An Intronic Downstream Enhancer Promotes 3′ Splice Site Usage of a Neural Cell-specific Exon

The human nonmuscle myosin heavy chain B gene contains a 30-nucleotide alternative exon, N30, that is included in the mRNA from neural cells but is skipped in all other cells. We have previously identified an intronic distal downstream enhancer (IDDE) region that is required for neural cell-specific inclusion of N30. In this study, we investigated the mechanism by which the IDDE promotes N30 exon usage. In vitro splicing analysis using neural cell nuclear extracts and two-exon pre-mRNA substrates, which consist of the N30 exon and either the upstream (E5) or downstream (E6) exon, demonstrates that the IDDE activates upstream E5-N30 splicing by facilitating early prespliceosome complex formation. The IDDE has no effect on N30-E6 splicing where the IDDE resides. Inspection of splice site consensus sequences shows that a polypyrimidine (Py) tract preceding N30 is suboptimal for U2AF binding. Optimizing the Py tract completely relieves the requirement for the IDDE in E5-N30 splicingin vitro. In transfected cells, the wild-type minigene transcripts, which consist of three exons, E5, N30, and E6, undergo neural cell-specific and IDDE-dependent alternative splicing of N30. Optimizing the Py tract in minigenes also completely relieves the requirement for the IDDE in N30 inclusion. Furthermore, overexpression of the truncated U2AF65, which contains the arginine and serine dipeptide-rich domain and linker domain, but lacks the RNA binding domain, selectively inhibits the IDDE-mediated N30 inclusion in mRNA from the wild-type minigene in a dominant negative fashion. These results support the hypothesis that the IDDE facilitates the recognition of the 3′ splice site preceding N30 by a network of protein-protein interactions implicated in the recruitment of U2AF to a suboptimal Py tract. The human nonmuscle myosin heavy chain B gene contains a 30-nucleotide alternative exon, N30, that is included in the mRNA from neural cells but is skipped in all other cells. We have previously identified an intronic distal downstream enhancer (IDDE) region that is required for neural cell-specific inclusion of N30. In this study, we investigated the mechanism by which the IDDE promotes N30 exon usage. In vitro splicing analysis using neural cell nuclear extracts and two-exon pre-mRNA substrates, which consist of the N30 exon and either the upstream (E5) or downstream (E6) exon, demonstrates that the IDDE activates upstream E5-N30 splicing by facilitating early prespliceosome complex formation. The IDDE has no effect on N30-E6 splicing where the IDDE resides. Inspection of splice site consensus sequences shows that a polypyrimidine (Py) tract preceding N30 is suboptimal for U2AF binding. Optimizing the Py tract completely relieves the requirement for the IDDE in E5-N30 splicingin vitro. In transfected cells, the wild-type minigene transcripts, which consist of three exons, E5, N30, and E6, undergo neural cell-specific and IDDE-dependent alternative splicing of N30. Optimizing the Py tract in minigenes also completely relieves the requirement for the IDDE in N30 inclusion. Furthermore, overexpression of the truncated U2AF65, which contains the arginine and serine dipeptide-rich domain and linker domain, but lacks the RNA binding domain, selectively inhibits the IDDE-mediated N30 inclusion in mRNA from the wild-type minigene in a dominant negative fashion. These results support the hypothesis that the IDDE facilitates the recognition of the 3′ splice site preceding N30 by a network of protein-protein interactions implicated in the recruitment of U2AF to a suboptimal Py tract. arginine and serine dipeptide-repeat nonmuscle myosin heavy chain nucleotide(s) intronic distal downstream enhancer phenylmethanesulfonyl fluoride dithiothreitol polypyrimidine reverse transcription-polymerase chain reaction downstream control sequences small nuclear RNA small nuclear ribonucleoprotein heterogeneous nuclear ribonucleoprotein kilobase pair Alternative splicing of pre-mRNA is a fundamental mechanism for regulating eukaryotic gene expression. In many cases, alternative RNA splicing contributes to developmentally regulated and cell type-specific patterns of gene expression. Although a great deal of information is available concerning the general constitutive splicing reactions of simple splicing units, the molecular basis for alternative splice site selection is not well understood (for review see Refs.1Sharp P.A. Cell. 1994; 77: 805-815Abstract Full Text PDF PubMed Scopus (447) Google Scholar, 2Staley J.P. Guthrie C. Cell. 1998; 92: 315-326Abstract Full Text Full Text PDF PubMed Scopus (909) Google Scholar, 3Lopez A.J. Annu. Rev. Genet. 