Exonic Splicing Enhancer-Dependent Selection of the Bovine Papillomavirus Type 1 Nucleotide 3225 3′ Splice Site Can Be Rescued in a Cell Lacking Splicing Factor ASF/SF2 through Activation of the Phosphatidylinositol 3-Kinase/Akt Pathway

Abstract

Bovine papillomavirus type 1 (BPV-1) late pre-mRNAs are spliced in keratinocytes in a differentiation-specific manner: the late leader 5′ splice site alternatively splices to a proximal 3′ splice site (at nucleotide 3225) to express L2 or to a distal 3′ splice site (at nucleotide 3605) to express L1. Two exonic splicing enhancers, each containing two ASF/SF2 (alternative splicing factor/splicing factor 2) binding sites, are located between the two 3′ splice sites and have been identified as regulating alternative 3′ splice site usage. The present report demonstrates for the first time that ASF/SF2 is required under physiological conditions for the expression of BPV-1 late RNAs and for selection of the proximal 3′ splice site for BPV-1 RNA splicing in DT40-ASF cells, a genetically engineered chicken B-cell line that expresses only human ASF/SF2 controlled by a tetracycline-repressible promoter. Depletion of ASF/SF2 from the cells by tetracycline greatly decreased viral RNA expression and RNA splicing at the proximal 3′ splice site while increasing use of the distal 3′ splice site in the remaining viral RNAs. Activation of cells lacking ASF/SF2 through anti-immunoglobulin M-B-cell receptor cross-linking rescued viral RNA expression and splicing at the proximal 3′ splice site and enhanced Akt phosphorylation and expression of the phosphorylated serine/arginine-rich (SR) proteins SRp30s (especially SC35) and SRp40. Treatment with wortmannin, a specific phosphatidylinositol 3-kinase/Akt kinase inhibitor, completely blocked the activation-induced activities. ASF/SF2 thus plays an important role in viral RNA expression and splicing at the proximal 3′ splice site, but activation-rescued viral RNA expression and splicing in ASF/SF2-depleted cells is mediated through the phosphatidylinositol 3-kinase/Akt pathway and is associated with the enhanced expression of other SR proteins.

Removal of introns from pre-mRNAs by splicing is a crucial step during eukaryotic and viral gene expression. Serine/arginine-rich (SR) proteins and small nuclear ribonucleoproteins are cellular splicing factors that are involved in intron removal (28). The SR proteins modulate constitutive splicing and are also crucial mediators of regulated alternative splicing (5) that act through binding to regulatory elements in a pre-mRNA and recruiting general splicing machinery to the regulatory elements (19). ASF/SF2 (alternative splicing factor/splicing factor 2), a prototype member of the SR protein family, is an essential splicing factor that was originally characterized as promoting the use of the 5′ splice site closest to the 3′ splice site (17, 27).

ASF/SF2 is an essential gene, and its disruption results in loss of cell viability (49) and, in Caenorhabditis elegans, defects in early embryonic development (32). ASF/SF2 regulates the alternative splicing of many eukaryotic and viral pre-mRNAs in vitro and in overexpression studies, including CD45 (30), β-tropomyosin (16), fibronectin EDI (10), simian virus 40 large T and small t antigens (17), adenovirus E1A pre-mRNAs (49, 51), and HIV tat pre-mRNAs (46).

Regulation of RNA splicing by ASF/SF2 is dependent on the expression level of ASF/SF2 and its phosphorylation status (13, 16, 24, 41, 53). However, the observations from in vitro or in overexpression studies may not truly reflect the function of ASF/SF2 under physiological conditions. Although a recent study shows that ASF/SF2 does affect alternative 5′ splice site selection of simian virus 40 early pre-mRNA in DT40 cells under physiological conditions (51), how ASF/SF2 affects alternative 3′ splice site selection under physiological conditions remains largely unexplored.

Bovine papillomavirus type 1 (BPV-1) is a wart-inducing small DNA tumor virus whose late gene expression involves alternative splicing and alternative polyadenylation. The primary viral RNA transcripts expressed from the late promoter have a common late leader 5′ splice site at nucleotide 7385 but use an alternative 3′ splice site at either nucleotide 3225 (proximal) or nucleotide 3605 (distal) for RNA splicing. A proximal 3′ splice site is the 3′ splice site closest to an upstream 5′ splice site over an intron. The spliced RNAs are then polyadenylated at either of two polyadenylation sites (A_E and A_L). It is known that splicing at the proximal 3′ splice site leads to the production of L2 mRNA while the splicing at the distal 3′ splice site results in the production of L1 mRNA (3). However, selection of the distal 3′ splice site is restricted to the granular cell layer of the epidermis while selection of the proximal 3′ splice site occurs in both the granular and spinous layers (3). This differentiation-specific alternative splicing is presumably related to the suboptimal features of both the proximal 3′ and distal 3′ splice site (59, 60).

Recently, we have shown that selection of the two alternative 3′ splice sites is regulated by five viral cis elements (Fig. 1); three exonic splicing enhancers, SE1, SE2, and SE4; and two exonic splicing suppressors, ESS1 and ESS2 (55, 58, 60). SE1, ESS1, and SE2 are positioned between the proximal and distal 3′ splice sites, and SE1 and SE2 each have at least two ASF/SF2 binding sites, although others may exist (56). Our previous studies showed that SE1 and SE2 synergistically promote the selection of the proximal 3′ splice site over the suppression by the ESS1 (56, 59), presumably by interacting with SR proteins, including ASF/SF2, since mutation of either one of the two ASF/SF2 binding sites within the exonic splicing enhancers switches the selection to the distal 3′ splice site (56, 60).

