A novel polypyrimidine tract-binding protein paralog expressed in smooth muscle cells.

Polypyrimidine tract-binding protein (PTB) is an abundant widespread RNA-binding protein with roles in regulation of pre-mRNA alternative splicing and 3'-end processing, internal ribosomal entry site-driven translation, and mRNA localization. Tissue-restricted paralogs of PTB have previously been reported in neuronal and hematopoietic cells. These proteins are thought to replace many general functions of PTB, but to have some distinct activities, e.g. in the tissue-specific regulation of some alternative splicing events. We report the identification and characterization of a fourth rodent PTB paralog (smPTB) that is expressed at high levels in a number of smooth muscle tissues. Recombinant smPTB localized to the nucleus, bound to RNA, and was able to regulate alternative splicing. We suggest that replacement of PTB by smPTB might be important in controlling some pre-mRNA alternative splicing events.

The importance of post-transcriptional mechanisms of gene regulation has been emphasized by the relatively modest number of genes in the human genome (1,2). Alternative splicing, RNA editing, and alternative translational initiation all allow for more than one protein isoform to be produced by individual genes. Alternative splicing is the most prevalent of the posttranscriptional mechanisms for producing protein isoforms. Conservative estimates predict that one-to two-thirds of human genes are alternatively spliced, and some of these genes have the potential to produce thousands of isoforms (reviewed in Refs. [3][4][5][6][7]. Regulation of alternative splicing involves the interaction of cellular trans-acting factors with specific cis-acting regulatory elements within a target pre-mRNA (8,9). These regulatory interactions influence the recognition of splice sites by the splicing machinery. Such regulation can be positive, involving activator factors and enhancer sequences. Conversely, repressor proteins can mediate their influence via silencer elements. Although some model systems of regulated splicing involve the presence or absence of a single regulatory protein, the majority of examples appear to be more complex, with regulatory decisions being achieved by particular combinations of regulatory factors, each of which is expressed more widely than the splicing event that is being regulated (7)(8)(9).
The heterogeneous nuclear ribonucleoproteins (hnRNPs) 1 are a group of abundant and widespread nuclear proteins with diverse roles in pre-mRNA and mRNA function, including the regulation of alternative splicing (10,11). Polypyrimidine tractbinding protein (PTB) 1 (reviewed in Refs. 12 and 13), also known as hnRNP-I, is a prominent member of this family. PTB was originally identified as a potential splicing factor due to its ability to bind to polypyrimidine tracts at 3Ј-splice sites (14 -16). However, it was subsequently recognized to act as a splicing repressor at particular splice sites (17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31). PTB also plays roles in nuclear pre-mRNA 3Ј-end processing (32,33), cytoplasmic internal ribosomal entry site-driven translation (34), mRNA localization (35), and regulation of mRNA stability (36). Consistent with these varied roles, PTB can shuttle between the nucleus and cytoplasm, but is predominantly localized in the nucleus (37). The optimal RNA binding sequence for PTB (UCUU in a pyrimidine-rich context) is found within silencer elements that act by binding PTB (26). These elements are often found within the 3Ј-splice site polypyrimidine tract; and in some cases, PTB acts by directly competing for binding to the polypyrimidine tract with the splicing factor U2AF 65 (23,27). However, PTB-binding sites are also found in other locations in the region of PTB-regulated exons, so it may also be able to inhibit splicing in other ways (reviewed in Ref. 13).
Consistent with its expression pattern, most PTB-mediated repression of specific exons is widespread. Regulated selection of the exons occurs in a small subset of tissues where the repressive action of PTB is either absent or in some way modulated. PTB exists in two major alternatively spliced isoforms, termed PTB1 and PTB4, which arise from skipping or inclusion, respectively, of exon 9, which encodes a 26-amino acid insert. A minor isoform, PTB2, is produced by inclusion of exon 9 using an internal 3Ј-splice site, giving a 19-amino acid insert. In at least one case, these isoforms have differential activity, with PTB4 being more repressive upon ␣-tropomyosin exon 3 than PTB2 or PTB1 (29). In addition to the PTB isoforms, at least two paralog genes, nPTB/brPTB and ROD1, with ϳ70% amino acid identity to PTB have been identified. nPTB/brPTB (38,39) is expressed predominantly in neuronal cells, whereas ROD1 is expressed mainly in hematopoietic cells (40). Cells expressing nPTB or ROD1 tend to express less PTB; and in the case of the alternative N1 exon of c-src, nPTB is less repressive than PTB, contributing to the neuronal selection of the N1 exon (39). The effect of nPTB on other splicing events is similar to that of PTB (29). Although no functional data have been reported on ROD1, a reasonable proposal is that, in neuronal or hematopoietic cells expressing nPTB or ROD1, many PTBrepressed splicing events will be unaffected, whereas a specific subset will be altered. Alterations in the expression of the alternatively spliced isoforms of PTB or in the expression of paralog genes therefore provide one way in which PTB activity can be modulated.
