pubs.acs.org/Langmuir © 2009 American Chemical Society
Formation of Supported Lipid Bilayers on Silica Particles Studied Using Flow Cytometry :: Gustav Nordlund, Rosa Lonneborg, and Peter Brzezinski* Department of Biochemistry and Biophysics, Centre for Biomembrane Research, Stockholm University, SE-10691 Stockholm, Sweden Received October 31, 2008. Revised Manuscript Received January 14, 2009 Silica colloidal particles with functionalized surfaces are used, for example, in studies of membrane proteins or for drug delivery, where novel applications are based on the use of particles covered by lipid membrane bilayers. The mechanism by which such supported lipid bilayers are formed on spherical support is not fully understood. Here, we present results from studies of this process using a new method based on flow cytometry. The approach enabled us to detect particle populations coated and uncoated with lipids in the same sample according to the vesicle:particle surface area ratio. The data suggest that DOPC lipid vesicles efficiently break upon interaction with the silica colloidal particle surface; only a small fraction of the adsorbed vesicles remain unbroken. Furthermore, the data support earlier observations showing that formation of the lipid bilayer at the surface is a cooperative process, where bilayer formation is catalyzed by previously bound membrane fragments.
Introduction Biomembranes with well-defined lipid compositions deposited on solid surfaces are used, for example, as model systems for cell membranes and in functional studies of membrane proteins.1-6 Recent developments in nanofabrication of solid nanosize particles with well-defined dimensions, composition, and architecture have presented us with new technology to manufacture membrane-coated colloidal particles as biomimetic cellular systems.7,8 Such membrane-coated particles are stable and have a wide range of applications, for example, as vectors for drug delivery.9 The mechanism by which surfacesupported lipid bilayers are formed at solid surfaces is not fully understood.7,8 Using quartz crystal microbalance and atomic force microscopy, studies of formation of supported lipid bilayers at planar silica surfaces have shown that the pathway of bilayer formation depends on the membrane composition. For example, for zwitterionic lipids (used in the present study) the vesicles initially adsorb to the surface, after which breaking and bilayer formation occur at a certain critical coverage density.10 Results from more recent studies on silica nanoparticles11 (using cryotransmission electron microscopy) showed that bilayer formation occurs upon *Corresponding author: e-mail
[email protected], phone +46 8 163280; fax +46 8 153679. (1) Lin, W. C.; Blanchette, C. D.; Ratto, T. V.; Longo, M. L. Methods Mol. Biol. 2007, 400, 503–513. (2) Mechler, A.; Praporski, S.; Atmuri, K.; Boland, M.; Separovic, F.; Martin, L. L. Biophys. J. 2007, 93, 3907–3916. (3) Paulick, M. G.; Wise, A. R.; Forstner, M. B.; Groves, J. T.; Bertozzi, C. R. J. Am. Chem. Soc. 2007, 129, 11543–11550. (4) Reimhult, E.; Kumar, K. Trends Biotechnol. 2008, 26, 82–89. (5) Salafsky, J.; Groves, J. T.; Boxer, S. G. Biochemistry 1996, 35, 14773– 14781. (6) Watts, T. H.; Brian, A. A.; Kappler, J. W.; Marrack, P.; McConnell, H. M. Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 7564–7568. (7) Richter, R. P.; Berat, R.; Brisson, A. R. Langmuir 2006, 22, 3497–3505. (8) Troutier, A. L.; Ladaviere, C. Adv. Colloid Interface Sci. 2007, 133 1–21. (9) Carmona-Ribeiro, A. M. J. Liposome Res. 2007, 17, 165–172. (10) Richter, R.; Mukhopadhyay, A.; Brisson, A. Biophys. J. 2003, 85, 3035–3047. (11) Mornet, S.; Lambert, O.; Duguet, E.; Brisson, A. Nano Lett. 2005, 5, 281–285.
