Surfactant protein A binds Mycoplasma pneumoniae with high affinity and attenuates its growth by recognition of disaturated phosphatidylglycerols.

Surfactant Protein A (SP-A) is an abundant, multifunctional lectin that resides within the bronchoalveolar compartment of the lung and plays an important role in the innate immunity of the organ. Mycoplasma pneumoniae is a human pathogen that resides in the same compartment as SP-A, and we examined the interaction between the two. Preparations of human and rat SP-A recognized the mycoplasma with high affinity in the presence of Ca(2+), exhibiting apparent K(')(d) values in the nanomolar range. Membranes prepared from the microbe also bound human and rat SP-A with similar characteristics and affinity to the intact cells. The ligand for SP-A was insensitive to proteolysis. Lipid extracts prepared from the mycoplasma, bound SP-A with high affinity when examined by ligand blot analysis. These lipid extracts were also potent competitive inhibitors (IC(50) = 0.2 nM) of human SP-A binding to mycoplasma membranes. The major lipid ligands for the protein identified by mass spectrometry are a group of disaturated phosphatidylglycerols. The addition of SP-A to cultures of M. pneumoniae markedly attenuated the growth of the organism assessed by colony formation, metabolic activity, and DNA replication. The bacteriostatic effects of SP-A were reversed by dipalmitoylphosphatidylglycerol. These findings demonstrate that human SP-A can play a direct role in antibody-independent immunity to M. pneumoniae by interacting with lipid ligands expressed on the surface of the organism and implicate SP-A in the immediate host response to the bacteria.

and proteins that reduces surface tension in the alveoli during expiration (2). A growing body of evidence demonstrates that SP-A plays an important role regulating the innate host defense system within the lung (3). Functional deletion of the SP-A gene from the mouse genome provides important evidence establishing the regulatory role of the protein in innate immunity in vivo (4 -6). The interaction of SP-A with immune cells can occur through ligation of a number of cell surface receptors that include the lipopolysaccharide-binding protein, CD14 (7), a 210-kDa protein, SPR 210 (7), a complement binding protein, C1qR (8), CD91/calreticulin (9), and SIRP␣ (10). In addition to recognition of immune cells, the protein can bind a variety of pulmonary pathogens, including Pseudomonas, Escherichia, Mycobacterium, Mycoplasma, Candida, Aspergillus, Histoplasma, and Influenza species (summarized in Ref. 3). For some organisms, SP-A functions as an opsonin and enhances phagocytosis (11,12). The interactions of SP-A with microorganisms and host cells are complex and incompletely understood. In some cases the ligation of microbes (or derived ligands) and immune effector cells leads to amplification of inflammatory responses (13,14), whereas in others the inflammatory responses are markedly attenuated (7,15,16). The emerging picture is consistent with the nature of the foreign ligand and the predisposed state of the responding immune cell, both playing important roles in dictating the final inflammatory response of the host. In addition to modulating the actions of inflammatory cells SP-A and SP-D can also exert a direct antimicrobial effect upon several Gram-negative bacteria (17,18) and Histoplasma capsulatum (18) by increasing membrane permeability. The effects upon Gram-negative cells appear to be mediated via surface lipopolysaccharide binding. The surface-binding molecule for Histoplasma is not known but could be a glycoconjugate.
In this report we have focused on the interactions of SP-A with Mycoplasma pneumoniae. Unlike Gram-negative bacteria mycoplasma lack lipopolysaccharides and are likely to have novel ligands for SP-A. M. pneumoniae is an important human pathogen that causes primary atypical pneumonia and other airway diseases, including tracheobronchitis and pharyngitis. Infection by the organism can also seriously exacerbate asthma (19,20). Recent evidence also demonstrates the presence of the organism in a subset of chronic asthmatics suggesting that it may have a contributing role in the etiology of the disease (21). Several studies have previously implicated SP-A as a modulator of macrophage-dependent killing of the mouse pathogen M. pulmonis (22)(23)(24), but little current information is available regarding the direct interactions of SP-A with the human pathogen M. pneumoniae. The purpose of this study was to 1) elucidate the interaction between SP-A and M. pneumoniae, 2) determine if membranes derived from the organism could also interact with SP-A, 3) characterize the types of mycoplasma ligands recognized by SP-A, and 4) evaluate the direct effects of the protein on mycoplasma growth. Our findings demonstrate high affinity interactions between SP-A and M. pneumoniae, dependent upon disaturated phosphatidylglycerols present in the membrane. In addition, SP-A directly inhibits the growth of the bacteria, and this action is reversed by disaturated phosphatidylglycerols. These findings demonstrate that SP-A plays an important role in controlling the antibodyindependent host response to mycoplasma.

