Two Proteins Essential for Apolipoprotein B mRNA Editing Are Expressed from a Single Gene through Alternative Splicing
2002; Elsevier BV; Volume: 277; Issue: 15 Linguagem: Inglês
10.1074/jbc.m111337200
ISSN1083-351X
AutoresGeoffrey S.C. Dance, Mark P. Sowden, Luca Cartegni, Ellen M. Cooper, Adrian R. Krainer, Harold C. Smith,
Tópico(s)RNA Research and Splicing
ResumoApolipoprotein B (apoB) mRNA editing involves site-specific deamination of cytidine to form uridine, resulting in the production of an in-frame stop codon. Protein translated from edited mRNA is associated with a reduced risk of atherosclerosis, and hence the protein factors that regulate hepatic apoB mRNA editing are of interest. A human protein essential for apoB mRNA editing and an eight-amino acid-longer variant of no known function have been recently cloned. We report that both proteins, henceforth referred to as ACF64 and ACF65, supported APOBEC-1 (the catalytic subunit of the editosome) equivalently in editing of apoB mRNA. They are encoded by a single 82-kb gene on chromosome 10. The transcripts are encoded by 15 exons that are expressed from a tissue-specific promoter minimally contained within the −0.33-kb DNA sequence. ACF64 and ACF65 mRNAs are expressed in both liver and intestinal cells in an approximate 1:4 ratio. Exon 11 is alternatively spliced to include or exclude 24 nucleotides of exon 12, thereby encoding ACF65 and ACF64, respectively. Recognition motifs for the serine/arginine-rich (SR) proteins SC35, SRp40, SRp55, and SF2/ASF involved in alternative RNA splicing were predicted in exon 12. Overexpression of these SR proteins in liver cells demonstrated that alternative splicing of a minigene-derived transcript to express ACF65 was enhanced 6-fold by SRp40. The data account for the expression of two editing factors and provide a possible explanation for their different levels of expression. Apolipoprotein B (apoB) mRNA editing involves site-specific deamination of cytidine to form uridine, resulting in the production of an in-frame stop codon. Protein translated from edited mRNA is associated with a reduced risk of atherosclerosis, and hence the protein factors that regulate hepatic apoB mRNA editing are of interest. A human protein essential for apoB mRNA editing and an eight-amino acid-longer variant of no known function have been recently cloned. We report that both proteins, henceforth referred to as ACF64 and ACF65, supported APOBEC-1 (the catalytic subunit of the editosome) equivalently in editing of apoB mRNA. They are encoded by a single 82-kb gene on chromosome 10. The transcripts are encoded by 15 exons that are expressed from a tissue-specific promoter minimally contained within the −0.33-kb DNA sequence. ACF64 and ACF65 mRNAs are expressed in both liver and intestinal cells in an approximate 1:4 ratio. Exon 11 is alternatively spliced to include or exclude 24 nucleotides of exon 12, thereby encoding ACF65 and ACF64, respectively. Recognition motifs for the serine/arginine-rich (SR) proteins SC35, SRp40, SRp55, and SF2/ASF involved in alternative RNA splicing were predicted in exon 12. Overexpression of these SR proteins in liver cells demonstrated that alternative splicing of a minigene-derived transcript to express ACF65 was enhanced 6-fold by SRp40. The data account for the expression of two editing factors and provide a possible explanation for their different levels of expression. apolipoprotein B expressed sequence tag group of overlapping clones reverse transcription serine/arginine-rich Chinese hamster ovary RNA recognition motif Mammalian mRNAs can be post-transcriptionally modified by site-specific adenosine or cytidine deaminases in a process known as mRNA editing (1.Smith H.C. Gott J.M. Hanson M.R. RNA. 1997; 3: 