This Article Abstract Full Text (PDF) All Versions of this Article: biophysj.104.051201v1 87/6/4188    most recent Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Alonso, C. Articles by Zasadzinski, J. A. Search for Related Content PubMed PubMed Citation Articles by Alonso, C. Articles by Zasadzinski, J. A.

More Than a Monolayer: Relating Lung Surfactant Structure and Mechanics to Composition

Coralie Alonso ^*, Tim Alig ^*, Joonsung Yoon ^*, Frank Bringezu , Heidi Warriner ^* and Joseph A. Zasadzinski ^*

^* Departments of Chemical Engineering and Materials, University of California, Santa Barbara, California 93106-5080 USA; and Institute of Medical Physics and Biophysics, University of Leipzig, D-04103 Leipzig, Germany

Correspondence: Address reprint requests to Joseph A. Zasadzinski, Tel.: 805-893-4769; Fax: 805-893-4731; E-mail: gorilla{at}engineering.ucsb.edu.

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL RESULTS AND DISCUSSION CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES

 
Survanta, a clinically used bovine lung surfactant extract, in contact with surfactant in the subphase, shows a coexistence of discrete monolayer islands of solid phase coexisting with continuous multilayer "reservoirs" of fluid phase adjacent to the air-water interface. Exchange between the monolayer, the multilayer reservoir, and the subphase determines surfactant mechanical properties such as the monolayer collapse pressure and surface viscosity by regulating solid-fluid coexistence. Grazing incidence x-ray diffraction shows that the solid phase domains consist of two-dimensional crystals similar to those formed by mixtures of dipalmitoylphosphatidylcholine and palmitic acid. The condensed domains grow as the surface pressure is increased until they coalesce, trapping protrusions of liquid matrix. At 40 mN/m, a plateau exists in the isotherm at which the solid phase fraction increases from 60 to 90%, at which the surface viscosity diverges. The viscosity is driven by the percolation of the solid phase domains, which depends on the solid phase area fraction of the monolayer. The high viscosity may lead to high monolayer collapse pressures, help prevent atelectasis, and minimize the flow of lung surfactant out of the alveoli due to surface tension gradients.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL RESULTS AND DISCUSSION CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES

 
Mammalian lung surfactant is a complex mixture of lipids and proteins, the main purpose of which is to reduce the surface tension in the lungs during breathing (Schürch et al., 1976, 1978; Goerke, 1998). Lower surface tensions lead to a greatly reduced work of breathing and greater stability of the alveoli against collapse or overinflation. A deficiency of lung surfactant, common in premature infants, results in neonatal respiratory distress syndrome (NRDS) (Notter, 2000). Early administration of exogenous lung surfactant is an effective treatment for NRDS; however, supplies of human lung surfactant are limited and currently used clinical replacement surfactants are derived from animal sources (Notter, 2000). Survanta is one of the formulations approved by the U. S. Food and Drug Administration to treat respiratory distress syndrome (RDS) and is widely used in U.S. hospitals. Survanta is an organic extract of minced bovine lung supplemented with 1,2-dipalmitoyl-sn-glycero-3-phosphatidylcholine (DPPC), palmitic acid (PA), and tripalmitin (Bernhard et al., 2000). The additional lipids alter certain measures of in vitro surfactant performance, especially the ability to reduce surface tension to near zero on film compression and promote reproducible hysteresis loops on repeated expansion and compression cycles (Notter, 2000).

The latest clinical trials comparing Survanta to Infasurf (calf lung lavage extract) or Curosurf (minced porcine lung extract) show modest differences in neonate mortality or major complications such as bronchopulmonary dysplasia, intraventricular hemorrhage, and necrotizing enterocolitis (Halliday, 1995; Bloom et al., 1997; Clark et al., 2001), despite the fact that the clinically used surfactants vary significantly in both lipid and protein composition (Bernhard et al., 2000). This suggests either that the detailed lipid and protein composition of surfactant is unimportant, or that the currently used clinical surfactants are still far from optimal. Because no replacement surfactant has yet been shown to be comparable to the complete, native surfactant (including all the surfactant proteins), it appears that there still is room for improvement. In vitro, adding SP-B and/or SP-C to Survanta improves the monolayer collapse and respreading behavior (Ding et al., 2001); in vivo, SP-B supplementation of Survanta improves oxygenation and lung compliance in animal models of RDS (Mizuno et al., 1995; Walther et al., 1997, 2000). To provide better answers to the question of optimal surfactant formulation requires a better understanding of how Survanta and other animal extract surfactant films operate at the air-water interface. One way to do this is by relating the roles of the individual lipid and protein species in Survanta to those in simpler model systems. This should provide the understanding necessary to optimize existing animal-extract surfactants as well as to create a purely synthetic replacement surfactant for treatment of NRDS and ARDS.

The generally accepted mechanism by which lung surfactants operate requires that the different lipids and proteins phase separate within the film, and certain components are removed or "squeezed out" as the surface pressure is raised during expiration (Notter, 2000). These components may be lost to the subphase or stored in an associated "reservoir" in the vicinity of the monolayer to facilitate readsorption when the surface pressure is lowered on inspiration (Schürch et al., 1995, 1998; von Nahmen et al., 1997; Ding et al., 2001). Other mechanisms of exchange with the monolayer have been proposed including partial collapse of the fluid phase (Piknova et al., 2002), or exchange with surfactant in solution or exchange with surfactant associated with the interface during the entire breathing cycle (Schürch et al., 1995, 1998; von Nahmen et al., 1997; Takamoto et al., 2001; Zasadzinski et al., 2001; Ding et al., 2003).

