Both Group IB and Group IIA Secreted Phospholipases A2 Are Natural Ligands of the Mouse 180-kDa M-type Receptor*

Snake venom and mammalian secreted phospholipases A2 (sPLA2s) have been associated with toxic (neurotoxicity, myotoxicity, etc.), pathological (inflammation, cancer, etc.), and physiological (proliferation, contraction, secretion, etc.) processes. Specific membrane receptors (M and N types) for sPLA2s have been initially identified with snake venom sPLA2s as ligands, and the M-type 180-kDa receptor was cloned from different animal species. This paper addresses the problem of the endogenous ligands of the M-type receptor. Recombinant group IB and group IIA sPLA2s from human and mouse species have been prepared and analyzed for their binding properties to M-type receptors from different animal species. Both mouse group IB and group IIA sPLA2s are high affinity ligands (in the 1–10 nm range) for the mouse M-type receptor. These two sPLA2s are expressed in the mouse tissues where the M-type receptor is also expressed, making it likely that both types of sPLA2s are physiological ligands of the mouse M-type receptor. This conclusion does not hold for human group IB and IIA sPLA2s and the cloned human M-type receptor. The two mouse sPLA2s have relatively high affinities for the mouse M-type receptor, but they can have much lower affinities for receptors from other animal species, indicating that species specificity exists for sPLA2 binding to M-type receptors. Caution should thus be exerted in avoiding mixing sPLA2s, cells, or tissues from different animal species in studies of the biological roles of mammalian sPLA2s associated with an action through their membrane receptors.

Secreted phospholipases A 2 (PLA 2 s, 1 phosphatide 2 acylhydrolase, EC 3.1.1.4) form a growing family of Ca 2ϩ -dependent enzymes that release free fatty acids and lysophospholipids from glycerophospholipids (1)(2)(3)(4). To date, five different sPLA 2 s referred to as group IB, IIA, IIC, V, and X sPLA 2 s have been characterized in mammals. The main common properties of these sPLA 2 s are their relatively low molecular mass (13)(14)(15)(16), the presence of many disulfide bridges in their structure, and a low selectivity for phospholipids with different polar head groups and fatty acid chains (5).
Group IB sPLA 2 is known as the pancreatic-type sPLA 2 . It was originally found in large amounts in the pancreas and then proposed to function in the digestion of dietary lipids (6). Later, this enzyme was identified and cloned in other tissues such as lung, spleen, kidney, and ovary (3,7), and it has now been proposed to be involved in various physiological and pathophysiological responses such as cell proliferation (8), cell contraction (9,10), lipid mediator release (11), acute lung injury (12), and endotoxic shock (13).
Group IIA sPLA 2 is also referred to as the inflammatory-type sPLA 2 , since it is highly expressed in the plasma and synovial fluids of patients with various inflammatory diseases such as rheumatoid arthritis, acute pancreatitis, Crohn's disease, and endotoxic shock (3, 14 -16) as well as in various cancers (17,18). The group IIA sPLA 2 has been shown to participate in the production of lipid mediators of inflammation (3,19,20) and in the destruction of pathogenic microorganisms (21). Recent data using group IIA sPLA 2 -deficient mice have suggested, however, that this sPLA 2 may not play a pivotal role in the progression and/or pathogenesis of inflammatory processes, at least in the mouse (22,23). The mouse group IIA sPLA 2 (mGIIA) 2 has also been proposed to have a role in cell proliferation (24) and more recently to act as a tumor suppressor gene in a mouse model of colorectal cancer (25,26).
Much less is known about the regulation and the biological roles of the more recently cloned group IIC, V, and X sPLA 2 s. Group IIC sPLA 2 has been cloned in rat and mouse (27) but appears to be a nonfunctional pseudogene in humans (28). Group V sPLA 2 is highly expressed in heart (29) and is detected in murine macrophages and mastocytes, where it is proposed to play an important role in lipid mediator production in place of the group IIA sPLA 2 (22,23). Group X sPLA 2 was recently cloned in human and has structural features that resemble those of group IB and group IIA sPLA 2 s (30). It is expressed in the immune system, suggesting possible roles related to inflammation or immunity.
