Same System−Different Results: The Importance of Protein

Anal. Chem. 2009, 81, 3042–3050

Same System-Different Results: The Importance of Protein-Introduction Protocols in Langmuir-Monolayer Studies of Lipid-Protein Interactions Wilhelm R. Glomm,*,† Sondre Volden,† Øyvind Halskau, Jr.,‡ and Marit-Helen G. Ese§ Ugelstad Laboratory, Department of Chemical Engineering, Norwegian University of Science and Technology (NTNU), N-7491 Trondheim, Norway, Department of Biomedicine, University of Bergen, Jonas Lies vei 91, N-5009 Bergen, Norway, and SINTEF Energy Research, N-7465 Trondheim, Norway For studies of protein-lipid interactions, thin films at the air-water surface are often employed as model systems for cell membranes. A convenient manner in which to study these interactions is the Langmuir technique, which allows for formation of monolayer phospholipid films together with a choice of where and how to introduce proteins, according to the desired response variable. Here, a distinction has been made between different interaction protocols and it is also commented upon to what extent introduction of protein to a solution prior to spreading of a lipid film affects the results. This paper describes commonly used methods when working with Langmuir monolayers as membrane mimics and compares the results of four different experimental protocols: formation of a lipid film on top of a protein-containing subphase, injection of protein under an existing, semicompressed phospholipid film (surface pressure 5 mN/ m), and deposition of a protein solution on top of a lipid film contained at either surface pressure 0 mN/m or at surface pressure 5 mN/m. Results obtained from Langmuir isotherms and Brewster angle microscope clearly differentiate between these methods and give insight into under which conditions and at which interfaces the protein interactions are predominant (protein-air or protein-lipid). This paper addresses fundamental issues involving the use of Langmuir monolayers for the study of protein-lipid interactions. Langmuir phospholipid monolayers at the air-water surface are well-established model systems for biological membranes, in that they mimic the interface between the phospholipid and the surrounding aqueous medium.1-4 The molecular arrangement, * To whom correspondence should be addressed. Fax: +47 73 59 40 80. E-mail: [email protected]. † Norwegian University of Science and Technology (NTNU). ‡ University of Bergen. § SINTEF Energy Research. (1) Caetano, W.; Ferreira, M.; Tabak, M.; Sanchez, M. I. M.; Oliveira, O. N.; Kruger, P.; Schalke, M.; Losche, M. Biophys. Chem. 2001, 91, 21–35. (2) de Souza, N. C.; Caetano, W.; Itri, R.; Rodrigues, C. A.; Oliveira, O. N.; Giacometti, J. A.; Ferreira, M. J. Colloid Interface Sci. 2006, 297, 546–553. (3) Vernoux, N.; Maniti, O.; Besson, F.; Granjon, T.; Marcillat, O.; Vial, C. J. Colloid Interface Sci. 2007, 310, 436–445.

3042

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

orientation, and composition at the surface can be well controlled by varying the molecular area and the surface pressure. Additionally, the physicochemical properties of the subphase, such as pH, ionic strength, and temperature, are tunable. These features make Langmuir monolayers a powerful tool with which to study interactions with phospholipid molecules self-organized in welldefined structures according to lipid composition, ionic characteristics of the monolayer, and the surface pressure. A large number of biological reactions take place at interfaces where the main constituents are lipids and proteins. The interaction between proteins and Langmuir phospholipid monolayers can be studied using a number of methods. Among these methods, dissolving/including the protein in the subphase followed by spreading of amphiphilic reagents on the air-water surface is perhaps the most common.1,2,4-9 However, the application of this method is restricted for rare/expensive proteins, as the protein needs to be distributed throughout the entire subphase (∼250 mL for the Minitrough used in this study). Moreover, as proteins are amphiphilic, their presence in the subphase prior to spreading of the phospholipids might result in a mixed protein-lipid monolayer, depending on the concentration and surface activity of the protein. Thus, the interactions studied with this method are not necessarily between proteins from the subphase and a phospholipid monolayer only. Girard-Egrot et al.10,11 have proposed another method of forming protein-lipid films wherein less protein is required. In this method, glycolipid is added to a protein solution to form protein-glycolipid vesicles, and the vesicle solution is spread onto the surface of the subphase using a microsyringe. Although this method requires less protein than that of spreading (4) Wang, X. L.; He, Q.; Zheng, S. P.; Brezesinski, G.; Mohwald, H.; Li, J. B. J. Phys. Chem. B 2004, 108, 14171–14177. (5) Boussaad, S.; Dziri, L.; Arechabaleta, R.; Tao, N. J.; Leblanc, R. M. Langmuir 1998, 14, 6215–6219. (6) Wang, X.; Zhang, Y.; Wu, J.; Wang, M.; Cui, G.; Li, J.; Brezesinski, G. Colloid Surf., B: Biointerfaces 2002, 23, 339–347. (7) Wang, X. L.; Zhang, H. J.; Cui, G. C.; Li, J. B. J. Mol. Liq. 2001, 90, 149– 156. (8) Ihalainen, P.; Peltonen, J. Langmuir 2003, 19, 2226–2230. (9) Rodland, I.; Halskau, O.; Martinez, A.; Holmsen, H. Biochim. Biophys. Acta: Biomembranes 2005, 1717, 11–20. (10) Girard-Egrot, A. P.; Chauvet, J. P.; Boullanger, P.; Coulet, P. R. Langmuir 2001, 17, 1200–1208. (11) Girard-Egrot, A. P.; Godoy, S.; Chauvet, J. P.; Boullanger, P.; Coulet, P. R. Biochim. Biophys. Acta: Biomembranes 2003, 1617, 39–51. 10.1021/ac8027257 CCC: $40.75  2009 American Chemical Society Published on Web 03/24/2009

