Identification of a domain (155-183) on CD36 implicated in the phagocytosis of apoptotic neutrophils.

Clearance of apoptotic neutrophils by macrophages is a crucial event following the resolution of acute inflammation. CD36, together with alphavbeta3, has been identified as one of the adhesion molecules on the surface of macrophages implicated in the clearance of polymorphonuclear leukocytes. The domain on CD36 implicated in the phagocytosis of aged neutrophils remains to be elucidated. In this study, COS cells transfected with human CD36 cDNA had a significantly higher capacity to phagocytose human apoptotic neutrophils compared with murine CD36 cDNA. Moreover, monoclonal antibodies 10/5 or OKM5 (epitopes identified on amino acids 155-183) but not monoclonal antibody 13/10 (epitope identified on amino acids 30-76) inhibited phagocytosis of apoptotic neutrophils by COS cells transfected by human CD36. Swapping the human CD36 155-183 domain from human to murine CD36 (human-murine CD36 chimera) imparted to murine CD36-transfected COS cells an increased capacity to phagocytose apoptotic neutrophils. Conversely, when the murine domain 155-183 was inserted in human CD36, a decreased phagocytic capacity was observed. In addition, a synthetic peptide(155-169) but not its scrambled form significantly inhibited phagocytosis. These results identify for the first time a functional domain encompassing amino acids 155-183 on human CD36 implicated in the recognition and phagocytosis of apoptotic neutrophils.

CD36 is a 88-kDa multifunctional adhesive glycoprotein expressed by platelets, monocytes, microvascular endothelial cells, mammary epithelial cells, erythroblasts, and several tumor cell lines (1). It has been identified as one of the receptors for collagen type I (2) and thrombospondin (TSP) 1 (3,4). Moreover, CD36 mediates the cytoadhesion of Plasmodium falciparum-infected erythrocytes to brain post capillary venular endothelium, a factor that contributes to the pathogenicity of the P. falciparum malaria (5). Recently CD36 has been shown to act as a receptor for oxidized low density lipoproteins, and thus, it may be implicated in the evolution of atherosclerosis (6). This adhesion molecule has also been reported to act as a signaling molecule capable of mediating an oxidative burst in monocytes (7) and activated platelets (8) by a process that may be associated with tyrosine phosphorylation (9). Recent observations show that E-selectin, expressed by activated endothelial or transfected L cells, is capable of significantly increase the number of CD36 molecules expressed by monocytes (10).
Clearance of apoptotic or senescent neutrophils (PMN) by macrophages is a crucial process in events such as hemostasis, wound healing, and tissue regeneration. Indeed, efficient macrophage phagocytosis of senescent neutrophils undergoing constitutive apoptosis is likely to be critical for successful resolution of inflammatory responses (11). To prevent damage to surrounding tissues, PMN undergo spontaneously apoptosis or programmed cell death (12), which leads to recognition and phagocytosis by macrophages (13). CD36, together with the ␣v␤ 3 integrin, has been identified as one of the adhesion molecules on the surface of macrophages implicated in the clearance of PMN (14). Apoptotic T cells, following the resolution of viral infections, are also recognized by these two adhesion molecules (15). Although increasing evidence supports the role of CD36 in phagocytosis of apoptotic cells, very little is known concerning the structural domain on the CD36 molecule implicated in this recognition process. The epitope for the mAbs that blocked phagocytosis of apoptotic PMN (14,16) was recently located in our laboratory to a region on CD36 encompassing amino acids 155-183 (17). In this study, human-murine chimeric CD36 constructs together with anti-CD36 mAbs and peptides derived from this region were used to show the functional role of the 155-183 domain.

MATERIALS AND METHODS
The cDNA coding for human and mouse CD36 were generously given by Dr Brian Seed (5) and Dr. Gerda Endemann (6), respectively. Lipo-fectAMINE reagent was from Life Technologies, Inc. Restriction endonucleases were from New England Biolabs (Beverly, MA) or Life Technologies, Inc. The COS-7 cell line was obtained from the European Collection of Animal Cell Cultures (Porton Down, Salisbury, Wiltshire, UK). Culture media and supplements were from Life Technologies, Inc. Other reagents were from different sources and were of the highest purity available.
