Identification and Characterization of Two Novel Splicing Isoforms of ...

4 downloads 92782 Views 894KB Size Report
Dec 6, 2005 - Cardiovascular Research Center and Department of Biomedical ..... This new isoform, which we call .... The asterisk over placenta represents.
0021-972X/06/$15.00/0 Printed in U.S.A.

The Journal of Clinical Endocrinology & Metabolism 91(2):569 –579 Copyright © 2006 by The Endocrine Society doi: 10.1210/jc.2004-1957

Identification and Characterization of Two Novel Splicing Isoforms of Human Estrogen-Related Receptor ␤ Wei Zhou, Zhilin Liu, Jianbo Wu, Jing-hua Liu, Salman M. Hyder, Eric Antoniou, and Dennis B. Lubahn Departments of Biochemistry (W.Z., J.-h.L., D.B.L.), Animal Sciences (Z.L., E.A., D.B.L.), and Child Health (D.B.L.), Dalton Cardiovascular Research Center and Department of Biomedical Sciences (J.W., S.M.H.), University of Missouri Center for Phytonutrient and Phytochemical Studies (W.Z., J.-h.L., D.B.L.), University of Missouri, Columbia, Missouri 65211 Context: Estrogen-related receptor ␤ (ERR␤) was one of the first two orphan nuclear receptors reported and is believed to play important roles in estrogen-regulated pathways. Embryo lethality of ERR␤-null mice indicated that ERR␤ is essential for embryo development.

RT-PCR analysis showed that short-form hERR␤ has a wide distribution in the 24 of 27 human tissues and cell lines tested, whereas hERR␤2 and hERR␤2-⌬10 were only expressed in testis and kidney. The three human ERR␤-splicing isoforms have different transcriptional activities when measured on an estrogen response elementdriven luciferase reporter in transfection assays. The localization of a nuclear localization signal of short-form hERR␤ was also determined. Interestingly, the F domain of hERR␤2 alters the function of the nuclear localization signal. Therefore, the ERR␤ isoforms are likely to have diverse biological functions in vivo, and characterizing the three isoforms of ERR␤ will lead to an understanding of the multiple levels of gene regulation involved in steroid receptor-signaling pathways in humans and may provide novel therapeutic targets for human diseases. (J Clin Endocrinol Metab 91: 569 –579, 2006)

Objective: Two novel splicing isoforms of human (h) ERR␤, hERR␤2⌬10 and short-form hERR␤, were identified during the cloning of previously reported hERR␤-hERR␤2. We aim to investigate the functional differences of these three human ERR␤-splicing isoforms. Results and Conclusions: A genomic sequence comparison within and flanking the ERR␤ genes of eight species demonstrated that short-form hERR␤ lacks an F domain and is the matched homolog of mouse and rat ERR␤ proteins in humans. However, hERR␤2-⌬10 and the previously reported hERR␤2 isoforms are primate specific.

E

been shown that ERR␤ can repress the transcriptional activity of glucocorticoid receptor, but not progesterone receptor (PR) in a cell type-dependent manner (5). ERR␤ is also believed to be involved in estrogen-regulated pathways because it can bind the estrogen response element (ERE), activate transcription independent of exogenous ligands and share coactivators with ER␣ and -␤ (6, 7). It has also been shown that micromolar concentrations of tamoxifen and its active metabolite, 4-hydroxytamoxifen as well as diethylstilbestrol can suppress the transcription activity and coactivator interaction of ERR␤ (8 –10). Additionally, three isoflavones (genistein, daidzein, and biochanin A) and one flavone (6,3⬘,4⬘-trihydroxyflavone) were shown to act as agonists of ERR␣ and ERR␤ (11). Rat ERR␤ (NR3B2, GenBank accession no. X51417) was the first ERR␤ gene reported (2, 12). Mouse ERR␤ (no. X89594) was cloned in 1996 (13), encoding a polypeptide of 433 amino acids (aa), the same length as rat ERR␤. In 1999, Chen et al. (12) identified human (h) ERR␤2 (GenBank no. AF094517, reviewed as NM_004452 in GenBank) as the human homolog of ERR␤, which encodes a polypeptide of 500 aa, with 67 extra aa at the C terminus compared with rat ERR␤ and mouse ERR␤. In this study we report the identification and tissue distribution of two novel mRNA alternative splicing isoforms of human ERR␤, short-form hERR␤ and hERR␤2-⌬10 as well as the tissue distribution of the previously known isoform hERR␤2. Short-form hERR␤, which is identified for the first time, is the actual human ortholog of mouse ERR␤ (no. X89594) and rat ERR␤ (no. X51417). Transfection assays showed that these three splicing isoforms have different tran-

STROGENS HAVE IMPORTANT and diverse regulatory functions in mammals. Estrogen exposure is also a known risk factor for breast and uterine cancer. Traditionally, it has been thought that estrogen acts only through estrogen receptors ␣ and ␤ (ER␣ and ER␤; NR3A1 and NR3A2). However, another subfamily, the estrogen-related receptors (ERR), exists in the nuclear receptor superfamily and shares sequence similarity, target genes, coregulatory proteins, some man-made ligands, and action sites with the ERs (reviewed in Ref. 1). This subfamily contains three members: ERR␣, -␤ and -␥. Recent studies have shown that ERRs may play important roles in physiology and pathology. ERR␤ was one of the first two orphan nuclear receptors reported by Giguere et al. in 1988 (2) using reduced stringency hybridization and is essential for embryo development. Targeted disruption of the ERR␤ gene in mice resulted in severely impaired placental formation, and the embryo died at 10.5 d post coitum (3). ERR␤ is expressed during mouse mammary gland development, and the expression of the estrogen-inducible pS2 gene, a human breast cancer prognostic marker, can be regulated by ERR␤ (4). It has also First Published Online December 6, 2005 Abbreviations: aa, Amino acid; CMV, cytomegalovirus; E2, 17␤estradiol; ER, estrogen receptor; ERE, estrogen response element; ERR␤, estrogen-related receptor ␤; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; h, human; luc, luciferase; NLS, nuclear localization signal; ORF, open reading frame; PR, progesterone receptor; PRE, progesterone response element; SV40, simian virus 40; TK, thymidine kinase. JCEM is published monthly by The Endocrine Society (http://www. endo-society.org), the foremost professional society serving the endocrine community.

569

570

Zhou et al. • Novel Isoforms of Human ERR␤

J Clin Endocrinol Metab, February 2006, 91(2):569 –579

scriptional activities on an ERE-driven luciferase reporter, but do not affect the ability of PR to activate progesterone response element (PRE)-driven luciferase activity. The nuclear localization signal (NLS) of short-form hERR␤ was also determined. Surprisingly, the F domain of hERR␤2 inhibits the function of the NLS. Our results will allow researchers to study the appropriate tissue-specific isoform of human ERR␤ and better understand the regulation mechanisms of nuclear receptors by alternative splicing. Materials and Methods Materials Human testis total RNA was obtained from OriGene Technologies, Inc. (Rockville, MD). Human Total RNA Master Panel II, which includes 20 tissue RNAs: cerebellum, brain (whole), fetal brain, fetal liver, heart, kidney, lung, placenta, prostate, salivary gland, skeletal muscle, spleen, testis, thymus, trachea, uterus, colon with mucosal lining, small intestine, spinal cord and stomach, was obtained from BD Clontech (Palo Alto, CA). Human fetal heart total RNA, human adult ovary total RNA, human adult breast total RNA, human cervix total RNA, additional human kidney, uterus (postmenopause), skeletal muscle, and placenta total RNA were obtained from Stratagene (La Jolla, CA). A third placenta total RNA was purchased from BioChain Institute, Inc. (Hayward, CA). Human uterus (premenopause) total RNA was purchased from Chemicon International, Inc. (Temecula, CA). Total RNA of human breast cancer cell lines MCF-7 and C4-12-5 (14), human endometrial adenocarcinoma line Ishikawa, as well as the total RNA of mouse testis were prepared using Tri-Reagent (Sigma-Aldrich Corp., St. Louis, MO) following manufacturer’s protocols. All of the above anonymous human RNAs and human kidney tissue lysate described in the following Western blot section were properly handled according to the University of Missouri-Columbia institutional review board. 17␤-Estradiol (E2) and progesterone were obtained from Sigma-Aldrich Corp. Materials and reagents used in Western blot and immunofluorescence confocal microscopy will be described in detail in later sections.

