Mixed Adsorption Mono- and Multilayers of ß-Lactoglobulin Fibrils and Sodium Polystyrene Sulfonate

Abstract

The formation of beta-lactoglobulin (BLG)/sodium polystyrene sulfonate (PSS) complexes decelerates the change in the surface properties of the mixed solutions with the surface age and increases the steady-state dilational surface elasticity in a narrow PSS concentration range. At the same time, the changes in the surface properties are accelerated in the dispersions of BLG fibrils with and without PSS due to the influence of small peptides coexisting with fibrils. A decrease in the peptide concentration as a result of the dispersion purification leads to slower changes in the surface properties at low PSS concentrations. The increase in the polyelectrolyte concentration results in an increase in the steady-state surface elasticity due to the fibril/PSS complex formation and in very slow changes in the surface properties if the polyelectrolyte exceeds a certain critical value. The latter effect is a consequence of the formation of large aggregates and of an increase in the electrostatic adsorption barrier. The consecutive adsorption of BLG fibrils and PSS leads to the formation of regular multilayers at the liquid–gas interface. The multilayer properties change noticeably with an increase in the number of layers from four to six in agreement with previous results on the multilayers of PSS with an oppositely charged synthetic polyelectrolyte, presumably due to the heterogeneity of the first PSS layer. The dynamic elasticity of the multilayers approaches 250 mN/m, indicating that they can effectively stabilize foams and emulsions.

1. Introduction

The interactions between amyloid fibrils and polyelectrolytes have been studied intensively over the last few decades [1,2,3,4,5,6,7,8,9]. The interest in these systems is caused by the applications of protein/polyelectrolyte complexes in the food industry [2], biocatalysis [10], and biosensoring [11,12], for the separation and purification of proteins [6,13,14], and in drug delivery [7,11,15]. Information on the interactions of polyelectrolytes with protein aggregates, in particular with amyloid fibrils, is more limited. Most attention in this field is paid to the inhibition of the growth of amyloid fibrils by polyelectrolytes, mainly by polysaccharides, with the aim of creating drugs against amyloidosis and connected severe diseases [16,17,18]. At the same time, recent studies have shown that the complexes of fibrils with polyelectrolytes can have other important applications in medicine and various technologies. For example, the addition of polysaccharides (low-acetylated gellan gum and κ-carrageenan) to amyloid fibril hydrogels and aerogels improves their mechanical properties and enhances their resistance to compression [19], while the addition of polysaccharides to fibril-based bioplastics films increases their tensile strength and water resistance [20].

In spite of the growing interest in thin films containing amyloid fibrils, especially in the adsorption layers in emulsions and foams [21,22,23,24,25,26,27,28,29,30,31,32], information on the interactions of amyloid fibrils with polyelectrolytes at liquid–fluid interfaces is quite scarce. Only recently, Peydayesh et al. have shown that the coacervation of protein (β-lactoglobulin, BLG) amyloids and linear polyelectrolytes (hyaluronic acid) at the interface can result in the formation of a thin layer with a highly ordered asymmetric structure [33]. The self-assembled layer with the amyloid fibrils occupying one side of the layer and the polysaccharide covering the other side can obviously find various applications.

The high surface activity of amyloid fibril/polyelectrolyte complexes gives a possibility to use them for the stabilization of foams and emulsions, but this application is limited by the lack of information on the organization of the complexes in the surface layer for most of these systems. Moreover, the high stability of liquid-phase dispersions with fibrils in the adsorption layer compared with those stabilized by native proteins is not always clear. Although it is generally accepted that dynamic surface properties, in particular surface viscoelasticity, are the main parameters determining the properties of liquid disperse systems [23,26,34,35,36], the surface properties of fibril dispersions and native protein solutions frequently coincide [35,36]. It has been shown recently that this is the consequence of the high concentrations of polypeptides of relatively low molecular weight in the dispersions. The elasticity of the fibril layers with a decreased concentration of the admixture exceeded the value for the unpurified system [37]. On the other hand, for some proteins, the dynamic surface elasticity of the fibril dispersions proved to be higher than that of the corresponding native protein solutions [24,38,39]. The dynamic elasticity of spread and adsorbed mixed layers of amyloid fibrils and polyelectrolytes has not been determined yet to the best of our knowledge. Information on the properties of regular multilayers of protein fibrils and polyelectrolytes at liquid–gas interfaces is even rarer [40].