Snake and insect venoms also contain a large diversity of sPLA 2 s (31,32). Most venom sPLA 2 s are potent toxins that exert many effects including neurotoxicity and myotoxicity (31,32). Two main types of high affinity sPLA 2 receptors have been identified using venom sPLA 2 s, including OS 1 and OS 2 purified from Taipan snake venom (33). N-type sPLA 2 receptors were first identified in rat brain membranes (34). These receptors have high affinities for neurotoxic sPLA 2 s such as OS 2 and the bee venom sPLA 2 (bvPLA 2 ) but not for nontoxic sPLA 2 s such as OS 1 , suggesting that N-type receptors contribute to the neurotoxic effects (34 -36). M-type sPLA 2 receptors were first identified in skeletal muscle cells (37) but are also expressed in other tissues (33). They consist of a single 180-kDa subunit and recognize with high affinity OS 2 and OS 1 but not the neurotoxic bvPLA 2 . The M-type receptor was later also identified by using as ligand the group IB sPLA 2 and was then proposed to have an essential role in the various biological effects produced by the group IB sPLA 2 (8,38). The M-type receptor has now been cloned in different species (39 -42) and found to belong to a novel family of membrane receptors comprising the macrophage mannose receptor, the dendritic cell receptor, and the endothelial lectin receptor (43,44). The protein domains involved in the sPLA 2 -receptor binding have been elucidated (38,45,46). Knock-out mice for the M-type receptor have now been generated (13). The real endogenous ligands of the M-type receptor, however, remain elusive. Available data on the binding properties of mammalian sPLA 2 s to the M-type receptor appear contradictory, depending on the animal origin of both the sPLA 2 and the M-type receptor. For instance, rat and bovine M-type receptors can bind both porcine (pGIB) or rat (rGIB) group IB sPLA 2 s with a high affinity (K d values of ϳ1 nM) but do not associate with rat (rGIIA) or rabbit (rbGIIA) group IIA sPLA 2 s, suggesting that group IB sPLA 2 s, but not group IIA sPLA 2 s, are the physiological ligands of the M-type receptor (38). On the contrary, the rabbit M-type receptor associates with high affinity with both pGIB and human group IIA sPLA 2 (hGIIA) with K d values of 1-10 nM, suggesting that both types of sPLA 2 s may be the natural ligands of the M-type receptor (39). Finally, the cloned human M-type receptor binds with very weak affinities group IB and group IIA sPLA 2 s, suggesting that neither of these two sPLA 2 s are physiological ligands of this receptor (40).
In an effort to clarify this situation, we have prepared various native and recombinant group IB and group IIA sPLA 2 s and analyzed their binding properties to M-type receptors of different animal species. We found that mouse group IB sPLA 2 and mouse group IIA sPLA 2 are recognized by the mouse Mtype receptor, indicating that both types of sPLA 2 s are probably true endogenous ligands of the M-type receptor in mouse. We also provided evidence that this binding may occur in vivo, since sPLA 2 s and the M-type receptor are co-expressed in several mouse tissues.
Cloning and Preparation of Recombinant Mouse Group IB sPLA 2 -A search for homology in genome data bases with sPLA 2 protein sequences led to the identification of an I.M.A.G.E. Consortium cDNA clone (identification number 315430, 5Ј; GenBank TM accession no. W12659) from the house mouse with a strong similarity to known pancreatic group IB sPLA 2 s. The obtained cDNA clone was sequenced and found to contain a cDNA insert of 555 base pairs that encodes for the full-length mouse group IB sPLA 2 (mGIB) cDNA. The full-length cDNA including the prepropeptide sequence was subcloned into the baculovirus transfer vector pVL 1392 and transfected into Spodoptera frugiperda cells (Sf9; ATCC CRL 1711) using the BaculoGold TM trans-fection kit (Pharmingen). After two rounds of virus amplification into Sf9 cells, Trichoplusia ni High Five insect cells (Tn5) were used for the production of recombinant sPLA 2 s, since preliminary experiments have shown a 3-fold higher yield of sPLA 2 production compared with Sf9 cells. Furthermore, preliminary experiments indicated that infection of cells with mGIB baculovirus in growth medium containing fetal bovine serum resulted in the secretion of a mixture of proenzyme and mature forms of mGIB, since the sPLA 2 activity of cell supernatant can be increased by the addition of trypsin, while fully activated mGIB was recovered from supernatant of cells that were cultivated in protein-free Insect-Xpress medium (BioWhittaker) (not shown). Large scale sPLA 2 productions were thus performed with Tn5-infected cells (2.10 6 cells/ml) grown in spinner culture bottles in protein-free Insect-Xpress medium for 5 days. Cell-free supernatants of infected cells (1 liter) were diluted twice in 1% (v/v) acetic acid and incubated batchwise for 2 h, at 4°C, and under continuous agitation with 150 ml of SP Sephadex C-25 gel (Amersham Pharmacia Biotech), which had been preequilibrated with 1% acetic acid. The gel was washed with 1% acetic acid and 1% acetic acid containing 100 mM ammonium acetate. Bound proteins were then eluted stepwise with 1% acetic acid containing 350 mM ammonium acetate. sPLA 2 containing fractions were lyophilized and applied to a C18 Beckman reverse phase HPLC column (10 ϫ 250 mm, 19.6 ml, 5 m, 100 Å). Elution was performed using an acetonitrile linear gradient in 0.1% trifluoroacetic acid, 10 -60% acetonitrile for 40 min at a flow rate of 4.5 ml/min. The peak containing sPLA 2 activity was lyophilized and applied to the Spherogel TSK SP-5PW HPLC column as indicated above for the purification of pGIB. The sPLA 2 peak was finally applied on a C18 Nucleosil TM reverse phase HPLC column (4.6 ϫ 250 mm, 4.2 ml, 5 m, 300 Å) that was eluted using an acetonitrile linear gradient in 0.1% trifluoroacetic acid, 15-25% acetonitrile for 10 min followed by 25-45% acetonitrile for 100 min at 1 ml/min. The final yield for the production of mGIB was 1.5 mg of purified sPLA 2 /liter of cell medium.