Cloning of human ERR␤ isoforms by RT-PCR A two-step RT-PCR system was applied to clone the three isoforms of the hERR␤ gene. Briefly, the first RT step was carried out in a 10-␮l volume containing 1 ␮g total RNA, a final concentration of 1 mm dNTPs (Invitrogen Life Technologies, Inc., Carlsbad, CA), 20 mA260 units random hexamer primer (Roche, Indianapolis, IN), 20 U avian myeloblastosis virus reverse transcriptase (Roche), 12.5 U ribonuclease inhibitor (Roche), and 2 ␮l 5⫻ single-strength avian myeloblastosis virus reverse transcriptase buffer (Roche). The reaction was performed at 42 C for 60 min. The second PCR amplification step was carried out with the Expand High Fidelity PCR System (Roche) using primers specifically designed for different isoforms in a “touch-up-then-down” PCR program: 2 min at 94 C, then 20 cycles for 40 sec at 94 C, 40 sec at 65 C, with the temperature increasing by 0.5 C every cycle, 2 min at 68 C, followed by 25 cycles for 40 sec at 94 C, 45 sec at the final annealing temperature of 55 C, 2 min at 68 C, and a final extension step at 68 C for 7 min. The whole open reading frame (ORF) of short-form hERR␤ (GenBank accession no. AY451389) was cloned from human fetal heart total RNA (Stratagene) with primers hERRB2f261 (5⬘-act ttg agg cca gag gtg atc cag tga ttt-3⬘) and hERR2r1690 (5⬘-cgg tct gtc cgt ttg tct gtc tgt agg t-3⬘). The whole ORFs of hERR␤2 mRNA (GenBank no. NM_004452 or AF094517) and hERR␤2-⌬10 (GenBank accession no. AY451390) were cloned from human testis total RNA (OriGene Technologies, Inc.) with primer pair hERRB2f261 and hERRB2r2057 (5⬘-gcc aga tgt tac atg gtg agc cag aga t-3⬘). Negative controls were performed using RNA without the reverse transcriptase step. Amplified DNA products were ligated into pGEM-T vector (Promega Corp., Madison, WI), and clones corresponding to each isoform were selected by checking the insert fragment sizes. Finally, sequences of each isoform clones were verified using the 377 DNA Sequencer with BigDye version 3.1 chemistry (Applied Biosystems, Foster City, CA) at the DNA core of University of Missouri-Columbia.

Determination of tissue-specific mRNA isoform expression To detect the distribution patterns of short-form hERR␤, hERR␤2, and hERR␤-⌬10 in different human tissues/cell lines, two-step RT-PCR was used again. Six pairs of primers were applied separately to each RT reaction product to investigate the distribution patterns of the three isoforms: hERRB2f963 (5⬘-acc aag att gtc tca tac cta ctg gt-3⬘) and hERRB2r1265 (5⬘-ctc ctc atc cat gat gta gtc ctc-3⬘) were used to detect the existence of hERR␤ (any isoform) in targeting tissue/cell line, hERRB2f1334 (5⬘-gct caa ggt gga gaa gga gga gtt tgt g-3⬘) and hERRB2r1868 (5⬘-ctt gac att ctt tca tcc ttg gga gat cct-3⬘) were used to detect the existence of both hERR␤2 and hERR␤2-⌬10; hERR2f1328 (5⬘-caa gaa gct caa ggt gga gaa gga gga g-3⬘) and hERR2r1690 were used to specifically amplify short-form hERR␤; hERRB2f1607 (5⬘-ctt cct gga gat gct gga ggc caa ggt t-3⬘), which covers the boundary of exons 9 and 11, and hERRB2r2151 (5⬘-tct gct aga ggg gct ctg aag tga ggt c-3⬘) were used to detect hERR␤2-⌬10 specifically; hERRB2f1565 (5⬘-cta tag cgt caa act gca ggg caa agt g-3⬘) and hERRB2r1833 (5⬘-ctg ctc ttg gcc aac ctg ccc tct-3⬘), which covers the boundary of exons 10 and 11, were used to specifically amplify hERR␤2; glyceraldehyde-3-phosphate dehydrogenase (GAPDH) forward (5⬘-acc cac tcc tcc acc ttt g-3⬘) and GAPDH reverse (5⬘-ctc ttg tgc tct tgc tgg g-3⬘) were used as a positive control to amplify the housekeeping gene GAPDH. For total RNAs, including testis, kidney, placenta, uterus, and skeletal muscle from different sources, two pairs of primers designed for amplifying the N-terminal part of hERR␤ were also used in addition of above six pairs of primers: hERRBf261 (5⬘-act ttg agg cca gag gtg atc cag tga ttt-3⬘) and hERRBr753 (5⬘-ggc agc tgt act caa tgt tcc ctt gga t-3⬘); hERRB2f315 (5⬘-ctc aga ggg ctg ctg aac agg atg tc-3⬘) and hERRBr753 (5⬘-ggc agc tgt act caa tgt tcc ctt gga t-3⬘). The RT reaction was produced as described above. During the amplification step, each reaction tube contained 1 ␮l RT reaction product (equal to 0.1 ␮g total RNA), a final concentration of 0.2 mm deoxy-NTPs (Invitrogen Life Technologies, Inc.), 0.5 U Expand High Fidelity Enzyme Mix (Roche), and single-strength PCR buffer in a final volume of 10 ␮l. A touch-down PCR protocol was used: 2 min at 94 C, then 10 cycles for 20 sec at 94 C, 30 sec at 65 or 60 C, with the temperature decreasing by 0.5 C every cycle, 45 sec at 72 C, followed by 25 cycles for 20 sec at 94 C, 30 sec at the final annealing temperature (60 C or 55 C), 45 sec at 72 C, and a final extension step at 72 C for 7 min. Negative controls were performed using RNA without the reverse transcriptase step. The amplified DNA was fractionated electrophoretically on a 2% agarose gel, stained with ethidium bromide, and visualized under UV light.

In silico bioinformatics approach Genome Browser screenshots and genome sequences were taken from http://genome.ucsc.edu (15, 16).

Cell culture ER-negative MCF-7 cells, C4-12-5, were maintained in complete medium consisting of phenol red-free Eagle’s MEM (Sigma-Aldrich Corp.) supplemented with insulin (6 ng/ml), HEPES (10 mm), and 5% charcoalstripped calf serum (Invitrogen Life Technologies, Inc., Gaithersburg, MD). The wild-type PR-positive parental T47D breast cancer cell line was maintained in phenol red-free DMEM:Ham’s F-12 (Invitrogen Life Technologies, Inc., Carlsbad, CA), supplemented with 10% fetal calf serum (JRH Bioscience, Lenexa, KS). COS-1 cells were maintained in DMEM supplemented with 10% fetal bovine serum (Invitrogen Life Technologies, Inc.). Ishikawa cells were maintained in complete medium consisting of phenol red-free Eagle’s MEM (Sigma-Aldrich Corp.) with insulin (6 ng/ml) and HEPES (10 mm), supplemented with 10% fetal bovine serum (Invitrogen Life Technologies, Inc.).