FIG. 1.

Schematic diagram of a BPV-1 late minigene expression vector and its pre-mRNA. Shown on the top is an expression vector containing either a simian virus 40 (SV40) or a cytomegalovirus (CMV) immediate-early (IE) promoter-driven BPV-1 late minigene (open box) with an early poly(A) (A_E) signal. Numbers below the lines are the nucleotide positions in the BPV-1 genome. The vector expresses the late pre-mRNA that has a large deletion in intron 1, removing all of the early promoters. Expression of the blasticidin (Bsd) resistance gene in the vector was controlled by both the bacterial EM7 and cytomegalovirus immediate-early promoters. Below the minigene is the minigene transcript, with the relative positions of SE1, ESS1, SE2, SE4, and ESS2 as well as the nucleotide positions of the 5′ splice site and 3′ splice site shown. Primer pairs used for reverse transcription-PCR are indicated by arrows under the transcripts and named by the location of their 5′ ends. A BPV-1 antisense probe synthesized for the RNase protection assay and its protection products are shown at the bottom of the figure by size in nucleotides (in parentheses).

In this study, we utilized a DT40-ASF cell line in which the endogenous chicken ASF/SF2 has been disrupted and replaced by a human ASF/SF2 controlled by a tetracycline-repressible promoter, creating a useful genetic system for studying the function of the human SR protein ASF/SF2 in vivo (50, 51). We stably expressed BPV-1 late pre-mRNAs in DT40-ASF cells and found that both viral RNA expression and the use of the proximal 3′ splice site in the cells were significantly reduced in the presence of tetracycline but could be rescued by activation of the phosphatidylinositol 3-kinase/Akt kinase pathway.

MATERIALS AND METHODS

Plasmid construction.

Since the BPV-1 late promoter has no activity in cell cultures, a pCMV/Bsd plasmid expression vector (Invitrogen) was used to construct a BPV-1 late minigene driven by either a simian virus 40 or a cytomegalovirus promoter. For plasmid p3086 (pCBG1) construction, the BglII-EcoRI fragment of the pCMV/Bsd vector, containing a blasticidin resistance gene under the control of both a cytomegalovirus promoter and a bacterial EM7 promoter, was ligated to the BamHI-EcoRI fragment of plasmid p2540 (CCB418), an expression vector containing a simian virus 40 promoter-driven BPV-1 late minigene with deletion of all early promoters in intron 1. The resulting plasmid, p3086, had a BPV-1 minigene with a simian virus 40 promoter and an A_E site.

To construct plasmid p3087 (pCBG2), the XhoI-XbaI fragment of the pCMV/Bsd vector was inserted at the EcoRI site of plasmid p3231(CCB458-2) (56) containing a cytomegalovirus promoter and both A_E and A_L sites. The resulting plasmid, p3088 (pCBG3), was digested to produce an XmnI-MluI fragment to replace the MluI-XmnI fragment of p3086, generating plasmid p3087, which is similar to p3086 but has a BPV-1 late minigene driven by a cytomegalovirus IE1 promoter.

Stable transfection.

DT40-ASF cells, a genetically engineered cell line expressing hemagglutinin (HA)-tagged human ASF/SF2 controlled by a tetracycline-repressible promoter as the only source of ASF/SF2 (50), were maintained in complete RPMI 1640 medium (Gibco-BRL) containing 10% fetal bovine serum (HyClone), 1% chicken serum (Sigma), 2 mM l-glutamine, and 50 μM 2-mercaptoethanol. The cells were transfected with 10 μg of each plasmid described above by electroporation with an Electro Cell Manipulator (ECM630; BTX) and selected in the medium described above with the addition of 12 μg of blasticidin (Invitrogen) per ml. The cell line selected with stable BPV-1 expression from plasmid p3086 transfection was designated DT40-ASF-BPV-1-1, and that from plasmid p3087 transfection was designated DT40-ASF-BPV-1-2.

ASF/SF2 depletion, anti-IgM-B-cell receptor cross-linking, reverse transcription-PCR, and immunoblotting analysis.

To deplete ASF/SF2, DT40-ASF-BPV-1-1 and -2 cells were incubated in complete RPMI 1640 medium containing 1 μg of tetracycline per ml (Sigma) for scheduled amounts of time. The cells, at a concentration of 10⁷/ml, were stimulated with 20 μg of anti-chicken immunoglobulin (IgM) μ chain antibody (Bethyl Laboratory Inc.) for 1 h to cross-link B-cell receptors. For the wortmannin inhibition assay, DT40-ASF-BPV-1-1 cells were treated with tetracycline for 48 h and then starved in serum-free RPMI 1640 medium containing 1 μg of tetracycline per ml for an additional 24 h. The starvation of cells, at a concentration of 10⁷/ml, was followed by treatment with 100 nM wortmannin (Calbiochem) for 1 h before the addition of anti-IgM for another 1 h. Total cell RNA was then isolated from DT40-ASF-BPV-1 cells with TRIzol reagent (Gibco-BRL) by following the manufacturer's instructions.

