Portal de Periódicos da CAPES

The SR splicing factors ASF/SF2 and SC35 have antagonistic effects on intronic enhancer-dependent splicing of the beta -tropomyosin alternative exon 6A

1997; Springer Nature; Volume: 16; Issue: 7 Linguagem: Inglês

10.1093/emboj/16.7.1772

1460-2075

Maria E. Gallego,

Viral Infections and Immunology Research

Article1 April 1997free access The SR splicing factors ASF/SF2 and SC35 have antagonistic effects on intronic enhancer-dependent splicing of the β-tropomyosin alternative exon 6A Maria E. Gallego Corresponding Author Maria E. Gallego Centre de Génétique Moléculaire, Centre National de la Recherche Scientifique, 91190 Gif-sur-Yvette, France Search for more papers by this author Renata Gattoni Renata Gattoni Institut de Génétique et de Biologie Moléculaire et Cellulaire, CNRS/INSERM/ULP, 67405 Illkirch, France Search for more papers by this author James Stévenin James Stévenin Institut de Génétique et de Biologie Moléculaire et Cellulaire, CNRS/INSERM/ULP, 67405 Illkirch, France Search for more papers by this author Joelle Marie Joelle Marie Centre de Génétique Moléculaire, Centre National de la Recherche Scientifique, 91190 Gif-sur-Yvette, France Search for more papers by this author Alain Expert–Bezançon Alain Expert–Bezançon Centre de Génétique Moléculaire, Centre National de la Recherche Scientifique, 91190 Gif-sur-Yvette, France Search for more papers by this author Maria E. Gallego Corresponding Author Maria E. Gallego Centre de Génétique Moléculaire, Centre National de la Recherche Scientifique, 91190 Gif-sur-Yvette, France Search for more papers by this author Renata Gattoni Renata Gattoni Institut de Génétique et de Biologie Moléculaire et Cellulaire, CNRS/INSERM/ULP, 67405 Illkirch, France Search for more papers by this author James Stévenin James Stévenin Institut de Génétique et de Biologie Moléculaire et Cellulaire, CNRS/INSERM/ULP, 67405 Illkirch, France Search for more papers by this author Joelle Marie Joelle Marie Centre de Génétique Moléculaire, Centre National de la Recherche Scientifique, 91190 Gif-sur-Yvette, France Search for more papers by this author Alain Expert–Bezançon Alain Expert–Bezançon Centre de Génétique Moléculaire, Centre National de la Recherche Scientifique, 91190 Gif-sur-Yvette, France Search for more papers by this author Author Information Maria E. Gallego 1, Renata Gattoni2, James Stévenin2, Joelle Marie1 and Alain Expert–Bezançon1 1Centre de Génétique Moléculaire, Centre National de la Recherche Scientifique, 91190 Gif-sur-Yvette, France 2Institut de Génétique et de Biologie Moléculaire et Cellulaire, CNRS/INSERM/ULP, 67405 Illkirch, France The EMBO Journal (1997)16:1772-1784https://doi.org/10.1093/emboj/16.7.1772 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions Figures & Info Exons 6A and 6B of the chicken β-tropomyosin gene are mutually exclusive and selected in a tissue-specific manner. Exon 6A is present in non-muscle and smooth muscle cells, while exon 6B is present in skeletal muscle cells. In this study we have investigated the mechanism underlying exon 6A recognition in non-muscle cells. Previous reports have identified a pyrimidine-rich intronic enhancer sequence (S4) downstream of exon 6A as essential for exon 6A 5′-splice site recognition. We show here that preincubation of HeLa cell extracts with an excess of RNA containing this sequence specifically inhibits exon 6A recognition by the splicing machinery. Splicing inhibition by an excess of this RNA can be rescued by addition of the SR protein ASF/SF2, but not by the SR proteins SC35 or 9G8. ASF/SF2 stimulates exon 6A splicing through specific interaction with the enhancer sequence. Surprisingly, SC35 behaves as an inhibitor of exon 6A splicing, since addition to HeLa nuclear extracts of increasing amounts of the SC35 protein completely abolish the stimulatory effect of ASF/SF2 on exon 6A splicing. We conclude that exon 6A recognition in vitro depends on the ratio of the ASF/SF2 to SC35 SR proteins. Taken together our results suggest that variations in the level or activity of these proteins could contribute to the tissue-specific choice of β-tropomyosin exon 6A. In support of this we show that SR proteins isolated from skeletal muscle tissues are less