Expression and reporter vector constructs pcDNA3.1⫹zeo short-form hERR␤ (AY451389), pcDNA3.1⫹zeo hERR␤2 (NM_004452), and pcDNA3.1⫹zeo hERR␤2-⌬10 (AY451390) were constructed by subcloning the related ORFs (described above in the RT-PCR section) from pGEM-T vector (Promega Corp.) into pcDNA3.1⫹zeo expression vector (Invitrogen Life Technologies, Inc.) with NotI and ApaI restriction sites. 5⬘-Myc (c-myc epitope: Glu-GlnLys-Leu-Ile-Ser-Glu-Glu-Asp-Leu)-tagged short-form hERR␤, hERR␤2,

Zhou et al. • Novel Isoforms of Human ERR␤

hERR␤2-⌬10, and short-form hERR␤ deletion constructs (Fig. 8; 101– 433, 169 – 433, 210 – 433, 1–210, and 1–169 aa) were constructed by standard molecular biology methods in pcDNA3.1⫹ vector (Invitrogen Life Technologies, Inc.). Mouse ERR␤ ORF (AF094518, reviewed as NM_011934) (13) was amplified from mouse testis total RNA using forward primer 5⬘-ctg aac cga atg tcg tcc gaa g-3⬘ and reverse primer 5⬘-aaa gaa aat gcg ggt gac aga tga g-3⬘. The mouse ORF was inserted into pGEM-T vector (Promega Corp.) and later subcloned into pcDNA3.1⫹zeo expression vector (Invitrogen Life Technologies, Inc.) by NotI and ApaI restriction sites. hERR␥ (NM_001438) ORF was amplified from human kidney multiple choice cDNA (Origene Technologies, Inc.) using forward primer (5⬘-gcg cgc tag cgc aca tgg att cgg tag aac ttt gc-3⬘) and reverse primer (5⬘-gcg cgg atc cgt cag acc ttg gcc tcc aac att tc-3⬘), then cloned into pcDNA3.1⫹zeo expression vector (Invitrogen Life Technologies, Inc.) by NheI and BamHI restriction sites. pCMX-hERR␣ (2) and vERE-thymidine kinase (TK)-luciferase (luc), which contains one copy of ERE (AGG TCA CAG TGA CCT; underlined nucleotides are core sequences of vERE), were provided by Dr. Vincent Giguere. The full length ER␣ expression vector has been described previously (17). The PvuII-SmaI fragment of pPRE/GRE.E1b.CAT was excised and inserted into the SmaI site of pGL3Basic from Promega Corp. pPRE/ GRE.E1b has two copies of the consensus PRE linked to the TATA element from E1b (provided by Dr. Zafar Nawaz, Creighton University, Omaha, NE). Renilla luciferase vector pRL-simian virus 40 (SV40)luc vector and pRL-cytomegalovirus (CMV)luc vector were obtained from Promega Corp.

Transient transfection and luciferase assay Two days before transfection, C4-12-5 cells were seeded in 24-well plates in phenol red-free medium, then transfected with different vectors as indicated using Plus and Lipofectamine reagents (Invitrogen Life Technologies, Inc.). Fifty nanograms of ER␣ and ERR expression vectors, 0.5 ␮g vERE-tk-luciferase reporter vector, and 20 ng Renilla luciferase control pRL-SV40luc vector (Promega Corp.) were used. After 12–16 h, transfected cells were then treated with ethanol vehicle (EtOH) and 100 nm E2, respectively. After 24-h incubation, cells were rinsed with PBS twice and lysed to measure the luciferase level using the dual luciferase assay kit (Promega Corp.). T47-D cells were transfected as follows. Cells were grown in DMEM supplemented with 10% fetal bovine serum and plated at 3 ⫻ 105 cells/ well in Falcon six-well dishes in 5% dextran-coated charcoal-stripped serum 24 h before transfection with the indicated plasmids using Superfect reagent (QIAGEN, Valencia, CA) according to the manufacturer’s guidelines. Cells were washed with PBS and incubated in DMEM: Ham’s F-12 and 5% serum in the presence of hormones as indicated. Cells were lysed after 20 h, and luciferase activity was measured using the dual luciferase reporter assay system (Promega Corp.). Experiments were performed in triplicate and repeated at least twice. Data were normalized to Renilla luciferase (pRL-CMV plasmid or pRL-SV40 plasmid, Promega Corp.) activity.

Western blot Anti-ERR␤ (rat ERR␤ aa 4 –100, GenBank accession no. X51417) mouse monoclonal antibody was purchased from R&D Systems, Inc. (Minneapolis, MN). Anti-actin (pan) polyclonal antibody was obtained from Cytoskeleton, Inc. (Denver, CO). Normal human tissue kidney lysate was obtained from ProSci, Inc. (San Diego, CA). C4-12-5 and Ishikawa cells cultured in six-well plates were transfected respectively with 1 ␮g empty vector pcDNA3.1⫹zeo, short-form hERR␤, hERR␤2, or hERR␤2-⌬10 using Plus and Lipofectamine reagents (Invitrogen Life Technologies, Inc.). Total cell lysates were collected 48 h later. Protein concentrations were determined by bicinchoninic acid protein assay kit (Pierce Chemical Co., Rockford, IL). Total cell lysate protein were then loaded on 10% SDS-PAGE gel and subjected to electrophoresis, membrane transfer, and Western blot. Western blot images were collected using a FujiFilm LAS-3000 imaging system (Tokyo, Japan).

J Clin Endocrinol Metab, February 2006, 91(2):569 –579

571

Immunofluorescence assay COS-1 or Ishikawa cells were grown as a single layer on glass coverslips in six-well plates. Cells were transfected with 1 ␮g (each) of short-form hERR␤, hERR␤2, hERR␤2-⌬10, 5⬘-Myc-tagged short-form hERR␤, 5⬘-Myc-tagged hERR␤2, 5⬘-Myc-tagged hERR␤2-⌬10, and 5⬘Myc-tagged short-form hERR␤ deletion construct expression plasmids using Plus and Lipofectamine reagents (Invitrogen Life Technologies, Inc.). Cells were fixed in 100% methanol for 10 min at ⫺20 C. Fixed cells were washed twice with 1⫻ PBS (Mediatech, Inc., Herndon, VA) and once with 1⫻ PBS containing 0.1% Triton X-100 (Sigma-Aldrich Corp.), 10 min each wash. Then cells were incubated with blocking buffer [1⫻ PBS containing 3% BSA, 3% goat serum, 0.1% Micr-O-Protect (Roche)] for 1 h at 37 C. After three 10-min washes in 1⫻ PBS containing 0.1% Triton X-100, cells were incubated with primary antibody, anti-ERR␤ monoclonal antibody (R&D Systems, Inc.) for short-form hERR␤, hERR␤2, and hERR␤2-⌬10 transfection; or anti-Myc polyclonal antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) for 5⬘-Myc-tagged short-form hERR␤, hERR␤2, hERR␤2-⌬10, or 5⬘-Myc-tagged short-form hERR␤ deletion construct transfection at a dilution of 1:100 in 1⫻ PBS containing 10% blocking buffer at 37 C for 1 h. Cells were washed three times with 1⫻ PBS (0.1% Triton-100) for 10 min each wash. Secondary antibody, AlexaFluor 488 goat antimouse IgG(H⫹L) (Molecular Probes, Sunnyvale, CA; Invitrogen Life Technologies, Inc.) for short-form hERR␤, hERR␤2, and hERR␤2-⌬10 transfection, AlexaFluor 568 goat antirabbit IgG(H⫹L) (Molecular Probes; Invitrogen Life Technologies, Inc.) for 5⬘-Myc-tagged construct transfection, was diluted at 1:50 in 1⫻ PBS containing 10% blocking buffer. Nuclear dye TO-PRO-3 (Molecular Probes; Invitrogen Life Technologies, Inc.) was also diluted at a final concentration of 10 ␮m with second antibody. Cells were incubated with second antibody and nuclear dye for 1 h at 37 C. Coverslips were washed three times in 1⫻ PBS (0.1% Triton X-100) and mounted on glass slides in Prolong Gold Antifade Reagent (Molecular Probes; Invitrogen Life Technologies, Inc.) and sealed with clear nail polish. At least 250 positive cells were scored for subcellular localization of short-form hERR␤, hERR␤2, and hERR␤2-⌬10. Images were obtained with a Bio-Rad Radiance 2000 confocal system (Bio-Rad Laboratories, Inc., Hercules, CA) coupled to an Olympus IX70 inverted microscope (Olympus, New Hyde Park, NY) at Molecular Cytology Core, University of Missouri-Columbia.

