Confocal Raman Microscopy Probing of Temperature-Controlled

Oct 8, 2012 - Rong Cai , Sofiene Abdellaoui , Jay P. Kitt , Cullen Irvine , Joel M. Harris , Shelley D. Minteer , and Carol Korzeniewski. Analytical C...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/ac

Confocal Raman Microscopy Probing of Temperature-Controlled Release from Individual, Optically-Trapped Phospholipid Vesicles Jonathan J. Schaefer, Chaoxiong Ma, and Joel M. Harris* Department of Chemistry, University of Utah, 315 South 1400 East, Salt Lake City, Utah 84112-0850, United States S Supporting Information *

ABSTRACT: Control of permeability of phospholipid vesicle (liposome) membranes is critical to their applications in analytical sensing, in fundamental studies of chemistry in small volumes, and in encapsulation and release of payloads for site-directed drug delivery. Applications of liposome formulations in drug delivery often take advantage of the enhanced permeability of phospholipid membranes at their gel-to-fluid phase transition, where the release of encapsulated molecules can be initiated by an increase in temperature. Despite numerous successful liposome formulations for encapsulation and release methods to study the kinetics, this process has been limited to investigations of bulk vesicle dispersions, which provide little or no information about the vesicle membrane structure and its relationship to the kinetics of transmembrane transport. In this work, confocal Raman microscopy is adapted to study temperature-dependent release of a model compound, 3-nitrobenzene sulfonate (3-NBS), from individual optically trapped phospholipid vesicles, while simultaneously monitoring structural changes in the vesicle membrane reported by vibrational modes of phospholipid acyl chains and the local environment of the encapsulated compound. The confocal geometry allows efficient excitation and collection of Raman scattering from a single vesicle, while optical trapping allows more than hour-long observations of the same vesicle. With window factor analysis to resolve component spectra, temperature-controlled release of 3-NBS through vesicle membranes composed of pure 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) was measured and compared to transport through a lysolipidcontaining membrane specifically formulated for efficient drug delivery.

P

mismatch in lipid packing and bilayer thickness that leads to greater contact between the acyl chains and water leading to an increase in membrane permeability.18 This phenomenon was first suggested as a means of targeting the delivery of liposomeencapsulated drugs by Yatvin et al.,20 where the local temperature of a specific site could be elevated above the membrane phase transition temperature to maximize sitedirected drug release. Building on this idea, Needham et al.21 developed a temperature-sensitive membrane composition that contained a small fraction of single acyl-chain lysolipid, which provided a more effective thermal release of liposomeencapsulated drugs in treating tumors in mice.22 The improvement in efficacy over membrane preparations lacking lysolipid was attributed to faster and greater drug release at the tumor site. Despite clinical and therapeutic successes of liposomeencapsulated drug formulations, methods to study their temperature-controlled release have been limited to studies of bulk dispersions of vesicles in large volume samples. These methods rely on changes in fluorescence,22,23 light absorption,24 or spin-relaxation times25 to determine the fraction of encapsulated molecules that are released. Little information

hospholipid vesicles or liposomes are attracting considerable interest as models for biological cells,1 as a means of isolating single biomolecules for studies of conformational dynamics,2−4 as containers to study the reactivity of encapsulated enzymes5,6 or other chemical reactions in small volumes, which can be initiated by vesicle fusion7,8 or through control of membrane permeability at a thermal phase transition.9 Liposomes have been used as sensing structures in chemical analysis,10 where, for example, the interior volume encapsulates a dye that can be released into solution upon analyte binding.11 Lipid vesicle formulations are also employed as biocompatible vehicles for drug delivery,12−15 where encapsulation of cytotoxic drugs prevents damage to healthy tissue, allowing therapeutic concentrations to be delivered with minimal side effects. Liposome encapsulation is also being developed for nonmedical uses including consumer products, agricultural formulations, and fermentation processes.16 In these applications, the controlled encapsulation and release of molecules is critical to the fundamental process under investigation or the targeted application of the liposome preparation. Enhanced permeability of phospholipid vesicle membranes at their main phase transition temperatures (Tm) was first reported by Papahadjopoulos et al.17 Lipid membranes become permeable near their melting transition due to the formation of grain boundaries between gel and fluid phases in the lipid bilayer.18,19 Between the gel and fluid phases there is a © 2012 American Chemical Society

Received: August 15, 2012 Accepted: October 8, 2012 Published: October 8, 2012 9505

