Molecular Cloning of Pancreatic Group I Phospholipase A2 Receptor*

We have recently reported that mammalian pancre- atic group I phospholipase A, (PLA,-I) has its specific receptor (PLA, receptor) on a variety of mammalian cells and that various biological responses are elicited by PLA,-I via this receptor. In this study, we cloned cDNAs encoding a protein corresponding to the bovine PLA, receptor purified from the corpora lutea on the basis of its partial amino acid sequences. The identity of a protein encoded by the cloned cDNA with the bovine PIA, receptor was verified by a transient expression experiment using COS-7 cells. Interestingly, the deduced primary structure of the PLA, receptor (1,463 amino acid residues) exhibits a close relatedness throughout the molecule to that of the macrophage mannose recep- tor, a unique member of Ca2+-dependent (C-type) animal lectin family, in spite of their functional diversity. Based on this sequence similarity between these two receptors, the domain organization of the PLA, receptor could be tentatively assigned as follows; 10 extracellular domains including 8 tandem repeats homologous to C- type carbohydrate-recognition domains (CRDs) and a single transmembrane region followed by a short cyto- plasmic tail. The results of transient expression experiments for mutant PLA, receptors supported this assign- ment and furthermore suggested the region responsible

the group I1 PLA2 has been described in connection with inflam-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
to the GenBankTMIEMBL Data Bank with accession number(s) D16109.
The nucleotide sequence(s) reported in this paper has been submitted Shionogi Research Laboratories, Shionogi and Co., Ltd., 12-4, Sagisu mation. However, the findings that PLA2-I is present in various kinds of nondigestive organs have raised the possibility that this protein might have some unknown physiological functions besides nutrient digestion (2, 3) and thus led us to search for them. As a result, we discovered the existence of specific binding sites for a mature form of PLA2-I on a wide range of cell types and identified various cellular responses via this PLA2-I binding site, i.e. PLA2 receptor (4); PLA2-I is found to induce DNA synthesis (4, 51, contraction (61, chemokinetic cell migration (7), and eicosanoid production (8) as receptor-mediated reactions. The studies regarding PLA2-I as a ligand for its specific receptor have been actively broadening our perspective on the physiological function of PLA2-I.
For understanding molecular mechanisms leading to these cellular responses mediated by PLA2 receptor, its structural information is indispensable. As a first step toward this end, we already reported the purification of the PLA2 receptor from the bovine corpus lutea, which clarified that the bovine PLA2 receptor is composed of a single glycosylated polypeptide chain with an apparent molecular mass of 190 kDa (9).
Here, we report the molecular cloning of a cDNA encoding the bovine PLA2 receptor, which allows us to deduce the complete primary structure of the receptor. We also describe some intriguing structural characteristics of the PLA2 receptor and the properties of the recombinant wild-type PLA2 receptor and its mutants.

EXPERIMENTAL PROCEDURES
Determination of Partial Amino Acid Sequences of the Bovine PLA, Receptor-Bovine PLA, receptor was purified to homogeneity from corpora lutea as described previously (9). The purified receptor protein was dissolved in 10 m M HCI containing 20 m M octyl thioglucoside and subjected to NH,-terminal analysis with a protein Sequencer (Applied Biosystems 477A). For determination of internal amino acid sequences of the receptor, the purified receptor was digested with lysylendopeptidase (Wako Chemicals), and the resultant digested fragments were separated by high performance liquid chromatography on a C8 column. Each isolated fragment was analyzed with the protein Sequencer, Preparation of a cDNA Fragment Encoding the NH2-terminal Region of the Bovine PLA, Receptor-On the basis of the determined NH2terminal sequence of the receptor (ETAAWAVTPERLREWQDKFI, where the residues are indicated in single-letter codes for amino acids and X refers to an unidentified residue), we designed four oligonucleotide primers (primers 1-41 for amplification of a cDNA fragment encoding the NH,-terminal region by the polymerase chain reaction (PCR) (10). Primers 1 and 2 (primer 1, 5'-GAGACCGCNGCNTGGGC-3'; primer 2,5'-ACCGCNGCNTGGGCNGT-3'; N = A/GlC/T) are sense ones corresponding to the NH,-terminal residues 1-6 and 2-7, respectively.
