R. Liuzzia,
S. Gallierb,
S. Ringlerb,
S. Caserta*ac and
S. Guidoac
aDipartimento di Ingegneria Chimica dei Materiali e della Produzione Industriale (DICMAPI), Università di Napoli Federico II, P.le Tecchio, 80, 80125, Napoli, Italy. E-mail: sergio.caserta@unina.it
bDanone Nutricia Research, Uppsalalaan 12, Utrecht 3584 CT, The Netherlands
cConsorzio Interuniversitario Nazionale per la Scienza e Tecnologia dei Materiali (INSTM), UdR INSTM Napoli Federico II, P.le Tecchio, 80, 80125, Napoli, Italy
First published on 7th November 2016
Phospholipids, which are amphiphilic biomolecules composed of a polar head group and two nonpolar fatty acid tails, play a central role in cellular and body functions. The most common phospholipid is phosphatidylcholine (PC), which is also widespread used as a surfactant in colloidal systems, such as oil-in-water (O/W) emulsions, for several applications ranging from food to cosmetics. In these systems, PC tends to arrange at the interface between polar and nonpolar phases. Specific identification of these molecules at the interface is still an ambitious task. In this work, we propose immunofluorescence and confocal microscopy as valuable tools for the localization of PC at the interface of O/W emulsions, where phospholipids are used as surfactants. The protocol required the use of a primary antibody (TEPC-15), specific for choline, and a secondary fluorescently-labelled antibody, which selectively binds to the primary antibody. By using this approach, we were able to visualize choline-based phospholipids in cell membranes, lipid-based emulsions and in PC films and to estimate molecule concentration by image analysis techniques. We also investigated the interference of proteins on the staining procedure. We attributed this interference to possible molecular interactions close to the interface, which were found to depend on protein concentration. This innovative approach can have a relevant impact in a wide range of interfacial engineering applications, such as tuning emulsion microstructure and stability.
The experimental determination of the interfacial concentration of PC, as well as of other surface-active ingredients, is an ambitious and still open task.15 An ideal technique should enable the specific identification of PC at the interface, even in the presence of other species, including different phospholipids. To our best knowledge, however, there are no methods or techniques that allow one to determine in a specific way the presence and the concentration of PC at the interface in a complex system. Pendant drop tensiometry16–19 and shear rheology20 are the most common techniques to investigate the role of phospholipids at the interface. Although these techniques give important parameters21 specific for the whole system, they are not useful for the quantification of molecules at the interface.
Microscopy could represent a valid tool to obtain a general overview of the system morphology, also to investigate chemical composition, and in particular to localize and quantify interfacial concentration of specific molecules. In particular, confocal laser scanning microscopy (CLSM) has been successfully used to investigate structure and lateral organization of molecules at the droplet interface22–24 by using phospholipid analogs fluorescently labeled on the head group or acyl chain.25,26 Although this procedure allows the visualization of phospholipid analogs, certain fluorescent probes may dramatically affect the properties of the individual lipid molecules and could alter the interfacial microstructure inducing local phase separation, if used at high concentration. For this reason, concentrations are usually kept very low to minimize the impact of the phospholipid analogs on the system.27
One of the most specific and well-established methods in cell biology, which is often used to localize specific molecules with high spatial resolution, is immunofluorescence microscopy.28,29 Immunofluorescence requires the use of fluorescently labeled antibodies, which specifically bind to the molecule of interest, thus enabling their localization by confocal microscopy. Detection can occur in a direct way, by targeting antigens with a single fluorophore-labeled antibody, or in an indirect way through binding of a fluorophore-labeled secondary antibody targeted to the primary one.28,30 Protocols for immunofluorescence assays are not straightforward and depend on the sample and on the molecules to detect. Fixing of cells, staining procedures, washing steps to remove unbound antibody are some of the fundamental steps