We have been investigating two alternative splicing events that are regulated in smooth muscle (SM) cells. In ␣-actinin, a SM-specific exon is repressed by PTB in non-SM cells, leading to inclusion of the mutually exclusive alternative non-muscle (NM) exon (see Fig. 1A) (28,41). In ␣-tropomyosin (␣-TM), exon 2 is included only as a result of repression of the mutually exclusive exon 3 in SM cells (see Fig. 1C) (42). This repression is mediated in part by high affinity PTB-binding sites on either side of exon 3 (21,26). In vitro splicing experiments have shown that PTB mediates a low level of exon 3 repression in nonmuscle extracts (29). However, in vivo, full repression is observed only in SM cells. In this respect, PTB-mediated repression of ␣-TM exon 3 differs from all other characterized splicing events regulated by PTB.
In an attempt to understand how the tropomyosin and actinin splicing events are regulated, we investigated the expression of PTB isoforms in dedifferentiating rat aorta smooth muscle (RASM) cells. We found no change in the ratio of PTB1 and PTB4. However, in a number of SM tissues, we detected expression of a novel PTB paralog that is distinct from PTB, nPTB/brPTB, and ROD1. We refer to the new paralog as smPTB due to its initial identification and high levels of expression in SM tissues. smPTB is ϳ70% identical to PTB and has additional 36-and 22-amino acid inserts in two of the linker regions separating RNA recognition motif (RRM) domains. Recombinant smPTB binds RNA in vitro and has splicing inhibitory activity. We propose that expression of smPTB may play a role in switching subsets of alternative splicing events in SM and other cells.

EXPERIMENTAL PROCEDURES
Analysis of Rat Aorta SM and Tissue RNAs-Primary RASM cells were isolated by enzymatic dispersion and cultured as described (43,44). Total RNA from SM cells and tissues was made using TRI reagent (29). Splicing patterns of the different genes were analyzed by reverse transcription (RT)-PCR. Oligo(dT) was used for reverse transcription by avian myeloblastosis virus reverse transcriptase (45). PCR was carried out with a 32 P-end-labeled primer using a hot start of 92°C for 3 min, followed by enzyme addition at 80°C. Thirty cycles were carried out at 94°C for 30 s, annealing temperature (variable) for 30 s, and 72°C for 60 s, with a final extension at 72°C for 2 min. A 55°C annealing temperature was used for primer sets 5Ј-vin/3Ј-vin, TM1/TM4, P30/31 (smPTB ϩ PTB) (see Fig. 4), P37/Pd3Ј2 (smPTB ϩ PTB) (see Fig. 2), P37/P38 (PTB-specific), and P12/P38 (human PTB plasmid markers). A 58°C annealing temperature was used for primer set P32/P33 (smPTBspecific). A 60°C annealing temperature was used for primer set EF1a5Ј/3Ј-act. In each case, the 5Ј-primer is stated first. For the PTB plasmid marker PCR, 10% Me 2 SO was included. Products were analyzed on 4% denaturing polyacrylamide gels. The sequences of the oligonucleotides used for PCR are as follows: Cloning-Constructs were prepared by standard cloning techniques (45). Full-length smPTB and PTB were cloned into vectors for in vivo transfection (pCMVSPORT), localization (pEGFPN1 or pEGFPC1), and overexpression in Escherichia coli (pET21). Rapid amplification of cDNA ends (RACE) was used to amplify smPTB in two halves using