Diverse genetic regulatory motifs required for murine adenosine deaminase gene expression in the placenta.

Murine adenosine deaminase (ADA) is a ubiquitous purine catabolic enzyme whose expression is subject to developmental and tissue-specific regulation. ADA is enriched in trophoblast cells of the chorioallantoic placenta and is essential for embryonic and fetal development. To begin to understand the genetic pathway controlling Ada gene expression in the placenta, we have identified and characterized a 770-base pair fragment located 5.4 kilobase pairs upstream of the Ada transcription initiation site, which directs reporter gene expression to the placenta of transgenic mice. The expression pattern of the reporter gene reflected that of the endogenous Ada gene in the placenta. Sequence analysis revealed potential binding sites for bHLH and GATA transcription factors. DNase I footprinting defined three protein binding regions, one of which was placenta-specific. Mutations in the potential protein binding sites and footprinting regions resulted in loss of placental expression in transgenic mice. These findings indicate that multiple protein binding motifs are necessary for Ada expression in the placenta.

Murine adenosine deaminase (ADA) is a ubiquitous purine catabolic enzyme whose expression is subject to developmental and tissue-specific regulation. ADA is enriched in trophoblast cells of the chorioallantoic placenta and is essential for embryonic and fetal development. To begin to understand the genetic pathway controlling Ada gene expression in the placenta, we have identified and characterized a 770-base pair fragment located 5.4 kilobase pairs upstream of the Ada transcription initiation site, which directs reporter gene expression to the placenta of transgenic mice. The expression pattern of the reporter gene reflected that of the endogenous Ada gene in the placenta. Sequence analysis revealed potential binding sites for bHLH and GATA transcription factors. DNase I footprinting defined three protein binding regions, one of which was placenta-specific. Mutations in the potential protein binding sites and footprinting regions resulted in loss of placental expression in transgenic mice. These findings indicate that multiple protein binding motifs are necessary for Ada expression in the placenta.
Adenosine deaminase (ADA) 1 is a purine catabolic enzyme that converts adenosine and deoxyadenosine to inosine and deoxyinosine, respectively (1). The enzyme is present throughout the evolutionary phyla, and the amino acid sequence is highly conserved from bacteria to humans (2,3), suggesting it is a key enzyme in purine metabolism. In mice, the levels of ADA differ as much as 10,000-fold among various tissues with the highest levels occurring in the deciduum, chorioallantoic placenta, tongue, esophagus, forestomach, and proximal small intestine; intermediate levels are found in the thymus and low levels in other tissues (4). In addition, Ada expression is subject to developmental regulation at the maternal-fetal interface prenatally and in the gastrointestinal tract postnatally (4 -8).
Although the physiological role that ADA plays in these tissues is not clear, it is thought that ADA prevents accumulation of the cytotoxic metabolite deoxyadenosine, which interferes with deoxynucleotide metabolism and/or disrupts cellular transmethylation reactions (9 -14). These disturbances result in apoptosis or disrupt cell differentiation.
Mounting evidence indicates that maintaining high levels of ADA in the placenta is essential for embryonic and fetal development. During placentation in mice, ADA appears at 7.5 days post coitum (dpc) in trophoblast giant cells surrounding the embryo and in diploid trophoblast precursor cells in the ectoplacental cone (6 -8). The expression of Ada increases as the placenta develops. At 13.5 dpc when the placenta is mature, ADA appears in all trophoblast lineages, including trophoblast giant cells surrounding the gestation site, spongiotrophoblast cells in the junctional zone, and syncytial trophoblast cells in the labyrinth zone. By this time, greater than 95% of the ADA enzymatic activity found at the gestation site resides in the trophoblasts of the placenta (7,22). Expression in the placenta persists until term, suggesting that ADA is functionally important in placental physiology. This importance is revealed by a series of mouse genetic experiments. ADA-deficient mice die perinatally of severe liver damage, brought on by profound disturbances in purine metabolism (15,16). The concentrations of adenosine and deoxyadenosine increase 3-and 1000-fold, respectively, in ADA-deficient fetuses, while levels of inosine decrease. These metabolic disturbances are not evident until 12.5 dpc, coinciding with the lack of Ada expression in the placenta (15). By genetically restoring Ada expression specifically to the placenta, most of the purine metabolic disturbances in the fetus are prevented (17). Deoxyadenosine levels were lowered over 30-fold in ADA-deficient fetuses and severe fetal liver damage was prevented. As a result, ADA-deficient mice were rescued from the perinatal lethality (17,18). Therefore, activating Ada gene expression in the placenta is critical for proper fetal development.
