Portal de Periódicos da CAPES

Regulation of alternative pre-mRNA splicing by the ERK MAP-kinase pathway

2001; Springer Nature; Volume: 20; Issue: 15 Linguagem: Inglês

10.1093/emboj/20.15.4194

1460-2075

Susanne Weg‐Remers,

RNA Research and Splicing

Article1 August 2001free access Regulation of alternative pre-mRNA splicing by the ERK MAP-kinase pathway Susanne Weg-Remers Corresponding Author Susanne Weg-Remers Forschungszentrum Karlsruhe der Helmholtz-Gemeinschaft, Institut für Toxikologie und Genetik, Postfach 3640, 76021 Karlsruhe, Germany Search for more papers by this author Helmut Ponta Helmut Ponta Forschungszentrum Karlsruhe der Helmholtz-Gemeinschaft, Institut für Toxikologie und Genetik, Postfach 3640, 76021 Karlsruhe, Germany Search for more papers by this author Peter Herrlich Peter Herrlich Forschungszentrum Karlsruhe der Helmholtz-Gemeinschaft, Institut für Toxikologie und Genetik, Postfach 3640, 76021 Karlsruhe, Germany Universität Karlsruhe, Institut für Genetik, Postfach 3640, 76021 Karlsruhe, Germany Search for more papers by this author Harald König Harald König Forschungszentrum Karlsruhe der Helmholtz-Gemeinschaft, Institut für Toxikologie und Genetik, Postfach 3640, 76021 Karlsruhe, Germany Search for more papers by this author Susanne Weg-Remers Corresponding Author Susanne Weg-Remers Forschungszentrum Karlsruhe der Helmholtz-Gemeinschaft, Institut für Toxikologie und Genetik, Postfach 3640, 76021 Karlsruhe, Germany Search for more papers by this author Helmut Ponta Helmut Ponta Forschungszentrum Karlsruhe der Helmholtz-Gemeinschaft, Institut für Toxikologie und Genetik, Postfach 3640, 76021 Karlsruhe, Germany Search for more papers by this author Peter Herrlich Peter Herrlich Forschungszentrum Karlsruhe der Helmholtz-Gemeinschaft, Institut für Toxikologie und Genetik, Postfach 3640, 76021 Karlsruhe, Germany Universität Karlsruhe, Institut für Genetik, Postfach 3640, 76021 Karlsruhe, Germany Search for more papers by this author Harald König Harald König Forschungszentrum Karlsruhe der Helmholtz-Gemeinschaft, Institut für Toxikologie und Genetik, Postfach 3640, 76021 Karlsruhe, Germany Search for more papers by this author Author Information Susanne Weg-Remers 1, Helmut Ponta1, Peter Herrlich1,2 and Harald König1 1Forschungszentrum Karlsruhe der Helmholtz-Gemeinschaft, Institut für Toxikologie und Genetik, Postfach 3640, 76021 Karlsruhe, Germany 2Universität Karlsruhe, Institut für Genetik, Postfach 3640, 76021 Karlsruhe, Germany *Corresponding author. E-mail: [email protected] The EMBO Journal (2001)20:4194-4203https://doi.org/10.1093/emboj/20.15.4194 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Differential gene expression through alternative pre-mRNA splicing is crucial to various physiological and pathological conditions. Upon activation of B and T lymphocytes during an immune response, variant isoforms of the cell surface molecule CD44 are generated by alternative pre-mRNA splicing. We show here that in primary mouse T cells as well as in the murine LB-17 T-cell line upregulation of variant CD44 mRNA species upon T-cell activation requires activation of the MEK–ERK pathway. By employing mutant signaling molecules and a novel luciferase-based splice reporter system we demonstrate that the Ras–Raf–MEK–ERK signaling cascade, but not the p38 MAP-kinase pathway, activates a mechanism that retains variant CD44 exon v5 sequence in mature mRNA. The findings demonstrate that a highly conserved pleiotropic signaling pathway links extracellular cues to splice regulation, providing an avenue for tissue-specific, developmental or pathology-associated splicing decisions. Introduction By alternative pre-mRNA splicing, functionally different proteins can be generated from one and the same gene. Alternative splice sites in pre-mRNA are utilized under the influence of developmental stage-, sex- or tissue-specific differentiation (Chabot, 1996; Lopez, 1998; Smith and Valcarcel, 2000). Alternative splicing is thus an important mechanism of differential gene expression. Inappropriate splicing is involved in the development of various diseases, a fact documented by an increasing number of examples (Lopez, 1998; Philips and Cooper, 2000). Establishment of stage- or