a phospholipid monolayer on top of a protein solution, more complex experimental steps are needed and the tuning of the protein/lipid ratio in their complex films is not straightforward. Also, the method proposed by Girard-Egrot et al.10,11 still results in a mixed protein-lipid film, and thus contributions from the protein interacting with a lipid film are hard to extract. Introducing the protein after the formation of a phospholipid monolayer at the air-water surface is a conceptually different approach of studying interactions between proteins and phospholipid films. There are two main routes within this approach: (i) injecting protein into the subphase below the phospholipid monolayer3,12 and (ii) spreading a protein solution on top of the film-covered subphase.13 The latter was first reported by Yin et al. in 2005,13 where a solution of human serum albumin was spread on top of a subphase covered with a layer of octadecylamine. While both approaches involve observing what happens as a protein is introduced to a phospholipid film, route (i) is perhaps more in sync with the use of Langmuir monolayers as a model for half of a cell membrane, as the proteins are approaching the phospholipid film from the surrounding aqueous medium, whereas route (ii) introduces the protein to the hydrophobic tails of the phospholipids, which are not exposed in an intact biological membrane. Considering the vast literature on protein-lipid interactions using the Langmuir technique, it is of great interest to know whether and how the method for studying protein-lipid interactions affects the results. In this study, we have studied the interaction between a zwitterionic phospholipid and bovine serum albumin (BSA) using four different methods (illustrated in Scheme 1): (i) spreading the phospholipid film on top of a proteincontaining subphase (hereafter referred to as Subphase), (ii) spreading the protein solution on top of the subphase covered with the phospholipid in a 2-D gaseous state (hereafter referred to as Top Π ) 0), (iii) spreading the protein solution on top of a precompressed (to Π ) 5 mN/m) phospholipid film (hereafter referred to as Top Π ) 5), and (iv) injecting the protein solution into the subphase under the phospholipid film precompressed to Π ) 5 mN/m (hereafter referred to as Injected). The surface pressure of 5 mN/m was chosen based on three criteria: (1) the protein should be introduced to a complete monolayer of the phospholipid but (2) not so tightly packed as to exclude intercalation, and (3) the remaining trough area after monolayer formation should be sufficient to characterize the effect of protein interaction upon compression as well as expansion. MATERIALS AND METHODS Materials. Bovine serum albumin (BSA) and 1-oleoyl-2stearoyl-sn-glycero-3-phosphocholine (SOPC) were purchased from Sigma Aldrich. Chloroform (p.a.) was obtained from VWR International. NaCl, KCl, Na2HPO4, KH2PO4, and NaOH were purchased from Merck. All chemicals were used without further treatment. Phosphate buffered saline (PBS) was prepared by adjusting an aqueous solution of NaCl (137 mM), KCl (2.7 mM), Na2HPO4 (10 mM), and KH2PO4 (1.8 mM) to pH 7.4 using small quantities of NaOH (1 M). (12) Kiss, E.; Dravetzky, K.; Hill, K.; Kutnyanszky, E.; Varga, A. J. Colloid Interface Sci. 2008, 325, 337–345. (13) Yin, F.; Kafi, A. K. M.; Shin, H. K.; Kwon, Y. S. Thin Solid Films 2005, 488, 223–229.

Langmuir Isotherms. Surface pressure-area isotherms were recorded using a KSV Langmuir Minitrough doublebarrier system (KSV LTD, Finland) using the manufacturer’s own software at 25 °C. The trough is made of Teflon with barriers of Delrin. Prior to introduction of the film material, the surface was swept with a vacuum pump-connected to a Pasteur pipet. Chloroform was used as a spreading solvent for the phospholipid samples. Introduction of SOPC was done by carefully spreading 15 µL of a 1 mg/mL lipid solution onto the surface using a 25 µL Hamilton syringe. The spreading solvent was allowed to evaporate for at least 15 min before compression was initialized. Film compression and relaxation were carried out with a constant barrier speed at 5 mm/ min while an electrobalance recorded the surface pressure. All systems studied here were run through one compression-relaxation cycle under these conditions. As a reference, the surface pressure of a film-free surface (or the surface pressure of a BSAcontaining subphase in the case of the Subphase protocol) was used. Brewster Angle Microscopy. Visualization of the monolayers was done by means of Brewster angle microscopy (BAM) using a KSV BAM 300. The instrument consists of a standard 10 mW He-Ne laser, emitting p-polarized light at a wavelength of λ ) 632.8 nm by use of a high-quality Glan-Thompson polarizer. The reflected light is imaged to a computer controlled CCD camera with 768 × 494 pixels through a 10× magnification objective, yielding a spatial resolution of ∼2 µm. A black wedge-shaped glass plate is placed at the bottom of the trough to reflect any light transmitted through the subphase out of the optical axis and to minimize the convection on the trough. Protein Interaction Protocols. Briefly, BSA was introduced to the phospholipid film by the following four methods (see also Scheme 1 for a visual representation): Subphase. The phospholipid was spread on top of a subphase containing BSA (50 nM). Top Π = 0. In this protocol, first reported by Yin et al.,13 50 µL of a 25 µM BSA solution (in PBS), corresponding to 50 µg BSA (or 0.5 nM BSA in the trough), was carefully spread on the surface 15 min after SOPC had been deposited at the interface, with the barriers completely open and SOPC in a 2-D gaseous state. Within 60 min following spreading of BSA onto the surface, compression was initiated. During the 60 min equilibration period, surface pressure as a function of time (hereafter referred to as Π(t)) was recorded. Top Π = 5. In a slight adaptation of the previous protocol, the SOPC film was compressed to Π ) 5 mN/m before the BSA was introduced. After the spreading of BSA, the film was kept at a constant area corresponding to ΠSOPC ) 5 mN/m for 60 min while recording Π(t). After 60 min, compression was initiated. Injected. The SOPC film was compressed to Π ) 5 mN/m. After the target pressure was reached, the trough area was kept constant and 0.5 mL of a 25 µM BSA solution (in PBS), corresponding to a total BSA concentration in the trough of 50 nM, was injected under the SOPC film from outside the barriers using an L-shaped syringe needle (see also Scheme 1). After recording Π(t) for 60 min, compression was initiated. Injecting 0.5 mL of PBS buffer did not affect the surface pressure or the BAM focus. For all protocols, surface pressure-area (Π-A) isotherms and Π(t) curves (where applicable) were recorded Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

