A Quantitative Fluorescence Microscopy-based Single Liposome Assay for Detecting the Compositional Inhomogeneity Between Individual Liposomes

Bioengineering

Your institution must subscribe to JoVE's Bioengineering section to access this content.

Fill out the form below to receive a free trial or learn more about access:

 

Summary

This protocol describes the fabrication of liposomes and how these can be immobilized on a surface and imaged individually in a massive parallel manner using fluorescence microscopy. This allows for the quantification of the size and compositional inhomogeneity between single liposomes of the population.

Cite this Article

Copy Citation | Download Citations | Reprints and Permissions

Münter, R., Andresen, T. L., Larsen, J. B. A Quantitative Fluorescence Microscopy-based Single Liposome Assay for Detecting the Compositional Inhomogeneity Between Individual Liposomes. J. Vis. Exp. (154), e60538, doi:10.3791/60538 (2019).

Abstract

Most research employing liposomes as membrane model systems or drug delivery carriers relies on bulk read-out techniques and thus intrinsically assumes all liposomes of the ensemble to be identical. However, new experimental platforms able to observe liposomes at the single-particle level have made it possible to perform highly sophisticated and quantitative studies on protein-membrane interactions or drug carrier properties on individual liposomes, thus avoiding errors from ensemble averaging. Here we present a protocol for preparing, detecting, and analyzing single liposomes using a fluorescence-based microscopy assay, facilitating such single-particle measurements. The setup allows for imaging individual liposomes in a massive parallel manner and is employed to reveal intra-sample size and compositional inhomogeneities. Additionally, the protocol describes the advantages of studying liposomes at the single liposome level, the limitations of the assay, and the important features to be considered when modifying it to study other research questions.

Introduction

Liposomes are spherical phospholipid-based vesicles that are heavily used both in basic and applied research. They function as excellent membrane model systems, because their physiochemical properties can be easily manipulated by varying the lipid components making up the liposome1,2. Also, liposomes constitute the most used drug delivery nanocarrier system, offering improved pharmacokinetics and pharmacodynamics as well as high biocompatibility3.

For many years, liposomes have primarily been studied using bulk techniques, giving only access to ensemble average read-out values. This has led the majority of these studies to assume that all liposomes in the ensemble are identical. However, such ensemble-averaged values are only correct if the underlying dataset is uniformly distributed around the mean value, but can represent a false and biased conclusion if the dataset includes multiple independent populations, for example. Additionally, assuming the ensemble mean to represent the whole population can overlook the information harbored within the inhomogeneity between liposomes. Only recently have quantitative assays emerged that are able to probe single liposomes, revealing large inhomogeneities between individual liposomes with respect to important physicochemical properties including liposome size4, lipid composition5,6, and encapsulation efficiency7, highlighting the importance of studying liposomes at the single liposome level.

A research area where ensemble averaging of liposome properties has been shown to bias results is studying liposome size-dependent protein-membrane interactions8,9. Traditionally, researchers studying such processes have been restricted to preparing liposomes with different ensemble average diameters by extrusion through filters with different pore sizes9. However, extracting the diameter of individual liposomes using single liposome assays has revealed large population overlaps, with liposomes extruded using 100 nm and 200 nm filters displaying up to 70% overlap in their size distribution4. This could severely bias bulk measurements of liposome size-dependent protein-membrane interactions10. Performing the membrane-protein interaction studies using the single liposome assay, researchers instead took advantage of the size-polydispersity within the sample, allowing them to study a wide range of liposome diameters within each single experiment, facilitating new discoveries of how membrane curvature and composition can affect protein recruitment to membranes4,11,12. Another field where the application of single liposome assays has proven instrumental is in mechanistic studies of protein-mediated membrane fusion13,14. For such kinetic measurements, the ability to study individual fusion events alleviated the need for the experimental synchronization of the fusion process, allowing new mechanistic insights that would otherwise have been lost in the spatiotemporal averaging done in bulk ensemble measurements. Additionally, single liposomes have been used as a membrane scaffold, allowing the measurement of individual proteins and offering new knowledge on transmembrane protein structural dynamics15,16. Furthermore, such proteoliposome-based setups made it possible to study the function of individual transmembrane transporters17 and pore-forming protein complexes18 as well as the mechanism of bioactive membrane-permeabilizing peptides19. Single liposomes have also been used as soft matter nanofluidics with surface-immobilized single liposomes serving as chambers for enzymatic reactions in volumes of 10-19 L, increasing the throughput and complexity of the screening assays with minimal product consumption20.

Recently, single liposome assays have been used for characterizing drug delivery liposomes at a previously unprecendented level of detail. Researchers were able to quantify significant inhomogeneities in the amount of polymer attached to the surface of individual liposomes21. The single liposome assays also allowed measurements of drug delivery liposomes in complex media, such as blood plasma, revealing how elements anchored to the liposome surface through lipid anchors can be susceptible to dissociation when liposomes are exposed to conditions mimicking those experienced during in vivo circulation22. Overall, the versatility and usefulness of the single liposome assays are substantiated by the great variety of problems these setups have been employed to address, and we envision that the methodology will continue to be developed and find use in new scientific fields.

