Dynamic surface properties of lysozyme solutions. Impact of urea and guanidine hydrochloride

Colloids and Surfaces B: Biointerfaces 129 (2015) 114–120

Contents lists available at ScienceDirect

Colloids and Surfaces B: Biointerfaces journal homepage: www.elsevier.com/locate/colsurfb

Dynamic surface properties of lysozyme solutions. Impact of urea and guanidine hydrochloride M.M. Tihonov, O.Yu. Milyaeva, B.A. Noskov ∗ Department of Colloid Chemistry, St. Petersburg State University, Universitetsky pr. 26, 198504 St. Petersburg, Russia

a r t i c l e

i n f o

Article history: Received 12 January 2015 Received in revised form 8 March 2015 Accepted 11 March 2015 Available online 19 March 2015 Keywords: Protein adsorption Surface dilational rheology Lysozyme Protein denaturation Guanidine hydrochloride Urea

a b s t r a c t Urea and guanidine hydrochloride (GuHCl) have different influence on surface properties of lysozyme solutions. The increase of GuHCl concentration leads to noticeable changes of kinetic dependencies of the dynamic surface elasticity and ellipsometric angles while the main effect of urea reduces to a strong drop of the static surface tension. The difference between the effects of these two denaturants on the surface properties of other investigated globular proteins is significantly weaker and is mainly a consequence of a different extent of the globule unfolding in the surface layer at equal concentrations of the denaturants. The obtained results for lysozyme solutions are connected with the strongly different denaturation mechanisms under the influence of urea and GuHCl. In the former case the protein preserves its globular structure in the adsorption layer at high urea concentrations (up to 9 M) but without tightly packed interior of the globule and with a dynamic tertiary structure (molten globule state). On the contrary, the increase of GuHCl concentration leads to partial destruction of the protein tertiary structure in the surface layer, although this effect is not as strong as in the case of previously studied bovine serum albumin and ␤-lactoglobulin. © 2015 Elsevier B.V. All rights reserved.

1. Introduction The X-rays crystallography and NMR spectroscopy give detailed structures of the globules of numerous proteins. The relations between these unique and stable structures on one hand, and protein functions or the properties of protein solutions on the other hand are the main problems of the classic protein physics [1,2]. Meantime, disordered or partly disordered protein structures are widespread in biological and industrial systems but so far have been investigated to a significantly less extent [2–5]. These can be, for example, intrinsically disordered proteins [2,5], some precursors of protein aggregates [6–8] or proteins unfolded due to the changes of environment, in particular, under the influence of chemical denaturants [4]. The intensive studies of the most of these disordered protein structures have started only recently and the obtained information has rather been limited until now [5–8]. At the same time, the disordered or partly disordered protein structures are a well known subject in surface chemistry. Proteins are frequently used in food, pharmaceutical and cosmetic industries to stabilize foams and emulsions. The classic point of view is

∗ Corresponding author. Tel.: +7 8124284093; fax: +7 9052003331. E-mail address: [email protected] (B.A. Noskov). http://dx.doi.org/10.1016/j.colsurfb.2015.03.034 0927-7765/© 2015 Elsevier B.V. All rights reserved.

that the protein molecules are entirely unfolded at the liquid–gas interface. According to Langmuir and Schaefer “the protein layer structure is like that of a net made to float on the surface of water by corks (hydrophobic groups of side chains) distributed over the surface of the net” [9]. The subsequent studies admitted also a possibility of the adsorption of some intact globules at high protein concentrations [10,11]. The advent of the neutron reflectometry led to a reconsideration of the classic ideas and indicated the preservation of the globular structure of the main model proteins at the liquid surface [12–17]. This conclusion, however, is not generally accepted. Different authors using similar experimental techniques give sometimes opposite answers to the question on the destruction or preservation of the protein globular structure at the liquid surface. For example, studies of the X-ray reflectivity allowed the authors to conclude that lysozyme globules are entirely unfolded in the surface layer [18–20] while the neutron reflection studies does not confirm strong changes of the tertiary structure of this protein in the course of adsorption [12,13,16]. Another sensitive method, the external reflection FTIR spectroscopy, gives evidence of noticeable changes of the secondary structure of lysozyme in the surface layer but does not allow direct estimation of the tertiary structure [21]. The lack of sufficiently reliable information on the protein tertiary structure in the adsorption layer is obviously caused by an extremely limited number of suitable experimental techniques.

