Acylation of pulmonary surfactant protein-C is required for its optimal surface active interactions with phospholipids.

This study investigates the importance of thioester-linked acyl groups in lung surfactant protein C (SP-C) in facilitating interactions with phospholipids that yield functionally important surface active behaviors. Native SP-C, palmitoylated at cysteine residues at positions 5 and 6, was isolated from bovine lung surfactant by liquid chromatography. Deacylated SP-C (dSP-C), unchanged in composition and sequence from SP-C but having a decreased α-helical content in films with dipalmitoyl phosphatidylcholine (DPPC) of 52 versus 70%, was obtained by treatment with 0.1 M sodium carbonate buffer at pH 10. Surface activity was studied for SP-C and dSP-C combined with column-purified phospholipids (PPL) from calf lung surfactant or with synthetic phospholipids (DPPC or a synthetic phospholipid mixture (SPL) containing 50:35:15, DPPC:egg phosphatidylcholine:egg phosphatidylglycerol). Interfacial measurements included surface pressure time adsorption isotherms for dispersed surfactants with diffusion minimized, dynamic surface pressure area isotherms and respreading for films in the Wilhelmy balance, and overall surface tension lowering at physiologic cycling rate in oscillating bubble experiments. Dispersions of PPL:SP-C and SPL:SP-C rapidly adsorbed to high equilibrium surface pressures of 47-48 mN/m, significantly better than corresponding dispersions containing dSP-C. The adsorption of PPL:dSP-C was essentially unchanged from that of PPL alone, and the adsorption of SPL:dSP-C was improved only slightly over SPL alone. In Wilhelmy balance studies, dynamic respreading was significantly improved over phospholipids alone in films of SP-C plus PPL, SPL, or DPPC. Respreading was improved less markedly by dSP-C in corresponding films with SPL or DPPC and not at all in films with PPL. Maximum surface pressures were also higher in cycled films of SP-C versus dSP-C combined with PPL or SPL. In bubble experiments (37°C, 20 cycles/min), dispersions of PPL:SP-C and SPL:SP-C reached low minimum surface tensions of <1 and 5 mN/m, respectively, whereas PPL:dSP-C and SPL:dSP-C only reached minima of ~20 mN/m as did PPL and SPL alone. Acylation in SP-C is crucial for its interactions with phospholipids over the full spectrum of adsorption and dynamic surface behaviors important for lung surfactant.

Pulmonary surfactant, a complex mixture of phospholipids with smaller amounts of specific proteins and neutral lipids, is essential for normal lung function (1,2). Among its apoprotein constituents, surfactant protein C (SP-C) 1 is the most hydrophobic, with a high content of valine, leucine, and isoleucine (3,4). SP-C is known to interact with phospholipids in bilayers, vesicles, and films and to improve phospholipid adsorption and dynamic film behavior (see Refs. 5-8 for review). However, the importance of specific molecular structural regions in SP-C for the generation of functionally important surface active behaviors in combination with phospholipids is only now being clarified.
Mature SP-C is an N-terminally truncated monomeric polypeptide, 35 amino acids long, made from a 21-kDa precursor, pro SP-C (3,4,7,8). The 35-residue chain has an extremely hydrophobic C-terminal region representing two-thirds of the molecule and a less hydrophobic N-terminal moiety. An additional structural characteristic of mature SP-C is that it is a true lipoprotein. Saturated palmitoyl moieties are linked as thioesters to adjacent cysteine residues at positions 5 and 6 in the majority of animal species studied (9,10), except for canine SP-C, which has one palmitoylated cysteine at position 5 (11). Acylation through thioester linkages is found in several proteins of cellular and viral origin (12). This structural modification can be biologically important through effects on cell membrane processes (13,14) and in the case of SP-C may play a role in intracellular trafficking in type II cells (15,16). Investigated in the current study is the specific importance of SP-C acylation in facilitating functionally important surface behaviors in combination with lung surfactant phospholipids.