A common feature of almost all lung surfactants and model mixtures (Longo et al., 1993; Lipp et al., 1996, 1998; Dluhy et al., 1998; Nag et al., 1998; Discher et al., 1999a,b; Diamant et al., 2000; Ding et al., 2001, 2002a, 2003; Takamoto et al., 2001; Diemel et al., 2002a,b; Piknova et al., 2002; Cruz et al., 2004) is coexistence between a semicrystalline, solid phase (often called liquid condensed (LC) phase in the monolayer literature), and a disordered, fluid phase (often called liquid expanded (LE)) (McConnell, 1991), especially at high surface pressures. Such phase coexistence is important to controlling the surface viscosity and other mechanical properties of the films (Ding et al., 2002a; Alonso and Zasadzinski, 2004), which, in turn, promote high collapse pressures (Diamant et al., 2000; Gopal and Lee, 2001; Lu et al., 2002).

The primary lipid component of Survanta (Bernhard et al., 2000), as is the case for all lung surfactants (Notter, 2000), is saturated dipalmitoylphosphatidylcholine. The fraction of DPPC and other saturated PCs ranges from 40% for Infasurf, to 60% for Curosurf, to 70% for Survanta (Bernhard et al., 2000). From single component and simple lipid mixture phase studies (McConnell, 1991; Bringezu et al., 2001; Lee et al., 2002), solid phases are formed from saturated lipids such as DPPC, as well as saturated fatty acids and fatty alcohols (Bringezu et al., 2001; Lee et al., 2002) and probably tripalmitin. Palmitic acid and DPPC form stable crystalline monolayer phases at lower temperatures than DPPC alone (Bringezu et al., 2001; Ding et al., 2002a; Lee et al., 2002), which increases the fraction of solid phases at a given temperature and surface pressure (Ding et al., 2002a). Survanta is the only animal-derived surfactant that has a significant fraction of free fatty acid and triglycerides (from 10 to 20% by weight relative to DPPC).

Adsorption from the subphase and respreading are generally associated with the fluid phases of surfactant, and hence the unsaturated lipids, although surfactant proteins SP-B and SP-C also play a significant role (Poulain et al., 1996). SP-C is likely involved in maintaining a reservoir of material available for respreading in the vicinity of the interface (von Nahmen et al., 1997; Lipp et al., 1998; Cruz et al., 2000; Ding et al., 2001, 2003; Takamoto et al., 2001). SP-B influences the nature of monolayer collapse, promoting a reversible monolayer folding (Lipp et al., 1998; Ding et al., 2001), alteration of solid-liquid coexistence morphology, and formation of three-dimensional structures in the films (Lipp et al., 1998; Schürch et al., 1998; Perez-Gil, 2002; Ding et al., 2003; Cruz et al., 2004), and is important to the formation of tubular myelin surfactant and adsorption from the subphase (Poulain et al., 1996; Notter, 2000). The surfactant proteins are primarily segregated to the fluid phases of the monolayers, and SP-B is often associated with anionic lipids such as phosphatidylglycerols (Lipp et al., 1998; Cruz et al., 2000, 2004; Ding et al., 2003). The SP-B level in Survanta is 10% and the SP-C level is 50% of native bovine surfactant (Bernhard et al., 2000). Survanta contains no cholesterol (Bernhard et al., 2000); the quantity, distribution, and function of cholesterol in both native and replacement surfactants is still an open question (Discher et al., 1999a; Orgeig and Daniels, 2001).

AFM images show coexistence of discrete monolayer islands of solid phase coexisting with continuous multilayer "reservoirs" of fluid phase adjacent to the air-water interface immediately after adsorption. This is in contrast to the coexistence of monolayer fluid and solid phases formed by deposition from organic solution. The multilayer reservoirs of fluid phase are present at low surface pressure and appear to result from the adsorption from the surfactant in the subphase (Winsel et al., 2003). The fluid phase multilayer patches are removed from the monolayer at high surface pressures. The condensed domains grow as the surface pressure is increased until they coalesce, trapping protrusions of liquid matrix at the boundaries between the condensed domains. The solid phase remains primarily as a monolayer up to collapse. Subsequent expansion and compression cycles lead to phase coexistence between solid and fluid phases at lower surface pressures, the fraction of solid phase at a given surface pressure increases with the number of expansion and compression cycles.

Grazing incidence x-ray diffraction shows that Survanta separates into at least two phases, one or more of which are two-dimensional crystals. One crystal population is similar to those formed by mixtures of dipalmitoylphosphatidylcholine and palmitic acid (Bringezu et al., 2001; Lee et al., 2002), as expected from the composition of Survanta. On the first compression, at 40 mN/m, a plateau exists in the isotherm at which the solid phase fraction increases from 60 to 90%. At or just before this plateau, the surface viscosity diverges. As in simple lipid and lipid/protein films (Ding et al., 2002a; Alonso et al., 2004; Alonso and Zasadzinski, 2004), the viscosity is driven by the jamming of the solid phase domains, which depends on the solid phase area fraction of the monolayer. The high viscosity at low surface tension is partly responsible for the high surface pressure at monolayer collapse (Lu et al., 2002), and also likely minimizes atelectasis as well as the loss of lung surfactant from the alveoli due to surface tension gradients (Bangham et al., 1979; Orgeig and Daniels, 2001; Alonso et al., 2004; Alonso and Zasadzinski, 2004). The solid phase lipids promote two important features of a functioning lung surfactant—high surface pressure at collapse and a high surface viscosity. The fluid phase lipids, in combination with the small fractions of SP-B and SP-C, promote adsorption to the interface and formation of reservoirs of fluid phase lipids attached to the film at low surface pressures. Subsequent cycles increase the fraction of solid phase, leading to the viscosity transition occurring at lower surface pressures. Increasing the temperature moves the viscosity transition to higher surface pressures, again consistent with the transition being driven by a minimum solid phase fraction (Ding et al., 2002a).