M.M. Tihonov et al. / Colloids and Surfaces B: Biointerfaces 129 (2015) 114–120

It has been shown recently that the dilational surface rheology can give additional information on the protein conformation at the liquid–gas interface [22–25]. This approach is based on a strong difference between kinetic dependencies of the dynamic surface elasticity for globular protein solutions and for solutions of non-globular or unfolded proteins. In the former case the kinetic dependencies are monotonic and resemble the corresponding results for aqueous dispersions of charged solid nanoparticles. In the latter case the kinetic dependencies are similar to those for solutions of amphiphilic polymers where the dynamic surface elasticity goes through a strong local maximum and approaches relatively low values close to equilibrium in agreement with the theory of the surface viscoelasticity of polymer solutions [25]. Measurements of the dynamic surface elasticity of the mixed solutions of bovine serum albumin (BSA) and ␤-lactoglobulin (BLG) with guanidine hydrochloride and urea as a function of surface age allowed tracing the globule unfolding and the formation of the distal region of the surface layer [22–24]. In this work the developed approach is applied to lysozyme solutions. Although lysozyme is one of the most frequently studied model proteins, the information of the structure of its adsorption layers at the liquid–gas interface is still rather controversial [11–13,16,18–21]. It has a more rigid globule than BSA and BLG and belongs to the group of “hard” proteins [26]. The molecule of lysozyme with the molecular weight of 14,300 Da consists of 129 aminoacid residues. The closed packed globule stabilized by four disulfide bonds has the size of 4.5 × 3.0 × 3.0 nm and consists of two main domains [27]. Lysozyme tertiary structure is relatively stable against the action of high temperatures [28] and chemical denaturants [28–30], in particular of urea and GuHCl. These substances are frequently used in the studies of the protein unfolding, however, the mechanism behind the denaturating power of urea and GuHCl is still not well understood [1,3,28]. Their effect on lysozyme globules is strongly different [28]. Urea has almost no influence on the secondary structure and the protein preserves its globular conformation up to high urea concentrations (>10 M). On the contrary, the Raman spectroscopy indicates the unfolding of lysozyme in concentrated solutions of GuHCl (>5 M). The distinctions in the denaturating mechanism of urea and GuHCl result in different influence on the amyloid fibril formation in lysozyme solutions. GuHCl accelerates this process at high concentrations but urea does not display any noticeable effect [31]. Studies of the dynamic surface properties of the mixed solutions of lysozyme and these two denaturing agents with strongly different impact on lysozyme globules give a possibility to elucidate further the relation between these properties and the protein conformation at the liquid–gas interface. This is the main aim of this work. The data on neutron reflection from the adsorbed layer of unfolded lysozyme molecules facilitate this task [16]. To the best of our knowledge the corresponding information for lysozyme/urea solutions has not been determined yet. Another aim is to estimate the degree of the destruction of lysozyme tertiary structure in the surface layer at different concentrations of the denaturants. 2. Materials and methods Lysozyme (Sigma–Aldrich, Germany) was used as received. Lysozyme solutions of required concentrations in phosphate buffer at pH = 7 were prepared by dilution of the solution with concentration of 0.05 mM. GuHCl and urea (Sigma–Aldrich, Germany) were used as received. These substances were dissolved in a small quantity of phosphate buffer before the addition to the protein solution. The volume of the solution was then increased up to the required value.pH of all the solutions was adjusted to 7 by adding components of the Na2 HPO4 –NaH2 PO4 buffer mixture (Sigma–Aldrich, Germany). The ionic strength of all the solutions was 0.04 M. The

115

solutions were prepared using triply distilled water. The glass apparatus was used in the last two steps of the distillation. The surface tension of the pure buffer solution was 72.8 mN/m. All lysozyme solutions were used without storing and measurements of the surface properties were started in several minutes after preparation of the fresh solution. All the measurements were carried out at 20 ± 1 ◦ C. The dynamic dilatational surface elasticity was measured by the oscillating ring [32,33] and oscillating barrier [22,23,33] methods. In the first case the surface of the solution under investigation was periodically expanded and compressed as a result of oscillations of a glass ring along its axis. The ring was partly immersed into the liquid with its axis perpendicular to the liquid surface and its internal surface was grounded to improve wetting. The ring oscillations led to regular oscillations of the liquid surface area and surface tension of the solution as a result of periodical changes of the meniscus shape at the internal surface of the ring. The surface tension of the investigated liquid was measured inside the ring by Wilhelmy plate method. The main advantage of the oscillating ring technique consists in almost pure dilational deformations of the liquid surface and thereby in a negligible contribution of shear stresses to experimental results. The relative amplitude and frequency of the solution surface area oscillations were 10% and 0.1 Hz, respectively. In the case of the oscillating barrier method the liquid surface area oscillated due to the sinusoidal motion of a hydrophobic barrier gliding back and forth along polished brims of a Langmuir trough. The induced oscillations of the surface tension were also determined by Wilhelmy plate method. The application of the oscillating barrier method was possible only at the oscillation frequencies less than approximately 0.2 Hz when the length of surface longitudinal waves far exceeded the length of the Langmuir trough and the surface tension oscillations in the center of the trough were uniform. Therefore the oscillation frequency of the barrier was 0.1 Hz and the relative amplitude was 3%. The real εr and imaginary εi components of the dilational dynamic surface elasticity ε were calculated from the amplitudes of oscillations of the surface tension ı and surface area ıS, and the phase shift ϕ between the oscillations of these two quantities by the following relation ε=

d Sı Sı cos ϕ + i sin ϕ = εr + iεi = d ln S ıS ıS

(1)

