Characterization of Submicron Bubbles Formed by ... - ACS Publications

Apr 29, 2019 - †School of Materials Science and Engineering, ‡School of Chemical and Biomolecular Engineering, ... E-mail: paul.russo@mse.gatech.e...
0 downloads 0 Views 4MB Size
Article Cite This: J. Phys. Chem. B XXXX, XXX, XXX−XXX

pubs.acs.org/JPCB

Characterization of Submicron Bubbles Formed by the Hydrophobin Cerato-ulmin Andrew Gorman,†,∥ Xujun Zhang,†,∥ Bailey Risteen,‡,∥ Christopher J. Tassone,⊥ and Paul S. Russo*,†,∥,§ †

School of Materials Science and Engineering, ‡School of Chemical and Biomolecular Engineering, §School of Chemistry and Biochemistry, ∥Georgia Tech Polymer Network, GTPN, Georgia Institute of Technology, Atlanta, Georgia 30332, United States ⊥ Stanford Synchrotron Radiation Laboratory, Stanford Linear Accelerator Center, Stanford, California 94025, United States

J. Phys. Chem. B Downloaded from pubs.acs.org by UNIV OF SUSSEX on 05/06/19. For personal use only.

S Supporting Information *

ABSTRACT: Cerato-ulmin is a fungal hydrophobin protein with a high surface activity due to its amphipathic nature. When aqueous dispersions are gently agitated by hand, cerato-ulmin (CU) assembles into cylindrical bubbles visible in an optical microscope. After approximately 1 h the larger micron-sized bubbles rise out of the solution, leaving only submicron particulates, which persist indefinitely. Dynamic light scattering experiments show that these persistent particles shrink when positive air pressure is applied to the suspension and expand when vacuum is applied. Small-angle X-ray scattering at ambient pressure suggests an extended core−shell structure, consistent with small air-filled bubbles stabilized by a protein film. A comparison between the SAXS of the persistent submicron bubbles and AFM of the buoyant larger bubbles found immediately after agitation show that both have similar film thickness of 13−15 nm or five protein molecules. The extended shapes confirm the solid-like properties of these CU membranes, even in submicron particulate structures, consistent with microtensiometry results on interfacial CU membranes.



INTRODUCTION Hydrophobins are fungal proteins with low molecular weight (∼10 kDa) known for their amphipathic properties and ability to self-assemble at hydrophobic−hydrophilic interfaces.1−3 Wessels coined the term after investigating the basidiomycete species Schizophyllum commune and discovering cysteine-rich proteins with a high amount of hydrophobicity.4 These surface-active proteins can reduce surface tensions in aqueous solutions from 72.8 mN/m to values of 25−45 mN/m depending on concentration and type of hydrophobin.5,6 Hydrophobins are divided into two categories: class I and class II. Both classes have similar hydropathy patterns and contain eight conserved cysteine residues that provide four disulfide bonds. All hydrophobins self-assemble into amphipathic films at a hydrophobic/hydrophilic interface. Most proteins in water fold in a manner to create a core of hydrophobic residues with the hydrophilic residues on the surface, but Hakanpaa7 found the class II hydrophobin HFBII to be globular with a central β barrel structure having two loop regions of aliphatic side chains. These chains create a “hydrophobic patch” that aligns at the interface, resulting in a natural Janus particle containing a hydrophobic and a hydrophilic surface.8 The four disulfide bonds provide the essential stability. They are not involved in assembly but rather maintain the protein’s folded structure and the monomer’s solubility in aqueous solutions.9 Similar work © XXXX American Chemical Society

