| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Messenger Ribonucleic Acid Isoforms Generated by Alternative Splicing and Promoter Usage1
EMBL (C.G., G.F.V.S.-B., F.G.), D-69117 Heidelberg, Germany; and National Diagnostic Center, National University of Ireland (C.G., P.N.), Galway, Ireland
Address all correspondence and requests for reprints to: Dr. Frank Gannon, EMBL, Postfach 10.2209, Meyerhofstrasse 1, D-69012, Heidelberg, Germany. E-mail: gannon{at}embl-heidelberg.de
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
(cER) messenger RNA (mRNA) variants in addition to the previously
described form (isoform A). Whereas one of the new variants (isoform B)
presents a 5'-extremity contiguous to the 5'-end of isoform A, the two
other forms (isoforms C and D) are generated by alternative splicing of
upstream exons (C and D) to a common site situated 70 nucleotides
upstream of the translation start site in the previously assigned exon
1 (A). The 3'-end of exon 1C has been located at position -1334
upstream of the transcription start site of the A isoform (+1). Whereas
the genomic location of exon 1D is unknown, 700 bp 5' to this exon were
isolated by genomic walking, and their sequence was determined. The
transcription start sites of the cER mRNA isoforms were defined. In
transfection experiments, the regions immediately upstream of the A–D
cER mRNA isoforms were shown to possess cell-specific promoter
activities. Three of these promoters were down-regulated in the
presence of estradiol and ER protein. It is concluded,
therefore, that the expression of the four different cER mRNA
isoforms is under the control of four different promoters. Finally,
RT-PCR, S1 nuclease mapping, and primer extension analysis of
these different cER mRNA isoforms revealed a differential pattern of
expression of the cER gene in chicken tissues. Together, the results
suggest that alternative 5'-splicing and promoter usage may be
mechanisms used to modulate the levels of expression of the chicken
ER gene in a tissue-specific and/or developmental stage-specific
manner. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Estrogen exerts its potent physiological effects by binding to its
cognate receptors, the estrogen receptors (ER), intracellularly. To
date, two nuclear ERs (ER and ERß), encoded by different genes in
a tissue-specific manner, have been identified in mammals (6, 7, 8, 9, 10).
Although it is highly probable that the ERß form is also present and
expressed in some tissues in oviparous vertebrates, its existence
remains to be demonstrated in these species. The ER and -ß
proteins belong to the steroid/thyroid hormone/retinoic acid receptor
family whose members act as ligand-inducible transcription factors
(11). Receptors of this family are characterized by a unique modular
structure. Discrete functional domains (named A–F), which include
regions required for DNA binding, ligand binding, and transcriptional
activation have been highly conserved within the family (12).
One approach to help answer the question of how the ER may account to a
large extent for the pleiotropic effects of its ligand in a wide range
of physiological processes, is to study the manner in which the
expression of the hormone receptor gene is controlled in an organism to
ensure that the correct amount of the protein is available, in the
correct cells, at the correct time of development. The human (h) ER
gene has recently been shown to be a very complex genomic unit,
exhibiting alternative splicing and promoter usage in a tissue-specific
manner. Six hER messenger RNA (mRNA) isoforms (A–F hER mRNAs)
are generated by the alternative splicing of five upstream exons
(1B–1F) to a common site upstream of the translation start codon
(12A ). Similarly, at least three ER mRNA forms have been identified
in the rat (13, 14, 15). These data suggest that alternative 5'-splicing
and promoter usage is probably a general feature of the ER genes in
mammals. In contrast, the existence of such a mechanism in oviparous
species remained to be elucidated when we began this work.
The chicken (c) ER gene has been characterized previously at both
the complementary DNA (cDNA) (16, 17) and the genomic (18) level. As
for the hER gene, its coding region is split into eight exons (18).
The 5'-untranslated and flanking sequences of the cER cDNA were
highly homologous with promoter and exon 1A sequences in human (18).
These data suggested a conservation of the ER gene organization and
expression through evolution. In this present study, we demonstrate
that, similar to the hER gene, the chicken ER gene is a complex
genomic unit, exhibiting alternative splicing and promoter usage. The
four cER mRNA isoforms (A–D cER mRNAs) that were isolated differ
in their 5'-untranslated regions (5'-UTRs) but code for the same
receptor protein. Controlled by different promoters, these transcripts
revealed a differential pattern of expression of the cER gene in
chicken tissues.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Rapid amplification of cDNA ends (RACE)
5'-RACE was performed as described by Frohman et al.
(19). Primer RI (5'-GTACTAGACATCCTCTCACGA-3'), located in exon 2, was
used for RT. Primers RII (5'-ATGGATGAAGGGTGAGAGCTG-3'; located in exon
1) and nested primer RIII (5'-GCTGCTTGACCCAAAAGATTCA-3'; located in
exon 1) were used for the first and second amplifications,
respectively, with the anchor primer
(5'-GGCCACGCGTCGACTAGTACGGGIIGGGIIGGGIIG-3'; for approximate primer
locations, see Fig. 2B
). PCR products were subcloned in the TA cloning
vector pCRTM2.1 (Invitrogen, San Diego, CA). Multiple
isolates were sequenced using a Pharmacia sequencing kit
(Pharmacia Biotech, Piscataway, NJ) to first check the
identity and then to confirm the accuracy of the sequences.
|
for approximate primer locations).