Here we describe a fluorescence microscopy-based single liposome assay that allows individual liposomes to be studied in a high-throughput manner (Figure 1). To illustrate the method, we use it to quantify the size and compositional inhomogeneity between individual liposomes within an ensemble. The assay employs fluorescence microscope imaging of single liposomes immobilized on a passivated glass surface. We first describe the critical steps in the liposome fabrication process that ensures proper fluorescent liposome labeling and immobilization. Then, we describe the surface preparation needed to facilitate liposome immobilization before outlining the procedure for ensuring appropriate liposome surface densities. We discuss the microscopy parameters important for acquiring high-quality images and delineate how to perform simple data analysis, allowing the extraction of liposome size and compositional inhomogeneity. This generic protocol should provide a good basis for the interested researcher to develop the assay further for his or her specific research interest.

Subscription Required. Please recommend JoVE to your librarian.

Protocol

1. Liposome Preparation

NOTE: Briefly, preparation of liposomes usually includes three crucial steps: 1) preparation of dry lipid films of the desired lipid composition; 2) rehydration of the lipids for formation of liposomes; and 3) controlling the size and lamellarity of the liposome population.

  1. Weigh out the lipids and dissolve them in tert-butanol:water (9:1) in glass vials.
    1. Dissolve POPC (1-palmitoyl-2-oleoyl-glycero-3-phosphocholine; MW = 760 g/mol) to 50 mM.
    2. Dissolve cholesterol (MW = 387 g/mol) to 25 mM.
    3. Dissolve DOPE-Atto488 (1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-Atto488; MW = 1,316 g/mol) to 0.1 mM.
    4. Dissolve DOPE-Atto655 (1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-Atto655; MW = 1,368 g/mol) to 0.1 mM.
    5. Dissolve DSPE-PEG2000-biotin (1,2-distearyl-sn-glycero-3-phosphatethanolamine-N-[biotinyl(polyethylene glycol)-2000]; MW = 3,017 g/mol) to 0.1 mM.
      NOTE: Heat lipids to 55 °C and use magnetic stirring in order to ensure complete dissolution of the lipids. Alternatively, use a sonication bath. Unused lipid stocks can be stored at -20 °C for several months.
  2. Mix the lipid stocks prepared in step 1.1 to a molar ratio of POPC:cholesterol:DOPE-Atto488:DOPE-Atto655:DOPE-PEG-biotin 68.95:30:0.5:0.5:0.05, by adding 138 µL of POPC, 120 µL of cholesterol, 500 µL of each fluorescently labeled lipid, and 50 µL of DSPE-PEG-biotin to a fresh glass vial.
    NOTE: The exact liposome composition can easily be modified in order to address the specific question of interest. See discussion for more detail.
  3. Loosen the lid of the glass vial, and snap-freeze the vial in liquid nitrogen.
  4. Lyophilize the frozen lipid mixture overnight.
  5. Add 1 mL of 200 mM D-sorbitol buffer (sorbitol buffer) to the dry lipids.
  6. Heat the mixture to 45 °C and expose to magnetic stirring for at least 1 h.
    NOTE: The buffer should reflect the specific question that is being addressed (e.g., physiological conditions for studying membrane-protein interactions, or a specific clinically approved buffer for studying drug delivery liposomes). However, if a specific buffer is not required for the study, a buffer without ions can be applied for rehydration in order to reduce the multilamellarity of the liposomes.
  7. Freeze the lipid suspension by dipping the vial in liquid nitrogen, and wait until the suspension is completely frozen.
  8. Dip the frozen suspension in a heating bath at 55 °C until the mixture is completely thawed.
  9. Repeat steps 1.7 and 1.8 until the liposome suspension has been exposed to a total of 11 freeze/thaw cycles.
    NOTE: Repeated freeze/thaw cycles have shown to reduce liposome multilamellarity23, which is paramount for the accuracy of the single liposome assay, as multilamellar liposomes will skew the fluorescence intensity versus liposome size ratio of the liposomes. The multilamellarity is usually inherently low when including more than 0.5% PEGylated lipid in the formulation (such as commonly done in liposomes for drug delivery)24.
  10. Extrude the liposome suspension once through an 800 nm polycarbonate filter using a mini extrusion kit. Follow the manufacturer's instructions for assembly of the extrusion kit (see Table of Materials).
  11. Store the liposomes at 4 °C overnight.

2. Surface Preparation of Imaging Chamber

  1. Prepare bovine serum albumin (BSA; 1 mg/mL), BSA-biotin (1 mg/mL) and streptavidin (0.025 mg/mL) in 10 mM HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) 95 mM NaCl buffer (HEPES buffer).
    NOTE: To prevent liposomal burst, the sorbitol buffer used for rehydration and the HEPES buffer used for surface preparation should be isotonic. It is thus recommended to check the osmolarity of the buffers before the experiment.
  2. Mix 1,200 µL of BSA and 120 µL of BSA-biotin and add 300 µL of the mixture to each well in an 8 well slide for microscopy.
    NOTE: Here we use a commercially available 8 well microscope slide with a glass bottom (see Table of Materials), but the protocol can easily be adapted to customized microscope chambers.
  3. Incubate the slide for 20 min at room temperature (RT).
  4. Wash the slide 8x with 300 µL of HEPES buffer.
    NOTE: Be careful not to leave the wells without buffer for more than a few seconds, as drying out the surface will damage it. Furthermore, be careful not to scratch the surface with the pipette tip as this will also damage it. Thus, when aspirating buffer from a well, do it from the edge or corner.
  5. Add 250 µL of streptavidin and incubate for 10 min at RT.
  6. Repeat the 8 washes described in step 2.4.
  7. Store the microscope slide with 300 µL of sorbitol buffer in each well at 4 °C. Evaporation of the solvent will damage the surface, so put the microscope slide in a Petri dish and seal it with parafilm unless the prepared slide is used immediately.