This study is devoted to the investigation of interactions of BLG fibrils with the typical synthetic polyelectrolyte—sodium polystyrene sulfonate (PSS)—at the liquid–gas interface. The main aim is to determine the conditions corresponding to high dynamic surface elasticity and thus to the possible formation of liquid-phase dispersions of high stability. The pH 2 value of the bulk fibril dispersion corresponds to opposite charges of the fibrils and polyelectrolyte. Another aim consists of the preparation of a multilayer with alternating fibril and PSS layers. To the best of our knowledge, the formation of fibril/polyelectrolyte multilayers has been studied in detail so far only at the solid–liquid interface [41].

2. Materials and Methods

2.1. Materials

BLG (Mw ≈ 18,300 Da) and thioflavin T (ThT) were purchased from Sigma-Aldrich (Darmstadt, Germany). The protein solutions in water, in sodium chloride solutions and phosphate buffer were prepared by the dilution of a stock solution, which had been stored in a refrigerator for no longer than three days. The buffer was prepared by mixing stock solutions of Na₂HPO₄ and NaH₂PO₄ (Sigma Aldrich, St. Louis, MO, USA). Deionized water was distilled three times in a glass apparatus before use. Sodium chloride (Vekton, St Petersburg, Russia) was preliminarily heated in a muffle furnace at about 750 °C for the elimination of possible organic impurities. The protein fibrils were prepared by heating a concentrated protein solution of pH 2 as described elsewhere [37]. After heating, the solution was immediately cooled and placed into an ice-water bath. The obtained dispersion was stored in a refrigerator at 4 °C. Further fibril purification consisted of the centrifugation and careful replacement of the supernatant weak hydrochloric acid at pH 2 [37].

The concentrated dispersions of protein fibrils were stored in a refrigerator not more than for two months and used for the preparation of the investigated diluted dispersions. The final concentration of the BLG fibril dispersions was determined by UV-Vis spectroscopy using a UV-1800 instrument from Shimadzu (Kyoto, Japan). The absorption spectra were measured in a wavelength range from 250 to 650 nm for solutions with known concentrations of native BLG. After that, the spectrum intensities around 280 nm were used to obtain calibration dependence. The dispersions of BLG fibrils and native BLG solutions have the same spectra, giving us the possibility to determine the protein concentration using the determined calibration line. PSS (Mw ≈ 70,000 Da, Sigma-Aldrich) was used as received.

The mixed BLG/PSS solutions and fibril dispersions (FBLG) with PSS were prepared by mixing BLG solutions or FBLG with equal volumes of PSS solutions. The concentrations of the BLG solution or FBLG before mixing were 0.0034 or 0.034 mass %. The PSS mass concentration was two times higher than in the final mixed solutions or dispersions.

2.2. Methods

The Wilhelmy plate method was used to determine the surface tension with the use of a platinum plate, which was roughened to improve the wetting. The error limits of the measurements were ±0.2 mN/m. Surface pressure $\Pi$ of the solution and its surface tension $\gamma$ are connected by the following relation:

$$\Pi ={\gamma }_0-\gamma$$

where ${\gamma }_0$ is the surface tension of the pure solvent.

The dilational dynamic surface elasticity was measured by the oscillating barrier method using an ISR instrument (KSV NIMA, Espoo, Finland) as described elsewhere [38,39]. In brief, two Teflon barriers moved along polished brims of a Teflon Langmuir trough with a frequency of 0.03 Hz and an amplitude of 2–10% relative to the distance between the barriers. The oscillations of the surface tension induced by the barrier motion were detected by a Wilhelmy plate. Complex dilational dynamic surface elasticity E can be defined as the ratio of the complex amplitudes of surface tension oscillations $\delta \gamma$ and oscillations of relative surface area $\delta \ln A$:

$$\mathsf{{\rm E}}=Re\left(\mathsf{{\rm E}}\right)+i Im\left(\mathsf{{\rm E}}\right)=\delta \gamma /\delta \ln A$$

The elasticity modulus equals the ratio of the oscillation amplitudes, while the phase shift between the oscillations of the two parameters provides the phase angle and thus allows the determination of the real, $Re\left(\mathsf{{\rm E}}\right),$ and imaginary, $Im\left(\mathsf{{\rm E}}\right),$ components of the dynamic surface elasticity. The accuracy of the surface elasticity determination was close to ±5%. Only the real part of surface elasticity was analyzed as its imaginary part was proved to be insignificant in comparison to the real part.

A null ellipsometer (Optrel-GBR, Berlin, Germany) with a laser wavelength of 623.8 nm was used to estimate the local changes in the surface concentration in the course of the adsorption. The measurements were taken at an angle close to the Brewster angle. In the case of a single solute, ellipsometric angle ∆ is expected to be proportional to its surface concentration [42].