Primers 3 and 4 (primer 3,5'-ATGAAGATRYRYl'TRTCYTG-3'; primer 4,5'-TTGATGAAGATRYRY"RTC-3'; R = N G , Y = TIC) are antisense ones corresponding to residues 16-22 and 17-22, respectively. Total cellular RNA was prepared from bovine corpora lutea by the method of Chomczynski and Sacchi (11) hybridized with the 62-bp probe is shown. The nucleotide sequence corresponding to an NH,-terminal translated region is given in bold letters with FIG. 1. Nucleotide sequence around the exon included in the genomic clone pGE-1. The nucleotide sequence around the region the deduced amino acid residues (the NHz-terminal residue of the native receptor is numbered +l). The most 5'-end residue in the isolated cDNA probe for the subsequent' cDNA library screening is indicated by the line shown above the nucleotide sequence. The splicing donor site (GT) is clones is numbered +1, and the residues located upstream to this 5'-end residue are indicated by lowercase letters. The region (128 bp) used as a underlined.
3 using the resultant cDNA mixture under the following thermal cycling conditions: 15 cycles of 1 min at 94 "C/1 min at 37 "C followed by an additional 15 cycles of 1 min at 94 "C/1 min at 50 "C/1 min at 72 "C. The PCR products of the expected size (65 base pairs (bp)) were purified by polyacrylamide gel electrophoresis and subjected to the second-round PCR with primers 2 and 3 under the following conditions: 30 cycles of 1 min at 94 "C/1 min at 50 "C/1 min at 72 "C. The amplified products of the expected size (62 bp) were purified and subcloned by the dideoxythymidine-tailed vector method (12). Some clones carrying the 62-bp insert were randomly picked up and subjected to DNA sequencing using a sequencing kit (Sequenase Version 2.0, U. S. Biochemicals). One-third of the sequenced clones were found to code for the NH,-terminal sequence of the PLA, receptor.

Screening and Isolation of Clones Encoding the Bovine PLAz
Receptor-Using the 62-bp cDNA fragment encoding the NH,-terminal region of the PLA, receptor labeled with 32P by PCR (13), we first screened A phage bovine genomic library (Clontech). A A phage DNA of one positive clone (pGE-1) was digested with HindIII, and the insert genomic DNA fragment was subcloned for sequencing. The predicted exon region in this genomic clone (residues 257-384 in Fig. 1) was amplified by PCR and used as a probe for screening of bovine cDNA libraries. Some positive clones carrying partial PLA, receptor cDNA were obtained from bovine placenta cDNA library (oligo(dT1 + randomprimed, Clontech). Since the 3"portion of the PLA, cDNA was not found in all the isolated clones, some cDNA libraries were rescreened with a cDNA probe generated from the most downstream region of the isolated clones. We eventually found some clones carrying the 3'-portion of the cDNA in a homemade plasmid cDNA library (cDNAs > 2.5 kilobase pairs) constructed from MDBK cells (bovine kidney-derived cell line) with a plasmid cloning kit (Life Technologies, Inc.). Two clones (pPL-1 and pMD-10, obtained from the bovine placenta and the homemade MDBK plasmid cDNA libraries, respectively) were completely sequenced with a Sequenase Version 2.0 sequencing kit (U. S. Biochemicals) or by the chemical method of Maxam and Gilbert (14).
RNA and DNA Blotting Analyses-%tal cellular RNAs were routinely prepared by the method of Chomczynski and Sacchi (11) and further purified with a PolyATractTM kit (Promega). Bovine brain poly(A)+ RNAs were purchased from Clontech. RNAs to be examined were run on a 1% agarose gel containing 2.2 M formaldehyde and then transferred onto a nylon membrane by standard methods (15). A calf thymus genomic DNA blot was prepared by conventional procedures (15). After cross-linking of RNA or DNA onto the membrane by ultraviolet irradiation, hybridization was performed at 65 "C overnight as described by Church and Gilbert (16). Probes used for the blotting analyses were labeled with 32P by a random priming method (F'rime-ItTM, Stratagene). After the hybridization, membranes were routinely washed with 30 m~ sodium citrate,3 m~ NaCI,O.l% SDS a t 65 "C twice, and the hybridization signals were detected by autoradiography with Kodak 0-Mat x-ray film using two intensifylng screens a t -70 "C.
