Identification of the binding domain for secretory phospholipases A2 on their M-type 180-kDa membrane receptor.

The rabbit muscle (M)-type receptor for secretory phospholipases A2 (sPLA2s) has a large extracellular domain of 1394 amino acids, composed of an N-terminal cysteine-rich domain, a fibronectin-like type II domain, and eight carbohydrate recognition domains (CRDs). It is thought to mediate some of the physiological effects of mammalian sPLA2s, including vascular smooth muscle contraction and cell proliferation, and is able to internalize sPLA2s. Here, we show by site-directed mutagenesis that OS1, a snake venom sPLA2, binds to the receptor via its CRDs and that deletion of CRD 5 completely abolishes the binding of sPLA2s. Moreover, a receptor lacking all CRDs but CRD 5 was still able to bind OS1 although with a lower affinity. Deletion of CRDs 4 and 6, surrounding the CRD 5, slightly reduced the affinity for OS1, thus suggesting that these CRDs are also involved in the binding of OS1. The M-type sPLA2 receptor and the macrophage mannose receptor are homologous and are predicted to share the same tertiary structure. p-Aminophenyl-α-D-mannopyranoside bovine serum albumin, a known ligand of the macrophage mannose receptor, binds to the M-type sPLA2 receptor essentially via CRDs 3-6.

The rabbit muscle (M)-type receptor for secretory phospholipases A 2 (sPLA 2 s) has a large extracellular domain of 1394 amino acids, composed of an N-terminal cysteine-rich domain, a fibronectin-like type II domain, and eight carbohydrate recognition domains (CRDs). It is thought to mediate some of the physiological effects of mammalian sPLA 2 s, including vascular smooth muscle contraction and cell proliferation, and is able to internalize sPLA 2 s. Here, we show by site-directed mutagenesis that OS 1 , a snake venom sPLA 2 , binds to the receptor via its CRDs and that deletion of CRD 5 completely abolishes the binding of sPLA 2 s. Moreover, a receptor lacking all CRDs but CRD 5 was still able to bind OS 1 although with a lower affinity. Deletion of CRDs 4 and 6, surrounding the CRD 5, slightly reduced the affinity for OS 1 , thus suggesting that these CRDs are also involved in the binding of OS 1 . The M-type sPLA 2 receptor and the macrophage mannose receptor are homologous and are predicted to share the same tertiary structure. p-Aminophenyl-␣-D-mannopyranoside bovine serum albumin, a known ligand of the macrophage mannose receptor, binds to the M-type sPLA 2 receptor essentially via CRDs 3-6.
Secretory phospholipases A 2 (sPLA 2 s) 1 are structurally homologous enzymes that have been isolated from a large number of biological sources, including mammalian tissues as well as insect and snake venoms (1)(2)(3). At least six different sPLA 2 s have been found in mammalian tissues. The pancreatic sPLA 2 and the inflammatory sPLA 2 are well characterized enzymes, while the others have only been purified (4) or cloned (5)(6)(7) recently. The pancreatic sPLA 2 has long been thought to act only as a digestive enzyme (8). However, the presence of this sPLA 2 in several non-digestive tissues has been demonstrated (9 -11), and it is now thought to play a role in airway and vascular smooth muscle contraction (12,13) as well as in cell proliferation (14). The inflammatory sPLA 2 has been purified and cloned from several sources (15,16) and is believed to play a central role in inflammatory processes (reviewed in Refs. [17][18][19][20]. It is secreted by a large number of cell types in which its expression is strongly up-regulated by inflammatory cytokines. Its concentration in various extracellular fluids is dramatically increased in several inflammatory diseases. Moreover, this enzyme has potent proinflammatory activities (21). sPLA 2 s are also found in abundance in snake and bee venoms. Besides their role in prey digestion, the snake venom sPLA 2 s can have neurotoxic, myotoxic, anticoagulant, and proinflammatory effects (22)(23)(24)(25). High affinity receptors for these enzymes have been characterized (26 -29). They are apparently involved in their biological effects. A first type of sPLA 2 receptors called N-type sPLA 2 receptors (for neuronal) has been initially identified in rat brain (26) and then in other tissues (28,29) using OS 2 , a novel neurotoxic sPLA 2 , purified from the Taipan snake (Oxyuranus scutellatus scutellatus) venom (26). It recognizes several other neurotoxic sPLA 2 s with high affinity while non-neurotoxic venom sPLA 2 s display much lower affinities. A second type of sPLA 2 receptor called M-type sPLA 2 receptor has been initially characterized in rabbit skeletal muscle (27), using OS 2 and OS 1 , another sPLA 2 purified from the Taipan snake venom (26). Very interestingly, this receptor binds the porcine pancreatic sPLA 2 as well as the human inflammatory sPLA 2 with high affinities (30), suggesting that these endogenous sPLA 2 s might be its physiological ligands. M-type sPLA 2 receptors were subsequently characterized in fibroblasts and other tissues using the porcine pancreatic sPLA 2 as a ligand (14,31).
