Fluorinated Vesicles Allow Intrabilayer Polymerization of a

Universite´ de Nice, Parc Valrose, 06108 Nice, Cedex 2, France; Institut de Ge´ne´tique et de. Biologie Mole´culaire et Cellulaire, CNRS/INSERM/UL...
0 downloads 0 Views 216KB Size
2872

Langmuir 2001, 17, 2872-2877

Fluorinated Vesicles Allow Intrabilayer Polymerization of a Hydrophobic Monomer, Yielding Polymerized Microcapsules Marie Pierre Krafft,*,† Laurent Schieldknecht,† Pascal Marie,† Franc¸ oise Giulieri,‡ Marc Schmutz,§,# Nathalie Poulain,⊥ and Evelyne Nakache⊥ Chimie des Syste` mes Associatifs, Institut Charles Sadron (UPR CNRS 22), 6, rue Boussingault, 67 083 Strasbourg, Cedex, France; Unite´ de Chimie Mole´ culaire, Universite´ de Nice, Parc Valrose, 06108 Nice, Cedex 2, France; Institut de Ge´ ne´ tique et de Biologie Mole´ culaire et Cellulaire, CNRS/INSERM/ULP, BP 163, 67404 Illkirch Cedex, C.U. Strasbourg, France; and ISMRA, Universite´ de Caen, 6, Bd Mare´ chal Juin, 14050 Caen, Cedex, France Received November 29, 2000. In Final Form: January 29, 2001 Intrabilayer polymerization of hydrophobic monomers has been attempted as a way to strengthen the structure of vesicles and producing polymer microcapsules. However, no clear evidence has been provided that demonstrates the formation of polymerized vesicles. On the contrary, it has recently been shown that polymerization of hydrophobic monomers within a vesicular bilayer did not yield the expected capsules but led to the formation of hybrid surfactant-polymer particles constituted by a polymer latex lump attached to the vesicle. We now report that use of highly ordered, microcompartmentalized fluorinated vesicles, i.e., made from fluorinated lipids, allows to achieve true intrabilayer polymerization. We have studied the thermally induced free radical polymerization of isodecyl acrylate (ISODAC) in small unilamellar vesicles (SUVs) made from a perfluoroalkylated phosphatidylcholine (F-PC) and compared it to polymerization of ISODAC in vesicles made from standard egg phospholipids (EggPC). Cryogenic transmission electron microscopy (cryo-TEM) confirmed that extended polymer/bilayer phase separation occurred in the EggPC vesicles. On the other hand, no evidence of phase separation was observed in the case of F-PC vesicles. The polymer, poly(isodecyl acrylate) (poly(ISODAC)), was homogeneously distributed within the bilayer. In addition, the rate of polymerization in F-PC vesicles, as monitored by 1H NMR, was higher than in EggPC vesicles. The molecular weight of poly(ISODAC), as determined by size exclusion chromatography (SEC), was smaller when obtained in F-PC than in EggPC vesicles. The internal fluorinated core present in F-PC vesicles significantly reduces the space available for polymerization, the monomer being excluded from the central core and confined in the two tight, nonexpandable lipophilic regions of the vesicles. Such conditions of confinement, which likely result both in an increase in local monomer concentration and in probability of polymerization termination steps, may explain the observed higher reaction rate and lower polymer molecular weight.

Introduction Vesicles, i.e., closed spherical structures made from bilayers of amphiphilic molecules, have a wide scope of applications as microreservoirs in the biomedical field and material sciences. The potential of vesicles is strongly hampered, however, by their intrinsic frailty. Bilayer polymerization has been proposed as a means of immobilizing and reinforcing vesicle structure, thus obtaining hollow polymer capsules. Vesicle-forming polymerizable amphiphiles have been widely investigated with this purpose.1-3 An alternative approach to producing polymer capsules is to achieve polymerization of hydrophobic monomers solubilized within the vesicle bilayer. †

