| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Embryo; |
Departments of Animal Sciences3 and Veterinary Pathobiology,4 University of Missouri, Columbia, Missouri 65211
ABSTRACT
Distal-less 3 (DLX3), a homeodomain transcription factor required for placental development in the mouse, modestly transactivates hCG-alpha subunit gene (hCGA) expression in human choriocarcinoma cells. Because hCG and interferon-tau (IFNT) are expressed in trophectoderm of primates and ruminants, respectively, we have tested the hypothesis that DLX3 regulates the genes for IFNT (IFNT). A bovine IFNT1 promoter (–457 to +66), linked to a luciferase (luc) reporter, was transactivated approximately 20-fold by overexpressing DLX3 in human JAr cells. Elimination of a potential DLX3-binding site (–54 GATAATGAG –46) by either truncation or mutagenesis abolished this effect. A sequence (–59 to –44) encompassing this site bound DLX3 specifically. Coexpression of DLX3 and ETS2, which is known to be a key regulator of IFNT expression, increased reporter activity by more than 250-fold, whereas deletion of the established ETS2 site (–79 to –70) eliminated the ability of DLX3 to transactivate the gene. Conversely, mutation of the DLX3 site significantly reduced the transactivational effects of ETS2. Both DLX3 and ETS2 are coexpressed in JAr cells and in an IFNT-producing, bovine trophoblast cell line, CT-1. The two can be immunoprecipitated together as a complex from CT-1 cells, and RNAi-mediated, partial knockdown of DLX3 expression reduced the production of IFNT by approximately 50+. Together, these results suggest that DLX3 has a central role in controlling IFNT gene expression by associating with ETS2 on the IFNT promoter.
embryo, gene regulation, interferon-
, transcription factor, trophoblast
Interferon- (IFNT), which is produced by the early trophoblast of ruminant species such as cattle and sheep before placentation is initiated, acts in a paracrine manner on the maternal endometrium to create an environment favorable for establishment of pregnancy [1–5]. Its best-established function is to intervene in the normal cyclic regression of the maternal corpus luteum, an event that would usually occur at the end of the estrous cycle if the animal were not pregnant. Accordingly, IFNT acts in a manner analogous to that of hCG in the human, but rather than being targeted directly to the corpus luteum and having a luteotropic effect, it acts locally on the uterine endometrium and prevents the pulsatile release of the luteolytic hormone, prostaglandin F2
. As with hCG, the rapid up-regulation of IFNT expression at a time when the corpus luteum is wavering on the point of regression is essential if the pregnancy is to be maintained, whereas a less-than-robust production may contribute to embryonic loss [6]. Hence, an understanding of the transcriptional control mechanisms operating on these genes is likely to provide an insight regarding why some pregnancies are successful and why others fail.
The genes for IFNT (IFNT) have been cloned from several species within the suborder Ruminantia [7–10], in which they are exclusively expressed in trophoblast. Unlike the genes encoding the related IFNA (interferon-alpha, IFNA) and IFNB (interferon beta, IFNB), the IFNT genes are not responsive to virus, although the 5' sequences upstream of the transcriptional start sites, which contain the main regulatory regions, are highly conserved across species through at least 400 bp [9]. A yeast single-hybrid screen, used to pinpoint proteins that bind within the proximal promoter region of IFNT, identified the transcription factor ETS2 as a likely regulator of IFNT expression in Day 13 ovine conceptuses [11]. ETS2 is the member of a large family of transcription factors that are characterized by their conserved ETS DNA-binding domain, which binds to a specific DNA motif, characterized by a core GGAA sequence. Overexpression of ETS2, but not of related transcription factors of the ETS family, transactivated a luciferase (luc) reporter under the control of the IFNT upstream regulatory region containing an ETS2-binding site (–79 to –70) in the human choriocarcinoma cell, JAr. Mutation of this DNA motif within the core GGAA sequence markedly reduced both basal and ETS2-induced activity [12]. Importantly, CSF1, a growth factor present in uterine secretions, was able to activate ETS2 via the Ras/MAP kinase signal transduction pathway and to enhance the up-regulation of IFNT promoters by ETS2 [13]. In addition, protein kinase A-mediated transactivation of the IFNT-regulatory region is strictly dependent on the ETS2-binding site [14]. Finally, several ovine IFNT "pseudogenes" that are poorly expressed lack a consensus ETS2-binding sequence in the proximal promoter element [11]. Replacement of the "mutated" and, presumably, defective ETS2-binding site in the pseudogene with 22 bp of active ETS2-binding site restores full promoter activity [15]. For these reasons, ETS2 is now generally regarded as being the pivotal transcription factor responsible for control of IFNT expression in trophectoderm.
Following the initial discovery that ETS2 regulates the IFNT, ETS2 has been shown to play a role in the regulation of several other genes that are characteristically expressed in trophectoderm, including the hCG and hCGβ subunit genes (hCGA and hCGB) [16, 17]. The fact that deletion of the murine Ets2 gene leads to placental failure and embryonic death as the result of defects in the development of the ectoplacental cone and reduced expression of matrix metalloproteinases [18] suggests that ETS2 is a key factor in specifying many aspects of trophoblast function [19].