efficient for exon 6A stimulation than SR proteins isolated from HeLa cells. Introduction The removal of introns from pre-mRNA precursors is an essential and often regulated step in the expression of eukaryotic genes. Although there has been much progress in the last few years in defining the molecular mechanisms and components involved in constitutive pre-mRNA splicing (reviewed in Lamond, 1993; Moore et al., 1993), the mechanisms by which splice sites are selected remain obscure (Black, 1995). Specific recognition of the correct 5′- and 3′-splice sites is a fundamental step in the regulation of alternative splicing. The splice sites of most alternatively spliced exons are close to the consensus sequences and cannot, therefore, be the sole determinants of splice site selection. Clearly other elements, in cis and in trans, are needed to define which splice site will be chosen by the splicing machinery in a given cell environment. Studies of vertebrate genes aimed at understanding the regulation of alternative splicing have led to the identification of sequences essential for the tissue-specific choice of alternative exons (Mardon et al., 1987; Hampson et al., 1989; Streuli and Saito, 1989; Helfman et al., 1990; Libri et al., 1990; Black, 1992; Lavigueur et al., 1993; Sun et al., 1993; Watakabe et al., 1993; Gooding et al., 1994; Del Gatto and Breathnach, 1995; Ryan and Cooper, 1996). Notable among these, purine-rich elements identified in several vertebrate exons, called splicing enhancers, are required for efficient splicing of the resident exons. In general, exonic splicing enhancers appear to activate splicing of the upstream intron (Lavigueur et al., 1993; Watakabe et al., 1993; Xu et al., 1993). These enhancer sequences contain contiguous repeats of the motif GARGAR (R = purine). Specific binding of SR proteins has been shown for enhancers present in exons of several genes. A direct correlation between SR protein binding and activation of splicing has been established for ASF/SF2 in the case of the growth hormone gene (Sun et al., 1993) and SRp40 and SRp55 in the case of the troponin T gene (Ramchatesingh et al., 1995). The SR proteins are essential splicing factors implicated in early steps of the spliceosome formation pathway (Krainer et al., 1990; Fu and Maniatis, 1992b). In fact, commitment of a pre-mRNA to splicing requires binding of a specific SR protein to the pre-mRNA (Fu, 1993; Kohtz et al., 1994; Staknis and Reed, 1994). Studies of in vitro protein–protein interactions and using the yeast two-hybrid system show that the individual RS domains of the SR proteins interact with themselves and with each other (Wu and Maniatis, 1993; Kohtz et al., 1994). In particular, SC35 and ASF/SF2 have been shown to interact with both the 70 kDa protein of the U1 snRNP and the 35 kDa subunit of the splicing factor U2AF (Wu and Maniatis, 1993). These obervations suggest a role of these proteins in early 5′- and 3′-splice site interactions. They function as both essential and alternative splicing factors in vivo and in vitro (Fu, 1995). At least for some of them, alternative splicing activity can be antagonized by increasing the concentration of the general RNA binding protein hnRNP A1 (Mayeda et al., 1992, 1993). This has led to the idea that some cases of tissue-specific splicing may be regulated by simply altering the concentrations or activities of constitutive pre-mRNA splicing factors. In contrast, genetic studies in Drosophila melanogaster have identified several trans-acting proteins that specifically modulate alternative splicing. Regulatory proteins have been identified that can either activate or repress the use of specific 3′-splice sites involved in sex determination (Hodges and Bernstein, 1994). We have used the chicken β-tropomyosin gene as a model to understand the mechanisms underlying the tissue-specific choice of mutually exclusive exons. This gene generates a number of different protein isoforms as a result of tissue-specific alternative splicing of its pre-mRNA. It contains three pairs of mutually exclusive exons. In two of these, the exon choice is related to the use of different promoters or polyadenylation signals. The third pair of exons (6A and 6B) are mutually exclusive and selected