Langmuir 2009, 25(8), 4601–4606
interaction of vesicles either by direct adsorption to the silica particles or by interaction of preformed patches on the particles via so-called “active edge” effects.12 Furthermore, these results showed that the bilayer follows the surface contour. To further investigate the mechanism by which silica surfaces are covered by lipid membranes, we present results from studies of formation of supported lipid bilayers on silica particles with a diameter of 600 ( 30 nm. To monitor lipid adsorption to the particles, we developed a novel approach based on the use of flow cytometry. The technique allows rapid analysis of a large particle population and statistical analysis of the adsorption process. Flow cytometry has traditionally mainly been used in immunology studies to monitor fluorescent antibodies, but the number of applications has increased in recent years.13-15 This method is based on analysis of light scattering, which is related to size,16 and the fluorescence of particles with diameters above 150 nm.16 In the present study we investigated adsorption of phosphatidylcholine (zwitterionic lipid) vesicles, supplemented with a small fraction of lipids labeled with a fluorescent probe (fluorescein), to solid silica particles over a wide range of vesicle:particle surface area ratios (Av/Ap, the nomenclature is adopted from ref 17).
Experimental Section Solid Silica Particles. Monodisperse silica particles were obtained as a dry powder from Tokuyama Soda Co., Tokyo, Japan. The particle diameter was 600 ( 30 nm (determined using (12) Reviakine, I.; Brisson, A. Langmuir 2000, 16, 1806–1815. (13) Aharoni, A.; Thieme, K.; Chiu, C. P.; Buchini, S.; Lairson, L. L.; Chen, H.; Strynadka, N. C.; Wakarchuk, W. W.; Withers, S. G. Nat. Methods 2006, 3, 609–614. (14) Hedhammar, M.; Stenvall, M.; Lonneborg, R.; Nord, O.; Sjolin, O.; Brismar, H.; Uhlen, M.; Ottosson, J.; Hober, S. J. Biotechnol. 2005, 119, 133– 146. (15) Rieseberg, M.; Kasper, C.; Reardon, K. F.; Scheper, T. Appl. Microbiol. Biotechnol. 2001, 56, 350–360. (16) Shapiro, H. M. Practical Flow Cytometry, 4th ed.; Wiley-Liss: Hoboken, NJ, 2003; p l, 681 pp. (17) Carmona-Ribeiro, A. M.; De Moraes Lessa, M. Colloids Surf., A 1999, 153, 355–361.
Published on Web 3/5/2009
DOI: 10.1021/la8036296
4601
Article
Nordlund et al.
dynamic light scattering, Nano Z, Malvern Instruments), and the particle density was 2 g/cm3. The particles were dispersed by sonication at a concentration of 3 mg/mL in 100 mM KCl, 25 mM HEPES, pH 7.4. Phosphatidylcholine Vesicles. 1,2-Dioleoyl-sn-glycero-3phosphocholine (DOPC) (Avanti Polar Lipids) and N-(fluorescein-5-thiocarbamoyl)-1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine (fluorescein-DHPE) (Avanti Polar Lipids) at a molar ratio of 200:1 were mixed in 100 mM KCl, 25 mM HEPES, pH 7.4 to a final lipid concentration of 3.8 mM (forming a lipid suspension). The lipid solution was sonicated on ice for 10 min (Sonicator XL, heat systems equipped with a titanium probe) to form vesicles, and the sample was then centrifuged (2880g, 15 min, 4 °C) to remove lipid aggregates from the vesicle solution. The diameter of the vesicles was determined to be 25 ( 5 nm using dynamic light scattering (Nano Z, Malvern Instruments). The vesicle size remained constant for ∼1 day. Membrane Coating of Solid Silica Particles. Particles and vesicles were mixed at different ratios (see text), keeping the particle concentration constant and varying the concentration of vesicles. The samples were then stirred at 22 ( 1 °C for 1 h. To remove excess vesicles, the samples were centrifuged at 4500g for 15 min, and the pellet containing particles was washed in a solution composed of 100 mM KCl, 25 mM HEPES at pH 7.4 two times by repeated centrifugation and dilution. The samples were then analyzed using a flow cytometer. Vesicle Rupture. To establish whether or not the vesicles rupture upon interaction with the particles, coating was performed with DOPC vesicles containing fluorescein-tagged lipids (fluorescein-DHPE at a ratio of 1:1000) or DOPC vesicles that were prepared in 250 μM fluorescein (Sigma-Aldrich), 100 mM KCl, 25 mM HEPES, pH 7.4 (i.e., fluorescein was found inside the vesicles). The vesicles were run three times on a PD-10 desalting column (GE HealthCare) to remove free fluorescein from the bulk solution. The fluorescence per vesicle was determined by measuring the total amount of phosphorus18,19 (to determine the lipid content) and the fluorescence of each sample (using a fluorometer, Cary Eclipse). Coating of silica particles, at a vesicle to particle surface area ratio (Av/Ap) of 15 (used in ref 11), was performed as described above, and the fluorescence of the coated particles was measured using flow