Expressions of the PLAz Receptor and Its Mutants-A cDNA clone which carries the full-length coding region of the PLA, receptor was reconstructed from pPL-1 and pMD-11 by connecting the 5'-portion of pPL-1 and the 3"portion of pMD-11 at the NcoI site (at the nucleotide residue 799, Fig. 2 4 ) . The resulting full-length cDNA insert was placed under the control of a strong eukaryotic promoter, SRa (171, in a transient expression vector carrying SV40 origin. This PLA, receptor ex-pression plasmid (2 pg) was introduced into COS-7 cells (2 x lo5 cells in a 9.6-em2 plastic dish) with lipofectin reagent (Life Technologies, Inc.). Three days after the transfection, PLA,-I binding activity was determined with lz5I-1abeled porcine PLA,-I as a ligand as described previously (4). A dissociation constant (K,) was calculated from a Scatchard plot of the binding data. The expression plasmids for the following three mutant PLA, receptors were constructed: COOH-terminal deletion mutants devoid of amino acid residues 1373-1443 and 1402-1443 were designated as the mutants-MI and -I, respectively; an NH,-tenninal deletion mutant lacking amino acid residues 29-276 was designated as the mutant-CT.
For generation of COOH-terminal deletion mutants, PCR was used for introducing a stop codon and a new restriction recognition site. The PCR was carried out with a wild-type cDNA upstream primer (5'-GGCTCTAATCTMTAACAATC-3') and a mutagenic downstream primer (for the mutant-MI, 5'-ACTGGCGGCCGCTCA?TLTCCTGGAT-GC'IT-3'; for the mutant-I, 5'4ATGGCGGCCGCCTAGTACATG-CAGAAGGA-3') using the wild-type cDNA as a template to obtain a mutated cDNA fragment (Gly-1373 or Lys-1402 was converted into a stop codon for the mutants-MI or -I, respectively; the 3'-end of the amplified DNA was flanked by Not1 site for convenience of the following construction). The amplified DNA was digested withXbaI andNotI and then replaced with the XbaVNotI region in the wild-type PLA, receptor expression plasmid. The structures of these mutant PLA, receptor expression plasmids were confirmed by DNA sequencing for avoiding artificial errors during cloning and/or PCR.
The expression plasmid for the NH,-terminal deletion mutant, the mutant-CT, was constructed from wild-type PLA, receptor expression plasmid by removal of the region XhoI-PuuII (corresponding to nucleotide residues 426-1167). The wild-type PLA, receptor expression plasmid was digested with XhoI and PuuII, treated with the Klenow fragment of Escherichia coli DNA polymerase I, and then tailed with a Sal1 linker (5'-CGGTCGACCG-3') (18) for making the in-frame connection.
Transient expression experiments were performed as described for the wild-type expression plasmid. In the case of analyses of mutant PLA, receptors, the culture media as well as the transfected cells were subjected to PLA2-I binding assays, and, furthermore, the internalization of PLA,-I via the receptor was monitored as described by Hanasaki and Arita (4). In brief, the binding sites on transfected COS-7 cells were first saturated with lZ5I-labeled porcine PLA,-I a t 4 "C for 2 h. After washing of the cells with phosphate-buffered saline twice, the internalization was induced by raising the incubation temperature to 37 "C. Forty min after the temperature shift, radioactivities internalized into the cells were measured, and a percentage of internalized PLA,-I to the initially bound one was calculated.

Isolation a n d Characterization of cDNAs Encoding Bovine
PLA, Receptor-In order to make library screening straightforward, we first prepared a cDNA fragment (62 bp) encoding the NH2-terminal part of the receptor by PCR. Using this short cDNA probe, we screened a bovine genomic library (5 X lo5 clones) and isolated one positive clone (pGE-1). The sequence analysis of pGE-1 revealed that the 62-bp cDNA sequence was interrupted by an intron (Fig. I), and a neighboring down-Primary Structure of Phospholipase A2 Receptor 5899 stream exon was not included in this clone. Hence, we next screened bovine cDNA libraries (about lo6 clones per each library) with a new cDNA probe generated from the predicted exon region (128 bp, indicated by the line above the sequence in Fig. 1) or the most downstream region of the isolated partial cDNA clones as described under "Experimental Procedures." We eventually isolated some cDNA clones (i.e. pPL-1 and pMD-10 as described under "Experimental Procedures") which collectively provided the complete structural information of the PIA2 receptor cDNA (Fig. 2).
