Originally published as Biophys J. BioFAST on August 12, 2005.
doi:10.1529/biophysj.104.052290
Biophysical Journal 89:3159-3173 (2005)
© 2005 The Biophysical Society
Interaction between Lipid Monolayers and Poloxamer 188: An X-Ray Reflectivity and Diffraction Study
Guohui Wu *,
Jaroslaw Majewski 
,
Canay Ege *,
Kristian Kjaer 
,
Markus Jan Weygand and
Ka Yee C. Lee *
* Department of Chemistry, the Institute for Biophysical Dynamics and the James Franck Institute, the University of Chicago, Chicago, Illinois; Manuel Lujan Jr. Neutron Scattering Center, Los Alamos National Laboratory, Los Alamos, New Mexico; and Materials Research Department, Risø National Laboratory, Roskilde, Denmark
Correspondence: Address reprint requests to Ka Yee C. Lee, Dept. of Chemistry, the Institute for Biophysical Dynamics and the James Franck Institute, the University of Chicago, 5735 S. Ellis Ave., Chicago, IL 60637. E-mail: kayeelee{at}uchicago.edu.
 |
ABSTRACT
|
|---|
The mechanism by which poloxamer 188 (P188) seals a damaged cell membrane is examined using the lipid monolayer as a model system. X-ray reflectivity and grazing-incidence x-ray diffraction results show that at low nominal lipid density, P188, by physically occupying the available area and phase separating from the lipids, forces the lipid molecules to pack tightly and restore the barrier function of the membrane. Upon compression to bilayer equivalent pressure, P188 is squeezed out from the lipid monolayer, allowing a graceful exit of P188 when the membrane integrity is restored.
|
INTRODUCTION
|
|---|
In victims of electrical shock or lightning injury, altered membrane properties in response to strong electric field exposure appear to be directly relevant to tissue injury mechanisms. Electrical injuries to cell membrane can result in the following events: thermal burns secondary to Joule heating (1
–3), permeabilization of cell membranes (4), and/or denaturation of macromolecules such as proteins (5). Electroporation (suggesting actual pore formation) and electropermeabilization (likely due to the loss of the innate lipid packing density) are the two terms commonly used to describe the biophysical process of enhanced membrane permeability due to electric field-driven reorganization of lipids in the lipid bilayer by supraphysiologic electric fields (6–8). Contemporary electroporation theory indicates that polar water molecules are pulled by Kelvin polarization stress into transient defects in the lipid packing order within a bilayer, leading to quasistable or stable pore formation (9–12).
A break in the integrity of the plasma membrane immediately compromises its essential role as a barrier, thereby severely affecting its control over the exchange of materials between the cell and its surrounding environment, which eventually causes cell death. Sealing of permeabilized cell membranes is therefore of vital importance to medical treatment of electrical shock victims.
The successful restoration of membrane transport properties through surfactant poloxamer 188 was first demonstrated in 1992 (2). Poloxamer 188 (P188, mol wt = 8400 g/mol), a member of a triblock copolymer family called poloxamers, or pluronics, has the structure of poly(ethylene oxide)-poly(propylene oxide)-poly(ethylene oxide) (PEO-PPO-PEO, with 80 wt % PEO content), and has a firmly established medical safety record since 1957. As a result of its amphiphilic nature, P188 has the capability to interact with a cell membrane and is found to alleviate as well as reverse damages caused by electrical trauma. It was demonstrated that P188 can seal cells against loss of carboxyfluorescein dye after electroporation (2,13–18). In the intervening years, P188 has been determined as a successful sealing agent for various permeabilized cells, such as skeletal muscle tissue after heat shock and intense irradiation (19,20), red blood cell from sickle cell disease (21), etc. Although these results show that P188 is effective in restoring damaged cell membranes, the underlying mechanisms have barely been studied until recently (22).
Besides being used as a sealing agent, poloxamers have also received much attention during the past decade for their application in drug delivery. It is known that liposomes (used as drug carriers) can be modified with poloxamers to prolong their circulation time in the blood stream as well as to prevent their flocculation, fusion, or binding (13–18). Although the mode of interaction between poloxamer and liposome has been hypothesized, it remains poorly understood.
Recently there have been studies to measure the ability of poloxamers to interact with model membranes (22,23), however, no effort has yet been made to elucidate the lipid/poloxamer interaction at the molecular level, which is critical in improving the design of poloxamers for medical applications. To gain insight into the mechanisms of interaction between poloxamer and damaged membrane, we have used a Langmuir lipid monolayer at the air-water interface as a model system and have performed a series of x-ray reflection (XR) and grazing incidence x-ray diffraction (GIXD) experiments to gain information at the molecular level.
The Langmuir lipid monolayer serves as a good model for the outer leaflet of the cell membrane. By controlling the surface area available for the lipid monolayer, the intact membrane (with tightly packed lipid molecules at the bilayer equivalent pressure of 30–35 mN/m (24,25), as well as the disrupted portion of the membrane post-trauma (with low lipid density) can be effectively mimicked. The low-density lipid film at the air-water interface represents the hydrophobic-aqueous environment that the poloxamer sees on a damaged membrane. Such a hydrophobic-aqueous environment can be found both at the highly curved lipid portion that constitutes the pore, and in the portion of the membrane that has enhanced permeability due to a reduction in lipid density. Both sites have enhanced exposure of the lipid hydrocarbon tail to the aqueous environment, into which the poloxamer, being amphiphilic in nature, can possibly insert.
