Design and Applications of a Fluorescent Labeling Technique for Lipid and Surfactant Preformed Vesicles.

Amphiphilic molecules such as surfactants, lipids, and block copolymers can be assembled into bilayers and form vesicles. Fluorescent membrane labeling methods require the use of dye molecules that can be inserted into the bilayers at different stages of synthesis. To our knowledge, there is no generalized method for labeling preformed vesicles. Herein, we develop a versatile protocol that is suitable to both surfactant and lipid preformed vesicles and requires no separation or purification steps. On the basis of the lipophilic carbocyanine green dye PKH67, the methodology is assessed on zwitterionic phosphatidylcholine vesicles. To demonstrate its versatility, it is applied to dispersions of anionic or cationic vesicles, such as a drug administrated to premature infants with respiratory distress syndrome, or a vesicle formulation used as a fabric softener for home care applications. By means of fluorescence microscopy, we then visualize the interaction mechanisms of nanoparticles crossing live cell membranes and of surfactants adsorbed on a cotton fabric. These results highlight the advantages of a membrane labeling technique that is simple and applicable to a large number of soft matter systems.

Introduction

Amphiphilic molecules such as double-tailed surfactants, lipids, or block copolymers can assemble in aqueous solutions into bilayers and form various structures including liquid-crystal smectic or cubic phases, multiconnected sponge analogues, or closed vesicles. Membrane composition, structure, and elastic properties have earned widespread attention because membrane-based colloids are implemented in many industrial and medical applications.[1] In chemistry, vesicles enter into topical home care and personal care formulations through their ability to adsorb onto various substrates.[2] In nanomedicine, an important application relates to drug delivery and the use of liposomes as nanocarriers in anticancer chemotherapies.[3] Under the impetus of biology, there has been recently a significant increase of novel fluorescence techniques that allow the visualization of membranes via optical microscopy and follow complex phenomena in real time.[4−7] The interest for fluorescence microscopy has further increased with the recent discovery of super-resolution imaging techniques that allow to go beyond the diffraction limit with optics.[8]

In this context, it is important to control the membrane labeling techniques. For surfactants, the mono- and bilayer labeling takes advantage of fluorescently tagged amphiphiles[9−11] or solvatochromic surfactant dyes,[12] the latter having the property to change the emission wavelength depending on their environment. With surfactants, however, it is often found that the dye insertion modifies the overall assembly properties, inducing a disruption of micellar or membrane structures. With lipids, the most common labeling methods are based on lipid film hydration[5,6,13,14] or electroformation[5,15] in which a small proportion of fluorescently modified lipids is added beforehand. This technique is suitable for the production of giant unilamellar vesicles and has shown excellent results.[5−7,13] The previous methods are however not suitable for living cell membranes or preformed vesicles. Here, we use the term “preformed vesicles” to describe surfactant or lipid compartments obtained using various chemical or biological processes and which structure must be preserved upon staining. Preformed vesicles are, for instance, present in cosmetics or home care formulations.[16−18] In biology, they are found in lung fluids or they are the exosomes secreted by live cells associated with regular trafficking functions.[19−21] In recent literature, we have not found any vesicle labeling protocol that is simple enough and suitable for a broad range of amphiphilic molecules, including surfactants and lipids. After evaluating several biology-based staining methods, we show here an easy and versatile solution to this problem.

In cellular biology, membrane labeling is commonly achieved using lipophilic carbocyanine dyes.[22,23] These dyes intercalate spontaneously and noncovalently into the bilayers, thanks to their aliphatic pending chains, and rapidly stain the outer cellular membrane through lateral diffusion.[24,25] With time, the dye molecules also spread to intracellular organelles as a result of lipid exchange. Some of the well-known examples are the carbocyanine fluorescent dyes PKH (e.g. PKH26 and PKH67) and DiO, DiI, DiD, and DiR which cover the visible and infrared emission spectrum.[21,22,26−28] Direct labeling using lipophilic dyes can however lead to significant issues. To speed up the insertion of fluorescent molecules into the membrane, large concentrations of lipophilic molecules and the use of specific solvents are prescribed, leading to further complex separation protocols.[20,21,27] For preformed vesicles, the difficulty is augmented because the vesicles and the nonsolubilized dye aggregates have comparable sizes and densities. In such cases, more sophisticated separation techniques based on sucrose gradient or on chromatography are to be implemented.[21]