Preparation of Recombinant Human Group IIA sPLA 2 s-AV12 fibroblast recombinant hGIIA (47) was provided by Dr. Ruth Kramer (Lilly). Recombinant hGIIA from insect Tn5 cells was prepared as above for mGIB and using the full-length hGIIA cDNA (48). Membrane-bound hGIIA sPLA 2 activity (3) was extracted from pelleted baculovirus-infected Tn5 cells for 30 min with 100 ml of phosphate-buffered saline buffer containing 1 M KCl and combined with the hGIIA sPLA 2 activity of cell-free supernatant. The pooled medium was diluted twice with water and incubated batchwise for 1 h at 4°C and under continuous agitation with 40 ml of heparin-Sepharose CL-6B gel (Amersham Pharmacia Biotech), which had been preequilibrated with 20 mM Tris, pH 7.4, containing 140 mM NaCl. The gel was washed with the equilibration buffer and 20 mM Tris pH 7.4 containing 200 mM NaCl. Bound proteins were then eluted stepwise with 20 mM Tris, pH 7.4, containing 1 M NaCl. sPLA 2 -containing fractions were dialyzed and loaded as described above for mGIB on a TSK SP-5PW column and C18 Nucleosil TM large pore column. The final yield for the production of hGIIA was about 0.3 mg of purified sPLA 2 /liter of cell medium.
Preparation of Native and Recombinant Mouse Group IIA sPLA 2 s-Native mGIIA was acid-extracted from BALB/c mouse intestine tissue and then loaded on a Bio-Gel P100 column as described previously (49). The Bio-Gel P100 sPLA 2 fractions were further purified on TSK SP-5PW column and C18 Nucleosil TM large pore column as described above. For production of mGIIA in human 293 fibroblast cells, the mGIIA cDNA (50) was cloned into the expression vector pRc/CMV (Invitrogen Corp.) and stably transfected into 293 cells. A neomycinresistant clone was selected for its high level of expression and then used for cell extraction of recombinant mGIIA as described above for intestinal mGIIA. Recombinant mGIIA from baculovirus-infected Tn5 cells was prepared essentially as described above for hGIIA. The final yield for the production of mGIIA in Tn5 cells was about 0.1 mg of purified sPLA 2 /liter of cell-infected medium.
Protein Characterization Techniques-Protein concentrations of sPLA 2 s were determined using the molar absorbance coefficients calculated from protein sequence. Ion spray mass spectrometry analysis was performed on a simple quadrupole mass spectrometer equipped with an ion spray source and using polypropylene glycol to calibrate quadrupole. N-terminal sequences were determined by automated Edman degradation of sPLA 2 with an Applied Biosystems sequencer equipped with an on-line phenylthiohydantoin-derivative analyzer. sPLA 2 activity assays were performed using Escherichia coli membranes as substrate (51).
Binding Studies-Crude microsomal membranes from cells and BALB/c adult mouse tissues were prepared as described previously (37,52). Membrane preparations containing bovine, rat, mouse, and rabbit M-type receptors were obtained from Madin-Darby bovine kidney cells (ATCC CCL 22), rat aortic smooth muscle cells (A7r5; ATCC CRL 1444), mouse embryo fibroblast cells (NIH 3T3; ATCC CRL 1658), and rabbit skeletal muscle cells (37). COS cell membranes expressing the human cloned M-type receptor were obtained as described (40). Recombinant expression of the mouse M-type receptor was performed in COS cells after cloning of its full-length cDNA from NIH 3T3 cells according to published sequence (41) and transfection into COS cells as for the human M-type receptor. All binding experiments were performed under equilibrium binding conditions using as ligand 125 I-OS 1 labeled to a specific activity of 3000 -3500 cpm/fmol as described by Lambeau et al. (37). Briefly, membranes, 125 I-OS 1 , and competitors were incubated at 20°C in 0.5 or 1 ml of buffer (140 mM NaCl, 0.1 mM CaCl 2 , 20 mM Tris, pH 7.4, and 0.1% BSA). Incubations were started by the addition of membranes and filtered after 90 min of incubation through GF/C glass fiber filters presoaked in 0.5% polyethyleneimine. Cross-linking experiments were performed with 50 M suberic acid bis-N-hydroxysuccinimide ester (Sigma) as described previously (37,52).