Results Identification of a novel human ERR␤ splicing variant form, hERR␤2-⌬10

During our RT-PCR amplification of the full length of hERR␤2 (GenBank accession no. NM_004452) ORF, we obtained two DNA bands (Fig. 1A). Sequence analysis of these two bands showed that although the larger band had the expected sequence of hERR␤2, the shorter band revealed a splicing jump from the ninth exon to the 11th exon (Fig. 1B). This variant splicing resulted in a frame shift and moved the stop codon to the 12th exon. This new isoform, which we call hERR␤2-⌬10, encodes a putative polypeptide of 508 aa (Fig. 1C). Its predicted last 76 aa, which start at the variant splicing site, are totally different from the last 68 aa of hERR␤2 because of a frame shift (Fig. 1D). Identification of short-form hERR␤

The new isoform, hERR␤2-⌬10, has an altered F domain that might potentially affect the ligand binding and AF-2 domains. Based upon this hypothesis, we began to look for its homolog in other species. The examination of human, chimpanzee, dog, rat, mouse, chicken, fugu, and zebrafish genome sequences within and flanking the ERR␤ gene revealed another unexpected result: exons 10, 11, and 12 in hERR␤2 do not have a comparable homologous region in the rat, mouse, chicken, fugu, or

572

J Clin Endocrinol Metab, February 2006, 91(2):569 –579

FIG. 1. Three splicing isoforms of human ERR␤. A, Amplification of the ORFs of hERR␤2, hERR␤2-⌬10, and short-form hERR␤. The ORFs of hERR␤2 and hERR␤2⌬10 were amplified from human testis total RNA using primer set hERRB2f261 and hERRB2r2057 (see Materials and Methods) in duplicate (lanes 1 and 2). The ORF of shortform hERR␤ was amplified from human fetal heart total RNA using primer set hERRB2f261 and hERR2r1690 (see Materials and Methods) in duplicate (lanes 3 and 4). B, Schematic diagram of hERR␤ variant mRNA splicing isoforms. hERR␤2 mRNA was drawn according to GenBank accession no. NM_004452. hERR␤2-⌬10 and short-form hERR␤ mRNAs were drawn according to the cDNA fragments amplified in A (GenBank accession no. AY451389 and AY451390). Numbers inside the rectangular boxes represent the exon numbers, and numbers below indicate the boundaries of exons and introns. Number 336 indicates the position of the start codon, and numbers 1836, 2046, and 1634 indicate the positions of the last aa of the ORFs. Arrowheads and associated numbers 261, 2057, and 1690 mark the approximate locations of the PCR primers used in A. For comparison convenience, the nucleotide numbers in hERR␤2 mRNA are still used in hERR␤2-⌬10 and short-form hERR␤ diagrams. The dark rectangular box of short-form hERR␤ indicates that this part of the sequence is located in the intron between exons 9 and 10 of hERR␤2. C, The aa sequence comparison between ERRs and ERs (modified from Ref. 12). The rectangular boxes represent the different domains (A/B, C, D, E, and F), and numbers above rectangular boxes indicate the number of aa residues. The percent homologies with hERR␤2 are indicated inside the boxes. D, Cterminal aa sequence comparison of hERR␤2, hERR␤2-⌬10, and short-form hERR␤. The first 420 aa, which are identical among the three isoforms, are not shown. The aa in dark shading are identical among the three isoforms. The F domains of hERR␤2 and hERR␤2-⌬10 are underlined.

Zhou et al. • Novel Isoforms of Human ERR␤

Zhou et al. • Novel Isoforms of Human ERR␤

zebrafish genome (Fig. 2, A and B). This indicates that no homolog isoform of similar aa content exists in rat, mouse, chicken, or fish. Even if a similar alternative splicing event should happen, it would introduce a stop codon just 24 bp downstream (Fig. 2C, putative stop codons, TAG, TGA, and TAG, shown as bold characters in mouse, rat, and chicken genome sequences comparable to the 10th exon of hERR␤2). Even though dog genome has comparable sequence (64% identity in nucleotide) to the 10th exon of hERR␤2, the putative aa encoded by this region only have 23% identity to that of the 10th exon of hERR␤2 encoded. Furthermore, dog genome does not have comparable sequences to the 11th and 12th exons of hERR␤2 (data not shown). This surprising result indicates that the F domains of hERR␤2 and hERR␤2⌬10 are primate specific. However, to investigate whether such splicing events occur in other species, RT-PCR experiments were performed with mouse testis and kidney total RNA with mouse-specific primers, but there is no evidence

J Clin Endocrinol Metab, February 2006, 91(2):569 –579

573

supporting a similar alternative splicing event in mice (data not shown). Detailed investigation of the genomic sequence comparisons revealed that the genomic sequences following the 3⬘ end of hERR␤2 exon 9 are highly conserved throughout the six species’ genomes (Fig. 2C). The first six nucleotides of the intron between hERR␤2 exons 9 and 10, gtgtga (or gtctga and gtttga in dog, chicken, and Fugu genomes), encode for the last aa, valine, of mERR␤/rERR␤ and the stop codon (shown as underlined characters in Fig. 2C). This led us to hypothesize that a shorter human ERR␤ isoform may exist that does not splice to exon 10 or 11, but extends the length of exon 9, like the last exons in mouse and rat. This shorter hERR␤ would encode a protein of 433 aa, the same length as mERR␤ and rERR␤. We hypothesized that hERR␤2 and hERR␤2-⌬10 are tissue-specific variant splicing isoforms of hERR␤. We designed a forward primer in exon 8, hERR2f1328, and two reverse primers in the intron between exons 9 and 10,

FIG. 2. Comparison of ERR␤ genomic sequences. This figure is modified from Genome Browser screenshots in http://genome.ucsc.edu (15, 16). A, The genomic region spanning the whole hERR␤2 mRNA sequences (Genome Browser, human, May 2004 assembly, position chr14: 75,900,000 –76,050,000); B, the zoom-in region only covering hERR␤2 mRNA exons 9 –12 (position chr14:76,034,000 –76,038,000). Only exons of hERR␤2 are shown (black boxes over the conservation graph labeled with exon numbers 1–12). In both A and B, the conservation diagram indicates the distribution of evolutionarily conserved regions along the genomic sequences. The gray shadow in chimpanzee, dog, mouse, rat, chicken, fugu, and zebrafish indicates the specific genomic regions in these species that are homologous to human genomic sequences. C, The detailed genomic sequences of the region spanning from hERR␤2 exon 9 3⬘ terminal to exon 10 5⬘ terminal. The exon bases (according to hERR␤2 mRNA sequences) are in capital letters, and the intron bases are in lowercase. The middle part of the intron sequences are not shown because of space limitation.