Herein, we develop a versatile labeling membrane protocol that is suitable to both surfactant and lipid preformed vesicles and requires no separation or purification steps. The methodology is based on direct staining using the PKH67 lipophilic green dye (Scheme ). The protocol design and optimization are first assessed using dipalmitoylphosphatidylcholine (DPPC) preformed vesicles. The DPPC vesicle features such as the size and charge are thoroughly investigated and found to be in agreement before and after the labeling treatment. The protocol is then extended to two types of vesicle dispersions: (i) an exogenous pulmonary surfactant drug administrated to premature infants with respiratory distress syndrome and (ii) a vesicle formulation used as a conditioner in home care applications. These vesicles differ by their surface charges, the former being anionic and the latter being cationic. In the end, we show the benefits of such a staining protocol in two applications, namely, the role of pulmonary surfactant in the particle internalization in live cells and the surfactant deposition on cotton fabric.

Scheme 1

Chemical Structure of PKH67 Amphiphilic Dye

The molecule is made of fluorescent head groups Z1 and Z2 emitting in the green (502 nm) emission and two aliphatic chains that are inserted into the bilayer (see Materials and Methods for details).

Results and Discussion

PKH67 Behavior in Polar Solvents

Dynamic light scattering was used to study the PKH67 dye solubility in solvents such as ethanol, deionized (DI) water, phosphate-buffered saline (PBS), and Diluent C. In ethanol, no aggregates nor micelles are observed with light scattering, indicating a good solubility in this medium. Diluent C is the commercially available solvent for carbocyanine PKH dyes. It is used primarily to maintain the cell viability while maximizing the dye solubility in the culture medium. As Diluent C components may change the vesicle properties (it contains surfactants, proteins and complexing agents), its use requires post-labeling treatment such as purification and separation steps. Alternative solvents to Diluent C are water and PBS, the latter being the reference buffer in many biological studies. Figure a–c displays the intensity distributions of 5 μM PKH67 dispersions in DI water, PBS, and Diluent C, respectively. The size distributions are obtained from the deconvolution of the scattered light autocorrelation function and reveal the existence of nanometer (for DI water and Diluent C) to micrometer (for PBS) sized aggregates. Obtained by fitting, the continuous lines show maxima at 400 nm, 1.40 μm, and 430 nm. The results suggest that the dyes are partially soluble in either of the solvents, a result that is explained by the hydrophobicity of the fluorescent molecule (Scheme ). From the aggregate sizes, it is inferred that the PKH67 solubility in DI water is comparable to that in Diluent C and better than that in PBS. Because the membrane labeling efficiency is linked to this solubility, we conclude that DI water can be an alternative solvent to Diluent C for staining vesicles.

Figure 1

Dynamic light scattering size distributions of 5 μM PKH67 dispersions using (a) DI water, (b) PBS, and (c) Diluent C as a solvent.

Incorporating PKH67 Dyes into Lipid Vesicles

Systematic studies were performed by mixing DPPC vesicles at a concentration of 1 g L–1 in DI water together with dye dispersions of increasing concentrations. We followed the scheme displayed in Figure for the sample preparation and increased the PKH67 molar concentration from 0.2 to 5 μM. These concentrations correspond to a dye-to-lipid ratio varying between 1:300 and 1:7000. It was found that at a ratio of 1:1400, the dyes did not alter the vesicular structure and the vesicles were fluorescent. The vesicle properties prior and after labeling were further examined using fluorimetry, light scattering, and microscopy.

Figure 2

Schematic representation of the protocol used for labeling DPPC vesicles in DI water.

Figure a shows the fluorescence excitation (λex = 490 nm) and emission (λem = 502 nm) spectra for a 50 nM PKH67 aqueous suspension and for a PKH67/DPPC dispersion prepared in DI water as described above. In particular, the dye-to-lipid ratio was set at 1:1400. The fluorescence signal observed for DPPC/PKH67 is 40 times higher than that of PKH67 alone at the same concentration. The low fluorescence level found for pure PKH67 dispersion is explained by the phenomenon of aggregation-caused quenching.[29,30] Aggregation-caused quenching takes place when poorly soluble molecules associate with micelles and aggregates, resulting in concentration-induced emission quenching. In contrast, the strong signal obtained in the presence of lipids suggests that the dyes have been incorporated in the membranes and are spatially dispersed. There, the distance between the dye molecules is estimated at 30 nm, preventing the self-quenching to take place. In both cases, the emission occurs at the same wavelength, showing that the presence of DPPC molecules does not alter the PKH67 photophysical properties.