required for a correct visualization. It is worth mentioning that in the case of liquid samples washing steps are not trivial. Consecutive concentration and dilution steps by sample centrifugation can possibly achieve partial washing, even if this protocol could potentially lead to loss of interfacial material. Without the washing steps, the quality of the acquired images decreases, since a diffused background signal, due to the presence of unbound randomly distributed fluorophore-labeled antibody, leads to a poor signal-to-noise ratio. The fine-tuning of the experimental protocol depends on the sample type and the specific application.31 This method, to our knowledge, has never been applied to the investigation of multiphase fluids, such as emulsions. A commercial anti-PC antibody, TEPC-15, has been already used for recognition of PC in plaques from subgingival and supragingival in patients with periodontal diseases.32 Its specificity has been tested by immunodiffusion and it has been demonstrated that the hapten binding is specific for lipids carrying phosphorylcholine.33 Efficiency of TEPC-15 can vary based on the arrangement of phospholipids in the system, for example, it recognizes poorly PC in liposome or lipoproteins, but it is highly specific for lipid emulsions, cell membranes of some bacteria (e.g. Streptococcus pneumococci), oxidized phospholipids, and lipid monolayers.34–37 Different interactions of this antibody with PC and phosphatidylethanolamine monolayer has been also reported by Urbaneja et al.36
The aim of this work is to propose immunofluorescence detection as a new approach to localize choline-based phospholipids at the interface of O/W emulsions. In this paper, experiments on two different cell lines were carried out, as a control, and then the technique was applied to phospholipid-based emulsions and PC film. The effect of the presence of proteins on TEPC-15 staining efficiency was also investigated and quantified.
Double staining was achieved by using a fluorescent dye for neutral lipids, Nile Red (9-diethylamino-5H-benzoalpha-phenoxazine-5-one; Sigma-Aldrich, St Louis, USA), added to the oil phase at a concentration of 0.03 μg mL−1 before emulsification process, and fluorescently acyl chain labeled phospholipids 2-(6-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)hexanoyl-1-hexadecanoyl-sn-glycero-3-phosphocholine (NBD C6-HPC, N-3786, Invitrogen) in the final emulsion at a concentration of 0.1 mg mL−1 added just before images acquisition. Preparation of the stock solution of Nile Red was achieved by dissolving 0.0005 g of Nile Red powder in 120 μL of ethanol by vortexing. In the second step, solution was added to sunflower oil at 40 °C for 1 h under magnetic stirring. After this time, ethanol was let to evaporate under vacuum. Stock solution of NBD C6-HPC was obtained by dissolution of phospholipids in ethanol at a concentration of 1 mg mL−1. Images were acquired using a CLSM (Zeiss Pascal). Excitation of the samples was provided by an Ar laser at wavelengths of 488 nm and a He–Ne at a wavelength of 543 nm. Images were acquired at different magnifications using a 63× oil immersion objective and samples were identified in the fluorescence mode. In order to improve image quality, the random movement of the droplets in the sample was limited by mixing the sample with a drop of agarose solution (0.5%, 161-3101, Bio Rad) previously added on the microscope slide.
Cells were fixed, labeled using TEPC-15 and Rb α-mouse 4888 antibodies and imaged using CLSM. As expected, a clear fluorescence signal is detected on the membrane of the cells, while dark nuclei are visible in the center (yellow arrows). High signal to noise ratio was obtained thanks to extensive washing of the samples with PBS, which allowed the removal of non-specific staining. Representative orthogonal images of the z–y and x–z planes of the confocal z-stack throughout the entire thickness of HaCaT cells are reported in Fig. 1C. Fluorescence signal follows the curvature of cells membrane, as indicated by white arrows, which suggests the specificity of the antibody for the membrane. Even if many membrane antibodies, such as those belonging to the anti-cadherin family41 are commonly used for immunofluorescence membrane staining, TEPC-15 is highly specific for PC, and in principle could be used even to quantify PC concentration and its interaction with other components or molecules.
Once the TEPC-15 staining efficiency was tested on cells, the antibody was used to visualize phospholipids in artificial systems, in particular in O/W emulsions.