3043

Scheme 1. Summary of Protein-Introduction Protocolsa

a Situation before (left column) and after (right column) the introduction of protein to the lipid film. Arrows with blue outline indicate compression prior to the situation shown. Protein solutes are depicted as dissolved in the subphase (green spheroids), interacting with hydrophobic lipid chains or the air-water interface (red spheroids), and interacting with lipid headgroups (blue spheroids). Injected: Protein is introduced by an L-shaped syringe under the precompressed film. Subphase: Protein is predissolved in the subphase, and has access to the air-water surface prior to compression, thus forming mixed lipid-protein films. Top Π ) 0: Protein is introduced onto an incompletely formed film, where lipid molecules are largely in a 2-D gaseous phase. Top Π ) 5: Protein is introduced onto a 2-D lipid film precompressed to an ordered monolayer.

for proteins in the absence of SOPC for comparison. Between runs, the trough was washed with ethanol and ultrapure (MilliQ) water and care was taken to mechanically remove protein from the trough surface by wiping it with an ethanol-soaked tissue paper. RESULTS AND DISCUSSION Surface Pressure-Area Profiles. Protein-Only Isotherms. Surface pressure-trough area (Π-A) isotherms of BSA-only introduced by the three major methods described here, present in the subphase, spread on top of the subphase and injected into the subphase, denoted Subphase, Top, and Injected, respectively, are shown in Figure 1. All three methods of introducing BSA resulted in strong film formation with surface pressures exceeding 3044

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

25 mN/m at full compression. Thus, spreading of any other filmforming agent after BSA has been introduced will lead to mixed monolayer behavior. The protein films formed by dissolving in or injecting BSA into the subphase are Gibbs monolayers (i.e., monolayers that are formed by substances soluble in the bulk), whereas spreading BSA on top of the subphase more closely resembles the formation of a Langmuir monolayer, with the caveat that the film material remains soluble in the subphase, and also that the spreading solvent (in this case PBS buffer) is not chemically inert with respect to film material and subphase, and that it does not evaporate at a higher rate than the subphase. When a protein solution (50 nM) is used as the subphase, the equilibrium between the bulk and surface has been reached. It

Table 1. Surface Pressure Increase from Introduction of Protein and Width of Hysteresis Loop in a Single Compression-Decompression Cycle ∆ΠBSA (mN/m)a

system SOPC BSA in Subphase BSA spread on Top BSA Injected Subphase Top Π ) 0 mN/m Top Π ) 5 mN/m Injected

Single Component Films N/A N/A 5.0 ± 0.4 6.7 ± 0.5

Protein-Phospholipid Films 15 ± 3c 4.7 ± 0.3 1.6 ± 0.3 2.4 ± 0.8

∆Ahysteresis (cm2)b 5.6 ± 0.8 42 ± 4 44 ± 5 37 ± 8 N/Ad 36 ± 2 8±2 15 ± 3

a Increase in surface pressure 60 min following introduction of BSA. Measured at Π ) 10 mN/m. c Refers to surface pressure increase upon addition of SOPC to BSA-containing subphase. d Initial surface pressure following addition of SOPC exceeded 10 mN/m. b

Figure 1. Surface pressure-film area (Π-A) isotherms of BSA at the air-water surface. BSA was introduced either as dissolved in the subphase (Subphase, [BSA] ) 50 nM), spread on top of the subphase (Top, [BSA] ) 0.5 nM), or injected into the subphase (Injected, [BSA] ) 50 nM).