Studying Ada gene expression in the placenta will not only enhance our appreciation of placental ADA, but also contribute to our knowledge of trophoblast differentiation. ADA is an early marker of trophoblast cell differentiation. Unlike other trophoblast-specific genes, whose expression is restricted to one cell type (19,20,21), ADA appears in all trophoblast lineages. Thus, the Ada gene may contain regulatory elements operable at all stages of trophoblast differentiation and in all three cell types, representing a good system to study trophoblast gene regulation. Starting with 6.4-kb 5Ј-flanking sequences, which direct chloramphenicol acetyltransferase (CAT) gene expression to the placenta prenatally (and forestomach postnatally) in * This work was supported in part by National Institutes of Health Grant DK46207 and HD34130. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The  1 The abbreviations used are: ADA, adenosine deaminase; dpc, days post coitum; CAT, chloramphenicol acetyltransferase; kb, kilobase pair(s); bp, base pair(s); TSE, trophoblast-specific element; TSEB, trophoblast-specific element-binding protein; bHLH, basic helix-loop-helix; hPL, human placental lactogen. transgenic mice (22), we have localized the placenta-specific regulatory element to a 770-bp fragment 5.4 kb upstream of the Ada transcription initiation site using transgenic mice as an assay system. The speed and the efficiency of this analysis were increased by examining expression of transgenes in placentas 14 days following zygote microinjection. In situ hybridization showed CAT reporter gene expression in the same cell types that expressed ADA. Sequence analysis and DNase I footprinting experiments revealed both general and placenta-specific protein binding sites within this region. Deletions and sitespecific mutations of potential protein binding sites within this 770-bp fragment resulted in loss of placenta-specific expression, suggesting that multiple factors function together to ensure proper Ada expression in the placenta.

EXPERIMENTAL PROCEDURES
Transgene Construction-The 6.4CAT transgene is described by Winston et al. (22) as the construct ADACAT. This construct was subcloned into the BamHI site of Bluescript KS ϩ II vector (Stratagene). Deletions were prepared by appropriate restriction digest of this parental plasmid. In short, 5.5CAT, 2.7CAT, and PCAT were generated by ClaI, NcoI, and XbaI digestion of 6.4CAT. 3.1PCAT, 2.0PCAT, FP3⌬PCAT, 1.8PCAT, 1.3PCAT, and 0.77PCAT were generated by ligation of PCAT to various restriction fragment from the 6.4-kb Ada flanking sequence. To generate mutant constructs, a 770-bp XbaI-PstI fragment was subcloned in Bluescript KS ϩ II vector. The 5Ј deletion was generated according to the instructions of the Erase-a-Base system (Promega). The FP1 deletion and GATA mutations were generated using the Muta-Gene phagemid in vitro mutagenesis system, version 2 (Bio-Rad). The mutagenic oligonucleotides, FP1⌬ (GTTGAAGA-GGAAACAGCGGTACCACGGGTCATCATGAGTTTTG) and GATAm (GGAGGAACATAAACTTAATCATCTTTGTCATC), were synthesized by the Institute of Molecular Genetics Nucleic Acids Core Laboratory, Baylor College of Medicine. The mutations were confirmed by nucleotide sequencing prior to ligation to PCAT as 5Ј⌬240PCAT, FP1⌬PCAT, and GATAmPCAT. PCAT contained murine Ada sequences from the XbaI site at Ϫ750 bp to the NcoI site at ϩ90 bp ligated to CAT coding region followed by the SV40 small t intron and late poly(A) signals. Prior to microinjection, prokaryotic vector sequences were removed by restriction digestion.
