SREC-II, a New Member of the Scavenger Receptor Type F Family, Trans-interacts with SREC-I through Its Extracellular Domain*

The scavenger receptor expressed by endothelial cells (SREC) with an extremely large cytoplasmic domain, was originally identified in a human endothelial cell line. In this study, we have cloned a second isoform named SREC-II and shown that there is a heterophilic interaction between SREC-I and -II at their extracellular domains. The cDNA for murine SREC-II encodes an 834-amino acid protein with 35% homology to SREC-I. Similar to SREC-I, SREC-II contains multiple epidermal growth factor-like repeats in its extracellular domain. However, in contrast to SREC-I, SREC-II had little activity to internalize modified low density lipoproteins (LDL). A Northern blot analysis revealed a tissue expression pattern of SREC-II similar to that of SREC-I with predominant expression in human heart, lung, ovary, and placenta. Mouse fibroblast L cells with no tendency to associate showed noticeable aggregation when SREC-I was overexpressed in these cells, whereas overexpression of SREC-II caused only slight aggregation. Remarkably, intense aggregation was observed when SREC-I-expressing cells were mixed with those expressing SREC-II. Deletion of almost all of the cytoplasmic receptor domain had no effect on the receptor expression and cell aggregation, indicating that solely the extracellular domain is involved in cell aggregation. The association of SREC-I and -II was effectively suppressed by the presence of scavenger receptor ligands such as acetylated LDL and oxidized LDL. These findings suggest that SREC-I and -II show weak cell-cell interaction by their extracellular domains (termed homophilic trans-interaction) but display strong heterophilic trans-interaction through the extracellular epidermal growth factor-like repeat domains.

Scavenger receptors are defined by their ability to bind and metabolize modified low density lipoproteins (LDLs) 1 such as acetylated LDL (AcLDL) and oxidized LDL (OxLDL) (1). There are several different classes of scavenger receptors in mammalian cells, and their relative contributions to lipid metabolism in pathophysiological conditions such as atherosclerosis have been the subject of intense investigation (2). The first scavenger receptors to be purified and cloned were class A scavenger receptors (SR-A) expressed in macrophages (3,4). In experimental model animals for atherosclerosis such as apoE (Ϫ/Ϫ) mice (5) or LDL receptor (Ϫ/Ϫ) mice (6), the absence of SR-A induced a significant reduction in atherosclerotic lesion size. This demonstrates the involvement of SR-A in in vivo atherosclerosis. SR-A also participates in host defense activities by recognizing and mediating the endocytosis of pathogenic substances (7)(8)(9). Some forms of bacterial endotoxin are SR-A ligands, and SR-A-mediated uptake of endotoxin in vivo can play a significant role in endotoxin clearance (10). Therefore, SR-A-deficient mice exhibit an enhanced susceptibility to pathogens (5). Scavenger receptor class B type I (SR-BI) was initially identified in an expression cloning study that used AcLDL as the ligand (11) and now is recognized as a physiologically relevant HDL receptor (12). SR-BI facilitates the cellular uptake of cholesteryl esters from the hydrophobic cores of HDL by a mechanism that markedly differs from that of other lipoprotein receptors. It does not involve the internalization of the intact lipoprotein particle and its subsequent degradation. SR-BI mediates the selective uptake of cholesteryl esters from the hydrophobic core of the lipoprotein into the cell but not the apoprotein at the surface of the HDL. Adenovirus-mediated overexpression of SR-BI in the liver leads to reduced plasma HDL levels and increased cholesterol secretion into bile (13). SR-BI-null knockout mice (14) showed an increase in abnormally large, apoE-enriched HDL-like particles in the plasma. Most recently, it has also been demonstrated that SR-BI expressed on the surface of endothelial cells mediates endothelial nitric oxide synthase, resulting in increased nitric oxide production that may be essential for the athero-protective properties of HDL (15,16). These and other findings demonstrate that scavenger receptors, in addition to having a role in the metabolism of modified LDLs, have a plurifunctional pathophysiology.