More recently, the M-type sPLA 2 receptor has been cloned in rabbit, bovine, and human species (30,(32)(33)(34). The cloned receptors are homologous to the macrophage mannose receptor, a protein involved in the endocytosis of glycoproteins and microorganisms (35,36), as well as to DEC-205, a protein recently cloned in dendritic cells and involved in the presentation of antigens (37). Interestingly, all of these proteins, which may constitute a new family of receptors, are predicted to share the same structural organization, i.e. a large extracellular region composed of an N-terminal cysteine-rich domain, a fibronectinlike type II domain, eight (30,35) or ten (37) repeats of a carbohydrate recognition domain (CRD), followed by a unique transmembrane domain and a short intracellular C-terminal domain. This latter domain contains a consensus sequence for the internalization of ligand-receptor complexes and is thought to confer to these receptors their endocytic properties (30,31,33,37,38).
The purpose of this paper is to identify by site-directed mutagenesis techniques the specific domain in the large extracellular part of the rabbit M-type sPLA 2 receptor (1394 residues) that is responsible for the binding of sPLA 2 s. as template the cDNA encoding for the rabbit M-type sPLA 2 receptor subcloned into the expression vector pRc/CMV (Invitrogen). The template used in mutagenesis experiments was shortened in the 3Ј-noncoding region by removing 713 base pairs between 2 SexAI sites located at positions 4556 and 5269 (30). Deletions were obtained by the gapped duplex method (40) with 40-mer oligonucleotide mutagenic primers. Single-stranded DNA was produced using R408 helper phage in Escherichia coli JM101 cells (Stratagene). The oligonucleotides used are 5Ј-GGGCAGAATCCCCACTTCTCCCACTTGAGGACCCGCTCGG-3Ј for ⌬NF, 5Ј-GCATCAGAAAGGACATGGCCTGGGCAGAATCCCCAC-TTCTC-3Ј for ⌬12, 5Ј-TTGGGTCTCACATCTCTTGGTGGGCAGAAT-CCCCACTTCT-3Ј for ⌬14, 5Ј-TTGGGTCTCACATCTCTTGGAGCAT-CAGAAAGGACATGGC-3Ј for ⌬34, 5Ј-TTCTCTATGACCCAAAATGT-TTTAATTTTGCATATCCATTCAC-3Ј for ⌬5, 5Ј-TCATGTCCAGCAG-AATCCTGTGGAATTTTGCATATCCATTCAC-3Ј for ⌬56, 5Ј-ACAGT-GTGAATATCTGCCTTAATTTTGCATATCCATTCAC-3Ј for ⌬58, 5Ј-TCATGTCCAGCAGAATCCTGTCTCTTACAAAATACTGGGCA-3Ј for ⌬6, and 5Ј-ACAGTGTGAATATCTGCCTTTCTCTTACAAATACTGG-GCA-3Ј for ⌬68. The ⌬X-5 mutant was obtained by a combination of the mutants ⌬14 and ⌬68 using a unique KpnI site in the cDNA encoding for the receptor, located in the region corresponding to CRD 5. The 3Ј-nucleotide region of ⌬68 was excised using KpnI and inserted into ⌬14 after removal of its 3Ј part with KpnI. All mutations have been verified by DNA sequencing.
Expression of Wild-type and Mutant M-type sPLA 2 Receptors in Eukaryotic Cells and Binding Assays-Wild-type and mutant receptor cDNAs were introduced into COS cells. This cell line was chosen because it lacks endogenous M-type sPLA 2 receptor. Half-confluent cells were transfected using the DEAE-dextran procedure with 50 g of DNA per 175-cm 2 Petri dish. 3 days later, the cells were scraped, homogenized by sonication, and assayed for 125 I-OS 1 binding as described previously except that appropriate isotopic dilution of the ligand was made to perform saturation curves (27).