tissue-specific splice patterns requires instruction of cells by extracellular cues, i.e. by soluble or cell-associated factors. Changes in alternative splicing of several target pre-mRNA transcripts upon stimulation by growth factors have been reported (Shifrin and Neel, 1993; Fichter et al., 1997; Smith et al., 1997; Liu and Kaczmarek, 1998). Also cytokines (Mackay et al., 1994), hormones (Chalfant et al., 1995; Xie and McCobb, 1998), drugs (Yao et al., 1996), membrane depolarization in neurons (Zacharias and Strehler, 1996) or antigenic stimulation of the T-cell receptor (TCR) (Screaton et al., 1997; König et al., 1998; Lynch and Weiss, 2000) causes splice changes. These extracellular cues are thought to act through the activation of signaling cascades that pass the information to the nucleus. As a result the relative abundance and/or activation state of splice factors (e.g. SR-proteins or hnRNPs) may be changed, which influences spliceosome assembly at alternative splice sites through interaction of the splice factors with cis-active regulatory elements on the pre-mRNA (Smith and Valcarcel, 2000). Even though evidence exists that protein kinase C (PKC) and Ras may regulate alternative splicing (Fichter et al., 1997; Smith et al., 1997; König et al., 1998; Lynch and Weiss, 2000), signaling components between the cell surface and the nuclear splicing machinery have not yet been identified. The CD44 family of type I transmembrane proteins provides a very interesting model to study the regulation of alternative splicing, since CD44 gene expression is extensively regulated by this mechanism and since the resulting CD44 isoforms are known to be involved in tumor progression, immune responses and embryonal development (Günthert et al., 1991; Arch et al., 1992; Sherman et al., 1998). CD44 is encoded by one single gene that spans ∼50 kb. The smallest isoform, CD44 standard (CD44s), is composed of a constant extracellular domain, encoded by exons 1–5 and constant transmembrane and intracellular domains, encoded by exons 15–19 in man and exons 16–20 in mice, respectively (Screaton et al., 1992; Tölg et al., 1993). Variant CD44 isoforms contain additional extracellular domains introduced by alternative splicing of exon 6 (v1–2) to exon 14 (v10) in man, and exon 6 (v1) to exon 15 (v10) in mice, respectively. Inclusion of variant domains influences binding of the CD44 molecule to ligands like hyaluronate (Stamenkovic et al., 1991; Bennett et al., 1995a; van der Voort et al., 1995; Sleeman et al., 1997) and growth factors (Bennett et al., 1995b; Sherman et al., 1998; van der Voort et al., 1999; Herrlich et al., 2000). Therefore, expression of different variant CD44 isoforms changes the functional properties of cells. CD44s is widely expressed in most tissues, whereas expression of variant CD44 isoforms is restricted to certain normal cell types, particularly to proliferating epithelia, and to many malignant tumors (Naor et al., 1997). Moreover, activation of T lymphocytes by injection of allogeneic lymphocytes into adult rats (Arch et al., 1992) or by TCR stimulation by an anti-CD3 antibody or by phorbol-ester treatment (Koopman et al., 1993) results in the generation of alternatively spliced variant CD44 isoforms. As the activating extracellular stimuli for T cells are known, T cells provide a suitable model to investigate signal-regulated alternative splicing of CD44. The small GTPase Ras has been shown to be activated in response to stimulation of the antigen receptor in T cells (Downward et al., 1990) and has been characterized as the central molecule in T cells coupling upstream signaling to a number of different signal transduction pathways including the three major mitogen-activated protein kinase (MAP kinase) cascades. The activity of these MAP kinases, the extracellular signal-related kinase (ERK), the Jun N-terminal kinase (JNK) and p38, is regulated through distinct, hierarchically organized modules of kinases consisting of a MAP kinase kinase (MAPKK, MKK or MEK) and a MAP kinase kinase kinase (MAPKKK). Effectors for Ras include the serine/threonine kinase Raf-1, which via MEK activates ERK, and the GTPase Rac, which via different MKKs activates JNK and p38 (Izquierdo et al., 1993; Genot et