on the structure of class I hydrophobins yielded the same explanation for the existence of the hydrophobic patch.10 Hydrophobins reverse the polarity of surfaces.11,12 In nature, they have a variety of functions such as aiding the growth of aerial hyphae and fruit bodies,4,9 attaching to hydrophobic surfaces,13 and providing gas channels in hyphae.14 Their surface-active properties make hydrophobins useful for a variety of commercial applications: stabilization of emulsions in the food industry,15−17 functionalizing surfaces,18 aiding in drug delivery,19 increasing alignment of semiconducting polymers,20 and modifying electrode surfaces.21 Class I hydrophobins assemble into an amphipathic film composed of 5−12 nm amyloid “rodlets”, small cylindrical structures that are not to be confused with the cylindrical bubbles studied in this work. Class I hydrophobins have a larger hydrophobic patch than their class II brethren, which may be responsible for their ability to form rodlets or increased insolubility.22 These films are highly insoluble and only disassociated by trifluoroacetic acid (TFA) or formic acid.4,23 The class II hydrophobins assemble into films at the air/water interface but are more soluble than class I hydrophobins. Received: February 20, 2019 Revised: April 8, 2019

A

DOI: 10.1021/acs.jpcb.9b01673 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

was observed with a Leica DMR optical microscope equipped with a Canon EOS6D digital camera (5472 × 3648 pixels). Light Scattering. The DLS experiments were conducted on a custom-built apparatus with a rotating arm detector at scattering angles between 30 and 120° and a Coherent OBIS 660 laser operating at wavelength λo = 660 nm. The system features an ocular that enables its user to view precisely what the photomultiplier tube will measure; thus, the likelihood of dust or other unwanted particulate matter in the small detected volume can be assessed in meaningful terms. The ability to test for particulates is especially important in the present case because the samples can neither be centrifuged (the bubbles would rise) or filtered (the strongly surface-active protein would bind). Light intensity was measured with a Pacific Precision Instruments (Irvine, California) wide-range photometer/preamplifier/discriminator, which drives an ALV-5000 digital autocorrelator. The ALV software contains a suite of programs to analyze experimental data with cumulant analysis, multiexponential fitting, and inverse Laplace transform functions. Samples were connected with tubing to a Harvard apparatus PHD 2000 syringe pump to control the pressure. An Extech 407910 manometer measured a relative pressure with respect to the surrounding atmosphere. Measurements were conducted at scattering angles between 30 and 120° at 15° increments. Samples for DLS were prepared by mixing the stock solution with Type 1 water (see above) directly in clean, dust-free glass vials and gently agitated by hand through a rocking motion to create the bubble dispersion. Small Angle X-ray Scattering. SAXS experiments were carried out at the Stanford Synchrotron Radiation Light source (SSRL), beamline 1−5. The beam is highly collimated with a spot size of 300 μm × 300 μm and an approximate flux of 1011 photons/s. The sample-to-detector distance was 869 mm and a Rayonix 165 CCD camera was used as the detector. The incident energy was 15 keV. Samples were prepared by gently hand agitating a solution of cerato-ulmin in a glass vial then injecting it into a custom-built flow cell with Kapton films separated by 1 mm. Measurements were taken 1 h after insertion to allow for the larger bubbles seen in Figure 1 to rise out of the exposure volume. The experiment was completed at ambient pressure. The spectrum is the summation of 20 runs, each of 60 s exposure. For transmission correction and background subtraction, Nanopure water was exposed in the same manner as the sample. Samples and the solvent were centered in the beam by seeking the minimum transmittance. Data reduction was performed using the Nika package for Igor Pro (version 6.37); fitting was done using SasView and the Irena package for Igor Pro.38 Atomic Force Microscopy. An ICON Dimension scanning probe microscope (Bruker) operating in tapping mode with a silicon tip (RTESP, Bruker) was used to analyze the surface morphology of the collapsed bubbles. An agitated solution of 2 mg/mL cerato-ulmin was drop-cast on a glass slide and allowed to dry overnight.

