Identification of Domains Conferring Ligand Binding Specificity to the Prostanoid Receptor

To identify domains conferring ligand binding specificity to prostanoid receptors, we constructed a series of chimeric receptors by successively replacing the regions from the carboxyl-terminal tail of mouse prostacyclin (prostaglandin I (PGI)) receptor (mIP) with the corresponding regions of the mouse PGD receptor (mDP). The mIP receptor expressed in COS 7 cells bound [3H]iloprost, a PGI2 analog, and [3H]PGE1 with K d values of 13 and 27 nm, respectively. This receptor did not bind [3H]PGD2, [3H]PGE2, and [3H]PGF2α. The mDP receptor bound only [3H]PGD2 with a K d value of 43 nm. The chimeric IPN-VII/DPCreceptor with replacement of the carboxyl tail of the mIP receptor with that of the mDP receptor showed 12–16-fold higher affinities for [3H]iloprost and [3H]PGE1 than the mIP receptor. The region extending from the sixth transmembrane domain to the carboxyl terminus of the mIP receptor was next replaced with the corresponding region of the mDP receptor. This chimeric IPN-V/DPVI-C receptor acquired the ability to bind [3H]PGD2 and [3H]PGE2 without decreasing the affinities of the mIP receptor to [3H]iloprost and [3H]PGE1. These binding characteristics did not change when the fourth and fifth transmembrane domains of the mIP receptor were further replaced with the corresponding regions of the mDP receptor. However, when the first extracellular to second intracellular loop of the mIP receptor containing the third transmembrane domain was further replaced with those of the mDP receptor, the affinities for [3H]PGE1, [3H]PGE2, and [3H]iloprost were markedly decreased, whereas that for [3H]PGD2was increased by about 2-fold. [3H]PGF2αshowed no affinity for the mIP, mDP, and all the chimeric receptors. These results suggest that the sixth to seventh transmembrane domain of the mIP receptor confers the specificity of this receptor to bind selectively to PGE1 and not to PGE2 and that the third transmembrane domain of the mDP receptor confers the selective binding of PGD2 to this receptor.

Prostaglandins (PGs) 1 contain prostanoic acid as a central structural element. PGs have two structural features in the prostanoic acid framework. First, they have functional groups on the cyclopentane ring, which classifies them into four types, D, E, F, and I. Second, they are classified into three series, 1, 2, and 3, by the number of double bonds in the side chains. Additionally, another cyclooxygenase product, thromboxane A 2 has an oxane ring instead of the cyclopentane ring. These prostanoids act on eight types and subtypes of the receptors. They are the PGD receptor (DP), the EP 1 , EP 2 , EP 3 , and EP 4 subtypes of the PGE receptor, the PGF receptor (FP), the PGI receptor (IP), and the thromboxane A 2 receptor (TP) (1)(2)(3)(4). These receptors can recognize the structural differences of prostanoid molecules. The binding affinities of these receptors to prostanoid molecules are determined primarily by the cyclopentane ring structures of ligands. For example, the DP receptor shows the highest affinities to PGD 2 and PGD 1 , but affinities to other prostanoids are at least 2 orders of magnitude less. One exception is the IP receptor, which shows the affinity to PGE 1 almost comparable to PGI analogs such as iloprost. This receptor, however, can bind PGE 2 with much lower affinity, suggesting that the IP receptor can discriminate a difference in the side chains.
We have cloned cDNAs for all of these types and subtypes of the mouse prostanoid receptors (5)(6)(7)(8)(9)(10)(11)(12)(13). These studies revealed that the prostanoid receptors belong to the G protein-coupled rhodopsin type receptor superfamily. They have several regions conserved specifically among them. These conserved regions may participate in the construction of binding domains for structures common to prostanoid molecules, whereas the other regions may confer specificity for ligand binding. For example, the arginine in the seventh transmembrane domain, which is conserved in all of the prostanoid receptors, was proposed to be the binding site for the carboxyl group of prostanoid molecules (5,14,15). In fact, Funk et al. (16) have shown that a point mutation at this arginine residue in the human TP receptor results in loss of ligand binding activity. However, structural domains of the prostanoid receptors conferring specificity for ligand binding are as yet unknown.