in a tissue-specific manner independent of the promoter or polyadenylation site choice. Exon 6A is incorporated in fibroblasts and smooth muscle cells, while exon 6B is skeletal muscle specific (Libri et al., 1989a,b). Several sequences involved in the choice of exons 6A and 6B have been identified both in vivo and in vitro. Sequences at the 3′-end of the intron between exons 6A and 6B and within exon 6B repress the use of the exon 6B 3′-splice site in non-muscle cells (Libri et al., 1990; Gallego et al., 1992). A pyrimidine-rich region present in the intron downstream of exon 6A (S4) has been shown to be essential for recognition of the 5′-splice site of exon 6A (Balvay et al., 1992; Gallego et al., 1992). In this study we sought to identify possible factors that interact with the S4 pyrimidine-rich sequence to facilitate exon 6A splicing. Preincubation of HeLa nuclear extracts with an excess of RNA containing this sequence specifically inhibits exon 6A recognition by the splicing machinery, suggesting that a specific factor is being sequestered so as to inhibit splicing. The S4 sequence can be replaced by a purine-rich sequence (P3AS, similar to the exonic enhancer sequences) as well as its complementary polypyrimidine sequence (P3S) with no change in the splicing efficiency of exon 6A. The fact that the purine-rich sequence stimulates exon 6A splicing prompted us to test the possible role of SR proteins in enhancer-dependent activation of exon 6A. ASF/SF2 stimulates exon 6A splicing through specific interaction with the intronic enhancer sequence. In contrast, the protein SC35 counteracts the stimulatory effect of ASF/SF2 on exon 6A splicing. These results demonstrate that the ratio of the proteins ASF/SF2 to SC35 influences the splicing efficiency of exon 6A in vitro. This leads to the hypothesis that variations in the level of these proteins could contribute to exon 6A exclusion in muscle cells. Indeed, we show that SR proteins prepared from skeletal muscle cells support exon 6A splicing poorly as compared with a HeLa SR protein preparation. Results A nuclear factor specifically recognizes intron sequences responsible for exon 6A recognition in non-muscle cells We have previously identified a pyrimidine-rich sequence of 33 nt in the intron starting 37 nt downstream of the 5′-splice site of exon 6A that is essential for efficient recognition of this splice site (S4; see Figure 3A). Previous studies support the idea that exon 6A activation by the S4 sequence involves specific interactions with trans-acting factors present in HeLa cell nuclear extracts (Balvay et al., 1992; Gallego et al., 1992). To test this hypothesis we have examined the ability of an RNA containing exon 6A and downstream intron sequences to specifically interact with factors present in HeLa nuclear extracts, using native gel electrophoresis and binding competition. 32P-Labeled RNAs were incubated with HeLa nuclear extracts and the resulting complexes were separated on native acrylamide gels. As shown in Figure 1, RNAs containing exon 6A and 200 nt of downstream intron sequences (712 Pmac) associate with extract proteins leading to retarded mobility. The same RNA molecule lacking the S4 sequence (33 nt) does not form a complex after incubation with HeLa nuclear extracts (Δ4 Pmac, deletion of the S4 region). The same result was obtained using RNAs where the 5′- (Δ10) or 3′-half (Δ11) of the S4 sequence was deleted. We have previously shown that both of these deletions (Δ10 and Δ11) weaken the use of the exon 6A 5′-splice site. As a control we analyzed complex formation using an RNA (Δ6 Pmac) where deletion of 25 nt of intron sequences (Δ6) just upstream of the S4 sequence has no effect on exon 6A splicing efficiency (Balvay et al., 1992; Gallego et al., 1992). This RNA gives a band of retarded mobility similar to wild-type RNA (Figure 1, Δ6 Pmac). These results suggest that HeLa nuclear extracts contain a factor(s) that interacts with the 33 nt intronic sequence defined by S4. Figure 1.HeLa cell nuclear extracts contain a factor(s) that specifically binds to the S4 intron sequence. 