Institut Charles Sadron. Universite´ de Nice. § Institut de Ge ´ ne´tique et de Biologie Mole´culaire et Cellulaire. ⊥ Universite ´ de Caen. # Present address: Institut Charles Sadron (UPR CNRS 22), 6, rue Boussingault, 67 083 Strasbourg, Cedex, France. * Corresponding author: Tel (33) 3 88 41 40 60; Fax (33) 3 88 40 41 99; E-mail [email protected]; http://www-ics.u-strasbg.fr/. ‡

(1) Fendler, J. H. In Surfactants in Solution; Mittal, K. L., Lindman, B., Eds.; Plenum Press: New York, 1984; p 1947. (2) Ringsdorf, H.; Schlarb, B.; Venzmer, J. Angew. Chem., Int. Ed. Engl. 1988, 27, 113. (3) O’Brien, D. F.; Ramaswami, V. In Encyclopedia of Polymer Science and Engineering; Mark, H., Bikales, N., Overberger, C., Eds.; John Wiley & Sons: New York, 1989; Vol. 17, p 108.

Since the pioneering work of Murtagh and Thomas,4 researchers have examined the polymerization of such hydrophobic monomers.5-8 The most investigated monomers included styrene, divinylbenzene, 1-methacryloyloxybutane, 1,2-bis(methacryloyloxyethane), and isodecyl acrylate. Mostly used amphiphiles were dioctadecylammonium bromide or chloride, mixtures of dodecyltrimethylammonium bromide and sodium dodecylbenzenesulfonate (which form spontaneous vesicles), and sodium ethylhexyl phosphate. Polymerization was thermally or photochemically induced. The initiators tested were either hydrophilic (potassium persulfate (K2S2O8), 2,2′-azobis(2-methylpropionamidine) dihydrochloride (V-50), cumene hydroperoxide/iron(II) (CHP/Fe II)) or lipophilic (azobis(isobutyronitrile) (AIBN), 2,2′-dimethoxy-2-phenylacetophenone (DMPA)). In all these studies, it was assumed, without direct proof, however, that the polymer chains remained confined within the bilayer while growing and that a hollow polymer capsule was eventually formed. On (4) Murtagh, J.; Thomas, J. K. Faraday Discuss. Chem. Soc. 1986, 81, 127. (5) Kurja, J.; Nolte, R. J. M.; Maxwell, I. A.; German, A. L. Polymer 1993, 34, 2045. (6) Poulain, N.; Nakache, E.; Pina, A.; Levesque, G. J. Polym. Sci., Part A: Polym. Chem. 1996, 34, 729. (7) Morgan, J. D.; Johnson, C. A.; Kaler, E. W. Langmuir 1997, 13, 6447. (8) Hotz, J.; Meier, W. Langmuir 1998, 14, 1031.

10.1021/la001658t CCC: $20.00 © 2001 American Chemical Society Published on Web 04/07/2001

