| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* International Max Planck Research School for Molecular Biology, University of Goettingen, Goettingen, Germany; Experimental Biophysics Group, Max Planck Institute for Biophysical Chemistry, Goettingen, Germany; and Institute of Biophysics/Biotec, Dresden University of Technology, Dresden, Germany
Correspondence: Address reprint requests to Petra Schwille, Experimental Biophysics Group, Max Planck Institute for Biophysical Chemistry, Am Fassberg 11, 37077, Goettingen, Germany. Tel.: 49-351-463-40328; Fax: 49-351-463-40342; E-mail: pschwil{at}gwdg.de.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
1.5 x 10–8 cm2/s. Released ODN has the mean diffusion constant D = 1.1 ± 0.2 x 10–6 cm2/s, which signifies that ODN is dissociated from cationic lipids. In contrast to earlier studies, we report that phosphatidylethanolamine can trigger ODN release from the carrier in the full absence of anionic phosphatidylserine in the target membrane and that phosphatidylethanolamine-mediated release is as extensive as in the case of phosphatidylserine. The presented methodology provides an effective tool for probing a delivery potential of newly created lipid formulations of CL-ODN complexes for optimal design of carriers. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
; Scherer and Rossi, 2003; Marshall, 2000, 2002). Reduced immunogenicity and safety of cationic lipid-based delivery systems compared to viral vectors motivate significant efforts focused on overcoming the notoriously low efficiency of these carriers. A huge potential in adaptability of the biophysical parameters of the active lipid formulation offers promising prospects for creating an individually designed nonviral carrier, optimized to a specific purpose. Toward this goal, the structure, morphology, and functions of the cationic lipid-based complexes have been extensively investigated by various experimental and theoretical methodologies in an attempt to discover correlation between the physical and chemical properties of the complex and its delivery potential. In particular, it has been shown that internal nanostructure of the cationic lipid-DNA (CL-DNA) particle (Koltover et al., 1998, 2000), its total surface charge (Koltover et al., 1999; Xu et al., 1999) and lipid charge density (Lin et al., 2003), and the DNA accessibility to enzymatic degradation after assembling of the complex (Zhang et al., 1997; Ferrari et al., 2001) are important parameters that affect transfection capability of the carrier.
A current understanding of the cellular machinery of DNA/ODN delivery by the cationic complexes suggests that the particles are internalized via endocytosis (Zelphati and Szoka, 1996a; Zuhorn et al., 2002a,b; Zuhorn and Hoekstra, 2002), and cellular membrane lipids, primarily anionic phosphatidylserine (PS), mediate DNA/ODN dissociation from the cationic lipids and escape from endosomes (Zelphati and Szoka, 1996a,b; Nakanishi and Noguchi, 2001; Noguchi et al., 1998). Given that DNA/ODN should dissociate from the carrier and be released into the cell cytoplasm for carrying out their functions, comprehensive knowledge of the release mechanisms, which has not been attained yet, is essential for designing efficient carriers.
Although extensive research has been focused on the characterization of CL-DNA complexes, only a few studies have investigated biophysical and biochemical properties of cationic lipid-ODN (CL-ODN) particles and their interactions with cellular components (Jääskeläinen et al., 1998; Meidan et al., 2000; Zelphati and Szoka, 1996b; Zelphati and Szoka, 1997; Weisman et al., 2004). The similarities and differences between the DNA and ODN complexes are not well identified, providing no evidence that lipid formulation effective for DNA delivery will give the same results when used for ODN transport. Indeed, in their recent cryo-transmission electron microscopy and small angle x-ray scattering studies, Weisman and coauthors (2004) have shown that the internal nanostructure of CL-ODN complexes differs from the structure of CL-DNA particles. Specifically, for lipid compositions forming lamellar structure of the complexes, which represent one type of possible arrangement of these particles, a CL-DNA complex was shown to exhibit a monolayer lattice of parallel helixes, whereas a CL-ODN complex shows no ordering of intercalated ODN molecules, oriented fairly randomly between the lipid bilayers (Weisman et al., 2004). This finding indicates that the short, flexible ODN molecules presumably bind cationic lipids of the carrier differently than do the more rigidly structured DNA helices. The differences in binding of DNA and ODN within the complex might result in different behavior of ODN molecules while interacting with cellular membrane lipids, possibly leading to altering barriers for ODN release from the carrier.