EXPERIMENTAL PROCEDURES
Purification of SP-A and Antibodies-Human SP-A was purified from alveolar proteinosis fluid as described previously (25). Rat SP-A was purified from CHO-K1 cells or Sf9 cells expressing the recombinant protein, by published methods (26). Polyclonal antibodies to the human and rat proteins were independently generated in rabbits, and the IgG fractions were isolated using protein A-Sepharose. For ELISA, the primary antibodies were conjugated to horseradish peroxidase.
Purification of M. pneumoniae-M. pneumoniae cells were grown in SP-4 medium (27) at 37°C in an atmosphere of 95% air, 5% CO 2 . The cells were harvested by centrifugation at 8,000 ϫ g ϫ 15 min at 4°C and washed three times with phosphate-buffered saline (PBS). The resultant pellet was resuspended in 2 ml of PBS and layered upon a four-step sucrose gradient consisting of 5 ml of 60%, 7 ml of 52%, 7 ml of 48%, and 7 ml of 42% sucrose. The gradient was centrifuged in an SW28 rotor at 20,000 ϫ g av ϫ 15 min. Cells were harvested from the 48/52% interface, diluted in PBS, and washed free of sucrose by centrifugation. The viability of the cells assessed by colony counting was essentially identical before and after the gradient purification. The purification of the organisms by density gradient centrifugation was essential, because large amounts of protein precipitate co-sediment with mycoplasma in the initial centrifugation step and interfere with binding measurements and quantification of microbial protein content.
M. pneumoniae Membrane Preparation-Mycoplasma membranes were prepared by mixing 1 ml of gradient-purified bacteria resuspended in PBS, with 3 ml of distilled water and incubation at 4°C for 30 min. The hypotonically swollen cells prepared in the previous step were disrupted by six 30-s bursts of sonication on ice with each burst interrupted by a 30-s cooling period. The sonicated mixture was centrifuged at 100,000 ϫ g at 4°C for 1 h. The pellet was resuspended in PBS, homogenized, and centrifuged at 100,000 ϫ g at 4°C for 1 h. The final pellet was resuspended in PBS and homogenized, and aliquots were frozen at Ϫ20°C prior to use.
Direct Binding of SP-A to M. pneumoniae in Solution-Human or rat SP-A at various concentrations was mixed with mycoplasma (40 g/ml total cell protein) in 50 l of buffer A (20 mM Tris, pH 7.4, 150 mM NaCl) supplemented with either 5 mM CaCl 2 or EGTA and 2% w/v bovine serum albumin (BSA). The binding was carried out for 30 min at room temperature. The reaction was layered on 300 l of buffer A with CaCl 2 containing 3% BSA and centrifuged at 10,000 rpm for 10 min at 4°C. The cell pellets were resuspended in buffer A with CaCl 2 , before relayering on buffer A with CaCl 2 containing 3% BSA, and again centrifuging. The bound SP-A was eluted from the cell pellets in 100 l of buffer A containing 5 mM EGTA and 2% BSA, at room temperature, for 30 min. The eluted material was centrifuged at 10,000 rpm for 10 min at 4°C. Eluted SP-A present in the supernatant was quantified by ELISA using polyclonal antibodies against either human or rat SP-A.
Binding of SP-A to M. pneumoniae Membranes on Solid Phase-Mycoplasma membranes (250 ng of membrane protein in 100-l aliquots) were adsorbed to microtiter wells in 0.1 mM NaHCO 3 , pH 9.6, at 4°C overnight. Following adsorption, the wells were blocked with buffer A containing 2% BSA for 1 h at 37°C. Human and rat SP-A at various concentrations were added to each well in buffer A with 2% BSA and either 5 mM CaCl 2 or EGTA as indicated. The binding reaction was performed for 1 h at 37°C. Unbound protein was removed by washing the plate twice with buffer A containing 0.1% BSA and 5 mM CaCl 2 or EGTA. The bound SP-A was detected by ELISA using 10 g/ml HRPconjugated primary antibody against either human or rat protein. Orthophenylenediamine (1 mg/ml) was used as the color development reagent for detecting the bound HRP-conjugated antibody.
