing of SRSF3 pre-mRNA produces another mRNA isoform consisting of the entire seven exons. This PTC-containing SRSF3 isoform (SRSF3-PTC) should be degraded through nonsense-mediated mRNA decay (NMD), a surveillance mechanism that decomposes PTC-containing mRNAs. However, this isoform is considered to contribute autoregulation of its own mRNA expression in a cell cycle-specific manner (12). At the same time, it was reported that SRSF3-PTC mRNA was translatable into a truncated SRSF3 protein (SRSF3-TR) partially lacking the RS domain in murine B-cell lymphoma K46 cells (12). However, the biological function of SRSF3-TR remains to be elucidated.
In this study, we showed that significant amounts of SRSF3-TR were translated from SRSF3-PTC mRNA in colon cancer cells exposed to an oxidant (sodium arsenite). Moreover, we suggest that SRSF3-TR may function as a positive regulator of oxidative stress-stimulated production of IL-8 through regulation of c-JUN expression. The present study may provide a new insight into the biological function of PTC-containing mRNAs of the SR protein family genes.
MATERIALS AND METHODS
Cell culture and subcellular fractionation.
HCT116 cells were cultured in McCoy's 5A medium (GIBCO, Grand Island, NY) supplemented with 5% (vol/vol) heat-inactivated fetal bovine serum and antibiotics at 37°C in 5% CO2. After HCT116 cells were incubated in lysis buffer (in mM: 10 Tris·HCl, pH 7.4, 100 NaCl, and 2.5 MgCl2, with 40 μg/ml digitonin) for 10 min, lysates were centrifuged at 2,060 g for 8 min at 4°C and supernatants were collected as cytosolic extracts. The remaining pellets were washed twice with the lysis buffer and further lysed with RIPA buffer (in mM: 10 Tris·HCl, pH 7.4, 150 NaCl, 1 EDTA, and 1 DTT, with 0.1% SDS and 1% NP-40) containing a protease and phosphatase inhibitor cocktail (Sigma-Aldrich, St. Louis, MO). After centrifugation at 21,000 g for 10 min at 4°C, supernatants were collected as nuclear extracts. Cells lysed with RIPA buffer were used as whole cell extracts (18).
A cDNA library was prepared from HCT116 cells, and the human SRSF3 mRNAs [Ensembl Transcript ID: ENST00000373715 (SRSF3-FL) and ENST00000477442 (SRSF3-PTC)] were amplified by PCR using a primer set: 5′-AAAAAAGGATCCATGCATCGTGATTCCTGTCCATTG-3′ (forward; BamHI site is underlined) and 5′-AAAAAAGATATCCTATTTCCTTTCATTTGACCTAGA-3′ (reverse; EcoRV site is underlined). The amplified products were separated with a gel extraction kit (Qiagen, Hilden, Germany) and cloned into the mammalian expression vector pCMV-3Tag (Stratagene, La Jolla, CA). FLAG was appended to the NH2 termini of SRSF3. All constructs were confirmed to have the expected sequence by DNA sequencing. These plasmids were transfected with FuGENE HD (Promega, Madison, WI) according to the manufacturer's instructions.
Small interference RNAs.
We used a small interference RNA (siRNA) (Stealth RNAi, Invitrogen, Carlsbad, CA) targeting exon 4, 5′-TCAACTAGCCCTTTCAGCGTCATGT-3′ (SRSF3-PTC siRNA), to selectively silence SRSF3-PTC mRNA and SRSF3-TR protein. To inhibit the NMD system, an siRNA targeting up-frameshift-1 (UPF1; Invitrogen), 5′-UAAACUCGUCACCAAGGUAACU-3′, was introduced to HCT116 cells. The Stealth RNAi negative control (Invitrogen) was used as a control siRNA. HCT116 cells were transfected with 10 nM of one of these siRNAs with Lipofectamine RNAiMAX (Invitrogen).
Quantitative real-time reverse transcription-PCR.
Total RNAs were extracted from HCT116 cells with an RNAiso plus (Takara, Otsu, Japan). One microgram of isolated RNA was reverse-transcribed with ReverTra Ace reverse transcriptase (TOYOBO, Osaka, Japan). SRSF3-FL, SRSF3-PTC, IL-8, IL-1β, IL-6, IFN-γ, and TNF-α mRNA levels were measured with SYBR Green Master Mix and the 7500 real-time system (Applied Biosystems, Foster City, CA). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA and 18S were measured as internal controls for normalization. With the ΔΔCt method (where Ct is threshold cycle), the data are presented as the fold change in gene expression relative to controls measured in the untreated or control siRNA-treated cells. The sequences of primer sets used are shown in Table 1.
