Prostanoid receptors of murine NIH 3T3 and RAW 264.7 cells. Structure and expression of the murine prostaglandin EP4 receptor gene.

Prostaglandin endoperoxide H synthase-1 (PGHS-1) is expressed constitutively in murine NIH 3T3 cells and RAW 264.7 cells. PGHS-2 is inducibly expressed in these cells following stimulation with serum or bacterial lipopolysaccharide (LPS), respectively. Reverse transcription-polymerase chain reaction (RT-PCR) analysis established that a variety of G protein-linked and peroxisomal proliferator-activated prostanoid receptors are expressed in both of these cell types. The levels of the EP2 and EP4 prostaglandin E2 (PGE2) receptors and the prostaglandin I2 receptor were changed in these cells by serum or LPS stimulation. Quantitative RT-PCR indicated that the mRNA for the murine EP4 receptor, the butaprost-insensitive PGE2 receptor that couples to Gs, increases 1.5-3-fold in response to serum (NIH 3T3) or LPS (RAW 264.7) with a time course approximating the induction of PGHS-2 expression. To study expression of the EP4 receptor we isolated the mouse EP4 receptor gene; the gene is 10 kilobase pairs (kb) in length and, like other known prostanoid receptor genes, contains three exons and two introns. The first intron is 0.5 kb and is located 16 base pairs (bp) downstream of the translational start site. This is a different location than that of the first introns of other prostanoid receptor genes. The second intron is located immediately following the sixth transmembrane domain at the same position as the second intron of the thromboxane A2 receptor, prostaglandin D2 receptor, prostaglandin I2 receptor, and one of the PGE2 (EP1) receptor genes. A major transcriptional start was detected at −142 bp upstream of the translational start. There are a variety of putative cis-acting elements within 1.5 kb upstream of the translational start site and within the first intron. Promoter analyses of the EP4 receptor gene promoter in RAW 264.7 cells indicated that there is a constitutive negative regulatory region between −992 and −928 bp, a constitutive positive region between −928 and −554 bp, and an LPS/serum-responsive region between −554 and −116 bp.

The nomenclature regarding EP2 and EP4 receptors has been somewhat confusing and has recently undergone some changes. Narumiya and co-workers (22) isolated a cDNA for a murine PGE 2 receptor that coupled to the activation of adenylate cyclase and that they called the EP2 receptor. However, this receptor differed from the PGE 2 receptor defined pharmacologically as the EP2 receptor in that it responded to butaprost only very weakly. More recently, a cDNA for a butaprostsensitive PGE 2 receptor was cloned (16); this latter receptor is the pharmacologically defined EP2 receptor, and that originally cloned by Narumiya and others is now called the EP4 receptor (22,25,26). mRNA for the mouse EP4 subtype has been detected in thymus, lung, spleen, ileum, and mastocytoma P-815 cells by Northern blotting analysis (22). The mouse EP4 receptor subtype is located on chromosome 15 (65).
In studies designed to relate expression of the two PGHS isozymes to that of G protein-linked and PPAR␥ prostanoid receptors in NIH 3T3 and RAW 264.7 cells, we used RT-PCR to determine which receptors are present in both quiescent (serum-starved) and serum-or bacterial lipopolysaccharide (LPS)-stimulated cells. Both EP2 and EP4 receptor mRNA levels were increased by cell stimulation, whereas IP receptor mRNA levels were decreased. We describe here investigations on the structure of the EP4 receptor including the isolation and characterization of the EP4 receptor gene and the transcriptional regulation of this gene.
Cell Culture and RNA Isolation-Murine NIH 3T3 cells were cultured in DMEM containing 2% FCS and 8% CS in a water-saturated 7% CO 2 incubator. Serum stimulation of quiescent, serum-starved NIH 3T3 cells was performed as described previously (29). Murine RAW 264.7 cells were cultured in DMEM containing 10% FCS in a watersaturated 5% CO 2 atmosphere. Stimulation of the cells was performed by adding LPS (200 ng/ml, final concentration) to the culture medium. NIH 3T3 cells and RAW 264.7 cells were isolated by scraping the cells from the culture dishes with a rubber policeman and then sedimenting the cells by centrifugation at 500 ϫ g for 5 min. Total RNA was isolated using Trizol reagent according to the instructions of the manufacturer.
