| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Laboratoire d'Imagerie Moléculaire et Nano-Bio-Technologie, IECB, UMR-CNRS 5471, Université Bordeaux I, 33607 Pessac Cedex, France
Correspondence: Address reprint requests to Alain R. Brisson, E-mail: a.brisson{at}iecb.u-bordeaux.fr.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSION AND PERSPECTIVESSUPPLEMENTARY MATERIALACKNOWLEDGEMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSION AND PERSPECTIVESSUPPLEMENTARY MATERIALACKNOWLEDGEMENTSREFERENCES |
|---|
; McConnell et al., 1986), have already found widespread use as mimics of cell membranes (Watts et al., 1984; Sackmann, 1996; Salafsky et al., 1996; Reviakine et al., 1998; Milhiet et al., 2002; Yip et al., 2002) and are attractive as building blocks for biotechnological applications (Sackmann, 1996; Boxer, 2000; Reviakine and Brisson, 2001; Kam and Boxer, 2003; Larsson et al., 2003). It is only in the recent years that a detailed image of the structural intermediates of the formation of SLBs is emerging from both experimental (Rädler et al., 1995; Keller and Kasemo, 1998; Jass et al., 2000; Keller et al., 2000; Reviakine and Brisson, 2000; Johnson et al., 2002; Reimhult et al., 2003; Richter et al., 2003; Benes et al., 2004) and theoretical studies (Seifert and Lipowsky, 1990; Seifert, 1997; Zhdanov et al., 2000), though many issues about the formation of SLBs remain to be explored. We do not have, at present, a comprehensive understanding of the driving forces of the SLB-formation process (Nollert et al., 1995; Reviakine and Brisson, 2000; Zhdanov and Kasemo, 2001; Reimhult et al., 2002b; Richter et al., 2003). In particular, the role of the solid support in the rupture of vesicles and the promotion of SLB formation (Keller and Kasemo, 1998; Starr and Thompson, 2000; Reimhult et al., 2003; Richter and Brisson, 2004) remains poorly understood.
Silica-based surfaces and mica are the prototypes of solid supports that are currently used to form solid-supported lipid bilayers. Notably, previous studies have provided indications for some differences in the SLB-formation process on these supports, such as the role of calcium (Reviakine and Brisson, 2000; Richter et al., 2003) and the influence of the vesicle size on rupture (Reviakine and Brisson, 2000; Reimhult et al., 2002b). A detailed comparison of the SLB-formation process between these surfaces is, however, still lacking.
A multitude of techniques, such as fluorescence microscopy (Nollert et al., 1995; Johnson et al., 2002), quartz-crystal microbalance with dissipation monitoring (QCM-D) (Keller and Kasemo, 1998; Keller et al., 2000; Reimhult et al., 2002a,b, 2003; Richter et al., 2003; Seantier et al., 2004), atomic force microscopy (AFM) (Jass et al., 2000; Reviakine and Brisson, 2000; Richter et al., 2003; Tokumasu et al., 2003; Seantier et al., 2004), surface plasmon resonance (SPR) (Keller et al., 2000), and ellipsometry (Benes et al., 2004) has been used to investigate the SLB-formation process. We have recently illustrated the complementary nature of AFM and QCM-D for the analysis of vesicle adsorption and SLB formation on silica (Richter et al., 2003). Giving access to spatially resolved structural information on the nanometer scale and overall adsorption and reorganization dynamics, respectively, the combination of both techniques has enabled the identification and characterization of several different SLB-formation pathways stressing the importance of electrostatic interactions in the SLB-formation process. Others have pointed out the complementary aspects between QCM-D and optical surface-sensitive methods, such as SPR (Keller et al., 2000) or ellipsometry: while optical methods allow measuring the dry mass of the adsorbed lipids, QCM-D provides not only the hydrated mass, but also a direct mean (the dissipation) to distinguish between different phases of the adsorbate—surface-bound vesicles or bilayer patches (Keller et al., 2000). As such, the combination of QCM-D and SPR has allowed establishing the concept of local critical vesicular coverage (Keller et al., 2000). Recent work has also demonstrated that in situ measurements on mica can be performed both with ellipsometry (Benes et al., 2002) and with QCM-D (Richter and Brisson, 2004) in a reproducible manner, opening up for the joint use of QCM-D, ellipsometry, and AFM on this support.
In this study, we combine sequential measurements by QCM-D and AFM on identical supports, complemented by ellipsometry, to characterize the vesicle adsorption and the SLB formation on mica. As in a previous study on silica (Richter et al., 2003), we employ vesicles with varying net charges, from positively charged dioleoyltriammonium-propane (DOTAP) to mixtures of zwitterionic dioleoylphosphatidylcholine (DOPC) and negatively charged dioleoylphosphatidylserine (DOPS). This allows for a quantitative comparison between the lipid deposition on mica and on silica and gives new insight in the role of the solid support, the lipid composition of the vesicles, and the calcium-EDTA balance in the SLB-formation process.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSION AND PERSPECTIVESSUPPLEMENTARY MATERIALACKNOWLEDGEMENTSREFERENCES |
|---|
) was synthesized and kindly provided by the group of M. Textor (ETH, Zurich, Switzerland). Other chemicals were purchased from Sigma (St. Louis, MO). Ultrapure water with a resistivity of 18.2 M
was used (Maxima, USF ELGA, Trappes, France). Muscovite mica disks of 12 mm diameter were purchased from Metafix (Montdidier, France). QCM-D sensor crystals (5 MHz), reactively sputter-coated with 50 nm silicon oxide, were purchased from Q-SENSE (Gothenburg, Sweden). Low-viscosity epoxy glue (EPOTEK 377) for the mica gluing was purchased from Gentec Benelux (Waterloo, Belgium).
