Langmuir 2007, 23, 7995-8002
7995
Surface Properties of Class II Hydrophobins from Trichoderma reesei and Influence on Bubble Stability Andrew R. Cox,* Florence Cagnol, Andrew B. Russell, and Martin J. Izzard UnileVer R&D Colworth, Colworth Park, Sharnbrook, Bedfordshire MK44 1LQ, United Kingdom ReceiVed February 15, 2007. In Final Form: April 19, 2007 We report the remarkable surface behavior of class II hydrophobin proteins HFBI and HFBII from Trichoderma reesei and the resulting effect that these proteins have on the stability of air bubbles to the process of disproportionation. The surface properties were studied using surface tensiometry and surface shear rheology. Surface tensiometry data show that hydrophobins are very surface active proteins, reducing the surface tension to approximately 30 mN m-1. The rate at which the hydrophobins adsorb at the surface may also be related to the self-assembly behavior in aqueous solution. We further show that hydrophobins form air/water surfaces with high elasticity, the magnitude of which is well in excess of that of surface layers formed by other common proteins used as foam or emulsion stabilizers. The measured surface properties translate to the stability of bubbles with adsorbed hydrophobin, and in this study, we demonstrate the ability of hydrophobin to have a dramatic effect on the rate of disproportionation in some simple bubble dissolution studies.
Introduction Hydrophobins are a family of surface-active proteins produced naturally by filamentous fungi. They are small proteins (7-9 kDa) and are characterized by the presence of eight cysteine residues forming four disulfide linkages. Hydrophobins are remarkable for their high surface activity and the fact that they self-assemble at hydrophobic-hydrophilic surfaces or interfaces to form a robust, amphipathic layer. This behavior is important for their role in nature, which is to assist fungi in the growth of aerial structures.1 Two classes of hydrophobins have been distinguished on the basis of aqueous solubilities and hydropathy profiles:2 Class I hydrophobins (e.g., SC3 from Schizophyllan commune) form highly insoluble aggregates which can only be dissolved with strong acids.3 Class II hydrophobins (e.g., HFBI and HFBII from Trichoderma reesei) are more readily solubilized and can be dissolved in aqueous solution to concentrations of at least 100 mg mL-1. Current knowledge on hydrophobins and their functionality is comprehensively reviewed by Linder et al.4 Utilization of the properties of hydrophobins in applications such as coatings, emulsion stabilization, and separation technologies has been discussed.5 Lumsdon et al. have also demonstrated the ability of some hydrophobins to improve the dispersibility of hydrophobic particles in water and to stabilize a simple oil in water (o/w) emulsion.6 In terms of the aeration behavior of solutions containing hydrophobin, however, no study has been reported demonstrating the use of this class of proteins for the stabilization of air/water (a/w) interfaces to create stable bubble dispersions and foams. In this current work, some of the physical properties of air/water interfaces stabilized by the class * To whom correspondence should be addressed. E-mail: andrew.cox@ unilever.com. (1) Wessels, J. G. H. AdV. Microb. Physiol. 1997, 38, 1-45. (2) Wessels, J. G. H. Annu. ReV. Phytopathol. 1994, 32, 413-437. (3) de Vocht, M. L.; Scholtmeijer, K.; van der Vegte, E. W.; de Vries, O. M. H.; Sonveaux, N.; Wosten, H. A. B.; Ruysschaert, J. M.; Hadziioannou, G.; Wessels, J. G. H.; Robillard, G. T. Biophys. J. 1998, 74, 2059-2068. (4) Linder, M. B.; Szilvay, G. R.; Nakari-Setala, T.; Penttila, M. E. FEMS Microbiol. ReV. 2005, 29, 877-896. (5) Hektor, H. J.; Scholtmeijer, K. Curr. Opin. Biotechnol. 2005, 16, 434439. (6) Lumsdon, S. O.; Green, J.; Stieglitz, B. Colloids Surf., B 2005, 44, 172178.