gene-specific primers from day 0 rat aorta RNA. 3Ј-RACE (45) was carried out as a first round with oligo(dT) and P32 (see below) as the forward gene-specific primer (30 cycles of 94°C of 30 s, 55°C for 30 s, and 72°C for 60 s). This was followed by nested PCR with a forward gene-specific primer (P10) under the same cycle conditions. For 5Ј-cDNA, SMART RACE (Clontech) was used. Reverse transcription was carried out using Moloney murine leukemia virus reverse transcriptase with oligo(dT) plus a SMART oligonucleotide (P77). The first round of PCR was carried out (30 cycles of 94°C for 30 s, 58°C for 30 s, and 68°C for 60 s) with a 5Ј-primer mixture at a ratio of 5:1 P79/P78 plus the gene-specific 3Ј-reverse primer P33 (see above). Advantage GC2 polymerase (Clontech) was used. This was followed by 25 cycles (94°C for 30 s, 60°C for 30 s, and 68°C for 2 min) of nested PCR with a 5Ј-primer (P80) and a gene-specific primer (P11 or P41). The oligonucleotides used for RACE were as follows: P10, 5Ј-TTCCTCAAGCTGC-AGGCTTGGCCA-3Ј; P11, 5Ј-GGGAGCCTGAGGTGACCATTGAGC-3Ј; P32, 5Ј-ATGACAGTCAGCCCTCTCCGGTC-3Ј; P41, 5Ј-GAGCTGAGG-CCATGTTCTGGACCTGGACCGGA-3Ј; P77, 5Ј-AAGCAGTGGTAACA-ACGCAGAGTACGCGGG-3Ј; and P80, 5Ј-AAGCAGTGGTAACAACGC-AGAGT. The sequence of the complete rat smPTB open reading frame has been deposited in the GenBank TM /EBI Data Bank (accession number AY223520).
In Situ Hybridization-35 S-Labeled sense and antisense RNA probes corresponding to mouse smPTB amino acids 100 -182 were transcribed in vitro and hybridized to day 10, 14, and 15 mouse embryo sections as described (46).
Localization of smPTB-PAC-1 or HeLa cells grown on coverslips were transiently transfected, using LipofectAMINE (Invitrogen), with PTB or smPTB tagged with green fluorescent protein (GFP) at the C or N terminus. After 24 or 48 h, the coverslips were inverted on a microscope slide, and GFP was visualized using a Zeiss fluorescent microscope with an MC100 camera attached for photographs.
Expression and Purification of Recombinant Proteins-C-terminally His 6 -tagged recombinant proteins were expressed and purified as previously described (46) with one modification for smPTB. To solubilize smPTB, it was necessary to re-extract the pellet after lysis using the non-detergent sulfobetaine (1 M; NDSB-201) in lysis buffer (47).
UV Cross-linking-High specific activity [␣-32 P]UTP-labeled RNA probes (20 fmol) were incubated with recombinant protein in 12 mM Hepes, pH 7.9, 100 mM KCl, 3% glycerol, 0.1 mM EDTA, 0.3 mM dithiothreitol, and 50 g/ml E. coli rRNA for 20 min at 30°C. For competitive binding experiments, PTB4 and smPTB were premixed before addition of the RNA. Heparin was added to 0.25 mg/ml, and the reaction was left for 5 min at room temperature. Samples were irradiated on ice at 254 nm in a Spectrolinker cross-linker with a controlled energy dose of 1.92 mJ. RNA was digested with RNases T1 (1 unit/l) and A (0.4 mg/ml), and the samples were run on SDS-polyacrylamide gels.
Cell Culture, Transfection, and Analysis of Cellular RNA-PAC-1 cells were grown in Dulbecco's modified Eagle's medium containing 10% fetal calf serum. Transient transfection was carried out using Lipo-fectAMINE; total RNA was isolated using TRI reagent; and RT-PCR for ␣-actinin was carried out as previously described (29).