In this work, we aim to gain deeper insights into the mechanisms controlling ODN dissociation from the carrier and release across endosomal membranes. Toward this end, we use laser scanning confocal microscopy (LSM) and fluorescence correlation spectroscopy (FCS) to characterize interaction between fluorescently labeled CL-ODN complexes and cellular membrane lipids, resulting in ODN release. As accurate manipulation of lipid membranes in intact cells is difficult to achieve, we use giant unilamellar vesicles (GUVs) to model cellular membranes and monitor ODN release across the target membranes of controllable lipid compositions. The efficiency of ODN release by specific membrane lipids is assessed via LSM analysis of labeled ODN transport inside the corresponding GUVs. Using FCS provides a tool for quantitative estimation of ODN dissociation from the cationic lipids upon the transfer across the target membrane and accumulation within the GUVs.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
; Dimitrov and Angelova, 1988). This technique has been proven to yield a high percentage of unilamellar vesicles with negligible lipid oxidation (Kahya et al., 2003; Scherfeld et al., 2003). GUVs were formed in water medium using a 400-µl flow chamber, which allowed for injecting solutions into the medium surrounding the GUVs after their formation. For GUV formation, an alternative voltage of 1.2-V amplitude and a 10-Hz frequency was applied for 1 h at 60°C. For visualization by LSM, 0.1 mol % of DiO or DiI lipid analog dye was added to each lipid mixture. For modeling the exoplasmic leaflet of the plasma membrane, GUVs were prepared using equimolar fractions of dioleoyl-phosphatidylcholine (DOPC), sphingomyelin, cholesterol, and ganglioside GM1, added in a small amount, similar to previous studies by others (Dietrich et al., 2001; Kahya et al., 2003). For the experiments, aimed to assess a role of aminophospholipids in the ODN release, GUVs were composed of PS, phosphatidylethanolamine (PE), and phosphatidylcholine (PC) in the molar fractions 0.15/0.25/0.6, respectively, as relevant to the phospholipid composition of endosomal membranes (Urade et al., 1988; Kobayashi et al., 2002). Although PE and PC lipids in these mixtures were always used as unsaturated species—dioleoyl-phosphatidylethanolamine (DOPE) and DOPC, respectively—PS was used both as dipalmitoyl-phosphatidylserine (DPPS) and as 1,2-dioleoyl-sn-glycero-3-phospho-L-serine (DOPS), which reproduces the distribution of saturated and unsaturated phospholipid molecular species in cells for these specific types of lipids (Blom et al., 2001; Urade et al., 1988). In addition, using PS in two different modifications allows for probing the influence of the acyl group on PS lipid activity in the ODN release. GUVs composed of variable amounts of DOPC, DPPS (DOPS), and/or DOPE were used for investigating the role of each of these cellular lipid species in the ODN release.
|
). DOTAP/DOPE lipid mixtures were prepared at 5 mM lipid concentration by dissolving the lipids in chloroform. Liposomes were formed by a hydration method similar to that described by others (Kahya et al., 2001). The final solution was stored at 4°C until use under argon atmosphere. For preparation of CL-ODN complexes, labeled ODNs (100 pmol for LSM experiments and 1 pmol for FCS measurements) were added to the unlabeled liposome solution at 2:1 cationic/anionic molar charge ratio and incubated for 20 min. After 20 min, the preparation was diluted to the total volume of 340 µl, injected into the GUV flow chamber, and allowed to diffuse evenly throughout the sample at room temperature.
Laser scanning confocal fluorescence microscopy
LSM imaging was performed using a Zeiss 510 Meta laser scanning microscope (Zeiss, Jena, Germany) with argon ion (488 nm, 30 mW, at 40% of maximum power output) and helium-neon (633 nm, 5.0 mW, used at full power) lasers. A water immersion objective 40x/1.2W (Zeiss) was used in the epidetection configuration with an adjustable pinhole set at 76 µm, corresponding to a <0.7-µm-thick confocal image slice of the sample. A band-pass filter transmitting 505–550 nm (Zeiss) was used in the experiments with DiO, while a long-pass filter at 650 nm (Zeiss) was applied to separate Cy-5 fluorescence.
Evaluation of ODN release using confocal LSM images
For evaluation of ODN release from the carriers upon interaction with the target membrane lipids, we analyzed the confocal images and estimated the number of GUVs filled with labeled ODNs inside their lumen. Once ODN molecules were transported inside the GUVs due to interaction of CL-ODN complexes with the membrane lipids, released ODNs were not able to cross the membrane bilayers in the opposite direction and remained accumulated inside GUVs for extensive periods of time. This allows for convenient comparison of the extent of ODN release for various compositions of the target membranes via counting a fraction of the ODN-filled GUVs in the corresponding set of the confocal images. Each GUV sample typically contained 1000 vesicles ranging in size from 10 to 100 µm, which provides a reliable basis for statistics. All GUV samples were evaluated after incubation with CL-ODN complexes for 2 h. Due to this, a bias in counting the fractions of filled GUVs, caused by possible vesicle-to-vesicle fusion events during the interaction with the complexes, was excluded. After 2 h incubation with CL-ODN complexes, the microscope objective was positioned on a randomly selected group of vesicles using 488 nm illumination. After that, a single snapshot image of the selected region was taken with the 633-nm laser line (5 mW power). The procedure was repeated several times to provide an image set (usually 15–20 images) containing 1000 vesicles for the given sample. Each experiment was repeated four times, corresponding to evaluation of 4000 vesicles in total for every studied lipid mixture. Thereafter, each set of images was processed by the following algorithm. First, a total number of vesicles in the image were calculated. Next, an intensity threshold was set at 100 ADU above the outside background to separate and count ODN-filled GUVs. The given threshold value was determined based on analysis of the fluorescent signals inside the GUVs for all images. The fraction of ODN-filled GUVs was determined for each membrane composition by calculating a percentage ratio of ODN-filled vesicles versus the total number of the vesicles in the set of images (as counted after the sample incubation with the complexes). Although not strictly exact for absolute evaluation of ODN transfer, this criterion allows for robust comparison of the relative efficiency of ODN release for different lipid compositions. We note that in control experiments with an equivalent amount of 100-pmol free labeled ODNs, some vesicles (<5%) were permeable to free ODNs. However, in contrast to the active ODN transport by CL-ODN complexes, a steep increase of ODN concentration inside the GUV lumen compared to the outside solution was never built up by free diffusion of ODN into leaky vesicles. For CL-ODN complexes used here, a fraction of free ODN molecules, present in the solution after formation of the complexes, was undetectable in LSM images. Therefore, the effect of the small proportion of leaky vesicles on quantifying GUVs containing high levels of accumulated ODNs was negligible because leaky vesicles were discarded during application of the intensity threshold.
