(Received for publication, January 22, 1997, and in revised form, April 9, 1997)
From the Gladstone Institute of Cardiovascular Disease, the Cardiovascular Research Institute, and the Department of Pathology, University of California, San Francisco, California 94141-9100
The species and tissue specificity of
apolipoprotein (apo) B mRNA editing is determined by the expression
of apoB editing catalytic polypeptide 1 (APOBEC-1), the cytidine
deaminase that catalyzes apoB mRNA editing. To understand the
molecular mechanisms that regulate the transcription of APOBEC-1, we
characterized rat APOBEC-1 cDNA and genomic DNA. cDNA cloning
and RNase protection analysis showed two alternative promoters for the
tissue-specific expression of APOBEC-1 in the liver and intestine,
Pliv and Pint. Both promoters lack a TATA
box, and Pint belongs to the MED-1 class of promoters,
which initiate transcription at multiple sites. We also identified two
allelic forms of the APOBEC-1 gene from the characterization of two rat
APOBEC-1 P1 genomic clones, RE4 and RE5. The RE4 allele is 18 kilobases
long and contains six exons and five introns, whereas the RE5 allele
contains an additional ~8 kilobases of intron sequences and an extra
exon encoding a 5-untranslated region; however, the APOBEC-1
transcripts from the two alleles appear to have similar, if not
identical, functions. Transgenic mouse studies showed that
Pliv was preferentially used in the liver, kidney, brain,
and adipose tissues, whereas Pint was preferentially used
in the small intestine, stomach, and lung. Our results suggest that the
tissue-specific expression of APOBEC-1 is governed by multiple
regulatory elements exerting control over a single coding sequence. The
presence or absence of these regulatory elements may determine the
tissue-specific expression of APOBEC-1 in other mammalian species.
Apolipoprotein (apo)1 B mRNA, which is synthesized abundantly in the liver and intestine, encodes two forms of apoB protein that are crucial for lipoprotein metabolism (1). The full-length apoB, apoB-100, is required for the synthesis and assembly of very low density lipoproteins (VLDL) by the liver. VLDL are converted to low density lipoproteins (LDL) in which apoB-100, the sole apolipoprotein component of LDL, mediates the uptake and removal of LDL from the circulation by the LDL receptor pathway. Mutations in apoB-100 or in the LDL receptor that disrupt ligand-receptor interactions, resulting in high LDL cholesterol levels and an increased risk of atherosclerosis (2, 3). The second form of apoB, apoB-48, is essential for the synthesis and transport of large triglyceride-rich chylomicrons. ApoB-48 is generated from an apoB mRNA that has been modified by a unique post-transcriptional mechanism, namely apoB mRNA editing. A specific cytidine residue in apoB mRNA, C6666, is deaminated to a uridine. This C to U change generates an in-frame stop codon, which leads to the production of apoB-48, the amino-terminal 48% of apoB-100 (1).
ApoB editing has both tissue and species specificity. The intestine of all mammals edit apo-B mRNA, but lower species such as avians and amphibians lack sequence determinants in their apoB sequence and editing activity in their intestines and synthesize only apoB-100. In some mammals, such as dogs, horses, mice, and rats, the liver also possesses editing activity (4). All other mammals examined, including humans, synthesize only apoB-100 in the liver. The editing of apoB mRNA by the liver has important metabolic consequences. In mice and rats, VLDL-containing apoB-48 is not converted to LDL, and therefore in these species, the low levels of LDL and the resistance to hypercholesterolemia when fed a high-cholesterol, high-fat diet are partly due to the synthesis of apoB-48-containing VLDL (4).
The probable basis for this species- and tissue-specific expression of apoB mRNA editing was suggested by the identification of the cytidine deaminase that catalyzes this process, APOBEC-1, an acronym for apoB editing catalytic polypeptide 1. APOBEC-1 alone is insufficient to catalyze apoB mRNA editing and requires auxiliary protein(s). Theoretically, the tissue and organ specificities could be determined by the expression of either APOBEC-1 or the auxiliary protein(s). However, auxiliary protein activity is found in most organs and tissue extracts, while APOBEC-1 is selectively expressed. In rabbits and humans, APOBEC-1 is mainly expressed in the small intestine, with low abundance in the colon (rabbit) and only trace amounts in other tissues (5-7). APOBEC-1 is expressed in a large number of organs in mice and rats. In rats, APOBEC-1 mRNA is found at high levels in the spleen and small intestine and at lower levels in the liver, colon, kidney, and lung (8). In mice, mRNA expression is highest in the small intestine, liver, and spleen, followed by the kidney, lung, muscle, and heart (9). The reason for the species-specific differences in the expression of APOBEC-1 is not clear. It could relate to differences in transcriptional regulatory sequences of the APOBEC-1 gene from different species or in concentrations of the trans-acting protein factors acting on the APOBEC-1 gene in the livers and other tissues of different species. In addition to its tissue specificity, apoB mRNA editing is also regulated during development by hormones and diet (10-13).
Since APOBEC-1 is the key enzyme in apoB mRNA editing, the mechanism of APOBEC-1 gene expression has important functional implications. Discernment of the species- and tissue-specific regulation of APOBEC-1 expression could be significant for apoB lipoprotein metabolism. For example, if humans expressed low levels of APOBEC-1 in the liver, as do mice and rats, less apoB-100 VLDL would be synthesized, lowering plasma levels of LDL. To understand how species-specific, tissue-specific, developmental stage-specific, and hormonal regulatory signals of APOBEC-1 gene are integrated, it will be crucial to identify the cis-acting elements involved in each aspect of the regulatory mechanism. In the present studies, we identified two promoters involved in tissue-specific regulation of APOBEC-1 gene and tested their functions in transgenic mice. We also identified two allelic forms of the rat APOBEC-1 gene.