3. Liposome Immobilization

  1. Dilute the liposome suspension to about 20 µM total lipid in sorbitol buffer. The procedure in section 1 will yield a liposome suspension of approximately 10 mM total lipid.
  2. Place the slide on the microscope and focus on the surface of the chamber using the increased laser reflection signal from the glass/buffer interface as a guide.
  3. Wash the chamber 4x with 300 µL of fresh HEPES buffer.
  4. Add 10 µL of diluted liposome stock (20 µM total lipid) to a microcentrifuge tube.
  5. Take out 100 µL of buffer from the chamber, add to the microcentrifuge tube prepared in step 3.4, mix properly, and put the 110 µL back into the chamber.
  6. Put the specimen under the microscope, and use a rudimental microscope setting capable of detecting signal from the liposome fluorophore(s) to observe the liposomes being immobilized on the surface.
  7. Aim for a liposome surface density that is simultaneously sparse enough to indentify individual liposomes and dense enough that it facilitates high-throughput measurements. Typically using a 50 µm x 50 µm field of view, 300−400 liposomes per frame is optimal (Figure 2). This can usually be achieved within 5−10 min.
    NOTE: If only rapidly moving fluorescent particles are detected in the field of view, an absence or too low concentration of any of the components critical for the immobilization process (BSA-biotin, streptavidin, DOPE-PEG-biotin) might be the cause. If no liposomes are detected it might either be related to a low concentration of liposomes, improper settings for the fluorescence detection, or potentially imaging with a focal plane not at the glass surface.
  8. Once an appropriate liposome surface density is reached, wash the chamber 3x with 200 µL of HEPES buffer.
    NOTE: Be aware that the immobilization kinetics also depend on liposome properties such as size and charge and should thus always be optimized for each liposome formulation.

4. Image Acquisition

NOTE: This section will depend a lot on the microscope system available to the researcher performing the experiment. Thus, overall guidelines on how to perform the imaging will be described. However, the exact settings and how to apply them will vary between the different microscope setups. For example, some systems allow choosing any emission filter combination desired, while other microscopes are equipped with specific, preset filters.

  1. Set up the microscope for imaging single liposomes. To ensure optimal image quality and subsequent data analysis use a high greyscale resolution and a pixel scheme that allows for oversampling individual liposomes. A bit depth of 16 and at least 1,024 x 1,024 pixels for a 50 µm x 50 µm area is recommended. If available, line-averaging can beneficially be applied to reduce noise (e.g., 3 scans per line).
  2. Select an excitation laser power that both ensures nonsignificant frame-to-frame bleaching and strong enough signal to clearly discriminate individual liposomes from the background. The optimal setting will depend on the specific microscope used as well as the fluorophore combination. Make sure that detectors are not saturated, as this will bias intensity quantification.
  3. Imaging both the DOPE-Atto488 and DOPE-Atto655 fluorophores in the liposomes requires imaging multiple channels. Thus, make sure each channel is imaged sequentially to avoid cross-excitation. For example, first take one image by exciting at 488 nm and reading emission at 495−560 nm. Thereafter, take another image by exciting at 633 nm but reading emission only at 660−710 nm. The specifics will depend on the microscope system (e.g., which lasers and filters) available.
  4. Make sure to cover different areas of the surface, acquiring at least 10 images of the sample, thus imaging at least 3,000 individual liposomes. If the surface density is lower than 300 liposomes/frame, acquire more images. Make sure to refocus the microscope for every new image.
    NOTE: For two-channel imaging, make sure to name the image files so that pairs of images with the same liposomes in different fluorescence channels can easily be identified during data analysis.
  5. In order to quantify the experimental uncertainty relating to the measurement of compositional inhomogeneity, image the same area of liposomes before and after refocusing (see Figure 3 and description in the text).
  6. Export the images from the microscope software as .tiff files. Export the two channels of the same liposomes individually.

5. Data Analysis

NOTE: Specially developed automated 2D Gaussian fitting routines have previously been employed6,11,12. However, to increase the applicability of the method a data analysis process that can be easily implemented in all laboratories is described.

