ARTICLE pubs.acs.org/Langmuir
Self-Assembly of Hydrophobin and Hydrophobin/Surfactant Mixtures in Aqueous Solution Xiaoli L. Zhang,† Jeffrey Penfold,*,†,‡ Robert K. Thomas,† Ian M. Tucker,§ Jordan T. Petkov,§ Julian Bent,|| Andrew Cox,|| and I. Grillo^ †
Physical and Theoretical Chemistry Laboratory, Oxford University, South Parks Road, Oxford, U.K. STFC, Rutherford Appleton Laboratory, Chilton, Didcot, OXON, U.K. § Unilever Research and Development Laboratory, Port Sunlight, Quarry Road East, Bebington, Wirral, U.K. Unilever Research Laboratories, Sharnbrook, MK44 1LQ, Beds, U.K. ^ Institute Laue Langevin, 6 Rue Jules Horowitz, F-38042 Grenoble, Cedex 09, France
)
‡
bS Supporting Information ABSTRACT: The self-assembly of the protein hydrophobin, HFBII, and its self-assembly with cationic, anionic, and nonionic surfactants hexadecylterimethyl ammonium bromide, CTAB, sodium dodecyl sulfate, SDS, and hexaethylene monododecyl ether, C12E6, in aqueous solution have been studied by small-angle neutron scattering, SANS. HFBII selfassembles in solution as small globular aggregates, consistent with the formation of trimers or tetramers. Its self-assembly is not substantially affected by the pH or electrolytes. In the presence of CTAB, SDS, or C12E6, HFBII/surfactant complexes are formed. The structure of the HFBII/surfactant complexes has been identified using contrast variation and is in the form of HFBII molecules bound to the outer surface of globular surfactant micelles. The binding of HFBII decreases the surfactant micelle aggregation number for increasing HFBII concentration in solution, and the number of hydrophobin molecules bound/micelle increases.
’ INTRODUCTION Biosurfactants are of much current interest because of their properties relating to biosustainability and biodegradation.1,2 A wide range of different biosurfactants have been studied and include the glycolipids, such as the rhamnolipids, sophorolipids, and mannostylerythitol lipids,3 and lipopeptides and proteins,4 such as surfactin and hydrophobin. Their novel adsorption and self-assembly properties lead to many interesting potential applications, and yet many of these aspects are poorly characterized or understood. Furthermore, many potential applications will involve mixtures with a range of surfactants and other proteins, and both the surface adsorption and self-assembly properties of such mixtures have not been extensively explored. Hence in this article we report on the self-assembly of the protein hydrophobin and its self-assembly in combination with the anionic, cationic, and nonionic surfactants of sodium dodecyl sulfate, SDS, hexadecyltrimethyl ammonium bromide, CTAB, and hexaethylene monododecyl ether, C12E6. Hydrophobin is a small (∼7 to 10 kDa), highly surface active globular protein that is produced by filamentous fungi.5 Hydrophobin exists as two major types, HFBI and HFBII, where HFBI is the more hydrophobic species. This work is focused entirely on the more water-soluble HFBII, which has been produced here via r 2011 American Chemical Society
fermentation of a modified yeast. The crystal structure of HFBII has been established6 and is characterized by a conserved pattern of eight cysteine residues that make four intermolecular disulfide bridges. This results in a globular protein conformation with a central β-barrel structure and a small R-helix segment, which is very compact and robust.4 Its functionality and surface activity arise primarily from a hydrophobic patch consisting of side-chain residues of leucine, valine, and analine, which occupies ∼20% of the surface area of the protein, as shown in Figure 1. The biological function of hydrophobin is diverse and is strongly related to its strong surface activity.4 Hence, much of the attention to date has inevitably focused on investigating its surface adsorption properties7 and, in particular, self-assembly at interfaces.8,9 More recently, we have used neutron reflectivity, NR, to study the adsorption of hydrophobin and hydrophobin/ surfactant mixtures at the airwater interface.10 In the measurements on hydrophobin/surfactant mixtures, it was observed for a range of different surfactants that the surface adsorption was dominated by hydrophobin at surfactant concentrations below Received: May 31, 2011 Revised: July 20, 2011 Published: July 20, 2011 10514
dx.doi.org/10.1021/la2020226 | Langmuir 2011, 27, 10514–10522
Langmuir
ARTICLE
In this article, SANS has been used to study the self-assembly of the hydrophobin, HFBII, and the self-assembly of mixtures of HFBII with the cationic, anionic, and nonionic surfactants hexadecylterimethyl ammonium bromide, CTAB, sodium dodecyl sulfate, SDS, and hexaethylene monododecyl ether, C12E6.