A buffer solution made of 150 mM NaCl, 3 mM NaN3, and 10 mM HEPES, pH 7.4, was prepared in ultrapure water, and either 2 mM EDTA or 2 mM CaCl2 were added as indicated in the text. Small unilamellar vesicles (SUVs) of desired lipid mixture were prepared by sonication as described earlier (Richter et al., 2003). Before use, vesicle suspensions were diluted to 0.1 mg/mL if not otherwise stated.
Quartz crystal microbalance with dissipation monitoring (QCM-D)
QCM-D measurements (Rodahl et al., 1995) were performed with the Q-SENSE D300 system equipped with an Axial Flow Chamber (QAFC 302) (Q-SENSE). Briefly, upon interaction of (soft) matter with the surface of a sensor crystal, changes in the resonance frequency, f, related to attached mass (including coupled water), and in the dissipation, D, related to frictional (viscous) losses in the adlayer are measured with a time resolution of better than 1 s.
The QCM-D sensor crystals were coated with mica and verified to operate stably and reliably according to a previously described protocol (Richter and Brisson, 2004). Briefly, mica sheets were glued to the QCM-D sensors using epoxy glue. The glued mica sheets were cleaved until sufficiently thin mica layers and stably operating sensors were obtained.
Measurements were performed in exchange mode (Richter et al., 2003), if not otherwise stated. The exchange mode allows following processes of adsorption and surface adlayer changes in situ while sequentially exposing different solutions to the supports. In this mode the fluid in the measurement chamber is generally still. Occasionally, flow mode was employed, i.e., the solution was continuously delivered to the measurement chamber (flow speed 80 µL/min) by the aid of a peristaltic pump (ISM832A, Ismatec, Zürich, Switzerland) (Richter et al., 2005). Resonance frequency and dissipation were measured at several harmonics (15, 25, 35 MHz) simultaneously. The working temperature was 24°C.
If not stated otherwise, changes in dissipation and in normalized frequency (
fnorm = fn/n, with n being the overtone number) of the third overtone (n = 3, i.e., 15 MHz) are presented. Adsorbed masses, m, are calculated according to the Sauerbrey equation (Sauerbrey, 1959), m = –Cfnorm, with the mass sensitivity constant C = 17.7 ng·cm–2·Hz–1 for 5 MHz sensor crystals. The equation has been demonstrated to be valid, within 5% error, for lipid bilayers or adsorbed nonruptured SUVs on rigid sensor coatings with a thickness ranging from several nanometers (such as evaporated gold or sputtered silica; Keller and Kasemo, 1998; Reimhult et al., 2002b) to several micrometers (such as glued mica sheets; Richter and Brisson, 2004).
For transfer of QCM-D sensors with adsorbed material from the QCM-D chamber to the AFM, the sensors were unmounted with the aid of a suction holder (Meni CUP, Menicon Pharma, Illkirch-Graffenstaden, France), ensuring that the sample remained permanently covered with liquid.
Ellipsometry
Ellipsometry is an optical technique based on the measurement of changes in the ellipsometric angles, and 
(Cuypers et al., 1983; Tompkins, 1993), of elliptically polarized light upon reflection off a planar surface. These changes are sensitive to the presence of thin deposited films and, consequently, the method allows monitoring adsorption phenomena in situ, at the solid-liquid interface, with a mass resolution in the range of 5–10 ng·cm–2. The employed null-ellipsometer setup has a time resolution of 10–15 s and is described in detail elsewhere (Cuypers et al., 1983; Corsel et al., 1986). In this study we consider the change in the angle only, which is roughly proportional to the (dry) lipid mass adsorbed to mica (Benes et al., 2004).
Measurements were performed in an open cuvette system (Corsel et al., 1986), at room temperature. Mica disks were glued on an aluminum slide over a hole (8 mm diameter) using melted wax (Benes et al., 2002). The backside of the mica disks was rendered opaque with emery paper before gluing onto the aluminum slide. Uniform mica surfaces were obtained by cleavage of the front side with adhesive tape and immediately mounted in the buffer-filled cuvette. The buffer (
3 mL) was stirred with a magnetic stirrer (1000 rpm). Samples were pipetted at appropriate concentrations into the buffer. Rinsing was realized by injecting 30 mL of buffer (injection rate, 1 mL/s) while simultaneously withdrawing excess liquid.
Preparation of PLL-g-PEG-coated AFM tips
Oxide-sharpened silicon nitride cantilevers with a nominal spring constant of 0.06 N/m (Digital Instruments, Santa Barbara, CA) were rinsed in water and ethanol, blow-dried with nitrogen, and exposed to ultraviolet/ozone (BHK, Claremont, CA) for 10 min (Richter et al., 2003). The cleaned tips were transferred immediately into a solution of 1 mg/mL PLL-g-PEG in 10 mM HEPES, pH 7.4. After 30 min of immersion, the tips were withdrawn, rinsed with ultrapure water, and blow-dried with nitrogen. PLL-g-PEG-modified tips were stored on Gel-Pak (Scotch, CA) under nitrogen and used within four weeks.