Compared to Class I, the class II hydrophobins exhibit less variation in amino acid sequence length between the conserved cysteine residues.24 Some class II hydrophobins such as HFBI assemble into a honeycomb structure at the air/water interface.25 Rather than requiring a strong acid, class II hydrophobins such as HFBI and HFBII are dissolved by ethanol or SDS.26,27 This makes them useful in applications requiring reversible processes or easily removable coatings.28 Class II hydrophobins act as stabilizers for bubbles and foams and retain their self-assembly characteristics and ability to be redispersed after bubble disintegration. Cerato-ulmin is a class II hydrophobin produced by the fungi Ophiostoma novo-ulmi and Ophiostoma ulmi, which have been implicated in Dutch elm disease.29 Cerato-ulmin contains 73 amino acid residues and a molar mass of 7626 g·mol−1. Similar to other class II hydrophobins, CU is highly conserved regarding the spacing of the eight cysteine amino acids. The secondary structure contains 15% α-helix and 50% β-sheet with all cysteine residues creating four disulfide bonds.30 It is a highly surface-active protein and lowers the surface tension of water perceptibly even at concentrations of 0.03 μg/mL.27 At a higher concentration of 50 μg/mL, cerato-ulmin decreases surface tension from 72.8 to 25 mN/m. When a suspension of CU in pure water is agitated gently by hand, the protein assembles into fibrils or rods on the microscopic scale. Long ago, Miller and co-workers showed that these “fibrils” or “rods” are actually bubbles by observing their expansion, contraction, and wrinkling when exposed to pressure changes.31 While confirming these results and extending them to a range of salt and pH conditions, Zhang et al.32 used dynamic light scattering (DLS) to show that submicron structures persist in suspension, even after the large and usually cylindrical bubbles formed immediately after hand agitation have risen to the surface. The evidence presented in this paper strongly suggests that these persistent submicron particulates are bubbles with a film thickness comparable to that of the larger bubbles and that they share a similar extended shape. Thus, the solid-like character of CU protein membranes33−35 is extended to a new size regime. The strength of CU films suggests its utility for a variety of encapsulation applications.



METHODS

Cerato-ulmin. As described elsewhere32 the cerato-ulmin was a gift from Wayne Richards of the Canadian Forest Service. It was produced by a strain of C. ulmi (RDT2) and purified by the methods of Takai and Richards36 and Stevenson et al.37 Stock solutions were prepared in scintillation vials precleaned with Type I water from Barnstead Nanopure or Millipore Milli-Q purification systems. To ascertain the absence of particulates, the water itself can be viewed optically and tested by photometric measurement in the light scattering instrument (see below). The purified CU is added to these precleaned vials, along with the requisite amount of clean water. Cerato-ulmin has 73 amino acid residues with a molecular weight of 7626 g·mol−1. The sample integrity was verified through MALDI (Supporting Information). Investigators who wish to request samples of CU are encouraged to write the authors. Optical Microscopy. A sample of 0.1 mg/mL cerato-ulmin was gently agitated by hand through a rocking motion in a rectangular glass vial supplied by Vitrocom Inc. The sample



RESULTS AND DISCUSSION Dynamic Light Scattering. When agitated by hand, aqueous dispersions of CU assemble into “rod” or “fibril” bubbles on the micron scale (Figure 1).32 Such large bubbles cause the solution to appear cloudy. After 1 h the large bubbles rose out of the solution as the solution changed from cloudy to clear until there were no longer large bubbles in the scattering volume, as seen by DLS B

DOI: 10.1021/acs.jpcb.9b01673 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

normalized, baseline-subtracted correlation function, g2(t) = 1 + βe−2γt, where β (0 < β < 1) is an instrumental factor, γ = q2D is the electric field autocorrelation function decay rate, q is the scattering vector magnitude, and D is the diffusion coefficient. kT The hydrodynamic radius is R h = 6bπη where kb is the Boltzmann constant, T is the Kelvin temperature, and η is the viscosity (Figure 3). Accordingly, autocorrelation functions

Figure 1. Cerato-ulmin bubbles created through gentle hand agitation of solution with a rocking motion (scale bar 100 μm).

results in Figure 2. Measurements were erratic (e.g., ±50%) until about 1.5 h, but thereafter reasonably precise (±10%) Figure 3. Normalized, baseline-subtracted DLS correlation functions measured at a scattering angle of 90° for cerato-ulmin solution (c = 0.1 mg/mL) agitated by hand at different pressures. Inset: Semilog plot of normalized, baseline-subtracted DLS correlation function.

