to the Inner Leaflet of Liposomal Membranes - ACS Publications

Biomacromolecules 2002, 3, 324-332

324

Mimicking a Cytoskeleton by Coupling Poly(N-isopropylacrylamide) to the Inner Leaflet of Liposomal Membranes: Effects of Photopolymerization on Vesicle Shape and Polymer Architecture Oliver Stauch,*,† Thomas Uhlmann,† Margret Fro¨hlich,† Ralf Thomann,‡ Mahmoud El-Badry,†,§ Yong-Keun Kim,† and Rolf Schubert† Department of Pharmaceutical Technology, Hermann-Herder-Strasse 9, Albert-Ludwigs-Universita¨t Freiburg, D-79104 Freiburg im Breisgau, Germany, and Freiburger Materialforschungszentrum und Institut fu¨r Makromolekulare Chemie, Stefan-Meier-Strasse 21, Albert-Ludwigs-Universita¨t Freiburg, D-79104 Freiburg im Breisgau, Germany Received September 24, 2001; Revised Manuscript Received December 14, 2001

Networks of N-isopropylacrylamide (NIPAM) copolymers, coupled to spherical phospholipid bilayers, are suitable as a model for the study of the interaction between the cytoskeleton and cellular membranes, as well as for promising new drug delivery systems with triggerable drug release properties and improved stability. In this article, we describe a simple preparation technique for liposomes from egg phosphatidyl choline (EPC) encapsulating a cross-linked NIPAM-TEGDM copolymer skeleton (tetraethylene glycol dimethacrylate, TEGDM) which is coupled only to the inner monolayer by a novel membrane anchor monomer. Polymerization in the lipid vesicles was initiated at the inner membrane surface by the radical initiator 2,2-diethoxy-acetophenone (DEAP) permeating through the membrane from the outside. The effects of photopolymerization and polymer formation on vesicle shape and membrane integrity were studied by transmission electron microscopy (TEM), cryo-TEM, and atomic force microscopy (AFM). Upon UV irradiation, approximately 100% of the vesicles contained a polymer gel and only occasional changes in the spherical shape of the liposomes were observed. The architecture of the polymer network inside the liposomal compartment was determined by the conditions of the photopolymerization. Composite structures of polymer hollow spheres or solid spheres, respectively, tethered to spherical membrane vesicles were produced. The increased stability of the polymer-tethered lipid bilayers against solubilization by sodium cholate, compared to pure EPC vesicles, was determined by radiolabeling the lipid membrane. Introduction Phospholipid bilayers are commonly employed as model systems for biological membranes. However, structure and dynamics of cell membranes highly depend on various interactions between membrane lipids and proteins, which are present in approximately equal quantities by weight.1 Furthermore, the bilayer couples to the cytoskeleton on the cytoplasmatic side. This highly flexible network of peripheral protein filaments attached to the inner leaflet by anchoring to integral membrane proteins is best understood for the human erythrocyte and can be regarded as the key for the unique mechanical stability and viscoelastic properties of the membrane during its lifetime of about 120 days.2-7 In the last two decades, the erythrocyte’s cytoskeleton has inspired numerous strategies to achieve a stabilization of phospholipid bilayers: vesicles made from photopolymer* To whom correspondence may be addressed: tel, (0761) 203-6360; fax, (0761) 203-6366; e-mail, [email protected]. † Department of Pharmaceutical Technology, Albert-Ludwigs-Universita ¨t Freiburg. ‡ Freiburger Materialforschungszentrum und Institut fu ¨ r Makromolekulare Chemie, Albert-Ludwigs-Universita¨t Freiburg. § Current address: Assiut University, Assiut, Egypt.

izable phospholipids,8-13 liposomes coated with polymer fragments,14-19 or a hydrophobic polymer scaffold in the interior of the liposomal membrane.20 The physicochemical properties as well as the potential use of these modified bilayers as novel drug delivery systems have been discussed. However, much fewer attempts have been made to get closer to the composite structure in which a quasi-two-dimensional hydrophilic network is linked to lipophilic membrane anchors in the inner liposomal membrane leaflet, mimicking the cytoskeleton of cellular plasma membranes.21 Polymerization of liposome-encapsulated hydrophilic monomers by UV light turned out to become the most important method for the creation of polymer-gel containing vesicles.21-23 Surprisingly little work has been done so far to elucidate the effects of UV-radiation-induced radical polymerization inside the liposomes on vesicle shape and membrane integrity. These phenomena, however, should be regarded as the crucial limitations to their use as model system for fundamental biophysical research as well as for the role as a drug delivery system. In this article we introduce a simple method to obtain liposomes encapsulating a flexible, cross-linked NIPAMTEGDM copolymer network attached only to the inner

10.1021/bm015613y CCC: $22.00 © 2002 American Chemical Society Published on Web 01/29/2002

