Bmc Molecular Biology Ets-2 and C/ebp-beta Are Important Mediators of Ovine Trophoblast Kunitz Domain Protein-1 Gene Expression in Trophoblast

Background: The trophoblast Kunitz domain proteins (TKDPs) constitute a highly expressed, placenta-specific, multigene family restricted to ruminant ungulates and characterized by a C-terminal "Kunitz" domain, preceded by one or more unique N-terminal domains. TKDP-1 shares an almost identical expression pattern with interferon-tau, the "maternal recognition of pregnancy protein" in ruminants. Our goal here has been to determine whether the ovine (ov) Tkdp-1 and IFNT genes possess a similar transcriptional code.


Background
The trophoblast Kunitz domain proteins are a family of closely related gene products uniquely expressed by trophoblast of ruminant ungulates [1,2]. Each protein is characterized by the presence of a well-conserved, carboxyl-terminal ~65 amino acid domain, typically found in the Kunitz family of serine peptidase inhibitors. The Kunitz domain is an evolutionary conserved module, existing in C. elegans, D. melanogaster, and all known vertebrates, that generally functions as a serine peptidase inhibitor, although functions unrelated to peptidase inhibition have also been documented [3][4][5][6]. This independently folding unit is preceded by one or more "repeat" sequences of ~80 amino acids of unknown function, which have similar but not identical sequences [1,2]. TKDPs represent one of the most abundant classes of secretory proteins produced by the conceptus during the peri-implantation period and throughout gestation [1,2,7,8]. Although the function of these proteins remains unclear, their unusually high expression at the interface between the maternal and fetal systems and the presence of conserved Kunitz units at their carboxyl termini strongly suggest an important role during pregnancy that relates to the evolution of the non-invasive placental type encountered in ruminant species.
Although expressed abundantly, the concentrations of different TKDP mRNAs do not remain constant over the course of pregnancy. Instead, expression of each TKDP appears to be quite tightly regulated and confined to specific stages of fetal development. For example, ovTKDP-1 and its bovine ortholog (boTKDP-1) are expressed significantly for just a few days preceding the period when the trophoblast begins to adhere tightly to the uterine wall [2] In this respect, the temporal expression pattern of TKDP-1 and that of IFN-τ, the major signal for maternal recognition of pregnancy in cattle and sheep are remarkably similar. For example, both TKDP-1 and IFN-t are secreted by the trophoblast mononuclear cells during days 13-21 of pregnancy in sheep [7]. In view of the physiological importance of IFN-τ in ensuring the survival of the conceptus [9,10] and the fact that both TKDP-1 and IFN-τ are up-regulated massively and simultaneously at a particular critical time when many embryos are lost, it seems likely that ovTKDP-1 has some function important for mediating maternal-conceptus interaction during the peri-attachment/implantation period of conceptus development.
The strikingly similar spatial-temporal expression pattern of TKDP-1 and IFN-τ raises the possibility that these genes (Tkdp-1 and IFNT) representing these proteins are controlled by very similar transcriptional mechanisms and possibly share common transcription factor binding elements in their promoter regions. The major focus of the current research has been to gain insight into the control of Tkdp-1 gene expression and to determine whether the mechanisms involved resembled those associated with the regulation of the IFNT. One important goal has been to elucidate whether or not there is a common theme to the transcriptional regulation of genes that are first expressed as the trophoblast begins to differentiate functionally early in pregnancy and secrete its signature gene products. The outcome of this research could, therefore, have implications to the establishment and/or maintenance of pregnancy and the specification of trophoblast function in all mammals.
The data presented here demonstrates Ets-2 having an important role in activating the transcription of the ovTkdp-1 gene in in vitro cell culture systems. However, unlike the IFNT promoter, which directly interacts with Ets-2 via consensus Ets-binding site, the Ets-2 effect on the Tkdp-1 promoter is indirect and mediated by protein-protein interaction with C/EBP-β. In other words, Ets-2dependent transactivation of genes with nearly identical spatial and temporal expression patterns can be mediated by different mechanisms. reduced significantly when the promoter was shortened to 558 bp. Promoter activity increased again in the 352, 254, 192 and 140 bp constructs, but was never completely restored to that of the 1000 bp construct. These data suggest that there are two enhancer/activator elements present, one distal, between -558 and -1000 bp, and the other more proximal, lying between 140 and 82 bp. There may also be a cryptic repressor element lying between the 558 bp and 352 bp promoter fragments, although another explanation for the poor activity of the 558 construct is that it folds in such a way that it is less easily transcribed than its longer and shorter relatives. Since the 140 bp construct was the shortest one able to drive Luc expression effectively, contained the proximal enhancer element, and included several relevant transcription factor binding sites (see below) our efforts became focused on this control region of the gene. Fig. 3 illustrates putative transcription factor binding sites within 140 bp upstream of the tsp of the Tkdp-1 promoter that can also be identified within the boIFNT1 proximal promoter (-126 to +50) and have been implicated in the basal expression of IFNT genes in vitro [11,12]. No Ets-2/ AP-1 composite enhancer motif comparable to the one essential for IFNT gene transcription is detectable within the proximal 140 bp of the Tkdp-1 promoter, although a number of potential Ets-factor binding elements are present. None of these sites possesses a sequence identical to the Ets-2-binding element found within the IFNT promoter (-79 ACAGGAAGTG -70), although each does possess the central GGA motif that is an invariable feature of all Ets-factor-binding sites [13,14] (Fig. 3 and additional data file 1 - Fig. 1). A possible AP-1 element is present in the Tkdp-1 promoter, but its sequence is different from that of the one placed just upstream of the Ets-2-binding site on all known IFNT genes (TGAGAGA versus TGAG/ CTCA) [11].

