Insulin Regulates Alternative Splicing of Protein Kinase C b II through a Phosphatidylinositol 3-Kinase-dependent Pathway Involving the Nuclear Serine/Arginine-rich Splicing Factor, SRp40, in Skeletal Muscle Cells*

,

Insulin regulates levels of protein kinase C (PKC) 1 ␤II mRNA in skeletal muscle by a novel mechanism that signals the activation of splice sites in the last intron of the pre-mRNA. Exon inclusion results in an mRNA that now encodes the C-terminal region of PKC␤II affecting subcellular localization and substrate specificity of the kinase. The terminal PKC␤Ispecific exon with its 3Ј-untranslated region is spliced to the PKC␤II-specific exon via exon inclusion such that a stop codon is introduced at the splice site, and as a result, the PKC␤I exon becomes part of an extended 3Ј-untranslated region of PKC␤II mRNA (1,2). Therefore, PKC␤II and PKC␤I differ only by their C-terminal 52-50 amino acids, respectively. In contrast to PKC␤I, increased expression of PKC␤II results in activation/ inactivation of the mitogen-activated kinase cascade (3), glycogen kinase synthase 3␤ (4), TLS/Fus (5), insulin receptor signaling (6), cyclin-dependent kinase (CDK)-activating kinase, 2 as well as cell proliferation (8 -10), protein trafficking (11), apoptosis, and glucose transport (12,13).
Pre-mRNA splicing occurs on nuclear spliceosomes, a macromolecular complex consisting of small nuclear ribonucleoproteins, proteins associated with heterogeneous nuclear RNA, and other splicing factors including serine-arginine-rich (SR) proteins (14,15). Exon splicing is highly regulated, and numerous consensus sequences that bind specific factors participate in the control of tissue-specific or developmentally controlled splicing via SR protein-RNA and protein-protein interactions (16). SR and SR-like proteins are characterized by a modular composition with one or more RNA recognition motifs and an arginine and serine domain (RS domain) in which the serine residues can be highly phosphorylated. The RS domain is responsible for protein-protein interactions and nuclear localization (17)(18)(19). SR and SR-like proteins have been implicated in 5Ј-splice site recognition and in the communication of splice sites caused by a network of SR proteins (20). They can bind to exon enhancer motifs that are often purine-rich sequences that promote the use of suboptimal splice sites (21). Their interaction with exon enhancers results in a concentration-dependent influence on alternative splicing (22)(23)(24)(25). Several SR protein kinases have been reported, including a U1 snRNP 70K-associated kinase, SR protein kinase (SRPK1), lamin B receptor kinase, and a family of CDC2-like kinases (10, 26 -28). Both hyper-and hypophosphorylation of SR proteins has been shown to influence splicing (29 -31), and the interaction of SR protein kinases with SR proteins can also influence their subcellular localization (32,33). However, at this time the regula-tion of SR protein kinases by peptide hormone-activated signal transduction pathways has not been demonstrated to our knowledge.
The precise mechanisms by which SR proteins govern alternative splicing are under investigation in many laboratories. One model proposes that different concentrations of spliceosomal proteins in different cell types cause alternative processing of pre-mRNAs. Evidence for this mechanism is based on the variable expression levels of some SR proteins in tissues as a function of cell growth or differentiation (24, 34 -36). Another model proposes the existence of cell and/or developmental specific splicing factors that modulate splice site selection. For example, the female-specific expression of Drosophila transformer protein determines the sexual fate of the fruit fly by directing splicing decisions (37,38). In addition, our recent finding that insulin regulated 5Ј-splice site selection of the PKC␤II-specific exon within minutes after it binds to cell surface receptors suggested a third possibility. SR proteins could regulate alternative splicing via a receptor-linked signaling pathway responding to metabolic change rather than to a change in growth or development (39,40).