Biomacromolecules, Vol. 3, No. 2, 2002 325

Mimicking a Cytoskeleton inside Liposomes Scheme 1. Synthesis of Reactive Membrane Anchor Molecule DOGM (6)

monolayer by a stable anchor monomer (1,2-distearyl-3octaethylene glycol ether methacrylate, DOGM (6), Scheme 1) homogeneously distributed in the uncharged egg yolk phosphatidylcholine bilayer. The stable fixation of the anchoring monomer to the lipid membrane was achieved by two alkyl chains sticking between the phospholipid molecules. A nonionic, hydrophilic spacer with eight ethylene oxide residues was introduced to avoid the permanent contact between polymer network and lipid membrane. Similarly to biological cells, the individual flexibility of the membrane and the skeletal network should be maintained. Vesicles of various size (100 up to 1500 nm in diameter) and lamellarity were loaded with a mixture of the comonomer N-isopropylacrylamide (NIPAM, 1 M) and the cross-linker tetraethylene glycol dimethacrylate (TEGDM, 0.025 M). As the initiation of radical polymerization by UV irradiation is dependent on the presence of a radical former, the formation of polymer chains and the architecture of the polymer network were supposed to be influenced by using the lipid membrane as a selective permeation barrier. NIPAM homopolymers as well as some copolymers are well-known for their sudden phase separation above a lower critical solution temperature (LCST) in aqueous solution.14,17,24,25 When grafted onto liposomal membranes from the outside, thermal contraction of the polymer causes membrane defects, leading to a release of entrapped materials.15-17 During this reversible collapse, the liposomes were not destroyed and the polymer chains remained anchored to the membrane.14 However, studies of this phenomenon in our system showed this to be more complex because the NIPAM-TEGDM copolymer is highly crosslinked and fixed to the inner vesicle membrane by numerous anchoring groups. While this fascinating phenomenon is subject to additional investigations, the present work focuses on the preparation technique and the characterization of vesicles with two types of an inner polymer skeleton, i.e., a three-dimensional network that filled the whole space in the vesicles and a thin, quasi-two-dimensional polymer layer underneath the membrane, respectively. To avoid possible deformation of the vesicle shape due to the phase transition of the tethered polymer upon heating, the preparation and characterization were performed below the LCST. All vesicles were subject

to extensive analysis by a combination of transmission electron microscopy (TEM), cryo transmission electron microscopy (cryo-TEM), and atomic force microscopy (AFM) with the application of various staining and preparation techniques. Experimental Section Materials and Methods. Synthesis of a Reactive Membrane Anchor Molecule (DOGM). The preparation of 1,2distearyl-3-octaethylene glycol glycerol ether methacrylate (DOGM) is outlined in Scheme 1. In detail, 1,2-distearylbenzyl glycerol ether (2) was prepared by Williamson ether synthesis. To a solution of 10.88 g (59 mM) of 3-benzyloxy1,2-propanediol (1) in 100 mL of absolute dioxane, 5.7 g (240 mM) of sodium hydride was added at room temperature and later 79.6 g (240 mM) of stearyl bromide. The mixture was refluxed for 4 days and evaporated. The raw product was extracted with 500 mL of diethyl ether/water (1:1). The organic phase was separated, washed three times with water, and dried (Na2SO4). After evaporation, the product was isolated and purified by column chromatography on silica gel 60 (12 × 5 cm, cyclohexane/diethyl ether, 10:1). The yield was 60%. The removal of the benzyl protective group of 2 was achieved by hydrogenation. A 218 g (318 mM) portion of 2 and 5% (m/m) palladium-carbon catalyst were dissolved in 100 mL of n-hexane/dioxane (1:1), and the mixture was stirred, evaporated, and filled with hydrogen gas. After removal of the catalyst by filtration (glass filter D4), white crystals of 1,2 distearyl glycerol ether (DSGE) (3) were obtained and purified by evaporation and recrystallization from petroleum ether, yielding 92%. To introduce the spacer group, DSGE (91.4 g/640 mM) was etherified with a surplus of dichlorodiethyl ether as described. 1,2Distearyl-3-diethylene glycol chloride glycerol ether (4) was purified by chromatography on silica gel 60 (cyclohexane/ diethyl ether, 4:1) as yellow-white, waxy product to yield 62%. The PEG chains were enlarged by the etherification of 4 in absolute dioxane with a 10-fold surplus of hexaethylene glycol (reflux for 8 days). 1,2-Distearyl-3-octaethylene glycol glycerol ether (DOGE) (5) was obtained after purification by chromatography on silica gel (ethyl acetate/ methanol, 9:1) as a yellow-white, waxy product, yielding