Detection of alternative splicing of the viral RNA was carried out by reverse transcription-PCR with the primer pair Pr7250 (5′-AATTATTGTGCTGGCTAGAC-3′) and Pr3715 (5′-TTTCAGCACCGTTGTCAGCAACTGTG-3′). Because of the low level of viral RNA transcripts expressed in the cells, another internal primer pair, Pr7345 (5′-CAATGGGACGCGTGCAAAGC-3′) and Pr3690 (5′-GAACCAGGTGGTGGTGCAGTTCTCG-3′), was used for nested PCR with diluted (1:3) reverse transcription-PCR products. Total BPV-1 late RNA transcripts and spliced chicken α-actin mRNA were examined by reverse transcription-PCR with the primer pair Pr3609 (5′-TGCTAACCAGGTAAAGTGCT-3′) and Pr3715 (size of the product, 107 bp) or chimeric T7/BPV-1 Pr3738 (5′-TAATACGACTCACTATAGGGA/CAGTATTTGTGCTTGTCCTT-3′) (size of the product, 151 bp) for BPV-1 and the primer pair Pr325 (5′-CCTAGTGAAGGCTGGCTTCG-3′) and Pr654 (5′-GGTGCCAGAT CTTCTCCATG-3′) for chicken α-actin.

All reverse transcription-PCR cycles were determined to produce linear amplification and relatively unsaturated products. The same amount of cells from each treatment (indicated in Fig. 5A and B and Fig. 6B and C) were lysed in 2× SR protein sample buffer (250 mM Tris-HCl [pH 6.8], 4% sodium dodecyl sulfate, 20% glycerol, 10% β-mercaptoethanol, 2 mM sodium orthovanadate, 2 mM phenylmethylsulfonyl fluoride, 40 μM leupeptin/aprotinin) or in Akt buffer A (18), resolved by electrophoresis on a sodium dodecyl sulfate-7.5% polyacrylamide gel, transferred to a nitrocellulose membrane, and immunoblotted for HA-tagged ASF/SF2, phosphorylated SR proteins, SC35, B-cell receptors, phospho-Akt (Ser 473), or Akt with monoclonal anti-HA (Babco), mAb104 (American Type Culture Collection), anti-SC35, polyclonal anti-chicken IgM μ chain (Bethyl Laboratory Inc.), polyclonal phospho-Akt (Ser473), or Akt (New England Biolabs) antibodies, respectively.

FIG. 5.

Anti-IgM-B-cell receptor cross-linking increases expression of phosphorylated SR proteins in DT40-ASF-BPV-1-1 cells lacking ASF/SF2. (A and B) Protein samples from DT40-ASF-BPV-1-1 cells with or without 72-h tetracycline (Tet) treatment or tetracycline plus an additional 1 h of anti-IgM-B-cell receptor cross-linking were collected and blotted with mAb104 (A) or with monoclonal antibody anti-SC35 (B). Antibody mAb104 specifically detects SR proteins with phosphoepitopes. The same membrane was then stripped and reprobed with monoclonal anti-HA antibody (middle blots in panels A and B) for HA-tagged ASF/SF2 or anti-chicken IgM for B-cell receptors (BCR; bottom blots in panels A and B). The ratios of hyperphosphorylated (x○P ) versus phosphorylated SRp30s (○P) (A) or phosphorylated (○P) versus unphosphorylated SC35 (B) were calculated from each lane based on the band intensity. (C) GST-ASF/SF2 and GST-SC35 (0, 0.04, 0.16, and 0.64 μM) were tested for their ability to complement splicing-deficient S100 extract for the in vitro splicing of BPV-1 pre-mRNAs. Arrows indicate the pre-mRNA and the spliced mRNAs. The lengths (in nucleotides) of HaeIII-digested pBR322 DNA size markers are shown.

FIG. 6.

Anti-IgM-B-cell receptor cross-linking stimulates splicing of BPV-1 pre-mRNAs through the phosphatidylinositol 3-kinase/Akt pathway. DT40-ASF-BPV-1-1 cells cultured in medium containing 1 μg of tetracycline (Tet) per ml for 48 h were starved in serum-free, tetracycline (1 μg/ml)-containing medium for an additional 24 h before a 1-h treatment with 100 nM wortmannin (Wort). Some of the cells were stimulated with anti-IgM for another 1 h in the presence of wortmannin. Total cell RNA was extracted, and splicing of BPV-1 pre-mRNA, total BPV-1 RNA, and α-actin pre-mRNA was evaluated in parallel by reverse transcription-PCR (A) as described for Fig. 4. Protein samples from DT40-ASF-BPV-1-1 cells with various treatments were collected in Akt buffer A (18) or SR protein sample buffer, blotted with phospho-Akt (Ser473) antibody, and reprobed with Akt antibody (B) or blotted with mAb104 (C).

RNase protection assay.

