Conjugated Polyelectrolyte

Dec 17, 2009 - Our studies further underscore the potential application of lipid/conjugated polymer complexes in light-harvesting devices. Our work fo...
18 downloads 0 Views 2MB Size
pubs.acs.org/Langmuir © 2009 American Chemical Society

Assembly of Zwitterionic Phospholipid/Conjugated Polyelectrolyte Complexes: Structure and Photophysical Properties An Thien Ngo and Gonzalo Cosa* Department of Chemistry, McGill University, 801 Sherbrooke Street West, Montreal, QC, H3A 2K6, Canada, and Centre for Self-Assembled Chemical Structures (CSACS). Received October 28, 2009. Revised Manuscript Received November 26, 2009 We report on the formation of complexes between zwitterionic phospholipid vesicles and an anionic fluorescent conjugated polyelectrolyte and the effect of mono- and divalent cations on the photophysical properties of these complexes. Our goal is to gain an understanding of the interplay of morphology and exciton transport in these complexes, information that is critical to designing efficient lipid/conjugated polymer-based sensors. Our studies further underscore the potential application of lipid/conjugated polymer complexes in light-harvesting devices. Our work focuses on the negatively charged conjugated polyelectrolyte poly[5-methoxy-2-(3-sulfopropoxy)-1,4-phenylenevinylene] (MPS-PPV) and its interaction with the zwitterionic lipid dioleoylphosphatidylcholine (DOPC). We utilize monovalent and divalent cations as a tool to control and explore the interaction of MPS-PPV with lipids. We show that Ca2þ ions promote the complexation of zwitterionic lipids and MPS-PPV in comparison to Naþ ions. The addition of increasing amounts of zwitterionic phospholipids in the form of vesicles gradually disrupts MPS-PPV aggregates albeit vesicle structure is preserved in Naþ buffer. Lipid complexation and the resulting MPS-PPV aggregate disruption produces an intensity enhancement and blue shifting of the MPS-PPV emission peak. In the absence of Ca2þ, the intensity enhancement and blue shift reach a plateau at larger than a 10:1 lipid/MPS-PPV monomer mole ratio. In the presence of Ca2þ, a plateau is reached at equimolar concentrations of MPS-PPV and lipid. Vesicle particle coalescence and agglomerate formation are observed herein. Lipid complexation and concomitant MPS-PPV shielding is shown to diminish the quenching of MPS-PPV emission by water-soluble quencher methyl viologen. FRET experiments conducted with membrane-intercalating acceptor dye DiD further underscore the large lipid/polymer interaction mediated by Ca2þ. We observe efficient light harvesting and MPS-PPV-amplified emission quenching in Ca2þ buffer and to a lesser extent in Naþ buffer. Our results highlight how the interplay of a zwitterionic lipid, cations, and buffer, in combination with the conjugated polyelectrolyte MPS-PPV, provides rich diversity in architecture and photophysical properties.

Introduction Conjugated polymers are organic macromolecules with a backbone of alternating single and double bonds functionalized with side groups to facilitate their solubility in organic solvents.1,2 When functionalization involves ionic side groups, water-soluble conjugated polymers or conjugated polyelectrolytes are obtained.2 Conjugated polymers in general are characterized by their high emission quantum yield, large molecular weight, and rapid migration of charge carriers and excitons along the π-conjugated backbone. Given these unique properties, these novel materials hold great promise as building blocks in the construction of electroluminescent3,4 and photovoltaic devices5,6 and in the development of sensors.7-9 The potential applications of conjugated polymers in optoelectronic devices have sparked a significant number of studies on *To whom correspondence should be addressed. E-mail: gonzalo.cosa@ mcgill.ca. Tel: 514-398-6932. Fax: 514-398-3797. (1) Thomas, S. W.; Joly, G. D.; Swager, T. M. Chem. Rev. 2007, 107, 1339–1386. (2) Jiang, H.; Taranekar, P.; Reynolds, J. R.; Schanze, K. S. Angew. Chem., Int. Ed. 2009, 48, 4300–4316. (3) Burroughes, J. H.; Bradley, D. D. C.; Brown, A. R.; Marks, R. N.; Mackay, K.; Friend, R. H.; Burns, P. L.; Holmes, A. B. Nature 1990, 347, 539–541. (4) Ho, P. K. H.; Kim, J.-S.; Burroughes, J. H.; Becker, H.; Sam, F. Y. L.; Brown, T. M.; Cacialli, F.; Friend, R. H. Nature 2000, 404, 481–484. (5) Sariciftci, N. S.; Smilowitz, L.; Heeger, A. J.; Wudl, F. Science 1992, 258, 1474–1476. (6) Gunes, S.; Neugebauer, H.; Sariciftci, N. S. Chem. Rev. 2007, 107, 1324– 1338. (7) Swager, T. M. Acc. Chem. Res. 1998, 31, 201–207. (8) Chen, L.; McBranch, D. W.; Wang, H.-L.; Helgeson, R.; Wudl, F.; Whitten, D. G. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 12287–12292. (9) Scholes, G. D.; Rumbles, G. Nat. Mater. 2006, 5, 683–696.

6746 DOI: 10.1021/la904100q

their exciton migration properties. It has been shown that exciton migration along the polymer backbone is very dependent on polymer conformation. Rapid exciton migration is normally encountered in polymer chains that are highly folded and/or aggregated, where through-space energy transfer is a viable transport mechanism, but it is not common in extended polymer chains, where through-bond energy transfer is the dominant exciton transport mechanism.10 Exciton migration along the polymer backbone has a direct effect on the photophysical properties of a conjugated polymer, such as its emission quantum yield and its emission wavelength.11,12 It also affects the exquisite amplified sensing properties that characterize conjugated polymers.8,13 Rapid exciton migration to low-energy sites, nonemissive traps, or molecular quenchers located along a polymer backbone accounts for polymer red emission, low emission quantum yield, and amplified fluorescence quenching1,7 or superquenching.14,15 A number of studies have been conducted toward modulating the exciton migration and, concomitantly, the emissive and (10) Tan, C.; Atas, E.; Muller, J. G.; Pinto, M. R.; Kleiman, V. D.; Schanze, K. S. J. Am. Chem. Soc. 2004, 126, 13685–13694. (11) Smith, A. D.; Shen, C. K.-F.; Roberts, S. T.; Helgeson, R.; Schwartz, B. J. Res. Chem. Intermed. 2007, 33, 125–142. (12) Schwartz, B. J. Annu. Rev. Phys. Chem. 2003, 54, 141–172. (13) Jiang, H.; Zhao, X.; Schanze, K. S. Langmuir 2007, 23, 9481–9486. (14) Achyuthan, K. E.; Bergstedt, T. S.; Chen, L.; Jones, R. M.; Kumaraswamy, S.; Kushon, S. A.; Ley, K. D.; Lu, L.; McBranch, D.; Mukundan, H.; Rininsland, F.; Shi, X.; Xia, W.; Whitten, D. G. J. Mater. Chem. 2005, 15, 2648–2656. (15) The term amplified fluorescence quenching or superquenching has been coined to define the marked fluorescence quenching in conjugated polymers or conjugated polyelectrolytes, respectively.