574

J Clin Endocrinol Metab, February 2006, 91(2):569 –579

hERR2r1680 and hERR2r1690 (Fig. 1B), to determine whether this putative shorter form exists. RT-PCR was conducted using human testis, fetal heart, and breast cancer cell lines MCF-7 and C4-12-5 (14) total RNA. DNA bands of the expected sizes appeared in all the RNA that we tested. Sequencing analysis of the cDNA fragment amplified from human fetal heart verified our hypothesis (data not shown). Later, the whole ORF of this isoform, named short-form hERR␤, was amplified from human fetal heart total RNA and verified by sequencing. Tissue distribution expression patterns of the three mRNA isoforms of hERR␤ by RT-PCR

In total, five pairs of primers were designed to distinguish the tissue distribution expression patterns of the three hERR␤ isoforms (Fig. 3 and Materials and Methods). Among them, primer hERRB2f1607 (5⬘-ctt cct gga gat gct gga ggc caa ggt t-3⬘), is designed to specifically amplify hERR␤2-⌬10. This primer covers the boundary of exons 9 and 11. The first 25 nucleotides of forward primer hERRB2f1607 belong to the 3⬘ end of exon 9, and the last three nucleotides belong to the 5⬘ end of exon 11. This design ensures that only the cDNA of hERR␤2-⌬10, but not hERR␤2 or short-form hERR␤, can be amplified (Fig. 3B). For the same strategy, hERRB2r1833

FIG. 3. Detecting the tissue distribution patterns of three hERR␤ isoforms by RT-PCR. A, Primer sets used for detecting different hERR␤ isoforms in human tissues/cell lines. The numbers in primer names coordinate with their positions as represented in Fig. 1B. B, RTPCR results of representative tissues. The asterisk over placenta represents one of the three RNA sample sources we examined.

Zhou et al. • Novel Isoforms of Human ERR␤

(5⬘-ctg ctc ttg gcc aac ctg ccc tct-3⬘), which covers the boundary of exons 10 and 11, was used to specifically amplify the cDNA of hERR␤2. To avoid possible individual differences in the RT-PCR results, we reexamined tissue RNAs, including testis, kidney, placenta, uterus, and skeletal muscle, which showed representative expression patterns of three splicing isoforms of human ERR␤, with more total RNA sample sources from different companies. Two pairs of primers designed to amplify the N terminal of ERR␤ cDNA (described in Materials and Methods) were used in addition to above five pairs of primers to determine the expression patterns of the three splicing isoforms of human ERR␤ (data not shown). RT-PCR results showed that although short-form hERR␤ mRNA is widely expressed in the 24 of 27 tissues/cell lines tested, hERR␤2 and hERR␤2-⌬10 mRNA are only detectable in testis and kidney (Fig. 3B and Table 1). There were no splicing isoforms of human ERR␤ detected in uterus and Ishikawa cells under our RT-PCR conditions. Three different sources of placenta total RNA showed different results: two samples showed no existence of any isoform of human ERR␤ (Fig. 3B), and one sample has short-form hERR␤ expression (Table 1 and data not shown).

Zhou et al. • Novel Isoforms of Human ERR␤

J Clin Endocrinol Metab, February 2006, 91(2):569 –579

575

TABLE 1. Tissue distribution patterns of the three hERR␤ isoforms by RT-PCR

Testis (2) Breast Kidney (2) Placenta (3) Prostate Uterus (3) Ovary Cervix Skeletal muscle (2) Fetal liver Fetal brain Fetal heart Heart Spleen Salivary gland Stomach Lung Thymus Trachea Small intestine Spinal cord Brain Cerebellum Colon MCF-7 C4-12-5 Ishikawa

Short-form hERR␤

hERR␤2

hERR␤2-⌬10

⫹ ⫹ ⫹ ⫺/⫹/⫺ ⫹ ⫺ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫺

⫹ ⫺ ⫹ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺

⫹ ⫺ ⫹ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺

⫹, The expression of the isoform was detected. ⫺, No expression was detected under the conditions of our RT-PCR experiments (see Materials and Methods). ⫺/⫹/⫺, Different results came from three different sample RNA sources: one sample showed the presence of short-form hERR␤ mRNA and the other two samples showed none. The numbers in parentheses represent the number of different sample sources of a certain tissue total RNA we examined.

Different transcriptional activation of vERE-controlled reporter gene by the three mRNA isoforms of hERR␤

When transiently expressed in C4-12-5 (MCF-7-derived, ER-negative) (14) cells, the three splicing isoforms of hERR␤ showed statistically different transcriptional activation on a reporter gene controlled under the basal TK promoter and single copy of vERE (Fig. 4). Short-form hERR␤ gives the strongest activation among three isoforms, hERR␤2-⌬10 gives an intermediate activation, and hERR␤2 gives no activation, which is not different from the empty control vector, pcDNA3.1⫹zeo. Mouse ERR␤, the homolog of short-form hERR␤, showed similar activation on vERE-TK-Luc reporter as short-form hERR␤. All ERR activation on vERE-TK-Luc reporter did not change when treated with 100 nm E2. Transfection of the three mRNA isoforms of hERR␤ in T47D cells did not influence the transcriptional activity mediated by endogenous PR in response to progesterone

Using pRL-CMV Renilla luciferase reporter as an internal transfection efficiency control, transient transfection and luciferase assay in T47D cells revealed that neither short-form hERR␤ nor hERR␤2-⌬10 isoform inhibited endogenous PR transcriptional activity in response to progesterone (Fig. 5A). However, the Renilla luciferase activity of the internal control plasmid pRL-CMV was about 2- to 3-fold higher when trans-

FIG. 4. C4-12-5 (MCF-7-derived, ER-negative) cells were transiently transfected with 50 ng expression vectors as indicated, 0.5 ␮g vERETK-luciferase reporter vector, and 20 ng control pRL-SV40luc vector. Results are expressed as the mean ⫾ SD from three independent experiments, each with triplicate samples. Fisher’s pairwise comparisons were used for statistical analysis, with an individual error rate of 0.01.

fected with hERR␤2 than when transfected with short-form hERR␤, hERR␤2-⌬10, or empty vector pcDNA3.1⫹zeo (data not shown). For this reason, we switched to the pRL-SV40 Renilla luciferase reporter as an internal control when performing hERR␤2 transfection analysis in T47D cells (Fig. 5B). The hERR␤2 isoform did not influence PR activity when pRL-SV40 Renilla luciferase reporter was used as an internal control (Fig. 5B). Protein expression of three splicing isoforms of human ERR␤ in transient transfected cell lines and human kidney tissue lysate

Western blot revealed three major bands representing, respectively, short-form hERR␤, hERR␤, and hERR␤2-⌬10 in C4-12-5 and Ishikawa cells transiently transfected with coordinated expression plasmids (Fig. 6A). In agreement with RT-PCR results (Table 1), C4-12-5 cells expressed endogenous short-form hERR␤ (Fig. 6, mock lanes), whereas Ishikawa cells did not express ERR␤ protein (Fig. 6A, mock lane). A shadow band seen in hERR␤2-⌬10 transfection in Ishikawa cells was sometimes observed. Western blot of human kidney lysate showed two major bands (Fig. 6B): a faster migrating band the size of short-form hERR␤, and slower migrating band, which may represent longer isoforms, hERR␤2 and/or hERR␤2-⌬10, that with only an eight-aa difference in length are too close in size to distinguish. This Western result is in agreement with our RT-PCR results (Table 1) where all three isoforms are expressed. Different subcellular localizations of three splicing isoforms of human ERR␤

To investigate the subcellular localizations of the three human ERR␤ isoforms, immunofluorescence confocal mi-

576

J Clin Endocrinol Metab, February 2006, 91(2):569 –579

FIG. 5. Transcriptional activity of PR in response to progesterone when transfected with short-form hERR␤, hERR␤2, and hERR␤2-⌬10 in T47D cells. No significant differences were observed. A, T47D cells were transfected with empty vector pcDNA3.1⫹zeo, short-form hERR␤, or hERR␤2-⌬10 (1 or 2 ␮g), 2 ␮g 2⫻PRE-luc reporter, and 0.2 ␮g pRLCMV-luc internal control reporter. Transfected cells were treated with vehicle ethanol (C; f) or 10⫺8 M progesterone (P; o). B, T47D cells were transfected with empty vector pcDNA3.1⫹zeo, or hERR␤2 (1 or 2 ␮g), 2 ␮g 2⫻PRE-luc reporter, and 0.2 ␮g pRL-SV40-luc internal control reporter. Transfected cells were treated with vehicle ethanol (C; f) or 10⫺8 M progesterone (P; o). Results are expressed as the mean ⫾ SD. Experiments were performed in triplicate and repeated at least twice.