cytometry. Flow Cytometry. The flow cytometric analyses were performed on a FACS Calibur instrument (BD Biosciences, San Jose, CA). The same instrument settings were used throughout the measurements. Fluorescence was detected through a 530 ( 15 nm (green) band-pass filter. For each cell sample data from 10 000 events were collected. Each sample was gated in order to remove the background contribution of the buffer, which was presumably due to presence of salt crystals. The liquid vector was BD FACSFlow Sheath Fluid (BD Biosciences, Canada). Surface Area Calculations. The Av/Ap ratio was calculated by dividing the total vesicle surface area (Av) by the total particle surface area (Ap) in each sample. To determine the total vesicle surface area, the number of lipid molecules present in one 25 nm vesicle was calculated by dividing the total membrane volume of one vesicle by the volume of one phosphatidylcholine molecule, which is 1.25 nm3.20 4π312:53 4πð12:5 -5Þ3 3 3
!
=
1:25
ð1Þ
(18) Fiske, C. H.; Subbarow, Y. J. Biol. Chem. 1925, 66, 375–400. (19) Chen, P. S.; Toribara, T. Y.; Warner, H. Anal. Chem. 1956, 28, 1756– 1758. (20) Huang, C.; Mason, J. T. Proc. Natl. Acad. Sci. U.S.A. 1978, 75, 308– 310.
4602
DOI: 10.1021/la8036296
Here we have assumed that the thickness of the membrane is 5 nm. The details of the Av/Ap calculation can be found in the Supporting Information.
It should be noted that the size distributions of the vesicles (25 ( 5 nm) and the silica particles (600 ( 30 nm) introduce an error in the Av/Ap ratio that is ∼20% of its value.
Results and Discussion The buffer concentration, pH, and ionic strength used in our studies were based on previously reported successful strategies for coating silica particles.11 The ionic strength used (∼100 mM) is adjusted to allow the vesicles and particles to approach, interact, and form supported lipid bilayers. Unfortunately, under the same conditions the particles have an increased tendency to form aggregates (see discussion below). An initial analysis of the silica particles using the flow cytometer showed that under the conditions chosen for coating there were two major populations. These populations differed in their light scattering properties, mainly with respect to the side-scattered light (Figure 1) but also to some extent in the forward scattered light (not shown). Side scattering is a relatively complex phenomenon, which depends, for example, on surface roughness and internal granularity. In addition, in the submicron range side scattering is also determined by the particle size.16,21,22 Forward scattering can be regarded as a rough measurement of size, but it is also affected by the refractive index and should not be considered as an absolute measurement of size, especially not when the size of the object is close to the wavelength of the laser light.16 Taken together, the difference between the two populations in side- and forward-scattered light and the tendency of the particles to form aggregates under the conditions used indicate that one of the populations consists of monodisperse particles (green in Figure 1) and the second population consists of aggregated particles (red in Figure 1). It should be noted that the “fluorescence” of the uncoated particles originates from light that is scattered by the particles into the fluorescence detector although the particles in themselves are not fluorescent. The flow cytometer detects fluorescence originating from lipid vesicles adsorbed to the silica particles. However, the data cannot discriminate between vesicles that are broken and adsorbed to the surface (cf. supported lipid bilayer) and those that are intact and just glued to the particle surface. To resolve this issue and determine the fraction of vesicles that are broken upon interaction with the particle surface, we used the following approach. Two vesicle samples were prepared, one having fluorescein dye on the inside and one containing lipids tagged with fluorescein. To normalize the fluorescence signal, we determined the total amount of lipids in each sample (by determining the amount of phosphorus; see Experimental Section) and calculated the fluorescence per lipid molecule, which is proportional to the fluorescence per vesicle. The ratio of the fluorescence originating from vesicles containing (internal) fluorescein prior to reaction with the silica particles (F pre int ) and that originating from vesicles containing the fluoresceintagged lipids (F pre tag ) (Figure 2) was (both values were determined at the same vesicle concentration) pre pre =Fint = 0:10 Ftag
ð2Þ
(21) Hercher, M.; Mueller, W.; Shapiro, H. M. J. Histochem. Cytochem. 1979, 27, 350–352. (22) Zarrin, F.; Risfelt, J. A.; Dovichi, N. J. Anal. Chem. 1987, 59, 850– 854.