Monoclonal Antibodies-Anti-CD36 monoclonal antibodies 10/5 and 13/10, were produced in our laboratory using for immunization a CD36expressing recombinant vaccinia virus (18). 2 These antibodies were further characterized by immunoprecipitation, Western blots, and fluorescence-activated cell sorting analysis of human platelets. mAbs 10/5 and 13/10 were found to be IgG2a and IgG2b, respectively, as deter-* This work was supported by the Ministère de l'Education Supérieure et de la Recherche ACC-SV9, Association pour la Recherche sur le Cancer (subvention 6586) (to J. L. M.) by Wellcome Trust Grant 039737 (to Y. R.), and by a National Institutes of Health Grant 1R01-HL42943 (to L. L. K. L.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
CD36 Constructs-Chimeric proteins were generated by introduction of endonuclease sites into the human and mouse cDNA without altering the coding sequence as described (17). Human and murine cDNA cloned in the eukaryotic vector pCDM8 were digested with the appropriate restriction endonuclease to allow exchange of sequences between these two species and the ligation products transformed into E. coli MC1061/P3.
Peptides-Linear peptides derived from the CD36 sequence were synthesized by Neosystems (Strasbourg, France) and purified by high performance liquid chromatography to a purity of Ͼ95% as analyzed by amino acid analysis and mass spectrometry. The numbering of the peptides corresponds to amino acid sequence of CD36 and is as follows: 148 -164, YQNQFVQMILNSLINL, and 155-169, MILNSLINKSKSSM. Control peptide for region 155-169 was identical to the authentic peptide, except that its sequence was scrambled. Stock solutions (14 mM) were prepared in water, and appropriate dilutions were made in Iscove's modified Dulbecco's medium, 10% autologous serum, just before each experiment.
COS Cell Transfection-COS cells monolayers cultured in 60-mm diameter dishes at 70% confluence were transfected with 2 g of DNA and 6 l of LipofectAMINE in DMEM according to the manufacturer's instructions.
Following 24 h of transfection, the cells were detached with trypsin/ EDTA, pelleted, and resuspended with DMEM, 10% fetal calf serum, seeded into 24-well plates (10 5 cells/well), and cultured in DMEM, 10% fetal calf serum for an additional period of 24 h. After 48 h of transfection, the expression of CD36 in the transfected cells was tested by anti-CD36 mAbs and flow cytometry with a FACSscan® using a LysisII software.
Cells-Neutrophils (Ͼ98% pure, Wright-Giemsa) were isolated from fresh, citrated blood from healthy donors, by dextran sedimentation and platelet-poor plasma Percoll density gradient centrifugation, aged for 24 h at 37°C, 5% C0 2 in Iscove's modified Dulbecco's medium containing 10% autologous platelet-rich plasma-derived serum. Apoptosis following 24 h of incubation was assessed in cytocentrifuge preparations stained with Wright-Giemsa (14). Human monocyte-derived macrophages were prepared by standard methods from adherent blood mononuclear cells by culture for 8 days in 24-well plates in Iscove's modified Dulbecco's medium containing 10% autologous platelet-rich plasmaderived serum, as described (14).
Phagocytosis Assays-Following 48 h of transfection, COS cells were washed with DMEM and then apoptotic PMN (aPMN) were added to the wells (2.5 ϫ 10 6 aPMN/well). Following coincubation between the transfected cells and aPMN for 3 h at 37°C, in Iscove's modified Dulbecco's medium, 10% autologous platelet-rich plasma-derived serum, noningested apoptotic neutrophils were then washed with cold DMEM. Subsequently, cells were detached with 150 l of trypsin/ EDTA, and a cytocentrifuge preparation was made from each well. These were then fixed with 2% glutaraldehyde, stained for myeloperoxidase to reveal ingested apoptotic neutrophils, and counterstained with Haemalum. The proportion of COS cells ingesting aPMN was assessed microscopically by counting 200 COS cells/slide.