Atomic force microscopy
AFM measurements were performed in liquid using a Nanoscope IV-Multimode (Veeco, Dourdan, France), equipped with a J-scanner (120 µm). Before use, the tapping mode fluid cell was washed in successive baths of ethanol and ultrapure water, followed by extensive rinsing in ethanol and blow-drying in a stream of nitrogen. Tubings and O-ring were sonicated in ethanol and water, rinsed with ethanol, and blow-dried in nitrogen.
For AFM investigations subsequent to QCM-D measurements, mica-coated QCM-D sensors, covered with the sample, were attached to Teflon-coated (Bytac, Norton, OH) metal disks using double-sided tape (TESA, Hamburg, Germany) and installed on the AFM scanner. In all other cases, mica disks were glued to Teflon-coated metal disks using the epoxy glue. Uniform surfaces were obtained by cleavage with adhesive tape and immediately covered with buffer solution. The AFM was equilibrated for 5–30 min before imaging.
Images were recorded in contact mode or tapping mode as indicated. Contact mode images were acquired at scanning rates of 4–8 Hz while manually adjusting the force to a minimum (<200 pN). Tapping mode images were acquired at scanning rates of 1–2 Hz. A resonance frequency 7 kHz was chosen with low free amplitude (0.4–0.6 V) and minimum load upon scanning. Images were first-order plane-fitted except otherwise stated.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSION AND PERSPECTIVESSUPPLEMENTARY MATERIALACKNOWLEDGEMENTSREFERENCES |
|---|
), and to some extent by AFM on mica (Reviakine and Brisson, 2000), and was chosen as a reference system in our study. The overall kinetics of the SLB-formation process on mica by QCM-D (Fig. 1) reveals a two-phase behavior. In the first phase the frequency decreases beyond –25 Hz (fmin = –39 Hz) whereas the dissipation increases (Dmax = 1.6 x 10–6), evidencing the adsorption of a substantial amount of nonruptured vesicles. Final shifts in frequency (ffin = –25 Hz) and dissipation (Dfin < 0.2 x 10–6) after the second phase match values previously obtained on silica (Keller and Kasemo, 1998; Richter et al., 2003) and indicate the presence of an SLB. AFM imaging of the final state of the surface adlayer revealed a flat surface devoid of defects on areas of 25 µm2 and more and confirmed the presence of a confluent ideal bilayer (Fig. 6 B).
|
|
; Richter et al., 2003). We believe that the distribution of vesicles and bilayer patches revealed in the image closely resembles the true distribution and is not significantly altered by the imaging process (see also below).
|
25 and 28 nm (for details of the size determination see the Supplementary Material), respectively, close to the size expected from SUVs. Notably, the size of both vesicles and SLB patches was heterogeneous, i.e., both small and large patches and vesicles were present, indicating a rather weak dependence of the rupture tendency on the vesicle size over the size distribution observed here. In general, lipid patches exhibited a circular shape. Both vesicles and bilayer patches were found not to move laterally over the time range of 10 min and more, except at relatively high imaging forces. AFM images acquired at low lipid coverage (c.f. Fig. 2 B) did only exceptionally, presumably due to tip artifacts, reveal vesicles or bilayer patches that disappeared upon imaging.
To verify whether lipid material desorbs from the mica surface, we monitored the SLB-formation process by ellipsometry. As this optical technique senses the dry mass, net desorption of material would be directly detected as an increase in the ellipsometric angle . After interrupting the SLB-formation process at a lipid coverage corresponding to 30% of a complete SLB (Fig. 2 D), remained stable, confirming that lipid material did not desorb. Therefore, we interpret the QCM-D response after the rinsing step as shown in Fig. 2 A as due to the decomposition of surface-bound vesicles into bilayer patches, the increase in frequency, and the decrease in dissipation being caused by the loss of water associated with and the changes in the rigidity of the lipid structures. Consequently, the QCM-D data indicate that the typical time for the spontaneous transition of isolated mica-bound vesicles into bilayer patches is in the range of minutes to hours. This is a remarkable difference compared to isolated vesicles on silica that remained stable over days under identical conditions (Richter et al., 2003).
Stability and mobility of small SLB patches
To investigate the stability of bilayer patches, a certain area was imaged repetitively with varying forces (Fig. 3). Although almost all SLB patches remained immobile and kept their initial shape, connections between a few distant patches (asterisks in Fig. 3 B) were established, likely induced by the AFM tip that was temporarily operated at an elevated force. The resulting patches clearly reshaped into circular objects (Fig. 3 C), the area of which corresponded to the sum of the areas of the initial patches, confirming that coalescence events were tracked (for details see the Supplementary Material). The center of mass of the merging patches remained approximately fixed during the merger event (Fig. 3 B). Given the image acquisition time of 2 min and the fact that connections are visible over several scan lines, corresponding to 0.2 s each, the time needed for coalescence and reshaping can be estimated to be in the range of a few seconds to a few minutes. We note that, in contrast to our observation on mica, small SLB patches of stable noncircular shape were observed on silica (Richter et al., 2003).
|
|
f increased above –25 Hz (to –16 Hz) and D decreased to 0.4 x 10–6, indicating that the major part of adsorbed vesicular material was transformed into bilayer patches. Corresponding AFM images showed indeed extended domains of lipid bilayers coexisting with a small number of vesicles (Fig. 5 B) and ellipsometry experiments indicated that no significant desorption occurred in similar conditions (Fig. 5 C).
|
; Muresan and Lee, 2001).