Three P1 plasmid clones (RE4, RE5, and RE6) were identified by screening a rat P1 genomic library (Genome System) using primer sets R235 and R320 (the sequences of all oligonucleotides used in this study are shown in Table I). The insert size of the P1 plasmid clones was determined by restriction enzyme digestion at NotI and SalI sites in the vector. The existence of the rat APOBEC-1 gene in these clones was confirmed by Southern blotting with probes generated from rat APOBEC-1 cDNA (kindly provided by Dr. Donna Driscoll, Cleveland Clinic Foundation). To characterize the genomic structure of APOBEC-1 in RE4, several BamHI genomic fragments derived from RE4, which hybridized to APOBEC-1 cDNA, were subcloned into the plasmid pBS.SK+. (Stratagene, La Jolla, CA). One of these subclones, RE4.B2.0K, which contains Exon 2 and its flanking sequences, was used as a probe for the restriction fragment length polymorphism experiments. These genomic subclones, as well as the RE4 P1 genomic clone, were sequenced with primers corresponding to rat APOBEC-1 cDNA sequences. The number and size of exons and the positions of exon-intron junctions were determined by comparing genomic DNA and cDNA sequences. The distance between exons was determined by the extra-long (XL) polymerase chain reaction (PCR) method (XL-PCR kit, CLONTECH) with exon-specific primers under the conditions suggested by the manufacturer and the following PCR program: 1 min at 94 °C; 16 cycles of 15 s at 94 °C, 10 min at 68 °C; 12 cycles of 15 s at 94 °C, 10 min at 68 °C with an auto extension of 20 s per cycle; 10 min at 72 °C. The promoter sequences were determined by oligonucleotide walking using a Taq Dye-Deoxy Terminator automatic sequencing kit (Applied Biosystems).
|
The structures of the RE5 and RE6 genomic clones were determined by
restriction mapping and XL-PCR analysis. The intestinal promoter
sequence and the sequence 5 of the promoter in RE5 were determined by
sequencing a 6.5-kilobase (kb) XhoI fragment subcloned into
pBS.SK+. Exon 1a was located by subcloning a 4.7-kb
XhoI/BamHI fragment into pZero vector
(Invitrogen) and analyzing the subclone by XL-PCR. XL-PCR was used to
determine the genomic structure of the RE5-type APOBEC-1 gene 5
to
Exon 1a. The distance between Exon 1 and Exon 1a was determined by
direct PCR amplification of Lewis rat genomic DNA with primers
hybridizing to these two exons (REP.L, L8-A) using the Expand Long
Template PCR system (Boehringer Mannheim) with the following
modification. To achieve optimal amplification of these genomic
fragments, we used PCR Buffer 1 (50 mM Tris-HCl, pH 9.2, 16 mM (NH4)2SO4, 1.75 mM MgCl2) provided with this kit and added
dimethyl sulfoxide to a final concentration of 4% in the reaction.
Lewis rat genomic DNA (200 ng) was used as the template. After 1 min at
94 °C, the PCR reaction was performed for 10 cycles of 10 s at
94 °C, 30 s at 55 °C, 8 min at 68 °C; 20 cycles of 10 s at 94 °C, 30 s at 55 °C, 8 min at 68 °C, with
an extension of 20 s at 68 °C in each cycle. A final 10-min
extension at 68 °C was added after the last cycle. The genomic
sequences immediately upstream of Exon 1 and Exon 2 of the Lewis rat
APOBEC-1 gene were cloned by conventional Taq PCR with
primers derived from the RE4 and RE5 genomic clones (RE.PL-2S and
RE48A; RE5.DS-1 and RE4+5.A) and directly ligated into pCR2.1 vector
(Invitrogen), and their sequences were determined by automatic
sequencing.
Sprague-Dawley, Fischer, and Lewis rat tails were obtained from Harlan Sprague-Dawley. Genomic DNA was extracted after proteinase K digestion. Additional Sprague-Dawley genomic DNA samples were obtained from CLONTECH. The genomic DNA was digested with BamHI. Randomly primed DNA probes were generated from RE4.B2.0K and used for Southern blot analysis.
RNA Preparation and Northern Blot AnalysisTotal RNA was prepared from rat and transgenic mouse tissues by homogenizing tissues in RNA STAT-60 total RNA/mRNA isolation reagent (Tel-Test "B," Inc.) followed by chloroform extraction. For Northern blot analysis of rat APOBEC-1, a cDNA fragment containing the entire coding sequence of rat APOBEC-1 was used to generate randomly labeled probes. Total RNA (20 µg) from different rat tissues was fractionated on a 1% agarose/formamide gel and transferred to a nylon membrane. The blot was first hybridized to a 32P-labeled APOBEC-1 probe to detect APOBEC-1 mRNA and then stripped and reprobed with a 32P-labeled glyceraldehyde-3-phosphate dehydrogenase (GAPDH) probe.
Riboprobe Templates and RNase Protection AssayThe
riboprobe APOBEC-5 is a 5
RACE product cloned in a
pBS.SK+ plasmid, containing sequences from 8 to 318 nucleotides (nt) of the liver transcript. Linearization of this plasmid
at the EcoRI site and transcription with T7 polymerase
generated a 340-nt riboprobe. pGEM3.REP (a gift from Dr. Donna
Driscoll) contains the entire coding sequence of rat APOBEC-1, 33 nt of
5
-UTR, and 50 nt of 3
-UTR. Digestion of this plasmid at the
SspI site and transcription by SP6 polymerase generated a
230-nt riboprobe, APOBEC-3
, which protects a 202-nt fragment at the 3
end of APOBEC-1 mRNA. A genomic fragment containing 118 bp of Exon
2 and 203 bp of its 5
-flanking sequence was cloned by PCR with primers
RE.PI.9S and RE.PE-1. The resulting 321-bp PCR product was cloned into a pCR2.1 vector. Transcription of the plasmid linearized at a BamHI site in the vector with T7 polymerase generated the
riboprobe Int.St for the mapping of intestinal start site of
APOBEC-1. A 230-nt GAPDH probe was generated by digestion of the
plasmid pTRI-GAPDH-mouse (Ambion) with DdeI followed by
transcription with SP6 polymerase. Antisense riboprobes were
synthesized by in vitro transcription in the presence of
[
-32P]cytidine 5
-triphosphate (3000 Ci/mmol,
Amersham).