1105-1123PubMed Google Scholar). Editing of coding sequences or RNA splice sites can alter the primary sequence of a protein, often with profound physiological consequences. Adenosine deamination is catalyzed by a family of enzymes known as ADARs (adenosinedeaminases active on RNA) that recognize select adenosines within double-stranded regions of substrate mRNAs and convert them to inosine (2.Gerber A.P. Keller W. Trends Biochem. Sci. 2001; 26: 376-384Abstract Full Text Full Text PDF PubMed Scopus (193) Google Scholar). Editing of cytidine by deamination is catalyzed by enzymes known as CDARs (cytidine deaminases active on RNA) and results in a change to uridine. Of the many CDAR-like proteins reported (reviewed in Ref. 2.Gerber A.P. Keller W. Trends Biochem. Sci. 2001; 26: 376-384Abstract Full Text Full Text PDF PubMed Scopus (193) Google Scholar) only CDD1 (3.Dance G.S. Beemiller P. Yang Y. Mater D.V. Mian I.S. Smith H.C. Nucleic Acids Res. 2001; 29: 1772-1780Crossref PubMed Scopus (29) Google Scholar) and APOBEC-1 (4.Teng B. Burant C.F. Davidson N.O. Science. 1993; 260: 1816-1819Crossref PubMed Scopus (499) Google Scholar) are known to edit mRNA. The best characterized enzyme, APOBEC-1, catalyzes the deamination of nucleotide C6666 of apolipoprotein B mRNA, resulting in a codon change from glutamine to STOP (5.Chen S.H. Habib G. Yang C.Y. Gu Z.W. Lee B.R. Weng S.A. Silberman S.R. Cai S.J. Deslypere J.P. Rosseneu M. Gotto M. Li W.-H. Chan L. Science. 1987; 238: 363-366Crossref PubMed Scopus (537) Google Scholar). Thus, apoB1expression generates mRNAs that encode either apoB100 (unedited) or apoB48 (edited) protein isoforms (5.Chen S.H. Habib G. Yang C.Y. Gu Z.W. Lee B.R. Weng S.A. Silberman S.R. Cai S.J. Deslypere J.P. Rosseneu M. Gotto M. Li W.-H. Chan L. Science. 1987; 238: 363-366Crossref PubMed Scopus (537) Google Scholar). Although apoB100 and apoB48 have similar lipid carrying capacities in blood, apoB48 containing lipoprotein particles are cleared more rapidly from the circulation and do not associate with the lipoprotein (Lp(a)). Hence, elevated levels of apoB48 containing lipoproteins reduce atherogenic risk (6.Teng B. Blumenthal S. Forte T. Navaratnam N. Scott J. Gotto Jr., A.M. Chan L. J. Biol. Chem. 1994; 269: 29395-29404Abstract Full Text PDF PubMed Google Scholar,7.Wu Y. Teng B.B. Brandt M.L. Piedra P.A. Liu J. Chan L. J. Surg. Res. 1999; 85: 148-157Abstract Full Text PDF PubMed Scopus (6) Google Scholar). Unlike the ADARs, APOBEC-1 cannot edit mRNA without interactions with auxiliary proteins (4.Teng B. Burant C.F. Davidson N.O. Science. 1993; 260: 1816-1819Crossref PubMed Scopus (499) Google Scholar). A 64-kDa human protein, cloned and identified as either ACF (APOBEC-1complementation factor) (8.Mehta A. Kinter M.T. Sherman N.E. Driscoll D.M. Mol. Cell. Biol. 2000; 20: 1846-1854Crossref PubMed Scopus (221) Google Scholar) or ASP (APOBEC-1 stimulating protein) (9.Lellek H. Kirsten R. Diehl I. Apostel F. Buck F. Greeve J. J. Biol. Chem. 2000; 275: 19848-19856Abstract Full Text Full Text PDF PubMed Scopus (146) Google Scholar), together with APOBEC-1 function as a minimal editosome in vitro. Several studies have suggested that apoB mRNA is edited by a multiprotein complex of 27 S called the C to U editosome (10.Harris S.G. Sabio I. Mayer E. Steinberg M.F. Backus J.W. Sparks J.D. Sparks C.E. Smith H.C. J. Biol. Chem. 1993; 268: 7382-7392Abstract Full Text PDF PubMed Google Scholar, 11.Blanc V. Navaratnam N. Henderson J.O. Anant S. Kennedy S. Jarmuz A. Scott J. Davidson N.O. J. Biol. Chem. 2001; 276: 10272-10283Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar). The complexity of the editosome may be attributed to several identified proteins that associated with APOBEC-1 and/or the apoB mRNA editing site to