AFM images acquired after interrupting the SLB formation at an even later stage, showed an apparently confluent bilayer in which local defects persisted (Fig. 6 A). Elongated defects of several micrometers in length and down to a few nanometers in width could be observed, which were particularly frequent after incubation with SUV solutions at low concentration (10 mg/mL). Notably, the apparent local radius of curvature of the bilayer boundary at the ends of some of these defects (arrowheads in Fig. 6 A) was in the range of 10 nm. Given our experimental conditions, we estimate that such small radii are at the limit of resolution. It is therefore not clear, whether the apparent topography reflects a truly continuous bilayer boundary or whether, instead, it marks the endpoint of a zone (dotted line in Fig. 6 A) at which two patch boundaries appose with each other without coalescing, the gap between them being small enough to remain unresolved by AFM. Whatever the true topography, some of these defects remained stable over hours, again indicating the presence of forces that inhibit shape equilibration (Benvegnu and McConnell, 1992).
Subsequent reincubation as well as incubation without interruption led to bilayers that were close to defect-free, as seen from the AFM images (Fig. 6 B). To verify whether defects persisted that were invisible to AFM, some DOPS-containing bilayers were incubated with the protein annexin A5 and the growth of two-dimensional crystals was followed as described elsewhere (Reviakine et al., 1998). Round-shaped crystals were observed as expected for undisturbed crystal growth (Reviakine et al., 1998) (not shown), whereas elongated discontinuities in the supporting bilayer are expected to locally prevent crystal growth thereby distorting the crystal's shape.
Influence of the vesicle charge and the presence of calcium on the SLB formation on mica
The vesicle charge and the presence of calcium were previously shown to have a pronounced influence on the lipid deposition pathways on silica (Richter et al., 2003). For comparison, we followed the global kinetics of lipid deposition on mica for vesicles of varying charge in the presence of the calcium chelator EDTA and in the presence of 2 mM calcium ions, respectively, by QCM-D. The vesicle charge was varied by mixing appropriate amounts of positively charged (DOTAP), zwitterionic (DOPC), and negatively charged (DOPS) lipids. For a quantitative comparison of the lipid deposition process, the minimum in frequency, fmin, and the maximum in dissipation, Dmax, were determined. SLB formation, if occurring, was additionally characterized by the time, tSLB, required to attain stable values in frequency, ffin, and dissipation, Dfin. Furthermore, the stability of adsorbed isolated vesicles was checked by interrupting the SLB formation at low coverage. Table 1 and Figs. 7 and 8 summarize the obtained results.
|
|
|
; Benes et al., 2004). In contrast, SLB formation was observed on silica (Richter et al., 2003). Rather small amounts of negative charges (20% DOPS) were sufficient to completely inhibit lipid adsorption. Similarly, vesicle adsorption was hampered on silica with increasing negative vesicle charge, although inhibition occurred only at considerably higher negative charge (50% DOPS) (Richter et al., 2003). These observations suggest that the electrostatic repulsion between negatively charged vesicles and the support is stronger with mica than with silica. In conjunction, we note that no SLBs could be formed on mica with DOPS-containing SUVs. For comparison, SLB formation was inhibited on silica only at a DOPS content of 33% and more (Richter et al., 2003).
QCM-D responses indicated the formation of SLBs for all types of positively charged SUVs investigated. These observations are qualitatively similar to silica. It is, however, notable that the formation of mica SLBs with pure DOTAP-SUVs (Fig. 7 C) exhibited very low dissipation values (Dmax < 0.5 x 10–6) throughout the whole SLB-formation process. Corresponding values on silica were slightly higher (Dmax = 0.8 ± 0.1 x 10–6) (Richter et al., 2003), indicating that the tendency toward immediate rupture of isolated DOTAP vesicles is even more pronounced on mica.
The formation of a supported vesicular layer (SVL) with DOPC-SUVs (Fig. 7 B) deserves some additional attention. The adsorption of these vesicles was partly reversible (Reviakine and Brisson, 2000; Benes et al., 2004), which is exceptional for the vesicle-surface interactions that we have so far investigated on silica and on mica. The frequency shift of |fmin| = 103 Hz, indicates a high surface coverage (Keller and Kasemo, 1998). Interestingly, the QCM-D responses after the addition of calcium (Fig. 7 B) suggest that the SVL is converted into a complete supported lipid bilayer, even though no vesicles were present in solution. Furthermore, the maximum dissipation shift for the mica-supported vesicular layer, Dmax = 10 x 10–6, is remarkably high, resulting in a ratio between dissipation and frequency of Dmax/|fmin| = 0.10 x 10–6·Hz–1. This is significantly more than on silica (0.06 x 10–6·Hz–1) (Richter et al., 2003), and indicates that DOPC vesicles become less flattened on mica than on silica (Reimhult et al., 2002b). For simplicity, we have compared the ratio of Dmax/|fmin|. (Note, however, that in addition to the flattening of the vesicle, the intervesicle interaction at high surface coverage can affect D/f. This potential artifact can be avoided by comparing the D/|f|-values at low vesicular coverage (Reimhult et al., 2002b). Corresponding values are 0.14 x 10–6·Hz–1 and 0.08 x 10–6·Hz–1 (at n = 7) on mica and silica, respectively, supporting our conclusions that DOPC vesicles become more flattened on silica.)