It is interesting to note that pMD-10 contained a 114-bp insert at the nucleotide residue number of 2114 ( Fig. 2 A ) , whereas other isolated clones did not carry such an insert. This short insert, probably generated through alternative splicing, does not interrupt the open reading frame and thus results in a n insertion of 38 amino acid residues at the middle of the receptor protein (Fig. 2 B ) . APCR-assisted analysis showed that the cDNA bearing this insert was minor in the MDBK cDNA library (data not shown).
Another alternative form of the PIAz receptor cDNA was found in the MDBK cDNA library; a cDNA clone (pMD-11) carried a 3"noncoding region shorter than that of pMD-10 by 1 kb (Fig. 2 A ) . Poly(A) tails of pMD-10 and -11 are preceded by possible polyadenylation signals (A'ITAAA and AATGAA, respectively), and thus these clones were probably derived from differentially processed mRNAs using alternative polyadenylation sites.
The predicted primary structure of the bovine PLAz receptor is composed of 1,463 amino acid residues and includes all partial amino acid sequences obtained from lysylendopeptidasedigested fragments of the purified receptor (Fig. 2 A ) . The calculated molecular mass of the mature receptor is 166.7 kDa, which is consistent with the apparent molecular mass of the deglycosylated receptor protein (about 150 kDa) (9). There are several potential N-linked glycosylation sites (Asn-X-(Ser or Thr)-X, Pro is inhibited at the position of X (19)), which also agrees well with the highly glycosylated nature of the native receptor (9).
Structural Relatedness of the PLAz Receptor with the Mannose Receptor-We subjected the nucleotide sequence encoding the PIAz receptor to a computer-assisted homology search against DNA data bank (GenBank release No. 73.0) with FASTA program (20). The highest homology scores were given for macrophage mannose receptors, unique members of the Ca2+-dependent (C-type) animal lectin family (21,22). The sequence similarity was not restricted in a specific region but observed throughout the molecule, and the overall identity between the amino acid sequences of the bovine PLA, receptor and human mannose receptor is 29%. Although the sequence identity may not be remarkably high, the structural relatedness between these two receptors is confirmed by the conserved distribution of Cys and Trp; 51 out of a total 55 Cys and 47 out of total of 62 Trp in PLAz receptor are found at corresponding positions in mannose receptor. Thus, these two receptor sequences can be aligned co-linearly according to the conserved positions of Cys and Trp. This structural similarity between the P L A Z and mannose receptors thus identified allowed us to tentatively assign the domain organization of the PLAz receptor by analogy with that of the mannose receptor (21). The mannose receptor is a unique member of C-type animal lectins; it has eight tandem carbohydrate-recognition domains (CRDs), whereas all the other C-type lectins identified so far carry a single CRD per polypeptide (21). The possible domain organization of the PLAz receptor is schematically shown in Fig. 3 together with a hydropathy plot of this receptor. The hydropathy plot pointed out the presence of two major hydrophobic regions in the PLA2 receptor, which is consistent with the ten-tative assignment of the domain organization of the PLAZ receptor; these two regions correspond to a secretory signal sequence and a membrane-spanning region as shown in Fig. 3. In summary, the PLA, receptor can be described to include 13 domains as follows: a signal sequence (20 amino acid residues); a Cys-rich head domain (146 amino acid residues); a fibronectin type I1 repeat-like domain (54 amino acid residues); eight tandem CRD-like domains (about 145 amino acid residues per domain); a membrane-spanning domain (29 amino acid residues); a cytoplasmic domain (42 amino acid residues). Among these domain designations, the terms "Cys-rich" and "fibronectin type I1 repeat-like" are only descriptions of their structural characteristics, and their functional implications are not known even in the case of the mannose receptor (23).