II hydrophobins HFBI and HFBII were investigated with a view to determining the potential for use of hydrophobin protein as an effective aerating agent. The formation and stability of colloidal dispersions, including emulsions and foams, are related to the processes used to create the dispersion and the physicochemical properties of the surfaceactive agents utilized. There have been some studies regarding the physical properties of both class I and class II hydrophobins. Notably, for the case of class II hydrophobins, the surfaceadsorbed film formed by each of the hydrophobins HFBI and HFBII at an a/w surface have been shown by atomic force microscopy (AFM) studies to have highly ordered structures, with the protein molecules packed in a crystalline layer.7 Lumsdon et al.6 measured the a/w surface tension (γ) of such a HFBII film and found an equilibrium value of 45 mN m-1 at the surprisingly low protein solution concentration of about 1 µM. Askolin et al.8 have measured the a/w surface tension of both HFBI and HFBII, obtaining values of around 40 and 35 mN m-1, respectively, for concentrations between about 3 and 30 µM. Although more quantitative analysis is required to properly understand the adsorption and surface behavior of both HFBI and HFBII, these observations certainly demonstrate the powerful surface activity of these hydrophobins in terms of reducing surface tension. However, such properties do not necessarily indicate the ability of either HFBI or HFBII to provide long-term stability to a surface or interface. It is well-known in the literature that the effectiveness of polymeric molecules, such as proteins, at stabilizing foams against disproportionation is related to the surface rheology of the layer adsorbed to the surface of the bubbles.9,10 Specifically, it is the elastic modulus of the interface which will determine the ability of an adsorbed layer to retard bubble disproportionation,11 (7) Paananen, A.; Vuorimaa, E.; Torkkeli, M.; Penttila, M.; Kauranen, M.; Ikkala, I.; Lemmetyinen, H.; Serimaa, R.; Linder, M. B. Biochemistry 2003, 42, 5253-5258. (8) Askolin, S.; Linder, M.; Scholtmeijer, K.; Tenkanen, M.; Penttila, M.; de Vocht, M. L.; Wosten, H. A. B. Biomacromolecules 2006, 7, 1295-1301. (9) Murray, B. S.; Ettelaie, R. Curr. Opin. Colloid Interface Sci. 2004, 9, 314-320. (10) Wilde, P. J. Curr. Opin. Colloid Interface Sci. 2000, 5, 176-181. (11) Kloek, W.; van Vliet, T.; Meinders, M. J. Colloid Interface Sci. 2001, 237, 158-166.
10.1021/la700451g CCC: $37.00 © 2007 American Chemical Society Published on Web 06/20/2007
7996 Langmuir, Vol. 23, No. 15, 2007
with increased stability being seen at greater moduli. Determination of the surface rheology of the adsorbed layer formed by HFBI and HFBII was therefore a key aspect of this current study, since information with respect to both the surface tension and surface rheology will provide some good indicative information about the potential of hydrophobins to act as foam stabilizers. In addition, the ability of HFBII, at low concentrations (less than 1 mM), to form very stable bubbles in comparison with that of other proteins is demonstrated, and the observed stability is related to the adsorption and surface properties of this unique protein. Experimental Section Materials. Class II hydrophobins (HFBI and HFBII) were obtained from VTT Biotechnology (Espoo, Finland), having been prepared as previously described.12,13 Briefly, both hydrophobins were made by fermentation of a T. reesei culture, extracted, and subseqently purified. HFBI was stored in powder form, and HFBII was stored, frozen, in aqueous solution. Both HFBI and HFBII were used as received without further purification. The molar masses of HFBI and HFBII are 7.5 and 7.2 kg mol-1, respectively. R1-Casein (>70%) and κ-casein (>90%) were obtained from Fluka, U.K. β-Casein (lyophilized, 90%), β-lactoglobulin (A and B, 90%), sodium dodecyl sulfate (SDS; 98.5%), hydrochloric acid (98%), and sodium hydroxide (98%) were obtained from Sigma-Aldrich, U.K. All of the bovine milk proteins were stored frozen and used as received without further purification. Citric acid (99.5%) was obtained from BDH. Solutions were made using distilled water (Elga Water purifier). Solution Preparation. Due to the surface-active nature of HFBI and HFBII, some care was required with respect to solution preparation. Filamentous hydrophobin aggregates were noticed to form when solutions were subjected to shear, e.g., mixing. This has been observed previously,14,15 and we attribute this to (i) a “skin” of hydrophobin forming at the air/water surface and then breaking up in the shear field, but not redissolving quickly, and (ii) incorporation of small bubbles if the shear was great enough to form a vortex and introduce air into the aqueous phase. Subjecting the solution to brief sonication (10-20 s in an ultrasound bath) immediately before any subsequent measurement led to rapid redissolution of the observed macroscopic aggregates. Electrophoretic Mobility and Simultaneous Particle Size. The electrophoretic mobility, ue, and particle size of HFBII in solution at 20 °C were measured using a Malvern Zetasizer (Malvern Instruments, U.K.), a technique that uses laser Doppler velocimetry and dynamic light scattering to determine these two parameters, respectively. A 0.1 wt % solution (ca. 140 µM) of HFBII was prepared and then filtered using a single-use Sartorius filter (0.2 µm pore size). Measurements were taken across a range of solution conditions from pH 9 to pH 3 and then back to pH 9. The instrument adjusted the pH by automatic titration (using hydrochloric acid or sodium hydroxide). Solutions were subject to mixing before measurement. Equilibrium Surface Tensiometry. The surface saturation concentration was determined using a Kruss K100 tensiometer with a critical micelle concentration (cmc) add on (Kru¨ss, Germany). The surface tension (γ) was effectively measured as a function of the HFBI and HFBII concentration from 0 to approximately 0.1 M protein. The tensiometer doses a predefined quantity of hydrophobin solution into the measurement dish and measures the surface tension until a predetermined termination point is reached, which is set by the user through the software. The solution is then gently stirred, a further quantity of hydrophobin solution from the stock then dosed into the solution, and the process repeated until the final hydrophobin concentration is reached. Preferably, one sets the termination point (12) Linder, M. B.; Selber, K.; Nakari-Setala, T.; Qiao, M. Q.; Kula, M.-R.; Penttila, M. Biomacromolecules 2001, 2, 511-517. (13) Bailey, M. J.; Askolin, S.; Horhammer, N.; Tenkanen, M.; Linder, M.; Penttila, M.; Nakari-Setala, T. Appl. Microbiol. Biotechnol. 2002, 58, 721-727. (14) Torkkeli, M.; Serimaa, R.; Ikkala, I.; Linder, M. Biophys. J. 2002, 83, 2240-2247. (15) Wang, X. Q.; Graveland-Bikker, J. F.; De Kruif, C. G.; Robillard, G. T. Protein Sci. 2004, 13, 810-821.