Chimeric receptors have been used to determine the regions involved in various functions of the receptors. For example, this approach was used to determine the regions involved in selective agonist and antagonist binding in adrenergic receptors (17,18). Chimeric receptors were also used to identify the binding site of non-peptide antagonists to the neurokinin receptors (19 -22) and to the angiotensin receptors (23) and the G protein activation sites of the muscarinic and ␤ 1 -adrenergic receptors (24). These results show that this approach has been useful in locating functional domains of various receptors.
To identify the domains conferring the ligand binding specificity to the prostanoid receptors, we have constructed chimeric receptors from the mIP and mDP receptors in this study. This strategy is based on the high homology of their amino acid sequences as well as common signal transduction. The prostanoid receptors can be functionally grouped into three categories: the relaxant receptors, the contractile receptors, and the FIG. 1. A membrane topology model of the mIP receptor. The model is based on hydrophobicity analysis of the mIP receptor according to the methods of Kyte and Doolittle (38). Solid circles indicate the residues that are identical to those of the mDP receptor. Sites for replacement in chimeric receptors are shown, and restriction endonucleases used for construction are indicated (see "Experimental Procedures").

FIG. 2. Diagrams of the mIP, mDP, and chimeric receptors (panel A) and strategy for construction of chimeric receptors (panel B).
Panel A, the part of receptors derived from the mIP receptor is shown by an open box, and that from the mDP receptor is shown by a closed box. Panel B, PCR products corresponding to the mIP sequence are shown by bold lines above each box of chimeric receptor cDNA and those to the mDP sequence by bold lines below each box. Sequences of primers used in PCR are shown in Table  I. Numbers in parentheses indicate nucleotide numbers of the 5Ј-and 3Ј-termini of each fragment corrected for the residue numbers in the mIP receptor cDNA. Restriction sites used for construction are indicated.
inhibitory receptor (14). The relaxant receptors, consisting of the IP, DP, EP 2 , and EP 4 receptors, mediate increases in cAMP and induce smooth muscle relaxation. The contractile receptors, consisting of the TP, FP, and EP 1 receptors, mediate calcium mobilization and induce smooth muscle contraction. The EP 3 receptor is an inhibitory receptor that mediates decreases in cAMP and inhibits several biological processes such as neurotransmission, gastric acid secretion, and water reabsorption. Sequence homology among these functionally related receptors is higher than that among the three separate groups (25). The amino acid sequences of the mIP and mDP receptors, which belong to the same relaxant receptor group, show 58% identity in the transmembrane domains ( Fig. 1), and both couple to the same G protein, Gs. Chimeric mIP/DP receptors were expressed in the COS 7 cells, and their ligand binding properties were examined.
The Chimeric IP N-V /DP VI-C Receptor-BspHI sites at equivalent positions in the mIP and mDP receptor cDNAs (Fig. 2B) were utilized to construct this chimeric receptor. Fragment D-2 was excised from pCMX-mDP by digesting with BspHI and BamHI, and fragment I-2 was excised from pCMX-mDP by digesting with SphI and BspHI. Both excised fragments were ligated into the SphI and BamHI sites of pCMX-mIP.
The Chimeric IP N-VII /DP C Receptor-Fragments I-1 and D-1 were amplified by PCR with the primer pairs shown in Table I to have a BamHI site (Fig. 2B). Fragment I-1 was digested with SphI and BamHI, and fragment D-1 was digested with BamHI and BspEI. Both digested fragments were ligated into the SphI and BspEI sites of pCMX-IP N-V /DP VI-C .