modulate the efficiency of apoB mRNA editing (10.Harris S.G. Sabio I. Mayer E. Steinberg M.F. Backus J.W. Sparks J.D. Sparks C.E. Smith H.C. J. Biol. Chem. 1993; 268: 7382-7392Abstract Full Text PDF PubMed Google Scholar, 11.Blanc V. Navaratnam N. Henderson J.O. Anant S. Kennedy S. Jarmuz A. Scott J. Davidson N.O. J. Biol. Chem. 2001; 276: 10272-10283Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar, 12.Lau P.P. Chen S.H. Wang J.C. Chan L. Nucleic Acids Res. 1990; 18: 5817-5821Crossref PubMed Scopus (52) Google Scholar, 13.Navaratnam N. Shah R. Patel D. Fay V. Scott J. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 222-226Crossref PubMed Scopus (86) Google Scholar, 14.Schock D. Kuo S.R. Steinburg M.F. Bolognino M. Sparks J.D. Sparks C.E. Smith H.C. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 1097-1102Crossref PubMed Scopus (48) Google Scholar, 15.Lau P.P. Zhu H.J. Nakamuta M. Chan L. J. Biol. Chem. 1997; 272: 1452-1455Abstract Full Text Full Text PDF PubMed Scopus (87) Google Scholar, 16.Greeve J. Lellek H. Rautenberg P. Greten H. Biol. Chem. 1998; 379: 1063-1073Crossref PubMed Scopus (38) Google Scholar). The interactions, functions, and regulation of the auxiliary proteins are of interest not only in the study of the mechanism of apoB mRNA editing but also because apoB mRNA editing is a regulated process during tissue development (17.Funahashi T. Giannoni F. DePaoli A.M. Skarosi S.F. Davidson N.O. J. Lipid Res. 1995; 36: 414-428Abstract Full Text PDF PubMed Google Scholar) and in response to metabolic or hormonal perturbation (18.Van Mater D. Sowden M.P. Cianci J. Sparks J.D. Sparks C.E. Ballatori N. Smith H.C. Biochem. Biophys. Res. Commun. 1998; 252: 334-339Crossref PubMed Scopus (21) Google Scholar, 19.McCahill A. Lankester D.J. Park B.S. Price N.T. Zammit V.A. Mol. Cell Biochem. 2000; 208: 77-87Crossref PubMed Google Scholar). Mehta et al. (8.Mehta A. Kinter M.T. Sherman N.E. Driscoll D.M. Mol. Cell. Biol. 2000; 20: 1846-1854Crossref PubMed Scopus (221) Google Scholar) reported ACF as a 586-amino acid protein in intestine, whereas Lellek et al. (9.Lellek H. Kirsten R. Diehl I. Apostel F. Buck F. Greeve J. J. Biol. Chem. 2000; 275: 19848-19856Abstract Full Text Full Text PDF PubMed Scopus (146) Google Scholar) reported an identical intestinal 586-amino acid protein, ASP (henceforth referred to as ACF64), and a 594-amino acid variant (ACF65) isolated from liver cDNAs. The role of ASP in apoB mRNA editing had not been determined, nor was it known how these virtually identical proteins were encoded. We show that human ACF64 and ACF65mRNAs are encoded by the same gene through alternative splicing of exon 12 and that their protein products support equivalent levels of editing in cells. A promoter region was identified immediately upstream of exon 1, and several ESTs were identified that suggested that the ACF transcript may be subject to further alternative and nonexclusive alternative splicing events. cDNA sequences for ACF64 (AF209192) and ACF65 (AJ272079) were aligned to the Celera (www.celera.com) human genomic data base by BLAST analysis and identified an incomplete 53-kb contig on chromosome 10 (identification number hCT18761). This contig was used to perform a BLAST search of the Public Chromosome 10 data base (www.sanger.ac.uk) and identified the bacterial artificial chromosome (BAC) bA449016, which contains the entire ACF64/ACF65 cDNA sequence. Sequence data reported in this manuscript were produced by the Chromosome 10 Sequencing Group at the Sanger Center (ftp.sanger.ac.uk/pub/human/sequences/Chr_10/unfinished_sequence/bA449016). Restriction enzyme-digested human genomic DNA was resolved through 1% agarose and transferred to a nylon membrane (Stratagene). Probes for exons 11 and 12 were