’ EXPERIMENTAL DETAILS
Figure 1. Structure of the hydrophobin monomer, HFBII, showing the hydrophobic patch in gray (drawn using RasMol and the Protein Data Bank).
the surfactant cmc (critical micellar concentration) and by the surfactant for surfactant concentrations greater than the surfactant cmc. These observations point to the important role that the competition provided by solution self-assembly in hydrophobin/surfactant mixtures plays in determining the adsorption behavior. Although the self-assembly of hydrophobin has been studied,11,12 there are no reports on the self-assembly of hydrophobin/surfactant mixtures, and this is the focus of this article. Kisko et al.11 used small-angle X-ray scattering, SAXS, to study the solution selfassembly of both HFBI and HFBII. The predominant structure is in the form of hydrophobin tetramers, and the self-assembly is driven largely by the hydrophobic interaction. Hence, the tetramers are stable over wide temperature and pH ranges and are largely insensitive to added electrolyte. This structural picture was also confirmed by the SAXS and size exclusion chromatography measurements of Torkkeli et al.12 In contrast, the small lipopeptide surfactin self-assembles in solution to form slightly larger aggregates with aggregation numbers in the range of 15 to 20.13 The compact, robust structures of hydrophobin are in marked contrast to those formed by more unstructured, disordered proteins such as β-casein. This can result in a very different pattern of self-assembly, especially when mixed with surfactants. Moitzi et al.14 recently reported on the effect of temperature and pH on the self-assembly of β-casein by cryo-TEM and SAXS. Ellipsoidal structures that had aggregation numbers that varied between 4 and 40 and that strongly depended on the temperature and pH were reported. The ability of anionic and cationic surfactants to induce protein unfolding has been studied in BSA/surfactant mixtures1517 and in lysozyme/surfactant mixtures.16,18 Chodankar et al.15,17 used SANS to probe the protein unfolding of BSA induced by SDS and CTAB. Lysozyme/SDS and BSA/SDS self-assembly were probed by RuizPena et al.16 using conductivity measurements. Stenstam et al.19 have discussed in more detail the relative contributions of the hydrophobic and electrostatic interactions in lysozyme/SDS complex formation.
Small-Angle Neutron Scattering, SANS. The SANS measurements were made on a range of different diffractometers, on the LOQ diffractometer20 at the ISIS pulsed neutron source at the Rutherford Appleton Laboratory, and on the D11 and D22 diffractometers21 at the Institut Laue Langevin, Grenoble, France. The measurements on the LOQ were made using the white beam time-of-flight method in the wave vector transfer, Q range of 0.008 to 0.25 Å1, where the wave vector transfer, Q, is defined as Q = 4π sin θ/λ, 2θ is the scattering angle, and λ is the neutron wavelength. The measurements on D11 were made using a wavelength of 6 Å (Δλ/λ ≈ 10%) and two different detector/collimation distance combinations (2.5/5.5 m, 10.0/10.5 m) to cover the Q range of ∼0.003 to 0.27 Å1. The measurements on D22 were made using a wavelength of 8 Å (Δλ/λ ≈ 10%) and two different detector/ collimation distance combinations (3.0/5.5 m, 11.0/11.2 m) to cover the Q range of ∼0.004 to 0.25 Å1. The samples were contained in Starna 1 mm path length quartz spectrophotometer cells and maintained at a temperature of 30 C. On each of the diffractometers, typical measurement times were ∼10 to 30 min, and the nature of the HFBII protein is stable at 30 C and over these lapse times. The data were corrected for background scattering, detector response, and the spectral distribution of the incident neutron beam and converted to an absolute scattering cross-section (I(Q) in cm1) using standard procedures.22,23 In SANS, the scattering cross section, or scattering intensity, for colloidal aggregates in solution can be written as24 Z 2 3 ð1Þ IðQ Þ ¼ N ðFp ðrÞ Fs Þexp iQr d r v where Fp and Fs are the aggregate and solvent scattering length densities and N is the number of aggregates per unit volume. For the surfactant micelles, in the absence of HFBII, the micelle structure is determined by analyzing the scattering data using a standard well-established model for globular micelles.24 For a solution of globular polydisperse interacting particles (micelles), the scattering intensity can be written in the decoupling approximation24 as IðQ Þ ¼ n½SðQ ÞjÆFðQ ÞæQ j2 þ ÆjFðQ Þj2 æQ jÆFðQ ÞæQ j2