It is well known that insulin binding to its receptor activates at least three kinase pathways that can signal to the nucleus (41,42). Insulin-induced mitogen-activated protein kinase (MAPK) activation is associated with mitogenic signaling of insulin, and insulin-induced signal transducers and activators of transcription or JAK/signal transducers and activators of transcription pathways lead to nuclear transcriptional activator and repressor activation involved in cell differentiation (43,44). In contrast, insulin activation of the phosphatidylinositol 3-kinase pathway is associated with metabolic signaling, consistent with the observation that insulin regulates PKC␤II exon inclusion independent of cell growth and differentiation. Therefore, the possibility that a PI 3-kinase-dependent signaling pathway could alter the phosphorylation of post-transcriptional regulatory factors such as SR proteins as a step in the regulation of PKC␤II expression was examined.
Our studies focused initially on SRp40 for the following reasons. One, it was first described as an early response gene (HRS/SRp40). Two, SRp40 concentrations were increased by insulin in the regenerating liver where it is induced as a delayed early gene. Three, SRp40 levels are transcriptionally up-regulated by insulin. Four, SRp40 effects on exon inclusion have been demonstrated previously for the alternative splicing of fibronectin mRNA (45). Finally, increased SR protein concentrations during development, cell differentiation, and cell proliferation determine alternative splicing decisions (24,25,46). In our case, however, insulin regulation of PKC␤II exon inclusion in BC3H-1 myocytes and L6 myotubes occurs within 15 min, prior to SRp40 transcriptional up-regulation and increases in its concentration. This suggested that if SRp40 was involved in the insulin-induced alternative splicing that results in PKC␤II mRNA, there must be another mechanism that influences SRp40 activity, other than changes in concentration.
In the present study, we provide evidence to support SRp40 involvement in the regulation of PKC␤II exon inclusion by insulin via its increased phosphorylation by a PI 3-kinase-dependent pathway.

EXPERIMENTAL PROCEDURES
Cell Culture-Rat L6 skeletal myoblasts (obtained from Dr. Amira Klip, The Hospital for Sick Children, Toronto, Canada) were grown on ␣-MEM supplemented with 10% fetal bovine serum to confluency in 100-mm or 6-well plates. Myoblasts were fused into myotubes by changing media to ␣-MEM supplemented with 2% fetal bovine serum for 4-days post-confluency. The extent of differentiation was established by observation of multinucleation of 85-90% of cells. For experiments, myotubes were incubated in ␣-MEM with 0.1% bovine serum albumin (BSA) for 6 h and placed in phosphate-buffered saline with 0.1% BSA just prior to treatment with insulin.
Preparation of Nuclear Extracts-L6 myotube nuclear extract was prepared from cells treated with or without insulin for 30 min as described by Dignam et al. (47).
Immunoprecipitation of SRp40 -L6 myotubes were collected by centrifugation, and pellets were lysed in 20 volumes of 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, and 50 mM Tris, pH 8.0, with protease inhibitors as follows: benzamidine HCl, 16 g/ml; aprotinin, 10 g/ml; leupeptin, 10 g/ml; phenylmethylsulfonyl fluoride, 1 mM. Cells were placed on ice for 30 min, and insoluble material was pelleted at 12,000 ϫ g for 10 min at 4°C. An aliquot (500 l) of lysate was incubated at a final concentration of 1 g/ml with anti-SRp40 polyclonal antibody followed by agitation at 4°C for 2 h. A 40-l aliquot of protein A-Sepharose beads in a 1:1 suspension with the lysis buffer was added and incubated again for 1 h. After centrifugation at 10,000 ϫ g, beads were washed twice with 1 ml of lysis buffer containing 0.5 M NaCl, followed by a wash in lysis buffer with no NaCl. After adding 50 l of SDS-PAGE sample buffer, the precipitate was boiled for 5 min, centrifuged at 1000 rpm for 5 min before loading on the gel, followed by Western blot analysis.