Acylation is known to affect the structure of SP-C and to influence its biophysical interactions with phospholipids (17)(18)(19)(20)(21)(22)(23). However, previous work has not addressed the importance of SP-C acyl groups over the full range of surface active behaviors relevant for lung surfactant, which include adsorption together with surface tension lowering and respreading in dynamically cycled films (5). Previous investigations on the functional import of SP-C acylation have also involved only model systems with synthetic phospholipids (17)(18)(19)(20)(21)(22)(23), as opposed to direct studies utilizing the complete mix of phospholipids in endogenous lung surfactant. Moreover, a number of prior studies on the effects of acylation have utilized recombinant SP-C or related synthetic peptides lacking natural acylation (18,19,20,23), which are not directly comparable with native SP-C.
The present study investigates the importance of acylation in purified native bovine SP-C versus deacylated bovine SP-C (dSP-C) for the generation of adsorption and dynamic film behaviors of functional significance for lung surfactant. SP-C and dSP-C are studied in combination with the complete mix of calf lung surfactant phospholipids (PPL) obtained by gel permeation chromatography, in addition to correlative studies with dipalmitoyl phosphatidylcholine (DPPC) and a complex synthetic phospholipid mixture (SPL). Phospholipid:apoprotein mixtures are studied for the time course of adsorption following dispersion in the aqueous phase, for dynamic respreading and surface pressure area (-A) isotherm in spread films cycled in the Wilhelmy balance, and for overall surface tension lowering ability of dispersions pulsated at rapid (physiologic) rate in an oscillating bubble apparatus. For this entire set of surface behaviors, deacylation is shown either to substantially reduce or completely remove the beneficial effects of SP-C on phospholipid surface activity, indicating that acylation is essential for the surface active function of this hydrophobic surfactant protein.

MATERIALS AND METHODS
Synthetic Phospholipids-DPPC, egg phosphatidylcholine, and egg phosphatidylglycerol were purchased from Avanti Polar Lipids, Inc. (Alabaster, AL). Purity was Ͼ99% as supplied, and the existence of single spots was verified on thin-layer chromatography with solvent system C of Touchstone et al. (24). All organic solvents used with phospholipids and other surfactants were high pressure liquid chromatography grade (VWR Scientific, Philadelphia, PA), and distilled, deionized water (Milli-Q UV Plus system, Millipore Corp., Bedford, MA) was used to form all subphases in adsorption, film, and bubble studies (10 mM HEPES, 1.5 mM CaCl 2 , 150 mM NaCl, pH 7.0).
Separation of PPL-PPL were purified by column chromatography from calf lung surfactant extract (CLSE). CLSE was obtained by organic extraction of large surfactant aggregates pelleted at 12,000 ϫ g from bronchoalveolar lavage from the lungs of freshly killed calves, as detailed previously (25). An initial low speed spin (250 ϫ g) was applied to remove cellular debris prior to pelleting of lung surfactant. Separation of PPL from CLSE was by gel permeation column chromatography on an LH-20 column (Pharmacia Biotech Inc.) with chloroform:methanol:0.1 N HCl (47.5:47.5:5) (25). The protein content of PPL was below the limits of detection by the trichloroacetic acid/Amido Black assay (26).
Purification of Hydrophobic SP-C-SP-C was purified from CLSE by successive stage liquid chromatography as described previously, with several modifications including the use of Silica C8 (J. T. Baker, Phillipsburg, NJ) rather than LH-20 or LH-60 (27,28). Briefly, CLSE was applied to a preparative C8 column and eluted with chloroform:methanol:0.1 N HCl (47.5:47.5:5) to yield crude SP-B, SP-C, and surfactant phospholipids. As determined by UV absorbance at 254 nm and 17% acrylamide SDS-polyacrylamide gel electrophoresis (29), fractions containing SP-C were pooled and dialyzed against 2:1 chloroform:methanol, reduced in volume, and further purified by chloroform:methanol: HCl elution through a second C8 column. Purity of the final SP-C isolate was determined by SDS-polyacrylamide gel electrophoresis and N-terminal amino acid sequence analysis. Quantitative amino acid compositional analysis (30) also assessed purity and was used to determine the concentration of SP-C. Final SP-C preparations contained no detectable SP-B by Western blot analysis using SP-B-specific antisera, and no detectable phospholipids by microanalysis for phosphorus (31).