In the presence of calcium, SLBs were formed over the entire range of lipid mixtures investigated. Notably, SLBs could be formed from vesicles containing as much as 80% DOPS. A detailed examination revealed a tendency of tSLB and |fmin| to decrease toward higher negative charges (Fig. 8). This evidences that increasing DOPS content facilitates SLB formation on mica in the presence of calcium. Notably, the opposite was observed on silica. The influence of calcium on the SLB-formation process for DOTAP-containing vesicles was minor.
Isolated mica-bound vesicles that contained only DOPC or small amounts of DOTAP were found to be stable over minutes and hours (Fig. 7 A), similar to the results obtained on pure silica (Keller et al., 2000). Isolated vesicles from all lipid mixtures containing 20% and more DOPS, however, slowly but measurably transformed into SLB patches.
Influence of the AFM tip on the imaging of lipid vesicles and SLB patches
Lipid structures such as surface-bound vesicles or bilayer patches are known to be easily modified by interactions with the AFM tip (Jass et al., 2000; Reviakine and Brisson, 2000; Richter and Brisson, 2003; Liang et al., 2004). During initial investigations with nonmodified tips it became apparent that the lipid material can mediate strong tip-sample interactions (Richter and Brisson, 2003), which renders the imaging of heterogeneous lipid structures such as coexisting bilayer patches and vesicles unstable and prone to artifacts, in particular, for low scan speeds and small image sizes. As demonstrated by the AFM images shown here, the modification of the AFM tip with PLL-g-PEG resulted in an improved control in imaging such lipid structures down to image sizes of 1 µm and less. When forces exerted by the AFM tip were adjusted to a minimum, vesicles and bilayer patches could be imaged intact (c.f. Fig. 5 B). Alternatively, increased forces could be employed to induce the rupture of vesicles (arrowheads in Fig. 2, B and C), the coalescence of bilayer patches (c.f. Figs. 3 and 4), or the displacement of lipid structures (not shown).
Imaging in tapping mode provided least disturbance of the lipid material, though images in contact mode also could be obtained with only minor artifacts, such as occasional tip-induced vesicle rupture. In particular, small bilayer patches appeared (almost) circular and immobile both in tapping and contact mode, indicating that small lateral forces did not significantly alter the location or overall shape of the patches.
Effects of the AFM tip on the apparent size of the bilayer patches and vesicles
We observed small but significant variations in the apparent size of the bilayer patches, depending on the applied force (see Supplementary Material). The apparent radius and height of the bilayer patches were observed to decrease by 5 and 0.2 nm, respectively, when increasing the normal force by 200 pN from the minimum force required for proper tracking of the surface. Similarly, the height of surface-bound vesicles was observed to vary by a few nanometers. This indicates that the forces exerted by the tip can induce lateral or normal compression of the lipid assemblies (as well as the tip coating). Next to tip-convolution effects, the sample deformation, thus, renders the determination of the correct size of adsorbed vesicles and bilayer patches a complex task, as discussed in the Supplementary Material.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSION AND PERSPECTIVESSUPPLEMENTARY MATERIALACKNOWLEDGEMENTSREFERENCES |
|---|
; Reviakine and Brisson, 2000; Johnson et al., 2002; Richter et al., 2003) and theoretical (Seifert and Lipowsky, 1990; Seifert, 1997; Zhdanov et al., 2000) studies have shown that adsorbed lipid vesicles can transform into bilayer patches via at least three different mechanisms: i), spontaneous rupture of isolated vesicles; ii), rupture of vesicles induced by high vesicle density (critical vesicular coverage); and iii), rupture of vesicles induced by the active edge of a bilayer patch. We show here that, by varying the vesicle composition and the calcium-EDTA balance, all above-described mechanisms can be reproduced on mica, and that significant differences exist between mica and silica.
Rupture of isolated vesicles
The time from adsorption to rupture of isolated vesicles, i.e., surface-bound vesicles at sufficiently low coverage, can vary widely. Whereas we found DOTAP to rupture faster than what can be resolved (1 s) by the employed techniques, DOPC vesicles were stable (Fig. 7 A) in 2 mM CaCl2 over experimental timescales (1 h). Both extremes can be understood from a thermodynamic perspective of the interaction between vesicle and support and have been described before (Lipowsky and Seifert, 1991; Seifert, 1997; Reviakine and Brisson, 2000).
We report here on an intermediate time range (minutes to hours) that is needed for DOPS-containing vesicles to rupture in the presence of calcium, which highlights the importance of the kinetic perspective (Zhdanov and Kasemo, 2001). The parameters that govern the rupture dynamics remain to be elucidated and will be discussed further below.