The imaginary part of the complex dynamic surface elasticity of the solutions under investigation proved to be much less than the real part. Therefore, only the results for the real part are discussed below. The experimental errors of both the oscillating ring and oscillating barrier methods are mainly determined by the errors of surface tension measurements and are less than ± 5%. However, the distinctions between the kinetic dependencies of surface properties can be higher due to the insufficient reproducibility of the induction period (cf. next section). A null-ellipsometer Multiskop (Optrel GBR, Germany) at a single wavelength of 632.8 nm was applied to estimate the adsorbed amount using a fixed compensator (±45◦ ) and a 2-zone averaging nulling scheme. The scheme of this apparatus has been described in detail elsewhere [34]. All the ellipsometric measurements were performed near the incidence angle of 50◦ close to the Brewster angle because this condition ensured the highest sensitivity of the ellipsometric angles to the properties of the adsorption layer. The elliptically polarized light consists of two components with the electric vectors oscillating parallel and perpendicular to the plane of incidence. The reflection at the interface results in different changes of the phase and amplitude of these two components. These changes depend on the optical properties of the interface and can be characterized by two ellipsometric angles  and  connected with the Fresnel reflectivity coefficients of

116

M.M. Tihonov et al. / Colloids and Surfaces B: Biointerfaces 129 (2015) 114–120

the parallel and perpendicular components, rp and rs , respectively [34]: rp = tan ei rs

90 80

(2)

surf = s − 0 = ksurf 

(3)

70

|ε|, mN/m

 is very insensitive to the optical properties of a thin film at the air/liquid interface and only the data on the ellipsometric angle  is presented below. The difference surf between the ellipsometric angles  for the investigated solution s and pure water 0 is approximately proportional to the adsorbed amount  for a single uniform isotropic layer separated by sharp interfaces from the two ambient mediums [34,35]

60 50 40 30 20 10 0 0

5

10

15

20

π, mN/m

where ksurf =

100

g()(εsurf − ε1 )(εsurf − ε2 ) εsurf

(4)

and where g() is a coefficient that depends only on the angle of incidence and the properties of the bulk phase, is the wavelength, is the density of the adsorption layer, εsurf , ε1 and ε2 are the dielectric constants of the layer and two ambient mediums (air and aqueous solution), respectively. If the conformation of the adsorbed macromolecules does not change in the surface layer and there are no other adsorbed components except the solvent, one can apply De Feijter relation to calculate the adsorbed amount [36]. However, these assumptions are incorrect for the system under investigation and the ellipsometric data is presented below only to characterize qualitatively the adsorption kinetics. The experimental error of the ellipsometric angle surf is about ± 0.1◦ . At the same time, the distinctions between different kinetic dependencies for the same solution may exceed this value due to possible fortuitous mechanical disturbances in the course of long measurements. The size of the particle in the bulk solution was controlled by dynamic light scattering (DLS) using a Zetasizer ZS Nano analyzer (Malvern Instruments, United Kingdom). The measurements were carried out at a scattering angle of 173◦ . 3. Results The dynamic surface properties of lysozyme solutions and mixed solutions of the protein with GuHCl and urea were measured as a function of surface age and denaturant concentration at a constant lysozyme concentration of 0.005 mM and at pH = 7. The surface properties at this low concentration change for a few hours and one can observe different steps of the adsorption layer formation. The adsorption of charged lysozyme globules at pH = 7 is slower than the adsorption of neutral amphiphilic polymers at similar concentrations due to an electrostatic adsorption barrier. The adsorption of first protein molecules results in the additional charge at the interface and thereby in the repulsion of other adsorbing molecules of similar charge. The kinetic dependencies of the surface properties for pure lysozyme solutions are characterized by a long induction period of the order of magnitude of 1 h. This effect has been already discussed in the literature and is thought to be connected with the peculiarities of the equation of state for the adsorption layer of this protein [37,38]. The duration of the induction time is characterized by poor reproducibility like in Refs. [37,38] leading to significant distinctions between the kinetic dependencies of the dynamic surface properties at the same protein concentration. If these data on the dynamic surface elasticity are plotted as a function of the surface pressure, the scatter of the data decreases strongly and does not exceed ± 5 mN/m (Fig. 1). This finding implies that the insufficient reproducibility of the dynamic surface properties