Inhibition of NMD activity and chemiluminescence-based NMD assay.
To inhibit the NMD system, HCT116 cells were treated with 50 or 100 μg/ml cycloheximide for 6 h or with 10 nM UPF1 siRNA for 48 h. Plasmids containing pCI-neo β-globin-wt (WT), pCI-neo β-globin-NS39 (NS39), and pCI-neo-firefly were kindly provided by Dr. Andreas E. Kulozik (University of Heidelberg) and used for chemiluminescence-based NMD assay (3). Twenty-four hours after pretreatment of HCT116 cells with control or SRSF-PTC siRNA, they were cotransfected with pCI-neo-firefly and WT or NS39 and then incubated for another 24 h. These cells were left untreated or treated with 100 μM arsenite for the indicated times. Cell lysates were prepared, and Renilla and firefly luciferase activities were measured with a dual-luciferase assay system (Promega) according to the manufacturer's protocol. The transfection efficacy was assessed by firefly luciferase activity.
Measurement of IL-8 concentration in culture medium.
After treatment of HCT116 cells with control or SRSF3-PTC siRNA for 24 h, they were exposed to 100 μM arsenite for the indicated times and supernatants were collected. Cell debris was removed by centrifugation at 10,000 g for 5 min at 4°C, and the resultant supernatants were subjected to ELISA measurement of IL-8 with a commercially available ELISA kit (Biolegend, San Diego, CA) according to the protocols provided by the manufacturer.
Whole cell lysates were prepared in RIPA buffer (in mM: 10 Tris·HCl, pH 7.4, 150 NaCl, 1 EDTA, and 1 DTT, with 0.1% SDS and 1% NP-40) containing a protease and phosphatase inhibitor cocktail (Sigma-Aldrich). The extracted proteins were separated by SDS-PAGE and transferred to a polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA). After blocking with 5% nonfat dry milk, the membranes were incubated overnight at 4°C with mouse monoclonal anti-SRSF3 (1:1,000 dilution; Sigma-Aldrich), anti-phospho-JUN-NH2-terminal protein kinase (JNK) (1:1,000; Cell Signaling Technology, Danvers, MA), anti-JNK (1:1,000; Cell Signaling Technology), anti-phospho-p38 mitogen-activated protein kinase (MAPK) (1:1,000, Cell Signaling Technology), anti-p38 MAPK (1:1,000, Cell Signaling Technology), anti-phospho-extracellular signal-regulated protein kinase ERK (1:1,000, Cell Signaling Technology), anti-ERK (1:1,000, Cell Signaling Technology), anti-phospho-p65 (1:1,000, Cell Signaling Technology), anti-p65 (1:1,000, Cell Signaling Technology), anti-UPF1 (1:1,000; Bethyl Laboratories, Montgomery, TX), or anti-GAPDH (1:5,000; Santa Cruz Biotechnology, Santa Cruz, CA) antibody. Signal intensities of anti-c-JUN-reactive bands were quantified with ImageJ software (http://rsb.info.nih.gov/ij/).
HCT116 cells growing on a multichamber culture slide (Matsunami Glass Industries, Osaka, Japan) were rinsed briefly in phosphate-buffered saline (PBS) and then fixed in 4% paraformaldehyde in PBS for 10 min at room temperature. The fixed cells were permeabilized in PBS containing 0.1% Triton X-100 for 10 min. After blocking with 4% Block ACE (Snow Brand Milk Products, Tokyo, Japan) at room temperature for 1 h, slides were incubated overnight at 4°C with mouse anti-FLAG (1:2,000 dilution; Sigma-Aldrich) and rabbit anti-TIAR (1:1,600; Cell Signaling Technology) antibodies in PBS containing 1% Block ACE and 0.1% Triton X-100. After washing, the slides were treated with a 1:1,000 dilution of Alexa Fluor 488-conjugated anti-mouse IgG (Invitrogen) and a 1:1,000 dilution of Alexa Fluor 555-conjugated anti-rabbit IgG. Nuclei were counterstained with TO-PRO-3 (Molecular Probes, Eugene, OR). After washing, the slides were mounted in Vectashield mounting medium (Vector Labs, Burlingame, CA) and examined by confocal laser-scanning microscopy (FluoView FV1000; Olympus, Tokyo, Japan).
Promoter activity assay.
