Distinct Promoters Induce APOBEC-1 Expression in Rat Liver and Intestine

From the Medizinische Kernklinik und Poliklinik, Universitäts-Krankenhaus Eppendorf, Hamburg, Germany.

Correspondence to Jobst Greeve, Medizinische Klinik, Universitäts-Krankenhaus Eppendorf, Martinistrasse 52, 20246 Hamburg, Germany.

	Abstract

Top Abstract Introduction Methods Results Discussion References

 
Abstract—The expression of apolipoprotein (apo) B can be modulated by mRNA editing, a unique posttranscriptional base change in the apo B mRNA. Apo B-48, the translation product of edited apo B mRNA, is not a precursor of the atherogenic low density lipoproteins and lipoprotein(a). In humans and various other mammals, the apo B mRNA is edited in the intestine but not in the liver, which exclusively secretes apo B-100–containing lipoproteins as precursors for low density lipoprotein formation. In species such as the rat, mouse, dog, and horse, apo B mRNA is also edited in the liver, resulting in low plasma levels of low density lipoprotein. Editing of the apo B mRNA is mediated by the apo B mRNA–editing enzyme complex, of which the catalytic subunit APOBEC-1 is not expressed in the liver of species without hepatic editing. To understand the molecular basis for liver-specific expression of APOBEC-1 and the editing of hepatic apo B mRNA, the expression pattern and genomic organization of the rat APOBEC-1 gene have been characterized. The rat APOBEC-1 gene contains 6 exons and 2 promoters with distinct activities. The expression of APOBEC-1 in the rat liver is the result of a promoter located upstream, with tissue-specific exon use and alternate splicing within the 5'-untranslated region of APOBEC-1 mRNA encoded by exon 2. In addition to the liver, this promoter also induces APOBEC-1 expression in the spleen, lung, kidney, heart, and skeletal muscle. The promoter located downstream belongs to a new class of TATA-less promoters and is responsible for the abundant expression of APOBEC-1 in the intestine. Mapping of the transcriptional start sites and deletion analysis of the promoter regions by using luciferase as the reporter gene have defined the regulatory elements of both promoters. The downstream, intestine-specific promoter contains a negative regulatory element between -1100 and -500, which appears to restrict its activity to the intestine. The upstream, liver-specific promoter of the rat APOBEC-1 gene induces APOBEC-1 expression and editing of apo B mRNA in human hepatoma HuH-7 and Hep G2 cells. Understanding the molecular basis for the liver-specific expression of APOBEC-1 in the rat promises new strategies to induce APOBEC-1 expression in the human liver for the reduction of atherogenic lipoprotein levels by hepatic apo B mRNA editing.

Key Words: apo B mRNA editing • hepatic expression • APOBEC-1 • TATA-less promoter

	Introduction

Top Abstract Introduction Methods Results Discussion References

 
Apolipoprotein B, the essential protein core component of atherogenic lipoproteins, exists in two forms.¹ Apo B-100, a 512-kDa protein, is the integral structural protein of VLDLs, which are secreted by the liver.^{1 2} After triglyceride hydrolysis, most of the VLDL remnants are rapidly taken up by the liver, but some are further metabolized into LDLs that contain apo B-100 as the only apoprotein.^{1 2} Elevated plasma levels of LDL represent a major condition for the development of atherosclerosis and coronary artery disease.³ The second apo B form, apo B-48, consists of the amino-terminal 48% of the full-length apo B-100 protein and is the essential protein component of chylomicrons secreted by the intestine.⁴ The generation of apo B-48 is due to mRNA editing, ie, a specific posttranscriptional base change from C to U in the apo B mRNA that creates a premature translational stop codon (UAA) from a glutamine codon (CAA).^{4 5 6} In humans and many other mammalian species, apo B mRNA is extensively edited in the intestine, which synthesizes only apo B-48, whereas the apo B mRNA remains completely unedited in the liver, leading to the exclusive synthesis of apo B-100.⁷ In some mammals such as the dog, horse, rat, and mouse, apo B mRNA is also edited in the liver, and the hepatic secretion of apo B-48–containing VLDL is accompanied by low levels of apo B-100–containing LDL.⁷ Thus, editing of apo B mRNA in the liver appears to be a genetic mechanism to limit the plasma levels of atherogenic lipoproteins.⁷

Editing of apo B mRNA is mediated by a multicomponent enzyme complex termed apo B mRNA editing enzyme.^{4 8 9 10} Its catalytic subunit APOBEC-1 (apo B editing catalytic polypeptide 1) is a cytidine deaminase with a novel RNA binding motif.^{11 12 13 14 15 16} APOBEC-1 requires additional "auxiliary" proteins for efficient apo B mRNA editing.^{11 12 13 14 15 16} In mice and rats, APOBEC-1 is expressed in many organs, the highest expression being found in the small intestine, spleen, and liver, followed by the kidney, lung, skeletal muscle, and heart.^{11 17} Editing of apo B mRNA in the rat and mouse liver can be modulated by developmental, hormonal, or nutritional stimuli, most of which apparently do not modulate the expression of APOBEC-1.^{18 19 20 21 22 23 24 25} Thus, apo B mRNA editing in the rat or mouse liver is dependent on both the abundance of APOBEC-1 and other trans-acting factors yet to be conclusively identified.^{26 27 28} In humans and rabbits, which do not edit hepatic apo B mRNA, APOBEC-1 is mainly expressed in the small intestine and, to some extent, in the colon (rabbit) but in no other organ.^{14 29} The lack of APOBEC-1 limits the editing of apo B mRNA in the liver of mammals that do not edit hepatic apo B mRNA.

Adenovirus-mediated expression of APOBEC-1 in the rabbit liver in vivo leads to editing of hepatic apo B mRNA and consequently to a substantial reduction of circulating plasma LDL levels in normal and Watanabe heritable hyperlipidemic rabbits.^{30 31} In mice transgenic for human apo(B) and human apo(a), hepatic overexpression of APOBEC-1 by adenovirus-mediated gene transfer lowers Lp(a) levels in plasma.³² Thus, induction of apo B mRNA editing in the human liver appears to be a promising genetic approach to reduce elevated plasma levels of the atherogenic LDL or Lp(a).³³ In APOBEC-1 transgenic mice and rabbits, however, in which APOBEC-1 expression was driven by the promoter of the human apo E gene, hepatocellular dysplasia and carcinoma developed in nearly every animal.³⁴ Naturally, APOBEC-1 can be expressed in the liver of horses, dogs, rats, and mice without obvious oncogenic effects.^{7 11 17} The molecular basis for the hepatic expression of APOBEC-1 in these mammalian species is therefore of great interest. In the current investigation, the genomic organization of the rat APOBEC-1 gene and its expression pattern were studied. Two distinct promoters induce APOBEC-1 expression in the rat liver and intestine. The upstream, liver-specific promoter of the rat APOBEC-1 gene has the capacity to induce APOBEC-1 expression and apo B mRNA editing in human hepatoma cells.

	Methods

Top Abstract Introduction Methods Results Discussion References

 
5' RACE and RT-PCR of APOBEC-1 mRNA
Total RNA was prepared from rat liver and intestine by CsCl density gradient ultracentrifugation essentially as described previously.^{7 30} Poly(A)⁺ RNA was prepared from total RNA by using oligo(dT) spin columns (Oligotex) essentially as described by the manufacturer (Qiagen). First- and second-strand cDNA synthesis with subsequent adapter ligation was performed with a commercially available kit (Marathon cDNA amplification kit, Clontech). In brief, first-strand cDNA was synthesized from 1 µg poly(A⁺) RNA at 42°C for 1 hour by using Moloney murine leukemia virus reverse transcriptase and a modified lock-docking oligo(dT) primer. For second-strand synthesis, the cDNA was incubated for 1.5 hours at 16°C with Escherichia coli DNA polymerase I, E coli DNA ligase, and Rnase H. Subsequently, T4 DNA polymerase was added, and the incubation was continued for 45 minutes. After phenol-chloroform extraction and ethanol precipitation, the double-stranded cDNA was ligated to the provided adapters (Marathon cDNA adapter) at 16°C for 16 hours by using T4 DNA ligase. The 5' ends of APOBEC-1 mRNA were amplified from the adapter-ligated cDNA pool by 5' RACE-PCR using Advantage KlenTaq-1 polymerase, the adapter-specific primer AP1, and 1 of the following rat APOBEC-1 antisense oligonucleotides: DA113 (+16 to -3 from ATG: CTGTCTCGGAACTCATCTT), DA115 (-4 to -31 from ATG: GCTCTCTGTGTCTCTCGACTCCTTCCTC), and DA1115 (-32 to -61 from translation start site ATG: GGGAGCTTATGTTGCGGCGATGGCTGCTGC). Thermal cycling was performed in a Perkin-Elmer DNA thermal cycler 480 using the following touchdown program: 1 minute at 94°C, 5 cycles (94°C for 30 seconds, 72°C for 4 minutes), 5 cycles (94°C for 30 seconds, 70°C for 4 minutes), and 27 cycles (94°C for 20 seconds, 67°C for 4 minutes). The products of the 5' RACE were analyzed on 2% agarose gels stained with ethidium bromide. The generated bands were excised from the gel, purified on QIAuick columns (Qiagen), and ligated into the pGEM-T vector (Promega). Plasmid DNA of individual recombinant clones was sequenced on an automated sequencer using the ABI PRISM dye terminator cycle sequencing ready reaction kit (Perkin-Elmer) with a vector-specific primer. RT-PCR of APOBEC-1 mRNA from rat liver and intestine was performed by using the antisense primer APOBEC III (CAGATGGGGGTACCTTGGCCAATGAGC, +519 to +493) and the sense primer APOBEC I (GAGGAAGGAGTCCAGAGACACAGAGAGC, -31 to -4 in the 5' UTR) or the antisense primer APOBEC V (TCCCAGAAGTCATTTCAACCCTGT, +699 to +670, spanning the stop codon TGA) and the sense primer APOBEC IV (GCTCATTGGCCAAGGTACCCCCATCTG, +493 to +519) essentially as described.³⁰