Fig. 1. Modulus of the dynamic surface elasticity of lysozyme solutions (0.005 mM) vs. surface pressure. Different markers and colors correspond to different independent measurements. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

of pure lysozyme solutions is connected with some kinetic effects. Although the dynamic light scattering did not discover any particles in the solution with the size greater than the globule diameter (∼4 nm), it is possible that the system is not entirely homogeneous and slight natural convection in the trough can influence the kinetics of the adsorption layer formation. For example, one can assume that the adsorption layer contains some surface aggregates (islands) at low surface pressures. The agreement between different kinetic dependencies of the ellipsometric angle  is better but also worse than that of solutions of other globular proteins (Fig. 1S of the supporting information) [22]. Note that the addition of GuHCl and urea even at low concentrations (0.5–2 M) improved significantly the reproducibility of the kinetic dependencies of the dynamic surface elasticity and surface tension (Fig. 2S and 3S of the supporting information). In spite of the insufficient reproducibility the kinetic dependencies of the surface properties of lysozyme solutions agree qualitatively with the corresponding results in the literature [39]. All the kinetic dependencies are monotonic and the dynamic surface elasticity reaches a plateau value close to equilibrium (∼80 mN/m). These data are typical for solutions of globular proteins [22–24] and also resemble the kinetic dependencies for aqueous dispersions of charged rigid nanoparticles [40]. This can indicate that lysozyme preserves its globular structure in the surface layer. The kinetic dependencies of the surface properties of mixed lysozyme/GuHCl solutions are similar to the results for solutions of the pure protein at GuHCl concentrations up to 2 M (Fig. 2S of the supporting information). The induction period decreases gradually with the increase of GuHCl concentration. Simultaneously the rate of change of the surface properties increases and they approach equilibrium values within the time of experiment. The same behavior has been observed earlier for solutions of other globular proteins and can be explained by the increase of the solution ionic strength with the increase of GuHCl concentration [22,23]. At the same time, one can observe a slight decrease of the maximum values of the modulus of the dynamic surface elasticity from ∼80 up to ∼65 mN/m (Fig. 2). This effect is opposite to the expectations for solutions with the increased ionic strength and to the observations for solutions of other globular proteins. It can indicate a slight loosening of the lysozyme globular structure. According to Perriman et al. this concentration range corresponds to the change of the globule orientation in the surface layer [16]. The transition from GuHCl concentration of 2 M to 4 M results in strong changes of the kinetic dependencies of the dynamic surface elasticity. The induction period disappears entirely and the changes

M.M. Tihonov et al. / Colloids and Surfaces B: Biointerfaces 129 (2015) 114–120

(b)100

(a) 100 90 80 70 60 50 40 30 20 10 0

|ε|, mN/m

|ε|, mN/m

117

0

5

10

15

90 80 70 60 50 40 30 20 10 0 0

20

5

π, mN/m

10

15

20

π, mN/m

Fig. 2. Modulus of the dynamic surface elasticity of lysozyme/GuHCl solutions vs. surface pressure at GuHCl concentrations of 0.5 (a) and 2 M (b). Different markers and colors correspond to different independent measurements. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

of the surface properties are accelerated strongly (Fig. 4S of the supporting information). One can also observe a local maximum of the surface elasticity, which is characteristic for the unfolding of protein globules (Fig. 3a) [22–25]. Similar kinetic dependencies with a local maximum were also obtained for the GuHCl concentration of 6 M (Fig. 3b and Fig. 4S of the supporting information). The results in Figs. 2 and 3 indicate that the unfolding of lysozyme globules in the surface layer starts in the concentration range of GuHCl from 2 to 4 M, i.e. at higher concentrations than in the bulk phase where the process occurs at concentrations close to 1 M [16,28]. This distinction means that the water–air interface has a stabilizing influence on lysozyme globules unlike the case of BSA and BLG solutions where the interface has the opposite effect and the globule destruction occurs at lower concentrations than in the bulk phase [22,23]. Perriman et al. also concluded on the basis of X-ray and neutron reflectivity data that lysozyme globules have higher stability at the interface as compared with their stability in the bulk phase the bulk phase [16]. These authors assumed the selective adsorption of folded lysozyme molecules similar to the case of the selective adsorption of BLG monomers from the solution containing BLG dimers or the renaturation of lysozyme molecules in the surface layer. As a result one can observe the unfolded lysozyme molecules in the bulk phase at lower GuHCl concentrations than in the surface layer. Although urea is not an electrolyte, its influence at relatively low concentrations (<4 M) on the dynamic surface properties of lysozyme solutions is similar to the effect of GuHCl (Fig. 4 and Fig. 3S of the supporting information). The induction period disappears and the rate of change of the surface properties increases noticeably with the increase of denaturant concentration. The maximal value of the surface tension also decreases indicating some slight changes