Rarely, if ever, does a single transcription factor operate alone. Most act in a combinatorial manner to provide a response that is tissue specific and regulated over time. Because ETS2 is expressed widely in adult bovine tissues (see GenBank Accession numbers BE758238.1, AW668843.1, and BE479476.1), and because IFNT expression is restricted to a single epithelial cell layer for just a few days in early pregnancy, other transcription factors must be involved in fine-tuning IFNT production. Here, we have studied a second transcription factor, the homeobox protein distal-less 3 (DLX3), a member of the distal-less family of non-Antennapedia homeobox genes, which is an established contributor to trophoblast function in mammals [20–24]. DLX3 is expressed in human and mouse trophoblast cells [25], and Dlx3–/– mice die between Embryonic Days 9.5 and 10 because of placental defects and abnormal placental vasculogenesis [22]. The protein itself consists of 287 amino acids and contains a centrally placed, DNA-binding homeodomain (Fig. 1B). It recognizes a specific sequence with a central TAAT core motif [26]. Two "signature" genes of trophoblast have been found to be regulated by DLX3—namely hCGA [24] and the gene for murine 3β-hydroxysteroid dehydrogenase VI (Hsd3b6) [23]—although others have noted that Hsd3b6 expression is not correlated with that of DLX3 in the rodent placenta, particularly in giant cells, where most progesterone synthesis occurs [20]. These observations led us to examine the potential role of DLX3 in controlling expression of the IFNT and to determine whether it acts in a combinatorial manner with ETS2.
|
Reporter Gene Constructs and Expression Plasmids
Bovine IFNT1 (boIFNT1)-reporter constructs, –49luc, –126luc, and –457luc (containing the gene control region –49 to +50, –126 to +50, and –457 to +66, respectively), have been described previously [11, 12]. The –67luc construct was generated from mutated –126luc (–126µAP1) by XbaI digestion and self-ligation [13]. Site-directed mutagenesis of the homeobox (DLX3)-binding site at –53 to –50 on the –126luc reporters was achieved with the primers 5'-aag GCC Ttg agt acc gtc ttc (uppercase letters show the mutation sites) and 5'-aag GCC Tga aaa ttt ctc tct ca and standard PCR procedures [13]. Mutagenesis of the ETS2-binding site at –79 (–126µETS2) has been described previously [12]. The mutated sequences are shown in Figures 2 and 3, respectively. The double mutation of the ETS2- and DLX3-binding sites on –126luc (–126µETS2/µDLX3) (Fig. 4) was generated by site-directed mutagenesis on the –126µETS2 reporter plasmid by using the primers 5'-cta gtc gtg aga gag aaa ttt tcg gGC CTt gag tac cgt ctt ccc and 5'-ggg aag acg gta ctc aAG GCc cga aaa ttt ctc tct cac gac tag and Quickchange Site-Directed Mutagenesis kit (Stratagene). Fidelity of all constructs was verified by DNA sequencing.
|
|
|
The DLX3 expression plasmid, Dlx3/pCI-neo, was a gift from Dr. Maria Morasso (Developmental Skin Biology Unit, National Institute of Arthritis and Musculoskeletal and Skin Diseases, Bethesda, MD) [24]. The ETS2 expression plasmids (pCGNEts2) have been described previously [11].
Cell Cultures and Transfections
JAr cells (HTB-144; American Type Culture Collection) were maintained in RPMI 1640 supplemented with 10% fetal bovine serum (FBS), penicillin (50 U/ml), and streptomycin (50 µg/ml). CT-1 cells [27] obtained from Dr. Alan Ealy (University of Florida, Gainesville, FL) were cultured in STO feeder cells conditioned medium (90% Dulbecco modified Eagle medium/F-12, 10% FBS, 1% antibiotic-antimycotic, and 10 ng/ml of fibroblast growth factor-2) on a Matrigel (Becton Dickinson)-coated surface [28]. Both DLX3 and ETS2 expression plasmids (2.6 and 0.3 µg, respectively) and 0.5 µg of the boIFNT1 promoter-luc reporter constructs were transfected into JAr cells either in 60-mm dishes (2 x 105 cells/dish) or in six-well plates (1 x 105 cells/well with half amount of plasmid DNA described above) by the calcium phosphate method [11]. The plasmid phRG-TK (10 ng; Promega) or 50 ng of pRSVLTR-βgal were included for normalization of transfection efficiency. The expression plasmid for the constitutively active catalytic subunit (RSV-PKA) has been described previously [14]. Both PKA and ETS2 expression plasmids (1.3 and 0.2 µg, respectively) and 0.5 µg of the IFNT-luc reporter were transfected in six-well plates to study PKA effects. Total amounts of transfected DNA were kept constant by adding corresponding empty vectors. Luc reporter assays were conducted 36 h after transfection. Enzyme assays for analyses of transfection experiments were carried out as described previously [12, 13]. The activities of both firefly and Renilla luciferases were measured with a dual-luciferase reporter assay system (Promega). Firefly luc activity was normalized to either Renilla luciferase or β-galactosidase activity, depending on which control was used.
