Exon Recognition in Vertebrate Splicing

INTRODUCTIONThe ProblemThe average vertebrate gene consists of multiple small exons (average size, 137 nucleotides) separated by introns that are considerably larger(1Hawkins J.D. Nucleic Acids Res. 1988; 16: 9893-9908Crossref PubMed Scopus (441) Google Scholar). Thus, the vertebrate splicing machinery has the task of finding small desired exons amid much longer introns. The splice site consensus sequences that drive exon recognition are located at the very termini of introns(2Green M.R. Annu. Rev. Cell Biol. 1991; 7: 559-599Crossref PubMed Scopus (553) Google Scholar, 3Moore M.J. Query C.C. Sharp P.A. Gestland R. Atkins R. The RNA World. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY1993: 303-357Google Scholar). Despite the discriminatory challenge faced during exon recognition in large multiexon premessenger RNAs, vertebrate splice sites are short and poorly conserved. In fact, splice site sequences in mammals are less conserved than their yeast counterparts despite the fact that only a minority of genes in Saccharomyces cerevisiae have introns; and those genes that are split by introns usually have only a single intron(4Guthrie C. Science. 1991; 253: 157-163Crossref PubMed Scopus (314) Google Scholar, 5Ruby S. Abelson J. Trends Genet. 1991; 7: 79-85Abstract Full Text PDF PubMed Scopus (184) Google Scholar). Thus, vertebrate splicing contends with a more complex specificity problem via recognition of less precise consensus sequences. Any mechanism for the orchestration of splicing in multiexon vertebrate genes must provide an explanation for this puzzle.Part of the solution of the puzzle comes from the observation that individual splice sites are not independently recognized consensus sequences. In both yeast and vertebrate splicing, interactions between 5′ and 3′ splice sites and the factors that recognize them have been observed during the earliest steps of spliceosome assembly(4Guthrie C. Science. 1991; 253: 157-163Crossref PubMed Scopus (314) Google Scholar, 5Ruby S. Abelson J. Trends Genet. 1991; 7: 79-85Abstract Full Text PDF PubMed Scopus (184) Google Scholar, 6Jamison S.F. Crow A. Garcia-Blanco M.A. Mol. Cell. Biol. 1992; 12: 4279-4287Crossref PubMed Scopus (80) Google Scholar, 7Lammond A.L. Konarska M.M. Sharp P.A. Genes 1: 532-543Crossref PubMed Scopus (106) Google Scholar, 8Kuo H.-C. Nasim F.H. Grabowski P.J. Science. 1991; 251: 1045-1050Crossref PubMed Scopus (130) Google Scholar, 9Michaud S. Reed R. Genes 7: 1008-1020Crossref PubMed Scopus (140) Google Scholar, 10Robberson B.L. Cote G.J. Berget S.M. Mol. Cell. Biol. 1990; 10: 84-94Crossref PubMed Scopus (547) Google Scholar, 11Rosbash M. Seraphin B. Science. 1991; 16: 187-190Scopus (115) Google Scholar, 12Wu J.Y. Maniatis T. Cell. 1993; 75: 1061-1070Abstract Full Text PDF PubMed Scopus (618) Google Scholar). Usually these interactions are depicted as occurring between the 5′ and 3′ splice sites across an intron. Experimentally, such interactions have been observed with in vitro splicing precursor RNAs having naturally short or artificially shortened introns. It is difficult to extrapolate initial interactions between the factors that recognize the 5′ and 3′ splice sites flanking a small vertebrate intron to introns that can naturally be 100 kilobases in length, especially given the likelihood that such introns will contain sequences that are as good a match to consensus splice sites as the actual utilized sites.Exon DefinitionModels that invoke pairing between the splice sites across an exon, as contrasted with pairing across an intron, are useful perspectives of splice site pairing for the splicing of pre-mRNAs with large introns and small exons. Such an exonic perspective of splice site recognition has been termed “exon definition”(10Robberson B.L. Cote G.J. Berget S.M. Mol. Cell. Biol. 1990; 10: 84-94Crossref PubMed Scopus (547) Google Scholar). This review discusses exon definition and contrasts it with intron-oriented perspectives that are more useful when considering splicing in lower eukaryotes with small introns. The basic exon definition model proposes that in pre-mRNAs with large introns, the splicing machinery searches for a pair of closely spaced splice sites in an exonic polarity (Fig. 1). When such a pair is encountered, the exon is defined by the binding of U1 and U2 snRNPs ( 1The abbreviations used are: snRNPsmall nuclear ribonucleoproteinSRarginine- and serine-rich splicing factorshnRNPheterogeneous nuclear ribonucleoprotein. )and associated splicing factors, including the 3′ splice site recognizing factors U2AF and SC35 and the 5′ splice site-recognizing factor ASF/SF2(2Green M.R. Annu. Rev. Cell Biol. 1991; 7: 559-599Crossref PubMed Scopus (553) Google Scholar, 13Eperon I.C. Ireland D.C. Smith R.A. Mayeda A. Krainer A.R. EMBO J. 1993; 12: 3607-3617Crossref PubMed Scopus (169) Google Scholar, 14Fu X.