Binding of SP-A to Trypsin and Proteinase K-digested M. pneumoniae Membranes in Solid Phase-M. pneumoniae membranes were digested with either 1 mg/ml Trypsin or 1 mg/ml Proteinase K at 37°C for 1 h. Trypsin activity was quenched with 1 mg/ml soybean trypsin inhibitor, and proteinase K activity was quenched with 1 mg/ml phenylmethylsulfonyl fluoride. The protease-treated and untreated control membranes were centrifuged at 100,000 ϫ g at 4°C for 1 h. The total lipid phosphorus of digested membranes was determined (28) and used to quantify recovery. Membranes containing 1 nmol of lipid phosphorus were coated on each microtiter well in 0.1 mM NaHCO 3 , pH 9.6, at 4°C overnight. Aliquots containing 630 ng/ml human and rat SP-A were added to each well in buffer A containing 2% w/v bovine serum albumin and either 5 mM CaCl 2 or EGTA as indicated. The binding was carried out for 1 h at 37°C. The bound SP-A was detected by 10 g/ml HRP-conjugated antibody against either human or rat SP-A and quantified by ELISA with 1 mg/ml orthophenylenediamine as the color-developing agent.
Binding of SP-A to Lipids Extracted from M. pneumoniae Membranes Separated by Two-dimensional Thin Layer Chromatography-Lipids were extracted from M. pneumoniae membranes (29) and separated by two-dimensional thin layer chromatography on SIL G plates (Macherey Nagel). Total lipids (equivalent to 40 nmol of lipid phosphorus) were applied to the plate and separated in a first dimension consisting of chloroform:methanol:NH 4 OH (65:35:8). The plate was next air-dried and exposed to vapor from methanol:acetic acid (90:10) for 10 min to neutralize the NH 4 OH. The second dimension was developed in chloroform:methanol:acetic acid:water (50:25:8:2.5). The plate was air-dried overnight. Lipid components were visualized by either 0.2% w/v 8-anilino-1-naphthalenesulfonic acid (ANSA) staining or orcinol staining for glycolipids. To determine binding of SP-A to the lipid, purified protein (10 g) was suspended in 1 ml of buffer A containing 2% bovine serum albumin and 5 mM CaCl 2 , and added to the plate and incubated for 1 h at room temperature. The plate was washed with buffer A containing 0.1% bovine serum albumin and 5 mM CaCl 2 three times. The bound SP-A was detected by using 10 g/ml HRP-conjugated primary antibody against the protein and 1 mg/ml 3,3Ј-diaminobenzidine as the color development reagent.
Lipid Competition for SP-A Binding to Solid Phase Mycoplasma Membranes-Lipids extracted from M. pneumoniae membranes and dipalmitoylphosphatidylcholine (16:0/16:0-PtdCho, Avanti Polar Lipids, Inc.) were hydrated in 20 mM Tris, pH 7.4, 150 mM NaCl buffer at 37°C for 30 min. The lipids were probe-sonicated to make unilamellar vesicles. Aliquots containing 32.5 ng of SP-A were incubated with various concentrations of vesicles in buffer A with 2% BSA and 5 mM CaCl 2 at room temperature for 30 min. The mixtures were added to M. pneumoniae membranes (250 ng of total protein), coated onto microtiter wells, and incubated for 1 h at 37°C. The plate was washed with buffer A containing 0.1% w/v bovine serum albumin and 5 mM CaCl 2 three times. The SP-A bound to membranes was detected with 10 g/ml HRP-conjugated primary antibody against the protein and quantified by ELISA.
