0013-7227/00/$03.00/0 Endocrinology Copyright © 2000 by The Endocrine Society
Vol. 141, No. 2 Printed in U.S.A.
Two Estrogen Receptor (ER) Isoforms with Different Estrogen Dependencies Are Generated from the Trout ER Gene* ¨ L ME ´ TIVIER, GILLES FLOURIOT, FARZAD PAKDEL, RAPHAE YVES VALOTAIRE
AND
Equipe d’Endocrinologie Mole´culaire de la Reproduction, UPRES-A Centre National de la Recherche Scientifique 6026, Universite´ de Rennes I, 35042 Rennes, France ABSTRACT A characteristic of all estrogen receptors (ER) cloned from fish to date is the lack of the first 37– 42 N-terminal amino acids specific to the A domain. Here we report the isolation and characterization from trout ovary of a full-length complementary DNA (cDNA) clone encoding an N-terminal variant form of the rainbow trout ER (rtER). Sequence analysis of open reading frame of this cDNA predicts a 622-amino acid protein. The C-terminal region of this protein, from amino acid position 45 to the end, was very similar to the previously reported rtER (referred to as the short form, or rtERS). In contrast, this novel rtER cDNA (referred to as the long form, or rtERL) contains an additional in-frame ATG initiator codon that adds 45 residues to the N-terminal region of the protein. This new N-terminal region may represent the A domain of ER found in tetrapod species. The first 227 bp of this new cDNA were similar to the 3⬘-end intronic sequence of the rtER gene intron 1. These data together with S1 nuclease, primer extension, and RT-PCR experiments demonstrate that the rtERL
represents a second isoform of rtER that arises from an alternative promoter within the first intron of the gene. Transcripts encoding both rtER forms were expressed in the liver. In vitro translation of the rtERL cDNA produced 2 proteins with molecular masses of 71 and 65 kDa, whereas rtERS cDNA produced 1 65-kDa protein. Interestingly, Western blot analysis with a specific antibody against the C-terminal region of rtER revealed 2 receptor forms of 65 and 71 kDa in trout liver nuclear extracts, in agreement with the presence of the 2 distinct classes of rtER messenger RNA in this tissue. Functional analysis of both rtER isoforms revealed that although rtERS consistently exhibited a basal (estrogen-independent) trans-activation activity that could be further increased in the presence of estrogens, the novel isoform rtERL is characterized by a strict estrogen-dependent transcriptional activity. These data suggest that the additional 45 residues at the N-terminal region of rtERL clearly modify the hormoneindependent trans-activation function of the receptor. (Endocrinology 141: 571–580, 2000)
E
In mammals, in addition to ER␣ and - forms generated from two distinct genes, several isoforms of these two subtypes were recently reported. These are transcribed either by alternative exon splicing or usage of different promoters of a single gene (6 –10). The existence of multiple forms of the receptor may explain the pleiotropic actions of estrogens in diverse tissues or species. In fish, estrogens play a key role in gonadal sex differentiation, liver vitellogenesis, and fat metabolism. Although the liver and the ovary appear to be major targets of estradiol, the liver from previtellogenic female fish contains much higher ER levels than the ovary (11). For this reason, ER complementary DNA (cDNA) from several fish species were cloned from the liver and, in all of them, the A domain was lacking compared with that in mammals, birds, or amphibian ER (11–15). We report here the cloning and functional characterization of a new ER isoform possessing an extra 45 amino acid residues in-frame with the previously cloned rainbow trout ER.
STROGENS play crucial functions in growth, differentiation, and homeostasis of male and female reproductive organs, but also in nonreproductive tissues, such as bone, liver, and cardiovascular system (1). The effects of estrogens are mediated by at least two receptors, namely estrogen receptor ␣ (ER␣) and ER, which are members of the nuclear receptor family (2). Ligand-activated ER regulates gene expression directly by binding to a specific ciselement called the estrogen-responsive element (ERE) or indirectly by interaction with other transcription factors implicated in specific gene transcriptional pathways. On the basis of sequence homologies, ER was divided into six distinct regions (A–F), which have been shown to represent functionally independent domains. The ER trans-activation functions, namely AF1 and AF2, were mapped respectively at the N-terminal (A/B) and C-terminal (E/F) regions. Interactions between these regions and specific coactivators, possessing histone acetyltransferase activity may result in the modification of target gene expression (3–5). Received July 1, 1999. Address all correspondence and requests for reprints to: Dr. Farzad Pakdel, Equipe d’Endocrinologie Mole´culaire de la Reproduction, UPRES-A Centre National de la Recherche Scientifique 6026, Universite´ de Rennes I, 35042 Rennes Cedex, France. E-mail:
[email protected]. * This work was supported by the Centre National de la Recherche Scientifique and a fellowship from Ministe`re de l’Enseignement Supe´rieur et de la Recherche (to R.M.).
Materials and Methods Preparation of messenger RNA (mRNA) Total RNA was extracted from liver and stage II ovaries (mean follicle diameter, 1.2 mm) of adult rainbow trout by the guanidium thiocyanatephenol-chloroform method. Polyadenylated RNA was purified by Dynabeads oligo(deoxythymidine) (DynAl, Chantilly, VA) according to the manufacturer’s protocols.
571
572
TWO DIFFERENT PROTEIN ISOFORMS FROM TROUT ER GENE
Construction and screening of a trout ovary cDNA library Two micrograms of polyadenylated RNA isolated from ovaries containing vitellogenic follicles with a mean diameter of 1.2 mm were used to prepare double stranded cDNA as described previously (16) with Amersham Pharmacia Biotech’s cDNA synthesis system (Arlington Heights, IL). The cDNA was ligated to the EcoRI/NotI linker (Pharmacia) and then phosphorylated by polynucleotide kinase for 45 min at 37 C. Double-stranded cDNA was fractionated by gel filtration on a Sepharose 4B column (Pharmacia). The largest fractions were pooled and ligated to gt10 vector, which had been dephosphorylated and digested with EcoRI (Stratagene, La Jolla, CA) at a molar ratio of 1:1 (cDNA:phage DNA). The ligation products were packaged in vitro as recommended by the manufacturer (Promega Corp., Madison, WI; Biotec’s packaging system). Recombinants (1.1 ⫻ 107) were obtained, and approximately 5 ⫻ 105 recombinant phages were screened. Duplicate filters (Hybond N, Amersham Pharmacia Biotech) were hybridized with either 32Plabeled rainbow trout ER (rtER) cDNA under high stringency hybridization conditions (17) or with a 600-bp SacI/SacI fragment of rat ER hormone-binding domain (18). With this latter probe, hybridization was performed with lower stringency conditions using 25% formamide instead of 50% as described by Thomas (17). Several positive clones were obtained, and one of them, O2–512, which was positive with both probes and contained a 3.5-kb cDNA fragment, was purified. The cDNA fragment was subcloned into the EcoRI site of pBluescript II KS (Stratagene) and sequenced on both strands. Nucleotide sequences were analyzed by the Genejockey computer program. The corresponding sequences have GenBank/EMBL Data Bank accession no. AJ242740 and AJ242741.