dx.doi.org/10.1021/ac302346n | Anal. Chem. 2012, 84, 9505−9512

Analytical Chemistry

Article

distilled and filtered with a Barnstead (Boston, MA) NANOpure II with a minimum resistivity of 18 MΩ·cm. Vesicle Preparation. For pure DPPC liposomes, approximately 1 mg of DPPC (Tm = 41 °C) was pipetted from a chloroform solution and dried under a stream of nitrogen in a glass vial. The glass vials were placed under mild vacuum (150 mTorr) for 3 h to remove any traces of solvent. Lipid sample vials were sealed and stored at −15 °C until used. Lysolipidcontaining, temperature-sensitive liposomes were prepared in a similar manner using a lipid mixture of DPPC, MSPC, and DSPE-PEG2000 in a molar ratio of 90:10:4. For either composition, lipid films were hydrated in 100 mM phosphate buffer (pH 7.2) containing 50 mM 3-NBS to produce a lipid concentration of 1 mg/mL. The lipid dispersion was hydrated at ∼51 °C, well above the lipid phase transition temperature, 42 °C, for at least 30 min; a 200 μL aliquot of the rehydrated solution was passed 11 times through a polycarbonate membrane (1.0 μm pores) also at 51 °C. The vesicle solution was diluted 100-fold by transferring 15 μL of the vesicle dispersion into 1.485 mL of 100 mM phosphate buffer (pH 7.2) containing 50 mM NaCl to balance osmotic pressure. A 70 μL aliquot of diluted vesicle dispersion was transferred to an aluminum well cell and heated to initiate release. Characterizing Vesicle Lamellarity. To test lipid vesicles for their average lamellarity, lipid films of pure DPPC and of lysolipid-containing DPPC mixtures were prepared with the addition of 0.5 mol % of a fluorescent-labeled lipid, NBD-PE. The lipids were rehydrated in 100 mM phosphate buffer without added 3-NBS and then extruded as above. After extrusion, 50 μL of 1 mg/mL vesicle dispersions were diluted into 2.95 mL of 100 mM phosphate buffer (pH = 7.2) in a 3.5 mL quartz cuvette. Fluorescence measurements were made on the labeled vesicle dispersions using a Hitachi F-7000 fluorescence spectrophotometer (Tokyo Japan), while the outer leaflet label was quenched in situ with a reducing agent, according to a previously published procedure.34 Specifically, the diluted vesicle solution was excited with 470 nm radiation, while fluorescence emission is monitored at 540 nm. Sodium dithionite quencher (150 μL of 1 M Na2S2O4 in 1 M Tris pH 10) was then added to the cuvette and the loss of fluorescence intensity due to reduction of the NBD label was measured at 540 nm for 3 min. The relative loss of fluorescence emission reports the fraction of lipid present in the outer bilayer leaflet, which can be related to the average lamellarity of the vesicles in the sample.34 Confocal Raman Microscopy. The confocal Raman microscope has been described in detail in a previous publication.35 Sample temperature control was achieved using an aluminum well cell constructed in house with a 4 mm cylindrical opening at the top, which transitions to a 1 mm opening at the bottom. A No. 1 glass coverslip (150 μm thickness) was attached to the bottom of the sample cell using 5 min epoxy. The aluminum sample cell was covered by a copper block, mounted on a temperature-controlled stage consisting of a silver base with Peltier stacks a DC supply (Technical Video).27 Temperature was measured in the base of the sample cell using a thermocouple (model 52 II, Fluke). Optically trapped vesicles within the aluminum well cell were subjected to ∼3 °C/min ramp until the membrane phase transition was reached, which was determined by observing structural changes in the temperature-dependent Raman spectra.27,28

about the structure of the lipid bilayer or mechanism of release is obtained, and the measurement of release kinetics is limited by slow heat transfer into large volume samples. Opticaltrapping confocal Raman microscopy26 is a powerful tool to investigate temperature-dependent structural changes in the lipid membranes of individual phospholipid vesicles.27 From the variation in Raman spectra versus temperature, structural changes in the phospholipid bilayer can be resolved through the phase transitions of the membrane.27,28 It is also possible to exploit the small sampling volume of confocal Raman microscopy to determine the permeability of a vesicle membrane for uptake29 or release30 of molecules into or out of the vesicle interior or the association of molecules with the vesicle membrane.31,32 The combination of these capabilities, observing temperature-dependent structural changes while monitoring the escape of encapsulated species, makes opticaltrapping confocal Raman microscopy well suited to investigating the kinetics and mechanism of temperature-controlled release from liposomes. In this work, optical-trapping confocal Raman microscopy is employed for the first time to study temperature-dependent structural changes in membranes of individual liposomes, while monitoring release of an encapsulated model compound, 3nitrobenzene sulfonate (3-NBS). Optical trapping of an individual vesicle by the focused laser beam allows hour-long observations of changes in local composition and structure of the vesicle within a small confocal volume of 0.6 fL.33 Within this volume, the local concentrations of lipid and encapsulated 3-NBS in a single vesicle are high and readily detected, while the overall sample is a very dilute vesicle dispersion (10 μg/mL lipid), which prevents aggregation and other interactions between vesicles during the experiment. Vesicles are observed in a small-diameter, aluminum cell that provides efficient heat transfer to the solution and control of the sample temperature. The structure of the lipid bilayer and along with temperaturedependent release of 3-NBS are simultaneously measured using Raman spectroscopy and resolved by window factor analysis, revealing changes both in the membrane that accompany an increase in permeability and in the local environment of the encapsulated 3-NBS. The temperature-controlled release of 3NBS through vesicle membranes composed of pure 1,2dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) is compared to release through lysolipid-containing vesicle membranes, in order to elucidate differences in the mechanism of release between these two membrane compositions.

