Protein−Lipid Interactions at the Air−Water Interface - American

Jun 17, 2010 - homogenization of milk,βlg adsorbs to the interface of lipid fat globules and ... face of fat globules during homogenization and stabi...
1 downloads 0 Views 1MB Size
pubs.acs.org/Langmuir © 2010 American Chemical Society

Protein-Lipid Interactions at the Air-Water Interface Ann Junghans,† Chloe Champagne,†,‡ Philippe Cayot,‡ Camille Loupiac,‡ and Ingo K€oper*,†,§ †

Max Planck Institute for Polymer Research, Mainz, Germany, ‡Equipe EMMA, Universit e de Bourgogne, Dijon, France, and §School of Chemical and Physical Sciences, Flinders University of South Australia, Adelaide, Australia Received January 4, 2010. Revised Manuscript Received May 27, 2010

Protein-lipid interactions play an important role in a variety of fields, for example in pharmaceutical research, biosensing, or food science. However, the underlying fundamental processes that govern the interplay of lipids and proteins are often very complex and are therefore studied using model systems. Here, Langmuir monolayers were used to probe the interaction of a model protein with lipid films at the air-water interface. The protein β-lactoglobulin (βlg) is the major component in bovine milk serum, where it coexists with the milk fat globular membrane. During homogenization of milk, βlg adsorbs to the interface of lipid fat globules and stabilizes the oil-in-water emulsion. pH and ionic strength of the subphase had a significant effect on the surface activity of the protein. Additionally, by using lipids with different charges, it could be shown that the interactions between βlg and a phospholipid layer were driven by hydrophobic as well as by electrostatic interactions. βlg preferentially interacted with phospholipids in an unfolded state. This could be either achieved by denaturation at the air-water interface or due to electrostatic interactions that weaken the intramolecular forces of the protein.

Introduction The interactions of proteins with lipid structures are important in various fields. For example, fundamental processes in food science include the interaction of proteins with a variety of interfaces, for example, water-oil or water-air. A detailed understanding of these processes can help in the formulation of novel food products. However, the interactions are often very complex, and thus model systems are utilized to facilitate detailed studies. As a model protein, the globular whey protein β-lactoglobulin (βlg) can be used. βlg is the major component in bovine milk serum1 and belongs to the lipocalin family that is known for specific transporter activity.2 βlg is a small globular protein with 162 residues and a molecular weight of 18.4 kDa with an elliptical diameter of 39 A˚.3,4 At room temperature, neutral pH, and physiological conditions, the native protein is in a dimeric state,5 while the monomeric form is predominant at concentration below 2 mg/mL and at low pH values.6 Additionally, the pH has an effect on the stability and structure of the protein. At lower pH, the protein core is very rigid and exhibits a stable conformation,7,8 while at higher pH, βlg is more flexible and easier to unfold. *Corresponding author. E-mail: [email protected]. (1) Palmer, A. H. The preparation of a crystalline globulin from the albumin fraction of cow’s milk. J. Biol. Chem. 1934, 104, 359-372. (2) Jost, R. Functional Characteristics of Dairy Proteins. Trends Food Sci. Technol. 1993, 4 (9), 283-288. (3) Verheul, M.; Pedersen, J. S.; Roefs, S.; de Kruif, K. G. Association behavior of native beta-lactoglobulin. Biopolymers 1999, 49 (1), 11-20. (4) Kontopidis, G.; Holt, C.; Sawyer, L. Invited Review: {beta}-Lactoglobulin: Binding Properties, Structure, and Function. J. Dairy Sci. 2004, 87 (4), 785-796. (5) McKenzie, H. A.; Ralston, G. B.; Shaw, D. C. Location of sulfhydryl and disulfide groups in bovine .beta.-lactoglobulins and effects of urea. Biochemistry 1972, 11 (24), 4539-4547. (6) Zimmerman, J. K.; Barlow, G. H.; Klotz, I. M. Dissociation of betalactoglobulin near neutral pH. Arch. Biochem. Biophys. 1970, 138, 101-109. (7) Shimizu, M.; Saito, M.; Yamauchi, K. Emulsifying and structural properties of beta-Lactoglobulin at different pHs. Agric. Biol. Chem. 1985, 49 (1), 189-194. (8) Ragona, L.; Pusterla, F.; Zetta, L.; Monaco, H. L.; Molinari, H. Identification of a conserved hydrophobic cluster in partially folded bovine beta-lactoglobulin at pH 2. Fold. Des. 1997, 2 (5), 281-290. (9) Bos, M. A.; Nylander, T. Interaction between beta-lactoglobulin and phospholipids at the air/water interface. Langmuir 1996, 12 (11), 2791-2797.