32P-Labeled RNAs were incubated with HeLa cell nuclear extracts in splicing buffer for the indicated times and the resulting complexes were separated on a 4% non-denaturing acrylamide gel. For each transcript the first lane corresponds to RNA incubated in the absence of nuclear extract. The incubation times are indicated. A schematic diagram of the different RNAs used in the gel mobility assay are shown on the right. Δ4 is a deletion of the S4 sequence (33 nt). Δ10 and Δ11 are deletions of the 5′- and 3′-half respectively of the S4 sequence. Δ6 is a deletion of 25 nt of intron sequence just upstream of the S4 sequence. Download figure Download PowerPoint Figure 2.Specific inhibition of exon 6A splicing by competitor RNA containing the S4 sequence. 32P-Labeled pre-mRNA substrates were incubated with HeLa cell nuclear extracts under standard splicing conditions, either in the absence or presence of unlabeled competitor RNA (S4 Pmac). The substrate in the left panel (6A–7) contains β-tropomyosin exon 6A (63 nt), 194 nt of the downstream intron fused to 90 nt of the intron upstream of exon 7 and 38 nt of exon 7. The pre-mRNA 6A–Δ4–7 is equivalent to 6A–7 but lacks the 33 nt corresponding to the S4 sequence. The substrate on the right consists of β-tropomyosin exons 6B and 7 and the complete intron between them. Splicing reactions were carried out for 2 h. The products of the reaction were separated on a 7% denaturing polyacrylamide gel. Schematic representation of the splicing reaction substrate, intermediates and products are indicated on the right. Boxes represent exons and lines represent introns. Download figure Download PowerPoint Figure 3.Two different sequences can replace S4 for exon 6A activation. (A) Diagrammatic representation of the pre-mRNA containing exons 6A and 7. The intron location of the enhancer sequence S4 and its distance to the exon 6A 5′-splice site are indicated. The nucleotide composition of the sequences used to replace S4 (P3S, P3AS and S5) are presented at the bottom of the diagram. (B) Splicing efficiency of exons 6A–7 of the S4 mutant pre-mRNAs. The names of the different pre-mRNAs are indicated at the top. 6A–7, wild-type pre-mRNA; 6A–Δ4–7, pre-mRNA lacking the S4 sequence; for pre-mRNAs 6A–P3S–7, 6A–P3AS–7 and 6A–S5–7 sequence S4 was replaced by sequences P3S, P3AS and S5 respectively. Reaction mixtures were incubated under splicing conditions in 60% nuclear extract for the indicated times. The position of the pre-mRNA, products and intermediates of the reaction are indicated on the right. (C) Splicing of mutant pre-mRNAs in the presence of S4 competitor RNA. The nature of the pre-mRNA tested as well as the amount of competitor RNA added to the reaction are indicated at the top. Reaction mixtures were incubated for 2 h under standard splicing conditions with 35% nuclear extract. Download figure Download PowerPoint To confirm the specificity of this interaction, competition experiments using wild-type and mutant RNAs were performed. 32P-Labeled wild-type RNA (712 Pmac) was incubated with HeLa nuclear extracts in the presence of different competitor RNAs. Addition of unlabeled wild-type RNA 712 Pmac or Δ6 Pmac reduced complex formation, while addition of Δ4 Pmac had very little effect on complex formation (data not shown). Taken together these results show that the S4 intron sequence specifically recognizes a protein(s) present in HeLa nuclear extracts. Competitor S4 RNA specifically inhibits in vitro splicing of exon 6A If the binding of a factor(s) to the S4 sequence is essential for exon 6A activation, then addition of an excess of unlabeled RNA containing the S4 sequence should specifically inhibit splicing of exon 6A. To test this hypothesis, we have determined the splicing efficiency of exons 6A–7 in the presence of increasing amounts of S4-containing RNA. The reference pre-mRNA used as substrate contains exon 6A, 194 nt of the downstream intron, 90 nt of the intron upstream of exon 7 and 38 nt of exon 7. The RNA used as competitor (S4 Pmac, 160 nt) is equivalent to 712 Pmac but lacking exon 6A and the first 37 nt of the intron downstream. Addition of an excess of S4 Pmac RNA to the splicing reaction strongly inhibited the splicing of β-tropomyosin exons 6A–7, as attested by the decrease in accumulation of mRNA and lariat intron (Figure 2, 6A–7). Similar results have been obtained when an RNA containing only the 33 nt