Overexpression of SRp40 in L6 Rat Skeletal Muscle Cells-L6 myoblasts were stably transfected with pCMV5 (empty vector control) or HRS/SRp40 in pCMV5 (46), using a calcium phosphate/DNA precipitate for 16 h. Cells were then washed twice with phosphate-buffered saline and placed in media for 18 h prior to selection of stable transfectants selected in the presence of 750 g/ml G418. Cells were selected and grown in bulk cultures. Overexpression was demonstrated by Western blot analysis using anti-SRp40 and mAb104 antibodies, described above. Transient overexpression was accomplished using Lipofectin TM (Life Technologies, Inc.) and was normalized by co-transfecting ␤-galactosidase (50).
Transient Transfection of Antisense-Antisense 2Ј-O-methoxyethyl oligonucleotides (20 mers) were designed to bind to a putative SRp40binding region (5Ј-TGGGAGCTTGGCTTGA-3Ј) located 351 bases downstream from the first ␤II exon 5Ј-splice site. The sequence of the antisense was 5Ј-ATTCAAGCCAAGCTCCCCAGC-3Ј. As a control, 4 base mismatches were introduced as indicated, 5Ј-ATTCCAGGCAACCTC-CAAGC-3Ј. Antisense (50 and 100 nM) was introduced into cells using Lipofectin TM transfection for 3 h. Cells were then placed in ␣-MEM with 2% fetal bovine serum overnight prior to treatment with insulin. Total RNA was isolated using Stat-60, and RT-PCR analysis was performed as described below. The transfectivity of L6 myotubes was shown to be Ͼ60% (51).
RT-PCR Analysis-Total RNA (1 g) was used to synthesize first strand cDNA using an oligo(dT) primer and Superscript II reverse transcriptase. Inclusion of the PKC␤II-specific exon was detected using an upstream sense primer corresponding to the C4 kinase domain, common to both PKC␤I and -␤II (5Ј-GTTGTGGGCCTGAAGGG-GAACG-3Ј), and an antisense primer to the -␤IV5 exon common to both transcripts, (5Ј-TGCCTGGTGAACTCTTTGTCG-3Ј). The PCR products would be 159 bp for PKC␤I, 374 bp for PKC␤II (where the first splice site was activated, SSI), and 510 bp for PKC␤II (where the second splice site, SSII, was activated). We found that insulin activated two 5Ј-splice sites in a time-dependent manner in some experiments (39). This assay allows for relative comparison of both PKC␤II versus PKC␤I mRNA levels in the same reaction. Following 35 cycles of a two-step PCR amplification program (95°C, 30 s; and 58°C, 2 min) using Taqplatinum DNA polymerase (PerkinElmer Life Sciences), 50% of the PCR was resolved by electrophoresis on 1.2% agarose gels containing 0.05% ethidium bromide at 120 V for 60 min. An additional set of primers was also used to evaluate only PKC␤II mRNA in some experiments. A sense primer corresponding to the coding region of the ␤II exon (5Ј-CACCCGCCATCCACCAGTCCT-3Ј) was used with antisense corresponding to -␤IV5 as described above. The resulting products would be 227 bp. This assay offers increased sensitivity for detecting the ␤II exon since PKC␤I mRNA was not amplified. For each experiment, ␤-actin was determined to compare RNA levels between samples. PCR products were visualized using a Kodak Digital Analysis System 120. The assay was verified against a competitive RT-PCR system, and equivocal results were obtained with the three PCR methods (39).
Materials-The SRp40 cDNA construct was kindly provided by Dr. Taub (46). Tissue culture media were purchased from Life Technologies, Inc. Fetal bovine serum was from Atlanta Biologicals (Norcross, GA). Porcine insulin was obtained from Sigma. Stat-60 was from Molecular Research Center, Inc. (Cincinnati, OH). The reagents used for polyacrylamide gel electrophoresis were from GradiGels (North Ryde, Australia). Antibody to the phosphoepitope of SR proteins (mAb104) was obtained from hybridoma cells (CRL 2067, ATCC). Anti-PKC␤II (polyclonal antibody), anti-rabbit and anti-mouse IgG and IgM antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-SRp40 (polyclonal antibody) was provided by Dr. Taub. LY294002 was obtained from Calbiochem. LY379196 was provided by Eli Lilly (Indianapolis, IN). ECL reagents were from Amersham Pharmacia Biotech (Arlington Heights, IL). Antisense oligonucleotides were synthesized by ISIS Pharmaceuticals (Carlsbad, CA). PCR primers were synthesized by MSW BIOTECH, Inc. (High Point, NC). Primers for ␤-actin were obtained from CLONTECH (Palo Alto, CA). Superscript II reverse transcriptase was from Life Technologies, Inc. Taq-platinum polymerase was from PerkinElmer Life Sciences. Lipofectin TM was from Promega. All other biochemicals and reagents were purchased from the usual vendors.