Polymerization of a Hydrophobic Monomer

the other hand, it was recently shown that, contrary to the conclusions of previous investigations, polymerization did not produce the expected hollow capsules.9 What actually was occurring was a complete phase separation between the polymer chain and the bilayer membrane. The resulting hybrid polymer-surfactant particle presented a peculiar “parachute”-like morphology with a polymer latex bead attached to the vesicle. One of our general goals is to assess the influence of fluorinated amphiphiles on the structure and properties of colloidal systems.10 For example, short single-chain surfactants derived from phosphocholine were shown to form very stable vesicles, while their hydrogenated analogues only form micelles.11 Nonchiral single-chain surfactants derived from dimorpholinophosphate formed microtubules, although chirality is generally considered as indispensable in the case of hydrogenated tubuleforming amphiphiles.12 The presence of semifluorinated alkanes (FnHm diblocks) in the bilayer of SUVs made from dipalmitoylphosphatidylserine strongly modified the kinetics of Ca2+-induced fusion in these vesicles.13 FnHm diblocks are also highly efficient stabilizers of oxygencarrying fluorocarbon emulsions (blood substitutes).14 Generally speaking, fluorinated amphiphiles provide a stronger driving force for self-organization than their hydrogenated counterparts, resulting in increased stability and other distinctive properties.10,15-18 Another of our objectives is to exploit the structural compartmentalization induced in colloidal systems by the unique combination of hydrophobicity and lipophobicity of fluorinated chains to generate new properties. For example, monolayers made from phospholipids and FnHm diblocks were shown to exhibit a reversible vertical phase separation: upon compression, the diblock is progressively expelled upward and forms a second organized layer on top of a phospholipid-only monolayer.19 Such up-and-down “lift” monolayers were utilized to control two-dimensional polymerization of 10,12-pentacosadiynoic acid.20 We hypothesized that use of vesicles made of fluorinated amphiphiles could change the outcome of the polymerization reaction because of greater ordering and because the lipophilic space in which polymerization is expected to take place is limited by covalently bound moieties. In addition, there is the possibility of creating a double shell made of two concentric capsules. We have therefore examined the thermally induced free radical polymerization of isodecyl acrylate (ISODAC) in SUVs made from a perfluoroalkylated phosphatidylcholine (F-PC) (Scheme (9) Jung, M.; Hubert, D. H. W.; Bomans, P. H. H.; Frederik, P. M.; Meuldijk, J.; van Herk, A. M.; Fischer, H.; German, A. L. Langmuir 1997, 13, 6877. (10) Krafft, M. P.; Riess, J. G. Biochimie 1998, 80, 489. (11) Krafft, M. P.; Giulieri, F.; Riess, J. G. Angew. Chem., Int. Ed. Engl. 1993, 32, 741. (12) Giulieri, F.; Krafft, M. P.; Riess, J. G. Angew. Chem., Int. Ed. Engl. 1994, 33, 1514. (13) Krafft, M. P.; Ferro, Y. Polym. Prepr. (Am. Chem. Soc., Div. Polym. Chem.) 1998, 39, 938. (14) Krafft, M. P.; Riess, J. G.; Weers, J. G. In Submicronic Emulsions in Drug Targeting and Delivery; Benita, S., Ed.; Harwood Academic Publ: Amsterdam, 1998; p 235. (15) Riess, J. G.; Fre´zard, F.; Greiner, J.; Krafft, M. P.; Santaella, C.; Vierling, P.; Zarif, L. In Liposomes-Non-Medical Applications; Barenholz, Y., Lasic, D., Eds.; CRC Press: Boca Raton, FL, 1996; Chapter 8, p 97 and references included. (16) Kunitake, T. Angew. Chem., Int. Ed. Engl. 1992, 31, 709. (17) Elbert, R.; Folda, T.; Ringsdorf, H. J. Am. Chem. Soc. 1984, 106, 7687. (18) Liang, K.; Hui, Y. J. Am. Chem. Soc. 1992, 114, 6588. (19) Krafft, M. P.; Giulieri, F.; Goldman, M.; Fontaine, P. Manuscript in preparation. (20) Wang, S.; Lunn, R.; Krafft, M. P.; Leblanc, R. M. Langmuir 2000, 16, 2882.

Langmuir, Vol. 17, No. 9, 2001 2873 Scheme 1. Chemical Structures of the Vesicle-Forming Perfluoroalkylated Phosphatidylcholine (F-PC, a) and Egg Phosphatidylcholine (EggPC, b); R ) Unsaturated Alkyl Chain