Transgenic Mice-Each transgene was isolated by agarose gel electrophoresis and was purified by the QIAEX II procedure (QIAGEN Inc.). DNA was resuspended at 3 ng/l in 10 mM Tris-HCl, pH 7.4, 0.1 mM EDTA and was microinjected into the male pronucleus of fertilized FVB/N oocytes (23). Fourteen days following DNA injection, gestation sites were separated into embryos and placentas for CAT assays, and genomic DNA was prepared from the fetal yolk sacs for genotyping. When necessary, lines were established by mating with ICR mice, and DNA was obtained from tail biopsies of 4-week-old offspring.
DNA Hybridizations-Transgenics were identified and transgene copy number was determined by DNA dot blot as described (22). For Southern blots, DNA was transferred to Zeta-Probe membrane (Bio-Rad) and hybridization was done as described (24).
Tissue Extracts and CAT Assays--Tissue extracts were prepared and CAT assays were performed as described (22). Typically, the placenta and embryo extracts were incubated in a solution containing 385 mM Tris-HCl, pH 7.8, 33 mM [ 14 C]chloramphenicol, and 1 mM unlabeled acetyl-CoA at 37°C, and the reactions were stopped by adding ethyl acetate. Substrate and acetylated products were resolved by thin layer chromatography (Sigma) and visualized by autoradiography. The specific radioactivities of the substrate and the products were measured using a Betascope 603 Blot Analyzer (Betagen).
In Situ Hybridizations-Embryos and placentas were fixed in 4% paraformaldehyde, phosphate-buffered saline overnight and processed for in situ hybridization as described (25). RNA probes were labeled with [␣-35 S]UTP (1000 Ci/mmol, Amersham). The ADA probes were generated from a 310 bp fragment from the 5Ј end of the cDNA. The CAT probes were generated from the first 250 bp of the coding sequence. Samples were hybridized overnight at 60°C and were treated as described (25). Slides were dipped in Kodak NTB-2 emulsion and exposed for 5 days. Sections were viewed and photographed with a Leitz Diaplan microscope (26). The micrographs shown in Fig. 2 are double exposures with the red representing hybridization signal using dark-field optics with a red filter, and the blue representing nuclei stained with Hoechst 33258, viewed with epifluorescence optics.
DNA Sequencing-The 770-bp XbaI to PstI fragment was subcloned into Bluescript KS ϩ II vector for sequencing by the dideoxy method utilizing the Sequenase 2.0 kit (U. S. Biochemical Corp.) following manufacturer's instructions. Each strand was sequenced by a series of oligonucleotides that were synthesized by the Institute of Molecular Genetics Nucleic Acids Core Laboratory, Baylor College of Medicine. Nucleotide sequence comparisons were performed with the GAP program, and transcription factor consensus sequences were sought using the Findpatterns program both from the Genetics Computer Group program, University of Wisconsin (27).
Nuclear Extracts and DNase I Footprinting-Nuclei were prepared from mid-gestation murine ICR placentas (12.5-15.5 dpc) or livers removed from 8 -12-week-old murine ICR females as described (28), and nuclear extracts were prepared by the procedure of Dignam et al. (29). For footprinting, the 770-bp Ada placental regulatory element was subdivided into a XbaI to SacI 510-bp fragment and a SacI to PstI 258-bp fragment, and each fragment was subcloned into Bluescript KS ϩ II vector. The plasmid was linearized with an enzyme that leaves a 5Ј overhang and calf intestinal phosphatase (alkaline)-treated. The fragment was released by a second digest utilizing an enzyme that produces a 3Ј overhang and was gel-purified. Fragment was end-labeled by polynucleotide kinase (Boehringer Mannheim) utilizing [␥-32 P]ATP (3000 Ci/mmol). The radiolabeled fragment was incubated with either placenta or liver nuclear extract for 45 min at room temperature in Dignam D buffer containing 2 g of poly(dI-dC) in a total volume of 50 l. 0.8 units of DNase I were added for 2 min at room temperature. Reactions were stopped, and DNA samples were resolved on either 6% (500-bp fragment) or 8% (250-bp fragment) polyacrylamide urea sequencing gels run in 1 ϫ TBE.