The Chimeric IP N-IV /DP V-C Receptor-Fragments I-3 and D-3 were amplified by PCR with primer pairs shown in Table I to have an SpeI site (Fig. 2B). Fragment I-3 was digested with SphI and SpeI, and fragment D-3 was digested with SpeI and BspEI. Both digested fragments were ligated into the SphI and BspEI sites of pCMX-IP N-V /DP VI-C .
The Chimeric IP N-III /DP IV-C Receptor-Fragments I-4 and D-4 were amplified by PCR with the primer pairs shown in Table I. In the D-4 fragment, the PvuII site was introduced (Fig. 2B). Fragment I-4 was digested with PstI and PvuII, and fragment D-4 was digested with PvuII and PstI. Both digested fragments were ligated into the PstI sites of pCMX-IP N-V /DP VI-C .
The Chimeric IP N-II /DP III-C Receptor-Fragment D-5 was amplified by PCR with the primer pairs shown in Table I. In the D-5 fragment, the PstI site was introduced (Fig. 2B). Fragment D-5 was digested with PstI and ligated into the PstI sites of pCMX-IP N-V /DP VI-C .
The Chimeric IP N-I /DP II-C Receptor-Fragments I-6 and D-6 were amplified by PCR with the primer pairs shown in Table I to have HaeII (Fig. 2B). Fragment I-6 was digested with Asp718 and HaeII, and fragment D-6 was digested with HaeII and BspEI. Both digested fragments were ligated into the Asp718 and BspEI sites of pCMX-IP N-V /DP VI-C .
Ligand Binding Studies-For transient expression of each prostanoid receptor, COS 7 cells cultured in 15-cm dishes were transfected with 20 g of plasmid DNA by the lipofection method (27). After culture for 60 h, the cells were harvested, and crude membranes were prepared as described (11). Briefly, harvested COS 7 cells were homogenized using a Potter-Elvehjem homogenizer in a solution containing 25 mM Tris-HCl (pH 7.5), 250 mM sucrose, 10 mM MgCl 2 , 1 mM EDTA, and 0.1 mM phenylmethylsulfonyl fluoride. The homogenate was centrifuged at 800 ϫ g for 1 min. The supernatant was collected and centrifuged at 100,000 ϫ g for 1 h. The pellet was suspended in 20 mM MES (pH 6.0) containing 10 mM MgCl 2 and 1 mM EDTA (the suspension buffer), and used as crude membranes. Binding assays were performed essentially as described previously (11). For Scatchard analysis, 50 g of crude membranes was incubated in the suspension buffer with various concentrations of [ 2 in the presence of various concentrations of PGD 1 or PGD 2 . The incubation was terminated by the addition of 2 ml of the ice-cold suspension buffer, and the mixture was rapidly filtered through GF/C filters (Whatman). The filter was then washed with 5 ml of the ice-cold suspension buffer three times. The radioactivity on the filter was measured in 5 ml of Clear-Sol scintillation mixture (Nakalai Tesque, Kyoto, Japan). Nonspecific binding was determined in the presence of a 1,000-fold excess of unlabeled ligands in the incubation mixture. K i values were calculated from IC 50 values of radioligand binding as described previously (20).