Figure 3

(a) Fluorescent excitation and emission spectra of a 50 nM PKH67 aqueous dispersion in the presence and absence of DPPC vesicles. The lipid concentration is 50 mg L–1 and the temperature is 25 °C. The arrow indicates that with DPPC, the fluorescence signal is enhanced by a factor 40. (b) Dynamic light scattering size distribution of PKH67 labeled and unlabeled DPPC vesicles at 1 g L–1. (c,d) Phase-contrast and fluorescence optical microscopy images showing DPPC vesicles adsorbed at a PDADMAC-coated glass substrate. (e) Merge of (c,d) showing colocalization.

Labeled and unlabeled vesicles were further studied using dynamic light scattering, zeta potential measurements, and optical microscopy. Figure b displays the size distribution obtained for a 1 g L–1 dispersion with and without PKH67. In the two experiments, the distribution peak is around 0.8–1 μm. In dilute solutions, the vesicles observed in 60× phase-contrast microscopy are identified as separated and immobile objects adsorbed at the glass surface (Figure c). With fluorescence, the vesicles appear as bright spots colocalized with those observed in phase contrast (Figure d). An analysis of the merge image (Figure e) leads to the result that more than 80% of the treated vesicles have been labeled and that their size is not modified by the staining. These findings confirm that the above protocol is efficient and does not change the overall vesicle geometry.

Labeling Biological Lipid and Industrial Surfactant Vesicles

The previous method was applied to vesicles of different origin and surface charge. Vesicles made from biological lipids coming from the pulmonary surfactant drug Curosurf or those made from the commercial double-chain surfactant TEQ were assessed. We recall that Curosurf is a lung fluid substitute used for the treatment of respiratory distress syndrome.[31−33] Its composition comprises a wide variety of lipids (Scheme ) which impart an average negative charge to the membrane (Table ). The TEQ surfactant in turn is a quaternary ammonium esterquat entering topical softener formulations. In Supporting Information S1 and S2, cryogenic transmission electron microscopy (cryo-TEM) images confirm that both Curosurf biological lipids and TEQ esterquat surfactants associate locally into bilayers with a thickness of 4–5 nm and that on a larger scale, the bilayers form unilamellar and multivesicular vesicles.[31]Figure a,b display the size distributions for PKH67-labeled and PKH67-unlabeled Curosurf and TEQ dispersions, respectively, obtained from light scattering at 1 g L–1. In this assay, the Curosurf vesicle size is controlled via extrusion through a 100 nm pore polycarbonate membrane.[34] The data indicate that there is a good superimposition of the distributions for the two dispersions. Electrophoretic mobility experiments show that the zeta potential remains unchanged with and without labeling, the values being around −20 and +60 mV for Curosurf and TEQ, respectively (Table ). In Figure c–e,f–h, phase-contrast and fluorescent microscopy images of native nonextruded Curosurf and TEQ are presented. The figures reveal well-contrasted spherical objects of average size 1 μm. As for DPPC, the vesicles appear as bright spots colocalized with those observed in phase-contrast mode. Of them, 85% are fluorescent and their size distribution remains unchanged. These findings confirm that the staining protocol developed for zwitterionic DPPC vesicles can be extended to anionic lipids from biological origin and to cationic surfactants coming from an industrial source. In the next sections, we take advantage of this labeling technique to investigate the interaction mechanisms of nanoparticles crossing live cell membranes and of surfactants adsorbed on cotton fabric.

Scheme 2

Chemical Formulae of (a) DPPC, (b) Most Common Lipids Present in Curosurf Formulation, and (c) Ethanaminium, 2-Hydroxy-N,N-bis(2-hydroxyethyl)-N-methyl-esters Abbreviated as TEQ; Curosurf Contains Zwitterionic (PC, Sphingomyelin, and Phosphatidylethanolamine) and Anionic Lipids (Phosphatidylinositol and PG); Its Composition Is Available in Supporting Information Table S1.