al., 1996, 1998; Zhang et al., 1999; Dong et al., 2000; Weiss et al., 2000; reviewed by Genot and Cantrell, 2000; Chang and Karin, 2001). We have shown previously that phorbol-ester stimulation or introduction of activated Ras into the mouse T-lymphoma cell line LB-17 can induce usage of CD44 exon v5 in RNA expressed from a minigene. The retention of exon v5 sequence depended on signal-responsive splice regulatory elements within the exon (König et al., 1998). The goal of the present study is to understand how extracellular stimuli, e.g. activation of the TCR by antigen, are linked to the regulation of alternative splicing. Since the activation of MAP-kinase pathways is a necessary step downstream of Ras in T-cell activation (Ward et al., 1997; DeSilva et al., 1998), we investigated whether MAP-kinase activation is involved in the upregulation of CD44 variant expression, and we found this to be the case. In a model T-cell line into which a novel luciferase-based splice reporter system had been incorporated, the Ras–Raf–MEK–ERK signaling pathway was sufficient and necessary to establish the inclusion of the CD44 v5 exon sequence in mature mRNA. The findings show that extracellular cues can be converted into nuclear changes in splice patterns through a highly conserved cellular protein kinase cascade. Furthermore, they implicate signaling-induced phosphorylation events as regulatory principle in alternative pre-mRNA splicing. Results Inhibition of MEK–ERK pathway activation interferes with upregulation of CD44 variant expression in primary T cells Activation of T lymphocytes leads to changes in alternative splicing of different target genes including CD44 (Arch et al., 1992; Screaton et al., 1997; Lynch and Weiss, 2000). To test which CD44 variants are induced after TCR stimulation we activated primary mouse T lymphocytes with anti-CD3ϵ/anti-CD28 antibodies and measured CD44 variant expression by exon-specific RT–PCR 6 h after induction (Figure 1). We found increased expression of splice variants carrying single variant exons, namely CD44 v3, CD44 v4, CD44 v5, CD44 v6, CD44 v7 and CD44 v8 (Figure 1B and D, compare lanes 4–9), and the appearance of transcripts carrying two and three variant exon sequences, respectively, e.g. CD44 v4/5 and v5/6 (Figure 1B and D, compare lanes 5 and 6). These changes in alternative splicing also occurred when protein synthesis was blocked by emetine (data not shown), suggesting regulation by preformed factors. A similar splice pattern mainly of isoforms carrying single variant exons was obtained by treatment of the cells with the phorbol-ester 12-O-tetradecanoylphorbol-13-acetate (TPA) mimicking TCR stimulation (Figure 1C and D, compare lanes 4–9). Figure 1.Induction of CD44 variant expression by anti-CD3ϵ/anti-CD28 antibodies in murine primary T cells is repressed by inhibition of the ERK pathway. (A) Simplified scheme of the murine CD44 gene organization indicating the positions of PCR primers (arrows) used for exon-specific RT–PCR. Open boxes represent constant exons, filled boxes denote variant exons, thin lines represent intervening intron sequences. (B–F) Exon-specific RT–PCR analysis of CD44 variant expression. Murine primary T cells were either left untreated (B), induced with TPA (40 ng/ml) (C) or stimulated by anti-CD3ϵ/anti-CD28 antibodies coated to plates for 6 h (D–F). In addition, cells were treated either with the solvent dimethylsulfoxide (DMSO) (D), the MEK inhibitor U0126 (E) or the p38 inhibitor SB203580 (F). For exon-specific RT–PCR 5′ primers hybridizing to either the 5′ constant region (C13, lane 1) or to the variant exons (lanes 2–11) were used together with a primer (C2A) recognizing the 3′ constant region. Asterisks mark an unspecific PCR product (König et al., 1996). Identity of bands in lanes 5, 6 and 11 was confirmed by sequencing. The lower band in (D) lane 11 represented a PCR artifact that was only seen in some experiments. (G) Activation of ERK and p38 MAP-kinase pathways: phosphorylation of ERK 1/2 (p42/p44) and p38 was examined by western blotting with phosphospecific antibodies after pre-treatment with U0126 or SB203580 for 15 min and subsequent induction with anti-CD3ϵ/anti-CD28 antibodies for 10 min. (H) Activation of the