Electrophoretic Mobility Shift Assay
Recombinant DLX3 protein was prepared by means of the DLX3 expression plasmid, Dlx3/pCI-neo, which carries the T7 promoter at the transcription start site [24]. The expression plasmid was linearized by NotI and was used as DNA template for in vitro transcription and translation (Promega L5040 TNT-coupled wheat germ extract system). Nuclear extracts were prepared from transfected and nontransfected JAr cells using the Nuclear Extract Kit (Active Motif). The JAr cells had been grown on 10-cm culture dishes. Cells in three dishes had been transfected with 10 µg/dish of Dlx3/pCI-neo plasmid DNA by means of Lipofectamine/Plus reagent (Invitrogen). Cells from three other dishes had not been transfected. A double-stranded 16-mer oligonucleotide, corresponding to nucleotide sequence –59 to –44 of the boIFNT1 promoter, was end-labeled with [
-32P]ATP. The DNA probe (20 fmol, 13 879 cpm) was mixed with 2.5 µl of the in vitro-translated DLX3 and nuclear extracts from the JAr cells (7.4 µg from transfected cells and 13.4 µg from nontransfected cells). For competition binding assays, double-stranded competitor DNA (200-fold molar excess, 4 pmol) was added before incubation with the labeled probe. The DNA-binding conditions have been described elsewhere [12]. To determine that the DNA-protein complexes contained DLX3, approximately 0.1 µg of rabbit anti-DLX3 antibody [24, 29] or the same amount of an unrelated immunoglobulin G (IgG; anti-biotin antibody from Cell Signaling Technology) were added to the reaction mixture before the addition of labeled probe.
Cells were extracted in Passive Lysis Buffer (Promega) or RIPA Buffer (10 mM Tris-HCl [pH 7.2], 1 mM EDTA, 1% Triton X-100, 0.1% SDS, 0.1% sodium deoxycholate, and 100 mM NaCl). After centrifugation to remove particulates in the extracts, soluble protein (40–50 µg/lane) was analyzed by 10–12.5% SDS-PAGE. Protein in the gels was transferred to polyvinylidene difluoride membranes (Immobilon-P; Millipore). DLX3 was identified by using anti-DLX3 antibody (diluted 1:2000) from Dr. Morasso [24, 29]. The ETS2 proteins were identified on blots by use of a rabbit anti-ETS2 antiserum [17] raised against the amino-terminal fragment (amino acids between 1 and 326) of human ETS2 fused to glutathione S-transferase. The second antibody was alkaline phosphatase-conjugated anti-rabbit IgG used as described previously [12] or horseradish peroxidase-conjugated anti-rabbit or anti-sheep IgG and detected with the Phototope-HRP Western Blot Detection System (Cell Signaling Technology). Sheep anti-DLX3 polyclonal antiserum was obtained from ewes that were immunized with a purified glutathione S-transferase fusion protein containing amino acids 197–287 of the murine DLX3 protein. The IgGs of anti-ETS2 and anti-DLX3 antibodies were affinity purified using the Aminolink Plus Immobilization Kit (Thermo Scientific).
The JAr and CT-1 cells were grown on coverslips and Matrigel-coated coverslips, respectively, placed in six-well tissue culture plates. After fixation (4% paraformaldehyde/PBS, 15 min) and permeabilization (1% Triton X-100/PBS, 30 min), cells on coverslips were incubated for 1 h with either 5% goat or donkey serum (depending on the secondary antibody used), followed by overnight incubation with primary antibodies (rabbit anti-ETS2 at 1:200, sheep anti-DLX3 at 1:40, mouse anti-IFNT monoclonal antibody [mAb] at 1:5) at 4°C. Secondary antibody staining was performed with fluorescein (Alexa Fluor [AF] 568, 488, and 647)-labeled goat anti-rabbit (AF 568), goat anti-mouse (AF 488), and/or donkey anti-sheep (AF 488 or AF 647) antibodies (Molecular Probes) at a 1:500 dilution. Nuclei were labeled with 4'-6-diamidino-2-phenylindole (Molecular Probes) or TO-PRO-3 iodide (Molecular Probes). Images were captured with a Bio-Rad Radiance 2000 coupled to an Olympus IX70 inverted microscope or with a Zeiss LSM 510 two-photon confocal system. A hybridoma line producing an mAb against bovine IFNT was produced by the University of Missouri-Columbia Cell and Immunology Core (http://www.biotech.missouri.edu/cic/) by using recombinant bovine IFNT1a. The mAb was purified from the hybridoma culture supernatant on a protein A-agarose column (Bio-Rad) with a BioLogic Fast Protein Liquid Chromatography (Bio-Rad). The protein A column was washed with 3.2 M NaCl and 1.6 M glycine (pH 9.0), and IgG mAb was eluted with 100 mM sodium citrate (pH 3.0). The eluted fractions were neutralized with 100 mM Tris-HCl (pH 9.0) and then dialyzed in 1x PBS, 0.02% sodium azide, 1 mM PMSF, and 1 mM EDTA.
Ovine Day 15 conceptuses were fixed in 4% paraformaldehyde for 4 h. Tissues were paraffin-embedded, and sections (thickness, 5 µm) were transferred to charged glass slides. After deparaffinization and hydration, antigen retrieval was conducted by exposing the sections to boiling "antigen unmasking solution" (citrate acid-based buffer; Vector Laboratories) for 5 min. Slides were allowed to cool for approximately 20 min at room temperature and then treated with a solution of 1.5% hydrogen peroxide in absolute methanol for 10 min at room temperature to quench endogenous peroxidases. The slides and sections were then incubated with serum-blocking solution (Vector Laboratories) for 30 min. Rabbit anti-DLX3 antibody diluted at 1:200 in the serum-blocking solution was used as primary antibody. Normal rabbit serum was used as negative control. Incubation with primary antibody was carried out overnight at 4°C. Bound IgG was detected with the Vectastain Elite ABC kit and ImmPACT DAB (diaminobenzidine; Vector Laboratories) according to the manufacturer's instructions. One adjacent section was stained with hematoxylin and eosin.