■

EXPERIMENTAL SECTION Reagents and Materials. The following phospholipids were purchased from Avanti Polar Lipids (Alabaster, AL): 1,2 dipalmitoyl- sn -glycerol-3-phosphocholine (DPPC); 1-stearoyl2-hydroxy- sn -glycero-3-phosphocholine (MSPC); 1,2 distearoyl- sn-glycero-3-phosphoethanolamine-N-[methoxy(polyethylene glycol)-2000] (DSPE-PEG2000); and N-(7nitrobenz-2-oxa-1,3-diazol-4-yl)-1,2-dihexadecanoyl-sn-glycero3-phosphoethanolamine (NBD-PE). Lipids were dissolved in chloroform, stored at −15 °C and used without further purification. A vesicle extruder (Avanti Polar Lipids) was equipped with polycarbonate membranes having 0.6 and 1.0 μm pore sizes from Nucleopore (Pleasanton, CA). Sodium phosphate and sodium chloride were supplied by Mallinckrodt (Paris, KY). 3-Nitrobenzene sulfonate (98%) was acquired from Sigma-Aldrich (St. Louis, MO) and used without further purification. Water used in these experiments was quartz 9506

dx.doi.org/10.1021/ac302346n | Anal. Chem. 2012, 84, 9505−9512

Analytical Chemistry

Article

Individual vesicles were optically trapped with ∼25 mW of excitation radiation at the focus, and Raman scattered light from the vesicle was integrated for 60 s on the CCD. Raman spectra were background and white-light corrected in MATLAB (Math Works) using programs described in a previous publication.30 Baseline fitting was performed using a seventh order polynomial fit in each of four sections while excluding the peaks.36 To correct for variations in light collection efficiency and motions of the vesicle within the optical trap as the sample was heated, Raman spectra of a single vesicle were normalized to the C−N symmetric stretch of the phospholipid headgroup at 718 cm−1, which remains unchanged through the lipid phase transition.37,38 The relative concentration of 3-NBS in a vesicle was determined by integrating the scattering intensity of the NO2 symmetric stretch over the range 1352−1359 cm−1. Data Analysis. Raman spectra were decomposed into component contributions that evolved with time (and temperature) using a combination of evolving factor analysis (EFA)39,40 and window factor analysis (WFA).41,42 Details of the mathematical steps are found in Supporting Information. Computations were performed in MATLAB (Mathworks) using the factor analysis toolbox, version 2.10, based on Malinowski’s algorithms.43 Uncertainties in the reported composition versus time vectors, C, were determined from the variance-covariance matrix, VC, which predicts the error of mapping the resolved spectra, A, back onto the original data, D, having a measurement variance σ2, where VC = (ATA)−1 σ2.44

■

RESULTS AND DISCUSSION Temperature-Controlled Release from Pure DPPC Vesicles. To investigate the structural changes associated with temperature-dependent release of a tracer through a vesicle membrane of pure phospholipid, individual 1.0 μm DPPC vesicles were optically trapped using ∼25 mW laser power at 647.1 nm. The vesicle was trapped above a glass coverslip inside an aluminum cell on a temperature controlled stage, while Raman scattered light was collected through the same objective used to focus the laser beam. As a control sample, Raman spectra of a DPPC vesicle without a tracer were acquired as the temperature is raised from 23 to 46 °C; significant changes in the Raman spectra are observed (Figure 1A) consistent with disordering of the acyl chains,28,45 as the membrane transitions from the gel to fluid phase. At the melting transition, lipid membranes become permeable to species that are otherwise membrane-impermeable because of defects formed at the boundaries between gel- and fluid-phase domains.17−22 A Raman active tracer used to explore this phenomenon was 3-nitrobenzene sulfonate (3-NBS), which exhibits strong scattering from a trigonal ring breathing mode at 998 cm−1 and NO2 symmetric stretch at 1359 cm−1. In addition to exhibiting identifiable Raman scattering, the sulfonate group in 3-NBS is negatively charged over a wide range of solution pH, so that it remains encapsulated indefinitely at temperatures well below the main phase transition temperature. DPPC vesicles were prepared with 50 mM 3-NBS in the aqueous interior and diluted 100-fold into a buffer free of 3-NBS; a single, optically trapped vesicle shows the addition of wellresolved Raman scattering from 3-NBS (see Figure 1B). Temperature-controlled release from optically trapped vesicles is initiated by increasing the temperature of the sample from 23 to 46 °C and measuring release of 3-NBS from scattering of the NO2 symmetric stretch at 1359 cm−1 while simultaneously monitoring the state of the lipid bilayer through

Figure 1. Raman spectra of an optically trapped DPPC vesicle (A) comparing 23 °C (black line) and 45 °C (red line) with peak assignments showing structural changes occurring over the main phase transition and (B) encapsulating a 50 mM solution of 3-NBS, including peak assignments.