JNK MAP-kinase pathway: phosphorylation of JNK (arrow) was examined by western blotting with an anti-phospho-JNK antibody 10 min, 1 h, 2 h and 6 h after induction with anti-CD3ϵ/anti-CD28 or after treatment with TPA (40 ng/ml) for 10 min as a positive control. Download figure Download PowerPoint The change in alternative splicing upon TCR activation is obviously mediated by signal transduction processes. We therefore investigated the major signaling cascades that are likely to be activated after anti-CD3ϵ/CD28 co-stimulation. Indeed, the co-stimulatory treatment led to phosphorylation of the ERK1/2 and p38 MAP kinases, as determined by western blot analysis with phosphospecific antibodies (Figure 1G, compare lanes 1 and 2). JNK phosphorylation could not be detected during the relevant time period after co-stimulation (Figure 1H, lanes 1–5). Phorbol-ester treatment, in contrast, resulted in JNK activation (Figure 1H, compare lanes 1 and 6). Our data on ERK and p38 MAP-kinase activation, which are consistent with previously published data (Zhang et al., 1999; Weiss et al., 2000), would be compatible with their involvement in the TCR-dependent upregulation of CD44 variant expression. To investigate whether activation of the ERK or the p38 MAP-kinase pathway, or of both, accounts for TCR-dependent induction of CD44 variant expression, we blocked either the activation of MAP kinase ERK kinase (MEK1/2) with the specific inhibitor U0126 (DeSilva et al., 1998; Favata et al., 1998), or p38 kinase activity with the inhibitor SB203580 (Lee et al., 1994; Ward et al., 1997) (Figure 1G). Inhibition of the MEK pathway by U0126 led to a significant reduction of variant exon-specific RT–PCR products (compare Figure 1D and E, lanes 4–9). Inhibition of p38 signaling with SB203580, however, did not interfere with the generation of CD44 variants (Figure 1F). Quantitative interpretation of the exon-specific RT–PCR is limited even though it was conducted under non-saturated conditions. Nevertheless, the data suggest that the MEK–ERK pathway is involved in the generation of variant CD44 mRNAs upon T-cell activation by TCR and CD28 stimulation. A novel splice reporter system in a T-cell model to study regulated alternative splicing of CD44 To dissect the signaling pathways regulating alternative splicing, we made use of a cell line in which we can modulate signaling by transfection experiments. We chose the transfectable murine T-lymphoma cell line LB-17 2.3, for which we have already shown that CD44 splice regulation resembles that in primary T cells. LB-17 cells express CD3 only very weakly (Zahalka et al., 1993 and our own observations) but expression of endogenous CD44 splice variants can be induced by phorbol-ester treatment (König et al., 1998), as in primary T cells. In this cell line, we previously established a minigene splice reporter containing CD44 exon v5, which recapitulates upregulation of exon v5 inclusion after phorbol-ester stimulation (König et al., 1998). Similarly to primary T cells, phorbol-ester treatment of LB-17 cells led to phosphorylation of ERK, JNK and p38 (Figure 2). Thus, LB-17 cells behave like primary T cells with respect to activation of MAP-kinase pathways. Figure 2.Phosphorylation of ERK, JNK and p38 after phorbol-ester treatment of LB-17 mouse T-lymphoma cells. LB-17 cells (5 × 106 per well) were induced with TPA (40 ng/ml) for 10 min or treated with an equivalent volume of DMSO (solvent) as control, harvested and subjected to SDS–PAGE followed by western blotting with phospho specific and pan-antibodies against ERK (A), JNK (B) and p38 (C). Download figure Download PowerPoint In order to rapidly and quantitatively measure changes of alternative splicing, we developed a novel luciferase-based splice reporter system by modifying the CD44 v5 minigene pETv5 (Figure 3A). The principle of the luciferase reporter gene, pETCatEBLucv5, is illustrated in Figure 3A. Translation only gives rise to a luciferase fusion protein if CD44 exon v5 is included in the reporter gene mRNA. To show that the reporter gene translates exon v5 inclusion into luciferase activity, we transfected the construct into LB-17 cells and stimulated the cells with TPA. We then analyzed the splicing pattern of