Langmuir 2009, 25(8), 4601–4606
Nordlund et al.
Article
Figure 1. Side scattering and fluorescence of the silica particles without and with lipid vesicles. (a) The concentration of particles was 0.4 mg/mL. The largest fraction of the population is monodisperse (green), and a small part of the population forms aggregates (red). (b) Particles mixed with DOPC/fluorescein-DHPE vesicles at a vesicle to particle surface area ratio of 290. Upon addition of the vesicles, the subpopulation of monodisperse particles (green) decreased relative to that of the aggregates (red). The small residual particle population with the lowest intensity of side-scattered light is seen to the left of the green peak in Figure 3. The population colored in green is that corresponding to the green peak in Figure 3.
Figure 2. Determination of the fraction unbroken vesicles upon binding to the silica particles. The fluorescence of vesicles containing fluorescein (flu)-labeled (tagged) lipids (white) and that of vesicles with enclosed (trapped) fluorescein (dark) is shown to the left (“before coating”). To the right is shown fluorescence of the silica particles after interaction with vesicles containing fluorescein-labeled lipids (white) and vesicles containing enclosed fluorescein (dark). Note that the bars are normalized to the highest level of vesicle fluorescence and particle fluorescence (different normalization factors), respectively. The fluorescence of vesicles and particles cannot be compared directly because different types of instruments were used in the studies. When evaluating the data, we only consider ratios of the numbers (see text). In the next step, each of the two vesicle samples was allowed to react with silica particles (we choose an Av/Ap ratio of 15 because this ratio was shown to result in efficient coating of silica particles11). The fluorescence originating from the particles that had reacted with vesicles containing the dye, F post int , and those that had reacted with vesicles containing tagged lipids, F post tag , was then determined using flow cytometry, and the ratio was found to be post post =Fint = 8:3 Ftag
ð3Þ
It should be noted that the flow cytometer only detects fluorescence from the silica particles (in this calculation the monodisperse population) and not from the fluorescently labeled liposomes in solution. Thus, we study only membranes (lipids) that are adsorbed to the silica particle surfaces. Langmuir 2009, 25(8), 4601–4606
The fraction of intact, unbroken vesicles adsorbed to the silica particles can be estimated from the relative fluorescence contribution from vesicles with internally trapped fluorescein. If the fluorescence per vesicle was the same independently on whether the fluorophore was in the membrane or inside the post post vesicle, this value would be F post int /(F int + F tag ). Because these fluorescence values are not the same, we scale F post tag by the ratio of the fluorescence from internalized fluorescein and pre that from fluorescein-labeled lipids, F pre int /F tag , determined independently on a vesicle sample (eq 2) prior to measurements on the silica particles. We obtain the following expression for the fraction of intact, unbroken vesicles: post Fint post Fint
þ
pre post Fint Ftag pre Ftag
¼
1 1þ
post pre Ftag Fint post pre Ftag Fint
= 1:2%
ð4Þ
The data are consistent with earlier observations showing that those vesicles that interact with the silica particle surface break effectively.11 The above-described estimation is done assuming that the fluorescein-DHPE fluorescence was the same in the vesicle membrane and in the membrane on the particle surface. Considering that the surrounding environment of the fluorophore is essentially the same in both cases (water on both sides of the membrane; there is ∼10 A˚ water layer between the membrane and the silica surface23-25), we consider this assumption to be valid. Furthermore, we assume that the fluorescein fluorescence is the same independently if the dye is on the inside or outside of the vesicle. This assumption most likely leads to an underestimation of the F pre int value because of fluorescence self-quenching due to the higher fluorescein concentration on