Endothelial cells express several distinct scavenger receptors. In our previous work, we cloned a novel scavenger receptor from cultured human umbilical vein endothelial cells, termed SREC (scavenger receptor expressed by endothelial cells) (17), which have been renamed SREC-I here. SREC-I has no significant homology to other types of scavenger receptors but has unique domain structures. It contains 10 repeats of EGF-like cysteine-rich motifs in the extracellular domain. Recently, the structure of SREC-I was shown to be similar to that of a transmembrane protein with 16 EGF-like repeats encoded by the Caenorhabditis elegans gene ced-I, which functions as a cell surface phagocytic receptor that recognizes apoptotic cells (18). In the course of study elucidating the functional characteristics of SREC-I, we have cloned a novel gene that shows significant homology to SREC-I by searching the expressed sequence tag (EST) data base. In this paper, we describe the sequence and functional characterization of this unique protein termed SREC-II. Our results suggest that although SREC-II has little scavenger receptor activity, it undergoes a heterophilic trans-interaction with SREC-I in which the interaction takes place at the extracellular domains of these two proteins.
Cloning and Sequencing of Murine SREC-II cDNA-The nucleotide sequence of human SREC-I was used as a probe to search the Gen-Bank TM /EMBL/DDBJ data base of bacterial artificial chromosome-(BAC) using the TBLASTN program. Using the BAC clone of bd3-6 (accession number AC000096) as a probe, the partial nucleotide sequences of cDNA encoding peptides with homology to SREC-I were found in the EST data base of GenBank TM . Then, the EST clone, AA266135, was used as a probe to screen a murine lung Lambda cDNA library (Stratagene). LambdaZAPII bacteriophages were plated at a density of ϳ5 ϫ 10 4 plaque forming unit/100-mm plate. They were transferred to a nylon membrane (Biodyne A, Pall) and then hybridized using a 32 P-labeled probe as described. The DNA sequence was determined using the method of Sanger et al. (21) using Big dye terminator cycle sequencing kit and an Applied Biosystems model 377 DNA sequencer.
Expression of Murine SREC-I and -II in CHO and L Cells-The EcoRI/NotI fragment of the SREC-I cDNA and the EcoRI/XbaX fragment of the SREC-II cDNA were subcloned into the pcDNA3, and expression plasmids termed pcDNA3-SREC-I and pcDNA3-SREC-II, respectively, were obtained. CHO and L cells were transiently transfected with either pcDNA-SREC-I or -II using LipofectAMINE (Invitrogen) according to the manufacturer's instructions.
Uptake of Modified LDL-CHO cells grown near to confluent in 24-well plates were mock transfected or transfected with either pcDNA3-SREC-I or pcDNA3-SREC-II and incubated at 37°C for 48 h. Then, cells were further incubated with 2 mg/ml DiI-AcLDL or DiI-OxLDL for 4 h. After washing and cell solubilization by SDS, cellassociated fluorescence was measured by a spectrophotometer (Hitachi F-2000) at an excitation wavelength of 514 nm and an emission of 550 nm. The data are shown as the means Ϯ S.D. of four independent experiments.
Cell Aggregation Assay-L cells (4 ϫ 10 5 cells/well) in 6-well plates were mock transfected or transfected with either pcDNA3-SREC-I or -II and incubated at 37°C for 48 h. Cell aggregation assay was performed essentially as described (see Ref. 22). Briefly, cultured L cells were released from the plates with trypsin (0.25%)-EDTA (1 mM) treatment and were suspended in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum at 1 ϫ 10 6 cells/ml in polystyrene tubes. Then, tubes were gently rotated at 37°C for 1 h. The extent of cell aggregation was viewed microscopically and photographed. For quantitation, cell clusters of more than four cells were considered as aggregated ones. The data are shown as the means Ϯ S.D. of four independent experiments.