the minigene mRNAs and, in parallel, measured luciferase activity. RT–PCR analysis of RNA from cells transfected with the luciferase-containing minigene pETCatEBLucv5 (Figure 3B, lanes 1 and 2) gave very similar results when compared with the original pETv5 minigene (Figure 3B, lanes 3 and 4). Whereas the smaller PCR fragment, indicating exon skipping, predominated without induction, TPA treatment resulted in a shift to the larger fragment, indicating exon inclusion. This shift in exon inclusion was accompanied by a 2-fold increase of luciferase activity compared with non-treated cells (Figure 3C), suggesting that luciferase activity reliably reflects exon inclusion. Figure 3.A luciferase-based splice reporter assay for exon inclusion of CD44 v5. (A) Schematic representations of the original pETv5 splice reporter construct (König et al., 1998), the luciferase splice reporter construct (pETCatEBLucv5) and the construct used as a control for transcriptional activation (pETLuc). The scheme for pETCatEBLucv5 illustrates the principle of the assay. Skipping of the v5 exon sequence leads to a reading frame that ends before the luciferase coding sequence (STOP). Inclusion of the v5 exon generates a reading frame for a luciferase fusion protein, ending after the luciferase coding sequence (Stop). Arrow, RSV LTR promoter; ATG, start codon for the fusion protein; gray boxes, insulin exon 2 and 3 sequences; open boxes, CD44 exon v5, and luciferase sequence, respectively; black boxes, SV40 poly(A) signal. Thin black lines, intron sequences; gray lines indicate splice pattern. (B) RT–PCR analysis of LB-17 cells transiently transfected with splice reporter constructs. Cells were harvested and assayed 24 h after transfection with either 2 μg of the splice reporters pETCatEBLucv5 (lanes 1 and 2) or pETv5 (König et al., 1998) (lanes 3 and 4) and treatment with DMSO or TPA (40 ng/ml) for 6 h as indicated. RT–PCR analysis of minigene transcripts resulted in a short fragment (pETv5: 244bp; pETCatEBLucv5: 253 bp) when CD44 exon v5 was skipped, and in a large fragment (pETv5: 354 bp; pETCatEBLucv5: 370 bp) when CD44 exon v5 was included. Identity of PCR fragments was confirmed by sequencing. Schematic drawings of the splice reporter constructs indicating the positions of the primers (arrows) used for PCR are given below the corresponding lanes (see legend to A). M, DNA molecular size marker (100 bp ladder; Gibco-BRL). (C) Detection of luciferase activity. LB-17 cells were transiently transfected with 2 μg of the luciferase splice reporter construct pETCatEBLucv5 or the control reporter for promoter activation, pETLuc. Eighteen hours after transfection, the cells were treated with TPA (40 ng/ml) or DMSO for 6 h before they were harvested and subjected to luciferase assays. Luciferase light unit counts of stimulated cells were divided by counts measured in unstimulated cells to obtain values of fold induction. Each bar represents data of three independent experiments and standard deviations are indicated on top of the bars. Download figure Download PowerPoint To exclude the possibility that increased transcription from the splice reporter plasmid induced by TPA might account for the increase in luciferase activity, we cloned the luciferase coding sequence immediately downstream of the Rous sarcoma virus (RSV) promoter within the vector backbone of the splice reporter (pETLuc, Figure 3A). In LB-17 cells transfected with this construct, luciferase activity after stimulation with TPA was only marginally higher than in unstimulated cells (Figure 3C). To rule out the possibility that varying amounts of primary transcripts expressed from the transfected plasmid changed the ratio of exon skipping and exon inclusion, we transfected increasing amounts of pETCatEBLucv5 into the LB-17 cells. Although increasing amounts of transfected plasmid gave rise to increasing transcript levels, no shift from exon skipping to exon inclusion was observed in the RT–PCR assay (data not shown). Thus, the luciferase reporter construct reliably monitors changes in exon v5 inclusion and recapitulates regulation of alternative splicing of the endogenous CD44 exon v5. Regulation of CD44 exon v5 inclusion by the Raf–MEK–ERK signaling cascade In