X-ray beams from intense and well collimated synchrotron sources have been used as a molecular probe for direct structural information on the organization of amphiphilic molecules at the air-water interface (26–29). Our earlier work has demonstrated that XR and GIXD can help gain insight into the interaction between lipid and poloxamer (30). Here we report detailed structural parameters for pure lipid and lipid/P188 monolayers at different surface pressures. To the best of our knowledge, these studies are the first complete in situ investigations providing quantitative information on the insertion of P188 into lipid monolayers, as well as the effect of P188 on the packing of lipid molecules. Our data also help to pinpoint the location of P188 in the lipid matrix.
|
MATERIALS AND METHODS
|
|---|
Materials
1,2-Dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) and 1,2-dipalmitoyl-sn-glycero-3-[phospho-rac-(1-glycerol)] (DPPG) were purchased in powdered form from Avanti Polar Lipids (Alabaster, AL) and used without further purification. Monolayer spreading solutions were prepared by dissolving the solid lipid in either pure chloroform (for DPPC) or 9:1 v/v chloroform-methanol (for DPPG) to obtain a concentration of 0.5 mg/ml. Throughout the experiments, ultrapure water (resistivity 
18 M
cm) obtained from a Milli-Q UV Plus system (Millipore, Bedford, MA) was used as the subphase.
P188 solution of 200 mg/ml was prepared by adding poloxamer (BASF, Parisippany, NJ) and ultrapure water to a vial containing a magnetic stir bar, and then left to mix on a stirplate for 0.5 h to ensure complete dissolution. The poloxamer solution was kept at 4°C before use, and made fresh every 2 weeks.
Experimental configuration
Surface-pressure area (Langmuir) isotherms were obtained with a home-built Langmuir trough equipped with a Wilhelmy plate (31). In a typical isotherm experiment, the lipid molecules dissolved in chloroform were deposited at low surface density at the air-water interface, leaving an initial surface pressure of 0 mN/m. The solvent was allowed to evaporate for 5–15 min. P188 was then injected into the subphase by inserting a syringe through the air-water interface, and the final P188 concentration was 50 µM in the subphase. All experiments were performed at 30°C. For isotherm measurements the compression was achieved with two barriers moving at a rate of 6.3 mm2/s. Depending on the initial amount of material deposited at the interface, this translates to a lipid molecular area reduction rate between 0.04 and 0.06 Å2·s–1.
Experiments were also performed at a synchrotron source. All x-ray scattering experiments were carried out at the BW1 (undulator) beam line at HASYLAB (Hamburg, Germany) using a dedicated liquid surface diffractometer (29,32–34) with an incident x-ray wavelength of 
1.3 Å. A thermostated trough equipped with a Wilhelmy balance for measuring the surface pressure (
) and a barrier for changing the surface area was mounted on the diffractometer.
Two types of x-ray scattering experiments were performed. The first set of experiments, referred to as pretreatment experiments in our previous work (22), was designed to study the pure lipid monolayer, as well as the interaction between P188 and lipid in a mixed film. The second set of experiments, aimed at studying the structure of P188 at the air-water interface, was performed in a manner similar to the first but with no lipid present. For both sets of experiments, after the sample was prepared but before x-ray data were collected at least 40 min were allowed for the trough enclosure to be flushed with helium to reduce the scattering background and to minimize beam damage during x-ray scans; x-ray data were collected afterward.
Both XR and GIXD were carried out to obtain out-of-plane as well as in-plane electron density distribution information on these samples. X-ray scattering theory and the liquid diffractometer used here have been described previously (26–29). The reflectivity data were analyzed using a kinematical approach (26,35). The electron density 
(z), which is laterally averaged over both the ordered and disordered parts in the footprint of the beam, can be modeled by a stack of homogeneous slabs ("boxes") with each having a constant electron density and thickness. The interfaces between boxes are smoothed with a Gaussian function of standard deviation 
to account for roughness at the boundary due to thermally excited capillary waves and atomic roughness of the interface (36,37).
The GIXD intensity resulting from a powder of 2-D crystallites may be represented by its projection onto the 
axis to yield Bragg peaks or onto the 
axis to yield Bragg rods, where qxy and qz are the horizontal and vertical components of the scattering vector q, respectively (26,35,38). The intensity distribution along a Bragg rod can be analyzed to infer the direction and magnitude of the molecular tilt in the crystalline part of the monolayer. In this work we model the lipid tail by a cylinder of constant electron distribution with adjustable parameters being the tilt angle of the cylinder from the vertical, the lateral tilt direction, the length of the cylinder, Lc (i.e., the length of the part of the molecule that scatters coherently), and the vertical root mean-square displacement, 
(Debye-Waller factor), in the crystallites. The horizontal positions of the Bragg peaks () allow for the determination of the spacings 
for the 2-D lattice of the alkyl chains. The width of Bragg peaks, corrected for the instrumental resolution, gives the finite size of the crystalline domains in the direction of the reciprocal scattering vector (the 2-D crystalline coherence length Lxy) (39).
The dimensions of the footprint of the incoming x-ray beam on the liquid surface were 2 x 50 mm2. As a precaution against beam damage, in GIXD the trough was translated by 0.025 mm horizontally across the x-ray beam, in the direction along the barrier compression at every step of the scan; in XR, the sample was completely renewed by occasional translation of the full 2-mm width of the beam. Remeasuring part of the reflectivity curve before and after translation afforded a check of the reproducibility.
|
RESULTS
|
|---|
Isotherm
The effect of P188 on the phase behaviors of pure DPPC and DPPG monolayers was studied via isotherm measurements, and results are shown in Fig. 1, A and B, respectively. In the pretreatment experiments (22), the lipid was spread at a high area/lipid molecule ( = 0 mN/m), and P188 was added to the water subphase. For both lipid monolayers, the addition of P188 to the subphase resulted in the partitioning of P188 to the air-water interface, causing an instantaneous increase in surface pressures from 0 to 22 mN/m; after waiting for 40 min to simulate conditions at the synchrotron source, the surface pressure further rose to 26 mN/m. Under this condition, fluorescence microscopy revealed surface heterogeneities at the length scale as large as 15 mm as discussed in Appendix A. The lipid/poloxamer system was then compressed until collapse. Fig. 1, A and B, shows that as the DPPC monolayer was compressed to 36 mN/m and beyond, or DPPG to 38 mN/m and beyond, the isotherms of the poloxamer-pretreated monolayers reverted to those of the pure lipids and the surface heterogeneities were eliminated, suggesting that P188 is eliminated from the system. Although the isotherms provide macroscopic information on the lipid-poloxamer interaction as a function of surface pressure, XR and GIXD experiments are needed to provide details at the molecular level regarding the fate of the poloxamer at high surface pressures, and to pinpoint the location of the poloxamer in the lipid matrix as well as the subsequent effect on lipid ordering upon P188 insertion.