Coimmunoprecipitation Analysis
The CT-1 cells, grown as monolayers in T75 flasks, were removed by scraping and were collected by centrifugation. The cell pellet was washed with ice-cold 1x PBS, and cells were lysed in buffer containing detergents and protease inhibitors [15]. Particulate material was removed by centrifugation at 16 060 x g. Freshly isolated protein (1 mg) was incubated overnight with either 5 µg of affinity-purified anti-DLX3 antibody or purified control sheep IgG (Santa Cruz Biotechnology). After centrifugation, supernatants were mixed with 50 µl of protein G-agarose beads (Santa Cruz Biotechnology) for 6 h and analyzed by Western blotting with affinity-purified rabbit ETS2 (1:1000) antibody (sc-351; Santa Cruz Biotechnology) as the detection reagent.
The JAr cells (in 60-mm dishes) were transfected with 3 µg of pCGNEts2 and 3 µg of Dlx3/pCI-neo plasmid DNAs by using Lipofectamine/Plus reagents. For each set of reactions, the extracts were prepared from six dishes in a lysis buffer containing detergents and protease inhibitors [16]. Cell lysates were cleared by centrifugation at 16 060 x g, and 1 mg of freshly isolated protein was incubated overnight with either 5 µg of affinity-purified anti-ETS2 antibody (sc-351) or purified nonspecific IgG (Santa Cruz Biotechnology). After centrifugation, supernatant solutions were mixed with 50 µl of swollen, prewashed, protein G-agarose beads for 6 h. After washing, bound immune complexes were eluted in nonreducing sample buffer [17] at 80°C for 15 min. Samples were analyzed in 12% SDS gel. The immune complexes formed in the presence of ETS2 antibody were detected on Western blots with affinity-purified sheep DLX3 antibody. Control immunoprecipitations were carried out with preimmune sheep IgG and analyzed in parallel. These controls also indicated the absence of IgG contamination in samples eluted from protein G-agarose beads. A 25-µg sample of protein extracted directly from the cells was analyzed in parallel to determine the relative amounts of ETS2 and DLX3 present. The DLX3 antibody was used at 1:1000 in 5% (wt/vol) nonfat dried milk dissolved in Tris-buffered saline added with 0.1% (vol/vol) Tween 20. The DLX3 bands were visualized by chemiluminescence with reagents purchased from either Cell Signaling Technology or Thermo Scientific.
Chromatin Immunoprecipitation Analysis
Chromatin immunoprecipitation (ChIP) analysis on chromatin purified from CT-1 cells was conducted as described previously [17]. The chromatin, prepared from approximately 107 cells, was sheared by sonication and treated with a protein G-agarose bead slurry (Santa Cruz Biotechnology) to remove components that bound nonspecifically. One fifth of the precleared chromatin was saved as "total input" control. The remaining chromatin solution was divided into five portions: untreated ("no antibody" control), treated with 2 µg of sheep affinity-purified DLX3 antibody, treated with 2 µg of rabbit anti-ETS2 antibody [17], treated with 2 µg of purified nonspecific IgG, and treated with 2 µg of RNA polymerase II antibody (Active Motif). The chromatin complexes were collected on protein G-agarose, processed as described elsewhere [17], and were used as template for each PCR reaction. The primers used in the ChIP assay were 5'-tga caa acc caa att tta ttg gga aa (forward) and 5'-tct gat gat gat cgt tct aag caa gg (reverse) and were designed to amplify a region of the IFNT proximal promoter (–188 to +3) containing the ETS2/DLX3 enhancer region. The PCR conditions were as follows: 95°C for 3 min for one cycle, and then 32 cycles of 94°C for 20 sec, 55°C for 30 sec, and 72°C for 30 sec. The PCR products were visualized by ethidium bromide staining following electrophoresis in 2% agarose.