PCR Detection of Murine Prostanoid Receptor mRNAs-Mouse lung and kidney were minced and homogenized with Trizol reagent, and total RNA was isolated as described above. Reverse transcription with oligo(dT) priming was used to generate cDNAs from 10 g of total RNA extracts of serum-starved and serum-stimulated NIH 3T3 cells and control and LPS-stimulated RAW 264.7 cells using reverse transcriptase (SuperScript II) and the protocol of Life Technologies, Inc. The following primers were used for PCR amplification of the resulting cDNA (PCR conditions were 94°C for 120 s and then 27 cycles of 95°C  for 30 s, 60°C for (22,23) were synthesized, and cDNA fragments (412 and 423 bp) for the mouse EP4 receptor were amplified with these primers using RT-PCR. The amplified fragments were labeled with [␣-32 P]dCTP using the random primer method and were used as probes for screening the FIXII mouse genomic library. Plaques were transferred onto nitrocellulose membranes and hybridized with the probes.
Determination of the EP4 Receptor Gene Structure-Genomic DNAs corresponding to the EP4 receptor were isolated from positive phage clones following digestion with SacI, and the SacI inserts were subcloned into PUC 19 as described by Sambrook et al. (34). The gene structures were characterized by restriction enzyme digestion and sequencing of the PUC19 constructs. Sequencing was performed according to the dideoxy method with Sequenase (version 2.0) and the protocol described by the manufacturer. cis-Acting elements in the promoter and the first intron were identified by computer analysis using the Wisconsin Sequence Analysis Package with the eukaryotic transcriptional factor data base (Genetics Computer Group, Inc.).
Primer Extension Analyses-Two oligonucleotides complementary to different regions of the EP4 receptor gene (WS 309; 5Ј-CAACCTCAGC-CATCAGTCTCTTC-3Ј (Ϫ10 bp to ϩ13 bp) and WS 310; 5Ј-CTCCAAC-CTCAGCCATCAGTCTC-3Ј (Ϫ7 bp to ϩ16 bp) were synthesized, labeled with [␥-32 P]ATP, and used for primer extension analysis. Labeled primers were annealed to 20 g of total RNA from serum-starved or -stimulated NIH 3T3 cells at 70°C for 10 min and elongated with SuperScript II reverse transcriptase (10 units) at 42°C for 1 h. The reaction mixture was treated with RNase A, extracted with phenol/ chloroform, and precipitated with ethanol. After centrifugation the pellets were dissolved in 10 l of loading buffer and loaded onto a 6% sequencing gel along with a sequence ladder. The gel was dried and exposed to Kodak XAR film. The positions of the radioactive bands were determined by reference to a sequence ladder.
Expression in RAW 264.7 Cells of a Luciferase Plasmid Containing EP4 Promoter Fragments-Various fragments from the promoter region of the EP4 receptor gene were prepared and ligated into the pGL3 luciferase basic plasmid; these constructs are designated pGLep4-1 to -5. The plasmids used for transfection were purified using Qiagen Tip-100 or -500. RAW 264.7 cells were subcultured in 35-mm plastic culture dishes at a density of 50 ϫ 10 4 cells/ml 1 day prior to the transfection. A luciferase plasmid (5 g) and 1 g of pSV-␤-galactosidase plasmid were transfected into RAW 264.7 cells for 3 h with 400 g/ml DEAE-dextran and 50 mM Tris-HCl (pH 7.3) in 1 ml of DMEM. Transfected cells were washed with phosphate-buffered saline and cultured in DMEM containing 0.2% CS. After 24 h in culture, the transfected cells were stimulated with 200 ng of LPS/ml and 16% FCS for 12 h. The luciferase activities were measured with a luciferase assay system as described by the manufacturer using a Turner model TD-20e Luminometer. Protein levels used in the luciferase assay were measured with Bio-Rad protein assay reagents. The ␤-galactosidase activities were monitored with a ␤-galactosidase assay kit as described by the manufacturer monitoring the absorbance at 420 nm.

