Portal de Periódicos da CAPES

Characterization of the Effects of Mutations in the Putative Branchpoint Sequence of Intron 4 on the Splicing within the Human Lecithin:cholesterol Acyltransferase Gene

2000; Elsevier BV; Volume: 275; Issue: 24 Linguagem: Inglês

10.1074/jbc.m910197199

1083-351X

Min Li, P. Haydn Pritchard,

Monoclonal and Polyclonal Antibodies Research

We have previously identified a point mutation (intervening sequence (IVS) 4: T → C) in the branchpoint consensus sequence of intron 4 of the lecithin:cholesterol acyltransferase (LCAT) gene in patients with fish-eye disease. To investigate the possible mechanisms responsible for the defective splicing, we made a series of mutations in the branchpoint sequence and expressed these mutants in HEK-293 cells followed by the analysis of pre-mRNA splicing using reverse transcriptase-polymerase chain reaction as well as LCAT activity assay. The results reveal that 1) the mutation of the branchpoint adenosine to any other nucleotide completely abolishes splicing; 2) the insertion of a normal branch site into the intronic sequence of the natural (IVS4-22c) or the branchpoint (IVS4-20t) mutant completely restores splicing; 3) the natural mutation can be partially rescued by making a single nucleotide change (G → A) within the branchpoint consensus sequence; and 4) other single base changes, particularly around the branchpoint adenosine residue, significantly decrease the efficiency of splicing and thus enzyme activity. Surprisingly, the nucleotide transversion at the last position of the branchpoint sequence (i.e. IVS4-25a or -25g) results in a 2.7-fold increase in splicing efficiency. Therefore, these observations clearly establish the functional significance of the branchpoint sequence of intron 4 for the splicing of the human LCAT mRNA precursors. We have previously identified a point mutation (intervening sequence (IVS) 4: T → C) in the branchpoint consensus sequence of intron 4 of the lecithin:cholesterol acyltransferase (LCAT) gene in patients with fish-eye disease. To investigate the possible mechanisms responsible for the defective splicing, we made a series of mutations in the branchpoint sequence and expressed these mutants in HEK-293 cells followed by the analysis of pre-mRNA splicing using reverse transcriptase-polymerase chain reaction as well as LCAT activity assay. The results reveal that 1) the mutation of the branchpoint adenosine to any other nucleotide completely abolishes splicing; 2) the insertion of a normal branch site into the intronic sequence of the natural (IVS4-22c) or the branchpoint (IVS4-20t) mutant completely restores splicing; 3) the natural mutation can be partially rescued by making a single nucleotide change (G → A) within the branchpoint consensus sequence; and 4) other single base changes, particularly around the branchpoint adenosine residue, significantly decrease the efficiency of splicing and thus enzyme activity. Surprisingly, the nucleotide transversion at the last position of the branchpoint sequence (i.e. IVS4-25a or -25g) results in a 2.7-fold increase in splicing efficiency. Therefore, these observations clearly establish the functional significance of the branchpoint sequence of intron 4 for the splicing of the human LCAT mRNA precursors. small nuclear ribonucleoprotein particle message RNA precursor intervening sequence lecithin:cholesterol acyltransferase fish-eye disease human embryonic kidney polymerase chain reaction reverse transcription base pair(s) branchpoint sequence The removal of introns from pre-mRNA involves the recognition of the 5′-splice site, the branchpoint sequence, and the 3′-splice site by five small nuclear ribonucleoprotein particles (snRNPs), i.e. U1, U2, U5, and U4/U6, and a large number of non-snRNP splicing factors (1.Moore M.J. Query C.C. Sharp P.A. Gesteland R.F. Atkins J.F. The RNA World. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY1993: 303-357Google Scholar, 2.Madhani H.D. Guthrie C. Ann. Rev. Genet. 1994; 28: 1-26Crossref PubMed Scopus (318) Google Scholar, 3.Reed R. Curr. Opin. Genet. Dev. 1996; 6: 215-220Crossref PubMed Scopus (249) Google Scholar, 4.Will C.L. Luhrmann R. Curr. Opin. Cell Biol. 1997; 9: 320-328Crossref PubMed Scopus (249) Google Scholar). The functional significance of the splice sites has been well established by the studies of a number of naturally occurring mutations and of site-directed mutagenesis at these exon/intron junctions (5.Green M.R. Ann. Rev. Genet. 1986; 20: 671-708Crossref PubMed Scopus (423) Google Scholar, 6.Padgett P.A. Grabowski P.J. Konarska M.M. Steiler S. Sharp P.A. Ann. Rev. Biochem. 1986; 55: 1119-1150Crossref PubMed Google Scholar, 7.Krawczak M. Reiss J. Cooper D.N. Hum. Genet. 1992; 90: 41-54Crossref PubMed Scopus (1123) Google Scholar, 8.Kodolitsch Y.V. Pyeritz R.E. Rogan P.K. Circulation. 1999; 100: 693-699Crossref PubMed Scopus (20) Google Scholar). It is estimated that up to 15% of all point mutations causing human genetic disease result from mRNA splicing defects, most of which involve a single nucleotide substitution at the splice sites (5.Green M.R. Ann. Rev. Genet. 