Fluorescence correlation spectroscopy
FCS analysis was performed according to the procedure outlined in detail elsewhere (Rigler et al., 1993; Schwille et al., 1997; Haustein and Schwille, 2003). In FCS, experimentally obtained temporal correlations of fluorescence signal from minute molecule ensembles are evaluated by mathematical modeling and a numerical curve fitting provides the so-called diffusion time of molecules, 
d, from which the diffusion coefficient D can be calculated. Here, the data-fitting was performed using a three-dimensional Brownian diffusion model (Schwille et al., 1997; Haustein and Schwille, 2003):
 | (1) |
is time, d is the diffusion time, and z0 and 
0 determine effective focal volume size in axial and lateral directions, respectively. The last factor in Eq. 1 characterizes the photophysical dynamics of the used dye (Cy-5), with dark fractions T and characteristic flickering times T (Widengren et al., 1995). Veff is determined in calibration measurements. The diffusion coefficient is calculated using the relationship 
.
The FCS measurements were performed using a commercial microscope ConfoCor2 (Zeiss) and a home-built FCS setup. The commercial ConfoCor2 microscope was used in the configuration described elsewhere (Kahya et al., 2003). The home-built setup was assembled in the epifluorescence configuration using the inverted Olympus IX70 microscope (Olympus, Tokyo, Japan). A helium-neon laser (633 nm, 14mW; Research Electro-Optics, Boulder, CO) was coupled onto the back aperture of the 40x/1.2W water-immersed objective (UPlanApo, Olympus, Hamburg, Germany) by the dichroic mirror Q 645 LP (AHF, Tübingen, Germany). The fluorescence emission was filtered by a band-pass filter (AHF, model 675DF45) and projected onto a spatially adjustable 100-µm core diameter optical fiber (Thorlabs, Newton, NJ), which core dimension provides a pinhole. The light collected by the optical fiber is directed to an avalanche photodiode (SPCMCD3017, Perkin Elmer Optoelectronics, Vaudreuil, Quebec, Canada) and analyzed using a commercial correlator card (ALV, Langen, Germany).
In all FCS experiments, GUVs were unlabeled and visualized using a halogen lamp with differential interference contrast optics. Positioning in lateral and axial directions inside the GUV lumen was achieved such that the measurement volume was just below the widest cross section of the vesicle.
FCS analysis of CL-ODN complex and free ODN diffusion
Due to significant differences in their molecular weights, CL-ODN complexes and free ODN molecules are characterized by largely different diffusion constants. (A diffusion coefficient of a molecule species in solution in the first approximation is proportional to the third root of the molecular weight (Schwille et al., 1997)). Based on this fact, they can be discriminated by FCS analysis, which has been recognized as a reliable method for measuring diffusion constants of molecular species with high precision (Rigler et al., 1993; Schwille et al., 1997; Haustein and Schwille, 2003). Therefore, FCS is used here to assess the release of ODN from the complexes upon interaction with the target membranes by measuring the diffusion constants of the molecular species inside lumens of ODN-filled GUVs and comparing the obtained values to the diffusion constants of free ODNs and of the CL-ODN complexes in solution. Importantly, as the effective measurement volume in the FCS experiments is 2 orders of magnitude smaller than the average GUV volume, the FCS measurements are strictly confined within the GUV lumen, thereby assessing exclusively the diffusion of fluorescent species accumulated inside the GUV. Additionally, the amplitude value of the correlation curve G(0), which is inversely proportional to the particle concentration, gives an estimate of ODN concentration in the measurement volume. For each lipid composition of the GUV membranes, mean diffusion coefficients for ODN within GUVs were determined from a series of at least 10 independent FCS measurements in different vesicles, and the error was calculated as a half of the difference between the largest and smallest measured value. The diffusion coefficient of free ODN in water solution was calculated by taking the mean of eight independent measurements.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
; Kahya et al., 2003). For visualization, the GUV membrane bilayer was labeled with the fluorescent lipid analog DiO dye (emitting green fluorescence) and the CL-ODN complexes contained ODNs covalently labeled with Cy-5 (emitting a red light signal). In control experiments, a water solution of free ODN molecules, not complexed with the cationic lipid carrier, was injected into the GUV medium. Fig. 1 A shows a representative confocal image, collected 2 h after injection of CL-ODN complexes to the GUVs of the given lipid composition. As seen in Fig.1 A by the distribution of red fluorescence, no significant amount of free ODNs is observable in the GUV surrounding medium due to the packaging of ODNs into discrete particles, which are visible as red dots in solution or yellow dots when attached to the green GUV membrane. After adding CL-ODN complexes to the GUV sample, some vesicles fused together over time, altering a distribution of the initially evenly spaced GUVs into a network, as seen in overview (Fig. 1 B). Due to this observation, no estimation of a number of vesicles in the confocal images was done before the injection of the complexes, and image processing was always performed for the images, collected 2 h after the injection of the CL-ODN solution. As demonstrated in Fig.1 B, for the given lipid composition only a small number of GUVs accumulated a clearly distinguishable concentration difference of labeled ODN within their lumen with respect to the outside solution, accounting for <5% of a total number of analyzed GUVs. In all images, GUVs were counted as ODN-filled if they had an averaged fluorescent signal of 100 ADU above the background, or higher throughout their lumen, similar to the three vesicles in Fig. 1 A. In contrast, the two remaining vesicles in Fig. 1 A were discharged by the applied intensity threshold (<100 ADU). Additionally, the diffuse nature of fluorescence inside the GUV lumen in Fig.1 A indicates that ODN molecules are likely no longer condensed within the CL-ODN complexes. Also, the accumulation in one GUV was notably independent of the accumulation in the neighboring GUVs.