Preparation of dSP-C-Native SP-C was deacylated by treatment for 12 h in 0.1 M sodium carbonate, pH 10, at 25°C (17,32). Following alkaline treatment, dSP-C was dialyzed against 2:1 chloroform:methanol at 4°C for 4 h, reduced in volume, and the deacylated protein purified by C8 liquid chromatography with chloroform:methanol:HCl as described above. Fractions containing dSP-C were pooled and dialyzed against 2:1 chloroform:methanol at 4°C for 8 h (350 ϫ volume, two changes). The composition and sequence of dSP-C was verified to be unchanged from SP-C by quantitative amino acid analysis and by N-terminal sequence analysis.
Fourier Transform Infrared Spectrometry Measurements and Secondary Structural Determination in the Presence of DPPC-Secondary structural analyses of SP-C and dSP-C were performed by attenuated total reflection Fourier transform infrared (FTIR) spectrometry as described by Baatz et al. (27). Approximately 16 g of purified protein (in 2:1 chloroform:methanol solution) were mixed with a 20 molar excess of DPPC (in chloroform solution) and then spread and dried onto a trapezoidal germanium internal reflection element. Attenuated total reflection FTIR spectra were acquired on an ATI RS-1 FTIR spectrometer (Madison, WI) equipped with a mercury-cadmium-telluride detector at a resolution of 4 cm Ϫ1 at ambient temperature. A total of 512 scans were signal averaged. Fourier self-deconvolution (K ϭ 1.9) was used to identify component band peak centers of the Amide I band (1700 -1600 cm Ϫ1 ). The centers were used as initial estimates for curve-fit analysis by the PeakFit program (version 4, Jandel Scientific, San Rafael, CA) with a 5% variance in peak centers allowed during computing iterations. Line shapes consisting of 50% Gaussian and 50% Lorentzian character were used for all component bands. The percent helical structure in SP-C or dSP-C in the presence of DPPC was calculated from the area of the ␣-helical peak (ϳ1653-1660 cm Ϫ1 ) divided by the sum total of fitted component bands.
Reconstitution of Surfactant Phospholipid-Protein Mixtures-For surface activity studies, DPPC, SPL (50:35:15 DPPC:egg phosphatidylcholine:egg phosphatidylglycerol by weight), or PPL were combined with either SP-C or dSP-C. The content of SP-C or dSP-C was fixed at 1.3% by weight relative to phospholipid, equal to the total hydrophobic protein content of CLSE (25). SP-C or dSP-C in chloroform:methanol (1:1) were combined with phospholipids in chloroform, dried under nitrogen, and either dissolved in hexane:ethanol (9:1, v:v) for film studies or dispersed in buffer (10 mM HEPES, 1.5 mM CaCl 2 , 150 mM NaCl, pH 7.0) for adsorption or bubble measurements. Dispersion was by probe sonication on ice for three 15-s bursts with a Heat Systems sonicator (model W-220F).
Adsorption Methods-Adsorption experiments were performed at 37°C in a Teflon dish with a stirred subphase to minimize diffusion resistance (33,34). Measurements were initiated immediately following injection of 5 ml of surfactant dispersion (2.5 or 4.0 mg of phospholipid/5 ml) from a hypodermic syringe into a 35-ml buffered subphase. Surface pressure was monitored as a function of time from the force on a sandblasted Wilhelmy slide dipped in the surface.
Wilhelmy Balance Methods-A modified Wilhelmy surface balance with a Teflon ribbon barrier designed to maximize confinement of films at high surface pressure was used to measure the dynamic -A behavior of spread surfactant films (35)(36)(37)(38). Surfactants dissolved in hexane: ethanol (9:1, v:v) were added dropwise from a syringe at the air-water interface to a uniform "surface excess" initial concentration of 15 Å 2 / phospholipid molecule. After a 10-min pause for solvent evaporation, surface pressure was measured as a function of area during seven continuous compression-expansion cycles between 448.6 (100% area) and 103.2 cm 2 (23%) at a constant cycling speed of 5 (23 Ϯ 1°C) or 1.5 min/cycle (37 Ϯ 0.3°C). Humidity was held near the fully saturated value by open dishes of water and dampened blotting paper in the balance chamber. Surface pressure in the confined film was measured by a platinum slide and force transducer, and a second slide outside the ribbon monitored for leakage, which was absent in all reported results. Dynamic respreading was assessed by calculated areas between compressions 2 and 1 (2/1) and 7 and 1 (7/1) on the -A isotherm, as defined previously by Wang et al. (37). A calculated isotherm area of zero between compressions 2/1 or 7/1 indicated complete respreading between the designated cycles, and respreading decreased as isotherm area increased (37). In cases where the -A curve for compression 2 (or 7) crossed compression 1 to reach higher maximum surface pressures, isotherm area calculations included only the region where surface pressure was below that of compression 1 (37). Respreading calculations based on isotherm areas agree conceptually with results from collapse plateau ratio criteria developed by Notter and co-workers (36,38) but are applied more readily to compression isotherms with varying slopes or collapse surface pressures (37).
