Contrasting Behavior of Zwitterionic and Cationic Polymers Bound to

Langmuir 2007, 23, 7539-7544

7539

Contrasting Behavior of Zwitterionic and Cationic Polymers Bound to Anionic Liposomes A. A. Yaroslavov,*,† T. A. Sitnikova,† A. A. Rakhnyanskaya,† Yu. A. Ermakov,‡ T. V. Burova,§ V. Ya. Grinberg,§ and F. M. Menger*,⊥ Department of Chemistry, M. V. LomonosoV Moscow State UniVersity, Leninskie Gory, Moscow 119992, Russia, A. N. Frumkin Institute of Physical Chemistry and Electrochemistry, RAS, Leninsky Prospect 31, Moscow 119991, Russia, Institute of Organo-Element Compounds, RAS, VaViloV Street 28, Moscow 119991, Russia, and Department of Chemistry, Emory UniVersity, Atlanta, Georgia 30322 ReceiVed March 5, 2007. In Final Form: May 2, 2007 Zwitterionic polymers were prepared by quaternizing polyvinylpyridine (DP ) 1100) with bromoacids (Br(CH2)nCOOH, where n ) 1, 2, 3, and 5). The resulting polymers were then added to unilamellar liposomes composed of egg lecithin or dipalmitoylphosphatidylcholine admixed with 20 mol % of cardiolipin (a phospholipid with two negative charges). These systems were compared (along with polyethylvinylpyridinium chloride, a polycation) by light scattering, electrophoretic mobility, fluorescence, and high-sensitivity differential scanning calorimetry. The external zwitterionic polymers induce no flip-flop of cardiolipin from the inner leaflet to the outer leaflet as does the polycation. Aside from this similarity, the four zwitterionic polymers all behave differently from each other toward the anionic liposomes: (a) For n ) 1, there is no detectable interaction between the polymer and the liposomes. (b) For n ) 2, electrostatic attraction induces polymer-liposome association (reversed by the addition of NaCl) that maintains the original negative charge on the liposome. Aggregation of the liposomes accompanies polymer adsorption. (c) For n ) 3, electrostatic binding also occurs along with aggregation. However, the binding is so strong that NaCl is unable to induce polymer/liposome dissociation. (d) For n ) 5, there is polymer binding and NaCl-promoted dissociation but no substantial aggregation. These differences among the closely related polymers are discussed and analyzed in molecular terms.

Introduction The interaction of synthetic water-soluble polymers with cell surfaces and lipid bilayer vesicles (liposomes) as cell mimetics has been intensively investigated for the last 30 years. Different aspects of such interactions have been discussed, including the composition and structure of interfacial complexes;1-3 polymerinduced structural rearrangements in cell and liposomal membranes;1,4-12 conformational transitions in the adsorbed polymers;13-15 effect of polymers on membrane perme* Corresponding author. E-mail: [email protected] (A.A.Y.); [email protected] (F.F.M.). † M. V. Lomonosov Moscow State University. ‡ Frumkin Institute of Physical Chemistry and Electrochemistry, RAS. § Institute of Organo-Element Compounds, RAS. ⊥ Emory University. (1) de Kruijff, B.; Rietveld, A.; Telders, N.; Vaandrager, B. Biochim. Biophys. Acta 1985, 820, 295. (2) Takahashi, H.; Matuoka, S.; Kato, S.; Ohki, K.; Hatta, I. Biochim. Biophys. Acta 1991, 1069, 229. (3) Bordi, F.; Cametti, C.; Gili, T.; Gaudino, D.; Sennato, S. Bioelectrochemistry 2003, 59, 99. (4) Laroche, G.; Pezolet, M.; Dufourcq, J.; Dufourc, E. J. Prog. Colloid Polym. Sci. 1989, 79, 38. (5) . Mittler-Neher, S.; Knoll, W. Biochim. Biophys. Res. Commun. 1989, 162, 124. (6) Yaroslavov, A. A.; Efimova, A. A.; Lobyshev, V. I.; Ermakov, Yu. A.; Kabanov, V. A. Membr. Cell Biol. 1997, 10, 683. (7) Raudino, A.; Castelli, F. Macromolecules 1997, 30, 2495. (8) Ikeda, T.; Yamaguchi, Y.; Tazuke, S. Biochim. Biophys. Acta 1990, 1026, 105. (9) Denisov, G.; Wanaski, S.; Luan, P.; Glaser, M.; McLaughlin, S. Biophys. J. 1998, 74, 731. (10) Kabanov, V. A. Makromol. Chem., Macromol. Symp. 1986, 1, 101. (11) Zakharova, O. M.; Rosenkranz, A. A.; Sobolev, A. S. Biochim. Biophys. Acta 1995, 1236, 177. (12) Macdonald, P. M.; Crowell, K. J.; Franzin, C. M.; Mitrakos, P.; Semchyschyn, D. J. Solid State Nucl. Magn. Reson. 2000, 16, 21. (13) Takahashi, H.; Yasue, T.; Ohki, K.; Hatta, I. Mol. Membr. Biol. 1996, 13, 233. (14) Lee, S.; Iwata, T.; Ouagi, H.; Aoyagi, H.; Ohno, M.; Anzai, K.; Kirino, Y. Biochim. Biophys. Acta 1990, 1151, 76.