that begin to decay at lower lag times will have higher diffusion coefficients and lower hydrodynamic radii. The submicron structures were smallest when overpressure (1.2 bar) was applied during agitation, aging, and measurement. In contrast, when the solutions were prepared, aged, and measured under mild vacuum (0.8 bar) the correlation functions decayed more slowly and exhibited distinct nonexponential character. The slow secondary decay at long lag times potentially represents even larger structures, but the apparent sizes associated with the slower mode are sufficiently high that artifacts such as number fluctuations may interfere.39 Two-exponential fitting with a slow and fast decay mode yielded satisfactory agreement with the experimental data, but for the reason just mentioned the slow-mode decay is not considered further. One might also expect a very fast decay for the individual free proteins in solution, but it does not appear in the experimental data. Because scattering intensity scales with the sixth power of size, the intensity contribution from the bubbles is so large comparatively that the intensity contribution from the free proteins is negligible. The number of free proteins in the solution is not high enough to offset the scattering intensity discrepancy between small and large scatterers. Figure 4 shows a summary of the pressure behavior of the fast decay mode cerato-ulmin structures. Each point is the zero-angle approximation of exponential fits to correlation functions for the seven scattering angles measured at each pressure; the uncertainty bars represent the standard error from extrapolating to a zero-angle measurement. When structures were created, held, and measured at a partial vacuum, the size distribution was much larger; apparently, the decreased pressures allowed for a broader range of sizes to

Figure 2. Hydrodynamic radius from DLS decreasing quickly after agitation ceases. Estimated uncertainties: ±50% until 1.5 h, ±10% thereafter.

measurements suggested particle radii of ∼200 nm. These measurements were performed at a single scattering angle, θ = 90°, corresponding to qRh = 4 at long times, where q is the scattering vector magnitude and Rh,app is the apparent hydrodynamic radius. Therefore, the radii are regarded as apparent values. The long-term stability of the scattering structures is a subject of additional study using multiple-angle DLS, but it suffices to state that submicron structures remain suspended for at least several days. To determine whether the structures are bubbles or solid particles, solutions were agitated by hand while held at various applied pressures. The solutions were held at those pressures for 1 h prior to and during DLS measurement. The difference in the resulting structures can be appreciated even from the C

DOI: 10.1021/acs.jpcb.9b01673 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

Figure 4. Hydrodynamic radius versus pressure applied during agitation, a 1 h hold and DLS measurement of a c = 0.1 mg/mL cerato-ulmin solution. Each point is the zero-angle extrapolation of the fast decay mode from multiple DLS runs.

Figure 5. DLS of cerato-ulmin bubbles shows a decrease in hydrodynamic radius as an increased pressure is applied followed by an increase in radius as the pressure decreases.

develop as the solution was agitated by hand. The increased size distribution creates a larger effect of apparent radius with angle so the zero-angle extrapolation standard error became more pronounced. When structures were created, held for 1 h, and measured at applied overpressures, the bubbles were not only smaller but also more uniform as the apparent radius was less affected by measuring angle (see Supporting Information). The results so far indicate that pressure affects the size of the structures formed during agitation, but the structures may yet be solids. To further investigate, a cerato-ulmin solution was agitated by hand at atmospheric pressure and then subjected to overpressure and partial vacuum. After each pressure change the solution was allowed to equilibrate for 1 h. A syringe pump increased the pressure by 0.1 bar three times before returning to atmospheric conditions. Finally, a partial vacuum was created by the syringe pump and the pressure was decreased to 0.9 bar. Each data set was analyzed with cumulants, exponential, and CONTIN fitting algorithms at scattering angles of 45, 60, 75, and 90°. The simplest fitting procedure that produced random residuals was the double exponential fit with a fast and slow decay mode for the larger and smaller bubbles. Again, only the apparent sizes from the fast decay mode are used. Figure 5 shows the pressure experiment results at 45° scattering angle with the two-exponential fit. With increased pressure the bubbles shrunk from 519 to 57 nm. Once the pressure was brought down to atmospheric pressure again, the bubbles partly recovered their size and increased to 229 nm before a further expansion to 531 nm in a partial vacuum of 0.9 bar. These results suggest that the submicron structures are bubbles and not solids. Small Angle X-ray Scattering. To examine the persistent structures in greater detail, small-angle X-ray scattering experiments probed a scattering vector magnitude range 0.17 < q < 8 nm−1 corresponding to characteristic lengths of 0.8 < d < 37 nm where q is the scattering vector and d is the characteristic length. The scattering envelope can be divided into three regimes; see Figure 6. The low-q regime (q < 0.7 nm−1) shows a distinct peak at q = 0.4 nm−1 relating to a characteristic length of d = 15 nm. A Guinier regime extends to ∼2 nm−1. The Guinier plot in this regime (Figure 6 inset) gives a radius of gyration Rg = 1.1 nm. If the protein is