S1 nuclease assay and primer extension Modified S1 nuclease protection and primer extension procedures were used as previously described (19, 20). These two methods involve the use of biotinylated, single stranded DNA templates to prepare highly labeled, single stranded DNA probes or long primers by extension from specific primers with the T7 DNA polymerase in the presence of [␣-32P]deoxy-CTP (3000 Ci/mmol). These probes or long primers are then hybridized with the appropriate RNA sample and subjected to either S1 nuclease digestion or reverse transcriptase extension, respectively. The origin of probe A template was a genomic PCR product obtained by amplification of the regions from ⫺424 in intron 1 to ⫹508 in exon 2 (see Fig. 2A). The DNA fragment was subcloned in the PCRTM2.1 vector (Invitrogen, San Diego, CA) downstream of T7 and upstream of M13 reverse primer. A PCR reaction was then performed using a biotinylated T7 primer with M13 reverse primer. To generate the template used to make probe B, a PCR reaction was performed using pBluescript KS plasmid containing the entire rtER short form (rtERS) cDNA, with a biotinylated T7 primer and primer I (5⬘-CTATATAAGGGTGTGGACGAGG-3⬘). Finally, the template for the long primer ⌺ was obtained by PCR using the 5⬘-biotinylated primer III (5⬘-ATGTACCCTGAGGAGACACGCGGA-3⬘) and primer I. All biotinylated PCR products were bound to streptavidin-coated magnetic beads (DynAl) as recommended by the manufacturer, and the nonbiotinylated DNA strands were removed in 0.1 m NaOH. The S1 probes (A and B) and the long primer ⌺ were obtained by extending primer II (5⬘-TCTCCAGGTAGTATGACTGGCTGG-3⬘) annealed to the different biotinylated, single stranded templates. After elution of the single stranded DNA probes by alkaline treatment and magnetic separation, 105 cpm of the probe or primer were coprecipitated with 30 g total RNA and then dissolved in 20 –30 l hybridization buffer [80% formamide, 40 mm PIPES (pH 6.4), 400 mm NaCl, and 1 mm EDTA (pH 8)], denatured at 70 C for 10 min and hybridized overnight at 55 C. The S1 digestions and the reverse transcriptase extension were carried out, and the samples were electrophoresed through denaturing polyacrylamide-urea gels. Transcription start sites were identified by the PEETA (primer extension, electrophoresis, elution, tailing and amplification) method, as recently described (21). Briefly, the extension product of interest, obtained from primer ⌺ (see Fig. 2A) extension, was cut out and eluted from the gel with 500 l 0.5 m ammonium acetate, 10 mm magnesium acetate, 1 mm EDTA (pH 8), and 0.1% SDS followed by a Qiaex II gel
Endo • 2000 Vol 141 • No 2
extraction (QIAGEN, Valencia, CA) as recommended by the manufacturer. Poly(C) tail was added to the extension product at 37 C for 30 min using 50 U terminal deoxynucleotidyl transferase, the corresponding buffer [100 mm cacodylate buffer (pH 6.8), 1 mm CoCl2, and 0.1 mm dithiothreitol; Promega Corp.] and 4 mm deoxy-CTP in a final volume of 20 l. The reaction was stopped by heating for 10 min at 65 C. One microliter of the reaction was subsequently submitted to 35 rounds of PCR amplification using an oligo(dG) adapter-primer (5⬘-GGCCACGCGTCGACTAGTACGGGIIGGGIIGGGIIG-3⬘) and primer IV (5⬘GAGTTTTGGGGGGGAGATGACGTG-3⬘), which hybridizes between primer ⌺ and the exon 2 acceptor site. The PCR product was then separated by electrophoresis on a 1.5% agarose gel, eluted using a Qiaex II gel extraction kit (QIAGEN) and subcloned in the TA cloning vector pCRTM2.1 (Invitrogen). The DNA product was sequenced by the dideoxy chain termination method.
RT-PCR analysis RT and PCR amplifications were performed using the Titan one-tube RT-PCR kit (Roche Molecular Biochemicals). Briefly, 5 g total RNA from estradiol (E2)-treated trout liver were reverse transcribed (30 min at 50 C) using 25 pmol downstream oligonucleotide, P3 (5⬘-CGCTGTCCTGTGCTCCAGG-3⬘), located in exon 5 of rtER. Specific cDNAs were amplified using the common primer P3 in combination with either primer P1 (5⬘-ATTGTGATGCGAAGCCAGAT-3⬘), located in exon 1, or primer P2 (5⬘-TGTGAATGTGATGCTGGTCA-3⬘), located in exon 2a. PCR parameters were an initial denaturation (94 C for 2 min), followed by 40 cycles of 94 C for 30 sec (denaturation), 58 C for 30 sec (annealing), and 68 C for 45 sec (primer extension). Amplified products (1/10th of each reaction) were separated on 1% agarose gel electrophoresis and stained with ethidium bromide. The gel was also transferred onto a nylon membrane (Hybond N, Amersham Pharmacia Biotech) and hybridized with a labeled rtERS cDNA probe.
Expression of short and long forms of rtER (rtERS and rtERL) in rabbit reticulocyte lysate The in vitro translation reaction was performed using 1 g pBluescript II KS plasmids containing the entire coding region of both ER cDNAs and T7 RNA polymerase in a rabbit reticulocyte lysate (methionine depleted) with 20 Ci [35S]methionine. Reactions were carried out at 30 C for 60 –120 min, as recommended by the supplier (TNT Reticulocyte Lysate System, Promega Corp.). Translation products of synthetic mRNA were analyzed by SDS-PAGE and autoradiography.