of the S4 sequence was used as competitor (data not shown). To determine whether this inhibition is specifically associated with the presence of the enhancer sequence, competition experiments were performed on pre-mRNAs lacking the enhancer sequence (S4). The splicing efficiency of this pre-mRNA (6A–Δ4–7) is 4- to 5-fold lower than that of the wild-type pre-mRNA (6A–7). However, a measurable level of splicing independent of the presence of the enhancer sequence was still detected. We expect this level of splicing to be unaffected by the competition experiments described above. In fact, addition of the competitor RNA S4 Pmac had only a limited effect on the splicing efficiency of pre-mRNA lacking the intronic enhancer sequence S4 (Figure 2, 6A–Δ4–7). This result implies that the competitor RNA does not behave as a general inhibitor of splicing. We have, as well, determined the splicing activity of β-tropomyosin exons 6B–7 (this RNA does not contain the S4 enhancer sequence) in the presence of the same amounts of competitor S4 Pmac RNA. As shown in Figure 2 (right panel, 6B–7), addition of 1.5 pmol (60-fold molar excess) of S4 Pmac RNA has no effect on the splicing efficiency of β-tropomyosin exons 6B–7. Furthermore, neither human β-globin nor adenovirus pre-mRNA splicing efficiency was affected by the presence of RNA competitor S4 Pmac (data not shown). These results suggest that a titratable factor present in HeLa cell nuclear extracts is able to stimulate exon 6A recognition through interaction with the S4 sequence. A purine-rich sequence can replace the S4 sequence for exon 6A 5′-splice site activation We asked whether a sequence rich in the GAR elements, characteristic of exonic enhancer sequences, could replace the S4 sequence for splicing activity. The nucleotide composition of the sequences used to replace S4 are shown in Figure 3A. The results showing the effect of these substitutions on the splicing efficiency of exons 6A–7 are presented in Figure 3B. Pre-mRNA 6A–P3AS–7, in which the S4 enhancer sequence was substituted by a purine-rich sequence, gave a similar or higher level of splicing than the wild-type transcript (Figure 3B, 6A–P3AS–7). Replacement of S4 by the pyrimidine-rich sequence P3S (which is complementary to the P3AS sequence) fully restored splicing efficiency between β-tropomyosin exons 6A and 7 (Figure 3B, 6A–P3S–7). Similar results have been obtained after transfection of myoblasts with a DNA containing equivalent changes to the S4 sequence (Balvay et al., personal communication). In contrast, deletion of the S4 sequence (6A–Δ4–7), or replacement of the S4 sequence by the S5 sequence (6A–S5–7), led to a decrease in splicing efficiency (Figure 3B). Splicing efficiency of 6A–Δ4–7 and 6A–S5–7 was reduced 5-fold and 3-fold respectively compared with that of 6A–7. Note that the S5 sequence, which corresponds to the intronic 32 nt sequence immediately downstream of the S4 sequence, can be deleted with no effect on exon 6A recognition in vivo (Balvay et al., 1992). Similar results have been obtained in vitro (Gallego et al., unpublished observations). The fact that three different sequences could act as enhancers for activation of the β-tropomyosin exon 6A 5′-splice site raised the question of whether they are recognized by a common factor. To answer this question, we tested the effect of addition of competitor RNA S4 Pmac on the splicing efficiency of pre-mRNAs containing P3AS and P3S as enhancer sequences. The results in Figure 3C show that splicing efficiency of both pre-mRNAs (6A–P3AS–7 and 6A–P3S–7) was reduced in the presence of an excess of RNA containing the S4 sequence, as attested by the strong decrease in accumulation of the mRNA. However, higher concentrations of competitor RNA (between 20 and 40%) are needed to reach an equivalent level of inhibition when compared with the wild-type pre-mRNA 6A–7 (Figure 2). Thus, we conclude that the wild-type enhancer S4 Pmac can titrate a factor(s) needed for enhancer function of the P3AS as well as P3S sequences. In a similar way, an excess of competitor RNA containing either the P3S or the P3AS sequence inhibited splicing of exons 6A–7 independent of the nature of the enhancer sequence (data not shown). Therefore, these results suggest