326

Biomacromolecules, Vol. 3, No. 2, 2002

53%. The reactive methacryl group was finally added by esterification with methacryloyl chloride. A solution of methacryloyl chloride in absolute THF (10 mL) was added at 0 °C within 15 min to a solution of 1 g of DOGE (1.1 mM) and 0.64 g of triethylamine (6 mM) in 80 mL of absolute THF. The mixture was stirred for 48 h at room temperature. After evaporation, petroleum ether/water (300 mL, 1:1) was added. The water phase was washed five times with petroleum ether. After the evaporation of the united organic phase and chromatography on silica gel (ethyl acetate), a colorless, waxy 1,2 distearyl-3-octaethylene glycol glycerol ether methacrylate (DOGM) (6) was obtained, yielding 55%. 1H NMR (CDCl3): δ 0.9 (t/6H), 1.2-1.4 (m/ 64H), 1.9 (s/3H), 3.4 (t/2H), 3.4-3.7 (m/37H), 4.3 (t/2H), 5.5 (d/1H), 6.1 (d/1H). All chemicals were purchased from Sigma Chemicals and used as received. Compounds for the Vesicle Preparation. Pure egg yolk phosphatidyl choline (EPC) for liposome preparation was a generous gift from Lipoid GmbH, Germany. N-Isopropyl acrylamide (NIPAM, Acros Organics, Belgium) was used as comonomer after purification by recrystallization from a toluene/n-hexane mixture.26 The cross-linking agent tetraethylene glycol dimethacrylate (TEGDM, Ro¨hm, Germany) and the initiator 2,2-diethoxyacetophenone (DEAP, Sigma Chemicals, St. Louis, MO) were used as received. Differential Scanning Calorimetry (DSC). The miscibility of EPC/DOGM was determined by DSC (Microcal MC 2 DSC microcalorimeter). Mixtures of EPC (1 mg/mL) with various amounts of DOGM (24, 12, 6, 3 mol %) in THF were dried by rotary evaporation in a round-bottom flask, producing a clear, thin film. After suspension in water, the samples were analyzed at a heating rate of 1 °C min-1 in the temperature range of 10-100 °C. Vesicle Preparation and Entrapping of Monomer Solution. The preparation of variously sized unilamellar liposomes was achieved by a detergent removal technique using N-octyl-β-D-glucopyranoside (octylglucosid, Fluka, Switzerland) for vesicles approximately 200 nm in size and n-octyl tetraoxyethylene monoether (C8E4, Bachem Biochemica, Heidelberg, Germany) for liposomes up to 1000 nm in size. In brief, a thin film of a mixture of EPC, 2 mol % DOGM and detergent was prepared under reduced pressure. The lipid/detergent ratio was 0.2 mol/mol. The film was suspended in 1 or 2 mL of an aqueous solution of 1 M NIPAM and 0.025 M TEGDM depending on the size of the dialysis chamber. The final lipid concentration was 20 mM. The mixed micelle solutions were dialyzed for 12 h against demineralized water (1000 mL of water/mL of mixed micelles) at room temperature using a commercially available dialysis chamber (MiniLipoprep, Diachema AG, Switzerland) and a highly permeable cellulose dialysis membrane (MW cutoff 10 000 Da, Diachema AG). Water was replaced after 1, 3, and 5 h, respectively. Upon dialysis, vesicles were formed from mixed micelles after approximately 90 min as monitored by light scattering. The loss of NIPAM and TEGDM concentration in the vesicles was estimated by UV spectrometry to be less than 50%. After dialysis, the residual amount of detergent and nonencapsulated comonomer in the preparation was negligible.

Stauch et al. Table 1. Vesicle Sizes According to Different Preparation Techniquesa

d (nm) preparation method

measurement

ZetamasterS

cryo-TEM

detergent removal

before polym after polym pure lipid before polym after polym pure lipid

182 ( 2 188 ( 7 168 ( 3 233 ( 9 240 ( 12 199 ( 7

182 ( 13 184 ( 17 173 ( 15 229 ( 19 237 ( 21 207 ( 17

extrusion method

a Polymer-entrapping vesicles were prepared by detergent removal using octylglucosid or extrusion (200 nm pore). Polymerization was made after 3 h of incubation with initiator. Size measurements before and after polymerization by dynamic light scattering (ZetamasterS) at 20 °C and cryo-TEM, mean ( standard deviation. Little change in size was observed when measured before or after polymerization. However, for pure lipid vesicles (98 mol % EPC, 2 mol % DOGM) particle size was somewhat smaller.

Method of Photopolymerization. The monomer-entrapping vesicle dispersion was transferred into a 5 mL roundbottom quartz polymerization tube (wall thickness 2.3 mm, Thoma, Freiburg, Germany) and mixed with an aliquot of 2,2-diethoxyacetophenone (DEAP) to give a final initiator concentration of 10-5 M. The samples were degassed at 200 mbar for 2 min and flushed with argon. Degassing and flushing was repeated five times. UV irradiation was performed for 3 h using two parallel low-pressure mercury vapor lamps (Sylvania G15T8, Osram Sylvania, Danvers, MA) with an UV output of 3.6 W each and maximum emission at 253.7 nm constructed by Bresemann+Schorpp, Heiligenstadt, Germany. The distance between the lamp and the sample was 5 cm; the solution was stirred magnetically to ensure a homogeneous irradiation. The maximum sample temperature during the polymerization process was 30 °C. Evaluation of Initiator Penetration Rate. Four polymerization tubes were filled with 1 mL of a dispersion of monomer-entrapping large unilamellar vesicles obtained by detergent removal of C8E4. The samples were degassed, flushed, and mixed with an aliquot of DEAP as described. Three of the tubes were kept in darkness at 4 °C for initiator incubation time of 1, 3, and 12 h, respectively, before the UV irradiation was started. The fourth tube was irradiated immediately after adding the initiator. The photopolymerization was performed as described. The vesicles were analyzed by negative staining TEM and AFM as described below to study the presence or absence and the architecture of an inner polymer network. Determination of Lipid Content, Vesicle Size, and Lamellarity of the Samples. The final phospholipid concentrations of all samples, determined by their phosphorus content, were 18-20 mM.27 The average liposome size before and after the polymerization was estimated by dynamic light scattering (Malvern ZetamasterS, Malvern Instruments Ltd, UK) and cryo-TEM (see Table 1). Due to the high polydispersity of the liposomes prepared by detergent removal of C8E4, the average diameter of these samples was not determined. The lamellarity of vesicles obtained by detergent removal was 1.2 for octylglucosid and 1.5 for C8E4, respectively, as determined by TEM.28 After polymerization, these values were slightly increased.