1998; 32: 279-305Crossref PubMed Scopus (535) Google Scholar). Considerable insight into the regulation of alternative splicing has been gained from studies of genes identified inDrosophila. The most extensively characterized of these involve Drosophila sex determination, for which examples of both positive and negative regulation of splicing were found (for review see Refs. 3Lopez A.J. Annu. Rev. Genet. 1998; 32: 279-305Crossref PubMed Scopus (535) Google Scholar and 4MacDougall C. Harbison D. Bownes M. Dev. Biol. 1995; 172: 353-376Crossref PubMed Scopus (47) Google Scholar). In vertebrates, however, much less is known about the mechanisms and cellular factors involved in regulated alternative splicing. Studies of vertebrate genes aimed at understanding the regulation of alternative splicing have led to identification of a number of pre-mRNA features that influence alternative splice site selection. These include the sequence of the 5′ and 3′ splice sites, branch point sequence and location, exon size, intron size, and specific RNA sequences (enhancers or repressors) located in exons or introns. Many parameters can affect splicing in ways that are complex and not readily predictable. cis-Acting RNA sequences can enhance or repress utilization of alternative splice sites. Elements that promote splicing of an adjacent splice site are known collectively as splicing enhancers and are classified by their location as exonic or intronic enhancers. Exonic enhancers are often rich in purine residues (5Watakabe A. Tanaka K. Shimura Y. Genes Dev. 1993; 7: 407-418Crossref PubMed Scopus (307) Google Scholar, 6Lavigueur A. La Branche H. Kornblihtt A.R. Chabot B. Genes Dev. 1993; 7: 2405-2417Crossref PubMed Scopus (272) Google Scholar, 7Lynch K.W. Maniatis T. Genes Dev. 1995; 9: 284-2938Crossref PubMed Scopus (149) Google Scholar). These exonic splicing enhancers bind to members of the SR protein family, which contain a characteristic arginine and serine dipeptide repeat (RS)1 domain and RNA binding domains (for review see Refs. 8Fu X.-D. RNA (NY ). 1995; 1: 663-680PubMed Google Scholar and 9Manley J.L. Tacke R. Genes Dev. 1996; 10: 1569-1579Crossref PubMed Scopus (601) Google Scholar). For exonic enhancers, which can recruit SR proteins, there is evidence that the enhancer-protein complex functions by stimulating spliceosome assembly at the upstream 3′ splice site. A large number of exons have this enhancer element. Only a few of them, however, appear to be a target for cell type-specific regulation. In addition to exonic enhancers, a number of splicing events are controlled by intronic splicing enhancers (10Black D.L. Cell. 1992; 69: 795-807Abstract Full Text PDF PubMed Scopus (147) Google Scholar, 11Huh G.S. Hynes R.O. Genes Dev. 1994; 8: 1561-1574Crossref PubMed Scopus (134) Google Scholar, 12Sirand-Pugnet P. Durosay P. Brody E. Marie J. Nucleic Acids Res. 1995; 23: 3501-3507Crossref PubMed Scopus (106) Google Scholar, 13Carlo T. Sterner D.A. Berget S.M. RNA (NY ). 1996; 2: 342-353PubMed Google Scholar, 14Kawamoto S. J. Biol. Chem. 1996; 271: 17613-17616Abstract Full Text Full Text PDF PubMed Google Scholar, 15Del Gatto F. Plet A. Gesnel M.-C. Fort C. Breathnach R. Mol. Cell Biol. 1997; 17: 5106-5116Crossref PubMed Scopus (82) Google Scholar, 16Gallego M.E. Gattoni R. Stévenin J. Marie J. Expert-Bezancon A. EMBO J. 1997; 16: 1772-1784Crossref PubMed Scopus (106) Google Scholar, 17Hedjran F. Yeakley J.M. Huh G.S. Hynes R.O. Rosenfeld M.G. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 12343-12347Crossref PubMed Scopus (47) Google Scholar, 18Lim L.P. Sharp P.A. Mol. Cell. Biol. 1998; 18: 3900-3906Crossref PubMed Scopus (92) Google Scholar, 19Cooper T.A. Mol. Cell. Biol. 1998; 18: 4519-4525Crossref PubMed Scopus (43) Google Scholar). These sequences have been found in the introns downstream of small and/or cell type-specific exons, and their presence is required for the splicing of these exons. However, with a few exceptions (16Gallego M.E. Gattoni R. Stévenin J. Marie J. Expert-Bezancon A. EMBO J. 1997; 16: 1772-1784Crossref PubMed Scopus (106) Google Scholar, 20Min H. Chan R.C. Black D.L. Genes Dev. 1995; 9: 2659-2671Crossref PubMed Scopus (169) Google Scholar, 21Min H. Turck C.W. Nikolic J.M. Black D.L Genes Dev. 1997; 11: 1023-1036Crossref PubMed Scopus (279) Google Scholar, 22Chou M.-Y. Rooke N. Turk C.W. Black D.L. Mol. Cell. Biol. 1999; 19: 69-77Crossref PubMed Scopus (213) Google Scholar, 23Philips A.V. Timchenko L.T. Cooper T.A. Science. 1998; 280: 737-740Crossref PubMed Scopus (687) Google Scholar), the proteins that mediate enhancer effects remain unidentified. Moreover, how these enhancers affect spliceosome assembly is not known. We have been using the human nonmuscle myosin heavy chain (NMHC)-B gene as a model system to study the regulatory mechanisms responsible for neural cell-specific alternative splicing of pre-mRNA. TheNMHC-B gene encodes a polypeptide of approximately 200 kDa. A dimer of this gene product, together with two pairs of light chains (20 and 17 kDa), constitutes a myosin molecule that demonstrates force-generating ATPase activity when it interacts with actin. TheNMHC-B gene is expressed ubiquitously in most cell types (24Kawamoto S. Adelstein R.S. J. Cell Biol. 1991; 112: 915-924Crossref PubMed Scopus (177) Google Scholar); however, in neural cells, specific forms of NMHC-B are generated by cassette-type alternative splicing. NMHC-B has been shown to play an important role in neural cell migration and adhesion in NMHC-B-depleted mice (25Üren D. Hwang H.-K. Hara Y. Takeda K. Kawamoto S. Tullio A.N., Yu, Z.-X. Ferrans V.J. Tresser N. Grinberg A. Preston Y.A. Adelstein R.S. J. Clin. Invest. 2000; 105: 663-671Crossref PubMed Scopus (36) Google Scholar). Alternative splicing occurs at two different locations in the pre-mRNA (26Takahashi M. Kawamoto S. Adelstein R.S. J. Biol. Chem. 1992; 267: 17864-17871Abstract Full Text PDF PubMed Google Scholar). One alternative exon (N30), consisting of a 30-nt coding sequence, is located between the constitutive exons E5 and E6. The 10 amino acids encoded by the N30 exon is located near the ATP-binding region of the molecule and includes a serine residue that has been shown to be phosphorylated by mitogen-activated protein kinase and brain-specific cyclin-dependent protein kinase 5 (27Pato M.D. Sellers J.R. Preston Y.A. Harvey E.V. Adelstein R.S. J. Biol. Chem. 1996; 271: 2689-2695Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar), suggesting a specific role of this isoform in a signal transduction pathway. The other alternative exon, consisting of a 63-nt coding sequence, is located between constitutive exons E15 and E16, which is near the actin-binding region. Inclusion of these two alternative exons is restricted to neural cells in mammals and birds. The two alternative splicing events, however, appear to be regulated differentially by agonist stimulation in cultured neural cells as well as during brain development (28Itoh K. Adelstein R.S. J. Biol. Chem. 1995; 270: 14533-14540Abstract Full Text Full Text PDF PubMed Scopus (73) Google Scholar). We have focused on N30 regulation, since some culture systems maintain regulation of N30 splicing (14Kawamoto S. J. Biol. Chem. 1996; 271: 17613-17616Abstract Full Text Full Text PDF PubMed Google Scholar, 28Itoh K. Adelstein R.S. J. Biol. Chem. 1995; 270: 14533-14540Abstract Full Text Full Text PDF PubMed Scopus (73) Google Scholar). Previously, we demonstrated that a minigene encompassing E5-N30-E6 reproduced cell type-specific and differentiation state-dependent regulation of N30 inclusion, using a transient minigene transfection system (14Kawamoto S. J. Biol. Chem. 1996; 271: 17613-17616Abstract Full Text Full Text PDF PubMed Google Scholar). In addition, the previous study established that neural cell-specific N30 inclusion requires a cis-acting intronic enhancer sequence that is located 1.5 kb downstream of exon N30. This intronic enhancer region, 142 nucleotides (nts) in length, is