A first investigation on an O/W emulsion with droplets surrounded by a phospholipid layer was carried out and reported in Fig. 2A. The staining of ELP emulsion was obtained using Nile Red to stain the core of the droplets and a fluorescently labeled phospholipids, NBD-C6-HPC, to visualize phospholipids in the mixture. As expected, NDB-C6-HPC was located at the interface of the oil droplets suggesting that phospholipids, due to their amphiphilic nature, tend to localize spontaneously at the oil-water interface. NBD-C6-HPC was homogenously distributed at the surface of the oil droplets. Once the presence of phospholipids at the interface was confirmed, the localization of PC was investigated using TEPC-15 staining. First, control experiments were carried out on emulsions without phospholipids (see Fig. S5 of the ESI†), where, as expected, no fluorescence signal was detected at the droplet interface. Only a diffused background signal was visible, given the presence of secondary antibodies in the bulk, due to the absence of washing steps. Results of tests carried out on ELP emulsions incubated with TEPC-15 and Rb α-mouse 488 are reported in Fig. 2B. As in Fig. 2A, the fluorescence signal is located on the surface of the droplets. Unlike the fluorescence signal from NBD-C6-HPC (Fig. 2A), fluorescence pattern of TEPC-15 (Fig. 2B, white arrows) is distributed in a non-homogeneous way, presenting irregular spots and non-stained regions along the interface, as visible in the inset. This result may be due to a non-homogeneous distribution of the PC at the interface, or the presence of other components that interfere with the staining procedure. In both cases of Fig. 2, staining efficiency did not seem to be influenced by droplet size. It is worth mentioning that in the case of liquid emulsions it was not possible to rinse the samples, as in the case of cell samples (Fig. 1). This limitation leads to a non-negligible background signal, which must be taken into account for quantitative image analysis. Further consequence of the lack of the washing step is the presence of bright spots in Fig. 2A, possibly due to an excess of phospholipid analogs in the bulk, and in Fig. 2B, which could be due to stained PC suspended in the aqueous phase, or even to unbound secondary antibody. Moreover, the addition of ELP to water can limit the adsorption of phospholipids at the interface of the oil droplets and can result in the likely formation of phospholipid aggregates in the aqueous phase. This situation can further complicate when other components are added to the system, such as, for example, proteins.
The effect of proteins on phospholipid staining was investigated (Fig. 3). PMP was added to the final ELP emulsion and left to equilibrate for about 10 min under gentle stirring, before following the usual staining and imaging protocol. Four samples were analyzed at different PMP concentration, specifically at 0%, 0.03%, 0.05% and 0.07%. 3D diagrams in the lower pictures show the fluorescence intensity spatial profile of the dotted square regions of each image (centered around a droplet). It can be observed that at 0% PMP, as in Fig. 2B, the droplet interface is stained, even if in a non-homogeneous way. A similar result was obtained for the sample containing 0.03% PMP. Based on a visual comparison, the 3D diagrams do not show significant difference compared to the 0% PMP. The higher background signal at 0.03% and 0.05% PMP is due to the fact that the corresponding images were acquired closer to the glass slide, where laser reflection from the surface is more significant. In the sample containing 0.05% PMP, several droplets show part of the interface unstained, as also visible from the fluorescence intensity plot of the dotted square region. Some droplets are not stained at all. A possible explanation for this effect could be some interactions between PC and proteins at the interface, which could become significant as soon as protein concentration is higher than a given value. In this scenario proteins could for example hide phospholipid epitopes specific for TEPC-15. When different molecules with emulsifying properties are mixed in the same system, it is possible that they form a interfacial multi-layer.14 Phospholipids at the oil-water interface form monolayer or multi-lamellar structures, depending on the conditions, helping the stabilization of the system. When proteins are added as emulsifiers, they assemble at the interface and rearrange their structure to form a stabilizing adsorbed layer of variable thickness. When mixed together, interactions between phospholipids and proteins become more complex. In particular, competitive displacement among molecules at the interface can happen. These interactions influence the stability of the system. It is also demonstrated that protein adsorption at interface is enhanced when a small amount of emulsifier is added on the droplet surface.42 This enhancement is not observed in the case of a thicker monolayer, thus encouraging a displacement of the proteins. Nature of the interface also plays a critical role for the interactions. In the ELP emulsion at 0.07% PMP only a background signal is detectable, whereas droplets are completely unstained. The fluorescence intensity plot does not show any peaks at the interface either. At a concentration value ranging between 0.05% and 0.07%, the interaction between proteins and phospholipids becomes critical for the antibody staining procedure. In order to quantify the staining efficiency as a function of PMP concentration, a detailed analysis has been done on the images for samples with different PMP concentration.