It should be noted that the adsorption of the injected poloxamer to the air-water interface is a time-dependent process. Fig. 2 shows the change with time in surface pressure of a DPPC/P188 film. The lipid was spread at a high area/DPPC molecule (ADPPC) at 109.6 Å2, resulting in a zero surface pressure. Upon the injection of P188 into the subphase, the surface pressure rose almost instantaneously from 0 to 22 mN/m. After that, the surface pressure increased slowly to 26 mN/m in 40 min (see Appendix C).
View larger version (13K):
[in this window]
[in a new window]
|
FIGURE 2 Change in surface pressure for a P188-treated DPPC monolayer on a water subphase at 30°C with time. P188 was introduced into water subphase at time = 0 when the pure lipid film gave a surface pressure of 0 mN/m.
|
|
As a control, the isotherm of pure P188 at the air-water interface was measured. P188 was injected into the subphase directly without any phospholipid monolayer spread at the air-water interface. After 40 min ( = 26 mN/m), the barrier compression was started and isotherm data were recorded. A weak dependence of surface pressure on area is observed: increases from 26 to 29 mN/m with barriers compressed from a maximum trough area of 438 cm2 to the minimum of 109 cm2.
GIXD and XR of poloxamer 188
X-ray scattering experiments were performed on pure P188 films with the same experimental condition as that for isotherm measurements except that x-ray measurements were carried out during the compression of P188 film. XR and GIXD were measured at = 26, 27, and 29 mN/m, corresponding to the maximum, intermediate, and minimum trough area, respectively. Bragg peaks were absent for all the investigated conditions when only polymer was present in the water subphase.
The x-ray reflectivity data for pure P188 in a water subphase at 30°C are shown in Fig. 3. The reflectivity profile of P188 is very similar to that of the bulk water subphase, and can be fitted by a one-box model, with a thickness of 21.6 ± 1.0 Å and normalized electron density of 1.04 ± 0.01. (All electron densities reported are normalized to that of the water subphase, water = 0.334e–/Å3.) Although the increase of surface pressure upon the addition of P188 clearly shows the presence of the polymer at the air-water interface, the low electron density contrast between P188 and water causes a very small deviation in the reflectivity of P188 (R) from that of the free surface of water (the Fresnel curve augmented with interfacial roughness 3 Å, giving a reflectivity 
). The low contrast is seen in the fitted parameter P188/H2O = 1.04 ± 0.01. The fitting results also show that the P188-water interface ( = 7.9 ± 1.0 Å) is much rougher than the air-P188 interface ( = 3.0 ± 0.1 Å ). Such roughness precludes the observation of any fringe corresponding to 22 Å thick P188 layer in the XR data.
Grazing-incidence x-ray diffraction
Pure DPPC
GIXD measurements provide in-plane structural information on the crystalline, diffracting portion of the monolayer. Fig. 4, A and B, shows Bragg peaks for a pure DPPC and P188-treated DPPC monolayer, respectively, at various areas/lipid molecule. The structural parameters are summarized in Table 1. In Fig. 4 A, a pure DPPC monolayer at ADPPC = 63 Å2 exhibits no Bragg peak, indicating the lack of any lipid ordering. Compression to ADPPC = 58 Å2 results in a similarly flat profile, though some features (still within the noise level) are becoming discernable.
View larger version (29K):
[in this window]
[in a new window]
|
FIGURE 4 Bragg peaks from GIXD on a water subphase at 30°C for (A) pure DPPC and (B) DPPC/P188 films at various ADPPC. For clarity, the data have been offset vertically. The two GIXD Bragg peaks observed for DPPC at 40 mN/m indicate a distorted hexagonal packing of the lipid tails in a 2-D unit cell with parameters a = b = 5.05 Å,  = 116.6°. The Miller indices {h, k} (distorted hexagonal lattice) are indicated for each peak. In B, the diffraction peak observed at ADPPC = 107 Å2 corresponds to a highly condensed lipid phase comparable to pure DPPC at ADPPC = 47 Å2, despite the large nominal ADPPC. The molecular packing parameters used in the fitting are listed in Table 1.
|
|
At ADPPC = 48 Å2, two Bragg peaks are observed at qxy = 1.38 Å–1 and 1.46 Å–1. The pattern thus consists of two low-order reflections, indicating a distorted-hexagonal cell (see Appendix D). The integrated intensity of the Bragg peak at 
is roughly twice that of the peak at 
This higher intensity results from coincident four reflections 
(simplified as {10,01}) versus two reflections 
(simplified as 
) and leads to the assignment of the 
reflection to the peak and the reflection to the 
peak. The spacings (
) are 
and 
giving rise to a distorted hexagonal unit cell with axes 
and = 
and a unit cell area (containing one chain), 
of 23.1 Å.
When resolved in both qxy and qz (cf. Figs. 5 A and 6 below), the peak at qxy = 1.38 Å–1 can be seen to consist of two barely resolved peaks, indicating that the lattice is oblique (see Appendix E). However, in this report this splitting is ignored for simplicity, leading to a distorted-hexagonal lattice, as indicated above. This holds true for the rest of DPPC films investigated in our system.