ability;10,16-19 and aggregation, fusion, and disruption of cells and liposomes in the presence of polymers.2,4,20-22 The above-cited work focused mainly on cationic and nonionic polymers. Cationic polymers were investigated in part because of their strong binding to negative lipid membranes.23 Another reason was the ability of polycations to complex with oligonucleotides and DNA molecules, thus favoring intracellular gene delivery.24-29 Nonionic polyethylene oxide/polypropylene oxide block copolymers (Pluronics) were found to increase the permeability of cell membranes toward noncharged antitumor drugs in the anthracycline series.19,30 Unfortunately, polycations are often characterized by a high level of cytotoxicity,31 while (15) Fukushima, K.; Sakamoto, T.; Tsuji, J.; Kondo, K.; Shimozawa, R. Biochim. Biophys. Acta 1994, 1191, 133. (16) Uchida, D. A.; Irvin, C. G.; Ballowe, C.; Larsen, G.; Cott, G. R. Exp. Lung. Res. 1996, 22, 85. (17) Krylov, A. V.; Kotova, E. A.; Yaroslavov, A. A.; Antonenko, Yu. N. Biochim. Biophys. Acta 2000, 1509, 373. (18) Kozlova, N. O.; Bruskovskaya, I. B.; Okuneva, I. B.; Melik-Nubarov, N. S.; Yaroslavov, A. A.; Kabanov, V. A.; Menger, F. M. Biochim. Biophys. Acta 2001, 1514, 139. (19) Glazunova, O. O.; Korepanova, E. A.; Efimov, V. S.; Smirnov, A. I.; Vladimirov, Yu. A. Membr. Cell Biol. 1998, 12, 401. (20) Osanai, S.; Nakamura, K. Biomaterials 2000, 21, 867. (21) Oku, N.; Shibamoto, S.; Ito, F.; Gondo, H.; Nango, M. Biochemistry 1987, 26, 8145. (22) Yaroslavov, A. A.; Kiseliova, E. A.; Udalykh, O. Yu.; Kabanov, V. A. Langmuir 1998, 14, 5160. (23) Kabanov, V. A.; Yaroslavov, A. A.; Sukhishvili, S. A. J. Controlled Release 1996, 39, 173. (24) Behr, J. P.; Seances, C. R. Soc. Biol. Fil. 1996, 190, 33. (25) Sobolev, A. S.; Rosenkranz, A. A.; Smirnova, O. A.; Nikitin, V. A.; Neugodova, G. L.; Naroditsky, B. S.; Shilov, I. N.; Shatski, I. N.; Ernst, L. K. J. Biol. Chem. 1998, 273, 7928. (26) Kabanov, V. A.; Kabanov, A. V. AdV. Drug DeliVery ReV. 1998, 30, 49. (27) Lemieux, P.; Vinogradov, S. V.; Gebhart, C. L.; Guerin, N.; Paradis, G.; Nguyen, H. K.; Ochietti, B.; Suzdaltseva, Y. G.; Bartakova, E. V.; Bronich, T. K.; St. Pierre, Y.; Alakhov, V. Y.; Kabanov, V. A. J. Drug Targeting 2000, 8, 91. (28) Zou, S. M.; Erbacher, P.; Remy, J. S.; Behr, J. P. J. Gene Med. 2000, 2, 128. (29) Yamazaki, Y.; Nango, M.; Matsuura, M.; Hasegawa, Y.; Hasegawa, M.; Oku, N. Gene Ther. 2000, 7, 1148.

10.1021/la700637d CCC: $37.00 © 2007 American Chemical Society Published on Web 06/06/2007

7540 Langmuir, Vol. 23, No. 14, 2007

YaroslaVoV et al.