Molecular Cloning and Sequencing of cDNA for SREC-II-To
identify sequences similar to the human SREC-I gene, we examined the human and murine genome databases prepared in BAC. We identified the BAC clone of b562f10 (accession number AC007731) and bd3-6 (AC000096) on human chromosome 22 and murine chromosome 16, respectively, as sequences showing significant similarity to the extracellular domain of SREC-I. We further searched the EST data base for tags that are homologous to these two fragments, and we found two human (AI401171 and AA234558) and three murine (AA245496, AA266135 and AA024015) ESTs that encode novel peptides with significant amino acid homology to SREC-I. Several clones were isolated from a murine lung cDNA library using AA245496 as a probe. Sequence analysis of the longest cDNA demonstrated that three of the murine EST clones were fragments of the same gene. The murine cDNA comprising 3,310 base pairs consists of an open reading frame of 2,502 base pairs, ending with a TAA stop codon at nucleotides 2,598 -2,600, followed by a 3Ј-poly(A) tail. The first ATG triplet was determined by alignment with SREC-I, and the putative signal peptide was detected by hydropathy analysis (23). Because the protein encoded by the cDNA showed 35% overall homology to SREC, we have termed it SREC-II, and thereafter SREC was renamed SREC-I. Subsequently, the human SREC-II sequence was determined by searching the vicinity of the human chromosome 22q11 containing sequences homologous to murine SREC-II. Considering the GT-AG rule, we estimated the exonintron junctions, and in analogy with the murine sequence the amino acid sequence of human SREC-II was determined. The homology between murine and human SREC-II, as revealed by the GENETYX-MAC program, is 81.6%. Fig. 1A shows the amino acid sequences deduced from murine and human SREC-I and -II cDNA. The predicted translation product of murine SREC-II encodes a protein of 834 amino acids. It has two hydrophobic regions located at the N terminus and in the middle of the protein, which may serve as a signal sequence and a transmembrane domain, respectively, suggesting that SREC-II is a type I transmembrane protein. Indeed, the N-terminal half of the molecule has two potential N-glycosylation sites (Fig. 1B). As in the case of SREC-I, this putative extracellular domain is composed of 10 EGF-like repeats, five of which (segments 1, 2, 4, 6, and 9) fit the consensus sequence for the EGF-like domain (CXCXXXXXGXXC) exactly (17). The putative transmembrane domain located at the middle of the protein is followed by a cytoplasmic domain rich in serine (12.6%) and proline (12.6%). Moreover, the intracellular portion contains 13 potential phosphorylation sites for protein kinases A, C, or G, suggesting that some biological signals are delivered via this domain. It is noteworthy that the intracellular domain of SREC-II is rich in positively charged residues (i.e. arginine and lysine). The arginine ϩ lysine content is 14.7% in murine SREC-II and 16.4% in human SREC-II. Considering these structural similarities, we suggest here that both SREC-I and -II are members of the type F scavenger receptor family (2). A high degree of homology (about 50%) was observed between the extracellular domains of SREC-I and -II. On the other hand, the intracellular domains were less homologous, and only about 20% homology was revealed by the GENETYX-MAC program. These results suggest that although the extracellular domains share some biological functions, different signals are delivered by the intracellular domains.

Primary Structures of Murine and Human SREC-I and -II-
Distribution of SREC-II-The tissue distribution of human SREC-II was compared with that of human SREC-I by Northern blot analysis. As shown in Fig. 2, transcripts of SREC-I and FIG. 1. Primary structures of SREC-I and -II. A, primary amino acid sequences of murine and human SREC-I and -II. Gaps are inserted into the sequences for optimal alignment. Residues that are the same in all four proteins are shown in black, and residues that are the same in three -II were both detected at around 3.5 kb and exhibited similar tissue distributions. Both receptors were expressed predominantly in heart, placenta, lung, kidney, spleen, small intestine, and ovary.
Uptake of Modified LDL by CHO Cells Expressing SRECs-To verify the biological functions of SREC-II, we first compared the ability of modified LDL uptake between SREC-I and -II. The expression plasmid containing either murine SREC-I or murine SREC-II cDNA (pcDNA3-SREC-I or pcDNA3-SREC-II) was transfected into CHO cells, and the transfectants were incubated with DiI-AcLDL (Fig. 3A). As a positive control, SR-BI was also transfected. As reported previously, expression of SREC-I in CHO cells caused approximately a 24-fold increase in DiI uptake measured by fluorescence intensity. In contrast, little increase was detected when SREC-II was expressed in the cells. The expression levels of SREC-I and -II were approximately the same as verified by Myc-tagged receptors followed by Myc antibody staining (data not shown). We also measured the uptake of DiI-OxLDL and found that although there was a 12-fold increase in the uptake of DiI in SREC-I-expressing cells, no increase was detected in SREC-IIexpressing cells (Fig. 3B). These results indicate that whereas SREC-I bound to and internalized modified LDL efficiently, SREC-II-mediated uptake was marginal, implying that SREC-II had little scavenger receptor activity.