1986; 20: 671-708Crossref PubMed Scopus (423) Google Scholar). To date, only a few mutations in the branchpoint consensus sequence of introns have been reported to cause human genetic diseases (9.Rosenthal A. Jouet M. Kenwrick S. Nat. Genet. 1992; 2: 107-112Crossref PubMed Scopus (254) Google Scholar, 10.Webb J.C. Patel D.D. Shoulders C.C. Knight B.L. Soutar A.K. Hum. Mol. Genet. 1996; 5: 1325-1331Crossref PubMed Scopus (57) Google Scholar, 11.Ris-Stalpers C. Verleun-Mooijman M.C. de Blaeij T.J. Degenhart H.J. Trapman J. Brinkmann A.O. Am. J. Hum. Genet. 1994; 54: 609-617PubMed Google Scholar, 12.Brand K. Dugi K.A. Brunzell J.D. Nevin D.N. Santamarina-Fojo S. J. Lipid. Res. 1996; 37: 1213-1223Abstract Full Text PDF PubMed Google Scholar, 13.Burrows N.P. Nicholls A.C. Richards A.J. Luccarini C. Harrison J.B. Yates J.R. Pope F.M. Am. J. Hum. Genet. 1998; 63: 390-398Abstract Full Text Full Text PDF PubMed Scopus (69) Google Scholar). However, the growing evidence indicates that the branchpoint sequence can be of essential importance for accurate and efficient splicing of human nuclear pre-mRNA. A series of elegant genetic experiments in both yeast and mammalian cell lines have shown that base pairs exist between the branchpoint sequence and a conserved region (5′-AUGAUG-3′) in U2 snRNA (14.Parker R. Siliciano P.G. Guthrie C. Cell. 1987; 49: 229-239Abstract Full Text PDF PubMed Scopus (323) Google Scholar, 15.Wu J. Manley J.L. Genes Dev. 1989; 3: 1553-1561Crossref PubMed Scopus (212) Google Scholar, 16.Zhuang Y. Weiner A.M. Genes Dev. 1989; 3: 1545-1552Crossref PubMed Scopus (201) Google Scholar). The branchpoint sequence is absolutely conserved in yeast (UACUAAC; the underlined A indicates the branchpoint nucleotide). Mutations of the yeast branchpoint adenosine significantly reduce or abolish splicing (17.Langford C.L. Klinz F.-J. Donath C. Gallwitz D. Cell. 1984; 36: 645-653Abstract Full Text PDF PubMed Scopus (207) Google Scholar). By contrast, the branchpoint sequence exhibits only a weak consensus in mammals (YNYURAY; where Y = pyrimidine, R = purine, and N = any nucleotide), and mutations of the branchpoint adenine in mammalian introns are known to only result in moderately reduced splicing efficiency due to the utilization of cryptic branchpoint sequences (18.Padgett R.A. Konarska M.M. Aebi M. Horning H. Weissmann C. Sharp P.A. Proc. Natl. Acad. Sci. U. S. A. 1985; 82: 8349-8353Crossref PubMed Scopus (67) Google Scholar, 19.Reed R. Maniatis T. Cell. 1985; 41: 95-105Abstract Full Text PDF PubMed Scopus (247) Google Scholar, 20.Ruskin B. Greene J.M. Green M.R. Cell. 1985; 41: 833-844Abstract Full Text PDF PubMed Scopus (146) Google Scholar). To compensate for the poorly conserved branchpoint sequence, most mammalian introns contain another cis-acting element, polypyrimidine tract, between the branchpoint sequence and the 3′-splice site. Early during the spliceosome assembly, the specific binding of U2AF to the polypyrimidine tract has been shown to promote the base pairing of U2 snRNA with the branchpoint sequence (21.Valcarcel J. Gaur R.K. Singh R. Green M.R. Science. 1996; 273: 1706-1709Crossref PubMed Scopus (229) Google Scholar). However, much less is still known about the mechanism of the selection of the branchpoint adenosine, which is usually located 18–40 nucleotides upstream of the 3′-splice site, in the mammalian intron with respect to the base pairs between the U2 snRNA and the pre-mRNA. The human lecithin:cholesterol acyltransferase (LCAT) is encoded by a single gene that is located on chromosome 16 and composed of six exons (22.Teisberg P. Gjone E. Olaisen B. Ann. Hum. Genet. 1975; 38: 327-331Crossref PubMed Scopus (45) Google Scholar, 23.McLean J. Wion K. Drayna D. Fielding C. Lawn R. Nucleic Acids Res. 1986; 14: 9397-9406Crossref PubMed Scopus (109) Google Scholar). Mutations in the LCAT gene have been demonstrated to underlie either familial LCAT deficiency or fish-eye disease (FED) (24.Kuivenhoven J.A. Pritchard P.H. Hill J. Frohlich J. Assmann G. Kastelein J. J. Lipid. Res. 1997; 38: 191-205Abstract Full Text PDF PubMed Google Scholar). We have previously identified a point mutation in intron 4 of the LCAT gene in patients with FED (25.Kuivenhoven J.A. Weibusch H. Pritchard P.H. Funke H. Benne R. Assmann G. Kastelein J.J. J. Clin. Invest. 1996; 98: 358-364Crossref PubMed Scopus (68) Google Scholar). This point mutation resides in a 100% matched branchpoint consensus sequence (CCCTGAC versus YNYTRAY), and it has been shown to result in the complete intron retention. Similar results have been observed when two other novel mutations (i.e. T → G or T → A) were introduced into the same site of the natural mutation (intervening sequence (IVS) 4: T → C), suggesting that the highly conserved thymidine residue in the branchpoint sequence is essential for the LCAT pre-mRNA splicing (26.Li M. Kuivenhoven J.A. Ayyobi A.F. Pritchard P.H. Biochim. Biophys. Acta. 1998; 1391: 256-264Crossref PubMed Scopus (11) Google Scholar). LCAT gene has an unusually high conserved sequence (27.Hixson J.E. Driscoll D.M. Birnbaum S. Britten M.L. Gene. 1993; 128: 