Inhibition of M. pneumoniae Growth by SP-A-We evaluated the effect of SP-A on M. pneumoniae growth by colony counting, culture acidification, and [ 3 H]thymidine incorporation. For the colony counting assay, 1-10 g/ml SP-A was incubated with 10 4 cells/ml M. pneumoniae in SP-4 medium (27) at 37°C for 1-5 days. For each day, an aliquot of the culture was spread on a pleuropneumonia-like organism (PPLO) plate. The PPLO plates were further incubated for another 7 days at 37°C, and the colonies were counted. In the acidification assay, varying concentrations of SP-A were incubated with 10 4 cells/ml of M. pneumoniae in SP-4 medium, in microtiter wells, for 5 days at 37°C with 5% CO 2 . The changing color of SP-4 medium due to acidification was measured by spectrophotometry at a wavelength of 550 nm. In the radioactivity assay, cells (10 4 or 10 3 cells/ml) were inoculated into SP-4 medium supplemented with [ 3 H]thymidine (1 Ci/ml) and incubated 5 or 6 days. The cultures were harvested by centrifugation at 10,000 ϫ g ϫ 10 min and washed by recentrifugation twice in PBS containing 1% BSA. The final pellets were resuspended in PBS with 1% Triton X-100, and the radioactivity was quantified by liquid scintillation spectrometry. We confirmed that the [ 3 H]thymidine was incorporated into macromolecular DNA by precipitation with 5% trichloroacetic acid. The acid treatment of mycoplasma revealed that Ͼ90% of the [ 3 H]thymidine was precipitable. The antagonism of the SP-A effect upon mycoplasma growth by lipids was measured by adding 50 nmol of the total lipid extract from the bacteria or 50 nmol of dipalmitoylphosphatidylglycerol to the cultures and measuring the effects upon [ 3 H]thymidine incorporation.
HPLC and Mass Spectrometry of Mycoplasma Lipids-The spot identified on TLC as the major ligand for SP-A was scraped from plates and extracted from the silica gel by the method of Bligh and Dyer (29). The resultant preparation was subjected to reversed phase HPLC to purify phospholipid molecular species and was carried out on a 150 ϫ 2.0-mm Columbus C18 column (Phenomenex, Rancho Palos Verdes, CA). The HPLC analysis used two solvents. Solvent A contained methanol/water (20:80, v/v) with 1 mM ammonium acetate. Solvent B was methanol made to 1 mM ammonium acetate. The gradient was developed from 20% to 100% solvent B in 35 min and remained at 100% solvent B for 15 min. The HPLC effluent was monitored at 220 nm using a photodiode array detector. 1-min fractions were collected from the HPLC and subjected to ligand blot analysis by TLC and direct mass spectrometric analysis. The individual HPLC fractions were analyzed by electrospray ionization in the negative ion mode by infusion of each sample into an API-III ϩ tandem quadrupole mass spectrometer (Applied Biosystems, Thornhill, Ontario, Canada). Tandem mass spectrometry was carried out using argon as the collision gas at a thickness of 235 ϫ 10 13 molecules/cm 2 . Nitrogen was used as the curtain gas with a collisional offset potential of 40 eV for product ion formation.

SP-A Binds with High Affinity to M. pneumoniae Cells and
Membranes-Our initial studies examined the interaction of SP-A with the human pathogen M. pneumoniae. We devised a binding assay that utilized purification of the organisms through an albumin step gradient and detection of the bound protein by ELISA. The results presented in Fig 1A demonstrate that human SP-A binds the mycoplasma in a concentration-dependent and saturable manner. The binding shows an absolute requirement for Ca 2ϩ and is completely inhibited by EGTA. Scatchard analysis of the binding data gave an estimated 31 ϫ 10 3 binding sites per cell with an apparent KЈ d value of 7.8 nM. Similar findings were obtained with respect to the Ca 2ϩ requirement and saturability when the rat protein was examined for interaction with the mycoplasma (Fig. 1B). The rat protein bound with an apparent KЈ d value of 8 nM and an estimated 52 ϫ 10 3 binding sites per cell.
To more fully evaluate the binding of SP-A to mycoplasma we developed a solid phase binding assay using membranes adsorbed to microtiter wells. The binding of the human SP-A to the solid phase membranes also showed dependence upon Ca 2ϩ and inhibition by EGTA as demonstrated in Fig. 2. The apparent affinity of the binding was higher using the membrane system, than that found for the binding of intact organisms in solution. For human SP-A the apparent KЈ d value is 0.3 nM. In these experiments we also compared the binding of SP-A derived from normal human subjects to that from alveolar proteinosis patients. There was no significant difference in the binding activity of the SP-A from either source. Examination of the binding of the rat SP-A produced in either CHO-K1 cells or Sf9 cells gave comparable results as shown in Fig. 3. The estimated KЈ d value for rat SP-A is 0.9 nM. From these data we conclude that SP-A from both human and rat sources binds M. pneumoniae membranes with high affinity and provides a useful cell free system for biochemical dissection of the process.