A total of 40 µg of RNA was coprecipitated with 1-5 × 105 of radiolabeled riboprobe. The substrate and probe RNA were redissolved in 30 µl of hybridization buffer (80% formamide, 40 mM PIPES, pH 6.4, 200 mM sodium acetate, and 1 mM EDTA), denatured at 85 °C for 3 min, and hybridized at 60 °C for 18 h. The unhybridized RNAs were digested with 10 units of RNase ONE (Promega) in 300 µl of RNase ONE digestion buffer (10 mM Tris-HCl, pH 7.5, 5 mM EDTA, 200 mM sodium acetate) for 1 h at 30 °C. The reaction was terminated by adding 30 µl of Stop Solution (10% SDS, 1.0 mg/ml yeast tRNA), and the protected RNA fragments were precipitated by ethanol and analyzed on a 6 or 8% polyacrylamide, 8 M urea gel, followed by autoradiography. Specific protected RNA fragments were quantitated by Scanalytics Ambis Radioisotopic Imaging System. Final values were expressed as a ratio of APOBEC-1/GAPDH to correct for differences in RNA content.
Reverse Transcription (RT)-PCR and RACE-PCRFirst-strand
cDNA was generated from DNase I-treated total rat RNA by reverse
transcription with Superscript RT (Life Technologies). Random primers
were used for reverse transcription except in 3 RACE, where an
oligo(dT)-containing adapter primer (Life Technologies) was used to
prime the synthesis of cDNA. After RNase H digestion, one-tenth of
the reverse transcription reaction was used directly for PCR
amplification except for the 5
RACE (see below). The resulting PCR
products were analyzed on agarose gels and directly cloned to PCR II
vector by TA cloning unless otherwise specified.
To map the 5 end of the rat intestinal APOBEC-1 by 5
RACE, the
cDNAs was purified with GlassMAX Spin Cartridge (Life Technologies) and modified by adding a polydeoxycytidine sequence to the 3
end with
terminal deoxynucleotidyl transferase. cDNA clones representing the
5
termini of APOBEC-1 were amplified by PCR with an anchor primer that
hybridized to the poly(dC) sequence and with antisense APOBEC-1 primers
(RE514A and RE153A) in two sequential nested PCR reactions.
The 5 end of rat liver APOBEC-1 cDNA was cloned from rat liver
5
-RACE ReadyTM cDNA (CLONTECH).
PCR was performed with 2 µl of cDNA, the 5
anchor primer, and
APOBEC-1 antisense primer RE514A for the primary reaction and RE153A
for the secondary reaction. The resulting PCR products were cloned into
pBS.SK+ vector. The 3
terminus of APOBEC-1 was determined
by 3
RACE. Oligo(dT)-primed cDNA was amplified with an antisense
primer hybridizing to the oligo(dT) anchor primer and sense primers
hybridizing to rat APOBEC-1 gene (RE153S).
Exon 1a and alternatively spliced forms of APOBEC-1 were cloned by reverse transcriptase-PCR from Lewis rat liver RNA with a sense primer hybridized to Exon 1 (REP.L) and an antisense primer hybridized to Exon 4 (RE153A). The expression of Exon 1a in rats with different APOBEC-1 genotypes and in transgenic mice was determined by reverse transcriptase-PCR with REP.L primer and a 32P-labeled RE153A primer.
Transgenic MiceP1 plasmids containing either rat APOBEC-1
genomic sequence RE4 or RE5 were digested with SalI and
NotI to release the genomic insert. The 52-kb RE4 and 78-kb
RE5 inserts were separated from the vector sequence on 1%
low-melting-point agarose gels by pulsed field gel electrophoresis
using the Bio-Rad CHEF-PRIII system. These fragments were purified by
gelase (Epicenter) digestion and dialyzed against Microinjection Buffer
P (10 mM Tris-HCl, pH 7.4, 0.2 mM EDTA, 100 mM NaCl, 30 µM spermine, 70 µM
spermidine). The constructs were injected into F2 C57/BL6XSJL murine
zygotes at a concentration of 2-3 ng/µl. Transgenic mice were
identified by Southern blot analysis with APOBEC-1 cDNA probe.
Three independent lines of transgenic mice for each construct were
established, and the expression of rat APOBEC-1 was determined by RNase
protection assay with APOBEC-3 riboprobe.
Primer extension of apoB was carried out essentially as described previously with the following modification (14). The rat and mouse liver cDNA were prepared by reverse transcription with Superscript reverse transcriptase. The primers used to amplify rat and mouse apoB were M49-rat and M50-rat and M49-mouse and M50-mouse.
Lipoprotein AnalysisPlasma samples from Lewis or Fisher rats fasted for 6 h were analyzed for lipoprotein contents by agarose gel electrophoresis, ultracentrifugation followed by SDS-polyacrylamide gel electrophoresis, anti-apoB Western blot, and fast protein liquid chromatography, as described previously (14).