generated by PCR and radiolabeled with [α-32P]dCTP using the RTS RadPrime system (Invitrogen). The blots were hybridized to probe in ExpressHyb (CLONTECH) at 60 °C, washed, and autoradiographed. Cell lines were obtained from ATCC (Manassas, VA) and maintained as recommended. Primary hepatocytes were obtained from Dr. S. Strom (University of Pittsburgh). The transfections were performed using Fugene 6 (Roche Molecular Biochemicals), and the total cellular RNA was isolated as previously described (20.Sowden M. Hamm J.K. Spinelli S. Smith H.C. RNA. 1996; 2: 274-288PubMed Google Scholar). Alternatively spliced variants of ACF were detected by RT-PCR using oligo(dT) primed first strand cDNA (20.Sowden M. Hamm J.K. Spinelli S. Smith H.C. RNA. 1996; 2: 274-288PubMed Google Scholar) and HSACF64/ACF65–5′ and HSACF64/ACF65–3′ to amplify exons 11 and 12. Exons 8–10 were amplified with E4Δ5E6 and HindIII5′. ΔE5 was specific for the exon 8/10 junction, and ACF65WT was specific for the inserted 24 nucleotides in ACF65. ACF mRNA sequences were analyzed for high score SR protein motifs (21.Liu H.X. Chew S.L. Cartegni L. Zhang M.Q. Krainer A.R. Mol. Cell. Biol. 2000; 20: 1063-1071Crossref PubMed Scopus (183) Google Scholar, 22.Liu H.X. Zhang M. Krainer A.R. Genes Dev. 1998; 12: 1998-2012Crossref PubMed Scopus (422) Google Scholar) using curated data sets for the score matrices. Only scores above the following thresholds are considered to have potential for exonic splicing enhancer (ESE) function: SF2/ASF-1.956, SC35–2.383, SRp40–2.670, and SRp55–2.676 (see Fig. 6). A reporter minigene construct containing a 2.3-kb PCR product encompassing exons 11 and 12, and the intron was amplified from the BAC clone (bA449016) and subcloned into pcDNAIII V5-His (Invitrogen). To ensure detection of the ACF64/65 spliced variant mRNAs encoded by the transfected minigene, an ACF-specific and a vector-specific primer (T7) were used in the PCR. Expression plasmids for SR proteins have been described (23.Caceres J.F. Misteli T. Screaton G.R. Spector D.L. Krainer A.R. J. Cell Biol. 1997; 138: 225-238Crossref PubMed Scopus (327) Google Scholar). ApoB RNA editing efficiency was determined upon RT-PCR-amplified human apoB reporter RNA transcripts by poisoned primer extension analysis (24.Sowden M.P. Smith H.C. Biochem. J. 2001; 359: 697-705Crossref PubMed Scopus (16) Google Scholar). An ACF64 cDNA was PCR-amplified from a rat liver cDNA library (Stratagene) and subcloned into p cDNAIII-V5 (Invitrogen). ACF65 cDNA was derived from ACF64 by run-around PCR (25.Fisher C.L. Pei G.K. BioTechniques. 1997; 23: 570-574Crossref PubMed Scopus (293) Google Scholar). A 1.8-kb PCR-amplified human genomic DNA fragment containing exon 1 as well as promoter subfragments were subcloned into p GL3Basic (Promega). The cells were transfected in triplicate in 24-well cluster dishes with 0.5 μg of reporter DNA and 0.2 ng of DNA encoding Renilla luciferase (pRL-CMV; Promega) as a transfection control and harvested after 48 h. Luciferase activity was determined using the dual luciferase assay kit (Promega). BLAST analyses using cDNAs for ACF64 (AF209192) and ACF65 (AJ272079) of the public and private genomic data bases were performed. The Celera data base contained a 53-kb aligned contig (identification number hCT18761) that mapped to chromosome 10, position 10q11.21 at 47.76–47.85 m but that lacked 135 nucleotides of cDNA sequence. BLAST analysis of the Sanger Center's unpublished chromosome 10 data base with this contig identified a single BAC clone (bA449016) that mapped to the identical region but contained all ACF64 and ACF65 cDNA sequences. The exon/intron structure of the ACF gene was determined by pairwise BLAST (www.ncbi.nlm.nih.gov/BLAST) of bA449016 sequence with ACF64/ACF65 cDNA