Effects of over-expressing Ets family transcription factors on the transcriptional activity of the Tkdp-1 promoter in JEG-3 cells
Although expressed widely in adult and embryonic tissues, the Ets family transcription factor Ets-2 is crucial for placental development [15]. It is also present in ovine conceptus tissue at the time that IFN-τ is produced and is known to support expression of several genes that are restricted to trophoblast [16][17][18][19][20][21][22], including IFNT [23]. Since the Tkdp-1 gene possesses several Ets-like sites within its promoter region, transient transfection experiments were performed to test the effects of over-expressing several Ets family members on the transcriptional activity of the 1000 bp promoter-reporter construct in JEG-3 cells (Fig. 4A). In order to provide roughly comparable expres-Detection of ovTKDP-1, IFN-τ, Ets-2 and C/EBP-β mRNA in sheep trophectoderm Figure 1 Detection of ovTKDP-1, IFN-τ, Ets-2 and C/EBP-β mRNA in sheep trophectoderm. RT reactions were performed on sheep trophoblast RNA collected from different days of pregnancy. After binding to oligo dT primer, reverse transcription and PCR amplification with primers specific for each message were performed. A) Comparison of ovTKDP-1 and IFN-τ expression during the peri-implantation period of conceptus development. Lanes 1-3: positive control PCR with TKDP-1, IFN-τ and S25 plasmid templates, respectively; lanes 4-9: PCR with days 14,15,16,17,19  To examine which part of the Tkdp-1 promoter is responsive to Ets-2, an equal amount of three truncated Tkdp-1 promoter-reporter constructs, 1000 bp, 352 bp and 140 bp, was co-transfected into JEG-3 cells with the Ets-2 expression plasmid. Ets-2 over-expression increased Luc activity of all three constructs quite similarly (Fig. 4B), although the 140 bp promoter was slightly more responsive than the 1000 bp promoter (Fig. 4B: 28-fold for 1000 versus 34-fold for 140). These data suggested that the region of the promoter targeted by Ets-2 is within 140 bp of the tsp. We also tested the effect of Ets-2 over-expression on the transcriptional activity of the promoter-less Luc reporter vector (pGL-2 Basic) (data not shown). As expected, Ets-2 had no significant effect on the reporter expression in absence of a promoter, indicating that the Ets-2-mediated activation of the Tkdp-1 promoter-Luc construct is a promoter-specific effect.
Ets-2-mediated activation of the Tkdp-1 promoter was dose-dependent (Fig. 4C). Differences in fold-activation between experiments probably reflect the quality of the plasmid DNA and other variables not easily controlled in transfection studies. A time-course experiment was also performed as a preliminary test to determine whether the Ets-2-responsiveness of the ovTkdp-1 promoter is early or delayed. An early response is indicative of a direct effect while a delayed response would suggest Ets-2-mediated up-regulation of some other transcription factor necessary for expression of the Tkdp-1 gene. When JEG-3 cells, cotransfected with Ets-2 and the 1000 bp ovTkdp-1 promoter constructs, were harvested at various time points after transfection (data not shown), Ets-2-responsiveness of the Expression of a series of Tkdp-1 promoter-Luc reporter constructs transfected into JEG-3 cells

Ras/MAPK-mediated activation of theTkdp-1 promoter in 3T3 cells
Ets-2 and the related transcription factor Ets-1 bind to specific DNA elements via their Ets domains, but often require activation through Ras/MAPK-dependent phosphorylation at specific residues in order to participate effectively in target gene transcription [11,16,[24][25][26][27][28]. To test whether the 1000 bp Tkdp-1 promoter-reporter construct was similarly responsive to MAPK activation, it was co-transfected with a constitutively active Ras expression plasmid in presence and absence of Ets-2 in mouse fibroblast 3T3 cells. These mouse fibroblasts were used because, unlike choriocarcinoma cells, they possess low endogenous levels of Ets-2 and activated Ras [11], so that the effect of over-expression is easily demonstrable. Ras in combination with Ets-2, but not by itself, caused a 21-fold activation of the 1000 bp promoter construct (Fig. 5). This stimulation of Tkdp-1 promoter activity was reduced approximately 96% by treating the transfected cells with the MAPK kinase (MEK1) inhibitor PD98059. If the codon encoding the Ras/MAPK target site, Thr 72, was mutated, Tkdp-1 promoter activity was reduced by about 50% compared to that obtained with the Ras and wild type Ets-2 combination. These data show that the Tkdp-1 promoter can be up-regulated by the Ras/MAPK pathway acting through Ets-2.