View larger version (42K):
[in this window]
[in a new window]
|
FIGURE 6 (A) Scattering intensity as a function of in-plane scattering vector component qxy for different qz intervals (intensity integrated over the vertical scattering vector qz in 10 successive 0.1-Å–1-wide qz-windows, as indicated) for the P188-treated DPPC film on a water subphase at 107 Å2/DPPC molecule and 30°C. (B) Background subtracted Bragg rod intensity distribution along qz vector integrated over the qxy range of the Bragg peaks. The rod was fitted (solid line) by approximating the coherently scattering part of the DPPC tail by a cylinder of a constant electron density.
|
|
The corresponding full widths at half-maximum height (FWHM) of the two peaks are 
and 
both of which exceed the instrumental resolution of 
The intrinsic FWHM can be obtained from the equation
 | (1) |
A simple model assumes that the monolayer consists of 2-D crystallites that are perfect and have a finite average size Lxy, the coherence length. Using the Scherrer formula (39), we can calculate the coherence length 
in the two crystallographic directions {10,01} and {
}, and they were found to be 54 Å and 240 Å, respectively.
The Bragg rod profiles, integrated over the qxy region corresponding to the two reflections, are shown in Fig. 5 B. The {1
}Bragg rod achieves its maximum intensity near 
indicating that the molecular axis lies in a plane nearly parallel to the distortion direction of hexagonal unit cell. We analyzed the {10,01} and {1} Bragg rod intensity profiles by approximating the fatty acid tail with a cylinder of constant electron density (26). Our analysis shows that the lipid molecule has a tilt of 30.0° from the surface normal and the projected tilt direction on the water subphase is 11.9° from the nearest neighbor. The effective coherently scattering length of the DPPC molecule, Lc, is 18.9 ± 1.0 Å. The tilt direction is in a nonsymmetry direction, as in the rest of the DPPC films, which again indicates that the rectangular symmetry is broken and an oblique 2-D crystal structure was formed instead (41).
When the pure DPPC film is compressed to ADPPC = 47 Å2, Bragg peaks occur almost at the same positions as those at ADPPC = 48 Å2 (Fig. 4 A), and subject to experimental error, the calculated d-spacings are almost identical. Tilt angle of the chains from the vertical remains large (28.3°) even at = 40 mN/m. Previous work on DPPC has also shown large tilt angles at high lateral pressures (41,42), which can be understood via a model where the carbonyl group of the ester at the C2 position pulls the attached chain toward the water subphase (41). This suggests that the packing of DPPC molecules is mainly limited by the headgroup which maintains a large cross-sectional area at high surface pressures. After the headgroups are compressed to tight contact with each other, there remains ample space for the alkyl tails to assume large tilt angles relative to the surface normal, and the size of unit cell of the alkyl chain cannot decrease further due to the limitation of headgroup packing.
DPPC/P188 film
The Bragg peaks obtained for the P188-treated DPPC film at different surface pressures are shown in Fig. 4 B. Upon spreading, the nominal area for DPPC is ADPPC = 107 Å2. At such a large area/molecule and in the absence of P188, the DPPC film is expected to be in the gas/liquid expanded coexistence phase, with the surface pressure at 0 mN/m and with no Bragg peak present. Upon the introduction of P188 in the subphase, the surface pressure rose from 0 to 26 mN/m without any compression, and two distinct Bragg peaks are observed at 
and The calculated d-spacings are 4.59 Å and 4.30 Å, which are identical to those found in Bragg peaks obtained from a pure DPPC monolayer at a much lower ADPPC. The occurrence of precisely two Bragg peaks resulting from the {10,01} and {1} reflections indicates a distorted hexagonal unit cell with axes 
and an interfacial area/molecule of 46.7 Å2. Since the size of the unit cell obtained for DPPC/P188 film is almost identical to that found for pure DPPC at ADPPC = 46.1 Å2, coupled to the fact that pure P188 does not form ordered structure, Bragg peaks in the DPPC/188 film have to come from the DPPC crystals, and P188 does not interfere with the packing of the ordered DPPC molecules. Instead, P188 is incorporated in the disordered phase of the monolayer. The observed Bragg peaks at a large area/DPPC molecule can only be explained in one way: P188, phase-separated from DPPC, physically occupies a portion of the available surface area, and the smaller area left for the lipid forces DPPC molecules to form a condensed phase.
Fig. 6 depicts the diffracted intensity as a function of the in-plane scattering vector component qxy for different qz intervals (Fig. 6 A), as well as the Bragg rod profile of the DPPC/P188 film (Fig. 6 B). The Bragg rod results show that the chains are tilted by 30.8 ± 1.0° in a direction 12.4 ± 1.0° from the nearest neighbor, and the chain length is 19.0 ± 1.0 Å.
At ADPPC = 63 Å2, the Bragg peaks and Bragg rods of the P188-treated DPPC film are almost identical to those of pure DPPC monolayer at ADPPC = 48 Å2 (Figs. 4 B and 7 A, respectively). Two Bragg peaks resulting from {10,01} and {1} reflections are observed at and The d-spacing values d10,01 = 4.62 Å and 
indicate a distorted hexagonal unit cell with axes 
Similar to the pure DPPC case, the {1} Bragg rod achieves its maximum intensity at and the Bragg rod profile of the {10,01} reflection gives a tilt angle of 30.0° toward the direction of 12.2° from the nearest neighbor, with the coherently scattering length of the molecule, Lc, of 19.0±1.0 Å. Compared to pure DPPC, these results again confirmed that P188 does not affect the lipid packing of the condensed phase. The unit cell and the molecular tilt found in the condensed phase of DPPC are preserved in this mixed system, suggesting that P188 is completely excluded from the condensed region of the film.