Insulin Activation of PI 3-Kinase Results in Exon Inclusion-
Since insulin regulates the alternative splicing of PKC␤II mRNA independent of growth and differentiation, we proposed that an insulin-stimulated PI 3-kinase pathway was involved. An insulin-sensitive p85/p110 PI 3-kinase is recruited to protein phosphotyrosine residues in response to activation of the insulin receptor tyrosine kinase producing phosphatidylinositol 3,4,5-trisphosphate which is necessary for metabolic and some mitogenic actions of insulin (41,53,54), and several protein serine/threonine kinases downstream of PI 3-kinase are implicated in the regulation of glucose transporter recruitment, glycogen synthesis, protein synthesis, and gene transcription, making this pathway a likely candidate for regulating post-transcriptional events such as alternative splicing (55).
To examine involvement of the insulin-stimulated PI 3-kinase pathway, L6 myotubes were pretreated with LY294002, a specific PI 3-kinase inhibitor (52), prior to insulin addition. RT-PCR analysis was used to evaluate changes in PKC␤ pre-mRNA splicing and PKC␤II mRNA production (Fig. 1). Insulin stimulated exon inclusion as evidenced by the presence of the ␤II-specific exon in the mature mRNA as reported previously (39), and LY294002 inhibited this insulin effect. To verify that LY294002 blocked PI 3-kinase, 2-deoxyglucose uptake, also dependent upon the insulin-induced PI 3-kinase pathway, was assessed and inhibited greater than 60% (Table  I).
SR Proteins Are Expressed by L6 Myotubes-Involvement of SR proteins in regulating alternative mRNA splicing has been well-established (20), and members of this family are characterized by a C-terminal domain with extensive arginine/serine motifs that are hyperphosphorylated. Since the phosphorylation state of SR proteins has also been proposed to regulate their function (53), we examined which SR proteins were expressed in L6 myotubes, using a monoclonal antibody (mAb104) raised to the phosphodomain of SR proteins to analyze nuclear extracts from control and insulin-treated cells. At least seven different SR proteins showed increased phosphorylation in response to insulin treatment, including SRp75, SRp55, SRp40, and SRp30a/b (Fig. 2). In control cells which were serum-starved, and represent basal conditions for splicing of the -␤II exon, only three major SR proteins, corresponding to SRp30a/b, SRp55, and SRp75, were detected with mAb104. Thus, although the phosphorylation states of seven SR proteins increased with insulin treatment, a significant change in SRp40 was observed over basal conditions and therefore was likely involved.
SRp40 Was Phosphorylated by a PI 3-Kinase-dependent Pathway in Response to Insulin-Although the L6 cells were  not dividing and represented fully differentiated insulin-responsive skeletal muscle cells, it was possible that insulin increased SRp40 phosphorylation state as well as SRp40 concentration. To examine this possibility, SRp40 was immunoprecipitated in lysates from cells treated with insulin and LY294002 using a polyclonal antibody developed for HRS/ SRp40 (46). The immunoprecipitates were analyzed by Western blot analysis probed with mAb104. As shown in Fig. 3A, following insulin treatment for 15 and 30 min, SRp40 phosphorylation increased. Pretreatment of cells with LY294002 blocked insulin effects on SRp40 phosphorylation. The cell lysates were also analyzed directly by Western analysis using anti-SRp40 antibody. In this, insulin treatment for up to 30 min was shown to have no discernible effect on SRp40 protein levels (Fig. 3B). SRp40 protein levels were also unaltered by LY294002 pretreatment.