Mimicking a Cytoskeleton inside Liposomes

Cryo-TEM. Cryo-TEM was performed on a Zeiss CEM 902 with cryo-stage (Zeiss, Oberkochen, Germany) as previously described,29 applying an acceleration voltage of 80 keV. Grids for specimen preparation were prepared according to Fukami and Adachi (1965).30 In short, copper grids (200 mesh, Science Services, Munich, Germany) were coated with a holey film prepared from Triafol-BN-foil (Merck KgaA, Darmstadt, Germany) and subsequently vapor deposited with carbon. Finally, the Triafol film was removed by washing the grid with ethyl acetate to obtain a pure holey carbon film (hole size 2-12 µm). At ambient conditions, a drop of the vesicle dispersion was deposited on the grid and the liquid was removed with blotting paper. The specimens were instantly shock-frozen by plunging them into liquid ethane and cooled to 90 K by liquid nitrogen. After freezing, the remaining ethane was removed using blotting paper. Finally, the specimens were transferred to the microscope using a cryo-transfer holder (Zeiss, Oberkochen, Germany). Examinations were carried out at a constant temperature of 90 K. Zero-loss filtered images were taken under low-dose conditions, i.e., using the minimal dose focusing device. Transmission Electron Microscopy (TEM). For TEM, the dispersion of EPC vesicles with an inner polymer network was diluted to 5 mM lipid. One drop was deposited onto a carbon-coated grid previously rendered hydrophilic by glow discharge in air/vacuum.31 TEM observations were performed using a LEO 912 Omega microscope and applying an acceleration voltage of 120 keV. Negative staining was performed with 2% AM, pH 7.0 in distilled water.32 Positive staining of polymer entrapping liposomes was obtained by treatment with RuO4 vapor prepared by oxidation of ruthenium trichloride with sodium hypochlorite.33 The specimens were stained by vapor phase reaction only for a relatively short time of 10-12 min. The reaction was terminated by adding a 10% aqueous solution of ascorbic acid. Finally, the specimens were dried again for 12 h at 30 °C under reduced pressure (600 mbar). Ultrathin Sectioning. For this purpose, the liposomes (lipid concentration 20 mM) were embedded in a watersoluble melamine resin (nanoplast, Bachhuber, Ulm, Germany). Sectioning (approximately 50 nm) was performed with an ultramicrotome (Ultracut E, Reichert and Jung) at ambient conditions. Atomic Force Microscopy (AFM). Atomic force microscopy experiments were performed with a Nanoscope III scanning probe microscope (Digital Instruments, Inc., USA). Height and phase images were obtained simultaneously, while operating the instrument in the tapping mode under ambient conditions. For tapping mode/height mode analysis, the sample was diluted to lipid content of 5 mM. One drop was deposited onto a commercial mica carrier and vacuum dried (20 °C, 200 mbar, 30 min). Tapping mode/phase mode measurements were applied to thin sections of the embedded sample. Preparation of Radiolabeled Vesicles. Three different kinds of radiolabeled vesicles were prepared: (a) vesicles with membrane-anchored polymer, (b) vesicles entrapping a polymer but not grafted to the membrane, and (c) vesicles without polymer but containing the reactive anchor monomer


in the membrane. Vesicles were obtained by an extrusion method. In brief, a thin film of a mixture of 98 mol % EPC, 2 mol % DOGM, and 37 kBq L-R-dipalmitoyl-[dipalmitoyl1-14C] (14C-DPPC, Du Pont, Dreieich, Germany) was prepared by drying the THF solution under reduced pressure, but for sample b, no DOGM anchor was added. The film was suspended with 2 mL of the same aqueous monomer solution as already described, and sample c was dispersed with 2 mL of isoosmotic aqueous solution of NaCl instead. To increase the trapping efficiency for the soluble monomers, the suspension was frozen in liquid nitrogen and thawed in a water bath of 37 °C; freezing and thawing were repeated five times.34 Oligolamellar vesicles were generated by 21fold extrusion through a polycarbonate membrane (pore size 200 nm, Corning Costar, Bodenheim, Germany) using a LiposoFast device (Avestin, Ottawa, Canada). Twenty one extrusion steps were performed to ensure complete sizing of lipid suspension (an odd number to avoid mixing of the sized liposomes with remaining initial multilamellar vesicles). The removal of the nonencapsulated comonomer was achieved by gel permeation chromatography (GPC, Sepharose 4B-CL, Pharmacia Fine Chemicals AB, Sweden) with demineralized water as eluent. Photopolymerization of samples a and b was as described; sample c was not exposed to UV and no initiator was added. The lipid concentration of the samples was determined to be 100 mM. Determination of 14C-DPPC Release after Addition of Sodium Cholate. Membrane solubilization of vesicles in the presence of various concentration of sodium cholate was evaluated by an ultracentrifugation method.35 In brief, mixtures of 1 mL of liposomes (lipid concentration 500 µM) and various cholate concentrations were ultracentrifuged in thick-wall 4-mL polycarbonate tubes for 210 min at 140000g, 20 °C, using a Beckman LE 80 ultracentrifuge and a fixed angle rotor 50.4 (Beckman, Munich, Germany). The release of 14C-DPPC as a membrane marker was determined in the supernatants with a β-counter (Tri-Carb, Canberra Packard GmbH, Germany). Results and Discussion The TEM studies were focused on the existence of vesicle shape anomalies which were expected to occur as a result of the UV-light-induced polymer growth and the interaction of an inner three-dimensional polymer network grafted to the vesicle membrane. Further, the architecture of the entrapped polymer networks was analyzed by the combination of positive staining TEM and AFM applied to ultrathin sections of the liposomes. Finally, the investigation of a stabilizing effect of this polymer skeleton against membrane solubilization by surfactant was performed by radiolabeling the bilayer. Linear NIPAM homopolymers and copolymers are well-known for their sudden collapse above a LCST in aqueous solution.14,17,24,25 When grafted onto liposomal membranes from the outside, thermal contraction of the polymer causes membrane defects, leading to a release of entrapped materials.15-17 In our system, however, the vesicles are filled with a cross-linked NIPAM-TEGDM copolymer which is attached to the inner membrane by numerous anchor