Aliquots of the PCR reactions were electrophoresed on a 1% agarose
gel. A unique band approximately 700 bp from the HindIII
vectorette library was isolated, subcloned in the pCRTM2.1 vector, and
sequenced.
|
clone containing 3 kb of
sequence upstream of the 5'-end of the cER cDNA (18), regions from
-503 to +183, -824 to -181, and -2463 to -1452, respectively [all
numbering is according to Nestor et al. (18) unless
otherwise stated]. For each construction, the two synthetic primers
used for the amplification [pAI
(5'-TGCGCTGGTACCTCTTTTACATTCTTCAATTTCTG-3') and pAII
(5'-AGTGCGAAGCTTCAACAGCAAGATCGGCAGCTGG-3'), pBI
(5'-GTTCATGGTACCAGCTGATTCCATTCAAGTTCTTTATTTT-3') and pBII
(5'-AGTGCGAAGCTTCACTGCAGGGCAGCCAGAG-CTCCACAG-3'), or pCI
(5'-GGTCATGGTACCCCCTTGATTTTATTCTACCAATAAATA-3') and pCII (5'-
AGTGCGAAGCTTGAGATAAGCCTGTTTAAATAACAC-3'), respectively] were designed
to introduce 5' KpnI and 3' HindIII restriction
sites at the ends of the PCR products. The amplified fragments were
directionally cloned into the polylinker of the pGL2 basic plasmid
(Promega Corp., Madison, WI) upstream of the luciferase
reporter gene. A reporter plasmid pGL2 pD containing 700 bp of putative
promoter region upstream of exons 1D was also constructed using the DNA
fragment cloned by the genomic walking technique. The reporter plasmid containing the apoprotein very low density apolipoprotein II (apoVLDLII) gene promoter fragment from -900 to +1455, called Apo-CAT (CAT, chloramphenicol acetyltransferase), was a gift from M. Evans (Morgantown, WV) (20).
ER expression vector preparation
An expression vector pSG cER was made by directionally
cloning the complete cER coding region from +158 to +2038 into the
parental expression vector pSG5 (21) as follows. Four microliters of
cDNA template (see RT-PCR section below for preparation) were amplified
using primers EVI (5'-ACGTAGAATTCACTGCCAGCTGCCGATCTTGC-3'; +158 to
+179) and EVII (5'-GCGTAGGATCCCGCTGCTGGGTTTCTCATACCAT-3'; +2038 to
+2015). The amplification was performed using the Expand high fidelity
PCR system from Boehringer Mannheim (Indianapolis, IN) following the
manufacturer’s instructions. Two primers, EVI and EVII, were designed
to introduce 5'-EcoRI and 3'-BamHI restriction
sites at the ends of the PCR product. The amplified receptor cDNA was
then directionally cloned into the polylinker
(EcoRI/BamHI) of the expression vector pSG5
downstream of the T7 promoter.
Cell culture and transient transfection assays
Chicken embryo fibroblast (CEF) cells (a gift from T. Graf,
Heidelberg, Germany) were maintained in DMEM supplemented with 5% FCS,
1% chicken serum, 10 mM HEPES (pH 7.4), 100 U/ml
penicillin, and 100 µg/ml streptomycin at 37 C in a 5%
CO2 incubator. Chicken hepatocellular carcinoma (LMH) cells
(American Type Culture Collection, Manassas, VA) were
grown in Weymouth’s MB/251 medium with 10% FCS,
L-glutamine, 100 U/ml penicillin, and 100 µg/ml
streptomycin. CEF cells were transiently transfected using the
DNA/calcium phosphate coprecipitation method (22). LMH cells were
transiently transfected as described by Binder et al. (23).
In all transfection studies, 6-cm dishes containing 5 x
105 cells were transfected with a total of 10 µg DNA/dish
[5 µg reporter plasmid and 0.1 µg reference plasmid EF-1-CAT
(24) and carrier DNA to 10 µg]. Medium was changed 6 h before
transfection. After 16-h incubation with the DNA-calcium phosphate
precipitate, the medium was aspirated, and cells were washed twice with
PBS, and fresh, serum-stripped, phenol red-free medium was added.
Transfected cells were cultured for 24 h in the absence or
presence of 10 nM 17ß-estradiol before
harvesting for luciferase and CAT assays. Luciferase assays were
performed as described by Brasier et al. (25) on 20% of the
lysate. CAT activity was assayed with the enzyme-linked immunosorbent
assay kit from Boehringer Mannheim using 20% of the lysate. The
reporter gene activity values were normalized for transfection
efficiency according to the activity of the cotransfected EF-1-CAT
control. The activity of the reporter gene was expressed relative to
the activity obtained using the promoterless reporter plasmid (pGL2
basic) in the same experiment.