These results indicated that in non-dividing L6 myotubes, insulin affects SRp40 function via phosphorylation rather than increasing concentrations of the factor, and SRp40 phosphorylation was blocked by an inhibitor of PI 3-kinase.
SRp40 Overexpression Mimicked Insulin Effects on Splice Site Selection-To establish further SRp40 involvement, SRp40 overexpression experiments were carried out (46). The co-transfection of HRS/SRp40 and a fibronectin minigene in H35 cells was correlated to HRS-mediated regulation of EIIIB exon inclusion, and this correlated to induction of HRS protein and fibronectin EIIIBϩ transcripts in developing liver. By analogy, we stably overexpressed SRp40 cDNA in L6 myoblasts to determine whether overexpression would mimic insulin effects on splicing. As shown in Fig. 4, A and B, increased SRp40 levels resulted in the inclusion of the PKC␤II-specific exon as determined by RT-PCR analysis, and the effect was analogous to that observed for insulin treatment. Basal levels of PKC␤II mRNA were negligible in serum-starved L6 myotubes, but upon stimulation with insulin for 30 min, inclusion of the 216-bp PKC␤II exon was evident. Transient overexpression of SRp40 also mimicked insulin-induced PKC␤II exon inclusion, and the splicing observed in the presence of transient SRp40 expression was also blocked by LY294002 (Fig. 4C).
Protein levels of SRp40 and PKC␤II were also analyzed to determine that the transfected cDNA was expressed and to determine whether the increase in PKC␤II mRNA resulted in newly synthesized protein. The transfection of cells with SRp40 constructs increased levels of the protein, Ͼ5-fold, over endogenous SRp40 levels (Fig. 5A). The phosphorylation of SRp40 in response to insulin also increased 5-fold over non-insulin-stimulated levels as determined using mAb104 to detect phosphorylated SRp40 (Fig. 5B). SRp40 overexpression increased levels of PKC␤II protein expression in a manner consistent with insulin treatment (Fig. 5C).
Effect of Antisense Oligonucleotide Targeted to Putative SRp40-binding Site in the Intron Spanning ␤II-␤I Exons-Since SRp40 mimicked insulin to enhance exon inclusion, the cis-elements involved in the regulation are likely to occur in the exon or intron sequence proximal to the splice site. An SRp40binding motif, TGGGAGCTTGGCTTGA, downstream from the PKC␤II exon 5Ј-splice site was identified from sequence analysis. This site is similar to a cis-element predicted earlier, TGGGAGCNNRGCTCGY, with a 2-bp difference at the 3Ј-end (58). To determine if this sequence might be involved in insulinstimulated splicing, an antisense oligonucleotide was designed. The modification used to synthesize the oligonucleotide ensured that it was RNase H-resistant and would not result in destabilizing the pre-mRNA. A 2Ј-O-methoxyethyl oligonucleo- tide was targeted to bind to the potential SRp40-binding site in the intron spanning ␤II-␤I exons as shown in Fig. 6A. This antisense sequence blocked insulin-stimulated exon inclusion in a dose-dependent manner (Fig. 6, B and C). A control oligonucleotide containing a 4-bp mismatch failed to block exon inclusion in the presence of insulin and confirmed the specificity of the antisense for the target sequence. Thus, by blocking SRp40 protein interaction with the element with the 2Ј-Omethoxyethyl oligonucleotide, insulin-induced splicing was di-rected away from exon inclusion to the alternative product, PKC␤I mRNA.

Effects of Cycloheximide, SRp40 Overexpression, and Antisense Oligonucleotides on Cellular 2-Deoxyglucose Uptake-To
show that the functional consequences of SRp40 overexpression correlated to PKC␤II-mediated metabolic changes induced by insulin action (51), PKC␤II-mediated response on glucose uptake by L6 cells was examined.