observation of C−C and C−H vibrations of the lipid acyl chains (Figure S-1, Supporting Information). Factor analysis techniques are used to resolve temperature-dependent changes in acyl chain order to compare with release of 3-NBS from the vesicle. The relevant frequency regions (the C−C stretch, CH2 twist, CH2 bend, and the C−H stretch) in the DPPC Raman spectra for membrane structure analysis are indicated in Figure S-1. These extracted sections of the time-dependent spectra are collected into a data matrix that is tested by factor analysis to determine the number of principal components based on an Ftest of the reduced eigenvalues.43 The number of significant components found was two, consistent with the DPPC vesicle transitioning from the gel to fluid phase. To extract pure spectra from the two-component data matrix, evolving factor analysis,39,40 is used to determine the time window in which each component exists along the time evolution of the experiment. By removing spectra that contain one of the two components from the data matrix, the influence of that component and its time-dependence can be discovered in the overall data set using window factor analysis.41−43 This process is carried out for both components to determine the time-dependent concentrations of gel- and fluid-lipid phases plotted in Figure 2. The corresponding spectra are determined 9507

dx.doi.org/10.1021/ac302346n | Anal. Chem. 2012, 84, 9505−9512

Analytical Chemistry

Article

Figure 3. Resolved Raman spectra of the gel-phase (black line) and fluid-phase (red line) of the DPPC membrane, with resolved 3-NBS in the interior vesicle solution (black line) and associated with the membrane (red line) shown as an inset.

CH2 bending region, where a change in peak area at 1455 cm−1 is also consistent with chain decoupling.28 Significant conformational disorder is also evident in the CH2 stretching region at 2847 and 2880 cm−1, where the shift to higher frequencies at 2849 and 2882 cm−1 and a loss in intensity from the 2880 cm−1 vibration consistent with chain decoupling.45 The time-evolution of the fractions of gel and fluid phases are plotted in Figure 2B, as the sample exceeds the melting transition temperature. The phase transition event is rapid, with most of the evolution in the component fractions occurring within 120 s. The time-dependence of structural changes in the DPPC membrane can be compared to the kinetics of release of 3-NBS from the vesicle interior, where the 3-NBS data are subjected to the same factor-analysis technique. In the inset to Figure S-1, the NO2 symmetric stretching region of 3-NBS is isolated for analysis and principal component analysis showed the presence of two time-varying components determined by an F-test of the reduced eigenvalues.43 Resolution of time-varying components by window-factor analysis revealed two populations of 3-NBS, one exhibiting an NO2 symmetric stretching frequency at 1359 cm−1 (the same as 3-NBS in aqueous solution) and a second population shifted to 1352 cm−1 (see Figure 3). The population exhibiting higher-frequency NO2 stretching exists at temperatures below the phase transition (see Figure 2C) and converts to the lower NO2-frequency population at a rate comparable to the passage of the lipid membrane through the phase transition. The 3-NBS molecules exhibiting the lower NO2 frequency then remain associated with the phospholipid vesicle for over 30 min, as 3-NBS is slowly released. This behavior of 3-NBS in the vesicle, as it is heated through the melting transition, is likely due to 3-NBS partitioning into vesicle membrane, from which it is slowly released into the external solution. It is known that polar and ionic molecules exhibit low solubility in solid-ordered (gel-phase) lipid membranes, but that they become more soluble and membrane-permeable above the melting transition in the liquid-disordered or fluid phase.50,51 This observation suggests a mechanism for the release kinetics where, at the melting transition, 3-NBS first associates with the vesicle membrane and then slowly permeates the bilayer to be released into solution. A simple system, to test whether lowering of the NO2

Figure 2. Slow release of 3-NBS from a DPPC vesicle. (A) The timedependent temperature ramp and corresponding relative concentrations of (B) the gel (black line) and fluid (red line) phases of a DPPC vesicle, encapsulating C) 3-NBS in the interior free solution (black line), which becomes membrane associated (red line) before exiting the vesicle. The maxima in the relative concentration profiles are scaled to unity. The uncertainties in the relative concentrations for the gel and fluid phases of DPPC are 3% and for free solution and membrane associated 3-NBS are 6%.

from the data matrix by linear-least-squares. The resulting resolved spectra of the gel and fluid phases are plotted in Figure 3 and show significant structural changes through the phase transition. As the vesicle passes through the phase transition (Tm = 41.6 °C),28,46 structural changes in C−C stretching region (Figure 3) are apparent in the decrease in scattering intensity from outof-phase and in-phase trans-conformers at 1065 and 1126 cm−1,47 as the acyl chains become disordered in the fluid phase.27,28,38 Gauche conformations in the acyl chains show increasing C−C stretching intensity at 1080 cm−1,48 exceeding the intensity of trans-conformers above the phase transition. Because trans chain sections shorter than eight C−C bonds do not produce significant Raman scattering at 1065 and 1126 cm−1,49 the decrease in trans-mode intensity is likely due to gauche conformations occurring near the middle of the acyl chains. Gauche conformations are also responsible for a decrease in intensity of CH2 twisting at 1296 cm−1, which broadens and shifts to higher frequency (1302 cm−1), indicating less chain coupling and freedom of twisting motions.45 Decreased interchain interactions are also apparent in the 9508