now designated as an intronic distal downstream enhancer (IDDE) region (corresponding to fragment d in Ref. 14Kawamoto S. J. Biol. Chem. 1996; 271: 17613-17616Abstract Full Text Full Text PDF PubMed Google Scholar). In the present study, we attempted to find out how the IDDE promotes N30 exon usage during the splicing reaction usingin vitro splicing and minigene transfection systems. The experimental data shown here support the hypothesis that the IDDE facilitates recognition of the 3′ splice site preceding the N30 exon by a network of protein-protein interactions implicated in the recruitment of U2AF to a suboptimal polypyrimidine (Py) tract. For in vitro splicing substrates, defined portions of the genomic DNA (see Fig.1 C) were amplified and connected by recombinant PCR using the appropriate synthetic primers and were introduced into pBluescriptIISK(−) (Stratagene), which contains the T7 promoter. Unique restriction enzyme sites were engineered at the junctions between the different portions of the genomic DNA.Figure 1Schematic diagrams of the humanNMHC-B gene surrounding alternative exon N30 and the constructs used in this study. A, native gene.Rectangles and horizontal lines in the diagram indicate exons and introns, respectively. E5 and E6 are constitutive exons, and N30 and R18 are alternative exons. Inclusion of R18 in the mRNA is always associated with N30 inclusion. However, N30 inclusion occurs without R18 inclusion, and inclusion of N30 alone is the dominant form of neural mRNA. Inclusion of both N30 and R18 can be found to a small extent in a limited number of neural cell lines but not in neural tissues of animals. The IDDE is located between nts 1544 and 1685 downstream of N30 in the native gene. Exon size and the IDDE are not drawn to scale. B, minigene. Minigenes are constructed by inserting the human NMHC-B genomic DNA fragments indicated into the intron between exons E2 and E3 of the rat preproinsulin gene (PPI). Transcription of the minigene is driven by the Rous sarcoma virus long terminal repeat (RSVLTR). Arrows above E5 andPPIE3 indicate the location of the primers used for RT-PCR.C, in vitro splicing substrates. A blank space between the horizontal lines in E5-N30 and E5-E6 indicates a deletion.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Construction of a wild-type minigene (Fig. 1 B) has been described previously (corresponding to minigene C in Ref. 14Kawamoto S. J. Biol. Chem. 1996; 271: 17613-17616Abstract Full Text Full Text PDF PubMed Google Scholar). The constructs for in vitro splicing substrates and minigenes bearing the mutations m1 and m2 were generated by recombinant PCR. For U2AF65 expression, a plasmid pCS3+MT (29Roth M.B. Zahler A.M. Stolk J.A. J. Cell Biol. 1991; 115: 587-596Crossref PubMed Scopus (267) Google Scholar), which contains the cytomegalovirus promoter/enhancer, Myc epitope tags, and SV40 poly(A) signal, was used as a host vector. This vector was a gift from Dr. Yongsok Kim (NHLBI, National Institutes of Health). The nuclear localization signal of the SV40 large T antigen was introduced following the 5th copy of the Myc epitope. The full-length coding sequence for human U2AF65 was obtained from HeLa cell mRNA by the reverse transcription-polymerase chain reaction (RT-PCR) and introduced downstream of the nuclear localization signal. Truncated U2AF65 fragments were further amplified by PCR and introduced into the same vector. The expressed proteins contain five copies of the Myc epitope and a nuclear localization signal at the amino terminus. All of the constructs were verified by DNA sequencing. Nuclear extracts from Y79 cells were prepared according to protocols previously described (30Ausubel F.M. Brent R. Kingston R.E. Moore D.D. Seidman J.G. Smith J.A. Struhl K. Curr. Protocols Mol. Biol. 1990; 2: 12.1.1-12.1.9Google Scholar). Cells are washed in phosphate-buffered saline and pelleted by centrifugation for 10 min at 1850 × g. Packed cells are quickly resuspended in 5× the packed cell volume of the hypotonic buffer (10 mm HEPES, pH 7.9 at 4 °C, 10 mm KCl, 1.5 mm MgCl2, 0.4 mmphenylmethanesulfonyl fluoride (PMSF) and 0.5 mmdithiothreitol (DTT)) and are centrifuged for 5 min at 1850 ×g. The cells are resuspended in the hypotonic buffer