328


Figure 1. Microcalorimetric scans of EPC/DOGM liposomes dispersed in demineralized water: (a) 24 mol % DOGM; (b) 12 mol %; (c) 6 mol %; (d) 3 mol % DOGM. Scans were performed at a temperature range of 10-100 °C; the diagram focuses only on the part with a signal (25-70 °C).

molecules. To avoid possible deformation of the vesicle shape due to the LCST effect, preparation and analysis were performed at temperatures below the phase transition of NIPAM-TEGDM copolymer of 33 °C as determined by its cloud point. Differential Scanning Calorimetry (DSC). To determine the miscibility of the anchor molecule DOGM in the EPC membrane, microcalorimetric scans were performed to investigate phase separation in mixtures of different EPC/ DOGM ratios (Figure 1). As the phase transition temperature (Tc) from EPC liposomes (Tc from -15 to -7 °C)36 is outside of the measuring range, the main signal at 54 °C is assigned to an endothermic phase transition of DOGM. The intensity decreased with the reduction of DOGM (Figure 1a-c) and finally disappeared at a content of 3 mol % addition in the lipid membrane (Figure 1d). Therefore, all liposomes with polymer skeleton used in this study were prepared with an uncritical addition of 2 mol % DOGM to ensure a homogeneous distribution of the anchor monomer in the EPC membrane. Determination of Vesicle Size. The vesicle size determined by dynamic light scattering and cryo-TEM was not significantly altered after photopolymerization; however, compared to pure EPC/DOGM liposomes, the diameter was approximately 10-20% larger (Table 1). This might be a result of the difference in the osmotic pressure between the monomer solution inside the vesicle and the surrounding water phase. During the vesicle formation by detergent dialysis or extrusion, respectively, a water influx affects the vesicle swelling. However, upon polymerization, the osmotic pressure inside the vesicle is strongly reduced. Therefore, almost no difference in the pressure inside and outside the vesicle remains and the vesicle size did not markedly change within a time scale of days. An alteration in size was not estimated for the vesicles prepared by detergent removal of C8E4 due to little homogeneity in size. For cryo-TEM, a representative population of approximately 100 vesicles was analyzed. As the nonencapsulated monomer fraction was

Stauch et al.

separated from the monomer-entrapping liposomes before photopolymerization either by GPC or dialysis, the vesicles were not linked together. TEM Studies of Vesicle Shape and Entrapping Efficiency of Polymer Gel. Upon photopolymerization, changes in the spherical shape of the liposomes were observed only occasionally. More than 90% of the vesicles could not be distinguished from pure EPC liposomes in cryo-TEM as they were perfectly spherical and predominantly unilamellar (Figure 2). In only about 10% of all vesicles investigated in this study did shape anomalies occur, which were never observed in pure lipid preparations. As a result of the UVlight-induced polymer growth and the polymer-membrane interaction, budding processes occurred in the interior of vesicles. The buds showed spherical (Figure 3a,b), wormlike (Figure 3d), or cavity-exhibiting structures (Figure 3c). Although cryo-TEM provides the potential to view different morphologies of polymer entrapping liposomes “in situ” without the artifacts caused by drying and staining, polymers display little scattering and thus show little contrast in the TEM. Hence, for the visualization of the composite structure of polymer-containing lipid vesicles and for the determination of the trapping efficiency, a negative staining technique (aqueous solution of ammonium molybdate) and a positive staining by treatment with vapor of ruthenium tetroxide were employed. As a result, the polymer sphere can be clearly distinguished from the lipid envelope (Figure 4). The outmost shell represents the lipid bilayer; the internal sphere shows the polymerized NIPAM as a sphere of homogeneous density. The formation of spherical and wormlike-shaped buds was explained by the hypothesis that the vesicles were of bi- or oligolamellar origin (i.e., vesicles with two or more bilayers). We suggested a homogeneous distribution of monomer and initiator molecules in the aqueous phase inside a bilamellar vesicle (Scheme 2a). Upon UV irradiation individual polymer particles (Scheme 2bI) as well as coherent macromolecules or polymer aggregates (Scheme 2bII), respectively, may be formed. If the polymerization occurred predominantly in the interlamellar compartment between the first and the second membrane, the entrapped liposome was subject to dramatic shape changes. As a result of polymer growth and space requirement, unfavorable entropy changes may result in the breakdown of the original liposome into numerous smaller vesicles (Scheme 2bI and Scheme 2cI). On the other hand, the bizarre deformation of the vesicle can be stabilized by the entrapped polymer which is fixed to the membrane via numerous anchoring molecules (Scheme 2bII and Scheme 2cII) and thereby hindering the breakdown of the original bilayer. Atomic Force Microscopy (AFM). The AFM analysis (Figure 5) was performed on EPC/DOGM liposomes prepared by detergent removal of C8E4 after UV polymerization of encapsulated monomers (1 M NIPAM, 0.025 M TEGDM, 3 h incubation time of DEAP). The height profile and the section analysis as a result of a tapping mode/height mode study showed that the vesicle interior was predominantly filled with a polymer solid sphere (e.g., see Figure 5a). The rare budding processes were investigated by tapping mode/