dx.doi.org/10.1021/ac302346n | Anal. Chem. 2012, 84, 9505−9512

Analytical Chemistry

Article

stretching frequency is consistent with 3-NBS interacting with the amphiphilic lipid bilayer, is 3-NBS interacting with another amphiphilic structure, cetylpyridinium chloride (CPC) in 10 mM solution or ∼10-fold higher than its critical micelle concentration.52 The results (Supporting Information) show an equivalent shift of the NO2 symmetric stretch to 1352 cm−1 when 3-NBS associates with CPC micelles. The lower NO2 symmetric stretching frequency is thus consistent with transfer of 3-NBS into the amphiphilic environment of the phospholipid bilayer from which it is slowly released. The kinetics of temperature-induced release of 3-NBS were measured for a series of different DPPC vesicles (see Supporting Information). While the results consistently show a loss of 3-NBS after reaching the membrane melting transition, the rate of release varies from vesicle to vesicle. Release kinetics cluster into slower and faster rates, where the flux of 3-NBS through vesicle membrane ranges between 6.4 (±0.6) × 10−9 and 2.2 (±0.4) × 10−8 mol/sec·m2. To investigate the source of this variation, data from both slow and fast release were subjected to window factor analysis to compare component spectra and time variations. When comparing the previous example (Figures 2 and 3) from a slowly releasing vesicle, the time evolution of the lipid melting transition for the fasterreleasing vesicle is indistinguishable (see Figure 4). The corresponding resolved spectra from a faster releasing example also show equivalent spectra for the gel- and fluid-lipid phases and for the free-solution and membrane-associated 3-NBS components (see Figure S-4 in Supporting Information). The time-dependence of the 3-NBS, however, shows a somewhat faster evolution of the free-solution to membrane-associated population and a much faster escape of membrane-associated 3NBS from vesicle as shown in Figure 4 (compare with Figure 2). The variation rates of release could arise from differences in the density of defects in the membrane and corresponding rates of release; this hypothesis, however, does not explain the clustering of release rates into slow and fast groups. A more consistent explanation of both the variation in release rates and their clustering would be differences in vesicle lamellarity. While small (≤100 nm) vesicles formed by extrusion are predominantly unilamellar (single lipid bilayer),53 larger extruded vesicles exhibit a greater fraction of multilamellar structures.34 A test of the average lamellarity of the DPPC vesicles in this study was performed by quenching fluorescence of NBD labels on the outer bilayer leaflet of vesicles in dispersion. The results are consistent with an average lamellarity slightly less than two bilayers (see Supporting Information). Based on a recent fluorescence microscopy study of the size and lamellarity of individual DPPC vesicles,54 the DPPC vesicles in this sample should represent a distribution of lamellarity, bounded by unilamellar vesicles at the lower limit and concentric multilamellar vesicles at the upper limit. The escape of 3-NBS from DPPC vesicles shows a pattern consistent with such a distribution, where fastest release would arise from unilamellar vesicles with a single-bilayer separating the vesicle contents and the external solution. Slow release would correspond to migration through a multiple concentric bilayers, where an additional step of transfer into and out of the solution between the bilayers would account for the slower conversion of solution-phase to membraneassociated 3-NBS (compare Figures 2C and 4C). Intermediate release rates would likely correspond to oligolamellar vesicles,54−56 where smaller vesicles are encapsulated within a

Figure 4. Fast release of 3-NBS from a pure DPPC vesicle. (A) The time-dependent temperature ramp and corresponding relative concentrations of (B) the gel (black line) and fluid (red line) phases of a DPPC vesicle, encapsulating (C) 3-NBS in the interior free solution (black line), which becomes membrane associated (red line) before exiting the vesicle. The maxima in the relative concentration profiles were scaled to unity. The uncertainties in the relative concentrations for the gel and fluid phases of DPPC are 2% and for free solution and membrane-associated 3-NBS are 7%.

larger structure, where the surface area-to-volume ratio of the smaller interior vesicles is large and would generate a smaller kinetic barrier to membrane transfer and release. Temperature-Controlled Release from Mixed-Lipid, Temperature-Sensitive Liposomes. A liposome membrane formulation developed for temperature-controlled release21 contains a fraction of lysolipid (one acyl chain instead of two to generate pores or channels) and a PEG-chain phospholipid to improve vesicle stability; this liposome membrane composition produced faster and more consistent release than singlecomponent phospholipid membranes.22 At the melting phase transition, these temperature-sensitive liposomes were thought to release their contents near the melting phase transition through the formation of pores in the membrane.23 Vesicles composed with this temperature-sensitive mixture21 of DPPC, MSPC, and DSPE-PEG2000 in a molar ratio of 90:10:4 (see structures in Supporting Information) were prepared in an identical manner as the DPPC vesicles above. Individual 9509

dx.doi.org/10.1021/ac302346n | Anal. Chem. 2012, 84, 9505−9512

Analytical Chemistry

Article

vesicles were optically trapped, and temperature-controlled release of 3-NBS was investigated, along with the lipid membrane melting transition. The resulting Raman spectra of the lysolipid-containing liposomes are nearly identical to DPPC vesicles; the relevant C−C stretch, CH2 twist, CH2 bend, and C−H stretching regions of the temperature-dependent Raman spectra were isolated for membrane structure analysis; see Figure 5. These extracted sections were collected into a data