RT-PCR analysis
Three micrograms of total RNA from various chicken tissues were
reverse transcribed using 10 ng oligonucleotide I
(5'-TCAAATGGAAAATACAGGTGGC-3') located in the 3'-UTR (from +2124 to
+2102 in exon 8) of the cER gene. Specific cDNAs were amplified
using the internal common oligonucleotide II
(5'-TGTTGACATATGTGGCACTATAG-3', from +2088 to +2064) in combination
with the upstream exon 1 A–D specific primers, AI
(5'-CCAGTGCTCACCCTGCATTT-3'), BI (5'-TTTCAGCGTCCTTTCCCGTTAGC-3'), CI
(5'-ATCAAGTACGTATTTATGTGTG-3'), and DI
(5'-TAATGGCAACA-ACCTTCTGGG-3'), respectively. Approximate primer
locations are shown in Fig. 5. After 30 cycles of PCR amplification,
1% of the PCR was taken and reamplified using the nested common
oligonucleotide III (5'-CGCTGCTGGGTTTCTCATACCAT-3', from +2038 to +2015
in exon 8) and the upstream nested oligonucleotides AII
(5'-AGCCTCAGAATAGGTTCTGGTG-3'), BII (5'-GACTAGCAAGAATAAAGT-3'), CII
(5'-TGTCTTAGCTGCATGTCTGTAGAG-3'), and DII (5'-CAACCTTCTGGGATAAAT-3'),
respectively. Both rounds of amplification were performed using the
Expand long template PCR system (Boehringer Mannheim) as recommended by
the manufacturer. Five microliters from each reaction were
electrophoresed on a 0.8% agarose gel, transferred to a nylon membrane
(Hybond N+, Amersham, Arlington Heights, IL),
Southern blotted, and hybridized at 55 C with oligonucleotide IV
(5'-AGGGTCATGGTCATTGCTAATGGC-3', in exon 1), which is common to all
cDNAs. This primer had been previously end labeled using T4
polynucleotide kinase and [
-32P]ATP (3000 Ci/mmol)
(26). Positive PCR products were visualized by autoradiography.
|
-32P]deoxy-CTP (3000 Ci/mmol).
The S1 probe A, A/B, and C templates were PCR amplified with the
biotinylated T7 primer and the M 13 reverse primer from the plasmids A,
A/B, and C, respectively. Plasmids A and A/B were constructed as
follows. Genomic PCR fragments covering regions -169 to +318 (A) and
-503 to +318 (A/B) of the cER gene were amplified using standard
PCR reaction conditions with the oligonucleotide pairs S1
(5'-GACTAGCAAGAATAAAGT-3')/S2
(5'-CAGCTGAGGTCTGCTCAGAGT-CTCCAGCT-3') and S3
(5'-TCTTTTACATTCTTCAATTTCT-3')/S2, respectively (see Fig. 4
for
approximate primer locations). These DNA fragments were subcloned in
the pCRTM2.1 vector, downstream of T7, to yield plasmids A and A/B. The
plasmid C insert, which spans the cER mRNA C region from +65 to
+376, was RT-PCR amplified using 5'-oligonucleotide S4
(5'-TCCAGTGTTATTTAAACAG-3') and 3'-oligonucleotide S2 (see Fig. 4 for
approximate primer locations). The RT-PCR product was then also
subcloned in the pCRTM2.1 vector and sequenced.
The long primer T, A/B, C, and D templates were PCR amplified with the
respective biotinylated 5'-primers S5 (5'-ACTGCCAGCTGCCGATCTTG-3'; +157
to +177), S6 (5'-AGCCTCAGAATAGGTTCTGGTG-3'; +122 to + 144), S7
(5'-TGTACTGCTGTCTTAGCTGCATG-3'; +180 to +203, numbering according to
cER mRNA C), and S8 (5'-ACCTTCTGGGATAAATAGGCTGTT -3'; +17 to +40),
and the common 3' M 13 reverse primer from plasmid A, C, or D. Plasmid
D was constructed by subcloning the RT-PCR product that spans a cER
mRNA D region from +17 (primer S8) to +212 (primer S2) in the pCRTM2.1
vector.
All biotinylated PCR products were bound to streptavidin-coated magnetic beads (Dynal, Great Neck, NY) as recommended by the manufacturer, and the nonbiotinylated DNA strands were removed in 0.1 M NaOH. The S1 probes and the long primers were obtained by extending the S2 primer annealed to the corresponding biotinylated, single stranded template. After elution of the single stranded DNA probes by alkaline treatment and magnetic separation, 105 cpm probe or primer were coprecipitated with 100 µg total RNA and then dissolved in 20–30 µl hybridization buffer [80% formamide, 40 mM piperazine-N, N'-bis[2-ethane-sulfonic acid] (PIPES; pH 6.4), 400 mM NaCl, and 1 mM EDTA (pH 8)], denatured at 65 C for 10 min, and hybridized overnight at 55 C. The S1 digestions and the reverse transcriptase extension were carried out as previously described (26), and the samples were electrophoresed through a denaturing polyacrylamide-urea gel.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
gene genes
showed areas of extensive homology. In addition to the beginning of the
well conserved coding region, two other areas of high homology, from
-5 to -157 (69% homology) and from -290 to -409 (76% homology),
were found in the 5'-flanking sequence, as shown in Fig. 1 (numbering from chicken ER gene
sequences). The first of these highly conserved regions corresponded to
the proximal promoter previously characterized by Nestor et
al. (18). These two regions were separated by a region of
intermediate homology (42% homology).