	EXPERIMENTAL

TOP ABSTRACT INTRODUCTION EXPERIMENTAL RESULTS AND DISCUSSION CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES

 
Materials
Survanta is an organic solvent extract of minced bovine lungs produced for clinical use by Ross Laboratories (Columbus, OH). Survanta contains mainly phospholipids and two surfactant specific proteins, SP-B and SP-C (Bernhard et al., 2000). The extract is supplemented with DPPC, PA, and tripalmitin (Tanaka et al., 1986). Like all natural products, the composition varies, but according to the package insert, 1 ml of Survanta contains 25 mg of lipids, of which, 11–15.5 mg/ml are disaturated phosphatidylcholine (mainly DPPC), 0.5–1.75 mg/ml are triglycerides (mainly tripalmitin), and 1.4–3.5 mg/ml are free fatty acids (primarily PA). The primary anionic lipid is phosphatidylglycerol. There is negligible cholesterol in Survanta due to the extraction process. Of the phospholipids, 80% is phosphatidylcholine, of which, 80% is disaturated (Bernhard et al., 2000). There is <1% (by weight) of surfactant protein, of which 1.3 mg/(mmol phospholipid) is SP-B and 16 mg/(mmol phospholipid) is SP-C. The SP-B level is 10% and the SP-C level is 50% of native bovine surfactant (Bernhard et al., 2000). The hydrophilic proteins SP-A or SP-D (Notter, 2000) are removed from Survanta and all other clinical surfactants. For our experiments, Survanta was diluted in buffer (NaCl 150mM, CaCl₂ 2mM, NaHCO₃ 0.2mM, and pH = 7) to a lipid concentration of 1–2 mg/ml and was spread onto the same buffer solution in the Langmuir trough via a syringe. Some material spreads laterally as a drop of solution hits the surface but the film is mainly formed via adsorption of the material from the subphase. Typically, sufficient Survanta was added to raise the surface pressure to 10–20 mN/m before compression was initiated.

Methods
Isotherms were recorded with a custom-built Langmuir trough equipped with a Wilhelmy plate and two barriers providing a symmetric compression (Lipp et al., 1997; Ding et al., 2002a,b). The trough had a surface area of 260 cm², and a typical compression/expansion cycle took 20 min. Certain experiments at higher temperature were done using a ribbon trough with an area of 140 cm² to minimize leakage around the barriers. There was no appreciable difference in the isotherms taken with the different troughs or for cycle times ranging from a minimum of 5 min to a maximum of 40 min. The trough temperature was maintained by circulating water and a thermoelectric device to within ±1°C of the desired temperature.

A p-polarized He-Ne laser was used as the light source for the Brewster angle microscope. A polarizer was placed after the laser to ensure complete polarization. The incident light hits the surface at angle of 53.1° with the vertical, which is the Brewster angle for pure water. At the Brewster angle, the water surface does not reflect any light. Materials, such as surfactants, at the interface invalidate the Brewster condition due to variations in the refractive index caused by variations in the composition or density of the film (Hénon and Meunier, 1991). As only the film reflects the light, the features in the BAM images come from the interfacial film. To improve the contrast on the final images, the reflected light was passed through an analyzer and a diaphragm before being collected with a CCD camera. The magnifier placed after the analyzer allows the adjustment of the horizontal field of view over the range of 420–2800 µm.

For fluorescence imaging, a Nikon Optiphot with the stage removed was positioned above the ribbon trough. A 40x long working distance objective designed for fluorescent light was used for magnification. A dichroic mirror/barrier filter assembly directs the excitation light onto the monolayer (with a normal angle of incidence) and filters the emitted fluorescence. The emitted fluorescence was collected by the objective and detected via a silicon intensified target camera. Images were recorded directly to the computer via an Adaptec frame grabber. Contrast in the images was due to segregation of .5–1 mol% of the fluorescent lipid Texas Red-DHPE (Molecular Probes, Eugene, OR). The fluorescent probe segregates to disordered or fluid phases, which appear bright in images; the probe is absent from the solid or liquid condensed phases, which appear dark (McConnell, 1991).

Grazing incidence x-ray diffraction (GIXD) measurements were performed on a pure water subphase at 20°C using the liquid surface diffractometer on the beam line BW1 at HASYLAB, DESY (Hamburg, Germany). The synchrotron beam was made monochromatic by a beryllium (002) crystal ( = 1.304 Å). Experiments were performed at an angle of incidence of 0.85 _c (_c is the critical angle for total external reflection), which maximizes sensitivity for the surface and reduces background scattering from the subphase. The dimension of the footprint is 2 x 50 mm. To suppress capillary waves, a borosilicate glass block was inserted underneath the footprint area limiting the water depth to 300 µm. The diffracted signal was recorded along the vertical scattering vector component Q_z = (2/) sin_f, where _f is the vertical exit angle, using a linear position sensitive detector (OED-100-M, Braun, Garching, Germany). The measurements were performed by scanning over a range along the horizontal scattering vector component Q_xy (4/) sin _xy, where 2_xy is the horizontal diffraction angle. The horizontal resolution of 0.008 Å^–1 was determined by a Soller collimator mounted in front of the position sensitive detector. The vertical acceptance of the detector was 0 < Q_z < 1.27 Å^–1. The analysis of the in-plane diffraction data results in Q_xy- peak positions, which yield d-spacings that can be indexed by two Miller indices (hk) to describe the unit cell. From the peak positions of the in-plane and the out-of-plane scattering vector components it is possible to derive information about the tilt angle and the tilt direction. Assuming an exponential decay of the correlation as known from liquid crystals, it is possible to derive information about the coherence length, _xy = 2/Q_xy from the full width at half maximum of the horizontal vector component (Q_xy) after correction for the instrumental resolution (Helm et al., 1987). Detailed reviews of the principles of GIXD for the study of two-dimensional crystalline films at the air-liquid interface are available in the literature (Als-Nielsen and Möhwald, 1991; Als-Nielsen et al., 1994; Kaganer et al., 1999).