Table 1. Quaternization of PVP by Four Bromoacids and Ethyl Bromide to Give the Four PBs (PBn) and PEVP abbreviation

modifying agent

spacer in betaine group

aa

PB1 PB2 PB3 PB5 PEVP

ω-bromoacetic acid ω-bromopropionic acid ω-bromobutyric acid ω-bromocaproic acid ethylbromide

-(CH2)1-(CH2)2-(CH2)3-(CH2)5no betaine

0.98 0.98 0.55 0.73 0.95

a

R ) fraction of pyridines derivatized.

low-toxicity Pluronics demonstrate a poor affinity for biological membranes.32 In the present paper we describe the interaction of biological membranes with yet another class of ionic polymers, namely, those carrying both cationic and anionic units. They were prepared by the quaternization of poly(4-vinylpyridine) (PVP) with ω-bromocarboxylic acids to give polybetaines (PBs), in which each positive quaternized pyridinium group was covalently bound to a negative COO- group through a -(CH2)n- spacer. It was shown previously that PVP-based PBs form electrostatic complexes with negative linear polyions.33 PB binding to negatively charged biomembranes was, therefore, a reasonable expectation. PBs were coupled with small unilamellar liposomes, composed of neutral egg yolk lecithin (EL) or dipalmitoylphosphatidylcholine (DPPC), and doubly anionic negative diphosphatidylglycerol (cardiolipin, CL2-), as cell-mimetic species. The EL liposomal membrane was always in the liquid-crystalline state, as are membranes of native cells. Throughout the research, effects of PBs on the lipid membrane were compared with those caused by a conventional polycation, poly(N-ethyl-4-vinylpyridium bromide) (PEVP). In particular, the following items were investigated: (1) the adsorption of PBs on the surface of negatively charged liposomes; (2) structural rearrangements in the liposomal membrane induced by interacting PBs; and (3) the stability of the interfacial PB-lipid complexes in saline and in the presence of competing polyanions. Experimental Section 1. Materials. To obtain a series of PBs (I), a PVP fraction with a degree of polymerization (DP) equal to 1100 was prepared and then quaternized, as described in ref 34, with various ω-bromocarboxylic acids. For preparing PEVP (II), we followed the same procedure with identical PVP, but ethyl bromide was used as the modifying agent instead. The extents of quaternization (R) for all synthesized polymers were determined by IR spectroscopy measuring the ratio of intensities at 1600 and 1640 cm-1 (Table 1).35 Poly(acrylic acid) (PAA) from Aldrich, USA, with DP ) 70 was used as received. The concentrations of all PBs and PEVP are given in moles of quaternized units per liter, and those of PAA are given in moles of polymer units per liter. Natural lipids: anionic CL2- (III) and neutral EL (IV), as well as synthetic DPPC (IV) and N-fluorescein-iso-thiocyanyldipalmitoylphosphatidylethanolamine (FITC-DPPE) (V) were obtained from Sigma, USA, and used as received. (30) Venne, A.; Li, S.; Mandeville, R.; Kabanov, A. V.; Alakhov, V. Y. Cancer Res. 1996, 56, 3626. (31) Morgan, D. M.; Larvin, V. L.; Pearson, J. D. J. Cell Sci. 1989, 94, 553. (32) Krylova, O. O.; Melik-Nubarov, N. S.; Badun, G. A.; Ksenofontov, A. L.; Menger, F. M.; Yaroslavov, A. A. Chem.sEur. J. 2003, 9, 3930. (33) Izumrudov, V. A.; Domashenko, N. I.; Zhiryakova, M. V.; Davydova, O. V. J. Phys. Chem. B 2005, 109, 17391. (34) Fuoss, R. M.; Strauss, U. P. J. Polym. Sci. 1948, 3, 246. (35) Kirsh, Yu. E.; Pluzhnov, S. K.; Shomina, T. S.; Kabanov, V. A.; Kargin, V. A. Vysokomol. Soedin. 1970, 12A, 186 (in Russian).