  1. Load the corresponding pair of .tiff images of two different fluorescence channels in the same imaging field into the FIJI (FIJI Is Just ImageJ) software.
  2. In the Image menu, choose Color, and use the Merge Channel function to create a composite of the two channels.
  3. Observe if the liposomes imaged in two different channels display good colocalization or whether visible drift occurred.
    NOTE: In case of drift between the two fluorescence channels (recognized as a systematic and equal X-Y offset between the signal in the two channels), one of the frames can be translated using the Transform > Translate function in the Image menu of FIJI. However, care should be taken with such image manipulation. It is thus recommended to instead avoid drift as much as possible when imaging the liposomes.
  4. Make sure the ComDet plugin (v.0.3.6.1 or newer) is installed, or do this in the Plugins menu.
  5. Open the ComDet plugin to detect particles by going to the Plugins menu and choosing ComDet v.0.3.6.1 > Detect Particles.
  6. In ComDet, make sure to choose the Detect Particles on Both Channels Individually function. Set the Max Distance Between Co-localized Spots similar to the Approximate Particles Size. Usually, 4 pixels is appropriate for the settings described here.
  7. The signal-to-noise ratio should allow for detection of even dim particles with a low amount of fluorescence. In ComDet, set this ratio to 3 by setting the Sensitivity of Detection in both channels to Very Dim Particles (SNR = 3).
    NOTE: While this SNR value is usually appropriate, it might be necessary to set it higher (e.g., if there is a lot of image noise that may be identified as liposomes and lead to false positives).
  8. Make sure the boxes with Calculate Colocalization and Plot Detected Particles in Both Channels are checked, before pressing OK.
  9. After running the analysis, two pop-up windows with Results and Summary will show. Export the data table Results containing the colocalization data (particle coordinates and integrated intensity of each particle detected) to a data handling software of choice by saving the table as a .txt file and importing it into the software.
  10. Make sure that each liposome is only included once in the data set and not both the channel1/channel2 as well as channel2/channel1 ratio. Thus, filter the data so only Abs_frame = 1 is plotted in step 5.11.
    NOTE: By choosing Colocalized = 1, false positives from noise in the images can be removed from the analysis. However, any liposomes with only one of the two fluorescent components present will also be excluded from the analysis, thus potentially removing important data points from the analysis.
  11. Plot a histogram of the column with data containing the Intensity Ratio for each detected liposome.
  12. The degree of compositional inhomogeneity for the studied liposomal system is represented by the width of the intensity ratio distribution. To quantify the inhomogeneity, fit the intensity ratio histogram with a Gaussian function and extract the mean (µ) and standard deviation (sigma). See the Representative Results section.
  13. A value for the degree of inhomogeneity (DI) can now be calculated using the coefficient of variation defined as DI = sigma / µ.

6. Liposome Size Calibration

  1. Take out part of the liposome stock, and extrude 21x through a 50 nm polycarbonate filter as described in step 1.10.
  2. Dilute the liposomes by adding 10 µL of the liposome suspension to 800 µL of sorbitol buffer in a microcentrifuge tube.
  3. Transfer the diluted liposome sample to a polypropylene single-use cuvette.
  4. Measure the size using dynamic light scattering (DLS). Perform at least three independent runs to measure the size and polydispersity of the liposome suspension.
    NOTE: If necessary, use a more concentrated liposome suspension for the measurement. Alternatives to DLS (e.g., nanoparticle tracking analysis) can also be applied for measuring the size of the liposomes. A description of how to execute such particle-size determination is beyond the scope of this protocol.
  5. Image the calibration liposomes on the microscope using exactly the same experimental settings as defined in steps 4.1 and 4.2.
  6. Extract the integrated intensity for each calibration liposome in the calibration images: Extract the filtered results sheet containing liposome fluorescence intensities as described in steps 5.1−5.10. From the results table, extract the "IntegratedInt" column.
  7. Because the total integrated intensity of a liposome labeled in its membrane is proportional to the surface area of the liposome and is thus proportional to the square of its diameter, plot a square root intensity histogram of the fluorescence intensity of the calibration liposomes.
  8. Fit the integrated intensity histogram produced in step 6.7 with a log normal distribution and extract the average fluorescence intensity of the calibration liposomes.
  9. To determine the relation between the square root intensity (IntSqrt) and liposome size, calculate the correction factor (C) using the average liposome diameter (Dia) weighed by the number obtained from the DLS measurements: Dia = C x IntSqrt is equivalent to C = Dia / IntSqrt.
  10. Calculate IntSqrt values for the liposomes in the compositional inhomogeneity experiment and convert these to diameters by multiplying with the correction factor.
  11. Plot the intensity ratio value as a function of diameter for the compositional inhomogeneity liposomes, thus achieving the inhomogeneity as a function of liposome size for a population of liposomes spanning from approximately 50 nm−800 nm.

Subscription Required. Please recommend JoVE to your librarian.

Representative Results

Following the protocol described makes it possible to image single liposomes in a massive parallel manner (Figure 1). The successful surface immobilization of liposomes should be immediately apparent upon the addition of the liposome solution to the chamber (step 3.6 in the protocol) as diffraction limited intensity spots should appear in the image (Figure 1B and Figure 1C).

In order to achieve good statistics and exploit the high-throughput abilities of the assay, several thousand liposomes should be imaged. In order to do this in a reasonable number of images, it is recommended to immobilize enough liposomes to achieve a density of 300−400 liposomes per frame, as this will bring the number of images per sample down to ~10, thus limiting the number of images that has to be acquired and analyzed. A lower density will make the data analysis more time-consuming, while a higher density can make it challenging for the image analysis software to distinguish single liposomes. To get a visual impression of what 300−400 liposomes look like, see Figure 2A. It should be noted, however, that for some applications (e.g., membrane-protein interactions with a protein that tends to elicit strong background binding to the BSA surface) it is recommended to use a slightly lower density, such as in Figure 2A at the top right.