The Ligand for Human SP-A Is a Lipid-The high affinity interactions of SP-A with mycoplasma membranes provided a versatile system to elucidate the nature of the ligand expressed on the organism. In these experiments we first examined the protease sensitivity of the binding. The membranes were exposed to 1 mg/ml trypsin or 1 mg/ml proteinase K for 1 h at 37°C. The resultant membranes were treated to inactivate the proteases and washed by centrifugation and adsorbed onto microtiter wells and examined for retention or loss of binding activity. The proteolysis of the membrane proteins was confirmed by gel electrophoresis and staining. We also determined that the levels of proteases added to the reactions were sufficient to completely digest amounts of BSA comparable to the amount of mycoplasma membrane protein present. The amount of mycoplasma membrane in both untreated and protease-treated samples was made equivalent with respect to their phospholipid content. The results shown in Fig. 4 dem-onstrate that proteolysis was without any significant effect upon either human or rat SP-A binding to the membranes. These findings indicate that the mycoplasma ligand is either a lipid or an extremely protease resistant protein.
To test the possibility that the ligand for SP-A is a lipid, the mycoplasma membranes were subjected to extraction with organic solvents, and the lipid fraction was examined for binding activity. The lipids were first separated by two-dimensional thin layer chromatography in parallel on four thin layer plates. Following the separation, the plates were stained for either total lipid (ANSA stain) or glycolipid (orcinol stain), and immunostained for human SP-A binding. The results of the experiment are shown in Fig. 5. The mycoplasma contains at least eight well resolved polar lipids and unresolved neutral lipids, which migrate close to the solvent front as shown in glycolipids include dihexosyldiacylglycerol (G6) and monohexosyldiacylglycerol (G7). The G3/L4 spot has the expected migration of trihexosyldiacylglycerol. Although the G5 spot co-migrates with phosphatidylglycerol, independent experiments demonstrate that phosphatidylglycerol does not react with orcinol. Thus, the G5/L6 spot contains multiple components. The G1 spot migrates with complex glycosphingolipids. The plate incubated with human SP-A and antibody (Fig. 5C) shows one major and one minor lipid ligand. The major lipid ligand comigrates with phosphatidylglycerol and an unknown glycolipid component. The control plate is shown in Panel D and demonstrates there is no antibody reactivity without the addition of SP-A. These data clearly indicate that one molecular category of high affinity ligand for SP-A in M. pneumoniae consists of lipids.
We next tested whether the lipid ligand for SP-A derived from mycoplasma was capable of inhibiting the binding of the protein to the solid phase membrane preparation. In experiments shown in Fig. 6, varying concentrations of the mycoplasma total lipid fraction, in the form of unilamellar liposomes, were added as competitor for the membrane binding reaction. The lipid fraction was able to completely inhibit the binding of SP-A to the solid phase membranes. The IC 50 for inhibition was 0.2 nM based upon lipid phosphorus. In the same experiment we also determined the competition for membrane binding by the surfactant lipid, 16:0/16:0-PtdCho, which is known to be a high affinity lipid ligand for SP-A (30). Quite remarkably, the mycoplasma high affinity ligand was a more potent inhibitor of SP-A binding to the solid phase membranes than 16:0/16:0-PtdCho. The apparent IC 50 of the mycoplasma ligand exhibited more than 10-fold greater affinity for SP-A than 16:0/16:6-PtdCho in the competitive membrane binding assay. These results clearly indicate that M. pneumoniae expresses a lipid on its surface that has extremely high affinity for human SP-A.
SP-A Attenuates the Growth of M. pneumoniae-In the early stages of this study we immediately noticed that the addition of SP-A to the mycoplasma growth medium resulted in markedly reduced acidification (color change of pH indicator) of the medium over time in culture. We quantified this effect by three different methods. Using colony counting of plated bacteria we determined that SP-A addition to the cultures reduced the colony number during log phase growth by a factor of 10. In these studies, a fixed amount of SP-A was added at the beginning of the incubation, and no subsequent additions were made. These data were consistent with SP-A initially reducing the replicating cell number in the population. Although SP-A reduced the increase in cell number in growing cultures, it did not change the viability of the cells in the initial colony count. Consistent with this observation we also found that the propidium iodide staining of mycoplasma was not increased by SP-A treatment, whereas treatment with the antibiotic nuquinolone markedly increased staining. These findings indicate that SP-A has a bacteriostatic rather than bacteriocidal effect on mycoplasma. Because of the nature of colony counting, we sought to exclude SP-A-induced aggregation of the cells as the basis for reduction of colony number in log phase cultures. We  Fig. 7, demonstrate that SP-A has a concentration dependent effect upon the metabolism of the cells. At 25 g/ml SP-A, there was a 4-fold difference in the concentration of protons produced by the cultures after 5 days of growth. Heat denaturation of the SP-A completely eliminates the inhibitory effect upon microbial growth.