the inside than on the outside of the vesicles (a larger F pre int would result in an even smaller fraction unbroken vesicles; cf. eq 4). We also assume that the unruptured vesicles do not adsorb to membrane patches of surfaceadsorbed vesicles. Finally, we assume that the particle coverage is the same when using vesicles with fluorescein on the inside or vesicles with fluorescein-labeled lipids. Figure 3 shows side scattering and fluorescence of the monodisperse and aggregated particles for samples containing vesicles with different Av/Ap ratios. In this figure we show the aggregated (in red) and the monodisperse (in green) subpopulations separately. In addition, the entire population (sum of the red and green subpopulations) is shown in blue. Upon mixing the silica particles and lipid vesicles, the sidescattering pattern in the aggregate population changed (see also Figure 1b), which is presumably due to an increase in aggregate size and in the aggregation propensity. We believe that the aggregation occurs because lipids adsorbed to the particle surfaces mediate particle-particle interactions, which presumably facilitates aggregation at the ionic strength used in these measurements (see also Discussion). The side scattering of the monodisperse particle population was not affected by the addition of vesicles; however, a slight decrease in forward scattering could be observed (see Supporting Information). The fluorescence of both populations increased significantly upon addition of the fluorescent vesicles. (23) Bayerl, T. M.; Bloom, M. Biophys. J. 1990, 58, 357–362. (24) Johnson, S. J.; Bayerl, T. M.; McDermott, D. C.; Adam, G. W.; Rennie, A. R.; Thomas, R. K.; Sackmann, E. Biophys. J. 1991, 59, 289–294. (25) Koenig, B. W.; Krueger, S.; Orts, W. J.; Majkrzak, C. F.; Berk, N. F.; Silverton, J. V.; Gawrisch, K. Langmuir 1996, 12, 1343–1350.
DOI: 10.1021/la8036296
4603
Article
Nordlund et al.
Figure 3. Side scattering and fluorescence of samples with increasing vesicle to particle surface area ratios (Av/Ap). Side scattering (left column) and fluorescence (right column) histograms for samples with (Av/Ap) ratios of 0 (only particles, a, b), 0.15 (c, d), 0.74 (e, f), 1.5 (g, h), and 290 (i,j). In each histogram the monodisperse particle fraction (green), the aggregate fraction (red), the entire particle population (blue, which is the sum of the red and green populations), and the entire particle population from (a, b) (gray) is shown. Note that this entire particle population marked in gray is that observed under the measuring conditions in (a, b), but not under the measuring conditions corresponding to the other panels. In the other panels these data are given only as a reference. The arrow in (f) indicates the double peak of the monodisperse particle population in the 0.74 (Av/Ap) ratio sample (this sample was chosen as an example because the ratio of coated and uncoated particles in this case is close to one; see also Figure 4). In the sample with an Av/Ap ratio of 290 (Figure 3j) there was essentially no fluorescence in the range of that of the uncoated particles (cf. Figure 3b; note that the gray peak corresponding to the uncoated particles is identical to the blue peak in Figure 3b; it is shown here only for reference). Because the detection limit of the flow cytometer is around 150 nm, any residual amount of free fluorescent vesicles, possibly present in the samples, would not have any effect on the measurement. 4604
DOI: 10.1021/la8036296
To estimate the fraction of monodisperse particles that are coated and uncoated at each Av/Ap ratio, we gated the data such that it was split at a fluorescence value at which the peaks corresponding to the uncoated and the fully coated monodisperse particles (see the green area in Figure 3f) converge (cf. the dashed-dotted line at a fluorescence value of ∼30 AU in Figure 4a,b for the 0.74 Av/Ap ratio data) and then estimated the number of particles corresponding to each of the two peaks for the different Av/Ap ratios. The fluorescence histogram of the 0.74 Av/Ap ratio monodisperse particle population (indicated by an arrow in Figure 3f and shown in Figure 4b) has two peaks. As indicated above, the