to a final volume of 3× the original packed cell volume and allowed to swell for 10–15 min on ice. The swollen cells are homogenized with 10–15 strokes in a glass Dounce homogenizer to obtain >70% of cell lysis and the nuclei are centrifuged for 15 min at 3300 ×g. The nuclear pellets are resuspended in the low salt buffer (20 mm HEPES, pH 7.9, 20 mm KCl, 1.5 mm MgCl2, 0.2 mm EDTA, 25% glycerol, 0.4 mm PMSF, and 0.5 mm DTT) with a volume equal to 0.33–0.5× the packed nuclear volume. High salt buffer (0.33–0.5× the original packed nuclear volume), consisting of 20 mm HEPES, pH 7.9, 1.2 m KCl, 1.5 mmMgCl2, 0.2 mm EDTA, 25% glycerol, 0.4 mm PMSF, and 0.5 mm DTT, is added to the suspended nuclei, and then the entire suspension is homogenized with 3–5 strokes in the Dounce homogenizer. Following extraction for 30 min with stirring, the nuclear suspension is centrifuged for 30 min at 25,000 × g. The resulting supernatant is dialyzed against a buffer consisting of 20 mm HEPES, pH 7.9, 100 mm KCl, 0.2 mm EDTA, 20% glycerol, 0.4 mm PMSF, and 0.5 mm DTT for 2–4 h. The insoluble materials are removed by centrifugation. The aliquoted nuclear extracts are quickly frozen in liquid N2 and stored at −70 °C. All procedures are performed at 0–4 °C. Protein concentrations were determined by the Bradford method (Bio-Rad) using bovine serum albumin as a standard. Typically, protein concentrations of 7–10 mg/ml were obtained. Capped splicing RNA substrates, uniformly labeled with α-32PUTP (800 Ci/mmol, Amersham Pharmacia Biotech), were synthesized with T7 RNA polymerase (Stratagene) from the template plasmids linearized by the appropriate restriction enzymes. In vitro splicing assays were performed in 15-μl volumes containing 5 μl of nuclear extracts, 1 μl of a 15× concentrated splicing mix (4.5 mm MgCl2, 22.5 mm ATP, 75 mm creatine phosphate, and 75 mmdithiothreitol), 40 units of RNasin (Promega), and 2 ng of capped and labeled pre-mRNA substrates. Reaction mixtures were incubated at 30 °C for 0.5–4 h. The reaction was stopped by addition of proteinase K, followed by phenol/chloroform extraction and ethanol precipitation. The resulting RNAs were electrophoresed in denaturing urea-polyacrylamide (4.5, 6, and 10%) gels and autoradiographed. The debranching reaction was performed as described previously (31Ruskin B. Green M.R. Methods Enzymol. 1990; 181: 180-188Crossref PubMed Scopus (18) Google Scholar). For analysis of spliceosome complexes, the splicing reaction was performed under the same conditions as described above. The reaction was started by addition of the labeled pre-mRNA and incubated at 30 °C for the indicated times. The reaction was stopped by addition of heparin (final concentration 5 mg/ml) and placed on ice for 10 min prior to electrophoresis. for were by heparin to the reaction without the and then the were added on ice. The reaction was electrophoresed in a native with using a buffer containing mm and mm at for h. The gels were autoradiographed. The RNA were by reverse using The resulting were amplified by using primers and followed by the using primers and The were in DNA was performed using The minigenes were transfected into Y79 cells using and into HeLa cells by the method as described previously (14Kawamoto S. J. Biol. Chem. 1996; 271: 17613-17616Abstract Full Text Full Text PDF PubMed Google Scholar). For of a minigene with the U2AF65 indicated of the U2AF65 and the host vector and the wild-type minigene containing IDDE were into 2 × Y79 cells with μl of Y79 cells were on and in 2 mm containing for of RNA from transfected cells, analysis of from the minigenes by and of by using 5′ primers was performed as described previously (14Kawamoto S. J. Biol. Chem. 1996; 271: 17613-17616Abstract Full Text Full Text PDF PubMed Google Scholar). that required both protein and mRNA were in expressed U2AF65 proteins, the protein extracts were prepared by addition of buffer to the transfected cells. Protein were electrophoresed in denaturing gels and to A specific to the Myc epitope was used for with the In a previous study, we have defined an intronic enhancer region, which is required for neural cell-specific alternative splicing of the cassette-type exon, N30, in human NMHC-B pre-mRNA (see In the presence