We first identified a Bos taurus cDNA (GenBank accession number NM_001081622, encoding 287 amino acids) with a sequence closely similar (90% identity in nucleotides) to that of a human DLX3 transcript (NM_005220, also 287 amino acids). Coidentity was confirmed by comparing a 158-kb length of genomic DNA from a B. taurus genomic clone (clone name, RP42-221D7; GenBank accession number, AC136966) with the human DLX3 gene in terms of conserved sequence and intron/exon boundaries. Three small interfering RNA (siRNA) duplexes against the bovine sequence (NM_001081622) were designed by using the siRNA selection program at Whitehead Institute for Biomedical Research (http://jura.wi.mit.edu/bioc/siRNAext/home.php). Only one of these duplexes, r(UGG CUC AUC ACC UUC CAG A) dTdT and r(UCU GGA AGG UGA UGA GCC A) dAdA, was effective at knocking down DLX3 expression. The CT-1 cells at 30–40% confluency in six-well plates were transfected by using "lipitoid" reagent, a cationic oligopeptoid-phospholipid conjugate [30] with 50 nM siRNA duplexes specific for bovine DLX3 (Ambion) or a nonspecific control (Silencer Negative Control #1 siRNA; Ambion). The transfection method was essentially the same as described elsewhere [30], but with the following minor modifications. After the culture medium was removed, the cells were washed twice with 1 ml of OptiMEM (Invitrogen) before addition of the transfection cocktail (0.1 ml/well; lipitoid:siRNA charge ratio, 3:1) for 2 h. The transfection regent was removed, and growth medium was refreshed 24 h later. After three further days of culture, the cells were lysed and soluble proteins analyzed by Western blot analysis, with rabbit anti-DLX3 (1:2000) as the detecting antibody. The blots were then stripped (IgG Elution Buffer; Thermo Scientific) and reprobed sequentially with anti-IFNT antiserum (1:400) and anti-β-actin antibody (1:2500; Cell Signaling Technology), respectively. The anti-IFNT antiserum used was generated by immunizing rabbits with recombinant bovine IFNT through standard procedures [31]. Bound IgG was detected by chemiluminescence as described above.
Antiviral activity present in the culture media from cells transfected with siRNAs was measured in a cytopathic reduction assay as described previously [32]. Serial 3-fold dilutions of CT-1 medium samples, normalized by their cellular protein concentration, were incubated on Mardin-Darby bovine kidney cells challenged with vesicular stomatitis virus. Commercially available bovine IFNA2 (PBL InterferonSource) of known antiviral activity (1 x 106 IU) was used as the standard. The IFNT concentration was determined by 50% protection from virus-induced cell death.
Data are presented as the mean ± SEM. Data were analyzed by one-way ANOVA followed by a Tukey multiple-comparison test to compare selected pairs of experimental groups with the PRISM statistical analysis software (version 4; GraphPad Software, Inc.). Differences of P < 0.05 were considered to be significant.
Transactivation of IFNT Promoters by the Homebox Protein DLX3
The DLX3 binds to an AT-rich region within the proximal control region of the hCG subunit gene and, when ectopically expressed in choriocarcinoma cells, has a modest ability (
2-fold) to transactivate promoter constructs containing this sequence [24]. Because IFNT genes also contain a possible DLX3-binding site (Fig. 2B), with seven out of eight bases conserved relative to the consensus sequence [26], we examined whether the IFNT promoters can be up-regulated in response to overexpression of this homeobox transcription factor. We cotransfected the DLX3 expression plasmid and a –457 IFNT promoter-reporter (–457luc) into JAr cells. Luc activity increased approximately 25-fold (Fig. 2A) relative to a control without DLX3 overexpression, although considerable variation was found between experiments. DLX3 had an almost identical effect on a truncated reporter construct (–126luc), which contains the ETS2-binding site (–79 to –70) known to be essential for full IFNT expression [11–13]. Deletion of this site by further truncation to –67, however, reduced the ability of DLX3 overexpression to transactivate the luc reporter by almost 80% (5-fold vs. 24-fold). Further truncation designed to remove the DLX3-binding site itself (–49luc) reduced the effects of ectopically expressed DLX3 by approximately 98% and virtually ablated basal activity. Mutation of the core TAAT sequence of the putative DLX3-binding site (–53 to –50) in the –126luc reporter construct (–126µDLX3) provided reduced basal activity (55%) compared to that observed with the wild-type –126luc reporter and led to complete unresponsiveness to DLX3 overexpression (Fig. 2A). Together, these data show that DLX3 can transactivate the IFNT promoter, that the critical sequence for DLX3 responsiveness lies between –54 and –46, and that DLX3 effects might be at least partially dependent on the presence of the established ETS2-binding site placed approximately 20 bp distal to the putative DLX3-binding site.
Binding of DLX3 to the IFNT Gene Regulatory Region
We next determined whether DLX3 physically interacts with IFNT regulatory region by employing in vitro-translated DLX3 and a labeled duplex DNA probe representing the 16 bp of the IFNT regulatory region lying between base pairs –59 to –44 (Fig. 5A). Although a control protein (in vitro-translated luciferase) failed to bind the probe (Fig. 5B, lane 1), the recombinant DLX3 protein formed a complex (lane 2). A 200-fold molar excess of unlabeled probe competed efficiently with labeled probe in the binding reaction (lane 3), whereas an equivalent molar excess amount of a control oligonucleotide with its DLX3 binding core motif mutated (Fig. 5A, µ) was ineffective as a competitor (lane 4). Addition of anti-DLX3 antibody prevented formation of a specific complex between the wild-type probe and the protein (lane 5), whereas a control rabbit IgG had no effect (lane 6). The experiments were repeated with nuclear extracts from JAr cells in which DLX3 had been overexpressed (Fig. 5C) and from nontransfected cells (Fig. 5D). The results shown in Figure 5C essentially match those shown in Figure 5B. A specific complex with the same mobility as that formed with recombinant protein was visible on the gel, and its mobility was disrupted ("supershifted" in the case of Fig. 5C) by the addition of anti-DLX3. A longer (17x) exposure was needed to detect the complexes formed from nuclear extracts of nontransfected cells (Fig. 5D), presumably because the DLX3 concentration was much lower in these cells. Although additional bands were revealed by this extended exposure, a complex whose formation could be inhibited by the presence of excess unlabeled probe or supershifted by anti-DLX3 was still detectable (Fig. 5D, arrow). It is unclear why anti-DLX3 prevented complex formation with recombinant protein (Fig. 5B) but caused a supershift with endogenously expressed DLX3 (Fig. 5, C and D). Nevertheless, these results indicate that both constitutively expressed and ectopically expressed DLX3 bind specifically to the –54 to –46 of the IFNT regulatory region.