Langmuir 2010, 26(14), 12049–12053

The protein is easily accessible and has been used in the past in investigations at the air/water or oil/water interface.9 It has also been used in more applied studies on the stabilization of emulsions,10 on gelation,11 and on foaming12 as well as on aroma perception.13 In dairy products, the protein adsorbs to the interface of fat globules during homogenization and stabilizes oil-inwater emulsions.14,15 The protein coexists with milk fat globular membranes.16 Therefore, studies of the protein at the interface of phospholipids can give important information about processes occurring during protein-fat contacts in the processing of milk.17-19 The detailed interaction processes are yet not fully understood;4 however, a more detailed knowledge would yield valuable information for the optimization of, for example, food processing conditions. Here, systematic investigations of protein-lipid interactions have been performed using model membranes at the air-water (10) Klemaszewski, J. L.; Das, K. P.; Kinsella, J. E. Formation and Coalescence Stability of Emulsions stabilized by Different Milk-Proteins. J. Food Sci. 1992, 57 (2), 366-&. (11) Mulvihill, D. M.; Kinsella, J. E. Gelation of beta-Lactoglobulin - Effects of Sodium-Chloride and Calcium-Chloride on the Rhological and Structural Properties of Gels. J. Food Sci. 1988, 53 (1), 231-236. (12) Davis, J. P.; Doucet, D.; Foegeding, E. A. Foaming and interfacial properties of hydrolyzed beta-lactoglobulin. J. Colloid Interface Sci. 2005, 288 (2), 412-422. (13) Seuvre, A. M.; Diaz, M. A. E.; Voilley, A. Retention of aroma compounds by beta-lactoglobulin in different conditions. Food Chem. 2002, 77 (4), 421-429. (14) Cornell, D. G.; Patterson, D. L. Interaction of Phospholipids in Monolayers with Beta-Lactoglobulin Adsorbed from Solution. J. Agric. Food Chem. 1989, 37 (6), 1455-1459. (15) Walstra, P. Physical chemistry of milkfat globules. In Developments in Dairy Chemistry; Fox, P. F., Ed.; Applied Science Publishers: London, 1983. (16) Patton, S.; Keenan, T. W. The milk fat globule membrane. Biochim. Biophys. Acta 1975, 415, 273-309. (17) Boots, J. W. P.; Chupin, V.; Killian, J. A.; Demel, R. A.; de Kruijff, B. Interaction mode specific reorganization of gel phase monoglyceride bilayers by betalactoglobulin. Biochim. Biophys. Acta, Biomembr. 1999, 1420 (1-2), 241-251. (18) Martins, P. A. T.; Gomes, F.; Vaz, W. L. C.; Moreno, M. J. Binding of phospholipids to beta-Lactoglobulin and their transfer to lipid bilayers. Biochim. Biophys. Acta, Biomembr. 2008, 1778 (5), 1308-1315. (19) Lefevre, T.; Subirade, M. Interaction of beta-lactoglobulin with phospholipid bilayers: a molecular level elucidation as revealed by infrared spectroscopy. Int. J. Biol. Macromol. 2000, 28 (1), 59-67.

Published on Web 06/17/2010

DOI: 10.1021/la100036v

12049

Article

Junghans et al.

Figure 1. Pressure-time isotherms of β-lg at pH 3 and pH 7. A three-step adsorption of the protein to the air-water interface could be observed, especially within the first minutes (inset) and is sketched schematically.

interface of a Langmuir film balance. By monitoring the surface pressure, the influence of different types of lipids and of the pH value of the subphase on the surface activity of the protein has been investigated. To evaluate the influence of charge and electrostatic repulsion on the protein core stability, experiments have been performed with films of zwitterionic and negatively charged phospholipids in the presence of positively (pH 3) and negatively (pH 7) charged βlg.

