INTRODUCTION

Surfactant protein A (SP-A)¹ is an abundant glycoprotein associated with pulmonary surfactant and is synthesized and secreted primarily by type II epithelial cells in the alveolus (1). SP-A is also expressed and synthesized in cells of tracheal, bronchial, and tracheobronchial glands indicating that SP-A may be secreted in conducting airways (2). SP-A is a member of the Ca²⁺-dependent lectin family of proteins containing an amino-terminal collagen-like domain and a carboxyl-terminal carbohydrate recognition domain (CRD) (3) that are linked by a more hydrophobic domain. SP-A is thought to play an important role in the structural organization, biophysical activity and homeostasis of surfactant lipids, and to interact with alveolar type II epithelial cells and macrophages (4, 5, 6). The interaction of SP-A with type II epithelial cells has been associated with inhibition of phospholipid secretion (7), involvement in uptake of liposomal phospholipids (8), participation in recycling of phospholipids by type II cells (9), and secretion of colony-stimulating factors (10). SP-A has also been implicated in lung-specific host defense due to its interaction with a variety of pathogens (11, 12, 13, 14) and because it stimulates chemotaxis, phagocytosis, and secretion of colony-stimulating factors and reactive oxygen intermediates by alveolar macrophages (12, 15, 16, 17, 18, 19).

The activities of SP-A are thought to be mediated by specific receptors on the cell-surface of alveolar type II cells and macrophages (4, 5, 6). Recent studies with recombinant SP-A mutants and domain specific SP-A antibodies show that the CRD domain of SP-A is important in binding to at least one population of receptors on the surface of type II cells (20, 21). There are conflicting reports regarding the mechanisms of SP-A binding to macrophages. Pison et al. (6) suggested that SP-A binds to alveolar macrophages via its collagen-like domain. Malhotra et al. (22, 23) reported that SP-A interacts with the C1q receptor on U937 macrophage-like cells, again implicating the collagen-like domain of SP-A in this interaction. However, Schlepper-Schaffer and colleagues (24, 25) reported that SP-A bound to monocytes and alveolar macrophages through a mannose-dependent process that may involve the lectin domain of SP-A.

In the present work a novel receptor for SP-A was identified on macrophages and rat lung membranes. Polyclonal antibodies to the receptor inhibit SP-A binding to alveolar macrophages and type II cells and block SP-A-mediated inhibition of lipid secretion by type II cells. Our findings suggest that alveolar macrophages and type II cells share an SP-A receptor.

EXPERIMENTAL PROCEDURES

Materials

Most reagents were purchased from Sigma. The protease inhibitors Pentafloc SC and (RS)-2-carboxy-3-phenyl propionyl-Leu-OH (CBBL) were purchased from Boehringer Mannheim and Bachem Bioscience (King of Prussia, PA), respectively. N-Hydroxysuccinimidyl-6-(biotinamido)-hexanoate (Pierce), streptavidin, and HRP-conjugated streptavidin were purchased from Pierce. Maltose-Sepharose was prepared as described previously by Townsend and Stahl (26). SP-A, biotinylated as described previously (27) at pH 6.3 using N-hydroxysuccinimidyl-6-(biotinamido)-hexanoate, was kindly provided by Dr. Gary Ross (Children's Hospital, Cincinnati, OH). C1q was purchased from Calbiochem and was filtered through a 10-ml Sephadex PD-10 column into PBS prior to use. Carrier-free Na¹²⁵I and the ECL Western blot detection reagent kit were from Amersham Corp.

Purification of SP-A

Human SP-A was purified from unfractionated alveolar proteinosis fluid (APF) according to a previously described procedure with some modifications (28). APF was delipidated by dropwise addition into 1-butanol (0.5 ml of APF/25 ml of 1-butanol). The insoluble protein was recovered by centrifugation, dried under a constant stream of nitrogen, extracted twice with 5 mM Hepes, pH 7.5, containing 0.15 M NaCl, 20 mM -D-octylglucoside and extensively dialyzed in 5 mM Hepes, pH 7.5, containing 1 mM EDTA and 0.02% azide. Soluble SP-A was then recovered by centrifugation at 15,000 × g. The protein concentration was adjusted to 1 mg/ml and stored at 4 °C.

Iodination of Proteins

Radioiodination of proteins was performed as described previously using chloramine T (29, 30) except that iodination of SP-A and C1q was carried in 0.1 M potassium phosphate, pH 7.7, containing 10 µg of chloramine T and 0.15 mCi of Na¹²⁵I.