position of one of these peaks coincides with that of the particles alone (uncovered particles, light gray) while the other is closer to the peak corresponding to the largest number of vesicles (i.e., an Av/Ap ratio of 290, light yellow). This comparison indicates that the two peaks correspond to uncoated and partly coated particles, respectively. In this context it is worth noting that the peak corresponding to the partly coated particles (the abscissa value of the dark yellow peak maximum is ∼60% of that corresponding to the light yellow peak position) is well separated from that corresponding to the uncoated particles. In other words, at an Av/Ap ratio of 0.74 there are fully uncoated particles and particles coated at a fraction corresponding to the 60% maximum fluorescence value coexisting in the same sample. This scenario is also seen from a comparison of the green peaks in Figure 3d,f,h, which shows that the peak overlapping with that corresponding to the uncoated particles (cf. green peak in Figure 3b) becomes gradually smaller, while the peak overlapping with that corresponding to the fully covered particles (cf. green peak in Figure 3j) becomes gradually larger. The observation of two distinct peaks corresponding to two particle populations; those that are uncovered and those that are covered to a large extent;indicates that formation of the lipid bilayer at the silica particle surface is a cooperative process. Otherwise, we would observe one peak corresponding to a statistical distribution of particles that are progressively more covered (i.e., one peak with a gradually increasing fluorescence) with an increasing Av/Ap ratio. This indication of cooperativity is supported by previous observations indicating that vesicle breaking and formation of the bilayer is accelerated by a preexisting adsorbed lipid bilayer (cf. refs 11 and 12). In Figure 4c the fraction of coated (i.e., fluorescent) monodisperse particles is shown as a function of the Av/Ap ratio. As seen in the figure, initially this fraction rapidly increased with increasing amount of added vesicles to reach ∼80% of the maximum value at an Av/Ap ratio of 5. Then the fluorescence increased slowly toward the saturation value. The figure also shows the mean fluorescence of the coated particles as a function of the Av/Ap ratio. These data are consistent with the previously used Av/Ap ratio of ∼15 for efficient coating of silica particles.11 As seen in Figure 4b, the peak position for the 290 Av/Ap ratio is ∼110 AU, while that for an Av/Ap ratio of 0.74 is 65 AU. Assuming that for the Av/Ap ratio of 290 all monodisperse particles are fully covered by lipid monolayers, at an Av/Ap ratio of 0.74, a fluorescence value of 0.74 110 AU = 80 AU would be expected. The lower value of ∼65 AU is consistent with the observation of fluorescence from the fraction of particles that have formed aggregates. In other words, a fraction of the lipid vesicles (or membranes attached to the particle surface) are incorporated into the aggregates and are therefore not found in the monodisperse particle Langmuir 2009, 25(8), 4601–4606
Nordlund et al.
Article
Figure 4. Particle coating as a function of the vesicle to particle surface area ratio (Av/Ap). (a) Separation of coated (yellow area) and uncoated (gray area) monodisperse particles in the 0.74 Av/Ap ratio sample. The coloring of the plot reflects the amount of particles at each point, where red corresponds to the highest amount and blue the lowest amount of particles. (b) A fluorescence histogram showing the coated (yellow plot) and uncoated (gray plot) monodisperse particles in the 0.74 Av/Ap ratio sample. Also shown for comparison are the monodisperse particle populations in the samples with only particles (light gray, dashed line) and with a 290 Av/Ap ratio (light yellow, dashed line). (c) Fraction of monodisperse lipidcoated particles as a function of the Av/Ap ratio. The data for each Av/Ap ratio were analyzed as shown in (a, b), and the fraction of the population corresponding to the yellow peak has been plotted. Note that the error in the Av/Ap values is ∼20%. The graph also shows the mean fluorescence of the monodisperse fluorescein particles.