Isolation and Characterization of Rat APOBEC-1 Genomic Clones
A rat genomic EMBL-3 SP6/T7 library, prepared from adult male Sprague-Dawley rats, was purchased from Clontech (catalog No. RL1022). Recombinant phages (4x10⁶)were screened by hybridization with a full-length APOBEC-1 cDNA as described previously.³⁰ One hybridizing -clone was plaque purified for 3 rounds. Phage particles were purified from plaque lysate by using a rabbit polyclonal antibody, and phage DNA was purified on glass-milk (Lambda-Trap Plus, Clontech). The -DNA was digested with SacI, and the resulting restriction fragments of 6.2, 4.0, 3.8, and 1.2 kb were cloned in SacI-digested pT7/T3 (Pharmacia). The resulting recombinant plasmids were hybridized with rat APOBEC-1 oligonucleotides to align them to the cDNA sequence. The exon-intron junctions were sequenced by using the ABI PRISM dye terminator cycle sequencing ready reaction kit (Perkin-Elmer) with the following primers: DA113 (3' splice acceptor in exon 2); APOBEC I (5' splice donor in exon 2); DA17 (TTCCCACATGATTTGTAGATTTGGGA GACT, sense primer in intron 2 for sequencing exon 3 and its junctions); APOBEC XIV (AGTCCTTGCCGATTTCGAGGATCTGCG TG, antisense +392 to +363 for the 3' splice acceptor site of exon 4); APOBEC XVIII (AGCCGATACCCCCATGTAACTCTGTTT, sense +313 to +339 for the 5' splice donor site in exon 4); APOBEC III (for the 3' splice acceptor site in exon 5); APOBEC IV (for the 5' splice donor site in exon 5); APOBEC V (for the 3' splice acceptor site in exon 6), APOBEC XII (CACATCCTGTGGGCCACAGGGTTGAAATGA, sense +661 to +690 for sequencing exon 6); and APOBEC XIIa [ATCCGGCCTGTCATTCCTGGGTAGCATGCTA, sense; 3' position of poly(A) site for sequencing the 3' end of the 6.2-kb SacI fragment, ie, 3' end of the 15.2-kb genomic fragment]. The 5' flanking region of exon 1, the 5' splice donor site of exon 1, and the 5' flanking region of exon 2 were isolated by PCR "walking" using the commercially available PromoterFinder DNA walking kit (Clontech) with Clontech's Advantage polymerase mix exactly as recommended. The following gene-specific primers were used: for amplification of the 5' flanking region of exon 1, DAex13 for the primary PCR (TTCTTCCTGATCTAATTTAAGGGAGGAGC, antisense exon 1, -158 to -187 from ATG) and DAex15 for the nested PCR (GCCTTTCTGCCTTCAAAATCC AACTCTCAT, antisense exon 1, -188 to -217 from ATG); for amplification of the 5' splice donor site of exon 1 and intron 1, DAex1S51 for the primary PCR (CCTACTCCCGCTAC AGAACCACTG TGCCC, sense in exon 1, -246 to -218 from ATG) and DAex1S2 (ATGAGAGTTGGATTTTGAAGGCAGAAAGGC, sense in exon 1, -217 to -188 from ATG); for amplification of the 5' flanking region of exon 2, DA113 for primary PCR and DA115 for nested PCR; and for walking within intron 1 toward exon 1, DA43 (AGTGTATTAATGA GACTAAGAACTCACAA) for primary PCR and DA45 (CAGGGAACCCTAAATCCTCTG TAACACTCT) for nested PCR. The primary PCR was performed in a Perkin-Elmer DNA thermal cycler 480 using the following 2-step cycle parameters: 7 cycles (94°C for 25 seconds, 72°C for 4 minutes) and 32 cycles (94°C for 25 seconds, 67°C for 4 minutes), followed by an additional 4 minutes at 67°C. For nested PCR the following 2-step cycle parameters were used: 5 cycles (94°C for 25 seconds, 72°C for 4 minutes) and 22 cycles (94°C for 25 seconds, 67°C for 4 minutes), followed by an additional 4 minutes at 67°C. The amplified genomic fragments were isolated from 0.8% agarose gels and cloned into the pGEM-T vector. A portion (2.5 kb) of the 5' flanking region of exon 1, the exon-intron junction of exon 1, and 2.0 kb of the 5' flanking region of exon 2 were sequenced by using the ABI PRISM dye terminator cycle sequencing ready reaction kit (Perkin-Elmer). For sequencing of the 5' flanking region of exon 1, the following primers were used: DAex17 (GCCAAGTTACCCCACTAATTAGGATTTAGGG) and DAex16 (CCCAGGAGCA GTGCCGCATCCACTTTCTTC). For sequencing of the 5' flanking region of exon 2, the following primers were synthesized: DA117 (CTCCCTCAAGGTCTAAGTCTCCCACTAG GG) and DA23 (ACTCACTATA GGGCTCGAGCGGC).

Primer Extension Analysis of Transcriptional Start Sites in APOBEC-1 mRNA From Rat Liver and Intestine
Five micrograms of total RNA from rat liver and intestine or tRNA was denatured for 10 minutes at 70°C, annealed to radiolabeled oligonucleotide DAex15 or DA115 for 30 minutes at 42°C, and extended with 10 U avian myeloblastosis reverse transcriptase for 60 minutes. After addition of 4 U RNase A, the primer extension products were separated on an 8% polyacrylamide–7 mol/L urea sequencing gel. The DNA sequence of exon 1 and exon 2 was determined by using Sequenase (Amersham) and radiolabeled oligonucleotides DAex15 and DA115. The sequencing reactions were separated on the same sequencing gel along with the primer extension products to map the transcriptional start sites.