|
Dependency of Efficient DLX3 Transactivation on the ETS2-Binding Site at –79/–70 in the IFNT Regulatory Region
Although the –67luc promoter construct contains the intact –54/–46 DLX3-binding site, DLX3 transactivated this promoter much less efficiently (P < 0.001) than the –126luc and –457luc constructs (Fig. 2A). We therefore examined the effect of mutating the ETS2-binding site (–457µETS2) (Fig. 3C) on the ability of DLX3 to transactivate the –457 IFNT reporter. In this series of experiments, we found that DLX3 overexpression up-regulated wild-type –457luc by 11.6-fold relative to the basal activity (Fig. 3A). Mutation of the ETS2 site reduced basal activity by 73% and the effects of DLX3 overexpression by almost 90%. Similar cotransfection experiments with the –126luc and –126µETS2 constructs (Fig. 3B) provided further evidence that the DLX3 effects on the IFNT promoter requires the presence of the ETS2-binding site approximately two helical turns (20 bp) upstream of the DLX3 site.
We next examined whether the reverse was true—that is, whether ETS2 transactivation was dependent on an intact DLX3 site. In this series of experiments, ETS2 overexpression was conducted according to the protocol described by Ezashi et al. [11, 12] and led to an approximately 17-fold increase in luc expression from the wild-type –126luc construct (Fig. 4A). Mutation of the DLX3-binding site reduced (P < 0.001) both basal and ETS2-induced luc activity by approximately 65%. Although overall activity was reduced for the mutant construct, the fold-change between basal and ETS2-induced activity of the –126µDLX3 and wild-type –126luc constructs remained approximately the same (17.4-fold vs. 18.4-fold, respectively). When we cotransfected the ETS2 and DLX3 expression constructs with the –457luc reporter, luc activity in cell extracts increased more than 250-fold, compared with the 10-fold effect of DLX3 and the 18-fold effect of ETS2 observed when the two transcription factors were examined on their own (Fig. 4B). We noted similar DLX3-ETS2 synergistic activity with the –126luc, –721luc, and –1675luc reporters (data not shown). These results provide evidence that the relatively weak transactivational activity exhibited by both these transcription factors can be markedly enhanced when the two are expressed together.
We recently reported the role of the PKA signal transduction pathway in regulating IFNT expression through the activation of ETS2 [14]. Although overexpression of the catalytic subunit of PKA alone had a modest ability to up-regulate –126luc (5-fold) (Fig. 4C), coexpression with ETS2 led to a large increase in luc expression (500-fold) (Fig. 4C). When the binding sites for ETS2 were mutated in the construct (–126µETS2) (Fig. 4C), luc expression was markedly reduced (97%) relative to the wild-type reporter, consistent with our previous observation [14]. We then tested whether the DLX3-binding site was important for the PKA effects on IFNT expression. Mutation of the DLX3-binding site reduced the PKA/ETS2 effects by approximately 87%, and activity was virtually abolished when both the ETS2- and DLX3-binding sites were mutated (–126µETS2/µDLX3) (Fig. 4C). These results suggest that both DLX3 and ETS2 play crucial cooperative roles in controlling basal and PKA-induced IFNT expression.
Expression of ETS2 and DLX3 in JAr and CT-1 Cells and in Trophectoderm Cells of Ovine Conceptuses
Although the human genome lacks IFNT genes, JAr cells are permissive for expression of reporter constructs driven by transfected IFNT promoters, presumably because the cells express a suitable complement of required transcription factors, albeit in amounts that may not be fully optimal [33]. It is perhaps not surprising, therefore, that JAr cells express both DLX3 and ETS2, as determined by Western blot analysis of whole-cell extracts (Fig. 1, A and B) and by immunocytochemistry (Fig. 6A). Every cell observed in this experiment expressed both transcription factors within their nuclei.
|
We also examined DLX3 expression in the trophectoderm cell line, CT-1, which was originally derived from a bovine blastocyst outgrowth [27] and produces IFNT. Unfortunately, although CT-1 cells secrete IFNT constitutively, the cells are difficult to transfect, have demanding culture requirements, and proliferate slowly, thus making them unsuitable for the experiments described earlier with JAr cells. Both ETS2 and DLX3, however, were present in cell extracts derived from these cultures, as determined by Western blot analysis (Fig. 1, A and B), and were colocalized within the nuclei of most of the cells within the cultures (Fig. 6B). Not all CT-1 cells, however, were producing IFNT (Fig. 6C). Rather, IFNT was confined to the cytoplasm of a minority (approximately one third) of the CT-1 cells in the culture (Fig. 6C), and all of these cells also expressed both DLX3 and ETS2 in their nuclei (Fig. 6D). These data are consistent with the notion that coexpression of ETS2 and DLX3 is necessary for the IFNT expression but is not, in itself, sufficient to ensure IFNT expression by CT-1 cells. The controls for JAr and CT-1 cell immunocytochemistry are shown in Supplemental Figure 1 (Supplemental Fig. 1 and all other supplemental data are available online at www.biolreprod.org).