population. This is most likely the reason for the saturating behavior observed in Figure 4c where full coverage of the particles is reached only at Av/Ap ratios >1 (i.e., the fluorescence curve of monodisperse particles does not reach 100% at lower Av/Ap values). Note that the data discussed above indicate that of those vesicles that directly interact with the silica particle surfaces a great majority break. We had to use an ionic strength of ∼100 mM to minimize the electrostatic interactions between the lipid vesicles and the particles thereby facilitating coating. At this ionic strength the particles displayed an increased tendency to aggregate, presumably due to electrostatic shielding, which weakens repulsion between the particles. At low ionic strength the particles do not aggregate; however, such conditions do not facilitate coating and could not be used. A comparison of the data Langmuir 2009, 25(8), 4601–4606
shown in the right-hand side panels in Figure 3 shows that the fraction of aggregates first increases with increasing Av/Ap ratio, but at the highest Av/Ap ratio the fraction of aggregates is again lower. Thus, the data indicate that aggregation may be facilitated at low Av/Ap ratios (from 0.15 to 1.5, Figure 3) by binding of vesicles to aggregates, where the lipids then bind to other particles (or aggregates), thereby increasing the tendency toward aggregation.26 The decrease in the fraction of aggregates at high Av/Ap ratios is presumably due to a large fraction of particles being entirely coated by lipid membranes, which reduces the probability for aggregation.
(26) Troutier, A. L.; Delair, T.; Pichot, C.; Ladaviere, C. Langmuir 2005, 21, 1305–1313.
DOI: 10.1021/la8036296
4605
Article
Nordlund et al.
Vesicle adsorption to planar solid surfaces has been studied previously in great detail,7 and the current study, using flow cytometry, should be considered a complement to those presented earlier. For example, Keller et al.27 found that egg phosphatidylcholine vesicles initially adsorb to a SiO2 surface and then break first after reaching a critical surface density. This cooperativity during formation of the supported lipid layer must be due to interactions between the surface-bound vesicles. Brisson and colleagues7,10 identified different mechanistic pathways for vesicle deposition on planar silica supports depending on the lipid composition of the vesicles. In the case of neutral and weakly negatively charged vesicles breaking occurred after initial adsorption of intact vesicles at a specific density. This cooperative behavior may be further enhanced by the “active edge” effect10,12 where patches of lipid layers adsorbed at the particle surface grow by incorporating new vesicles. The process is driven by the hydrophobic effect due to the exposure of the hydrophobic portion of the lipid layer at the edges of the adsorbed layer. Adsorption of lipid layers onto particles has not been characterized as thoroughly as that to solid planar surfaces.8 Mornet et al.11 found that upon adsorption of vesicles to nanosize particles the vesicles appear to break independently at the surface, forming bilayer patches which incompletely cover the particles. The continued adsorption process was proposed to involve the “active edge” effect as outlined above. The data from this study indicate that vesicles interact during adsorption to the surface because of the cooperative nature of the process. We can rule out the possibility that intact vesicles first bind to the surface and then break only after reaching a (27) Keller, C. A.; Glasmastar, K.; Zhdanov, V. P.; Kasemo, B. Phys. Rev. Lett. 2000, 84, 5443–5446.
4606
DOI: 10.1021/la8036296
critical density. If this was the case, we would observe one peak with a gradually increasing fluorescence with an increasing number added vesicles. Instead, for example, at an Av/Ap ratio of 0.74 we observe two coexisting populations (Figures 3f and 4a,b), one of fully uncoated and one of partly coated particles. Consequently, a more likely scenario is that upon binding to the particle surface the vesicles break, and then formation of a continuous layer proceeds by interaction of the patches with newly adsorbed vesicles. The process may also involve lateral migration of the vesicles along the surface, although it is presently unclear how mobile these vesicles are after being bound to the surface.7 In conclusion, the results from this study indicate that lipid vesicles break upon interaction with the silica particle surface forming a supported lipid membrane layer. The process is cooperative such that breaking of the lipid vesicles and formation of the bilayer is stimulated by already formed membrane patches at the particle surface (cf. refs 11 and 12). Furthermore, the results indicate that the fraction vesicles that remain intact after interaction with the surface is very small. Acknowledgment. The project was supported by funding from the Centre for Biomembrane Research at Stockholm University and the Knut and Alice Wallenberg Foundation. :: Silica particles were supplied by Lennart Bergstrom at the Department of Inorganic Chemistry, Stockholm University, Stockholm, Sweden. Supporting Information Available: Details of the Av/Ap calculation. This material is available free of charge via the Internet at http://pubs.acs.org.
Langmuir 2009, 25(8), 4601–4606