RNase Protection Assay of Transcriptional Start Sites and Splice Variants of APOBEC-1 mRNA in Rat Organs
A rat genomic DNA fragment consisting of 158 nucleotides spanning 68 bp of the upstream promoter region and 90 bp of transcribed sequences of exon 1 was amplified by PCR with the oligonucleotides MS5 (GCGAGCTCGCCTGACAGCAA ACCCCAGCCA, -68 to -39 from the transcriptional start site in exon 1) and DAex15. A rat genomic fragment of 248 nucleotides spanning 94 bp of the promoter region of exon 2 and 154 bp of the transcribed sequences of exon 2 was amplified by PCR with the oligonucleotides MS16 (GCGAGCTCTCTCCAGGGAGGACG GAAATC, -166 to -148 from the transcriptional start site in exon 2) and DA115. The two PCR products were cloned into the pGEM-T Easy vector (Promega) and sequenced to identify their orientation. Two micrograms of plasmid DNA linearized with EcoRI was transcribed for 30 minutes at 37°C with 5 U T7 RNA polymerase in the presence of 500 µmol/L each ATP, GTP, and CTP and 12 µmol/L of [-³²P]UTP (specific activity, 800 Ci/mmol) to generate an antisense transcript of 228 bases of exon 1 and of 322 bases of exon 2. After digestion of plasmid DNA with 10 U RNase-free DNase I (Pharmacia) for 15 minutes at 37°C, the RNA transcripts were purified by phenol-chloroform extraction and S-300 microspin columns (Pharmacia). Total RNA was purified from 10 rat organs (liver, intestine, stomach, colon, spleen, lung, kidney, brain, heart, and skeletal muscle) by using Tri-Reagent and following the manufacturer`s (Molecular Research Center Inc) protocol exactly. Fifty micrograms of total RNA from each organ was coprecipitated with 4x10⁴ counts per minute of -³²P–labeled RNA probe and hybridized for 18 hours at 42°C in 20 µL of solution A of the ribonuclease protection assay II kit (Ambion Inc). After 200 µL of solution B containing 1:50–diluted RNase A and RNase T1 was added, the incubation was continued for 30 minutes at 37°C. After addition of 300 µL of solution D, the undigested RNA fragments were precipitated for 30 minutes at -20°C and subsequently resolved on a 6% denaturing polyacrylamide sequencing gel. A DNA molecular-weight marker (Boehringer Mannheim 100-bp ladder) and oligonucleotide DD3 with a length of 35 bases,³⁰ which had been labeled by using T4 polynucleotide kinase and [-³²P]ATP, were analyzed on the same gel. The semilogarithmic graph of molecular weight versus length of migration gave a straight line, from which the approximate lengths of the protected RNA fragments were deduced. For improved precipitation of smaller protected fragments, 250 µL of 100% ethanol was added to the precipitation mix, and the precipitation was extended to 1 hour at -20°C.

Generation of APOBEC-1 Promoter–Luciferase Reporter Constructs: 5' Flanking Region of Exon 1
Two fragments of the 5' flanking region of exon 1 of 2.5 and 2.0 kb that had been isolated by genomic PCR walking and cloned into the pGEM-T vector were isolated by SalI digestion and cloned into the XhoI site of pGL3-basic. The resulting promoter constructs, designated LP1 and LP2, were sequenced to confirm their orientation. LP1 extended from nucleotide position -2535 from the transcriptional start site to +92 (corresponding to position -188 from the translational start site ATG) and LP2 from -2089 to +92. Serial deletions were constructed by PCR by using the universal downstream oligonucleotide MS1 (CACAAGCTTCCTGACCAGAGAAAGCTGAGG, APOBEC-1 5' UTR antisense; -252 to -279 from translational start site ATG) and the following upstream primers: for LP3, MS2 (GAAGTCGACTTCCATTTTGATAGAGGTCAG, position -930 to -901 from the transcriptional start site); for LP4, MS3 (ACCGTCGACTGTTTTTCAGTCA TCTGGACT, position -501 to -472); for LP5, MS4 (AAGCGGCTGGATGTCGACAACA ACAATAAC, -155 to -126); for LP6, MS5 (GCGAGCTCGCCTGACAGCAAACCC CAGCCA, -68 to -39); and for LP7, MS6 (GCGAGCTCTTAAAGCCGAACCAGCAA, -25 to +1). The resulting promoter fragments were digested with HindIII and SalI (LP3 to LP5) or SacI (LP6 and LP7) and cloned into the respective sites of pGL3-basic. The generated constructs were sequenced to confirm the orientation and sequence. For the 5' flanking region of exon 2, 2.0 kb of the 5' flanking region of exon 2, which had been isolated by genomic PCR walking and cloned into the pGEM-T vector, was cloned into the XhoI site of pGL3-basic. The resulting construct, designated IP1, was sequenced to confirm its orientation and sequence. IP1 extended from nucleotide position -2070 from the transcriptional start site to +81 (corresponding to position -4 in the 5' UTR from the translational start site ATG). Serial deletions were constructed by PCR by using the universal downstream oligonucleotide MS11 (TGTAAGCTTCCTTCCTCGGGAGCTTATG, APOBEC-1 5' UTR antisense; -23 to -42 from the translational start site ATG, corresponding to +44 to +63 from the transcriptional start site) and the following upstream primers: for IP2, MS12 (TTTGTCGACTGATGATTGGTGTGGGAGTGC, -1173 to -1144 from the transcriptional start site); for IP3, MS13 (GAAGTCGACCCAACGTTCTGGCAACAGAAG, -587 to -558 from the transcriptional start site); for IP4, MS14 (CCTGTCGACTTCTTAGATATAGGAGCT, -324 to -298); for IP5, MS15 (GCGAGCTCTGGAATTTAACATCATCGATG, -311 to -293); for IP6, MS16 (GCGAGCTCTCTCC AGGGAGGACGGAAATC, -166 to -148); and for IP7, MS17 (TGGAGCTCAGGAAT GAGTCATCCTGT, -57 to -32). The resulting fragments were digested with HindIII and SalI (IP2 to IP4) or SacI (IP5 to IP7) and cloned into pGL3-basic. The generated constructs were sequenced in their entirety to confirm orientation and sequence. For IP mutants, the fragment -2070 to -1145 and the fragment -2070 to - 558 of the 5' flanking region of exon 2 were amplified by PCR using a vector-specific SP6 primer (ATTTAGGTGACACTATAGAATACTCAAG) and the antisense downstream primer MS22 (GCGAGCTCGCACTCCCACACCAATCATCAGTC, -1145 to -1173 from the transcriptional start site in exon 2) or MS23 (GCGAGCTCTCTTCTGTTGCCAGAACGTTGGG, -557 to -582 from the transcriptional start site in exon 2). The 2 PCR products were digested with SacI and fused directly in front of the promoter region of the construct IP5 by using the unique SacI site. The proper orientation of insertion was determined by DNA sequencing. These two constructs were designated IP1144–216 and IP556–216. For IPxLP mutants, the fragments -1167 to -266, -581 to -266, and -318 to -266 of the 5' flanking region of exon 2 were amplified by PCR using the universal antisense downstream primer MS22 (GCGAGCTCGCACTCCCACACCAATCATCAGTC, -266 to -289 from the transcriptional start site in exon 2) and either MS12, MS13, or MS14. The resulting PCR products were cloned into the pGEM-T easy vector. Fragments of the proper orientation were excised from the vector with SacI and NheI, gel purified, and ligated in front of the promoter fragment of LP5 by using the unique sites SacI and NheI in LP5. The resulting promoter constructs were designated IP(1167–266)-LP5, IP(581–266)-LP5, and IP(318–266)-LP5.

Transfection of Promoter Constructs
Human colonic carcinoma Caco-2 cells, kindly provided by Dr C. Courtelle, London, UK, were grown in Dulbecco's modified Eagle's medium containing 2 mmol/L glutamine and 20% FCS. Mouse fibroblast NIH/3T3 cells, kindly provided by Dr M. Spiess, Basel, Switzerland, and the human hepatoma HuH-7 cells, kindly provided by Dr J. Chowdhury, New York, NY, were grown in Dulbecco's modified Eagle's medium containing 2 mmol/L glutamine and 10% FCS. Cells (5x10⁵) were seeded per well of a 6-well plate. After 24 hours the cells were cotransfected with 2 µg/mL reporter plasmid and 0.1 µg/mL CMV–ß-gal by using the Lipofectin procedure (GIBCO) essentially as described.³⁰ In brief, cells were incubated for 6 hours with 1 mL of Optimem-1 (GIBCO) containing 2 µg reporter plasmid, 0.1 µg CMV–ß-gal as the internal-standard plasmid, and 15 µL Lipofectin. Subsequently, cells were grown in normal medium for 48 hours, washed twice with PBS, and lysed in 250 µL of reporter lysis buffer (Promega). The cell lysate was freeze-thawed and centrifuged in a Microfuge. After addition of 50 µL luciferase assay buffer (Promega), luciferase activity was determined in 10 µL of soluble cell extract by using a luminometer (Berthold Microlumat LB 96 P). ß-Gal activity was measured by luminometry using the Galacto-Light reporter gene assay for ß-gal (Tropix, BL300G). The luciferase activity of each well was normalized to ß-gal activity. All assays were performed in triplicate. At least 3 independent experiments, each in triplicate assay, were performed for each determination. The mean±SEM was calculated and expressed as relative luciferase activity, based on the luciferase activity of cells transfected with the promoterless plasmid pGL3-basic and the internal-standard plasmid CMV–ß-gal.