A Placenta-specific Regulatory Element Resides within a 770-bp Fragment in the 5Ј Flank of the Murine Ada Gene-We
have shown previously that a placental regulatory element lies within a 6.4-kb region immediately upstream of the murine Ada gene (22). We sought to identify the placenta-specific regulatory element by deletion analysis in transgenic mice. Transgenic mice were used because this approach provides a physiological assay system in which the complete array of necessary regulatory elements could be defined in a developmental context. To increase the pace of the analysis, we chose to study founder mice instead of generating transgenic lines since the same features of CAT expression were observed in both assays (Table I). Of nine 6.4CAT transgenic lines, seven showed high levels of CAT activity in the placentas and undetectable CAT activity in the adjoining embryos. Two lines with a single copy of transgene displayed very low levels of CAT activity in both placentas and embryos. All six 6.4CAT founders showed high levels of placenta-specific CAT expression. The level of CAT expression varied considerably among the 6.4CAT transgenic lines and founders and did not correlate with copy number. Thus, the data presented in Table I indicated that transgenic founders worked as well as the lines to define the Ada placentaspecific regulatory element. We next tested various 6.4CAT deletion constructs for their ability to target CAT expression in the placentas of founder mice 14.5 dpc (Fig. 1). Three 5Ј deletion constructs from 5.5 to 0.8 kb did not show CAT activity in the placentas, suggesting that the placental regulatory element resided at the 5Ј part of the 6.4-kb fragment. The smallest construct, PCAT, which contained the Ada promoter, was used as a basal transcription unit to identify the upstream placental regulatory element. Five additional deletion constructs narrowed the placental regulatory element to a 770-bp XbaI to PstI fragment located about 5.4 kb upstream of the Ada transcription initiation site, which in combination with PCAT was able to direct high levels of CAT expression in the placentas of transgenic mice. CAT activities in the adjoining embryos were undetected in most transgenic mice. In some cases (two 2.0PCAT mice and one 1.8PCAT mouse), embryos displayed low levels of CAT activity but were more than 10-fold less than that in the adjoining placentas. Additionally, one 2.0PCAT and two 0.7PCAT founders did not show any CAT activity in either placentas or embryos, presumably due to the effects of the integration sites. These data indicated that the 770-bp region contained the Ada placenta-specific genetic regulatory element.
The Ada Placental Regulatory Element Directs Reporter Gene Expression to the Same Cell Types as the Endogenous Ada Gene with Appropriate Developmental Timing-To confirm that CAT gene expression occurred in the same cell types as endogenous ADA, the cellular localization of CAT transcripts from the 2.0PCAT transgene (Fig. 1B) were identified by in situ hybridization. A transgenic male containing approximately 18 copies of the 2.0PCAT transgene was bred to non-transgenic ICR females. Placentas and embryos were isolated at 9.5 and 13.5 dpc. Day 13.5 tissues were genotyped from DNA obtained from extraembryonic membranes while 9.5-dpc transgenics were identified by the presence of CAT transcripts. Tissue sections were hybridized to either CAT or ADA sense or antisense radiolabeled RNA probes. Both CAT and ADA antisense probes produced strong signals in trophoblast cells (Fig. 2). At 9.5 dpc (Fig. 2, compare panels A and B), ADA and CAT transcripts were most abundant in the spongiotrophoblasts of the developing junctional zone. Hybridization was also observed in diploid precursors. In higher magnification of panels A and B (Fig. 2, C  and D), both ADA and CAT mRNA was observed in the secondary giant trophoblasts. At 13.5 dpc (Fig. 2, compare panels E and F), the strongest signals for each gene occurred in the spongiotrophoblasts of the junctional zone. Both CAT and ADA mRNA were also visible in the syncytial trophoblasts and the secondary giant trophoblasts in the labyrinth zone. Neither ADA nor CAT mRNA was detected by this procedure in embryo sections (data not shown), although very low levels of ADA enzyme activity is detected in embryos (6). These results demonstrate that the placental signal within the 2.0-kb BamHI to EagI fragment directs CAT expression to the ADA expressing trophoblasts in the placenta. CAT expression observed during placental formation suggested that this placental signal activates CAT expression at the proper time in development.