Samples for atomic force microscope (AFM) imaging were prepared using classical Langmuir-Blodgett deposition (Zasadzinski et al., 1994). Mica substrates, cleaved with adhesive tape immediately before use, were put into the subphase and held by tweezers perpendicular to the interface. Survanta was spread as described earlier and the barriers were compressed to the correct surface pressure. The substrate was pulled through the interface while the surface pressure was held constant via the motion of the computer-controlled barriers. At the plateau in the isotherm, where the change in area with change in surface pressure was large (Fig. 1), the pressure control feedback was difficult so films could not be prepared. The film-covered mica surfaces were transferred to a modified Nanoscope III AFM (Digital Instruments, Santa Barbara, CA) for imaging under ambient conditions. AFM imaging was done with a 150- x 150-µm (J) scanner in contact mode in air at ambient temperature. Silicon nitride tips with a spring constant of 0.12 N/m were used. Under the deposition and imaging conditions used here, the hydrophobic or air side of the film is facing outward from the mica and is the surface being imaged (Zasadzinski et al., 1994).

View larger version (25K): [in this window] [in a new window]   FIGURE 1  Surface pressure-trough area (-A) isotherms of Survanta adsorbed from saline suspension onto a saline buffer at 25°C. The first compression-expansion cycle is labeled 1. Consecutive compression-expansion cycles are labeled 2–4. The lower surface pressure at the beginning of the second cycle suggests that material is lost during the first compression. Dashed lines identify several zones common to all isotherms. Zone A is a homogeneous LE phase at low surface pressure. Zone B corresponds to phase coexistence between an LC phase and the LE phase (see Figs. 2–4). Zone C shows the squeeze-out plateau, starting at 38 mN/m, which accompanies the removal of the LE phases from the film, which leads to a less compressible condensed phase (zone D). At E, the film collapses; after the first cycle, the high surface pressure zones are reproducible, suggesting a minimal loss of solid phase.

View larger version (72K): [in this window] [in a new window]   FIGURE 2  Fluorescence microscopy images of Survanta films spread on saline buffer at 37°C during the first (column 1) and second compression (column 2) of the films. The fluorescent lipid Texas Red-DHPE (0.5–1 mol%) (Molecular Probes) was added to provide contrast in the images. The solid phase (see Fig. 3 and Table 1) appears dark and the liquid phase is bright. The surface pressure () is given to the left. On the first compression, the film is in the liquid expanded phase at low surface pressures (13 mN/m). Increasing the surface pressure leads to a steady increase in the fraction of solid phase domains; the fluid phase remains continuous up to the plateau pressure. At 55 mN/m, above the plateau pressure, the number of bright, white spots increases, suggesting a loss of fluid phase material to the subphase. On the second compression, the film has changed sufficiently in composition that the solid phase domains start to nucleate by 13 mN/m (column 2). At every surface pressure, the fraction of solid phase is higher, consistent with the shift of the isotherms toward smaller areas in Fig. 1. The collapse of the film occurs through the formation of bright cracks, one of which is visible in the image at 55 mN/m. At high surface pressure, there is less evidence of bright spots in the image, consistent with less fluid phase being available.

View larger version (23K): [in this window] [in a new window]   FIGURE 3  Contour plots (column 1) of the corrected x-ray intensities versus in-plane and out-of-plane scattering vector components Qxy and Qz for Survanta on water at 20°C. The corresponding Bragg peaks as function of Qxy are shown in column 2. The Bragg peaks were obtained by summarizing the measured x-ray intensities over Qz between 0 < Qz < 0.5 Å–1. From bottom to top the plots show the results obtained at 20, 30, 35, and 40 mN/m, respectively. At 20 mN/m, the contour plots show a significant intensity for Qz > 0, indicating a tilted phase. By 40 mN/m, the contour has moved to Qz = 0, indicating that the films are untitled. This shows that the solid phase is not pure DPPC, which remains tilted at 40 mN/m. The lattice parameters and phases are assigned in Table 1.

View larger version (66K): [in this window] [in a new window]   FIGURE 4  Atomic force microscopy images of Survanta transferred onto mica substrates by conventional Langmuir-Blodgett deposition at the surface pressure given at the left of each image. The low-resolution images are 25 x 25 µm, and the high-resolution images are 2.5- x 2.5-µm sections taken from the low-resolution images. The height trace in the third column corresponds to the white line on each high-resolution image. The brightness in the image increases with the height relative to the mica substrate. At 15 mN/m, the film consists of a continuous fluid phase separating circular domains of solid phase, consistent with the fluorescence images in Fig. 2. The fluid phase is 30 nm thicker than the solid phase domains, suggesting the adsorption of multilayers of lipid onto the fluid phase. At 30 mN/m, the morphology is similar, but the height difference has decreased and the solid domains appear less homogeneous and much rougher. As in the fluorescence images, the fraction of solid phase domains increases with increasing surface pressure. At 40 mN/m, the height difference between the solid and fluid domains decreases and the solid domains are packed closer together, although the fluid phase is still continuous. From 44–47 mN/m, the solid phase domains coalesce and the fluid phase is removed. At 44 mN/m, the solid and fluid phase both start to roughen, and the solid domains are no longer uniform circles; they appear to fuse to form more oblong structures. By 47 mN/m (see Fig. 7), most of the fluid phase has been removed; only small patches remain between the solid phase domains. Further compression to 70 mN/m does not change the monolayer significantly.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL RESULTS AND DISCUSSION CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES

 
Isotherms
Typical surface pressure-trough area (-A) isotherms for Survanta films spread from saline solution onto saline buffer at 25°C are shown in Fig. 1. The x axis is labeled as % trough area, as the absolute amount of Survanta that adsorbs to the interface is not fixed. Four consecutive compression-expansion cycles are shown, labeled from 1 to 4. All cycles present the same features; however, the first cycle, just after adding the surfactant, is offset to higher areas than subsequent cycles. A larger amount of material is lost to the subphase between the two first cycles as compared to later cycles (see Figs. 2 and 3). During the first compression, excess material is removed from the interface. However, as this Survanta film was formed via adsorption from the subphase, the first compression could also be a composition-refining process; that is, the excess material expelled from the interface may not have the same composition as what is left behind. In Fig. 1, dashed lines are plotted with the isotherms to distinguish five zones, labeled A–E, singled out by their slope.

View this table: [in this window] [in a new window]   TABLE 1  Lattice parameters (a,b,), chain tilt (t), area per chain perpendicular to the chain long axis (A0), and the projected area per chain (Axy) at the water surface of two distinct crystal phases observed in Survanta films

View larger version (49K): [in this window] [in a new window]   FIGURE 7  Brewster angle microscopy images of a Survanta film deposited on saline buffer at 25°C taken during the second compression. Each picture is 300 x 300 µm. The contrast is inverted from the fluorescence images—the solid phase is bright and the fluid phase is dark. At the plateau (Fig. 1, zone C) at 42 mN/m, the dark (fluid) phase is mainly eliminated in favor of the bright (solid) phase. The solid phase is beginning to coalesce, as in the fluorescence and AFM images. By 47 mN/m, the contrast is lost, indicative of a homogenous condensed monolayer. On further compression (zone D), bright white spots appear representing three-dimensional aggregates and their number increases up to the collapse pressure (66 mN/m).

Grazing incidence x-ray diffraction
GIXD is sensitive only to phases with crystalline order. Fig. 3 shows contour plots of the corrected x-ray diffraction intensities as a function of the in-plane scattering vector components Q_xy and the out-of-plane scattering vector component Q_z at lateral pressures of 20, 30, 35, and 40 mN/m. Starting in zone B of the isotherm, the diffraction pattern shows Bragg peaks, confirming that the dark domains in Fig. 2 have crystalline order. Considering the composition of Survanta, these domains are likely formed of DPPC and PA (Bringezu et al., 2001; Lee et al., 2002) and tripalmitin or DPPC plus tripalmitin, the three saturated lipid species known to crystallize at the air-water interface. The lattice parameters are given in Table 1. At 20 mN/m, four diffraction peaks at 1.428 Å^–1, 1.454 Å^–1, 1.482 Å^–1, and 1.509 Å^–1 can be extracted from the data. The analysis yields a Bragg rod maximum at Q_z = 0 Å^–1 for the peak at highest Q_xy value. All other reflections occur at Q_z positions above zero (0.11, 0.35, 0.45Å^–1), indicating tilt. As the surface pressure increases to 30 mN/m, the two reflections at lowest Q_xy merge. The two reflections at 1.513 Å^–1 and 1.489 Å^–1 are located at Q_z = 0 Å^–1, and the third reflection at 1.475 Å^–1 exhibits a Bragg rod maximum at 0.25 Å^–1. At the highest pressure investigated ( = 40 mN/m) only two reflections at 1.515 and 1.501 Å^–1 could be resolved, both having their Bragg rod maximum at Q_z = 0. The two distinct peaks at the highest surface pressure suggest that there are two distinct crystal lattices in the film at all surface pressures. Survanta is a complex mixture of bovine extract surfactants enriched with saturated DPPC, PA, and tripalmitin. These saturated lipids are known to form condensed phases at the air-water interface by themselves (Helm et al., 1987; Brezesinski et al., 1995; Teer et al., 2000; Lee et al., 2002) and in mixtures with each other and with unsaturated lipids (Bringezu et al., 2001). At low surface pressure, PA packs into a centered rectangular lattice with tilted chains in direction toward nearest neighbor (NN). With increasing pressure, a phase transition due to a change of the tilt direction from NN toward next nearest neighbor (NNN) occurs, before a second phase transition leads to an upright orientation of the molecules (Teer et al., 2000) at 22 mN/m. However, pure PA monolayers collapse at surface pressures <40 mN/m, so the reflections seen in Fig. 3 are not from pure PA crystals. Pure DPPC monolayers exhibit an oblique lattice that is only slightly distorted from centered rectangular symmetry. In this case, mismatch of the space requirements between the bulky hydrated PC headgroup and the two chains is compensated by a large tilt of the chains of 27° ( = 30 mN/m). Even at 45 mN/m, the chains are tilted (Brezesinski et al., 1995). However, incorporation of alkanes (Thoma and Möhwald, 1995; Brezesinski et al., 1996), fatty acids, or fatty alcohols (Bringezu et al., 2001; Lee et al., 2002) into the distorted chain lattice reduces the mismatch between head and tail, leading to decreased tilt angles and more stable crystals.