that the three enhancer sequences tested interact with a common set of proteins present in HeLa nuclear extracts. ASF/SF2 but not SC35 or 9G8 rescues splicing of exon 6A in the presence of an excess of S4 RNA The SR proteins have been shown to bind to purine-rich enhancer sequences present in the exons of several alternatively spliced genes (Lavigueur et al., 1993; Sun et al., 1993; Heinrichs and Baker, 1995; Ramchatesingh et al., 1995; Tacke and Manley, 1995; Wang et al., 1995). The fact that a purine-rich sequence can replace the S4 sequence for exon 6A activation led us to ask whether SR proteins participate in enhancer-dependent splicing of β-tropomyosin exon 6A. We initially tested the ability of the different enhancer sequences to bind SR proteins. Uniformly labeled RNAs corresponding to the three different enhancer sequences (S4, P3S and P3AS) and the S5 sequence were UV crosslinked in splicing buffer with purified total SR proteins from HeLa cell extracts (purified as previously described; Zahler et al., 1992). The results in Figure 4 show that the three enhancer sequences interact with the SR proteins and especially with the 30 kDa SR proteins. In contrast, the S5 sequence, which is not able to activate exon 6A 5′-splice site recognition, interacts poorly with the SR proteins. These results indicate a correlation between enhancer activity and the capacity to interact with the 30 kDa SR proteins. Of the known SR proteins, ASF/SF2, SC35 and 9G8 (Cavaloc et al., 1994) possess molecular weights of 30 kDa. To determine whether this binding was functionally significant, we have analyzed the ability of the recombinant proteins ASF/SF2, SC35 and 9G8 (each isolated from baculovirus-infected insect cells) to rescue splicing activity of exons 6A–7 in the presence of the RNA competitor S4 Pmac. The activity of the individual preparations was determined as their capacity to complement a cytoplasmic S100 fraction with a human β-globin splicing substrate. Figure 4.UV crosslinking of SR proteins to the three enhancer sequences. The different RNAs tested (S4, P3S, P3AS and S5) are indicated at the top of the Figure. Aliquots of 50 fmol RNA were incubated with the indicated amount of HeLa SR proteins in splicing buffer with ATP and creatine phosphate for 15 min at 30°C prior to irradiation. Crosslinked proteins were resolved by electrophoresis on a 10% polyacrylamide gel. M, 14C molecular weight markers. Download figure Download PowerPoint The results in Figure 5 show that addition of the protein ASF/SF2 in the presence of 0.8 pmol competitor RNA S4 Pmac strongly increased the splicing efficiency of pre-mRNAs containing any of the enhancer sequences (Figure 5, 6A–7, 6A–P3S–7 and 6A–P3AS–7). However, as expected, addition of ASF/SF2 protein had only a limited effect on the splicing efficiency of a pre-mRNA lacking an enhancer sequence (Figure 5, 6A–Δ4–7). Quantification data show that splicing of 6A–Δ4–7 is 27% stimulated after addition of ASF/SF2, while that of 6A–7 is 125% stimulated. Stimulation of 6A–P3S–7 and P3AS–7 pre-mRNAs by ASF/SF2 was roughly the same as that of 6A–7 pre-mRNA. From these results, we suggest that the level of activation observed on 6A–Δ4–7 is most likely associated with the role of the protein ASF/SF2 as a general splicing factor. However, these results also indicate that in addition to its general role, ASF/SF2 acts specifically through interaction with the enhancer sequences. In agreement with in vitro splicing experiments, UV crosslinking shows that purified ASF/SF2 binds to the three enhancer sequences, while no binding is observed on the S5 sequence (Figure 6). In contrast, neither the SC35 nor 9G8 proteins are able to rescue splicing of pre-mRNAs containing either of the enhancer sequences (data not shown). We conclude that the binding of ASF/SF2 to the enhancer intronic sequence downstream of exon 6A is functionally significant and responsible for activation of exon 6A splicing. Figure 5.ASF/SF2 stimulates splicing of enhancer-containing pre–mRNAs in the presence of competitor RNA. The names of the different pre-mRNAs used as substrates for the splicing reaction are indicated at the top. The presence or absence of 0.8 pmol competitor RNA (S4 Pmac) on the splicing reaction is