-D. Maniatis T. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 11224-11228Crossref PubMed Scopus (198) Google Scholar, 15Kohtz J.D. Jamison S.F. Will C.L. Zuo P. Luhrmann R. Garcio-Blanco M.A. Manley J.L. Nature. 1994; 368: 119-124Crossref PubMed Scopus (527) Google Scholar, 16Zuo P. Manley J.L. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 3363-3367Crossref PubMed Scopus (90) Google Scholar). Following definition of the exon, neighboring exons must be juxtaposed, presumably via interactions between the factors that recognize individual exons. Thus, from this perspective, assembly of the active vertebrate spliceosome consists of the sequential steps of exon definition and exon juxtaposition.Predictions of Exon DefinitionThe exon definition model offers predictions of pre-mRNA behavior. Several of these predictions have been tested in the last several years, and the results lend credence to an exonic perspective of splice site recognition.Exon SkippingExon-oriented and intron-oriented perspectives of splice site pairing predict different phenotypes resulting from mutation of splice sites bordering an internal exon (Fig. 2). Models invoking an initial pairing of splice sites across introns predict that such mutations should inhibit splicing of the intron in which they occur but should have minimal impact on the splicing of neighboring introns. In contrast, exon definition predicts that mutation of a splice site bordering an internal exon should depress recognition of the exon with concomitant inhibition of splicing of the adjoining intron, i.e. mutations in an intron will inhibit the splicing of two introns, the intron containing the mutation and the intron on the other side of the exon bearing the mutation. This hypothesis has been tested in vitro, where it was observed that mutation of a 5′ splice site depressed the removal of the upstream intron 20-fold(17Talerico M. Berget S.M. Mol. Cell. Biol. 1990; 10: 6299-6305Crossref PubMed Scopus (158) Google Scholar). The converse experiments have also been reported. Strengthening a naturally weak 5′ splice site of an internal exon by making it a better fit to the consensus site increased in vitro splicing of the upstream intron(8Kuo H.-C. Nasim F.H. Grabowski P.J. Science. 1991; 251: 1045-1050Crossref PubMed Scopus (130) Google Scholar, 18Grabowski P.J. Nasim F.H. Kuo H.-C. Burch R. Mol. Cell. Biol. 1991; 11: 5919-5928Crossref PubMed Scopus (36) Google Scholar). In vivo, mutant 5′ splice sites were genetically suppressed by second mutations that improved the 3′ splice site across the exon(19Carothers A.M. Urlaub G. Grunberger D. Chasin L. Mol. Cell. Biol. 1993; 13: 5085-5098Crossref PubMed Scopus (67) Google Scholar, 20Tsukahara T. Casciato C. Helfman D.M. Nucleic Acids Res. 1994; 22: 2318-2325Crossref PubMed Scopus (35) Google Scholar).Figure 2Predictions of the phenotype of mutation of the 5′ splice site bordering an internal exon. Exon pairing of splice sites predicts exon skipping or the activation of a proximal cryptic 5′ splice site (left), whereas intronic pairing of splice sites predicts intron inclusion or distal cryptic site activation (right).View Large Image Figure ViewerDownload Hi-res image Download (PPT)Mutation of vertebrate splice sites also leads to exon skipping. A survey of mammalian mutations available in the data base in the summer of 1994 indicated that over 100 splice site mutations have been characterized in disease gene DNA(21Nakai K. Sakamoto H. Gene (Amst.). 