The diffraction patterns observed for Survanta suggest a demixing of the condensed phases into a phase with lattice spacing and tilt very similar to that of 1:1 (mol/mol) DPPC:PA (phase I) and a second, somewhat more condensed phase, which likely contains the remaining DPPC and the tripalmitin (phase II). The three reflections at 1.428 Å^–1, 1.454 Å^–1, and 1.482 Å^–1 are assigned to the low order (10), (01), and (11) reflections of phase I at 20 mN/m, corresponding to a lattice with aligned chains tilted in nonsymmetry direction (oblique). For such intermediate tilt azimuth, the relation between the Bragg rod maxima is Q_z1 = Q_z2 + Q_z3 with Q_z1 being the peak with the largest Q_z value (Kaganer et al., 1999). Here, Q_z1 = 0.45 Å^–1, Q_z2 = 0.35 Å^–1, and Q_z3 = 0.11 Å^–1, which satisfies Q_z1 = Q_z2 + Q_z3 and suggests that the three reflections come from the same lattice. From the peak positions, a tilt of 19° is calculated for phase I at 20 mN/m with a projected area per chain, A_xy, of 21.5 Ų and the chain area perpendicular to the chain, A₀, of 20.4 Ų. The lattice spacings, tilt, projected area per chain, and area per chain are similar to that of a 1:1 (mol/mol) DPPC/PA cocrystal (Bringezu et al., 2001; Lee et al., 2002). With increasing pressure, the Bragg peak positions of the tilted phase shift to larger values, consistent with the DPPC/PA cocrystal. At 30 mN/m, the titled phase exhibits two reflections, the degenerate and the nondegenerate (02) reflections of a centered rectangular lattice with NN titled chains and a tilt of 11°. At 35 mN/m, this rectangular phase shows a decreased tilt angle of 8.4°. At 40 mN/m, the two reflections merge, and both exhibit their Q_z maxima on the horizon, indicating untilted chains. As is the case for 1:1 DPPC/PA at 40 mN/m, the lattice is hexagonal and the molecules are untilted with A_xy = 20.2 Ų (Lee et al., 2002).

The remaining peak, at 1.509 Å^–1 at 20 mN/m, which moves to 1.515 Å^–1 at 40 mN/m, is assigned to a second, untilted hexagonal phase. At 20 mN/m, the hexagonal lattice parameter is 4.81 Å and the area perpendicular to the chains A_xy = A₀ = 20.0 Ų. The lattice remains hexagonal and untilted as the surface pressure increases, and the lattice parameters slightly decrease to 4.79 Å and A_xy = A₀ = 19.9 Ų. This second phase might be the tripalmitin and any remaining DPPC, however, there are no established values of mixed tripalmitin/DPPC films available in the literature. There does not appear to be any pure DPPC crystals in the film as there are no nonzero Q_z components at this surface pressure, and pure DPPC crystals remain tilted at 40 mN/m. However, the diffraction clearly shows phase separation into one or more crystalline phases of the saturated lipids. Thus, the interfacial film consists of crystalline condensed domains coexisting with a disordered fluid phase, as is the case for simple lipid mixtures containing DPPC and PA (Bringezu et al., 2001).

Atomic force microscopy
To obtain higher-resolution images of the film and a quantitative estimation of the range of domain sizes and thicknesses, Survanta films were imaged by AFM after Langmuir-Blodgett deposition at various surface pressures. During the transfer, material located near the air-water interface can be trapped between the mica substrate and the interfacial film; therefore, AFM can also reveal the three-dimensional structure of the monolayer and the associated reservoir material. Although this complicates interpretation, AFM is the only tool that can map the three-dimensional film morphology, as many of the features seen in AFM are below the resolution of optical imaging techniques.

Representative AFM images at various stages of the first compression are shown in Fig. 4. The low-resolution image at low surface pressure (15 mN/m) shows a bright network along with discrete, dark circular and elliptical domains, similar to the fluorescence images in Fig. 2. In AFM images, bright means higher or thicker, and dark means thinner, as is quantified in the height traces in the third column. At 15 mN/m, the bright network is relatively smooth and is 15 nm higher than the circular domains. From the correspondence between the fluorescence and AFM images, the bright, continuous phases are likely fluid, whereas the discrete dark domains in both FM and AFM images are solid phase, as shown by the GIXD results. The much greater thickness of the fluid phase is likely due to material underneath the monolayer trapped during LB deposition. The thick fluid domains of the films formed by adsorption are quite different from AFM images of lipid or lipid/protein films spread from solvent, in which the entire film is a single monolayer at low surface pressure (Zasadzinski et al., 1994; von Nahmen et al., 1997; Lipp et al., 1998; Ding et al., 2001, 2003; Takamoto et al., 2001; Cruz et al., 2004). Adsorption likely occurs via the fluid phase and the multiple layers are associated with spreading from a multibilayer surfactant aggregate. This morphology is similar to films containing fluid phase lipids and SP-C that form multilayer patches in the fluid phase, but only at high surface pressure (von Nahmen et al., 1997; Ding et al., 2001; Takamoto et al., 2001). Most of the protein is located in the fluid phase, so if the proteins are involved in adsorption or stabilization of multilayer reservoirs (Schürch et al., 1995, 1998; von Nahmen et al., 1997; Ding et al., 2001; Takamoto et al., 2001), it is reasonable for the fluid phase to be the most likely location for adsorption. Quantitative Brewster angle microscopy has also shown 12-nm-thick films adsorbing at the air-water interface from lavage fluids deposited in the subphase (Winsel et al., 2003).