Western Blot Experiments-Transfected COS cells were rinsed three times with PBS, scraped in 20 mM Tris/HCl, pH 7.5, 2 mM EDTA and sonicated. Cell membranes were denatured at 95°C for 5 min and loaded onto a 7.5% SDS-polyacrylamide gel. After separation, the proteins were electrotransferred to Hybond-C Extra (Amersham). Nitrocellulose sheets were blocked with 5% fat-free milk in PBS, incubated with guinea-pig polyclonal antibodies (working dilution 1/5000) raised against the whole purified rabbit M-type sPLA 2 receptor (27), washed, and then incubated with peroxidase-conjugated anti-guinea pig-goat IgG (Cappel Res. Products) diluted 10,000 times. After extensive washings in PBS, 0.1% Tween 20, blots were revealed with the BM chemiluminescence Western blotting reagent (Boehringer Mannheim), and exposed to X-OMAT AR films (Eastman Kodak Co.) for 1 min.
Indirect Immunostaining of Transfected COS Cells-24 h after transfection, cells were dissociated and plated onto coverslips in 6-well plates. 3 days after transfection, the cells were fixed with 3% paraformaldehyde, permeabilized with 0.1% Triton X-100, washed, and then incubated for 1 h with the guinea pig M-type sPLA 2 receptor antiserum diluted 200 times in PBS. The cells were washed and then incubated for 30 min with anti-guinea pig fluorescein isothiocyanate-conjugated whole goat IgG (Sigma) diluted 100 times in PBS. After washing, the coverslips were mounted with Vectashield (Vector Laboratories) and observed through a confocal microscope (Leica).
Computational Sequence Analysis-Computational sequence analyses have been performed using gcg (41). Molecular weights of the mutant receptors have been computed with the peptidesort program. Multiple sequence alignments of protein sequences of CRDs of identical position have been performed using the pileup program, and their relative percentage of identity has been determined using the distances program.

Design and Expression of the Rabbit M-type sPLA 2 Receptor
Mutants-Ten mutants of the M-type sPLA 2 receptor have been designed for determination of the domain involved in the binding of sPLA 2 s to this receptor (Fig. 1). To prevent possible misfolding of expressed proteins, the boundaries of deleted portions were always placed at junction sites between predicted domains. Western blot experiments were performed to check that all of these mutants were correctly expressed in COS cells. The native receptor and all of the mutant receptors were visible as two bands of slightly different molecular weights (Fig. 2). These two bands probably result from differences in the glycosylation level of proteins when overexpressed in COS cells. Apparent molecular weights of native and mutated M-type sPLA 2 receptors are consistent with their calculated molecular weights, indicating that all mutants are efficiently translated. No protein band was detected in control mock-transfected cells.
Indirect immunostaining experiments were also performed to demonstrate that the wild-type and mutated receptors are targeted to the plasma membrane. Representative results are shown in Fig. 3. A strong labeling was observed at the cell surface for the native receptor and the various mutant receptors, indicating that all of these proteins are normally processed. The low level of staining, which is visible inside the cells, is probably due to receptors engaged in translational or posttranslational processes. Taken together, these results suggest that all these mutants are properly expressed in a way similar to that of the wild-type M-type sPLA 2 receptor. FIG. 2. Immunoblot of native and mutated rabbit M-type sPLA 2 receptors. Protein samples were separated on a 7.5% acrylamide gel, transferred electrophoretically to Hybond-C Extra membranes, and subjected to an immunolabeling using an anti-rabbit M-type sPLA 2 receptor guinea pig antiserum. The lanes are named according to the mutant receptor that has been loaded, and MOCK corresponds to membranes from mock-transfected COS cells. Because mutants in which large regions have been deleted were not labeled as well as the wildtype receptor, various amounts of membranes (3-120 g of proteins) have been loaded in the different lanes.