The 5′ flank of the human IL-8 gene was cloned into the pGL3-basic luciferase reporter vector (Promega). In brief, the first PCR was performed with human genomic DNA as a template. The IL-8 proximal promoter region (from −652 to +46 bp) was amplified with the following primer set: 5′-AAAAAGGTACCTAGTCTTATCTATTTCTAGATC-3′ (forward) and 5′-AAAAAGAGCTCGAAGCTTGTGTGCTCTGCTGTC-3′ (reverse). Subsequently, the amplified product was used as a template to generate deletion constructs consisting of the −229 to +46 bp, −99 to +46 bp, and −55 to +46 bp regions with the same reverse primer and one of the following forward primers: 5′-AAAAAGGTACCGAGCTTCAGTATTTTAAATGTATA-3′ for −229 to +46 bp, 5′-AAAAAGGTACCGCCATCAGTTGCAAATCGTG-3′ for −99 to +46 bp, and 5′-AAAAAGGTACCAGGGTGCATAAGTTCTCTAGTAG-3′ for −55 to +46 bp. The amplified products were subcloned into the pGL3-basic vector with KpnI and SacI restriction sites. To generate a mutated reporter plasmid directed to activator protein-1 (AP-1), PCR was performed with the KOD-Plus-Mutagenesis Kit (Toyobo) and pGL3 −229/+46 as the template. The two-point mutation in the AP-1 binding site (−126 TGACTCA −120 bp to −126 TATCTCA −120 bp) was prepared from pGL3 −229/+46 construct. To generate AP-1 reporter vector (pGL3-AP1), minimal promoter (minP) sequence, 5′-AGACACTAGAGGGTATATAATGGAAGCTCGACTTCC-3′, was subcloned into the pGL3-basic vector with restriction sites. Subsequently, AP-1 response element, 5′-TGAGTCAGTGACTCAGTGAGTCAGTGACTCAGTGAGTCAGTGACTCAG-3′, was subcloned into the pGL3-minP. pGL3-minP was used as negative promoter activity. HCT116 cells (2.0 × 104 cells) were cultured on 24-well plates, and then pGL3-luciferase constructs (100 ng) were cotransfected with pRL-CMV vector (50 ng) with FuGENE HD (Promega). Twenty-four hours after the transfection, cells were harvested, and the firefly and Renilla luciferase activities were measured with the Dual-Luciferase Reporter Assay System (Promega).
Electrophoretic mobility shift assay.
Nuclear proteins were prepared from HCT116 cells as previously described (17). A double-stranded oligonucleotide for the AP-1 consensus binding site (−132 to −108 bp, GTGTGATGACTCAGGTTTGCCCTGA) was radiolabeled by filling in a 5′ overhang with [α-32P]deoxycytidine 5′-triphosphate and Klenow fragment (New England Biolabs, Beverly, MA) and then purified with a QIAquick nucleotide removal kit (Qiagen). A mutated probe for the AP-1 site (AP1mt; GTGTGATATCTCAGGTTTGCCCTGA) was used as a non-self-competitor. Protein-DNA binding reactions were performed with 5 μg of nuclear extract proteins, 1 μl of 32P-labeled oligonucleotide (40,000 cpm), and 2 μg of salmon sperm DNA (Invitrogen) in 10 μl of 10 mM Tris·HCl buffer (pH 7.9) containing (in mM) 50 NaCl, 5 MgCl2, 10 EDTA, and 10 DTT, with 20% glycerol. After being incubated at room temperature for 30 min, protein-DNA complexes were resolved on a 5% nondenaturing polyacrylamide gel in 0.25× Tris-borate-EDTA buffer at 4°C. For self-competition experiments, the nuclear proteins were incubated with the unlabeled double-stranded oligonucleotide for 20 min at room temperature before addition of the 32P-labeled probe. Antibody perturbation experiments were performed by preincubating nuclear proteins with 1 μg of anti-c-JUN antibody for 30 min at room temperature before adding the 32P-labeled probe.
Chromatin immunoprecipitation assay.