Western blot analysis Western blots were performed with slight modifications of a previously described method (22). Routinely, 25–30 g protein in nuclear extracts from trout liver or whole cell extracts from yeast cells were fractionated on 10% polyacrylamide-SDS gels and transferred to Immobilon-P type polyvinylidene difluoride membranes (Millipore Corp., Bedford, MA). Blots were incubated for 12 h with the purified anti-rtER antibody (raised against C-terminal domains of rtER), diluted 1:100. Then the blots were incubated for 1 h with the secondary antibody, which was either goat antirabbit IgG-horseradish peroxidase phosphatase conjugated (dilution of 1:1000 of commercial stock, TEBU, Le Perray-en-yvelines, France) or goat antirabbit IgG-alkaline phosphatase conjugated (dilution of 1:5000 of commercial stock, Tropix, Bedford, MA). After several washes, the blots were revealed by either 4-chloro1-naphthol (Sigma) or the CSPD chemiluminescent detection system from a Tropix kit by autoradiography for 1–2 min. All incubations and manipulations were performed at 22⫺25 C as previously described (22).
Plasmid constructions, yeast expression, and -galactosidase assays The entire coding region of rtERL was amplified by PCR and inserted into the unique BamHI site of the yeast expression vector YEpucG, which was also used to express the entire coding region of rtERS as described previously (23). The reporter genes used in this study were 3EREc-CyclacZ (23) and FP3-EREp-Cyc-lacZ (24). Briefly, 3EREc-Cyc-lacZ reporter plasmid was constructed with three tandem copies of the consensus ERE
TWO DIFFERENT PROTEIN ISOFORMS FROM TROUT ER GENE linked to the yeast cytochrome (Cyc) proximal promoter located upstream the lacZ gene in the yeast vector PLGÆ178 (25). The reporter FP3-EREp-Cyc-lacZ was constructed with 100 bp of the trout ER gene promoter encompassing the major cis-acting element involved in the estrogen regulation of the gene and containing an imperfect ERE (26). This 100-bp fragment (FP3-EREp) was subcloned into the XhoI site of the Cyc promoter of PLGÆ178 yeast vector (24). All of the constructions were verified by sequencing. The yeast strain used in this study was BJ2168 (a leu2 trp1 ura3–52 prb1–1122 pep4 –3 prc1– 407 gal2; Yeast Genetic Stock Center, Berkeley, CA). As described previously (23), yeast cells were stably transformed with 2 g of each plasmid using a standard lithium acetate method and were selected by growth on complete minimal medium [0.13% dropout powder lacking uracil and tryptophan, 0.67% yeast nitrogen base, 0.5% (NH4)2SO4, and 1% dextrose]. -Galactosidase assays were performed as previously reported (23) in the presence of 17-estradiol or different hormones (for 4 h at 30 C) for the trans-activation assays. -Galactosidase activity was measured at 420 nm using the o-nitrophenyl -d-galactopyranoside substrate. The formation of colored product was quantified with a spectrophotometer. -Galactosidase activity was expressed in Miller units.
Results Cloning of a novel ER protein, rtERL
To identify a new subtype or isoform of rtER, a trout ovary cDNA library was constructed, and about 500,000 recombinants were screened with both the trout ER cDNA (11) and a 600-bp fragment corresponding to the hormone-binding domain of rat ER cDNA (18). After several rounds of screening with both probes, one clone strongly positive with both
573
probes was purified and analyzed. This clone, containing a 3.5-kb cDNA, was sequenced and showed exactly the same nucleotide sequence as the previously reported rtER cDNA (now referred to as rtERS), except the first 227 nucleotides. This novel cDNA form called rtERL (Fig. 1A), contains 171 nucleotides in the 5⬘-untranslated region and possesses one in-frame ATG codon, located at nucleotide 172, which adds 45 amino acid residues at the N-terminal region of the short isoform rtER (Fig. 1B). This cDNA open reading frame encodes a protein of 622-amino acid residues with a calculated molecular mass of 68.1 kDa. The complete cDNA sequence of rtERL was deposited at the EMBL database (accession no. AJ242740). rtERS and rtERL are generated from two distinct classes of transcripts
Previous work showed that rtERS is encoded by two transcripts, rtERS mRNA1 and mRNA2, generated from an alternative splicing that occurs between exon 1 and exon 2 (27) (Fig. 2A). This process does not modify the structure of the rtERS protein, as illustrated in Fig. 2A. The first 227 nucleotides of the novel rtERL cDNA, referred to as exon 2a, exhibited 100% sequence identity with the 3⬘-end of intron 1 of the trout ER gene (data not shown). To confirm the transcriptional initiation of a new class of rtER transcript within intron 1, primer extension and S1 nuclease assays were per-
FIG. 1. Two rtER cDNA isoforms differ in their 5⬘-nucleotide sequence regions. Nucleotide and deduced amino acid sequences of the first exon (exon 1 or 2a), a part of exon 2, and exon 10 of rtERL cDNA (A) and rtERS cDNA (B) are indicated. The ATG start codon for each isoform is shown in bold. The asterisk in A indicates the position of the transcription start site for rtERL mRNA determined by the PEETA method. The closed triangle indicates the donor splice site in exon 2a used to generate rtERS mRNA2 (see also Fig. 2A).