indicated by the signs + or −. The amounts of recombinant ASF/SF2 added to each reaction are indicated. Reaction mixtures were incubated under splicing conditions for 2 h. The splicing products and intermediates are represented on the right. Download figure Download PowerPoint Figure 6.Recombinant ASF/SF2 binds specifically to the three enhancer sequences. The different RNAs tested (S4, P3S, P3AS and S5) are indicated at the top of the Figure. Aliquots of 50 fmol RNA were incubated with the indicated amount of proteins in splicing buffer with ATP and creatine phosphate for 15 min at 30°C prior to irradiation. Crosslinked proteins were resolved by electrophoresis on a 10% polyacrylamide gel. The negative control lanes (S5) are overexposed relative to the others in order to emphasize the absence of protein interaction. Download figure Download PowerPoint SC35 inhibits splicing of exon 6A in an enhancer–specific manner We have tested the capacity of a SR protein preparation to rescue splicing activity between exons 6A and 7 in the presence of the competitor RNA S4 Pmac. Activity of the SR proteins was controlled with a β-globin or adenovirus splicing substrate after addition to S100 extract. Results presented in Figure 7 indicate that addition of SR proteins can rescue splicing of exons 6A–7 for pre-mRNAs containing either the wild-type enhancer sequence S4 (Figure 7, 6A–7) or the purine-rich enhancer P3AS (Figure 7, 6A–P3AS–7). As expected, no stimulation of splicing activity for a precursor RNA lacking an enhancer sequence (Figure 7, 6A–Δ4–7) was detected after addition of SR proteins. Surprisingly, splicing of the pre-mRNA containing the pyrimidine-rich enhancer sequence P3S was not activated after addition of the SR proteins (Figure 7, 6A–P3S–7). This result is in apparent contradiction to those described above showing that the SR protein ASF/SF2 is able to rescue splicing of exons 6A–7 independent of the nature of the enhancer sequence (Figure 5). Figure 7.The SR proteins stimulate exon 6A splicing in an enhancer-specific manner. The pre-mRNA substrates containing an enhancer sequence (6A–7, 6A–P3S–7 and 6A–P3AS–7) or lacking the S4 sequence (6A–Δ4–7) were incubated under standard splicing conditions. The absence or presence of 0.8 pmol competitor RNA (S4 Pmac) is indicated (− or +). The amount of SR proteins from HeLa cells added to each reaction is indicated in μg. After 2 h incubation, splicing products were separated in a 7% denaturing acrylamide gel. The position of the precursor, products and intermediates are shown on the right. Download figure Download PowerPoint The simplest hypothesis to explain this discrepancy would be the presence in the total SR preparation of a protein able to counteract the effect of ASF/SF2 on exon 6A activation. As mentioned above, the three enhancer sequences UV crosslinked a protein(s) of 30 kDa in a total SR protein preparation. We decided to determine whether the 30 kDa SR proteins SC35 and 9G8 could counteract ASF/SF2 for exons 6A–7 splicing. In fact, addition of recombinant SC35 protein to the in vitro splicing reaction strongly inhibited splicing of exons 6A–7 for pre-mRNAs containing the P3S enhancer sequence (Figure 8, 6A–P3S–7) and to a lesser extent that containing the wild-type enhancer sequence S4 (Figure 8, 6A–7). Addition of 0.6 μg SC35 reduced the splicing efficiency of 6A–7 and 6A–P3S–7 pre-mRNA by 55 and 75% respectively. This result may explain the inability of a total SR protein preparation to rescue splicing of exons 6A–7 in the presence of the P3S enhancer sequence. In contrast, only a limited effect was observed on the splicing efficiency of the pre-mRNA containing the enhancer sequence P3AS (15%), indicating that SC35 does not have a general inhibitory effect on exon 6A recognition (Figure 8, 6A–P3AS–7). We conclude that exon 6A splicing inhibition by SC35 is correlated with the nature of the enhancer sequence. Figure 8.Addition of SC35 to HeLa nuclear extracts inhibits splicing of exons 6A–7. Pre-mRNA subtrates containing each of the three enhancer sequences were incubated under splicing conditions in the presence of the indicated amounts of baculovirus recombinant SC35. Splicing of human β-globin was not inhibited after addition of equivalent o