RESULTS
Prostanoid Receptors in Murine NIH 3T3 cells and RAW 264.7 Cells-Prostaglandin endoperoxide H synthase-1 (PGHS-1) is expressed constitutively in both murine NIH 3T3 cells and RAW 264.7 cells; a second isozyme, PGHS-2 is induced as an immediate early gene when quiescent, serum-starved murine NIH 3T3 cells are treated with phorbol esters or serum (29,35,36) or when RAW 264.7 cells are treated with endotoxin (i.e. bacterial LPS (37)). We used RT-PCR to determine which G protein-linked and PPAR␥ prostanoid receptors are expressed in quiescent and serum-stimulated 3T3 cells and in control and LPS-stimulated RAW 264.7 cells, because these receptors might be expected to mediate downstream events associated with constitutive and/or inducible prostanoid production in these cells. PCR primers for the various receptors were developed, which permitted us to perform the amplification steps at relatively high temperatures, and such that convenient, unique restriction endonuclease sites were present near the midpoints of the amplified fragments. As summarized in Fig. 1, both quiescent and serum-stimulated 3T3 cells express EP4, FP, and IP receptor mRNA but lack detectable EP1, EP2, EP3, or TP receptor mRNA. Quiescent and LPS-stimulated RAW 264.7 cells express EP2, EP3, EP4, and IP receptor mRNA but lack EP1, FP, and thromboxane receptor mRNA. PPAR␥ receptor mRNA was present in RAW 264.7 but not in NIH 3T3 cells (data not shown); the levels of PPAR␥ in RAW 264.7 cells were unchanged by LPS treatments. We were unable to develop appropriate PCR primers for the murine DP receptor. The results of these initial RT-PCR experiments also indicated that the levels of mRNA for three of the prostanoid receptors were changed by serum and/or LPS stimulation. In both NIH 3T3 cells and RAW 264.7 cells, EP4 receptor mRNA levels were increased by cell stimulation, whereas IP receptor mRNA levels were decreased; in RAW 264.7 cells, EP2 receptor mRNA levels increased.
The EP4 receptor is the butaprost-insensitive PGE 2 receptor, which functions through G s to activate adenylate cyclase (22,25,26,28). Small but consistent increases in EP4 mRNA appeared to occur upon serum stimulation of 3T3 cells and LPS stimulation of RAW 264.7 cells (Fig. 1). A "three-band" quantitative RT-PCR method was developed to measure EP4 receptor mRNA levels (38), and this procedure was applied to both quiescent and serum-stimulated murine NIH 3T3 cells and control and LPS-treated murine RAW 264.7 cells; the threeband method is a modification of the two-band quantitative RT-PCR method developed by Clontech Laboratories, Inc. The results of experiments performed with RAW 264.7 cells are shown in Fig. 2. Briefly, the following cDNAs were amplified: (a) a 528-bp ␤-actin cDNA generated by reverse transcription of ␤-actin mRNA present in the total RNA fraction from RAW 264.7 cells; (b) a 423-bp EP4 receptor cDNA generated by reverse transcription of EP4 receptor mRNA present in the total RNA fraction from RAW 264. double-stranded cDNA containing a MIMIC cDNA sequence with appropriate EP4 receptor primer sequences at either end of the cDNA (EP4-MIMIC cDNA). The EP4-MIMIC cDNA was added just prior to the PCR amplification step as an external control to the cDNAs generated during the reverse transciptase step. By using EP4-MIMIC cDNA as an external control, it was possible to control for any changes in ␤-actin mRNA levels while at the same time quantifying EP4 receptor mRNA using the Clontech MIMIC two-band quantitation method. EP4 receptor mRNA levels were measured in RAW 264.7 cells before and after stimulation with LPS. In RAW 264.7 cells at zero time (just preceding the addition of LPS), the EP4 receptor band was of the same intensity as that of the EP4-MIMIC band at the M 8 dilution (15.625 ϫ 10 Ϫ3 amol/0.25 g of total RNA), and the intensity of the ␤-actin band was the same as that of the EP4-MIMIC band between the M 3 and M 4 dilution (Fig. 2, left panel). After stimulation with LPS for 2.5 h (Fig. 2, right  panel), the intensity of the EP4 band was the same as the intensity of the EP4-MIMIC at the M 7 dilution (31.25 ϫ 10 Ϫ3 amol/0.25 g of total RNA), whereas the intensity of the ␤-actin band was that of the EP4-MIMIC at the M 4 dilution. After normalizing for ␤-actin expression, the expression level of EP 4 receptor mRNA in RAW 264.7 cells was estimated to increase 1.5-fold 1 h following LPS stimulation and 3-fold 2.5 h after LPS stimulation. EP4 receptor mRNA from serum-starved and serum-stimulated murine NIH 3T3 cells was also quantified using the three-band quantitation system. One h following serum (16%) stimulation, EP4 receptor mRNA levels increased 2-fold, from 15.625 to 31.25 ϫ 10 Ϫ3 amol/0.25 g of total RNA (data not shown).