sequences. The ACF gene spans 82 kb and comprises 15 exons (Fig. 1) ranging from 45 to 274 bp (Table I). The introns range from 997 bp to 21.4 kb, and all splice junctions follow the consensus GT-AG motif. The currently described functional domains of ACF are present in separate exons (Table I). ACF65 cDNA differs from ACF64 by insertion of 24 nucleotides at position 1140 with respect to the start codon (9.Lellek H. Kirsten R. Diehl I. Apostel F. Buck F. Greeve J. J. Biol. Chem. 2000; 275: 19848-19856Abstract Full Text Full Text PDF PubMed Scopus (146) Google Scholar). Significantly, this insertion is identical to the 3′ 24 nucleotides of intron 11, suggesting that ACF65 and ACF64 are alternatively spliced variants of exon 12 (Fig. 1).Table IGene structure of ACFIntron 3′ sequenceExon terminiIntron 5′ sequence and size (bp)Exon number and size (bp)Codon at splice junctionProtein domainATAATCA- ∼ -TACTCAGgtatga- 213671//475′-Untranslated regionctagATAACAG- ∼ -AAATTATgtaagt- 8202//485′-Untranslated regiontcagCATGACT- ∼ -CAAAGAGgtgtaa- 290331-aAlternatively spliced exons located in EST clones only.//935′-Untranslated regiontcagTGAGCAA- ∼ -GGTCCAGgtagga- 88664//144QNone definedaaagGAATTTT- ∼ -GGAAAAGgtaagc- 654251-aAlternatively spliced exons located in EST clones only.//143Internal STOPgcagGAAAATG- ∼ -GTGAAAAgtgagt- 19956//135KRNP2 of RRM1gcagAATCGGT- ∼ -AAATTAGgtaagc- 55497//131RRNP1 of RRM1gcagAAATGGG- ∼ -CTACCAGgttagc- 77788//239GRRM2acagGAAGAAT- ∼ -AAACCAGgtagga- 74819//165GRNP2 of RRM3acagGTGCTGT- ∼ -TGGCAAGgtaagg- 427210//98KRNP1 of RRM3acagGTGCTGG- ∼ -GTTAGAGgtaaca- 1943/196711//274GNone definedctagAAATTTA- ∼ -TCCCCAGgtaggt- 273012 ASP variant//206QASP variantgtagGGGCTGC- ∼ -TCCCCAGgtaggt- 273012 ACF variant//182QACF variantgtagATATTAG- ∼ -CTGCAATgtgcgt- 99713//137IDouble-stranded RNA bindingccagCCACCCT- ∼ -TTCCCAGgtatgg- 291814//149GNone definedccagGATATGC- ∼ -AAAAGAA15//211None definednyagG/A-Consensus-C/AAGgtragt-The exons and their sizes, coding capacity, and sequences at their 5′ and 3′ termini are listed. Nucleotides shown in bold indicate the reading frame of each exon.1-a Alternatively spliced exons located in EST clones only. Open table in a new tab The exons and their sizes, coding capacity, and sequences at their 5′ and 3′ termini are listed. Nucleotides shown in bold indicate the reading frame of each exon. An alternative origin for ACF64 and ACF65 mRNAs would be gene duplication. To investigate this possibility, human genomic DNA was analyzed by Southern blotting using diagnostic restriction sites located within intronic sequences flanking exons 11 and 12 (Fig. 2). The exon 11 probe hybridized to only the predicted 2.6-, 4.3-, and 5.5-kb restriction fragments. Similarly, the exon 12 probe, containing the additional 24 nucleotides, hybridized to only the predicted 5.1- and 4.3-kb fragments and two fragments of 2.35 and 0.8 kb. The lack of hybridization of either probe to other restriction fragments indicated that there is only one gene that encoded both ACF64 and ACF65. Moreover, searches of the public and private human genomic DNA data bases with cDNA or identified genomic sequence failed to identify other loci that encode complete cDNA sequences. Despite extensive searching of GenBank's human EST data base, no indication of the transcribed sequence upstream of exon 1 was found. To prove that the identified ACF gene is functional and not an inactive pseudogene, the upstream sequence was evaluated for promoter activity. Exon 1 and the −1.73-kb region were subcloned into the promoter assay plasmid, pGL3Basic. Promoter activity was assayed in transfected HepG2, Caco2, and HeLa cells and normalized with respect to Renilla luciferase activity. The −1.7-kb ACF gene fragment showed a 147- or 427-fold increase