ð2Þ
where the averages denoted by ÆQæ are averages over particle size and orientation, n is the micelle number density, S(Q) is the structure factor, and F(Q) is the particle form factor. The micelle structure (form factor) is modeled using a standard coreshell model,24 where the form factor is FðQ Þ ¼ V1 ðF1 F2 Þ F0 ðQR1 Þ þ V2 ðF2 Fs Þ F0 ðQR2 Þ
ð3Þ
radii, Vi = 4πRi3/3, cos(QR)]/(QR)3, F1,
R1 and R2 are the core and shell F0(QRi) = F2, and Fs 3j1(QRi)/(QR) = 3[sin(QR) QR are the scattering-length densities of the micelle core and shell and the solvent, and j1(QRi) is a first-order spherical Bessel function. The micelle coreshell model24 comprises an inner core made up of the alkyl chains only and is constrained to space fill a volume limited by a radius, R1, which is the fully extended chain length of the surfactant, lc. For larger aggregation numbers, ν, volumes greater than that defined by R1 (as is found in this study) are accommodated by a prolate elliptical distortion with dimensions of R1, R1, and eR1 (where e is the elliptical ratio). The outer shell with dimensions of R2, R2, and eR2 contains the headgroups and the corresponding water of hydration. The interparticle 10515
dx.doi.org/10.1021/la2020226 |Langmuir 2011, 27, 10514–10522
Langmuir
ARTICLE
Table 1. Key Molecular Parameters Used in the Model Constraints molecular volume (Å3)
extended length (Å)
SDS alkyl chain
327
16.7
SDS headgroup CTAB alkyl chain
60 426
21.7
CTAB headgroup
142
C12E6 alkyl chain
326
C12E6 headgroup
348
HFBII
9190
D2O
30
component
16.7
interactions, S(Q), are included using the rescaled mean spherical approximation, RMSA, calculated for a repulsive screened Coulombic potential25,26 defined by the surface charge, z, the micelle number density, n, the micelle diameter, and the DebyeHuckel inverse screening length, k1 . Using the molecular constraints incorporated into the model and the known molecular volumes and neutron-scattering lengths, we find that the refinable model parameters are ν, z, and e. Correctly predicting the absolute scattering cross-section is also an important model constraint, and for the different contrasts measured, the model has to predict the absolute scattering to within an error in the absolute scaling of (10%. This same basic model has been used for the quantitative analysis of the scattering data for the HFBII/surfactant mixtures. In this case, the model that is consistent with the data is one with the HFBII molecules bound to the outside of the surfactant micelles. This is incorporated into the model with additional parameters, nHFBII and nhydr, that define the number of hydrophobin molecules bound to the outer shell of the micelles and the amount of associated solvent in that layer per surfactant molecule, with the outer dimension determined from simple space filling. The key molecular parameters used in the model constraints are summarized in Table 1. The scattering data from HFBII alone has been analyzed in two complementary ways. The data were analyzed using eq 2 for uniform polydisperse spheres, where the polydispersity is described as a Schultz distribution with a characteristic width σ, and assuming negligible interparticle interactions, S(Q) ≈ 1.0. From a Guinier analysis, a radius of gyration, Rg, was extracted for direct comparsion with the results of Kisko et al.11 In the Guinier approximation,27 the scattering intensity can be expressed as I(Q) ≈ NV2(Δp)2 exp(Rg2Q2/3) and the slope of ln I(Q) versus Q2 is related to Rg. Materials and Measurements Made. The deuterium-labeled surfactants (alkyl chain deuterium labeled) of SDS, CTAB, and C12E6 (abbreviated as d-SDS, d-CTAB, and d-C12E6) were synthesized using established synthesis routes.2830 Hydrogeneous SDS and CTAB (h-SDS and h-CTAB) were obtained from Sigma, and all of the