To determine
the structure of the rat APOBEC-1 gene and its regulatory sequences, we
obtained three bacteriophage P1 genomic clones by the PCR screening of
a Sprague-Dawley rat genomic library. The sizes of the DNA inserts in
these P1 clones, as determined by SalI/NotI
digestion, are: RE4, 52 kb; RE5, 78 kb; and RE6, 86 kb (Fig.
1A). The presence of the APOBEC-1 gene in
these genomic clones was confirmed by Southern blot analysis with a
full-length rat APOBEC-1 cDNA probe. The shortest clone, RE4, was
chosen for further characterization of the genomic structure of the rat
APOBEC-1 gene. The exon-intron organization and the exon-intron
junctions of the APOBEC-1 gene were determined by direct DNA sequencing of the RE4 P1 plasmid and its subclones with a series of
oligonucleotides spanning the length of APOBEC-1 cDNA. Intron
length was determined by XL-PCR (XL-PCR) amplification of genomic
clones with primers that hybridized to the 5 and 3
exon sequence
flanking the introns. These results yielded the APOBEC-1 genomic
structure diagrammed in Fig. 1B. Table II
displays a summary of the sizes of the exons and introns. The RE4 rat
APOBEC-1 gene contained six exons and five introns, spanning a total of
18 kb. The RE4 P1 plasmid contained 3.8 kb of 5
sequence and 30 kb of
3
sequence flanking the APOBEC-1 gene. Restriction mapping with a
rare-cutting endonuclease Sfi showed that RE5 and RE6 clones
contain more 3
sequences flanking the APOBEC-1 gene than RE4. At the
3
end, RE5 and RE6 are 22 and 21 kb longer than RE4, respectively
(Fig. 1A). At the 5
end, RE6 is 4 kb longer than RE4.
Although RE5 appears to have more 5
-flanking sequence, as will be
described, it is a longer allelic form of the rat APOBEC-1 gene, and
the RE5 P1 clone is missing the first exon of the rat APOBEC-1
gene.
|
The previously published rat APOBEC-1 intestinal
cDNA sequence contains 30 bp of 5-untranslated region (5
-UTR) and
105 bp of 3
-UTR (8). Because liver APOBEC-1 mRNA (~1.2 kb)
is significantly larger than intestinal mRNA (~1.0 kb) (6) it
seemed likely that the rat liver APOBEC-1 mRNA contains
additional 5
- or 3
-untranslated sequences.
We used 5 and 3
RACE and RNase protection analysis to analyze the
APOBEC-1 transcripts from the liver and small intestine of the
Sprague-Dawley rat. Using 5
RACE, we compared the 5
-UTR of rat liver
APOBEC-1 cDNAs with that of the published rat intestinal APOBEC-1
cDNA (8) and found that the rat liver APOBEC-1 cDNAs contained
additional 5
-UTR nucleotides. These results suggest that the longer
rat liver APOBEC-1 transcripts are initiated by a unique liver promoter
located upstream of the intestinal promoter.
We performed RNase protection analysis using the riboprobes depicted in
Fig. 2A to investigate the longer and shorter
forms of APOBEC-1 transcripts expressed in the liver and small
intestine of Sprague-Dawley rats. The antisense riboprobe APOBEC-5
allowed us to distinguish between the two forms of APOBEC-1
transcripts. The antisense riboprobe APOBEC-3
protects the same
fragments in both forms of APOBEC-1 mRNA. Consistent with the
previous findings by Teng et al. (8), rat APOBEC-1 was
expressed abundantly in the small intestine and at a lower level in the
liver (Fig. 2B, APOBEC-3
). The majority of the liver
transcripts were in the longer form, represented by the protected
fragment of 310 bp, whereas only a very small amount of intestinal
APOBEC-1 transcripts was in the longer form (Fig. 2B,
APOBEC-5
). Four predominant fragments ranging from 210 to 150 nt
were protected by the APOBEC-1 mRNA from the rat small intestine
(Fig. 2B).
To investigate the origin of these intestinal transcripts, we performed
5 RACE-PCR using rat intestinal RNA as a substrate. Eight different
clones were analyzed. The transcriptional start sites of the clones are
shown in Fig. 3A. Two clones
started in Intron 1 with the longer clone starting 59 bp upstream from
the beginning of Exon 2. Six clones had start sites in Exon 2 sequences, which may account for the protected bands between 150 and
180 nt observed in the RNase protection assay. Therefore, we
demonstrated that the intestinal APOBEC-1 promoter initiates
promiscuous transcription at multiple sites. The transcription
initiation sites for the intestinal APOBEC-1 mRNA transcripts were
collectively called Pint. The base encoding the first
nucleotide of the longest intestinal transcript was designated as
+1.
APOBEC-1 liver transcripts were also analyzed by 5 RACE. The liver
transcripts are initiated at three separate and distinct sites at the
beginning of Exon 1 (Fig. 3B). The promoter located at the
beginning of Exon 1 was designated Pliv, since it is
preferentially used in the liver. Thus, these results suggest that two
distinct promoters, Pint and Pliv, are used to
direct the tissue-specific expression of the rat APOBEC-1 gene.
We determined the sequence of ~670 bp of the 5-flanking sequence of
Pint (Fig. 3A) and ~1 kb of 5
-flanking
sequence Pliv (Fig. 3B). Computer analysis using
known consensus sequences of transcription factor-binding sites
revealed initiator sequences close to the intestinal and liver APOBEC-1
transcription initiation sites. Putative binding sites for AP-1 (
367
to
359), GATA (
465 to
449), NF-
B (
238 to
228), and C/EBP
(
169 to
160) were found in the 5
-flanking sequence of
Pint (Fig. 3A). Interestingly, a MED-1
(Multiple start site Element
Downstream) site GCTCCCG, which facilitates transcription
initiation at multiple sites in a new class of TATA-less promoters, was
found at nucleotide 128 bp downstream from the initiation site of the
longest intestinal transcript, suggesting that the intestinal APOBEC-1
promoter belongs to this family of promoters (15). In the 5
-flanking
sequence of Pliv, putative binding sites were found for
SP-1 (
77 to
86), RARE (
918 to
902), GATA (
924 to
918,
132
to
118), NF-
B (
379 to
369), and liver-enriched transcription
factors HNF-3 (
805 to
815), HNF-4 (
671 to
665), and C/EBP
(
990 to
981,
976 to
967,
689 to
680) (Fig.