Aggregation of L Cells Expressing SREC-I-Because certain EGF-like repeats have been shown to mediate homo-or heterophilic protein-protein interactions (24,25), we next examined the effects of SREC-I on cell-cell interaction employing the cell aggregation assay (see "Experimental Procedures"). Monolayer cultures of L cells transfected with or without pcDNA-SREC-I were dissociated into single cells. When the suspensions were gently shaken, L cells expressing SREC-I tended to aggregate. In this assay system, about 30% of the cells were microscopically aggregated after an hour of shaking (Fig. 4b). Remarkably, only cells taking up DiI formed large aggregations when the assay was performed after incubation with DiI-AcLDL (Fig.  4bЈ), suggesting that the aggregation occurred by a homophilic interaction between SREC-I on the respective L cell surface. In control L cells that were transfected with the pcDNA plasmid alone, about 10% spontaneous aggregation was microscopically  observed, and the aggregates usually contained only 2-5 cells (Fig. 4a).
Effect of SREC-I C-terminal Deletions on AcLDL Uptake and L Cell Aggregation-SREC-I contains a large cytoplasmic domain consisting of about 400 amino acids. The following investigation was performed to determine the influence of this domain on AcLDL uptake and L cell aggregation (Fig. 5). Deletions of SREC-1 with missing C-terminal fragments of 170 and 370 amino acids (corresponding to deletions of about half and almost all of the cytoplasmic domain) were created, and their functions were determined. The results showed that the cytoplasmic domain has practically no effect on either assessed function, indicating that the intracellular domain has no major effect on the described functions of SREC-I. Fig. 6A shows the homo-and heterophilic interaction of SRECs visualized by an aggregation assay. The extent of the SREC-II-mediated aggregation was significantly less than that of SREC-I. Only about 15% of L cells were aggregated under the same conditions as shown in Fig. 6B. To our surprise, intensive aggregation was observed when SREC-I-expressing cells were mixed with those expressing SREC-II, with about 70% of the cells forming relatively large aggregates. About half of the cells in aggregated cell clusters showed DiI fluorescence after incubation with DiI-AcLDL (note identical cell clusters with and without fluorescence in Fig. 6C). Because SREC-I-expressing cells fluoresced when treated with DiI-AcLDL and SREC-IIexpressing cells did not, this experiment showed that both SREC-I-and SREC-II-expressing cells are involved in the formation of these cell clusters. The heterophilic interaction between SREC-I and -II was therefore far more intensive than the homophilic interaction and thus mediated the drastic increase in cell-cell association.

Homophilic and Heterophilic Interaction of SREC-I and -II-
Inhibition of SREC-mediated Cell Aggregation by Modified LDL-Because AcLDL and OxLDL are well known ligands of SREC-I (17), we examined the effects of these ligands on cell aggregation induced by the heterophilic interaction between SREC-I and -II. Cell aggregation was nearly completely inhibited in the presence of either AcLDL (Fig. 7b) or OxLDL (Fig.  7c), suggesting that the binding sites for modified LDL on SREC-I (and similar binding sites on SREC-II if any are present), were identical to or overlapping with those responsible for the homophilic and heterophilic interactions. Mouse SREC-I cDNAs encoding the fulllength amino acid residues 1-664 (⌬C170), and amino acid residues 1-464 (⌬C370), respectively, were constructed, Myc-tagged at the C termini, and transfected into L cells. After transfection, DiI-AcLDL uptake (A) and L cell aggregation assay (B) were performed as described.