should be noted that the surface pressure is set to zero when only the subphase is present, and so the Π-A isotherm of BSA dissolved in the subphase will start at Π ) 0 mN/m. However, Figure 1 clearly shows that there is a film present at the surface prior to the introduction of any other film-forming species when a protein solution is used as the subphase. Thus, the initial surface pressure of Π ) 0 mN/m prior to compression of the film does not represent a film-free surface but rather a starting point where the initial protein film has been baseline-subtracted. It should be noted that of the three Π-A isotherms shown in Figure 1, the protein film resulting from BSA being dissolved in the subphase is more rigid and appears to reach a higher surface pressure upon maximum compression than the films formed by injecting BSA into the subphase or by spreading the protein on top of the subphase. The films resulting from injecting BSA or spreading it on top of the subphase resulted in very similar film behavior (Figure 1). With the use of both of these methods, BSA was introduced to a subphase consisting of buffer only, and so the surface pressure increase of introducing BSA relative to a filmfree surface could be determined, yielding an initial increase in surface pressure of 6.7 and 5.0 mN/m for BSA injected and BSA spread on top, respectively (Table 1). Considering the difference in protein concentration between these two methods (corresponding to bulk concentrations of 50 nM, same as for BSA dissolved in the subphase, and 0.5 nM for BSA injected into and spread on top of the subphase, respectively), the observation that the two isotherms (Figure 1, Injected and Top) are more or less identical is remarkable. Specifically, it means that despite the film-forming material being completely soluble in the subphase, it is possible to control the partitioning between bulk and surface by varying how and where the protein is introduced. Protein-Phospholipid Isotherms. Surface pressure-area (Π-A) isotherms of BSA-phospholipid films at the air-water surface as well as the compression isotherm of the phospholipid SOPC only are shown in Figure 2. A summary of isotherm data is listed in Table 1. From the Π-A isotherm of SOPC only, the mean molecular area (MMA), obtained from the inflection point marking the transition from a gaseous to a liquid state, was found to be 175 Å2/molecule. The collapse pressure of the pure SOPC film

Figure 2. Surface pressure-film area (Π-A) isotherms of the BSA-SOPC systems studied here at the air-water surface. The Π-A isotherm of the pure phospholipid (SOPC) is shown for comparison.

ΠC (i.e., the highest surface pressure to which a monolayer can be compressed without collapsing/forming multilayers) was found to be ∼43 mN/m. Introduction of proteins by any of the four protocols described above did not significantly affect the collapse pressure of the BSA-SOPC films (data not shown). When the protein is introduced by dissolution into the subphase prior to spreading of the phospholipid (Figure 2, subphase) and when the protein is spread on top of a subphase covered with phospholipid in a 2-D gaseous state (Figure 2 Top Π ) 0), the Π-A isotherms deviate significantly from the Π-A isotherm of SOPC only. From comparison with protein-only isotherms in Figure 1, it is apparent that the characteristics of the protein are dominating in the isotherms Subphase and Top Π ) 0 in Figure 2, indicating that BSA is present in the SOPC film. As shown above, introducing BSA by either method results in formation of a protein film at the air-water surface, indicating that both protocols (Subphase and Top Π ) 0) lead to the formation of a mixed protein-phospholipid film. In the Subphase protocol, the phospholipid is spread on top of a protein film. Thus, the initial surface pressure increase prior to compression of the film (15 mN/m, Table 1) emanates from addition of SOPC to a BSA-containing subphase rather than from Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

3045

the increase in surface pressure from introduction of BSA (∆ΠBSA) as for the other protocols. Because the phospholipid resides in a 2-D gaseous state, the protein is able to form a film at the air-water surface by occupying the vacant surface area when employing the protocol adapted from Yin et al.13 (Top Π ) 0). Introduction of BSA via the Top Π ) 0 protocol resulted in a ∆ΠBSA of 4.7 mN/m (Table 1), which confirms the presence of a second film-forming species at the surface. With the use of the Langmuir monolayer technique for the study of protein-lipid interactions using either of these protocols (Subphase and Top Π ) 0), it is important to keep in mind that the resulting Π-A isotherms do not describe interactions between protein in the subphase and a pure phospholipid monolayer. Rather, these two protocols allow the study of protein-lipid interactions within the mixed protein-phospholipid film. Moreover, it should be noted that as the Π-A isotherms in Figure 2 represent contributions from the phospholipid and either protein partitioned at the surface (Subphase) or spread on top of the subphase (Top Π ) 0), the use of mean molecular area (MMA) values should be abandoned, as the concentration of surface species is no longer directly quantifiable. The Π-A isotherms of BSA introduced to SOPC films compressed to Π ) 5 mN/m (Top Π ) 5 and Injected, Figure 2) display mostly lipid characteristics, indicating that the internal interactions in the phospholipid film are not severely disturbed. From the initial surface pressure increase following introduction of BSA, ∆ΠBSA, some interaction between the protein and the phospholipid film can be observed (Table 1). The observed ∆ΠBSA for the Injected system is only slightly higher than for Top Π ) 5 (2.4 vs 1.6 mN/m, Table 1) and thus very low considering that the amount of BSA in the Top Π ) 5 protocol is only 1% of the amount of protein introduced in the Injected protocol. However, while only a fraction of the protein introduced by injection reaches the surface and interacts with the SOPC film, BSA spread on top of an SOPC film compressed to Π ) 5 mN/m is probably retained as “islands” associated with the extended hydrocarbon chains without disturbing the SOPC film significantly, as indicated by the observation that the Top Π ) 5 compression isotherm retains the characteristics of SOPC. Compression-Expansion Hysteresis Curves. Film behavior during one compression-expansion cycle for the systems studied here is depicted in Figure 3. The compression-expansion cycle of the film formed by SOPC only is narrow, with no significant hysteresis, indicating that the film is stable and reversible under these conditions. When BSA is present in the subphase (Figure 3A Subphase) the compression-expansion curve shows a wide hysteresis loop. As discussed above, the compression is dominated by the protein, whereas the expansion part of the isotherm displays more of the phospholipid characteristics, indicating that some of the protein is expelled from the film at higher surface pressures. As the surface pressure upon complete expansion is still very high compared to that of the SOPC-only monolayer, BSA is still associated with the film following one compression-expansion cycle. Thus, the plateaulike region found in both the compression and (to a lesser extent) expansion isotherms is likely a result of BSA being solubilized into the subphase rather than from high compressibility of the protein domains in the mixed film. Moreover, the high initial surface pressure may indicate that 3046