|
gene, we showed recently the existence of a consensus
acceptor splice site in exon 1A for six alternative upstream
5'-noncoding exons (12A ). The sequence alignment showed that there was
also a good candidate acceptor site (gtgttttttttag/GTG) in exon 1A of
the cER gene at position +153/4 relative to the previously
determined transcription start site (18). These data allowed the
prediction that an upstream exon(s) would splice to this point if the
transcriptional regulation of the cER gene was similar to that of
the hER gene. To determine whether such spliced sequences exist, a
Zap cDNA library, prepared from laying hen oviduct tissue, was
screened. Whereas many clones were isolated that corresponded to the
previously determined cER mRNA, only one clone (clone 2) had 11 bp
of sequence upstream of position +154 divergent from the original mRNA,
such as would occur if splicing from an upstream exon took place at
this site (Fig. 2A).
SI nuclease mapping and primer extension analysis directly confirmed
the existence of such a splicing event. A single stranded DNA probe,
which mapped the 5'-extremity and promoter region of the cER gene
(probe A in Fig. 2B), was used for the S1 mapping experiment. As shown
in Fig. 2B, three protected fragments were detected after hybridization
of probe A with oviduct RNA followed by SI nuclease digestion. Their
5'-extremities were located 393, 224, and 70 nucleotides upstream from
the start of the cER open reading frame (ORF). No protected
fragments were seen with yeast total RNA used as negative control. The
5'-extremity of the medium sized fragment (indicated as the start site
in Fig. 2B) corresponded to a position that was also mapped by primer
extension (see primer A/B in Fig. 6A) and confirmed therefore the
transcription start site previously characterized by Nestor et
al. (18). The 5'-position of the longest fragment (the fragment
weakly detected) resulted from full protection of the cER-specific
sequence of the probe, indicating that some RNAs were produced from
transcription start sites further upstream than the transcription start
site previously described (18). Finally, the shorter fragment resulted
from a partial protection of the classical cER mRNA up to the
predicted splice site position (+153/4), 70 nucleotides upstream of the
translational start site of the cER gene. No extension product was
seen at the corresponding position from a primer extension analysis
using a long primer specific to sequences 3' to the splice site (see
primer T in Fig. 6A). Therefore, this result suggested that a
significant fraction of the cER mRNAs included 5'-extremities
different from the previously characterized 5'-end. The fact that an
activity was readily detectable by SI nuclease mapping at the splice
site position in exon 1A, indicated that the splicing from upstream
regions was a significant event in the expression of the cER gene,
at least in oviduct tissue. This prompted further investigation using
the more sensitive PCR-based RACE technique to amplify new 5'-mRNA ends
of the cER gene.
|
cDNA ends and the corresponding
5'-flanking regions mRNAs (see Fig. 2B for primer
location). Sequence analysis of the subcloned RACE products revealed
5'-cDNA ends of three additional cER mRNA isoforms, named B, C, and
D. In keeping with the previous results of the SI nuclease analysis,
the most frequently cloned isoform was identical to the previously
described cER cDNA (called cER cDNA A hereafter). Clones
containing B, C, and D cER cDNA sequences were much less abundant.
All of the 5'-cER cDNA ends contained common exon 1A sequences 3' to
the splice site, but diverged from each other immediately upstream from
this position (Figs. 1 and 3).
|
mRNA B isoform was identical to
exon 1A and its 5'-flanking region, indicating the existence of a
second promoter region (promoter B) directly upstream of the promoter
characterized previously by Nestor et al. (18) (called
promoter A hereafter). Corroborating these data were the previously
mentioned conservation between the human and chicken ER genes of a
second region of high homology (from -290 to -409) directly upstream
of the proximal promoter A (Fig. 1). In humans, this second region had
been previously shown to drive the transcription of the B hER mRNA
isoform (12A ). Comparison of this locus in the two species
revealed a sequence TAGATT found at the same position as the TATA box
motif in the hER B promoter (Fig. 1). Similarly, a CAAAC sequence
was found in the same location as the CCAAC motif in hER B promoter
(Fig. 1). Because these motifs and their relative locations are
accurately conserved in both species, they probably have functional
roles in the initiation of transcription of mRNA B isoform. For the
cER gene, transcriptional initiation was indeed detected downstream
of these motifs by an S1 nuclease mapping experiment (see Figs. 1 and 6B, probe A/B).