Table 1

Hydrodynamic Diameter (DH) and Zeta Potential (ζ) of the Native and Fluorescent DPPC, Curosurf, and TEQ Vesiclesa

	DH (nm)		ζ (mV)
vesicles	native	fluorescent	native	fluorescent
DPPC	1350	1080	0	0
pulmonary surfactant substitute Curosurf*	130	135	–20	–18
double-tail esterquatsurfactant TEQ	550	530	+60	+58

* indicates that the vesicles are extruded with a 100 nm pore polycarbonate membrane.

Figure 4

(a) Dynamic light scattering size distribution of 1 g L–1 extruded Curosurf vesicles with and without PKH67 dyes. The extrusion was made using a 100 nm pore polycarbonate membrane. (b) Similar to (a) for native TEQ vesicles. (c,d) Phase-contrast and fluorescence optical microscopy showing nonextruded Curosurf vesicles. (e) Merge of the patterns found in phase-contrast and fluorescence. (f–h) Same as (c–e) for TEQ quaternary ammonium esterquat surfactant.

Role of Pulmonary Surfactant in the Cellular Uptake of Nanoparticles

When inhaled, nanoparticles are able to reach the respiratory zone and enter in contact with the alveolar region.[35−37] Various scenarios of nanoparticles passing from the alveolar spaces toward the blood circulation have been examined, and in some instances, translocation has been demonstrated.[37,38] Our research group has recently explored the behavior of 50 nm particles and of surfactant substitutes in controlled physicochemical conditions. Three types of surfactant mimetics, including the exogenous substitute Curosurf, were assessed together with aluminum oxide, silicon dioxide, and latex nanoparticles.[34] The result that emerged from this survey (particles were selected to display different shapes and surface charge densities[39]) was the observation of the spontaneous nanoparticle/vesicle aggregation induced by Coulomb attraction. The aggregation was strongly enhanced for oppositely charged species. Labeling pulmonary surfactant vesicles, as shown previously, represents hence an opportunity to further study the aggregate structure. Figure a displays a high magnification view of a silica/Curosurf aggregate observed under phase-contrast (Figure a1), green (Figure a2), and red (Figure a3) emission. In this assay, the particles are aminated and fluoresce in the orange-red at 590 nm because of the rhodamine molecules present in the silica structure. The experimental conditions are a total concentration of 1 g L–1, a volumetric lipid-to-nanoparticle ratio of 2, and a temperature of 37 °C.[34] The merge image of Figure a2,a3 exhibits an excellent superimposition of the green and red channels over the entire object, indicating that the aggregates contain both fluorescent species (Figure a4). These results, together with those of Supporting Information S3, ascertain that the aggregates are made of vesicles and particles and that both are intertwined at the micron scale (Figure b).

Figure 5

(a) Silica nanoparticle/Curosurf aggregate observed under phase-contrast (a1), green (a2), and red (a3) illumination. The silica fluoresce in the orange-red at 590 nm and the vesicles are labeled with a green fluorescent dye (PKH67) emitting at 502 nm. The merge image of the red and green signals is shown in (a4). (b) Schematic representation of a nanoparticle/Curosurf aggregate derived from optical microscopy. (c) Bright field (c1) and confocal microscopy (c2,c3) of A549 alveolar epithelial cells incubated with silica nanoparticle/Curosurf aggregates at concentrations 0.10/0.015 g L–1. The nuclei are labeled in blue with DAPI. The white lines are the cell contours. (d) Three-color merge images showing cells (d1), plasma membrane (d2), and cytoplasm (d3). The bars in (d2,d3) are 2 and 4 μm, respectively. (e) Schematic representation of a cell with adsorbed and internalized silica nanoparticles.

In the alveolar spaces, inhaled nanoparticles enter first in contact with the lipid monolayer at the air–liquid interface and then with the hypophase.[40,41] This interaction leads to an aggregate formation comparable to the one discussed previously and is susceptible to modify the fate of the nanoparticles toward the alveolar cells located beneath the hypophase. In particular, the lipids are expected to mitigate the protein adsorption.[42] Here, the double labeling permits to visualize the vesicle and particle localization following the cellular uptake. A549 lung epithelial carcinoma cells were incubated for 4 h with the nanoparticle/vesicle aggregates shown in Figure a. Because of their size and density, the aggregates sedimented at the bottom of the Petri dish[36,43] and enter in contact with the cell layer. Figure c exhibits confocal images of a cluster of five cells in bright field (Figure c1) and in green and red fluorescence (Figure c2,c3, respectively). The nuclei are stained in blue with 4′,6-diamidino-2-phenylindole (DAPI). Figure d exhibits the merge image of the three fluorescent signals, together with close-up views of the plasma membrane and of the cytoplasm. These images demonstrate that in the presence of lipid membranes, the particles are internalized in the cells in a significant amount (see Supporting Information S4 for additional data). In this respect, the vesicular membrane seems not to have a protective role toward cells, as it is the case with supported lipid bilayers.[31] Yellow patches associated with nanoparticle/vesicle colocalization are essentially visible at the outer cellular range, but not internally or in a lesser amount. This suggests that after passing through the outer lipid barrier, a separation of the two components occurs and that the preformed aggregates break. Figure e illustrates schematically this process. To our knowledge, it is the first time that assays using membranes of biological origin reveal a separation mechanism for the organic and inorganic species in cells.