Construction of an APOBEC-1 Minigene
The XbaI-BamHI fragment of pSVL–APOBEC-1³⁰ containing the open reading frame of APOBEC-1 was cloned into the XbaI-BamHI site of pT7T3 (Pharmacia). Subsequently, the BamHI fragment of pSVK3 (Pharmacia), which contains the SV40 small T-antigen splice site and the SV40 early poly(A) site, was ligated into the BamHI site. Subsequently, the SalI fragment of LP1 containing 2.5 kb of the 5' flanking region of exon 1 was ligated into the SalI site. The orientation of the promoter fragment was confirmed by DNA sequencing. The resulting plasmid p–APOBEC-1 contains 2.5 kb of the 5' flanking region of exon 1, the complete open reading frame of APOBEC-1 followed by the SV40 small T-antigen splice site, and the SV40 early poly(A) site in the plasmid backbone of pT7T3. pSVL–APOBEC-1 and p–APOBEC-1 were transfected in human hepatoma HuH-7 cells and Hep G2 cells essentially as described.³⁰ Apo B mRNA was amplified by RT-PCR and assayed for editing by primer extension analysis essentially as described.^{7 30}

	Results

Top Abstract Introduction Methods Results Discussion References

 
5' RACE of APOBEC-1 mRNA From Rat Liver and Intestine
The 5' UTR of APOBEC-1 mRNA from liver and intestine was amplified by 5' RACE. Two major bands of 260 and 210 bp were generated from hepatic APOBEC-1 mRNA by using the antisense oligonucleotide DA1115 (Figure 1A, lane 1). 5' RACE of intestinal APOBEC-1 mRNA using the antisense oligonucleotide DA113 resulted in 3 bands of 120, 100, and 80 bp (Figure 1A, lane 2). Control 5' RACE without RT did not generate any product (Figure 1A, lanes 3 and 4). The 5' RACE products from liver and intestine were purified from a 2% agarose gel and cloned. The inserts of at least 6 individual recombinant clones for each purified 5' RACE product were sequenced. The 5' UTR of APOBEC-1 mRNA from rat liver contained unique sequences at its 5' end that were lacking in the 5' UTR of rat intestinal APOBEC-1 mRNA (Figure 1B). The 5' ends of both the 260- and the 210-bp 5' RACE products were identical. Fifteen recombinant clones of the 5' RACE products from rat liver ended within a region of 5 nucleotides (Figure 1B, one asterisk). The larger 5' RACE product of APOBEC-1 mRNA from rat liver contained an additional insertion of 51 nucleotides in comparison with the smaller 5' RACE product from rat liver (Figure 1B, boxed sequence). The 5' RACE products of APOBEC-1 mRNA from rat intestine contained as many as 85 nucleotides of the most 3' region of the hepatic 5' UTR of APOBEC-1 (Figure 1B, underlined sequence). The three 5' RACE products of APOBEC-1 mRNA from rat intestine differed by 20 nucleotides each at their 5' ends (Figure 1B; 5' ends of the products are indicated by 2, 3, and 4 asterisks). Similar results were obtained in 3 additional independent 5' RACE reactions using the primer DA115 in addition to DA1115 and DA113. RT-PCR of rat intestinal total RNA was performed using the primers APOBEC III and APOBEC I, or APOBEC V and APOBEC IV. These 2 RT-PCR products span the open reading frame of APOBEC-1. Both RT-PCRs generated a single major band that was isolated and cloned. DNA sequencing confirmed the rat APOBEC-1 cDNA sequence in rat liver and intestine as reported by Teng et al,¹¹ without evidence for additional alternate processing of APOBEC-1 mRNA.

View larger version (42K): [in this window] [in a new window]   Figure 1. 5' RACE of APOBEC-1 mRNA from rat liver and intestine. 5' RACE of APOBEC-1 mRNA was performed using total RNA from rat liver and intestine. Resulting 5' RACE products and their controls without reverse transcriptase were separated on 2% agarose gel. Photograph of ethidium bromide–stained gel is shown together with a synopsis of nucleotide sequences of the various 5' RACE products. The following symbols were used to indicate 5' ends of the 5' RACE products and splice variants of hepatic APOBEC-1 mRNA: *5' end of 5' RACE of APOBEC-1 from rat liver; **5' end of 5' RACE of APOBEC-1 from rat intestine (120-bp product); ***5' end of 5' RACE of APOBEC-1 from rat intestine (100-bp product); ****5' end of 5' RACE of APOBEC-1 from rat intestine (80-bp product), boxed insertion in the longer 5' RACE product of APOBEC-1 mRNA from rat liver; +proximal 3' splice acceptor site of exon 2; and ++distal 3' splice acceptor site of exon 2, underlined common 5' UTR of APOBEC-1 mRNA in rat liver and intestine ATG translational start codon of APOBEC-1.

Characterization of the Rat APOBEC-1 Gene
A -clone was isolated from an EMBL3 SP6/T7 rat genomic library that encoded a 15.2-kb genomic fragment of the rat APOBEC-1 gene (Figure 2A). Digestion with SacI generated 4 restriction fragments of 6.2, 4.0, 3.8, and 1.2 kb that were cloned and aligned with the APOBEC-1 cDNA sequence. The exon-intron junctions were determined by DNA sequencing. The isolated genomic fragment contained 5 exons and 0.7 kb of the 5' flanking region of the most 5'-located exon. This exon contained the translational start site for APOBEC-1 and the common 5' UTR of APOBEC-1 mRNA from both rat liver and intestine. The additional sequences of the 5' UTR of rat hepatic APOBEC-1 mRNA, however, could not be aligned with this 15.2-kb rat genomic fragment. The 5' flanking region of the 5' UTR of APOBEC-1 from rat liver and its 5' splice donor site were isolated by PCR walking. The rat APOBEC-1 gene contains an additional exon in the 5' position of exon 2, resulting in a total of 6 exons. Figure 2A illustrates the genomic organization of the rat APOBEC-1 gene, the isolated -clone, its 4 SacI fragments, and the 4 PCR walks that span the 5' end of the rat APOBEC-1 gene. In Figure 2B, the exon-intron organization and the exon-intron junctions are summarized. Exon 1 has a size of 122 nucleotides. Exon 2 contains two alternative 3' splice acceptor sites spaced by 51 nucleotides and thus encodes either 173 or 122 nucleotides, as was demonstrated by 5' RACE (Figure 1). Exon 3 contains 28 nucleotides, while exons 4 and 5 span 398 and 119 nucleotides, respectively. Exon 6 encodes 230 nucleotides up to the poly(A) signal.

View larger version (28K): [in this window] [in a new window]   Figure 2. Illustration of genomic organization of the rat APOBEC-1 gene. A, -Clone encoding 15.2 kb of the rat APOBEC-1 gene was isolated from a rat genomic library. Four SacI fragments were subcloned and the exon-intron organization of the APOBEC-1 gene was determined. The 5' end of the rat APOBEC-1 gene was isolated by genomic PCR walking. Genomic organization of the rat APOBEC-1 gene is illustrated. B, Sequences at exon-intron-junctions, exon sizes, and intron sizes are listed.

Approximately 2.5 kb of the 5' flanking region of exon 1 and 3.0 kb of the 5' flanking region of exon 2 were isolated by PCR walking and sequenced; the sequences are available from the European Molecular Biology Organization nucleotide sequence database under the accession numbers AJ006695 and AJ006696 (Figure 3A and 3B). The 5' flanking region of exon 1 contains TAAA as a variant TATA box at -16 to -13 (from the transcriptional start site) and a CACT motif as a putative CAT box at -101 to -98 (Figure 3A). Within the 5' flanking region of exon 1, a variety of consensus sequences for potential interactions with trans-acting factors were identified by computer alignment and are illustrated in Figure 3A. Notably, motifs for hepatic transcription factors were identified: liver activator protein at -964 to -956, HNF3/AP1 at -706 to -698, and C/EBP at -303 to -299 and -837 to -833. The 5' flanking region of exon 2 contains two CCACT motifs as variant CAT boxes at -131 to -126 and -263 to -259. Within the typical distance of the transcriptional start site (20 to 30 nucleotides), a classic TATA box, however, is lacking. A putative TATA box, however, is located upstream at position -307 to -304. A conserved downstream element of many TATA-less promoters, the MED-1 motif GCTCCC,³⁵ is located within exon 2 at +49 to +53 from the transcriptional start site. The consensus sequences for potential trans-acting factors within the 5' flanking region of exon 2 as well as the MED-1 sequence in exon 2 are illustrated (Figure 3B).