Northern Blot Analysis-A homemade and a commercial mouse Northern blot (CLONTECH Laboratories, Inc., catalog no. 7762-1) containing RNAs from various adult mouse BALB/c tissues were first probed with the random primed 32 P-labeled full-length mGIIA cDNA in 50% formamide, 5ϫ SSPE (0.9 M NaCl, 50 mM sodium phosphate, pH 7.4, 5 mM EDTA), 5ϫ Denhardt's solution, 0.1% SDS, 20 mM sodium phosphate, pH 6.5, and 250 g/ml denatured salmon sperm DNA at 50°C for 18 h. Blots were washed to a final stringency of 0.1ϫ SSC (30 mM NaCl, 3 mM trisodium citrate, pH 7.0) with 0.1% SDS at 55°C and exposed to Biomax MS Kodak films with an HE intensifying screen (Amersham Pharmacia Biotech). Northern blots were then stripped, checked for dehybridization, and hybridized under the same conditions as above with the entire coding sequence of mGIB. The integrity and relative quantities of RNAs were checked with the manufacturer's mouse ␤-actin probe (not shown).

Preparation of Native and Recombinant
Mammalian sPLA 2 s-Native or recombinant group IB and group IIA sPLA 2 s from different species have been prepared in order to analyze their binding properties to M-type receptors. The human pancreatic group IB sPLA 2 (hGIB) was purified to homogeneity as described previously (53) and migrates as a single band on SDSpolyacrylamide gel (Fig. 1). Highly purified pGIB was obtained after further purification of the commercially available pGIB preparation on a cation exchange HPLC column. Six fractions with sPLA 2 activity and molecular masses close to 14 kDa as determined by gel analysis (data not shown) were resolved. N-terminal sequence and electrospray mass spectrometry indicated that the major peak (molecular mass of 13,981 Da) corresponds to the authentic isoform ␣ of the porcine pancreatic sPLA 2 (theoretical molecular mass of 13,980 Da; Refs. 7 and 54). Recombinant mGIB was prepared using the baculovirus expression system (see "Experimental Procedures" for details) after cloning of its cDNA (GenBank TM accession no. AF097637) by screening public data bases with sPLA 2 sequences (30). The full-length cDNA was found to code for a protein of 146 residues containing a signal peptide sequence of 15 residues and a propeptide of seven residues ending with a single arginine, followed by a mature protein of 124 residues. The mature protein contains seven disulfide bridges located at positions that are typical of group IB sPLA 2 s and a pancreatic loop of six residues and displays 89 and 81% identity with rGIB and hGIB, respectively (7,55). Taken together, these properties indicate that the cloned cDNA codes for mGIB. Although SDSpolyacrylamide gel electrophoresis analysis indicates that the electrophoretic mobility of the recombinant mGIB protein is slightly faster than that of hGIB and pGIB (Fig. 1), the electrospray molecular mass (14,075 Da), the N-terminal sequence, and the amino acid composition of the recombinant mGIB protein (not shown) were found to be identical to those predicted from the cDNA sequence, indicating the successful recombinant expression of mGIB. hGIIA was obtained from mammalian fibroblasts (47) and from insect cells by using the baculovirus expression system. The two products were identical as checked by gel analysis (Fig. 1), mass spectrometry (13,904 Da), amino acid composition, and N-terminal sequence (not shown). Furthermore, their molecular masses are identical to the theoretical value calculated from the cDNA sequence, indicating that the recombinant sPLA 2 proteins are full-length proteins and are not modified after translation, despite the presence of a potential site of N-glycosylation in the protein sequence (30). The mGIIA purified from BALB/c mouse intestine tissue and the corresponding recombinant proteins produced in mammalian and insect cells have a similar electrophoretic mobility (Fig. 1) and also have the same molecular mass (13,958 Da), N-terminal sequence, and amino acid composition (not shown) as those predicted from the cDNA sequence, indicating the absence of any post-translational modification such as glycosylation in both native and recombinant mGIIA sPLA 2 s.
Binding Properties of Group IB and Group IIA sPLA 2 s to M-type Receptors from Different Species-125 I-OS 1 , a specific and very high affinity ligand of the M-type receptor (33), was used to analyze the binding properties of the various group IB and group IIA sPLA 2 s to M-type receptors from different species. Fig. 2 and Table I show the results obtained from competition binding experiments between labeled OS 1 and unlabeled sPLA 2 s to M-type receptors from mouse, rat, rabbit, human, and bovine species. OS 1 has a similar and very high affinity (100 -200 pM) for all receptors except for the human receptor, for which the observed K 0.5 value is only 4 nM (Table I). A common pharmacological property of the different M-type receptors is that 125 I-OS 1 binding is not inhibited by the neurotoxic bvPLA 2 , which was previously found to be a specific sPLA 2 ligand for rat and rabbit N-type sPLA 2 receptors (33,34,37).