cAMP/PKA-mediated activation of the Tkdp-1 promoter in JEG-3 cells
The cAMP/PKA signaling pathway has been shown to modulate the Ets-dependent activation of several trophoblast-specific genes, including the human chorionic gonadotropin (hCG)-α and β subunit genes [16,17] and IFNT [17]. Treatment of JEG-3 cells with increasing con-Possible transcription factor-binding sites in the proximal ovTkdp-1 and boIFNT1 promoters Figure 3 Possible transcription factor-binding sites in the proximal ovTkdp-1 and boIFNT1 promoters. The sequence of the Ets-AP-1 composite enhancer in the boIFNT1 promoter is also shown [23]. The tsp is designated +1.

+ Strand
centrations of the cAMP analog, 8-Br-cAMP, after co-transfecting the cells with the 1000 bp Tkdp-1 promoterreporter construct and the Ets-2 expression plasmid provided only a 2-3-fold enhancement of the Ets-2 effect (Additional data file 1 - Fig. 2). 8-Br-cAMP alone, i.e. in absence of over-expressed Ets-2, had no significant ability to up-regulate the promoter (data not shown).
One possibility of explaining why 8-Br-cAMP had such a minor effect on the Tkdp-1 promoter was that in JEG-3 cells PKA, the downstream target of cAMP, is present in limiting amounts. Accordingly, the effects of a constitutively active catalytic subunit of PKA were examined. As shown in Fig. 6A, over-expression of the constitutively active catalytic subunit of PKA led to a 120-fold up-regulation of the basal activity of the 1000 bp Tkdp-1 promoter in JEG-3 cells. When Ets-2 was also over-expressed, activity of the promoter was enhanced over 700-fold. The PKA and PKA+Ets-2 effects on the Tkdp-1 promoter activation were inhibited over 95% by expression of a PKA inhibitor (PKI) construct (Fig. 6A). Additionally, substitution of the plasmid expressing the active form of PKA with one with a mutated catalytic site (mutPKA) failed to up-regulate the Tkdp-1 promoter (Fig. 6A). Finally, treatment of trans-fected cells with a pharmacological PKA inhibitor, H89, caused dramatic reduction in promoter activity, both in absence and presence of Ets-2, in a concentration-dependent manner (Fig. 6B). These data indicate that the Tkdp-1 promoter is responsive to the cAMP/PKA signal transduction pathway and that Ets-2 acts synergistically with this pathway to up-regulate this promoter.

Indirect interaction between Ets-2 and the Tkdp-1 promoter
As mentioned previously, other than having the core GGA sequence, none of the putative Ets-like sites within the Tkdp-1 promoter showed a close match with either the consensus Ets-2-binding motif (C/A)(C/A)GGA(A/T)(A/ G) or the well defined, functional site (CAGGAAG) in the IFNT promoter. In order to determine whether any of these putative Ets-like sites within the Tkdp-1 promoter binds Ets-2, competition electrophoretic mobility shift assay (EMSA) was performed. The 140 bp minimal promoter was chosen because it had multiple potential Etslike sites (Fig. 3) and provided full Ets-2 responsiveness (Fig. 4B). Initial experiments in which we employed EMSA to demonstrate an interaction between recombinant Ets-2 and oligonucleotides representing these potential Ets-binding sequences failed to reveal the formation of any complexes (data not shown). As shown in Fig.  7, none of the six potential competitor oligonucleotides (lanes 5-10) was able to disrupt either the slower or faster migrating DNA-protein complexes formed between the 32 P-labeled Ets consensus oligonucleotide and ovEts-2-GST-fusion protein (lanes 1 and 3). The slight variation in intensities in the faster and slower migrating specific complexes across lanes 5-10 likely resulted from differences in gel loading. EMSA experiments employing labeled oligonucleotides representing the putative Ets-binding sites within the 140 bp promoter failed to provide specific complexes with recombinant Ets-2 (data not shown). Together, these results indicate that Ets-2 is unable to associate with the proximal promoter and that none of the putative Ets-binding sites present is likely to be functional.