RESULTS
Alternative Splicing Changes in RASM Cells-In view of the evidence for the involvement of PTB in controlling alternative splicing of ␣-TM and ␣-actinin and of the differential activity of the alternatively spliced PTB isoforms upon ␣-TM splicing (29), we decided to analyze expression of PTB in a well characterized model cell system in which these events are regulated. Freshly isolated RASM cells are initially highly differentiated, but, over the course of 4 -6 days in culture, become dedifferentiated and cease to express a number of genes associated with the differentiated contractile state (43). We analyzed alternative splicing of ␣-TM, ␣-actinin, and vinculin/meta-vinculin in RNA from RASM cells at day 0 or 4 in culture and from the PAC-1 pulmonary artery SM cell line (Fig. 1). RT-PCR analysis of ␣-actinin splicing indicated that, at day 0, the major product included the SM exon, with a small amount containing the larger NM exon (Fig. 1A). A small quantity of product corresponding to skipping of both the NM and SM exons was also observed. Like the SM isoform, this "double-skipped" product, which has not been reported before, would result in a nonfunctional EF-hand domain. By day 4, there were roughly equal amounts of the SM and NM isoforms, with a decrease in the amount of the skipped product. The cultured PAC-1 cells were similar to the day 4 cells, but with more NM than SM inclusion and no double-skipped product. RT-PCR analysis of vinculin ( Fig. 1B) showed that the meta-vinculin isoform was expressed as a minor isoform only in day 0 RASM cells and was undetectable in day 4 and PAC-1 cells. Mutually exclusive splicing of ␣-TM exons 2 and 3 produced products of identical size. To differentiate them, the radiolabeled PCR products were digested with XhoI, which cuts within exon 2 to produce a 145nucleotide product, or with PvuII, which cuts exon 3 products to produce a 150-nucleotide band (Fig. 1C). Double digests showed that the PCR product could be fully digested by both enzymes. The almost complete XhoI digestion and PvuII resistance of the day 0 PCR product showed that fully differentiated RASM cells predominantly expressed the exon 2-containing ␣-TM isoform. By day 4, PvuII digested a greater proportion of the PCR product compared with XhoI, indicating a substantial switch toward inclusion of exon 3 instead of exon 2. In comparison, in cultured PAC-1 cells, the majority of ␣-TM RNA contained exon 3. PAC-1 cells commonly show a greater degree of regulated splicing than observed here (21,42,48), but they served as a useful undifferentiated control sample. Taken together, the data indicate that the three alternative splicing events analyzed in RASM cells showed a substantial switch toward the non-SM pattern after 4 days in culture.
Identification and Expression of smPTB-Having observed the switch in alternative splicing of ␣-TM and ␣-actinin, both of which are regulated by PTB, we next analyzed expression of the PTB isoforms. RT-PCR was carried out using primers P37 and P38, which correspond to exons 8 and 11, respectively. This analysis allows the detection of alternative splicing of exon 9, which gives rise to the PTB1 and PTB4 isoforms. Unlike ␣-TM, ␣-actinin, and vinculin, alternative splicing of the PTB1 and PTB4 isoforms showed no significant changes between the day 0 and 4 RASM and PAC-1 samples (Fig. 2B). Therefore, despite the fact that PTB4 has been shown to be a more active repressor of ␣-TM exon 3 compared with PTB1 (29), changes in the ratio of the PTB isoforms do not cause the switch in ␣-TM and ␣-actinin splicing in dedifferentiating RASM cells. RT-PCR was not carried out under conditions that would allow quantitative analysis of absolute levels of expression. Nevertheless, we consistently observed that the levels of PTB products appeared to be lower in day 0 cells than in day 4 cells.
Further analysis of PTB expression using primers P37 and Pd3Ј2, which prime within exons 8 and 12, respectively, produced a strikingly different result. A novel band (labeled smPTB in Fig. 2, C and D) larger than PTB4 was the major PCR product in day 0 samples, but not in day 4 or PAC-1 samples. Cloning and sequencing of this PCR product showed that it was derived from a PTB-related gene that was distinct from PTB and the known paralogs ROD1 (40) and nPTB (38,39). We refer to this new PTB paralog as smPTB due to its high expression in a number of SM tissues (see below). RT-PCR using smPTB-specific primers P32 and P33 confirmed that it was expressed in day 0 RASM cells, but not in day 4 or cultured PAC-1 cells (Fig. 2D). smPTB was also expressed in other SM tissues such as the uterus and vas deferens.