Cox et al. to be the point at which the equilibrium surface tension is reached. We found, however, that, at very low concentrations in particular, equilibrium could take a number of hours to reach, as can be typical for proteins. Therefore, we set the termination point at each concentration to when either (i) the measurement time has passed 1200 s or (ii) the standard deviation of the surface tension of 10 subsequent measurements is less than 0.03. We found that this method determines the equilibrium surface tension well for high concentrations but may slightly overestimate the value at very low concentrations. However, quantification of the surface saturation point was not affected. Surface Shear Rheology. The surface shear viscous modulus (Gs′′) and elastic modulus (Gs′) at the a/w surface of solutions containing protein was measured using an AR-G2 rheometer (TA Instruments, U.K.). A Pt-Ir Du Nou¨y ring setup was used (13 mm diameter, Kru¨ss, Germany, flamed prior to each experiment) with a glass dish of 60 mm diameter. All measurements were made in oscillatory mode with the instrument bearing mode set at “soft”. The instrument was mapped prior to all experiments. Measurement of Gs′ and Gs′′ as a function of time at a constant oscillation frequency (1 Hz), oscillation amplitude (5 × 10-3 rad), and temperature (20 °C) was made as a function of both concentration and solution pH. Citric acid or sodium hydroxide was used to decrease or increase the pH, respectively. All solutions were sonicated in an ultrasound bath for 20 s immediately before the measurements were made. Bubble Stability and Optical Imaging of Bubbles. For the disproportionation studies of single bubbles with atmospheric air, 500 µL of protein solution of known concentration was pipetted into a well of a 76 × 26 × 1.35 mm glass microscope slide. The well was 15-18 mm in diameter with a depth between 0.6 and 0.8 mm. Single bubbles were then carefully blown using a Hamilton syringe (750 RN) equipped with a very fine needle. No coverslip was used to prevent the possibility of bubble compression or, in the case of hydrophobin, to prevent bubbles adhering to the glass. Bubbles were visualized using a Leica DMR microscope equipped with differential interference contrast illumination. The microscope was attached to a JVC KY-F75U 3CCD camera and connected to a computer through a FireWire connection. Images were taken using KY-Link software (Optivision, Leeds, U.K.). Bubbles were tracked during ripening experiments by manually moving the field of view, and bubble size analysis was carried out by reference to a stage micrometer. For non-quantitative analysis of hydrophobin-stabilized bubbles, intact bubbles were allowed to partially dehydrate over a period of between 5 and 10 min before optical micrographs were taken. Groups of bubbles were formed by blowing air into a hydrophobin solution using a 3.5 mL volume disposable plastic Pasteur pipet (VWR International). For the long-term stability experiments (over a period of several days) of bubble dispersions, clusters of bubbles were blown into a hydrophobin solution Hamilton syringe (750 RN) which was enclosed between two welled glass microscope slides (Marienfeld GmbH & Co., Germany, ref 13 200 02) to eliminate moisture loss. Long working distance 20× and 40× objectives were employed.