The Ada Placental Regulatory Element Fails to Direct Expression to the Forestomach-Since the original 6.4CAT construct also directs expression to the forestomach in adult mice (22), we wished to determine whether this placental regulatory element can function in the forestomach of adult mice. To address this question, a male founder harboring the 2.0PCAT transgene was mated to several non-transgenic ICR females. One pregnant female was sacrificed at 14.5 dpc to obtain pla- centas. The remaining females were allowed to go to term, and the resulting litters were sacrificed at 5 weeks. CAT activity was measured in extracts prepared from various tissues. CAT activity was apparent in transgenic placentas, but was not detected in the forestomach or other tissues of adult transgenic F1 mice (Fig. 3). The 2.0-kb BamHI to EagI fragment containing the Ada placental regulatory element appeared to lack element(s) essential for forestomach expression. This result was confirmed with a second 2.0PCAT line. Thus, the Ada placental regulatory element appeared to have only the capability of directing enhanced reporter gene expression to the placentas in transgenic mice.

The Ada Placental Regulatory Element Contains Sequence Motifs Homologous to Transcription Factor Binding Sites
Found in Other Placenta-specific Genes-To identify potentially important regulatory sequences within the Ada placental regulatory element, its sequence was determined and analyzed for homology with regulatory regions of other genes expressed in trophoblasts (Fig. 4). A significant stretch of homology was found with the human placental lactogen (hPL) gene enhancer (30) (Fig. 4B). With exception of a 12-bp gap in the middle, the sequence between nucleotides 57 and 103 on the Ada placental regulatory element showed 77% identity with footprinting region 3 in the hPL enhancer. Interestingly, the gap matched well with footprinting region 4 of the hPL enhancer (66% identity). However, no significant stretch of homology was found with the 4311 spongiotrophoblast enhancer region (31).
The 770-bp Ada placental regulatory element was rich in transcription factor binding sites including some playing a role in trophoblast-specific gene expression (Fig. 4A). Two sequence motifs similar to the GATA consensus sequence (WGATAR) (32) were observed at nucleotides 376 and 390 and two consensus cAMP response elements (33) were found at positions 26 and 598, respectively (Fig. 4). One trophoblast-specific element (TSE) consensus sequence (CCNNNGGG) (34) was at 336 and five basic helix-loop-helix (bHLH) sequence motifs (CANNTG) (35) representing potential binding sites for either Mash-2 (39) or Hxt (43) were also identified. Thus, this 770-bp placental regulatory element contained sequence motifs of potential functional importance.
A Placenta-specific Protein-binding Region Was Detected within the 770-bp Ada Placental Regulatory Element-To directly probe the protein binding sites in the 770-bp Ada placental regulatory element, DNase I footprinting experiments were performed in the presence of either placenta nuclear extracts or liver nuclear extracts. The Ada placental regulatory element was divided into a 510-bp XbaI to SacI 5Ј fragment and a 258-bp SacI to PstI 3Ј fragment for footprinting. A single footprint was detected on the 5Ј fragment that covered nucleotides 241-270 (FP1, Fig. 5, A and B) in placenta but not liver nuclear extracts. This footprint contained two sequences similar to the TSE consensus sequence CCNNNGGG (34), suggesting FP1 may represent an essential, placenta-specific regulatory region. The 3Ј fragment contained two footprints: FP2 (nucleotides 641-670) and FP3 (nucleotides 731-750); each one appeared in both liver and placenta extracts (Fig. 5, C and D). Thus the Ada placental regulatory element is capable of binding both placenta-specific and potentially general factors.