Small unilamellar EL/CL2- and DPPC/CL2- liposomes with a 20% molar content of negative CL2- headgroups were prepared as follows: First, the corresponding amounts of EL (or DPPC) and CL2- solutions in methanol were mixed in a flask, after which the solvent was carefully removed under vacuum. The resulting thin layer of lipid mixture was dispersed in a 10-2 M borate buffer, pH 9.2, with a Cole-Parmer 4700 ultrasonic homogenizer for 400 s. Liposome samples thus obtained were separated from titanium dust by centrifugation and used within 1 day. The diameter of liposomes, measured by photon correlation spectroscopy, was in the 70-100 nm range. EL/CL2- liposomes with FITC-DPPE label incorporated into the bilayer were prepared by the same procedure except that 0.5 wt % of FITC-DPPE was added to the lipid mixture solution before the solvent evaporation. 2. Methods. The mean hydrodynamic diameters of liposomes, D, and their complexes with PBs were determined by photon correlation spectroscopy at a fixed scattering angle (90°) in a thermostatic cell using an Autosizer IIc (Malvern, UK), equipped with a He-Ne laser, and a Malvern K7032090 autocorrelator. The software provided by the manufacturer was employed to calculate the D values. The average values from 10 consecutive measurements are reported. The electrophoretic mobility (EPM) of liposomes and their complexes with PBs was measured by laser microelectrophoresis in a thermostatic cell using a Zetasizer IIc (Malvern, UK) equipped with a He-Ne laser in conjunction with a Malvern K7032090 autocorrelator and the software provided by the manufacturer. The fluorescence intensities of FITC-labeled liposome suspensions were measured at λem ) 525 nm (λex ) 495 nm) using an F-4000 fluorescence spectrophotometer (Hitachi, Japan). The pH measurements were carried out using a PHM83 potentiometer with standard glass electrode 2040C (Radiometer, Denmark). Phase transitions in DPPC/CL2- liposomes were registered with a DASM-4 differential adiabatic scanning microcalorimeter (Design Bureau for Biological Instruments, Pushchino, Russia). The samples were prepared as follows: Suspensions of liposomes and PBs in a 10-2 M phosphate buffer, pH 7.5, were heated separately up to 55°C (above the phase transition temperature) and then mixed. The mixtures were kept at this temperature for 5 min, then cooled to a room temperature and placed into the microcalorimeter cell. The calorimetric measurements were carried out at the heating rate of 0.25°C/ min within the range of 5-55° All experimental runs were repeated more than once, and average values are reported. To prepare solutions, we used double-distilled water additionally treated by passing through a Milli-Q system (Millipore, USA) equipped with ion-exchange and adsorption columns and a filter to remove large particles.

Results and Discussion The study of complexation between liposomes and ionic polymers has a long history. It began shortly after the preparation and detailed description of spherical lipid vesicles by the Bangham group in the 1960s.36 Since then, different methods for studying the complexation have been developed, including dynamic light scattering,37,38 electrophoresis,22,39 fluorescence,38,40,41 and dif(36) Bangham, A. D. Chem. Phys. Lipids 1993, 64, 275. (37) Bordi, F.; Cametti, C.; Diociaiuti, M.; Gaudino, D.; Gili, T.; Sennato, S. Langmuir, 2004, 20, 5214. (38) Yaroslavov, A. A.; Efimova, A. A.; Lobyshev, V. I.; Kabanov, V. A. Biochim. Biophys. Acta 2002, 1560, 14.

BehaVior of Zwitterionic/Cationic Polymer-Liposomes

ferential scanning calorimetry.6,42,43 All these methods were also used in our research into assessing and controlling the interaction of negative liposomes with PBs. Figure 1 shows the dependences of particle sizes in EL/CL2liposome-PB systems, measured by dynamic light scattering, as polymer concentrations increased (curves 1-4). As follows from the figure, the addition of PB2 and PB3 caused drastic enlargement of particles in the system (curves 2 and 3), consistent with their binding to the liposome surface and inducing aggregation. In contrast to this, only slight changes in the particle sizes were observed when PB1 and PB5 solutions were mixed with the liposome suspension (curves 1 and 4). This did not allow a definite conclusion as to whether these latter two PBs complex with liposomes (i.e., whether there is complexation without aggregation or no complexation at all). To distinguish between the two cases, the fluorescence approach was applied. Since quaternized poly(vinylpyridinium) cations are known to be effective fluorescence quenchers, polymer complexation to liposomes can also be detected by monitoring the fluorescence intensity of a tag incorporated into the liposomal membrane. The results of the fluorescence experiment are given in Figure 2. The addition of PB1 to a suspension of FITC-labeled EL/CL2liposomes had only faint effects on the FITC fluorescence (curve 1), indicating negligible PB1-to-liposome association. Meanwhile, a rise in PB5 concentration was accompanied by a sharp decrease in the tag’s fluorescence, obviously due to the adsorption of PB5 onto the liposome surface (curve 4). Two other PBs (PB2 and PB3), already shown to adsorb on the surface of labeled liposomes, also quenched the tag’s fluorescence, but with different efficiencies (curves 2 and 3, respectively). Thus, there was no systematic correlation between the quenching effect of PBs and their chemical structure. PB2 and PB5 caused the fluorescence to diminish to lower levels than PB3, a PB with a spacer of intermediate length. Surface charge and particle size are known to be important factors controlling the migration of a charged particle in an electric field. EPM was, therefore, determined for the anionic liposomes with and without added PB (Figure 3). Cationic polymer PEVP is seen in curve 3 to bind to the liposomes, reduce their negative charge, and, above 0.3 mM, ultimately impart a positive charge to them. These changes correlate well with changes in the size of the PEVP-liposome complex particles shown by curve 5 in (39) Son, K. K.; Tkach, D.; Patel, D. H. Biochim. Biophys. Acta 2000, 1468, 11. (40) Barenholz, Y.; Hirsch-Lerner, D. Biochim. Biophys. Acta 1998, 1370, 17. (41) Ikeda, T.; Lee, B.; Yamaguchi, Y.; Tazuke, S. Biophys. Biochim. Acta 1990, 1021, 56. (42) Ito, Y.; Okuyama, T.; Kashiwagi, T.; Imanishi, Y. J. Biomater. Sci., Polym. Ed. 1994, 6, 707. (43) Kennedy, M. T.; Pozharski, E. V.; Rakhmanova, V. A.; MacDonald, R. C. Biophys. J. 2000, 78, 1620.