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

Figure 3. Compression-expansion isotherms of the BSA-SOPC systems studied here. The corresponding compression-expansion curve for the pure phospholipid is shown for comparison. Here, dashed lines represent the expansion part of the compression-expansion cycle for the BSA-SOPC systems, and the dotted line represents the expansion part of the SOPC-only cycle.

the phospholipid is not homogeneously dispersed in the protein film but rather resides as islands until surface area is made available due to protein being solubilized into the subphase. When BSA is injected below a phospholipid film compressed to Π ) 5 mN/m (Figure 3A, Injected), the resulting film emanates from BSA interacting directly with the SOPC film and not from a mixed protein-phospholipid film. The slope of both the compression and expansion isotherms are almost identical to those of the corresponding SOPC-only film, indicating that little or no protein is present in the film. Rather, the increase in surface pressure following injection is likely caused by BSA directly associated with the phospholipid headgroups. As can be seen from Figure 3A, the expansion isotherm of Injected displays an increase in surface pressure as the barriers approach full extension. This upturn was found to be highly reproducible and is consistent with protein from the subphase reaching the surface as the surface area becomes available following relaxation of the phospholipid film. Compression-expansion cycles of the two protocols wherein BSA is introduced by spreading on top of the subphase are shown in Figure 3B. When spreading a BSA solution on top of a subphase covered with SOPC in a 2-D gaseous state (Figure 3B Top Π ) 0), there is a significant hysteresis loop similar to what was found for protein dissolved in the subphase (Figure 3A, Subphase). During the compression part of the cycle, BSA dominates the film

properties, whereas the film properties more closely resemble those of SOPC-only during film expansion. As for Subphase, this indicates that BSA is solubilized into the subphase during compression. Interestingly, the Top Π ) 0 compression-expansion cycle displays an increase in surface pressure upon full extension as was also found for Injected, indicating that protein from the subphase migrates toward available surface area following relaxation of the film. There are several fundamental problems with the Top Π ) 0 protocol adapted from Yin et al.,13 including the issue that the protocol results in mixed protein-phospholipid film formation as discussed above. BSA does not satisfy the criterion of a suitable film material for the Langmuir technique in this case, as it is readily solublized in the subphase and thus not able to form a stable film. As seen for the Subphase and Top Π ) 0 systems, BSA is solubilized into the subphase upon compression, rendering quantitative control of the surface species impossible. Moreover, using a PBS solution to spread BSA onto a PBS subphase violates the criteria for a suitable spreading solvent.14 The spreading solvent should aid in the spontaneous dispersal of the film forming molecules and ensure that the film material resides in monomeric form at the air-water surface. To accomplish this, the spreading solvent must have a positive spreading coefficient on the subphase and have sufficient solvent power with respect to the film-forming species. Following introduction to the air-water surface, the spreading solvent should rapidly and completely evaporate from the system. Thus, a suitable spreading solvent must be volatile and evaporate within a reasonably short period of time (typically 15 min). If the spreading solvent is soluble in the subphase, some of the film-forming material might be solubilized into the subphase upon deposition. Since one of the film-forming materials in this protocol (BSA) is readily soluble in the subphase, this is most certainly a concern when using the Top Π ) 0 protocol. Here it should be mentioned that solubility and thus partitioning of macromolecules in general and proteins in particular can be tuned by the Hofmeister series or by varying the ionic strength and/or pH of the solution.15-21 By employing the “salting-out” effect, BSA could form a stable film using the Top Π ) 0 protocol. However, if the partitioning is significantly altered, this would also affect the other protocols. For example, using the Subphase protocol, “salting-out” would increase the surface concentration of BSA as well as induce aggregation, thus exacerbating the observed difference between this and protocols designed to introduce protein to a preformed phospholipid film. Figure 3B Top Π ) 5 shows the compression-expansion cycle following spreading of BSA onto an SOPC film precompressed to a surface pressure of Π ) 5 mN/m. While the film strongly resembles the SOPC cycle, the increase in surface pressure (14) Gaines, G. L., Jr. Insoluble Monolayers at Liquid-Gas Interfaces; Wiley: New York, 1966. (15) Chen, X.; Yang, T.; Kataoka, S.; Cremer, P. S. J. Am. Chem. Soc. 2007, 129, 12272–12279. (16) Hofmeister, F. Arch. Exp. Pathol. Pharmakol. 1888, 24, 247–260. (17) Ornebro, J.; Nylander, T.; Eliasson, A. C.; Shewry, P. R.; Tatham, A. S.; Gilbert, S. M. J. Cereal Sci. 2003, 38, 147–156. (18) Waziri, S. M.; Abu-Sharkh, B. F.; Ali, S. A. Biotechnol. Prog. 2004, 20, 526– 532. (19) Zhang, Y. J.; Cremer, P. S. Curr. Opin. Chem. Biol. 2006, 10, 658–663. (20) Zielenkiewicz, A. J. Therm. Anal. Calorim. 2007, 89, 893–897. (21) Zielenkiewicz, A.; Zielenkiewicz, W. J. Therm. Anal. Calorim. 2005, 80, 407–411.