Different Sequence Motifs in the Ada Placental Regulatory Element Contribute to Ada Placenta-specific Expression-Sequence analysis and DNase I footprinting experiments indicated several sequence motifs in the Ada placental regulatory element with potential importance. The first 240 bp of the Ada placental regulatory element contained sequence homologous to the human placental lactogen enhancer and (four) consensus bHLH binding sites (Fig. 4). However, we were unable to detect any protein binding in this region by the DNase I footprinting experiments (Fig. 5). To examine the importance of this region, the 5Ј 240-bp sequence of 0.7PCAT was deleted and its ability to direct placental expression was tested in transgenic mice (Fig. 6). Five founder mice were generated containing different copies of the 5Ј⌬240PCAT transgene. No CAT activity was detected in either placentas or embryos of two founders, and low placenta-specific CAT activity appeared in another two mice. In one case, low levels of CAT activity were detected in the embryo but not in the placenta. These data indicated that the 5Ј 240-bp fragment is required to confer reliable placentaspecific expression in transgenic mice.
A strong placenta-specific protein binding region (FP1) detected in the Ada placental regulatory element implied its functional importance. To determine whether FP1 was a critical part of the Ada placental regulatory element, 18 bp of the FP1 region in the Ada placental regulatory element were deleted and the mutant construct FP1⌬PCAT was tested in transgenic mice. The resulting four founders with copy numbers ranging from 3 to 70 did not display CAT activity in either placentas or embryos (Fig. 6). These data strongly indicated that the placenta-specific protein/DNA interaction in the FP1 region is essential for Ada expression in the placenta.
Two consensus GATA sites were identified 7 bp apart in the Ada placental regulatory element. GATA factors were reported to be involved in regulation of human chorionic gonadotropin ␣ subunit gene and mouse placental lactogen gene expression in the trophoblast cells (36,37). To examine whether GATA motifs are required for Ada expression in the placenta, G 3 C point mutations were introduced into both GATA sites in the Ada placental regulatory element, and the effect of these mutations on placental expression was tested in transgenic mice (Fig. 6). Of four transgenic mice, two showed placenta-specific CAT expression compared to that in the embryos but the level of expression was low to barely detectable in the placenta; one showed very low but detectable CAT activity in both placenta and embryo; and one did not show any CAT activity. These results suggested that one or both GATA motifs are required for enhanced Ada expression in the placenta.
The Ada placental regulatory element contained motifs not  (Fig. 5). B, homology between the Ada placental regulatory element (middle) and human placental lactogen enhancer footprinting region DF3 (lower bold underlined sequence), DF4 (upper bold underlined sequence).
only for placenta-specific factors, but for potential ubiquitous factors as well. One region, FP3, was tested for its contribution to Ada expression in the placenta. The FP3 deletion construct FP3⌬PCAT, which had a 34-bp deletion from the 3Ј end of the Ada placental regulatory element, was unable to direct CAT expression in five transgenic mice (Fig. 6). Therefore, the se-quence motifs in FP3, which bind proteins present in both livers and placentas, are required for placenta-specific expression of Ada.
In summary, the mutational data demonstrated that both placenta-specific motifs (FP1), and potential ubiquitous motifs (FP3), as well as GATA motifs and possible bHLH motifs, are required for Ada expression in the placenta.