1a). Isodecyl acrylate was chosen as the monomer because it forms elastomeric homopolymers (Tg < 0 °C).6 In addition, isodecyl chains may increase interactions with the lipid matrix and thus minimize phase separation phenomena. For comparison, polymerization of ISODAC was also achieved in vesicles made from egg phosphatidylcholine (EggPC) (Scheme 1b). The morphology and average size of vesicles were studied by cryogenic transmission electron microscopy (cryo-TEM) and quasi-elastic light scattering (QELS). Experimental Section Materials. F-PC, synthesized according to ref 21, was a gift from Prof. J. G. Riess (MRI Institute, Medical Center, University of California at San Diego). EggPC came from Lipoid GmbH (Ludwigshaffen, Germany). ISODAC (99%, Aldrich) was distilled under reduced pressure before use to eliminate the polymerization inhibitor (hydroquinone). The oil-soluble thermal initiator azobis(isobutyronitrile) (AIBN, Aldrich) was used as received. Vesicles were prepared in purified Millipore water (Milli-Q system). Vesicle Preparation. A solution of lipids (F-PC or EggPC) in CHCl3 was evaporated in a rotavapor to obtain a thin film. This film was hydrated at room temperature with an appropriate amount of purified water and homogenized with a low-energy mixer (Ultra-Turrax, Ika-Labortecknik, Staufen, Germany). A lower lipid concentration (4.3 × 10-3 mol L-1) was employed for electron microscopy studies, as compared to that used for polymerization kinetics studies (5.4 × 10-2 mol L-1). Vesicles were obtained by submitting the resulting coarse dispersion to sonication (Branson B-30 sonifier, 15 mm titanium probe, 30 min, power 5, 50% pulse) in a Rosette cell thermostated at 25 °C. Polymerization Process. The dispersion of vesicles (F-PC or EggPC) was placed into a three-necked round-bottomed flask equipped with a magnetic stirrer, a condenser, and a septum to collect samples. AIBN was solubilized in ISODAC prior to addition ([ISODAC]:[AIBN] ) 50:1). The ISODAC/AIBN mixture was then added to the vesicle dispersion ([ISODAC]:[lipid] ) 1:1) through the septum with a syringe at 25 °C under air. The solution was stirred for 24 h at room temperature to allow diffusion of ISODAC into the vesicle bilayers. Purified nitrogen was bubbled into the solution for 45 min. The dispersion was then heated to 60 °C. Completion of the reaction took ca. 40 min for F-PC vesicles and 90 min for EggPC vesicles (see NMR-monitored kinetics studies below). Polymer Isolation and Purification. After polymerization, the dispersion of vesicles ([lipid] ) 5.4 × 10-2 mol L-1) was extracted three times with CHCl3. The combined extracts were dried overnight with MgSO4. After filtration, the volume of solution was reduced to 20 mL. Poly(ISODAC) was reprecipitated three times in EtOH (140 mL). It was then filtered and dried in a vacuum oven. The yield of polymerization was 98% in F-PC vesicles and 95% in EggPC vesicles. (21) Santaella, C.; Veirling, P.; Riess, J. G. New J. Chem. 1991, 15, 685.