Figure 5. Resolved Raman spectra of the gel (black line) and fluid (red line) phases of a mixed lipid vesicle. The single-component spectrum of 3-NBS (inset) is consistent with its presence in an aqueous solution.

matrix and analyzed by factor analysis to determine the number of principal components, which was again two, consistent with a single transition from a gel to fluid phase. Window factor analysis was used to determine the time-dependent concentrations of gel and fluid phases, as plotted in Figure 6; despite the presence of lysolipid and PEG-chain DSPE in the membrane, the time- and temperature-dependent evolution of the gel and fluid phases are indistinguishable from pure DPPC (compare Figures 2B and 6B). The resolved Raman spectra of the two gel and fluid phases in the C−C stretch, CH2 twist, CH2 bend, and C−H stretching regions are also equivalent to DPPC (compare Figures 3 and 5). Although the time- and temperature-dependent acyl-chain structure of the mixed lipid membrane cannot be distinguished from that of pure DPPC by Raman spectroscopy, the kinetics of 3-NBS release differ significantly. The Raman scattering from the NO2 symmetric stretching mode (see Figure S-7) shows no frequency shift with increasing temperature; indeed, factor analysis indicates only a single time-varying component in the Raman spectra. The NO2 symmetric stretching frequency of this single component indicates 3-NBS remains in aqueous solution and does not detectably associate with the vesicle membrane, even after the main phase transition of the mixed lipid vesicle is reached. This implies that escape of 3-NBS does not occur by diffusion through an amphiphilic membrane environment as in pure DPPC vesicles (see above), but through pores or channels that preserve an aqueous solution environment, as suggested in previous discussions of the effect of lysolipid on vesicle release.23 At the melting transition temperature, a rapid loss of 3-NBS is observed, the time dependence of which is plotted in Figure 6C. With the mixed-lipid vesicles, a consistently fast release of 3-NBS was observed where nearly all of the 3-NBS is released in less than 120 s, as fast as the lipid phase transition. This release of 3-NBS from the mixed lipid vesicles is very

Figure 6. Release of 3-NBS from a mixed-lipid vesicle. (A) The timedependent temperature ramp and corresponding relative concentrations of (B) the gel (black line) and fluid (red line) phases of the optically trapped lysolipid-containing vesicle membrane, encapsulating (C) 3-NBS in the interior solution, which escapes without detectable association with the vesicle membrane. Maxima in the relative concentration profiles were scaled to unity. Uncertainties in relative concentrations for gel and fluid phases of the mixed-lipid vesicle are 4% and 3% for the free solution 3-NBS.

reproducible (see Figure S-8, Supporting Information), much more so than from pure DPPC vesicles. The temperaturesensitive mixed vesicles were tested for lamellarity by quenching of the fluorescence of NBD labels on their outer leaflet and found to be indistinguishable from a unilamellar population (Supporting Information), which is consistent with their uniform release kinetics. From the rate of 3-NBS escape, the average initial flux through the mixed-lipid vesicle membrane is 8.0 (±0.6) × 10−8 mol/sec·m2, which is nearly 4× greater than the fastest flux measured through unilamellar DPPC membranes. This result, together with the Raman spectroscopic data showing 3-NBS remains in aqueous solution, suggests that the mixed-lipid membrane forms water-filled channels or pores at the phase transition temperature; these pores lead to a more rapid loss of the vesicle payload than diffusion through the lipid bilayer. If channels through which 3-NBS passes exhibit water-like diffusional pathways, then the measured flux of 3-NBS can be used with Fick’s first law to estimate the total area of such 9510

dx.doi.org/10.1021/ac302346n | Anal. Chem. 2012, 84, 9505−9512

Analytical Chemistry

■

water-filled pores. Using an estimated free diffusion coefficient for 3-NBS in water from the Stokes−Einstein equation, D ∼ 7 × 10−6 cm2/sec, the total pore area in the vesicle membrane responsible for the measured flux would be 3 × 10−16 cm2, assuming that 3-NBS can freely diffuse in water through the pores. This predicted pore area is, however, unrealistically small and equivalent to a single pore with a diameter of only 0.2 nm, which is insufficient to pass a molecule even as small as 3-NBS. Thus, while the spectroscopic results indicate that 3-NBS remains in a water-like environment during its escape, the insufficient pore size suggests that 3-NBS does not pass unhindered through water-filled pores. A possible explanation for hindrance is that water within the pores could be ordered, significantly slowing the diffusion of 3-NBS. Indeed, molecular dynamics simulations suggest that the diffusivity of water decreases inside of nanometer sized pores.57,58 Another possibility for hindered diffusion within water-filled pores is that the pore channels are transient in nature,59,60 so that the passage of molecules is limited by a small fraction of time that pores are open.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research was supported in part by the National Science Foundation under Grant CHE-0957242.