Mutation of the C/EBP binding site in theTkdp-1 minimal promoter reduces its activity
When the CCAAT/enhancer element (TTATGCAAT) of the 140 bp Tkdp-1 promoter-reporter construct was mutated (TTATcCccc) both the basal transcriptional activity and the effects of Ets-2, PKA and Ets-2+PKA were markedly reduced (~90 %) compared to the wild type promoter in JEG-3 cells (Fig. 9A). That the C/EBP-binding site mutation had no additional effect on Ets-2-dependent than on basal promoter activation was somewhat surprising, However, basal promoter activity is quite high in JEG-3 cells, e.g. compared to the non-trophobast cell line, NIH3T3, possibly because of the higher content of endogenous Ets2 and other essential transcription factors present in these trophoblast cells. Accordingly, the effect of the mutation on basal activity may be difficult to distinguish from the effect observed when Ets-2 is expressed ectopically. Reduction in the basal promoter activity and the Ets-2 or Ets-2+Ras responsiveness was also observed in the 3T3 mouse fibroblast cell line when the CCAAT/ enhancer element was mutated (Fig. 9B), and, the effect of the mutation when Ets-2 was over-expressed was more pronounced. Thus, mutation of the CCAAT/enhancer element largely eliminated the promoter's ability to respond to Ets-2 and the cAMP/PKA or the Ras/MAPK signal transduction pathways. These data along with our earlier observation (Fig. 8A) that C/EBP-α, -β or -δ has no effect on the reporter gene expression by its own, suggest that the effects of PKA, Ets-2, Ets-2+PKA and Ets-2+Ras on the 140 bp Tkdp-1 promoter are mediated through the C/EBPbinding site, conceivably through a direct interaction of Ets-2 with C/EBP-β.

C/EBP-β directly interacts with theTkdp-1 promoter
In order to identify which of the three C/EBP family members used in the transfection experiments ( Fig. 8A and 8B) interacts with the 140 bp promoter in vitro, we performed EMSA with a 32 P-labeled double-stranded (ds) oligonucleotide possessing the C/EBP-binding site of the Tkdp-1 promoter, JEG-3 cell extracts, and antibodies specific for C/EBP-α, -β, and -δ (Fig. 10). As shown in lane 1, a specific DNA-protein complex was formed between the CCAAT/ enhancer element and JEG-3 cell extracts. Excess unlabeled competitor oligonucleotide containing the wildtype C/EBP-binding site (TTATGCAAT; lane 7) inhibited formation of the specific complex when it was added to Effects of Ets-2 and activated Ras on the Tkdp-1 1000 bp pro-moter activity in NIH3T3 cells  Luciferase Fold Induction the reaction mixture. However, an unlabeled oligonucleotide possessing a mutated C/EBP-binding site (TTATc-Cccc; lane 8) failed to compete in the binding reaction. The addition of an antiserum directed against individual C/EBPs only provided a major "supershift" when the anti-C/EBP-β reagent was employed (lane 4). There was a minor effect noted with anti-C/EBP-α (lane 3). Neither anti-C/EBP-δ (lane 5) nor non-immune rabbit IgG (lane 6) had any effect. Thus C/EBP-β would appear to be the major isoform of C/EBP that binds the TTATGCAAT element present in the 140 bp minimal promoter of the Tkdp-1 gene.

Interaction between C/EBP-β and Ets-2 in JEG-3 cells
Co-immunoprecipitation (Co-IP) and oligonucleotide pull-down experiments were performed to demonstrate the in vivo interaction between C/EBP-β and Ets-2 in JEG-3 cells. For the former, endogenous Ets-2 in JEG-3 cell extracts was allowed to form an immune complex with a rabbit polyclonal Ets-2-specific immunoglobulin, collected on Protein-A beads and then subjected to western blot analysis with C/EBP-β-specific antiserum (Fig. 11A, The data indicate that endogenous C/EBP-β is associated with Ets-2 in JEG-3 cell extracts. No C/EBP-β protein was detected when non-immune rabbit IgG was substituted for the anti-Ets-2 antiserum (lane 4). As positive control, endogenous C/EBP-β was collected with C/ EBP-β-specific antiserum and the western blot developed with the same antiserum (lane 2).
When a streptavidin-bound, biotinylated ds oligonucleotide representing the C/EBP-binding region of the Tkdp-1 promoter was used as a bait to collect proteins present in JEG3 cells, the oligonucleotide was able to trap endogenous C/EBP-β (lane 3, Fig. 11B). This ability to bind C/ EBP-β was inhibited in the presence of an excess of unlabeled wild type competitor oligonucleotide, but not by unlabeled mutated competitor (lanes 5 versus 4). The C/ EBP-binding site containing biotinylated oligonucleotide was able to trap Ets-2 as well as C/EBP-β (lane 3, Fig. 11C), even though it did not possess a known Ets-binding sequence. The Ets-2-specific bands were not detected when excess unlabeled, wild-type competitor oligonucleotide was included in the reaction mixture (lane 4, Fig. Effects of Ets-2 and activated PKA on the Tkdp-1 1000 bp promoter activity in JEG-3 cells   11C). These data provide conclusive evidence that Ets-2 associates with C/EBP-β while the latter is bound to its CCAAT-binding-sequence. The possibility that Ets-2 bound to the CCAAT/enhancer element directly is unlikely in view of earlier data (Fig. 7).

Mutation of the AP-1-binding site does not affect the Ets-2-responsiveness of the Tkdp-1 promoter
The 140 bp Tkdp-1 promoter possesses a conserved AP-1 element (-99 to -93) immediately downstream of the CCAAT/enhancer element (-117 too -109). In order to determine the functional importance of this site for transcriptional activation of the promoter, site-directed mutagenesis followed by transient transfection experiments were performed. As demonstrated in Fig. 12A, mutation of the AP-1 element caused up to 90% reduction in basal activity and 74% reduction in PKA-dependent promoter activation. This mutation did not affect the Ets-2-mediated ovTkdp-1 promoter activation, although it reduced the effect of the Ets-2 plus PKA combination on the promoter by up to 65%. Together these data indicate that the AP-1 element adjacent to the C/EBP-β-binding site is important in regulating transcription from the Tkdp-1 minimal promoter and for activation of transcription by the cAMP/PKA signaling pathway, but is not involved in the recruitment of Ets-2 to the promoter.