3B).
Detailed
restriction enzyme mapping and XL-PCR analysis of the three genomic
clones showed that RE6 was identical to RE4 but had additional 5 and
3
sequences. Direct DNA sequencing and XL-PCR demonstrated that the P1
plasmid RE5 had the same exon-intron organization as the P1 plasmids
RE4 and RE6 from Exon 2 through Exon 6. However, 5
of Exon 2 in RE5
(Fig. 1A), XhoI and BamHI digestion
indicated a restriction fragment length polymorphism (RFLP). DNA
sequencing of this region revealed that the P1 plasmid RE5 is identical
to the P1 plasmid RE4 up to
433 bp of the Pint. The
sequences completely diverged (Fig. 3A) 5
of this point, resulting in the loss of a BamHI site in the RE5 P1 plasmid
clone that is present in RE4 and RE6 clones at
497 bp (Fig.
3A). XL-PCR using a sense Exon 1 primer and an antisense
Exon 2 primer resulted in 4-kb PCR products from RE4 and RE6 but no PCR
product from RE5, suggesting that Pliv and Exon 1 are
missing in the RE5 P1 clone (data not shown). This result was confirmed
by Southern blot analysis with an Exon 1-specific probe (Fig.
4).
We explored two possible explanations for these findings. One
possibility is that there are two allelic forms of the rat APOBEC-1 gene. Since the P1 library was constructed with genomic DNA from Sprague-Dawley rats, an outbred strain, these two genes could coexist
in the library. The second possibility is that RE5 is a chimeric clone
generated from two unrelated genes during the construction and
propagation of the P1 library; however, compared with yeast artificial
chromosome libraries, this kind of recombination event is extremely
rare in P1 libraries (16, 17). To investigate whether RE5 was a cloning
artifact or a bona fide APOBEC-1 gene, we performed RFLP analysis on
several Sprague-Dawley rat genomic DNA samples obtained from
CLONTECH, taking advantage of the polymorphic BamHI site that existed in RE4 and RE6 clones but not in the
RE5 clone. A 2.0-kb BamHI fragment subcloned from the RE4
clone (RE4.B2.0K) was used as a probe. This probe hybridized to a 2-kb
BamHI fragment in the RE4 and RE6 clones and to a 2-kb
BamHI fragment in two Sprague-Dawley rat genomic DNA samples
(Fig. 5A, SD-1 and SD-2). It also
hybridized to a 10-kb BamHI fragment in the RE5 P1 clone and
to genomic DNA from one of the Sprague-Dawley rats (Fig. 5A, SD-3). These results indicated that RE5 was a bona fide form of rat APOBEC-1 gene. To confirm this conclusion, we also analyzed genomic
DNA from two inbred rat lines, Lewis and Fischer. The Lewis rats
exhibited the same BamHI RFLP pattern as the RE5 P1 plasmid,
and the BamHI RFLP for the DNA from Fischer rats was identical to that of the RE4 P1 plasmid (Fig. 5B). The
similarity of the Lewis APOBEC-1 gene to the RE5 P1 plasmid was not
limited to the loss of a BamHI site. DNA sequence analyses
of 384 bp of the genomic sequences (778 to
384) in the region of
sequence divergence in the Lewis rat APOBEC-1 gene were identical to
those of the RE5 P1 plasmid. Therefore, two allelic forms (the RE4/RE6 allele and the RE5 allele) of the APOBEC-1 gene exist in rats.
Structure, Expression, and Function of the RE5 Genotype
The observation that Pliv is missing from the RE5 APOBEC-1 genomic clone raised the question of whether APOBEC-1 was expressed in the liver of rats with the RE5-allelic form of the APOBEC-1 gene (e.g. Lewis rats). Since liver expression of APOBEC-1 is essential for hepatic apo-B mRNA editing (6, 18, 19), a lack of APOBEC-1 expression in the rat liver would have profound consequences on apo-B mRNA editing and therefore on lipoprotein metabolism. Rats lacking hepatic apo-B editing would be expected to have elevated levels of plasma LDL. If, on the other hand, rats with the RE5-like gene expressed APOBEC-1 in the liver, then it is likely that these rats contain the Pliv sequences that are missing from the RE5 P1 plasmid.