DISCUSSION
In this paper, we have isolated a new member of the type F scavenger receptor family, named SREC-II. By definition, SREC-II may not belong to the scavenger receptor family because of the fact that unlike SREC I, SREC-II does not recognize AcLDL or OxLDL, which are typical ligands for scavenger receptors. We propose that SREC-I is a probable natural ligand for SREC-II and vice versa. L cells are commonly used in experiments demonstrating trans-receptor/receptor interactions, i.e. cadherins (26,27) nectins (22,28) and the neural adhesion molecule L1 (29). Employing this system, we have demonstrated that the SREC-I/II heterophilic trans-interaction is more intense than the SREC-I/I homophilic trans-interaction. SREC-I and -II are coexpressed in the human umbilical vein endothelial cell line (HUVEC) (data not shown). The similar tissue distribution patterns of SREC-I and -II provide some support for the existence of a SREC-I and -II heterophilic interaction in vivo.
It is not surprising that scavenger receptors function as adhesion molecules. Gordon and colleagues (30) demonstrated that SR-A mediates divalent cation-independent adhesion of murine macrophages to plastic plates used for tissue culture. SR-A also mediates macrophage association to activated lymphocytes (31) and glucose-modified basement membranes (32). CD36, a member of the SR-B family, is well known as an adhesion receptor for thrombospondin-1 (33,34) and erythrocytes infected by plasmodium falciparum (35,36). LOX-1, another member of the scavenger receptors expressed by endoassay. Phase-contrast light micrographs of typically aggregated L cells (a) and the fluorescence of cell-associated DiI-AcLDL (aЈ) are shown. thelial cells, can support rolling and adhesion of leukocytes at later stages of leukocyte-endothelium interaction (37). SREC-I and -II, which belong to the same class F SR family, may be an example exhibiting direct receptor-receptor trans-interaction between separate cells through extracellular domains.
The SR-F receptor family is also characterized by the presence of extremely large cytoplasmic domains when compared with other scavenger receptor families. A computer analysis revealed that some of these serine residues are potential phosphorylation sites for protein kinases A, C, and G. These structural features indicate that the cytoplasmic domains play a role in transducing intracellular signals other than receptor endocytosis. Indeed, we have preliminary observations that certain serine residues in the cytoplasmic domain of SREC-I are phosphorylated when expressed in L cells (data not shown). Moreover, when SREC-I is overexpressed in L cells, the cells form long processes 2 days after transfection. This effect only occurs when the cytoplasmic domain is present. These observations suggest that the cytoplasmic domain of SREC-I is involved in the regulation of intracellular cytoskeletal organization (data not shown). SREC-I and -II may transduce different signals due to low sequence similarity and different potential phosphorylation sites. It should also be noted that the cytoplasmic domain of SREC-II, unlike that of SREC-I, is rich in positively charged residues such as arginine and lysine. This may indicate that the cytoplasmic domains of SREC-I and -II interact with different cellular proteins.
The physiological function of the trans-interaction of SREC-I and -II are totally unknown at present. It is possible that homoand heterophilic interactions between SREC-I and -II play a role in the adherence of endothelial cells and the adhesion of circulating monocytes to endothelial cells in some inflammatory conditions (38). Although the cytoplasmic domain is not required for SREC-mediated cell-cell interactions, it is possible that formation and dissociation of the trans-interaction between SREC-I and -II delivers a signal to the cytoplasmic domains and transduces this signal to intracellular components such as the cytoskeletal system. Further studies are needed for elucidating the pathophysiological function, inter-, and intracellular signaling mediated by SRECs.
OxLDL promotes atherogenesis by supplying cholesterol and by generating agents that activate macrophages, smooth muscle, and endothelial cells. Moreover, it is well known that OxLDL induces a variety of cellular responses in vascular wall cells and circulating cells. For example, OxLDL is a chemoattractant for endothelial cells and T-lymphocytes (39,40). It increases the expression of VCAM-1 and P-selectin and the release of endothelin and interleukin-8 (41)(42)(43)(44). Furthermore, it potentiates the growth of macrophages and smooth muscle cells (45)(46)(47). In this study, we have demonstrated that OxLDL effectively dissociates the heterophilic trans-interaction of SREC-I and -II. We therefore propose that another pathological function of OxLDL is the ability to disrupt intercellular receptor/receptor interactions, as was shown to be the case with SREC in this study.