Rabbit M-type sPLA 2 Receptor-M-type sPLA 2 receptors have been first characterized using the snake venom sPLA 2 s OS 1 and OS 2 , purified from the Taipan snake (26), and OS 1 has become the canonical ligand for M-type sPLA 2 receptors (29). Thus, 125 I-OS 1 was used to determine the sPLA 2 binding domain in the rabbit M-type sPLA 2 receptor. Fig. 4 and Table I show the results obtained from typical 125 I-OS 1 binding equilibrium experiments on membranes from wild-type and mutant receptors. The maximum number of binding sites (B max ) is routinely comprised between 0.5 and 5 pmol/mg of protein, indicating that all mutants are expressed at similar levels as compared with the wild-type M-type sPLA 2 receptor ( Fig. 4 and Table I). Analysis of the affinities (K d values) shows that 125 I-OS 1 binds to the ⌬NF mutant with almost the same affinity as compared with the wild-type receptor. These data clearly indicate that the N-terminal cysteine-rich and the fibronectin-like type II domains are not involved in 125 I-OS 1 binding. In agreement with this conclusion, a similar deletion of these two domains in the bovine M-type sPLA 2 receptor did not modify the binding properties of porcine pancreatic sPLA 2 (32).
Deletion of CRDs 1-4 (⌬12, ⌬34, and ⌬14 mutants) only results in 2-6-fold reduced affinities for 125 I-OS 1 (Table I). This suggests that the sequence comprising CRDs 1-4 is involved but is not essential for 125 I-OS 1 binding to the receptor.
Deletion of CRDs 6 -8 (⌬6 and ⌬68 mutants) also leads to lower affinities of 125 I-OS 1 for the receptor (Table I). These deletions decreased the affinity for 125 I-OS 1 by 10-and 12-fold, respectively. These changes indicate that the CRDs 6 -8 sequence is also not crucially involved in 125 I-OS 1 binding, although CRD 6 is likely to play a role in 125 I-OS 1 binding.
A more dramatic effect was observed in different mutants lacking CRD 5 (⌬58, ⌬56, and ⌬5, see Table I), demonstrating that this CRD is directly involved in 125 I-OS 1 binding. All these mutants have completely lost their ability to bind 125 I-OS 1 . It was then interesting to test whether CRD 5 expressed alone was able to bind 125 I-OS 1 . The ⌬X-5 construct, in which all CRDs but CRD 5 have been deleted, was still able to bind 125 I-OS 1 , although with a 40-fold reduced affinity (Table I). Taken together, these data show that the main domain of the extracellular region of the rabbit M-type sPLA 2 receptor involved in 125 I-OS 1 binding is CRD 5, while CRDs 3, 4, and 6 provide further interactions to increase the affinity of 125 I-OS 1 .
CRDs Involved in Mannose-BSA Binding-The interaction of the M-type sPLA 2 receptor with sPLA 2 s is clearly a proteinprotein interaction (30). However, besides its sPLA 2 binding property, this receptor is also able to recognize sugars such as mannose and galactose (30). Indeed, based on the similarity between the rabbit M-type sPLA 2 receptor and the macrophage mannose receptor, we previously observed by competition experiments against sPLA 2 binding that the M-type receptor also binds mannose-BSA (30). Mannose-BSA is a typical ligand of the mannose receptor (42) and is known to bind to this receptor mainly via CRDs 4 and 5 (43). It was thus interesting to determine which CRDs on the M-type sPLA 2 receptor were involved in mannose-BSA binding.
The  ing by mannose-BSA to ⌬NF and ⌬12 mutants are very close to K 0.5 values obtained with the native receptor (Table II). This demonstrates that the N-terminal cysteine-rich domain, the fibronectin-like type II domain, and CRDs 1 and 2 are not involved in the binding of mannose-BSA to the M-type sPLA 2 receptor. Conversely, mannose-BSA was not able to inhibit 125 I-OS 1 binding on the ⌬14, ⌬34, and ⌬X-5 mutant forms of the M-type receptor. These results are in accordance with those establishing the importance of CRD 4 for the binding of mannose-BSA to the macrophage mannose receptor (42). Since 125 I-OS 1 had no measurable affinity on mutated receptors lacking CRD 5, it was not possible to determine the importance of this latter CRD in the binding of mannose-BSA to the M-type sPLA 2 receptor.
The affinity of mannose-BSA for the ⌬6 and ⌬68 mutants of the M-type sPLA 2 receptor was decreased 13-and 19-fold, respectively, as compared to its affinity toward the wild-type receptor (Table II). This suggests that CRDs 6 -8 are not crucially involved in the binding of mannose-BSA to the M-type sPLA 2 receptor but that these latter CRDs contribute to the binding of mannose-BSA. The involvement of these CRDs in the binding of a glycoprotein such as invertase and in the binding of mannan has already been described in the case of the macrophage mannose receptor (43,44). The CRD 4 structure of the macrophage mannose receptor is sufficient by itself to bind invertase and mannan, but it does it with a weak affinity and needs the presence of the CRDs 5-8 to bind with an affinity identical to that of the native receptor (43,44).