574
TWO DIFFERENT PROTEIN ISOFORMS FROM TROUT ER GENE
Endo • 2000 Vol 141 • No 2
FIG. 2. The novel rtERL mRNAs are generated from transcription start points located within the previously assigned intron 1. A, Schematic representation of rtER gene and transcripts. The positions of the transcription start points (tsp) for rtERS mRNAs (tsp ⫹1) and rtERL mRNAs (tsp ⫹1⬘) are noted. The first 227 nucleotides of rtERL cDNA are boxed as exon 2a in the previously defined intron 1 (30). S1 probes A and B were prepared for S1 nuclease mapping assays. For the primer extension analysis, a long primer complementary to part of exon 2 (primer ⌺) was used. B, Thirty micrograms of total RNA prepared from liver of untreated (⫺E2) or estradiol-treated (⫹E2) trout were hybridized to the labeled probes (left panel, probe A; middle panel, probe B) and then treated with S1 nuclease as described in Materials and Methods. The same amount of yeast RNA was used as a negative control. The S1 nuclease-resistant hybrids were separated on a 4% polyacrylamide gel adjacent to DNA mol wt markers and free probes (left and middle panels). The origins of the different protected fragments are schematically represented at the right of each panel: the transcription start sites of rtERL mRNAs (probe A, protected fragments from 500 – 600 bases in size) and rtERS
TWO DIFFERENT PROTEIN ISOFORMS FROM TROUT ER GENE
formed with RNA from liver, a tissue in which estrogens have been reported to have important regulatory functions, such as the induction of vitellogenin (28). The two single stranded DNA S1 probes, A and B, were designed to be specific for rtERL and rtERS mRNAs, respectively (see Materials and Methods for the preparation of the probes). These two probes shared a common region 3⬘ to the acceptor splice site of exon 2 (Fig. 2A). After hybridization of probe A with liver RNA and S1 nuclease digestion, several specific protected fragments were detected. No protected fragments were seen with yeast RNA, which was used as a negative control. The two smallest fragments (366 and 407 bp) corresponded to a protection of rtERS mRNA 1 and 2 that remained homologous to probe A as far as the two acceptor splice sites at the 3⬘-end of intron 1, and then diverged in their 5⬘-ends from probe A complementary sequences. Protected fragments, approximately 500 – 600 bases in size, indicated the existence of a distinct class of rtER transcripts (rtERL mRNAs) that are generated from intron 1 (Fig. 2B, probe A). It is worth noting that several transcription start sites were detected for this new class of transcripts. Likewise, the detection of a fragment corresponding in size to a total protection of rtER-specific sequences of probe A (without the 88 last nucleotides of the probe that correspond to the vector sequence) indicated the existence of rtERL transcripts that could start further in 5⬘. The results of the S1 nuclease mapping analysis with probe B showed, in addition to the expected protected fragment detected at the acceptor splice site of exon 2 (due to a partial protection of probe B by rtERL mRNA and rtERS mRNA 2, 366 bases in size), two other protected fragments that corresponded to the transcription start sites of rtERS mRNA 1 (Fig. 2B, probe B). The sizes of these fragments were 475 and 490 bases. The existence of two classes of rtER transcripts was strengthened by a primer extension experiment using primer ⌺, which was designed to hybridize rtER mRNAs in a region downstream from the splice site position in exon 2 and was thus able to be extended to all of the 5⬘-extremities of rtER transcripts. The results showed several extension products corroborating in size to the transcription start sites of rtERL mRNA (fragment of 584 bases), rtERS mRNA 2 (fragments from 470 – 490 bases), and rtERS mRNA 1 (fragments from 510 –530 bp) determined above by S1 nuclease assay (Fig. 2B, primer ⌺). The identity of the longest extension product was further confirmed by the PEETA (primer extension, electrophoresis, elution, tailing, and amplification) method (21). Although 10 nucleotides shorter, the sequence of this extension product was similar to the 5⬘-extremity of the novel rtERL cDNA clone. The existence of two classes of rtER transcripts was finally confirmed by RT-PCR analysis (Fig. 3). Single stranded cDNAs were synthesized from liver RNA using a rtER gene-specific primer (P3) located in exon 5. The
575
two rtER transcripts were amplified by PCR using specific 5⬘-primers for exon 1 (P1) or exon 2a (P2) in combination with the common downstream primer (P3) in exon 5 (Fig. 3). The size of the amplified cDNAs was as expected (Fig. 3B, left side), and Southern blot hybridization with labeled rtERS cDNA confirmed the specificity of these PCR products (Fig. 3B, right side). Additionally, the cDNA fragment amplified from P2/P3, which was subcloned and partly sequenced, showed a sequence identical to that of the rtERL cDNA clone (data not shown). Together, these results showed that trout liver expresses two classes of rtER transcripts potentially encoding two different rtER protein isoforms. Quantitative analysis of the S1 nuclease mapping and primer extension results indicated that rtERS mRNAs are the main rtER transcripts expressed in liver and that both classes of rtER mRNAs are up-regulated by estradiol (compare lanes Liver ⫺E2 and Liver ⫹E2 in Fig. 2B). rtERS and rtERL protein forms are produced in trout liver
As previously described (22), Western blot analysis performed with an antibody directed against the C-terminal region of rtER demonstrated in liver nuclear extracts two receptor forms, migrating as 65 and 71 kDa (Fig. 4A, lane 2). The expression levels of both receptors were virtually undetectable in the liver of untreated male trout (Fig. 4A, lane 1), whereas they were strongly induced by estradiol. Analysis of the proteins synthesized by in vitro translation from cDNA rtERL and rtERS in rabbit reticulocyte lysate also evidenced two protein bands migrating at the same mol wt as those detected by Western blot analysis in liver extract (Fig. 4B, lanes 2 and 3). Moreover, transformed yeast with an expression vector containing rtERS cDNA exhibited only a 65-kDa protein form (Fig. 4C, lane 2), whereas the rtERL transformed yeast expressed the 71-kDa protein as the major ER form (Fig. 4C, lane 3). These data indicate that both classes of rtER transcripts (rtERS and rtERL mRNAs) detected in the liver are indeed translated in this tissue, giving rise to different ER isoforms (Fig. 4). Hormonal specificity and transcriptional activity of the novel rtER isoform
As rtERL and rtERS differ in their N-terminal region, which is involved in the AF1 trans-activation function of ER, functional studies with both isoforms were performed in a cell context sensitive to AF1. Yeast cells were selected for this characteristic. The entire coding region of rtERL cDNA was subcloned into the yEpucG yeast expression vector and transformed in yeast cells containing an estrogen-responsive reporter gene construct (see Materials and Methods). These cells were grown in liquid culture in the absence or presence of different hormones (10⫺6 m) for 4 h at 30 C. Cells were
mRNA 1 (probe B, fragments of 475 and 490 bases), the positions of the acceptor splice site in exon 2 used to generate rtERS mRNA1 (probes A and B; **, protected fragments of 366 bases), and the acceptor splice site in exon 2a for rtERS mRNA2 (probe A; *, protected fragment of 407 bases) are also indicated. Primer extension was performed by hybridization of 30 g total RNA to the labeled long primer ⌺ complementary to exon 2 (as described in Materials and Methods). After RT, the extension products were separated on a sequencing gel and autoradiographed (right panel, primer ⌺). The same amount of yeast RNA was used as a negative control. Arrows indicate the positions of the extension products obtained for rtERL mRNA (fragment of size 584 bases), rtERS mRNA 2 (fragments from 470 – 490 bases in size), and rtERS mRNAs mRNA 1 (fragments from 510 –530 bases in size). The sequence of the highest mol wt fragment was determined by the PEETA method and confirms its specificity for rtERL mRNAs. ⬍ and ⬍⬍ indicate undigested S1 probe and the total protection of rtER-specific sequences, respectively.