EP4 Receptor Gene Structure-As a prelude to examining EP4 receptor gene expression, it was necessary to characterize the gene for this receptor. Approximately 8 ϫ 10 5 plaques from a FIXII mouse genomic library were screened using two different probes amplified from the RNA of murine NIH 3T3 cells by RT-PCR. Two positive clones, FIXEP 4 -3-1 and 3-2 were isolated. FIXEP 4 -3-1 and FIXEP 4 -3-2 contained 15-and 22-kb inserts, respectively. These inserts proved to be overlapping clones with FIX/EP 4 -3-2 containing the 5Ј-end of the EP4 receptor gene. Both of the genomic clones were digested using SacI and subcloned into PUC 19. The intron/exon structure of the EP4 receptor was determined by nucleotide sequencing and restriction mapping. The results are presented diagrammatically in Fig. 3. The EP4 receptor gene is about 10 kb in length and contains three exons and two introns. The first, 5Ј-most intron is 0.5 kb, and the second intron is 7.0 kb. The exonintron junction structures conform to the GT-AG rule ( Table I).
The first intron is located 16 bp downstream and the second intron is 951 bp downstream of the translational start site (ATG).
Transcriptional Start Sites in EP4 Receptor mRNA-Transcriptional start sites were determined by primer extension analysis (Fig. 5). Using two different primers (WS309 and WS310), we identified a transcriptional start site at Ϫ142 bp (counting from the translational start site). No differences were observed in the position of this transcriptional start site determined using RNA from either serum-stimulated or quiescent NIH 3T3 cells.
EP4 Receptor Gene Promoter Analysis-We made five different constructions containing various EP4 promoter sequences coupled to luciferase cDNA and expressed these by transient transfection of RAW 264.7 cells (Fig. 6). Under both serumstarved and serum/LPS-stimulated conditions (Fig. 6, dark and hatched bars), luciferase activity increased in going from pGLep4-1 to pGLep4-4, suggesting that between Ϫ4200 and Ϫ992 there are negative regulatory elements in the promoter. In contrast, luciferase activity was decreased when the EP4 receptor promoter region between Ϫ928 and Ϫ554 (i.e. pGLep4-4 versus -5) was removed, indicating that this part of the promoter contains positive regulatory sites. After serum/ LPS stimulation (Fig. 6, hatched bars), luciferase activities increased between 20 and 90% with all of the constructs. When the luciferase activities observed with the control pGL3 plasmid are subtracted from the luciferase activities obtained with the other constructs, the differences between nonstimulated and stimulated conditions become much more obvious, with increases ranging from 1.2-to 2.7-fold over the pGL3 control. Most notably after subtracting the values for the pGL3 control, the luciferase activity in pGLep4-5 was stimulated 2.7-fold by serum/LPS. These results suggest that there are positive regulatory elements responsive to serum/LPS stimulation between Ϫ554 and Ϫ116.

DISCUSSION
There appear to be two distinct prostaglandin biosynthetic systems that can coexist in cells (2, 47, 48). One system uses the constitutive PGHS-1 as the initiating enzyme and functions primarily in the ER to produce prostanoids that act extracellularly through G protein-linked receptors (5, 6) to mediate well known physiological "housekeeping" effects of prostanoids (2, 3). A second system, involving the inducible PGHS-2, ap-pears to produce prostanoids that act, at least in part, at the level of the cell nucleus (2, 3), perhaps through nuclear receptors such as PPAR␥ (7)(8)(9)(10)(11) in the early stages of cell replication or differentiation. In the work described in this report, we first used RT-PCR to identify those G protein-linked and PPAR␥ prostanoid receptors that are expressed in quiescent and activated murine NIH 3T3 and RAW 264.7 cells. This led us to examine the structure and regulation of expression of the gene for the EP4 receptor, the butaprost-insensitive, PGE 2 receptor which is coupled to G s . Overall, these studies are potentially important in determining which receptors are involved in mediating prostanoid responses associated with cell replication and differentiation.
Both NIH 3T3 cells and RAW 264.7 cells expressed mRNAs for several G protein-linked prostanoid receptors. PPAR␥ was expressed in RAW 264.7 cells but not NIH 3T3 cells. Assuming that at least some of the prostanoids formed via PGHS-2 act on nuclear targets in association with cell replication, the finding that the PPAR␥ was present only in RAW 264.7 cells suggests that this receptor is not uniformly involved in mediating nuclear actions of prostanoids.
The EP4 and IP receptors were the only receptors expressed by both NIH 3T3 and RAW 264.7 cells. In each of these cell types, stimulation with serum or LPS, respectively, led to small but consistent increases in EP4 receptor mRNA expression and consistent decreases in IP receptor expression. In both cell types a peak in EP4 receptor mRNA expression occurred at about the same time as the peak level of PGHS-2 mRNA (29,37,49,50). This observation raises the possibility that the increases in EP4 receptor expression are important in mediating effects of PGE 2 related to cell replication and/or differentiation. The EP4 receptor is coupled to cAMP synthesis and is probably a cell surface receptor (22).