The radioactive antisense BPV-1 and chicken α-actin RNA probes were transcribed in vitro in the presence of [α-³²P]GTP from DNA templates prepared by PCR. For BPV-1 DNA template preparation, the chimeric 3′ primer T7/BPV-1 Pr3738 (described above) was used in combination with 5′ primer Pr3121 (5′-ATCATCTTACTTCACATTA/CGCTTTGGTGACGAGGCAG-3′). For chicken α-actin DNA template preparation, another primer set, Pr325 (described above) and the chimeric T7/α-actin primer Pr717 (5′-TAATACGACTCACTATA/GGGCCTCAGTGAGCAGGGTGG-3′), was used. Three nanograms of the probe (specific activity, 35,000 cpm/ng) was hybridized overnight to 100 μg of total DT40-ASF-BPV-1-1 cell RNA or yeast tRNA in hybridization buffer and then digested by RNases A and T₁ as described in the RPA III kit instructions (Ambion). Yeast tRNA was used as a negative RNA control and also as a reference for the efficiency of the RNase digestion of the free probe. Protected RNA fragments were separated on an 8% polyacrylamide-8 M urea gel, and autoradiographic data were captured with a Molecular Dynamics Storm 860 PhosphorImager and analyzed with ImageQuant software.

In vitro splicing assay.

In vitro splicing of ³²P-labeled BPV-1 late pre-mRNA 4 (see Fig. 2 in reference 57) was carried out in a volume of 25 μl with 7 μl of HeLa cell cytosolic S100 extract (36), the indicated amounts of glutathione S-transferase (GST) fusion protein (GST-ASF/SF2 or GST-SC35) (37), and 20 fmol of ³²P-labeled pre-mRNA, followed by incubation at 30°C for 2 h as described previously (35). The products were analyzed by electrophoresis on a 9% polyacrylamide-7 M urea gel.

FIG. 2.

Correlation of ASF/SF2 depletion to splicing of the BPV-1 pre-mRNAs. (A) HA-tagged ASF/SF2 in DT40-ASF-BPV-1-1 cells was immunoblotted with anti-HA-11 antibody at the indicated time points after treatment of the cells with tetracycline (Tet)-containing medium. The same membrane was then stripped and reprobed with anti-chicken IgM for detection of B-cell receptors (BCR) as an internal control. (B and C) To analyze splicing of the BPV-1 pre-mRNAs, total cell RNA was isolated at each time point from DT40-ASF-BPV-1-1 cells incubated in tetracycline-containing medium. The viral RNA was analyzed by reverse transcription-PCR with the sense primer, Pr7250, combined with a ³²P-labeled antisense primer, Pr3715 (B), and then by nested PCR with another internal primer pair, Pr7345 and Pr3690 (C). BPV-1 L1 cDNA and a 100:1 mixture of L2L (spliced at the nucleotide 3225 3′ splice site) and L2S (spliced at the nucleotide 3605 3′ splice site) cDNAs were used as controls (C). A series of the mixtures of L2L and L2S cDNAs were also amplified by PCR for comparison of proximal versus distal 3′ splice site usage (D). Chicken α-actin mRNAs (216 bp) were detected by reverse transcription-PCR (B) with primer sets as described in the text. Labeled reverse transcription-PCR products were separated by electrophoresis on an 8% polyacrylamide-8 M urea gel, and unlabeled reverse transcription-PCR products were separated by electrophoresis on a 1.5% agarose gel. The numbers along the right side are the spliced products, indicated by the nucleotide positions of the splice junction, and along the bottom of panel C are ratios of splicing at the proximal versus distal 3′ splice sites.

RESULTS

Depletion of ASF/SF2 from cells decreases exonic splicing enhancer-dependent splicing of BPV-1 late pre-mRNAs at a proximal 3′ splice site.

To investigate the role of ASF/SF2 in exonic splicing enhancer-dependent splicing of the BPV-1 late transcripts in vivo, we constructed two expression vectors containing BPV-1 late minigenes driven by either a simian virus 40 or a cytomegalovirus promoter (Fig. 1). This allowed us to determine if RNA polymerase II promoter structure could modulate viral RNA alternative splicing and ASF/SF2 effects on exonic splicing enhancers (10, 11).

The two constructs were transfected into DT40-ASF cells, a genetically engineered chicken B-cell line in which the endogenous chicken ASF/SF2 has been disrupted and replaced by HA-tagged human ASF/SF2 controlled by a tetracycline-repressible promoter as the only source of ASF/SF2 (50, 51). Two stable cell lines (DT40-ASF-BPV-1-1 and -2) expressing the respective BPV-1 late minigenes were obtained. In these cell lines, the viral RNAs expressed from the BPV-1 late minigenes were spliced at the proximal 3′ splice site with no detectable unspliced, full-length pre-mRNAs, while selection of the distal 3′ splice site for the viral RNA splicing was minimal, usually below the detection level of a one-time reverse transcription-PCR assay (lanes 2 of Fig. 4A and B) or RNase protection assay (see Fig. 4D, lane 2).

FIG. 4.

Splicing of the BPV-1 pre-mRNAs at the proximal 3′ splice site in cells lacking ASF/SF2 can be rescued by anti-IgM-B-cell receptor cross-linking. Cell lines established from transfection of each expression vector were treated as indicated in each panel. For tetracycline (Tet) treatment, cells were incubated in culture medium with 1 μg of tetracycline per ml for 72 h, and some of the cells treated with tetracycline were stimulated with anti-chicken μ chain IgM for 1 h. Two cell lines, DT40-ASF-BPV-1-1 (A, C, and D) and DT40-ASF-BPV-1-2 (B) were used. The spliced products of the viral pre-mRNAs were analyzed in total-cell RNA by reverse transcription-PCR (A and B) and then by nested PCR (C) as described for Fig. 2. Total viral RNA and the spliced chicken α-actin mRNAs (216 bp) were detected by reverse transcription-PCR (A and B) with primer sets described in the text. Shown on the left of each panel are the sizes (in base pairs) of DNA markers, and on the right the splicing patterns or the sizes of reverse transcription-PCR products are shown. (D) RNase protection assay. Total cell RNA (100 μg) extracted from DT40-ASF-BPV-1-1 cells with or without tetracycline or tetracycline plus anti-IgM treatment was hybridized overnight to the probe depicted in Fig. 1. Yeast tRNA served as a negative control but was not subjected to RNase A and T₁ digestion in lane 6. The protected products at 514 nucleotides (nt) are the products spliced at the proximal 3′ splice site (ss). Numbers on the left are the sizes of a 100-bp ladder DNA.