(b) 100

(a) 100 90 80 70 60 50 40 30 20 10 0

|ε|, mN/m

|ε|, mN/m

of the globular structure. The main difference between the surface properties of lysozyme/GuHCl and lysozyme/urea solutions in this concentration range consists in the decrease of the surface tension to lower values in the latter case. At higher concentrations of the denaturants the kinetic dependencies of surface properties display more significant differences. The surface tension of lysozyme/urea solutions changes during the whole time of experiment (250 min) while for lysozyme/GuHCl solutions it reaches a constant value within the error limits in about 30 min after the surface formation. The most important thing is that the dependencies of the dynamic surface elasticity on the surface age and surface pressure do not display any noticeable local maxima at urea concentrations lower than 8 M (Fig. 5 and Fig. 5S of the supporting information). At higher concentrations one can observe slight and gradual decrease of the surface elasticity at high surface pressures but even in this case the obtained results differ strongly from the data for lysozyme/GuHCl solutions (cf. Figs. 3 and 5). In the case of BSA and BLG solutions the distinctions in the influence of these denaturants on surface properties were not so significant [22–24]. The addition of both urea and GuHCl at concentrations higher than a certain critical value to BSA and BLG solutions resulted in non-monotonic kinetic dependencies of the dynamic surface elasticity with a strong local maximum. This effect gave evidence of the globule unfolding in the surface layer and was a little weaker for the solutions with urea due to its weaker denaturating strength. The results of the given work indicate that the addition of urea does not lead to lysozyme unfolding but can change the protein tertiary structure to a certain extent. The kinetic dependencies of the ellipsometric angle surf corroborate this conclusion (Fig. 6). All the dependencies for lysozyme/urea solutions are similar and any deviations from the mean values are close to the error limits. These data do not indicate

0

5

10

π, mN/m

15

20

90 80 70 60 50 40 30 20 10 0 0

5

10

15

20

π, mN/m

Fig. 3. Modulus of the dynamic surface elasticity of lysozyme/GuHCl solutions vs. surface pressure at GuHCl concentration of 4 (a) and 6 M (b). Different markers and colors correspond to different independent measurements. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

118

M.M. Tihonov et al. / Colloids and Surfaces B: Biointerfaces 129 (2015) 114–120

(b) 100

90 80 70 60 50 40 30 20 10 0

|ε|, mN/m

|ε|, mN/m

(a) 100

0

5

10

15

20

90 80 70 60 50 40 30 20 10 0

0

25

5

π, mN/m

10

15

20

25

π, mN/m

Fig. 4. Modulus of the dynamic surface elasticity of lysozyme/urea solutions vs. surface pressure at urea concentrations of 1 (a) and 2 M (b). Different markers and colors correspond to different independent measurements. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

(b) 100

90 80 70 60 50 40 30 20 10 0

|ε|, mN/m

|ε|, mN/m

(a)100

0

5

10

15

20

25

90 80 70 60 50 40 30 20 10 0

0

5

π, mN/m

10

15

20

25

π, mN/m

Fig. 5. Modulus of the dynamic surface elasticity of lysozyme/urea solutions vs. surface pressure at urea concentrations of 4 (a) and 8 M (b). Different markers and colors correspond to different independent measurements. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

formation do not exceed the error limits. Secondly, the equilibrium value of surf decreases significantly with the increase of GuHCl concentration. The corresponding increase of the refractive index of subphase is not sufficient to explain the observed changes of surf and one has to assume strong changes of the properties of the adsorption layer probably due to the protein unfolding. According to Perriman et al. the lysozyme unfolding in the surface layer is accompanied by the increase of the adsorbed amount and the thickness of the adsorption layer [16]. The local density of aminoacid residues decreases and the solvent concentration increases in the

5.0

5.0

4.5

4.5

4.0

4.0

3.5

3.5

3.0

3.0

Δsurf, deg

Δsurf, deg

strong changes of the adsorption layer structure and the structure of the protein globules. The dynamic surface elasticity and surface tension prove to be more sensitive to the changes in the protein tertiary structure (Fig. 5 and Fig. 5S) than the ellipsometric angle and therefore the adsorbed amount in this case. This is a consequence of the proximity of the adsorbed amount to the value for a saturated monolayer. At the same time, the addition of GuHCl influences strongly the kinetic dependencies of the ellipsometric angle (Fig. 7). Firstly, it accelerates strongly the adsorption kinetics and at concentrations higher than 2 M all the changes within 10–15 min after the surface

2.5 2.0 1.5

2.5 2.0 1.5 1.0

1.0 0.5

0.5

0.0

0.0 0

10000 20000 30000 40000 50000 60000

t, s Fig. 6. Ellipsometric angle surf vs. time for lysozyme/urea solutions at urea concentrations of 0 (violet crosses), 2 (black squares), 6 (red circles) and 8 M (green diamonds). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

0

10000 20000 30000 40000 50000 60000

t, s Fig. 7. Ellipsometric angle surf vs. time for lysozyme/GuHCl solutions at GuHCl concentrations of 0 (violet crosses), 0.5 (black squares), 2 (red circles), 4 (green diamonds) and 6 M (blue hexagons). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

M.M. Tihonov et al. / Colloids and Surfaces B: Biointerfaces 129 (2015) 114–120

119

layer leading to the decrease of the mean refractive index and thereby of the angle surf according to relations (1) and (2).