Occasionally, when acquiring images it is hard to get the whole field of view in proper focus, as illustrated in Figure 2B. Such issues might indicate that the sample plate is tilting, which can arise from the plate not being placed properly on the specimen holder on the microscope. Also, if the liposomes seem large and blurry, the buffer in the chamber might have evaporated and the surface dried out. It is especially important to keep this issue in mind when imaging for long time periods or when performing measurements at temperatures higher than RT. To illustrate the difference between properly stored and dried out liposomes, an image of the same sample before and after drying out the chamber is shown in Figure 2C.

After having ensured optimal image quality, the integrated intensity for each liposome in the two imaging channels can be extracted following the steps in section 5 of this protocol. Doing this will create a list of unique intensity values, allowing us to calculate the intensity ratio for each individual liposome. The compositional inhomogeneity is evaluated from intensity ratio histograms, typically revealing a Gaussian distribution around a mean ratio value (Figure 3A). Notice that if strong deviations from a Gaussian distribution are observed (Figure 3B), it indicates a detection sensitivity issue for at least one of the imaging channels, suggesting that a subset of liposomes showing a weaker signal have been excluded. This could be due to both the detection limit of the imaging system, or excluding liposomes in the data analysis (e.g., when applying a certain minimum threshold; see step 5.7).

Calculating the DI value as described in steps 5.11−5.13 in the protocol will provide a quantitative measure of the compositional inhomogeneity between the individual liposomes of the preparation. For the liposomal system studied here, DI = 0.23 ± 0.01 (Figure 3C). This value can be used to systematically compare how varying liposome properties or preparation methods affect the compositional inhomogeneity. To conceptualize the meaning of the DI value, we refer to the normal distribution displayed by the intensity ratio histogram, meaning they will obey the empirical 68-95-99.7 rule. This rule describes the percentage of a population that falls within one, two, and three standard deviations around the mean value. For the DI value of 0.23 ± 0.01 found here, it means that 32% of the liposomes in the population will have an intensity ratio that differs by more than 23% from the mean molar ratio of the ensemble (Figure 3D).

For the control experiment imaging the same liposomes before and after refocusing (section 4.6), the same data analysis and result plotting as for the actual experiment is performed. Doing this allows for the quantification of the experimental uncertainty of the DI value, which was found to be DIuncertainty = 0.10 ± 0.01 (Figure 3E). Although the DIuncertainty value will depend on the employed imaging system, it was found that confocal microscope setups consistently give rise to DIuncertainty values of around 0.105,6. The DI value of DI = 0.23 ± 0.01 found for the liposomal system here is more than twice the experimental uncertainty, which suggests the presence of significant compositional inhomogeneity between the individual liposomes of the preparation.

Performing the size calibration experiment described in section 6 will allow the arbitrary liposome intensity values to be transformed to physical diameters in nanometers. The method has been validated against other imaging techniques25 such as cryogenic electron microscopy, which has an ability to optically resolve the nanometer-sized liposomes, although with much lower throughput. Imaging the control sample using the exact same microscope settings as used for the actual experiments, it is now possible to correlate the mean liposome intensity to the mean liposome diameter determined by DLS (Figure 4A). The next step is to depict the liposome intensity distribution, which based on the physical constraints against creating extremely small liposomes, typically will display a log normal distribution (Figure 4B). If the distribution is not well described by a log normal distribution (most often due to missing liposomes with low intensity values) it means that a part of the liposome population was excluded either due to low microscope detection sensitivity or overly strict parameters during data analysis (Figure 4C). It is important to catch and address these issues to ensure unbiased data interpretation. Next, the intensity values for each individual liposome in Figure 4B can be converted to actual diameters using the correction factor determined in Figure 4A (Figure 4D). After determining the size of each liposome, the DI as a function of diameter can now be investigated by plotting the intensity ratio versus diameter for each individual liposome (Figure 4E). This funnel-like data structure corroborates earlier findings that the DI value increases when the liposome size decreases6. Importantly, the symmetrical increase in the spread of intensity ratios around a mean value suggests that there is no systematic size-dependent variation in the average composition of liposomes.

Figure 1
Figure 1: The single liposome assay. (A) Liposomes are formulated with an equimolar content of the fluorescently labeled lipids DOPE-Atto488 and DOPE-Atto655 and immobilized on a BSA surface using a biotin/streptavidin linkage. (B) The immobilized liposomes are imaged using a confocal microscope, allowing for detection of the fluorescence intensities from the single liposomes. Scale bars = 8 µm. (C) Zoom of the green area in (B) depicted as surface intensity plot for two adjacent liposomes displaying inverted fluorescent signal between the two channels, illustrating the concept of compositional inhomogeneity. Please click here to view a larger version of this figure.