Effects of Monoclonal Antibodies-Monocyte-derived macrophages or transfected COS-7 cells were incubated with mAbs (final concentration, 50 g/ml) for 30 min at room temperature. Cells were then washed twice with 2 ml of DMEM followed by the addition of 2.5 ϫ 10 6 aPMN/ well. Phagocytosis assay was carried out as above.
Effect of Peptides-The effect of CD36-derived peptides on the phagocytosis of aPMN by monocyte-derived macrophages or hCD36-transfected COS cells was determined using the standard protocol with peptides being added to the PMN suspension at the concentration indicated prior to its addition to the COS cell monolayer.
Statistical Analysis-Statistical analysis was performed using unpaired Student's t test.

RESULTS
Expression of Chimeric Constructs of CD36 -COS cells constitutively expressing ␣ v ␤ 3 as shown by mAb LYP18 (results not shown) were transfected with either human wild type CD36, mouse wild type CD36 or the chimeric constructs. Transient expression (48 h post transfection) of CD36 on the surface of COS cells was measured through the binding of mAbs (10/5 and 13/10) directed against different domains of CD36. COS cells transfected with mCD36 cDNA in which the 155-183 domain is replaced by its human counterpart (mhm 155-183) expressed the epitope for the mAb 10/5 but not for 13/10. Cells transfected with the reverse construct, that is the replacement in hCD36 cDNA of the human 155-183 domain by its murine counterpart (chimera designated as hmh 155-183), were recognized by 13/10 but not by 10/5. Transfected cells had similar percentage of expression with wild type hCD36 or chimeric constructs (Fig. 1). The level of expression of mCD36 on the surface of COS cells was similar to that observed for hCD36 as measured by an anti-Nak a antiserum (data not shown).
PMN Phagocytosis by Mouse/Human Chimeric Constructs-Wild type human CD36-transfected COS cells show a 3-fold increase (18.57 Ϯ 3.77%, n ϭ 8) in their capacity to internalize and ingest aPMN compared with COS cells transfected with wild type murine CD36 (7.55 Ϯ 2.08%, n ϭ 5) (Fig. 4). Human CD36 cDNA transfection did not elicit phagocytosis of freshly isolated (nonapoptotic) neutrophils (5.89 Ϯ 2.17%, n ϭ 5). Differences in phagocytic capacity of COS cells resulting from wild type human or murine CD36 cDNA transfection were exploited to locate the functional domain on CD36 implicated in the recognition and internalization of aPMN. For that purpose, chimeric constructs of human and murine CD36 were generated and subsequently transfected in COS cells. COS cells transfected with a mCD36 cDNA in which the 155-183 domain was replaced by its human counterpart (chimeric construct mhm 155-183) had a capacity to phagocytose aPMN (15.4 Ϯ 4.3%, n ϭ 5), which was significantly higher compared with wild type murine CD36 (p Ͻ 0.01). In contrast, COS cells transfected with a chimeric construct in which the human CD36 domain 155-183 was replaced by its murine counterpart (hmh 155-183 CD36) had a significantly reduced (p Ͻ 0.01) capacity to phagocytose apoptotic aPMN (7.79 Ϯ 4.3%, n ϭ 5) similar to mock transfected cells (Fig. 4).
Phagocytosis of aPMN Is Inhibited by Peptides Derived from Human CD36 -A peptide derived from amino acids 155-169, at a final concentration of 500 M, significantly inhibited the phagocytosis of aPMN by hCD36-transfected COS cells and monocyte-derived macrophages (by 46.63 Ϯ 4.19%, n ϭ 3 and 27.178 Ϯ 6.06%, respectively). No effect was observed for the peptide 155-169 in its scrambled form or by the peptide derived from the region 148 -164 (Fig. 5). DISCUSSION In the present study we have identified, for the first time, an important functional domain on CD36 involved in the phagocytosis of apoptotic neutrophils. The role of this domain in mediating phagocytosis of human apoptotic neutrophils, in tandem with ␣ v ␤ 3 , is supported by a number of lines of evidence: 1) COS cells transfected with mCD36 were observed to have a minimal capacity to phagocytose human aPMN compared with hCD36. Inserting the human CD36 155-183 domain into mCD36 imparts to the chimeric CD36 a capacity to phagocytose aPMN. 2) Conversely, inserting the mCD36 155-183 domain in the hCD36 abolishes its phagocytic capacity. 3) Monoclonal antibodies directed against the 155-183 domain but not mAbs directed toward the N-terminal domain block phagocytosis. 4) A peptide derived from hCD36 amino acids 155-169 but not its scrambled form blocks phagocytosis.