To investigate these two possibilities, we examined the tissue
distribution of APOBEC-1 mRNA, hepatic editing, and plasma lipoprotein profiles of Lewis (RE5 allele) and Fischer (RE4 allele) rats. Northern blot analysis showed that APOBEC-1 was expressed in the
liver and all other tissues examined in both Lewis and Fischer rats
(Fig. 6). The identities of the other larger bands seen
in the spleen and small intestine are not known. They do not correlate
with the size of any of the alternatively spliced forms of APOBEC-1
that we have identified. The extent of hepatic apo-B mRNA editing
was similar in the two strains (Lewis, 56.2%; Fischer, 59.9%). Fast
protein liquid chromatography analysis showed no significant difference
in lipoprotein profiles between Lewis and Fischer rats (data not
shown). To determine if the RE5 P1 plasmid was expressed in transgenic
mice, we generated transgenic mice using the 52-kb RE4 and 78-kb RE5 P1
inserts, since RE4 contains both Pliv and Pint
and RE5 contains only Pint. Three transgenic mouse lines
from each construct were established and analyzed for the rat APOBEC-1
expression in the liver and various other tissues by RNase protection
assay with the APOBEC-3 riboprobe. The hepatic editing of mouse apo-B
mRNA in RE4 transgenic mice was consistent with the hepatic
expression of the APOBEC-1 transgene. Apo-B mRNA was completely
edited in RE4 transgenic mice by the liver expression of APOBEC-1,
while the editing of apo-B mRNA in RE5 transgenic mice was similar
to that of nontransgenic mice (nontransgenic mice: 69.3 ± 2.2%,
n = 4; RE4 mice: 93.0 ± 1.6%, n = 5; RE5 mice: 62.0 ± 7.0%, n = 3), indicating
the rat RE5 P1 plasmid was either not expressed hepatically or was
expressed poorly in these transgenic mice. These results indicated that the APOBEC-1 gene in Lewis rats contained regulatory elements for its
hepatic expression and these elements were missing in the RE5 P1
genomic clone (Fig. 7, A and
B).
By 5 RACE-PCR, we found that the APOBEC-1 mRNA from the Lewis rat
liver contained Exon 1, which has a nucleotide sequence identical to
the RE4-type Exon 1, but transcription started 11 nt upstream from the
transcription initiation site of RE4 liver mRNA. We cloned and
sequenced the genomic sequence surrounding the start site from Lewis
rat genomic DNA and found that the Lewis APOBEC-1 gene had the
identical sequence as RE4 in the 5
-flanking region of
Pliv. However, the Pliv in the Lewis gene was
~12.4 kb upstream of the Pint (Fig. 8).
Additional DNA sequences encoding another liver-specific exon, Exon 1a,
were found in this gene and the RE5 genomic clone but were absent from
the RE4 allele. By restriction mapping and direct DNA sequencing, we
found that the rat APOBEC-1 RE4 allele was missing ~8.3 kb of
internal sequences present in the RE5 allele (Fig. 8).
Alternate Splicing of the Rat RE5 APOBEC-1 Allele
At least
four alternatively spliced forms of APOBEC-1 mRNA were expressed in
Lewis rat liver (Fig. 9). The first transcript contains
all of the exons. The second transcript contained a deletion of the
noncoding exon E1a (E1a), which should not affect the translation of
the APOBEC-1 protein. However, in the alternatively spliced form
(
E1a,E2), the translational start codon was spliced out; in
E1a,E3, the start codon was preserved but out-of-frame, and
translation of the sequence would produce an out-of-frame protein that
terminated shortly after initiation. The next in-frame ATG codon was
codon 144, which, if used to initiate translation, would produce a
truncated, nonfunctional APOBEC-1 protein. In a total of 7 clones
analyzed, 1 expressed all exons, 4 the
E1a form, 1/7 the
(E1a,E3)
form, and 1/7 the
(E1a,E2) form.
Functions of Pliv and Pint in Directing Tissue-specific Gene Expression in Transgenic Mice
Rats, like mice, express APOBEC-1 in multiple tissues. Therefore, we considered the possibility that promoters that direct the expression of the APOBEC-1 gene in the liver and small intestines also direct the expression of APOBEC-1 in other tissues. In RE4 transgenic mice, APOBEC-1 was expressed in all of the tissues examined, closely mimicking the expression pattern of APOBEC-1 in rats, with the highest expression levels in the small intestine, stomach, and spleen and relatively lower levels in the liver, kidney, brain, lung, and heart (Fig. 7A). In RE5 transgenic mice, the high levels of APOBEC-1 expression in the small intestine and the stomach were retained, but the expression in the spleen decreased, and the expression of APOBEC-1 mRNA in liver, kidney, and brain was dramatically reduced (Fig. 7A). These results indicate that RE5 lacks the regulatory elements for APOBEC-1 expression in these organs. In adipose tissue, APOBEC-1 expression was also detected at a high level in RE4 transgenic mice but at a much reduced level in RE5 transgenic mice (data not shown). The same expression patterns were observed in all three transgenic mouse lines for each construct. These results indicate that the Pliv directs the tissue expression of APOBEC-1 in the liver, kidney, brain, and adipose tissues, while the downstream Pint primarily directs intestine, stomach and, lung expression (Fig. 7B). Since spleen APOBEC-1 mRNA was only reduced by 50% in RE5, it seems that the spleen APOBEC-1 transcription started from both of the promoters, or alternatively, that different cell types in the spleen used different promoters. Furthermore, the expression of APOBEC-1 in tissues that predominantly utilized the upstream Pliv was not completely abolished in RE5 mice, suggesting that Pint was operating at low levels in these organs.
The tissue-specific expression of apo-B mRNA alters lipoprotein metabolism and apparently lowers plasma LDL concentrations in species with hepatic apo-B mRNA editing (14). Although the editing enzyme complex consists of APOBEC-1 and one or more additional auxiliary proteins, the expression of the catalytic subunit of APOBEC-1 determines a tissue's ability to express editing activity. Modulation of APOBEC-1 expression at the transcriptional level is a central mechanism for the regulation of apo-B mRNA editing by nutrients and hormones and during development (11, 12, 20). This study shows that in rats, two distinct promoters were used to direct tissue-specific expression of APOBEC-1 and that two allelic forms of the APOBEC-1 gene exist. The complexity of APOBEC-1 transcripts originated from both tissue-specific exon use and from alternative splicing.