Isolation of Macrophages

Rat bone marrow-derived macrophages (RBMM) were prepared as described previously (29). Rat alveolar macrophages (RAM) were isolated by lung lavage with Ca²⁺- and Mg²⁺-free Hanks' balanced salt solution. Cells were suspended in binding buffer (Hanks' balanced salt solution containing 20 mM Hepes, 20 mM TES, pH 7.4, 1 mM CaCl₂, 0.2 mM MgCl₂ and 1% BSA). U937 cells were grown in RPMI containing 10% fetal bovine serum. For isolation of the SP-A receptor, U937 cells were cultured in 175-mm³ flasks to a density of 100 million cells per flask in a volume of 100-150 ml. For studies with surface-labeled cells, U937 cells were iodinated using glucose oxidase-lactoperoxidase as described previously (31). Cell viability was >98%.

Isolation of Rat Alveolar Type II Epithelial Cells

Type II cells were isolated from 200-300-g male Sprague-Dawley rats essentially as described by Dobbs et al. (32) except that type II cells were cultured in Dulbecco's modified Eagle's medium containing 5% rat serum for 24 h. To suspend adherent type II cells the plates were rinsed with phosphate-buffered saline containing 5 mM EDTA and incubated in phosphate-buffered saline-EDTA for 15 min at 4 °C. Subsequently, the cells were scraped off the plates, centrifuged, and suspended in binding buffer. Cell viability was >90%.

Binding Assays

Binding of radiolabeled proteins to RBMM and U937 cells was performed as described previously (34). Binding to alveolar macrophages and type II cells was performed in binding buffer at a concentration of 2.5 to 3.5 million cells/ml in a volume of 0.3 ml. After 1 h on ice the cells were pelleted at 7,000 × g for 5 min, washed once in binding buffer, resuspended in 100-µl aliquots, and counted in a gamma counter. Nonspecific binding was determined in the presence of 10 mM EDTA.

Electrophoretic Methods

Except where indicated, SDS-PAGE was performed under reducing conditions. For all ligand blotting procedures samples were mixed in reducing Laemmli SDS-PAGE buffer and loaded on gels within 30 min. All other samples were boiled for 3-5 min prior to electrophoresis. Fractionated proteins were electrophoretically transferred to nitrocellulose by semidry blotting. The anode buffer was composed of 0.3 M Tris, pH 10.4, and 20% methanol. The cathode buffer contained 25 mM Tris, 40 mM -amino-n-caproic acid, pH 9.4, and 0.1% SDS. Transfers were performed for 40 min at 170 mA for large gels or 100 mA for mini gels.

Detection of SP-A Receptor on Nitrocellulose Blots

Blots were initially blocked for 1-2 h at room temperature with TBS-T (25 mM Tris, pH 7.6, 0.146 M NaCl, 0.1% Tween 20 containing 5% BSA). All ligand blotting procedures were performed in TBS-T supplemented with 1 mM CaCl₂.

Blots were incubated with unlabeled SP-A at a concentration of 1 µg/ml at 4 °C overnight, washed, then probed with rabbit anti-human mannose receptor antibody (1:200) (33). After incubation with HRP-conjugated secondary antibody (1:1000) the membranes were washed and developed using 3,3-diaminobenzidine according to standard methods (33).

This method was initially developed by Dr. Gary Ross (Children's Hospital, Cincinnati, OH).² The blots were incubated at 4 °C overnight with 1.3 µg/ml bSP-A. Blots were then washed three times, 10 min each, with copious amounts of buffer. To detect bound bSP-A the blots were incubated for 1 h with HRP-conjugated streptavidin at a dilution of 1:50,000 from a stock solution of 1 mg/ml. The blots were washed three times in buffer and developed by chemiluminescence. In some experiments blots were blocked sequentially with streptavidin and D-biotin prior to addition of bSP-A in order to determine the presence of endogenous biotinylated proteins.

Blots were incubated overnight at 4 °C with 1 µg/ml radioactive ligand, washed as described earlier, and exposed to autoradiographic film.

Blots were probed with rabbit anti-rat lung SP-R210 (described below) at a dilution of 1:3000 in blocking buffer for 1-3 h at room temperature or overnight at 4 °C. The blots were washed three times for 10 min each with TBS-T and probed with secondary HRP-conjugated goat anti-rabbit IgG at a dilution of 1:20,000 for 2-4 h at room temperature. The blots were washed with TBS-T, briefly washed with water, and developed using chemiluminescence.

Detergent Extraction of Cells

Cells were pelleted and resuspended in homogenization buffer (HB: 25 mM Tris, pH 7.5, containing 5 mM MgCl₂, 1 mM EGTA, 1 mM PMSF, 1 mM o-phenanthroline, 10 µg/ml leupeptin, 50 µg/ml bacitracin, 3 µg/ml chymostatin, 10 µg/ml bestatin, 600 µg/ml Pentafloc SC, 20 µM phosphoramidon, 20 µM (RS)-2-carboxy-3-phenylpropionyl-Leu-OH, and 1 mg/ml DNase I) at a ratio of 1 ml/10-20 million cells. After addition of Triton X-100 to 1%, cells were maintained on ice for 30 min and used as whole cell extracts or microcentrifuged for 10 min to separate detergent-insoluble from detergent-soluble membranes.