Results and Discussion Electrophoretic Mobility and Simultaneous Particle Sizing. The electrophoretic mobility of HFBII in solution as a function of pH is shown in Figure 1. We calculate the isoelectric point of this protein to be at pH 4.8, and in neutral solution the charge is negative. The data were obtained starting from pH 9 and reducing this to pH 3. Subsequently, when the pH was increased to 9 again, little hysteresis was observed. The particle size was simultaneously measured and is plotted in the same figure. Close to the isoelectric point, there is a significant increase in particle diameter, which we believe to be due to charge neutralization, leading to subsequent molecular aggregation of hydrophobin in solution. However, even at a pH well above or below the isoelectric point, the measured particle size is significantly greater
Class 2 Hydrophobins from Trichoderma reesei
Figure 1. Electrophoretic mobility, ue, and particle size as a function of pH for a solution of 0.1 wt % HFBII. Filled circles represent measured ue from pH 9 to pH 3, and empty circles represent measured ue from pH 3 to pH 9. Times signs represent the particle diameter.
Figure 2. Equilibrium surface tension (γ) as a function of hydrophobin concentration: HFBI (filled circles), HFBII (empty circles).
than the individual molecular size and also the sizes of the proposed tetramer aggregates, as reported previously by Torkelli et al.14 The sizes of the aggregates measured in this case are, in fact, partly a result of the mixing (i.e., shear) that takes place prior to measurement. The mixing process leads to the formation of larger solution aggregates (and potentially small bubbles), many of which are optically visible. Such formation of larger solution species has been noted previously for both class I and class II hydrophobin proteins.13,14,16,17 Since the size distribution of these aggregates can be broad and also a function of the magnitude and form of shear, we therefore report electrophoretic mobility data and refrain from reporting values of the ζ potential, which can be determined through use of the Henry equation, requiring an input of particle size. Nevertheless, these data provide useful information with respect to the solution behavior in terms of molecular charge, potentially having a bearing on the adsorption behavior on surfaces or other colloidal particles. Reducing or, as far as possible, eliminating mixing in the experimental process would be a preferable route to determine the ζ potential in subsequent studies of this nature. Equilibrium Surface Tension. The change in the equilibrium surface tension as a function of the hydrophobin concentration, for both HFBI and HFBII (at pH 7), is shown in Figure 2. The curves for both hydrophobin protein samples are similar, suggesting that although there are some structural differences in (16) Wosten, H. A. B.; Devries, O. M. H.; Wessels, J. G. H. Plant Cell 1993, 5, 1567-1574. (17) Stroud, P. A.; Goodwin, J. S.; Butko, P.; Cannon, G. C.; McCormick, C. L. Biomacromolecules 2003, 4, 956-967.
Langmuir, Vol. 23, No. 15, 2007 7997
these two hydrophobin molecules as reported by Panaanen et al.7 and Ritva et al.,18 the resulting surface activity is broadly comparable. Such data are typically used to calculate a “critical micelle concentration”, or cmc, for the surface-active agent, i.e., the concentration at which the surface becomes saturated, and any additional surfactant will then lead to the formation of micelles in solution. In Figure 3 we show data from three repeats of the same measurement for HFBII. Although these show that many aspects of the data are quite reproducible, some points for discussion need to be made. To obtain a measure of the cmc, one looks for a clear break where the equilibrium surface tension becomes constant at a particular concentration of the surfaceactive component. This is not necessarily immediately apparent from the data presented in this figure. We emphasize three areas of interest in one of the plots (a-c). First, as the concentration of HFBII is increased, the equilibrium surface tension decreases (a) as one would predict. A break is reached (b) where the gradient of the curve changes, although does not necessarily become constant. We take this as the surface saturation concentration (ssc) for HFBII, for reasons which we will discuss later, and we determine this value to be 0.38 ( 0.09 µM, or 2.7 × 10-4 ( 6.9 × 10-5 wt %, where the error is based on the standard deviation from the mean value. Currently there are relatively few reproduced data reporting clear values of the ssc of hydrophobins in general. However, the values we determine for HFBII appear to be somewhat lower that those calculated from data reported by Askolin et al.8 The value we report, however, compares with an ssc of 0.2-0.3 µM for a class I hydrophobin, Sc3p, reported by Martin et al.19 One may envisage a lower value of the ssc for Sc3p, since class I hydrophobins are more hydrophobic than class II hydrophobins. We describe point b as a surface saturation concentration as opposed to a critical micelle concentration since we believe that small aggregates of HFBI or HFBII are likely to be present in solution before the surface is saturated with hydrophobin molecules. Previous reports have discussed the nature of small solution aggregates of HFBI and HFBII, particularly at relatively high solution concentrations of hydrophobin. Notably, Torkelli et al.14 have described the formation of tetramers of HFBII in solution above a concentration of 10 mg/mL (ca. 3 mM), whereas below this concentration HFBII is present as monomers or dimers. These concentrations are well above the scc that we determine by surface tensiometry. However, data from Szilvay et al.20 using Fo¨rster resonance energy transfer (FRET) to study molecular interactions indicate that aggregates may be present at much lower concentrations (