of the the mRNA from a minigene transfected into human neural Y79 cells can include the N30 exon, to the NMHC-B in the of the the minigene mRNA exon N30. Exon N30 is skipped in the from minigenes both with and without the IDDE in transfected cell as HeLa and cells (14Kawamoto S. J. Biol. Chem. 1996; 271: 17613-17616Abstract Full Text Full Text PDF PubMed Google (Fig. the IDDE neural cell on N30 inclusion. how the IDDE N30 recognition and promotes of the upstream and downstream introns, we established vitro splicing system for NMHC-B pre-mRNA using Y79 nuclear We Y79 cells as a of nuclear extracts since Y79 mRNA includes the N30 exon to a large to the and These Y79 cells have been used to the IDDE in a minigene transfection We the effects of the IDDE on in vitro splicing of pre-mRNA in the of three different two-exon substrates that consist of the N30 exon and either the upstream (E5) or downstream (E6) exon and E5-E6 A diagram of the pre-mRNA substrates is shown in 1 to the native gene and the minigenes used in the The nts of the intron between E5 and N30 was removed from the pre-mRNA for in vitro splicing in to obtain splicing since this is for the neural cell-specific N30 inclusion in the minigene The in vitro were incubated with Y79 nuclear extracts, and the resulting RNAs were by denaturing urea-polyacrylamide and the results are shown in The consisting of the N30 and E6 exons with and without the IDDE undergo splicing and in The N30 exon and the containing as well as the mRNA and the intron final are readily of the and final RNAs were from the size of the the of their their in different concentration and the effect of the with containing debranching activity not of of the from the pre-mRNA containing the IDDE are equal to from pre-mRNA the that there is no effect of the IDDE on the splicing of In pre-mRNA consisting of the E5 and N30 exons, in the of the is the the final are 2 in However, the IDDE downstream of N30 in the pre-mRNA of E5-N30 splicing. All and final are now readily 4 in the IDDE promotes of the intron between E5 and N30. We also pre-mRNA consisting of the two constitutive exons E5 and E6. In the of the this pre-mRNA is 10 in The 5′ splice sequence following E5 is readily used in the of the E5-E6 pre-mRNA. all these splicing reactions are out using two-exon without a exon, of the E5-N30 splicing in the of the IDDE is to the recognition of the 3′ splice site preceding the N30 the IDDE to promote the recognition of the 3′ splice signal of the N30 splicing is a reaction the has been well that assembly of the spliceosome a of complex 5 and protein A of are at different of spliceosome assembly in the A B C R. Curr. Genet. Dev. 1996; PubMed Scopus Google Scholar). at during the splicing reaction the IDDE we spliceosome assembly using the E5-N30 with and without the The splicing reaction at different was by native electrophoresis. form large of complex in which proteins bind to the is not a in the spliceosome assembly pathway. The prespliceosome complex in which and U2AF are to the 5′ and 3′ splice sites, be out from complex in this 15 the pre-mRNA containing the IDDE can form prespliceosome complex A in which is to the 3′ splice site The that this complex in an in a for complex A. of the however, of the A complex is at all in The B and C can be at this This be to the of splicing of NMHC-B pre-mRNA with used in vitro splicing substrates as pre-mRNA. The splicing of NMHC-B is with that of pre-mRNA not Although pre-mRNA the IDDE can recruit binding is a for has not been it appear that the IDDE facilitates spliceosome assembly at an early of complex A. The above data indicate that the 3′ splice site region preceding exon N30 is by spliceosome in the of the we focused on this 3′ splice region in For this the location of the branch site was determined by has been that some reverse at the of a template RNA are of this J. T. Nucleic Acids Res. 1997; PubMed Scopus Google Scholar). The RNA mixtures were and the resulting were amplified by PCR using the in A. This selectively the of the RNA to be but to the of the RNA (Fig. 4 A 4 A, The sequences of the PCR are shown in The shown are at the between the of the 5′ splice site and the branch site. In the reverse of at the branch point (Fig. 4 as