By serum-depriving L6 myotubes for longer periods (up to 18 h) before measuring glucose uptake, PKC␤II mRNA and protein levels were demonstrated to be low (Fig. 5C). Under these conditions, glucose uptake reflects a requirement for new protein synthesis if the rapid effect of insulin on splicing of PKC␤II is relevant to a physiological event. To demonstrate this requirement, L6 myotubes were serum-starved for 18 h prior to 2-deoxyglucose uptake (56). As shown in Table I, cycloheximide pretreatment blocked insulin effects on 2-[ 3 H]deoxyglucose uptake. The transient overexpression of SRp40 cDNA in L6 myotubes also mimicked insulin effects by increasing basal glucose uptake, and there was no further stimulation in the presence of insulin. Basal glucose uptake was also highly elevated in cells stably transfected with SRp40, and insulin had no further stimulatory effect (data not shown.) To demonstrate further the importance of the switch in PKC␤ isozymes, cells were transiently transfected with the antisense oligonucleotide (AS 34) shown to block splicing of the pre-mRNA (Fig. 6). At a concentration shown to block exon inclusion and the switch to PKC␤II, insulin effects on glucose transport were totally blocked. The control antisense oligonucleotide with the 4-bp mismatch (AS35) had no effect on glucose transport. Hence, under conditions where insulin stimulation of alternative splicing of PKC␤II mRNA was controlled, glucose uptake was linked to the regulation of SRp40 phosphorylation and PKC␤II splicing by insulin.
As a control, L6 myotubes were pretreated for 2 h with LY379196, which inhibits PKC␤I (IC 50 50 nM) and PKC␤II (IC 50 30 nM). At 30 nM, glucose uptake was blocked 50% of the full insulin effect consistent with the involvement of PKC␤II in insulin-stimulated glucose transport and previous studies demonstrating the effects of PKC␤ inhibitors (51,65). DISCUSSION Insulin is known to activate PI 3-kinase (41), and several downstream kinases are activated in a PI 3-kinase-dependent manner including Akt, PKC, and PKC␤II in skeletal muscle (51,(65)(66)(67). Insulin signaling to the nucleus by PI 3-kinase is not as well studied, but it is associated with activation of gene transcription in addition to its roles in glycogen and protein synthesis and glucose transport (55). Our study found increased phosphorylation of SR proteins by insulin-dependent PI 3-kinase pathways. This indicated that factors involved in 5Ј-splice site selection could be regulated by insulin signaling mechanisms.
The C-terminal portion of SR proteins contains a domain rich in serine and arginine residues that is highly phosphorylated. At least eight members of the SR family, including SRp40 and SF2/ASF, contain phosphoepitopes that are recognized by mAb104 (68). The finding that insulin treatment increased the phosphorylation state of at least seven proteins in nuclear extracts from skeletal muscle cells is consistent with the observation that the RS domains are highly phosphorylated in vivo (10,69).
We focused on HRS/SRp40, a splicing factor with a molecular mass of about 40 kDa that was originally shown to be induced by insulin in rat hepatocytes (71). By using immunoprecipitation of SRp40 followed by detection with an antibody to examine serine phosphorylation specifically, it became evident that insulin treatment increased the phosphorylation state of the SR protein rather than its concentration. The phosphorylation was blocked by pretreatment with LY294002 as was splicing. This indicated a role for PI 3-kinase in alternative splicing, and to our knowledge, this is the first report of a hormone signaling the phosphorylation of an SR protein.
Since the overexpression of trans-factors has also been useful in establishing their role in splicing, SRp40 cDNA was expressed in differentiated myotubes and resulted in exon inclusion. This was used as evidence to link the regulation of SR concentration to splice site selection in previous studies (46). Here, the overexpression was linked to increased RS domain phosphorylation, increased PKC␤II mRNA, increased PKC␤II protein, and to increased glucose transport. It is premature to suggest which downstream kinases or phosphatases are activated or inhibited by PI 3-kinase to result in increased SR protein phosphorylation. It is possible that insulin-activated kinases such as Akt or PKC could phosphorylate SR proteins (72).