4. Discussion The strong change of the kinetic dependencies of the dynamic surface elasticity with the increase of denaturant concentration in a narrow range is characteristic for solutions of globular proteins [22–25]. The kinetic dependencies at low concentrations of the denaturants are monotonic and similar to the data for dispersions of rigid nanoparticles. If the concentration exceeds a certain critical value, the dynamic surface elasticity starts to decrease after a local maximum and reaches relatively low values close to equilibrium. Approximately the same behavior was observed for solutions of flexible amphiphilic polymers [25,41]. According to the theory of surface dilational viscoelasticity of polymer solutions the relaxation of surface stresses occurs mainly due to the segment exchange between the proximal region of the surface layer and the distal one – the region of tails and loops [41]. If the affinity of the segments to the surface is sufficiently high, the adsorbed macromolecules at the beginning of adsorption have almost flat two-dimensional conformations without long loops and tails. The increase of the adsorbed amount in the course of adsorption leads to stronger repulsion between the segments and, consequently, to the increase of the surface elasticity. Gradually some loops and tails appear and the surface stresses can be relaxed at the expense of the segment exchange between the proximal and distal regions of the surface layer. The dynamic surface elasticity starts to decrease due to the increase of the number of loops and tails and thereby goes through a local maximum. The time resolution of conventional surface rheological methods allows observation of this local maximum of the surface elasticity at low concentrations of amphiphilic polymers like poly(N-isopropylacrylamide) or poly(vinylpyrrolidone) [41]. One can observe similar kinetic dependencies of the dynamic surface elasticity with local maxima also for solutions of non-globular proteins or unfolded globular proteins at high denaturant concentrations [25]. The main distinction between the results for polymer and protein solutions is that in the latter case the dynamic surface elasticity close to equilibrium is higher – between 10 and 30 mN/m as compared with 3–5 mN/m for polymer solutions. This difference can be connected with some additional restrictions of the protein dynamics in the surface layer due to some elements of the secondary structure or intramolecular disulfide bonds. The kinetic curves of the dynamic surface elasticity of lysozyme solutions also change their shape in a narrow concentration range of GuHCl beyond 2 M (Figs. 2S and 4S of the supporting information) and this effect is accompanied by strong acceleration of the adsorption kinetics. The ellipsometric data corroborate the idea of strong changes of the lysozyme tertiary structure in this concentration range (Fig. 7). At the same time, the decrease of the surface elasticity beyond the maximum is not as strong as in the case of BSA and BLG solutions and the surface elasticity decreases close to equilibrium only up to about 40 mN/m. This relatively high value indicates some restrictions of the conformational freedom of lysozyme molecules and can be explained by only partial protein unfolding in the surface layer. Hedoux et al. also came to a similar conclusion on the partial lysozyme unfolding in the bulk phase under the influence of GuHCl [28]. According to these authors the polypeptide chain of lysozyme at high concentrations of the denaturant cannot be represented by a model of random coil conformation. It is possible to assume that even at GuHCl concentrations higher than 2 M the conventional model of a concentrated proximal region of the surface layer and a dilute distal region of loops and tails protruding into the bulk phase does not describe lysozyme adsorption layer. Nevertheless, the neutron reflectivity shows the increase of the layer

Fig. 8. Schematic representation of the impact of GuHCl and urea on lysozyme tertiary structure.

