Cell Adhesion to Phosphatidylserine Mediated by a Product of Growth Arrest-specific Gene 6*

Gas6, a product of a growth arrest-specific gene 6, potentiates proliferation of vascular smooth muscle cells and prevents cell death of vascular smooth muscle cells. It has been also demonstrated that Gas6 is a ligand of receptor tyrosine kinases Axl, Sky, and Mer. Gas6 contains γ-carboxyglutamic acid residues, which are found in some blood coagulation factors and mediate the interaction of the coagulation factors with negatively charged phospholipid. In this study, we clarified that Gas6 specifically bound to phosphatidylserine and the binding was dependent on Ca2+ and γ-carboxyglutamic acid residues. Furthermore, we found that U937 cells, which express Gas6 receptor on their surfaces, adhered to phosphatidylserine-coated enzyme-linked immunosorbent assay (ELISA) plate only in the presence of Gas6 and Ca2+. U937 cells also bound to ELISA plate coated with phosphatidylinositol, but the binding was independent of Gas6 and Ca2+. On the other hand, U937 cells did not adhere to phosphatidylcholine- or phosphatidylethanolamine-coated ELISA plate even in the presence of Gas6 and Ca2+. These findings suggest that Gas6 may play a role in recognition of cells exposing phosphatidylserine on their surfaces by phagocytic cells, which is supposed to be one of the mechanisms for clearing dying cells.

Gas6 was initially identified as a product of a gene whose expression in fibroblasts increased during the growth-arrested state, although its biological activity was unclear (1). We purified Gas6 as a protein, which is released from cultured vascular smooth muscle cells (VSMCs) 1 and specifically potentiates VSMC proliferation mediated by Ca 2ϩ -mobilizing receptors (2). Another biological activity of Gas6 is the prevention of cell death of VSMCs and fibroblasts (3,4). On the other hand, Gas6 has been suggested as a ligand for receptor tyrosine kinases Axl, Sky, and Mer (5)(6)(7)(8). The members of this receptor subfamily, including Axl (also called Ufo and Ark) (9 -11), Sky (Rse, Brt, and Tyro3) (12)(13)(14)(15), and c-Eyk (Mer) (16, 17), contain a characteristic extracellular ligand-binding domain composed of two immunoglobulin-like motifs and two fibronectin type III motifs. Thus, the biological activities of Gas6, i.e. potentiation of cell proliferation and prevention of cell death, are very likely to be mediated by these receptor tyrosine kinases.
Gas6 has 46% amino acid identity to protein S, a serum protein negatively regulating blood coagulation. Similarly to protein S, Gas6 is composed of defined structural motifs: a Gla domain, four epidermal growth factor-like repeats, and a Cterminal domain (1). The Gla domain is rich in ␥-carboxyglutamic acid (Gla) residues, which are found in some factors in blood coagulation. Amino acid analysis of Gas6 revealed that Gas6 contains 11-12 Gla residues (2). The function of the Gla domain of protein S is thought to be mediation of its Ca 2ϩ -dependent binding to negatively charged phospholipids, which is important for the biological activity of protein S (18). The specific binding of Gas6 to its receptor on cell membrane has been shown to be dependent on Ca 2ϩ (19), suggesting that interaction of the Gla domain with Ca 2ϩ is important for the receptor binding. Moreover, Gla-deficient Gas6 lacks receptor binding activity and growth-potentiating activity, indicating that Gla residues of Gas6 are necessary for Gas6 activity (20).
In this study, we examined the interaction of Gas6 with phospholipids and found that it specifically bound to phosphatidylserine (PS) and that the interaction was dependent on Ca 2ϩ and Gla residues. Furthermore, we demonstrate that Gas6 functions as an adhesion molecule, which binds Gas6 receptorexpressing cells to PS-exposing surfaces.
Preparation of Recombinant Rat Gas6 and Gla-deficient Gas6 -CHO cells were transfected with rat Gas6 expression plasmid. Confluent CHO cells were cultured in protein-free culture medium PM-1000 (Eiken, Japan) in the presence of 4 M vitamin K 2 . Recombinant rat Gas6 was purified from the culture medium as described elsewhere (2). To prepare Gla-deficient Gas6, the CHO cells were cultured in PM-1000 in the presence of 1 M warfarin. The Gla-deficient Gas6 was prepared as described elsewhere (20).
Preparation of Axl-Fc Fusion Protein-cDNA fragment coding for the extracellular domain of human Axl (amino acid residues 1-438) was ligated in-frame with the polymerase chain reaction-amplified cDNA fragment of the Fc region (residues 216 -443) of human IgG 1 , with the spacer sequence Ser-Ser-Val-Pro-Gly. The fused cDNA was subcloned into PUC-SR␣ expression vector and transfected into COS-7 cells, using a liposome method. The serum-free conditioned medium of COS cells was collected for 3 days, and the Axl-Fc was purified with Protein A-Sepharose (Pharmacia Biotech Inc.).
