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 . Although expressed widely in adult and embryonic tissues, Ets-2 is required for placental development in the mouse  and supports expression of several genes, including the IFNT, whose expression provide a phenotypic "signature" for trophectoderm [15–21, 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 , but appears to be limiting for driving maximal expression of reporter genes from IFNT and hCG subunit gene promoters in human choriocarcinoma cells, such as JAr and JEG3, and in 3T3 mouse fibroblasts . Our first goal, therefore, was to determine whether Ets-2 might play a role, possibly a central one, in control of the Tkdp-1 gene expression, just as it does in controlling the activities of other signature genes of trophectoderm.
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 . 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. 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 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 several reasons to consider the possibility that a CCAAT-binding protein might regulate the Tkdp-1 gene. First, C/EBP-α, -β and -δ are expressed in human placental cells . Second, C/EBP-β transactivates multiple genes that are expressed by trophoblast cells, including the gene for the homeobox protein distal-less-3 (Dlx3) , a transcription factor that binds to the IFNT promoter and co-operatively 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 failure in placental development . 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/EBP-β, 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 [46–55], 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, 57–61]. As shown by site-directed 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 , and C/EBP-β is known to form complexes with c-Jun, 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)  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 (TK NNGNAAK ; where K = T or G and N = any nucleotide) , 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.