Preparation of SP-A Affinity Column

Purified SP-A was bound to maltose-Sepharose noncovalently via its CRD. SP-A was dialyzed in 25 mM Tris, pH 7.4 and mixed with maltose-Sepharose resin at a ratio of 0.6 mg of SP-A/ml of packed resin. After mixing, CaCl₂ was added to a final concentration of 1.5 mM. The supension was rotated for several hours at room temperature or overnight at 4 °C. The affinity column was washed to remove unbound material, resuspended in binding buffer and stored in the cold until used. Greater than 95% of added SP-A bound to the maltose-Sepharose.

Isolation of SP-A Receptor from U937 Cells

All subsequent steps were performed at 0-4 °C. U937 cells were washed once in RPMI and resuspended in 10 mM Tris, pH 8.0, containing 1 mM MgCl₂, 1 mM EGTA, 1 mM PMSF, 1 mM o-phenanthroline, 10 µg/ml soybean trypsin inhibitor, 10 µg/ml leupeptin, 10 µg/ml bestatin, 100 µg/ml bacitracin, 20 µM phosphoramidon, and 20 µM CBBL (solution A). The cells were left on ice until 75% of cells lysed (20-40 min) and then homogenized with 10 strokes in a Wheaton Ten-Broeck chamber. At this stage, 97% of cells lysed. The homogenate was then centrifuged at 800 × g and a total membrane fraction was collected from the supernatant at 100,000 × g for 1 h. The membranes were resuspended using a plastic Pasteur pipette and allowed to swell in 10 ml of above buffer for a few hours to overnight. The membranes were solubilized for one hour on ice by addition of 20% Triton X-100 to 2% final concentration and adjusted to 0.15 M in NaCl. Insoluble material was removed by centrifugation at 100,000 × g for 1 h. Solubilized membranes were rotated overnight with a 1 ml packed volume of gelatin-Sepharose to remove fibronectin and collagenases and then with a 3-ml packed volume of maltose-Sepharose for 6 h in the presence of 2 mM CaCl₂ to remove maltose-binding lectins. The unadsorbed protein was rotated overnight with a 3-ml packed volume of previously prepared SP-A-maltose-Sepharose. Unbound protein was eluted and the resin washed with 10 volumes of solution A containing 2% Triton X-100. The resin was then washed with 5 volumes of 25 mM Hepes, pH 7.4, containing 60 mM octylglucoside, 0.15 M NaCl, 2 mM CaCl₂, and protease inhibitors (solution B). The SP-R210 was then eluted in solution B adjusted to 0.25 mM in EGTA, 0.5 M in NaCl and pH 6.5. The SP-R210 fraction was dialyzed against 25 mM Hepes, pH 7.4, containing 10 mM octylglucoside, 0.15 M NaCl, 2 mM EDTA, 0.5 mM PMSF, 0.5 mM o-phenanthroline, and concentrated to 1.5 ml by osmosis against polyethylene glycol 20,000.

Partial Purification of SP-A Receptor from Rat Lung Membranes and Development of Polyclonal Antibodies

Rat lungs were perfused via the pulmonary artery with 10 mM Tris, pH 7.4, 150 mM NaCl, 2 mM MgCl₂ and 0.5 mM EGTA. The lungs were isolated and intratracheally instilled with 10-15 ml of ice-cold DNase I-free HB buffer containing 0.25 M sucrose. In all subsequent steps, HB buffer was used free of DNase I. The tracheas were then tied with surgical suture, and the lungs were placed on ice until homogenization. All subsequent steps were performed at 0-4 °C. The lungs were cleared of extraneous tissue and homogenized in a Tekmar tissue homogenizer (3 × 20 s, two-thirds of maximum speed) using 25 ml of HB/lung. The homogenate was filtered over two layers of gauze and centrifuged at 5,000 × g for 25 min. The post-nuclear supernatant was centrifuged at 45,000 × g for 2.5 h over a cushion of 50% sucrose. The membrane pellet was washed once in 50 ml of HB buffer, resuspended in 15 ml of HB without sucrose and allowed to swell for several hours to overnight. Lung membranes at a concentration of 3.5 mg/ml were solubilized on ice for 15 min by addition of 25% Triton X-100 to 1% final concentration, and Triton-insoluble membranes were collected by centrifugation at 45,000 × g over a cushion of 50% sucrose for 1 h. The insoluble fraction was further enriched in SP-A receptor by dilution to a final protein concentration of 1 mg/ml in the presence of 1 M KCl and 1% Triton X-100. This treatment removed peripheral membrane proteins and membrane-associated myosin. After 15-30 min on ice, the mixture was centrifuged for 1.5 h at 45,000 × g over a layer of HB containing 50% in sucrose. The pellet that passed through the sucrose layer containing the SP-R210-enriched fraction was resuspended in 3 ml of sucrose-free HB. To obtain polyclonal antibodies, the SP-R210 enriched extract was further purified by SDS-PAGE and its position identified by ligand blot analysis with bSP-A. The SP-R210 band was excised from the polyacrylamide gel and injected into rabbits in incomplete Freund's adjuvant according to standard protocols. The rabbit was boosted with antigen twice at 2 and 8 weeks after the initial injection.