Finally, DLX3 localization was examined in sections of Day 15 filamentous ovine conceptuses, which were collected when IFNT expression was high [34, 35]. The simple epithelial layer of trophectoderm was positive for DLX3 (Supplemental Fig. 2A), with relatively higher signal intensity in nuclei than in cytoplasm (Supplemental Fig. 2C), whereas sections incubated with normal rabbit serum were essentially devoid of signal (Supplemental Fig. 2, B and D).
Coimmunoprecipitation of DLX3 and ETS2 from JAr Cell Extracts
The data presented so far indicate that DLX3 and ETS2 can associate on IFNT promoters transfected into JAr cells and, by inference, on the endogenous promoters. Here, we examined whether DLX3 and ETS2 exist together as complexes in JAr and CT-1 cells. To examine this possibility, we conducted coimmunoprecipitation experiments. Figure 1D shows that JAr cell immunocomplexes isolated by using anti-ETS2 immunoglobulin contained DLX3. Similarly CT-1 cell immunocomplexes isolated using anti-DLX3 immunoglobulin contained ETS2 (Fig. 1C). This experiment demonstrates that in these trophoblast-derived cell lines, ETS2 and DLX3 form a complex stable enough to be captured by immunoprecipitation.
We then attempted to determine whether DLX3 was associated with ETS2 on the proximal (–188 to +3) promoter region of actively transcribed IFNT by using ChIP assays. Sheared chromatin was prepared from bovine CT-1 cells. Sheared DNA collected in immunocomplexes after addition of affinity-purified sheep anti-DLX3, rabbit anti-ETS2 and rabbit anti-RNA polymerase II immunoglobulin, respectively, was subjected to PCR analysis with specific primers designed to amplify the proximal control region of the IFNT (Supplemental Fig. 3). Although the ETS2 and RNA polymerase II antibodies immunoprecipitated DNA containing the IFNT proximal regulatory region, the immunocomplexes brought down with DLX3 antibody failed to do so. Either DLX3 did not associate with the regulatory region of the IFNT promoter or the antibody was unsuitable for the ChIP analysis employed here.
Effect of DLX3 Silencing on IFNT Production by CT-1 Cells
We next examined whether suppression of endogenous DLX3 expression by siRNA silencing would influence IFNT production in CT-1 cells. We employed lipitoid reagent that delivers siRNA transfection efficiently to cells that are otherwise difficult to transfect [30]. The DLX3 concentration was significantly reduced after testing one of the three DLX3 siRNAs at 50 nM (Fig. 7A). Importantly, the intracellular concentration of IFNT was simultaneously lowered by transfection of this DLX3 siRNA (Fig. 7B), as was the antiviral activity released into the medium (untreated control cells, 10.37 ± 0.26 x 105 IU; cells transfected with nonspecific siRNA, 8.82 ± 0.67 x 105 IU; cells transfected with DLX3 siRNA, 4.07 ± 1.41 x 105 IU), with data normalized to 90 µg of cell lysate. A second experiment provided essentially similar data, although IFNT production was considerably lower in all samples, including the untreated control.
|
The results of the present study indicate that two transcription factors that have been implicated previously in the development of the mouse placenta [18, 22] act together to transactivate the proximal promoter region of the genes encoding IFNT, a major gene product of the trophectoderm of cattle for a few days during the peri-implantation phase of pregnancy. The present study also extends the concept that the signature genes of trophoblast from different species, such as IFNT from cattle and hCGA of the human, depend on a common set of transcription factors for their regulation. The presence of such transcription factors, including DLX3 and ETS2, in the trophoblast-derived cell line, JAr, probably is what makes these cells permissive for basal expression of reporters driven by IFNT promoters. On the other hand, overexpression of these proteins, which clearly increases the quantity of the ETS2/DLX3 complex in the cells relative to that in nontransfected controls (Fig. 1D), is necessary to achieve maximal reporter gene activity. Thus, both ETS2 and DLX3, although important for promoter transactivation, normally are present at limiting concentrations in some, if not all, cells.
The observation that the ETS2 and DLX3 effects on the IFNT promoter are greater than additive indicates that the two transcription factors act together synergistically. The binding sites are only two helical turns apart, and it seems probable that the proteins are partners within the same transactivational complex. The failure of the DLX3 antibody to precipitate the region of the IFNT promoter containing the putative DLX3-binding site, however, weakens this hypothesis. On the other hand, it is clear that DLX3 can bind to the site (Fig. 1) and can interact with ETS2 in CT-1 cells (Fig. 1C). The most likely explanation for this contradiction is that the epitope recognized by the DLX3 antibody, located between amino acids 197 and 287, is buried within the transcriptional complex or otherwise inaccessible to antibody. Another alternative is that the avidity of the antibody for its target is insufficient to retain DLX3 through the stringent washing procedures. We predict that this antibody is inappropriate for ChIP assays.