Oscillating Bubble Methods-A pulsating bubble surfactometer (Electronetics, Inc., Amherst, NY) based on the design of Enhorning (39) was used to define the surface tension lowering ability of surfactant dispersions cycled at physiologic rate (20 cycles/min, 37°C). A small air bubble, communicating with ambient air, was formed in a surfactant dispersion in a pulsator-mounted sample chamber and pulsated immediately between maximum and minimum radii of 0.55 and 0.4 mm, respectively (50% area change). The interfacial pressure drop was measured by a precision transducer, and surface tension at minimum bubble radius was calculated as a function of pulsation time from Laplace's equation for a sphere (39). The absolute accuracy of calculations based on the spherical equation, even at low surface tensions where significant nonspherical deformations can occur, has been defined previously for the bubble apparatus (40).

RESULTS
Analysis of the secondary structure of SP-C and dSP-C in the presence of DPPC was measured for subsequent correlation with the effects of acylation on the surface activity of apoprotein:phospholipid mixtures. Curve fit analysis of the Amide I band from the attenuated total reflection FTIR spectra of native (acylated) SP-C yielded 70% of the sum total area as the 1657 cm Ϫ1 component band area associated with ␣-helical structure. The area of component bands corresponding to ␤-structure (1684, 1672, and 1628 cm Ϫ1 ) represented 24% of the total area, and less than 6% of the area originated from other "random" structures (1638 cm Ϫ1 ). After deacylation, the area of the ␣-helical component (1658 cm Ϫ1 ) was reduced to 52% of the sum total area of the fitted Amide I band, whereas 20% of the area was derived from bands corresponding to ␤-structures (1683, 1671, and 1630 cm Ϫ1 ) and 19% from bands associated with random structures (1644 cm Ϫ1 ). Deacylation thus caused a decrease in the ␣-helical structure and an increase in random structures in dSP-C versus SP-C.
The time adsorption of PPL and SPL phospholipids, alone and with 1.3% SP-C or dSP-C, is shown in Fig. 1. Dispersions of PPL:SP-C (Fig. 1A) and SPL:SP-C (Fig. 1B) had greatly improved adsorption compared with PPL or SPL alone. Both PPL:SP-C and SPL:SP-C reached plateau values of equilibrium surface pressure in the range of 47-48 mN/m (Fig. 1), equivalent to native calf surfactant (41) and CLSE (Fig. 1A). These much higher equilibrium surface pressures compared with PPL and SPL alone result largely from facilitation of phospholipid adsorption by SP-C, because the pure protein itself generated a surface pressure of only ϳ3.3 mN/m for the small apoprotein concentration present over 90 min (not shown on Fig. 1). In contrast to the substantial effects of SP-C in increasing phospholipid adsorption, dSP-C gave little improvement in adsorption in dispersions of PPL:dSP-C versus PPL (Fig. 1A) and a relatively minor facilitation of adsorption for SPL:dSP-C versus SPL (Fig. 1B). The equilibrium surface pressure reached by dispersions of PPL:dSP-C was essentially unchanged from PPL alone (Fig. 1A). SPL:dSP-C had a similar (nonequilibrium) surface pressure after 50 min of adsorption, despite being at higher subphase concentration than PPL (Fig. 1B). Even after 90 min of adsorption (not shown in Fig. 1B), dispersions of SPL:dSP-C only reached adsorption surface pressures of 32.3 Ϯ 0.8 mN/m.