The AP-1 element in the 140 bpTkdp-1 promoter interacts with C/EBP-β
Pull-down assays were conducted in JEG-3 extracts with a biotinylated, ds oligonucleotide containing positions -99 to -93 of the Tkdp-1 promoter (Fig. 12B and 12C). Some interactions between this element and several Jun family transcription factors (c-Jun, Jun B, Jun D and p39) could be demonstrated by western blotting (Fig. 12C). By contrast none of the Fos-related proteins (c-Fos, Fos B, Fra-1 and Fra-2) was pulled down from the JEG-3 cell extracts (negative results not shown). Unexpectedly, this AP-1 element-protein complex contained C/EBP-β, suggesting that whatever bound to this sequence also associated with C/EBP-β (Fig. 12B).

Expression of Ets-2 and C/EBP-β transcripts in conceptuses
RT-PCR analyses performed on RNA extracted from sheep conceptuses/trophoblast at days 14, 15, 16, 17 and 25 of pregnancy showed that C/EBP-β mRNA, like Ets-2 mRNA, is expressed relatively uniformly over these days of conceptus development (Fig. 1B and 1C). This expression pat-Competition EMSA with recombinant Ets-2 protein and a 32 P-labeled ds Ets-2 consensus oligonucleotide  tern is in sharp contrast to that of the Tkdp-1 gene, whose expression pattern is maximal around day 17 and then tails off by day 25.

Discussion
TKDP-1 is an abundant secretory product of the ruminant placental trophoblast cells during the peri-implantation stage of pregnancy, a period when the enlarging blastocyst starts to signal its presence to the mother in preparation for establishing an intimate contact with the uterine endometrium [2,7,35]. This stage of conceptus development is also marked by the production of IFN-τ, the protein responsible for "rescue" of the corpus luteum during early pregnancy in ruminant ungulate species, such as cattle, sheep, goat and deer. Like ovTKDP-1, IFN-τ is a product of trophoblast mononuclear cells, and its production rises sharply as the blastocyst expands and begins to elongate, presumably in response to growth factors, hormones and cytokines present in the uterine secretions of the mother [2,11,12]. Many of such factors operate through activation of established signal transduction pathways, including the Ras/MAPK and cAMP/PKA pathways [12,36,37]. Shutdown of expression after the trophoblast begins to adhere to the uterine wall might be due to lack of access of the adherent trophectoderm to these maternal factors, which probably originate from the uterine glands of the underlying endometrium. The striking similarity in the expression patterns of ovTKDP-1 and IFN-τ, which is illustrated in Fig. 1A, led us to speculate that the genes for these two proteins might be controlled by similar transcriptional mechanisms, involving the same transcription factors and signal transduction pathways.
The starting point for our study on Tkdp-1 transcriptional control was the Ets family transcription factor, Ets-2, a homolog for the viral oncogene of the avian leukemia virus E26 [38]. Although expressed widely in adult and embryonic tissues, Ets-2 is required for placental development in the mouse [39] and supports expression of several genes, including the IFNT, whose expression provide a phenotypic "signature" for trophectoderm [15][16][17][18][19][20][21]23]. The promoters of each of these genes possess functional Ets-2-binding sites which, when mutated, cause almost complete loss of reporter gene expression. Ets-2 transcripts are certainly present in the ovine conceptus (Fig.  1B), and Ets-2 protein is expressed in trophectoderm of bovine conceptuses of equivalent developmental stage [22], but appears to be limiting for driving maximal
Based on the fact that the Tkdp-1 promoter possessed several sites that might bind Ets-2 or one of its relatives, transient transfection experiments were performed in JEG3 choriocarcinoma cells to test the effects of various Ets family transcription factors on the Tkdp-1 promoter expression (Fig. 4A). As expected, Ets-2 proved most effective of the factors tested and up-regulated the Tkdp-1 promoter about as well as it did the IFNT promoter [12]. The ability of Ets-2 to drive expression from the Tkdp-1 promoter in 3T3 cells required an activated MAPK pathway (Fig. 5) and was at least partially dependent on the MAPK target residue, Thr 72, in the so-called "pointed" domain of the protein [23]. These results were entirely consistent with Ets-2 effects on the IFNT and hCG subunit gene promoters when they were examined in 3T3 cells [16,23,28]. In human choriocarcinoma cells, expression was highly responsive to over-expression of the catalytic subunit of PKA. A combination of PKA and Ets-2 over-expression could up-regulate the promoter over 700-fold (Fig. 6A). Again, in all these respects, the Tkdp-1 promoter behaved similarly to the IFNT, hCGα, and hCGβ promoters [11,16,17,23]. Finally, we have also demonstrated that the Ets-2 effects on the Tkdp-1 promoter are silenced by the expression of Oct-4 (Chakrabarty, A. and Roberts, R. M., unpublished results), another feature shared with the, hCGα, hCGβ, and IFNT genes [29,40,41]. These similarities suggest that all these trophoblast-expressed genes share a common transcriptional control mechanism, with a central role played by Ets-2 in each, which allows the genes to be up-regulated soon after the trophectoderm cell lineage first emerges.
In the case of the IFNT, hCGα and hCGβ genes, however, transactivation by Ets-2 is entirely dependent upon the ability of the transcription factor to bind directly to DNA at recognizable, although relatively diverse, Ets-2-binding sequences [11,16,17,23,42]. Strikingly, the proximal/ minimal Tkdp-1 promoter region between 82 and 140 bp from the tsp, which is responsive to Ets-2 ( Fig. 2 and 4B), seems physically incapable of forming a stable association Lucife ras e Fold Induction EMSA with 32 P-labeled C/EBP oligonucleotide (-117 to -109) and JEG-3 cell extracts (CE)