Other Methods

Methods for immunoprecipitation, phosphatidylcholine secretion, and collagenase digestion were performed as described previously (34, 35, 36). Protein concentration was measured using the BCA assay with BSA as the protein standard.

RESULTS

We initially used RBMM, a routine source of macrophages in our laboratory, to study the binding of SP-A to macrophages. As shown in Fig. 1 binding of SP-A to RBMM was saturable and specific. Assuming a multimeric molecular mass of 6.5 × 10⁵ for SP-A (4), analysis of the binding curve according to Klotz (37) revealed a K_d of 2.2 nM and 2.4 × 10⁵ sites/cell. Binding of SP-A occurred rapidly within the first 2 min, followed by a slower equilibration phase after 50 min (not shown). Binding of SP-A was Ca²⁺-dependent and trypsin-sensitive (Fig. 2A).

SP-A is a glycoprotein that contains carbohydrate groups with both high mannose and complex oligosaccharide moieties and it also binds to carbohydrates via its CRD domain (3, 38) in a Ca²⁺-dependent manner. Since other studies have suggested that the interaction of SP-A with macrophages is mannose-dependent (24, 25), we examined both the role of the mannose receptor in the interaction of SP-A with RBMM and binding of SP-A to surface carbohydrates. The mannose receptor ligands HRP and mannan failed to inhibit binding of SP-A to RBMM at all concentrations tested (Fig. 2B). Ovalbumin, Man-BSA, and monosaccharide ligands of the mannose receptor also did not inhibit binding of SP-A (not shown). Similar SP-A binding characteristics were obtained with U937 cells, a monocytic cell-line that does not express the mannose receptor (not shown). These findings demonstrate that the macrophage mannose receptor is not involved in SP-A binding. Ligands for other carbohydrate-specific macrophage receptors such as sialic acid, mannose 6-phosphate, -glucan, galactose, and hyaluronan and sulfated carbohydrates (39, 40) also did not inhibit the binding of SP-A (not shown). Maclura pomifera agglutinin, a lectin that interacts with cell-surface carbohydrates (41) did not inhibit SP-A binding (not shown) and HRP, a known ligand for the CRD domain of SP-A (42) also did not inhibit SP-A binding (Fig. 2B). Thus, neither the glycoconjugates of SP-A nor the carbohydrate binding activity of SP-A are necessary for SP-A binding to macrophages. Finally, the macrophage scavenger receptor ligands fucoidan and polyinosinic acid (43) also did not inhibit binding of SP-A indicating that SP-A accumulating in a potentially oxidative environment in alveolar proteinosis is not subject to clearance by the scavenger receptor (not shown).

The complement component C1q and SP-A are structurally homologous proteins (15). Both contain an amino-terminal collagen-like domain and a carboxyl-terminal globular domain. There are at least three different receptors for C1q that are widely distributed among cells of lymphoid and myeloid lineage (30, 44, 45). The studies of Tenner et al. (16) showed that SP-A could substitute for C1q in the enhancement of particle phagocytosis by monocytes. However, it was not clear whether SP-A and C1q shared the same receptor. The findings of Malhotra et al. (22, 23) suggested that SP-A and C1q shared the same receptor in U937 cells since iodinated SP-A bound to U937 cells in a C1q-like manner (22) and excess unlabeled SP-A inhibited the interaction of a purified C1q lymphocyte receptor with C1q in vitro (23). In Table I the characteristics of SP-A and C1q binding to RBMM are compared to clarify the role of the C1q receptor in SP-A binding. In contrast to SP-A binding, the binding of C1q was markedly increased by low ionic strength as previously observed (46). Trypsin and EDTA reduced the binding of SP-A by 84% while the binding of C1q was reduced only by 30-40%. The relative resistance of C1q binding to protease digestion has been previously observed (44, 46). Competition with excess unlabeled SP-A and C1q inhibited the binding of the respective ligands by 84-87% (Table I). However, cross-competition with excess unlabeled C1q appeared to enhance SP-A binding 2-fold, while excess unlabeled SP-A inhibited the binding of C1q by 50%. One possible explanation for these results is that interaction of SP-A with C1q glycoconjugates enhanced SP-A binding to the cells while at the same time inhibited the interaction of C1q with its receptor, likely due to steric hindrance. Although we cannot exclude binding of SP-A to the C1q receptor on macrophages, our results strongly suggest that SP-A and C1q have distinct binding sites on the surface of macrophages.