Langmuir, Vol. 23, No. 14, 2007 7541

Figure 1. Neutralization of the liposome surface charge by adsorbed PEVP induces liposome aggregation, but further increase of polycation concentration causes the particle size to decrease. In this polycation concentration range, polycation-liposome complex particles are stabilized against aggregation by the abundant positive charge of the adsorbed polycation. PB2 and PB5 (curves 1 and 2) also reduce the EPM but much less dramatically than does PEVP, and they never bring the plots into the positive domain. More pronounced EPM reduction in the case of PB2 is accompanied by aggregation of the resulting PB2liposome complex (cf. with curve 2 in Figure 1); meanwhile, only slight changes in particle size are observed for the PB5liposome complex with high negative surface charge (cf. with curve 4 in Figure 1). The data in Figures 1-3 can be summarized as follows: (a) For PB1, there is no detectable interaction between polymer and liposomes. (b) For PB2, polymer binds to the liposomes and induces aggregation. (c) For PB3, polymer also binds to liposomes and induces their aggregation. (d) For PB5, polymer binds but induces little aggregation. How can these results be rationalized in molecular terms? Binding of the PBs to anionic liposomes must be considered in terms of the “independence” of the positive and negative charges within the polymers. PB1 does not bind to the anionic liposomes because, presumably, the pyridinium cation and carboxyl anion are proximate to each other. Consequently, the anionic charges, residing as they do close to the pyridinium cations, reduce the electrostatic attraction between the pyridinium ion and the liposomal CL2- that would otherwise promote polymer binding. Polymer binding to the liposomes does, however, occur with PB2, PB3, and PB5, where the carboxylates are positioned more remotely from the betaine positive charges. Pyridinium/ CL2- contact can, in other words, occur with a diminished repulsive CL2-/carboxylate interaction. Liposome aggregation is observed with PB2 and PB3 owing to interliposomal “salt bridges”, probably small in number, connecting polymer carboxylates and those pyridinium units not engaged in association with the liposomal cardiolipin. The absence of aggregation with PB5 may stem, we surmise, from a more well-developed corona of negative charge surrounding the liposome/polymer complex that electrostatically impedes liposome/liposome association. It was shown previously that the adsorption of cationic polymers with high linear charge density, such as polylysine and ethylated PVP, induced lateral lipid segregation in a mixed membrane composed of neutral and anionic lipids.38,41,44 As a result, the initial membrane divided into two two-dimensional mi(44) Yaroslavov, A. A.; Kuchenkova, O. Ye.; Okuneva, I. B.; Melik-Nubarov, N. S.; Kozlova, N. O.; Lobyshev, V. I.; Kabanov, V. A.; Menger, F. M. Biochim. Biophys. Acta 2003, 1611, 44.

7542 Langmuir, Vol. 23, No. 14, 2007

Figure 1. Effect of polyelectrolyte concentration on the mean hydrodynamic diameter of EL/CL2- liposomes: PB1 (1), PB2 (2), PB3 (3), PB5 (4), and PEVP (5). Total lipid concentration, 1 mg/mL; [CL2-] ) 1.5 × 10-4 M.