DISCUSSION
As part of our effort to understand the physiological importance of placental ADA in prenatal development, we wish to identify signaling pathways that govern the temporal and cellular expression of the murine Ada gene during placental development. Our approach involved the use of a relatively convenient and rapid transgenic assay in which the expression of ADA/CAT reporter genes was determined in transgenic placentas 14 days following zygote microinjection. The data here represent the identification and detailed analysis of a 770-bp sequence located approximately 5.4 kb upstream of the transcription initiation site of the murine Ada gene, which, in combination with an 800-bp region containing the Ada promoter, is sufficient to target reporter gene expression to the placentas in transgenic mice. In the absence of the placental regulatory element, no detectable CAT activity is produced from the Ada promoter in either the placenta or embryo, and the combination of the two regions directs expression only to the placenta and not to other tissues. Because the Ada promoter can be highly activated in a variety of unrelated cell types (trophoblasts, gastrointestinal epithelium, uterine decidual cells, and thymocytes), we believe that trophoblast specificity lies with the 770-bp upstream regulatory element. Consistent with this view are unpublished data indicating that the 770-bp fragment can activate heterologous promoters in a placenta-specific manner. 2 The Ada placental regulatory element is capable of directing reporter gene expression in all trophoblast lineages and in a manner that coincides with the appearance of endogenous ADA transcripts in this cell lineage (6 -8).
These results indicate that the Ada placental regulatory element contains regulatory motifs capable of functioning in all three murine trophoblast lineages and at the correct developmental time.  DNase I footprinting experiments determined that protein(s) present in placenta nuclear extracts bind to the 770-bp fragment at three places, termed FP1, FP2, and FP3. FP1 binds proteins present in nuclear extracts from placenta but not liver, suggesting that the protein/DNA interaction is placenta-specific. Deletion of FP1 resulted in loss of placenta-specific expression, indicating that sequence motifs within this region are required for placenta-specific expression. Inspection of the FP1 sequence revealed that it contains two motifs similar to the consensus sequence for the trophoblast-specific element-binding protein, TSEB (34). This protein is present in the human choriocarcinoma cell line JEG-3 and binds the TSE that is part of the cis-regulatory sequence required for the enhanced expression of the human gonadotropin ␣ and ␤ subunit genes and the aromatase gene in that cell type (34,38). Although mouse placenta does not produce gonadotropin, it is possible that a murine equivalent of TSEB is participating in the regulation of the Ada gene in the mouse placenta. The other footprinting regions, FP2 and FP3, bind proteins present in both liver and placenta nuclear extracts, and they presumably involve widely distributed proteins that may play supporting roles in achieving tissue-specific gene expression. The importance of FP3 in the Ada placental regulatory element was revealed by a deletion mutant that removed 34 bp, including a portion of FP3, from the 3Ј end of the Ada placental regulatory element. Transgenic animals carrying this deleted transgene failed to show placenta-specific expression. This finding suggests that in addition to placenta-specific factors associated with FP1, other more generally expressed factors associated with FP3 are required for placenta-specific expression. These features of the Ada placental regulatory element are consistent with the concept that cell specificity is achieved in part by tissue-specific regulatory elements and in part by regulatory elements that bind more widely distributed proteins.
Although the footprinted regions represent good candidates for functional components of the placental regulatory element, the lack of observable footprints over other regions does not exclude the possibility that other factors of functional importance may bind in vivo. Moreover, since the entire placenta was used to prepare nuclear extracts, regulatory proteins present in small populations of cells, such as the giant cells, may not have been sufficiently concentrated to be detected in this assay. In view of these potential limitations of footprint assays, the sequence of the Ada placental regulatory element was closely examined to identify sequence motifs common to other trophoblast-specific genes. At the 5Ј end, the sequence from 57 to 103 bp shows high homology to the human placental lactogen enhancer footprinting regions DF3 and DF4. DF3 and DF4 bind nuclear proteins and mediate a synergistic trophoblast-specific enhancer activity in JEG-3 cells (30). The homology between the human placental lactogen enhancer and the Ada placental regulatory element suggests that similar factors may be involved in the placenta-specific expression of each gene. The potential importance of the human placental lactogen homology region is underscored by the fact that it lies within an essential 240-bp region at the 5Ј end of the Ada placental regulatory element. The Ada placental regulatory element also contains five binding sites for bHLH proteins (35). Two bHLH sites lie within the human placental lactogen homology region and may represent critical components of the conserved sequence. Two bHLH factors, Mash-2 and Hxt, are specifically implicated in gene regulation during trophoblast development (39 -44). Mash-2 is found in both trophoblast precursors and spongiotrophoblasts and is essential for spongiotrophoblast development (39 -41). Hxt is present in early trophoblasts and in differentiated giant cells and can induce commitment of blas-tomeres to trophoblasts (42)(43)(44). Although specific target genes of Mash-2 and Hxt have not been identified, available evidence favors an important role for these proteins in regulation of trophoblast gene expression. A potential role for bHLH factors in the regulation of Ada gene expression is consistent with the fact that four of the five bHLH sites found in the Ada placental regulatory element are present within the critical 240-bp fragment at the 5Ј end. Additional experiments are required to specifically assess the importance of bHLH proteins in the regulation of Ada gene expression in trophoblasts.