View larger version (24K):
[in this window]
[in a new window]
|
FIGURE 7 Bragg rod profiles for Bragg rods in the {10,01} and {1} directions for a DPPC/P188 film on a water subphase at (A) ADPPC = 63 Å2 and (B) ADPPC = 47 Å2, both at 30°C. The rods were fitted (solid line) by approximating the coherently scattering part of the alkyl chain by a cylinder of constant electron density.
|
|
For our particular scan, the integrated intensity observed for the Bragg peaks of DPPC/P188 film at ADPPC = 107 Å2 is 12% of that at ADPPC = 47 Å2. According to fluorescence microscopy (22), at ADPPC = 47 Å2, the DPPC/P188 film is mainly composed of condensed phase. Because GIXD is sensitive only to the ordered phase, this decrease in scattering intensity suggests that an ordered phase in DPPC/P188 film at ADPPC = 107 Å2 occupies 12% relative to that at ADPPC = 47 Å2. It should be noted that the area coverage reported here for DPPC only reflects the distribution of the portion of the film scanned by the x-ray beam; the actual number could fluctuate depending on the heterogeneity of the film. Nonetheless, at ADPPC = 107 Å2 there is a substantial increase in the area occupied by the condensed DPPC domains in the presence (12%) of P188 compared to the absence (0%) of P188.
For the DPPC/ P188 monolayer, at ADPPC = 47 Å2 (Figs. 4 B and 7 B), the integrated intensities observed with and without P188 are the same, indicating that both films have a similar amount of ordered phase. The in-plane coherence lengths of DPPC crystallites in pure DPPC and DPPC/P188 films at ADPPC = 47 Å2 are also almost identical (
). This suggests that at low area/molecule (or high lipid density), P188 no longer stays in the lipid film. This finding corroborates our isotherm data indicating that at high surface pressures (
), the phase behavior of the DPPC/P188 system reverts back to the pure DPPC.
DPPG and DPPG/P188
Similar phenomena related to tightening the packing of the lipid molecules have also been observed for DPPG. When the area/DPPG molecule (ADPPG) is 84 Å2, pure DPPG monolayer is in the liquid-expanded phase, and no Bragg peak is observed (see Fig. 8 A). With P188 in the subphase at the same nominal ADPPG = 84 Å2, one {1, 0} in-plane Bragg peak (in our notation, {1,0} indicates six coinciding reflections {h,k}={(1,0), (0,1), (
1), (0), (0,), (1,)}) is observed at 
(Fig. 8 B), indicating the existence of 2-D crystallites with a hexagonal unit cell: 
and 
per tail. Comparing this Bragg peak with that of pure DPPG at ADPPG = 39 Å2 (Fig. 8 A, Table 2), the dimensions of this unit cell of DPPG/P188 film at ADPPG = 84 Å2 are almost identical to those of the pure DPPG monolayer at ADPPG = 39 Å2, again demonstrating that P188 does not interfere with the packing of ordered lipid molecules. Similarly, identical lattice parameters for DPPG are obtained in the absence and presence of P188 at even lower ADPPG (see Table 2).
View larger version (16K):
[in this window]
[in a new window]
|
FIGURE 8 Bragg peaks at different packing densities from GIXD on a water subphase at 30°C of (A) pure DPPG and (B) DPPG/P188 films. For clarity, the data have been offset vertically.
|
|
A finite size of the crystalline domain leads to the broadening of Bragg peaks. The coherence lengths Lxy calculated from the FWHM of Bragg peaks are listed in Table 2, along with the orientation of these straight alkyl chains (chain length Lc, tilt direction, and tilt angle). There is a pronounced decrease in the coherence length Lxy of the DPPG/P188 film compared to that of pure DPPG. Comparing the intensities of Bragg peaks at 30 mN/m, the intensity of the DPPG/P188 peak is 66.4% of that of pure DPPG; this is due to the presence of P188 in the footprint and smaller coherence length of crystallites in the DPPG/P188 film.
X-ray reflectivity
Pure DPPC
XR measurements were carried out for both DPPC (Fig. 9) and P188-pretreated DPPC monolayers below and above the "squeeze-out pressure", the surface pressure beyond which isotherms of pure DPPC and P188/DPPC superimpose. The pure lipid monolayers were fitted with the simplest two-box model (see Table 3, Fig. 9): one for the headgroup region and one for the tail region, with only one root-mean-square (roughness) parameter to smear out the interfaces. During the refinement and model fitting the number of electrons in the tail region was fixed to 242 (value calculated from the chemical formula for DPPC) and the area/molecule was fixed to the measured value to reduce the number of fitting parameters.
View larger version (25K):
[in this window]
[in a new window]
|
FIGURE 9 (A) X-ray reflectivity data for a DPPC monolayer on a water subphase at various surface packing densities at 30°C. The solid lines are fits to the data using box models, as discussed in the text. (B) The corresponding normalized electron density profiles for XR data. For clarity, the data have been offset vertically for all panels. z = 0 signifies the start of the water subphase. The fitting parameters are listed in Table 3.
|
|
For DPPC at 21 mN/m, the fit gives a hydrocarbon layer thickness of 12.3 ± 0.3 Å, resulting in a molecular tilt of 49 ± 2° and a total film thickness of 20.9 Å. At 24, 30, and 40 mN/m, the thickness of the hydrocarbon layer increases to 13.0 ± 0.2 Å, 15.7 ± 0.2 Å and 16.1 ± 0.3 Å, respectively, since the tails are less tilted at higher lipid packing densities (with tilt angles of 46 ± 2o, 34 ± 2°, and 31 ± 2°, respectively). Comparing the XR results of DPPC at 30 mN/m to GIXD measurements, the tilt angles from XR are larger. This is due to the fact that XR measurements, unlike GIXD, are sensitive to both the crystalline and noncrystalline phases, with molecules in the latter having larger tilt angles as they are more loosely packed. The number of electrons in the headgroup region decreases inversely with surface pressure (due to the elimination of hydrating water molecules from the headgroup region upon tight packing), whereas the electron densities of both the headgroup and the alkyl tail increase with compression, as expected.