Double-Chain Surfactant Deposition on Cotton Fibers

Recently, we show that the PKH67-labeled vesicles made from the TEQ double-tailed surfactant could also be used to decipher the interaction mechanism with cellulose nanocrystals.[17] Obtained from natural cellulose,[44] cellulose nanocrystals are rod-shaped particles in the form of 200 nm laths. Being negatively charged, these anisotropic particles interact strongly with the TEQ cationic vesicles and form mixed aggregates comparable to those obtained with the silica nanoparticles and discussed in the previous section. Examples are given in Supporting Information S5.

To evaluate the performances of a home care product such as a fabric softener, it is more convenient to use the actual fabric materials for deposition assessment. Pertaining to the deposition on cotton, important issues are the amount of adsorbed surfactants and the morphology of the deposited layer, questions to which labeled vesicles could provide an answer.[45−48] In a typical experiment, a 1 × 1 cm2 piece of woven cotton fabric is treated with a dispersion containing TEQ fluorescent vesicles at 1 g L–1 during 10 min and then rinse with DI water. The PKH67 concentration was set at 6 μM for these assays, which is slightly higher than for DPPC and Curosurf vesicles. Figure a displays a phase-contrast microcopy view of a treated fabric submerged in water. There, the 200 μm sized yarns are recognizable and separated by large voids. Additional scanning electron microscopy (SEM) data confirm the woven structure (Supporting Information S6). In the emission mode (Figure b), the fibers exhibit an intense fluorescence signal, illustrating that TEQ surfactants have been adsorbed on the cellulose substrate in a significant manner. The bright spots (arrows) are assigned to large TEQ vesicles coating the cellulose substrate. The vesicle adsorption is explained by Coulomb attraction between the cationic vesicles and the anionic fibers. Control experiments performed with unlabeled surfactant show that in the present conditions, the fibers are not fluorescent and appear dark (Supporting Information S7). An analysis of the distribution of the fluorescence signal along the fiber cross section shows an enhanced signal on the edges, the intensity showing an M-shaped signal (Figure c,d and the inset in 6c). Such M-shaped signals are characteristic from fluorescence emission arising from the fluorescent surfaces or interfaces.[6,49] Concerning the nature of the absorbed layer, it could come either from supported lipid bilayers or from a supported vesicle layer. In the first scenario, the fibers would be coated with a single or with a stack of surfactant bilayers, whereas in the second scenario, the vesicles remain intact and stick to the cellulose fibers. Figure b,d actually suggest that both scenarios are here plausible. These two types of surfactant depositions are schematically presented in Figure e. In conclusion, the present technique allows a direct visualization of surfactants adsorbed on cotton fabric and should permit to assess the formulation deposition performance.

Figure 6

Phase-contrast (a,c) and fluorescence (b,d) microscopy images of a woven cotton fabric treated with fluorescent vesicles. Magnification are 20× for (a,b) and 40× for (c,d). In (b), fluorescent vesicles are indicated with yellow arrows. The inset in (c) displays the fluorescence profile along the line A depicted in (d). The fluorescence maxima are ascribed to the fiber edges. (e) Schematics of surfactant deposition on cotton fibers assuming either supported lipid bilayer or supported vesicle layer models.