Chromatin immunoprecipitation (ChIP) assays were performed with the Chromatin Immunoprecipitation Assay Kit (Millipore, Billerica, MA). Briefly, HCT116 cells were fixed with 1% formaldehyde in PBS for 10 min and then washed twice with ice-cold PBS. These cells were resuspended in SDS lysis buffer, incubated for 10 min on ice, and then sonicated. Immunoprecipitation was carried out overnight at 4°C with 3 μg of anti-c-JUN antibody. Normal rabbit IgG was used to detect nonspecific reactions. Immune complexes were collected with protein A agarose-salmon sperm DNA. Cross-linking between proteins and DNA was reversed according to the manufacturer's protocol. Protein-bound DNA was extracted with phenol-chloroform-isoamyl alcohol. The extracted DNA was amplified by PCR (32 cycles; denaturing at 98°C for 10 s, annealing at 55°C for 30 s, and extension at 72°C for 1 min) using the following primers: for IL-8 sequence between −568 and −401 bp, 5′-ATGTGCCCCTTCACTCTGTT-3′ and 5′-GGTGAAGATAAGCCAGCCAAT-3′; for IL-8 sequence between −263 and −111 bp, 5′-AAAAAGCATACAATTGATAATTCACC-3′ and 5′-GGCAAACCTGAGTCATCACA-3′. The nuclear chromatin DNA from HCT116 cells (input) was used as a positive control for PCR.
Production of SRSF3-PTC mRNA isoform after exposure to sodium arsenite.
SRSF3 generates two mRNA isoforms through alternative splicing, a major isoform (we describe here as SRSF3-FL) and a minor isoform (SRSF3-PTC) (Fig. 1A). When HCT116 cells were exposed to 100 μM arsenite or 500 μM hydrogen peroxide (H2O2) for 6 h, they significantly increased SRSF3-PTC mRNA levels without changing SRSF3-FL mRNA levels (Fig. 1B). Pretreatment with the antioxidant N-acetyl-cysteine (NAC, 10 mM) almost completely blocked both arsenite- and H2O2-stimulated SRSF3-PTC mRNA expression, suggesting that oxidative stress stimulates production of SRSF3-PTC mRNA isoform (Fig. 1B). After treatment of HCT116 cells with 100 μM sodium arsenite, SRSF3-PTC mRNA levels were increased, with a peak at 3 h, and remained elevated for at least 12 h without changing SRSF3-FL mRNA levels (Fig. 2A). The arsenite-stimulated production of SRSF3-PTC mRNA through alternative splicing was also detectable by RT-PCR using the primer set designed to amplify exons 3–5 (Fig. 2B). SRSF3-PTC mRNA should be degraded through NMD. At the same time, however, it was possible that SRSF3-PTC mRNA expressed in arsenite-treated cells would be translated into a truncated protein partially lacking the RS domain (SRSF3-TR) (Fig. 1A). Western blotting using an antibody against the NH2 terminus of SRSF3 showed that treatment of HCT116 cells with arsenite newly induced an immunoreactive protein with the predicted molecular size of SRSF3-TR (14 kDa), while it did not increase levels of constitutive, full-length SRSF3 protein (SRSF3-FL) (Fig. 2C). Treatment with SRSF3-PTC siRNA specifically targeting SRSF3 exon 4 completely blocked the arsenite-induced appearance of the immunoreactive protein band (Fig. 2D), indicating that the 14-kDa protein was likely to be SRSF3-TR.
NMD-mediated degradation of SRSF3-PTC mRNA.
mRNAs containing PTCs that are located >50 nucleotides upstream of the final exon-exon junction are considered to be targets for NMD (26). To examine whether SRSF3-PTC mRNA is actually regulated by NMD, we first disrupted NMD by knockdown of UPF1, which is an essential factor of NMD pathway (29). Reduction of UPF1 after treatment with UPF1 siRNA for 48 h (Fig. 3A) increased SRSF3-PTC mRNA levels fourfold (Fig. 3B). Inhibition of NMD by treatment with cycloheximide also increased SRSF3-PTC mRNA levels 30-fold (Fig. 3C). Thus SRSF3-PTC mRNA appeared to be a target for NMD. It has been shown that NMD is often inactivated under stressful conditions including hypoxia, amino acid starvation, and oxidative stress (8, 27, 35). Exposure to arsenite was also reported to inhibit NMD activity in mouse embryonic fibroblasts (MEFs) and Chinese hamster ovary (CHO) cells (35). To test whether arsenite inhibited NMD activity in HCT116 cells, we employed a chemiluminescence-based reporter system (3). Plasmids containing a fusion construct of an in-frame Renilla luciferase and human β-globin gene with or without a nonsense mutation at codon 39 (NS39 or WT, respectively) (25, 33) were cotransfected with firefly luciferase serving as a normalization control. NMD activity was monitored as the ratio between NS39 and WT chemiluminescences. As shown in Fig. 3D, after treatment with arsenite relative luciferase activity (NS39/WT) was significantly increased, probably because of impaired degradation of the nonsense-mutated human β-globin (NS39) transcripts by NMD. In fact, in the presence of transcription inhibitor actinomycin D (2.5 μg/ml), the half-life of SRSF3-PTC mRNA became longer, to >8 h in arsenite-treated cells compared with an ∼2-h half-life in untreated control cells (Fig. 3E).