The effect of SP-A on cell growth is dependent not only upon the concentration of the protein, but also on the ratio of the protein to the cell number of the culture. The preceding experiments used starting inocula of 10 4 cells/ml. In Fig. 8 we show the results of using a starting cell number of 10 3 /ml and different concentrations of human and rat SP-A. For these studies we used [ 3 H]thymidine incorporation into the macromolecular pool to evaluate the effects of the protein on replication. In these experiments the cells were grown for 6 days, which corresponds to mid log phase for the control cultures. The human SP-A markedly reduced the rate of [ 3 H]thymidine incorporation into DNA by the mycoplasma in a concentration-dependent manner. At 10 g/ml the human SP-A reduced the incorporation of radiolabel into DNA by more than 90%. In this same series of experiments we also demonstrated that rat SP-A produced in either CHO-K1 cells or baculovirus/Sf9 cells was capable of reducing the growth of the organism albeit to a lesser extent. These studies of M. pneumoniae growth clearly establish that SP-A reduces the cell number, culture metabolic rate, and replication of the organism. From the above findings we conclude that the high affinity interactions between SP-A and M. pneumoniae have potent antibacterial consequences.
The Major Lipid Ligands for SP-A in M. pneumoniae Membranes Are Disaturated PtdGros-We next sought to identify the major lipid ligands for SP-A present in mycoplasma membranes. Lipids were extracted from purified bacterial membranes and separated by TLC. The region of the plate corresponding to the major SP-A reactive component identified by FIG. 5. SP-A recognizes lipid components present in mycoplasma membranes. Lipids were extracted from mycoplasma membranes and separated by two-dimensional thin layer chromatography on 4 Merck Sil 60 plates. The first dimension was developed in chloroform:methanol:NH 4 OH (65:35:8), and the second dimension was developed in chloroform:methanol:acetic acid:water (50:25:8:5) as indicated by the arrows. In A, polar lipids were detected with 0.2% 8-anilino-1naphthalenesulfonic acid (ANSA). The L1, L3, and L6 components correspond to sphingomyelin, phosphatidylcholine, and phosphatidylglycerol, respectively. L2 co-migrates with complex glycosphingolipids, and L4 co-migrates with trihexosyldiacylglycerol. The identities of L5, L7, and L8 are not known. In B, glycolipids were detected with orcinol. G5, G6, and G7 co-migrate with phosphatidylglycerol, dihexosyl-, and monohexosyldiacylglycerol, respectively. The G3 component migrates at the expected position for trihexosyldiacylglycerol. The G1 component migrates at the position of complex glycosphingolipids. The G2 and G4 components are unknown. In C, 10 g/ml human SP-A was added to the plate in the presence of a blocking buffer, and the mixture was incubated for 1 h at room temperature. The bound human SP-A was detected with HRP-conjugated polyclonal antibody to SP-A. The minor reactive ligand designated A1 co-migrates with sphingomyelin (L1). The major reactive ligand, designated A2, co-migrates with phosphatidylglycerol (L6) and a hexose containing glycolipid (G5). In D, the control plate was treated identically to the SP-A immunoblot plate except that the SP-A was omitted. ligand blot analysis (designated A2 in Fig. 5) was scraped and extracted to isolate the lipids. The recovered lipid extract was next subjected to HPLC and mass spectrometry as shown in Fig. 9. Individual fractions from the HPLC were subjected to both TLC and mass spectrometry. The SP-A-reactive material in each fraction was detected on TLC by the ligand blot method. As shown in Fig. 9A only three fractions contained material that reacted with SP-A, and they were present in HPLC fractions 26, 29, and 32. These fractions contained different molecular species of phosphatidylglycerol (PtdGro). Mass spectrometry shown in Fig. 9B  To confirm the nature of the major ligands for SP-A identified by mass spectrometry we performed a reverse analysis by subjecting purified lipids to ligand blot reactions. In these experiments shown in Fig. 10, disaturated lipids (16:0/16:0-PtdGro, 16:0/18:0-PtdGro, and 18:0/18:0-PtdGro) and monounsaturated lipids (16:0/18:1-PtdGro and 18:0/18:1-PtdGro) were separated by TLC and subjected to the SP-A ligand blotting procedure. The disaturated PtdGros corresponding to those identified by mass spectrometry were strongly reactive with SP-A, whereas the monounsaturated-PtdGros were very weakly reactive. The more highly unsaturated 18:1/18:1-PtdGro did not bind SP-A nor did PtdGros derived from egg PtdCho synthesized by transphosphatidylation. These results confirm the findings from the HPLC/mass spectrometry analyses that disaturated PtdGros constitute novel high affinity ligands for SP-A.