To determine the role of ASF/SF2 in driving splicing at the proximal 3′ splice site, we depleted ASF/SF2 from the cells by adding tetracycline. Complete depletion of ASF/SF2 could be achieved by the presence of tetracycline for 48 h (Fig. 2A), as has been reported (51). Further analysis of ASF/SF2 depletion and BPV-1 late pre-mRNA splicing in the tetracycline-treated cells showed that the exonic splicing enhancer-dependent splicing at the proximal 3′ splice site was not blocked significantly until the tetracycline treatment was extended to 72 h (Fig. 2B), at which point the cell viability dropped to approximately 70% (data not shown). Significant reduction of the proximal 3′ splice site usage thus lagged 24 h behind the complete ASF/SF2 depletion (compare Fig. 2A and B), but splicing at the distal 3′ splice site became highly visible by reverse transcription-PCR and then nested PCR analysis at this time (Fig. 2C).

The ratio of the relative densities of the spliced products at the proximal 3′ splice site versus distal 3′ splice site dropped from 2.5:1 in the presence of ASF/SF2 (at time 0) to 1:1 in the absence of ASF/SF2 (at 72 h of tetracycline treatment). As a comparison, we performed a series of PCR amplifications with 1:1, 10:1, 100:1, and 1,000:1 mixtures of L2L cDNA (spliced at the nucleotide 3225 3′ splice site) and L2S cDNA (spliced at the nucleotide 3605 3′ splice site) (Fig. 2D) and then nested PCR on the 100:1 mixture for semiquantitative analysis because the estimated ratio of the spliced products at the proximal 3′ splice site versus the distal 3′ splice site in DT40-ASF-BPV-1-1 cells was less than 100:1 (comparing lanes 2 with BPV-1 L2 cDNA [L2L:L2S = 100:1] lanes in both Fig. 4A and B). A ratio of only 1.7:1 was obtained from the nested PCR on the 100:1 mixture.

In separate experiments, the nested PCR performed on a fixed dilution (1:3) of PCR products amplified from L2L and L2S mixtures (1:1, 10:1, 100:1 and 1,000:1) or on a serial dilution of the PCR products obtained from the 100:1 and 1,000:1 mixtures excluded the possibility that the apparent increase in the nucleotide 3605 splicing in the ASF/SF2-deficient DT40-ASF-BPV-1-1 cells might be due to an artifact of nested PCR resulting from changes in total template levels (data not shown). Therefore, the 1:1 ratio of the proximal to the distal 3′ splice sites in the cells with ASF/SF2 depletion at 72 and 96 h of tetracycline treatment implies a truly significant change in the usage of the alternative 3′ splice site. The cellular splicing factor ASF/SF2 thus plays an important role in exonic splicing enhancer-dependent selection of the proximal 3′ splice site.

B-cell activation mediates splicing of BPV-1 late pre-mRNAs in the presence or absence of ASF/SF2.

Activation of B cells through B-cell receptors leads to multiple cellular responses (4, 7), including transcription (34). Cross-linking of an anti-IgM antibody to B-cell receptors on DT40-ASF-BPV-1-1 cells in the presence of ASF/SF2 triggered exonic splicing enhancer-dependent splicing of the viral pre-mRNAs and expression of HA-tagged ASF/SF2. In the cells containing ASF/SF2, viral RNA splicing at the proximal and distal 3′ splice sites peaked at 30 to 60 min after anti-IgM-B-cell receptor cross-linking and declined to a basal level 4 h later (Fig. 3).

FIG. 3.

Splicing kinetics of the BPV-1 pre-mRNA in DT40-ASF-BPV-1-1 cells containing ASF/SF2 after cross-linking of anti-IgM to B-cell receptors (BCR). (A and B) Total cell RNA was isolated from DT40-ASF-BPV-1-1 cells at the indicated time (hours) after cross-linking of anti-chicken IgM to B-cell receptors. Splicing at the nucleotide 3225 3′ splice site in the late mRNAs was analyzed by reverse transcription (RT)-PCR with the same set of primers described for Fig. 2 and quantified with an Agilent 2100 bioanalyzer (Agilent Technology) (A). The usage of the nucleotide 3605 3′ splice site was detected further by nested PCR with another set of internal primers, Pr7345 and Pr3690 (B). BPV-1 L1 cDNA, L2-S cDNA, total DT40-ASF cell RNA, and BPV-1-transfected 293 cell RNA were used as controls. DNA size markers are shown on the left. The relative amounts of the amplified products in A and B are shown in C and D, respectively.