Mimicking a Cytoskeleton inside Liposomes


Figure 2. The same vesicle preparation observed by cryo-TEM (left) and negative staining TEM (right). Vesicles were obtained by a detergent removal technique using octylglucosid. The bar represents 200 nm. In cryo-TEM (left), budding was detected with about 10% of the vesicles (arrows). The remaining vesicles were predominantly unilamellar and spherical. In negative staining TEM (right), polymer particles could be detected in nearly all vesicles (arrows).

Figure 3. Cryo-TEM micrographs of the main budding phenomena: vesicle preparation (a-c) by detergent removal of octylglucosid and (d) by C8E4. The bar represents 100 nm. Such phenomena occurred approximately with the same frequency for all preparations. Scheme 2. Suggestion for a Possible Mechanism of Budding Phenomena and Shape Deformation upon Polymerization in Liposomes of Two Bilayersa

Figure 4. TEM micrographs of polymer-containing vesicles. The bar represents 100 nm. (a) Unilamellar liposome prepared by detergent removal of octylglucosid (negative stain by 2% AM). The outmost shell represents the lipid bilayer; the internal sphere shows the polymerized NIPAM as a sphere of homogeneous density. The gap between sphere and shell is an artifact by drying when preparing the negative staining. For comparison, panel b shows the result of positive staining of a polymer-containing oligolamellar vesicle (positive stain by RuO4 vapor). The oligolamellar structure resulted from the preparation by extrusion.

phase mode analysis of the ultrathin sections (Figure 5b). The AFM micrograph shows buds of higher hardness compared to the lipid lamellae. These might be pure polymer particles or small polymer-entrapping vesicles as shown in Scheme 2 (sketch bI and cI) in correspondence to the cryoTEM observations. Analysis of the Architecture of Entrapped Polymer Networks. For the evaluation of initiator penetration rate, liposomes prepared by UV irradiation after various incubation times (1, 3, and 12 h, respectively) of the radical former DEAP were analyzed for the presence of encapsulated polymer gel. The vesicles could not be distinguished either in cryo-TEM or in negative-staining TEM, as polymer particles were clearly detectable in vesicles of all preparations even when the polymerization was initiated by UV immediately after degassing, flushing, and mixing with DEAP.

a (b , c ) As a result of the UV-light-induced polymer growth and the I I polymer-membrane interaction, the enclosed vesicle may break down into numerous smaller vesicles. (bII, cII) Deformations can be stabilized by the entrapped polymer hindering the breakdown of the membrane.

This indicated that DEAP was able to penetrate through the lipid membrane in a time scale of only a few seconds. If no initiator was added to the monomer-entrapping liposomes, no polymer growth was observed after UV irradiation. To find out about the architecture of the polymer gels inside these vesicles, ultrathin sectioning was applied to the embedded samples. In AFM and positive-staining TEM investigations of the sections, two main types of polymer network were found: A more or less complete filling of the vesicles inside with the polymer (solid sphere) and a pNIPAM network attached as a thin layer underneath the inner surface of the lecithin bilayer (hollow sphere), respectively. The latter occurred exclusively in vesicles prepared

330


Stauch et al.

Figure 5. AFM analysis of polymer entrapping liposomes prepared by detergent removal of C8E4 and UV irradiation after 3 h of incubation with the initiator DEAP: (a) tapping mode/height mode analysis of a single polymer solid sphere, height profile, 1 (0 nm), 2 (3 nm), 3 (13 nm). (b) The density distribution becomes apparent in tapping mode/phase mode analysis applied to an ultrathin section of a vesicle containing encapsulated polymer buds. The bar represents 100 nm.