|
|
We also observed that a number of GUVs of exoplasmic leaflet composition accumulated ODNs and/or CL-ODN complexes bound to the membrane surface. As shown in Fig. 1 C, the homogenous distribution of ODN fluorescence on the GUV membrane indicates that bound complexes have significantly decondensed, but failed to transfer integrated ODNs into the GUV lumen. This is confirmed by quantitative analysis of the distribution of fluorescence signal on the GUV membrane and through the GUV lumen. Cy-5 fluorescence signal, and therefore corresponding ODN concentration, is more than five times higher on the GUV membrane, compared to the inside of the GUV lumen. The fluorescence peaks for Cy-5-labeled ODN (channel 1, red) are clearly distributed outside of the DiO peaks, which identify the membrane (channel 2, green). The intensity in the Cy-5 channel drops sharply toward the center of the plot, decreasing to nearly the background level within the GUV. Taken together, a described series of experiments clearly demonstrated that lipid membrane, enriched in cholesterol, DOPC, and sphingomyelin, is largely unsupportive to the CL-ODN complex fusion and ODN release.
In the next series of experiments we aimed to investigate interaction of the CL-ODN complexes with model endosomal membranes. According to the existing hypothesis, endocytosis is a prevailing route for cellular internalization of the cationic complexes (Zelphati and Szoka, 1996a; Zuhorn et al., 2002a,b; Zuhorn and Hoekstra, 2002). Moreover, the experiments described above provide indication that the CL-ODN complexes do not fuse with the external leaflet of the plasma membrane, giving additional indirect evidence that the particles rather need to be enveloped into the endosomes to gain the cellular entry. Therefore, CL-ODN complex interaction with the lipids present in the endosomal membrane is of key importance for ODN releasing from endosomal compartments.
The leaflets of the plasma membrane are known to differ substantially in lipid composition (Devaux, 1991). Specifically, the exoplasmic leaflet of the plasma membrane is enriched in cholesterol, glycosphingolipids, and PC, but is generally devoid of both negatively charged PS and overall neutral PE. In contrast, the cytoplasmic leaflet does not include large amounts of sphingolipid and cholesterol, but in addition to PC contains high fractions of aminophospholipids PS and PE. Although the asymmetric lipid distribution in the leaflets of the plasma membrane is tightly regulated, this asymmetry is likely disrupted during endocytosis, with significant amounts of PS and PE entering the luminal leaflet of the endosome (Devaux, 1991; Xu, and Szoka, 1996; Devaux, 2000). Also, cationic drugs have been shown to cause the disruption of lipid asymmetries in the cells, and for in vitro liposome systems (Moreau et al., 1997). Therefore, an exact understanding of the roles of cellular aminophospholipids, PS and PE, in disruption of the CL-ODN complexes and ODN release is highly significant for engineering of the effective ODN carriers.
To characterize specific roles of major endosomal lipids, coming into a close contact with the CL-ODN complexes (i.e., PC, PS, and PE) in ODN release, we designed a series of experiments where these lipid species were subsequently excluded and/or exchanged in the cellular relevant proportions in model membranes. First, GUVs were prepared from pure DOPC lipids. For this sample, we observed low ODN transfer, <5%, similar to the endoplasmic leaflet composition GUVs. Next, GUVs composed of DPPS/DOPE/DOPC and of DOPS/DOPE/DOPC at the mole fractions 0.15/0.25/0.6 were incubated with the CL-ODN complexes. The particular selection of the lipid species and their relative amounts in these compositions is a carefully reasoned choice. Particularly, the molar proportion and the selection of acyl groups of PS and PE lipids reflects a distribution found in the endosomal membranes in cells (Blom et al., 2001; Urade et al., 1988). The additional rationale for investigating two otherwise analogous compositions with a substitution of DPPS by DOPS is to verify whether varying the length and saturation of the hydrophobic tails may have a significant effect on the activity of PS in the ODN release. As shown in Fig. 2 A, after 2 h 35% of DPPS/DOPE/DOPC GUVs accumulated significant amounts of labeled ODNs within their lumen, which exceeded the 100-ADU intensity threshold across the lumen of the vesicle. GUVs with extensive amounts of membrane-bound, yet not fused, complexes (as observed in the experiments with the model plasma membrane exoplasmic leaflet GUVs) were not detected. The qualitative visual assessment of the confocal images, showing the distribution of labeled ODNs in the GUV lumen as opposed to ODN concentration on the GUV membrane and in the outside solution, is enhanced by quantitative analysis of the fluorescence intensity profiles across selected individual GUVs. Fig. 2 A clearly demonstrates that the intensity of Cy-5-labeled ODNs within the GUV lumen of the less bright vesicle significantly exceeds the outside signal (150 ADU on average), indicating that concentration of ODNs within the vesicle is considerably higher. The same results were obtained for the analogous lipid composition, DOPS/DOPE/DOPC, where the level of ODN transfer across the membrane stayed about the same, 35% (data not shown).
To probe the specific roles of PS and PE in ODN release, we prepared membrane compositions containing various amounts of either PS or PE lipids, keeping the proportion of DOPC constant. Correspondingly, the fraction of the anionic PS lipids in the initial composition was fully replaced by the neutral PE lipids in a DOPE/DOPC mixture at the mole fraction 0.4/0.6. We observed that the ODN transfer across the membrane remains at about the same high level as in the previous case, 35% (Fig. 2 B). Considering that this finding is in contrast with the earlier results of others for ODN release by DOPE/DOPC liposomes (Zelphati and Szoka. 1996b), the given observation was examined in more detail. Specifically, we investigated GUVs of only neutral lipids DOPE/DOPC with a variable DOPE fraction (exact data are given in line 2 of Table 1). In general, we observed that an increase in DOPE fraction results in a rise of ODN release. The role of PS lipids in ODN release was verified by preparing membranes, devoid of PE but containing PS, DPPS/DOPC and DOPS/DOPC compositions at 0.15/0.85 molar fractions. Again, in these mixtures PS was included in two modifications, with varying acyl groups. For both compositions we observed about the same degree of ODN transfer, 10%.