576
TWO DIFFERENT PROTEIN ISOFORMS FROM TROUT ER GENE
Endo • 2000 Vol 141 • No 2
FIG. 3. RT-PCR analysis of the rtERS and rtERL mRNA isoforms. A, The figure illustrates diagrammatically the locations of the oligonucleotides used as primers (indicated as P1, P2, and P3) to amplify by RT-PCR the cDNA of rtERS and rtERL transcripts from E2-treated trout liver RNA. Approximate locations of oligonucleotides are shown by short arrows. Note that primer P3 is common to both transcripts, whereas P1 and P2 are specific for each mRNA isoform. The expected sizes of the amplified cDNAs from P1/P3 and P2/P3 are 969 and 903 bp, respectively. Different exons of rtER gene (1–10) and the positions of the translation initiation codons (ATG) in exon 2a and exon 2 are indicated. B, Ethidium bromide staining of the liver RT-PCR products amplified with P1/P3 or P2/P3 primers (left side). The PCR products were also hybridized in Southern experiments with labeled rtER cDNA probe (right side), confirming the specificity of the PCR products. Additionally, the cDNA fragment amplified from P2/P3 (fragment of 903 bp in size) was subcloned and sequenced. Its extreme 5⬘-end sequence corresponded to rtERL cDNA sequence. The standard markers loaded in lane M are indicated in base pairs at the left of each panel.
FIG. 4. Two protein isoforms of rtER are produced in vivo and in vitro. A, Twenty-five micrograms of liver nuclear extracts from an untreated (⫺E2, lane 1) or a 48-h E2-treated (⫹E2, lane 2) male trout were resolved on 10% SDS-PAGE, transferred to nitrocellulose, and revealed by specific anti-rtER antibody. B, Using the TNT reticulocyte lysate system from Promega Corp., in vitro transcription and translation were performed with 1 g Bluescript plasmid, empty (vector) or containing either rtERS or rtERL cDNA isoform used as template, T7 RNA polymerase, and 20 Ci [35S]methionine in a 25-l final reaction volume. Five microliters of radioactive translation product were denatured and subjected to a 10% SDS-PAGE. The gel was dried and autoradiographed for 2 h. C, Whole cell extracts were prepared from yeast cells expressing each rtER isoform. Twenty micrograms of extract from control (vector, lane 1), rtERS and rtERL transformed yeasts (lanes 2 and 3, respectively) were resolved on 10% SDS-PAGE and immunorevealed by specific anti-rtER antibody. The positions of standard proteins (in kilodaltons) are indicated at the right of each panel.
harvested and lysed, and -galactosidase activity was then measured. As shown in Fig. 5, the activation of reporter gene by the rtERL was estrogen specific, and as expected, no other hormones could stimulate transcriptional activity of the receptor (Fig. 5A). Similar results were observed for rtERS expressed in yeast Saccharomyces cerevisiae (Fig. 5B). How-
ever, interestingly, in the absence of exogenously added estrogens, rtERL did not exhibit any transcriptional activity (Fig. 5A), whereas rtERS consistently showed a significant transcriptional activity representing 10 –15% of the maximum activity obtained with estradiol (Fig. 5B, see also Fig. 6).
TWO DIFFERENT PROTEIN ISOFORMS FROM TROUT ER GENE
577
FIG. 5. Steroid specificity of the novel rtER isoform expressed in yeast. Yeast cells were stably transformed with both an expression vector containing either rtERL cDNA (A) or rtERS cDNA (B) and a -galactosidase reporter plasmid containing 3EREs. The recombinant cells were grown in suspension medium containing 10⫺6 M E2, estrone (E1), estriol (E3), diethylstilbestrol (DES), progesterone (Pg), dexamethasone (Dex), testosterone (Test), thyroid hormone (T3), 4-hydroxytamoxifen (OHT), ICI 164384 (ICI), or no hormone (EtOH). After 4 h of incubation, -galactosidase activity was measured and expressed as a percentage of the maximum activity obtained with E2. Values represent the mean ⫾ SD from at least three independent experiments. Asterisks indicate a significant difference between controls (EtOH) and hormonal treatments (*, P ⬍ 0.05; **, P ⬍ 0.01; ***, P ⬍ 0.001; by Student’s t test).
The data illustrated in Fig. 6 represents the dose-response curve obtained in the -galactosidase assay from recombinant yeast cells expressing either rtERS or rtERL after exposure to various estradiol concentrations. The lacZ reporter genes used for these assays were 3ERE-Cyc, containing three ERE linked to the yeast cytochrome c-proximal promoter (Fig. 6A), and FP3-EREp-Cyc, which consisted of a 100-bp fragment of trout ER gene containing an imperfect ERE linked to the yeast cytochrome c-proximal promoter (Fig. 6B). Both reporter genes showed very low -galactosidase activity in the absence of ER (⬍0.2 U; data not shown). The level of -galactosidase expression was increased 50- to 100-fold
(22 U, Fig. 6A; 13 U, Fig. 6B) by the presence of rtERS in the absence of hormone (Fig. 6, A and B). Upon addition of estradiol, the expression of both reporter plasmids was further enhanced to reach almost 250 U (Fig. 6A) or 40 U (Fig. 6B). In contrast, in the absence of E2, the rtERL isoform did not exhibit any basal transcriptional activity (Fig. 6), even with the reporter plasmid FP3-EREp-Cyc, whereas rtERS showed a very high E2-independent transcriptional activity (30% of E2 induction; Fig. 6B). However, in the presence of ligand, the stimulation of reporter gene expression by rtERL was roughly half that achieved by rtERS. These data indicate that although the novel rtER isoform (rtERL) is a potent
578
TWO DIFFERENT PROTEIN ISOFORMS FROM TROUT ER GENE
Endo • 2000 Vol 141 • No 2
trans-activator of estrogen target genes, this isoform exhibited a clearly different hormone-independent trans-activation function. Discussion
FIG. 6. Transcriptional activity of rtER isoforms tested on different reporter genes. Yeast cells containing either reporter plasmid 3ERECyc-lacZ (A) or FP3-EREp-Cyc-LacZ (B), expressing short (rtERS) or long (rtERL) isoforms, were grown in medium in the presence of increasing concentration of estradiol. After 4-h incubation, -galactosidase activity was measured in yeast extracts. Values represent the mean ⫾ SD from at least four independent experiments. Asterisks indicate a significant difference between rtERS and rtERL (P ⬍ 0.01).