In an additional series of experiments we tested the biological relevance of disaturated PtdGros as receptors for SP-A. We utilized the growth arrest of mycoplasma by SP-A (see Fig. 8) as a measure of its biological function and tested the ability of di-C16:0-PtdGro to antagonize this effect. The results of these studies are presented in Fig. 11. M. pneumoniae cultures were grown under standard conditions and treated with either lipids derived from the bacteria, di-C16:0-PtdGro, or SP-A in various combinations. The data show that inclusion of either total bacterial lipid or di-C16:0-PtdGro has no effect on the growth of mycoplasma. In contrast, the addition of SP-A significantly attenuates the growth of the organism. Addition of either the total bacterial lipid preparation or di-C16:0-PtdGro reverses the inhibition of growth effected by SP-A. These data clearly demonstrate that saturated PtdGros provided in the medium can antagonize the binding and activity of SP-A as an antimicrobial agent. Collectively these data provide strong evidence that saturated-PtdGros present in mycoplasma membranes act as receptors for SP-A. DISCUSSION M. pneumoniae is an important human pathogen, and the full details of the host interaction and response to the organism are not well understood. In this report we sought to elucidate the interactions between SP-A and the bacteria. SP-A is an abundant protein within the bronchoalveolar environment inhabited by M. pneumoniae, and it interacts with bacteria, fungi, and viruses as a specialized pulmonary component of the innate immune system (3).
These studies provide clear evidence that SP-A from human and rat sources binds M. pneumoniae with high affinity.
The SP-A binding shows an absolute requirement for Ca 2ϩ , which is typical of the interactions of the protein for most carbohydrate and lipid ligands. The apparent K d Ј values for human SP-A and mycoplasma and its isolated membranes are 7.8 and 0.3 nM, respectively. The apparent K d Ј values for rat SP-A and mycoplasma and its isolated membranes are 8.0 and 0.9 nM, respectively. We believe that the differences in apparent K d Ј values for intact organisms and membranes are due to the different assay systems employed. For intact organisms, both the processing time and the sample dilution are much greater than those used for the isolated membranes. It is likely, under the conditions employed for the intact cells, that greater dissociation of the SP-A ligand complex will occur during processing after the reaction is terminated. In this respect, the membrane binding measurements probably best estimate the true affinity of the SP-A ligand complex, because they can be executed relatively quickly.
Using isolated membranes, we demonstrated that human SP-A derived from either normal individuals or alveolar proteinosis patients behave identically in the binding reactions. The lung lavage material from alveolar proteinosis patients is a convenient source of human SP-A, but the protein displays some covalent cross-linking anomalies evident as a non-reducible dimeric form of the protein observed by gel electrophoresis under denaturing conditions (31). This finding validates the use of the alveolar proteinosis protein in these studies. We also compared the binding properties of recombinant rat SP-A produced in CHO-K1 cells and baculovirus/Sf9 cells and found that the proteins behaved similarly to each other and to the human protein. This indicates that the large collection of mutant forms of SP-A produced with the baculovirus/Sf9 system should be useful for examining structure-function relationships for interactions between SP-A and the microbial ligand.