Negative control experiments, with injection of solution of naked, nonlipid-enveloped, Cy-5-labeled ODNs to each of the samples, clearly demonstrated that ODNs remained largely outside vesicles. Fig. 2 C shows a representative image for the negative control experiment for the DPPS/DOPE/DOPC mixture. The fluorescence signal in the leaky vesicles was always at the same level or lower compared to the outside solution, as seen for the vesicle marked by the arrow in Fig. 2 C.
An additional interesting observation, which we report here, is that during CL-ODN complex interaction with GUV samples containing PS and/or PE lipids, a binding of the complex to the membrane induced formation of budding vesicles in the contacting membrane. Usually, the vesiculation was more strongly expressed for PE-enriched membranes. The induced vesicles often were large enough to resolve the details of their internal structure. Fig. 3 shows that ODN can be accumulated within a vesicle, budding from a site of the complex merger to the GUV membrane. This likely indicates that released ODN remained trapped within the membrane of the associated vesicle, giving indirect evidence that free ODNs cannot cross the membrane bilayer.
|
1.5 x 10–8 cm2/s. The measured value for the diffusion coefficient of free ODN in solution is 1.3 ± 0.1 x 10–6 cm2/s. Because CL-ODN particles form rather inhomogeneous aggregates, varying in molecular weights, data-averaging over measurement series would not be statistically applicable for providing data errors in such experiments. However, we report that the diffusion of CL-ODN complexes was 100-fold slower compared to free ODN molecules.
|
40-fold higher with respect to the outside solution. The diffusion coefficients measured for ODNs diffusing inside DOPE/DOPC vesicles (at the mole fraction 0.25/0.75) and DOPC vesicles are correspondingly 1.5 ± 0.4 x 10–6 cm2/s and 1.6 ± 0.4 x 10–6 cm2/s, with representative FCS curves, shown in Fig. 4, F and G, respectively. The diffusion constants from all measurements inside ODN-filled GUVs, regardless of the lipid composition of the interacting membranes, are in good agreement with the mean diffusion coefficient of free ODNs in solution.
Therefore, based on the obtained FCS data we report that ODN molecules diffusing in the GUV lumen after release from the CL-ODN complexes are largely dissociated from the cationic lipids of the carrier. Additional important evidence of truly monomerically ODN diffusion in the GUV lumen is also given by the pronounced shoulder in the fast time range of the FCS curves seen in Fig. 4 A (control ODN) and in Fig. 4, D–G. This short fast range decay in the temporal correlation reflects on the specificity of the Cy-5 photophysics related to its isomerization reactions (Widengren and Schwille, 2000). If many fluorescent units diffuse in aggregates, as is the case for the cationic particles (Fig. 4 A, CL-ODN complexes), the pronounced photophysics, which is visualized as fast flickering of single dye molecules, is statistically disguised. Note that the curve in Fig. 4, C–G, cannot be directly compared by d to those in Fig. 4 A, as d is a setup-determined parameter and the measurements were performed on separate setups, with different observation volumes.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
). This conclusion was based on FRET experiments showing that anionic liposomes, when added to the CL-ODN complexes containing labeled ODNs, were able to release ODN into the solution. To investigate this premise in more detail, and to probe the specific roles of other common cellular lipids such as DOPC, DOPE, sphingomyelin, and cholesterol in interaction with CL-ODN complexes, we used giant unilamellar vesicles as a model membrane system, which allows us to directly visualize ODN transport. In contrast to the investigations based on using liposomes to model interaction of CL-ODN complexes with cellular membranes, a methodology that relies on working with GUVs provides a convenient approach not only to probe the mechanisms of action for contacting CL-ODN complexes and membrane lipids, but also to visualize and monitor the release process in real time under the optical microscope. In general, our data show that ODN release from the CL-ODN complexes during complex interaction with the membrane lipids strongly depends on the lipid composition of a target membrane, and that the different cellular lipids differently contribute to this effect. The principal novelty of our observations with GUV membrane systems, compared to earlier studies based on using small liposomes, is that we directly observe transfer of ODN molecules across lipid bilayers. Although modeling interaction with biological membranes by liposomes allows for investigation of ODN release, such an approach fails to assess the state of lipid bilayers during interaction with the CL-ODN complexes and provides no information on interaction between released oligonucleotides and contacting membranes. By using GUVs we demonstrated for the first time that model biological membranes remained intact after interaction with CL-ODN complexes and that the lipid bilayer boundaries were capable of keeping released ODN inside the GUV lumen, with no ODN leakage for an extended time. On the other hand, free ODN molecules, when injected into GUV samples, remained mostly outside GUVs and were not able to cross lipid bilayers. This is an important observation as it unambiguously confirms that cationic lipid carriers are capable of mediating ODN penetration across lipid bilayers, otherwise impermeable to free oligonucleotides.