primary T cells inclusion of exon v5 sequence in mature mRNA required the activation of MAP-kinase pathways. To confirm this result for LB-17 cells and to individually test the participation of signaling components of these pathways we co-transfected the luciferase reporter gene together with dominant-active versions of signaling components into LB-17 cells and determined their effect on splicing. Activation of the ERK and the JNK pathways by these constructs were monitored in parallel by co-transfections with luciferase-based Gal-Elk and Gal-Jun reporter systems (Hibi et al., 1993; Kortenjann et al., 1994). Co-transfection of an expression construct encoding a constitutively active form of Ras, HaRas L61 (Medema et al., 1991), induced CD44 exon v5 usage similarly to TPA stimulation (Figure 4A and B; see also König et al., 1998). HaRas L61 activated the ERK pathway (Gal4-luciferase was activated by co-transfected Gal-Elk; Figure 4C) and to a minor extent the JNK pathway (Gal4-luciferase was activated by co-transfected Gal-Jun; Figure 4D) and also led to phosphorylation of p38 (not shown). Co-transfection of constructs expressing dominant-active mutants of Raf (Raf-BxB) (Bruder et al., 1992), MEK1 (MEK1-DD) (Mansour et al., 1994) and a MEK–ERK fusion construct (MEK1–ERK2-LA) generating a constitutively active ERK kinase (Robinson et al., 1998) had only a marginal effect on JNK but activated the Gal-Elk reporter 10- to 50-fold (Figure 4C) and induced exon inclusion of CD44 exon v5 as demonstrated by upregulation of luciferase activity (Figure 4A) and by a shift from the small (–v5) to the large RT–PCR fragment (+v5) in the RT–PCR assay (Figure 4B). Figure 4.Inclusion of CD44 exon v5 is induced by activation of the ERK MAP-kinase pathway. (A) Detection of CD44 v5 exon inclusion by luciferase activity. LB-17 cells were co-transfected with 2 μg of the pETCatEBLucv5 splice reporter construct and 2 μg of an expression vector for constitutively active HaRas L61, Raf-BxB, MEK1-DD or MEK1-ERK2-LA (filled bars) or the corresponding empty vectors (open bars). Each bar represents values of three independent experiments. Standard deviations are indicated on top of the bars. (B) Detection of CD44 v5 exon inclusion by RT–PCR analysis. LB-17 were co-transfected with 2 μg of pETv5 and 2 μg of constitutively active signaling mutants or vector controls as indicated (pRC-CMV, pkRSPA, pcDNA3, pCMV5). RT–PCR bands were quantified densitometrically after scanning using the Fuji Aida programme. For calculating percentage relative exon inclusion the value of % exon inclusion of control vectors was set to 100%. (C and D) To test for activation of the ERK and JNK pathways, LB-17 cells were co-transfected with 0.2 μg of a construct expressing the Gal-Elk fusion protein or 0.8 μg of a Gal-Jun expression vector and 0.8 μg of the pG5.E4Δ38lux3 luciferase reporter plus 2 μg of control vectors (left bars, open) or expression constructs for the signaling mutants indicated (right bars, filled). Luciferase assays were performed 24 h after transfection. Download figure Download PowerPoint HaRas L61 and to some extent the MEK–ERK fusion and Raf-BxB activated, in addition to the ERK pathway, the JNK pathway as indicated by Gal-Jun activation (Figure 4D). To dissect the relative contributions of MAP-kinase pathway activation, we measured the effect of activated Ras and activated Raf on v5 exon inclusion in the absence or presence of the MEK inhibitor U0126. U0126 inhibited induction of v5 exon inclusion upon expression of activated Ras or activated Raf by 55–70% (see Figure 5A; RT–PCR not shown). The U0126 inhibitor effectively blocked Gal-Elk (Figure 5B) but not Gal-Jun activation (Figure 5C). We interpret this result to indicate a major contribution of MEK–ERK signaling and a minor contribution of most likely JNK or p38 to the regulation of alternative splicing of CD44 v5 by Ras signaling. Figure 5.Ras- and Raf-induced alternative splicing depends on MEK activity. (A) LB-17 cells were co-transfected with 2 μg of pETCatEBLucv5 and 2 μg of expression vectors for constitutively active HaRas L61 and Raf-BxB, respectively, or corresponding empty vectors (pRC-CMV or pkRSPA). U0126 (10 μM) was added 3 h after transfection. Cells