A B C
with Ets-2 in vitro (Fig. 7). The most obvious conclusion is that Ets-2 must form an association with a second transcription factor that binds within this critical region, and, when so positioned, is able to influence transcription from the Tkdp-1 gene promoter about as efficiently as it does when it binds directly to the IFNT promoter.
The next goal, therefore, was to elucidate the mechanism by which Ets-2 controls Tkdp-1 gene expression without a direct interaction with the DNA. Our initial focus was on a conserved CCAAT/enhancer element located between 117 and 109 bp up-stream of the tsp (Fig. 3). There are sev-eral reasons to consider the possibility that a CCAATbinding protein might regulate the Tkdp-1 gene. First, C/ EBP-α, -β and -δ are expressed in human placental cells [29]. Second, C/EBP-β transactivates multiple genes that are expressed by trophoblast cells, including the gene for the homeobox protein distal-less-3 (Dlx3) [43], a transcription factor that binds to the IFNT promoter and cooperatively enhances its Ets-2 responsiveness (Ezashi, T. and Roberts, R. M.; unpublished data). Third, recent experiments have suggested a crucial role for C/EBP-α and -β in embryogenesis, since deletion of both genes results in mortality around embryonic day 10-11 due to gross Importance of the AP-1 element (-99 to -93) for the Tkdp-1.140 bp promoter activity in JEG-3 cells  Lucifer as e Fold Induction failure in placental development [44]. Fourth, C/EBP-α and -β have been demonstrated to interact with several Ets family members, including Elk-1, Fli-1, Ets-1 [30,32,33,45]. Fifth, C/EBP proteins, particularly C/EBPβ, are responsive to both cAMP/PKA and MAPK pathways [34,46]. Activation of the cAMP/PKA pathway promotes C/EBP-β gene transcription, nuclear translocation, recruitment of transcriptional co-activators (e.g. CBP/p300) and phosphorylation at specific serine residues [46][47][48][49][50][51][52][53][54][55], while the major effect of the MAPK pathway on C/EBP-β is mediated through phosphorylation at specific threonine residues, which leads to its activation [55,56].
When over-expressed in JEG-3 cells, none of the C/EBP isoforms -α, -β, and -δ had an effect on the 140 bp Tkdp-1 minimal promoter. In presence of Ets-2 (Fig. 8A), a modest increase in promoter activity was observed. When Ets-2+C/EBP-β-transfected cells were treated with 250 μM Br-cAMP, a synergistic increase in promoter activity was observed, particularly in presence of C/EBP-β (Fig. 8B). Thus, in JEG-3 cells, maximal effects of the C/EBP proteins on the Tkdp-1 promoter require the participation of both Ets-2 and the cAMP/PKA signaling pathway. Transient transfection experiments and EMSA indicated C/EBP-β was the most potent of the C/EBP isoforms that bound to the CCAAT/enhancer element ( Fig. 8A and 8B).
Further importance of the CCAAT/enhancer element in controlling transcriptional activity of the 140 bp Tkdp-1 promoter was demonstrated by mutational analysis (Fig.  9A and 9B). Not only was an intact CCAAT element required for C/EBP-β binding and basal promoter activity, but also for Ets-2-and Ets-2+PKA-mediated activation (Fig. 9A). Thus, C/EBP-β bound to the CCAAT/enhancer element is likely to be the target for the cAMP/PKA signaling pathway and also responsible for recruiting Ets-2. Further support for the importance of C/EBP-binding site in Ets-2-dependent activation of the Tkdp-1 promoter came from co-transfection experiments in NIH3T3 cells, where the mutated promoter demonstrated markedly reduced Ets-2 and Ets-2+Ras effects (Fig. 9B). Co-IP and pull-down assays confirmed a direct association between C/EBP-β and Ets-2 proteins in JEG-3 cells (Fig. 11A, B and 11C). Together, these experiments indicate that C/EBP-β binds to the Tkdp-1 promoter between -117 and -109 and that Ets-2 is recruited to this site via protein-protein interaction with C/EBP-β. It should be noted here that due to the unavailability of an appropriate ovine trophoblast cell line, further demonstration of the in vivo association between C/EBP-β and Ets-2 proteins in context of the ovTkdp-1 promoter using the chromatin immunoprecipitation assay was not possible.
Our next focus was on a putative AP-1 element at -99 to -93, adjacent to the CCAAT/enhancer motif in the Tkdp-1 promoter (Fig. 3), since AP-1 family transcription factors, especially c-Jun and c-Fos are known to interact with both Ets and C/EBP proteins [49,[57][58][59][60][61]. As shown by sitedirected mutagenesis and transfection experiments, this AP-1 site appeared crucial for maintaining basal activity of the Tkdp-1 promoter in JEG-3 cells, as well as for the full PKA responsiveness of the promoter. On the other hand, mutation of the AP-1 site had no effect on Ets-2 responsiveness (Fig. 12A). In other words, PKA-dependent activation of the Tkdp-1 promoter may be mediated by some factor bound to the AP-1 site adjacent to the CCAAT/ enhancer motif. Pull down experiments with a broad spectrum antibody that recognizes multiple members of the Jun family demonstrated that some of these proteins did indeed bind to this site (Fig. 12C). However, C/EBP-β itself was also a part of the protein complex (Fig. 12B).
Since C/EBP proteins bind to DNA as either homo-or heterodimers [62], and C/EBP-β is known to form complexes with c-Jun [63], it is tempting to speculate that the AP-1 site adjacent to the CCAAT element in the Tkdp-1 promoter recruits a heterodimer consisting of C/EBP-β and a member of the Jun family of proteins.
A search of the recent draft of the bovine genome sequence (see website) [64] allowed us to identify the putative bovine ortholog of the Tkdp-1 gene (between positions 424335-407548 on contig no. NW_928686 on chromosome 13). There is 89% sequence identity between the ovine and bovine genes over the 140 bp region directly upstream of the transcription start site determined for the ovine gene (Additional data file 1 - Fig. 3). The single base mutation in the CCAAT element in the bovine gene (TTAcGCAAT versus TTATGCAAT) would likely not destroy the core C/EBP-binding motif (TKNNG-NAAK ; where K = T or G and N = any nucleotide) [65], while the single base mutation in the AP-1-binding site (TGACcCA versus TGACTCA) might well reduce its ability to interact with typical AP-1 proteins. The high degree of conservation in the C/EBP-β binding sequence maintained over the 17 million or so years since the lineage leading to modern day cattle and sheep diverged provides an additional argument for the importance of the CCAAT motif.