CRD was originally defined as a minimal functional unit responsible for carbohydrate binding in C-type lectins and known to contain 14 invariant residues (24). Although all of these invariant residues are not conserved in the CRD-like domains of the PLAz receptor, some of them are located at the conserved positions. Most importantly, 4 Cys, which are known to form two disulfide bonds in the CRD, are completely conserved in CRD-like domains of the PLAz receptor (Fig. 41, implying the tertiary structural similarity between the CRD and CRD-like domain. In contrast, some invariant residues, such as those directly involved in Ca2+ binding in the CRD (251, are completely replaced in the CRD-like domain (Fig. 41, suggesting functional diversity between these domains. It is interesting to note that a canonical internalization signal sequence (NPXY (26); Fig. 2 A ) is present in the cytoplasmic domain as expected from an endocytotic behavior of the PIAz receptor (4). This is also coincident with the case of the mannose receptor (21).
Genomic DNA and RNA Blotting Analyses for the PLAz Receptor-Genomic DNA blotting analysis showed a single band in each lane of digested genomic DNAs with different restriction enzymes, suggesting that this receptor gene is a single copy (Fig. 5 A ) . No related genes were detected by this genomic DNA blotting analysis under the hybridization conditions employed here. Although we should use the full-length cDNA probe to search for genes encoding proteins related to the PLAz receptor, the use of a long cDNA probe gave multiple bands on the blot because of the presence of many long introns in the PLAz receptor gene and consequently made it difficult to discriminate the signals of the related gene(s) from those of the PLAz receptor gene since we did not know the complete genomic structure of this gene.
RNA blotting analysis revealed that multiple mRNA species different in size were detected on the blot of the poly(A)+ RNAs isolated from MDBK cells (Fig. 5B 1: the two bands (4.8 and 2.9 kb) were stronger than the remaining three bands (8.5,6.2, and 5.4 kb). Although the larger transcripts could be detected by any probes derived from the cDNA, the 2.9-kb mRNA was not visualized by the 3'-end cDNA probe on the RNA blot (Fig. 5B), suggesting that this short mRNA was devoid of a region corresponding to the 3"portion of the PLAz receptor cDNA. These observations supported the notion that the long transcripts were derived by alternative polyadenylation and suggested that the short transcript encoded a PLA, receptor-related protein lacking, at least, the CRD-like domains 7 and 8, and the membrane-spanning domain. In fact, we obtained evidence of the occurrence of alternative polyadenylation during the cDNA cloning (pMD-10 and pMD-ll), and these clones were likely copies of 6.2-and 5.4-kb mRNAs, respectively. Furthermore, it is interesting to note that the poly(A)+ RNAs isolated from the bovine brain contained a significant amount of the PLA, receptor transcripts larger than 4.8 kb but lacked the 2.9-kb transcript (Fig. 5B).   4. Alignment of the primary structures corresponding to CRD-like domains 1-8. We aligned the deduced amino acid sequences of the CRD-like domains 1-8 so as to demonstrate the conservation of the distribution of Cys and Trp in these domains. Fourteen invariant residues found in C-type CRDs are shown in the topmost row. In the lowermost row, the conserved or preferential residues in the CRD-like domains of the PLA, receptor are displayed in uppercase or lowercase letters, respectively. The numbers in parentheses indicate the amino acid residue numbers at the NH, termini of respective CRD-like domains. Four Cys residues corresponding to those making two disulfide bonds in C-type CRDs are highlighted in bold letters. Arrows indicate the positions of 5 residues directly involved in the CaZ+ binding of a mannose binding protein (25). The first 2 of these 5 residues are not involved in 14 invariant residues, but are conserved as GldGln and AsplAsn, respectively, in C-type CRD.