Table I. Comparison of SP-A and C1q binding to rat bone marrow macrophages Binding of radioactive ligands to cells was performed as described under ``Experimental Procedures.'' For all experiments 0.5 µg of labeled ligand was used. Competition experiments were performed in the presence of a 100-fold molar excess of unlabeled ligand. Binding experiments with C1q were performed in binding buffer adjusted to 0.1 M NaCl unless indicated otherwise. Each value is the mean ± S.D. from triplicate determinations and is representative of values from two to three separate experiments. Inhibitor Ligand bounda 125I-SP-A 125I-C1q ng None (0.15 M NaCl) 14.1  ± 1.56 1.26  ± 0.43 None (0.1 M NaCl) 14.7  ± 0.89 24.5  ± 1.8 Unlabeled SP-A 1.84  ± 0.51 11.82  ± 0.38 Unlabeled C1q 31.1  ± 8.02 3.58  ± 0.01 EDTA (1.5 mM) 2.25  ± 0.49 15.61  ± 2.66 Trypsin (8 µg/ml) 2.1  ± 0.36 18.22  ± 3.71 a Similar results were obtained with U937 cells.

We initially sought to determine the presence of an SP-A receptor on RBMM and U937 cell extracts by ligand blot analysis. As shown in Fig. 3, both RBMM and U937 cells express a high molecular mass SP-A binding protein of approximately 210 kDa (SP-R210). The binding was abolished in the presence of excess unlabeled SP-A and in the presence of EDTA (not shown).

Since both RBMM and U937 cells express a similar size protein, we sought to purify the SP-R210 from U937 cells since these cells can be grown in large numbers in culture. We developed an affinity method of noncovalently immobilized ligand since covalently immobilized SP-A on CNBr-activated Sepharose was inactive in binding SP-R210 (not shown). SP-A was bound to maltose-Sepharose via its CRD binding site. In our initial efforts to purify SP-A binding proteins we attempted small scale experiments using cell-surface iodinated cells. As shown in Fig. 4A, a protein of approximately 155 kDa bound to the affinity column. This protein, however, was degraded to lower molecular mass fragments within a day after isolation. Subsequently, we determined that the successful isolation of the receptor was critically dependent on the presence of the cell-surface endopeptidase inhibitors phosphoramidon and CBBL (47). As shown in Fig. 4B, in the presence of phosphoramidon a major protein of 210 kDa bound to the affinity column in agreement with the ligand blot analysis (SP-R210). Under conditions of low ionic strength, additional components of 66 kDa and 45 kDa components also bound to the affinity column (Fig. 4B). These proteins were not present in normal ionic strength conditions (not shown). The 66- and 45-kDa components observed in low ionic strength were similar in size and elution behavior to C1q binding proteins previously observed by Erdei and Reid (44) in U937 cells indicating that immobilized SP-A may also interact with C1q receptors. SP-R210 was eluted in the presence of 60 mM octylglucoside and 0.5 M salt at acidic pH 6.3 in order to render the SP-A insoluble on the affinity matrix and selectively elute the receptor. As shown in Fig. 5 a single protein of approximately 210 kDa was eluted under these conditions. Approximately 66 µg of SP-R210 were isolated from 90 mg of solubilized membrane protein from 4.3 × 10⁹ U937 cells. Our purification represents an approximate yield of 22% based on an estimated 2.4 × 10⁵ binding sites/cell.

Since both SP-A and the mannose receptor (MR) belong to the Ca ²⁺-dependent lectin family of proteins (3, 30), we hypothesized that antibodies to the MR would cross-react with the CRD domain of SP-A. As shown on the first lane of Fig. 6, A and B, both SP-A and MR antibodies bind to the monomeric and dimeric forms of alveolar proteinosis SP-A (36). Treatment of SP-A with collagenase generates collagenase-resistant fragments with reduced molecular masses of 42 and 21 kDa that represent the CRD domain of SP-A (36). As shown on the second lane of Fig. 6, A and B, both antibodies recognize the 42- and 21-kDa fragments (lanes 1-4).

In the present study we have used MR antibodies and an immunoligand blot analysis procedure in an preliminary effort to assess the mechanism of SP-A binding to its receptor. Under nonreducing conditions SP-R210 migrates near the top of a 5-20% gradient gel indicating a disulfide-dependent multisubunit structure (not shown). The unreduced SP-R210 multimer was blotted on nitrocellulose and was incubated with unlabeled SP-A. As shown in Fig. 6C the SP-A*SP-A receptor complex was detected using the MR antibodies suggesting that at least a portion of the CRD domains in SP-A do not interact with the receptor implicating the collagen-like domain of SP-A in binding to the receptor.