A recent study has shown that the functional repertoire of human CD36 includes the capacity to promote, together with ␣ v ␤ 3 , phagocytosis of apoptotic cells on COS and Bowes mela-noma cells (16). Moreover, it has been reported that the cooperation of CD36 and the ␣ v ␤ 3 integrin is crucial for recognition and phagocytosis of aPMN (22)(23)(24). Our results are in line with these findings but also show that mCD36 cDNA does not have the same capacity as hCD36, when transfected in COS cells expressing ␣ v ␤ 3 , to phagocytose human apoptotic neutrophils. Indeed, it is of interest to note that human and murine macrophages may use different receptors for the recognition and phagocytosis of aPMN (25). Differences in phagocytosis capacity of COS cells, resulting from wild type human or murine CD36 transfection, together with epitope mapping of blocking mAbs have allowed us to identify a domain (amino acids 155-183) on hCD36 implicated in the recognition and internalization of aPMN.
The 155-183 domain is adjacent to a region encompassing amino acids 139 -155 that has been reported to represent a part of the OKM5 epitope and that is implicated in the initial binding to TSP (26). It is important to note that domain 139 -155 is identical in both the human and mouse proteins with the exception of residues 146 and 152 (17). Thus, phagocytosis of apoptotic neutrophils is dependent on the 155-183 region, but residues in the adjacent 139 -155 sequence may be important as well as observed for mAb binding (17,26). Strong evidence is available to show that interaction between aPMN-and monocyte-derived macrophages or CD36-transfected cell lines is mediated by TSP (14,16). According to the two step model of Leung et al. (26), the 139 -155 domain might be implicated in the initial binding of TSP that would consequently give rise to a high affinity binding site (amino acids 93-110) on CD36. Exposition of this second site allows further consolidation of the CD36-TSP interaction. Human TSP, in the same way as blocking mAbs, may initially interact with CD36 via a discon- tinuous site involving amino acids 139 -155 and 155-183 domains. Replacing the 155-183 domain, in human CD36, by its murine homologue may either affect the stability of the TSP-CD36 interaction and/or the transmission of necessary outside-in signals (9) to the macrophage and prevent the recognition and phagocytosis of aPMN. Alternatively, any changes in the sequence of 155-183, from human to murine, may locally affect the conformation of CD36 and consequently alter its capacity of phagocytosis. A peptide (155-169) derived from region 155-183 inhibits, in contrast to its scrambled form or a peptide derived from region 148 -164, phagocytosis of aPMN by hCD36-transfected COS cells. Differences in functional reactivity between these two peptides may conceivably be due to one (155-169) but not the other (148 -164), acquiring a conformation that is closer to the native state of this region. The work of Asch et al. implicates amino acids 87-99 in the binding of TSP and shows that this binding is linked to the phosphorylation of Thr 92 (27). Moreover, the group of Silverstein uses recombinant proteins to show the role of the 93-120 domain in binding TSP (28). The current hypothesis, modified following results obtained in this study, would suggest that TSP interacts in a first stage with a conformational site on CD36, represented by region 139 -183, prior to its binding to a high affinity site most probably represented by region 87-120 (26).
The data presented in this study indicate that the functional domain encompassing amino acids 155-183 plays a critical role in the phagocytosis of aPMN. Moreover, recent data from our laboratory suggest that region 155-183 is a multifunctional domain implicated as well in the cytoadherence of P. falciparum-infected erythrocytes to CD36 (29), as a signaling molecule implicated in inducing platelet aggregation and secretion (18) and in oxidized LDL binding (30). The domain 155-183 is coded by a single exon (exon VI) (31), and it is possible that this exon delineates an independent, structural, and functional domain of human CD36, directly implicated in cell-cell interactions.