Published on Web 12/17/2009

Langmuir 2010, 26(9), 6746–6754

Ngo and Cosa

Figure 1. Structures of the MPS-PPV polymer, the acceptor dye DiD, and the zwitterionic lipid DOPC used in this study.

sensing properties of conjugated polymers. In the specific case of conjugated polyelectrolytes, these properties can be tuned by altering the degree of aggregation of the chromophore segments along the conjugated polyelectrolyte backbone via either solvent polarity,11 ionic strength,11,16-18divalent cations,11,13,19 or complexation with surfactants.20-24 In the last case, it has been shown that conjugated polyelectrolyte aggregates are disrupted upon the formation of micelles with surfactants, resulting in a significant, up to 20-fold increase in emission quantum yield and a blue shift in emission.25 Recently, the complexation of conjugated polyelectrolytes with phospholipids and the changes that they impart to the polymer emission properties have been exploited in novel lipasesensing platforms26,27 and sensor arrays for the detection and differentiation of different cell lines28 as well as in devising novel biocidal agents.26,29,30 The interplay of electrostatic and hydrophobic interactions between conjugated polyelectrolytes and lipids is crucial to the success of these devices. For example, (i) the transformation of zwitterionic phosphocholine lipids into negatively charged phosphatidic acids by lipases upon hydrolysis of the phosphoester bond reduces lipid/polymer interactions and renders polymers nonemissive; (ii) the desorption of negatively charged conjugated polyelectrolytes from a positively charged fluorescence-quenching Au nanoparticle scaffold by the membranes of some cell lines enables their specific detection; and (iii) the specific polyelectrolyte/lipid interaction enables in situ singlet oxygen sensitization on the cellular membrane of bacteria. Herein, we set out to characterize the interaction of the zwitterionic lipid dioleoylphosphatidylcholine (DOPC) with the negatively charged conjugated polyelectrolyte poly[5-methoxy-2-(3sulfopropoxy)-1,4-phenylenevinylene] (MPS-PPV, Figure 1). Our (16) Fan, Q. L.; Zhou, Y.; Lu, X. M.; Hou, X. Y.; Huang, W. Macromolecules 2005, 38, 2927–2936. (17) Cabarcos, E. L.; Carter, S. A. Macromolecules 2005, 38, 10537–10541. (18) Cabarcos, E. L.; Carter, S. A. Macromolecules 2005, 38, 4409–4415. (19) Jiang, H.; Zhao, X.; Schanze, K. S. Langmuir 2006, 22, 5541–5543. (20) Al Attar, H. A.; Monkman, A. P. J. Phys. Chem. B 2007, 111, 12418–12426. (21) Burrows, H. D.; Tapia, M. J.; Fonseca, S. M.; Pradhan, S.; Scherf, U.; Silva, C. L.; Pais, A. A. C. C.; Valente, A. J. M.; Schillen, K.; Alfredsson, V.; Carnerup, A. M.; Tomsic, M.; Jamnik, A. Langmuir 2009, 25, 5545–5556. (22) Chen, L.; McBranch, D.; Wang, R.; Whitten, D. Chem. Phys. Lett. 2000, 330, 27–33. (23) Chen, L.; Xu, S.; McBranch, D.; Whitten, D. J. Am. Chem. Soc. 2000, 122, 9302–9303. (24) Dalvi-Malhotra, J.; Chen, L. J. Phys. Chem. B 2005, 109, 3873–3878. (25) Treger, J. S.; Ma, V. Y.; Gao, Y.; Wang, C.-C.; Wang, H.-L.; Johal, M. S. J. Phys. Chem. B 2008, 112, 760–763. (26) Chemburu, S.; Ji, E.; Casana, Y.; Wu, Y.; Buranda, T.; Schanze, K. S.; Lopez, G. P.; Whitten, D. G. J. Phys. Chem. B 2008, 112, 14492–14499. (27) Liu, Y.; Ogawa, K.; Schanze, K. S. Anal. Chem. 2007, 80, 150–158. (28) Bajaj, A.; Miranda, O. R.; Kim, I.-B.; Phillips, R. L.; Jerry, D. J.; Bunz, U. H. F.; Rotello, V. M. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 10912–10916. (29) Corbitt, T. S.; Sommer, J. R.; Chemburu, S.; Ogawa, K.; Ista, L. K.; Lopez, G. P.; Whitten, D. G.; Schanze, K. S. ACS Appl. Mater. Interfaces 2008, 1, 48–52. (30) Ding, L.; Chi, E. Y.; Chemburu, S.; Ji, E.; Schanze, K. S.; Lopez, G. P.; Whitten, D. G. Langmuir 2009, 25, 13742–13751.

Langmuir 2010, 26(9), 6746–6754

Article

goal is to understand better the interplay of exciton transport and lipid/conjugated polyelectrolyte complex morphology, information that is critical to designing efficient lipid/polymer sensors and liposome beacons.31 Our studies were conducted in the presence of monocations (Naþ) and dications (Ca2þ) where hydrophobic and both hydrophobic and electrostatic interactions are expected, respectively, between the lipids and polymers (Ca2þ is known to bind to zwitterionic lipids conferring on them cationic properties).32-36 Studies with different cations thus provide unique opportunities to tune and study the binding of lipids with the anionic polyelectrolyte MPS-PPV. We show that MPS-PPV readily intercalates within vesicle lipid membranes, an effect markedly enhanced by Ca2þ over Naþ. A number of experiments support this conclusion, including the order of lipid to polymer addition, vesicle rupture experiments, vesicle size studies via dynamic light scattering (DLS), and MPS-PPV emission spectra and emission quantum yields under increasing DOPC/MPS-PPV monomer molar ratios and emission-quenching experiments both with hydrophilic electron acceptors (methyl viologen MV2þ) and lipophilic energy acceptors (DiD). Novel zwitterionic phospholipid/conjugated polyelectrolyte supramolecular assemblies are produced with unique spectroscopic characteristics. These studies underscore the potential of employing conjugated polyelectrolytes to understand better the interaction of lipids with charged polymers and ultimately highlight the potential application of lipid/polymer complexes in light-harvesting devices.