Two Allelic Forms of the Rat APOBEC-1 GeneTwo different allelic forms of the rat APOBEC-1 gene, RE4 and RE5, were identified from the analysis of three rat APOBEC-1 P1 genomic clones. As shown in Fig. 8, Lewis rats possess the RE5-type APOBEC-1 gene, which has a 8.3-kb genomic fragment that is missing from the RE4-type APOBEC-1 gene, which is the allele present in Fisher rats. This 8.3-kb fragment contains an extra exon (Exon 1a) that is missing from the RE4 allele of the APOBEC-1 gene. We did not detect any functional differences between the two allotypes of the rat APOBEC-1 gene when we examined editing activity and lipoprotein profiles in Lewis and Fisher rats.
The comparison of the two rat APOBEC-1 alleles with the mouse APOBEC-1
gene (9) allows us to speculate on the evolution of the APOBEC-1 gene
(Fig. 8). The mouse gene contains eight exons, all of which are found
in the liver transcripts, whereas the intestinal transcripts contain
the last five. The rat APOBEC-1 RE5 and RE4 alleles contain seven and
six exons, respectively, all of which can be expressed in the liver.
The first exon of the rat gene, which is identical in the rat RE4 and
RE5 alleles, is shorter than that of the mouse; however, the sequences
in common are 87% identical to the first exon of the mouse gene (Exon
1a). Exon 1a, which is only found in the RE5 allele of the two rat
APOBEC-1 alleles, is 68% identical to the mouse Exon 3. This finding
suggests that, of the two rat alleles, RE5 is more closely related to
the structure of the mouse gene than RE4. The first exon expressed in
mouse intestine is Exon 4, which is 154 bp longer than its rat
counterpart (rat Exon 2). In both rat and mouse genes, the last four
exons are expressed in the liver and intestine transcripts. The
difference in the size of rat and mouse APOBEC-1 mRNA (~1 and 2 kb, respectively) is largely due to the long 3-UTR encoded by the last
exon of the mouse APOBEC-1 gene but absent in the rat genes.
A comparison of the Pint region also indicates that the RE5 allele is more closely related to the mouse APOBEC-1 gene than the RE4 allele. The genomic sequence for the RE4 and RE5 alleles diverge dramatically 433 bp upstream of Pint. Alignment of the mouse intestine promoter and the two rat Pint promoters revealed a high degree of conservation of sequences in this region between the mouse gene and the RE5 allele, whereas the RE4 allele has no similarity to the mouse gene or the RE5 allele in this region. The loss of Exon 1a and its surrounding sequences has no apparent consequence for the function of Pliv or Pint in directing tissue-specific expression of APOBEC-1, since rats with the RE4 (Fischer) and RE5 alleles (Lewis) show the same tissue-specific expression of mRNA. The effect of these deletions on the developmental or hormonal regulation of these genes is not known. The human APOBEC-1 gene is not expressed in the liver, and the previously characterized human intestinal APOBEC-1 transcripts do not contain sequence homology to the liver exons of the rat gene. Although there are other possibilities, these results suggest that the sequences containing the liver promoter and exons are lost from the human genome and therefore APOBEC-1 is not expressed in the human liver.
Dual Promoters for Tissue-specific Expression of the APOBEC-1 GeneThe rat APOBEC-1 gene is expressed in multiple tissues and
organs. The intestine, stomach, and spleen have the highest APOBEC-1 expression level, followed by the kidney, liver, lung, heart, and
brain. This suggests that the APOBEC-1 gene, unlike ubiquitously expressed housekeeping genes, has tissue-specific regulation. The
regulation of tissue-specific transcription requires two complementary components: the trans-acting transcription factors, which
are enriched in a given tissue, and the cis-acting binding
sites for these factors in the genomic DNA sequences surrounding the
gene, namely, promoters and enhancers (21). By analyzing the APOBEC-1 transcripts from different rat tissues, we located two distinct promoters for tissue-specific gene expression of the rat APOBEC-1 gene.
The function of these promoters was directly demonstrated by transgenic
mice studies. The upstream promoter, Pliv, which is found
in the RE4 transgene, directs APOBEC-1 expression in the liver, kidney,
adipose tissues, and brain. Rat APOBEC-1 expression was dramatically
reduced in these tissues of RE5 transgenic mice, which lack
Pliv. These results indicate that for the RE4 allele, all
necessary regulatory elements for APOBEC-1 expression in the liver,
kidney, adipose tissue, and brain are contained in the 3.8-kb DNA
sequence 5 of the APOBEC-1 gene and/or in intron 1. Further deletion
and site-directed mutagenesis studies will be necessary to pinpoint the
location of these tissue-specific elements. Several potential binding
sites for tissue-enriched transcription factors were found in this
regulatory region, such as those for HNF-3, HNF-4, and C/EBP.
The downstream promoter, Pint, is preferentially used in
the small intestine and stomach. Because the RE4 and RE5 transgenes contain only 433 bp in common 5 of the intestinal start sites, it is
likely that the tissue-specific elements for intestinal expression of
APOBEC-1 lie within these 433-bp 5
-flanking sequences. A less likely
possibility is that intestinal elements are present in introns or 3
of
the APOBEC-1 gene; such as the well characterized hepatic control
region of the apo-E gene is 15 kb downstream from its transcription
initiation site (22, 23).
It is worth noting that the two rat APOBEC-1 promoters are used preferentially, but not exclusively, in different tissues. Results from both the transgenic mouse study and the analysis of transcripts in different rat tissues support this notion. For example, the RE5 transgene, which lacks Pliv, supports low-level APOBEC-1 transcription in the liver. The relative strength of a promoter in a given tissue is determined by the activity of tissue-specific transcription factors bound to the promoter or enhancer elements. However, tissue-specific gene expression in many cases is not determined by a single trans-acting factor present in that tissue, but rather by the coordinated function of many tissue-specific and ubiquitous transcription factors. The liver and intestine are both derived from primitive endoderm and share similarity in the expression of a large number of transcription factors, including C/EBP and HNF-4 (24-26). However, the expression of certain HNF-3 family members is much higher in the liver than in the small intestine (27, 28). Although only trace amounts of apo-B mRNA are expressed in the spleen, high levels of APOBEC-1 mRNA (8) and protein (11) are found in this tissue. Pliv and Pint are equally active in the spleen. Binding sites for GATA, a transcription factor that is essential for gene expression in hematopoietic cells (29), are found in both promoters, suggesting that APOBEC-1 may be expressed in this lineage of cells.