YaroslaVoV et al.

Figure 3. Effect of polyelectrolyte concentration on the EPM of EL/CL2- liposomes: PB2 (1), PB5 (2), and PEVP (3). Total lipid concentration, 1 mg/mL; [CL2-] ) 1.5 × 10-4 M.

Figure 2. Changes in the relative fluorescence intensity of labeled EL/CL2- liposomes in the presence of polyelectrolytes: PB1 (1), PB2 (2), PB3 (3), and PB5 (4). Total lipid concentration, 1 mg/mL; [CL2-] ) 1.5 × 10-4 M.

crophases: one enriched with neutral lipid and other with anionic lipid, the latter being segregated due to ionic contacts with adsorbed polycation units. An ability of polycations to induce lateral segregation in lipid membranes is of great interest from a biomedical point of view. Domain formation among lipids, for example, is accompanied by the creation of temporal defects in the lipid bilayer, thereby accelerating the membrane transport of small bioactive compounds.45 To check whether PBs are capable of provoking similar rearrangements in the lipid membrane, we exploited highsensitivity differential scanning calorimetry applied to liposomes composed of neutral DPPC and CL2-. Before studying the liposome/PB system, the phase transition temperatures for individual single-component DPPC and CL2- liposomes were determined. For the former, it was found to be 40.5 °C, which was in accordance with previously published data46 on the transition temperature for DPPC. We did not find information for CL2- liposomes in the literature. Our measurements showed (45) Ohno, H.; Shimidzu, N.; Tsushida, E.; Sasakawa, S.; Honda, K. Biochim. Biophys. Acta 1981, 649, 221. (46) Bach, D. In Biomembrane Structure and Function; Chapman, D., Ed.; Verlag Chemie: Basel, Switzerland, 1984, pp 1-141.

Figure 4. Calorimetric curves of DPPC/CL2- liposomes and their complexes with polyelectrolytes: (A) liposomes (1), liposomes + PEVP (2); (B) liposomes (1), liposomes + PB1 (2); (C) liposomes (1), liposomes + PB2 (2). Total lipid concentration, 1 mg/mL; [CL2-] ) 1.5 × 10-4 M; [PB] ) [PEVP] ) 5 × 10-4 M.

no phase transition in the range between 5° and 55 °C, indicating that the corresponding transition for CL2- probably lies below 5 °C. The binary DPPC/CL2- liposomes, prepared by sonication of a water-lipid suspension, were characterized by a rather wide transition profile with a maximum at 33.2°C and a shoulder at 37.1 °C (curves 1 in Figure 4A-C), apparently reflecting the coexistence of two types of mixed DPPC/CL2- phases with different DPPC-to-CL2- ratios. Concurrently, CL2- molecules are distributed nearly equally between both membrane leaflets, as was shown by previous electrophoresis experiments.38 To prepare liposome-polymer samples, the dispersions of DPPC/CL2- liposomes and polymer solutions were preheated above the liposomal membrane phase transition temperature, mixed with polyelectrolyte at 2[CL2-]/[quaternized unit] ) 1.6, and then cooled to room temperature. Thus, the excess of polyelectrolyte initially interacted with liposomal membranes in