The fluorescence signal of the interface was calculated over a corona surrounding the droplet, as described in the methods section, and normalized with respect to the MGL calculated over the entire image.
In Fig. 4A the normalized interfacial MGL is reported, for each of the samples considered, as a function of the droplet size. No significant differences are visible, for each data series, for different droplet diameters, in the range 2–11 μm. The higher signal is measured in the case of 0% PMP, while a decreasing emission is obtained from more concentrated samples. A linear regression is added to each plot in order to visualize the data trend, which is independent on droplet size. Only in the case of 0.05% PMP a slightly decreasing trend is observed (green line), probably due to the fact that PMP concentration is almost close to a critical point for protein–phospholipid interaction. This could also be related to the non-uniform fluorescence distribution along the drop interface, shown in Fig. 3 for the same sample. When PMP concentration was increased to 0.07%, only a background signal was visible, while droplet interfaces are not labeled, with the interfacial MGL resulting almost 0. The normalized interfacial MGL, averaged of all the droplets considered (about 10 for each sample), is reported in Fig. 4B as a function of PMP concentration. The intensity decreases exponentially with PMP concentration. This preliminary result can be considered as a measure of the sensitivity of the staining efficiency, suggesting that even slight differences in protein concentration in the emulsions can significantly modify the molecular interactions at the interface, affecting the selectivity of the antibody and the fluorescence signal. The value of 0.07% could be identified as a threshold value above which the phospholipids in the system cannot be visualized by TEPC-15 staining.
![]() | ||
Fig. 4 Analysis of fluorescence signal in ELP emulsions at 0%, 0.03%, 0.05% and 0.07% of PMP, reported in Fig. 3. Fluorescence intensity of droplet interface normalized respect the MGL of the entire image, was measured for droplets of different sizes, in the range of 2–11 μm, and plotted as a function of the droplet diameter (A). Linear regressions were added as guide for eyes. No significant trend with droplet size was observed for samples at 0% (black), 0.03% (red), and 0.07% (blue) of PMP. Only in the case of the sample containing 0.05% (green) PMP, a slight decreasing trend is observed, in analogy with the complex situation described in the respective image reported in Fig. 3. (A) In the right chart (B) the normalized MGL of the interface calculated over all the droplets, was plotted as a function of PMP concentration. The intensity decreases exponentially with PMP concentration (f = 11.4![]() |
We tried to improve staining efficiency of TEPC-15 by increasing antibody concentrations in ELP emulsions containing proteins. In these experiments ELP and PMP were mixed together, respectively 0.10 g and 5.63 g before adding the oil phase, the emulsion was first stained with TEPC-15 (1:
10) and then incubated with secondary antibody (Rb α-mouse 488) at different dilutions, 1
:
100, 1
:
20 and 1
:
10. In Fig. 5 three images of the corresponding dilutions are reported. Even in the case of increased antibody concentration, droplets are not stained. Indeed, most of the fluorescence signal is detected in the bulk. Analogous results were obtained increasing primary antibody concentration (data not shown). We can conclude that the interference effect of the proteins is strong enough to hide PC epitopes even in the case of high antibody concentration.
![]() | ||
Fig. 5 ELP + PMP emulsions stained with TEPC-15 (1![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Other experiments were carried out trying to understand the reasons of this poor efficiency, since interactions between phospholipids and proteins can occur either in the bulk or directly at the interface. In order to investigate the effective interaction mechanism and the limiting step, we also added proteins at different steps of the antibodies staining procedure. Results were promising, but deeper analysis of the data (data not shown) is necessary. For this reason, the specificity of TEPC-15 and the effect of proteins on staining procedure, were also verified on PC film.