Rupture induced by critical vesicular coverage
The pathway of vesicle rupture induced by the cooperative action of neighboring adsorbed vesicles has been described and characterized in detail on silica (Keller et al., 2000; Reimhult et al., 2002a,b; Richter et al., 2003). Similar to silica (Keller et al., 2000), isolated vesicles containing pure DOPC or DOPC/DOTAP (4:1) were observed to be stable when adsorbed on mica in the presence of calcium (c.f. Fig. 7 A). The formation of SLBs at higher surface coverage evidences that the influence of neighboring vesicles is necessary to induce the rupture of the first vesicles, indicating that the phenomenon of critical vesicular coverage is also present on mica (Reviakine and Brisson, 2000).
Coexistence of several vesicle rupture mechanisms
Once a few vesicles have ruptured by one of the two above-described mechanisms, further vesicle rupture can also be induced by the active edge of bilayer patches, enhancing the growth of bilayer patches and, eventually, leading to the formation of an SLB.
As mentioned above it takes minutes to hours for isolated DOPS-containing vesicles to rupture (c.f. Fig. 2). On the other hand, the time needed to form an SLB is in the range of a few minutes with the employed lipid concentration (100 µg/mL). This raises the question whether the combination of rupture of isolated vesicles and edge-induced rupture is sufficient to explain the SLB formation at such short timescales. If vesicle adsorption is much faster than the rupture of isolated vesicles, the surface density of intact vesicles may locally become elevated. This suggests in alternative that rupture induced via the critical vesicular coverage coexists with the rupture of isolated vesicles. Future theoretical treatments, similar to the ones described earlier by Zhdanov et al. (2000) may allow distinguishing between these two scenarios. Also, as the time needed to establish the critical vesicular coverage is dependent on the bulk lipid concentration, this experimental parameter provides the means for controlling the balance between the rupture of isolated vesicles and rupture induced by the critical vesicular coverage.
We note that the presence of a local minimum in frequency and a local maximum in dissipation, as commonly observed by QCM-D, is a necessary but not a sufficient condition to identify the critical vesicular coverage as the unique mechanism for the formation of the first bilayer patches. Additional AFM imaging, allowing for the examinations on the level of a single vesicle, or interrupted QCM-D measurements, allowing for the investigation of the overall stability of the adsorbed lipid material, are required to ascertain that the critical vesicular coverage is the only rupture-initiating mechanism present.
The role of shape changes in bilayer patches
AFM data indicate that the interactions between lipids and the mica support are sufficiently strong in order for SLB patches and bound vesicles to be stably located in the presence of small and transient forces. On the other hand, the interactions are sufficiently weak to allow for movements and/or reshaping in the presence of permanent forces, such as the line tension of small noncircular patches (c.f. Fig. 3) (Muresan and Lee, 2001). This sliding motion has important consequences, as reshaping patches can catalyze mergers with neighboring patches (c.f. Fig. 4) or vesicles and thereby enhance vesicle rupture via active bilayer edges.
It may be tempting to postulate that such an effect can be of considerable value to fill up defects in the bilayer during the final stage of SLB formation. In the light of our observation that the propensity for shape changes is decreased (at least locally) at higher coverage (Fig. 6 A) such a conclusion appears premature. However, as demonstrated earlier for silica (Richter et al., 2003), close to defect-free SLBs can be formed without reshaping being present. A potential origin for the decreased speed of shape relaxation may be the presence of contaminations in the bilayer or on the solid support.
Taken together, our results demonstrate that the timescales associated with the different mechanisms of vesicle rupture can vary. As the mechanisms of vesicle rupture can coexist and exhibit some degree of interdependence, a multitude of SLB-formation pathways can occur.
Comparing silica and mica: the role of calcium
We observed several pronounced differences in the deposition of lipid vesicles and in the SLB formation on mica and silica supports, despite the fact that both supports are highly hydrophilic and usually possess an overall negative charge. (Surface charges 0.1 electron per nm2 have been reported for both mica (Pashley, 1981a,b) and silica (Bolt, 1957; Considine and Drummond, 2001), with significant variations as a function of electrolyte concentration, pH, and surface preparation (Toikka and Hayes, 1997; Morigaki et al., 2002; Penfold et al., 2002).) For simplicity, we distinguish two situations, namely the absence and the presence of calcium.
In the absence of calcium, repulsive interactions between the negatively charged support and the negatively charged vesicles as well as attractive interactions between the support and positively charged vesicles are stronger on mica than on silica. This underlines the important role of electrostatic interactions and suggests that mica may be more strongly negatively charged than silica.
We found that vesicles made of DOPC become less flattened on mica than on silica. In opposition, the attractive van der Waals forces are expected to be higher on mica than on silica. (Nonretarded Hamaker constants of A = 2 x 10–20 J and 0.8 x 10–20 J have been reported for two surfaces of mica and silica, respectively, interacting in water (Israelachvili, 1992).) This may indicate that electrostatic double layer forces are sufficiently strong to outweigh the van der Waals forces, despite the small surface potential of DOPC under the employed conditions (Egawa and Furusawa, 1999). Alternatively, additional short-ranged forces may substantially influence the vesicle-support interaction. (Silica and mica are known to exhibit remarkable differences in their short-range interactions (Chapel, 1994; Israelachvili and Wennerström, 1996) the molecular origin of which is still under debate.) Although it remains difficult to identify the exact nature of the interaction forces, this example is illustrative for the complexity of the vesicle-support interactions.