Expressions of the Recombinant PLA2 Receptor and Its Mutants-In order to examine the identity of the encoded protein by the cDNA with the native receptor protein, we tried to check the functional properties of the encoded protein. Transient expression experiments revealed that a substantial PLA2-I binding activity was detected in COS-7 cells transfected with the PLA, receptor expression plasmid whereas COS-7 cells transfected with the vehicle vector did not show any specific PLA,-I binding under the same assay conditions (Fig. 6). The quantitative analysis of the PIA,-I binding to COS-7 cells expressing the PLA2 receptor revealed that the Kd between the recombi-nant PLA, receptor and PLA2-I was 1.8 I", indicating that the recombinant PLA2 receptor had almost the same affinity to PIA2-I as the native receptor (Fig. 6). Therefore, we concluded that the cDNA isolated in this study actually encoded the functional PLA, receptor protein which is probably indistinguishable from the native one at least in terms of the PLA2-I binding.
We further intended to verify our tentative assignment of the domain organization of the PLAz receptor by transient expression experiments using various mutant PLA, receptor expression plasmids. The first issue to be examined was the discrimination of extracellular domains from intracellular and and the actual nucleotide sequence of the insert and its deduced amino acid sequence are shown in panel B . The amino acid sequences elucidated by the protein chemical analyses are underlined. Some characteristic amino acid sequence motifs are highlighted in the deduced primary structure: a stippled box, Asn in a possible N-linked glycosylation site; a wavy underline, a consensus endocytosis signal. As described in the text, two possible polyadenylation sites, through which two distinct PLAz receptor transcripts are generated, are indicated by open boxes. membrane-spanning ones. In order to test whether our tentative assignment of the membrane-spanning and cytoplasmic domains is correct, we constructed two types of expression plasmids directing the synthesis of deletion mutant PIA2 receptors; the mutant-MI lost both the membrane-spanning region and the cytoplasmic domain shown in Fig. 3, and the mutant-I was devoid of only the cytoplasmic domain. According to our tentative assignment, the recombinant mutant-MI was expected to be secreted into culture medium whereas the mutant-I was not. Therefore, PIA2-I binding activities in both transfected cells and the culture media were assayed in these experiments. Fig. 7 summarizes the results of PIA2-I binding assays of these mutants, indicating that the recombinant mutant-MI was secreted without lowering the binding affinity to PLA2-I. In contrast to the case of the mutant-MI, the PIA2-I binding activity of the mutant-I was detected mainly on the transfected cells with a slight decrease in the binding affinity whereas a small but significant amount of the PJ.A2-I binding activity was also found in the culture medium (Fig. 7). These results were coincident with the tentative assignment of the domain organization of the PLA2 receptor. Furthermore, these results suggested that some leakage of the mutant-I from cell surface might occur during culture because of destabilization of the membrane anchoring caused by the complete loss of the cytoplasmic domain. Thus, the decline in the PLA2-I binding affinity of the mutant-I might be ascribed to an artifact caused by leakage of the receptor-PIA2-I complex from cell surface during the binding assay.
The second issue investigated was the determination of the region responsible for the PLA2-I binding in the extracellular domains. We constructed a n expression plasmid for a deletion mutant PIA2 receptor (the mutant-CT) lacking most of the Cys-rich domain, the complete fibronectin type I1 repeat-like domain, and a part of the CRD-like domain 1, introduced this expression plasmid to COS-7 cells, and then assayed the PLA2-I binding ability of the transfected cells. As also shown in Fig. 7, the deletion of these NHz-terminal domains did not affect either the binding affinity or subcellular localization of the receptor protein. Therefore, it could be concluded that a region carrying the CRD-like domains was responsible for PLAz-I binding.
The third issue examined was the identification of a domain responsible for the receptor internalization on PLAZ-I binding. According to our tentative assignment, an endocytosis signal sequence is located in the cytoplasmic domain and thus the mutant-I was expected to be incapable of undergoing the internalization. The transient expression experiments revealed that the mutant-I receptor on COS-7 cells internalized only 0.8% of the cell surface-bound PLAz-I during 40 min at 37 "C whereas about 60% of the ligand was internalized by the wild-type receptor under the same conditions, indicating that the mutant-I carried virtually no endocytotic ability.