View larger version (63K): [in this window] [in a new window]   Figure 3. Rat APOBEC-1 promoter sequences. A, Sequences of upstream liver-specific promoter LP. Nucleotide sequence of 5' flanking region of exon 1 (1.3 kb) is illustrated. Consensus sequences for putative trans-acting factors, potential CAT boxes, and the variant TATA box within the 5' flanking region of exon 1 were identified by computer search and are underlined. Transcribed sequences of exon 1 are marked by uppercase letters. B, Sequences of downstream intestine-specific promoter IP. Nucleotide sequence of 1.3 kb of the 5' flanking region of exon 2 is illustrated. Consensus sequences for potential trans-acting factors, CAT box of the 5' flanking region of exon 2, and MED-1 motif within exon 2 were identified by computer search and are underlined. Transcribed sequences of exon 2 are in uppercase letters.

Primer Extension Analysis of the Transcriptional Start Sites for APOBEC-1 in Rat Liver and Intestine
Using radiolabeled antisense oligonucleotide DAex15, which corresponds to the -188 to -217 region of the 5' UTR of rat hepatic APOBEC-1 encoded by exon 1, we generated a discrete primer extension product exclusively from total hepatic RNA but not from total intestinal RNA or tRNA control (Figure 4A). The DNA sequence analysis of genomic DNA with radiolabeled DAex15 was separated side by side along with the primer extension assay on the same sequencing gel (Figure 4A). By alignment with the genomic sequence, the putative transcriptional start site of APOBEC-1 mRNA in rat liver was mapped to nucleotide position -279 from the translational start codon ATG (Figure 4A and 4B). This putative transcriptional start site maps 22 nucleotides upstream from the 5' RACE of rat hepatic APOBEC-1 mRNA that ended at nucleotide position -257 (Figure 4B). Using radiolabeled oligonucleotide DA115, which corresponds to the -4 to -31 region of the 5' UTR of APOBEC-1 encoded by exon 2, we generated a major primer extension product from intestinal RNA only and not from hepatic RNA or tRNA (Figure 4C). A minor extension product was visible in both intestinal and hepatic RNA but not in the tRNA control. Alignment of the major extension product from rat intestinal APOBEC-1 mRNA with the DNA sequence of exon 2 mapped the putative transcriptional start site for APOBEC-1 mRNA in rat intestine to nucleotide position -77 from ATG (Figure 4C and 4D). In comparison, the 2 larger 5' RACE products of APOBEC-1 mRNA from rat intestine extended to nucleotide positions -85 and -64 from ATG, respectively (Figure 4D).

View larger version (57K): [in this window] [in a new window]   Figure 4. Primer extension analysis for determination of transcriptional start sites. A, Transcriptional start sites within exon 1 were determined by primer extension analysis in total RNA from rat liver and intestine and tRNA control by using radiolabeled oligonucleotide DAex15. DNA sequence analysis of exon 1, determined from cloned rat genomic DNA by using radiolabeled DAex15, was separated on 8% polyacrylamide–7 mol/L urea sequencing gel side by side with primer extension products. DNA sequence of exon 1 is shown as the complementary sense sequence. B, Transcriptional start sites as determined by 5' RACE and primer extension analysis are illustrated together with genomic sequence of exon 1. C, Transcriptional start sites within exon 2 were determined by primer extension analysis in total RNA from rat liver and intestine and tRNA control by using radiolabeled oligonucleotide DA115. DNA sequence analysis of exon 2, determined from cloned rat genomic DNA using radiolabeled oligonucleotide DA115, was separated on 8% polyacrylamide–7 mol/L urea sequencing gel side by side with primer extension products. DNA sequence of exon 2 is shown as the complementary sense sequence. D, Transcriptional start sites as determined by 5' RACE and primer extension analysis are illustrated together with genomic sequence of exon 2.

Ribonuclease Protection Assay for Transcriptional Start Sites of APOBEC-1 in Various Rat Organs
The transcriptional start sites of APOBEC-1 in exons 1 and 2 and the splice sites in exon 2 were further characterized by ribonuclease protection assay. An antisense RNA probe of exon 1 was transcribed containing a 158-bp genomic fragment that spans approximately the distal 70 bp of the promoter and 90 bp of the transcribed sequences of exon 1. This transcript has a length of 232 nucleotides owing to transcribed vector sequences (Figure 5A). With the use of this probe, an RNA fragment of 90 bases was protected in total RNA from the liver, intestine, colon, spleen, lung, kidney, heart, and skeletal muscle but not in the stomach or brain (Figure 5A). The strongest signals were generated in the spleen, lung, and kidney. In the liver and intestine, relatively weak protection was observed (Figure 5A). With the use of total RNA from the heart, the protected fragment was slightly larger (Figure 5A). No fragments of these sizes were protected when an equal amount of yeast total RNA was used (Figure 5A). However, 2 large bands were generated in the same quantities in both yeast total RNA and RNA from all organs tested. Increased RNase concentration or extended digestion time did not diminish these 2 large fragments in yeast or rat RNA. Identical results were obtained in 4 independent ribonuclease protection assays using various preparations of this antisense RNA probe encoded by exon 1 of rat APOBEC-1.

View larger version (40K): [in this window] [in a new window]   Figure 5. Ribonuclease protection assay for determination of transcriptional start sites. A, For exon 1, its antisense RNA was synthesized in vitro in the presence of 12 µmol/L [-32P]UTP from a genomic DNA fragment that was generated by PCR with oligonucleotides MS5 and DAex15 and which spans the distal 70 bp of the promoter region and 90 bp of transcribed sequences of exon 1. This antisense RNA (4x104 cpm) was hybridized for 18 hours at 42°C with 50 µg of yeast total RNA or 50 µg total RNA from rat organs (liver, intestine, stomach, colon, spleen, lung, kidney, brain, heart, or muscle). Subsequently, RNA was digested in buffer B with RNase A and RNase T1 using a commercially available RNase protection assay kit. Protected RNA fragments were resolved on denaturing 6% polyacrylamide sequencing gel. As a control for integrity of the RNA probe, radiolabeled probe was hybridized with 50 µg of yeast total RNA and incubated in buffer B without addition of RNases. Portion (1/50) of this RNA control was loaded onto the gel. After electrophoresis, the gel was dried under vacuum and exposed to autoradiography. Representative autoradiogram is shown together with an illustration of RNA antisense probe and protected fragment. Migration of kinased DNA marker and approximate length of protected fragment are indicated. B and C, For exon 2, its antisense RNA was synthesized in vitro in the presence of 12 µmol/L [-32P]UTP from a genomic DNA fragment which was generated by PCR with oligonucleotides MS16 and DA115 and which spans 94 bp of the promoter region of exon 2 and 154 bp of transcribed sequences of exon 2. This antisense RNA (4x104 cpm) was hybridized for 18 hours at 42°C with 50 µg of yeast total RNA or 50 µg total RNA from rat organs (liver, intestine, stomach, colon, spleen, lung, kidney, brain, heart, or muscle). RNA was digested in buffer B with RNase A and RNase T1 using a commercially available RNase protection assay kit. After precipitation, protected RNA fragments were resolved on denaturing 6% polyacrylamide sequencing gel. As a control for integrity of the RNA probe, radiolabeled probe was hybridized with 50 µg of yeast total RNA and incubated in buffer B without addition of RNases. Portion (1/50) of this control RNA probe was loaded onto the gel. The dried gel was exposed to autoradiography. Representative autoradiogram is shown together with an illustration of RNA antisense probe and protected fragments. Migration of kinased DNA marker and approximate lengths of protected fragments are indicated (B). For improved precipitation of smaller fragments, RNA was precipitated in 95% ethanol, and protected fragments were resolved on denaturing 8% polyacrylamide sequencing gel (C). Representative autoradiogram is shown together with an illustration of RNA antisense probe and protected fragments corresponding to the 3 distal transcriptional start sites in exon 2. Migration of kinased DNA marker and approximate lengths of protected fragments are indicated.