The morphology of the adsorption layers was investigated by atomic force microscopy (AFM) using the instrument NTEGRA Prima (NT-MDT, Moscow, Russia). The adsorbed layer was transferred from the liquid surface onto a freshly cleaved mica plate using the Langmuir–Schaeffer method and then dried in a desiccator. The AFM studies were conducted in a semicontact regime, utilizing a cantilever with an approximate curvature radius of 10 nm.

The macroscopic morphology of the adsorption layers at the air–water interface was determined by Brewster angle microscopy (BAM) using a BAM 1 instrument (Nanofilm Technology, Main, Germany) equipped with a 10 mV He–Ne laser.

The size of the protein particles and their ζ-potential in the bulk phase were determined by dynamic light scattering (DLS) using a Zetasizer ZS Nano analyzer (Malvern Instruments, Malvern, UK). The measurements were carried out at a scattering angle of 173°.

The PSS/FBLG duplex layers were formed in the Langmuir trough by the consecutive adsorption of PSS and FBLG. Firstly, PSS solution with mass concentration of 1% was poured into the Langmuir trough. After waiting for one hour, it was exchanged for water (pH = 2). The water taken in a volume seven to eight times larger than the trough volume was pumped through the trough by two Unipan peristaltic-type pumps (Warszawa, Poland). After that, the FBLG of the mass concentration of 0.017 mass % was pumped three times through the trough. A repetition of this procedure allowed us to obtain a multilayer with a given number of PSS/FBLG duplex layers.

3. Results and Discussion

The rate of adsorption and thus the kinetic dependencies of the surface properties of protein solutions differed strongly at pH 2 and 7 for the same protein; however, the corresponding dependencies of the dynamic surface elasticity on surface pressure almost did not depend on the solution pH (Figure S1 of the Supplementary Materials). While the protein charge determines the adsorption kinetics, it does not influence noticeably the organization of the protein molecules in the surface layer.

The addition of PSS to protein solutions at pH 2 can decelerate strongly the changes in surface properties, with the surface age indicating the formation of protein/PSS complexes. Despite a decrease in the total charge of the complexes with increasing PSS concentration, the surface properties change slower in this case, presumably due to the formation of large aggregates in the bulk phase (Figures S2 and S3 of the Supplementary Materials). The deceleration of the adsorption kinetics starts when the polyelectrolyte mass concentration exceeds 20% of that of the protein. The deceleration rate depends on the total protein content, and at higher protein concentrations, the adsorption rate starts to change noticeably at lower PSS concentrations (Figure S4 of the Supplementary Materials). At the same time, the increase in the total protein concentration results in the appearance of a slight opalescence of the aqueous dispersion, indicating an increase in the concentration of relatively large aggregates.

The formation of PSS/protein complexes changes the structure of the adsorption layer and it becomes more rigid. The dynamic surface elasticity exceeds the results for native protein solutions at a given surface pressure at the ratios of BLG/PSS mass concentrations approximately in the range 1–2.5 at a BLG concentration of 0.0017 mass % (Figure 1). For example, at a BLG/PSS ratio of 2.5, the surface elasticity of the mixed solution almost exceeds two times the value of the native protein solution. A further increase in the PSS concentration decreases the surface elasticity again, and at BLG/PSS ratios of 50 or lower than about 0.05, all the dependencies of the dynamic surface elasticity on surface pressure almost coincide within the error limits with the results for native protein solutions. In this case, the excess of PSS molecules of low surface activity does not influence the static surface properties, and the activity of the complexes is presumably close to that of the native protein.

The changes in the surface properties with the surface age are strongly accelerated for fibril dispersions compared with the native protein solutions of the same mass concentrations (Figure 2). This unexpected effect was explained by the impact of admixtures of peptides of low molecular mass and high surface activity [24,37,43]. The heating of the protein solution in the course of the fibril preparation leads to the denaturation and partial destruction of protein molecules, with the formation of small protein fragments. As a result, the kinetic dependencies of unpurified fibril dispersions approach the data for solutions of small peptides with a relatively high diffusion coefficient. The purification of the fibril dispersions by centrifugation and the subsequent replacement of the supernatant by water decelerates strongly the changes in the surface properties with the surface age (Figure 2). At the same time, even a careful purification of BLG dispersions by this method can be insufficient, and the admixtures can still influence the static surface properties [37]. As a result, the dependencies of the surface elasticity on surface pressure almost coincide for the dispersions of purified and unpurified fibrils and native protein solutions at a low pH (Figure S5 of the Supplementary Materials).