Materials and Methods Materials. HEPES acid, methyl viologen (MV2þ), and a 0.25 wt % solution of poly[5-methoxy-2-(3-sulfopropoxy)-1,4phenylenevinylene] (MPS-PPV) in distilled deionized water were purchased from Sigma-Aldrich (Oakville, ON, Canada). Sodium bicarbonate, sodium chloride, calcium chloride, and sodium hydroxide were from Fisher Scientific Canada (Ottawa, ON, Canada). Sephacryl S-500 HR was from GE Healthcare Biosciences (Piscataway, NJ). 5- and 6-Carboxyfluorescein (mixed isomers) and 1,10 -dioctadecyl-3,3,30 ,30 -tetramethylindodicarbocyanine perchlorate (DiD) were purchased from Invitrogen (Burlington, ON, Canada). Dioleoylphosphatidylcholine (DOPC) was acquired from Avanti Polar Lipids (Alabaster, AL). Water was purified by a Millipore Milli-Q system. All buffers that were utilized had the same ionic strength: 10 mM in HEPES, pH 7.3. The Ca2þ buffer was 10 mM in CaCl2, and the Naþ buffer was 30 mM in NaCl. Vesicle Preparation. DOPC vesicles in either Ca2þ buffer or Naþ buffer were prepared as follows. DOPC powder was dissolved in chloroform to a final concentration of 40-60 mg/mL. Aliquots of this solution were transferred to individual glass vials. The solvent was evaporated while rotating the sample vial to create a thin film on the vial wall. Films were placed under vacuum for at least 1 h to remove excess solvent. Lipid films were then hydrated to a final concentration of 20 mM lipids with buffer and subjected to 15 cycles of freeze-thaw-sonicate, where freeze indicates 3 min in dry ice, thaw indicates 1 min in a water bath of 40-50 C, and sonicate indicates 5 min of sonication at 40-50 C. (31) Ngo, A. T.; Karam, P.; Fuller, E.; Burger, M.; Cosa, G. J. Am. Chem. Soc. 2007, 130, 457–459. (32) Kharakoz, D. P.; Khusainova, R. S.; Gorelov, A. V.; Dawson, K. A. FEBS Lett. 1999, 446, 27. (33) Marra, J.; Israelachvili, J. Biochemistry 1985, 24, 4608–4618. (34) Lis, L. J.; Lis, W. T.; Parsegian, V. A.; Rand, R. P. Biochemistry 1981, 20, 1771–1777. (35) McManus, J. J.; Raedler, J. O.; Dawson, K. A. J. Phys. Chem. B 2003, 107, 9869–9875. (36) McManus, J. J.; Radler, J. O.; Dawson, K. A. J. Am. Chem. Soc. 2004, 126, 15966–15967.

DOI: 10.1021/la904100q

6747

Article The resulting solution was translucent and colorless. The lipid mixtures were subsequently extruded 15 times at room temperature through a 100 nm pore size polycarbonate membrane unless otherwise noted.37 Absorption and Emission Spectra. Steady-state fluorescence spectroscopy was carried out using a Cary Eclipse spectrophotometer. Absorption spectra were recorded using a Cary 300Bio (Varian) UV-vis spectrophotometer in double-beam mode. For all steady-state absorption and emission experiments, the solutions were placed in 1 cm  1 cm quartz cuvettes. All emission spectra were taken at 450 nm unless otherwise stated. Samples were diluted 50- to 100-fold in buffer so that the absorbance at 450 nm was between 0.05 and 0.1. These absorbances correspond to an MPSPPV concentration in monomer units of 1.610-5 to 3.210-5 M, where the experimentally determined extinction coefficient for MPS-PPV is 3500 M-1 cm-1 per monomer unit at 450 nm. Titration of MPS-PPV with Lipid Aliquots. Solutions of MPS-PPV with an absorbance of 0.05 to 0.1 were prepared in Naþ or Ca2þ buffer. Aliquots of a solution of extruded vesicles, prepared as described above and diluted as needed, were added to the MPS-PPV and thoroughly mixed. Fluorescence spectra were taken immediately after vesicle addition. Approximately 5 min elapsed between adding each aliquot. Emission was corrected for sample dilution. Titration of Lipid with MPS-PPV Aliquots. Freshly prepared 100 nm vesicles were combined with aliquots of MPS-PPV in order to achieve a series of solutions 1  10-5 M in lipids (0.1 nM in vesicles) and with MPS-PPV concentration in the range of 1  10-7 to 1  10-4 M in monomer units. The fluorescence spectra of these samples were next measured together with the spectra of free polymer at the same concentration. Free polymer emission was corrected by the scattering from a buffer sample, whereas the emission from the lipid/polymer complex was corrected for scattering from a 1  10-5 M lipid vesicle solution with no polymer.

Blending MPS-PPV and Lipid via Freeze-Thaw Cycles and Extrusion. Dry lipid films were hydrated with MPS-PPV

solutions in either Naþ or Ca2þ buffers at the appropriate DOPC/ MPS-PPV monomer mole ratio. Samples were subjected to 10 freeze-thaw-sonicate cycles (3 min in dry ice/1 min between 40 and 50 C/5 min of sonication at 40-50 C). Immediately prior to analysis by fluorescence spectroscopy, each sample was extruded by passing it 20 times through a 100 nm polycarbonate membrane in a Mini-Extruder (Avanti Lipids Corp.). Dynamic Light Scattering (DLS). Aliquots of sample were added to buffer that had been filtered through 0.45 μm hydrophilic syringe filters. The intensity-intensity time correlation of scattered 532 nm laser light was measured at 25 C, perpendicular to the incident beam, using a Brookhaven BI-9000 AT digital correlator and BI-200SM goniometer system. The optical density filter and detector slit were adjusted to yield a scattering intensity of 10-100 kcps. The normalized intensity-intensity time correlation function was transformed to the normalized electric field time correlation function via the Siegert relation and evaluated via the CONTIN regularization algorithm. Vesicle Stability and Release of Vesicle Content. The effect on vesicle stability following MPS-PPV addition was evaluated by monitoring the leakage of carboxyfluorescein. Aliquots of lipid vesicles were added to each of the following solutions: buffer, 0.04% w/v Triton X-100 in buffer, and MPS-PPV 2.010-5 M in buffer, where buffer denotes 150 mM NaCl, 10 mM HEPES, pH 7.3. Immediately after addition, solutions were mixed well and fluorescence spectra were obtained at 450 nm excitation on a Cary Eclipse spectrofluorometer. The lipid vesicle sample consisted of extruded DOPC vesicles, a fraction of which were loaded with carboxyfluorescein (37) Liposomes, 2nd ed.; Torchilin, V. P., Weissig, V., Eds.; Oxford University Press: New York, 2003; Vol. 264.