Materials and Methods For all preparation steps, ultrapure water, filtered with a Millipore device (Billerica, MA), was used. Langmuir Isotherms. Isotherms were measured on a Nima (Biolin Scientific, V€astra Fr€ olunda/Sweden) BAM Trough (AM = 925 cm2), with symmetric Delrin twin barriers. The trough was equipped with a Nima surface pressure sensor (PS4) that uses the Wilhelmy plate technique to determine the change in surface tension of the air-water interface in the presence of surfactant molecules. In general, measurements were repeated at least three times for any given set of parameters (type of lipid, pH, etc.) with a very good agreement between the different measurements. Kinetics. The surface pressure was recorded as a function of time at a fixed surface area. Lipids were spread dropwise onto the aqueous subphase using a syringe (Hamilton, Bonaduz, Switzerland). Fifteen minutes after spreading, the protein (2 mg/L) was inserted into the subphase with a pipet (Eppendorf, Hamburg, Germany). Preliminary experiments have shown that no agitation was necessary to enable proper mixing of βlg in the subphase. For measurements at smaller molecular area and therefore higher initial surface pressure, lipids were spread as described above. After a waiting time of 15 min to allow for evaporation of chloroform and relaxation of the system, the surface area was reduced to the desired value. The barrier compression speed was first 10 A˚2/(molecule min) and was lowered to 0 A˚2/(molecule min) to prevent a decrease in surface pressure due to the stopping of the barriers. The protein was added immediately after the selected molecular area was obtained. Protein. β-Lactoglobulin was purified from whey protein powder (bipro, Davisco, Geneva, Switzerland). The powder was slowly dissolved in water (73.5 g of whey powder/L) under constant agitation for at least 2 h. Trichloroacetic acid (Serva, Heidelberg, Germany) was used as precipitant (3.1 g in 20 mL of H2O per 100 mL of protein) solution. The mixture was centrifuged (10 000 rcf, 30 min, 20 °C), and the supernatant was dialyzed 12050 DOI: 10.1021/la100036v

(cutoff = 3500 Da) under constant agitation three times for 2 h against water and three times for 2 h against PBS. Sodium azide (0.2 g/L) (Acros, Geel/Belgium) was added to avoid any microbial degradation of the sample. The concentration and purity of the protein was probed by UV-vis spectroscopy (Lambda900, Perkin-Elmer, Waltham, MA). The absorbance of the diluted protein was measured in the range of 350-200 nm. The shape of the resulting peak gives information about the quality of the sample, as “shoulders” imply impureness. The protein concentration was determined using the absorbance intensity at 278 nm and ε(0.1%) = 0.96. The concentration was about 20 mg/mL, with slight variations for every sample batch, due to small variations in dialysis time and amount of material in the dialysis cell. A microcalorimeter DSC III (Setaram Instrumentation, France) has been used in order to determine denaturation temperature (Td) and enthalpy (ΔH) of protein unfolding for βlg in solution. The temperature was varied between 25 and 110 °C, with a temperature gradient of 0.5 °C min-1. About 600 mg of protein solution (15 g/L of protein) was placed in a hermetically closed cell. PBS buffer solution (0.1 M) was used as a reference. From the obtained thermograms, Td and ΔH were calculated by integration of the measured curves. Lipids. 1,2-Di-O-phytanyl-sn-glycero-3-phosphocholine (DPhyPC) and 1,2-di-O-phytanyl-sn-glycerol (DPhyPG) were obtained from Avanti Polar Lipids (Alabaster, AL). Lipids were dissolved in chloroform (2 mg/mL) and spread onto the aqueous subphase.

Results The surface activity of the pure protein has been investigated as a function of time and at pH 3 and pH 7 (Figure 1). The surface pressure increase was more prominent at the higher pH value. A similar approach has been recently reported, showing that electrostatic interactions play a major role in the interaction of βlg and membranes.20 Yet, the authors have performed experiments at different pH values and used different lipids. We will discuss the differences between the two approached in more details later on. A three-step adsorption process, which has not been reported before, was visible in the isotherms. The process can be interpreted as an initial rapid diffusion of the surface active protein to the interface, leading to a first increase in surface pressure. (20) Zhang, X. Q.; Ge, N.; Keiderling, T. A. Electrostatic and hydrophobic interactions governing the interaction and binding of beta-lactoglobulin to membranes. Biochemistry 2007, 46 (17), 5252-5260.

Langmuir 2010, 26(14), 12049–12053

Junghans et al.

Figure 2. Differential scanning calorimetry of βlg at pH 7 and pH 3. The thermograms clearly show that the protein structure is more compact and stable at pH 3 compared to pH 7, as denaturation temperature and enthalpy increase at this pH.