Specifically, we observe that ODN transfer into the lumen of GUVs composed of an equimolar mixture of DOPC, sphingomyelin, and cholesterol (mimicking the exoplasmic leaflet of the plasma membrane) is low (<5%, Table 1). Consistent with cellular studies by others, suggesting that the complexes are internalized primarily via endocytosis (Zuhorn et al., 2002a,b), our data with the model endoplasmic leaflet membrane might provide indirect indication that the CL-ODN complexes need to be enveloped into endosomes to gain cellular entry, because they do not actively fuse at the external leaflet of the plasma membrane. At the same time, we clearly observed that whereas a pure DOPC lipid composition remains largely unsupportive for ODN transfer (<5%, Table 1), similar to cholesterol- and SM-containing membranes, availability of aminophospholipid species, i.e., PS and/or PE, significantly enhances the ODN transport. In a most general context, aminophospholipids may stimulate the transfer of ODN across the membrane through their tendency to promote membrane fusion with the CL-ODN complexes (Devaux, 1991) and due to the propensity of these lipids (particularly PE) to form nonbilayer structures.
Another important observation, which we report here, is that the degree of ODN transfer across the membranes, which contain either negatively charged PS (DPPS/DOPC and DOPS/DOPC compositions) or overall neutral PE (DOPE/DOPC composition) aminophospholipids, is similar. Additionally, for the membrane containing PE but devoid of PS, an increase of DOPE fraction in the membrane results in a higher level of ODN transfer across the bilayer. In other words, we have observed that PE lipids can mediate ODN release from the cationic carriers without PS interference, and that they can do this as efficiently as anionic PS lipids. This observation and the recent cellular studies of others (Lin et al., 2003), showing that the transfection efficiency of the DOTAP/DOPE lipid-DNA complexes is independent of their membrane charge density, indicate that PE lipid species, being overall neutral aminophospholipids, have an important function in ODN release, other than direct mediation of electrostatic interactions with the cationic lipids.
Whereas in the main our observations regarding the active promoting role of anionic PS lipids in ODN release from the carriers support existing knowledge (Xu and Szoka, 1996; Zelphati and Szoka, 1996a,b), what we report here about the PE capability to release ODN from the CL-ODN complexes in the absence of anionic lipids is a new finding. In contrast to the earlier data by others (Xu and Szoka, 1996; Zelphati and Szoka, 1996a,b), we show that PE lipids by themselves can induce ODN release from the CL-ODN complexes to the same extent as anionic PS lipids. This is a significant result as it revises the prevailing concept of the DNA/ODN release mechanisms, which suggests that the presence of the anionic lipids in the contacting membrane is a primary prerequisite for ODN release (Xu and Szoka, 1996; Zelphati and Szoka, 1996a,b). The given concept was developed based on FRET measurements of labeled DNA/ODN molecules released from the complexes upon addition of various composition liposomes. Specifically, it was observed that addition of PS/DOPE/DOPC (1:2:1) liposomes to the CL-ODN complexes induced significant release of labeled nucleotides, 80% (as estimated by fluorescence quenching assay), whereas DOPE/DOPC (2:1) liposomes were releasing only a minor amount of ODNs, 7% (Zelphati and Szoka, 1996b). Importantly, in the studies of Szoka's group, releasing activity of various anionic lipids (PS, PS, PI, PA) was always probed in triple mixtures of anionic lipid/DOPE/DOPC in the presence of a high fraction of PE in the liposome compositions (1:2:1 fractions, which equate to 50% PE). The double composition of PS/DOPC was not investigated. Therefore, we assume that in thus designed experiments the observed strong release of ODN by the triple-mixture liposomes might be a combined effect of both anionic lipids and PE. However, further investigation is needed to resolve the discrepancy between our data and the earlier observations by others regarding PE releasing activity, considering that we applied assays and measurement methodologies different from those of Szoka's group. So far the vigorous research being focused on understanding of the DNA/ODN-releasing actions is primarily for anionic lipids (Tarahovsky et al., 2004). By reporting our finding that PE can be the same efficient trigger of ODN release as PS, we suggest that more extensive investigation is also desired to comprehend a multifaceted role of membrane PE lipids in this process.
The diffuse nature of the ODN fluorescence in the GUV lumen signifies that released ODN is no longer aggregated into the original complex. However, this visual indication does not give a strict confirmation that ODN inside the vesicle is truly dissociated from the cationic lipids. By measuring the diffusion coefficients of ODN transferred within the GUV lumen and comparing them to the diffusion constant of free ODN in solution, we provide a quantitative confirmation that released ODN molecules dissociate from the carrier lipids.
A mechanism of ODN release from the cationic complexes in the absence of negatively charged anionic lipids in the barrier membrane is not understood and needs further investigation. A pertinent explanation of PE supporting activity in the ODN dissociation from the cationic lipids might be the specificity of its amine group. Given that the positive charge of the amine group in PE is relatively less shielded, it might be rather active in interaction with negatively charged molecular groups, such as phosphate groups in ODN. As a result, the amine group of PE can possibly compete with the cationic lipids of the carrier for the ODN phosphate groups, thus assisting in disassembly of the complex and allowing the ODN to dissociate from the carrier. This proposed mechanism of the PE amine group involvement in interaction with phosphate groups of ODNs also agrees with the observations, reported by others, showing that PE promotes destabilization of the CL-DNA complexes (Wong et al., 1996) and that the CL-DNA complexes prepared with PE helper lipids disassemble more readily than those prepared with PC lipids (Harvie et al., 1998).
Finally, data presented in this study support a growing trend in literature (Devaux, 2000; Cans et al., 2003), indicating that lipid-lipid interactions and lipid self-reorganization play highly active roles in diverse processes, which were previously thought to be governed exclusively by protein action.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
S.B. acknowledges a research fellowship from Alexander von Humboldt Foundation and thanks Dr. R.Winkler-Oswatitsch for general support. Financial support was additionally provided by a Volkswagen Foundation grant to P.S. S.P.G. was supported by an extended grant from the German-American Fulbright Commission and a stipend from the Max Planck Institute for Biophysical Chemistry.
|
FOOTNOTES |
|---|
Svitlana Berezhna's present address is Dept. of Molecular Biology, The Scripps Research Institute, La Jolla, CA 92037.