■

REFERENCES

(1) Torchilin, V.; Weissig, V. Liposomes, A Practical Approach, 2nd ed.; Oxford University Press: Oxford, 2003. (2) Boukobza, E.; Sonnenfeld, A.; Haran, G. J. Phys. Chem. B 2001, 105, 12165. (3) Okumus, B.; Wilson, T. J.; Lilley, D. M.; Ha, T. Biophys. J. 2004, 87, 2798. (4) Rhoades, E.; Cohen, M.; Schuler, B.; Haran, G. J. Am. Chem. Soc. 2004, 126, 14686. (5) Chiu, D. T.; Wilson, C. F.; Karlsson, A.; Danielsson, A.; Lundqvist, A.; Strömberg, A.; Ryttsén, F.; Davidson, M.; Nordholm, S.; Orwar, O.; Zare, R. N. Chem. Phys. 1999, 247, 133. (6) Cheng, Z.; D’Ambruos, G. D.; Aspinwall, C. A. Langmuir 2006, 22, 9507. (7) Chiu, D. T.; Wilson, C. F.; Ryttsen, F.; Stromberg, A.; Farre, C.; Karlsson, A.; Nordholm, S.; Gaggar, A.; Modi, B. P.; Moscho, A.; Garza-Lopez, R. A.; Orwar, O.; Zare, R. N. Science 1999, 283, 1892. (8) Kulin, S.; Kishore, R.; Helmerson, K.; Locascio, L. Langmuir 2003, 19, 8206. (9) Bolinger, P.-Y.; Stamou, D.; Vogel, H. J. Am. Chem. Soc. 2004, 126, 8594. (10) Edwards, K. A.; Baeumner, A. J. Talanta 2006, 68, 1421. (11) Ishimori, Y.; Rokugawa, K. Clin. Chem. 1993, 39, 1439. (12) Drummond, D. C.; Zignani, M.; Leroux, J.-C. Prog. Lipid Res. 2000, 39, 409. (13) Medina, O. P.; Zhu, Y.; Kairemo, K. Curr. Pharm. Des. 2004, 10, 2981. (14) Wang, G. In Drug Delivery, Principles and Applications; Wang, B., Siahaan, T. J., Soltero, R. A., Eds.; John Wiley: Hoboken, NJ, 2005. (15) Samad, A.; Sultana, Y.; Aqil, M. Curr. Drug Delivery 2007, 4, 297. (16) Barenholz, Y.; Lasic, D. D. Handbook of Nonmedical Applications of Liposomes; CRC Press: Baco Raton, FL, 1996. (17) Papahadjopoulos, D.; Jacobson, K.; Nir, S.; Isac, I. Biochim. Biophys. Acta, Biomembr. 1973, 311, 330. (18) Marsh, D.; Watts, A.; Knowles, P. F. Biochemistry 1976, 15, 3570. (19) Kim, D. H.; Costello, M. J.; Duncan, P. B.; Needham, D. Langmuir 2003, 19, 8455. (20) Yatvin, M. B.; Weinstein, J. N.; Dennis, W. H.; Blumenthal, R. Science 1978, 202, 1290. (21) Needham, D.; Anyarambhatla, G.; Kong, G.; Dewhirst, M. W. Cancer Res. 2000, 60, 1197. (22) Kong, G.; Anyarambhatla, G.; Petros, W. P.; Braun, R. D.; Colvin, O. M.; Needham, D.; Dewhirst, M. W. Cancer Res. 2000, 60, 6950. (23) Mills, J. K.; Needham, D. Biochim. Biophys. Acta, Biomembr. 2005, 1716, 77. (24) Mills, J. K.; Needham, D. Biochim. Biophys. Acta, Biomembr. 2005, 1716, 77. (25) Ponce, A. M.; Viglianti, B. L.; Yu, D.; Yarmolenko, P. S.; Michelich, C. R.; Woo, J.; Bally, M. B.; Dewhirst, M. W. J. Natl. Cancer Inst. 2007, 99, 53. (26) Cherney, D. P.; Harris, J. M. Annu. Rev. Anal. Chem. 2010, 3, 277. (27) Fox, C. B.; Myers, G. A.; Harris, J. M. Appl. Spectrosc. 2007, 61, 465.

■

CONCLUSIONS Optical-trapping confocal Raman microscopy was found to be an informative tool for studying the release of observing encapsulated molecules from phospholipid vesicles. The method reports structural changes in vesicle membranes that accompany the release, as well as the microenvironment of the encapsulated molecules during the process. Using this technique, it was discovered that pure DPPC liposomes exhibited two-step release of 3-NBS at the main phase transition. The NO2 symmetric stretching frequency revealed a change in environment that corresponds to the association of 3-NBS with the vesicle membrane at the main lipid phase transition temperature followed by slow release. The variation in the release rates is consistent with the distribution of lamellarity of 1 μm DPPC vesicles prepared by extrusion. In contrast, release of 3-NBS from lysolipid-containing, temperature sensitive liposomes was 4× faster than from DPPC vesicles, with no detectable variation in rate from one vesicle to the next, which agreed with the vesicle population being unilamellar. The NO2 symmetric stretch of 3-NBS in the lysolipid-containing vesicles exhibited only a single time-varying component, indicating that 3-NBS does not associate with the vesicle membrane during its release. The aqueous solution environment of 3-NBS observed during its release from the mixed-lipid vesicles suggests that the membrane forms waterfilled pores through which the molecule escapes. The rate of release is slow, however, relative to diffusion through nanometer pores filled with liquid water, suggesting that transport may be hindered by ordering of water molecules in the pores or that pores are transient and open only a small fraction of time. Further applications of this methodology could be used to investigate the kinetics of analyte-responsive release from vesicles11 or phospholipase-catalyzed degradation of phospholipid membranes leading to controlled release targeted at inflammation or cancerous tissue.61