As shown in Fig. 7, binding of SP-A to alveolar macrophages and type II epithelial cells was also saturable and specific. As with RBMM, the binding of SP-A to RAM and type II cells was also Ca²⁺-dependent (Fig. 7) and trypsin-sensitive (not shown). However, in contrast to RAM, the Ca²⁺-independent binding of SP-A to type II cells represented 50% of total binding indicating the presence of additional binding sites for SP-A on type II cells (Fig. 7A). As shown in Table II the affinity of SP-A binding to RAM (1.8 ± 0.02 nM) was 4-fold higher than type II cells (7.7 ± 0.3 nM) and was similar to the affinity constant with RBMM (2.2 nm). Our estimated K_d for SP-A binding to type II cells is similar to the value reported by Suwabe et al. (48) of approximately 7.1 nM (4.6 µg/ml) using freshly isolated type II cells in suspension. Earlier studies (4, 5), however, reported a much higher affinity of 0.5-1.02 µg/ml with adherent type II cells indicating that the physical state of the cells may be important for SP-A binding in vivo. The number of SP-A binding sites per cell was significantly higher in RAM (9.7 ± 2.9 × 10⁵) than in both type II cells (6.7 ± 1.9 × 10⁵) and RBMM (2.4 × 10⁵). The number of SP-A binding sites on type II cells was previously estimated at between 1.4 × 10⁵ and 3.1 × 10⁵ sites/cell using suspended or adherent cells, respectively (5). Although measurements of binding constants are inherently dependent on the binding assays, it is clear from our observations and those of others that alveolar cells contain an abundant pool of high affinity binding sites for SP-A. It remains to be determined whether differences in the binding constants between macrophages and type II cells observed in vitro are physiologically relevant.

Table II. Binding constants of 125I-SP-A binding to cells Binding of 125I-SP-A to cells was performed as described under ``Experimental Procedures.'' To calculate binding constants, binding data for respective cell types in Figs. 1, 7, and 8 were analyzed according to Klotz (37). Each value was the mean ± S.D. from separate experiments as indicated. Cell type Kd Binding sites nM × 105/cell Type II (n = 4) 7.7  ± 0.3 6.7  ± 1.9 RAM (n = 4) 1.8  ± 0.02 9.7  ± 2.9 RBMM (n = 1) 2.2 2.7

We next attempted to identify and purify the SP-A binding site(s) from rat lung membranes. In initial studies using ligand blot analysis we determined that lung membranes also contained a high molecular mass SP-A binding protein of approximately 210 kDa (not shown). However, we also found that the putative lung SP-R210 was insoluble in Triton X-100. As such, we could not apply the affinity method used for U937 cells. Instead, we obtained a membrane preparation enriched in lung SP-R210 by differential detergent and high salt extraction. As with U937 cell isolations, the presence of membrane endopeptidase inhibitors was critical for the stability of the lung SP-R210 (see methods). As shown in Fig. 8 (lane 1), a major protein of about 210 kDa was enriched in the detergent and salt membrane extracts. The 210-kDa protein was an SP-A-binding protein as determined by ligand blot analysis using bSP-A (Fig. 9, lane 1) and was also present on type II and macrophage cell extracts (Fig. 9, lanes 2, 3, and 4). Another major component of the membrane extracts at about 43 kDa (Fig. 8) was determined to be actin by immunoblot analysis (not shown). Using amino acid analysis, our enriched lung membrane preparation from the lungs of 10 rats contained 200 µg of SP-R210.

Polyclonal antibodies against the gel-purified SP-R210 were raised in rabbits (see ``Experimental Procedures''). The SP-R210 antibody recognized a 210-kDa protein on detergent and salt extracts of isolated rat lung membranes (Fig. 8, lane 2) as well as on cell extracts from RBMM, U937, RAM, and type II cells (Fig. 10). In addition, the antibody specifically immunoprecipitated a 210-kDa protein from metabolically labeled U937 cells and RBMM (Fig. 11). In contrast to RBMM and U937 cell extracts, SP-R210 was resistant to detergent solubilization in alveolar macrophages and type II cells (Fig. 10). The basis and physiological significance of detergent insolubility of SP-R210 in the lung is not yet known.

Antibodies to SP-R210 inhibited binding of radiolabeled SP-A to RAM and type II cells by 80 and 70%, respectively (Fig. 12). Similar findings were observed with RBMM. SP-A is known to inhibit phospholipid secretion from type II cells, and this process is thought to be receptor-mediated (7). In Table III the role of SP-A and its receptor in TPA-stimulated surfactant secretion was tested using the SP-A receptor antibody. The antibody blocked the SP-A-mediated inhibition of phosphatidylcholine secretion in TPA-stimulated cells, suggesting a role for SP-R210 in the regulation of surfactant homeostasis. Similar results were obtained using rat SP-A (not shown). Interestingly, the presence of anti SP-R210 antibodies alone enhanced the secretion of phosphatidylcholine in both stimulated and unstimulated cells, suggesting that the antibody may inhibit the effects of endogenous SP-A (Table III).