For exon 2, an antisense RNA probe was synthesized in vitro containing 94 bp of the distal promoter region of exon 2 and 154 bp of transcribed sequences of exon 2. Together with the transcribed multiple cloning site, the transcript has a length of 322 nucleotides. The 248 nucleotides of exon 2 contain both the proximal and distal splice acceptor sites as well as the various transcriptional start sites within exon 2, as deduced from 5' RACE. With the use of this antisense RNA probe, several fragments were protected in total RNA from the liver, intestine, spleen, lung, kidney, and to a lesser degree, the stomach, colon, heart, and skeletal muscle (Figure 5B). The RNA fragment of 154 nucleotides corresponds to the major proximal splice acceptor site of exon 2. The fragment of 144 nucleotides appears to be due to a minor proximal splice acceptor site in exon 2. The fragment of 103 nucleotides represents the distal splice acceptor site in exon 2. These 3 protected fragments that represent splice variants of APOBEC-1 transcripts initiated in exon 1 are present in the spleen, liver, lung, and to a minor extent, the colon, kidney, heart, and skeletal muscle; however, they are abundant in the intestine, where the exon 1–specific antisense transcript was hardly protected (Figure 5A and 5B). In the intestine, an additional RNA fragment of 80 nucleotides was protected (Figure 5B). Moreover, a very faint band of 250 nucleotides was specifically protected in the intestine (Figure 5B). By improved precipitation of smaller RNA fragments and separation on an 8% polyacrylamide sequencing gel, 2 additional protected RNA fragments of 60 and 40 nucleotides, respectively, could be demonstrated exclusively in the intestine (Figure 5C). The 3 fragments of 80, 60, and 40 nucleotides correspond to the 3 intestinal transcriptional start sites as demonstrated by 5' RACE (Figure 1A and 1B). No protected fragment was generated when yeast total RNA was used (Figure 5B and 5C). Identical results, as shown in Figure 5B and 5C, were obtained in 3 other independent ribonuclease protection assays.

Functional Characterization of the 5' Flanking Regions of Exons 1 and 2
The promoter activities of the 5' flanking regions of exons 1 and 2 were investigated in transfection studies with luciferase as a heterologous reporter. Serial deletions of the 5' flanking regions of exons 1 and 2 were constructed by PCR, cloned into the pGL3-basic vector, and sequenced. The constructs of the 5' flanking region of exon 1 were designated LP1 to LP7. LP1 and LP2 ended at nucleotide position +92 from the transcriptional start site, and LP3 to LP7 ended at nucleotide position +24 from the transcriptional start site. The constructs of the 5' flanking region of exon 2, designated IP1 to IP7, extended to nucleotide position +81 (IP1) or +63 (IP2 to IP7) from the intestinal transcriptional start site and therefore contained the MED-1 motif that is located at position +49 to +53.

The results of transfection studies for these promoter constructs are illustrated in Figure 6. LP1 (-2535 to +92) exerted a 2.5-fold stronger promoter activity in HuH-7 cells and a 3-fold stronger promoter activity in NIH/3T3 cells compared with Caco-2 cells (Figure 5A). LP2 (-2089 to +92) and LP3 (-924 to +28) showed a gradual loss of promoter activity in all 3 cell lines. LP4 (-495 to +28) and LP5 (-140 to +28) demonstrated an increase in promoter strength in all 3 cell lines. LP5, with slightly more promoter activity than LP1, still contained the TAAA motif at -16 to -13 as the presumptive variant TATA box and the CACT motif at -101 to -98 as the putative CAT box. LP6 (-63 to +28), which lost the putative CAT box at -101 to -98, had strongly reduced promoter activity. LP7 (-20 to +28), in which the putative TATA box at -16 to -13 was the only remaining promoter element, lost any promoter activity and was similar to the promoterless control construct pGL3-basic. The construct LP1, in which the promoter fragment of LP1 was inserted in the opposite direction in pGL3-basic, exhibited less relative luciferase activity than did pGL3-basic (data not shown).

View larger version (40K): [in this window] [in a new window]   Figure 6. Transfection of APOBEC-1 promoter–luciferase reporter constructs in HuH-7, Caco-2, and NIH/3T3 cells. A, 5' Flanking region of exon 1 was truncated from the 5' end by PCR and cloned into promoterless construct pGL3-basic to generate constructs LP1 to LP7. B, 5' Flanking region of exon 2 was truncated from 5' end by PCR and cloned into promoterless construct pGL3-basic to generate constructs IP1 to IP7. Lengths and positions of various promoter constructs are illustrated. C, Constructs IP1144–216 and IP556–216 were generated by ligating fragments -1144 to -216 and -556 to -216 of the 5'-flanking region of exon 2 (IP) to promoter construct IP5. D, Crossover mutants IP(1167–266)-LP5, IP(581–266)-LP5, and IP(318–266)-LP5 were generated by ligating regions -1167 to -266, -581 to -266, and -318 to -266 of the 5' flanking region of exon 2 (IP) to promoter construct LP5. All constructs were transfected into HuH-7, Caco-2, and NIH/3T3 cells together with CMV–ß-gal. After 48 hours activities of luciferase and ß-gal were determined by luminometry. Activity of luciferase was normalized for activity of ß-gal and expressed as relative luciferase activity based on activity of promoterless construct pGL3-basic, to which the value of 1.0 was assigned. Mean±SEM of at least 3 independent experiments in triplicate assay each are given. Values in C and D represent mean±SEM of 4 additional independent determinations in triplicate assay each, even for constructs IP1, IP5, and LP5.

In contrast, IP1 extending from position -2070 of the transcriptional start site of APOBEC-1 mRNA in rat intestine to position +81 (corresponding to position -4 in the 5' UTR from the translational start codon ATG) exhibited an 4-fold higher promoter activity in human colonic carcinoma Caco-2 cells than in HuH-7 cells (Figure 6B). Truncations of the 5' flanking region of exon 2 in IP2 (-1167 to +63), IP3 (-581 to +63), and IP4 (-318 to +63) had no major effect on promoter activity in Caco-2 or HuH-7 cells. In NIH/3T3 cells, all constructs (IP1 to IP7) had very low promoter activities (<5), close to the activity of the promoterless construct pGL3-basic. In IP5 (-216 to +63) and IP6 (-161 to +63), a 2-fold increase in promoter activity was observed in Caco-2 cells and a 5-fold increase in HuH-7 cells compared with the effect of IP1. The shortest construct IP7 (-52 to +63) demonstrated a 4-fold reduction of promoter activity in Caco-2 cells and nearly a 6-fold reduction in HuH-7 cells compared with the effect of IP6. In Caco-2 cells, IP7 still exhibited a comparably high relative luciferase activity. The construct IP2, in which the promoter fragment of IP2 was inserted in the opposite direction in pGL3-basic, exhibited less relative luciferase activity than did the promoterless construct pGL3-basic (data not shown).

The addition of insulin (10 nmol/L), T3 (50 nmol/L), dexamethasone (100 nmol/L), or glucagon (5 nmol/L) to the cell culture medium of transfected HuH-7 cells did not significantly alter the reporter gene expression of any of the generated promoter constructs (LP1 to LP7 or IP1 to IP7) (data not shown, 3 independent experiments in triplicate assay each).

The regions -1144 to -216 and -556 to -216 were deleted from the 5' flanking region of exon 2 (IP) by fusing the promoter fragments -2070 to -1145 and -2070 to -557 directly in front of the promoter construct IP5, thus generating the constructs IP1144–216 and IP556–216. In addition, the regions -1167 to -266, -581 to -266, and -318 to -266 of the 5' flanking sequences of exon 2 were ligated in front of the promoter construct LP5 to generate the constructs IP(1167–266)-LP5, IP(581–266)-LP5, and IP(318–266)-LP5.

In a second set of experiments, these 5 additional constructs and the basic constructs IP1, IP5, and LP5 were transfected in HuH-7, Caco-2, and NIH/3T3 cells to determine their promoter activities. The mean±SEM of 4 independent experiments are each shown in Figure 6C and 6D. In comparison with IP1, IP1144–216 showed an increased promoter activity similar to that of IP5 in both HuH-7 and Caco-2 cells. This increase was most pronounced in HuH-7 cells, in which IP1144–216 had an 3-fold stronger promoter activity than IP1 (Figure 6C). IP556–216 lost promoter strength in both HuH-7 and Caco-2 cells compared with IP1144–216 and exerted promoter activities similar to those of IP1. In NIH/3T3 cells, all 4 constructs (IP1, IP5, IP1144–216, and IP556–216) had low promoter activities, yet in relative terms IP5 and IP1144 to 216 were 2-fold stronger than IP1 or IP556–216. Increased amounts of Caco-2 cells (1 to 2x10⁶) seeded in a 6-well plate and resulting in a very densely packed confluent monolayer of microscopically differentiated Caco-2 cells at day 2 after transfection did not result in higher reporter gene expression of the promoter constructs IP1, IP2, or IP3 (data not shown). The crossover mutant IP(1167–266)-LP5 had markedly reduced promoter activity in all 3 cell lines compared with the construct LP5 (Figure 6D). The mutants IP(581–266)-LP5 and IP(318–266)-LP5 demonstrated a gradual increase of promoter activity in all 3 cell lines compared with IP(1167–266)-LP5 (Figure 6D). IP(318–266)-LP5, however, still had less promoter activities than did the baseline construct LP5 (Figure 6D).