in luciferase expression relative to pGL3Basic control in HepG2 and Caco2 cells, respectively (Fig. 3A). There was no detectable promoter activity in HeLa cells. To further delimit where the cell type-specific promoter activity resides, deletion constructs were evaluated that contained either exon 1 through −0.33 kb or exon 1 plus the distal −1.4-kb upstream sequence but in which the exon 1-proximal −330 bp had been deleted. The −330-bp region proximal to exon 1 retained significant promoter activity in both HepG2 and Caco2 cells (Fig. 3B). However, deletion of the upstream 1.4-kb sequence resulted in a 2-fold less active promoter in Caco2 cells (180-fold induction) compared with a promoter of equivalent strength in HepG2 cells (115-fold induction). The −330-bp promoter fragment showed no activity in HeLa cells. The reverse orientation of the −330-bp or −1.4-kb fragments had barely detectable promoter activity in HepG2 cells (Fig. 3A). These data indicate that a functional promoter is located in the −0.3-kb region that is active in liver and intestinal cell lines but not cervical carcinoma cells. Moreover, there are likely intestinal cell-specific regulatory elements in the upstream −1.4-kb fragment. Analysis of the −1.73-kb region using TESS and NSITE (searchlauncher.bcm.tmc.edu/seq-search/gene-search.html) revealed multiple potential transcription factor binding sites (Fig. 3B). Within the −330-bp promoter region, putative binding sites were identified for general, liver-specific, and hormonally induced transcription factors. Lellek et al. (9.Lellek H. Kirsten R. Diehl I. Apostel F. Buck F. Greeve J. J. Biol. Chem. 2000; 275: 19848-19856Abstract Full Text Full Text PDF PubMed Scopus (146) Google Scholar) reported the presence of ACF64 and ACF65 in intestinal and liver cDNA libraries, respectively, whereas Mehta et al.(8.Mehta A. Kinter M.T. Sherman N.E. Driscoll D.M. Mol. Cell. Biol. 2000; 20: 1846-1854Crossref PubMed Scopus (221) Google Scholar) identified ACF64 cDNAs from intestinal and kidney libraries. To determine whether there is tissue-specific expression of each variant in human cells, an RT-PCR strategy was employed to coamplify both mRNAs. Oligo(dT)-primed first strand cDNA was amplified with primers HSACF64/ACF65 5′ and HSACF64/ACF65 3′, which are located in exons 11 and 12, respectively. PCR products of 224 bp for ACF65 and 200 bp for ACF64 were detected in primary human hepatocytes, HepG2, hepatoma cells, and Caco2 intestinal cells at an approximate ratio of 4:1. (Fig. 4). These data are the first demonstration that the spliced variants are simultaneously expressed in liver and intestinal cell types. ACF64 is necessary and sufficient as an auxiliary factor for complementing APOBEC-1 to edit apoB mRNA (8.Mehta A. Kinter M.T. Sherman N.E. Driscoll D.M. Mol. Cell. Biol. 2000; 20: 1846-1854Crossref PubMed Scopus (221) Google Scholar). However, the function of ACF65 has not been determined (9.Lellek H. Kirsten R. Diehl I. Apostel F. Buck F. Greeve J. J. Biol. Chem. 2000; 275: 19848-19856Abstract Full Text Full Text PDF PubMed Scopus (146) Google Scholar). The eight-amino acid insertion (EIYMNVPV) constitutes a tyrosine phosphorylation site (NETPHOS; www.cbs.dtu.dk/services/NetPhos/) that might alter complementation activity and/or subcellular localization of ACF65. CHO cells were transfected with expression vectors encoding epitope-tagged ACF64 or ACF65, together with APOBEC-1 and apoB mRNA expression vectors. CHO cells were selected for these analyses because they do not express detectable levels of ACF by Western blot compared with HepG2 or Caco2 cells (data not shown), nor are they capable of supporting high levels of APOBEC-1-dependent editing in the absence of exogenous ACF64 or ACF65 (4% editing; Fig. 