Table III. Role of SP-A receptor in phospholipid secretion Release of [3H]PC from prelabeled type II cells (35) was determined 3 h after exposure to 100 nM TPA in the presence or absence of 10 µg/ml SP-A. In experiments using polyclonal antisera, the cells were incubated with preimmune antiserum (PI) or SP-A receptor antiserum (aSP-R210) for 30 min at a dilution of 1:50 followed by treatment with TPA in the presence or absence of SP-A. Data are average values ± S.D., n = 12.  Treatment [3H]Phosphatidylcholine secretion % Control 0.7  ± 1.1 TPA 4.3  ± 1.3a TPA + SP-A 2.1  ± 1.1b TPA + SP-A + PI 2.8  ± 0.9c TPA + SP-A + aSP-R210 7.1  ± 2.3d,e aSP-R210 4.2  ± 1.2f,g TPA + aSP-R210 5.3  ± 2.0h a p > 0.0001 compared to control. b p > 0.0001 compared to TPA. c p > 0.004 compared to TPA. d p > 0.0001 compared to TPA. e p > 0.0001 compared to TPA + SP-A + PI. f p > 0.0001 compared to control. g p > 0.0001 compared to TPA + SP-A + aSP-R210. h p > 0.0001 compared to control.

Recent studies in our laboratory have demonstrated that SP-A promotes the attachment and uptake of BCG by RBMM and human monocytes. The SP-A-mediated enhancement of BCG uptake by macrophages was blocked in the presence of a 1:100 dilution of the polyclonal antibodies against the SP-A receptor (60).³ Based on these findings we can conclude that alveolar type II cells and macrophages express the same or a similar receptor for SP-A.

DISCUSSION

The ability of SP-A to alter the function of macrophages and type II cells suggests that these cells express specific cell-surface SP-A receptors. Using in vitro binding assays and ligand blot analysis we identified a distinct 210-kDa membrane protein (SP-R210) on macrophages, type II cells and isolated rat lung membranes. The SP-R210 was purified to apparent homogeneity from U937 macrophage membranes and rat lung membranes. Polyclonal antibodies against the rat lung SP-R210 inhibited the Ca²⁺-dependent binding of SP-A to type II cells and macrophages indicating that SP-R210 is expressed on the cell-surface. The ability of antibodies against SP-R210 to inhibit SP-A function demonstrate that SP-R210 is a cell-surface receptor involved in the SP-A-mediated regulation of surfactant homeostasis and phagocytosis of pathogens.

The mechanism of SP-A binding to cells is not completely understood. In the present work we were able to purify the SP-A receptor from macrophages using SP-A immobilized via its carbohydrate binding site. Furthermore, we could detect the purified receptor by immunoligand blotting using SP-A and antibodies specific to the CRD domain of SP-A (Fig. 6) suggesting that domains other than the carbohydrate binding site of SP-A are important for binding to the receptor. In agreement with previous findings (18), we also observed that collagen V inhibited SP-A binding to macrophages suggesting that the collagen-like domain is involved in SP-A binding (not shown). We also determined that carbohydrate-specific receptors were not involved in SP-A binding (Fig. 3) to macrophages. In agreement with our conclusions, earlier findings (28, 49, 50) indicated that the glycoconjugates of SP-A and the carbohydrate binding site of SP-A were not required for SP-A binding to type II cells.

Studies with mutant recombinant SP-A (20, 21) and recombinant SP-A (24, 25) have at the same time shed some light into the mechanism of SP-A binding to cells and introduced new discrepancies in conclusions derived with native SP-A molecules. Using glycosylation-deficient mutants McCormack et al. (20) confirmed that SP-A glycosylation is not required in cell binding. Recombinant rat SP-A deficient in the hydroxylation of prolines oligomerized to a lesser extend than native SP-A and had a significantly reduced ability to compete with native SP-A in binding to type II cell membranes, to inhibit phospholipid secretion and to aggregate liposomes (20), indicating that the intact collagen-like domain participates in SP-A function. Using epitope mapping analysis with monoclonal antibodies, Kuroki et al. (21) determined that the ability of SP-A to inhibit phospholipid secretion by type II cells resided in the CRD domain of SP-A. In contrast to earlier conclusions that the CRD binding site is not involved in SP-A binding, however, mutation of Glu¹⁹⁵ and Arg¹⁹⁷ to Gln¹⁹⁵ and Asp¹⁹⁷ in the CRD binding site of hydroxyproline-deficient SP-A converted SP-A to a galactose-specific lectin with lower affinity for type II cells and decreased ability to inhibit phospholipid secretion (51). Schlepper-Schaffer and colleagues (24, 25) found that binding of recombinant SP-A to macrophages was partially mannose-dependent. The recombinant SP-A used in their studies was generated in Chinese hamster ovary cells transfected with a vector containing the ₃ gene of SP-A (24, 25) that was shown to form only partial SP-A oligomers and was inefficient in enhancing phagocytosis of bacteria compared to purified human alveolar proteinosis SP-A. Interestingly, the recombinant SP-A generated using both ₂ and ₃ SP-A genes partially restored the ability of SP-A to direct phagocytosis of bacteria (18). Thus, it appears that the state of SP-A aggregation may be an important determinant in regulating the mechanism of SP-A binding to cells. In addition, binding of SP-A is Ca²⁺-dependent. Additional characterization of mutant SP-A molecules with respect to these parameters will further the understanding of SP-A binding to cells.