Submitted on March 18, 2004; accepted for publication October 7, 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Blom, T. S., M. Koivusalo, E. Kuismanen, R. Kostiainen, P. Somerharju, and E. Ikonen. 2001. Mass spectrometric analysis reveals an increase in plasma membrane polyunsaturated phospholipid species upon cellular cholesterol loading. Biochemistry. 40:14635–14644.[CrossRef][Medline]
Cans, A., N. Wittenberg, R. Karlsson, L. Sombers, M. Karlsson, O. Orwar, and A. Ewing. 2003. Artificial cells: unique insights into exocytosis using liposomes and lipid nanotubes. Proc. Natl. Acad. Sci. USA. 100:400–404.
Devaux, P. F. 1991. Static and dynamic lipid asymmetry in cell membranes. Biochemistry. 30:1163–1173.[CrossRef][Medline]
Devaux, P. F. 2000. Is lipid translocation involved during endo- and exocytosis? Biochimie. 82:497–509.[Medline]
Dietrich, C., L. A. Bagatolli, Z. N. Volovyk, N. L. Thompson, M. Levi, K. Jacobson, and E. Gratton. 2001. Lipid rafts reconstituted in model membranes. Biophys. J. 80:1417–1428.
Dimitrov, D. S., and M. I. Angelova. 1988. Lipid swelling and liposome electroformation mediated by electric fields. Bioelectrochem. Bioenerg. 19:323–333.[Medline]
Ferrari, M. E., D. Rusalov, J. Enas, and C. J. Wheeler. 2001. Trends in lipoplex physical properties dependent on cationic lipid structure, vehicle and complexation procedure do not correlate with biological activity. Nucleic Acids Res. 29:1539–1548.
Harvie, P., F. M. P. Wong, and M. B. Bally. 1998. Characterization of lipid DNA interactions. I. Destabilization of bound lipids and DNA dissociation. Biophys. J. 75:1040–1051.
Haustein, E., and P. Schwille. 2003. Ultrasensitive investigations of biological systems by fluorescence correlation spectroscopy. Methods. 29:153–166.[CrossRef][Medline]
Jääskeläinen, I., B. Sternberg, J. Mönkkönen, and A. Urtti. 1998. Physicochemical and morphological properties of complexes made of cationic liposomes and oligonucleotides. Int. J. Pharm. 167:191–203.[CrossRef]
Kahya, N., E.-I. Pecheur, W. P. De Boeij, D. A. Wiersma, and D. Hoekstra. 2001. Reconstitution of membrane proteins into giant unilamellar vesicles via peptide-induced fusion. Biophys. J. 81:1464–1474.
Kahya, N., D. Scherfeld, K. Bacia, B. Poolman, and P. Schwille. 2003. Probing lipid mobility of raft-exhibiting model membranes by fluorescence correlation spectroscopy. J. Biol. Chem. 278:28109–28115.
Kettling, U., A. Koltermann, P. Schwille, and M. Eigen. 1998. Real time enzyme kinetics of restriction endonuclease EcoRI monitored by dual-color fluorescence cross-correlation spectroscopy. Proc. Natl. Acad. Sci. USA. 95:1416–1420.
Kobayashi, T., M. H. Beuchat, J. Chevallier, A. Makino, N. Mayran, J. M. Escola, C. Lebrand, P. Cosson, and J. Gruenberg. 2002. Separation and characterization of late endosomal membrane domains. J. Biol. Chem. 277:32157–32164.
Koltover, I., T. Salditt, J. O. Raedler, and C. R. Safinya. 1998. An inverted hexagonal phase of cationic liposome-DNA complexes related to DNA release and delivery. Science. 281:78–81.
Koltover, I., T. Salditt, and C. R. Safinya. 1999. Phase diagram, stability, and overcharging of lamellar cationic lipid-DNA self-assembled complexes. Biophys. J. 77:915–924.
Koltover, I., K. Wagner, and C. R. Safinya. 2000. DNA condensation in two dimensions. Proc. Natl. Acad. Sci. USA. 97:14046–14051.
Lin, A. J., N. L. Slack, A. Ahmad, C. X. George, C. E. Samuel, and C. R. Safinya. 2003. Three-dimensional imaging of lipid gene-carriers: membrane charge density controls universal transfection behavior in lamellar cationic liposome-DNA complexes. Biophys. J. 84:3307–3316.
Marshall, E. 2000. Biomedicine: gene therapy on trial. Science. 288:951–957.
Marshall, E. 2002. News of the week: gene therapy a suspect in leukemia-like disease. Science. 298:34–35.
Meidan, V. M., J. S. Cohen, N. Amariglio, D. Hirsch-Lerner, and Y. Barenholz. 2000. Interaction of oligonucleotides with cationic lipids: the relationship between electrostatics, hydration and state of aggregation. Biochim. Biophys. Acta. 1464:251–261.[Medline]
Moreau, C., J. C. Sulpice, P. F. Devaux, and A. Zachowski. 1997. Drug-induced transmembrane lipid scrambling in erythrocytes and in liposomes requires the presence of polyanionic phospholipids. Mol. Membr. Biol. 14:5–12.[Medline]
Nakanishi, M., and A. Noguchi. 2001. Confocal and probe microscopy to study gene transfection mediated by cationic liposomes with a cationic cholesterol derivative. Adv. Drug Deliv. Rev. 52:197–207.[CrossRef][Medline]
Noguchi, A., T. Furuno, C. Kawaura, and M. Nakanishi. 1998. Membrane fusion plays an important role in gene transfection mediated by cationic liposomes. FEBS Lett. 433:169–173.[CrossRef][Medline]
Rigler, R., Ü. Mets, J. Widengren, and P. Kask. 1993. Fluorescence correlation spectroscopy with high count rates and low background: analysis of translational diffusion. Eur. Biophys. J. 22:169–175.