6748 DOI: 10.1021/la904100q

Ngo and Cosa (CF-DOPC), where the unencapsulated carboxyfluorescein was removed by gel filtration using a column packed with Sephacryl S-500 HR (preequilibrated upon elution of DOPC liposomes). The final range of [DOPC] was 6.0  10-8 to 1.7  10-3 M. The concentration of carboxyfluorescein encapsulated in the CF-DOPC vesicles was 43 mM in 108 mM NaCl, 10 mM HEPES, pH ∼7. The lipid content in the extruded, filtered vesicles was determined from the known carboxyfluorescein/lipid mole ratio and the absorption of the encapsulated carboxyfluorescein. Namely, all of the fractions corresponding to the vesicle elution were collected, and the carboxyfluorescein absorbance was measured. Knowing that each 100 nm DOPC vesicle contains ca. 76 600 lipids37 and that the carboxyfluorescein trapped within the 2.2  10-18 L volume vesicle is ca. 43 mM, we can readily determine that the ratio of carboxyfluorescein to lipid is 57:77 in a 100-nm-diameter vesicle. Self-quenching is expected at this concentration of carboxyfluorescein packed within vesicles.38 Energy Transfer to a Membrane-Embedded Dye. Lipid vesicles were prepared as described above, where the lipid solution in chloroform was supplemented with lipophilic dye DiD prior to lipid film formation. Six samples were prepared where the fraction of dye was gradually increased from 0:1500 to 10:1500 DiD/ DOPC mole ratio. Lipid films containing increasing DiD/DOPC mole ratios were hydrated in Ca2þ or Naþ buffer, and upon freeze-thaw-sonication cycles (but no extrusion to minimize sample loss), the resulting vesicles were added to MPS-PPV. Solutions with a constant DOPC/MPS-PPV monomer mole ratio of 1 but increasing DiD/DOPC ratios were obtained. We note that under our experimental conditions we are working under a large excess of donor and as such FRET is not expected to be quantitative. Samples were excited at 440 nm to minimize the direct excitation of the membrane-embedded dye. The Forster radius (R0) for the MPS-PPV/DiD donor/acceptor pair was calculated according to eq 1,39 where κ2 is related to the orientation factor for the transition dipoles of donor and acceptor, taken to be 2/3 given the random orientation expected between donors and acceptors, n is the refractive index, φD is the donor emission quantum yield, and J(λ) is the overlap integral calculated from the normalized, corrected fluorescence intensity of the donor (FD(λ)) and the extinction coefficient of the acceptor (εA(λ)). 2 R0 ¼ 48:785  10 -25

Z

¥

JðλÞ ¼

31=6 ! ΦD K2 JðλÞ5 ðcmÞ n4

ð1Þ

FD  εA  λ4  dλ

0

Results and Discussion Absorption and Fluorescence of DOPC-MPS-PPV Complexes. We first monitored changes in the MPS-PPV emission peak and emission intensity upon addition of zwitterionic lipid DOPC. The experiments were conducted on MPS-PPV solutions of equal ionic strength containing either 30 mM NaCl or 10 mM CaCl2. Figure 2 shows the emission spectra for MPS-PPV when adding increasing amounts of DOPC in the form of 100 nm vesicles. A steady intensity increase reaching a 5- to 6-fold enhancement is accompanied by a gradual shift in the emission peak from ca. 565 to 544 nm as lipids are added to a polymer solution in buffer. Both the peak shift and intensity increase over a free polymer plateau at a DOPC/MPS-PPV monomer ratio in the (38) Klausner, R. D.; Kumar, N.; Weinstein, J. N.; Blumenthal, R.; Flavin, M. J. Biol. Chem. 1981, 256, 5879–5885. (39) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 3rd ed.; Springer: New York, 2006.

Langmuir 2010, 26(9), 6746–6754

Ngo and Cosa

Article

Figure 2. Titration of MPS-PPV with DOPC vesicles and vice versa. Background-subtracted MPS-PPV emission spectra obtained upon increasing vesicle addition in (A) 30 mM NaCl, 10 mM HEPES, pH 7.2, with 1.5  10-5 M MPS-PPV and (B) 10 mM CaCl2, 10 mM HEPES, pH 7.2, with 1.5  10-5 M MPS-PPV. (A, B) Normalized to the emission of free MPS-PPV in Naþ and Ca2þ buffers, respectively. (C) Intensity enhancement (9) and peak shift (0) vs the DOPC/MPS-PPV mole ratio measured in Naþ buffer. The experimental points correspond to the ratio of intensity at the MPS-PPV emission maximum measured with free polymer (i.e., no lipids (Io) and with increasing lipid concentration (I) for data from (•••) the titration of MPS-PPV with DOPC vesicles, (green ---) the titration of DOPC vesicles with MPS-PPV, (() MPS-PPV blended with DOPC at a DOPC/MPS-PPV ratio of 12. (D) Intensity enhancement (red b) and peak shift (red O) vs the DOPC/MPS-PPV mole ratio measured in Ca2þ buffer. The experimental points correspond to (•••) the titration of MPS-PPV with DOPC vesicles, (green ---) the titration of DOPC vesicles with MPS-PPV, and (red () MPS-PPV blended with DOPC at a DOPC/MPS-PPV ratio of 12. The lines connecting the experimental points are a visual aid.

range of 10 to 100 (Figure 2A,C). Results obtained upon addition of 100 nm vesicles prepared in solutions containing Ca2þ are remarkably different, whereupon the intensity enhancement and peak shift occur over a short range of DOPC/MPS-PPV values in the vicinity of a mole ratio value of 1 (i.e., at equimolar concentrations of DOPC and MPS-PPV monomers, Figure 2B,D). Herein the peak shift is steep, and it qualitatively reflects the contributions of two species, a red and a blue one, with the latter contributing a larger proportion as the lipids are added to the polymer (Figures 2B, 6, and 7). The MPS-PPV emission (λmax) originally at ca. 570 nm in Ca2þ buffer shifts to a value as low as 540 nm at DOPC/MPS-PPV mole ratios larger than 1. Furthermore, there is merely a 2.5-fold MPS-PPV emission enhancement in the Ca2þ buffer in comparison to the ca. 5-fold emission enhancement in the Naþ buffer solution upon adding DOPC vesicles to polymer. Remarkably, the order of addition (i.e., lipid vesicles to polymer, as above, or polymer to lipid vesicles; see Scheme 1 and text below) affects the spectroscopic changes. When MPS-PPV is added to a solution of freshly prepared 100 nm vesicles in Naþ buffer, the emission of the polymer increases by ca. 7-fold in the initial points when the DOPC/MPS-PPV mole ratio is larger than 10 (Figure 2C). A similar phenomenon is observed when the polymer is added to vesicle solutions prepared in a 10 mM Ca2þ solution, albeit the changes occur over a small concentration range (Figure 2D). A 9-fold intensity enhancement is observed upon addition of MPS-PPV to a vesicle solution when the DOPC/ MPS-PPV ratio is larger than 1. The value for the intensity enhancement readily drops with decreasing DOPC/MPS-PPV ratio as more polymer is added to the vesicle-containing solution. Intensity enhancements of 10-fold both for Naþ and Ca2þ are also obtained when MPS-PPV is included in the hydration buffer Langmuir 2010, 26(9), 6746–6754

Scheme 1. Cartoon illustrating the effect that the order of addition (vesicles to polymer or polymer to vesicles) has on the particles formed in Ca2þ. Note that particles with a small amount of polymer will be very emissive (green shadow)

utilized in preparing the vesicles and mechanically blended with the lipids (Figure 2C,D, diamonds). Freeze-thaw cycles and extrusion lead to membrane rupture and healing in the solution containing MPS-PPV, ensuring intimate mixing of the polymer with the lipids. A number of conjugated polyelectrolyte-micellar surfactant interactions have been described, which in all cases lead to a significant intensity enhancement and emission blue shift for the conjugated polyelectrolyte upon polymer aggregate disruption.20-25 In the specific case of conjugated polyanion MPS-PPV, studies DOI: 10.1021/la904100q