1994; 141: 171-177Crossref PubMed Scopus (255) Google Scholar). Four phenotypes were observed: exon skipping, activation of a cryptic splice site, creation of a pseudo-exon within an intron, and intron retention, in ratios of 51, 32, 11, and 6%, respectively. The most frequent phenotype was exon skipping. Exon skipping is a predicted phenotype from an exon perspective because mutation of the splice site at one side of an exon should inhibit pairing of splice sites across exons and inhibit recognition of the exon. Rejection of the exon leads directly to exon skipping.The observation of exon skipping strongly indicates that splice sites are recognized as exonic pairs. It is presumably this dependence upon a pair of sites that minimizes recognition of isolated cryptic sites within large vertebrate introns. Occasionally, mutation of human genes has created a strong splice site deep within an intron. Such created sites have been observed to be utilized via the activation of a nearby cryptic splice site of the opposite polarity to create a pseudo-exon from within an intron. Again, the observation is that only pairs of splice sites can be recognized and that cryptic splices in introns can only be activated by creation of a nearby site of the opposite type in an exonic polarity.Occasionally, mutation of an internal splice site results in intron retention. Exon definition would not predict intron retention, except perhaps for very small introns. Of the splice site mutations mentioned above, only 6% caused intron retention. Four of the included introns were very short, and three were terminal introns, suggesting abrogation of exon definition modes of recognition when introns are very small or at the ends of pre-mRNAs (see below). Three examples involved large internal introns and cannot be explained by current exon perspectives.Exon Size MaximumIn addition to exon skipping, the other major phenotype resulting from mutation of a splice site in a human gene is activation of a cryptic site of the same type. The activated cryptic site always lies close to the mutated site, suggesting that splice sites are acceptable only if they reside close to a site of the opposite polarity and that, therefore, internal vertebrate exons may have a size maximum imposed in part by the splicing machinery. Fig. 3 indicates the size distribution of 1600 primate internal exons. Of these exons, only 3.5% are longer than 300 nucleotides and less than 1% are longer than 400 nucleotides, indicating that large internal exons are rare. In vitro, the assembly of ATP-dependent spliceosomes is inhibited if internal exons with strong constitutive splice sites are internally expanded to greater than 300 nucleotides(10Robberson B.L. Cote G.J. Berget S.M. Mol. Cell. Biol. 1990; 10: 84-94Crossref PubMed Scopus (547) Google Scholar). In vivo, expansion of internal exons residing in vertebrate genes with moderate to large introns has two phenotypes: activation of internal cryptic splice sites within the expanded exon to create small exons or skipping of the entire exon (see below). ( 2D. A. Sterner, T. Carlo, and S. M. Berget, unpublished data. )These phenotypes are consistent with splicing-imposed restriction on exon length. Presumably, such a size limitation helps explain why cryptic splice sites located inside of long vertebrate introns are not occasionally misrecognized to create large internal exons when the normal sites are mutated. A few spectacularly long vertebrate internal exons exit; the mechanism whereby such exons bypass restrictions on exon length is unknown.Figure 3Internal exon size distribution. Length distribution of 1600 primate internal exons from a library normalized to represent highly related exons only a single time (top) (library kindly provided by D. Searles, University of Pennsylvania) or 194 alternative vertebrate cassette exons (bottom) compiled by Stamm et al.(46Tian M. Maniatis T. Genes 8: 1703-1712Crossref PubMed Scopus (132) Google Scholar) or by S. Smith and T. A. Cooper (Baylor College of Medicine).View Large Image Figure ViewerDownload Hi-res image Download (PPT)Exon Size MinimumSimultaneous recognition of splice sites bordering an exon also suggests that a minimal separation between the sites might be required to prevent steric hindrance between the factors that recognize individual sites. When a constitutively recognized internal exon was internally deleted below 50 nucleotides it was skipped by the in vivo splicing machinery(23Dominski Z. Kole R. Mol. Cell. Biol. 1991; 11: 6075-6083Crossref PubMed Scopus (178) Google Scholar). Increasing the strength of the splice sites alleviated problems in recognition, suggesting that exon size and splice site strength are additive factors in exon recognition(24Dominski Z. Kole R. Mol. Cell. Biol. 1992; 12: 2108-2114Crossref PubMed Scopus (71) Google Scholar). Some very small natural internal exons