Although the principal structural characteristics of the complete coding sequence of the first ER cloned from a fish (11, 29) showed a remarkable homology in the DNA- and ligand-binding domains with those of other species, the lack of the N-terminal A domain was also reported. This structure was confirmed by genomic clones (30), and more recently after the cloning of ER from several other fish species (12–15). To date, the function of this domain corresponding to the first 37– 42 amino acids of ER is unclear. However, this region is well conserved during evolution from amphibian to mammals (i.e. 87% similarity between human ER and chicken ER, and 60% similarity between human ER and Xenopus ER). In search of new forms of rtER, a rainbow trout ovary cDNA library was constructed and screened. We found a full-length cDNA in which the 5⬘-end nucleotide sequences (exon 2a) are completely different from the first exon of the previously reported cDNA clone (rtERS). The nucleotide sequences of the 5⬘ extremity of the novel cDNA possess an in-frame ATG initiator codon that adds 45 amino acid residues at the Nterminal extremity of the receptor. This region may therefore account for the A domain of ER in fish. As the comparison of this A domain of rtERL with that of tetrapod species shows no obvious homology (data not shown), one can assume that the conservation of this region within ER occurred late during the evolutionary process of the animal species. Accordingly, it would be of great interest to find out if this domain is conserved among fish species. Moreover, structural analysis of this domain could provide valuable information concerning its function in ER action. The sequence homologies and phylogenetic relationships of various ERs suggest that rtER should be considered as an ER␣-like receptor, and this is also true for most other teleost ERs described to date (14). It is not clear why screening of the trout ovary cDNA library with the hormone-binding domain (HBD) of rat ER cDNA did not allow to clone a trout ER form. However, different hypothesis can be formulated. As these experiments were performed under low stringent conditions, the HBD of trout ER-like might have lower sequence homology to rat ER HBD than that expected in comparison to this domain among ER␣ cloned from different species. In other words, the ligand-binding domain of ER may have been less conserved during evolution compared with the ␣-subtypes. In fact, phylogenetic analysis seems to indicate that ER␣ has been more conserved during evolution and that ER would represent a phylogenetically newer protein. However, the presence of ER subtype in fish cannot be excluded, because at least in anguillomorph species, -like ERs are expressed (13, 14). Thus, in future studies, the highly conserved DNA-binding domain of mammalian ER cDNA should be used to isolate this ER subtype in trout. Analysis of the protein synthesized from both rtER isoform cDNAs revealed two protein bands at 65 and 71 kDa, which were also detected in liver nuclear extracts using a specific antibody raised against the C-terminal region of the
TWO DIFFERENT PROTEIN ISOFORMS FROM TROUT ER GENE
receptor. Although the rtERS cDNA produces only a 65-kDa protein, expression of the novel rtERL cDNA in yeast S. cerevisiae demonstrated the production of a major 71-kDa protein that reacts strongly with the specific anti-rtER antibody. The hepatic endogenous 71-kDa protein form is probably encoded by the rtERL transcript, which is also present in this organ. However, we failed to detect ER proteins or transcripts in the ovary by classical assays such as Western or Northern blots (data not shown). Further experiments are needed to determine whether this is due to a lack of sensibility (probably much lower level of ER in this tissue) or if the receptor is only expressed during certain critical steps of ovarian maturation. On the other hand, unlike the liver, the ovary is not a homogenous tissue, and therefore, ER may be expressed only in some specific cells. To answer these questions and to investigate cell- and/or stage-specific expression of these two forms of rtER, in situ hybridization and/or immunocytochemistry with specific probes or antibodies for each ER form, S1 nuclease assays, and/or RT-PCR will be necessary. Yeast expression systems were used by several groups to demonstrate that steroid receptors can function as a liganddependent transcriptional activator as in mammalian cells (31–35). These data also indicated that the basic transcription apparatus has been conserved among eukaryotes ranging from yeast to mammals. A previous study from our laboratory showed that the short isoform of rtER expressed in yeast exhibited low (10 –20% of the total activity), but consistent, hormone-independent transcriptional activity compared with human ER␣ (23). As the N-terminal region of these two receptors exhibits very low homology, and the AF1 activity of ER seems to be predominant in yeast (36, 37), we have speculated that this could be a consequence of the differences in the structure/activity of AF1 functions within each receptor. Interestingly, expression of the novel rtERL in yeast showed very low, if any, hormone-independent activity compared with that of rtERS. In a recent work, the region responsible for rtER AF1 activity was mapped in the B domain, and in addition, we showed that further deletion of the A domain from human ER induces a ligand-independent activity, as found for rtERS. Therefore, these data suggest that the presence of the A domain at the N-terminal region of ER represses this hormone-independent activity of the receptor (Me´tivier, R., F. Petit, Y. Valotaire, and F. Pakdel, in preparation). Wahli and collaborators (38) showed that in the liver of Xenopus two forms of ER, a full-length and an N-terminal truncated version, were produced. These two isoforms are generated from a single mRNA species by use of two ATG initiator codons within the same reading frame. Interestingly, the N-terminal truncated ER form lacks the first 41 amino acids corresponding to the A domain. It is, however, unknown whether specific transcripts encoding this N-terminal truncated ER isoform also exist in Xenopus, as at least four ER mRNA species were detected in the liver (39). Recently, a study in chickens also showed the presence of an A domain-truncated ER generated from a distinct hepatic mRNA species (Gannon, F., personal communication). These data together with the results reported here demonstrate that in addition to a full-length ER, the existence in all oviparous