6749

Article

with micellar cationic surfactant 1,2-dioleoyl-3-trimethylammonium propane (DOTAP)22-24 reveal an association stoichiometry of a 0.6/1 mole ratio between DOTAP and MPS-PPV. The spectroscopic changes that we report upon addition of nonmicellar surfactant DOPC are qualitatively in agreement with those previously published. The results in Naþ solution point to weak interactions between the zwitterionic lipids and polyanion MPS-PPV, interactions that are hydrophobic but not electrostatic in nature.31 A large excess of lipids is required to achieve complete binding to the polymer as revealed from the changes in intensity and emission peak with lipid content. Under these conditions, a large fraction of the lipids are not directly involved in complex formation with MPS-PPV. The presence of Ca2þ in solution dramatically enhances the polymer-lipid interaction. Roughly equimolar concentrations of DOPC in Ca2þ result in the saturation of polymer with lipids, as reflected by the lack of any further changes in the MPS-PPV emission peak or emission intensity at DOPC/MPS-PPV mole ratios larger than 1. Ca2þ has been shown to confer cationic properties to a zwitterionic lipid when interacting with polyanions, as revealed by the structure of complexes of DNA, dipalmitoylphosphatidylcholine (DPPC), and Ca2þ.35,36 Our results in the quaternary system Ca2þ, MPS-PPV, DOPC, and buffer are in line with those observations where the binding of the zwitterionic lipid to MPS-PPV is significantly enhanced as a result of electrostatic interactions, much as would be expected for a cationic micellar detergent. Strong lipid/polymer interaction in Ca2þ buffer leads to quantitative polymer uptake by the vesicles in 1:1 stoichiometry with no free lipid fraction. A high density of MPS-PPV in lipids is expected. The close proximity of polymer strands in turn enables rapid exciton transport to nonemissive sites, limiting the fluorescence enhancement to ca. 2-fold (Scheme 1, bottom first step). In contrast, the addition of polymer to a large excess of DOPC (Scheme 1, top, first step) or a blending of the polymer and DOPC via freeze-thaw-sonication under otherwise identical conditions produces a ca. 9-fold increase compared to that for free MPS-PPV in Ca2þ buffer. The order of the vesicle-to-polymer addition plays a key role in the Ca2þ solution, whereupon lipid/polymer complexes formed upon adding vesicles to MPS-PPV do not rearrange as an increasing number of vesicles are added beyond the equimolar lipid/polymer concentrations. The polymer molecules are trapped in close proximity to each other; as a result, the exciton may migrate efficiently to nonemissive traps, concomitant with lowintensity enhancement. Long incubation times or mechanical blending distribute the polymer more widely within the lipid, increasing the emission enhancement. This situation may be contrasted with that in the absence of Ca2þ, where the order of addition produces only small differences in the intensity enhancement, indicating that a significant fraction of lipids are uncomplexed at equimolar lipid/polymer concentrations. Fluorescence Quenching Studies. To further explore the complexation of MPS-PPV and lipids, we next studied the quenching of MPS-PPV when added to a great excess of DOPC (Scheme 1, top, first step) in the form of lipid vesicles. At a DOPC/MPS-PPV mole ratio of 140:1, the MPS-PPV emission intensity was monitored upon addition of increasing amounts of methyl viologen (MV2þ), a well-known MPS-PPV quencher that may not cross lipid bilayers. Figure 3 illustrates the Stern-Volmer fluorescence quenching plots both in Naþ and Ca2þ buffers. The Stern-Volmer quenching constant (Ksv) that we calculated for MV2þ was ca. 6-fold larger in Naþ- than in Ca2þ-containing solutions with values of 4.0 103 and 0.7 103 M-1, respectively. 6750 DOI: 10.1021/la904100q

Ngo and Cosa

Figure 3. Stern-Volmer fluorescence quenching plots for MPSPPV embedded within DOPC vesicles. DOPC vesicles were added to the MPS-PPV solution in a single aliquot (DOPC/MPS-PPV mole ratio of 140:1) and mixed thoroughly prior to quenching with methyl viologen. (9) MPS-PPV 1.5  10-5 M in 30 mM NaCl (Ksv = 4.0  103 M-1) and (red b) MPS-PPV 1.5  10-5 M in 10 mM CaCl2 (Ksv = 0.7  103 M-1). Also shown are the linear fit to the initial data points for both quenching experiments. Both samples were prepared in 10 mM HEPES, pH 7.3 buffer.

These values are markedly smaller than the values for MPS-PPV when free in Naþ and Ca2þ buffers (Ksv = 1.13  106 and 9.40  105 M-1, respectively; data not shown). The Stern-Volmer constant (Ksv) is defined as the product of the quenching constant kq and the decay lifetime in the absence of quencher τo; see eq 2. The effect of ions and lipid vesicles on Ksv can be understood in terms of how they individually affect kq and τo. Io ¼ 1 þ kq  τo  ½MV2þ  I

ð2Þ

On the basis of our experiments, the association of the polymer with a large excess of lipids results in a ca. 10-fold increase in the fluorescence quantum yield (φf) and therefore a 10-fold increase in the decay lifetime τo for Naþ and Ca2þ buffers, respectively (where φf is directly proportional to τo). Given the increase in τo with lipid complexation, the marked reduction in the Ksv value for MV2þ quenching in going from free polymer to membraneembedded polymer (250- and 1500-fold reductions in Naþ and Ca2þ buffers, respectively) can be solely attributed to a drastic decrease in kq. A similar argument (a drop in kq) may be used to rationalize the 6-fold-larger drop in Ksv for Ca2þ versus Naþ buffers. A number of conclusions can be made at this point: (i) the Naþ and Ca2þ ions have a negligible effect on the value of Ksv for free MPS-PPV in solutions with medium ionic strength;40 (ii) as has been shown previously, association with surfactants (nonmicellar surfactants in our study) hinders the polymer complexation with MV2þ, reducing the quenching by this species;22 and (iii) for DOPC, the presence of Ca2þ significantly hinders the interaction of the polymer with MV2þ when compared to Naþ. This would be consistent with the stronger binding of the lipids to the polymer in the presence of Ca2þ over Naþ, in agreement with the MPS-PPV titration studies with increasing DOPC described in Figure 2. (40) At lower ionic strength, a 20-fold increase in Ksv is observed for MPS-PPV quenching with MV2þ in 1 mM CaCl2 versus 1 mM LiCl, where the reduced screening of electrostatic interactions combined with Ca2þ-mediated polymer aggregation at this low ionic strength may account for this result . Smith, A. D.; Shen, C. K.-F.; Roberts, S. T.; Helgeson, R.; Schwartz, B. J. Res. Chem. Intermed. 2007, 33, 125–142.