We then repeated the experiment in cells pretreated with tetracycline for 72 h to completely deplete ASF/SF2. Anti-IgM-B-cell receptor cross-linking also stimulated splicing of the viral pre-mRNAs in two different cell lines lacking ASF/SF2 (Fig. 4), but had no effect on the constitutive splicing of abundant chicken α-actin pre-mRNAs (bottom panels of Fig. 4A and B). The increased viral RNA splicing induced by cross-linking in the cells lacking ASF/SF2 correlated with an increased amount of total viral RNA transcripts (middle panels of Fig. 4A and B) and could be blocked by actinomycin D in combination with α-amanitin (data not shown), suggesting cross-linking-mediated viral RNA transcription and splicing.

In addition, the viral RNA splicing enhanced by anti-IgM-B-cell receptor cross-linking in the cells lacking ASF/SF2 also correlated to a slightly increased selection of the proximal over the distal 3′ splice site (Fig. 4C). The ratio of proximal versus distal 3′ splice site usage for the viral RNA splicing in tetracycline-treated (72 h) cells dropped to 1 from 2.2 in the cells without tetracycline treatment, as we showed in Fig. 2C. However, anti-IgM-B-cell receptor cross-linking brought this ratio back a little, to 1.4 (Fig. 4C), suggesting that the rescued proximal 3′ splice site usage was not due to the presence of ASF/SF2.

RNase protection assay analysis (Fig. 4D) showed that 72-h tetracycline treatment of DT40-ASF-BPV-1-1 cells had almost no effect on chicken α-actin RNA expression but caused an approximately ninefold reduction of the viral RNA spliced at the proximal 3′ splice site (compare lane 3 with lane 4, Fig. 4D). In contrast, anti-IgM-B-cell receptor cross-linking on the cells lacking ASF/SF2 rescued the RNA splicing at the proximal 3′ splice site in full or even increased it a little (compare lane 2 with lane 4, Fig. 4D). The data suggest that the cellular splicing factor ASF/SF2 was not involved in the rescue of the exonic splicing enhancer-dependent splicing of the BPV-1 pre-mRNAs at the proximal 3′ splice site by anti-IgM-B-cell receptor cross-linking.

B-cell activation and expression of phosphorylated SR proteins.

To further understand how B-cell activation through anti-IgM-B-cell receptor cross-linking could rescue viral RNA expression and exonic splicing enhancer-dependent splicing at the proximal 3′ splice site in the cells lacking ASF/SF2, protein samples prepared from the DT40-ASF-BPV-1-1 cells were blotted for analysis of phosphorylated SR protein expression with the monoclonal antibody mAb104, which recognizes phosphoepitopes of all species of SR proteins. Two major SR protein bands were detected by mAb104 Western blotting in the cells with and without tetracycline treatment, SRp30s (including 9G8 and SRp30a [ASF/SF2], b [SC35], and c) and SRp55 (Fig. 5A). Surprisingly, two strong SR protein bands, one representing SRp40 and another representing hyperphosphorylated SRp30s, were also detected in the cells with tetracycline plus anti-IgM treatment (Fig. 5A). Expression of hyperphosphorylated versus phosphorylated SRp30s in the cells lacking ASF/SF2 was greatly enhanced (approximately 4.5-fold) by cross-linking.

One of the increased SRp30s was identified as SC35 (SRp30b), which had an approximately twofold increase in its phosphorylated version in the anti-IgM-B-cell receptor cross-linked cells (Fig. 5B). These data suggest that the increased expression of phosphorylated SRp30s, especially SC35, and SRp40 might be responsible for the cross-linking-induced viral RNA expression and splicing at the proximal 3′ splice site. This was not predicted because there is no conserved SC35 binding site and only one potential SRp40 binding site (ACGGG) (31) in the sequence between the two alternative 3′ splice sites.

To confirm whether SC35 stimulates viral RNA splicing at the proximal 3′ splice site, an in vitro S100 complementary assay (36) was conducted with the GST fusion proteins GST-ASF/SF2 and GST-SC35. The S100 extract in this assay contains almost all essential splicing factors but an insufficient amount of SR proteins for RNA splicing. The addition of ASF/SF2 or SC35 to the splicing-deficient S100 extract gives efficient splicing, and thus this assay has been widely used to study splicing of different pre-mRNAs (8, 14, 29, 37).

If the rescued viral RNA splicing at the proximal 3′ splice site in the cells lacking ASF/SF2 was due to the increased expression and phosphorylation of SC35 or another SR protein, the same protein should have the capacity to complement the S100 extract and to confer efficient splicing of the viral pre-mRNAs containing this 3′ splice site in vitro. As shown in Fig. 5C, both GST-ASF/SF2 and GST-SC35 complemented the splicing-deficient S100 extract for splicing of the BPV-1 pre-mRNAs containing the proximal 3′ splice site in a dose-dependent manner (Fig. 5C), although the splicing activity of GST-SC35 was weaker than that of GST-ASF/SF2. The data suggest that other SR proteins, especially SC35 and SRp40, may compensate in vivo for deficiencies in ASF/SF2 for the selection of the proximal 3′ splice site.

The phosphatidylinositol 3-kinase/Akt pathway is involved in B-cell activation-mediated viral RNA splicing.

Based on reports that cross-linking of anti-IgM to B-cell receptors on B cells activates the phosphatidylinositol 3-kinase/Akt pathway (1, 18, 39), we hypothesized that the enhanced viral RNA transcription and pre-mRNA splicing by anti-IgM-B-cell receptor cross-linking in the cells lacking ASF/SF2 might relate to the activation of this pathway. Western blot analysis of Ser473-phosphorylated Akt in DT40-ASF-BPV-1-1 cells with tetracycline or tetracycline plus anti-IgM treatment reproducibly showed that anti-IgM-B-cell receptor cross-linking increased Akt phosphorylation in the cells lacking ASF/SF2 (Fig. 6B).