579
species of an N-terminal-truncated ER isoform lacking the A domain. This truncated ER exhibits significant hormoneindependent transcriptional activity compared with that of the full-length ER variant or with human ER␣ activity (23) (this study). Further experiments should be performed using naturally occurring estrogen-responsive reporter genes to elucidate whether these isoforms, which differ only in their N-terminal region, may activate target genes in a different manner. In parallel, it would be interesting to examine possible differential properties of these ER isoforms for their interaction with recently identified coactivators for steroid receptors (40 – 42; for review, see Ref. 3). Although the physiological significance of this truncated ER form is unclear, the fact that this isoform was only found in the liver of different oviparous species may implicate a role in vitellogenesis. The process of vitellogenesis, common to all oviparous species (birds, reptiles, amphibians, and fishes), is characterized by hepatic production of yolk lipoproteins and their massive deposition within the oocytes, thus providing the main nutritional reserves necessary for embryo development (28). Vitellogenin, a complex glycophospholipoprotein synthesized by the liver under estrogenic induction, is transported via the bloodstream to the ovary, where it is sequestered by growing oocytes in a cell-specific fashion. Therefore, in addition to its detoxification and biotransformation activity, the liver of oviparous vertebrates also represents a reproductive organ in which the ER gene is highly expressed. One can speculate that unlike mammals, in all oviparous animals the presence of a constitutively active ER form (rtERS) may direct the hepatic tissue toward its reproductive function during organogenesis. This hypothesis may be reinforced by the fact that the constitutive activity of rtERS can be stimulated by the orphan receptor COUP-TF1 (24), which plays fundamental roles in embryogenesis and the control of cell differentiation (43, 44). In vertebrates, COUP-TF activity has also been associated with the transcriptional regulation of several liverexpressed genes (45– 47), including the trout ER gene (24, 48). Acknowledgments We are grateful to Dr. O. Kah for reviewing the manuscript, and to Drs. B. S. Katzenellenbogen P. Chambon, and D. Metzger for providing the yeast expression vector and reporter plasmid. We also thank Drs. G. Kuiper and J. A. Gustafsson for rat ER cDNA.
References 1. Korach KS 1994 Insights from the study of animals lacking functional estrogen receptor. Science 266:1524 –1527 2. Katzenellenbogen BS, Korach KS 1997 A new actor in the estrogen receptor drama-enter ER-. Endocrinology 138:861– 862 3. Glass CK, Rose DW, Rosenfeld MG 1997 Nuclear receptor coactivators. Curr Opin Cell Biol 9:222–232 4. Grant PA, Sterner DE, Duggan LJ, Workman JL, Berger SL 1998 The SAGA unfolds: convergence of transcription regulators in chromatin-modifying complexes. Trends Cell Biol 8:193–197 5. Moras D, Gronomeyer H 1998 The nuclear receptor ligand-binding domain: structure and function. Curr Opin Cell Biol 10:384 –391 6. Friend KE, Ang LW, Shupnik MA 1995 Estrogen regulates the expression of several different estrogen receptor mRNA isoforms in rat pituitary. Proc Natl Acad Sci USA 92:4367– 4371 7. Schreihofer DA, Resnick EM, Soh AY, Shupnik MA 1999 Transcriptional regulation by a naturally occurring truncated rat estrogen receptor (ER), truncated ER product-1 (TERP-1). Mol Endocrinol 13:320 –329 8. Erenburg I, Schachter B, Mira R, Lopez Y, Lssowski L 1997 Loss of an estrogen receptor isoform (ER␣⌬3) in breast cancer and the consequences of its reex-
580
9. 10.
11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.
23. 24.
25. 26. 27. 28.
TWO DIFFERENT PROTEIN ISOFORMS FROM TROUT ER GENE
pression: interference with estrogen-stimulated properties of malignant transformation. Mol Endocrinol 11:2004 –2015 Chu S, Fuller PJ 1997 Identification of a splice variant of the rat estrogen receptor  gene. Mol Cell Endocrinol 132:195–199 Ogawa S, Inoue S, Watanabe T, Orimo A, Hosoi T, Ouchi Y, Muramatsu M 1998 Molecular cloning and characterization of human estrogen receptor cx: a potential inhibitor of estrogen action in human. Nucleic Acids Res 26:3505–3512 Pakdel F, Le Gac F, Le Goff P, Valotaire Y 1990 Full-length sequence and in vitro expression of rainbow trout estrogen receptor cDNA. Mol Cell Endocrinol 71:195–204 Tan NS, Lam TJ, Ding JL 1996 The first contiguous estrogen receptor gene from a fish, Oreochromis aureus: evidence for multiple transcripts. Mol Cell Endocrinol 120:177–192 Todo T, Shinji A, Yamauchi K 1996 Molecular cloning and characterization of Japanese eel estrogen receptor cDNA. Mol Cell Endocrinol 119:37– 45 Munoz-Cueto JA, Burzawa-Ge´rard E, Kah O, Valotaire Y, Pakdel F 1999 Cloning and sequencing of the gilthead sea bream estrogen receptor cDNA. DNA Sequence 10:75– 84 Xia Z, Patino R, Gale WL, Maule AG, Densmore LD 1999 Cloning, in vitro expression, and novel phylogenetic classification of a channel catfish estrogen receptor. Gen Comp Endocrinol 1133:60 –368 Gubler U, Hoffman BJ 1983 A simple and very efficient method for generating cDNA library. Gene 25:263–269 Thomas PS 1980 Hybridization of denatured RNA and small DNA fragments transferred to nitrocellulose. Proc Natl Acad Sci USA 77:5201–5205 Kuiper GGJM, Enmark E, Pelto-Huikko M, Nilsson S, Gustafsson JA 