Two other variants of B cER mRNA were isolated subsequently by
RT-PCR using a cER mRNA B-specific 5'-primer (BI) with a 3' primer
(S2) common to all cER mRNA isoforms (see Figs. 2B and 5 for primer
locations; Fig. 3). These two isoforms were produced by the splicing
out the sequence between -153 and +154 (B'' variant) or between +51
and +135 (B' variant; Figs. 1 and 3). Although the cER mRNA B''
variant, corresponding to splicing out the sequence between -153 and
+154, was a minor form detected by RT-PCR, it was interesting to note
that in humans, a similar splicing event between exons 1B and 1A, which
allowed removal of the intronic regions -168 to +163, was responsible
for the generation of isoform B in the hER gene (Fig. 1). Sequence
comparison revealed the conservation of the consensus donor and
acceptor splice sites needed to generate B ER mRNA in humans and the
isoform B'' variant in chickens.
In contrast to the 5'-ends of A and B isoforms, the 5'-extremities of
the two other cER mRNAs (C and D) were not homologous to any of the
hER mRNA isoforms A–F. Comparison of the 63 bp of RACE sequence
specific for the 5'-end of the C isoform with the 3-kb of known
5'-cER genomic sequence upstream of the transcription start of
isoform A (18) revealed that mRNA isoform C is formed by the splicing
of a new exon (exon 1C) that ends at -1334 to the acceptor site at
+154 in exon 1A (Figs. 3 and 4). A first
transcription start site (at +88 in the C cER sequence of Fig. 3)
was determined using primer and probe C (whose locations are shown in
Fig. 4) in primer extension and SI nuclease mapping experiments (Figs. 3 and 6, A and B). However, the major activity detected by the SI
nuclease probe C corresponded to a specific protection of the total
cER sequence of the probe, indicating that the major start site was
missed. Therefore a new SI probe, covering a sequence more 5', was used
to map the major transcription start site (data not shown). This site
was found to be located 284 nucleotides upstream of the translation
start codon (+1 in the C cER sequence of Fig. 3). This transcription
start site was also observed after a longer exposure of the primer
extension using primer C (data not shown). TATA- and CAAT-like
sequences were located at 27 and 74 bp, respectively, immediately
upstream from the most distal 5'-extremity, as shown in Fig. 3.
The 5'-cDNA end sequence of D cER mRNA isoform was not found in the
known 3-kb genomic region upstream of exon 1A (Figs. 3 and 4). Southern
blot hybridization and genomic PCR experiments confirmed that exon 1D
was further 5' (data not shown). To isolate the 5'-flanking genomic
region of this message, a rapid genomic walking technique was
performed, as described in Materials and Methods. A 700-bp
fragment was amplified containing the 5'-portion of exon 1D plus
additional new 5'-genomic sequence (Fig. 3). Whereas RT-PCR analysis
showed that sequences directly upstream of the 5'-extremity of the RACE
product were not transcribed (data not shown), both primer extension
and SI nuclease mapping experiments using RNA from different tissues
failed to determine the exact location of the 5'-end of cER isoform
D [Fig. 6A (primer D) and data not shown]. These data suggest the
possibility that the transcription unit D contains an further upstream
exon(s) that was not isolated during the 5'-RACE experiment. Indeed,
sequence analysis of the 5'-flanking region of exon 1D showed potential
splice acceptor sites. Nevertheless, an alternative possibility comes
from the fact that the region 5' to exon 1D is a relatively strong
promoter in transfection experiments (see below), and sequence analysis
showed the presence of putative transcription factor-binding sites in
this region (Fig. 3). Theses results strongly suggested that in at
least some tissues, the initiation of transcription of D cER mRNA
isoform occurs close to the 5'-extremity of the RACE product (which we
currently assigned +1 in the D cER sequence in Fig. 3).
A variant of cER mRNA D was isolated by RT-PCR using a cER mRNA
D-specific 5'-primer (DI) with a 3'-primer (S2) common to all cER
mRNA isoforms and located in exon 1A (see Figs. 2B and 5 for primer
locations). It contained an extra 19-bp insertion between the sequence
of exons 1D and 1A (Fig. 3). The 3'-part of this extra sequence
corresponded to the 11 bp of divergent sequence isolated during
screening of the chicken oviduct cDNA Zap library (see Fig. 2A).
This new sequence was also not located within the known 3 kb of genomic
sequence. The fact that the two first 5'-nucleotides of the 19-bp
insertion were not consensus splicing donor sequence (GT) suggested
that these 19 extra nucleotides were not contiguous to exon 1D, but
constituted an independent additional exon, named exon 2D.
The new preliminary 5'-genomic organization for the cER gene is
summarized in Fig. 4.
Structures, distributions, and levels of new cER mRNA
isoforms
As the new 5'-cER mRNA extremities have been described as a
consequence of their linkage to exon 1A, it was important to determine
whether the corresponding cER mRNA transcripts were identical to the
previously described isoform A at the 3'-end, thereby encoding the
common A–F regions (16, 17, 18). To investigate this, a PCR analysis was
performed on single stranded cDNAs synthesized from total RNA from
various tissues using a cER gene-specific primer (I) chosen from the
A cER mRNA gene 3'-untranslated sequence (exon 8; Fig. 5). The different cER cDNAs were
amplified by two rounds of PCR using a common 3'-primer (II) and a
nested primer (III) located upstream from primer I in exon 8, in
combination with 5' primers and nested primers specific for each of the
different cER mRNA 5'-extremities. The sizes of the amplified cDNAs
were as expected, and after Southern blotting, the hybridization of
these PCR products with various probes recognizing specifically the
eight coding exons of the A cER mRNA isoform demonstrated that
sequences encoding the regions A–F were identical for all cER mRNA
isoforms (Fig. 5 shows only the results obtained with the
oligonucleotide probe IV specific for exon 1A). This study also showed
the presence of shorter specific PCR products that, after probing and
sequencing, were shown to be due to deletions of exon 2 or exon 4.