The EP4 receptor gene was found to contain three exons and two introns. A similar overall structure has been found for TP (51), EP1 (52), DP (53), and IP (54) prostanoid receptor genes. In all of these cases, including that of the EP4 receptor, the second intron occurs at the end of the putative sixth transmembrane domain. However, the first intron of the EP4 receptor gene is located within the coding region of the receptor cDNA, whereas the first introns for other prostanoid receptor genes characterized to date are found on the 5Ј-side of the ATG translational start sites (51)(52)(53)(54). In addition, the first intron is relatively small (0.5 kb) in the case of the EP4 receptor gene. DP, EP2, EP4, and IP receptors are classified in the same phylogenetic cluster, CL-1, whereas TP and EP1 receptors are in cluster CL-2, and EP3 receptors are in cluster CL-3 (55). EP4 receptors are divided from other receptors of the CL-1 cluster at an early evolutionary stage. The overall organizations of the DP and IP receptor genes are more similar to the TP receptor gene than to the EP4 receptor gene. Thus, the EP3 receptor gene in cluster CL-3 may have the same structure as other previously characterized prostanoid receptor genes, and only the EP4 receptor gene (and perhaps the EP2 receptor gene) may have a structure different from that of other prostanoid receptor genes.
EP4 receptor mRNA is detectable and thus is apparently expressed constitutively in murine lung, spleen, and ileum (22); we find that it is also present in both quiescent and activated NIH 3T3 cells and RAW 264.7 cells. Consistent with this basal level expression, the promoter region of the EP4 receptor gene resembles that of a typical constitutive "housekeeping" gene having a GC-rich region (from Ϫ201 to Ϫ1150) highly populated with Sp1 sites and lacking a functional TATA box (56). 2 Although EP4 receptor mRNA appears to be expressed in many cells and tissues, 1.5-3-fold increases in EP4 receptor mRNA did occur in response to both serum (3T3 cells) and LPS (RAW 264.7 cells), indicating that the EP4 receptor gene is inducible. In the case of RAW 264.7 cells, which have the cell surface monocyte CD14 antigen, the LPS/serum LPS-binding protein complex binds to CD14 and elicits intracellular signals (57). When we examined the expression and induction by serum/LPS of five different constructs of the EP4 receptor gene promoter in murine RAW 264.7 cells, a maximal increase in luciferase activity was observed with the pGLep4-5 construct containing nucleotides Ϫ116 to Ϫ554 bp. There are Sp1 and AP2 elements and an E-box in this region. Although Sp1 sites are essential for CD14 expression (58), we have not resolved which specific response element(s) is essential for EP4 receptor gene expression.
The expression of PGHS-2 is stimulated by a number of agents including serum, tumor-promoting phorbol esters, cytokines, growth factors, cytotoxins, etc. (2). Several transcription factors have been identified as regulators of PGHS-2 gene expression including NF-IL6 (C/EBP␤) (59,60), c-Jun through v-Src (61, 62), C/EBP␦ (63), NF-B (60), and upstream stimulatory factor (64). These same response elements are in the EP4 receptor gene promoter. It will be important to determine if any of these elements are functional during the stimulation of expression of the EP4 receptor gene. FIG. 6. Expression of EP4 receptor promoter/luciferase plasmid constructs in transfected RAW 264.7 cells. RAW 264.7 cells were cotransfected with the indicated EP4 receptor promoter/luciferase plasmid constructs and with a plasmid encoding ␤-galactosidase and were subsequently analyzed for expression of luciferase and ␤-galactosidase activities as described under "Experimental Procedures." Each construct (pGLep4 -1 to -5) of luciferase plasmid contained a fragment of the EP4 receptor gene promoter that included nucleotides Ϫ116 to the indicated nucleotide as numbered in Fig. 4A. Luciferase activities were normalized to ␤-galactosidase activities. Black bars represent the luciferase activities in control, quiescent RAW 264.7 cells; hatched bars indicate luciferase activities in serum/LPS-treated RAW 264.7 cells. The S.E. is indicated by the error bars (n ϭ 3). The -fold increase was calculated by subtracting the control and stimulated values obtained for the the pGL3 background construct from the respective control and stimulated values obtained for the various EP4 promoter-containing constructs.