croscopy was used with anti-ERR␤ antibody. As expected, short-form hERR␤ and hERR␤2-⌬10 were primarily found located in the nucleus (96.5% and 97%, respectively; Fig. 7 and Table 2), similar to ER␣ (18, 19). However, we were surprised to find that hERR␤2, with an F domain different from that of hERR␤2-⌬10 (Fig. 1D), lost much of its ability to exclusively localize to the nucleus. More than 50% of stained COS-1 cells showed hERR␤2 localized mostly in cytoplasm (Table 2 and Fig. 7). Similar results were obtained in Ishikawa cells transiently transfected with these three constructs expressing the three human ERR␤ isoforms (data not shown). In addition, COS-1 cells transfected with 5⬘-Myc-tagged short-form hERR␤, hERR␤2, and hERR␤2-⌬10 constructs revealed the same localization patterns of the three human ERR␤ splicing isoforms when detected with anti-Myc antibody (data not shown).

Zhou et al. • Novel Isoforms of Human ERR␤

FIG. 6. Expression of three splicing isoforms of human ERR␤ in transient-transfected C4-12-5 and Ishikawa cells and in human kidney tissue lysate. A, C4-12-5 cells and Ishikawa cells growing in six-well plates were transfected with 1 ␮g (each) of short-form hERR␤, hERR␤2, and hERR␤2-⌬10 expression plasmids and empty vector (mock). Twenty micrograms of proteins were loaded on each lane. B, Three micrograms of C4-12-5 cell lysate transfected with expression plasmids and empty vector (mock) and 20 ␮g human kidney tissue lysate were loaded on each lane. SF␤, Short-form hERR␤; ␤2, hERR␤2; ⌬10, hERR␤2-⌬10.

The NLS of short-form hERR␤ locates in the D domain

To determine where the NLS of short-form hERR␤ is located, 5⬘-Myc-tagged short-form hERR␤ deletion mutants were constructed: 101– 433 aa (C, D, and E domains), 169 – 433 aa (D and E domains), 210 – 433 aa (E domain), 1–210 aa (A/B, C, and D domains), and 1–169 aa (A/B and C domains). From these data, the NLS is found in the D domain of short-form hERR␤. The constructs containing D domain (101– 433, 169 – 433, and 1–210 aa) all locate within the nucleus, whereas the constructs lacking D domain (210 – 433 and 1–169 aa) are not in the nucleus (Fig. 8A). Discussion

In this communication, we report the identification of two novel mRNA alternative splicing isoforms of human ERR␤, short-form hERR␤ and hERR␤2-⌬10. Tissue distribution studies, transfection assays, and subcellular localization studies of these two novel isoforms of human ERR␤ as well as a previously known isoform, hERR␤2, suggest that these ERR␤ splicing isoforms have diverse biological functions. Alternative RNA splicing is an important molecular mech-

Zhou et al. • Novel Isoforms of Human ERR␤

J Clin Endocrinol Metab, February 2006, 91(2):569 –579

FIG. 7. Subcellular localizations of three splicing isoforms of human ERR␤ in transient-transfected COS-1 cells. COS-1 cells growing as a single layer on coverslips were transfected with 1 ␮g (each) of shortform hERR␤, hERR␤2, and hERR␤2-⌬10 expression plasmids and empty vector (data not shown). The subcellular localizations of human ERR␤ isoforms proteins were determined by double-label indirect immunofluorescence, with anti-ERR␤ antibodies and nuclei staining with the far-red fluorescence dye TO-PRO-3.

anism for the creation of protein diversity derived from a single gene in eukaryotes. The completion of the draft of the human genome revealed that the human genome appears to have only 30,000 – 40,000 protein-coding genes (20, 21), which is only 2- to 3-fold more than the number of genes of Caenorhabditis elegans and Drosophila melanogaster. However, it is currently estimated that at least 35–59% of human genes are alternatively spliced (20, 22–24). This indicates that instead of increasing gene numbers to achieve greater complexity, gene alternative splicing may be used to regulate complex functions in advanced organisms. Many nuclear receptor genes undergo alternative RNA splicing (reviewed in Ref. 25). Multiple RNA splicing variants for both ER␣ and ER␤ genes have been reported (26 –28). N-terminal mRNA splicing variants have also been reported for ERR␣ and ERR␥ (29 –31), but no C-terminal splicing variants have been reported for ERR␣ and ERR␥ to date. For ERR␤, no RNA variant splicing isoform has been reported before this manuscript, except for the website www.ncbi.nih.gov/IEB/ Research/Acembly/av.cgi?db⫽human&c⫽Gene&l⫽ESRRB, based upon 26 sequences from 19 cDNA clones and products, predicted the existence of six different human ERR␤ protein isoforms produced by alternative splicing: a, b, c, d, e, and f, encoding 500, 245, 186, 128, 99, and 102 aa, respectively. The ERR␤ form a is actually the hERR␤2 reported by Chen et al. in 1999 (12). The remaining five isoforms are severely truncated proteins that only contain the N-terminal part of human ERR␤, and none of them is the alternatively spliced products demonstrated in this paper. These five isoforms are unlikely to have characteristics similar to the full-length ERR␤ protein. TABLE 2. Subcellular localization of three splicing isoforms of human ERR␤ overexpressed in COS-1 cells

Short-form hERR␤ hERR␤2 hERR␤2-⌬10

Nucleus only and N ⬎ C (%)

Cytoplasm only and C ⬎ N (%)

C⫽N (%)

96.5 31 97

1.9 50.3 1.5

1.6 18.7 1.5

N, Nucleus; C, cytoplasm. Over 250 stained cells (each) were scored.

577

This report describes three human ERR␤ RNA splicing variants at the C terminus. Additionally, the human ERR␤ gene is able to use human-specific genome sequence to produce RNA alternative splicing variants not found in other species. Therefore, humans/primates may potentially achieve more finely tuned regulation of ERR␤ function than other species. Although human, rat, and mouse all express the common ERR␤ mRNA form coding 433 aa, humans/primates specifically express the hERR␤2 and hERR␤2-⌬10 mRNA isoforms, coding for 500 and 508 aa, respectively. Our results on the tissue distribution of human ERR␤ isoforms showed that although hERR␤2 and hERR␤2-⌬10 expression is limited to testis and kidney, short-form hERR␤ is widely expressed in 24 of 27 tissue/cell lines we tested, including MCF-7 cell (Fig. 3B and Table 1). Three different sources of placenta total RNA gave different results: two samples revealed no expression of any ERR␤ isoform, and one sample expressed short-form hERR␤ mRNA. Only uterus and Ishikawa cells were free of ERR␤ detection under our RT-PCR conditions (Materials and Methods). Previous literature reported that ERR␤ expression is limited to only a few tissues. Giguere et al. (2) reported that by Northern blot, a 4.8-kb mRNA band was found in rat kidney, heart, testis, hypothalamus, hippocampus, cerebellum, and prostate, but not in human placenta and prostate. Using the same method, Pettersson et al. (13) detected a 4.3-kb mRNA band in undifferentiated F9 embryonic carcinoma cells, undifferentiated embryonic stem cells, and a few adult tissues, including kidney and heart. Chen et al. (12) cloned hERR␤2 cDNA from a human testis cDNA library, and their Northern blot showed that an approximately 5.5-kb transcript was expressed at a low level in heart, kidney, liver, skeletal muscle, and stomach. Interestingly, testis from which the cDNA was derived had no detectable signal with Northern analysis. Lu et al. (4) reported that ERR␤ is expressed during all stages of mammary gland development in the mouse. However, no ERR␤ was detected in the human breast cancer cell lines (including the MCF-7 cell line) and normal epithelial cell lines that they tested by Northern blot (4). Our observations differ from previous reports of ERR␤’s tissue distribution. However, we believe that this discrepancy is mainly due to the sensitivity of the detection method: Northern blot vs. RT-PCR. Generally, a standard Northern procedure is less sensitive than RT-PCR. This is consistent with the observation that testis from which hERR␤2 was cloned had no detectable hERR␤2 signal by Northern analysis (12). Using realtime quantitative PCR, Ariazi et al. (32) found that ERR␤ mRNA levels were quite low in human primary breast cancers (⬃1.2 ⫻ 103 copies/ng cDNA in tumor samples; n ⫽ 38) compared to ERR␣ (1.768 ⫻ 106 copies/ng cDNA in tumor samples; n ⫽ 38). This may explain why ERR␣, but not ERR␤, can be detected in human breast cancer cell lines using Northern blot (4). There also remains the possibility that different individuals have different mRNA expression patterns, such as the three different placenta total RNA samples with contradictory RT-PCR results (Fig. 3B and Table 1). Several studies have shown that the F domain of ER␣ can influence the receptor’s transcriptional activity and ligandbinding affinity (33–35). Not all nuclear receptors have the F domain, and the function of the F domain is still not clearly elucidated. Except for hERR␤2 and hERR␤2-⌬10, no other