1996 Cloning of a novel estrogen receptor expressed in rat prostate and ovary. Proc Natl Acad Sci USA 93:5925–5930 Flouriot G, Nestor P, Kenealy MR, Pope C, Gannon F 1996 An S1 nuclease mapping method for detection of low abundance transcripts. Anal Biochem 237:159 –161 Flouriot G, Pope C, Kenealy MR, Gannon F 1997 Improved efficiency for primer extension by using a long highly labeled primer generated from immobilized single-stranded DNA templates. Nucleic Acids Res 25:1658 –1659 Flouriot G, Brand H, Gannon F 1999 Identification of differentially expressed 5⬘-end mRNA variants by an improved RACE technique (PEETA). Nucleic Acids Res 27:e8 Pakdel F, Petit F, Anglade I, Kah O, Delaunay F, Bailhache T, Valotaire Y 1994 Overexpression of rainbow trout estrogen receptor domains in Escherichia coli: characterization and utilization in the production of antibodies for immunoblotting and immunocytochemistry. Mol Cell Endocrinol 10:81–93 Petit F, Valotaire Y, Pakdel F 1995 Differential functional activities of rainbow trout and human estrogen receptors expressed in the yeast Saccharomyces cerevisiae. Eur J Biochem 233:584 –592 Petit F, Me´tivier R, Valotaire Y, Pakdel F 1999 Synergism between an half site estrogen-responsive-element (ERE) and an imperfect ERE and cooperation with COUP-TFI are required for estrogen receptor (ER) to achieve a maximal estrogen-stimulation of rainbow trout ER gene. Eur J Biochem 259:385–395 Guarente L, Mason T 1983 Heme regulates transcription of the CYC1 gene of S. cerevisiae via an upstream activation site. Cell 32:1279 –1286 Le Dre´an Y, Lazennec G, Kern L, Saligaut D, Pakdel F, Valotaire Y 1995 Characterization of an estrogen-responsive element implicated in regulation of the rainbow trout estrogen receptor gene. J Mol Endocrinol 15:37– 47 Lazennec G, Huignard H, Valotaire Y, Kern L 1995 Characterization of the transcription start point of the trout estrogen receptor-encoding gene: evidence for alternative splicing in the 5⬘ untranslated region. Gene 166:243–247 Wallace R A 1985 Vitellogenesis and oocyte growth in non mammalian vertebrates. In: Browder LW (ed) Developmental Biology, vol 1, chapt 3. Plenum Press, New York, pp 127–177
Endo • 2000 Vol 141 • No 2
29. Pakdel F, Le Guellec C, Vaillant C, Le Roux MG, Valotaire Y 1989 Identification and estrogen induction of two estrogen receptors (ER) messenger ribonucleic acids in the rainbow trout liver: sequence homology with other ERs. Mol Endocrinol 3:44 –51 30. Le Roux MG, The´ze´ N, Wolff J, Le Pennec JP 1993 Organization of a rainbow trout estrogen receptor gene. Biochim Biophys Acta 1172:226 –230 31. Metzger D, White JH, Chambon P 1988 The human estrogen receptor functions in yeast. Nature 334:31–36 32. Schena M, Yamamoto KR 1988 Mammalian glucocorticoid receptor derivates enhance transcription in yeast. Science 241:965–967 33. McDonnell DP, Pike JW, Drutz DJ, Butt TR, O’Malley BW 1989 Reconstitution of the vitamin D-responsive osteocalcin transcription unit in Saccharomyces cerevisiae. Mol Cell Biol 9:3517–3523 34. Mak P, McDonnell DP, Weigel NL, Schrader WT, O’Malley BW 1989 Expression of functional chicken oviduct progesterone receptors in yeast (Saccharomyces cerevisiae). J Biol Chem 264:21613–21618 35. Wrenn CK, Katzenellenbogen BS 1993 Structure-function analysis of the hormone binding domain of the human estrogen receptor by region-specific mutagenesis and phenotypic screening in yeast. J Biol Chem 268:24089 –24098 36. Berry M, Metzger D, Chambon P 1990 Role of the two activating domains of the oestrogen receptor in the cell-type and promoter-context dependent agonistic activity of the anti-oestrogen 4-hydroxytamoxifen. EMBO J 9:2811–2818 37. Metzger D, Losson R, Bornert JM, Lemoine Y, Chambon P 1992 Promoter specificity of the two transcriptional activation functions of the human oestrogen receptor in yeast. Nucleic Acids Res 20:2813–2817 38. Claret FX, Chapel S, Grace´s J, Tsai-Pflugfelder M, Bertholet C, Shapiro D, Wittek R, Wahli W 1994 Two functional forms of the Xenopus laevis estrogen receptor translated from a single mRNA species. J Biol Chem 269:14047–14055 39. Weiler IJ, Lew D, Shapiro DJ 1997 The Xenopus laevis estrogen receptor: sequence homology with human and avian receptor and identification of multiple estrogen messenger ribonucleic acids. Mol Endocrinol 1:355–362 40. Jacq X, Brou C, Lutz Y, Davidson I, Chambon P, Tora L 1994 Human TAFII30 is present in a distinct TFIID complex and is required for transcriptional activation by estrogen receptor. Cell 79:107–117 41. Cavaille`s V, Dauvois S, L’Horset F, Lopez G, Hoare S, Kushner PJ, Parker MG 1995 Nuclear factor RIP140 modulates transcriptional activation by the estrogen receptor. EMBO J 14:3741–3751 42. Onate SA, Tsai S Y, Tsai MJ, O’Malley BW 1995 Sequence and characterization of a coactivator for the steroid hormone receptor superfamily. Science 270:1354 –1357 43. Qiu Y, Tsai SY, Tsai MJ 1994 COUP-TF: an orphan member of the steroid/ thyroid hormone receptor superfamily. Trends Endocrinol Metab 5:234 –239 44. Qiu Y, Krishnan V, Pereira FA., Tsai SY, Tsai MJ 1996 Chicken ovalbumin upstream promoter-transcription factors and their regulation. J Steroid Biochem Mol Biol 56:81– 85 45. Gaudet F, Ginsburg GS 1995 Transcriptional regulation of the cholesteryl ester transfer protein gene by the orphan nuclear receptor apolipoprotein A1 regulatory protein-1. J Biol Chem 270:29916 –29922 46. Hall RK, Sladeck FM, Granner DK 1995 The orphan receptors COUP-TF and HNF-4 serve as accessory factors required for induction of phosphoenopyruvate carboxykinase gene transcription by glucocorticoids. Proc Natl Acad Sci USA 92:412– 416 47. Power S, Cereghini 1996 Positive regulation of the vHNF1 promoter by orphan receptors COUP-TFI/Ear3 and COUPTFII/ARP1. Mol Cell Biol 16:778 –791 48. Lazennec G, Kern L, Valotaire Y, Salbert G 1997 The nuclear orphan receptors COUP-TF and ARP-1 positively regulate the trout estrogen receptor gene through enhancing autoregulation. Mol Cell Biol 17:5053–5066