Figure 6. Ultrathin sections of liposomes obtained by detergent removal of C8E4. Photopolymerization of the entrapped monomers was performed immediately after mixing with the initiator aliquot to influence the density of the inner polymer network; see text. (a) The outmost shell (bright contrast) represents the lipid bilayer; the dark thin layer underneath the membrane shows the quasi-two-dimensional polymer network. Preparation for TEM by positive staining with RuO4 vapor, the bar represents 200 nm. (b) AFM tapping mode/phase mode analysis of a unilamellar vesicle with a dense polymer shell underneath the membrane. Thickness of the polymer layer was approximately 50 nm.

by the immediate UV irradiation after mixing with the initiator aliquot and was found to be predominant and reproducible. Positive-staining TEM micrographs of liposome sections with polymer shells attached to the membrane are shown in Figure 6 (dark contrast of the polymer due to ruthenium tetroxide stain). By imaging these sections with AFM phase mode analysis (Figure 6c), the existence of the polymer hollow sphere was proved and, therefore, staining artifacts were excluded. According to the results above, we suggest that the polymerization process in the interior of phospholipid vesicles can be triggered by the presence of a radical former (see Scheme 3). If the penetration rate of the radical initiator through the vesicle membrane is known, the formation of polymer chains and hence the density of the polymer network can be triggered through its incubation time. Determination of Membrane Stabilization. One crucial difference between pure phospholipid vesicles and biological cells, e.g., the red blood cell, is the stability against membrane solubilization by surfactants. For erythrocytes, for example, the critical concentrations for bile acid induced solubilization

are about 10-14 mM sodium cholate,37 whereas solubilization of pure EPC membranes starts at approximately 4 mM sodium cholate.38 By creating a cross-linked network inside the liposomes, its stability in a solution of sodium cholate can only be improved significantly if the polymer is linked to the inner monolayer by the anchoring group DOGM (Figure 7 “membrane-anchored polymer”). In this case, the bile salt induced membrane dissolution starts at 8 mM sodium cholate. Compared with “pure EPC/DOGM vesicles” (formation of mixed micelles at 4.5 mM), stability was doubled. If the polymer inside the liposome was not fixed to the membrane (“polymer not anchored”), no increase in the stability against dissolution was observed (see Figure 7). Conclusion We report the preparation of phospholipid vesicles entrapping a polymer network which is attached to the inner monolayer by reactive membrane anchor molecules. Upon UV irradiation polymer-entrapping vesicles were predominantly spherical and unilamellar. Budding anomalies as a

Mimicking a Cytoskeleton inside Liposomes Scheme 3. The Influence of Polymer Architecture by the Membrane-Permeation Rate of the Radical Formera

a (a) Homogeneous distribution of radical former in the vesicle (f polymer solid sphere). (b) Radical former after immediate penetration through the liposomal membrane (f polymer hollow sphere).

Figure 7. Determination of membrane stabilization against sodium cholate by the release of 14C DPPC (fit sigmoide) at 20 °C. Liposomes with “membrane-anchored polymer” showed a significantly higher stability against bile-salt-induced dissolution, in contrast to pure EPC/ DOGM vesicles or polymer-entrapping vesicles without membraneanchoring groups (averaged results of four determinations, standard deviation was below 5%).

result of the UV-light-induced polymer growth in bilamellar vesicles occurred in only about 10% in each preparation analyzed for this study. Vesicles with polymer-tethered membranes revealed a significantly higher stability against solubilization by sodium cholate compared to pure lecithin vesicles and to vesicles entrapping a polymer network which was not coupled to the membrane. The permeation barrier of liposomal membranes for the penetration of the radical former provides a possible trigger mechanism for the formation of polymer chains by photopolymerization. As a result, the polymer density and hence the architecture of the polymer network inside the liposomal compartment can be a hollow sphere or a solid sphere, respectively, as determined by TEM and AFM analysis of ultrathin sections.