smPTB Is Expressed from an Intronless Gene-At the time that we identified smPTB, no corresponding sequences could be identified by BLAST searches of available expressed sequence tag or genomic data bases. However, using 3Ј-and 5Ј-RACE, full-length smPTB cDNA was isolated from day 0 RASM cell RNA. The open reading frame encodes a 588-amino acid protein with a predicted molecular mass of 63.7 kDa and with 53-74% amino acid identity to PTB, nPTB, and ROD1 (Fig. 3). Pairwise BLAST analyses showed that smPTB is more closely related to PTB than to either of the other genes. Subsequent to cloning the full-length smPTB cDNA, the corresponding rat gene sequence was identified using the ENSEMBL Trace Database, whereas the mouse gene was identified by BLAST analysis of annotated mouse genomic data bases and was located in contig 132920, corresponding to chromosome X A1.1. Both the rat and mouse genes are intronless. The mouse gene contains three possible polyadenylation addition signals giving a message size of 4.16, 5.09, or 6.53 kb. As determined by Northern blot analysis, the size of the rat smPTB mRNA is close to 6 kb (data not shown). smPTB has the same overall structural organization as PTB (Fig. 3), with four RRM domains and the same unusual fifth FIG. 1. Alternative splicing of ␣-actinin (A), vinculin (B), and ␣-tropomyosin (C) in SM cells. RNA harvested from day 0 (0) and day 4 (4) primary RASM cells and from the PAC-1 SM cell line was analyzed by RT-PCR using a one end-labeled PCR primer. The n lanes are PCR-negative controls with no template added. The sizes of amplified bands (bp) are shown to the left of each panel, and bands marked by an asterisk are SM-specific splicing products. For ␣-TM, the two spliced products, which are the same size, were differentiated by digestion with XhoI (X lanes) or PvuII (P lanes), which are specific for exon 2-and exon 3-containing products, respectively. The U lanes are undigested PCR products, whereas the XP lanes are double digests.

FIG. 2. Novel PTB in differentiated SM cells.
A, alternative splicing of PTB exons 9 and 11. Inclusion of exon 9 produces PTB4, whereas skipping leads to PTB1. Skipping of exon 11 produces mRNAs that encode truncated isoforms PTB1tr and PTB4tr. The positions of primers P37, P38, and Pd3Ј2 used for PCR in B and C are indicated. B, RT-PCR of RNA harvested from RASM cells at day 0 (0) or day 4 (4) or from PAC-1 cells. PCR primers P37 and P38 correspond to PTB exons 8 and 11, respectively. C, same as B, but using primers P37 and Pd3Ј2, corresponding to PTB exons 8 and 12, respectively, which also detect the novel smPTB paralog. D, same as B and C, but using primers specific for the smPTB paralog. Tissues from the uterus and vas deferens were also tested with smPTB-specific primers. Size markers are either HaeIII ØX174 (M lanes; sizes are indicated in bp) or PCR products from a mixture of PTB1, PTB2, and PTB4 plasmids.
␤-strand within RRM3 (49). The degree of identity to PTB within RRM1-4 is 78, 89, 71, and 75%, respectively. Within the characteristic RNP1 and RNP2 motifs, most amino acid changes are conservative (Fig. 3). PTB and nPTB have bipartite nuclear localization signals (37), the N-terminal half of which appears to be lacking in both smPTB and ROD1. smPTB contains sequences equivalent to PTB exon 9, which defines the PTB4 isoform. The larger size of smPTB is accounted for by additional inserts of 36 amino acids between RRM1 and RRM2 and 22 amino acids between RRM2 and RRM3. These two linkers are the most divergent regions both between smPTB and PTB, and also between rat and mouse smPTBs. Strikingly, the linker region between RRM3 and RRM4 is identical in smPTB and the other paralogs. This region was not observed in the NMR structure of PTB RRM3 and RRM4, suggesting that it is extremely flexible (49). The absolute conservation suggests that the linker serves an important role and perhaps takes up a defined structure upon RNA binding. PTB has three sites that can be cleaved by caspase-3 (50). The most efficiently cleaved site (LKTD 138 S) is conserved in nPTB and ROD1, but not in smPTB. However, the next major site (AAVD 170 A in PTB) is reasonably conserved in smPTB (SAVD 178 T), suggesting that it may also be a target of caspase-3 during apoptosis.