At 15 mN/m, the solid domains exhibit a hemispherical mound in the center of the domains. The measured thickness variations of >30 nm across the image demonstrates that even at low pressure, the film is not a simple monolayer, but rather a monolayer with an underlying reservoir of multilayers (Schürch et al., 1995, 1998; von Nahmen et al., 1997; Ding et al., 2001, 2003; Winsel et al., 2003).

At 30 mN/m, the images show isolated solid domains surrounded by a fluid matrix, which is 13 nm higher than the condensed phase and is still smooth. By comparison with the lower pressure image, this indicates a thinning and likely loss of reservoir material. The centers of the solid domains lose reservoir material and the hemispherical mounds on the solid phase domains dissipate. If a 5-min waiting period is imposed before film transfer, the solid domains completely lose the reservoir material (image not shown).

The corresponding low-resolution image (50 x 50 µm) shows the close correspondence between the AFM images and the fluorescence images (Fig. 2). As in the fluorescence images, the area fraction of the solid phase domains increases with increasing surface pressure. As the surface pressure increases to 40 mN/m, at the start of the plateau in the isotherm (Fig. 1, zone C), the morphology of circular solid domains in a continuous fluid phase matrix remains. The variation in height across the film decreases to 5 nm. As the surface pressure is increased to 44 mN/m, the fluid phase roughens and is again 16 nm higher than the center of the condensed areas. The condensed areas are rough as well and start to coalesce. For all films transferred at surface pressures corresponding to the plateau region in the isotherm (Fig. 1), distinct areas of solid and liquid are seen, but neither region is smooth. By the end of the plateau at 47 mN/m, no discrete solid domains are left. The film is flat with some small protrusions and is likely now a single monolayer. In the low-resolution image, the protrusions appear to trace out shapes similar to those of the solid phase domains at lower surface pressures. The film appears to be almost completely solid with only isolated patches of the reservoir material. The small protrusions are likely remnants of the fluid material that was ejected, possibly stabilized by SP-B (Lipp et al., 1998; Ding et al., 2001, 2003) or SP-C (von Nahmen et al., 1997; Takamoto et al., 2001). In earlier work, adding SP-B to Survanta helped to keep the fluid phase continuous, separating the solid phase domains at high surface pressures (Ding et al., 2001). At collapse at 70 mN/m, few bright spots remain, again apparently lining the boundaries between the compressed solid domains.

Fig. 5 helps quantify the morphological changes happening within the film at the plateau. The AFM images were digitally thresholded to enhance the contrast between liquid and condensed phases to more accurately quantify the condensed area fraction of the film. Below the plateau, the interfacial film is composed of circular condensed phase domains in a fluid matrix. As the surface pressure increases, the average diameter of the domains increases. By the end of the plateau, the domains coalesce, and the fluid phase is no longer continuous. The condensed area fraction at different surface pressures is listed in Table 2 and plotted in Fig. 6 (labeled A; solid circles). Two regimes are easily distinguished according to the slope. The solid fraction increases gradually with surface pressure up to 35 mN/m, then much more rapidly. The film is effectively all solid phase above the plateau pressure. The original fraction of DPPC in Survanta is 70%. Taking into account the fraction of PA and tripalmitin, which are also likely in the solid phase, the maximum solid phase fraction should be 85%. Hence, some of the fluid phase is pushed away from the interface to the subphase. The surfactant films below and above the plateau have different compositions, and both may be different from the composition of the surfactant in the subphase.

View larger version (21K): [in this window] [in a new window]   FIGURE 5  Relationship between the morphology of the solid domains in a Survanta film and isotherm features. AFM pictures were taken below (20 and 34 mN/m) and above (>40 mN/m) the plateau in the isotherm. The images were thresholded and the fractional area of the dark phase was determined to evaluate the solid phase fraction as a function of surface pressure (see Table 2 and Fig. 6). Below the plateau, the size of the solid domains (black areas) increases with the surface pressure. The plateau corresponds to the coalescence of these domains.

View larger version (15K): [in this window] [in a new window]   FIGURE 6  Solid area fraction (measured by AFM) and surface viscosity () plotted as a function of surface pressure. For low the solid fraction area increases steadily and increases quickly at the plateau pressure, indicating the percolation of the solid domains and the expulsion of fluid phase from the interface (see Fig. 4). increases exponentially (linear on this semilog plot) up to the plateau, then jumps by more than an order of magnitude at the plateau pressure. The viscosity transition is correlated with the large change in the solid fraction area and the jamming of the solid phase domains.

Surface viscosity
The consequences of the solid-fluid phase coexistence on the mechanical properties of the film are profound (Diamant et al., 2000; Ding et al., 2002a; Lu et al., 2002; Alonso and Zasadzinski, 2004). The surface viscosity of the film (; open circles) determined from the magnetic needle viscometer is plotted as a function of surface pressure along with the solid phase fraction (Ding et al., 2002a; Alonso and Zasadzinski, 2004) in Fig. 6. The different regimes observed correspond closely to the different zones identified on the isotherm (see Fig. 1). Below 10 mN/m (zone A) the surface viscosity is constant at a low value, similar to other continuous fluid phase monolayers (Ding et al., 2002a,b; Alonso and Zasadzinski, 2004). In zone B, between 10 and 38 mN/m, the film viscosity increases exponentially with surface pressure (linearly on the semilog plot in Fig. 6); the surface viscosity at 38 mN/m is about one order of magnitude greater than at 10 mN/m. However, at the plateau surface pressure (zone C), at which the condensed phase fraction shows a dramatic change, the viscosity suddenly increases by almost two orders of magnitude. At the end of the plateau, the film is so viscous that the needle cannot move even at the highest force available with the magnetic needle viscometer. The jamming of the solid domains is responsible for the sudden increase in the surface shear viscosity (Ding et al., 2002a,b; Alonso and Zasadzinski, 2004).