Figure 4. Time-dependent surface pressure (Π(t)) curves for the BSA-SOPC systems studied here, as well as for BSA introduced via the Top and Injected protocols. The curves shown here represent the first 60 min following introduction of BSA.

following deposition (Table 1) and the presence of a small hysteresis loop confirms that BSA is associated with the film. Also noteworthy is the absence of the small increase in surface pressure close to full expansion observed for Injected and Top Π ) 0. This indicates that with the use of the Top Π ) 5 protocol, there is no migration of protein from the subphase to the surface following relaxation of the phospholipid film, despite the amount of protein being identical to what is used in Top Π ) 0. Consequently, the protein might reside as islands on top of the lipid film, as the SOPC film is packed too tightly for BSA to penetrate into the subphase. The observed hysteresis could be the result of lateral migration of the protein from the points of deposition. Alternatively, the hysteresis could arise from mixed-film behavior upon expansion, as the phospholipid film becomes less densely packed, eventually allowing the protein to reach the aqueous surface. As this protocol involves introducing BSA to a hydrophobic interface, it is very likely that the protein denatures to maximize favorable interactions with the phospholipid tails.22 Time-Dependent Surface Pressure (Π(t)) Curves. The timedependent surface pressure behavior recorded during the 60 min equilibration time following protein introduction for the different protocols is shown in Figure 4. Note that this behavior is not shown for the Subphase protocol, as protein is present in the (22) Glomm, W. R.; Halskau, O.; Hanneseth, A. M. D.; Volden, S. J. Phys. Chem. B 2007, 111, 14329–14345.

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

3047

system prior to spreading the phospholipid, and because this protocol typically does not include any equilibration period beyond what is required for evaporation of the spreading solvent (∼15 min). Τhe Π(t) behavior of BSA spread on top of the subphase in the absence of SOPC is shown in Figure 4A (trace a) BSA spread on top). From trace a, it is apparent that BSA spread on top does not reach equilibrium within 60 min. As discussed above, the protocol does not satisfy the requirements for neither film-forming material (i.e., BSA is readily soluble in the bulk phase) nor spreading solvent (as the spreading solvent is the same as in the subphase). Thus, it is very likely that upon spreading BSA on the surface, some protein is solubilized into the subphase and that the slow surface pressure increase following the initial steep increase is caused by protein in the bulk migrating toward the surface. When spreading a BSA solution on top of a surface covered with SOPC in a 2-D gaseous state (Figure 4A, trace b), the initial surface pressure increase occurs at a lower rate but reaches a slightly higher surface pressure than BSA only after 60 min. The lower rate of surface pressure increase is probably due to the phospholipid forming an initial hindrance to the formation of a BSA monolayer. Consequently, BSA is probably not homogeneously distributed but rather resides as islands at the air-water surface. As some of the protein is solubilized upon spreading, it is not unreasonable to speculate that the higher surface pressure after 60 min for trace b is caused by protein from the bulk associating with the SOPC film. The Π(t) curve for the equilibration step in the Top Π ) 5 protocol is shown in Figure 4A, trace c). As can be seen, both the initial slope and the final surface pressure after 60 min are significantly lower than for BSA introduced to the pure subphase or to SOPC in a 2D gaseous state. In this case, the protein is spread on top of a hydrophobic interface, with no possibility to dissolve into the subphase. Thus, the increase in surface pressure is very likely due to the protein spreading out on the hydrophobic interface over time, away from the points of deposition and closer to the film balance. Moreover, the observation that the protein migrates from the initial islands of deposition supports the hypothesis stated above regarding denaturation/conformational changes in BSA deposited via the Top Π ) 5 protocol. The Π(t) curves of BSA injected into a film-free PBS subphase (trace d) and under an SOPC monolayer precompressed to Π ) 5 mN/m (trace e) are shown in Figure 4B. When BSA is injected into a film-free subphase, the system reaches equilibrium within ∼60 min. With introduction of BSA via the Injected protocol (trace e), the rate of surface pressure increase as well as the final surface pressure after 60 min are severely reduced, and the system does not appear to have reached equilibrium. Of the protocols studied here, Injected is the only one which is consistent with using Langmuir monolayers of phospholipids as a model for half of a cell membrane, in that the proteins are introduced to the phospholipid headgroups of a precompressed film. Brewster Angle Microscopy (BAM) Images. In recent years, Brewster angle microscopy has been frequently used in the characterization of monolayers at the air-water surface.6,23-26 (23) Glomm, W. R.; Ese, M. H. G.; Volden, S.; Pitois, C.; Hult, A.; Sjoblom, J. Colloid Surf., A: Physicochem. Eng. Asp. 2007, 299, 186–197. (24) He, Q.; Li, J. B. Adv. Colloid Interface Sci. 2007, 131, 91–98. (25) Sui, G.; Micic, M.; Huo, Q.; Leblanc, R. M. Colloid Surf., A: Physicochem. Eng. Asp. 2000, 171, 185–197.