Subcellular distribution of SRSF3-PTC mRNA isoform.
Although several models for the mechanism of NMD have been proposed (30), a pioneer round of translation in the cytoplasm is an essential step for recognition of PTC-containing mRNAs by NMD. Thus NMD degrades PTC-containing mRNAs mainly in the cytoplasm. Nuclear and cytoplasmic fractions were prepared before (0) or 3 and 6 h after treatment with arsenite, and the purity of each fraction was confirmed by RT-PCR using the primer set for GAPDH pre-mRNA (Fig. 4A). In cytoplasmic fractions an exon 4-containing transcript (SRSF3-PTC) appeared 3 h after exposure to arsenite, and its levels time-dependently increased (Fig. 4A). In nuclear fractions SRSF3-PTC mRNA was present before arsenite treatment, and its levels were not changed after exposure to arsenite (Fig. 4A). The time-dependent changes in subcellular distribution of SRSF3-PTC mRNA were quantified by quantitative real-time reverse transcription-PCR (qPCR) (Fig. 4B). qPCR demonstrated that treatment with arsenite increased SRSF3 PTC mRNA levels preferentially in the cytoplasm (20-fold increase), compared with those in the nucleus (<2-fold increase) (Fig. 4, B and C). These results suggested that, in the presence of arsenite, SRSF3-PTC mRNA appeared to escape from degradation through the NMD system and significant amounts of SRSF3-PTC mRNA were accumulated in the cytoplasm. As a consequence, SRSF3-TR could be transiently translated from SRSF3-PTC mRNA under oxidative stress.
Phosphorylation and subcellular localization of SRSF3-TR.
Phosphorylation of the RS domain determines the subcellular localization of individual SR proteins (5, 21). It has been shown that functional SRSF3-FL localizes in the nuclear speckles (4, 23). We next examined phosphorylation and subcellular localization of SRSF3-TR partially lacking the RS domain. Since the anti-SRSF3 antibody used was raised against the NH2 terminus of SRSF3, it could not distinguish between SRSF3-TR and SRSF3-FL. We therefore transfected plasmids encoding FLAG-tagged SRSF3-FL or -PTC and assessed phosphorylation of SRSF3-TR and SRSF3-FL by Western blotting using anti-SRSF3 and anti-FLAG antibodies (Fig. 5). The anti-FLAG antibody recognized FLAG-SRSF3-FL and -TR. Dephosphorylated FLAG-SRSF3-FL and endogenous SRSF3-FL appeared after a calf intestinal phosphatase (CIP) treatment, while we could not detect any shifted band of FLAG-SRSF3-TR after CIP treatment (Fig. 5, top). Endogenous SRSF3-TR was not detectable with the anti-SRSF3 antibody in mock-transfected or FLAG-SRSF3-FL- or FLAG-SRSF3-TR-transfected cells, and the anti-SRSF3 antibody recognized the dephosphorylation of endogenous or FLAG-tagged SRSF3-FL in these cells (Fig. 5, middle).
To examine the subcellular localization of SRSF3-FL and SRSF3-TR, we performed double immunofluorostaining with an antibody to FLAG and an antibody to a stress granule-associated TIAR protein, using FLAG-SRSF3-FL- or -TR-overexpressing HCT116 cells. Confocal laser-scanning microscopy showed that FLAG-SRSF3-FL was localized in nuclei (Fig. 6E), and a part of FLAG-SRSF3-FL was translocated to stress granules that appeared in the cytoplasm 6 h after exposure to arsenite (Fig. 6H′). In contrast, FLAG-SRSF3-TR was diffusely distributed in both the nucleus and the cytoplasm (Fig. 6I). At the same time, overexpression of FLAG-SRSF3-TR itself induced formation of stress granules in unstressed HCT116 cells (Fig. 6J), and a part of FLAG-SRSF3-TR was associated with them (Fig. 6L). After arsenite treatment, SRSF3-TR still remained in both stress granules and nucleus (Fig. 6L′).
Effect of SRSF3-TR on arsenite-stimulated IL-8 production.
It has recently been shown that SRSF3 acts as a negative regulator of inflammation (28). We also examined whether reduction of SRSF3-TR changed expression of mRNAs encoding IL-1β, IL-6, IL-8, interferon (IFN)-γ, and tumor necrosis factor (TNF)-α after exposure to arsenite using qPCR and found that arsenite-stimulated elevation of IL-8 mRNA levels was significantly attenuated in SRSF3-PTC siRNA-treated cells compared with control siRNA-treated cells (Fig. 7A). Consequently, SRSF3-PTC siRNA significantly decreased arsenite-stimulated IL-8 release (Fig. 7B).