This step is then followed by protein unfolding due to partial denaturation of the protein at the air-water interface visible as a second step in the isotherm and possible protein aggregation before reaching saturation (inset in Figure 1). An increase in surface pressure can be related to a saturation of formerly unsaturated bonds at the air-water interface. βlg as a globular protein has in its native state relatively little unsaturated bonds. Thus, in order to exhibit a pronounced increase in surface pressure, the protein structure has to at least rearrange, i.e., partially denature. Additional experiments using predenatured protein showed very similar isotherms in terms of the final pressure, yet without the second unfolding-related process (data not shown). The two processes were more pronounced at pH 7 than at pH 3. The differences in the first diffusion process to the interface as well as in the second surface pressure increase Δπ correspond to the compactness of the protein at the respective pH value. As described in the literature, the protein core is more rigid at lower pH values and exposes thus less hydrophobic parts, which makes the protein less surface active.21,22 The more rigid protein is also more difficult to unfold; thus, the denaturation part in the isotherm was slowed down. This led to the reduced effects at pH 3 (Figure 1). Here, the protein is more compact, leading to a smaller partial denaturation and thus to a smaller increase in surface pressure, when compared to the behavior at pH 7. The structure of the protein was verified using DSC (Figure 2). The measured thermograms clearly show that the protein structure is more compact and stable at the lower pH. Compared to pH 7, the denaturation temperature increased from 74.2 to 83.7 °C, and the denaturation enthalpy increased from 0.35 J/g of protein in solution to 0.57 J/g of protein in solution. This clearly indicates that the structure of the protein is more compact and stable at pH 3. Furthermore, it has been reported that βlg exists in a monomeric state at pH 3, while at pH 7 the dimer form is predominant.21,22 At pH 4.6, as been used by Zhang et al.,20 the protein structure is probably a mixture of monomers and dimers, as this (21) Kuwata, K.; Hoshino, M.; Forge, V.; Era, S.; Batt, C. A.; Goto, Y. Solution structure and dynamics of bovine beta-lactoglobulin A. Protein Sci. 1999, 8 (11), 2541-2545. (22) Uhrinova, S.; Smith, M. H.; Jameson, G. B.; Uhrin, D.; Sawyer, L.; Barlow, P. N. Structural changes accompanying pH-induced dissociation of the betalactoglobulin dimer. Biochemistry 2000, 39 (13), 3565-3574.

Langmuir 2010, 26(14), 12049–12053

Article

pH value is very close to the isoelectric point of the protein. However, these pH effects on the monomer-dimer equilibrium of the protein are probably negligible for Langmuir balance measurements as the protein concentration used was very low. Yet, they would have an effect close to the isoelectric point, where the protein net charge is close to zero. In the present results, the dominant effects leading to the observed changes in surface pressure are thus mainly linked to the differences in compactness and stability of the protein structure at pH 3 and pH 7, as seen in the DSC experiments. The interaction of βlg with lipid layers and the influence of the charge of the lipid have been investigated by using two different lipids. DPhyPC is zwitterionic, while DPhyPG is negatively charged, thus forming lipid layers of different charge. Additionally, lipid films of different surface pressure have been used to study the influence of the lipid density on the interactions with the protein. The lipids are derived from archaea bacteria and are know to form densely packed lipid layer, which are fluid at ambient conditions.23 Both lipids were spread at the air-water interface and showed typical pressure-area isotherms (Figure 3). At large molecular areas, the lipids are sparsely distributed at the interface. Upon reduction of the available area, the lipid film is compressed until it collapses at about 50 mN/m. The form of the isotherms was mainly governed by the interactions between the hydrophobic parts. Phase transitions, often seen in amphiphilic molecules, do not occur for the investigated lipids due to interdigitation of the side groups in the hydrocarbon chains. The different head groups only led to small differences in the phase behavior; the isotherms for DPhyPC were shifted to lower areas due to less electrostatic repulsion. Similarly, the pH of the subphase had only very little effect on the form of the isotherms. The interaction between βlg and the different lipids has been studied by adding the protein to lipid films at molecular areas of 190 A˚2 and at 100 A˚2. This was done at pH 3, where the protein is positively charged, and at pH 7, where it is negatively charged. Thus, both the electrostatic effects and the influence of lipid density have been probed. The pressure-time isotherms for the different combinations varied significantly (Figure 4). In general, the addition of the protein to a lipid film led to an increase in surface pressure. This increase depended both on the type of lipid and on the pH value of the subphase. At pH 3 and for films at large molecular areas, the surface pressure increase was similar to the isotherms observed without lipids (Figure 1). However, the addition of the protein to the negatively charged lipid DPhyPG led to an enhanced increase in surface pressure due to electrostatic interactions between the positively charged protein and the negatively charged lipids. At low molecular areas, the surface pressure increased by about 15 mN/m, while the increase for a DPhyPC film was only about 10 mN/m. At pH 7, the protein led to a pronounced increase in surface pressure when added to a dispersed lipid film, however, with no significant difference between the two lipids. The increase in surface pressure of a lipid film, when the protein is added, is due to the surface activity of the protein. In general, as seen before, the protein diffuses to the interface, partially denatures, and thus lowers the surface tension. The amplitude of the surface pressure increase is due to the degree of unfolding of the protein, a process that is governed by multiple parameters, (23) Franz, H.; Dante, S.; Wappmannsberger, T.; Petry, W.; de Rosa, M.; Rustichelli, F. An X-ray reflectivity study of monolayers and bilayers of archae lipids on a solid substrate. Thin Solid Films 1998, 327, 52-55.