295-298Crossref PubMed Scopus (13) Google Scholar). The intron 4 of LCAT gene contains only 83 base pairs, and the possible absence of cryptic branchpoint sequences in this short intron may render it highly sensitive to any alteration within the branchpoint region. Therefore, the branchpoint sequence in intron 4 of the LCAT gene may act as a model to identify the functional significance of conserved nucleotides contained within the branchpoint sequence of the human genes. In this study, we demonstrate that the mutations of the branchpoint adenosine residue in intron 4 of the human LCAT gene, like the natural mutation, completely abolish the splicing in the absence of cryptic branch sites in the intron. The point mutations introduced into other positions of the branchpoint sequence of the intron also have significant effects on the efficiency of splicing and thus the enzyme activity in vivo. We therefore establish the functional significance of the branchpoint sequence of intron 4 for the splicing of the human LCAT mRNA precursors. The LCAT minigene containing the full length cDNA (28.McLean J. Fielding C. Drayna D. Dieplinger H. Baer B. Kohr W. Henzel W. Lawn R. Proc. Natl. Acad. Sci. U. S. A. 1986; 83: 2335-2339Crossref PubMed Scopus (144) Google Scholar) and the wild type sequence of intron 4 (25.Kuivenhoven J.A. Weibusch H. Pritchard P.H. Funke H. Benne R. Assmann G. Kastelein J.J. J. Clin. Invest. 1996; 98: 358-364Crossref PubMed Scopus (68) Google Scholar) was first released from the pUC19-LCAT-intron 4 plasmid by the digestion of EcoRI and BamHI. TheEcoRI-BamHI DNA fragment was then subcloned into pcDNA3.1(−), a mammalian expression vector (Invitrogen, San Diego, CA) using standard procedures (29.Sambrook J. Frisch E.F. Maniatis T. Molecular Cloning: A Laboratory Manual. 2nd Ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY1989Google Scholar). The LCAT intron 4 mutants were constructed by the overlap-polymerase chain reaction (PCR) technique (30.Ho S.N. Hunt H.D. Horton R.M. Pullen J.K. Pease L.R. Gene (Amst.). 1989; 77: 51-59Crossref PubMed Scopus (6833) Google Scholar), and the pcDNA3.1-LCAT intron 4 wild type plasmid was used as the template. The upstream (5′-GGG AGA CCC AAG CTG GCT A-3′) and downstream (5′-CGT CGA GGC TGA TCA GCG G-3′) primers used in the mutagenesis are complementary to sequences that flank the 5′- and 3′-sites of the polylinker of the pcDNA3.1 vector, respectively. The sequences of site-directed oligonucleotide primers used to create the desired intron mutations are listed in TableI. In detail, the first PCR was performed with 100 ng of template DNA, 200 μm of each dNTP, 0.4 μm of each upstream primer and reverse mutagenic oligonucleotide, or 0.4 μm of each downstream primer and forward mutagenic oligonucleotide. The PCR was performed as follows: 30 cycles each of 30 s at 95 °C, 30 s at 56 °C, and 60 s at 72 °C with 2.5 units of pfu DNA polymerase (Stratagene, La Jolla, CA). The resultant two PCR fragments were gel-purified and used in a subsequent fusion reaction under the same PCR condition except for the extension time increased to 2 min. The final products purified from the agarose gel were digested with EcoRI andBamHI and then ligated into the pcDNA3.1 vector. The presence of the introduced intron mutations was confirmed by dideoxy sequencing.Table IOligonucleotide primers used to create the desired intron mutationsIVS4-19a5′-CTGCCCTGAA CCCTTCCAC-3′IVS4-19g5′-CTGCCCTGAG CCCTTCCAC-3′IVS4-20c5′-GCTGCCCTGCCCCTTCCA-3′IVS4-20g5′-GCTGCCCTGGCCCCTTCCA-3′IVS4-20t5′-GCTGCCCTGTCCCCTTCCA-3′IVS4-21c5′-AGCTGCCCTCACCCCTTCC-3′IVS4-21t5′-AGCTGCCCTTACCCCTTCC-3′IVS4-23a5′-CCAGCTGCCATGACCCCTT-3′IVS4-23g5′-CCAGCTGCCGTGACCCCTT-3′IVS4-24a5′-CCCAGCTGCACTGACCCCT-3′IVS4-24t5′-CCCAGCTGCTCTGACCCCT-3′IVS4-25a5′-CCCCAGCTG ACCTGACCCC-3′IVS4-25g5′-CCCCAGCTG GCCTGACCCC-3′IVS4-21a, -22c5′-CAGCTGCCCCAACCCCTTCC-3′IVS4-yeast5′-CCAGCTG TACTAAC CCCTTCC-3′The introduced mutations at the corresponding positions relative to the 3′-splice site of LCAT intron 4 are in boldface and the 7-bp consensus sequences are underlined. Open table in a new tab The introduced mutations at the corresponding positions relative to the 3′-splice site of LCAT intron 4 are in boldface and the 7-bp consensus sequences are underlined. Transient transfection of the LCAT minigene constructs into human embryonic kidney (HEK)293 cells was performed using the calcium phosphate co-precipitation method (31.Rosenthal N. Methods Enzymol. 