Conclusions

In this work, we report a fluorescent labeling technique based on the spontaneous insertion of lipophilic molecules into membranes made either from surfactants or lipids. The protocol is derived from methods developed for live cell imaging. Thanks to an accurate dosage of the dye molecules in solution, here the PKH67 lipophilic green fluorophore, issues related the removal of the excess dyes or to the use of a specific solvent are avoided. The PKH67 concentration is namely adjusted to that of the lipids in a way that the fluorescence signal remains high and that at the micron scale, the vesicular structures are not altered by the dye insertion. Optimum ratios are in the range of one dye molecule per 1400 lipids. The vesicular resilience upon dye addition is evaluated on DPPC vesicles using light scattering, zeta potential, and fluorescent microscopy. With the above dye-to-lipid ratio, we ensure moreover that phenomena such as aggregation-caused quenching do not occur and mitigate the fluorescence intensity. As anticipated, we also found that separation and centrifugation techniques usually necessary to remove the dye excess are not required here. To demonstrate its versatility, the protocol is applied to two other vesicular systems, a lipid drug administrated to premature infants with respiratory distress syndrome (Curosurf) and a surfactant formulation used as a fabric conditioner. In Curosurf, the vesicles are negatively charged, whereas in the surfactant formulation, they are positively charged. Interestingly, the amounts of dye added to optimize the fluorescence signals depend slightly on the system or on the electrostatic charge, around 1 μM for a lipid or surfactant solution at 1 g L–1. To complete this study, we illustrate the benefits of this labeling method in evaluating the internalization of silica nanoparticles in cells after their interactions with Curosurf vesicles. With the fabric softener formulation, we also investigate the deposition of surfactant on cotton fibers. In both examples, we are able to visualize using fluorescence microscopy the fate of the vesicles after their interactions with live cells or with cotton. These results allow us to establish the interaction patterns specific to each system and to assign the relevant nanoscale interaction mechanisms thanks to fluorescence microscopy.

Materials and Methods

PKH67 Fluorescent Dye

Developed for live cells, PKH linkers are lipophilic molecules that incorporate spontaneously into the lipid bilayer. Their photophysical properties cover the visible and infrared emission spectrum and the molecules are described as nontoxic effects toward cells.[20,22,26] The 1 mM PKH67 solution in ethanol was purchased from Sigma-Aldrich and used as received. The molecule is made of two fluorescent head groups and two aliphatic chains. Its chemical structure as provided by the supplier is shown in Scheme . PKH67 fluoresces in the green, its excitation, and emission wavelengths being at 410 and 502 nm, respectively. More information can be found in the website in Supporting Information S5 and S7.

Phospholipids and Surfactants

DPPC (Scheme a) was purchased from Sigma-Aldrich. It is a phosphatidylcholine (PC) lipid with a gel-to-fluid phase-transition temperature of 41 °C.[50] For vesicle preparation, DPPC was initially dissolved in methanol at 10 g L–1. After a 30 min solvent evaporation at low pressure and 60 °C, a lipid film was formed on the glass surface. It was hydrated with the addition of DI water at 60 °C and stirred at atmospheric pressure for another 30 min. DI water was added to obtain a 1 g L–1 dispersion. In these conditions, DPPC was found to form uni- and multilamellar vesicles.

Curosurf, also called poractant Alfa (Chiesi Pharmaceuticals, Parma, Italy), is an extract of whole mince of porcine lung tissue purified by column chromatography.[33,51] It is indicated for the rescue treatment of respiratory distress syndrome in premature infants and administered intratracheally at a dose of 200 mg kg–1. Curosurf was kindly provided by Dr. Mostafa Mokhtari and his team from the neonatal service at Hospital Kremlin-Bicêtre, Val-de-Marne, France. Curosurf is produced as an 80 g L–1 phospholipid and protein suspension where the phospholipids are assembled in the form of multilamellar anionic vesicles. It contains, among others, PC lipids, sphingomyelin, phosphatidylglycerol (PG), and the membrane proteins SP-B and SP-C.[32,34,40,52] Some of the lipids present in Curosurf are displayed in Scheme b (Supporting Information Table S1). According to the manufacturer, the pH of Curosurf is adjusted with addition of sodium bicarbonate at pH 6.2.[53] In the conditions used in this work, Curosurf is organized in anionic vesicles with a large size distribution (50 nm to 2 μm), as shown by cryo-TEM.[54] The gel-to-fluid transition of Curosurf membranes was measured and found at TM = 29.5 °C (Supporting Information S8).