Figure 2
Figure 2: Optimization and pitfalls. (A) The top left shows 29 liposomes immobilized in a 50 µm x 50 µm frame. These are too few liposomes per frame to exploit the possibility for high-throughput investigation of liposome inhomogeneity. The top right shows 123 liposomes per frame. This is a suboptimal density but can be used if many images are acquired. The bottom left shows 300−400 liposomes per frame. This is optimal for the protocol described here. The bottom right shows more than 1,500 liposomes per frame, which is too many to distinguish individual liposomes. There is a high risk that a single spot will actually be two liposomes immobilized within the same diffraction limited spot. (B) If the chamber is not placed properly in the specimen holder, getting the whole field of view in proper focus will be impossible. In the left micrograph the plate is tilting around the traverse axis, giving rise to the lower left corner being out of focus. In the right micrograph the plate is tilting around the longitudinal axis, giving rise to a heavier tilt. Both the lower left and upper right corner are out of focus. (C) If imaging for long time periods or at elevated temperatures the buffer might evaporate, leading to drying out of the chamber. We illustrated this scenario here by imaging liposomes before and after the chamber was left to dry for 3 min, and then rewetted by adding fresh buffer to the chamber. The resulting liposomes are larger, blurry, have lower fluorescence intensity, and appear to be spread out on the plate. Please click here to view a larger version of this figure.

Figure 3
Figure 3: Intensity ratio histograms for quantifying liposome compositional inhomogeneity. (A) The intensity ratio of DOPE-Atto488 and DOPE-Atto655 fluorescence for each individual liposome plotted as a histogram. (B) A truncated Gaussian distribution illustrating a potential sensitivity issue, either during imaging or in the data treatment step. (C) Fitting the intensity histogram with a Gaussian function makes it possible to extract the mean and standard deviation used to calculate the DI value. (D) A DI value of 0.23 can be translated to 32% of the population differing by more than 23% from the mean molar ratio of the ensemble. (E) Control experiment imaging the same liposomes before and after refocusing and then performing the same data treatment procedure as for the actual experiment. This allows the experimental error for the DI quantification to be determined. Please click here to view a larger version of this figure.

Figure 4
Figure 4: Converting liposome intensity to physical diameter in nm. (A) Intensity histogram of the 50 nm extruded calibration liposomes compared with the DLS data in order to determine the calibration factor. (B) A histogram of the IntSqrt value from the experimental liposomes typically yields a log-normal distribution due to the physical constraints against creating very small liposomes. (C) A non log-normal IntSqrt value histogram indicates either issues with detection sensitivity or that small liposomes were excluded during data treatment. (D) Liposome intensities in (B) are converted to actual liposome diameters using the correction factor. (E) The intensity ratio of each liposome plotted as a function of liposome diameter can now be displayed allowing the investigation of compositional inhomogeneity as a function of liposome size. Please click here to view a larger version of this figure.

Subscription Required. Please recommend JoVE to your librarian.

Discussion

It is important to note that while we describe in detail how the single liposomes assay can be used to study the compositional inhomogeneity between individual liposomes, the platform is very versatile. As previously shown and discussed in the introduction, the protocol can easily be adapted to study aspects of membrane-membrane fusion, protein-membrane interactions, or liposomal drug carrier characterization. For any scientific questions being addressed, the power of the single liposome assay lies in the ability to detect the individual components of the ensemble and thus having a quantitative readout, not biased by ensemble averaging effects.

The single liposome assay is optimally suited for surface imaging modalities such as total internal reflection or confocal microscope, where the Z-direction sectioning can eliminate unwanted background fluorescence from the solution above the liposome layer. However, this aspect is most important for studies where fluorescently labeled compounds, such as peptides, proteins, or other liposomes are added to the solution and remain there during imaging. For example, if the assay is used in the format described in detail here, where the fluorescent dyes are restricted to the immobilized liposomes, the assay could in principle be performed using more broadly available widefield microscopes, providing the detection system is sensitive enough.

A priori there are no limitations with respect to the method used to prepare the liposomes used in the assay. Here we describe in detail the use of a freeze-drying technique. However, previously chloroform-based lipid rehydration or ethanol injection-based techniques have been employed to fabricate liposomes used in the assay5,6. A critical parameter is the inclusion of a minute fraction of biotinylated lipids in the lipid mixture to ensure immobilization (0.05 mol% recommended). Another necessary element is to include a small amount of fluorescently labeled lipid. Overall, the amount of lipid-dye added should be kept as low as possible to both avoid quenching effects and to avoid the lipid-dye from significantly altering the physicochemical properties of the liposome. The lower limit of the amount of lipid dye is set by the concentration, where the stochastic variation in the number of individual lipid dyes per liposome becomes significant compared to the average amount of lipid-dyes (see Larsen et al.6 for an in-depth discussion). Going below this limit will introduce large uncertainties in the extracted intensities. Finally, the amount of lipid dye needs to be high enough for accurate detection of the single liposomes with a good signal-to-noise ratio. While this criterion of course heavily depends on the imaging system, we found that lipid dye concentrations between 0.05 and 0.5 mol% work well in the vast majority of cases.