1987; 152: 704-720Crossref PubMed Scopus (403) Google Scholar). The mammalian expression vectors, pcDNA3.1 and pcDNA3.1-LCAT cDNA as well as pcDNA3.1-LCAT IVS4 wild type, were used as negative and positive controls, respectively. For transfection, 28 μg of each of the plasmids was added to 100-mm Petri dishes containing approximately 106 cells in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) supplemented with 10% fetal bovine serum. The medium was replaced with serum-free Opti-MEM 24 h after transfection and incubated for another 48 h. Culture medium was collected for the LCAT activity assays, and total cytoplasmic RNA was prepared using the RNeasy Mini Kit (Qiagen) according to the manufacturer's protocol. LCAT activity in the collected media was determined using proteoliposome substrate as described previously (26.Li M. Kuivenhoven J.A. Ayyobi A.F. Pritchard P.H. Biochim. Biophys. Acta. 1998; 1391: 256-264Crossref PubMed Scopus (11) Google Scholar). The substrate contained apoA-I, [3H]cholesterol, and phosphatidylcholine at a molar ratio of 0.8:12.5:250. The reaction mixture containing 100 μl of HEK-293 transfection medium was incubated at 37 °C for 2 h, and the reaction was terminated by adding 1 ml of absolute ethanol. Unesterified cholesterol and cholesteryl esters were separated using thin-layer chromatography in petroleum ether/diethyl ether/acetic acid (70:12:1, v/v), and the radioactivity was determined by liquid scintillation counting. First-strand cDNA was synthesized from 1.0 μg of total cytoplasmic RNA in a 20-μl reverse transcription mixture containing 50 mmTris-HCl (pH 8.3), 75 mm KCl, 3 mmMgCl2, 1 mm dNTPs, 200 ng of oligo(dT)12–18, and 200 units of Moloney murine leukemia virus reverse transcriptase (RT; Life Technologies, Inc.) at 42 °C for 1 h. After reverse transcription, 10% of the reaction product was used for PCR amplification in a 50-μl final reaction volume (10 mm Tris-HCl (pH 8.3), 50 mm KCl, 1.5 mm MgCl2, 0.4 μm of primers that were derived from sequences of exon 2 (5′-GCT ACC GCA AGA CAG AGG AC-3′) and exon 5 (5′-GGC CAA TGA GGA AGA CAG GC-3′) of the LCAT gene, respectively, and 2 units of Taq DNA polymerase). To determine the efficiency of transient transfection, another primer pair complementary to the sequences (sense primer, 5′-GGG GTT CGA AAT GAC CGA CC-3′; antisense primer, 5′-CAG CTG GCA CGA CAG GTT TC-3′) within the neomycin resistance gene of the pcDNA3.1 vector were utilized to serve as an internal control. Thirty-five cycles of amplification were performed using a Perkin-Elmer DNA thermal cycler (System 2400). Each cycle consisted of a 30-s denaturation at 95 °C, a 30-s annealing at 62 °C, and a 30-s extension at 72 °C. The PCR products were separated on a 3% agarose gel, and the bands were visualized by ethidium bromide staining. To study the effects of various mutations in the branchpoint consensus sequence of intron 4 on the efficiency of LCAT pre-mRNA splicing in vivo, we first cloned the human LCAT minigene into the mammalian expression vector, pcDNA3.1 (Fig.1 A). Intron 4 of the human LCAT gene is small and consists of only 83 base pairs (Fig.1 B). The putative branchpoint sequence (CCCTGAC) is located −19 to −25 bases upstream of the 3′-splice site. For convenience, we designated the nucleotide at −19 relative to the 3′-splice site as the first position of the branchpoint sequence. Based on the wild type pcDNA3.1-LCAT minigene, we made a series of point mutations in the branch site using the synthetic oligonucleotides as described under "Experimental Procedures." Most of the site-directed mutagenesis involved the nucleotide transversion according to the consensus sequence YNYTRAY. We also made one plasmid construct in which the wild type branchpoint sequence CCCTGAC was replaced by the absolutely conserved yeast branch site sequence TACTAAC to test whether the invariant branchpoint sequence is the most preferred sequence for mammalian pre-mRNA splicing (32.Reed R. Maniatis T. Genes Dev. 1988; 2: 1268-1276Crossref PubMed Scopus (135) Google Scholar). The retention of intron 4 of human LCAT gene caused by the substitution of the thymine to other nucleotides in the putative branchpoint sequence (25.Kuivenhoven J.A. Weibusch H. Pritchard P.H. Funke H. Benne R. Assmann G. Kastelein J.J. J. Clin. Invest. 1996; 98: 358-364Crossref PubMed Scopus (68) Google Scholar, 26.Li M. Kuivenhoven J.A. Ayyobi A.F. Pritchard P.H. Biochim. Biophys. Acta. 1998; 1391: 256-264Crossref PubMed Scopus (11) Google Scholar) suggested that the mutations of the branchpoint adenosine residue also result in the defective splicing for this particular intron. As shown in TableII, the LCAT cDNA exhibited the highest LCAT activity. When the wild type intron 4 was inserted into cDNA, the enzymatic activity decreased to approximately 20–25% of the original, which was most likely due to the influence of the splicing process on the expression of the gene. However, like the natural mutation, IVS4-MUT, and its two related mutants, IVS4-MUT-1 and -MUT-2, substitution of the branchpoint adenine with a cytosine resulted in the absence of LCAT activity as compared with control transfections with the empty pcDNA3.1 vector. A similar result was also obtained when the branchpoint adenine was changed to guanosine or thymine. To determine further the effect of the branchpoint mutations on the LCAT pre-mRNA splicing in vivo, RT-PCR was performed. The wild type intron was efficiently spliced out of the pre-mRNA (Fig. 2, lane 3) giving a spliced RNA band corresponding to the LCAT cDNA (Fig. 2, lane 2), whereas no spliced RNAs could be detected for either the IVS4-MUT, -MUT-1, and -MUT-2 (Fig. 2, lanes 4–6) or the branchpoint mutants, IVS-20c, -20g, and -20t (Fig. 2,lanes 7–9). These observations are consistent with