2874

Langmuir, Vol. 17, No. 9, 2001

1H NMR Analysis. Spectra were obtained using a Bruker 200 MHz magnetic resonance spectrometer. Two milliliter size aliquots were sampled at regular intervals of time during the polymerization process. The analysis of the monomer to polymer conversion was based on the assay of residual monomer concentration. The samples were extracted as described above, evaporated, and dissolved in CDCl3. Conversion analysis was performed by relating the integrals of the vinylic protons (5.8, 6.2, and 6.4 ppm) to that of an internal reference composed of CHCl3 (1%) in CDCl3. Size Exclusion Chromatography (SEC). The absolute molecular weight distributions were determined by SEC coupled with MALLS (multiangle laser light scattering) and DRI detectors. A Shimadzu differential refractive index detector (model RID-6A) and a Dawn-DSP laser photometer operating at 632.8 nm from Wyatt Technology Corp. were used as the concentration detector and the molecular weight and size detector, respectively. A Shimadzu pump (model LC-6A), a 717 plus injector (Waters Associates), and five PL gel columns (three mixed B, 500 Å and 105 Å, Polymer Laboratories) fitted in series completed the equipment. The mobile phase was THF (analytical grade), and the flow rate was 1 mL min-1. Molecular weights were determined by using the dn/dc value ()0.065 mL g-1) which was determined from the surface area of the chromatogram peak. Quasi-Elastic Light Scattering (QELS). A Malvern Zeta Sizer 3000 HS was used for dynamic light scattering at a scattering angle of 90 °C. The temperature was 25 °C. The z-averaged hydrodynamic mean diameters (D h ) of vesicles were determined using the software provided by Malvern. Measurements were achieved on the 4.3 × 10-3 mol L-1 lipid-concentrated dispersions. Cryogenic Transmission Electron Microscopy (CryoTEM). A 400 mesh lacey carbon film copper grid was rendered hydrophilic by a mild glow discharge. A 5 µL aliquot of the vesicle solution was deposed and adsorbed for 2 min. Excess of sample solution was removed with a piece of filter paper (Whatmann 2 or 5) to leave a thin film. The grid was rapidly plunged into liquid nitrogen-cooled ethane and stored in liquid nitrogen until observation. The grid was transferred in a Gatan 626 cryo holder and observed in a CM12 Philips microscope equipped with an additional anticontamination device. Low-dose images were recorded on SO163 films and developed under standard conditions. Image Processing. The digitized image was analyzed with Visilog Image Analysis Software (Noesis, France) using filtering techniques (smoothing for removing noise and sharpening for enhancing low contrasted images), a morphological gradient method (the gradient image shows the edge of particles in the continuous phase), mathematical morphology functions22 (skeleton and thinning, edge detection method), and adaptive thresholding to locate the particles’ boundaries and interface contour and to extract their dimensions and bilayer thickness. Precision on measurements was (0.5 nm.

Results Characterization of the Vesicles Prior to Polymerization. Morphologies of F-PC and EggPC vesicles were determined using cryo-TEM, which allows direct observation of vesicles in their native state. Micrographs show that, in both cases, small unilamellar vesicles were formed (Figures 1a and 2a). A significant difference can be noted, however, in the aspect of the bilayers of the two types of vesicles. The bilayer of EggPC vesicles is seen as two concentric dark rings, as previously described.23 On the other hand, the F-PC vesicle bilayer appears as a single, thicker dark ring. In this case, the central fluorinated core constituted by the fluorinated tails strongly scatters electrons, forming a new zone of contrast in addition to that of polar heads. As a consequence, under optimum (22) Coster, M.; Chermant, J. L. Pre´ cis d’Analyse d’Images; Presses du CNRS: Paris, 1979. (23) Lepault, J.; Pattus, F.; Martin, N. Biochim. Biophys. Acta 1985, 820, 315.

Krafft et al.

Figure 1. Cryo-TEM micrographs of EggPC vesicles (a) before and (b) after polymerization. The classical bilayer can easily been seen in several vesicles. Arrowheads show (b) the lumps formed by poly(ISODAC), indicating polymer/bilayer phase separation (bar ) 50 nm).

Figure 2. Cryo-TEM micrographs of F-PC vesicles (a) before and (b) after polymerization. The bilayer is not observed due to the fluorinated core that scatters electrons in the same manner that the polar headgroups. No lumps are observed (b), indicating that no phase separation occurred during polymerization (bar ) 50 nm).

conditions, one could have expected the fluorinated bilayer to be visualized as three concentric rings. However, only one ring can be observed on our cryo-TEM picture. This is likely due to the fact that the narrow low-contrast region corresponding to the hydrogenated segment (only six carbons for FPC) is too small to be distinguishable. We hope that observation with a FEG instrument will allow the visualization of the fluorinated core and polar head zones. Vesicle profiles were studied by image processing. For EggPC vesicles, total bilayer thickness (db), polar head thickness (dph), and hydrocarbon chain thickness (dhc) were measured to be 6.0, 1.8, and 2.3 nm, respectively. These values are in qualitative agreement with db, dph, and dhc values reported for LR lamellar phases of egg phosphatidylcholine at maximum hydration (4.52, 0.9, and 2.72