thickness at GuHCl concentrations higher than 2 M [16]. The partial protein unfolding probably leads to the formation of some short loops and the segment exchange between the two region of the surface layer can lead to the relaxation of surface stresses but to the less extent than in the case of more unfolded macromolecules, for example, of BSA and BLG (Fig. 8). The higher stability of lysozyme globules and the formation of a more rigid adsorption layer can be connected with a special stabilizing role of disulfide bonds in this case. In the molecules of other proteins they connect neighboring aminoacid residues and thereby do not limit strongly the flexibility of unfolded macromolecules. In the case of lysozyme the disulfide bonds between the 6th and 127th and also between the 30th and 115th aminoacid residues make the globule more rigid [42] and consequently do not allow formation of long loops and tails in the surface layer. The results on the dynamic surface elasticity of lysozyme/urea solutions (Fig. 5) indicate only gradual changes of the adsorption layer structure with the increase of urea concentration and require another explanation. The kinetic dependencies of the ellipsometric angle do not show any significant changes of this structure and thereby any signs of the globule unfolding (Fig. 6). The decrease of the surface elasticity with urea concentration (Figs. 4 and 5) is probably connected with the softening of the globular structure without its destruction. At the same time, the strong drop of the surface tension at high urea concentrations (Fig. 9) implies the increase of the concentration of hydrophobic aminoacid residues in the proximal region of the surface layer close to the gas phase. This is possible if the globules are soft, the hydrophobic groups are mobile and some of them can move from the interior of the globule to its surface (Fig. 7). These features are consistent with the molten globule state of lysozyme molecules, which was assumed by Hedoux et al. in their study of the lysozyme denaturation in the bulk phase [28]. The Raman spectroscopy shows that urea almost does not influence lysozyme secondary structure but makes the tertiary structure more mobile. GuHCl and urea interact with different groups of protein molecules – hydrophobic in the former case and polar groups in the latter [3,28]. These distinctions result in different denaturation mechanisms and in significantly different effects of the two denaturants on the dynamic surface properties. Unlike urea, GuHCl leads to the hardening of the protein dynamics [28]. Although the unfolding releases some hydrophobic groups from the globule interior, they are able to interact between themselves and do not move to the boundary with the gas phase (Fig. 8). As a result the surface tension decreases to a less extent than in the case of lysozyme/urea solutions (Fig. 9). At high GuHCl concentrations (>3 M) the interactions

120

M.M. Tihonov et al. / Colloids and Surfaces B: Biointerfaces 129 (2015) 114–120

75

(a)

(b)

75 70

70

γ, mN/m

γ, mN/m

65 65

60 55

60 50 55

45 0

50

100

150

200

250

0

t, min

50

100

150

200

250

t, min

Fig. 9. Kinetic dependencies of the dynamic surface tension of lysozyme/GuHCl solutions (a) at GuHCl concentrations of 0.5 (black squares), 2 (red circles), 4 (green diamonds), 6 M (blue hexagons) and lysozyme/urea solutions (b) at urea concentrations of 2 (black squares), 4 (red circles), 6 (green diamonds), 8 M (blue hexagons). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

between hydrophobic groups can lead to lysozyme aggregation in the surface layer [16]. The difference between the interactions of urea and GuHCl with lysozyme globules can also explain different effects of these denaturants on the amyloid fibril formation. The partial unfolding of lysozyme is a prerequisite of this process [7]. Therefore the addition of GuHCl accelerates strongly the lysozyme fibril formation while urea has almost no effect [31]. 5. Conclusions Measurements of the kinetic dependencies of the dynamic surface properties and ellipsometric angle of lysozyme/urea and lysozyme/GuHCl solutions show that these two denaturants have different influence on the adsorption layer structure. The dynamic surface elasticity and surface tension decrease gradually with the increase of urea concentration but the kinetic dependencies of the ellipsometric angle surf and consequently of the adsorbed amount do not display any significant changes. These results are connected with a gradual transformation of lysozyme globules in the surface layer into the molten globule state similar to the corresponding transition in the bulk phase with the increase of urea concentration [28]. Lysozyme preserves its molten globule state in the course of adsorption and the interface has no specific effects unlike the case of BSA and BLG solutions [22–24]. On the contrary, the increase of GuHCl concentration leads to an abrupt transition from monotonic to non-monotonic kinetic dependencies of the dynamic surface elasticity of lysozyme solutions in a narrow concentration range and to strong changes of the ellipsometric data. These peculiarities indicate the beginning of the globule unfolding in the surface layer but the observed effect is not as strong as in the case of BSA and BLG solutions. The disulfide bonds between remote aminoacid residues of the protein chain restrict the mobility of the partially unfolded lysozyme molecule and do not allow formation of long loops and tails in the surface layer. The obtained results demonstrate the high sensitivity of the surface dilational rheological properties to protein conformations at the liquid–gas interface.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34]

Acknowledgements The work was financially supported by the Russian Foundation of Basic Research (RFFI No. 14-03-00670 a) and St. Petersburg State University (research grant No. 12.38.241.2014).

[35] [36] [37] [38]

Appendix A. Supplementary data

[39] [40]

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.colsurfb. 2015.03.034.

[41] [42]