previous reports (32.Reed R. Maniatis T. Genes Dev. 1988; 2: 1268-1276Crossref PubMed Scopus (135) Google Scholar, 33.Zhuang Y. Goldsten A.M. Weiner A.M. Proc. Natl. Acad. Sci. U. S. A. 1989; 86: 2752-2756Crossref PubMed Scopus (117) Google Scholar) that mutations of the highly conserved thymine at the fourth position of the branchpoint consensus sequence have a similar effect on splicing as do mutations at the branch site adenosine.Table IIThe effect of mutations in the branch site adenine on the activity of LCAT from medium of HEK-293 cellsnmol/ml/hpcDNA3.10.04 ± 0.01***LCAT cDNA23.49 ± 1.59***IVS4-WT5.15 ± 0.89IVS4–22c (MUT)0.03 ± 0.01***IVS4–22g (MUT-1)0.02 ± 0.01***IVS4–22a (MUT-2)0.02 ± 0.01***IVS4–20c0.04 ± 0.01***IVS4–20g0.04 ± 0.01***IVS4–20t0.04 ± 0.02***Confluent HEK-293 cells were incubated in serum-free Opti-MEM for 48 h. The medium was assayed for LCAT activity as described under "Experimental Procedures." Data shown represent means ± S.D. from duplicate measurements of triplicate transfections. *** p < 0.0001 compared to the intron wild type. Open table in a new tab Confluent HEK-293 cells were incubated in serum-free Opti-MEM for 48 h. The medium was assayed for LCAT activity as described under "Experimental Procedures." Data shown represent means ± S.D. from duplicate measurements of triplicate transfections. *** p < 0.0001 compared to the intron wild type. In our study, we have not observed any alternative splicing of LCAT pre-mRNA containing the natural and the branchpoint mutations. Examination of the intron sequence upstream from the 3′-splice site (Fig. 1 B) does not reveal any potential branch site that matches the mammalian YURAY consensus sequence, and this may be the reason why we observe a complete abolishment of splicing. To test this hypothesis, we inserted a normal branchpoint sequence CCCTGAC into the intronic sequence just six bases upstream of the mutated branch sites (Fig. 3). The results of the insertion of the branchpoint sequence are shown in Fig.4. The introduction of a normal branch site totally restored the enzyme activities from the natural and the branchpoint mutants, respectively. Interestingly, the insertion of the branch site into the branchpoint mutant IVS4-20t resulted in even higher LCAT activity than did the wild type intron. However, the insertion had no obvious effect on the intron 4 wild type (Fig.4 A). The efficiently spliced RNAs from the natural and the branchpoint mutants were revealed by RT-PCR (Fig. 4 B,lanes 6 and 8, respectively), which further confirmed the effect of the insertion of a normal branch site into the intron sequences on the splicing of these mutated LCAT mRNA precursors. These results suggested that the inserted branch site was utilized when the normal branchpoint sequence contained a mutation.Figure 4The effect of the introduction of a normal branch site into the intron sequences on the splicing of the LCAT mRNA precursors containing the natural (IVS4-MUT) and the branchpoint (IVS4-20t) mutations. A, enzymatic activity secreted by transiently transfected HEK-293 cells. Bar 1, IVS4-WT; bar 2, IVS4-BPS-WT; bar 3, IVS4-MUT;bar 4, IVS4-BPS-MUT, bar 5, IVS4-20t; bar 6, IVS4-BPS-20t. ***p < 0.0001 compared with the intron wild type. B, RT-PCR analysis of total cytoplasmic RNA from HEK-293 cells transfected with the LCAT intron 4 mutants containing the branchpoint sequence insertion. Lane 1, pcDNA 3.1 vector without insert; lane 2, LCAT cDNA;lane 3, IVS4-WT; lane 4, IVS4-BPS-WT; lane 5, IVS4-MUT; lane 6, IVS4-BPS-MUT, lane 7, IVS4-20t; lane 8, IVS4-BPS-20t; and M, molecular weight marker. BPS represents the inserted normal branchpoint sequence.View Large Image Figure ViewerDownload Hi-res image Download (PPT) We have previously proposed (25.Kuivenhoven J.A. Weibusch H. Pritchard P.H. Funke H. Benne R. Assmann G. Kastelein J.J. J. Clin. Invest. 1996; 98: 358-364Crossref PubMed Scopus (68) Google Scholar, 26.Li M. Kuivenhoven J.A. Ayyobi A.F. Pritchard P.H. Biochim. Biophys. Acta. 1998; 1391: 256-264Crossref PubMed Scopus (11) Google Scholar) that the natural mutation T → C at the fourth position of the branch site of intron 4 of the LCAT gene might not only destroy the base pairing itself with the fourth position of AUGAUG sequence in human U2 snRNA but also disrupt the immediate downstream weak G-U base pairing interaction between the branchpoint sequence and U2 snRNA (Fig.5 A, scheme b). This disruption would result in the inappropriate bulging of the adenosine residue in the branchpoint region and thus lead to the defective splicing. Therefore, if the G at the third position of the branchpoint sequence is altered to A, the Watson-Crick A-U base pair between the branch site and the AUGAUG sequence in U2 snRNA might compensate, at least in part, for the natural mutation in the branchpoint sequence (Fig. 5 A, scheme c). As shown in Fig.5 B, the alteration from the sequence CCCCGAC to CCCCAAC did indeed increase the LCAT activity, although it was still very low as compared with the wild type intron. However, a correctly spliced transcript could be clearly identified (Fig. 5 C, lane 5). By