The lipid headgroups of monolayers are of particular interest, since they constitute the interaction sites with the aqueous environment and peripheral membrane proteins (43). For a pure DPPC monolayer, noting that the bare PC headgroup has 164 electrons, the extra electrons associated with the headgroup could only be attributed to the hydrating water molecules. The number of water molecules in the hydration shell of each headgroup (Nwater) is listed in Table 3. From the fitted thickness of the headgroup box, as well as the area/molecule from the isotherm, the total volume of the headgroup region is known. If it is assumed that the partial volumes are additive, then the bare headgroup volume (Vo,head) can be calculated, and the value is listed in Table 3. For the component volume of each water molecule, the bulk value Vw = 30.0 Å3 has been used. It is worth noting that the calculated bare headgroup volumes from different surface pressures turn out to be almost constant, and agree well with reported values (44,45), within the error range of a few Å3.
We further note that M. Schalke analyzed reflectivity data for DPPC at 22°C (46), accounting for the lipid headgroup in terms of its three components (glycerol backbone, phosphate, and choline) described by partly overlapping volume-restricted distribution functions along the z axis (43,47). With additional information from FTIR and neutron reflectivity, a similar pressure dependence of headgroup hydration has been found.
DPPC/P188
 |
Strong corroborative evidence for phase separation of P188 from DPPC molecules can be found from the heterogeneity of the DPPC/P188 film observed with XR. For DPPC/P188 on pure water at ADPPC = 107 Å2 and 30°C, scanning different regions results in significantly different reflectivity profiles. The analysis shows that one region is mainly composed of P188 (Fig. 10 A), and the other is mainly composed of DPPC condensed-phase domains (Fig. 10 B). In Fig. 10 A, the XR curve shows no interference in the measured qz range and is very similar to those of pure P188 (Fig. 3), indicating that this region of the film is primarily composed of P188 that has a very low contrast in electron density compared to water. The data can be fitted with a one-box model for P188 at the interface, and the fitting results give a film thickness of 29.9 ± 1.4 Å, with an electron density 1.07 ± 0.01 times that of water. The mixed DPPC/P188 film has a larger thickness and an electron density comparable to that of pure P188 before compression (21.6 ± 1.0 Å, with electron density 1.04 ± 0.01 times that of water). This is possibly due to the presence of a DPPC molecule arranging in a staggering manner with P188 and thereby enhancing the overall average thickness of the surface layer. Alternatively, P188 in the presence of DPPC may exist in a more extended conformation compared to pure P188.
In Fig. 10 B, the XR curve shows the typical features of a layered structure, and can be fitted with a two-box model, one for the DPPC tail/P188 region and the other for the headgroup/P188 region. Since the electron density contrast between P188 and water is very small, a layer of P188 underneath the DPPC headgroup region is difficult to detect by XR, although it is likely that such a layer exists. Once again, the electron density contrast between P188 and the alkyl tail is small, whereas the contrast between P188 and the DPPC headgroup is significant. The existence of P188 in the DPPC layer can therefore be confirmed by the reduction of the electron density for the headgroup region of pure DPPC at a similar surface pressure.
The fit results for the mixed system are presented in Table 4. Results presented here are obtained by using the simplest possible model. We can fit the high qz range more precisely by including more parameters (e.g., different roughness on different interfaces). We have attempted such fittings with higher precision, but they do not give us any new information as far as the electron density and the box thickness are concerned.
When the x-ray beam falls within either region, a normal XR curve is obtained, as shown in Fig. 10. However, when the x-ray beam crosses the boundary of the two regions, poor overlaps in the reflectivity curve were observed (data not shown) when part of the qz range was remeasured before and after beam translation. Poor overlaps have been observed when the beam was moved from a P188-rich region (similar to Fig. 10 A) to a lipid-rich region (similar to Fig. 10 B), as well as in the opposite direction. Since changes in both directions have been observed, beam damage can be excluded as the potential cause for such poor overlaps, as it only alters the intrinsic feature of the XR curve to that of a bare water interface but can never induce lipid features, as observed in half of our cases showing poor overlaps. Furthermore, XR data measured on the same surface positions are reproducible.
Our analysis suggests that P188 inserts all the way to the end of the DPPC tail group. Despite attempts to divide the tail region into two boxes, least-square fitting always reverts back to parameters corresponding to only one box for the tail region. Although the electron density contrast between DPPC tail and poloxamer is low, neutron reflectivity experiments with specific site deuteration to afford better contrast are currently underway to elucidate the extent of poloxamer insertion into the tail region.
At 33mN/m the film remains heterogeneous, giving either lipid-rich or lipid-poor featured XR curves (see Table 4) depending on the location of the x-ray beam. To illustrate surface hererogeneity, two XR curves (not shown here but similar to Fig. 10 B) representing the two regions of the film were collected and analyzed. Similar to the case of 26 mN/m, the fitting parameters calculated from P188 surface coverage, which is estimated from the expansion of ADPPC in DPPC/P188 film compared to that of pure DPPC, serve as good starting points for fitting these XR data, and the resulting fitting parameters are listed in Table 4. Results from fitting indicate that the P188 content corresponds to 15% and 29%, respectively.
The XR data for both pure DPPC and DPPC/P188 films are shown in Fig. 11 A. Both the XR curve and the corresponding electron density profile of the DPPC/P188 film are identical to those of pure DPPC. These data corroborate the theory that at high surface pressures, P188 is squeezed out from the DPPC monolayer, as indicated earlier by our isotherm results. With the exclusion of P188 from the monolayer, the thickness of the former DPPC headgroup/P188 region also decreases and reverts back to the value of pure DPPC. It should be noted that for all measurements taken at this surface pressure, the results are independent of the location of the footprint of the x-ray beam. This is in contrast to observations obtained at higher areas/molecule (lower surface pressures) when P188 is present.