Langmuir 2010, 26(9), 6746–6754

Ngo and Cosa

Figure 4. (A) Titration of MPS-PPV with DOPC vesicles. Aliquots of extruded DOPC vesicles (100 nm) were added to a solution of MPS-PPV (1.6  10-5 M). (B) Titration of DOPC vesicles (1.6  10-5 M in lipids) with MPS-PPV in the presence and absence of calcium ions. Aliquots of MPS-PPV were added to a solution of extruded DOPC vesicles (100 nm). The weighted-average hydrodynamic diameter of the particles was monitored by dynamic light scattering (532 nm) after each successive aliquot addition and plotted here against the DOPC/MPS-PPV ratio. (9) 30 mM NaCl, 10 mM HEPES, pH 7.3; (red b) 10 mM CaCl2, 10 mM HEPES, pH 7.3.

Also (see below), Ca2þ promotes the formation of large multilamellar agglomerates that would isolate much of the polymer from MV2þ, further decreasing kq. In closing this section, we note that the Stern-Volmer fluorescence quenching plots exhibit downward curvature indicative of an unquenchable fraction. The number of data points collected, however, precludes us from obtaining an accurate estimate of the unquenchable fraction.39 Particle Size Studies. We next explored via dynamic light scattering (DLS) the effect of polymer addition on vesicle size. We recorded the vesicle particle sizes in Naþ and Ca2þ buffers upon adding lipids to a constant amount of MPS-PPV (Figure 4A and Scheme 1, bottom) and upon adding MPS-PPV to a constant amount of lipid (Figure 4B and Scheme 1, top). In the experiments conducted in Naþ buffer, the average particle size remained fairly constant at around 120 nm regardless of the DOPC/MPS-PPV ratio or the order of addition (i.e., vesicle to polymer or polymer to vesicle (Figure 4, black squares)). For those experiments conducted in Ca2þ buffer solutions, the lipid/ polymer interactions, in contrast, gave rise to the formation of large agglomerates that were 300-1000 nm in size and larger (Figure 4, red circles). Upon close examination of the experiments in Naþ, we observed no discernible trend in particle size as the vesicles were added to the polymer; the average particle size was between 110 and Langmuir 2010, 26(9), 6746–6754

Article

140 nm (Figure 4A). However, a close examination of Figure 4B reveals a gradual 10% increase in size, from 110 to 124 nm, as MPS-PPV is added to vesicles. This trend is within experimental error; nonetheless, one may expect that following MPS-PPV insertion within a vesicle bilayer, the vesicle should increase in size. When analyzing the results obtained following the addition of DOPC vesicles to polymer in Ca2þ (Figure 4A), we observe an increase in particle size giving rise to a fraction of micrometersized particles in addition to a fraction of ca. 100 nm particles. Two species are present, vesicles with polymers forming large micrometer-sized agglomerates and vesicles that are polymer-free and do not agglomerate (Scheme 1, bottom). Arguably, under the large polymer excess conditions initially prevailing in the MPSPPV titration with DOPC vesicles, all vesicles readily become coated with polymer given the high lipid/polymer affinity promoted by Ca2þ. With an increasing number of vesicles, past the 1:1 DOPC/MPS-PPV ratio, all polymer originally present is complexed with the lipids and new vesicles find no free polymer with which to interact. A second population of vesicles that are 100 nm in diameter is thus observed in addition to the vesicle/ polymer agglomerates whose size is in the micrometer range. When MPS-PPV was added to vesicles in Ca2þ buffer (Figure 4B), a steady increase in particle size was recorded. Particles in the micrometer size range were obtained at a DOPC/ MPS-PPV mole ratio equal to or smaller than 1; an increasing amount of polymer in solution led to particle agglomeration (Scheme 1, top). Dynamic light scattering studies indicate that the increased width of the spectra in Figures 2B and 6B (Ca2þ buffer) compared to that in Figures 2A and 6A (Naþ buffer) results from the presence of at least two species in the solution. Smaller particles (100 nm lipid vesicles), where the polymer is thoroughly intercalated and isolated within the vesicle membrane, would be expected to produce high-intensity blue emission upon polymer aggregate disruption. Lipid/polymer agglomerates formed at a 1:1 lipid polymer ratio, however, yield low-emission particles with a red-shifted emission peak.11,41 The order of addition in Ca2þ buffer (i.e., whether vesicles are added to excess polymer or vice versa) influences the sequence of structural changes in the particles upon association. The addition of vesicles to excess polymer (Scheme 1, bottom) may produce polymer-coated vesicles that gradually agglomerate; these particles do not efficiently break up MPS-PPV aggregates. The addition of small amounts of polymer to an excess of vesicles (Scheme 1, top) would produce intervesicle bridging leading to the rupture of the vesicle membrane,42 where the large excess of lipids ensures proper polymer aggregate disruption and a concomitant intensity enhancement and blue shift in emission. We may conclude that Ca2þ strengthens lipid/polymer complexation, resulting in vesicle agglomeration and rupture. These results are akin to those observed in complexes between cationic lipids and DNA, and underscore the role of Ca2þ in strengthening the lipid/polyanion interaction upon conferring zwitterionic lipids cationlike properties.42 Vesicle Rupture Studies. To determine whether vesicular membrane integrity is compromised by the addition of MPS-PPV in Naþ buffer, we conducted vesicle rupture studies in this media. On the basis of our experimental data, DOPC vesicles maintain their integrity upon MPS-PPV addition in Naþ buffer. DOPC (41) Kaur, P.; Yue, H.; Wu, M.; Liu, M.; Treece, J.; Waldeck, D. H.; Xue, C.; Liu, H. J. Phys. Chem. B 2007, 111, 8589–8596. (42) Kennedy, M. T.; Pozharski, E. V.; Rakhmanova, V. A.; MacDonald, R. C. Biophys. J. 2000, 78, 1620–1633.

DOI: 10.1021/la904100q

6751

Article

Ngo and Cosa

Figure 5. Fluorescence increase following the rupture of DOPC vesicles loaded with 43 mM carboxyfluorescein. (O) Upon adding vesicles to MPS-PPV 2  10-5 M and (9) upon adding vesicles to Triton X-100 0.04% w/v; under these conditions, 90% vesicle rupture occurs.43 Also displayed are control experiments showing (2) MPS-PPV 2  10-5 M emission enhancement following DOPC vesicle addition and (3) the emission of carboxyfluorescein when DOPC vesicles encapsulating 43 mM carboxyfluorescein are added to the buffer.