BehaVior of Zwitterionic/Cationic Polymer-Liposomes

the liquid state. The results for three systems containing cationic PEVP (taken as a control) and two PBs (PB1 and PB2) are represented by curves 2 in Figure 4A-C, respectively. Adsorption of PEVP on the outer leaflet of the liposomal membrane caused dramatic changes in the calorimetric curve: the melting peak became narrower and shifted to 40 °C, a temperature close to the melting point of the single-component DPPC liposomes (Figure 4A, curve 2). It means that most CL2molecules, initially located on both membrane leaflets, were involved in the complexation, leaving large sections of pure DPPC to melt near it normal transition temperature of 40 °C. This effect, namely, the polycation-induced transmembrane migration of negative lipid molecules (polycation-induced flipflop), was described for the first time in ref 47. Less pronounced thermotropic changes were observed as PB1 solution was added to DPPC/CL2- liposome suspension. A shift of the melting peak to higher temperatures was also registered in this case, but the resulting peak remained broad as before (Figure 4b, curve 2). A rather minor effect of PB1 on the transition profile was in agreement with the conclusion of a negligible PB1 binding to negative liposomes based on the light scattering data. Results for the liposome/PB2 system were intermediate between the two preceding cases. On one hand, the adsorption of PB2 on the liposome surface was accompanied by lateral lipid segregation, as follows from the data of Figure 4C, curve 2. At the same time, the coexistence of two nearly equal peaks in the calorimetric curve at 37.4 and 40 °C indicate incomplete microphase separation of CL2-. It is possible that only CL2- molecules from the outer membrane leaflet (without CL2- flip-flop) were involved in lateral segregation mediated by their complexation with PB2. The mechanism describing the adsorption of cationic polymers (e.g., PEVP) on the anionic liposomal membrane has been discussed earlier in a series of publications.38,44,48 A driving force for the adsorption consists of an interfacial electrostatic complex stabilized by salt bridges between polymer units and lipid headgroups. By potentiometric titration with a Br--selective electrode of PEVP-liposome complex suspensions with 2[CL2-]/ [quaternized unit] g 1, the fraction of quaternized PEVP units involved in electrostatic binding with liposomes was found to be about 90%.23 The residual units, being ionic pairs with strongly held Br- anions, are excluded from the complexation. Polycationto-negative liposome interaction is accompanied by the release of small counterions (Na+ from the CL2-, and Br- from the PEVP). The resulting entropy increase in the system no doubt contributes to the adsorption free energy.43 The binding of PBs to the negative liposome surface is, as we have argued, assisted by electrostatic interactions as well. However, here, the Na+ counterions from the outer CL2-, but not those from the macromolecular chain itself, are available for releasing. Therefore, the gain in the entropy of binding should be smaller than that in the case of PEVP (with the same DP). The difference in thermodynamics between PB and PEVP complexation should be revealed by the stability of the resulting complexes in water-salt media. Liposome/PB complexes should likely dissociate at lower salt concentrations than PEVP complexes. Complex dissociation, monitored by the fluorescence technique, revealed that the resulting liposome/PB complexes responded differently to an increase in salt concentration (Figure 5). Thus, injection of NaCl solution to suspensions of PB2- and PB5-containing complexes, in which fluorescence was partially (47) Yaroslavov, A. A.; Kul’kov, V. Ye.; Polinsky, A, S.; Baibakov, B. A.; Kabanov, V. A. FEBS Lett. 1994, 340, 121. (48) Yaroslavov, A. A.; Yaroslavova, E. G.; Rakhnyanskaya, A. A.; Menger, F. M.; Kabanov, V. A. Colloids Surf., B 1999, 16, 29.

Langmuir, Vol. 23, No. 14, 2007 7543

Figure 5. Changes in the relative fluorescence intensity of labeled EL/CL2- liposomes complexed with PBs after NaCl addition: PB2 (1), PB3 (2), and PB5 (3). Total lipid concentration, 1 mg/mL; [CL2-] ) 1.5 × 10-4 M; [PB] ) 2 × 10-4 M.

quenched, led to a recovery of the fluorescence intensity, the initial level being achieved at [NaCl] ) 0.05 M (curves 1 and 3). This indicates a complete dissociation of the PB2- and PB5coated liposomes due to a shielding effect of Na+ and Cl- ions. Hydrophobic forces apparently play only a small role here because the complexes would then remain stable when NaCl was added. As to the labeled liposome/PB3 complex, its fluorescence, and hence its stability, was nearly insensitive to the addition of NaCl (curve 2). Importantly, the PB3 complex is uniquely stable under physiological salt concentration, a fact that suggests future biomedical applications. The fluorescence data were consistent with the results of diameter measurements. The NaCl-induced dissociation of PB2and PB5-containing complexes decreased the sizes of particles in the suspensions down to the initial liposome diameter, whereas the addition of NaCl had only a minor effect on the size of liposome/PB3 complex aggregates (data not shown). What could be a reason for such unexpectedly high stability of the liposome-PB3 complex? As follows from Table 1, owing to the vagaries of organic synthesis, the PB polymers have different extents of quaternization (R): close to the ultimate for PB1 and PB2, but only 0.55 for PB3 and 0.73 for PB5. Rather high free pyridine contents in the latter two cases could, in principle, hydrophobically anchor the polymers to the liposomal membrane. However, as we demonstrated above, a salt readily displaces PB5 from its complex with liposomes. Moreover, we have been able to increase R for PB3 from 0.55 to 0.71 (equivalent to that of PB5) without a noticeable change in its binding properties. In particular, the latter’s complex with negative liposomes completely dissociated as the concentration of salt increased. Although not definitive, these arguments strongly suggested the absence of an “R” effect and forced us to consider a different version. It is known that six-membered cyclic structures exhibit an elevated thermodynamic stability. In principle, the betaine groups in PB3 can also assume six-membered rings due to internal salt bridges. Such ion-paired rings with mutually neutralized positive and negative charges could penetrate deeper into the lipid bilayer and thus help stabilize the liposome-PB3 complex, making it relatively insensitive to the salt concentration. Charge-mediated ring formation in PB3 effectively decreased the amount of available quaternized pyridinium units and diminished its

7544 Langmuir, Vol. 23, No. 14, 2007

YaroslaVoV et al.