Polymerization of a Hydrophobic Monomer

Langmuir, Vol. 17, No. 9, 2001 2875

Figure 3. Cryo-TEM micrograph of an half and half mixture of F-PC and EggPC vesicles. An EggPC bilayer is visualized as two concentric rings (arrowhead), while a F-PC bilayer is seen as a single, thicker dark ring (arrow) (bar ) 50 nm).

nm).24 The measured hydrocarbon chain thickness, dhc ) 2.3 nm, is significantly smaller than the maximal hydrocarbon chain thickness value (dmax) calculated by using the length of a fully extended hydrocarbon chain in an all-trans configuration (given by lc ) 1.5 + 1.265n, n ) number of carbons25), dmax ) ca. 4 nm, reflecting the fact that the hydrocarbon chains are in the fluid state. For F-PC vesicles, image processing only allowed measurement of total bilayer thickness, db ) 6.0 nm. This value is in good agreement with the maximal thickness (dmax ) 5.4 nm) of the fluorinated bilayer calculated by using the maximal length of the semifluorinated chain of F-PC (C6F13(CH2)6-), lc ) 1.8 nm, and assuming that dph ) 0.9 nm, according to ref 24. lc was calculated with the following equation: lc ) 2 + 1.265nH + 1.34nF (nH and nF: numbers of hydrogenated and fluorinated carbons).26 The fact that dmax is close to db, by contrast to what is observed in the case of EggPC vesicles, may suggest that, in the fluorinated vesicles’ bilayer, both the fluorinated termination and the hydrocarbon spacer are in an extended conformation. The fact that bilayer thickness was the same for EggPC and F-PC vesicles is purely fortuitous. Vesicle mean diameters were measured by QELS (Figure 4a,b). D h values were 31.9 nm (polydispersity 0.27) for EggPc vesicles and 41.3 nm (polydispersity 0.25) for F-PC vesicles. It was not possible to produce smaller vesicles by increasing sonication time or power. The diameter measured by QELS cannot be directly compared to that measured on cryo-TEM pictures because the latter method gives only a snapshot of the specimen, while QELS gives an average measurement of the bulk. Polymerization Process. Vesicles were incubated with ISODAC for 24 h. The lipophilic initiator AIBN was solubilized in the monomer prior to addition. At this step, it was verified that no polymer had been formed. After degassing the dispersion by bubbling nitrogen, polymerization was initiated by heating the dispersion at 60 °C. Reaction reached completion after ca. 40 min for F-PC and 90 min for EggPC (see above kinetics studies). Characterization of the Vesicles after Polymerization. Figures 1b and 2b show F-PC and EggPC vesicle morphologies after polymerization, as determined by cryoTEM. Many lumps can be seen in the bilayers of EggPC vesicles (Figure 1b). These lumps have moved apart the two lipid layers that form the bilayer without disrupting it. These micrographs are very similar to those reported for dioctadecylammonium bromide/styrene system.9 They strongly suggest that phase separation also occurred in the EggPC/ISODAC system. As reported by Jung et al., (24) Petrache, H. I.; Tristam-Nagle, S.; Nagle, J. F. Chem. Phys. Lipids 1998, 95, 83. (25) Tanford, C. In The Hydrophobic Effect; Duke, J. B., Ed.; Wiley: New York, 1980; p 42. (26) Giulieri, F.; Krafft, M. P. Colloids Surf. 1994, 84, 121.