The mouse APOBEC-1 gene also has two promoters. Nakamuta et
al. (9) described differential transcription start sites for liver
and intestine APOBEC-1 transcription in mice. They tested the activity
of 1.6-kb 5-flanking sequences of each transcription start sites in
three types of cells in vitro: HeLa (non-hepatic, nonintestinal epithelial cell line), Hepa (mouse hepatoma cell line),
and Caco-2 (human colon carcinoma cell line). The 5
-flanking sequence
of intestine transcription start site showed greatest tissue
specificity since it was active only in Caco-2 cells. The liver
promoter construct was expressed in all three cell lines and was
approximately 60-70% as active as the intestine promoter construct in
Caco-2 cells. The in vitro nature of these experiments and
the limited number of promoter sequences tested prevent definitive conclusions regarding the function of these mouse promoters. However, it is possible that, like the rat promoters, the mouse liver promoter may have low activity in the intestine, supporting our notion that
these two promoters are preferentially but not exclusively used in
different tissues. The functions of these mouse promoters in the
context of an intact mouse APOBEC-1 gene in vivo and their activity in other APOBEC-1 expressing tissues are not known.
Tissue-specific transcription factors apparently exert their functions by interacting with components of the basal transcription initiation complex. In most mammalian RNA polymerase II (Pol II) promoters, the assembly of such a complex starts with the binding of a general transcription factor TATA-binding protein to a TATA box ~20 to 30 bp upstream of the initiation site (30). However, in promoters lacking a TATA box, the function of TATA-binding protein and the DNA elements involved in basal transcription is not clear. Apparently, an initiator sequence can recruit TATA-binding protein-associated factors and Pol II to the promoter (31, 32). Basal transcription factor TFII-I has been shown to bind to some of the initiator elements (33). The transcription factor SP-1, which recognizes a GC-rich sequence and interacts with TATA-binding protein-associated factors, may facilitate the basal transcription of several TATA-less promoters (31). However, in some cases, such an initiator sequence is not sufficient to direct the transcription initiation at a specific site, and multiple start sites are used (15). Ince and Scotto (15) grouped this type of TATA-less promoters into a new class characterized by multiple start sites and a downstream MED-1 element. In the case of the P-glycoprotein gene, the prototype of this new class, the mutation of the MED-1 sequence decreased the transcription initiation downstream of the initiator to 25% of wild-type (15). However, little is known regarding the protein factors that bind to the MED-1 sequence or whether MED-1 and initiator sequence coordinately or independently regulate basal transcription.
Neither of the two rat APOBEC-1 promoters contains a TATA box.
Pliv probably relies on the function of the SP-1 and
initiator elements to recruit basal transcription factors. Three
transcription initiation sites close to each other are found in the
liver transcripts, a characteristic observed in this type of TATA-less
promoter. In contrast, Pint, the downstream promoter, lacks
an SP-1 element, and transcription from this promoter starts
promiscuously at least at eight different sites in a window of 126 bp.
A MED-1 site is found 128 bp downstream from the extreme 5 end of the
transcription initiation window. The position of the MED-1 as part of
the rat APOBEC-1 intestinal promoter site fits the consensus motif for this new class of promoters (15).
Interestingly, although the rat and mouse APOBEC-1 genes are similar in
their genomic structure, only the rat gene used this promiscuous
intestinal basal promoter. The mouse intestinal promoter contains a
TATA box at position 28 to
25 and it has a single intestinal
transcript (9). Intriguingly, the MED-1 sequence in the rat promoter is
conserved in the mouse gene but does not cause promiscuous
transcription initiation. Most likely, MED-1 acts independently to
start transcription at multiple sites only in the absence of a strong
promoter element such as the TATA box. An interesting question arises
from these analyses: if the same tissue-specific elements are used in
both rat and mouse APOBEC-1 genes, how do the tissue-specific
transcription factors work thorough the two distinct sets of basal
transcription machinery to regulate their functions? One possibility is
that a common component in these two types of initiation complexes acts
as the target for the tissue-specific transcription factors.
At least in the Lewis rat liver, the complexity of
APOBEC-1 transcripts is caused not only by alternative promoters and
multiple start sites, but also by alternative splicing of the 5 exons. The splicing donor of Exon 1 can be joined with the splicing acceptors of either of its three following exons. Exclusion of the noncoding Exon
1a gives rise to a shortened 5
-UTR. However, because this is the
normal APOBEC-1 liver form expressed from the RE4-type allele,
exclusion of Exon 1a probably has no significant effect on the
stability or translation of the transcript. However, exclusion of the
second and third exons will have a profound effect on APOBEC-1 function
because Exon 2 contains the translation start codon and deletion of
Exon 3 will jeopardize the original reading frame. Whether either of
these APOBEC-1 transcripts is translated is not known. But even if
translation occurs, the resulting APOBEC-1 protein would not have
editing activity.
We thank Kay Arnold, Isabelle Lee, and Stacy Taylor for technical help; Drs. Bruce Conklin, Robert H. Costa, Robert Mahley, and John Taylor for critical comments on this manuscript; Amy Corder, John Carroll, and Stephen Gonzales for graphics; Susannah White for manuscript preparation; and Gary Howard and Stephen Ordway for editorial support.