View larger version (19K):
[in this window]
[in a new window]
|
FIGURE 11 (A) X-ray reflectivity data for DPPC and DPPC/P188 films on a water subphase at ADPPC = 47 Å2. The solid lines are fits to the data using the box models discussed in the text. (B) The corresponding normalized electron density profiles for the XR data. z = 0 signifies the start of the water subphase. For clarity, the data have been offset vertically. The fitting parameters are listed in Table 4.
|
|
DPPG and P188-DPPG
We have also carried out similar XR experiments for DPPG and DPPG/P188 films (data not shown), and have obtained similar results as in the DPPC case. At low surface pressures ( < 38 mN/m) P188 remains inserted into the DPPG monolayer. The presence of the poloxamer results in similar surface heterogeneities to those observed for DPPC, giving rise to DPPG-rich and P188-rich regions in the surface film. Such heterogeneities are reflected by the different XR profiles observed at a given surface pressure (data not shown). At high surface pressure ( > 38 mN/m), P188 is squeezed out from the tail portion of DPPG monolayer. Although DPPC and DPPG have different headgroup charges, our observation shows that P188 insertion is not influenced by electrostatics. This should not be surprising, as P188 is a nonionic surfactant.
|
DISCUSSION
|
|---|
Our isotherm results demonstrate that P188 is a highly surface-active copolymer. With the presence of a lipid film at the air-water interface at 0 mN/m and 30°C, the injection of P188 into the subphase at a final concentration of 50 µM reduces the surface tension and increases the surface pressure from 0 to 26 mN/m in 40 min. The surface pressure changes sharply in the first few minutes after P188 injection; after 10 min, the change becomes gradual and slow. After 40 min, XR scans on P188 over time (with a 2-h gap) show small differences that are within the error range (data not shown). All x-ray scattering data presented here were taken at least 40 min after helium flushing, at which point the adsorption of P188 to the air-water interface is so slow that any residual change in the film structure over time can hardly be detected by XR. It is likely that the slow P188 adsorption into lipid monolayer slightly changes the occupancy of lipid-rich regions. Besides, there might also be a conformational change of P188 at the interface.
XR and GIXD measurements have been used to provide molecular-level information for the insertion of P188 into the DPPC and DPPG monolayers at the air-water interface. They both agree well with the postulate that P188 interacts with structurally disrupted membrane portions, and is effectively excluded from the once-damaged portion when the lipid packing density of a normal membrane has been reestablished. Recently, the squeezing-out of Gemini surfactants with anti-HIV activity from the phospholipid monolayer upon compression has also been reported (48).
It is clear from our GIXD data that P188 does not affect the molecular packing of the ordered phase of DPPC, suggesting that P188 is completely excluded from the ordered portion of the monolayer. Our findings also indicate that P188 and DPPC do not mix well with each other. Such demixing has been observed to occur almost instantaneously upon P188 adsorption to the lipid film via fluorescence microscopy (data not shown). Clear evidence of phase separation between P188 and the ordered lipid phase is provided by GIXD measurements taken 40 min after the introduction of the poloxamer.
The presence of P188 at the interface acts to force lipid molecules to pack more tightly, and the amount of P188 incorporated into the DPPC film is likely to be controlled by the surface free energy. When the surface pressure of DPPC monolayer is low (<36 mN/m), P188 adsorbs to the air-water interface and physically occupies a portion of the available area, causing the subsequent formation of 2-D crystallites of DPPC molecules. When the surface pressure is increased to >36 mN/m, the close contact between P188 and DPPC becomes unfavorable and P188 is entirely squeezed out from the DPPC tail region. In the context of membrane sealing, the insertion of poloxamer into a low-density lipid film at the air/water interface helps elucidate the mechanism by which the poloxamer seals damaged membranes in living cells. In effect, the low-density lipid film in our model system mimics the hydrophobic-aqueous environment found both at the highly curved lipid portion that constitutes the pore on a damaged membrane, and in the portion of the membrane that has enhanced permeability due to a reduction in lipid density. Just as in our model system, both of these sites in a damaged membrane have an enhanced exposure of the lipid hydrocarbon tail to the aqueous environment, into which the poloxamer inserts. Such an insertion results in the tightening of the lipid packing, as demonstrated in our GIXD results. This, along with the fact that the PEO portion of the polymer remains situated in the aqueous phase with a certain radius of gyration, helps restore the barrier function of the damaged membrane and reduces its permeability. Even when a pore exists on the membrane surface, we do not expect the poloxamer's sealing action to be brought about by its insertion into the water-water interface. Rather, by inserting into the highly curved regions around the pore, the poloxamer can help bring lipids on different sides of the pore together and displace water molecules out of the pore region, thus eliminating the pore formation and restoring the barrier function.
Phase separation in lipid/poloxamer systems has previously been inferred (49). Differential scanning calorimetry (DSC) has shown that the liposomal bilayer's main phase transition endotherm broadens with a shoulder at high poloxamer concentration, indicating the presence of two distinct entities. Moreover, the decrease in mean liposome radius implies the possible existence of mixed micelles. Nonetheless the DSC and size analysis results only indirectly point to a separation between poloxamer-rich and poloxamer-poor phases. DPPC ordering by P188 and the surface inhomogeneity observed here unambiguously confirm such phase separation. In the context of membrane sealing, P188 insertion increases the lipid packing density, which in turn helps arrest leakage, as observed in damaged cells (2,50). Although the presence of P188 would increase the lipid packing density, we do not expect the lipids to attain the gel state in real membranes due to the large fraction of unsaturated lipids present.
The incapability of P188 to remain in the system at high surface pressures can be beneficial in terms of its application as a membrane sealant. After electroporation, cells may activate a self-healing process, restoring the structural integrity of the lipid bilayer of the cell membrane. As the cell heals and the lipid packing of the membrane is restored, our data suggests that there is a mechanism through which P188 exits the membrane.