Wortmannin, a phosphatidylinositol 3-kinase inhibitor (18, 47), was able to block phosphorylation of Akt (Fig. 6B) and also inhibited anti-IgM-triggered splicing of the viral pre-mRNAs (Fig. 6A) at the proximal 3′ splice site in the cells lacking ASF/SF2. These data suggest that the phosphatidylinositol 3-kinase/Akt pathway was indeed involved in the cross-linking-induced, exonic splicing enhancer-dependent splicing of the viral pre-mRNAs at the proximal 3′ splice site and might be the mediator responsible for expression of the phosphorylated SR proteins stimulated by cross-linking in the cells lacking ASF/SF2 (Fig. 5A and B). Therefore, Western blotting was applied to examine expression of phosphorylated SR proteins in the presence of wortmannin. As shown in Fig. 6C, DT40-ASF-BPV-1-1 cells under starvation conditions displayed enhanced expression of phosphorylated SR proteins in response to anti-IgM-B-cell receptor cross-linking, although this response was weaker than in cells that were not starved (Fig. 5A). Interestingly, addition of wortmannin to tetracycline-treated DT40-ASF-BPV-1-1 cells strongly inhibited the cross-linking-induced expression of phosphorylated SR proteins. These results provide the first evidence that activation of the phosphatidylinositol 3-kinase/Akt pathway through anti-IgM-B-cell receptor cross-linking promotes expression of phosphorylated SR proteins.

DISCUSSION

We used a genetic system developed in the chicken B-cell line DT40 (50, 51) to study the function of the cellular splicing factor ASF/SF2 in in vivo splicing of BPV-1 late pre-mRNAs under physiological conditions. This genetically engineered cell line, DT40-ASF, expresses no chicken ASF/SF2 but does express human ASF/SF2 controlled by a tetracycline-repressible promoter and thus is a suitable system for analyzing how human ASF/SF2 regulates alternative splicing of BPV-1 late pre-mRNAs. However, the stable DT40-ASF-BPV-1 cell lines that we established express much less BPV-1 late RNA than do 293 cells transiently transfected by BPV-1 late-minigene expression vectors (56). Splicing of the viral RNAs at the distal 3′ splice site is also minimal in the stable DT40-ASF-BPV-1 cell lines and is usually under the detection level of first-run reverse transcription-PCR. As in 293 cells (56), however, the unspliced, full-length viral late RNAs are not detected in these stable cell lines.

Depletion of ASF/SF2 from the cells decreased exonic splicing enhancer-dependent splicing at the proximal 3′ splice site of the BPV-1 late pre-mRNAs, which led to a notably increased selection in a time-dependent manner, as detected by nested PCR, of a distal 3′ splice site, an alternative 3′ splice site commonly switched on when one of the two ASF/SF2 binding sites in BPV-1 SE1 or SE2 is mutated or deleted (56, 60). These data further support the idea that the coordinated interaction of cellular ASF/SF2 and the two intact viral exonic splicing enhancers under physiological conditions are important for antagonizing the suppression by the viral ESS1 of the selection of the proximal 3′ splice site for late pre-mRNA splicing (56, 59). We also evaluated the response to B-cell activation in cells deficient in ASF/SF2 and showed that through cross-linking of anti-IgM to B-cell receptors on DT40-ASF-BPV-1-1 cells, B-cell activation stimulated not only the phosphatidylinositol 3-kinase/Akt signal transduction pathway but also viral RNA transcription and splicing.

When ASF/SF2 was depleted from the DT40-ASF cells, a significant reduction in the proximal 3′ splice site usage lagged behind the ASF/SF2 depletion by approximately 24 h (Fig. 2). This could be due to the presence of residual ASF/SF2 below the detection threshold of Western blotting. Alternatively, the 24-h lag in response may be because other SR proteins complement ASF/SF2 in the early phase of depletion, as the function of many SR proteins is redundant (32).

In addition, depletion of ASF/SF2 from the cells appeared to reduce levels of the total viral RNA (Fig. 4A and B). Since the loss of ASF/SF2 is lethal and depletion of ASF/SF2 affects cell viability (32, 50), a low level of total viral RNA and thereby of viral RNA splicing is more likely due to the effects of ASF/SF2 deficiency on other cellular genes involved in RNA transcription and processing, even though we did not see a significant effect on the abundance of α-actin expression. When the viral RNA transcription is shut off due to ASF/SF2 deficiency, it may take time to deplete the BPV-1 late mRNAs spliced at the nucleotide 3225 3′ splice site completely due to the stability of those mRNAs, which have a half-life of approximately 12 h (data not shown). Compared to the BPV-1 late mRNAs, chicken α-actin mRNA has a relatively longer half-life (approximately 24 h; data not shown). It is also possible that the unspliced viral pre-mRNAs are very unstable following ASF/SF2 depletion (51). However, this interpretation does not explain the rescue of the BPV-1 late pre-mRNA splicing at the proximal 3′ splice site by anti-IgM-B-cell receptor cross-linking on DT40-ASF-BPV-1-1 cells lacking ASF/SF2.