To complement adsorption measurements on dispersed surfactants, dynamic -A isotherms were measured at room and body temperature for solvent spread films of DPPC, SPL, and PPL, with and without SP-C or dSP-C, cylcled on the Wilhelmy balance. Representative -A isotherms (37°C) showing the first, second, and seventh compressions and first expansion for films of PPL, PPL:SP-C, and PPL:dSP-C are given in Fig. 2.
Corresponding -A isotherms for SPL, SPL:SP-C, and SPL: dSP-C (not shown) were conceptually similar except for reduced respreading. Dynamic respreading for all films studied is given in Tables I (23°C) and II (37°C) based on isotherm area calculations (see "Materials and Methods"). In films of phospholipids alone, respreading was consistently ordered as PPL Ͼ SPL Ͼ DPPC (Tables I and II). The excellent respread- ing of films of PPL compared with even the complex SPL mixture was particularly evident at physiologic temperature (37°C , Table II). Despite the excellent respreading of PPL alone, combination with native SP-C improved respreading further in films of PPL:SP-C, and this acylated protein had even more significant effects on respreading in films of SPL: SP-C versus SPL and DPPC:SP-C versus DPPC (Tables I and  II). In contrast, dSP-C had almost no effect on the respreading of PPL and gave only minor increases in the respreading of SPL and DPPC (Tables I and II).
In addition to dynamic respreading, the surface tension lowering ability of phospholipid:apoprotein films during compression also varied depending on whether they contained SP-C or dSP-C. The maximum surface pressures reached as a function of cycle number for films of PPL and SPL, with and without SP-C versus dSP-C, are shown in Fig. 3. Films were spread initially to 15 Å 2 /phospholipid molecule and cycled at 1.5 min/complete cycle at 37°C. SP-C did not alter the maximum surface pressure in mixed films with PPL (Fig. 3A) and significantly raised the maximum surface pressure in films with SPL (Fig. 3B). In contrast, dSP-C reduced maximum pressures (increased minimum surface tensions) compared with phospholipids alone when combined with either PPL or SPL (Fig. 3, A and B).
The overall surface activity of mixtures of phospholipids with SP-C versus dSP-C was defined in oscillating bubble studies for physiologically relevant area compressions and physical conditions (Fig. 4). The minimum surface tension reached by lung surfactant films is a strong function of compression rate. Wilhelmy balance methodology does not permit film cycling rates in the range of normal breathing due to technical limitations, whereas oscillating bubble measurements at 37°C and high

Respreading of surfactant films on a Wilhelmy balance at 37°C
All films were initially spread in hexane:ethanol (9:1, v:v) to 15 Å 2 / phospholipid molecule and cycled continuously at 1.5 min/cycle at 37°C. Data are the means Ϯ S.E. for n ϭ 4 -5. Other details as in Table  I   humidity were at a rate of 20 cycles/min with a 50% area compression. Bubble measurements again found significant differences in activity between phospholipid mixtures containing 1.3% native SP-C versus dSP-C (Fig. 4). Dispersions of PPL:SP-C (Fig. 4A) and SPL:SP-C (Fig. 4B) reached minimum surface tensions of Ͻ1 and 4.6 Ϯ 0.8 mN/m, respectively, over 20 min of pulsation. In contrast, dispersions of PPL:dSP-C and SPL:dSP-C reached minimum surface tensions equivalent to the phospholipids alone (minima Ͼ20 mN/m; Fig. 4, A and B). Dispersions of PPL:dSP-C and SPL:dSP-C did achieve lower minimum surface tensions than the phospholipids alone early during pulsation, but this difference did not persist over the full 20-min pulsation period investigated (Figs. 4, A and B). DISCUSSION This study shows that palmitoylation in SP-C is crucial for its surface active interactions with phospholipids across a range of interfacial behaviors important for lung surfactant function. Apoprotein-phospholipid interactions were studied directly with the complete mix of phospholipids purified from calf lung surfactant, supplemented by additional experiments with DPPC and model synthetic phospholipids. To further maximize relevance for mammalian lung surfactant, the importance of acylation was investigated using purified bovine SP-C, rather than recombinant SP-C or related synthetic peptides that must be acylated synthetically (18 -20, 23, 42). Experiments utilized complementary interfacial techniques to assess both the adsorption and dynamic film properties of phospholipid-apoprotein mixtures. All of the surface activity results were consistent in showing that although native SP-C significantly enhanced the surface activity of phospholipids, its beneficial effects were either completely abolished or substantially decreased by deacylation.