The selective insertion of P188 into low- over high-density lipid film helps to clarify some of the earlier findings on the interaction between poloxamers and different lipid bilayers. It has been suggested (51) that poloxamers associate with relatively fluid eggPC bilayers, resulting in an increased particle size caused by the projection of PEO groups from the liposome surface. This was not observed when poloxamers were added to gel state DSPC liposomes (52). It has also been found by DSC (49) that phospholipid bilayer pretransition enthalpy was reduced when poloxamers interacted with the liquid-crystalline state, but not gel-state liposomes. These findings can now be understood in terms of lipid ordering tuning poloxamer insertion, where loose packing in the fluid bilayer promotes insertion whereas tight packing in the gel state inhibits insertion. Moreover, our data suggest that poloxamer interacts with liposomes via penetration into liquid crystalline state bilayers rather than adsorption onto liposome surfaces. Previous work has also reported similar results for peptide-lipid interactions with monolayer and bilayer models (53). Although our monolayer work can help shed light on bilayer studies, differences may exist in the conformation and orientation of the interacting poloxamer.
|
APPENDIX A
|
|---|
Surface heterogeneity of DPPC/P188 filmat ADPPC = 107 Å2
At large nominal area DPPC (ADPPC = 107 Å2, = 26 mN/m), we have performed fluorescence microscopy measurements, and have found that the adsorption of poloxamer at the air/water interface blackens out the monolayer initially in gas (dark)/liquid-expanded (bright) coexistence (see Fig. 12), and cuts the surface into regions with different gray levels (Fig. 13). These regions can be of the length scale of 15 mm.
View larger version (69K):
[in this window]
[in a new window]
|
FIGURE 12 Fluorescence micrograph of a DPPC monolayer at ADPPC = 110 Å2, showing the coexistence of gas (dark) and liquid-expanded (bright) phases. The width of the micrograph is 550 mm.
|
|
View larger version (19K):
[in this window]
[in a new window]
|
FIGURE 13 Fluorescence micrographs after P188 injection at a nominal area of ADPPC = 110 Å2. (A) dark-gray (DPPC-rich); (B) one example showing coexistence of dark and light gray; (C) light gray (DPPC poor). The width of each micrograph is 550 mm.
|
|
At intermediate surface pressures (e.g., ADPPC = 55 Å2, = 33 mN/m), similar surface heterogeneities have been observed. Fluorescence microscopy results suggest that the DPPC-rich and DPPC-poor regions can be as large as 15 mm in width. To complete one XR measurement, the sample stage is usually translated by 6 mm. Depending on the starting location of the x-ray footprint, there are variations in the distance the beam needs to be moved to locate the reported changes in the XR curves. To ensure that the observed heterogeneity is real, we have repeated the experiment on more than eight fresh samples of the DPPC/P188 system at ADPPC = 110 Å2 ( = 26 mN/m). For each sample, we have carried out multiple XR scans either by translating the beam or by staying at the same spot and measuring over time. Observations from these experiments confirm the existence of surface heterogeneities at this pressure. Similar systems of DPPC with other poloxamers of different molecular weight and hydrophobicity (not descirbed in this manuscript) have also been studied, and similar heterogeneous XR curves have been observed for every film.
In our experiments, we use the entire trough (approximately tens of cm) to model the packing around the damaged portion of the membrane (approximately submicron meter), therefore the length scale of poloxamer patches observed in our experiments are not representative of that found in a damaged membrane. We observe approximately millimeter poloxamer-rich patchts inserted at the surface when the starting lipid packing density is low. The existence of these large length scale patches is due to the fact that the initial lipid packing density is less than half of that expected for an intact membrane, and thus the poloxamer occupied 50% of the surface area upon insertion for ADPPC = 107 Å2 (the coverage of inserted poloxamer decreases as the degree of integrity loss decreases). Although the surface occupancy of the poloxamer in our model system should reflect that found in the damaged membrane of the cell, the physical length scale of the poloxamer patches does not have any bearing on the actual system.
|
APPENDIX B
|
|---|
Monitoring the beam damage
In the XR measurements of DPPC/P188 films at relatively low surface pressures, we have obtained different XR data when we move the footprint of the x-ray beam to different regions of the film. Beam damage could be a plausible cause for the observed behavior. To verify that the observed differences of the XR profiles were not caused by beam damage, we have measured XR as a function of time without moving the x-ray beam. If beam damage was indeed the culprit, identical result could not be achievable. Fig. 14 shows that two XR curves taken 2 h apart. Close examination indicates that the two curves are basically identical up to qz/qc = 25. Since the intensity minima at qz/qc 
10 greatly deviate from the Fresnel curve, they constitute the most sensitive location to reflect possible beam damages. The fact that these minima are identical excludes the possibility that there is any beam damage in the two curves taken 2 h apart.
View larger version (22K):
[in this window]
[in a new window]
|
FIGURE 14 (A) X-ray reflectivity data for a DPPC/P188 film on a water subphase at 40 mN/m with time elapse of 2 h in between. The solid lines are fits to the data using box models discussed in the text. (B) The corresponding normalized electron density profiles for XR data. z = 0 signifies the start of water subphase. For clarity, the data have been offset vertically.
|
|
The slight difference observed here can be due to a time effect intrinsic to the system. The fitting results show that the parameters for the thickness and electron density are identical for the two cases with only the roughness for the interfaces being different. (For the first measurement, the roughness is 3.87 Å, and after 2 h it increases to 4.76 Å.) As discussed earlier, the P188 is squeezed out from the DPPC film upon compression. One possible explanation for the increase in roughness with time is that P188, upon squeeze-out (where the relatively hydrophobic PPO portion becomes excluded from the alkyl chain region), rearranges under the headgroup and undergoes conformational changes with time.
|
APPENDIX C
|
|---|
All initial pressures reported in this study are 4 mN/m higher than in our previous published results (22). This is due to the fact that the pressures were recorded 5 min after P188 injection in previous experiments (22), whereas data reported here were taken 40 min after injection to match the waiting time needed in x-ray experiments for He flushing. A similar difference is also found in the squeeze-out pressure reported below.
|
APPENDIX D
|
|---|
TOP
ABSTRACT
INTRODUCTION