The novel SP-A receptor that we have characterized in the present report is different from previously identified SP-A-binding proteins. Using anti-idiotype antibodies Strayer et al. (52) have identified a 30-kDa protein that is present on both type II cells and on ciliated cells of conducting airways, but not on macrophages. Employing a similar approach Stevens et al. (53) have recently identified a 170-200-kDa SP-A-binding protein that is composed of 55-kDa subunits under reducing conditions. This protein is restricted to type II cells. The functional role of these proteins is not yet known. The SP-A receptor we report here has a reduced molecular mass of 210 kDa that oligomerizes to a higher molecular mass form under nonreducing conditions. In contrast to the two previous studies the SP-R210 is localized on both type II cells and macrophages.

The SP-A receptor is also distinct from the C1q receptor based on different binding characteristics between SP-A and C1q (Table I) although SP-A may interact with the C1q receptor (Fig. 5B). Modification of SP-A may promote SP-A binding to the C1q receptor. Malhotra et al. (22) reported similar binding characteristics between SP-A and C1q in U937 macrophages. In these studies iodinated SP-A with a high specific activity of 4.3 × 10⁷ cpm/µg was used. In our hands, iodinated SP-A with specific activities above 1 × 10⁶ cpm/µg led to SP-A binding that was not inhibitable by EDTA or excess unlabeled SP-A (not shown), indicating that modification of critical tyrosine residues may promote binding of SP-A to the C1q rather than the SP-A receptor. Interestingly, Geertsma et al. (54) reported that binding of SP-A labeled with fluorescein isothiocyanate (a lysine modification) occurred via the monocyte C1q receptor and was inhibited by C1q but not by SP-A. This fluorescein isothiocyanate-SP-A directed the phagocytosis of Staphylococcus aureus via the monocyte C1q receptor (54). The importance of amino groups for SP-A binding to type II cells was initially noted by Kuroki et al. (50) with reductively methylated SP-A. Biotinylation of SP-A at pH 8.0, another lysine modification, blocked the ability of SP-A to inhibit phospholipid secretion (28). However, when SP-A was biotinylated at pH 6.3, a condition that promotes biotinylation primarily at the amino-terminal amino group, the function of SP-A was preserved suggesting that internal lysine residues are important for SP-A binding to cells. Thus, chemical modification of SP-A with reporter molecules must be considered as a critical variable in studies of SP-A function.

The ability of SP-R210 polyclonal antibodies to block SP-A-dependent inhibition of phospholipid secretion in TPA-stimulated type II cells suggests that the SP-A receptor may interact with the protein kinase C signaling pathway. At least three different signaling pathways are known to stimulate phospholipid secretion in type II cells which include cAMP, protein kinase C, and phosphatidyl inositol-dependent phospholipase C-dependent pathways (55). One downstream pathway in surfactant secretion is thought to involve mobilization of actin (56) and actin phosphorylation (57) in response to stimulation of the cAMP-dependent signaling pathway. Our finding that SP-R210 may associate with the actin cytoskeleton could lead to mechanisms for SP-A-mediated inhibition of surfactant secretion by type II cells and for SP-A-mediated phagocytosis of pathogens.

In the present report we have shown that SP-R210 is a functional SP-A receptor. Binding of SP-A to SP-R210 leads to inhibition of TPA-stimulated secretion of phosphatidylcholine by type II cells. In a separate report it is shown that SP-A binds to BCG and the complex interacts with SP-R210 on the surface of macrophages (60).³ The effect of SP-A on stimulated surfactant secretion and in macrophage-pathogen interaction has been previously documented (7, 12, 16, 17, 18, 19). Additionally, SP-A is known to stimulate type II cells and macrophages to secrete granulocyte/macrophage-colony-stimulating factor (10), an important cytokine in lung surfactant homeostasis (58), and to promote chemotaxis, release of reactive oxygen intermediates and clearance of pathogens by macrophages (12, 16, 17, 18, 19). SP-A is also known to induce lymphocytes to secrete proinflammatory cytokines (59) suggesting a role of SP-A in cell-mediated immunity. We have already determined that SP-A binding to macrophages is up-regulated both in vitro and in vivo by activating agents (lipopolysaccharde, phorbol 12-myristate 13-acetate, interferon-), but down-regulated by dexamethasone, suggesting that the expression of the SP-A receptor depends on the functional state of the macrophage (30). We do not yet know whether modulation of SP-A binding by these agents is due to regulated expression of SP-R210 or another yet unidentified SP-A receptor. The availability of purified receptor and polyclonal antibodies obtained in the present work will be crucial in understanding the structure and function of SP-R210, in cloning the SP-A receptor and in understanding the mechanism of SP-A function.