Figure 6. SAXS pattern of c = 2 mg/mL cerato-ulmin bubble dispersion and fit with a combination of a core−shell cylinder model representing the assembled structure and Guinier regime mostly representing free protein in solution. Uncertainties are comparable in size to data points. Inset: Guinier plot.

spherical, its diameter D is given by 2R g

5 3

= 2.8 nm. This is

close to the size of a protein with the mass of cerato-ulmin, assuming a spherical shape, which is given by D = 23

3M 4πρp Na

= 2.6 nm assuming density ρp = 1.3 g/cm3

and mass M = 7618 g/mol.30,40 Finally, there is the high-q regime, where two peaks appear at q = 4 and 5.5 nm−1, corresponding to 1.57 and 1.14 nm, respectively. These last two peaks lie in a regime where baseline subtraction may impact the results; though intriguing because they may reveal internal details of the CU protein, we do not consider them further. The 2.8 nm diameter from the Guinier regime is much smaller than bubbles of any recorded size. Accordingly, a continuum approach was applied to understand the low-q features. Because it was not clear whether the submicron structures are spherical, two potential models were tried: a D

DOI: 10.1021/acs.jpcb.9b01673 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

thickness of t = 13.6 ± 3.2 nm. Thus, the inner cavity has a diameter of ∼40 nm. The thickness parameter is the most influential on the overall scattering pattern in this q-range and is the most reliable value as it falls directly within the characteristic distances of the measured q-range. The contribution from the length is minimal within the measured q-range, so changes in L have only minor effects on the simulated scattering envelope. Even a change in the length up to a factor of 8 has negligible impact on the scattering pattern in this q region (Supporting Information). Atomic Force Microscopy (AFM). To corroborate the conclusions from SAXS regarding film thickness of the persistent structures, AFM was used to measure whether the membrane thickness of the microscopic bubbles was of similar magnitude. An agitated sample with c = 2 mg/mL was dropcast onto a glass slide and allowed to dry overnight. After drying, AFM images show deflated and collapsed bubbles with a height twice the thickness of the bubble film (Figure 8). A

core−shell spherical form factor and a core−shell cylindrical form factor.41,42 After potential sizes are simulated and the data are fit with both models, while adding in the Guinier term, the better fit is for the cylindrical shape. Other simulated scattering spectra and additional fits appear in the Supporting Information. Apparently, the persistent submicron structures maintain an extended form, as do the microscopic bubbles seen soon after hand agitation (Figure 1). In quantitative terms, the expression used for this successful (until q ∼ 2 nm−1) fit is given in eqs 1−4, and Figure 7 shows

Figure 7. Diagram of the core−shell cylinder model with appropriate terms.

the core−shell cylinder model along with appropriate terms. The intensity of scattering from randomly oriented objects is proportional to the form factor, I(q) ∝ P(q). For the core− shell cylinder model, the form factor is given by Pcore − shell cylinder(q ,α) =

scale Vs

∫0

π /2

f 2 (q) sin α dα

(1)

Figure 8. Cerato-ulmin c = 2 mg/mL AFM images and height profiles for inflated and deflated bubbles: (A) height profile across collapsed bubbles; (B) height profile of inflated bubble; (C) AFM image of collapsed bubbles (scale bar = 5 μm); (D) AFM image zoomed in on one inflated bubble (scale bar = 1 μm).