The binding profiles of pGIB and hGIB to the various M-type receptors are very similar. Both sPLA 2 s display an affinity in the nanomolar range for mouse, rat, and bovine receptors, have a 50 -100-fold weaker affinity for the rabbit receptor, and bind very weakly to the human receptor (Table I). These results are in accordance with previous data obtained with pGIB and hGIB on mouse, rat, and bovine M-type receptors (38). The binding profile of mGIB is slightly different from that of pGIB and hGIB as it binds with high affinity to mouse and rabbit receptors but with an ϳ10-fold lower affinity to the rat receptor. However, as for pGIB and hGIB (40), mGIB is a very low affinity ligand of the human M-type receptor. These data indicate that the pancreatic-type sPLA 2 mGIB is probably a natu-  (Table I). However, this does not hold for hGIB and the human M-type receptor (Table I).
In contrast with the similar binding profiles observed with the group IB sPLA 2 s, the group IIA sPLA 2 s mGIIA and hGIIA were found to have very distinct binding properties ( Fig. 2 and Table I). Recombinant hGIIA sPLA 2 s prepared from mammalian or insect cells behave similarly. They do not bind to rat, bovine, and human M-type receptors; they associate with the mouse receptor but only with a very low affinity; and they bind with a nanomolar affinity to the rabbit receptor. Taken together, these results suggest that hGIIA is not a physiological ligand for the human receptor. However, the recombinant hGIIA proteins appear properly folded, since they can bind with high affinities to the rabbit M-type receptor. Similar to hGIIA, both native and recombinant mGIIA sPLA 2 s were found to recognize the rabbit receptor with a nanomolar affinity, while they do not bind to the human receptor. However, mGIIA was found to associate with a relatively high affinity of ϳ10 nM to the mouse M-type receptor, and it also binds to the rat and bovine receptors (Table I), indicating that mGIIA would be a second natural ligand for the M-type receptor in the mouse.
Altogether, both mGIB and mGIIA appear as ligands of the mouse M-type receptor endogenously expressed in NIH 3T3 cells (Fig. 2). To confirm this view, we investigated the binding properties of these sPLA 2 s to recombinant mouse M-type receptor expressed in transfected COS cells. The recombinant receptor was found to have the expected binding properties for venom sPLA 2 s, i.e. a high affinity for OS 1 (K 0.5 ϭ 0.3 nM) and no measurable affinity for bvPLA 2 , suggesting that this receptor is successfully expressed in COS cells (Fig. 3). Furthermore, labeled OS 1 was found unable to bind to mock-transfected cells (not shown), indicating that OS 1 binds specifically to the mouse M-type receptor. Fig. 3 shows that mGIB and mGIIA bind to the recombinant mouse M-type receptor with affinities similar to those observed in OS 1 competition assays on NIH 3T3 membranes (Fig. 2), clearly indicating that these two sPLA 2 s can bind to the cloned mouse M-type receptor.
Since the M-type receptor was found to share similarities with the macrophage mannose receptor that belongs to the C-type lectin superfamily, and since the rabbit M-type receptor was previously found to bind with nanomolar affinities various glycoconjugates of BSA (39) (i.e. to display lectin-like properties), it was of interest to analyze the binding properties of the different M-type receptors for the various glycoconjugated derivatives of BSA. We observed that only the rabbit receptor has a high affinity for the various BSA glycoconjugates, whereas the other receptors displayed either a much weaker affinity (human and mouse) or even no measurable affinity (rat), suggesting that lectin properties of the M-type receptor (such as binding of glycosylated BSA) have not been conserved between mammalian species and therefore would not be physiologically relevant. This view is in agreement with a previous observation that mannosylated BSA is unable to bind to the bovine M-type receptor (42).
Tissue Distribution of the Mouse M-type Receptor, mGIB, and mGIIA-The above binding data indicate that both mGIB and mGIIA bind to the mouse M-type receptor and therefore may be physiological ligands in the mouse. To strengthen this view, the tissue distribution of the M-type receptor was analyzed in BALB/c mice and then compared with that of mGIB and mGIIA, with the idea that a colocalization of the receptor with the expression sites of one or both sPLA 2 s would add further evidence for a physiological significance of the binding of mGIB or mGIIA to the mouse M-type receptor. The presence of the M-type receptor in various BALB/c mouse tissues was determined with the labeled ligand OS 1 , and the binding results are presented in Fig. 4 and Table II. The M-type receptor is expressed in various tissues and corresponds to a single family of binding sites with an equilibrium binding constant (K d ) close to 50 pM for labeled OS 1 (Table II and Fig. 4). Lung, colon, kidney, and salivary glands were found to contain the highest amounts of receptor, while binding sites for 125 I-OS 1 were absent in brain membranes. The maximal number of binding sites in the various tissues remains low as compared with that observed in the mouse fibroblast cell line NIH 3T3 (Table II), which was used as a source of mouse M-type receptor for the competition binding assays (Fig. 2). The tissue distribution of the M-type receptor protein shown here using 125 I-OS 1 binding is in agreement with the previous analysis of the M-type receptor transcript in mouse (41). Most notably, the highest amounts of transcripts were also found in lung and kidney, while lower levels were observed in heart and liver (41).