exist. Six and seven nucleotide exons are frequently found in muscle protein genes; N-CAM has a three-nucleotide exon. Although few very small exons have been studied, those that have suggest that very small exons require special enhancing sequences in addition to strong splice sites for inclusion(25Black D.L. Genes 5: 389-402Crossref PubMed Scopus (114) Google Scholar, 26Black D.L. Cell. 1992; 69: 795-807Abstract Full Text PDF PubMed Scopus (147) Google Scholar, 27Sterner D.A. Berget S.M. Mol. Cell. Biol. 1993; 13: 2677-2687Crossref PubMed Scopus (68) Google Scholar). Deletion of these elements causes exon skipping when the exon is small but not when it has been internally expanded to a more normal length. The small exon enhancers are located within the neighboring introns outside of the normal splice sites. It seems likely that such enhancers function as binding sites for splicing factors that artificially extend the exon domain during exon recognition.Terminal ExonsExon definition suggests that terminal exons, both first and last exons, will require special mechanisms for their recognition. First exons end with a 5′ splice site but have no processing signal at their beginning. They do, however, bear a modification at their beginning via the 7-methylguanosine cap attached to all polymerase II transcripts. The cap and nuclear proteins that bind the cap are essential for in vitro splicing of simple one-intron pre-mRNAs(28Izaurralde E. Lewis J. McGuigan C. Jankowska M. Darzynkiewicz E. Mattaj I.W. Cell. 1994; 78: 657-668Abstract Full Text PDF PubMed Scopus (425) Google Scholar). In two-intron pre-mRNAs, changing the guanosine cap to an adenosine cap depressed removal of the first intron in vitro but had only minimal impact on the second intron (29Ohno M. Sakamoto H. Shimura Y. Proc. Natl. Acad. Sci. U. S. A. 1987; 84: 5187-5191Crossref PubMed Scopus (97) Google Scholar). These results indicate both that pre-mRNAs are recognized segmentally in vitro and that the cap is essential for recognition and removal of the first intron. Or as stated from an exon perspective, first exons can be recognized via interactions between the factors that recognize caps and 5′ splice sites.Last exons begin with a 3′ splice site and terminate with a poly(A) site(30Manley J.L. Proudfoot N.J. Genes 8: 259-264Crossref Scopus (35) Google Scholar). They are often the largest exon in a vertebrate gene, with an average size of approximately 600 nucleotides(1Hawkins J.D. Nucleic Acids Res. 1988; 16: 9893-9908Crossref PubMed Scopus (441) Google Scholar, 31Brunak S. Engelbrecht J. Knudsen S. J. Mol. Biol. 1991; 220: 49-65Crossref PubMed Scopus (623) Google Scholar). Exon recognition predicts that factors recognizing 3′ splice sites interact with factors recognizing poly(A) sites to recognize last exons. Indeed mutation of 3′ splice sites inhibits the in vitro polyadenylation cleavage reaction(32Niwa M. Rose S.R. Berget S.M. Genes 4: 1552-1559Crossref PubMed Scopus (228) Google Scholar). Just as with first exons, mutation of the signal at the distal end of a 3-terminal exon, the poly(A) site, inhibits in vitro removal of proximal but not distal introns(33Niwa M. Berget S.M. Genes 5: 2086-2095Crossref PubMed Scopus (149) Google Scholar). These results suggest that splicing and polyadenylation factors interact across 3′-terminal exons. The mechanism of this interaction is unclear, although recent observations have suggested that U1 snRNPs or the U1 snRNP A protein are involved, either positively or negatively, via recognition of exon internal sequences upstream of the polyadenylation signal AAUAAA(34Boelens W.C. Jansen E.J. Ven Venrooij W.J. Stripeke R. Mattaj I.W. Gunderson S.I. Cell. 1993; 72: 881-892Abstract Full Text PDF PubMed Scopus (176) Google Scholar, 35Lutz C.S. Alwine J.C. Genes 8: 576-586Crossref PubMed Scopus (126) Google Scholar, 36Wasserman K.M. Steitz J.A. Genes 7: 647-659Crossref PubMed Scopus (91) Google Scholar).Exon Enhancer Sequences and Differential SplicingExon definition has proven to be a useful framework for considering differential splicing, especially those differential splicing events involving cassette exons that are differentially included. Generally, differentially recognized exons have either weaker splicing signals or a suboptimal length compared with constitutive exons (3Moore M.J. Query C.C. Sharp P.A. Gestland R. Atkins R. The RNA World. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY1993: 303-357Google Scholar, 37Stamm S. Zhang M.Q. Marr T.G. Helfman D.M. Nucleic Acids Res. 1994; 22: 1515-1526Crossref PubMed Scopus (87) Google Scholar) (Fig. 3), suggesting that the constitutive exon definition process is so strong as to be difficult to regulate unless the involved exon recognition signals are weak. Exon inclusion in these cases appears to be via recognition of special sequences by tissue or development-specific splicing factors(38Cooper T.A. Ordahl C.P. Nucleic Acids Res. 1989; 17: 6999-7011Crossref PubMed Scopus (59) Google Scholar, 39Dirksen W.P. Hampson R.K. Sun Q. Rottman F.M. J. Biol. Chem. 1994; 269: 6431-6436Abstract Full Text PDF PubMed Google Scholar, 40Hedley M.L. Maniatis T. Cell. 1991; 65: 579-586Abstract Full Text PDF PubMed Scopus (175) Google Scholar, 41Inoue K. Hoshijima K. Higuchi I. Sakamoto H. Shimura Y. Proc. Natl. Acad. Sci. U. S. A. 1992; 88: 683-687Google Scholar, 42Lavigueur A. La Branche H. Kornblihtt A.R. Chabot B. Genes 7: 2405-2417Crossref PubMed Scopus (272) Google Scholar, 43Mardon H.J. Sebastio G. Baralle F.E. Nucleic Acids Res. 1987; 15: 7725-7733Crossref PubMed Scopus (89) Google Scholar, 44Ryner L.C. Baker B.S. Genes 5: 2071-2085Crossref PubMed Scopus (124) Google Scholar, 45Streuli M. Saito H. EMBO J. 1989; 8: 787-796Crossref PubMed Scopus (88) Google Scholar, 46Tian M. Maniatis T. Genes 8: 1703-1712Crossref PubMed Scopus (132) Google Scholar, 47Watakabe A. Tanaka K. Shimura Y. Genes 7: 407-418Crossref PubMed Scopus (307) Google Scholar, 48Xu R. Teng T. Cooper T.A. Mol. Cell. Biol. 1993; 13: 3660-3674Crossref PubMed Scopus (178) Google Scholar, 49Yeakley J.M. Hedjran F. Morfin J.P. Merillat N. Rosenfield M.G. Emerson R.B. Mol. Cell. Biol. 1993; 13: 5999-6011Crossref PubMed Scopus (71) Google Scholar). One class of sequences commonly found associated with differential exons, referred to as exon enhancers, resides within the target exon. The existence of exon internal consensus sequences was initially surprising because of the constraints imposed upon such sequences by coding requirements. A family of such sequences, often purine-rich and coding for a wide variety of amino acids, has been observed to be important for recognition of weak exons. These sequences appear to be the binding site for a family of splicing factors known as SR proteins because of the arginine and serine repeats that characterize them(50Zahler A.M. J.A. Genes PubMed Scopus Google Scholar). In addition to binding exon sequences via their RNA binding the SR proteins also via their SR with U2AF the 3′ splice site and U1 snRNPs to the 5′ splice site via in J.Y. Maniatis T. Cell. 1993; 75: 1061-1070Abstract Full Text PDF PubMed Scopus (618) Google Scholar, 15Kohtz J.D. Jamison S.F. Will C.L. Zuo P. Luhrmann R. Garcio-Blanco M.A. Manley J.L. Nature. 1994; 368: 119-124Crossref PubMed Scopus (527) Google Scholar, D. Reed R. Mol. Cell. Biol. 1994; PubMed Scopus Google Scholar). Such recognition the SR proteins for proteins involved in exon across exons has been in that of U2AF to the 3′ splice site of an isolated exon is by the strength of the 5′ splice site the Grabowski P.J. Genes PubMed Scopus Google SR proteins have not been found in S. those yeast splicing proteins that are in known function to vertebrate proteins containing SR SR in their yeast N. M. Genes 8: PubMed Scopus Google Scholar, J. M. Genes 7: PubMed Scopus Google Scholar). an exon definition this may not be surprising in that with small introns, such as S. may not exon definition and may not or all of the SR In to suggest that pre-mRNAs with small introns the intron, than the exon, as the initial of pairing between splice C. Science. 1991; 253: 157-163Crossref PubMed Scopus (314) Google Scholar, 5Ruby S. Abelson J. Trends Genet. 1991; 7: 79-85Abstract Full Text PDF PubMed Scopus (184) Google Scholar, 11Rosbash M. Seraphin B. Science. 