Expression of smPTB-Expression of smPTB relative to PTB was monitored across a range of rat tissues using PCR primer pair P37/Pd3Ј2, corresponding to PTB exons 8 and 12, respectively. These primers also allow detection of alternatively spliced PTB isoforms in both exons 9 and 11, respectively. We have found exon 11 to be skipped in a small proportion of PTB mRNA, giving rise to the PTB1tr (where "tr" is truncated) and PTB4tr isoforms (Fig. 4). 2 3. Amino acid alignment of rat  PTB, smPTB, ROD1, and nPTB. The amino acid sequences of the four rat paralogs were aligned using Clustal, followed by some manual adjustment. Positions identical between the four genes are shown by asterisks, whereas conservative and semiconservative alterations in the RNP2 and RNP1 boxes are marked by colons and periods, respectively. The bipartite N-terminal nuclear localization signal of PTB is underlined. RRM domains are shown in blue, with the RNP2 and RNP1 boxes in red. RRMs are as defined by PROSITE, with the exception that RRM3 is extended to include the fifth ␤-strand (underlined LTKD(Y/F)) determined by NMR (49) .   FIG. 4. Comparison of PTB and smPTB expression in rat tissues. RNA from rat tissues was analyzed by RT-PCR using primers P37 and Pd3Ј2 (Fig. 2, A and C), which detect both PTB and smPTB. The no DNA lane is a PCR-negative control, whereas the PTB lane is a positive control containing PTB plasmid markers. The identities of the bands are shown to the right. nently in two SM tissues, the aorta (day 0) and uterus, where it was more abundant than PTB. It was also readily detected in the testis, thymus, skin, and lung (Fig. 4), although in these tissues, PTB was more abundant. It was not highly expressed in the stomach or small or large intestine, all of which contain SM cells. Longer exposures showed that smPTB expression could be detected in most tissues.
We next investigated smPTB expression by in situ hybridization, which allows direct analysis of expression without PCR amplification. Expression of smPTB was analyzed by in situ hybridization to day 10, 14, and 15 mouse embryo sections. This allowed us to conduct an unbiased survey of smPTB expression across multiple tissues at a stage of embryonic development when many SM cells express late markers of differentiation (e.g. SM myosin heavy chain). Whereas the sense smPTB control probe produced no signal, the antisense probes hybridized in a number of places, indicated by the light areas in the dark-field images (Fig. 5, lower panels). The most prominent signal was seen in the terminal bronchioles of the lung. The signal was readily detected at embryonic day 14.5 and was higher by embryonic day 15.5. Hybridization was also seen in the skin, intercostal muscles, and the venous plexus of the liver. In day 10 embryo sections, hybridization to the maternal uterus could be observed (Fig. 5, left panels). The lung, skin, and uterus are all tissues that showed relatively high levels of smPTB expression by RT-PCR (Fig. 4). The embryo sections did not include the aorta, which was one of the tissues with the highest smPTB signal upon RT-PCR (Fig. 4). Nevertheless, the in situ hybridization data were generally in agreement with the RT-PCR data and indicated that smPTB is differentially expressed in mouse embryos.
smPTB Is a Nuclear Protein That Binds RNA-To start to address the possible functions of smPTB, PAC-1 cells were transiently transfected with expression vectors for smPTB fused to GFP at either the C or N terminus (Fig. 6). Like PTB-GFP (12), GFP-smPTB localized almost completely to the nucleus, despite lacking sequences equivalent to the N-terminal half of the PTB bipartite nuclear localization signal (see above). Identical results were obtained with both N-and Cterminal GFP fusions and in HeLa cells. In a small number of PAC-1 cells, fluorescence was observed in the cytoplasm as well as the nucleus. These results demonstrate that smPTB is predominantly localized to the nucleus, consistent with a role in regulation of splicing, but that, like PTB, it may also be able to play additional cytoplasmic roles.