The development of a binding assay using solid phase mem- A, lipids identified as the major reactive spot for SP-A by ligand blot analysis on TLC plates were recovered from the silica gel and applied to a reversedphase HPLC column. The lipids were eluted using solvents A (methanol/water, 20:80, v/v, containing 1 mM ammonium acetate) and B (methanol made to 1 mM ammonium acetate) and a gradient that progressed from 20% B to 100% B over 35 min, followed by continued elution at 100% B for 15 min. The lipid eluting from the column was detected by A 220 nm . B, analysis of fraction 26 by negative ion mass spectrometry and collision-induced decomposition (inset) of m/z 721.5. C, analysis of fraction 29 by negative ion mass spectrometry and collision-induced decomposition (inset) of m/z 749.5. D, analysis of fraction 32 using negative ion mass spectrometry.
branes has simplified the approach to studying the specific features of the reaction. We find that large quantities of membranes can be frozen without significant loss of binding activity. The membrane coating of microtiter wells and SP-A binding are highly reproducible. We used the mycoplasma membrane system to examine the nature of the ligands that interact with SP-A. Treatment of the membranes with proteases failed to diminish the binding of the surfactant protein, indicating that the ligand was either a lipid or highly protease resistant protein. Competitive binding experiments clearly reveal that liposomes derived from the lipid fraction are extremely effective at blocking the interactions between SP-A and solid phase membranes with an IC 50 of 0.2 nM total lipid phosphorus. Quite remarkably, the affinity of the mycoplasma lipid ligands for SP-A is greater than that of 16:0/16:0-PtdCho, which is the major lipid present in pulmonary surfactant. Collectively these results indicate that a lipid component is a dominant, if not the only, class of ligand for human SP-A expressed on M. pneumoniae.
Lipid extraction, TLC, and ligand blot analysis revealed that the mycoplasma lipid fraction contains one abundant class of high affinity ligand for SP-A as well as a ligand that appears less abundant. We have previously reported similar interactions between SP-D, a protein closely related to SP-A, and mycoplasma membranes and lipids (32). The spectrum of lipids recognized by SP-D is distinctly different from that recognized by SP-A, but there is some overlap. The major SP-A ligand co-migrates with the most abundant polar lipid class of the organism, which is PtdGro, as well as a prominent glycolipid that co-migrates in the same location. Previous studies with rat SP-A indicate that egg-PtdGro is not a ligand for the protein. We also tested the direct binding of human SP-A to egg-PtdGro by ligand blot analysis, and it is also not a significant ligand for this protein. However, more detailed analysis of the major SP-A ligand by combined TLC, HPLC, and mass spectrometry now reveals that the critical lipids for high affinity interaction are disaturated PtdGros. The 16:0/ 16:0-PtdGro, 16:0/18:0-PtdGro, and 18:0/18:0-PtdGro species were identified as the bacterial ligands. The high affinity interaction was further confirmed by demonstrating that commercially available disaturated PtdGros are high affinity ligands for human SP-A. Among the disaturated PtdGros present in mycoplasma the 16:0/18:0-PtdGro appears to be the most abundant molecular species. The ligation of M. pneumoniae by SP-A is likely to have important consequences for the organism and its human host. Our binding studies reveal that SP-A attenuates the growth of the mycoplasma. The reduced growth was confirmed by colony counting, general metabolic activity, and DNA replication. The action of SP-A must be highly specific and effective insofar as the medium that mycoplasma is grown in is very complex and contains numerous sources of potential competitive inhibitors. Some of these potential inhibitors include 15% bovine serum, yeast extract, yeastolate, and protein hydrolysates. Both the concentration and stoichiometry of SP-A and mycoplasma also appear as important factors for the attenuation of growth. The effectiveness of SP-A is progressively reduced as the bacterial inoculum is increased. Our findings demonstrate that SP-A causes cell stasis rather than cell death. This effect is specific for SP-A, because the other prominent pulmonary collectin, SP-D, fails to alter the growth of mycoplasma.
Although SP-A has been reported to bind to many microorganisms, there are only a few reports of a direct effect of the protein upon microbial growth. van Rozendaal et al. (33)  unknown receptors. Our data now provide evidence that mycoplasma is also a target for the proteins, which recognize disaturated phospholipids on the cell surface. However, we observe bacteriostatic activity rather than bacteriocidal activity by SP-A. It seems likely that the direct antimicrobial effects of SP-A will be relevant for other organisms and serve as an additional function of the protein in conjunction with its opsonizing activity and modulation of inflammatory mediator production (3).
In summary, we have demonstrated that rat and human SP-A bind M. pneumoniae with high affinity. Bacterial lipids constitute the major molecular class of surface ligands for SP-A, and they bind with higher affinity than the major surfactant lipid. Human SP-A markedly attenuates the growth of mycoplasma and is likely to play an important role in controlling the antibody independent immunity to the bacteria in vivo.