Taken together, our observations, including the results of the in vitro S100 complementary assay shown in Fig. 5C, suggest that a complementary splicing system in the cells for the viral RNA splicing at the proximal 3′ splice site might exist in the early phase of the ASF/SF2 deficiency. An ongoing ASF/SF2 deficiency might detrimentally affect the splicing of the pre-mRNAs of those transcription and splicing factors, resulting in the reduced viral RNA transcription and splicing that appear approximately 24 h after the complete depletion of ASF/SF2. In this case, the residual viral RNA in the ASF/SF2 deficient cells either would be spliced predominantly at the distal 3′ splice site (Fig. 2C and Fig. 4C) or would remain unspliced for immediate degradation. However, activation by anti-IgM-B-cell receptor cross-linking of the cells lacking ASF/SF2 revived cell transcription and complementary splicing activity by bypassing the requirement for ASF/SF2 and rescuing the viral RNA transcription and splicing at the proximal 3′ splice site (Fig. 4).

In this regard, the expected ratio of the proximal versus distal 3′ splice site usage would not be proportional to that in the cells in the presence of ASF/SF2 because the cells lacking ASF/SF2 would not confer a functional BPV-1 exonic splicing enhancer for selection of the proximal 3′ splice site. Nevertheless, our finding of other SR proteins complementing ASF/SF2 in the cells lacking ASF/SF2 provides in vivo evidence for the functional redundancy of SR proteins.

RNA alternative splicing is often tightly regulated in a cell type- or differentiation-specific manner. The regulation of BPV-1 late pre-mRNA alternative splicing is a good model for elucidating this extremely complicated process in the expression of the genes of higher eukaryotes. The demonstration that ASF/SF2 under physiological conditions promotes recognition of a proximal 3′ splice site for BPV-1 late pre-mRNA splicing suggests that the differentiation-specific alternative 3′ splice site selection of BPV-1 late pre-mRNAs (3) may be attributed to varied SR protein expression and phosphorylation in different layers of the epidermis. Presumably, a higher level of phosphorylated ASF/SF2 and other SR protein expression in undifferentiated or intermediate differentiated keratinocytes facilitates selection of the proximal 3′ splice site, a common 3′ splice site also used for the expression of many BPV-1 early genes, whereas reduced expression of ASF/SF2 and other SR proteins in the most terminally differentiated keratinocytes leads to use of the distal 3′ splice site, resulting in viral L1 mRNA maturation. Examination of the SR protein expression profile in undifferentiated versus differentiated keratinocytes is currently under active investigation.

This report shows that the phosphatidylinositol 3-kinase/Akt signal transduction pathway is coupled to transcription (2, 6, 21, 22, 42) and also to splicing of the viral RNA and expression of phosphorylated SR proteins, especially SC35 (15) and SRp40 (45, 54). The expression of other SRp30s, such as SRp30c (44) and 9G8 (40), may be induced by anti-IgM-B-cell receptor cross-linking (Fig. 5A). Western blotting with an anti-9G8 antibody (8) excluded the possibility of 9G8's being triggered by cross-linking (data not shown), however. We were unable to identify SRp30c because an anti-SRp30c antibody was not available during the course of the study.

The finding that expression of phosphorylated SRp40 is enhanced by activation of the phosphatidylinositol 3-kinase/Akt pathway in DT40-ASF-BPV-1-1 cells is consistent with a report that insulin regulates SRp40 phosphorylation via the same pathway (38). However, even though SC35 was demonstrated in this report to be able to complement a splicing-deficient S100 extract and to promote splicing of BPV-1 late pre-mRNAs containing the proximal 3′ splice site in vitro, how expression of phosphorylated SC35 and SRp40 is triggered by anti-IgM-B-cell receptor cross-linking and how they are directly involved in rescue of BPV-1 RNA splicing in cells lacking ASF/SF2 remains to be elucidated.

Although a majority of the nuclear targets regulated by extracellular signals are transcription factors (Fig. 7), recent reports show that RNA processing is also a target of several signal transduction pathways, including the mitogen-activated protein kinase kinase 3/6-p38 (MKK3/6-p38)(48), ERK mitogen-activated protein kinase (26, 52), protein kinase C (33), and phosphatidylinositol 3-kinase pathways (38). Whether the activities of the SR protein kinases (20, 25), Clk/Sty (9, 12), and topoisomerase I (43) and the activity of RNA polymerase II (23) can be regulated by signal transduction pathways is not yet known. The finding of phosphatidylinositol 3-kinase/Akt-mediated expression of phosphorylated SR proteins in this study and in others (38) provides strong evidence that expression of phosphorylated SR proteins is also one of the major cellular responses to extracellular stimuli.

FIG. 7.

Signal transduction pathways in the regulation of RNA transcription (broken lines) and splicing (solid lines). Three major pathways associated with RNA splicing are phospholipase C (PLC)/protein kinase C (PKC) (33), mitogen-activated protein kinase (MAPK) (26, 48, 52), and phosphatidylinositol 3-kinase (PI3K)/Akt (38). Regulation of RNA splicing involves signal transduction-mediated phosphorylation of SR proteins (38) or subcellular distribution of splicing factors (48). NF-AT, nuclear factor of activated T cells; CREB, cyclic AMP-responsive element binding protein; SRF, serum response factor; TCF, ternary complex factor; SR, serine/arginine-rich protein.