The presence of calcium has a remarkable effect on the lipid deposition on mica as evidenced by the fact that SLBs can be formed at a DOPS content of 80% and more (Reviakine et al., 2001; Richter and Brisson, 2004). Recent results (Richter et al., 2005) indicate that, during SLB formation, DOPS molecules become distributed asymmetrically between the two SLB leaflets. Calcium's exceptional capacity to promote SLB formation correlates, thus, with our finding that DOPS molecules become enriched in the SLB's mica-facing side, suggesting the presence of a particular calcium-mediated interaction between mica and DOPS. It is tempting to identify the (re)distribution of DOPS within and in between (flip-flop) the leaflets of adsorbed vesicles as the determining parameter for the slow rupture kinetics that we observed for isolated DOPS-containing vesicles. Interestingly, Yaroslavov et al. (1994, 1998) have reported that a membrane-binding polycationic polymer can induce the redistribution of negatively charged lipids toward the outer leaflet of SUVs, although the mechanism of this polycation-induced flip-flop remains little understood. It appears conceivable that the calcium-mediated interaction between mica and DOPS, in conjunction with the stress that other vesicle-support interactions exert on the vesicle, may enhance lipid flip-flop in mica-bound vesicles, thereby promoting their rupture.
We note that the SLB-promoting effect of calcium is not restricted to DOPS but extends also to DOPC. Although calcium was observed to facilitate adsorption and SLB formation on silica, too, the effect was considerably less pronounced.
Mica-bound lipid material retained some degree of mobility, whereas bilayer patches were found to be immobile on silica. Could a difference in the roughness of the two supports explain this effect (Rädler et al., 1995)? The roughness of silica wafers employed in the reference study (Richter et al., 2003), was in the range of 0.1 nm (Richter and Brisson, 2003), i.e., well below what is generally considered the thickness of the water layer between the lipids and the solid support (Bayerl and Bloom, 1990; Johnson et al., 1991). We therefore refrain from attributing the observed differences entirely to roughness. Other effects may be induced by the nature of the interactions between the solid support and bilayer patches and/or their boundary.
This study, thus, underlines the important role of the solid support in determining the lipid deposition pathway. Whereas the investigations undertaken in this work identified important parameters that determine the lipid deposition pathway, the results also indicate that the interaction between lipids or lipid assemblies and the solid support can be rather complex. It is hoped that further detailed comparative studies on mica, silica, and other surfaces (such as titanium oxide) at varying ionic strength and calcium content will provide access to information on the quantitative as well as molecular level. In particular, such investigations may help to separate the contribution of vesicle-support, intervesicle and intrabilayer interactions, respectively, in the lipid deposition process and, in particular, to elucidate the molecular nature of the calcium-mediated interaction between DOPS and the support.
Combining QCM-D, AFM, and ellipsometry
We emphasize the importance of combining QCM-D, ellipsometry, and AFM to obtain the results reported in this study. The application of all techniques on an identical support allowed obtaining kinetic information about the adsorption, desorption, and rupture of vesicles with high time resolution as well as detailed information about local structural changes of the lipid assemblies. While the multitechnique approach gives an improved control over possible experimental artifacts, the combination on an identical support opens up for a detailed quantitative investigation of SLB formation or other self-assembly processes.
Influences of the AFM tip on the appearance of immobilized vesicles and SLB patches
The modification of the AFM tip with a lipid-repelling polymer proved essential to reproducibly obtain images of different coexisting lipid structures. Both SLB patches and vesicles could be imaged stably and without lateral displacement. Furthermore, controlled tip forces allowed investigating shape transitions such as patch movement, coalescence or vesicle rupture. Such events can potentially be exploited to quantify the influence of the tip on the apparent size of surface-bound vesicles and bilayer patches. In particular, the coalescence of patches can be exploited to determine the correct size of bilayer patches, and the rupture of adsorbed, flattened vesicles can be used to estimate the diameter of the corresponding nonflattened vesicle as well as to evaluate the influence of the tip on the apparent width of a vesicle (see Supplementary Material).
|
CONCLUSION AND PERSPECTIVES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSION AND PERSPECTIVESSUPPLEMENTARY MATERIALACKNOWLEDGEMENTSREFERENCES |
|---|
Our study provides evidence for important differences in the interaction of lipids with silica and mica, stressing the importance of the solid support, the calcium-EDTA balance, and electrostatic interactions in the SLB-formation process.
The improved imaging of surface-confined lipid structures demonstrated in this study together with the exploitation of the shape transitions (Supplementary Material) for size determinations open up for the reliable quantification of adsorbed lipid material and for the investigation, in detail, of the shape of surface-bound vesicles.
|
SUPPLEMENTARY MATERIAL |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSION AND PERSPECTIVESSUPPLEMENTARY MATERIALACKNOWLEDGEMENTSREFERENCES |
|---|
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSION AND PERSPECTIVESSUPPLEMENTARY MATERIALACKNOWLEDGEMENTSREFERENCES |
|---|
Benda and Martin Bene (Heyrovsk
Institute, Prague, Czech Republic) for help with its setup. We acknowledge Sylvie Bordère (ICMCB, Bordeaux, France) and Ilya Reviakine (Technical University Clausthal, Clausthal-Zellerfeld, Germany) for discussions around the shape of confined bilayer patches and the SLB-formation process, respectively. Ralf Richter was partly supported by the CNRS, the Ministère délégué à la Recherche (France), the Conseil Régional d'Aquitaine (France), and by European Community grant FP6-NMP4-CT2003-505868 "Nanocues". This research was supported by the Conseil Régional d'Aquitaine, the Fonds Européen de Développement Régional, and European Community grant FP6-NMP4-CT2003-505868 "Nanocues".