vesicles were loaded with 43 mM carboxyfluorescein, at which concentration significant self-quenching (ca. 10-fold intensity reduction) is expected.38 Vesicles encapsulating carboxyfluorescein were then added in aliquots to a solution of MPS-PPV in 150 mM NaCl, 10 mM HEPES, pH 7.3. Three control experiments were performed: (i) the addition of carboxyfluoresceinloaded vesicles to Triton X-100 (full rupture); (ii) the addition of carboxyfluorescein-loaded vesicles to buffer (no rupture); and (iii) the addition of empty vesicles to MPS-PPV (fluorescence enhancement due solely to polymer lipid complexation); see Figure 5. Vesicles added to Triton yielded an 8-fold increase in fluorescence over vesicles added to buffer as a result of the Tritonmediated vesicle rupture and concomitant carboxyfluorescein dilution. In contrast, vesicles added to MPS-PPV experienced only a slight intensity increase. The small increase in fluorescence that we measured for vesicles added to MPS-PPV equals the sum of the emission from control samples ii and iii above; see Figure 5. In conclusion, vesicles added to MPS-PPV in Naþ buffer retained an intact hollow structure capable of encapsulating carboxyfluorescein. Energy-Transfer Studies of Membrane-Embedded Dyes. A further understanding of the combined role of lipids and cations in the spectral properties of MPS-PPV comes from F€orster resonance energy transfer (FRET) studies with MPSPPV and membrane-intercalating dye DiD as the energy acceptor. FRET experiments underscore the large difference in lipid/ polymer interactions for Naþ versus Ca2þ buffers (Figure 6). They also highlight the potential application of lipid/polymer complexes in light-harvesting devices. We exploited FRET as a means to monitor the lipid/polymer proximity indirectly via both the fluorescence enhancement of the DiD dye monitored at 670 nm and the concomitant quenching of the MPS-PPV polymer emission. Our experiment consisted of adding DOPC vesicles to various MPS-PPV solutions where the DOPC/MPS-PPV mole ratio was kept constant at a value of 1 but with vesicles embedding an increasing amount of DiD (from a 0:1500 to 10:1500 DiD/DOPC mole ratio). As expected (Figures 6 and 2), the addition of DOPC vesicles with no DiD to MPS-PPV solutions resulted in an emission peak blue shift and 1.5-fold emission intensity enhancement relative to free polymer in Naþ. The addition of vesicle embedding increasing amounts of DiD 6752 DOI: 10.1021/la904100q

Figure 6. Energy transfer between MPS-PPV and lipid-embedded dye DiD. Experiments were conducted at a 1:1 DOPC/MPS-PPV monomer mole ratio in 10 mM HEPES, pH 7.3 buffer, where the concentration of MPS-PPV was 1.5  10-5 M. Excitation was carried out at 440 nm. (A) Solutions were 30 mM in NaCl; emission spectra were taken immediately after the addition of lipid vesicles to MPS-PPV solution. (B) Solutions were 10 mM in CaCl2; emission spectra were taken immediately after the addition of lipid vesicles to MPS-PPV solution. The intensity curves shown are relative to the intensity of free polymer (A) in Naþ or (B) in Ca2þ. Free polymer was 2-fold as intense in Naþ vs Ca2þ, and care should be taken in applying this factor when comparing intensities in spectra A and B.

leads to the quenching of MPS-PPV and the sensitization of the DiD emission. Energy transfer from the MPS-PPV polymer to the DiD dye solely accounts for the observed DiD emission because at the 440 nm excitation wavelength that we employed, absorption from DiD was negligible. A control experiment with DiD in DOPC and no MPS-PPV showed no DiD emission when exciting at 440 nm. The higher lipid/polymer affinity promoted by Ca2þ was evident from the increased polymer to DiD FRET recorded in Ca2þ over Naþ buffer, manifest both in a marked drop in emission at 540 nm and an increase in emission at 670 nm. (We note that the DiD emission quantum yield was the same in both Naþ and Ca2þ buffers as observed upon direct excitation.) Figure 6B displays the MPS-PPV emission intensity in Ca2þ buffer upon adding DOPC vesicles embedding increasing amounts of DiD to various MPS-PPV solutions. The intensities are displayed relative to those of the free polymer in the same buffer. At a 5:1500 DiD/DOPC ratio, ca. 50% of the MPS-PPV intensity is quenched when in Ca2þ compared to 40% when in Naþ buffer (where the extent of quenching is calculated from the polymer emission intensity in solutions containing lipids with no DiD vs lipids with DiD at a 5:1500 DiD/DOPC ratio). Under these conditions (5:1500:1500 Langmuir 2010, 26(9), 6746–6754

Ngo and Cosa

DiD/DOPC/MPS-PPV), each DiD is able to quench the emission of 300 monomers or 30 MPS-PPV quasi-chromophores assuming a chromophore length of 10 monomers.44 Doubling the DiD content to 10:1500 DiD/DOPC results in quenching of ca. 40 and 60% of the initial intensity for lipid/polymer assemblies in Naþ and Ca2þ, respectively. Notably, most of the photons emitted in the Ca2þ buffer under the 5:1500:1500 DiD/DOPC/MPS-PPV condition come from sensitized DiD, as can be directly observed from the peak intensities at 545 and 670 nm, a measure of the good light-harvesting properties of the polymer. Excitation spectra acquired at 670 nm are consistent with the FRET observations from Figure 6A,B where a band due to the absorption of energy donor MPS-PPV is readily observed (not shown). Given the low emission quantum yield (φD) for polymer in Ca2þ buffer with added lipids (φD = 0.009 for DOPC/MPS-PPV= 1) and despite the good spectral overlap between the donor emission spectrum and the acceptor absorption spectrum,31 the F€orster radius R0 values that we have calculated for energy transfer between MPS-PPV and DiD are in the range of 3.1 nm. In Naþ buffer, unless otherwise identical conditions are indicated, larger R0 values (3.5 nm) are calculated because the polymer has a ca. 2-fold larger φD value when membraneembedded. In Naþ buffer, the larger R0 does not translate to enhanced quenching as a result of the weaker polymer lipid binding, yet again reflecting that a large fraction of the lipids are not directly involved in complex formation with MPS-PPV under these conditions (Figure 6). Figure 7 illustrates that the characteristics of the lipid/polymer complexes in Ca2þ buffer are primarily a function of the lipid/ polymer ratio. Working with two different initial polymer concentrations (2.1  10-6 and 2.1  10-5 M), upon increasing the DOPC/MPS-PPV mole ratio with DiD-embedding DOPC vesicles, the trends in fluorescence enhancement and FRET are identical despite the fact that the solution in Figure 7A is 10-fold more concentrated than that in Figure 7B. In all cases, these experiments were conducted above the critical micelle concentration of the DOPC lipid.35 When adding increasing amount of vesicles with a 1:1500 DiD/ DOPC ratio, the relative quenching for MPS-PPV in Ca2þ buffer remained constant for 1:1, 10:1, and 100:1 DOPC/MPS-PPV mixtures although the DiD content increased by 2 orders of magnitudes overall. The results support the hypothesis that Ca2þ promotes the 1:1 binding of lipid to polymer where a large fraction of the lipids are directly involved in complex formation with MPS-PPV at DOPC/MPS-PPV mole ratios smaller than or equal to 1. Maximum FRET is observed at 1:1 DOPC/MPS-PPV, indicating that the amount of lipid (with its intercalated dye) closely associated with the polymer corresponds to 1 lipid per monomer. Subsequent additional lipid may be unable to bind to the polymer, and no further FRET is therefore expected. Our results show that the prevalent structure in Naþ buffer is that of lipid vesicles embedding MPS-PPV within their membranes, where a fraction of lipids are not directly involved in complex formation with MPS-PPV. Free MPS-PPV is aggregated in aqueous solutions, which gives rise to its low emission quantum yield and red emission. The action of the lipids lies in disrupting the polymer aggregates; a large excess of lipid is necessary to achieve the disruption. Importantly, no vesicle rupture takes place upon embedding the polymer within the membrane. Given that the lipids are zwitterionic, we may foresee that the interactions (43) de la Maza, A.; Parra, J. L. Colloid Polym. Sci. 1996, 274, 866–874. (44) Woo, H. S.; Lhost, O.; Graham, S. C.; Bradley, D. D. C.; Friend, R. H.; Quattrocchi, C.; Bredas, J. L.; Schenk, R.; M€ullen, K. Synth. Met. 1993, 59, 13–28.