The esterquat cationic surfactant ethanaminium, 2-hydroxy-N,N-bis(2-hydroxyethyl)-N-methyl-esters with saturated and unsaturated C16–18 aliphatic chains (Scheme c), abbreviated as TEQ hereafter, was provided by Solvay. Its gel-to-fluid phase-transition temperature is observed at 60 °C, and the counterions associated with the quaternized amines are methyl sulfate anions.[17] TEQ is a cationic surfactant widely used in the industry as the main component of hair and fabric conditioners. It forms unilamellar and multivesicular vesicles in water, as shown by cryo-TEM in Supporting Information S2.[17]

Silica Nanoparticles

The positively charged silica particles were synthetized using the Stöber synthesis.[31,55] Briefly, fluorescent silica seeds were prepared in three steps. The rhodamine derivative rhodamine red c2 maleimide (Aldrich) was first covalently bound to silica precursor (3-mercaptopropyl)-trimethoxysilane (MPS, Aldrich). The rhodamine–MPS compound was then mixed with a tetraethyl orthosilicate silica precursor (TEOS, Aldrich) and the Stöber synthesis was performed. With this approach, the dyes were covalently bound to the silica matrix. In a third step, a nonfluorescent silica shell was grown with TEOS to increase the particle size and prevent leakage of the dyes out of the particles. Functionalization by amine groups was then performed, resulting in a positively charged coating.[55] Aminated silica was synthesized at 40 g L–1 and diluted with DI water at pH 5. The hydrodynamic and geometric diameters were determined as DH = 60 nm and D0 = 41.2 nm, respectively. The fluorescence properties were characterized using a Cary Eclipse fluorimeter (Agilent), with leading excitation and emission peaks at 572 and 590 nm, respectively. A particle identity card displaying UV–vis spectrometry, fluorimetry, TEM, and light scattering data is provided in Supporting Information S9.

Cellulose Fibers

Woven cotton fabric (1 × 1 cm2) was used to investigate the softener surfactant deposition onto fibers. The fabrics are made of cotton yarns of diameters 300 ± 100 μm, each of them being constituted of approximately 20 cotton fibers of diameter 10–20 μm. A representative SEM image of the woven cotton fabric used in this work is provided in Supporting Information S6. All fabrics were first treated with DI water for 10 min and dried at 35 °C under air circulation before use.

Vesicle Labeling

The 1 mM commercial PKH67 stock solution in ethanol was first diluted 10 times using pure ethanol. The diluted stock (2 μL) was then mixed rapidly with 100 μL of DI water and vortexed for 10 s. It was then added to 100 μL of a DPPC or Curosurf vesicle aqueous dispersion at the concentration of 2 g L–1. The final dye concentration was then 1 μM. Rapid vortexing for ca. 10 s was followed to ensure the dye insertion in the lipid membranes. The dispersion was let to rest in the dark for 10 min. For labeling TEQ vesicles, the dye concentration used was six times larger than the previous one. In particular, 2 μL of the PKH67 stock solution was rapidly mixed with 100 μL of DI water and vortexed for ca. 10 s. This solution was then added to 100 μL of a TEQ vesicle dispersion at a concentration of 2 g L–1 and the mixture was vortexed.

Sample Preparation

To evaluate the PKH67 solubility, solutions were prepared at 5 μM in different solvents, such as ethanol, PBS, Diluent C, and water. Diluent C is a commercial product developed for cell labeling. It contains glycerol (50%), Triton X100 (0.15 wt %), bovine serum albumin (200 mg L–1), dithiothreitol (1 mM), ethylenediaminetetraacetic acid (EDTA) (0.1 mM), NaCl (250 mM), and Tris–HCl (10 mM). Pertaining to the vesicles, Curosurf, DPPC, and TEQ dispersions were prepared at room temperature using DI water and a concentration of 2 g L–1. Before use, the dispersions were characterized by dynamic light scattering and ζ-potential measurements. In some cases, the dispersions were extruded through a polycarbonate membrane following a protocol described in ref (34) to reduce the vesicle size dispersity. To study the interaction of silica nanoparticles with lipids, we use the method of continuous variation developed by Job to determine the stoichiometry of binding (macro)molecular species in solutions.[56] The method is here combined with static and dynamic light scattering, leading to Job scattering plots.[17,34,53,57] In brief, stock solutions (1 mL) were prepared in the same conditions of pH and concentration (1 g L–1) and mixed at the volumetric ratios X = VLP/VNP = 2, where VLP and VNP are the volumes of the surfactant and particle solutions. The particle and lipid concentrations (0.66 and 0.33 g L–1, respectively) correspond to identical surface area concentrations in membrane and in silica.[34]

Deposition of Labeled TEQ Vesicles on Cotton Fabrics

PKH67 (10 μL) was dispersed in 1500 μL of DI water and mixed rapidly with vortex. This mixture was then added in 1500 μL of TEQ prepared at 1 g L–1 and mixed rapidly with the vortex. Cotton fabric (0.03 g) cut in small pieces was immersed for 10 min in this dispersion.