Conclusion
The pattern of expression of the ovine trophoblast Kunitz domain protein-1 produced by the trophectoderm during the peri-implantation stage of embryo development is almost identical to that of IFN-tau and directed by the same transcription factor, Ets-2, in combination with the protein kinase A and mitogen-activated protein kinase signal transduction pathways, but does not involve direct binding of Ets-2 to promoter control elements. Instead, up-regulation of the gene is accomplished indirectly through protein-protein interactions with C/EBP-β. Most importantly, the work reported here demonstrates how Ets-2, a key transcription factor for trophoblast differentiation and function, can control expression of genes having similar spatial and temporal expression patterns via very different mechanisms.

Reverse transcription polymerase chain reaction (RT-PCR) analyses of sheep conceptus RNA
Animal husbandry and surgical procedures were performed according to the protocols approved by the Animal Care and Use Committee at the University of Missouri-Columbia [66,3097]. Expression patterns of TKDP-1, IFN-τ, Ets-2, C/EBP-β, and ribosomal protein S25 mRNAs were determined by RT-PCR on sheep conceptus/ trophoblast RNA collected on days 14,15,16,17,19 and 25 of pregnancy. RNA was extracted by using RNA STAT-60 reagent (Tel-Test, Friendswood, TX) [66]. RT reactions were performed on 1-2 μg RNA (pre-treated with deoxyribonuclease; Ambion, Inc., Austin, TX) by using Super-Script III ribonuclease Hreverse transcriptase (Invitrogen, Carlsbad, CA) in presence of 50 μM oligodeoxythymidine (dT) primer. RT reaction product (5 μl) was used for each 50 μl PCR reaction with 100 ng of sense and anti-sense primers (Additional data file 1 - Table 1) and PicoMaxx High Fidelity PCR Master Mix (Stratagene, La Jolla, CA). Twenty-five PCR cycles were used to amplify TKDP-1, IFN-τ and S25 messages, while 30-35 cycles were needed to visualize the Ets-2 and C/EBP-β-mRNAs. PCR products were resolved through 1.5% TAE (Tris-acetate-EDTA)-agarose gels at 100 volts for 1 h and stained with ethidium bromide. In order to confirm that RNA preparations were free of DNA, control RT reactions were performed without initial reverse transcription. As positive controls we employed 10 ng plasmid DNAs encoding the open reading frames for ovTKDP-1, ovIFN-τ4, Ets-2, C/ EBP-β and ribosomal protein S25.