This analysis also revealed the pattern of distribution of the
different cER mRNA isoforms in various estrogen-responsive tissues
from laying hen and adult cock. As shown in Fig. 5, all full-length
isoforms (A–D) could be detected in oviduct, ovary, and liver from
males and females and from testis. In contrast, B and C full-length
mRNA isoforms were absent from kidney, and the B isoform was not
detected in lung tissue.
As RT-PCR analysis does not allow the quantitative estimation of the
expression level of each different cER mRNA isoform in the various
RNA samples, primer extension and SI nuclease mapping analysis were
performed. The long primers used in the primer extension experiment
were designed to be either specific to each 5'-end of the cER mRNAs
to identify precisely the transcription start site of the cER
transcripts (primers A/B, C, and D, as shown in Fig. 4) or to be
complementary to a region common to all cER messages, which thus
allowed extension to all 5'-extremities (primer T, as diagrammatically
illustrated in Fig. 4). To compare directly the extension products of
the long primer T with the other primers used (primers A/B, C, and D),
all long primers were designed to possess the same first cER
sequences. In this way, a tissue distribution profile of the
combination of the various cER mRNA isoforms could be analyzed. As
shown in Fig. 6A, the results of this
experiment indicated that the cER mRNA A was the main form expressed
in the tissues where a specific extension of primer T was observed.
Unfortunately, as the signal due to the A cER mRNA isoform
(transcription start site and the incompletely extended products) was
much higher than those produced by the other isoforms, it thus masked
the accurate determination of the weaker signals and prevented proper
analysis of the pattern of expression of the other cER mRNA isoforms
in the different tissues tested. Hence, it was decided to use the
alternative SI nuclease mapping approach to ascertain this
analysis.
The SI probes (probes A/B, C, and D, as shown in Fig. 4) were designed
to be specific for each 5'-extremity of the different cER
transcripts. Moreover, due to their common region 3' to the splice site
position, each of these probes was also able to measure the residual
expression resulting from the sum of the expression of the other cER
isoforms. To distinguish between undigested probes and specific
protected fragments, all of the S1 probes contained additional
sequences in their 3'-extremities that came from the vectors used for
single strand probe preparation. Qualitative analysis of the SI
nuclease mapping results showed that the alternative splicing event,
which occurs 70 nucleotides upstream of the translational initiation
codon, takes place in tissues other than the oviduct (see large
asterisk in Fig. 6B). It also provided confirmation of the
splicing out of the sequence between +51 and +135 that takes place in
one of three B isoforms, as a protected fragment was detected by the S1
probe A/B at the position of the acceptor site of this splicing event
(see small asterisk in Fig. 6B).
The results of the SI nuclease mapping experiments were quantified by
densitometry and expressed as a percentage of the total cER mRNA
expression detected by each probe in oviduct tissue (Fig. 7). The cER A isoform represents the
major cER mRNA expressed in reproductive tissues, accounting for
approximately 40–60% of the total cER mRNA level, whereas its
expression was much lower in nonreproductive tissues. The B cER RNA
isoforms were a minor component in oviduct, liver, and ovary tissues.
Transcript C was one of the major forms expressed in the kidney (51%),
whereas lower levels (11–22%) of expression were measured in other
tissues, with the exception of lung and testis, where it was not
detected. Finally, all activities detected by the D cER
isoform-specific probe were found at the splice site position, 70
nucleotides upstream of the translational initiation codon, and none
included protected D cER mRNA-specific sequences (data not shown).
Likewise, a similar SI nuclease mapping pattern arose using a probe
(probe D'; as shown in Fig. 4) specific for the second cER mRNA D
variant that contained an additional 19 bp between exons 1D and 1A
(data not shown). We therefore concluded that D forms were only very
minor messages at least under the conditions studied. When summarized,
the data also revealed that the level of expression of A–D cER
mRNAs does not account for the total cER mRNA expression detected by
the probes A/B, C, and D in the different tissues tested. This
indicates the likely existence of further and as yet unidentified
cER mRNA isoforms.