The human IL-8 gene promoter contains a number of regulatory cis elements that are activated by the interaction with AP-1, CCAAT/enhancer-binding protein (C/EBP), and nuclear factor-κB (NF-κB) (15). The transcription start site of the IL-8 gene is located 101 bp upstream of the translation site. To define the minimal promoter of the IL-8 gene responsive to arsenite stress, we generated the luciferase reporter vector having the 5′-flank of the human IL-8 gene from −652 to +46 bp and then serially truncated fragments were constructed. As shown in Fig. 7C, the arsenite-dependent luciferase activity resided in the proximal promoter region from −229 to −99 bp. This region contains an AP-1 binding site, and the two-point mutation of this AP-1 site almost completely inhibited the arsenite-dependent promoter activity (Fig. 7C), indicating a crucial role of AP-1 and its binding site located in −126 to −120 bp for arsenite-induced IL-8 transcription. We also confirmed that SRSF3-PTC siRNA did not change the stability of IL-8 mRNA (Fig. 7D), suggesting that SRSF3-PTC siRNA might downregulate transcription of the IL-8 gene.
As shown in Fig. 8, SRSF3-PTC siRNA did not significantly change levels of IL-1β, IL-6, IFN-γ, and TNF-α mRNAs before and after treatment with arsenite. We also examined whether SRSF3-PTC siRNA affected survival or apoptosis after exposure to higher concentrations of arsenite, but we could not detect any changes in numbers of viable cells and terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling-positive cells between SRSF3-PTC and control siRNA-treated cells (data not shown).
Effect of SRSF3-PTC siRNA on arsenite-stimulated c-JUN protein induction.
AP-1 is a homo- or heterodimer composed of Jun and Fos family members. Among them, c-JUN and c-Fos together comprise an AP-1 heterodimer and induce IL-8 gene transcription (34). MAPKs, including ERK, JNK, and p38 MAPK, are involved in the IL-8 gene transcription (11). As shown in Fig. 9, the induction of c-JUN protein was significantly reduced 3, 6, and 9 h after exposure to arsenite in SRSF3-PTC siRNA-treated cells compared with that in control siRNA-treated cells. There were no differences in expression and phosphorylation of p38 MAPK, ERK, or c-Fos between SRSF3-PTC and control siRNA-treated cells (Fig. 9A). We also investigated the possibility that changes in NF-κB activation were involved in the attenuation of the IL-8 gene transcription in SRSF3-PTC siRNA-treated cells. We compared phosphorylation of an essential factor of NF-κB (p65) and confirmed that SRSF3-PTC knockdown did not change p65 phosphorylation after exposure to arsenite (Fig. 9A). These results suggested that the reduction of c-JUN protein might be responsible for the downregulation of arsenite-stimulated IL-8 gene transcription in SRSF3-PTC siRNA-treated cells.
Reduction of c-JUN binding to IL-8 promoter in SRSF3-PTC siRNA-treated cells.
To determine whether the AP-1 consensus sequence (TGACTCA) located at −126 and −120 bp in the 5′-flanking region was able to interact with c-JUN, we synthesized a 25-bp double-stranded oligonucleotide (−132 to −108 bp; GTGTGATGACTCAGGTTTGCCCTGA) and used it as a probe for electrophoretic mobility shift assay (EMSA). As shown in Fig. 10A, arsenite stress for 3 h strongly generated a complex that was weakly induced in cell lysates from cells not treated with arsenite, and SRSF3-PTC siRNA treatment reduced this arsenite-induced complex. Unlabeled self-oligonucleotide, but not the mutated AP-1 oligonucleotide (GTGTGATATCTCAGGTTTGCCCTGA), competed for formation of this complex. Moreover, preincubation with an anti-c-JUN antibody, but not rabbit IgG, supershifted this band and produced a more slowly migrated band (Fig. 10A). Together these results indicate that arsenite stress produced an interaction between activated c-JUN and the AP-1 binding element present in the 5′-flanking region of the human IL-8 gene.