The tissue distribution of mGIB and mGIIA was carried out FIG. 2. Binding properties of sPLA 2 s to the mouse M-type receptor from NIH 3T3 cells. Competition experiments between 125 I-OS 1 and unlabeled venom sPLA 2 s (A), group IB sPLA 2 from pigs, humans, and mice (B), and group IIA sPLA 2 from mice and humans (C) on mouse NIH 3T3 membranes. Membranes (50 g of protein/ml) were incubated in the presence of 125 I-OS 1 (50 pM) and various concentrations of unlabeled sPLA 2 s. All results are expressed as percentages of the specific binding measured in the absence of unlabeled sPLA 2 s. 100% corresponds to a 125 I-OS 1 specific binding of 2.7 pM. The nonspecific binding was determined in the presence of 30 nM unlabeled OS 1 and was below 20% of the total binding. by successively probing BALB/c mouse tissue Northern blots at high stringency with the two different cDNAs (Fig. 5). Transcripts of 0.8 kb coding for mGIB were detected at very high levels in pancreas and at lower levels in liver, lung, and spleen. No transcript was detected in other tissues such as intestine, heart, brain, skeletal muscle, kidney, and testis. This tissue distribution is in accordance with that previously described in mice (41) and is similar to the distribution observed in humans, with the exception of liver where no transcript was detected (7,30). Also in agreement with previous data (56,57), we found that very high amounts of the mGIIA transcript are present in intestine but not in other analyzed tissues with the exception of liver, where a very weak expression is observed. Taken together, these data indicate that different mouse tissues express both the mouse M-type receptor and mGIB or mGIIA.
Both mGIB and mGIIA sPLA 2 s Bind to the 180-kDa M-type Receptor Expressed in Mouse Colon-To finally confirm that the binding properties of mGIB and mGIIA observed on mouse M-type receptor endogenously expressed in the fibroblast NIH 3T3 cell line (Fig. 2) or transiently expressed in COS cells (Fig.  3) are similar in normal mouse tissues, we performed competition binding experiments as well as cross-linking experiments with mGIB and mGIIA on mouse colon membranes (Fig. 6). The K 0.5 values determined for mGIB (K 0.5 ϭ 1.3 nM) and mGIIA (K 0.5 ϭ 10 nM) appeared very similar to those measured  on NIH 3T3 membranes and on the recombinant M-type receptor (Figs. 6A, 2, and 3, respectively). Cross-linking experiments with labeled OS 1 on mouse colon membranes resulted in the labeling of a single band of about 180 kDa, that fits well with the previously determined molecular mass of the M-type receptor in various species (33,37,38,58). Furthermore, the labeling displayed the expected pharmacological profile (i.e. it was totally prevented by unlabeled OS 1 , mGIB, and mGIIA but not by unlabeled bvPLA 2 ).

DISCUSSION
The 180-kDa M-type receptor was initially identified using snake venom sPLA 2 s such as OS 1 (37). A first clue to the physiological function of the M-type receptor was provided when it was observed that the mammalian pancreatic group IB sPLA 2 , but not the inflammatory group IIA sPLA 2 , was a ligand of this M-type receptor with a K d value of 1 nM (8). However, other binding experiments to the rabbit M-type receptor suggested that both group IB and group IIA sPLA 2 s may be natural endogenous ligands of this receptor (39). On the other hand, the human M-type receptor was found to have very weak affinities for both group IB and group IIA sPLA 2 s (40). The situation was thus clearly confusing, and the physiological relevance of these binding data was difficult to evaluate because sPLA 2 s and M-type receptors from different animal species were used in many of these binding experiments (38,39).
This paper now presents data showing that mGIB and mGIIA sPLA 2 s are high affinity ligands of the mouse M-type receptor (Fig. 2) and therefore would be natural candidates to act as endogenous ligands of this receptor. This view is strengthened by the colocalization in mice of the two sPLA 2 s and of the M-type receptor. In particular, both mGIIA and the M-type receptor are expressed in small intestine and colon (Table II, Fig. 5, and Ref. 56). Although the affinity of mGIIA for the M-type receptor is not very high (K 0.5 close to 10 nM), this binding is likely to occur in small intestine and colon, because huge amounts of mGIIA transcript and protein are found in these tissues (49,56,57). The presence of mGIB and mGIIA sPLA 2 s in mouse serum or platelets is not well documented. However, since both group IB and group IIA sPLA 2 activities were detected in serum from other animal species (14), it is likely that these sPLA 2 s are also present in mouse serum and then may reach M-type receptors that are far away from cells producing sPLA 2 s.