Submitted on September 28, 2004; accepted for publication February 14, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSION AND PERSPECTIVESSUPPLEMENTARY MATERIALACKNOWLEDGEMENTSREFERENCES |
|---|
Benes, M., D. Billy, A. Benda, H. Speijer, M. Hof, and W. T. Hermens. 2004. Surface-dependent transitions during self-assembly of phospholipid membranes on mica, silica, and glass. Langmuir. 20:10129–10137.[CrossRef][Medline]
Benes, M., D. Billy, W. T. Hermens, and M. Hof. 2002. Muscovite (mica) allows the characterization of supported bilayers by ellipsometry and confocal fluorescence correlation spectroscopy. Biol. Chem. 383:337–341.[CrossRef][Medline]
Benvegnu, D. J., and H. M. McConnell. 1992. Line tension between liquid domains in lipid monolayers. J. Phys. Chem. 96:6820–6824.[CrossRef]
Bolt, G. H. 1957. Determination of the charge density of silica sols. J. Phys. Chem. 61:1166–1169.[CrossRef]
Boxer, S. G. 2000. Molecular transport and organization in supported lipid membranes. Curr. Opin. Chem. Biol. 4:704–709.[CrossRef][Medline]
Chapel, J.-P. 1994. Electrolyte species dependent hydration forces between silica surfaces. Langmuir. 10:4237–4243.[CrossRef]
Considine, R. F., and C. J. Drummond. 2001. Surface roughness and surface force measurement: a comparison of electrostatic potentials derived from atomic force microscopy and electrophoretic mobility measurements. Langmuir. 17:7777–7783.[CrossRef]
Corsel, J. W., G. M. Willems, J. M. M. Kop, P. A. Cuypers, and W. T. Hermens. 1986. The role of intrinsic binding rate and transport rate in the adsorption of prothrombin, albumin and fibrinogen to phospholipid bilayers. J. Colloid Interface Sci. 111:544–554.[CrossRef]
Cuypers, P. A., J. W. Corsel, M. P. Janssen, J. M. M. Kop, W. T. Hermens, and H. C. Hemker. 1983. The adsorption of prothrombin to phosphatidylserine multilayers quantitated by ellipsometry. J. Biol. Chem. 258:2426–2430.
Egawa, H., and K. Furusawa. 1999. Liposome adhesion on mica surface studied by atomic force microscopy. Langmuir. 15:1660–1666.[CrossRef]
Huang, N.-P., J. Vörös, S. M. De Paul, M. Textor, and N. D. Spencer. 2002. Biotin-derivatized poly(L-lysine)-g-poly(ethylene glycol): a novel polymeric interface for bioaffinity sensing. Langmuir. 18:220–230.[CrossRef]
Israelachvili, J. N. 1992. Intermolecular and Surface Forces. Academic Press, London, UK.
Israelachvili, J., and H. Wennerström. 1996. Role of hydration and water structures in biological and colloidal interactions. Nature. 379:219–225.[CrossRef][Medline]
Jass, J., T. Tjärnhage, and G. Puu. 2000. From liposomes to supported, planar bilayer structures on hydrophilic and hydrophobic surfaces: an atomic force microscopy study. Biophys. J. 79:3153–3163.
Johnson, J. M., H. Taekijp, S. Chu, and S. G. Boxer. 2002. Early steps of supported bilayer formation probed by single vesicle fluorescence assays. Biophys. J. 83:3371–3379.
Johnson, S. J., T. M. Bayerl, D. C. McDermott, W. A. Adam, A. R. Rennie, R. K. Thomas, and E. Sackmann. 1991. Structure of an adsorbed dimyristoylphosphatidylcholine bilayer measured with specular reflection of neutrons. Biophys. J. 59:289–294.
Kam, L., and S. G. Boxer. 2003. Spatially selective manipulation of supported lipid bilayers by laminar flow: steps toward biomembrane microfluidics. Langmuir. 19:1624–1631.[CrossRef]
Keller, C. A., K. Glasmästar, V. P. Zhdanov, and B. Kasemo. 2000. Formation of supported membranes from vesicles. Phys. Rev. Lett. 84:5443–5446.[CrossRef][Medline]
Keller, C. A., and B. Kasemo. 1998. Surface specific kinetics of l
–) and dissipation (solid line), for the deposition of SUVs composed of DOPC/DOPS (molar ratio 4:1) on mica in 2 mM CaCl2. Lipid exposure starts at 0 min. The peak in dissipation and the minimum in frequency indicate the presence of intact vesicles bound at an intermediate state. The final frequency shift of –25 Hz and the low final dissipation shift indicate the presence of an SLB. A rinse with buffer (arrow) does not affect the SLB.
, (
average number of charges per lipid molecule). Data for mica (•) and silica (