The direct correspondence between solid phase fraction and the surface viscosity is similar to other monolayers exhibiting solid-fluid phase coexistence (Ding et al., 2002a,b; Alonso and Zasadzinski, 2004). The film mechanical properties at coexistence are the two-dimensional analog of a three-dimensional suspension of solids in a liquid (Ding et al., 2002a,b; Alonso and Zasadzinski, 2004). In three-dimensional suspensions, as the solid volume fraction, , approaches the volume fraction for random close packing, _c, the viscosity diverges as ; µ is the steady shear viscosity of the dispersion, µ_o is the viscosity of the suspending liquid, and the exponent, a, varies from 1 to 2 depending on the interactions between the solids (Brady, 1993). The surface viscosity of lipid monolayers at coexistence for a variety of compositions, surface pressures, and temperatures diverges in a similar fashion: ; _o is the surface viscosity of the continuous, fluid phase, and is the effective surface viscosity of the monolayer (Ding et al., 2002a,b). A is the area fraction of the solid phase and A_c is a critical solid phase fraction at which the solid domains jam into each other and the viscosity diverges. Relatively small changes in the fluid/solid ratio can lead to order of magnitude changes in the surface viscosity near A_c (Ding et al., 2002a,b). Fig. 8 A shows the measured surface viscosity of Survanta as a function of the solid phase area fraction. The dotted line is a fit to with ₀ = 0.0003 mN-s/m (equivalent to surface poise) and A_c = 0.9.

View larger version (12K): [in this window] [in a new window]   FIGURE 8  (A) Surface viscosity () plotted as a function of the solid fraction area, A, for a Survanta film deposited on saline buffer at 25°C. The dotted line is a fit to with 0 = 0.0003 mN-s/m and Ac = 0.9 (Ding et al., 2002a). It was difficult to prepare films for AFM evaluation over the narrow range in surface pressure during which the solid phase fraction changed from 70 to 90% (see Table 2). (B) Surface shear viscosity () measured as a function of surface pressure () for a Survanta film deposited on saline buffer as a function of temperature for 20, 25, 30, and 37°C. The basic features of the surface viscosity response are the same at all temperatures. The surface pressure at which the viscosity transition occurs increases with increasing temperature, consistent with the shift of the plateau in the isotherm and the critical area fraction toward higher surface pressure with temperature. (C) Surface shear viscosity () of Survanta film at 25°C for three consecutive compression-expansion cycles. The viscosity transition occurs at a lower surface pressure for the second and third cycles, consistent with the differences in the isotherms in Fig. 1. The overall viscosity is higher as well. This is consistent with the loss fluid phase as shown in the isotherms in Fig. 1 and the images in Figs. 2 and 4.

Although temperature and composition affect the shear viscosity, it is primarily by their influence on the solid phase area fraction, A, and A_c. For Survanta, the isotherms in Fig. 1 are similar and show the same general features (zones A–E) at any temperature from 20 to 37°C. The slopes below and above the plateau get steeper as the temperature is decreased and the plateau surface pressure increases with increasing temperature: 38 mN/m at 25°C, 40 mN/m at 30°C, and 42 mN/m at 37°C. Lower temperatures increase the crystallinity of the film, thus the fraction of condensed phase increases when the temperature drops (Kaganer et al., 1999; Ding et al., 2002a). Hence, the surface pressure at which the surface viscosity increases should increase with increasing temperature, as shown in Fig. 8 B. At 20°C, jumps at 20 mN/m, at 35 mN/m at 25°C, at 40 mN/m at 30°C, and 43 mN/m at 37°C. In all cases, the jumps are associated with the plateau in the isotherm and the rapid increase in solid phase fraction. At all temperatures, the surface viscosity at the limiting surface pressure is so great that the needle in the viscometer cannot move even under the highest forces applied by the magnetic coils.

As with the isotherms, successive cycles also change the surface viscosity (Fig. 8 C) at 25°C; the surface pressure at the viscosity transition decreases by 10 mN/m from the first to the second cycle, and is relatively constant after the first cycle. This is consistent with the isotherms (Fig. 1) and fluorescence images (Fig. 2) that suggest a decrease in the fluid phase and an increase in the solid phase fraction on cycling. lifts off at 25 mN/m on the second and third compressions, well below the 38 mN/m surface pressure at which the surface viscosity changes on the first compression. This is consistent with the loss of fluid phase during the first cycle. The condensed fraction area is increased with respect to the fluid phase area at a given surface pressure on the second and subsequent cycles, which lowers the surface pressure at which the percolation threshold is reached. This shows again that the first compression removes any excess material from the interface and refines the composition for subsequent cycles. The composition variation then leads to important changes in the mechanical properties of the film.

	CONCLUSIONS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL RESULTS AND DISCUSSION CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES

 
Survanta contains 65% DPPC, 10% fatty acid (PA), 10% saturated tripalmitin, a relatively small fraction of unsaturated lipids, and 1% of the surfactant protein C. This balance of components is suitable to achieve two required properties of a good lung surfactant: the ability to reduce surface tension to value near zero and the ability to respread efficiently. These key features can be understood by the characteristic morphology of Survanta as a function of surface pressure: a distinct change from a continuous fluid phase separating discrete solid phase domains at low surface pressure to a percolated solid phase at higher surface pressures. Adsorption from an aqueous suspension leads to the formation of a multilayer reservoir of fluid phases at low surface pressure that are eventually removed at higher surface pressure. The DPPC and PA added to Survanta enhance the