A.V. Finkelstein, O.B. Ptitsyn, Protein Physics, Academic Press, Boston, 2002. V.N. Uversky, Chem. Rev. 114 (2014) 6557. J.L. England, G. Haran, Annu. Rev. Phys. Chem. 62 (2011) 257. K.A. Dill, J.L. McCallum, Science 338 (2012) 1042. J. Habchi, P. Tompa, S. Longhi, V.N. Uversky, Chem. Rev. 114 (2014) 6561. S.S.-S. Wang, K.-N. Liu, B.-W. Wang, Eur. Biophys. J. 39 (2010) 1229. S.-Y. Ow, D.E. Dunstan, Soft Matter 9 (2013) 9692. G. Navarra, F. Troia, V. Militello, M. Leone, Biophys. Chem. 177 (2013) 24. I. Langmuir, V.J. Schaefer, J. Am. Chem. Soc. 60 (1938) 1351. L.K. James, L.G. Augenste, Adv. Enzymol. Relat. Areas Mol. Biol. 28 (1966) 1. D.E. Graham, M.C. Phillips, J. Colloid Interface Sci. 70 (1979) 427. J.R. Lu, T.J. Su, R.K. Thomas, J. Penfold, J. Webster, J. Chem. Soc. Faraday Trans. 2 (94) (1998) 3279. J.R. Lu, T.J. Su, B.J. Howlin, J. Phys. Chem. B 103 (1999) 5903. A.W. Perriman, M.J. Henderson, S.A. Holt, J.W. White, J. Phys. Chem. B 111 (2007) 13527. D.J.F. Taylor, R.K. Thomas, J. Penfold, Adv. Colloid Interface Sci. 132 (2007) 69. A.W. Perriman, M.J. Henderson, C.R. Evenhuis, D.J. McGillivray, J.W. White, J. Phys. Chem. B 112 (2008) 9532. P.A. Wierenga, H. Gruppen, Curr. Opin. Colloid Interface Sci. 15 (2010) 356. C. Postel, O. Abillon, B. Desbat, J. Colloid Interface Sci. 266 (2003) 74. Y.F. Yano, T. Uruga, H. Tanida, H. Toyokawa, Y. Terada, M. Takagaki, H. Yamada, Langmuir 25 (2009) 32. Y.F. Yano, J. Phys.: Condens. Matter 24 (2012) 50310. M.D. Lad, F. Birembaut, J.M. Matthew, R.A. Frazier, R.J. Green, Phys. Chem. Chem. Phys. 8 (2006) 2179. B.A. Noskov, D.O. Grigoriev, A.V. Latnikova, S.-Y. Lin, G. Loglio, R. Miller, J. Phys. Chem. B 113 (2009) 13398. B.A. Noskov, A.A. Mikhailovskaya, S.-Y. Lin, G. Loglio, R. Miller, Langmuir 26 (2010) 17225. A.A. Mikhailovskaya, B.A. Noskov, E.A. Nikitin, S.-Y. Lin, G. Loglio, R. Miller, Food Hydrocolloids 34 (2014) 98. B.A. Noskov, Adv. Colloid Interface Sci. 206 (2014) 222. M. Malmsten, J. Colloid Interface Sci. 207 (1998) 186. M.A. Walsh, T.R. Schneider, L.C. Sieker, Z. Duauter, V.S. Lamzin, K.S. Wilson, Acta Crystallogr. D 54 (1998) 522. A. Hédoux, S. Krenzlin, L. Paccou, Y. Guinet, M.-P. Flament, J. Siepmann, Phys. Chem. Chem. Phys. 12 (2010) 13189. B. Ibarramolero, J.M. Sanchezruiz, Biochemistry 36 (1997) 9616. B. Chen, H. Zhang, W. Xi, L. Zhao, L. Liang, Y. Chen, J. Mol. Struct. 1076 (2014) 524. S. Emadi, M. Behzadi, Biochem. Biophys. Res. Commun. 450 (2014) 1339. J.J. Kokelaar, A. Prins, M.J. Gee, J. Colloid Interface Sci. 146 (1991) 507. B.A. Noskov, G. Loglio, R. Miller, Adv. Colloid Interface Sci. 168 (2011) 179. H. Motschmann, R. Teppner, in: D. Mobius, R. Miller (Eds.), Ellipsometry in Interface Science in Novel Methods to Study Interfacial Layers, Elsevier, Amsterdam, 2001, pp. 1–42. K. Tonigold, I. Varga, T. Nylander, R.A. Campbell, Langmuir 25 (2009) 4036. A. De Feijter, J. Benjamins, F.A. Veer, Biopolymers 17 (1978) 1759. S. Sundaram, J.K. Ferri, D. Vollhardt, K.J. Stebe, Langmuir 14 (1998) 1208. V.S. Alahverdjieva, D.O. Grigoriev, J.K. Ferri, V.B. Fainerman, E.V. Aksenenko, M.E. Leser, M. Michel, R. Miller, Colloids Surf. A 323 (2008) 167. D.E. Graham, M.C. Phillips, J. Colloid Interface Sci. 76 (1980) 227. P.A. Yazhgur, B.A. Noskov, L. Liggieri, S.-Y. Lin, G. Loglio, R. Miller, F. Ravera, Soft Matter 9 (2013) 3305. B.A. Noskov, A.V. Akentiev, A.Y. Bilibin, I.M. Zorin, R. Miller, Adv. Colloid Interface Sci. 104 (2003) 245. J.-Y. Chang, L. Li, FEBS Lett. 511 (2002) 73.