The view that both group IB and group IIA sPLA 2 s would operate as endogenous ligands of the M-type receptor in the mouse would not apply to humans, since both types of human sPLA 2 s were found unable to bind to the human M-type receptor (Ref. 40 and Table I). Another situation is found in the rat, since the rat M-type receptor binds rGIB but not rGIIA (38). The rabbit M-type receptor is unique in binding all the different mammalian group IB and group IIA sPLA 2 s analyzed so far (Table I). It is also the only one that has lectin-like binding properties (Table I). Altogether, the available data indicate that group IB sPLA 2 could behave as an endogenous ligand of M-type receptors at least in mouse, rat, and porcine (33) species. On the other hand, the group IIA sPLA 2 appears to serve as an endogenous ligand of the M-type receptor in mice but not in rats (38) and humans (40). It is then possible that rGIIA and hGIIA have their own receptors, distinct from the M-type receptor. Finally, whether the more recently characterized group IIC, group V, and group X sPLA 2 s are endogenous ligands of M-type receptors remains to be determined.  The CRD5 domain of the M-type receptor is centrally involved in sPLA 2 binding (46). It is therefore likely that the different binding properties of the M-type receptor in various animal species have to do with differences between these receptors in the CRD5 domain. Interestingly, we previously observed that, of the eight CRDs of the M-type receptor, the CRD5 domain is the least conserved among rabbit, mouse, bovine, and human receptors (46). On the other hand, we have also previously demonstrated that residues close to or within the Ca 2ϩ -binding loop domain of sPLA 2 are involved in binding to the M-type receptor (45). We particularly suggested that, besides glycine 30 and aspartate 49, which are perfectly conserved in sPLA 2 s and which are essential for binding, the identity of the residues at position 31 and possibly 34, which greatly varies in sPLA 2 s (59), may determine whether binding to the M-type receptor is possible or not. It is noteworthy that this view fits well with the fact that mGIIA binds with high affinity to the mouse M-type receptor, while rGIIA and hGIIA do not bind to this receptor (Table I and Ref. 38). Indeed, the Ca 2ϩ -binding loop domain of these three sPLA 2 s is perfectly conserved except at positions 31 and 34, where rGIIA and hGIIA, but not mGIIA, have the same residues (3,48,57). This observation, however, does not eliminate the possibility that residues located elsewhere in the sPLA 2 structure also contribute to the large differences in binding properties.
The physiological reason why mGIIA binds to the mouse M-type receptor while rGIIA and hGIIA do not bind to the respective rat and human M-type receptors is hard to understand. A tempting hypothesis would be that mGIIA is not the ortholog of rGIIA and hGIIA, i.e. that mGIIA has physiological functions that are distinct from those of rGIIA and hGIIA. In addition to differences in binding properties, several other lines of evidence would support this hypothesis. First, mGIIA, rGIIA, and hGIIA display relatively low levels of sequence identity compared with those observed between group IB sPLA 2 s. mGIIA has only 76 and 67% of identity with rGIIA and hGIIA, while mGIB has 89 and 81% of identity with rGIB and hGIB, respectively. Second, both hGIIA and rGIIA are expressed in many different tissues and cells (3,30), while the tissue distribution of mGIIA is thus far essentially restricted to intestine (where it is expressed at a very high level), with a low expression level in liver (Fig. 5) and the skin of new born mice (56). Furthermore, in hGIIA transgenic mice established with the complete hGIIA gene, the tissue distribution of hGIIA resembles that of hGIIA in humans, where it is expressed in many organs (60), in marked contrast with the endogenous expression of mGIIA (Fig. 5). In these transgenic mice, hGIIA was not expressed in the intestinal Paneth cells (60), whereas Paneth cells in wild-type mice are known to contain huge amounts of mGIIA (56,61,62). Since the transgenic mice were established with a transgene comprising a reasonably large 5Ј noncoding sequence of 1.6 kilobase pairs that contains transcriptional regulatory elements of the hGIIA gene (63,64), and since one would expect a conservation of the transcriptional regulatory elements between mGIIA and hGIIA genes if these latter were true orthologs, the difference observed between the endogenous expression of mGIIA and that of hGIIA in transgenic mice suggests that the transcriptional regulatory elements of the mGIIA and hGIIA genes are distinct and therefore supports the idea that mGIIA and hGIIA are not true orthologs. The third indication that mGIIA may not be the ortholog of rGIIA and hGIIA comes from physiological considerations. While there are many studies showing that expression of rGIIA and hGIIA is dramatically increased by proinflammatory cytokines, in many human inflammatory diseases, and in various rat models of inflammatory diseases (3), there is no clear evidence for an increased expression of mGIIA in inflammatory conditions in mouse cells or tissues. Thus far, only two studies have shown that the expression of mGIIA could be increased in intestine after injection of lipopolysaccharide (LPS) (65,66), but this increase was only modest as compared with the induction observed in LPS-treated rats (3,67). Finally, while many studies have shown that rGIIA and hGIIA play a role in lipid mediator release (3,68), mGIIA-deficient mice were found to have a normal inflammatory response that is mediated by mouse group V sPLA 2 (22,23); and while mGIIA has been proposed as a genetic modifier of colon tumorigenesis (25,26), the numerous attempts to demonstrate a similar role for hGIIA in human colorectal cancer were unsuccessful (69 -75).