Scherer, L., and J. Rossi. 2003. Approaches for the sequence specific knockdown of mRNA. Nat. Biotechnol. 21:1457–1465.[CrossRef][Medline]
Scherfeld, D., N. Kahya, and P. Schwille. 2003. Lipid dynamics and domain formation in model membranes composed of ternary mixtures of unsaturated and saturated phosphatidylcholines and cholesterol. Biophys. J. 85:1–11.
Schwille, P., J. Bieschke, and F. Oehlenschläger. 1997. Kinetic investigations by fluorescence correlation spectroscopy: the analytical and diagnostic potential of diffusion studies. Biophys. Chem. 66:211–228.[CrossRef][Medline]
Tarahovsky, Yu. S., R. Koynova, and R. C. MacDonald. 2004. DNA release from lipoplexes by anionic lipids: correlation with lipid mesomorphism, interfacial curvature, and membrane fusion. Biophys. J. 87:1054–1064.
Ulrich, A. S. 2002. Biophysical aspects of using liposomes as delivery vehicles. Biosci. Rep. 22:129–150.[CrossRef][Medline]
Urade, R., Y. Hayashi, and M. Kito. 1988. Endosomes differ from plasma membrane in the phospholipid molecular species composition. Biochim. Biophys. Acta. 946:151–163.[Medline]
Weisman, S., D. Hirsch-Lerner, Y. Barenholz, and Y. Talmon. 2004. Nanostructure of cationic lipid-oligonucleotide complexes. Biophys. J. 87:609–614.
Widengren, J., U. Mets, and R. Rigler. 1995. Fluorescence correlation spectroscopy of triplet states in solution: a theoretical and experimental study. J. Phys. Chem. 99:13368–13379.[CrossRef]
Widengren, J., and P. Schwille. 2000. Characterization of photoinduced isomerization and back-isomerization of the cyanine dye Cy5 by fluorescence correlation spectroscopy. J. Chem. Phys. 104:6416–6428.
Wong, F. M. P., P. Harvie, Y. P. Zhang, E. C. Ramsay, and M. B. Bally. 2003. Phosphatidylethanolamine mediated destabilization of lipid-based pDNA delivery systems. Int. J. Pharm. 14:117–127.
Wong, F. M. P., D. L. Reimer, and M. B. Bally. 1996. Cationic lipid binding to DNA: characterization of complex formation. Biochemistry. 35:5756–5763.[CrossRef][Medline]
Xu, Y., S. W. Hui, P. Frederik, and F. C. Szoka. 1999. Physicochemical characterization and purification of cationic lipoplexes. Biophys. J. 77:341–353.
Xu, Y., and F. C. Szoka. 1996. Mechanism of DNA release from cationic liposome/DNA complexes used in cell transfection. Biochemistry. 35:5616–5623.[CrossRef][Medline]
Zelphati, O., and F. C. Szoka. 1996a. Intracellular distribution and mechanism of delivery of oligonucleotides mediated by cationic lipids. Pharm. Res. 13:1367–1372.[CrossRef][Medline]
Zelphati, O., and F. C., Szoka. 1996b. Mechanism of oligonucleotide release from cationic liposomes. Proc. Natl. Acad. Sci. USA. 93:11493–11498.
Zelphati, O., and F. C. Szoka. 1997. Cationic liposomes as an oligonucleotide carrier: mechanism of action. J. Liposome Res. 7:31–49.
Zhang, Y.-P., D. L. Reimer, G. Zhang, P. H. Lee, and M. B. Bally. 1997. Self-assembling DNA-lipid particles for gene transfer. Pharm. Res. 14:190–196.[CrossRef][Medline]
Zuhorn, I. S., and D. Hoekstra. 2002. On the mechanism of cationic amphiphile-mediated transfection. To fuse or not to fuse: is that the question? J. Membr. Biol. 189:167–179.[CrossRef][Medline]
Zuhorn, I. S., R. Kalicharan, and D. Hoekstra. 2002a. Lipoplex-mediated transfection of mammalian cells occurs through the cholesterol-dependent clathrin-mediated pathway of endocytosis. J. Biol. Chem. 277:18021–18028.
Zuhorn, I. S., V. Oberle, W. H. Visser, J. B. F. N. Engberts, U. Bakowsky, E. Polushkin, and D. Hoekstra. 2002b. Phase behavior of cationic amphiphiles and their mixtures with helper lipid influences lipoplex shape, DNA translocation, and transfection efficiency. Biophys. J. 83:2096–2108.
This article has been cited by other articles:
 |
|
 |
|
  M. Wang, M. Law, J. Duhamel, and P. Chen Interaction of a Self-Assembling Peptide with Oligonucleotides: Complexation and Aggregation Biophys. J., October 1, 2007; 93(7): 2477 - 2490. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Koynova, L. Wang, and R. C. MacDonald An intracellular lamellar-nonlamellar phase transition rationalizes the superior performance of some cationic lipid transfection agents PNAS, September 26, 2006; 103(39): 14373 - 14378. [Abstract] [Full Text] [PDF]
|
|
| |||||||