Transfection of an APOBEC-1 Minigene Containing the 5' Flanking Region of Exon 1 in Human Hepatoma HuH-7 Cells
p–APOBEC-1, which contained the promoter fragment LP1 (-2535 to +28), the cDNA of rat APOBEC-1 from position -31 from ATG to position -699, the SV40 small T-antigen splice site and the SV40 early poly(A) site in the plasmid backbone of pT7T3, and the positive control plasmid pSVL–APOBEC-1, in which APOBEC-1 expression was driven by the SV40 late promoter, were transfected into the human hepatoma cell line HuH-7. Apo B mRNA from transfected and mock-transfected HuH-7 cells was amplified by RT-PCR and assayed by primer extension (Figure 7). Apo B mRNA from HuH-7 cells transfected with p–APOBEC-1 was edited in contrast to that of mock-transfected HuH-7 cells, yet the amount of apo B mRNA editing was considerably lower in comparison with HuH-7 cells transfected with pSVL– APOBEC-1. Similar results were also obtained in human hepatoma HepG2 cells.

View larger version (34K): [in this window] [in a new window]   Figure 7. Editing in human hepatoma HuH-7 cells transfected with APOBEC-1 minigene. Liver-specific promoter LP of the rat APOBEC-1 gene was cloned upstream of an APOBEC-1 expression cassette containing the open reading frame of APOBEC-1, SV40 small T-antigen splice site, and SV40 early poly(A)+ site to generate construct p–APOBEC-1. p–APOBEC-1 and pSVL–APOBEC-1, in which APOBEC-1 expression was driven by SV40 late promoter, were transfected in human hepatoma HuH-7 cells. Two days after transfection apo B mRNA was amplified by RT-PCR and assayed for editing by primer extension analysis. Autoradiograph of primer extension assay is shown. Positions of extension products for edited (U) and unedited (C) apo B mRNA are indicated.

	Discussion

Top Abstract Introduction Methods Results Discussion References

 
The present investigation demonstrates that the tissue-specific expression of APOBEC-1 is a consequence of differential promoter usage in the rat APOBEC-1 gene. The low expression of APOBEC-1 in rat organs other than the intestine, such as the liver, spleen, lung, kidney, or heart is due to a distinct promoter at the 5' end of the rat APOBEC-1 gene, which leads to tissue-specific exon use and alternate splicing in the 5' UTR of rat hepatic APOBEC-1 mRNA. The second downstream-located promoter specifically promotes strong expression of APOBEC-1 in the intestine but does not contribute significantly to the extraintestinal expression of rat APOBEC-1.

In the rat, APOBEC-1 is expressed not only in the intestine but also in the spleen, liver, kidney, lung, and heart, in contrast to the situation in humans or rabbits.^{11 14 29} Although most of these organs do not have influence on lipoprotein metabolism, the expression of APOBEC-1 in rat liver with hepatic apo B mRNA editing and the secretion of apo B-48–containing VLDL has an important impact on species-specific LDL levels.^{4 7} Given the importance of LDL for the development of atherosclerosis and coronary artery disease, this topic is of great physiological significance. The current investigation focused on the molecular basis of tissue-specific expression of APOBEC-1 the in rat liver and intestine, the 2 main sources for apo B–containing lipoproteins. The results demonstrate that hepatic expression of APOBEC-1 is due to a distinct promoter in the rat APOBEC-1 gene, which was designated LP. The intestine-specific promoter, termed IP, lacks a typical TATA box in the vicinity of the transcriptional start site but contains a hexanucleotide MED-1 motif (GCTCCC) within the 5' UTR downstream from the transcriptional start site.³⁵ This MED-1 sequence is characteristic for a new class of RNA polymerase II promoter without a canonical TATA box.³⁵ Overall, our results are in good agreement with 2 other very recent publications on the genomic organization and expression pattern of the rat APOBEC-1 gene.^{36 37} As in the mouse,¹⁷ the rat APOBEC-1 gene appears to have acquired an additional exon with a distinct promoter region that directs expression of APOBEC-1 in organs other than the intestine. This upstream-located promoter LP leads to tissue-specific exon use and alternate splicing within the 5' UTR of exon 2, without alteration of the open reading frame of APOBEC-1. According to the data presented, this promoter LP is a "classic" promoter with a canonical TATA box and far less tissue specificity than the downstream promoter IP, which appears to be very specific for the intestine. Qian et al³⁷ recently reported 2 allelic variants of the rat APOBEC-1 gene, of which the variant RE5 allele lacks the upstream liver-specific promoter. In RE5 transgenic mice, expression of rat APOBEC-1 was dramatically reduced in the liver, kidney, and brain in comparison with mice transgenic for the RE4 allele containing the upstream liver-specific promoter.³⁷ The presumptive absence of this less-tissue-specific promoter in the human APOBEC-1 gene most probably represents the molecular basis for the restriction of APOBEC-1 expression to the human intestine.^{29 38}

The transcriptional start sites for APOBEC-1 in the rat liver and intestine were determined by 5' RACE and primer extension analysis. In the liver, primer extension analysis and 5' RACE pinpointed the transcriptional start site exclusively within exon 1. Primer extension analysis mapped the transcriptional start site 22 nucleotides upstream from the longest 5' RACE product. In the intestine, 5' RACE resulted in 3 distinct products, all of which mapped to exon 2. The three 5' RACE products differed each in 20 nucleotides at the 5' end. Primer extension analysis of rat intestinal RNA yielded a major and a minor extension product within this region of exon 2. Ribonuclease protection assays confirmed the transcriptional start sites and the splice variants of APOBEC-1 in the rat liver and intestine and proved that the expression of APOBEC-1 in organs such as the spleen, liver, lung, kidney, and heart was driven exclusively by the upstream-located promoter LP. With the use of ribonuclease protection assays, the transcriptional start sites for intestinal APOBEC-1 mRNA were found to be located within exon 2, without any evidence for transcription initiation in exon 1. Three transcriptional start sites were found downstream from the distal splice acceptor site exactly as identified by 5' RACE. The fragments of 154, 144, and 103 nucleotides, which are protected in intestinal RNA by the exon 2–specific antisense RNA probe, indicate that the splice acceptor sites of exon 2 are additional sites of transcription initiation in the intestine. This evidence of multiple start site selection for the transcription of APOBEC-1 in the rat intestine further supports the classification of IP as a TATA-less promoter for which multiple transcription initiation is characteristic.³⁵

The transfection studies in human hepatoma HuH-7 cells and human colonic carcinoma Caco-2 cells further demonstrate that the 2 promoters of the rat APOBEC-1 gene have distinct activities. The liver-specific promoter LP in the 5' flanking region of exon 1 had much stronger activities in HuH-7 and NIH/3T3 cells than in Caco-2 cells. Its minimal promoter elements reside within a region 140 bp from the transcriptional start site, which contains the putative TATA box and the putative CAT box. The organization of LP in the rat is very similar to the upstream-located promoter of the mouse APOBEC-1 gene, which also contains a variant TATA box 20 nucleotides upstream from the transcriptional start site and an upstream CAT box.²⁹ In contrast to LP, IP exhibited a 4-fold increased promoter activity in Caco-2 cells in comparison with HuH-7 or NIH/3T3 cells. The constructs IP5 and IP6 had strongly increased promoter activity in both Caco-2 and HuH-7 cells. Negative regulatory elements within the region -1100 to -500 of IP were further demonstrated by using the deletion mutants IP1144–216 and IP556–216 as well as the crossover mutants IP(1167–266)-LP5, IP(581–266)-LP5, and IP(318–266)-LP5. This region appears to be a universal negative regulatory element restricting the expression of IP to the intestine. The expression of IP in human colonic carcinoma Caco-2 cells probably underestimates its promoter activity in the small intestine, for which no better-suited cell line was available for this study. The addition of insulin, T3, dexamethasone, or glucagon did not influence reporter gene expression of any of the generated constructs. Whether these results reflect the absence of hormonal regulation of the rat APOBEC-1 gene or are merely due to the tested promoter constructs lacking the appropriate hormonal response elements remains to be investigated. In accordance with the results presented, T3 has been shown to stimulate the editing of rat hepatic apo B mRNA without increasing the expression of APOBEC-1.²³ The modus of hormonal regulation of hepatic apo B mRNA editing in the rat liver appears to be complex, involving both APOBEC-1 and other trans-acting factors.²³ Our results are in some contrast to the recent published study of Hirano et al,³⁶ who did not observe a major difference in promoter activity of the intestine-specific promoter IP (REX1) between Caco-2 cells and rat hepatoma McArdle7777 cells nor increased promotor activities of their shorter constructs. Although the MED-1 motif was conserved in all promoter constructs of the current study (IP1 to IP7), the constructs of Hirano et al³⁶ ended exactly within this motif. Whether this or other differences explain this discrepancy remains to be studied.