Figure 4. Quasi-elastic light scattering (QELS) of (a) EggPC and (b) F-PC vesicles before (open circles) and after (solid circles) polymerization.

the polymer chains have formed latex beads that remained attached to the vesicles. By contrast, Figure 2b shows no evidence of lumps attached to F-PC vesicles, which suggests that in this case poly(ISODAC) is evenly distributed within the bilayer and that no phase separation has occurred. This difference is clearly illustrated in Figure 3, which shows a mixture of EggPC and F-PC vesicles after polymerization. The thickness of EggPC and F-PC vesicle bilayers was evaluated by image processing. The values for EggPC was found to be unchanged (db ) 6 nm, dph ) 1.8 nm, and dhc ) 2.3 nm), which is not surprising, taking into account that polymerization does not take place within the bilayer and has therefore no impact on its thickness. In the case of polymerized F-PC vesicles, the total bilayer thickness was found to be 5.5 nm, i.e., not significantly different from that measured before polymerization. This was a priori more surprising, since one would have expected an increase of bilayer thickness following polymerization (see discussion below). Mean diameters of polymerized vesicles were measured by QELS (Figure 4a,b). D h values were 95 nm (polydispersity 0.23) for EggPc vesicles and 110 nm (polydispersity 0.20) for F-PC vesicles. In both cases, polymerization has increased the vesicles’ mean diameter. Kinetics of Polymerization. The kinetics of ISODAC polymerization in F-PC and EggPC vesicles were inves-

2876

Langmuir, Vol. 17, No. 9, 2001

Figure 5. Conversion from ISODAC to poly(ISODAC) as a function of time. Polymerization was achieved in F-PC (solid circles) and EggPC (open circles) vesicles.

Krafft et al.

Figure 7. Differential molecular weight distributions of poly(ISODAC) synthesized in F-PC vesicles (solid circles) and EggPC vesicles (open circles), as obtained by size exclusion chromatography. Table 1. Average Molecular Weights and Polydispersity of Poly(ISODAC) Synthesized in F-PC or in EggPC Vesicles polymer

Mw (g mol-1)

Mn (g mol-1)

Mw/Mn

poly(ISODAC)/EggPC poly(ISODAC)/F-PC

2.90 × 9.7 × 104

2.24 × 3.4 × 104

1.30 ( 0.1 2.9 ( 0.1

106

106

ASTRA, are listed in Table 1. Results show that the polymer synthesized in F-PC vesicles (poly(ISODAC)/FPC) presents much smaller molecular weight (by a factor of 30) than the polymer synthesized in EggPC vesicles. It is also noteworthy that, in the case of F-PC vesicles, the polymer has a broader distribution and that chains with small masses constitute a large population of distribution. Discussion

Figure 6. Kinetics of ISODAC polymerization in F-PC (solid circles) and EggPC (open circles) vesicles. M0 and M are monomer concentrations at time t ) 0 and t, respectively.

tigated by measuring by 1H NMR the amount of monomer remaining in the reaction medium. The variation of monomer conversion percentage as a function of time is displayed in Figure 5. It shows that the overall polymerization rate is much higher for F-PC than for EggPC vesicles. Log M0-Log M varies linearly over time, showing that in both cases the partial order of the reaction with respect to monomer concentration is one (Figure 6). Polymer Characterization. Figure 7 presents the absolute molecular weight distributions of polymers calculated from the combined measurements of molecular weight (MALLS detector) and concentration (DRI detector) using ASTRA software.27 The different averaged molecular weights were calculated assuming that each slice of the chromatogram contains molecules of a single molecular weight or at least very narrow distributions of molecular weights.28 Mean values of molecular weights and polydispersity, and their standard deviations calculated using (27) Wyatt, P. J. Anal. Chim. Acta 1993, 272, 1. (28) Shut, D. W. J. Liq. Chromatogr. 1993, 16, 3371.