Two consensus sequences have been proposed for SRp40binding sites (58,73). Both are present in the introns flanking the ␤II-specific exon. The first sequence occurs prior to the 3Ј-pyrimidine tract (ACDGS). The second sequence we identified by sequence analysis is longer, occurs about 350 bp after the first 5Ј-splice site, and corresponds closely to one described (58). When the second site was targeted using antisense oligonucleotides, exon inclusion was inhibited. The use of antisense oligonucleotides for down-regulating gene expression is well documented where sequences are targeted to block translation or lead to destabilization of the message by RNase H or inhibit transcription by forming triplex structures within the promoter regions of DNA. Antisense oligonucleotides have also been used to restore splicing of mutated pre-mRNA in thalassemic ␤-globin and to redirect splice site selection for Bcl-xS versus Bcl-xL independent of down-regulating gene expression (59,60,74). Here, blocking one site for SRp40-RNA interaction provided mutual dependence of a downstream sequence with insulininduced changes in SRp40 phosphorylation. This finding is analogous to studies where antisense toward SF2/ASF-binding sites blocked splicing of bGH pre-mRNA in vitro (61).
The effect of the PI 3-kinase inhibitor, LY294002, to block glucose uptake has been reported (57), and its effect is consistent with a role for PI 3-kinase in insulin action (53) since this signaling pathway directly links downstream kinases with the recruitment of glucose carriers to the plasma membrane and results in increased glucose uptake. Next, we evaluated the effect of newly synthesized PKC␤II on insulin-stimulated glucose uptake. Cycloheximide blocked the recruitment of glucose carriers in adipose cells (62,63). Although there is a conflicting report in adipocytes (64), differences in the preparation and pretreatment of cells could be involved. In this study, cycloheximide treatment blocked insulin effects in serum-depleted myotubes.
We demonstrated that overexpression of SRp40 mimicked insulin to increase basal glucose uptake. This is consistent with the effects of increased PKC␤II concentrations that occur following SRp40 overexpression. Since increased SRp40 concentrations may be altering alternative splicing of other pre-mRNA in a nonspecific manner, this correlation should be interpreted tentatively. For example, the insulin receptor is also alternatively spliced, and the B form of the receptor is thought to signal more effectively (76). However, the ability of antisense oligonucleotides targeting the SRp40-binding site to block insulin effects on glucose transport suggests that the alteration in splicing alone is responsible for the increase in transport. Finally, LY379196, a PKC␤ inhibitor which blocks glucose uptake in primary mouse myotubes (65), also inhibited insulin effects on glucose uptake.
Our studies in cells support in vitro observations for an SRp40 role in splice site selection where addition of one or more SR proteins to in vitro deficient splicing extracts restored splice site selection in a concentration-dependent manner (7,70). The identification of SRp40 as a component of insulin-regulated splicing was defined by the following criteria: its ability to mimic insulin effects on PKC␤II splicing, inhibition of its phosphorylation state by LY294002, a compound that blocks insulin activation of PI 3-kinase, and the ability of 2Ј-O-methoxyethyl antisense oligonucleotides directed to a putative SRp40 site to block insulin effects on splicing as well as to block insulin effects on glucose transport. The demonstration of SRp40 as a factor regulated by a PI 3-kinase signaling cascade provides an additional mechanism for regulating alternative splicing. SRp40 phosphorylation correlates to alternative splicing of the ␤II exon in a manner analogous to insulin treatment and links a signaling pathway to exon inclusion events in vivo. Unlike systems of tissue-specific alternative splicing, however, the concentration of SRp40 did not change with insulin treatment, rather its phosphorylation state increased. Taken together, the multiple strategies used here to investigate SRp40 interactions in intact cells indicate a pivotal role for this trans-factor and PI 3-kinase in insulin-stimulated alternative splicing of PKC␤ pre-mRNA and subsequent effects of insulin on glucose transport.