ionic surfactants were recrystallized in ethanol/acetone mixtures before use. Hydrogeneous C12E6 (h-C12E6) was obtained from Nikkol and used as supplied. d-C12E6 was purified on a chromatography column before use. The purity of the surfactants was assessed by the surface tension and the requirement of the absence of a minimum in the surface tension at the cmc. HFBII was produced using an external fermentation company via a yeast fermentation route and was subsequently purified by a twophase extraction at Unilever Research, Vlaardingen, using procedures described elsewhere.31,32 The extracted and purified material was freeze dried before dissolution into the appropriate solvent for the SANS measurements. The SANS measurements were all made in D2O and in H2O (unless otherwise stated). High-purity water (Elga Ultrapure) was used, and D2O was obtained from Fluorochem. The pH was adjusted by the addition of HCl and NaOH. All glassware and the quartz sample cells
Figure 2. (a) Scattering intensity, I(Q), vs the scattering vector, Q, for (red) 1.0, (blue) 2.0, (green), 5.0 and (black) 10.0 mg/mL HFBII in D2O. The solid lines are model fits for polydisperse spheres with a radius of 20 ( 2 Å and a polydispersity, σ, of ∼0.35. (b) I(Q)/C vs Q for the data in Figure 1a showing the invariance in the form of scattering with the HFBII concentration. (c) Guinier plot, ln I(Q) vs Q2, for 5 mg/mL HFBII/D2O. The solid line is a model fit to Guinier’s law (see the text) for Rg ≈ 21 ( 1 Å. for the SANS measurements were cleaned using an alkali detergent (Decon 90) and rinsed thoroughly in high-purity water. All of the measurements were made at a fixed temperature of 30 C, which was 10516
dx.doi.org/10.1021/la2020226 |Langmuir 2011, 27, 10514–10522
Langmuir
ARTICLE
chosen to be in excess of the Krafft point of the ionic surfactants, 16 C for SDS, and 26 C for CTAB, which were well below the cloud point, 55 C, of the nonionic surfactant.33 The cmc values for the different surfactants are 8 103 M for SDS, 9 104 M for CTAB, and 9 105 M for C12E6.33 On the LOQ diffractometer, measurements were made for HFBII at concentrations of 1, 2, 5, and 10 mg/mL in D2O at pH 7 and at 5 mg/mL at pH 3, 7, and 10 in D2O and in 0.01 and 0.1 M NaCl. At an HFBII concentration of 5 mg/mL, measurements were also made with added CTAB at concentrations of 0.1, 1.0, 2.0, 5.0, and 10.0 mM for h-CTAB in D2O. On the D11 diffractometer, measurements were made for HFBII/ surfactant mixtures (CTAB, SDS, and C12E6) at a fixed surfactant concentration of 25 mM and HFBII concentrations of 2, 5, 10, and 20 mg/mL. The measurements were made for three different isotopic combinations: HFBII/h-surfactant/D2O, HFBII/d-surfactant/D2O, and HFBII/d-surfactant/H2O.
’ RESULTS AND DISCUSSION HFBII Self-Assembly. SANS measurements were made for HFBII in D2O at solution concentrations of 1, 2, 5, and 10 mg/ mL, and the data are shown in Figure 2a. In Figure 2a, the scattering for Q > 0.01 Å1 is consistent with small globular, noninteracting aggregates that scale with the solution concentration. In Figure 2b, the data, scaled by the solution concentration, are plotted and show an invariance in the scattering form factor with concentration. The solid lines in Figure 2a are model calculations for uniform polydisperse spheres with a mean radius of 20 ( 2 Å and a polydispersity of ∼0.35. This simple model provides a good description of the data and also confirms that at these relatively low concentrations there are no significant interaggregate interactions. From space-filling arguments, the mean size is consistent with the formation of trimers/tetramers. In Figure 2c, the scattering data for 5 mg/mL HFBII/D2O are plotted as ln I(Q) vs Q2, and the solid line is a fit to the Guinier expression, from which a radius of gyration, Rg, is obtained. The value of Rg obtained, 21 ( 1 Å, is similar to that reported by Kisko et al.11 for HFBII from SAXS measurements. They performed a more detailed interpretation of Rg on the basis of the known crystal structure and showed that the scattering was consistent with the formation of tetramers.11 The Rg value of 20 Å implies a sphere radius of