The addition of PSS to dispersions of unpurified fibrils results in a decrease in the rate of decrease in the properties as for native protein solutions, but this effect is weaker for unpurified fibril dispersions, presumably due to the influence of the peptide admixtures (Figure 3). Nevertheless, the adsorption of the polyelectrolyte/fibril complexes leads to a more rigid adsorption layer with a higher dynamic surface elasticity than for the systems without PSS at the ratios of BLG/PSS mass concentrations approximately in the range 0.33–5 and at a BLG concentration of 0.0017 mass % (Figure 4).

The purification of fibrils gives a possibility to decrease the influence of the admixture. In this case, one can observe an influence of PSS on the kinetic dependencies of the surface properties of purified fibril dispersions at lower concentrations of the polyelectrolyte (Figure 5). Changes in the surface tension for five hours do not exceed 1 mN/m, and the corresponding changes in the dynamic surface elasticity are less than 5 mN/m for the systems with the ratios of BLG/PSS concentrations less than 30. At ratios less than 5, the surface properties change only within error limits for the time of measurements. The almost constant surface properties do not mean adsorption cessation but only a slow adsorption rate. A slow change in the surface concentration can be observed by ellipsometry, even at the BLG/PSS concentration ratio close to one (Figure 6). The surface compression in this case in about 12 h after the beginning of measurements leads to a significant increase in the surface pressure and is accompanied by an increase in the dynamic surface elasticity due to the increase in the surface concentration (Figure S6 of the Supplementary Materials). AFM measurements show that the formation of a network of fibrils can occur at the interface, even at surface pressures close to zero. The surface compression results in an increase in the fibril surface concentration, and the formation of a more rigid network with a higher surface pressure and surface elasticity is observed (Figure 7). In the case of unpurified fibril dispersions mixed with PSS (Figure S7 of the Supplementary Materials), the small peptides of high surface activity and their complexes with PSS presumably hinder the surface network formation, and the influence of PSS on the kinetic dependencies of the surface properties becomes noticeable only at BLG/PSS concentration ratios less than 5. DLS studies show the formation of large aggregates in the bulk phase of BLG dispersions in the presence of PSS, and their mean size increases with an increase in the polyelectrolyte concentration, leading to opalescent dispersion (Figure S8 of the Supplementary Materials). These aggregates diffuse slowly to the interface, resulting in very slow changes in the surface properties. Moreover, an increase in the absolute value of the aggregate charge with an increase in PSS concentration beyond the isoelectric point also strongly decelerates the adsorption kinetics of the complexes (Figure S9 of the Supplementary Materials).

Although the formation of the fibril network leads to microscopic heterogeneities of the surface layer for dispersions of BLG/PSS aggregates (Figure 7), the application of BAM shows that the layer is homogeneous at a macroscopic scale (Figure S10 of the Supplementary Materials). The surface layer is fluid-like, and slight mechanical disturbances of the surface leave no traces after a few seconds.

The opposite signs of the charges of BLG and PSS molecules allow the formation of mixed regular multilayers of these two substances. The surface activity of PSS gives a possibility to form a multilayer in the absence of a chemically different anchoring layer. The PSS adsorption layer was formed in a 1 mass % aqueous solution. The corresponding steady-state ellipsometric angle Δ was close to 1 degree and thus about two times lower than for dilute fibril dispersions with a concentration of 0.017 mass % (Figure 8). The exchange of the PSS solution below the adsorption layer by water did not lead to noticeable changes in the ellipsometric angle. If the solution below the PSS layer was changed by a BLG fibril dispersion with a protein concentration of 0.017 mass %, the steady-state ellipsometric angle increased up to 2.2 degrees as a result of the fibril adsorption. The subsequent exchange of the fibril dispersion by a PSS solution increased Δ only a little, less than 10%. It is possible to assume that, in this case, PSS molecules are adsorbed below the BLG fibrils but form only a loose layer. The next exchange of the PSS solution by the fibril dispersion again increased Δ significantly. The repetition of this procedure a few times gave the possibility to obtain a structure of alternating layers of PSS and fibrils. The steady-state ellipsometric angle increased with the number of layers, and for 10 consecutive layers, i.e., 5 duplex layers of PPS and BLG fibrils, it approached 9 degrees (Figure 8).

The application of AFM shows that the concentration of fibrils at the interface increases with the numbers of the layers and the formed structure becomes denser (Figure 9). SEM images allow the observation of a multilayer morphology (Figure S11 of the Supplementary Materials). It is not homogeneous and consists of irregular domains with some gaps between them, where the multilayer thickness is presumably lower.