RESULTS
The mIP, mDP, and six chimeric receptors were expressed in COS 7 cells, and crude membranes were prepared for binding studies. Crude membranes were incubated with various concentrations of each of [ (Fig. 3). Saturation kinetics of these binding was obtained and subjected to Scatchard analysis. Representative analyses are shown in Fig. 4, and the results of several analyses are summarized in Table II (Table II). Binding was observed also with [ 3 H]PGD 2 and [ 3 H]PGE 2 , but their affinities were too low to be analyzed by the Scatchard analysis. This ligand binding specificity of the mIP receptor is consistent with previous reports on the cloned mIP receptor (11) and on native IP receptor in various cells (28,29). On the other hand, the mDP receptor showed a high affinity binding only to [ 3 H]PGD 2 with a K d value of 43 Ϯ 6 nM (Fig. 4, H and h, and Table II). This is also consistent with previous reports on the cloned mouse and human DP receptors (12,30) and on native human DP receptor (31). We then examined the binding properties of the chimeric receptors. The carboxyl tail of the mIP receptor was first replaced with that of the mDP receptor. This chimeric IP N-VII /DP C receptor showed a 12-16-fold increase in binding affinity to [ 3 (Fig. 4, B and b, and Table II (Fig. 4, C and c, and Table II). Similar ligand binding properties were shown by the chimeric IP N-IV /DP V-C and IP N-III /DP IV-C receptors, which have further substitution of the fifth and fourth transmembrane domains (Fig. 4, D and (Table II). Similar ligand binding specificity was exhibited by the chimeric IP N-I /DP II-C receptor (Fig. 4, G and g, and Table II).

H]iloprost and [ 3 H]PGE 1 without an appreciable increase in the binding of [ 3 H]PGE 2 and [ 3 H]PGD 2
The above results indicate that the mIP receptor may accommodate the cyclopentane ring structure of PGD and that it exerts its ligand binding specificity mainly by discriminating the structural difference in the ␣-side chain. To examine this hypothesis, PGD 1 binding was analyzed by competition binding studies on the mIP, mDP, and chimeric IP N-V /DP VI-C receptors using [ 3 H]PGE 1 or [ 3 H]PGD 2 as a radioligand (Fig. 5) is negligible. These binding properties indicate that the mIP receptor exerts its ligand binding specificity in two ways. One is the recognition of the configuration of the side chains. PGI has a unique configuration of the ␣-side chain because of the presence of an additional ring attached to the cyclopentane ring. It is believed that PGE 1 without a double bond in the ␣-side chain can mimic this configuration of the PGI molecule and bind to the IP receptor, but this is not achieved by PGE 2 . The other is the recognition of the cyclopentane ring structure. It appears that this receptor can accommodate the cyclopentane rings of I and E types of PG, but not D and F. However, the specificity of this receptor for the cyclopentane ring structure appears less strict than those of other prostanoid receptors including the mDP receptor. As shown, the mDP receptor can bind only [ 3 H]PGD 2 with high affinity, and the binding affinities for E, F, and I types of PG are much lower. This indicates that the mDP receptor has strict recognition of the cyclopentane ring structure. Therefore, questions addressed in this study were which region(s) of the IP receptor discriminate PGE 1 and PGE 2 , which region(s) of the IP receptor and how strictly they accommodate the cyclopentane ring, and which region(s) of the DP receptor determines the specific recognition of the cyclopentane ring of D type.
We examined these questions by successively replacing the regions of the mIP receptor with those of the mDP receptor from the carboxyl terminus. Replacement of the region extending from the sixth transmembrane to the carboxyl terminus of the mIP receptor resulted in loss of ligand binding specificity of the mIP receptor mentioned above. This chimeric IP N-V /DP VI-C receptor bound [ 3 H]PGD 2 and [ 3 H]PGE 2 as well as [ 3 H]iloprost and [ 3 H]PGE 1 . The fact that this chimera binds PGE 2 , whereas neither mIP nor mDP binds this prostanoid, suggests that the domains recognizing the ring structure and the side chain configuration of prostanoid molecules are located in different regions of the prostanoid receptors. Because such a change was not observed in the chimeric IP N-VII /DP C receptor, these results suggest that the sixth to seventh transmembrane domain is responsible for recognition of the side chain configuration; this region of mDP appears to accommodate both 1 and 2 series of the prostanoid molecules, whereas that of mIP appears more strict, discriminating a structural difference in the ␣-side chain between PGE 1 and PGE 2 . A more detailed analysis is required to locate an exact domain