1991; 16: 187-190Scopus (115) Google Scholar). In Saccharomyces pre-mRNAs have multiple small introns of less than 100 M.Q. Marr T.G. Nucleic Acids Res. 1994; 22: PubMed Scopus Google Scholar). In of the introns are less than 100 nucleotides and are often by large J.D. Nucleic Acids Res. 1988; 16: 9893-9908Crossref PubMed Scopus (441) Google Scholar, S.M. C. G. C. Nucleic Acids Res. 1992; PubMed Scopus Google Scholar). small introns in either inhibits splicing of the intron or cryptic sites within the expanded M. S.M. Mol. Cell. Biol. 1993; 13: PubMed Scopus Google Scholar, M. Berget S.M. Mol. Cell. Biol. 1994; PubMed Scopus Google Scholar). ( in genes with small exons, the exon leads to splicing, whereas in genes with small introns, the introns leads to These observations suggests that the pairing utilized is that the between two splice of splice sites in genes with small introns has a different phenotype than the same mutation in genes with large introns. In pre-mRNAs with small introns, mutation of an internal 5′ splice site not to exon skipping. the mutated intron is included in the and the splicing of neighboring introns is M. Berget S.M. Mol. Cell. Biol. 1994; PubMed Scopus Google Scholar). A in splicing signals between the two of introns has also been S.M. C. G. C. Nucleic Acids Res. 1992; PubMed Scopus Google Scholar, M. S.M. Mol. Cell. Biol. 1993; 13: PubMed Scopus Google Scholar). introns often the located between the and the 3′ splice site of vertebrate but not S. cerevisiae introns. small introns appear to have different signals and to be recognized than large pairing of splice sites across an exon may be to initial pairing across an intron. for the SR the vertebrate factors known to be required for splicing are found in yeast and are required as Several of also suggest that either the intron or exon can be the pairing during pre-mRNA recognition. mentioned expansion of an internal exon in a vertebrate gene can exon skipping. the same exons and their flanking splice however, are in a gene in which the introns flanking the expanded exon are the expanded exon is constitutively Chasin Mol. Cell. Biol. 1994; PubMed Scopus Google of the small introns the phenotype to exon skipping. These observations suggest that large exons are only a problem in genes with large introns, and more that the same splice sites can be recognized in either intronic or exonic polarity (Fig. complex in via exon definition that in lower eukaryotes via intron Large Image Figure ViewerDownload Hi-res image Download in also suggests multiple of pairing splice sites within the same Although genes fit two characterized as genes with small introns and large exons or as genes with small exons and large introns, are a of genes that have a suggesting that over part of their length the exon is the of recognition and over part of their length the intron is the of recognition. two such recognition mechanisms can within the same precursor RNA a in exon recognition or exon is one of the for exon exon definition suggests exons and their splice sites are initially recognized by the splicing it not a solution to the second in spliceosome assembly (Fig. 1). of exons across large vertebrate introns is a especially if exon skipping is to be is available as to such A likely interactions between the SR proteins to one exon with the SR proteins to an adjoining exon. In addition to the SR class of nuclear proteins found only in with large introns is the G. M.J. S. Annu. Rev. 1993; PubMed Scopus Google Scholar). one protein 5′ splice site recognition and is likely to have a major in differential splicing M.R. S. Y. Chabot B. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: PubMed Scopus Google Scholar, Stamm S. Helfman D.M. Krainer A.R. Science. 1994; PubMed Scopus Google Scholar). the SR the proteins contain both an recognition domain and a recognition the SR the available suggests that the proteins recognize intronic consensus sequences than exonic sequences. their to differentially recognize RNA sequences and their for intronic sequences, the proteins in both differential splicing and exon INTRODUCTIONThe ProblemThe average vertebrate gene consists of multiple small exons (average size, 137 nucleotides) separated by introns that are considerably larger(1Hawkins J.D. Nucleic Acids Res. 1988; 16: 9893-9908Crossref PubMed Scopus (441) Google Scholar). Thus, the vertebrate splicing machinery has the task of finding small desired exons amid much longer introns. The splice site consensus sequences that drive exon recognition are located at the very termini of introns(2Green M.R. Annu. Rev. Cell Biol. 1991; 7: 559-599Crossref PubMed Scopus (553) Google Scholar, 3Moore M.J. Query C.C. Sharp P.A. Gestland R. Atkins R. The RNA World. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY1993: 303-357Google Scholar). Despite the discriminatory challenge faced during exon recognition in large multiexon premessenger RNAs, vertebrate