Figure 7. Changes in the relative fluorescence intensity of labeled EL/CL2- liposomes complexed with PBs after PAA addition: PB2 (1), PB3 (2), and PB5 (3). Total lipid concentration, 1 mg/mL; [CL2-] ) 1.5 × 10-4 M; [PB] ) 2 × 10-4 M. Figure 6. Structures of liposome-PB2 (A) and liposome-PB3 (B) complexes (schematic presentation).

quenching effect (in comparison with PB2 and PB5), as was already noted (see Figure 2). Ion-paired carboxylate and pyridinium are obviously in equilibrium with disconnected ions so that both conformations of the betaine contribute in PB3-tonegative liposome interaction. Structures of the liposome/PB2 and liposome/PB3 complexes are schematically represented in Figure 6. Thus, only PB2 and PB5 were both electrostatically bound to negative liposomes but completely removed from the liposomal membrane with increasing NaCl concentration. It was shown previously that the cationic polymer, PEVP, can also be removed from the EL/CL2- liposome surface by the addition of NaCl. However, complete PEVP removal requires a much higher NaCl concentration of 0.25 M.48 As is reasonable, therefore, electrostatic complexes of negative liposomes with PBs (with no net charge on the latter) demonstrate a lower stability in aqueous salt media than do liposome complexes with a conventional polycation. The removal of cationic polymers from the negative surface can also be induced by competitive complexation with linear polyanions such as anionic polyacrylate (PAA-).38,44 We found that PAA- also removed electrostatically coupled PB2 and PB5 from the negative liposomal membrane. This was shown in the course of the following experiment: First, FITC-labeled negative liposomes were complexed with both PBs, and equilibrium levels of fluorescence quenching in these systems were achieved (ca. 1 min after mixing). Increasing amounts of PAA- were then added to liposome/PB complex samples, and changes in the FITC fluorescence intensity were monitored. The ultimate levels of fluorescence recovery are represented in Figure 7 (curves 1 and 3). It is seen that, in both systems, a 2-fold excess of PAAassured a recovery of fluorescence intensity up to the initial level. Therefore, complete removal of both PBs from the liposome surface could be achieved because of their association with PAAand the formation of PB/PAA- polycomplexes. In sharp contrast, the PAA--competition experiment with PB3 was unable to remove

PB3 from the negative membrane (curve 2), once again providing evidence of a special stabilization of the liposome/PB3 complex.

Conclusions Complexation of negative liposomes to zwitterionic derivatives of PVP takes place only if the length of the spacer separating cationic and anionic charges in the betaine units equals two methylene groups or more. The complexation is accompanied by a decrease in the surface charge of liposomes, aggregation, and lateral lipid segregation. Similar effects have also been revealed for the interaction of negative liposomes with a cationic polymer, PEVP. However, there are important differences between the behavior of the zwitterionic and cationic polymer systems: (i) Binding of PEVP to liposomes results in a reversal of the liposome surface charge from negative to positive. When PB binds to liposomes, the surface charge remains negative even in the excess of polymers. (ii) Adsorption of PEVP induces flipflop in the negative membrane. This results in irreversible disruption of membranes with high content of anionic lipid.22 In contrast, calorimetric data indicate no flip-flop in the negative liposome/PB system. (iii) Complexes of liposomes with PEVP are stable at physiological NaCl concentration (0.15 M). Electrostatic complexes of liposomes with PBs, on the other hand, completely dissociate under these conditions. However, the case of PB with a -(CH2)3- spacer between cationic and anionic groups is different in that the complexation could be retained in salt solutions apparently due to the penetration of internally ion-paired betaine groups into the liposomal membrane. PBs seem to have potential applications for biomedical purposes. Acknowledgment. The authors highly appreciate the support of some parts of this research by the Russian Foundation for Basic Research (Grant 05-03-33033) and the Fogarty International Research Cooperation Award (Grant TW05555). F.M.M. was supported by the National Institutes of Health. LA700637D