conferring this selectivity. The above results also suggest that the binding pocket of the mIP receptor for the cyclopentane ring of prostanoid molecules is localized in another region and can accommodate the cyclopentane rings of not only I and E but also D type, although we cannot exclude the possibility that the sixth to seventh transmembrane domain of the mDP receptor has contributed to accommodate the cyclopentane ring of D type in this chimeric receptor. Interest-ingly, the affinities for [ 3 H]PGE 1 and [ 3 H]iloprost were not changed by further replacement of the fourth and fifth transmembrane domains, suggesting that the binding domain of the cyclopentane ring in the mIP receptor localizes in a region containing the first to third transmembrane domain. We have examined if the binding specificity of the mIP receptor is determined solely by recognition of the side chain structure by analyzing the binding of PGD 1 to the mIP. As shown in Fig. 5, no appreciable binding of PGD 1 was observed in the mIP receptor, suggesting that if the mIP receptor can accommodate the cyclopentane ring of D type, the relative configuration between the cyclopentane ring and the side chains is also important in determining the ligand binding affinity. On the other hand, PGD 1 bound to the chimeric IP N-V /DP VI-C receptor, suggesting that the sixth to seventh transmembrane domain of the mIP receptor is also responsible for determining this binding specificity. Moreover, the facts that the affinity of PGD 1 for the chimeric IP N-V /DP VI-C receptor was 1 order of magnitude lower than that of PGD 2 and that this rank of binding is identical to that observed in the mDP receptor may indicate that the sixth to seventh transmembrane domain of the mDP receptor is responsible for determining these affinities.
A region determining the specificity of the mDP receptor was suggested by further replacement of the first extracellular to second intracellular loop of the mIP receptor with the corresponding region of the mDP receptor. This replacement resulted in loss of the binding of iloprost and E type of PGs but increased the binding affinity for PGD 2 . These observations indicate first that this region of the mIP receptor is indispensable for iloprost, PGE 1 , and PGE 2 binding, and second and more importantly that this region of mDP receptor may be responsible for the ligand binding selectivity of this receptor. Then, which domain in this region is responsible for this selectivity? The ligand binding pocket in most rhodopsin-type receptors for small molecules is formed by transmembrane domains. If this is also the case for the prostanoid receptors, we can assume that the third transmembrane domain is responsible for the above selectivity. Surprisingly, the third transmembrane domain has only four amino acids different between the mIP and mDP receptors (Fig. 1). If this domain is responsible, it would be intriguing to examine if any of these four amino acids has an important influence on the recognition of the cyclopentane ring. The functional groups at the 9-and 11positions of the cyclopentane ring of PG molecules are either oxo or hydroxy groups. One hypothesis is that these groups are involved in formation of hydrogen bonds to some amino acids of the prostanoid receptors. The vicinal hydroxyl groups of the catechol ring are shown to be involved in formation of hydrogen bonds to Ser 204 and Ser 207 of the ␤ 2 -adrenergic receptor (32)(33)(34). These issues may be tested by construction of more detailed chimeric receptors in this region.
This investigation has also revealed that the affinities for iloprost and PGE 1 of the mIP receptor are increased 12-and 16-fold by replacement of its carboxyl tail with that of the mDP receptor. There have been several reports concerning the effects of carboxyl tails of the rhodopsin-type receptors. We observed that alternative splicing of the EP3 and TP receptor in the carboxyl tail affects the specificity and efficacy of G-protein coupling (35,36) as well as the sensitivity to agonist-induced desensitization (37). However, none of them showed a change in ligand binding properties. Whether a difference in the carboxyl tail increases the ligand binding affinity remains to be tested because our IP N-VII /DP C receptor also contains the replacement of several residues of the seventh transmembrane domain (Fig. 1).  In summary, the present study has identified the domains of the mIP and mDP receptors which confer ligand binding specificities to each receptor. Continued application of molecular biology including introduction of a point mutation to the identified regions will provide a more detailed understanding of the molecular basis of ligand recognition by the prostanoid receptors and will help design more specific therapeutic agents.