There are two centrally located GATA sites in the Ada placental regulatory element. GATA factors are zinc finger transcription factors that play a pivotal role in hematopoiesis and erythroid gene expression (45)(46)(47)(48)(49)(50)(51). Two members of this gene family, GATA-2 and GATA-3, are found in trophoblasts of the placenta (37,52). The GATA motif is part of the trophoblastspecific enhancer that regulates the human chorionic gonadotropin ␣ subunit gene expression in JEG-3 cells (36). GATA factors are also involved in trophoblast-specific expression of mouse placental lactogen I gene in rat Rcho-1 cells (37). The possible involvement of GATA factors in regulating Ada expression in the placenta was investigated by introducing point mutations into each of the GATA motifs. Lack of CAT activity in placentas harboring such mutant transgenes suggests that GATA factors are required to activate the Ada gene in the placenta. GATA factors may be essential for activation of the Ada gene expression in the ectoplacental cone and/or maintaining Ada gene expression in all branches of the trophoblast population.
The evidence that Ada placental regulatory element is composed of a series of distinct genetic regulatory motifs suggests that a combination of multiple transcription factors are required for Ada expression in the mature placenta. These potential factors include bHLH factors, GATA factors, TSEB, and other ubiquitously expressed factors. Identification of these factors and characterization of their function in Ada expression are currently under way. It is possible that different set of transcription factors are required for Ada expression at different stages during placenta development. Some of these factors may be involved in activation of Ada in the ectoplacental cone, some may participate in maintaining high levels of Ada expression in one or multiple trophoblast lineages.
In addition to placental trophoblasts, the murine Ada gene is also expressed at high levels in several diverse cell types (4 -8): the keratinized epithelium of the tongue, esophagus, and forestomach; the absorptive epithelium of the small intestine; and the decidual cells of uterine stroma (4 -8). Ada is expressed at intermediate levels in the thymus and low levels in most other tissues (2,4). The results presented here indicate that the Ada placental regulatory element is capable of directing enhanced expression only to the placenta, suggesting that it is functionally independent from other tissue-specific regulatory elements. Since the original 6.4-kb 5Ј-flanking region also directs reporter gene expression to the forestomach postnatally (22), it appears that the placenta and forestomach regulatory elements are distinct. In this regard, recent experiments have identified a distinct and completely independent regulatory region located approximately 1 kb downstream of the placental regulatory element capable of directing expression specifically to the forestomach. 3 A regulatory region in intron 1 functions to direct enhanced CAT expression to the thymus (53)(54)(55)(56). This same region also produces low levels of expression in most tissues of transgenic mice, suggesting that it contains regulatory elements that contribute to the ubiquitous expression of the Ada gene (56). Recent results with the intron 1 regulatory region indicate that the ubiquitous elements do not require the thymus-specific elements to function (56). Thus, we have identified individual regulatory elements for enhanced expression in the placenta, forestomach, and thymus, as well as a regulatory region required for ubiquitous expression. Regulatory elements not yet identified include those for the tongue, esophagus, small intestine, and deciduum. We speculate that separate regulatory elements will be required for each of these remaining tissues.