DISCUSSION
An interesting finding in this study is the structural relatedness of the PIAz receptor to the macrophage mannose receptor, which greatly helped us to predict the domain organization of the PLAz receptor from the primary structure since the structure and function of the mannose receptor have been well studied to date (27,28). Since all results of the transient expression experiments for the mutant receptor proteins were consistent with known properties of the PLAz receptor and the tentatively assigned domain organization based on the structure of the mannose receptor, the domain organization schematically shown in Fig. 3 seemed to be reasonable. According to this assignment, the PLAz receptor is structurally grouped into class I transmembrane protein categorized by von Heijne (29); the PLAz receptor consists of a large extracellular NHz-terminal portion (about 95% of the molecule) and a small membranespanningkytoplasmic COOH-terminal portion. The short cytoplasmic tail must be responsible for transducing an extracellular signal (i.e. the binding of PLAZ-I) to some intracellular machinery. Although an obvious function of this cytoplasmic tail was to direct endocytosis of the ligand as demonstrated in this study, we could not find any other characteristic sequence motifs in this region.
The PLAz receptor specifically binds only a mature form of mammalian PLAZ-I, which does not carry any carbohydrate chains, in a Ca2+-independent manner (30). Since these functional properties are distinctively different from those of the mannose receptor (211, the structural relatedness of these two receptors was quite surprising. Furthermore, the results of the expression experiments of the mutant PIAz receptor devoid of the NHz-terminal domains indicated that the region including the CRD-like domains was responsible for binding of the ligand, i.e. PLAz-I, implying that CRD has been functionally converted into the PLAz-I recognition domain during evolution. These observations raised the question as to how the PLAz receptor achieves highly specific ligand recognition with the structural motifs diverged from carbohydrate binding proteins. This question from an evolutional point of view is very intriguing and will be discussed in detail elsewhere. The molecular cloning of the PLAz receptor made it possible to examine its expression at the mRNA level. Using the isolated cDNA probe, we found that a single PLAz receptor gene generated multiple transcripts distinguishable on the RNA blot probably through alternative processing after transcription as in the cases of many other receptor genes such as insulin receptor gene (31). A particularly interesting issue to be mentioned was the presence of the 2.9-kb mRNA which is too short to encode the PLA, receptor isolated from the bovine corpora lutea. Since we could detect this small transcript also in the poly(A)+ RNAs isolated from the corpora lutea (data not shown), the bovine corpora lutea might contain a certain protein structurally related to the PIAz receptor although we could isolate only a single type of PLA2-I binding protein from the membrane fraction of this tissue. Interestingly, this short transcript was not observed in the brain whereas other longer transcripts of the PLAz receptor were present in the bovine brain. These observations raised the question as to the function of the PLAZ receptor in the brain and the physiological implication of the 2.9-kb transcript. Although the 2.9-kb transcript was more abundant than the 5.4-or 6.2-kb PIAz receptor transcript in MDBK cells, we unfortunately failed to isolate a cDNA corresponding to this short transcript in the MDBK plasmid cDNA library where nearly full-length PLAz receptor cDNA clones were obtained. There is the possibility that the protein encoded by this short transcript is functionally related to the PLAz receptor as in the cases of some cytokine receptor genes which generate distinctive mRNAs encoding a membrane-bound and a soluble receptor as a consequence of alternative splicing (32,33). Although we did not enter into detail on the characterization of this short transcript in this study, the elucidation of the structure of the protein encoded by this 2.9-kb mRNA would give an insight into its function and the relationship to the PLA, receptor.
The isolation of the PIAz receptor cDNA provided a powerful tool for further investigation on the physiological function of this receptor on a molecular basis. For example, we can produce a soluble PLAz-I binding protein with the same binding affinity to PLAz-I as the native PLA, receptor by deletion of a COOHterminal part of the molecule as shown in the case of the mutant-MI in this study. The soluble PLAz-I binding protein is more suitable for further analyses on the PLAz-I recognition mechanism than the native receptor due to its high solubility even in the absence of detergents and might be used as a competitive inhibitor of the PLAz-I binding to the receptor. Furthermore, any kinds of structural alterations can now be introduced into the PIAz receptor by DNA recombinant technology, which would greatly facilitate investigation of the structure-function relationship of the PLAz receptor. On the other hand, an artificial expression of the PLAz receptor followed by close examination of changes in cellular response against PLAz-I will open a way to study the signal transduction pathway through which the PLAz-I binding evokes the cellular responses on a molecular basis. These courses of studies will shed much light on the structure and function of the PLAz receptor in the future.