5A). ACF64 and ACF65 increased the editing efficiency of the apoB reporter mRNA 17-fold (Fig. 5A). Furthermore, an equivalent level of promiscuous editing (Fig. 5A, bands labeled 1 and 2) was promoted by both proteins. Western blotting demonstrated that the epitope-tagged ACF64 and ACF65 were expressed at equivalent levels and that the expression of APOBEC-1 was also comparable between all transfectants (Fig. 5B). Editing efficiency is not affected by apoB mRNA abundance (20.Sowden M. Hamm J.K. Spinelli S. Smith H.C. RNA. 1996; 2: 274-288PubMed Google Scholar), and therefore the expression level of the apoB reporter RNA was not determined. The subcellular localization of epitope-tagged ACF65 was determined by immunocytochemistry in editing competent (McArdle RH7777) and incompetent (CHO) cell lines (data not shown) and showed equivalent nuclear and cytoplasmic localization to that of ACF64 (26.Yang Y. Sowden M.P. Smith H.C. J. Biol. Chem. 2000; 275: 22663-22669Abstract Full Text Full Text PDF PubMed Scopus (43) Google Scholar). The SR family of proteins consists of highly conserved serine/arginine-rich RNA-binding proteins that regulate alternative splice site selection (reviewed in Ref. 27.Graveley B.R. RNA. 2000; 6: 1197-1211Crossref PubMed Scopus (884) Google Scholar). Sequence analysis (21.Liu H.X. Chew S.L. Cartegni L. Zhang M.Q. Krainer A.R. Mol. Cell. Biol. 2000; 20: 1063-1071Crossref PubMed Scopus (183) Google Scholar, 22.Liu H.X. Zhang M. Krainer A.R. Genes Dev. 1998; 12: 1998-2012Crossref PubMed Scopus (422) Google Scholar) predicted several high score motifs for the SR proteins SC35, SRp40, SRp55, and SF2/ASF in exon 12 (Fig. 6A). To evaluate whether SR proteins are involved in regulating exon 12 alternative splicing, HepG2 cells were cotransfected with the exon 11–12 minigene (Fig. 6A) and an expression plasmid for each of the four SR proteins. RT-PCR was performed using primers specific for the ACF64/65 spliced variant mRNAs encoded by the transfected minigene. The PCR products differ by only 24 nucleotides; therefore, it is reasonable to assume that each mRNA was amplified with equal efficiency; hence the ratio of the two bands is a good semi-quantitative means of evaluating the level of alternative splicing within each sample (Fig. 6B). Compared with the vector control, SRp40 had the most significant effect on alternative splicing. Its expression promoted the inclusion of the 24 nucleotides at the 5′ end of exon 12. SRp55 increased the abundance of ACF65 mRNA slightly. SC35 marginally increased ACF64 mRNA. SF2/ASF had no effect on the ratio of ACF65/ACF64. SR proteins affect alternative splicing in vivoin a concentration-dependent manner (27.Graveley B.R. RNA. 2000; 6: 1197-1211Crossref PubMed Scopus (884) Google Scholar), and thus it is likely that the alteration of the ratio of endogenous SR proteins to mRNA transcript will affect the level of alternative splicing. The reduced ratio of ACF65/ACF64 mRNA expressed from the minigene construct compared with that observed on the endogenous transcript (4:1 ratio; Fig. 4) is likely due, therefore, to overexpression of the ACF reporter RNA. Recently, a deletion variant of ACF (ACFdel55) was reported (28.Chester A. Scott J. Anant S. Navaratnam N. Biochim. Biophys. Acta. 2000; 1494: 1-13Crossref PubMed Scopus (76) Google Scholar) for which neither a function nor an origin was described. From our described exon structure ACFdel55 lacks exon 9 and therefore likely results from alternative splicing of the ACF primary transcript. To evaluate this possibility, mRNA from primary human hepatocytes, HepG2 cells, and Caco2 cells was analyzed by RT-PCR. Primers specific for exons 8 (E4Δ5E6) and 10 (ACF HindIII3′) generated a 292-bp product indicative of exon 9 inclusion and, although