We have shown that polymerization of ISODAC in vesicles made from EggPC led to extended phase separation between the polymer and the lipid matrix. In the case of F-PC, on the other hand, (1) no evidence for latex bead formation was observed, (2) the rate of polymerization was higher, and (3) the molecular weight of the resulting polymer was lower. Vesicles prepared from F-PC have very special properties.10,15-18 Being both hydrophobic and lipophobic, fluorinated chains tend to segregate in water and form an internal fluorinated core within the bilayer that is surrounded by two lipophilic shells constituted by the hydrocarbon spacers (Scheme 2b). The lipophilic space between the fluorinated core and the hydrophilic headgroups is maintained by covalent bounds. This triple-layer hydrophobic structure is flanked by two hydrophilic layers on both outsides of the membrane. This topology was confirmed by X-ray diffraction of lamellar phases made of perfluoroalkylated phosphatidylcholines.29 Fluorinated vesicles have therefore a nonconventional nanocompartmentalized structure as compared to hydrogenated vesicles, which typically comprise a single lipophilic core flanked by two hydrophilic regions. Because of the presence of the fluorinated core, fluorinated vesicles are more stable and less permeable than hydrogenated vesicles.15,16 (29) McIntosh, T. J.; Simon, S. A.; Veirling, P.; Santaella, C.; Ravily, V. Biophys. J. 1996, 71, 1853.

Polymerization of a Hydrophobic Monomer Scheme 2. A Hydrogenated Bilayer (a) and a Fluorinated Bilayer (b), Schematically Represented

The results reported here indicate that due to their unique structure which constitutes a new ordered microheterogeneous medium, fluorinated vesicles withstand polymerization of a monomer inserted in their bilayers. It is likely that the hydrophobic monomer remains confined within the lipophilic shells of the bilayer and is excluded from the internal fluorinated core. The latter is strongly organized and maintains the integrity of the vesicles. The hydrogenated space that hosts the ISODAC monomer in F-PC vesicles cannot, contrary to EggPC vesicles, separate or expand. The following comments may further help understand the observed polymerization behavior. First, the reaction space available in FPC vesicles (dhc ) ca. 1 nm, calculated from the length of a fully extended six-carbon chain) is much smaller than in EggPC vesicles (dhc ) ca. 2.3 nm, as measured by image processing, or ca. 4 nm, as calculated on the basis of a fully extended 15-carbon chain). This leads to a higher local concentration of the monomers. One would thus expect the reaction rate to be higher, which is borne out by our experiments. Second, the

Langmuir, Vol. 17, No. 9, 2001 2877

extremely confined and fixed space in F-PC vesicles is likely to increase the probability of termination reactions. This is expected to lead to a lower polymer molecular weight, which is also in line with the experimental data. Taking into account the molecular weight of the polymer obtained in F-PC vesicles (Mw ) 9.7 × 104), its gyration radius is larger than the available space. Therefore, it is likely that, being in such confined and nonexpandable space, the polymer is forced to grow in a rodlike conformation. Finally, we have verified that poly(ISODAC) is soluble in hexane, but not in heavier hydrocarbons, such as tetradecane, or in perfluorooctane. This suggests that the hydrocarbon segment zone of the fluorinated bilayer, and not the fluorocarbon core (Scheme 2b), may constitute a good solvent for the polymer, while the 15-carbon chains of the hydrogenated bilayer (Scheme 2a) may be a poorer solvent, which may facilitate precipitation of the polymer. We can thus visualize the fluorinated bilayer as a confined medium that is able to direct and contain polymer growth into one major direction. Conclusions We have shown that small unilamellar vesicles made from a fluorinated phospholipid can be used as a matrix for intrabilayer polymerization of isodecyl acrylate, which was not possible with vesicles made from hydrogenated phospholipids. Our investigation of bilayer morphology, in particular by cryo-TEM, indicates nonambiguously that polymer chains do not form a phase-separated polymer bead attached to the vesicle, as in the case of hydrogenated vesicles, but are homogeneously distributed within the bilayer. To further characterize the polymer capsules, we are now attempting to cross-link the polymer and eliminate the fluorinated phospholipid matrix. Acknowledgment. The authors thank Prof. J. G. Riess (MRI Institute, Medical Center, University of California at San Diego) for helpful discussions. They gratefully acknowledge Alliance Pharmaceutical Corp. (San Diego, CA) for financial support. They also thank the Institut National de la Sante´ et de la Recherche Me´dicale, the Centre National de la Recherche Scientifique, and the Hoˆpital Universitaire de Strasbourg (M.S.). LA001658T