contrast, no band corresponding to the mature mRNA transcript was observed in the cells that expressed the natural mutant (Fig. 5 C, lane 4). To exclude the possibility that the effect of the double mutant (IVS4-21a, -22c) on the splicing of LCAT pre-mRNA resulted from the differences in the efficiency of transient transfection rather than specific base-pairing interaction, we assayed the relative abundance of each expressed construct by RT-PCR with a set of primers that amplify 502-bp fragment of the neomycin resistance gene (NeoR), which is under the control of the SV40 early promoter in the pcDNA3.1 vector. As shown in Fig.5 C, the amount of amplified NeoR gene product was similar in the cells transfected with each plasmid construct (lanes 6–10), suggesting that there were no differences in the transient transfection efficiency. To demonstrate further the functional role of the branchpoint sequence, we identified the effects of mutations of other nucleotides in the branchpoint consensus sequence on the efficiency of LCAT pre-mRNA splicing by LCAT activity assay and RT-PCR (Fig. 6). The mutants IVS4-19g and IVS4-21c (Fig. 6 A, bars 4 and5) were associated with the lowest LCAT activities followed by IVS4-19a (Fig. 6 A, bar 3). A dramatic reduction in LCAT enzyme activity was also observed with IVS4-23g (Fig.6 A, bar 8). The mutants IVS4-21t and IVS4-23a (Fig. 6 A, bars 6 and 7) also had markedly decreased LCAT activities compared with that of the wild type. However, the C → A transversion or C → T transition at the sixth position (IVS4-24a or IVS4-24t) of the consensus sequence had only a mild decreased enzyme activity (Fig. 6 A, bars 9and 10) as compared with wild type control. Interestingly, in contrast to these downstream mutations, the change at the last position of the branchpoint sequence, i.e. IVS4-25a or -25g, resulted in approximately a 2.7-fold increase in LCAT activity compared with the intron 4 wild type (Fig. 6 A, bars 11 and 12). The effects of these intron mutations on the efficiency of RNA splicing were further determined by RT-PCR. As expected, the normally spliced RNA could be detected in all of these mutants, but the efficiency of the splicing was completely different from one another. The mutant IVS4-19g and IVS4-21c had the least spliced RNAs (Fig. 6 B,lanes 6 and 7) followed by IVS4-19a and IVS4-21t (Fig. 6 B, lanes 5 and 8). The densities of the spliced RNAs for the mutants IVS-23a and 23g were less than that of wild type (Fig. 5 B, lanes 9 and10), and the levels of spliced RNA in IVS4-24a and -24t were similar to those of wild type control (Fig. 5 B, lanes 11 and 12). It was interesting to note that the RNA splicing of mutants IVS-25a and -25g was much more efficient than that of the wild type (Fig. 5 B, lanes 13 and14), approximately 50% of the transcripts were spliced. All of these observations are consistent with the corresponding LCAT activities for these mutants. Therefore, it appears that the efficiency of splicing is well correlated with the activity of LCAT secreted from the transfected HEK-293 cells into the medium. Another surprising observation in this study was that the exchange of the wild type branchpoint sequence of intron 4 with the invariant TACTAAC box of yeast did not increase the efficiency of splicing but had a mildly decreased spliced RNA (Fig. 6 B, lane 4) and LCAT activity (Fig. 6 A, bar 2). We have previously shown that a point mutation in a lariat branchpoint sequence of intron 4 of the LCAT gene (IVS4:T-22C) resulted in the intron retention causing a human inherited disorder, fish-eye disease (25.Kuivenhoven J.A. Weibusch H. Pritchard P.H. Funke H. Benne R. Assmann G. Kastelein J.J. J. Clin. Invest. 1996; 98: 358-364Crossref PubMed Scopus (68) Google Scholar). We have also demonstrated that the thymine residue at the fourth position of the branchpoint sequence in which the natural mutation occurs is crucial for the pre-mRNA splicing (26.Li M. Kuivenhoven J.A. Ayyobi A.F. Pritchard P.H. Biochim. Biophys. Acta. 1998; 1391: 256-264Crossref PubMed Scopus (11) Google Scholar). In this report, we extend our studies into the whole region of the branchpoint sequence and further demonstrate that the branchpoint sequence of intron 4, especially the branchpoint adenosine residue and the nucleotides around it, is essential for the efficient splicing of LCAT pre-mRNA. Previous studies have shown that the branchpoint sequence is not essential for splicing of mammalian introns, because deletion or mutation of the branchpoint sequence does not abolish splicing bothin vitro and in vivo (18.Padgett R.A. Konarska M.M. Aebi M. Horning H. Weissmann C. Sharp P.A. Proc. Natl. Acad. Sci. U. S. A. 1985; 82: 8349-8353Crossref PubMed Scopus (67) Google Scholar, 19.Reed R. Maniatis T. Cell. 1985; 41: 95-105Abstract Full Text PDF PubMed Scopus (247) Google Scholar, 20.Ruskin B. Greene J.M. Green M.R. Cell. 1985; 41: 833-844Abstract Full Text PDF PubMed Scopus (146) Google Scholar). This paradox has been explained by the utilization of a relatively inefficient nearby cryptic branchpoint sequence (20.Ruskin B. Greene J.M. Green M.R. Cell. 1985; 