Langmuir 2010, 26(9), 6746–6754

Article

Figure 7. Energy transfer between MPS-PPV and lipid-embedded dye (DiD) upon addition of vesicles to polymer. Two different initial MPS-PPV concentrations were tested: (A) 2.1  10-5 and (B) 2.1  10-6 M. Experiments were conducted at a 1:1500 DiD/ DOPC mole ratio in 10 mM HEPES, pH 7.3 buffer, 10 mM in CaCl2. DOPC/MPS-PPV mol ratios of 1:1 (red b) and 10:1 (green 2) in DOPC/MPS-PPV were tested (also shown is the emission of free MPS-PPV (9) and a control sample with DOPC and DiD but no MPS-PPV, blue -). Upon excitation at 440 nm, the fluorescence spectra show the same trend in enhancement and FRET at the two different lipid/polymer ratios.

between polymer and lipid are weak and mostly involve the hydrophobic backbone of the polymer, which readily embeds within the hydrophobic core of the lipid bilayer. It is critical to note that the weak nature of the interaction is very dynamic; regardless of whether lipids are added to polymers or vice versa, the structures obtained are similar. In polymer titrations with DOPC, the trends in particle size, emission intensity enhancement, and emission blue shift all indicate that the polymer association to the lipid bilayer is dynamic, enabling polymer exchange between loaded vesicles and freshly added, polymer-free vesicles. The addition of Ca2þ markedly affects the prevalent structure. It has long been known that Ca2þ binds to phosphatidylcholine bilayers.45 The number of lipid headgroups around each Ca2þ has been estimated to be two,35 three,46 or four,47 where it is believed that Ca2þ neutralizes the negatively charged phosphate moieties in the lipid headgroups,35 imparting a net positive charge to the bilayer. Consistent with Ca2þ conferring cationic properties upon zwitterionic lipids, its presence herein results in strong lipid/polymer (45) Bangham, A. D.; Dawson, R. M. C. Biochim. Biophys. Acta 1962, 59, 103– 115. (46) Tatulian, S. A. Eur. J. Biochem. 1987, 170, 413–420. (47) Boeckmann, R. A.; Grubmueller, H. Angew. Chem., Int. Ed. 2004, 43, 1021– 1024.

DOI: 10.1021/la904100q

6753

Article

interactions, as expected for the interaction of cationic lipid with MPS-PPV and as has been reported for the interaction of micellar cationic surfactants DOTAP and DTA with MPS-PPV.22-24 The new interactions promoted by Ca2þ are electrostatic in nature and give rise to agglomerate formation in solution most probably involving vesicle rupture. Although at present we have no direct information on the molecular structure of the complex between MPS-PPV and DOPC promoted by Ca2þ, it is known that in complexes among lipid, Ca2þ, and anionic polyelectrolyte DNA, the net positive charge on the bilayer causes the two lipid headgroups to reorient so that their positive amine groups are pointing toward the anionic polyelectrolyte.35,36 It is tempting to speculate that such an arrangement also prevails herein. Cryo-TEM and small-angle X-ray scattering (SAXS) will certainly enable us to understand this organization at the molecular level better. We may further speculate that in agreement with reports on DNA binding to cationic lipids48 and on Ca2þ-promoted DNA binding to zwitterionic lipids35,36 the polymer-bound lipid phase is multilamellar in nature. At the macromolecular level, the DOPC-MPSPPV complexes are agglomerates of polymer and lipid that rapidly increase in size with vesicle addition, reaching values in the micrometer range. The lipid/polymer complex stoichiometry involves a 1:1 DOPC/MPS-PPV monomer ratio. At approximately this stoichiometry, the spectroscopic changes reach a plateau. When lipid is added to polymer, the emission enhancement is 5-fold smaller than when polymer is added to excess lipid under otherwise identical conditions. This is consistent with polymers tightly arranged in multilamellar structures, where the structures formed are kinetically trapped. In effect, the subsequent addition of lipid in excess of a 1:1 ratio does not affect (48) Radler, J. O.; Koltover, I.; Salditt, T.; Safinya, C. R. Science 1997, 275, 810– 814.

6754 DOI: 10.1021/la904100q

Ngo and Cosa

polymer emission, and the newly added lipid vesicles remain intact in solution as observed from DLS studies.

Conclusions Our results underscore how the interplay of a zwitterionic lipid, conjugated polyelectrolyte MPS-PPV, cations, and buffer provides a rich diversity of architectures and photophysical properties for MPS-PPV. In these quaternary systems, different results are obtained simply on the basis of the order of addition and the cation employed. Interestingly, the addition of Ca2þsignificantly enhances the formation of lipid/polymer complexes consistent with Ca2þ, conferring cationic properties to DOPC. The multilamellar structures formed could play key roles in light-harvesting systems relying on lipid/polymer complexes.

Abbreviations Poly(phenylenevinylene) (PPV), poly[5-methoxy-2-(3-sulfopropoxy)-1,4-phenylenevinylene] (MPS-PPV), dioleoylphosphatidylcholine (DOPC), dipalmitoylphosphatidylcholine (DPPC), F€orster resonance energy transfer (FRET), dynamic light scattering (DLS), 1,10 -dioctadecyl-3,3,30 ,30 -tetramethylindodicarbocyanine perchlorate (DiD), and methyl viologen (MV2þ) Acknowledgment. G.C. is grateful to McGill University, the Natural Sciences and Engineering Research Council of Canada, the Canada Foundation for Innovation New Opportunities Fund, the Fonds Quebecois de la Recherche sur la Nature et les Technologies-Nouveaux Chercheur Program, and the Centre for Self-Assembled Chemical Structures for financial assistance. A.T.N. is also thankful to the McGill Chemical Biology Fellowship Program (CIHR) and Fonds Quebecois de la Recherche sur la Nature et les Technologies for postgraduate scholarships.

Langmuir 2010, 26(9), 6746–6754