Like many other transcription factors, ETS2 is relatively promiscuous in its ability to associate with other known transcription factors, including CEBPB (C/EBPβ) [37], POU5F1 (OCT4) [12], and CREBBP/EP300 (CBP/p300) [14, 38]. It associates with CEBPB to up-regulate the gene for trophoblast Kunitz domain protein 1 (TKDP1), which has an expression pattern almost identical to that of IFNT [37]. The POU5F1 is of particular interest, because it silences both the IFNT [12] and hCG promoters [39, 40], most likely through its ability to quench ETS2 effects by coupling its POU domain with a region near the amino-terminus of ETS2 [12]. Conceivably, POU5F1 binding prevents a productive association of ETS2 with other transcriptional activators, such as DLX3. Exactly how ETS2 and DLX3 associate and whether the process is direct or involves a bridging factor, possibly a coactivator, such as CREBBP/EP300, remain to be determined. Conceivably, a less stable but still functional transactivating complex involving ETS2, DLX3, and a coactivator can form even when DLX3 is not tethered to DNA. Such a model may explain why ETS2 can up-regulate luc expression with the same fold-efficiency from the –126µDLX3 as the wild-type promoter (Fig. 4A). It remains unclear whether ETS2 and DLX3 need to be covalently modified to form a productive association. The ETS2 is target for both the protein kinase A [14, 16, 17] and the Ras/MAP kinase signal transduction pathways [13, 41, 42], whereas DLX3 can be activated through protein kinase C [43]. Clearly, the proximal enhancer region of the IFNT can be subject to complex informational input, much of it originating from growth factors present in the maternal uterine environment in which the conceptus resides before it attaches and forms a true placenta [13, 14].
Although both ETS2 and DLX3 cause embryonic lethality when deleted from the mouse as a result of placental defects [18, 22], the factors are best known for their role in differentiation of extraplacental lineages. DLX3 is required for the proper formation of skin keratinocytes, branchial arches, teeth, and bone [44, 45]. Like DLX3 and its relatives, ETS2 and related family members appear to act as master regulatory switches during differentiation [46, 47]. In particular, ETS2 regulates the expression of some endogenous cytokeratins found in simple epithelia [48, 49] and of several genes involved in the turnover of extracellular matrix [50, 51]. Although a link between ETS2 and DLX3 in transcriptional control has not been reported previously for any known gene, our data show that the two transcription factors probably function together to regulate IFNT expression and might possibly partner to dictate other aspects of trophoblast development.
ACKNOWLEDGMENTS
The authors thank Dr. Maria Morasso for the DLX3 expression plasmid and rabbit anti-DLX3 antibody. Dr. Mark Roberson provided rabbit anti-DLX3 antibody. The authors also thank Drs. Alan Ealy and Neal Talbot for CT-1 cells and the cell culture protocol; Drs. Mark Hannink, Shrikesh Sachdev, and Jonathan Green as well as Ms. Tina Parks for their cooperation in producing and purifying sheep anti-DLX3 antibody; Dr. Michael Kubisch for the monoclonal anti-IFNT antibody; and Dr. Ronald Zuckermann for the lipitoid reagent.
FOOTNOTES
1Supported by National Institutes of Health grants HD21896 and HD42201 to R.M.R. 
Correspondence: 2Toshihiko Ezashi, 240h, Christopher S. Bond Life Sciences Center, Columbia, MO 65211-7310. FAX: 573 884 9394; e-mail: ezashit{at}missouri.edu
Received: 14 November 2007.
First decision: 24 November 2007.
Accepted: 27 February 2008.
REFERENCES
: a novel subtype I IFN1. Structural characteristics, nonubiquitous expression, structure-function relationships, a pregnancy hormonal embryonic signal and cross-species therapeutic potentialities. Biochimie 1998; 80:755–777.[Medline] genes encoding functional class II and trophoblast (oTP) -interferons. Mol Cell Endocrinol 1991; 76:161–171.[CrossRef][Medline] genes of the giraffe, a nonbovid species. J Interferon Cytokine Res 1996; 16:949–951.[Medline] gene expression by Ets-2. Proc Natl Acad Sci U S A 1998; 95:7882–7887.-interferon promoter by oct-4. Mol Cell Biol 2001; 21:7883–7891. (IFN-) gene promoters by growth factors that target the Ets-2 composite enhancer: a possible model for maternal control of IFN- production by the conceptus during early pregnancy. Endocrinology 2004; 145:4452–4460. expression in a trophoblast cell line. Mol Endocrinol 2008; 22:331–343. gene expressions are regulated through the short promoter region including Ets-2 binding site. Mol Reprod Dev 2005; 72:7–15.[CrossRef][Medline] by Ets-2 and cAMP-responsive element binding protein. Mol Endocrinol 2005; 19:1049–1066. and the homeodomain protein distal-less 3 are required for placental-specific expression of the murine 3β-hydroxysteroid dehydrogenase VI gene, Hsd3b6. J Biol Chem 2002; 277:7945–7954. subunit gene in choriocarcinoma cells. J Biol Chem 2001; 276:10016–10024. and transferrin expression as respective in vitro markers. Biol Reprod 2000; 62:235–247. genes: relative transcript abundance during early placental development, promoter sequences of genes, and biological activity of protein products. Endocrinology 2001; 142:2906–2915. administration on endometrium of nonpregnant ewes: a comparison with pregnant ewes. Endocrinology 2006; 147:2127–2137.-subunit of human chorionic gonadotropin by the embryonic transcription factor Oct-3/4. Mol Endocrinol 1997; 11:1651–1658.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