578

J Clin Endocrinol Metab, February 2006, 91(2):569 –579

Zhou et al. • Novel Isoforms of Human ERR␤

FIG. 8. NLS of short-form hERR␤. A, COS-1 cells growing as a single layer on coverslips were transfected with 1 ␮g (each) of expression vectors for 5⬘-Myctagged short-form hERR␤ deletion mutants. The subcellular localizations of human ERR␤ isoforms proteins were determined by double-label indirect immunofluorescence with anti-Myc antibodies and nuclei staining with the far-red fluorescence dye TO-PRO-3. B, Comparison of human ER␣ NLS and short-form hERR␤ NLS. The first two conserved basic stretches of ER␣ NLS are shown in yellow highlights. ER␣ NLS’s third basic stretch, which is not conserved through species, is shown in red dot shading. The similar basic stretches in short-form hERR␤ are shown by yellow highlights, and red vertical lines indicate identical aa between ER␣ and short-form hERR␤ basic stretches. Identical aa between the two basic stretches of ER␣ and shortform hERR␤ are underlined.

members of the ERR family have an F domain. Because hERR␤2 and hERR␤2-⌬10 have specific tissue distribution patterns, we speculate that hERR␤2 and hERR␤2-⌬10 isoforms have different transcriptional activities or binding affinities with unknown ligands and/or transcriptional factors than the short-form hERR␤ isoform. The result of transient transfection in C4-12-5 (MCF-7-derived, ER-negative) (14) cells supports this hypothesis: the three splicing isoforms of human ERR␤ exhibit differential activation of the firefly luciferase reporter when driven by single copy of ERE, with short-form hERR␤ the strongest and hERR␤2 the weakest (Fig. 4). Furthermore, transfection of human ERR␤ isoforms into T47D cells did not provide evidence that the transcriptional activity of PR through PRE in response to progesterone is blocked by any hERR␤ isoforms, in agreement with a previous report showing lack of inhibition of PR activity by rat ERR␤ in CV-1 and SK-N-MC cell lines (5). However, unexpectedly, the Renilla luciferase activity of the internal control plasmid pRLCMV is about 2- to 3-fold higher when transfected with hERR␤2 than when transfected with short-form hERR␤, hERR␤2-⌬10, or empty vector pcDNA3.1⫹zeo (data not shown). For this reason, we switched to the pRL-SV40 Renilla luciferase reporter as an internal control when performing hERR␤2 transfection analyses in T47D cells (Fig. 5B). Clearly, this differential interaction of hERR␤2 on the CMV promoter strengths the view that the ERR␤ isoforms could have differential biological activities. Subcellular localization studies of the three human ERR␤ splicing isoforms by immunofluorescence confocal microscopy

unveiled additional surprising information about the F domain function, in addition to its contribution to the differential biological activities of the three isoforms. Although short-form hERR␤ and hERR␤2-⌬10 proteins localize mainly in the nucleus (⬎95%), hERR␤2 loses its ability to localize exclusively within the nucleus; more than 50% of stained cells show that hERR␤2 preferentially localizes in the cytoplasm and not the nucleus (Fig. 7 and Table 2). Deletion constructs of short-form hERR␤ revealed that the NLS of short-form hERR␤ is located within the D domain (Fig. 8). hERR␤2’s F domain, which is totally different in aa sequences from hERR␤2-⌬10’s F domain, inhibits the function of the NLS of short-form hERR␤. This is the first example to our knowledge that the F domain can interfere with a nuclear receptor’s subcellular localization. It is possible that nuclear receptors may use this alternative splicing strategy to finely regulate their biological activity by changing their subcellular localization. We also compared the D domain (170 –210 aa) of short-form hERR␤ with the known NLS sequence (256 –303 aa) of human ER␣ (19). The first two basic stretches of ER␣ NLS, aa 256 –260 and 266 –271, have been shown to be conserved through human, mouse, rat, and chicken ER␣ (19). Interestingly, there are two similar basic stretches, 174 –178 and 184 –188 aa, in the D domain of short-form hERR␤ (Fig. 8B). ER␣ and short-form hERR␤ even share the same gap lengths (5 aa) between these two basic stretches and three of the gap aa are identical. Considering the overall low similarity (27%) of the D domains between ER␣ and short-form hERR␤, it is very likely that the

Zhou et al. • Novel Isoforms of Human ERR␤

sequences responsible for short-form hERR␤ nuclear localization are aa 174 –188. ERR␤ was one of the first two orphan nuclear receptors discovered more than 15 yr ago. However, unraveling its function is still in its infancy. With this paper we finally know the correct human isoform to study. Additionally, the lack of a known natural ligand has prevented full understanding of how ERR␤ is regulated in vivo and how ERR␤ regulates other genes. Additional characterization of the new appropriate ERR␤ isoforms will probably provide novel gene targets for future treatment of human diseases. Acknowledgments We thank Paula Whitehead Boettler and Dr. Brian Morin for their help in cloning ERR␥, and Leilani Castleman for her help in cloning mouse ERR␤. Received October 11, 2004. Accepted November 23, 2005. Address all correspondence and requests for reprints to: Dr. Dennis B. Lubahn, Room 110A ASRC, 920 East Campus Drive, University of Missouri, Columbia, Missouri 65211. E-mail: [email protected]. This work was supported by National Institute of Environmental and Health Sciences Grant P01-ES-10535 and Army Concept Proposal Award DAMD17-03-1-0561 (to D.B.L.) and in part by National Institutes of Health Grant CA-86916 (to S.M.H.).

J Clin Endocrinol Metab, February 2006, 91(2):569 –579

15. 16. 17. 18. 19. 20.

21.

22. 23. 24.