|
mRNA A, B, C, and D transcription units mRNA A, B, C, and D exhibited promoter activity,
cER genomic DNA fragments of these regions were inserted adjacent to
the luciferase gene in the pGL2 basic vector, as described in
Materials and Methods, thus generating the reporter vectors
pGL2 pA, pB, pC, and pD, respectively. These vectors were transiently
transfected into the CEF and hepatocellular carcinoma (LMH) cells
together with the expression vector pSG5, which contained the cER
cDNA (pSG5 cER), or not. It should be noted that in other studies,
low levels of cER were detected in LMH cells (23), whereas CEF cells
do not contain any detectable endogenous ER (29). The transfection
conditions were first checked in both cell types using a CAT reporter
gene driven by the estrogen-responsive promoter of the apoVLDLII gene
(Apo-CAT) (20). As shown in Fig. 8, the
results obtained with this chimeric gene reproduce the previously
reported capability of cER to trans-activate the
apoVLDLII promoter in the presence of hormone (20).
|
). In LMH cells, promoter A resulted in a 7-fold
induction, whereas no detectable increase in luciferase activity was
observed for any of the other promoter constructs (pGL2 pB, pC, or pD;
Fig. 8). We concluded from these results that promoters A–D were able
to drive the transcription of the reporter gene in a cell-specific
manner.
Also of interest is the fact that cotransfection of the cER
expression vector (pSG cER) in CEF cells resulted in a 3- to 4-fold
estradiol-dependent decrease in promoter A, B, and D
activities (Fig. 8). In LMH cells, a similar repressing effect in the
presence of hormone and receptor was detected for promoter A. These
results indicated that cER A, B, and D promoters may be subjected to
autoregulation despite the fact that computer-assisted analysis (30) of
all promoter sequences (Figs. 1 and 3) showed the absence of consensus
estrogen response elements. In contrast, two putative transcription
factor activating protein-1 (AP-1) target sites were found in
cER promoter sequences. The first was identified in the proximal
promoter B (see Fig. 1), whereas the second was located 81 nucleotides
upstream of the assigned transcription start site of the D mRNA (see
Fig. 3). These data were of particular interest given the fact that in
addition to its capability to bind to the estrogen response element, ER
can act in a protein/protein-related manner with the AP-1 complex (31)
and thereby potentially autoregulate its expression. To investigate the
functionality of these two putative target sites, the AP-1 sensitivity
of pGL2 pA (the AP-1 site located in the proximal promoter B was also
contained in pGL2 pA construction), pB, and pD reporter genes were
tested by adding the phorbol ester tetradecanoylphorbol 12-myristate
13-acetate (0.1 µM) to CEF cell culture medium.
For the three reporter genes, this led to a 3- to 5-fold induction of
luciferase activity (data not shown). However, as a similar fold
increase was seen with the control promoterless pGL2 basic vector, we
concluded that the observed effect was probably nonspecific, although
such comparisons of fold increases are open to difficulties in view of
the extremely low basal levels of the promoterless construct.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
gene has been shown to be
a complex genomic unit, exhibiting alternative splicing and consequent
differential promoter usage similar to that observed for the human and
rat ER genes (12A, 15). We report the isolation and characterization
of three new chicken ER mRNA isoforms (isoforms B, C, and D) in
addition to the previously described form A (16, 17, 18). Whereas cER
mRNA B variants presents a 5'-extremity contiguous to the 5'-end of
isoform A, the two other forms are generated by alternative splicing of
upstream exons (1C and 1D) to a common site, situated in the 5'-UTR of
exon 1A, 70 nucleotides upstream of the translation start site.
Therefore, similar to those of the mammalian species, all cER mRNA
isoforms encode a common protein composed of regions A–F, but diverge
in their 5'-UTR sequences. During the course of this study, other
cER mRNA isoforms were also detected that result from single exon
deletions within the protein-coding region of the message. For the
hER gene, a similar pattern of alternative splicing within the
primary ER transcript have been documented previously (32).
Analysis of 3 kb of known 5'-cER gene sequences upstream of the
transcription start site of the A isoform revealed that exon 1C ends
from -1334 relative to the major start point of transcription for
isoform A. To date, the genomic location of exon 1D is not known.
A high degree of homology in the human 5'-UTR sequences of the A and B
ER mRNA isoforms was found between the chicken and the human,
suggesting that a functional role for these regions has been preserved
during evolution. Sequence alignment of the 5'-UTR of cER mRNA C and
D isoforms with the known ER sequences from other mammalian species
showed no significant homology. The 5'-untranslated sequences of each
of the new cER cDNAs contain at least one ATG triplet followed by a
short ORF (sORF). This unusual feature, which has been also described
for some, but not all, of the hER mRNA isoforms (12A ), is found in
less than 10% of the vertebrate mRNAs characterized (33). The
significance of these sORFs in the different cER mRNA 5'-UTRs
remains to be elucidated, but similarly placed sORFs in other messages,
such as GCN4 and BCR/ABL oncogene mRNA, have been shown to be involved
in the translational control of their expression (34, 35). This
information suggests that alternative splicing might play a role in
regulating the levels of the cER by altering the turnover and/or the
translation efficiency of the encoded mRNAs.
The results on the distribution of the cER mRNA isoforms show a
differential pattern of expression of the cER gene in the chicken
tissues studied. As expected, the highest amount of cER mRNA was
detected in the oviduct. A relatively high amount was also present in
the other chicken female reproductive tissues tested, such as liver and
ovary. Liver is one of the main estrogen target tissues in oviparous
species, as the synthesis of vitellogenin, a precursor of the major
egg-yolk protein, is under the control of estradiol (36).