Thus, as for the macrophage mannose receptor, the CRD 4 structure of the M-type sPLA 2 receptor is involved in the binding of glycosylated ligands, but the CRD 5-8 structure is required to confer a full affinity. Although the same CRDs appear to be involved in the binding of mannose-BSA, several lines of evidence suggest that the amino acid residues within these CRDs, which are implicated in the binding process, are different in the two receptors. First, these two receptors, although predicted to share a similar overall structural organization, only display a low identity (28%) at the amino acid level (30). Second, the binding of mannose-BSA to the mannose receptor is strictly Ca 2ϩ dependent (42,45), while the binding of sPLA 2 s to the M-type receptor is Ca 2ϩ independent (46). We also observed that mannose-BSA is able to inhibit sPLA 2 binding in the absence of free calcium. 2 In fact, residues predicted to be involved in calcium and sugar binding in the CRDs of the mannose receptor are not conserved in CRDs 4 and 5 of the rabbit M-type sPLA 2 receptor (30). Third, although both receptors bind mannose-BSA, they display different specificities for several other glycoconjugates. For example, galactose-BSA, which is not a good ligand for the mannose receptor (42), binds to the M-type sPLA 2 receptor (30) while invertase and mannan, which avidly bind to the mannose receptor (42)(43)(44), do not compete with 125 I-OS 1 binding to the M-type receptor (30). Taken together, these observations make it unlikely that both types of receptors bind mannose-BSA through similar residues within their respective CRDs 4 -8.
Comparison of the Rabbit M-type sPLA 2 Receptor and Other Related Proteins-The data reported here indicate that the N-terminal cysteine-rich region and the fibronectin-like type II domain are not involved in sPLA 2 binding. While the function of the N-terminal cysteine-rich region still remains unknown, we have recently observed that the fibronectin-like type II domain binds to type I and IV collagens (47), suggesting that this domain in the M-type sPLA 2 receptor might play a role in cell adhesion. From the successive deletions of CRDs, it can be concluded that of the eight CRDs of the M-type sPLA 2 receptor, the key domain involved in sPLA 2 binding is CRD 5, although other CRDs (mainly CRDs 4 and 6) appear necessary to reach a high affinity. On the other hand, CRDs 1 and 2 are not required for binding.
The high conservation of CRD 4 between different animal species (Table III) has led to suggest that this CRD may have a central role in sPLA 2 binding (34). However, the data shown here indicate that CRD 5 rather than CRD 4 is most directly implicated in sPLA 2 binding.
Sequence comparisons show that among the CRDs of sPLA 2 receptors, CRD 5 is in fact the least conserved between species (Table III). The identification of CRD 5 as the most important CRD for sPLA 2 binding might explain the differences observed in the binding properties of different sPLA 2 s to M-type receptors from different species (30,33,34).
Recently, various sPLA 2 inhibitors have been discovered that are able to bind sPLA 2 s (48, 49) with binding properties (affinities and Ca 2ϩ dependence) similar to those observed for Mtype sPLA 2 receptors (46). They have been purified from the blood plasma of different snakes (48, 49). Moreover, the pulmonary surfactant protein A has been shown to recognize the sPLA 2 purified from the Trimeresurus flavoviridis venom and to inhibit its activity (50). Very interestingly, the primary structure of all of these proteins has sequence homology with several CRDs of the C-type lectin family and hence with CRDs  The percentage of identity between CRDs in identical positions in M-type sPLA 2 receptors from different species are given for CRDs 3 to 6, respectively (30,(32)(33)(34). Homologies between CRD 5 from M-type sPLA 2 receptors in different species are indicated in boldface. The sequence of CRD 6 of the rat sPLA 2 receptor is unknown. of M-type sPLA 2 receptors. The different CRDs 5 of M-type sPLA 2 receptors in different species have a homology of at least 69% (Table III), while the similarity between these CRDs 5 and CRDs of sPLA 2 inhibitors is less than 20%. Sequence comparison between these sPLA 2 inhibitors and CRDs of the M-type sPLA 2 receptors together with site-directed mutagenesis will probably help to better define residues that, within CRD 5, are crucial for sPLA 2 binding.