were harvested 24 h later and luciferase assays were performed. (B and C) ERK and JNK pathway activation was assayed after co-transfection of the Gal-Elk or Gal-Jun luciferase reporter system with 2 μg of expression plasmids for constitutively active HaRas L61, Raf-BxB or a control vector. Inhibitor treatment was performed as described in (A). Download figure Download PowerPoint A dominant-active form of MAP kinase kinase 6 (MKK6), MKK6-E, described to exclusively activate the p38 MAP kinase pathway (Raingeaud et al., 1996), activated only marginally the Gal-Jun and Gal-Elk reporters (Figure 6A and B) but efficiently phosphorylated a co-transfected Flag-tagged p38 protein (Figure 6C). MKK6-E, however, had only marginal effects on CD44 v5 exon inclusion (Figure 6D; RT–PCR not shown). Thus, p38 activation does not suffice for upregulation of alternative splicing of CD44 v5 in LB-17 cells. Figure 6.Inclusion of CD44 exon v5 can be induced by constitutively active Rac. (A and B) JNK kinase pathway (A) and ERK pathway (B) activity was determined by co-transfection of LB-17 cells with the Gal-Jun or Gal-Elk luciferase reporter system and 2 μg of plasmids expressing constitutively active Ras (HaRas L61), Rac (Rac L61) or MKK6 (MKK6-E) (right bars, filled), or with a corresponding empty vector (left bars, open). (C) Activation of p38 MAP kinase after co-transfection of MKK6-E. LB-17 cells were co-transfected with 20 μg of a Flag-p38 expression construct (lanes 2 and 3) and 20 μg of a plasmid expressing constitutively active MKK6-E or pcDNA3 as control. Flag-tagged p38 (arrow) has a lower electrophoretic mobility compared with endogenous p38 (compare with sample where a control vector has been transfected instead of the Flag-p38 construct, lane 1). (D) Activation of CD44 v5 exon inclusion upon expression of constitutively activated Rac. Two micrograms of the luciferase splice reporter pETCatEBLucv5 were co-transfected with 2 μg of plasmids expressing HaRas L61, Rac L61 or MKK6-E (filled bars), or with an empty expression plasmid (open bars). Cells were harvested 24 h after transfection and luciferase activity was determined. Download figure Download PowerPoint The small GTPase Rac has been described to activate the JNK and the p38 pathways (Lamarche et al., 1996; Genot et al., 1998). It is therefore possible that Ras acts through Rac onto JNK to regulate alternative splicing. In a co-transfection experiment with a dominant-active form of Rac (Rac L61), the JNK reporter indeed responded (Figure 6A), while the ERK pathway was not activated (Figure 6B). When co-transfected with the luciferase splice reporter pETCatEBLucv5, constitutively active Rac upregulated luciferase activity similarly to activated Ras (Figure 6D) and induced a shift from the small (−v5) to the large RT–PCR fragment (+v5) (data not shown). This result is consistent with the possibility that activation of the JNK pathway can suffice to upregulate CD44 exon inclusion in LB-17 cells, and that it may contribute to the induction of CD44 v5 exon usage by Ras. Taken together, we have shown that in both primary T cells and in the LB-17 T-lymphoma cell line induction of CD44 v5 exon inclusion is regulated by the ERK MAP-kinase pathway. Although activation of the JNK pathway may contribute to the regulation of CD44 alternative splicing in the LB-17 cell line, co-stimulation of primary T cells does not lead to JNK pathway activation. Thus, in activated T cells the ERK MAP-kinase pathway is the key pathway for regulation of alternative splicing of CD44, which is in agreement with experiments using pathway-specific inhibitors. Components of the ERK pathway actively contributing to exon inclusion are Ras, Raf, MEK and ERK. ERK-induced CD44 v5 exon inclusion depends on exonic silencer sequences To test whether sequences of CD44 v5 that we had previously identified to be necessary for phorbol-ester-activated exon inclusion (König et al., 1998) are also required for activation of v5 usage by the ERK pathway, we made use of v5 exon mutants. In these mutants exonic silencer sequences in the left part (L) or the middle part (M) of the exon were replaced by a heterologous sequence that confers splice silencing comparable to the wild- type sequence but