41: 833-844Abstract Full Text PDF PubMed Scopus (146) Google Scholar). However, our results reported here clearly demonstrate that this is not always the case. The single base substitutions of the adenosine reside in the branchpoint region of intron 4 of LCAT pre-mRNA completely abolish splicing as assayed by the enzyme activity and RT-PCR. Newman et al. (34.Newman A.J. Lin R.J. Cheng S.C. Abelson J. Cell. 1985; 42: 335-344Abstract Full Text PDF PubMed Scopus (89) Google Scholar) has previously shown that mutation of the adenosine residue in the yeast branchpoint sequence abolishes the RNA splicing by preventing the cleavage at the 5′-splice site. In this respect, the mutations of the branchpoint A as well as the natural mutation in intron 4 of the LCAT gene might have the same effect on the cleavage at the 5′-splice site. Because the intronic sequence contains a stop codon signal, the failure to detect any LCAT activity and the truncated protein (data not shown) in the culture medium indicates that the translation products are likely rapidly degraded within the cells (26.Li M. Kuivenhoven J.A. Ayyobi A.F. Pritchard P.H. Biochim. Biophys. Acta. 1998; 1391: 256-264Crossref PubMed Scopus (11) Google Scholar). Recently, another example of the branchpoint mutation that disrupts mRNA splicing in intron 9 of the low density lipoprotein receptor gene has been described in a patient with familial hypercholesterolemia. This mutation disrupts the branchpoint consensus sequence and also causes intron retention (10.Webb J.C. Patel D.D. Shoulders C.C. Knight B.L. Soutar A.K. Hum. Mol. Genet. 1996; 5: 1325-1331Crossref PubMed Scopus (57) Google Scholar). These observations, therefore, suggest that the branchpoint sequence plays an important role in the selection of the branchpoint nucleotide, although it is poorly conserved in mammals. To investigate the possible mechanisms for the defective splicing, we first inserted a normal branchpoint sequence CCCTGAC into the natural and the branchpoint mutants (IVS4-22c and IVS4-20t, respectively) six nucleotides upstream of the mutated branch sites (Fig. 3). The purpose of the insertion is to see whether this sequence could rescue the natural or the branchpoint mutation. As expected, we observed not only the complete restoration of the LCAT activities of both the mutants but also an increase in splicing efficiency of the branchpoint mutant compared with the intron 4 wild type construct. These results suggest that the sequence of intron 4 itself does not contain any potential cryptic branch sites. Once the branchpoint sequence had been mutated, there was no alternative to back up the splicing machinery so that the consequence of the natural and the branchpoint mutations was retention of the intron. At the moment, we do not know the reasons why the introduction of the normal branch site into the intronic sequence of the branchpoint mutant increased the efficiency of LCAT pre-mRNA splicing. It is possible that two potential branch sites within a single intron might compete with each other (33.Zhuang Y. Goldsten A.M. Weiner A.M. Proc. Natl. Acad. Sci. U. S. A. 1989; 86: 2752-2756Crossref PubMed Scopus (117) Google Scholar). For the intron wild type and the natural mutant, either of the two potential branchpoint sequences could be screened as the branch site signal, but the one that mostly matches the consensus sequence would be preferentially utilized. In the case of the branchpoint mutant, the inserted branch site becomes the only branchpoint sequence to be used in the pre-mRNA splicing, which may be one of the possible explanations for the increased efficiency of pre-mRNA splicing. The base pairing interaction between the branchpoint sequence and the sequence AUGAUG in U2 snRNA is one of the early essential events in pre-mRNA splicing (14.Parker R. Siliciano P.G. Guthrie C. Cell. 1987; 49: 229-239Abstract Full Text PDF PubMed Scopus (323) Google Scholar, 15.Wu J. Manley J.L. Genes Dev. 1989; 3: 1553-1561Crossref PubMed Scopus (212) Google Scholar, 16.Zhuang Y. Weiner A.M. Genes Dev. 1989; 3: 1545-1552Crossref PubMed Scopus (201) Google Scholar). In an attempt to test the hypothesis that the natural mutation causes intron retention as the result of interference with the binding of U2 snRNA to the branchpoint sequence (Fig. 5 A), we changed the naturally mutated branchpoint sequence from CCCC GAC to CCCC AAC (the underlined C is the natural mutation, and the bold letters represent the mutation we made, G → A). This alteration would produce a stronger Watson-Crick base pair between the branchpoint region and the AUGAUG sequence in U2 snRNA compared with the natural mutation (Fig. 5 A), which might, at least in part, suppress the effect of the natural mutation on the LCAT pre-mRNA splicing by increasing the ability to tolerate a mismatch at the adjacent position. As expected, the base change just one nucleotide upstream of the branchpoint adenosine residue did recover some of the LCAT activity in the medium and produce a clearly spliced RNA band, although these changes were relatively small compared with the wild type control (Fig. 5, B