It is well known that SP-C and other lung surfactant proteins mix with phospholipids and affect their surface activity (see Refs. 5-8 for review). Acylation in SP-C is thought to play a role in its biophysical behavior (18 -21, 27), but direct comparisons of the interactions of SP-C and dSP-C with the complex mix of phospholipids present in lung surfactant have not been done. Also, the interactions of SP-C and dSP-C with phospholipids have not been defined over the range of specific surface active behaviors and cycling rates important in lung surfactant function. This includes surface tension lowering in dynamically cycled interfacial films as well as during adsorption from the subphase. It also includes respreading from dynamically generated film collapse phases during cycling, which is conceptually distinct from adsorption in the absence of compression (5). The demonstration that deacylation of SP-C results in marked and consistent detriments in its ability to improve this range of surface active behaviors for PPL as well as synthetic phospholipids is direct evidence for the functional importance of this molecular modification in lung surfactant biophysics.
Adsorption and dynamic film properties were consistently different between corresponding mixtures of phospholipids with SP-C versus dSP-C. Mixtures of PPL:SP-C and SPL:SP-C adsorbed to final equilibrium surface pressures similar to extracted calf lung surfactant, whereas corresponding mixtures containing dSP-C had adsorption improved only slightly compared with PPL and SPL alone (Fig. 1). In interfacial films, the dynamic respreading of PPL and synthetic phospholipids was significantly enhanced by SP-C, whereas phospholipid films containing dSP-C exhibited substantially smaller improvements in respreading (Tables I and II). Maximum surface pressure was decreased in films of PPL and SPL with dSP-C versus SP-C in Wilhelmy balance studies at slow compression rates (Fig. 3). Dispersions of PPL or SPL with dSP-C also could not reduce surface tension below the ϳ20 mN/m values achieved by phospholipids alone in bubble studies at higher compression rate, whereas PPL:SP-C and SPL:SP-C dispersions reached low minimum surface tensions of Ͻ1 and 5 mN/m, respectively in oscillating bubble studies (Fig. 4).
The differences in adsorption and dynamic film properties found here for mixtures of phospholipids with SP-C versus dSP-C are consistent with prior work on recombinant and chemically acylated SP-C. Gü nther et al. (18) showed that mixtures of bovine SP-C and synthetic phospholipids were better able to inhibit plasmic cleavage of fibrinogen than corresponding mixtures with recombinant nonpalmitoylated human SP-C. Seeger et al. (19) reported that recombinant SP-C lacking covalently linked palmitoyl groups was less effective than native hydrophobic apoproteins at reversing plasma protein-induced inhibition of surfactant function. Creuwels et al. (20) showed that the presence of chemically linked palmitoyl chains in recombinant human SP-C affected the ability of the protein to facilitate the insertion of phospholipids from vesicles into monolayers and also affected film compressibility and apoprotein orientation. Qanbar and Possmayer (21) found no difference in the surface tension lowering ability of 7:3 DPPC:egg phosphatidylglycerol combined with native versus chemically deacylated SP-C in pulsating bubble studies, but the extremely high surfactant concentration studied (10 mg/ml) may have masked activity differences. The structure of SP-C isolated from the lung lavage of patients with alveolar proteinosis has also been shown to be modified by partial or complete removal of the covalently linked palmitate residues, possibly contributing to reduced surfactant function (43).
Acylation-related differences in ␣-helix structure between SP-C and dSP-C almost certainly contribute to the surface property differences measured here (Tables I and II and Figs. 1-4). The secondary structure of native (acylated) SP-C is primarily ␣-helical, although a wide range of percentages (60 -90%) has been reported (42, 44 -49). Our value of 70% helical structure for native bovine SP-C in DPPC films is consistent with previous results of 71% helix in SP-C by CD (44) and 23% nonhelical structures (first eight amino acids) in SP-C by NMR (49). Our result that bovine dSP-C has a reduced ␣-helical content of 52% also agrees with prior studies using different deacylation methodology but showing a similar magnitude of decrease in helical structure in dSP-C based on CD (45) and FTIR (48) measurements. Also consistent is the NMR result that removal of the two palmitoyl groups causes the helix to start at or near residue 15 rather than at residue 9 as in native SP-C, essentially leaving only the polyaliphatic moiety of SP-C in a helical confomation (Ͻ60% of the polypeptide in helical structure) (50).