The surface tension decreases at the adsorption of the first fibril layers from 52 to 45 mN/m and decreases further to 40 mN/m after the adsorption of the second PSS layer (Figure S12 of the Supplementary Materials). After that, the increase in the multilayer thickness due to the consecutive adsorption of PSS and fibrils does not influence noticeably the surface tension, indicating that only the composition and structure of the first two layers determine its value. At the same time, the changes in the dynamic surface elasticity with the number of layers at a frequency of about 0.03 Hz are different (Figure 10). The surface elasticity is almost constant in the course of the formation of the first three layers. In this case, relatively fast relaxation processes in the loose mixed layers presumably do not allow the development of noticeable surface stresses upon surface compression or expansion at the given deformation frequency. A few authors have already noticed that the properties of a relatively thin multilayer on the basis of PSS with two or three duplex layers, especially the dynamic elasticity, differ noticeably from the properties of a thicker multilayer [44,45,46]. A possible explanation of this peculiarity consists of a looser structure of the first duplex layers. The PSS monolayers are not homogenous and contain some microaggregates [47]. As a result, the first and second duplex layers have a relatively loose structure and the number of layers required to create a fully connected dense multilayer of high elasticity has to exceed four or five [46].

When the multilayer becomes thicker and denser, the surface stress relaxation proceeds slower and the dynamic surface elasticity starts gradually to increase with the increase in the number of layers and reaches about 150 mN/m at a deformation amplitude of 10% when five duplex layers are formed. For a thicker multilayer with more than two duplex layers, its response to compression and expansion becomes non-linear and the surface elasticity strongly increases at a decrease in the deformation amplitude, reaching values of 230 mN/m at an amplitude of 4% (Figure 10). The application of the general stress decomposition method, proposed previously [48,49], shows the appearance of an asymmetric response of the surface layer to dilation with an increase in the number of layers from five to ten (Figure S13 of the Supplementary Materials). This method gives a possibility to analyze a non-linear response of the surface layer to deformations and is based on the decomposition of the surface tension oscillations into four components, τ1, τ2, τ3 and τ4 (the definitions are given at the end of the Supplementary Materials), which take into account the contributions of higher harmonics. This contribution increases significantly with an increase in the layer thickness.

In spite of a relatively high density, the multilayers consisting of five duplex layers are fluid-like, and BAM images do not show any visible deformations after touching the layer by a thin needle. There were no indication of the formation of a solid-like multilayer as in the case of multilayers of PSS and poly(allylamine chloride), where the increase in layer thickness led to the formation of a solid surface phase [46]. At the same time, it is possible to observe some folds of the multilayer after its compression by more than two times (Figure 11). The subsequent expansion of the layer does not allow us to recover its homogeneity, and one can observe some regions of higher density in it.

The formation of relatively thick regular multilayers on the basis of amyloid fibrils can have various applications, for example, in the course of the preparation of durable microcapsules for directed drug delivery. Although PSS is not a suitable substance for medical applications, it can be exchanged presumably by a biocompatible polyelectrolyte. On the other hand, the mixed adsorption layers of fibrils and polyelectrolytes with a high surface elasticity can ensure the high stability of foams and emulsions.

4. Conclusions

The surface properties of the dispersions of BLG fibrils with PSS have some similarities with the properties of PSS/BLG solutions. The main distinctions consist of the formation of the complexes of two components and relatively large aggregates in a broader range of PSS concentrations, and in slightly higher values of the dynamic surface elasticity in the PSS concentration range corresponding to the complex formation. The fibrils/PSS dispersions are also characterized by faster changes in the surface properties with the surface age. This effect is connected with the influence of the admixture of small peptides of high surface activity in the dispersion. The purification of the dispersions results in a significant deceleration of the changes in surface properties and the enlargement of the range of the high surface elasticity where the AFM application allows an observation of a fibril network at the interface. This concentration range is also characterized by slow adsorption due to the formation of large aggregates close to the isoelectric point and the increase in the electrostatic adsorption barrier. The sequential adsorption of the oppositely charged polyelectrolyte and protein fibrils at the liquid–gas interface allows the formation of a regular multilayer consisting of a few fibril/polyelectrolyte duplex layers. The dynamic elasticity of the multilayer increases with the number of layers and reaches 230 mN/m for five duplex layers, a few times higher than for mixed solutions of the native protein and PSS. The properties of the multilayer with two duplex layers differ noticeably from the properties of thicker multilayers, presumably due to a more compact structure in the former case. The durability of the multilayer consisting of amyloid fibrils and polyelectrolyte can exceed the properties of the layers containing only oppositely charged polyelectrolytes. The method of the formation of fibril/polyelectrolyte multilayers proposed in this work can be used to produce new materials.