and C). The results reported here indicate that the strength of the base pairs between the pre-mRNA and U2 snRNA may be one of the critical factors in determining the efficiency of nuclear pre-mRNA splicing. This conclusion was further supported in the mutational analysis of the functional significance of other specific nucleotides in the branchpoint consensus sequence. In addition to the mutations at the fourth position of the branchpoint sequence, the alterations of the nucleotides, especially around the branchpoint A, i.e.IVS4-19a, -19g, and -21c have a dramatic effect on the splicing of the intron, probably because these mutations greatly reduce the ability of potential base pairing with the U2 snRNA. And if the branchpoint adenosine residue was not flanked by a base-paired residue on either its 3′- or 5′-site, such as in the case of IVS4-19a, -19g, and -21c, the branch nucleotide would not be properly bulged and, therefore, would affect the splicing. Recently, Pascolo et al. (35.Pascolo E. Seraphin B. Mol. Cell. Biol. 1997; 17: 3469-3476Crossref PubMed Scopus (24) Google Scholar) demonstrated that U2 snRNA does indeed base pair, with the nucleotides preceding and following the branchpoint adenosine residue that is constrained into a bulged conformation. The mutants IVS4-23g, -21t, and -23a also have significant effects on the splicing efficiency, probably through a similar mechanism. Interestingly, in contrast to these downstream mutations, the substitution of the last nucleotide from C → A or G transversion significantly enhanced the efficiency of LCAT pre-mRNA splicing up to 2.7-fold as compared with the wild type control, whereas the substitution of the sixth position in the branchpoint sequence appeared to have less effect on the splicing, which is consistent with the consensus sequence in which the nucleotide could be either a purine or a pyrimidine. These observations suggest that the potential base pairing between the first four nucleotides in the branchpoint region and the U2 snRNA are probably one of the major determinants in the splicing of nuclear pre-mRNA, and this might also explain why the branch site in mammalian introns could be more degenerate. The reasons for the markedly increased splicing efficiency in the two mutants, IVS4-25a and -25g, are not clear. However, these findings raise the possibility that a DNA polymorphism involving the branchpoint sequence might affect the efficiency of mRNA splicing and thus have significant clinical implications. A common C → T polymorphism that resides in the last position of the branchpoint consensus sequence of intron 9 of the low density lipoprotein receptor gene has been reported (10.Webb J.C. Patel D.D. Shoulders C.C. Knight B.L. Soutar A.K. Hum. Mol. Genet. 1996; 5: 1325-1331Crossref PubMed Scopus (57) Google Scholar). However, this polymorphism has been observed to have no effect either on splicing in vitro or on plasma lipid concentrations in the population. This is not surprising, because the last position of the consensus sequence for the mammalian branchpoint sequence could be either C or T. Another unexpected observation made in this study is that the substitution of the normal branchpoint sequence CCCUGAC with the yeast branch site UACUAAC sequence did not increase the efficiency of LCAT pre-mRNA splicing but was modestly decreased. This result is surprising, because it is in conflict with a previous study (33.Zhuang Y. Goldsten A.M. Weiner A.M. Proc. Natl. Acad. Sci. U. S. A. 1989; 86: 2752-2756Crossref PubMed Scopus (117) Google Scholar) in which the data showed that the optimal branchpoint sequence is the yeast UACUAAC sequence for mammalian introns. Although the UACUAAC has perfect complementarity with the sequence AUGAUG in U2 snRNA, our results from the transfection experiments do not seem to support the concept that the perfect match improves the efficiency of pre-mRNA splicing in vivo. However, the observation that the yeast UACUAAC box did not increase the efficiency of LCAT pre-mRNA splicing is consistent with the results that the mutations of the last position of the branchpoint sequence markedly increase the splicing efficiency even though the changes reduce the match within the consensus sequence. This would, at least in part, explain why splicing is less dependent on base pairing of the pre-mRNA with U2 snRNA in mammals. For the mammalian introns, it is likely that other protein-RNA and RNA-RNA interactions are required to maintain the pre-mRNA and the U2 snRNA in the correct conformation in the absence of one or two of these base pairs. In conclusion, by taking advantage of our in vivo assay for measuring the splicing efficiency, we are able to demonstrate that the branchpoint consensus sequence of intron 4 is essential for the splicing of LCAT pre-mRNA. With the increasing intron sequence data available, more intron mutations far from the splice sites will be found to be the underlying causes for the human genetic disorders. We are grateful to Dr. Ross T. A. MacGillivray and Dr. John S. Hill for suggestions during the experiments and critical reading of the manuscript.