The liver-specific promoter LP of the rat APOBEC-1 gene is sufficient to induce APOBEC-1 expression in human hepatoma HuH-7 and HepG2 cells. This experiment demonstrates that the lack of a liver-specific promoter within the human APOBEC-1 gene^{29 38} most probably explains the absence of APOBEC-1 expression in the human liver rather than the hypothetical assumption of specific repressors for APOBEC-1 expression. Further characterization of the liver-specific promoter of the rat APOBEC-1 gene in transgenic animals will help to elucidate how APOBEC-1 can be expressed in the liver without harmful effects. The mystery of APOBEC-1 as an oncogene must reside within the regulatory elements of the APOBEC-1 gene from species such as the rat, mouse, dog, or horse, which naturally express APOBEC-1 in the liver. Solving this mystery might revive the prospects for gene therapy with APOBEC-1, which appeared to be gloomy due to the oncogenic threats of APOBEC-1.³³

	Selected Abbreviations and Acronyms

 

APOBEC-1 = apo B editing catalytic polypeptide 1 CMV = cytomegalovirus ß-gal = ß-galactosidase MED = multiple start site element downstream PCR = polymerase chain reaction RACE = rapid amplification of cDNA ends RT = reverse transcription SV = simian virus UTR = untranslated region

	Acknowledgments

 
This study was supported by Deutsche Forschungsgemeinschaft Gr 973/2–2 (J.G.) and SFB 545 (J.G.) and by Bundesministerium für Bildung und Forschung (BMBF) Förderkennzeichen 01KV95090 (J.G.).

Received September 17, 1997; accepted January 28, 1998.

	References

Top Abstract Introduction Methods Results Discussion References

 

1. Kane JP. Apolipoprotein B: structural and metabolic heterogeneity. Annu Rev Physiol. 1983;45:637–650.[Medline] [Order article via Infotrieve]
2. Goldstein JL, Brown MS. A receptor-mediated pathway for cholesterol homeostasis. Science. 1986;232:34–47.[Free Full Text]
3. Steinberg D, Witztum JL. Lipoproteins and atherosclerosis: current concepts. JAMA. 1990;64:3047–3052.
4. Innerarity TL, Boren J, Yamanaka S, Olofson S-O. Biosynthesis of apolipoprotein B-48 containing lipoproteins. J Biol Chem. 1996;271:2353–2356.[Free Full Text]
5. Powell LM, Wallis SC, Pease RJ, Edwards YH, Knott TJ, Scott J. A novel form of tissue-specific RNA processing produces apo B-48 in the intestine. Cell. 1987;50:831–840.[Medline] [Order article via Infotrieve]
6. Chen SH, Habib G, Yang CY, Gu ZW, Lee BR, Weng SA, Silbermann SR, Cai SJ, Deslypere JP, Rosseneu M, Gotto AM, Li WH, Chan L. Apo B-48 is the product from a messenger RNA with an organ-specific in-frame translational stop codon. Science. 1987;238:363–366.[Abstract/Free Full Text]
7. Greeve J, Altkemper I, Dieterich JH, Greten H, Windler E. Apolipoprotein B mRNA editing in 12 different mammalian species: hepatic expression is reflected in low concentrations of apo B-containing plasma lipoproteins. J Lipid Res. 1993;34:1367–1382.[Abstract]
8. Greeve J, Navaratnam N, Scott J. Characterization of the apolipoprotein B mRNA editing enzyme: no similarity to the proposed mechanism of RNA editing in kinetoplastid protozoa. Nucleic Acids Res. 1991;19:3569–3576.[Abstract/Free Full Text]
9. Garcia ZC, Poksay KS, Boström K, Johnson DF, Balestra ME, Shechter I, Innerarity TL. Characterization of apolipoprotein B mRNA editing from rabbit intestine. Arterioscler Thromb. 1992;12:172–179.[Abstract/Free Full Text]
10. Hodges P, Navaratnam N, Greeve J, Scott J. Site-specific creation of uridine from cytidine in apolipoprotein B mRNA. Nucleic Acids Res. 1991;19:1197–1201.[Abstract/Free Full Text]
11. Teng BB, Burand CF, Davidson NO. Molecular cloning of an apolipoprotein B messenger RNA editing protein. Science. 1993;260:1816–1819.[Abstract/Free Full Text]
12. Navaratnam N, Morrison JR, Battacharya S, Patel D, Funahashi T, Giannoni F, Teng BB, Davidson NO, Scott J. The p27 catalytic subunit of the apolipoprotein B mRNA editing enzyme is a cytidine deaminase. J Biol Chem. 1993;268:20709–20712.[Abstract/Free Full Text]
13. Driscoll DM, Zhang Q. Expression and characterization of p27, the catalytic subunit of the apolipoprotein B mRNA editing enzyme. J Biol Chem. 1994;269:19843–19847.[Abstract/Free Full Text]
14. Yamanaka S, Poksay KS, Balestra ME, Zeng GQ, Innerarity TL. Cloning and mutagenesis of the rabbit apo B mRNA editing protein. J Biol Chem. 1994;269:21725–21734.[Abstract/Free Full Text]
15. Navaratnam N, Bhattacharya S, Fufino T, Patel D, Jaumuz AL, Scott J. Evolutionary origins of apo B mRNA editing: catalysis by a cytidine deaminase that has acquired a novel RNA-binding motif at its active site. Cell. 1995;81:187–195.[Medline] [Order article via Infotrieve]
16. Anant S, MacGinnitie AJ, Davidson NO. Apobec-1, the catalytic subunit of the mammalian apolipoprotein B mRNA editing enzyme, is a novel RNA-binding protein. J Biol Chem. 1995;270:14762–14767.[Abstract/Free Full Text]
17. Nakamuta M, Oka K, Krushkal J, Kobayashi K, Yamamoto M, Li WH, Chan L. Alternative mRNA splicing and differential promoter utilization determine tissue-specific expression of the apolipoprotein B mRNA editing protein (APOBEC-1) gene in mice. J Biol Chem. 1995;270:13042–13056.[Abstract/Free Full Text]
18. Davidson NO, Powell LM, Wallis SC, Scott J. Thyroid hormone modulates the introduction of a stop codon in rat liver apolipoprotein B messenger RNA. J Biol Chem. 1988;263:13482–13485.[Abstract/Free Full Text]
19. Baum CL, Teng BB, Davidson NO. Apolipoprotein B messenger RNA editing in rat liver: modulation by fasting and refeeding a high carbohydrate diet. J Biol Chem. 1990;265:19263–19270.[Abstract/Free Full Text]
20. Seishima M, Bisgaier CL, Davies SL, Glickman RM. Regulation of hepatic apolipoprotein synthesis in the 17-ethinyl estradiol-treated rat. J Lipid Res. 1991;30:941–951.
21. Sjöberg A, Oscarsson J, Boström K, Innerarity TL, Eden S, Olofsson SO. Effects of growth hormone on apolipoprotein-B (apo B) messenger ribonucleic acid editing, and apo B48 and apo B100 synthesis and secretion in the rat liver. Endocrinology. 1992;130:3356–3363.[Abstract]
22. Patterson AP, Tennyson GE, Hoeg JM, Sviridov DD, Brewer HB. Ontogenetic regulation of apolipoprotein B mRNA editing during rat and human development. Arterioscler Thromb. 1992;12:468–473.[Abstract/Free Full Text]
23. Funahashi T, Giannoni F, DePaoli AM, Skarosi SF, Davidson NO. Tissue-specific, developmental and nutritional regulation of the gene encoding the catalytic subunit of the rat apolipoprotein B mRNA editing enzyme: functional role in the modulation of apo B mRNA editing. J Lipid Res. 1995;36:414–428.[Abstract]
24. Lau PP, Cahill DJ, Zhu HJ, Chan L. Ethanol modulates apolipoprotein B mRNA editing in the rat. J Lipid Res. 1995;36:2069–2078.[Abstract]
25. Thorngate FE, Raghow R, Wilcox HG, Werner CS, Heimberg M, Elam MB. Insulin promotes the biosynthesis and secretion of apolipoprotein B-48 by altering the apolipoprotein B mRNA editing. Proc Natl Acad Sci U S A. 1994;91:5392–5396.[Abstract/Free Full Text]
26. Schock D, Kuo SR, Steinburg MF, Bolognino M, Sparks JD, Sparks CE, Smith HC. An auxiliary factor