To choose which lipid dye to include in the liposome, the first prerequisite is that the excitation and emission properties match the illumination and detection capabilities of the imaging system. Second, to increase the sensitivity and accuracy of the method we recommend using dyes with both high quantum yields and photostability, and we have previously successfully used dyes with different excitation wavelengths6,11. Care should be taken in choosing lipid anchors with similar physicochemical properties to ensure that lipid partitioning does not lead to artificially high inhomogeneity. For example, for this protocol, we have anchored both fluorophores using DOPE, while having one fluorophore on DOPE and another on DPPE could lead to fluorophore distribution heterogeneities not related to the inherent inhomogeneity in the liposome population. Especially for dual color imaging it is important to choose dyes with narrow excitation and emission spectra and the smallest spectral overlap as possible to avoid significant bleed-through, cross talk, and fluorescence resonance energy transfer (FRET) between dyes and channels. To avoid artifacts related to variation in the fluorophore environment we prefer to work with dyes, which have been proven experimentally to show extremely low membrane interaction propensity as determined by Hughes et al.26. Finally, if a specific liposome lipid composition is not a hard requirement for the experimental design, it can be beneficial to include up to 10 mol% of negatively charged lipids to ensure that individual liposomes repel each other and are efficiently immobilized as single liposomes12.

An advantage of the single liposome assay is the small experimental volume, which can greatly reduce the amount of expensive or rare compounds used per experiment. Here we describe the use of standard and commercially available chambers with an experimental volume between 150–300 µL. However, custom-made microscope chambers that can reduce the volume to a range of 50–80 µL can be used. Also, the liposome consumption is very low, with a final concentration around 2 µM total lipid.

A disadvantage of the single liposome assay is that it is hard to control the lipid concentration in the chamber due to the variations in immobilization tendencies described above. Additionally, with respect to applying the assay for studying membrane-protein interactions, it is challenging to determine the concentration or number of bound proteins directly from the fluorescence intensity.

When using the liposomes in the assay as membrane model systems (e.g., to study the membrane interaction of fluorescent compounds, peptides, or proteins) it is important to ensure low non-specific binding to the protein components that facilitates the immobilization. If this is not the case, high background signal can be detected that could prevent the detection of the specific binding to the liposomes. If high non-specific background is detected we have previously successfully reduced the background by changing from streptavidin to Neutravidin or avidin.

Performing single liposome detection using other techniques like flow cytometry could potentially offer an increased detection throughput compared to the microscope-based assay. However, given that flow cytometers are optimized for studying much larger and brighter samples like cells, the detection system of most is not sensitive enough to detect the relatively weak fluorescence from an individual liposome. So while new and more sensitive flow cytometry is being developed the microscope-based assay offers the best solution for elucidating liposome inhomogeneities when performed efficiently and monitoring thousands of liposomes per experiment.

Subscription Required. Please recommend JoVE to your librarian.

Disclosures

The authors declare no conflict of interest.

Acknowledgments

This work was funded by the Danish Council for Independent Research [grant number 5054-00165B].

Materials

Name Company Catalog Number Comments
8-well microscopy slides (µ slides) Ibidi 80827 Microscopy slides with glass bottom
Avanti Mini Extrusion kit Avanti Polar Lipids 610000 Consumables (Whatman filters) can be aquired from GE Healthcare
BSA Sigma A9418
BSA-Biotin Sigma A8549
Cholesterol Avanti Polar Lipids 700000 Traded trough Sigma
Computer with FIJI (Fiji Is Just ImageJ) ComDet plugin must be installed. Also, a data handling software (Excel, MatLab, OpenOffice, GraphPad Prism etc.) able to load .txt files will be needed to plot the data
DOPE-Atto488 Atto-Tech AD488-165
DOPE-Atto655 Atto-Tech AD655-165
DOPE-PEG-Biotin Avanti Polar Lipids 880129 Traded trough Sigma
D-Sorbitol Sigma S-6021
Freeze-dryer e.g. ScanVac Coolsafe from Labogene
Glass vials Brown Chromatography 150903 Glass vials that can resist snap-freezing in liquid nitrogen. The 8 mL version of the vials has a size that also fits with the syringes of the extrusion kit
HCl Honeywell Fluka 258148
Heating bath Capable of heating to minimum 65 °C
Heating plate with Magnet stirring Capable of heating to minimum 65 °C
HEPES Sigma H3375
Liquid nitrogen Including container for storage, e.g. Rubber-bath
Magnetic stirring bars VWR 442-4520 (EU)
Microcentrifuge tubes 1.5 mL Eppendorf 0030 120.086 (EU)
Microscope For the images in this protocol a Leica SP5 confocal microscope has been used
Na HEPES Sigma H7006
NaCl Sigma S9888
NaOH Honeywell Fluka 71686
POPC Avanti Polar Lipids 850457 Traded trough Sigma
Streptavidin Sigma S4762
tert-Butanol (2-methyl-2-propanol) Honeywell Riedel-de Haën 24127
Ultrapure water e.g. MilliQ