References 1. Giguere V 2002 To ERR in the estrogen pathway. Trends Endocrinol Metab 13:220 –225 2. Giguere V, Yang N, Segui P, Evans RM 1988 Identification of a new class of steroid hormone receptors. Nature 331:91–94 3. Luo J, Sladek R, Bader JA, Matthyssen A, Rossant J, Giguere V 1997 Placental abnormalities in mouse embryos lacking the orphan nuclear receptor ERR-␤. Nature 388:778 –782 4. Lu D, Kiriyama Y, Lee KY, Giguere V 2001 Transcriptional regulation of the estrogen-inducible pS2 breast cancer marker gene by the ERR family of orphan nuclear receptors. Cancer Res 61:6755– 6761 5. Trapp T, Holsboer F 1996 Nuclear orphan receptor as a repressor of glucocorticoid receptor transcriptional activity. J Biol Chem 271:9879 –9882 6. Xie W, Hong H, Yang NN, Lin RJ, Simon CM, Stallcup MR, Evans RM 1999 Constitutive activation of transcription and binding of coactivator by estrogenrelated receptors 1 and 2. Mol Endocrinol 13:2151–2162 7. Vanacker JM, Pettersson K, Gustafsson JA, Laudet V 1999 Transcriptional targets shared by estrogen receptor-related receptors (ERRs) and estrogen receptor (ER) ␣, but not by ER␤. EMBO J 18:4270 – 4279 8. Tremblay GB, Kunath T, Bergeron D, Lapointe L, Champigny C, Bader JA, Rossant J, Giguere V 2001 Diethylstilbestrol regulates trophoblast stem cell differentiation as a ligand of orphan nuclear receptor ERR␤. Genes Dev 15:833– 838 9. Tremblay GB, Bergeron D, Giguere V 2001 4-Hydroxytamoxifen is an isoformspecific inhibitor of orphan estrogen-receptor-related (ERR) nuclear receptors ␤ and ␥. Endocrinology 142:4572– 4575 10. Coward P, Lee D, Hull MV, Lehmann JM 2001 4-Hydroxytamoxifen binds to and deactivates the estrogen-related receptor ␥. Proc Natl Acad Sci USA 98:8880 – 8884 11. Suetsugi M, Su L, Karisberg K, Yuan Y, Chen S 2003 Flavone and isoflavone phytoestrogens are agonists of estrogen-related receptor. Mol Cancer Res 1:981– 991 12. Chen F, Zhang Q, McDonald T, Davidoff MJ, Bailey W, Bai C, Liu Q, Caskey CT 1999 Identification of two hERR2-related novel nuclear receptors utilizing bioinformatics and inverse PCR. Gene 228:101–109 13. Pettersson K, Svensson K, Mattsson R, Carlsson B, Ohlsson R, Berkenstam A 1996 Expression of a novel member of estrogen response element-binding nuclear receptors is restricted to the early stages of chorion formation during mouse embryogenesis. Mech Dev 54:211–223 14. Oesterreich S, Zhang P, Guler RL, Sun X, Curran EM, Welshons WV, Osborne

25. 26.

27. 28. 29. 30.

31.

32. 33.

34. 35.

579

CK, Lee AV 2001 Re-expression of estrogen receptor ␣ in estrogen receptor ␣-negative MCF-7 cells restores both estrogen and insulin-like growth factormediated signaling and growth. Cancer Res 61:5771–5777 Kent WJ, Sugnet CW, Furey TS, Roskin KM, Pringle TH, Zahler AM, Haussler D 2002 The human genome browser at UCSC. Genome Res 12:996 –1006 Karolchik D, Baertsch R, Diekhans M, Furey TS, Hinrichs A, Lu YT, Roskin KM, Schwartz M, Sugnet CW, Thomas DJ, Weber RJ, Haussler D, Kent WJ 2003 The UCSC genome browser database. Nucleic Acids Res 31:51–54 Ansell PJ, Espinosa-Nicholas C, Curran EM, Judy BM, Philips BJ, Hannink M, Lubahn DB 2004 In vitro and in vivo regulation of antioxidant response elementdependent gene expression by estrogens. Endocrinology 145:311–317 Welshons WV, Lieberman ME, Gorski J 1984 Nuclear localization of unoccupied oestrogen receptors. Nature 307:747–749 Picard D, Kumar V, Chambon P, Yamamoto KR 1990 Signal transduction by steroid hormones: nuclear localization is differentially regulated in estrogen and glucocorticoid receptors. Cell Regul 1:291–299 Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, Baldwin J, Devon K, Dewar K, Doyle M, FitzHugh W, Funke R, Gage D, Harris K, Heaford A, Howland J, Kann L, Lehoczky J, LeVine R, McEwan P, McKernan K, Meldrim J, Mesirov JP, Miranda C, Morris W, Naylor J, et al 2001 Initial sequencing and analysis of the human genome. Nature 409:860 –921 Venter JC, Adams MD, Myers EW, Li PW, Mural RJ, Sutton GG, Smith HO, Yandell M, Evans CA, Holt RA, Gocayne JD, Amanatides P, Ballew RM, Huson DH, Wortman JR, Zhang Q, Kodira CD, Zheng XH, Chen L, Skupski M, Subramanian G, Thomas PD, Zhang J, Gabor Miklos GL, Nelson C, et al 2001 The sequence of the human genome. Science 291:1304 –1351 Mironov AA, Fickett JW, Gelfand MS 1999 Frequent alternative splicing of human genes. Genome Res 9:1288 –1293 Croft L, Schandorff S, Clark F, Burrage K, Arctander P, Mattick JS 2000 ISIS, the intron information system, reveals the high frequency of alternative splicing in the human genome. Nat Genet 24:340 –341 Brett D, Hanke J, Lehmann G, Haase S, Delbruck S, Krueger S, Reich J, Bork P 2000 EST comparison indicates 38% of human mRNAs contain possible alternative splice forms. FEBS Lett 474:83– 86 Lou H, Gagel RF 2001 Alternative ribonucleic acid processing in endocrine systems. Endocr Rev 22:205–225 Flouriot G, Griffin C, Kenealy M, Sonntag-Buck V, Gannon F 1998 Differentially expressed messenger RNA isoforms of the human estrogen receptor-␣ gene are generated by alternative splicing and promoter usage. Mol Endocrinol 12:1939 – 1954 Petersen DN, Tkalcevic GT, Koza-Taylor PH, Turi TG, Brown TA 1998 Identification of estrogen receptor ␤2, a functional variant of estrogen receptor ␤ expressed in normal rat tissues. Endocrinology 139:1082–1092 Lu B, Leygue E, Dotzlaw H, Murphy LJ, Murphy LC, Watson PH 1998 Estrogen receptor-␤ mRNA variants in human and murine tissues. Mol Cell Endocrinol 138:199 –203 Yang N, Shigeta H, Shi H, Teng CT 1996 Estrogen-related receptor, hERR1, modulates estrogen receptor-mediated response of human lactoferrin gene promoter. J Biol Chem 271:5795–5804 Johnston SD, Liu X, Zuo F, Eisenbraun TL, Wiley SR, Kraus RJ, Mertz JE 1997 Estrogen-related receptor ␣1 functionally binds as a monomer to extended halfsite sequences including ones contained within estrogen-response elements. Mol Endocrinol 11:342–352 Heard DJ, Norby PL, Holloway J, Vissing H 2000 Human ERR␥, a third member of the estrogen receptor-related receptor (ERR) subfamily of orphan nuclear receptors: tissue-specific isoforms are expressed during development and in the adult. Mol Endocrinol 14:382–392 Ariazi EA, Clark GM, Mertz JE 2002 Estrogen-related receptor ␣ and estrogenrelated receptor ␥ associate with unfavorable and favorable biomarkers, respectively, in human breast cancer. Cancer Res 62:6510 – 6518 Montano MM, Muller V, Trobaugh A, Katzenellenbogen BS 1995 The carboxyterminal F domain of the human estrogen receptor: role in the transcriptional activity of the receptor and the effectiveness of antiestrogens as estrogen antagonists. Mol Endocrinol 9:814 – 825 Nichols M, Rientjes JM, Stewart AF 1998 Different positioning of the ligandbinding domain helix 12 and the F domain of the estrogen receptor accounts for functional differences between agonists and antagonists. EMBO J 17:765–773 Schwartz JA, Zhong L, Deighton-Collins S, Zhao C, Skafar DF 2002 Mutations targeted to a predicted helix in the extreme carboxyl-terminal region of the human estrogen receptor-␣ alter its response to estradiol and 4-hydroxytamoxifen. J Biol Chem 277:13202–13209

JCEM is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.