It is also worth noting that a small, but noticeable, amount of cER
mRNA was detected in nonreproductive and male tissues (lung, kidney,
and testis). Isoform A is the predominant message in the oviduct
(60%). This is in agreement with the fact that A cER cDNA was the
main form isolated from a cDNA library prepared from laying hen
oviduct. In liver and ovary, its level was reduced to 30–40% of the
total cER mRNA level and reached a low level in the nonreproductive
tissues. The tissue distribution pattern of the A isoform in the
chicken parallels that observed in the human, as A hER mRNA was also
the predominant isoform in female reproductive tissues (except ovary)
and was a minor form in the other tissues, such as liver, which is not
considered to be involved in mammalian reproduction (12A ). In the
chicken, the B isoform displayed the same profile as the A variant,
albeit at much lower levels. Likewise, the analogous hER B mRNA
isoform represented in the best case 11% of the total hER message
and maintained a tissue distribution pattern similar to that of the A
variant in all hER-positive tissues (12A ). One possible explanation
for this observation is that the two proximal promoters (pA and pB)
that control the respective expressions of A and B ER messages are
located very close to each other, and probably elements involved in the
regulation of one of these promoters will have an impact on both. ER
has been previously detected in avian kidney (37), and here it was
demonstrated that the C cER mRNA isoform was the main cER
transcript in this tissue, where it accounted for approximately half of
the total receptor expression detected. Isoform C also contributed
significantly to the female reproductive tissues, but was absent in
lung and testis. D cER mRNA expression was not detected in this or
any other tissue analyzed, using either SI nuclease mapping or primer
extension analysis, under the experimental conditions tested. However,
we showed, using the more sensitive RT-PCR method, that this transcript
is present at low levels in all chicken tissues tested. This result
does not exclude the possibility that it may be more highly expressed
in the presence of specific stimuli or at specific physiological stages
in some tissues not included in this study. Finally, it should be noted
that in most of the tissues analyzed, the sum of the expression levels
of the different cER mRNA isoforms identified does not account for
the total cER mRNA expression level detected with the different S1
probes tested. This is obvious for both lung and ovary tissues, where
roughly half (45–66%) of the total cER mRNA expression is
unexplained. This observation may result from the fact that the RACE
analysis was not performed on RNA from all tissues tested by SI
nuclease mapping, but only on oviduct and female liver RNA. Further
5'-RACE experiments are obviously required to identify and characterize
these putative cER mRNA isoforms.
Given the tissue distribution data, it is expected that the various
cER mRNA isoforms are controlled by different promoters in a
cell-specific manner. Transfection experiments confirmed this
hypothesis, but yielded the unexpected result that promoters B and D
were the most active promoters in CEF cells, and no or very low
promoter activity was detected in LMH cells, although these last cells
are reported to contain ER (23). The detection of very low levels of
expression of the reporter gene in LMH cells and the relative levels of
activity in the CEF cells could of course be due either to the origin
and physiological state of the cells or to the region of DNA used to
test the promoter activity, as it may contain cell-specific enhancer or
silencer elements.
The transfection results also demonstrated that estrogen down-regulated
the activities of A, B, and D promoters in the presence of the cER
protein. This supports the observation by Maxwell et al.
(17), that estrogen administration led to a decrease in chicken oviduct
ER mRNA levels in vivo. Analysis of promoter sequences
failed to identify any consensus estrogen-responsive elements or
conserved sequences between the different promoters, with the exception
of two putative transcription factor AP-1 target sites in the proximal
regions of promoters B and D. Therefore, these two sites appear to be
possible target sites for ER action (31), and the fact that promoters B
and D were the most active in CEF cells suggested a particular role for
AP-1 in this experiment. A preliminary experiment that was designed to
assess this hypothesis, however, was not conclusive, inasmuch as the
promoterless control showed an altered level of activity comparable to
the relative increase in the more highly active cER promoters.
Further studies, such as analysis of deoxyribonuclease I-hypersensitive
sites, followed by a more detailed promoter characterization using the
in vivo footprinting technique, would be informative in
identifying sequences involved in the cell-specific expression of the
cER mRNA isoforms. Conceivably, other cis-elements that
exist upstream of promoters A–D may also provide a mechanism for
hormone- and tissue-specific factors to regulate each isoform of cER
mRNA in a particular cell/tissue type or at specific stages of
development. Multiple promoters and differential splicing are
frequently used mechanisms to create diversity and flexibility in the
regulation of gene expression for other members of the
steroid/thyroid/retinoic acid receptor family (38, 39, 40). This complexity
would account in some part for the tissue and developmental differences
in the receptor level and thereby give rise to the pleiotropic effects
of their corresponding ligands in a wide range of physiological
processes.
|
Footnotes |
|---|
Received June 26, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and ß. Endocrinology 1997 138:863–870
and ß. Endocrinology 138:863–870 gene are generated by alternative splicing and promoter
usage. Mol Endocrinol, in press
-C
cER