3048

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

Figure 5. BAM images of SOPC (A-C) and BSA Injected (D-F) at the air-water surface taken at ascending surface pressures upon film compression. Surface pressures (in mN/m) are as follows: (A) 0, (B) 31.5, (C) 40.5, (D) 7.5, (E) 21.5, and (F) 25.5.

Figure 6. BAM images of BSA introduced to SOPC via the Subphase (A-C) and Injected (D-F) protocols at the air-water surface taken at ascending surface pressures upon film compression. Surface pressures (in mN/m) are as follows: (A) 15, (B) 20, (C) 25, (D) 6.5, (E) 16.5, and (F) 31.5.

Here, BAM has been utilized in order to provide complementary information about the topography of the films under the experimental conditions studied here. Because of the large number of images collected, only BAM images of four systems are shown: SOPC, 50 nM BSA injected into the subphase, Subphase, and Injected (protein concentration in the trough for both these protocols is 50 nM). BAM images of these four systems are shown in Figures 5 (SOPC and BSA) and 6 (Subphase and Injected). The corresponding compression isotherms can be found in Figures 1 (50 nM BSA injected) and 2 (SOPC, Subphase and Injected). For clarity, dark regions represent the pure (film-free) surface, while bright regions represent film-forming material. Prior to compression of the pure phospholipid film (Figure 5A, Π ∼ 0 mN/m), no film formation is evident, only a clean surface, as evidenced by the homogeneous dark region depicted in Figure 5A. Upon compression, a very homogeneous monolayer is eventually formed and further compressed (Figure 5B, Π ) 31.5 mN/m) until the film reaches the collapse pressure, as indicated by the “ripples” in the BAM image (Figure 5C, Π ) 40.5 mN/m). When only BSA is present in the subphase at a concentation of 50 nM, there is already a complete monolayer (26) Brandal, O.; Viitala, T.; Sjoblom, J. J. Dispersion Sci. Technol. 2007, 28, 95–106.

before compression is initiated (Figure 5D, Π ) 7.5 mN/m), confirming that the protein is partitioned at the surface. As the system is compressed, an increasingly thicker film is formed, as shown in Figure 5E (Π ) 21.5 mN/m) and Figure 5F (Π ) 25.5 mN/m). Thus, if a phospholipid is spread after the protein has been introduced to the system, the lipid is spread on top of a protein monolayer, forming a mixed protein-lipid film, as discussed above. Here, it should be noted that the BAM images were obtained for a system where BSA was injected into the subphase. The image shown in Figure 5D was captured after an equilibration time of 60 min. As the microscope is calibrated to a pure, presumedly film-free subphase, a reliable image of the completely relaxed system cannot be obtained if the protein is dissolved in the subphase. When spreading SOPC onto a subphase containing 50 nM BSA, there is already a complete, homogeneous monolayer formed before compression is initiated, as shown in Figure 6A (Π ) 15 mN/m). Compressing this monolayer leads to a progressively thicker film, as shown in Figure 6B (Π ) 20 mN/m) and Figure 6C (Π ) 25 mN/m). Comparison of the sequence in Figure 6A-C to the two sequences shown in Figure 5, it is apparent that compression of the Subphase system is more comparable to compression of BSA, which is consistent with the corresponding compression isotherms in Figures 1 and 2. Introduction of BSA to a precompressed (to Π ) 5 mN/m) SOPC film via the Injected protocol, a homogeneous phospholipid monolayer is already formed (Figure 6D, Π ) 6.5), and as the film is compressed, it becomes more dense (Figure 6E, Π ) 16.5), until it appears to collapse at Π ) 31.5 mN/m (Figure 6F). The images in Figure 6D-F are more comparable to the images obtained for the pure phospholipid film than for the pure protein film (Figure 5), which is consistent with the compression isotherms shown in Figure 2. Thus, using Brewster angle microscopy, we can observe that the film topography is very dependent on how protein is introduced to a phospholipid film. Historically, Langmuir monolayers were used to explore and characterize the phase behavior of surface active compounds upon an aqueous subphase. As knowledge accumulated and research fronts moved, increasingly more complex systems both with respect to composition and the constituent molecules were explored. For instance, focus has moved from investigating the effect of ionic strength and valency in the subphase,23,26-30 to relatively small organic molecules in pharmacological studies,30-35 to peptides and DNA/RNA, to large and complex biomolecules. Proteins and peptides and their interaction with lipid membranes (27) Ese, M. H.; Sjoblom, J. Colloid Surf., A: Physicochem. Eng. Asp. 1997, 123, 479–489. (28) Gundersen, S. A.; Ese, M. H.; Sjoblom, J. Colloid Surf., A: Physicochem. Eng. Asp. 2001, 182, 199–218. (29) Havre, T. E.; Ese, M. H.; Sjoblom, J.; Blokhus, A. M. Colloid Polym. Sci. 2002, 280, 647–652. (30) Agasoster, A. V.; Holmsen, H. Biophys. Chem. 2001, 91, 37–47. (31) Agasosler, A. V.; Tungodden, L. M.; Cejka, D.; Bakstad, E.; Sydnes, L. K.; Holmsen, H. Biochem. Pharmacol. 2001, 61, 817–825. (32) Broniec, A.; Gjerde, A. U.; Olmheim, A. B.; Holmsen, H. Langmuir 2007, 23, 694–699. (33) Hidalgo, A. A.; Caetano, W.; Tabak, M.; Oliveira, O. N. Biophys. Chem. 2004, 109, 85–104. (34) Jutila, A.; Soderlund, T.; Pakkanen, A. L.; Huttunen, M.; Kinnunen, P. K. J. Chem. Phys. Lipids 2001, 112, 151–163. (35) Oruch, R.; Hodneland, E.; Pryme, I. F.; Holmsen, H. Biochim. Biophys. Acta: Biomembranes 2008, 1778, 2165–2176.