It is known from literature that the interaction of TEPC-15 with phospholipids, when arranged as a monolayer, at the air–water interface, is dependent on the type of phospholipids used.36 It has also been demonstrated that some antibodies lose their efficiency due to a strong antibody–surface interaction, which can result in a modification of the antibody structure, depending on its flexibility.43 This modification can result in a poor recognition of the antigen. Efficiency is strongly dependent on the hydrophobicity of the surface and not on the geometrical attachment. Here, PC film was created on the bottom of a μ-slide well. After the evaporation of ethanol, images of the PC structure were acquired in brightfield mode (H) and confocal microscopy mode, as reported in Fig. 6. PC film is visible in brightfield as stratified domains (Fig. 6A). A slight reflection due to the proximity of the objective to the glass slide is visible in confocal microscopy, but no autofluorescence signal is displayed. After the addition of the TEPC-15, as expected, no fluorescent signal is detectable (Fig. 6B), while in brightfield PC structure is still visible. After completing the incubation with secondary antibody (Rb α-mouse 488), images show once again a fluorescent signal corresponding to the PC film (Fig. 6C). In particular, fluorescent lines are present in proximity of the domain edges. A dependence of the fluorescence signal from the focus position was observed, suggesting that PC domains do not form homogeneous film.
The film experimental setup was used to confirm the effect of proteins on the staining procedure. Experiments were carried out by adding PMP solution on PC film after antibody staining. Images of the same field of view were acquired at different times, respectively after staining (t0), immediately after the PMP addition (t1), 10 min (t2) and 25 min (t3) after the addition. In Fig. 7A confocal images are reported for each time value. In agreement with Fig. 6, fluorescent signal is very strong at t0 (in some case even saturating the image). After addition of PMP, the signal gradually decreases from t1 down to t3, in agreement with a dynamic interaction mechanism of the proteins at the interface. The effect is quantitatively analyzed in Fig. 7B, where image histograms data are reported. Histograms of each image (Fig. 7B) show a gradual decrease in the number of fluorescent pixel up to the intensity value of 200, where a step-up is found, which suggests some saturation effect, as also visible from the images. Even after 25 min from the PMP addition (t3), when the mean intensity is the lowest, there are still some saturated pixels. The MGL of each image is also reported (Fig. 7C). A strong decrease is evident especially going from t0 to t1. It is worth mentioning that only four images were acquired during the entire experiment for each field of view, and the sample was not exposed to the light source between consecutive acquisitions, in order to minimize any bleaching effect.
In this work, the immunofluorescence technique has been applied to O/W emulsions containing phospholipids (and proteins) in order to localize choline-based phospholipids at the droplet interface with TEPC-15 as a primary antibody. Efficiency of the staining has been verified by confocal imaging and image analysis. Tests on two different cell lines have confirmed the fluorescent staining of the antibody on cell membranes. In emulsions containing only phospholipids, PC is recognized from the antibody at the droplet interface, as confirmed by experiments carried out with phospholipid analogs. We proved that this method, based on TEPC-15 antibody, is efficient to stain PC even in presence of other phospholipids, as in the case of ELP emulsions. The presence of other molecules, in particular proteins, can induce the formation of complex molecular interactions at the interface of oil droplets which can seriously affect the antibody staining. Proteins hinder the efficiency of the staining procedure when more than 0.07% PMP is added to the emulsion. This value can be identified as the threshold value, above which PC is no more visible. Increasing antibody concentrations does not allow PC staining. Similar results have been obtained for PC films, where the fluorescence intensity of the field of view starts to decrease after the addition of proteins.
Studying the interactions between phospholipids and proteins is still a challenging task among researchers and deeper investigations are needed. In our case, proteins were found to interfere with antibody staining, likely by hiding the antigen-bounding site of PC.
The immunofluorescence methodology described in this work is applied for the first time, to our knowledge, to colloidal systems. Our results show that immunohistochemistry can represent a powerful tool to investigate molecule location in microstructured fluids. Given the wide diffusion of multiphase fluids, this result could be relevant in several industrial applications, such as food, cosmetic and pharmaceutical industries.44
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra13775j |
This journal is © The Royal Society of Chemistry 2016 |