To confirm the results obtained from EMSA and to reveal a crucial role of c-JUN in the reduction of IL-8 transcription, ChIP assay was employed with an anti-c-JUN antibody. As shown in Fig. 10B, c-JUN bound to the −263 to −111 bp region containing an AP-1 binding site in both control and SRSF3-PTC siRNA-treated cells, whereas c-JUN did not bind to the −568 to −401 bp region. As expected, exposure to arsenite increased the binding of c-JUN in both control and SRSF3-PTC siRNA-treated cells. qPCR measurement of the immunoprecipitated −263 to −111 bp region demonstrated that arsenite-induced c-JUN binding to the IL-8 promoter was partially, but significantly, attenuated in SRSF3-PTC siRNA-treated cells (Fig. 10C).
AP-1 binding sites are widely found in promoter regions of various genes, especially related to proliferation, apoptosis, survival, and tumorigenesis. AP-1 reporter assay was employed to examine whether SRSF-PTC siRNA generally affected AP-1 activity. As shown in Fig. 10D, treatment of HCT116 cells with SRSF3-PTC siRNA significantly inhibited arsenite-stimulated AP-1 reporter activity, suggesting that SRSF3-TR might modify AP-1-dependent cellular stress responses.
Alternative splicing generates protein isoform diversity and profoundly affects gene expression program. However, up to one-third of human alternative splicing events generate a PTC that would cause the resulting mRNA to be degraded by NMD (20), indicating the widespread coupling of alternative splicing and NMD (AS-NMD). Interestingly, AS-NMD events are found in all SR protein genes including SRSF3 (19). Furthermore, these events coincide with regions of extreme sequence conservation, termed “ultraconserved elements” (UCEs), suggesting the existence of evolutional significance of UCEs. There are 481 described UCEs, and 68% of them are transcribed, constituting a new category of noncoding RNAs (6). However, 47% of UCEs have been designated either exonic or probably exonic (2). SRSF3 gene contains a 577-bp UCE that includes entire exon 4 (data not shown). We found that sodium arsenite stimulated cytoplasmic accumulation of SRSF3-PTC mRNA isoform through inhibition of NMD activity, resulting in translation of a truncated protein, SRSF3-TR, in colon cancer cells (HCT116). Furthermore, arsenite stress facilitated nuclear translocation of SRSF3-TR, suggesting novel function(s) of the truncated protein in the nucleus under oxidative stress.
Jumaa and Nielsen reported that SRSF1 (also known as ASF/SF2) and SRSF2 (SC35) were involved in the exclusion of SRSF3 exon 4 (13, 14). The inclusion of exon 4 to mature SRSF3 mRNA was reported to occur under serum starvation (12), and the biological significance of exon 4 inclusion is considered to be an autoregulatory feedback loop by SRSF3 itself (13). Moreover, because of multiple PTCs in exon 4, a recent study (19) together with our findings have revealed that exon 4-including SRSF3-PTC mRNA is a target for NMD. To avoid production of the COOH-terminally truncated proteins that have potential dominant-negative or gain-of-function properties, PTC-containing mRNAs are expected to be degraded by NMD. However, several lines of evidence have shown that NMD activity is often inhibited by stresses, such as hypoxia, amino acid starvation, and generation of reactive oxygen species (8, 27, 35). Consistent with those reports, our results also suggest that the cytoplasmic accumulation of SRSF3-PTC mRNA may, at least in part, result from inhibition of NMD under arsenite-induced oxidative stress. In addition, expression of PTC-containing mRNA isoforms of other SR protein genes (SRSF6, SRSF7, and SRSF9) was also upregulated after exposure to arsenite (data not shown), suggesting that arsenite stress may extensively modify AS-NMD events. Arsenite stress may also change the alternative splicing pattern to preferentially produce the SRSF3-PTC variant. To address this issue, we prepared a minigene spanning SRSF3 exon 2 to exon 7. This minigene produced exon 4-containing transcript as a major transcript, and we could not detect any change in alternative splicing. Truncated SRSF3 is translated from cytosolic SRSF3-PTC mRNA. Arsenite increased SRSF3-PTC mRNA levels 20-fold in the cytoplasmic fractions (Fig. 4, B and C), while it did not change the major isoform (SRSF3-FL mRNA) levels (Fig. 2A). Arsenite treatment apparently stabilized SRSF3-PTC mRNA (Fig. 3E). These results suggest that regulation of mRNA stability may play an important role in the arsenite-induced accumulation of truncated SRSF3 protein.