have become especially popular studies. In contrast to smaller molecules, these molecules often have strong intrinsic affinities for interfaces of any kind, also the air-water interface. Thus, the at one time reasonable assumption that the subphase solvents did not have the ability to form films in their own right does not hold for many applications of this technique on protein-lipid systems. Specifically, the initial assumptions break down upon introduction of several film-forming species, of which at least one is soluble in the bulk. At this point, quantitative control of the surface species, one of the cornerstones of the Langmuir technique, is lost. Also, the bulk-to-film protein exchange kinetics must be taken into account on a system to system basis and also depend on protein introduction protocol. With the injection of a protein into the subphase, the time required for instilling equilibrium between surface and bulk-residing species will depend on protein molecular weight, hydrophobicity, and charge density, in addition to variables associated with the physicochemical environment. Carefully collected Langmuir data at different pHs, temperatures, and ionic strengths can be used to gain insight into the way a protein interacts with the lipids and penetrates the monolayer, especially if combined with other techniques.4,6 There has also been some interest in to what extent protein flexibility affects membrane- and surface affinity.22,36-38 Such flexibility will facilitate spreading along the plane of a surface, allowing the protein to maximize its interaction with the interface. This is one phenomenon among many that will affect Langmuir data. It should be emphasized, however, that it is challenging to separate the effect of spreading from other effects, such as intercalation and reorientation of nonspherical proteins. Still, considering the flexibility or softness of a protein is useful for selecting protocol and conditions. Although the best estimate of flexibility is assessment of dynamics by NMR or fitting of DSC data to models that give information about the availability of partially unfolded states at the experimental conditions,39,40 proteins with lower stability will in general be more flexible.22,41 As shown in this study, the obtained results are highly dependent on what protein-introduction protocol has been used. Great care must be taken to ensure that the conditions of the protocol are conducive to the interactions one wishes to study. For example, with the aim to use the Langmuir technique as a realistic model for interactions between proteins and (halfof) a cell membrane, the proteins must be introduced to a precompressed phospholipid film from the bulk phase in order to best imitate physiological conditions. Also, for proteins expected to readily spread along air-water surfaces, access to the interface must be minimized to ensure that the measurements are dominated by protein-monolayer interactions. The protocol named Injected fits these criteria, as it introduces the protein to a headgroup side of a precompressed phospholipid film, with the possibility of tailoring the state of the monolayer to physiological conditions (∼30 mN/m) or to any other desired packing. While (36) Engel, M. F. M.; Visser, A.; van Mierlo, C. P. M. Langmuir 2004, 20, 5530– 5538. (37) Galisteo, F.; Norde, W. Colloid Surf., B: Biointerfaces 1995, 4, 375–387. (38) Galisteo, F.; Norde, W. Colloid Surf., B: Biointerfaces 1995, 4, 389–400. (39) Andersson, A.; Maler, L. J. Biomol. NMR 2002, 24, 103–112. (40) Halskau, O.; Perez-Jimenez, R.; Ibarra-Molero, B.; Underhaug, J.; Munoz, V.; Martinez, A.; Sanchez-Ruiz, J. M. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 8625–8630. (41) Kamerzell, T. J.; Unruh, J. R.; Johnson, C. K.; Middaugh, C. R. Biochemistry 2006, 45, 15288–15300.

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

3049

the protocol named Top Π ) 5 succeeds in introducing the protein to a preformed film, the polypeptide is trapped at a hydrophobic interface, which deviates significantly from the use of Langmuir monolayers as a model for half of a cell membrane. It also allows the protein to spread along the air-water surface. On the other hand, if the goal is to study lateral interactions within a mixed phospholipid-protein film, protein should be introduced in such a way as to ensure the presence of the protein at the interface prior to compression of the nonsoluble film-forming species. Here, the protocols labeled Subphase and Top Π ) 0 fit this description, as both partition protein at the surface either prior to introduction of protein (in the case of Subphase) or while the phospholipid

3050

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

resides in a 2-D gaseous state (in the case of Top Π ) 0). Of these two protocols, the latter requires far less protein to fill vacant surface area at the air-water interface. ACKNOWLEDGMENT W.R.G. gratefully acknowledges support from the NFR/NTNU Grant 10287305, and Ø.H. Jr. acknowledges support from The Norwegian Cancer Society. Received for review December 23, 2008. Accepted March 5, 2009. AC8027257