The M-type receptor has been associated with a myriad of biological roles such as cell proliferation, cell contraction, cell migration, hormone release, and lipid mediator release (38,76). These effects are currently believed to be mediated by group IB sPLA 2 s but not by group IIA sPLA 2 s. However, since group IIA sPLA 2 s and cells from different animal species have been used in these previous studies, and since we now know that animal specificity is important for the M-type receptor interaction, some of these previous data may require reevaluation. For example, the mitogenic effect of pGIB on mouse fibroblasts was believed to indicate a specific action of group IB sPLA 2 s through binding to the mouse M-type receptor, since rGIIA and rbGIIA were found unable to bind to this receptor (8).
Since mGIIA now appears as a ligand of the mouse M-type receptor, this suggests that group IIA sPLA 2 s may also have mitogenic effects on these mouse fibroblasts that may be linked to the mouse M-type receptor. The recent targeted disruption of the M-type receptor gene in the mouse suggests that the M-type receptor plays a critical role in inflammatory processes induced by LPS and leading to endotoxic shock (13). M-type receptordeficient mice have a longer survival time than wild-type mice after challenge with LPS and are also resistant to the lethal effects of pGIB after sensitization with sublethal dose of LPS. The results of the present study indicating that mGIB is a natural ligand of the mouse M-type receptor fit well with the view that group IB sPLA 2 would play a role in processes leading to endotoxic shock after LPS challenge through binding to the mouse M-type receptor (13). mGIIA, which now appears as a second natural ligand of the mouse M-type receptor, is certainly not implicated in this resistance to endotoxic shock, since M-type receptor knock-out mice are also naturally deficient for mGIIA (13,25,65). Since several other mouse sPLA 2 s have now been identified (among them the mouse group V that has been shown to play a role in the release of lipid mediators of inflammation in place of mGIIA (22,23)), it will be important to analyze whether they can also act as natural ligands of the M-type receptor and whether they are involved in the inflammatory processes leading to endotoxic shock.
Besides a possible role of mGIIA in the production of inflammatory lipid mediators (20), mGIIA has been proposed to have bactericidal properties to protect the small intestine crypts from microbial invasion (62), and mice lacking mGIIA show an altered response to microbial infection (77). Whether the Mtype receptor may contribute to these effects remains to be analyzed. mGIIA was originally discovered as an intestinal protein called enhancing factor that increases the binding of the epidermal growth factor and synergizes with this latter to stimulate cell proliferation (24). The potential role of the Mtype receptor in these effects would be worth analyzing. More recently, mGIIA was identified as a genetic modifier of tumor formation in a mouse model of colorectal cancer (25,26). Mice carrying a deficient mGIIA gene were found to develop more intestinal adenomas than mice expressing a functional mGIIA (25). In addition, transgenic mice overexpressing mGIIA become resistant to intestinal tumorigenesis (26). While the mechanism by which mGIIA confers protection against adenoma formation is presently unknown (26), the colocalization of the M-type receptor with mGIIA in colon may suggest an implication of the M-type receptor in the resistance to intestinal tumorigenesis conferred by mGIIA.
In conclusion, this work has shown that both mGIB and mGIIA are probably physiological ligands of the mouse M-type receptor. This observation may be important to the understanding of the physiological and pathological roles of these sPLA 2 s in various processes such as cell proliferation, inflammation, and cancer. Furthermore, comparison of the binding properties of mouse and human sPLA 2 s to receptors from different animal species has indicated that sPLA 2 binding to M-type receptors is species-specific. Interestingly, species specificity of binding was previously observed for cytokines such as tumor necrosis factor and interleukin-1 (78 -80). In examining the biological effects of sPLA 2 s in further studies, it will be essential to use as much as possible sPLA 2 s extracted from the same animal as that used for in vitro or in vivo physiological studies. This work also suggests that mGIIA and hGIIA may not have similar functions and that new sPLA 2 s corresponding to the true orthologs of mGIIa and hGIIA may exist in mice and humans.