DOI: 10.1021/la100036v

12051

Article

Junghans et al.

Figure 3. Chemical structure and πS-AM isotherm of DPhyPC (left) and DPhyPG (right) compressed at 10 A˚2/(molecule min). Afterward, protein has been added to films at different molecular areas indicated by A = 190 A˚2 and B = 100 A˚2.

Figure 4. πS-time isotherm for the addition of βlg to lipid films of DPhyPC and DPhyPG at pH 3 (A) and pH 7 (B).

e.g., pH, charge, concentration, temperature, or interactions with lipids. In the present case, the interactions are influenced by the pH and thus the stability of the protein core as well as by electrostatic interactions. As seen by the DSC measurements, the pH has a dominant effect on the protein structure. The denaturation enthalpy is much higher at the lower pH; the protein is more compact and less easy to denature. At low lipid molecular areas, the protein hardly had to compete with the lipids, and the surface activity of the protein is similar to the situation without protein. Yet the pH-dependent stability of the protein plays an important role on the amplitude of the surface pressure increase. However, the electrostatic interactions between protein and lipids seemed to enhance the denaturation of the protein at or the attraction of the protein to the air-water interface, leading to a higher increase for DPhyPG. For more compressed films, the electrostatic interactions become more important. At lower molecular areas, more charges per area are present and the attraction of the protein should be enhanced. Yet, the lipid films were already relatively compact at a molecular area of 100 A˚2. The protein had thus to compete with 12052 DOI: 10.1021/la100036v

the lipids for space at the air-water interface in order to be able to unfold and denature. At pH 7, the electrostatic interactions are weak and hardly any protein could diffuse to the air-water interface, and the surface pressure changed only slightly. At pH 3, the protein is positively charged and attracted to the negatively charged DPhyPG. This led to a stronger increase in surface pressure upon protein addition. Here, the electrostatic interactions seemed to be high enough to attract the protein to the surface and probably weaken its structure to allow for denaturation. In contrast to previous studies, were the protein was present in dimeric form, the surface pressure increase Δπ at pH 3, where the protein is monomeric, did not depend on the lipid packing density. The increase in surface pressure Δπ was approximately the same at high and low molecular areas for a given lipid type. Probably, electrostatic interactions, more than the free waterair interface, induce protein unfolding, and only then are lipidprotein interactions possible. If dimers are present, this possibly interferes with this process; i.e., a dimer might be too big to penetrate the lipid layer and would therefore be unable to lower the surface tension. Langmuir 2010, 26(14), 12049–12053

Junghans et al.

Article

Conclusion A detailed analysis of protein-lipid interactions, especially for food-related systems, can give valuable information, for example, for the development of novel food processing techniques. The influence of the model protein β-lactoglobulin, one of the major components of milk serum, on the structure of lipid monolayers at the air-water interface of a Langmuir film balance has been investigated. βlg is a surface-active protein and probably denatures, at least partially, at the air-water interface. π-t isotherms showed that βlg approached the air-water interface in a tristep action. First, it quickly diffused to the surface; then it unfolded and aggregated. The degree of unfolding was influenced by the pH and therefore by the compactness of the globular molecule as well as by the charge of lipids. The protein core stability increased with decreasing pH, which could be observed in

Langmuir 2010, 26(14), 12049–12053

π-t isotherms. At pH 3, the protein lowered the surface tension considerable less than βlg at pH 7. This could be correlated to the more compact protein structure at the lower pH. This can be very valuable information for food industry, as this would enable induction of βlg-lipid interaction more directly, leading to more stable and durable emulsions. The interactions between βlg and phospholipids are driven by hydrophobic as well as by electrostatic interactions. The latter were investigated by exposing the protein at different pH values to zwitterionic DPhyPC and negatively charged DPhyPG lipids. The strongest interaction was observed between positively charged βlg at pH 3 and DPhyPG. The protein-lipid interaction with DPhyPC did not show a significant pH dependence and is probably governed by hydrophobic forces.

DOI: 10.1021/la100036v

12053