To examine the activities of smPTB in vitro, we overexpressed C-terminally His-tagged smPTB in E. coli. Recombinant smPTB migrated upon SDS-PAGE with an anomalously high mobility of 83 kDa compared with the expected size of 65 kDa. A similarly sized product was obtained by in vitro translation of smPTB in reticulocyte lysate (data not shown). Recombinant smPTB was analyzed for RNA binding by both UV cross-linking and electrophoretic mobility shift assays. RNA probes containing various regulatory elements from ␣-tropomyosin were used in UV cross-linking assays with recombinant smPTB and PTB4 (Fig. 7, A and B). Both PTB and smPTB cross-linked to RNA probes that contained the PTB-binding elements P3 and DY (probes 1-4), but not to probe 5, which contains a UGC motif regulatory element. Whereas PTB crosslinked with roughly equal efficiency to probes 1-4, smPTB cross-linked more efficiently to probes 1 and 4, which contain the DY regulatory sequence. These data confirm that smPTB is an RNA-binding protein. To compare the affinity with which smPTB and PTB bind to RNA, we carried out a competitive UV cross-linking assay. PTB4 and smPTB were premixed before incubation together with ␣-TM RNA probe 1. While one protein was held at a constant concentration of 0.5 M, the concentration of the other protein was varied. smPTB was readily able to displace PTB (Fig. 7C, lanes 1-6). In contrast, titration of PTB4 led to a more gradual increase in its own cross-linking signal, and there was little evidence of smPTB displacement (lanes 7-13). Similar results were seen with RNA probes 3 and 4 (data not shown). These data suggest that smPTB binds to the RNA probes tested with higher affinity compared with PTB.
smPTB Is a Splicing Repressor-We tested the activity of recombinant smPTB in a number of in vitro splicing assays, but were able to observe only nonspecific inhibitory activity (data not shown). At this stage, we do not know whether this is due to the recombinant protein lacking full activity or to the lack of an essential cofactor. We also tested the activity of smPTB as a splicing regulator by cotransfection with tropomyosin and actinin splicing reporter constructs. It had relatively modest effects on splicing of ␣-TM constructs in PAC-1 SM and other cell types (data not shown). This could be because there is already abundant PTB in these cells, and smPTB does not have a significantly different activity on this substrate. We also tested the effects of smPTB cotransfection with the pA ␣-actinin splicing reporter into HeLa cells. This reporter predominantly spliced to include the NM exon (Fig. 8). In control experiments, the pA reporter has been shown to be unresponsive to other overexpressed proteins, including ␤-galactosidase, hnRNP-C, hnRNP-L, and PTBtr1. 2 Consistent with previous results (29), overexpression of PTB led to skipping of both mutually exclusive exons (Fig. 8, lanes 5 and 6). Overexpression of smPTB had a similar effect, leading to enhanced skipping of both the NM and SM exons (lanes 3 and 4), although it was not as potent compared with PTB. This result establishes that smPTB has the ability to act as a splicing repressor, with activity similar (but not identical) to that of PTB. DISCUSSION The data reported here confirm the existence and tissuespecific expression of mRNA for a fourth PTB paralog, smPTB, in rat and mouse. We have also observed additional crossreactive bands in Western blots using anti-PTB antiserum and protein samples from day 0 RASM cells (data not shown). In the future, we aim to further analyze and verify smPTB expression using specific antisera raised against the recombinant protein. We have demonstrated that recombinant smPTB has various properties in common with PTB, including predominant nuclear localization (Fig. 6), RNA binding (Fig. 7), and splicing repression (Fig. 8). At present, it is not clear which alternative splicing events might be specific targets of endogenous smPTB. The correlation between smPTB mRNA expression and regulation of ␣-TM and ␣-actinin alternative splicing in primary RASM cells and its high expression in some SM tissues (Figs. 1, 2, and 4) initially suggested that it may play a key role in regulating these two splicing events. However, some SM tissues such as the intestine did not express smPTB, and cotransfection of smPTB with either ␣-TM or ␣-actinin reporters did not cause a significant switch toward the characteristic SM-specific splicing pattern of each gene. One possibility is that smPTB plays no role in these regulated splicing decisions. Other possibilities are that smPTB may simply replace the activity of PTB in these systems, that it may require a specific cofactor for activity in these systems, or that it may not be active in the presence of PTB. Currently, we cannot distinguish between these possibilities, although the fact that it binds RNA in vitro with higher affinity than PTB argues against the third possibility. In the future, it will be of interest to test the activity of smPTB using in vitro splicing and translation assays and also in cells after knockdown of endogenous PTB by RNA interference (31).  4) and PTB4 (lanes 5 and 6) expression plasmids. RNA was analyzed by RT-PCR using a labeled PCR primer. Lane 1, no template (n); lane 2, pA with no cotransfection; lane 7, size markers, with sizes (bp) shown to the right.