at much lower abundance, a product of 127 bp, the expected size for an mRNA in which exon 9 had been skipped (Fig. 7A). The size difference between the two PCR products precludes an accurate determination of their relative abundance, but it is likely that the ACFdel55variant is very poorly expressed. A primer (ΔE5) specific for the exon 8/10 junction created in ACFdel55 mRNA together with ACF65WT (specific for the alternatively spliced 24 nucleotides of exon 12) yielded a PCR product of 408 bp (Fig. 7B). The presence of another ACF variant that lacked exon 9 but included the 24 nucleotides at the 5′ end of exon 12 demonstrates that the two identified alternative splicing events of ACF RNA are not mutually exclusive. We propose that the exon 9-skipped ACF64 and ACF65 mRNAs be named ACF58 and ACF59 based upon their theoretical molecular masses. Human EST data base analyses revealed further ACF splice variants (Fig. 8). Some have an altered exon composition upstream of the ATG initiator codon in exon 4. AV655933 is a unique variant that includes an additional exon (exon 3), whereas AK000324,AV687201, AV698948, AV684607, and AV688086 all lack exon 2. The role of these alternative splicing events located in the 5′-untranslated region is unclear unless ACF translation is subject to regulation through elements contained within these exons. Variants AK000324 and AF271790include a novel exon 5, and both comprise the shorter variant of exon 12. Whether alternative splicing of these upstream exons regulates the downstream exon composition of ACF remains to be elucidated. The cloning of APOBEC-1 demonstrated that cytidine to uridine editing of apoB mRNA editing requires a cytidine deaminase whose activity on RNA depends on auxiliary protein factors (4.Teng B. Burant C.F. Davidson N.O. Science. 1993; 260: 1816-1819Crossref PubMed Scopus (499) Google Scholar). The inability of APOBEC-1 alone to edit apoB mRNA can be explained, in part, by its low affinity and nonselective binding to the editing site (29.MacGinnitie A.J. Anant S. Davidson N.O. J. Biol. Chem. 1995; 270: 14768-14775Abstract Full Text Full Text PDF PubMed Scopus (110) Google Scholar). Consequently, multiple auxiliary proteins have been implicated in apoB mRNA editing through their ability to interact with the apoB mRNA editing site and/or APOBEC-1 and in some cases, modulate editing activity (8.Mehta A. Kinter M.T. Sherman N.E. Driscoll D.M. Mol. Cell. Biol. 2000; 20: 1846-1854Crossref PubMed Scopus (221) Google Scholar, 9.Lellek H. Kirsten R. Diehl I. Apostel F. Buck F. Greeve J. J. Biol. Chem. 2000; 275: 19848-19856Abstract Full Text Full Text PDF PubMed Scopus (146) Google Scholar, 11.Blanc V. Navaratnam N. Henderson J.O. Anant S. Kennedy S. Jarmuz A. Scott J. Davidson N.O. J. Biol. Chem. 2001; 276: 10272-10283Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar, 14.Schock D. Kuo S.R. Steinburg M.F. Bolognino M. Sparks J.D. Sparks C.E. Smith H.C. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 1097-1102Crossref PubMed Scopus (48) Google Scholar, 15.Lau P.P. Zhu H.J. Nakamuta M. Chan L. J. Biol. Chem. 1997; 272: 1452-1455Abstract Full Text Full Text PDF PubMed Scopus (87) Google Scholar, 16.Greeve J. Lellek H. Rautenberg P. Greten H. Biol. Chem. 1998; 379: 1063-1073Crossref PubMed Scopus (38) Google Scholar). Recently, ACF was shown to comprise, together with APOBEC-1, the minimal functional editosome in vitro (8.Mehta A. Kinter M.T. Sherman N.E. Driscoll D.M. Mol. Cell. Biol. 2000; 20: 1846-1854Crossref PubMed Scopus (221) Google Scholar). Simultaneously, two variants referred to as ASP were described, the shorter of which is identical to ACF in primary sequence and function (9.Lellek H. Kirsten R. Diehl I. Apostel F. Buck F. Greeve J. J. Biol.
Referência(s)