The agreement between our dSP-C structural findings and those of previous studies using different deacylation methods (45,48,50) indicates that the marked differences found between SP-C and dSP-C in interfacial studies with phospholipids do not simply reflect method-dependent alterations in dSP-C. In addition, our dSP-C obtained by carbonate buffer treatment (17) of bovine SP-C was verified to be unchanged in amino acid composition and retained a monomeric form. Extensive dialysis removed the mild acidic and/or basic solvents used during preparation and purification, and subphase pH was adjusted to neutral during all biophysical experiments (see "Materials and Methods"). Our SP-C and dSP-C preparations were both highly delipidated, and this variable was also not confounding in interpretations of surface activity.
Differences in the length of the ␣-helix, as found for SP-C versus dSP-C, can affect both incorporation into and interaction with lipid bilayers (47,49 -52). For example, mismatches be-tween the length of transmembrane helices and the thickness of lipid bilayers with which they interact can cause phase separation of peptides and lipids (52). This may be an important factor for the interactions of positive charges at amino acid residues 11 and 12 (Lys and Arg, respectively, in bovine SP-C) with surfactant phospholipids. Creuwels et al. (53) have recently reported that the positively charged amino acid residues are important in the ability of SP-C to facilitate the binding of subphase phospholipid vesicles to preformed interfacial films. If deacylation alters the secondary structure associated with the region of these charged residues (50), then interactions with phospholipids will be affected, contributing to differences in surface active properties such as adsorption and respreading as measured here.
In addition to affecting surface activity by influencing the structure of SP-C itself, acyl groups on the protein can also directly interact with fatty chains on adjacent phospholipid molecules. The similarity of fatty acid chains is known to influence the ability of phospholipids to mix and interact in interfacial films (54), and the palmitoyl moieties of SP-C are ideally suited for interacting with many of the phospholipid constituents of lung surfactant including DPPC. Acylated, native SP-C lowers the gel to liquid crystal transition temperature of phospholipid bilayers (51,(55)(56)(57) and decreases transition enthalpy (17,47,51,(55)(56)(57). Chain-chain interactions between SP-C acyl groups and adjacent phospholipids very likely contribute to these effects. The lowering and broadening of the gel to liquid crystal phase transition by native SP-C indicates that the lipids surrounding the protein are in a more fluid state (57). This would be consistent with our finding that SP-C increases respreading in phospholipid films, because fluid-unsaturated phospholipids with low transition temperatures have been shown to improve the respreading of DPPC (36,58). Direct acyl chain interactions wth phospholipids might also help "anchor" SP-C in the bilayer or surface film, antagonizing phase separation, and permitting maximal interactions. Precise definition of the relative contributions of different molecular mechanisms to the surface active function of SP-C is complicated by their interdependence. The acyl chains themselves are part of the SP-C molecule, and their removal by necessity affects protein structure in addition to removing the potential for direct chain-chain interactions with phospholipids.
In summary, this study demonstrates that the acylation of SP-C is crucial for its surface active interactions with lung surfactant phospholipids. Native, acylated bovine SP-C greatly facilitated the adsorption of purified calf lung surfactant phospholipids and synthetic phospholipids. Dynamically cycled phospholipid:SP-C films also had improved respreading and surface tension lowering in the Wilhelmy balance, and dispersions of SP-C with surfactant phospholipids and synthetic phospholipids reached minimum surface tensions of Ͻ1 and Ͻ5 mN/m, respectively, at physiologic compression rates in oscillated bubble studies. In contrast, deacylated bovine SP-C either did not improve or gave only minor improvements in phospholipid adsorption and dynamic film respreading, particularly when combined with the complete mix of lung surfactant phospholipids, and was detrimental to surface tension lowering in Wilhelmy balance studies. Acylation in SP-C is essential for a spectrum of surface active behaviors relevant for lung surfactant function.