Binding of Gas6 to Phospholipids-The binding of Gas6 to phospholipids was examined by enzyme-linked immunosorbent assay (ELISA) * 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.
§ To whom correspondence should be addressed. Tel.: 81-6-458-5861; Fax: 81-6-458-0987; E-mail: toru.nakano@shionogi.co.jp. 1 The abbreviations used are: VSMC, vascular smooth muscle cell; (21). Phospholipids were dissolved in ice-cold ethanol (2 g/ml unless otherwise described), and 100 l of the individual phospholipids was used to coat each well of the ELISA plate (Corning). Wells coated with ethanol alone were used as a control. The wells were kept at 25°C for 18 h to evaporate the ethanol and blocked by incubation of the phospholipid-coated wells with Tris-buffered saline (TBS) (10 mM Tris-HCl, pH 7.4, 150 mM NaCl) containing 3% bovine serum albumin (BSA) for 1 h at room temperature. Then, the wells were washed with TBS containing 0.05% Tween 20 and incubated with Gas6 or Gla-deficient Gas6 in TBS containing 5 mM CaCl 2 unless otherwise described. The binding of Gas6 or Gla-deficient Gas6 was determined using rabbit anti-rat Gas6 IgG and peroxidase-conjugated anti-rabbit IgG (Chemicon International Inc.). The anti-Gas6 IgG detected Gas6 and Gladeficient Gas6 equally on ELISA and Western blotting (data not shown). The binding of Gas6 to PS is very firm. EDTA, PS, or a high concentration of salt, which all inhibited the binding, did not dissociate bound Gas6 from the PS-coated plate. Only a strong solubilizing reagent, such as SDS, guanidine HCl, or urea, dissociated the bound Gas6 (data not shown).
BIAcore Analysis-Axl-Fc fusion protein (0.7 g) was immobilized to the carboxymethyl dextran layer of a CM5 sensor chip on a BIAcore instrument (Pharmacia), using procedures described in the manufacturer's manual. Gas6 was dissolved in TBS containing 0.1% BSA and 2 mM CaCl 2 with or without phospholipids and passed over the immobilized Axl-Fc at a flow rate of 2 l/min for 20 min. The binding of the ligand with Axl-Fc was monitored in real time by the increase in the relative resonance unit on the sensorgram. The apparent association rate constants (k a ) and dissociation rate constants (k d ) were calculated using the manufacturer's software. The equilibrium dissociation con- Adhesion of U937 Cells to PS-ELISA plates were coated with phospholipids (0.5 g/well) as described above. The wells were blocked with 3% BSA in phosphate-buffered saline (PBS) (137 mM NaCl, 2.7 mM KCl, 8.1 mM Na 2 HPO 4 , 1.3 mM KH 2 PO 4 , pH 7.4) for 1 h and washed with PBS containing 0.05% BSA. The wells were filled with PBS, 0.05% BSA containing U937 cells (1 ϫ 10 5 cells/well) and various concentrations of Gas6 and CaCl 2 . The plate was sealed with a transparent film and incubated for 30 min at 37°C. Then, the plate was inverted and incubated for 15 min. During the incubation, unbound cells came off the plate. The cells on the plate were observed with a microscope, and adherent cells in a field were counted.

RESULTS AND DISCUSSION
First, we investigated the interaction of Gas6 with phospholipids. The wells on the ELISA plate were coated with various concentrations of phospholipids, including acidic phospholipids (PS and PI) and neutral phospholipids (PC and PE). The wells were incubated with Gas6, and Gas6 bound to the wells was detected with ELISA using anti-Gas6 IgG. As shown in Fig. 1A, Gas6 interacted with PS but not with PC, PE, or PI. Fig. 1A also demonstrates that the binding was dependent on the concentration of PS, with half-maximal binding occurring at approximately 60 ng/well. Since interaction of some Gla-containing coagulation factors with phospholipids is dependent on Ca 2ϩ , we examined the Ca 2ϩ dependence of the interaction of Gas6 with PS. As shown in Fig. 1B, approximately 2 mM Ca 2ϩ was required for maximal binding, whereas Mg 2ϩ did not substitute for Ca 2ϩ .
In the experiment shown in Fig. 1C, the wells coated with PS or those without coating were incubated with various concentrations of Gas6. Specific binding of Gas6 to PS, which was calculated by subtracting the values with non-coated wells from the values with PS-coated wells, was saturable, with half-maximal binding observed at approximately 0.7 g/ml Gas6.