However, additional characterization of ensembles of polymer-entrapping liposomes by static and dynamic light scattering and small-angle neutron scattering are currently under investigation, also considering possible effects on the vesicle’s shape due to the sudden phase separation of the grafted NIPAM-TEGDM copolymer above the LCST. Finally, we believe that the use of our system as a model for biological cell membranes provides new possibilities in fundamental biophysical research, e.g., the study of collective motions and viscoelastic properties in membranes by transverse nuclear spin relaxation measurements,39 as well as the determination of bending elastic constants of membranes interacting with the cytoskeleton.28,40 Acknowledgment. Financial support was provided by the Deutsche Forschungsgemeinschaft, Sonderforschungsbereich 428, Teilprojekt D4. The phospholipids were a generous gift from Lipoid GmbH, Ludwigshafen, Germany. We thank Professor Walther Burchard from the Institut fu¨r Makromolekulare Chemie Freiburg, Germany, for many helpful discussions. DOGM synthesis by D. Gutmayer is gratefully acknowledged. References and Notes (1) Yeagle, P. The Membranes of Cells; Academic Press: Orlando, FL, 1987. (2) Lux, S.; Shohet, S. B. Hospital Practice 1984, NoV., 77. (3) Shen, B. W. Ultrastructure and Function of Membrane Skeleton, In Red Blood Cell Membrane; Agre, P., Ed.; Marcel Dekker: New York, 1989; Chapter 10. (4) Nash, G. B.; Gratzer, W. B. Biorheology 1993, 30, 397-407. (5) Janmey, P. Cell Membranes and the Cytoskeleton. In Handbook of Biological Physics; Lipowsky, R., Sackmann, E., Eds.; Elsevier Science B.V., 1995; Vol. 1A. (6) Sackmann, E. Biological Membranes. Architecture and Function. In Handbook of Biological Physics; Lipowsky, R., Sackmann, E., Eds.; Elsevier Science B.V., 1995; Vol. 1. (7) Sackmann, E. Physical Basis of Self-Organization and Function of Membranes: Physics of Vesicles. In Handbook of Biological Physics; Lipowsky, R., Sackmann, E., Eds.; Elsevier Science B.V., 1995; Vol. 1A. (8) Regen, S. L.; Czech, B.; Singh, A. J. Am. Chem. Soc. 1980, 102, 6638-40 (9) Hub, H.-H.; Hupfer, B.; Koch, H.; Ringsdorf, H. Angew. Chem., Int. Ed. 1980, 19 (11), 938-940. (10) Fendler, J. Science 1984, 223, 888. (11) Krause, J.; Juliano, R. L.; Regen, S. J. Pharm. Sci. 1987, 76 (1), 1-5. (12) Okada, J.; Cohen, S.; Langer, R. Pharm. Res. 1995, 12 (4), 576582. (13) Chen, H.; Torchilin, V.; Langer, R. J. Controlled Release 1996, 42, 263-272. (14) Ringsdorf, H.; Sackmann, E.; Simon, J.; Winnik, F. M. Biochim. Biophys. Acta 1993, 1153, 335-344. (15) Kim, J. C.; Bae, S. K.; Kim, J. D. J. Biochem. 1997, 121, 15-19. (16) Hayashi, H.; Kono, K.; Takagishi, T. Bioconjugate Chem. 1998, 9, 382-389. (17) Kono, K.; Nakai, R.; Morimoto, K.; Takagishi, T. Biochim. Biophys. Acta 1999, 1416, 239-250. (18) Meyer, O.; Papahadjopoulos, D.; Leroux, J.-C. FEBS Lett. 1998, 421, 61-64. (19) Yamazaki, Winnik, F. M., Cornelius, Brash, Biochim. Biophys. Acta 1999, 1421, 103-115. (20) Hotz, J.; Meier, W. Langmuir 1998, 14, 1031-1036. (21) Ringsdorf, H.; Schlarb, B.; Tyminski, P. N.; O’Brian, D. F. Macromolecules 1988, 21, 671-677. (22) Torchilin, V. P.; Klibanov, A. L.; Ivanov, N. N.; Ringsdorf, H.; Schlarb, B. Makromol. Chem., Rapid Commun. 1987, 8, 457-460. (23) Meier, W. Chem. Soc. ReV. 2000, 29, 295-303. (24) Schild, H. G.; Tirrell, D. A. J. Phys. Chem. 1990, 94, 4352-4356. (25) Pan, Y. V.; Wesley, R. A.; Luginbuhl, R.; Denton, D. D.; Ratner, B. D. Biomacromolecules 2001, 2, 32-36.

332


(26) Wu, X.; Pelton, R. H.; Hamielec, A. E.; Woods, D. R.; McPhee, W. Colloid Polym. Sci. 1994, 272, 467-477. (27) Bartlett, G. R. J. Biol. Chem. 1959, 234, 466-468. (28) Althoff, G.; Stauch, O.; Vilfan, M.; Frezzato, D.; Moro, G. J.; Hauser, P.; Schubert, R.; Kothe, G. Submitted for publication in Phys. Chem. B. (29) Schmidtgen, M. C.; Drechsler, M.; Lasch, J.; Schubert, R. J. Microsc. 1998, 191, 177-186. (30) Fukami, A.; Adachi, K. J. Electron Microsc. 1965, 14, 112-118. (31) Dubochet, J.; Groom, M.; Mu¨ller-Neuteboom, S. In AdVances in Optical and Electron Microscopy; Cosslett, V. E., Barer, R., Eds.; Academic Press: London, 1982. (32) Bugelski, P. J.; Sowinski, J. M.; Kirsh, R. L. In Liposomes a practical approach; New, R. R. C., Ed.; Oxford University Press: New York, 1990; Chapter 4.1. (33) Sawyer, L. C.; Grubb, D. T. Polymer Microscopy; Chapman and Hall: London, 1987.

Stauch et al. (34) Mayer, L. D.; Hope, M. J.; Cullis, P. R.; Janoff, A. S. Biochim. Biophys. Acta 1985, 817, 193-196. (35) Schubert, R.; Schmidt, K.-H. Biochemistry 1988, 27, 8787. (36) New, R. R. C. Liposomes a practical approach; New, R. R. C., Ed.; Oxford University Press: New York, 1990. (37) Van der Meer, R.; Termont, D. S. M. L.; de Vries, H. J. Am. Phys. Soc. 1991, G142. (38) Schubert, R.; Beyer, K.; Wolburg, H.; Schmidt, K.-H. Biochemistry 1986, 25, 5263. (39) Althoff, G.; Heaton, N. J.; Gro¨bner, G.; Prosser, R. S.; Kothe, G. Colloids Surf., A 1996, 115, 31-37. (40) Althoff, G.; Frezzato, D.; Vilfan, M.; Stauch, O.; Schubert, R.; Vilfan, I.; Moro, G. J.; Kothe, G. Submitted for publication in Phys. Chem. B.

BM015613Y

to the Inner Leaflet of Liposomal Membranes - ACS Publications

Recommend Documents