Cell Culture

Adenocarcinomic human alveolar epithelial cells A549 (ATCC reference CCL-185) were grown in T75-flasks in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin. Exponentially growing cultures were maintained in a humidified atmosphere of 5% CO2 and 95% air at 37 °C. When the cell confluence reached 80%, cell cultures were passaged using trypsin–EDTA. All products used in cell culture came from Gibco, Life Technologies.

Light Scattering

The scattering intensity IS and hydrodynamic diameter DH were measured using the Zetasizer NanoZS spectrometer (Malvern Instruments, Worcestershore, UK). A 4 mW He–Ne laser beam (λ = 633 nm) is used to illuminate the sample dispersion, and the scattered intensity is collected at a scattering angle of 173°. The second-order autocorrelation function g(2)(t) is analyzed using the cumulant and CONTIN algorithms to determine the average diffusion coefficient DC of the scatterers. The hydrodynamic diameter is then calculated according to the Stokes–Einstein relation, DH = kBT/3πηDC, where kB is the Boltzmann constant, T the temperature, and η the solvent viscosity. Measurements were performed in triplicate at 25 °C after an equilibration time of 120 s.

Zeta Potential

Laser Doppler velocimetry (Zetasizer, Malvern Instruments, Worcestershore, UK) using the phase analysis light scattering mode and a detection at an angle of 16° was performed to determine the electrophoretic mobility and zeta potential of the different studied dispersions. Measurements were performed in triplicate at 25 °C, after 120 s of thermal equilibration.

Phase-Contrast and Fluorescence Optical Microscopy

Images were acquired on an IX73 inverted microscope (Olympus) equipped with a 60× objective. An Exi Blue camera (QImaging) and Metaview software (Universal Imaging Inc.) were used as the acquisition system. The illumination system “Illuminateur XCite Microscope” produced a white light, filtered for observing a green signal in fluorescence (excitation filter at 470 nm—bandwidth 40 nm and emission filter at 525 nm—bandwidth 50 nm). Thirty microliters of the vesicle dispersion (DPPC, Curosurf or TEQ) or nanoparticle/vesicle dispersions (silica particles/Curosurf or nanocellulose/TEQ) were deposited on a glass plate and sealed into a Gene Frame dual-adhesive system (Abgene/Advanced Biotech). In the case of DPPC and Curosurf vesicles, the glass slides were coated using a cationic polymer (PDADMAC) to improve their adhesion properties based on electrostatic interaction. Images were digitized and treated by the ImageJ software and plugins (http://rsbweb.nih.gov/ij/).

Confocal Microscopy

Images were acquired on an LSM 710 microscope (Zeiss) equipped with a 40× immersion objective and a temperature and carbon dioxide content controller. The A549 cells (ATTC) were first seeded on a glass slide in a six-well plate at 200 000 cells per well. After 48 h of growth in complete white DMEM, the cells were rinsed with PBS and incubated 4 h at 37 °C with a silica/Curosurf dispersion. The concentrations are 0.1 g L–1 for the particles and 0.015 g L–1 for Curosurf in white DMEM. Cells were then fixed with PFA and nuclei were stained in blue with DAPI. Finally, slides were sealed into a Gene Frame dual-adhesive system (Abgene/Advanced Biotech). Images were digitized and treated by the ImageJ software and plugins (http://rsbweb.nih.gov/ij/).

Absorbance and Fluorescence Measurements

A UV–visible spectrometer (SmartSpecPlus from BioRad) was used to measure the absorbance of PKH67 aqueous dispersions at 0.05 and 2 μM. The fluorescence properties of PKH67 at 0.05 and 2 μM and of DPPC/PKH67 at 0.05 g L–1/0.05 μM were characterized using a Cary Eclipse fluorimeter (Agilent) with PMT at 860, 980, and 860, respectively.