With some coagulation factors the Gla residues are known to be important for interactions with negatively charged phospholipids (23). We thus investigated the contribution of Gla residues to the binding of Gas6 to PS. As shown in Fig. 1C, no specific binding of Gla-deficient Gas6 to PS was observed, indicating that Gla residues are essential for the binding of Gas6 to PS.
Since the interaction of Gla-containing blood coagulation factors with negatively charged phospholipids remarkably increases the activity of the coagulation factors, we next examined the effect of PS on the binding activity of Gas6 to its receptor Axl. Table I summarizes the effect of PS and PC on k a , k d , and K d estimated with BIAcore. The most significant effect of PS was observed on the k a value. PS but not PC increased it by 6-fold, whereas PS increased the k d value by about 4-fold. As a result, PS decreased the K d value by approximately 30%. On the other hand, PC decreased k a and increased k d , resulting in an increase of K d by about 10-fold.
The above results demonstrated that Gas6 interacted with PS and the interaction facilitated receptor binding of Gas6. However, the enhancement of receptor binding is not remarkable enough to consider it as a major meaning of the interaction of Gas6 with PS. On the other hand, these findings also implied that Gas6 might be a divalent ligand and raised the possibility of its function as an adhesion molecule. The requirement of Gla residues for the interaction of Gas6 with PS suggests that Gas6 binds to PS via the N-terminal Gla domain. Furthermore, it has been reported that Gas6 interacts with its receptor via the C-terminal globular domain (24). Therefore, we hypothesized that Gas6 could connect the cells containing Gas6 receptor with the cells expressing PS on their surfaces.
To assess this possibility, we examined the interaction of monoblastic cells U937 with PS coated on ELISA plates. It has been reported that U937 cells express Axl, a receptor for Gas6 (25). When U937 cells were incubated with an ELISA plate coated with PS, they did not bind to the plate ( Fig. 2A). However, when they were incubated in the presence of Gas6, U937 cells adhered to the plate (Fig. 2B). In the experiment in Fig. 2, U937 cells and Gas6 were simultaneously incubated in PScoated wells. On the other hand, when PS-coated wells were first incubated with Gas6, washed, and then incubated with U937 cells, the cells also adhered to the wells (Fig. 2C). Furthermore, Gas6 stimulated the binding of U937 cells to PS when the incubation was carried out at 4°C instead of at 37°C (data not shown). From these results, it is suggested that Gas6 interacts with PS and Axl at the different sites in Gas6 molecule and mediates adhesion of Axl-expressing cells to PS-expressing surfaces. As shown in Fig. 3, binding of U937 cells to PS required Ca 2ϩ , which is also required for the interaction of Gas6 with PS (Fig. 1B) or receptor (20). Fig. 3 also shows that Gla-deficient Gas6, which does not bind to PS (Fig. 1C) or receptor (20), did not stimulate the binding of U937 cells to PS-coated plate. Fig. 4 shows the binding of U937 cells to the plates coated with several phospholipids. Binding of U937 cells to PS definitely required both Ca 2ϩ and Gas6. U937 cells also adhered to the PI-coated plate, whereas the adhesion did not require Ca 2ϩ or Gas6. U937 cells weakly adhered to the plates coated with PC or PE. However, the addition of Ca 2ϩ decreased the binding to PC or PE and Gas6 did not affect the binding.
PS is normally almost totally confined to the inner leaflet of the plasma membrane but has been reported to be exposed on apoptotic cells, senescent red blood cells, or activated platelets (26 -28). Therefore, PS is supposed to be one of the markers for phagocytic macrophages to identify dying cells. Our results demonstrate that binding of monoblastic cells U937 to PS is dependent on Gas6. Thus, Gas6-mediated adhesion may be one of the mechanisms for phagocytic cells to recognize dying cells, which express PS on their surfaces. It has been reported that classes A and B macrophage scavenger receptors may be involved in phagocytosis of apoptotic cells (29,30). As shown in Fig. 4, adhesion of U937 cells to PI is not dependent on Gas6 or Ca 2ϩ . Therefore, adhesion to PI, another negatively charged phospholipid, may be mediated by other mechanisms including those scavenger receptors.
In this study, we demonstrated that Gas6 links Axl-expressing cells to the PS-expressing surface and suggested that Gas6dependent adhesion may be involved in the recognition of dying cells by phagocytic cells. The Gas6-mediated adhesion is unique since it also stimulates tyrosine phosphorylation in receptor-expressing cells. The phosphorylation may stimulate phagocytic activity of the cells. However, further in vivo studies are necessary to clarify the physiological importance of the Gas6-mediated cell adhesion mechanism.