Overexpression of SRSF3-TR alone could induce stress granule formation. Interestingly, a part of SRSF3-TR was colocalized with a stress granule marker, TIAR. It was reported that SRSF3-ΔRS protein lacking the entire RS domain localized throughout the cytoplasm and nucleoplasm (4), suggesting that the RS domain has a crucial role in determining subcellular localization. In that case, stress granule formation was not documented (4). SRSF3-TR still has partial RS domain, which might be responsible for stress granule formation in the cytoplasm. Stress granules are formed to protect RNAs from various types of stressors, such as oxidants or heat shock. The stress granule formation itself may modify the subsequent stress responses including the arsenite-induced IL-8 production. It is possible that the ability of the SRSF3-TR construct to induce stress granules may be important for understanding the role of this protein. Further studies are needed to clarify this issue.
To disclose unknown functions of SRSF3-TR, we focused on proinflammatory cytokine responses to arsenite, since silencing of SRSF3 was reported to increase IL-1β secretion due to elevation of IL-1β and caspase-1 mRNA levels in addition to activation of caspase-1 (28). SRSF3-TR knockdown did not affect arsenite-stimulated expression of IL-1β, IL-6, IFN-γ, and TNF-α mRNAs, but it significantly suppressed arsenite-stimulated elevation of IL-8 mRNA levels and IL-8 release from HCT116 cells. IL-8 belongs to the C-X-C chemokine family and has proinflammatory effects. It has been shown that maximal IL-8 production requires a combination of derepression of the gene promoter, activation of protein kinase cascades (JNK, p38 MAPK, ERK, and NF-κB), and enhanced mRNA stabilization and translation efficiency (11). Among these regulators, AP-1 was found to be crucial for arsenite-stimulated promoter activity of the IL-8 gene in HCT116 cells. Furthermore, we found that SRSF3-TR knockdown decreased arsenite-stimulated c-JUN induction. Consistent with the reduction of c-JUN protein levels, EMSA and ChIP assay experiments also showed that SRSF3-TR knockdown partially inhibited the binding of the c-JUN/AP-1 complex to the AP-1-binding element. RNA-immunoprecipitation (RNA-IP) showed that SRSF3-TR bound to JUN mRNA, suggesting the possibility that SRSF3-TR may alter JUN mRNA stability. However, we could not detect any difference in the stability between control and SRSF3-PTC siRNA-treated cells (data not shown). To further investigate this issue, we established a cell line stably expressing FLAG-SRSF3-TR. Consistent with the results obtained in SRSF3-PTC siRNA-treated cells, c-JUN protein levels were significantly increased in FLAG-SRSF3-TR-overexpressing cells. Similarly, there was no difference in JUN mRNA stability between the FLAG-SRSF3-TR-expressing cell line and control cells (data not shown). The present study suggests that SRSF3-TR knockdown reduces arsenite-induced IL-8 production, at least in part, through regulation of c-JUN protein levels. Once c-JUN/AP-1 is activated, c-JUN is positively autoregulated by its product, c-Jun/AP-1 (1). Although it is still unknown what an initiator of c-JUN/AP1 activity is and how SRSF3-TR interacts with it, we have provided the first evidence that a functional truncated SR protein could be translated from an UCE- and PTC-containing splice variant of the SR protein family gene.
Finally, as mentioned above, NMD was often inhibited under certain conditions such as hypoxia and oxidative stress. Cancer tissues are exposed to similar microenvironmental conditions, and it has been shown that the impairment of NMD is involved in the oncogenic mechanism (9). In such situations, oxidative stress-induced SRSF3-TR might function as a potential mediator of inflammation. Further studies are needed to fully elucidate the molecular mechanism for the SRSF3-TR-dependent regulation of c-JUN expression. Although UCEs are considered to have evolutional importance, their biological significance is not fully understood. Our finding may provide a new insight on UCEs encoded in the SR protein family genes.
This work was supported by a grant from the Takeda Science Foundation (to K. Nishida) and Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology, Japan (no. 12016603 to K. Nishida and no. 22390146 to K. Rokutan).
No conflicts of interest, financial or otherwise, are declared by the author(s).
Author contributions: S.K., K.M., Y.K., and T.T. conception and design of research; S.K., K.N., H.K., C.N., K. Kita, K. Kurokawa, and Y.K. performed experiments; S.K., H.K., C.N., K. Kita, Y.A., K. Kurokawa, and Y.K. analyzed data; K.N. prepared figures; K.N. drafted manuscript; K.N. and K.R. edited and revised manuscript; K. Kajita, K.M., Y.K., T.T., and K.R. interpreted results of experiments; K.R. approved final version of manuscript.
We thank Dr. Andreas E. Kulozik (University of Heidelberg) for providing plasmids of pCI-neo β-globin WT and NS39, and pCI-neo firefly.
- Copyright © 2014 the American Physiological Society