DOWNLOAD MATERIALS LIST

References

  1. Chan, Y. H. M., Boxer, S. G. Model membrane systems and their applications. Current Opinion in Chemical Biology. 11, (6), 581-587 (2007).
  2. Veatch, S. L., Keller, S. L. Seeing spots: Complex phase behavior in simple membranes. Biochimica Et Biophysica Acta-Molecular Cell Research. 1746, (3), 172-185 (2005).
  3. Sercombe, L., et al. Advances and Challenges of Liposome Assisted Drug Delivery. Frontiers in Pharmacology. 6, 13 (2015).
  4. Hatzakis, N. S., et al. How curved membranes recruit amphipathic helices and protein anchoring motifs. Nature Chemical Biology. 5, (11), 835-841 (2009).
  5. Elizondo, E., et al. Influence of the Preparation Route on the Supramolecular Organization of Lipids in a Vesicular System. Journal of the American Chemical Society. 134, (4), 1918-1921 (2012).
  6. Larsen, J., Hatzakis, N. S., Stamou, D. Observation of Inhomogeneity in the Lipid Composition of Individual Nanoscale Liposomes. Journal of the American Chemical Society. 133, (28), 10685-10687 (2011).
  7. Lohse, B., Bolinger, P. Y., Stamou, D. Encapsulation Efficiency Measured on Single Small Unilamellar Vesicles. Journal of the American Chemical Society. 130, (44), 14372 (2008).
  8. Iversen, L., Mathiasen, S., Larsen, J. B., Stamou, D. Membrane curvature bends the laws of physics and chemistry. Nature Chemical Biology. 11, (11), 822-825 (2015).
  9. Bhatia, V. K., Hatzakis, N. S., Stamou, D. A unifying mechanism accounts for sensing of membrane curvature by BAR domains, amphipathic helices and membrane-anchored proteins. Seminars in Cell & Developmental Biology. 21, (4), 381-390 (2010).
  10. Bhatia, V. K., et al. Amphipathic motifs in BAR domains are essential for membrane curvature sensing. EMBO Journal. 28, (21), 3303-3314 (2009).
  11. Larsen, J. B., et al. Membrane curvature enables N-Ras lipid anchor sorting to liquid-ordered membrane phases. Nature Chemical Biology. 11, (3), 192 (2015).
  12. Larsen, J. B., et al. Membrane Curvature and Lipid Composition Synergize To Regulate N-Ras Anchor Recruitment. Biophysical Journal. 113, (6), 1269-1279 (2017).
  13. Yoon, T. Y., Okumus, B., Zhang, F., Shin, Y. K., Ha, T. Multiple intermediates in SNARE-induced membrane fusion. Proceedings of the National Academy of Sciences of the United States of America. 103, (52), 19731-19736 (2006).
  14. Fix, M., et al. Imaging single membrane fusion events mediated by SNARE proteins. Proceedings of the National Academy of Sciences of the United States of America. 101, (19), 7311-7316 (2004).
  15. Hatzakis, N. S., et al. Single Enzyme Studies Reveal the Existence of Discrete Functional States for Monomeric Enzymes and How They Are "Selected" upon Allosteric Regulation. Journal of the American Chemical Society. 134, (22), 9296-9302 (2012).
  16. Mathiasen, S., et al. Nanoscale high-content analysis using compositional heterogeneities of single proteoliposomes. Nature Methods. 11, (9), 931-934 (2014).
  17. Veshaguri, S., et al. Direct observation of proton pumping by a eukaryotic P-type ATPase. Science. 351, (6280), 1469-1473 (2016).
  18. Tonnesen, A., Christensen, S. M., Tkach, V., Stamou, D. Geometrical Membrane Curvature as an Allosteric Regulator of Membrane Protein Structure and Function. Biophysical Journal. 106, (1), 201-209 (2014).
  19. Kristensen, K., Ehrlich, N., Henriksen, J. R., Andresen, T. L. Single-Vesicle Detection and Analysis of Peptide-Induced Membrane Permeabilization. Langmuir. 31, (8), 2472-2483 (2015).
  20. Christensen, S. M., Bolinger, P. Y., Hatzakis, N. S., Mortensen, M. W., Stamou, D. Mixing subattolitre volumes in a quantitative and highly parallel manner with soft matter nanofluidics. Nature Nanotechnology. 7, (1), 51-55 (2012).
  21. Eliasen, R., Andresen, T. L., Larsen, J. B. PEG-Lipid Post Insertion into Drug Delivery Liposomes Quantified at the Single Liposome Level. Advanced Materials Interfaces. 6, (9), 1801807 (2019).
  22. M√ľnter, R., et al. Dissociation of fluorescently labeled lipids from liposomes in biological environments challenges the interpretation of uptake studies. Nanoscale. 10, (48), 22720-22724 (2018).
  23. Traikia, M., Warschawski, D. E., Recouvreur, M., Cartaud, J., Devaux, P. F. Formation of unilamellar vesicles by repetitive freeze-thaw cycles: characterization by electron microscope and P-31-nuclear magnetic resonance. European Biophysics Journal with Biophysics Letters. 29, (3), 184-195 (2000).
  24. Nele, V., et al. Effect of Formulation Method, Lipid Composition, and PEGylation on Vesicle Lamellarity: A Small-Angle Neutron Scattering Study. Langmuir. 35, (18), 6064-6074 (2019).
  25. Kunding, A. H., Mortensen, M. W., Christensen, S. M., Stamou, D. A fluorescence-based technique to construct size distributions from single-object measurements: Application to the extrusion of lipid vesicles. Biophysical Journal. 95, (3), 1176-1188 (2008).
  26. Hughes, L. D., Rawle, R. J., Boxer, S. G. Choose Your Label Wisely: Water-Soluble Fluorophores Often Interact with Lipid Bilayers. PLoS One. 9, (2), e87649 (2014).

Comments

0 Comments


    Post a Question / Comment / Request

    You must be signed in to post a comment. Please or create an account.

    Usage Statistics