■

Article

ASSOCIATED CONTENT

S Supporting Information *

Additional information and figures. This material is available free of charge via the Internet at http://pubs.acs.org. 9511

dx.doi.org/10.1021/ac302346n | Anal. Chem. 2012, 84, 9505−9512

Analytical Chemistry

Article

(28) Fox, C. B.; Uibel, R. H.; Harris, J. M. J. Phys. Chem. B 2007, 111, 11428. (29) Myers, G. A.; Harris, J. M. Anal. Chem. (Washington, DC, U.S.) 2011, 83, 6098. (30) Cherney, D. P.; Conboy, J. C.; Harris, J. M. Anal. Chem. 2003, 75, 6621. (31) Fox, C. B.; Horton, R. A.; Harris, J. M. Anal. Chem. 2006, 78, 4918. (32) Fox, C. B.; Harris, J. M. J. Raman Spectrosc. 2010, 41, 498. (33) Bridges, T. E.; Houlne, M. P.; Harris, J. M. Anal. Chem. 2004, 76, 576. (34) McIntyre, J. C.; Sleight, R. G. Biochemistry 1991, 30, 11819. (35) Houlne, M. P.; Sjostrom, C. M.; Uibel, R. H.; Kleimeyer, J. A.; Harris, J. M. Anal. Chem. 2002, 74, 4311. (36) Schaefer, J. J.; Fox, C. B.; Harris, J. M. J. Raman Spectrosc. 2012, 43, 351. (37) Wong, P. T. T. Annu. Rev. Biophys. Bioeng. 1984, 13, 1. (38) Gaber, B. P.; Yager, P.; Peticolas, W. L. Biophys. J. 1978, 21, 161. (39) Gampp, H.; Maeder, M.; Meyer, C. J.; Zuberbuehler, A. D. Talanta 1985, 32, 1133. (40) Gemperline, P. J.; Hamilton, J. C. J. Chemom. 1989, 3, 455. (41) Den, W.; Malinowski, E. R. J. Chemom. 1993, 7, 89. (42) Malinowski, E. R. J. Chemom. 1992, 6, 29. (43) Malinowski, E. R. Factor Analysis in Chemistry, 3rd ed.; Wiley: New York, 2002. (44) Draper, N. R.; Smith, H. Applied Regression Analysis; John Wiley and Sons, 2003. (45) Orendorff, C. J.; Ducey, M. W.; Pemberton, J. E. J. Phys. Chem. A 2002, 106, 6991. (46) Koynova, R.; Caffrey, M. Biochim. Biophys. Acta 1998, 1376, 91. (47) Levin, I. W. Adv. Infrared Raman Spectrosc. 1984, 11, 1. (48) Meier, R. J.; Csiszar, A.; Klumpp, E. J. Phys. Chem. B 2006, 110, 5842. (49) Meier, R. J. Polymer 2002, 43, 517. (50) Heimburg, T.; Jackson, A. D. Biophys. J. 2007, 92, 3159. (51) Blicher, A.; Wodzinska, K.; Fidorra, M.; Winterhalter, M.; Heimburg, T. Biophys. J. 2009, 96, 4581. (52) Anacker, E. W. J. Phys. Chem. 1958, 62, 41. (53) Lohr, C.; Kunding, A. H.; Bhatia, V. K.; Stamou, D. Methods Enzymol. 2009, 465, 143. (54) Heider, E. C.; Barhoum, M.; Edwards, K.; Gericke, K.-H.; Harris, J. M. Anal. Chem. 2011, 83, 4909. (55) Mui, B.; Chow, L.; Hope, M. J. Methods Enzymol. 2003, 367, 3. (56) Bagatolli, L. A.; Parasassi, T.; Gratton, E. Chem. Phys. Lipids 2000, 105, 135. (57) Breed, J.; Sankararamakrishnan, R.; Kerr, I. D.; Sansom, M. S. Biophys. J. 1996, 70, 1643. (58) Hartnig, C.; Witschel, W.; Spohr, E. J. Phys. Chem. B 1998, 102, 1241. (59) Marrink, S. J.; Jähnig, F.; Berendsen, H. J. Biophys. J. 1996, 71, 632. (60) Leontiadou, H.; Mark, A. E.; Marrink, S. J. Biophys. J. 2004, 86, 2156. (61) Davidsen, J.; Jørgensen, K.; Andresen, T. L.; Mouritsen, O. G. Biochim. Biophys. Acta, Biomembr. 2003, 1609, 95.

9512

dx.doi.org/10.1021/ac302346n | Anal. Chem. 2012, 84, 9505−9512