Constructs, antibodies and cell culture reagents
The transcription start point (tsp) of the Transfection results were expressed as average fold induction in Luc reporter activity ± SEM. Statistical test on differences in relative Luc activities (normalized values) were performed by one way ANOVA followed by Tukey's multiple comparison tests (GraphPad Prism version 4).

Electrophoretic mobility shift assay (EMSA)
EMSA was carried out with 32 P-labeled ds Ets-2 oligonucleotides (35 fmole) [69] (Additional data file 1 - Table 4) and 2 μg of ovEts-2-GST fusion protein [21] (with GST as control). Competition was provided by addition of either a 250-fold molar excess of unlabeled Ets-2 consensus oligonucleotide or each of six overlapping oligonucleotides spanning the -150 bp to + 20 bp region on the Tkdp-1 promoter ( Table 4 in additional data file 1 -data; tsp = +1). Competitor oligonucleotides were added prior to the addition of the labeled oligonucleotide to the binding reaction mixtures. Procedures for DNA binding reactions and electrophoresis have been described previously [23]. The specificity of Ets-2 binding was verified by the ability of 0.3 μg of Ets-2 antibody to retard the DNA-protein complex during electrophoresis.
EMSA was also performed with the labeled, ds CCAAT/ enhancer element consensus sequence (35 fmol) from the Tkdp-1 gene (between -117 and -109) and 10 μg of JEG-3 cell extracts prepared with 200 mM (pH 7.8) sodium phosphate buffer, 0.5% triton X-100, 0.5 mM dithiothreitol and 1× peptidase inhibitor cocktail (Sigma, St. Louis, MO). The specificity of DNA-protein complex formation was verified by adding 250-fold molar excess of unlabeled wild type or mutated C/EBP oligonucleotide (Additional data file 1 - Table 4). In order to identify which of the three C/EBP isoforms among -α, -β and -δ bound to the CCAAT/enhancer element from the Tkdp-1 gene, 0.4 μg of affinity purified rabbit polyclonal antibody specific to each these transcription factors was added to each reaction.

Co-immunoprecipitation reactions (Co-IP)
Rabbit IgG TrueBlot kit (eBioscience, Inc., San Diego, CA) was used forCo-IP experiments with JEG-3 cells treated with 250 μM 8-Br-cAMP for 24 h prior to cell lysis. Cell extracts were prepared with 200 mM (pH 7.8) sodium phosphate buffer, 0.5% triton X-100, 0.5 mM dithiothreitol and 1× each of the peptidase and phosphatase inhibitor cocktails. For immunoprecipitation (IP), 3 μg of each of the rabbit Ets-2 antibody, rabbit C/EBP-β antibody and non-immune rabbit IgG were used per mg of protein in cell extracts. Immunoprecipitated samples were resolved through 10% SDS-PAGE under denaturing conditions and protein bands were transferred onto PVDF membranes (Immobilon-P by Millipore Billerica, MA). Western blotting was performed with a 1:500 dilution of rabbit polyclonal C/EBP-β antibody in blocking solution (25 mM Tris-HCl, pH 7.3, 150 mM NaCl, 0.1% tween-20 and 5% non-fat dry milk). Following incubation with primary antibody, the blot was treated with horseradish peroxidase (HRP) conjugated anti-rabbit IgG (eBioscience, San Diego, CA) at 1:1000 dilution in blocking solution. Membranes were developed with the Phototype-HRP Western Blot Detection System from Cell Signaling Technology, Inc., Beverly, MA, and recorded on Kodak BioMax Light Film (Kodak, Rochester, NY) according to the manufacturers' instructions.

Pull-down assays
Pull-down assays from JEG-3 cell extracts were performed with biotinylated ds oligonucleotides representing C/EBP and AP-1-binding elements (Additional data file 1 - Table  5). Each pair of sense and anti-sense oligonucleotides was prepared and modified (3 / biotinylation of the sense strand) by MWG-Biotech (High Point, NC). Biotinylated ds oligonucleotide were incubated with 50 μl of immobilized streptavidin (Pierce Biotechnology, Rockford, IL) at 4°C for 1-2 h, followed by addition of 0.5 to 1.0 mg of JEG-3 cell extracts (prepared with 25 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.5% Triton X-100, 0.5 mM dithiothreitol and 1× each of the peptidase and phosphatase inhibitor cocktails (Sigma, Saint Louis, MO) at 4°C for overnight with gentle mixing. The immobilized DNA-protein complex was washed (3×) with cold 25 mM Tris-HCl, pH 7.5, 150 mM NaCl and boiled with 5× SDS-loading dye under denaturing condition and resolved through 10% SDS-PAGE gels. Proteins were transferred from the gel to the PVDF membrane as described above. Western blotting was performed with 1:750 dilution of each of the rabbit polyclonal Ets-2, C/EBP-β, pan-Jun and pan-Fos antibodies and developed by the Western-Star Systems from Applied Biosystems (Bedford, MA) according to the manufacturer's instruction.