ÄÅ É ÅÅqL cos α ÑÑÑ ÅÅÇ ÑÖ J1[qr sin α ] 2Ñ f (q) = 2(ρc − ρs )Vc sin ÄÅ É ÅÅqL cos cos α ÑÑÑ [qr sin α ] ÅÅÇ ÑÖ 2Ñ

where

small number of bubbles did not fully deflate even after 24 h, demonstrating the high strength of the bubble film, in agreement with macroscopic measurements.33−35 Height profiles were created with line cuts using the NanoScope analysis software. The inflated bubbles were 204 ± 28 nm in height (5 cuts). The collapsed bubbles measured 27 ± 2 nm (19 cuts); assuming that this height represents two bubble membrane layers (top and bottom), a single membrane would have a thickness of 14 ± 1 nm, which is 5 times larger than the 2.8 nm diameter of the protein from SAXS measurements. An odd number of proteins results in a hydrophilic exterior and hydrophobic interior.

+ 2(ρs − ρsolv )Vs sin[q(L+2t ) cos α /2] / [q(L+2t ) cos α /2]

J1[q(r +t ) sin α ] [q(r +t ) sin α ]

(2)

Combining with the Guinier equation P(q) =

2 2 I(q) = e−(R g q /3) I(0)

(3)

the custom model for the core−shell cylinder with a Guinier contribution is I(q) = p1 P(q)core − shell cylinder + p2 P(q)Guinier



(4)

CONCLUSIONS We have studied the structure of buoyant microbubbles and persistent submicron bubbles formed by a hand-agitated cerato-ulmin solution. Our main techniques have been DLS, SAXS, and AFM. SAXS suggests two scattering populations: individual proteins in free solution and bubbles that fit a cylindrical shell form factor. The individual subunit size correlates to a protein of the size expected for cerato-ulmin, and the SAXS pattern fits a cylindrical shell form factor with a bubble thickness of 13.6 ± 3.2 nm and outside diameter of 69

where p1 and p2 are the weights of the Guinier and core−shell cylinder functions, L and t are the length and thickness, r is the radius of the core, Rg is the radius of gyration, J1 is the firstorder Bessel function, ρ is the scattering length density, α is the angle between the axis of the cylinder and the scattering vector q, and the subscripts c, s, and solv represent the core, shell, and solvent, respectively. Because the bubbles are randomly oriented in solution the function is integrated over all angles. The fit gives an external diameter 2(r + t) of 69 ± 5 nm with a E

DOI: 10.1021/acs.jpcb.9b01673 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B ± 3.2 nm. AFM experiments on the microscopic bubbles dropcast on a silicon wafer suggest the bubble membrane thickness to be approximately 14 nm, in agreement with the SAXS fitting of the submicron bubbles. Using the protein size derived from the SAXS scattering of the free proteins in solution from the Guinier contribution, we conclude that the bubble film is five protein molecules thick. Combining these conclusions, we suggest a cross-sectional morphology of the persistent, submicron bubbles (Figure 9). Future studies should be



SAXS modeling details and alternative approaches; SAXS theoretical curves for spherical core−shell (Figures S9 and S10) and cylindrical core−shell form factors (Figures S11 and S12) (PDF)

AUTHOR INFORMATION

Corresponding Author

*Paul Russo. Address: 771 Ferst Dr. NW, Atlanta, GA 30332. Phone: 404-385-2607. E-mail: paul.russo@mse.gatech.edu. ORCID

Paul S. Russo: 0000-0001-6009-2742 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was primarily supported by the Gulf of Mexico Research Initiative (GoMRI) through the Consortium for Molecular Engineering of Dispersant Systems (CMEDS) at Tulane University, and by a Hightower Endowment to the Georgia Tech Foundation. Limited support from NSF 1609058 is acknowledged. The Stanford Synchrotron Radiation Laboratory is supported by the U.S. Department of Energy, Office of Science and Office of Basic Energy Sciences under Contract No. DE-AC02-76SF00515.

Figure 9. Suggested shapes of cerato-ulmin cylindrical bubbles. (Left) The cross section of the bubble is five proteins thick with the hydrophobic and hydrophilic sections of the protein aligning to minimize energy. (Right) This section of the bubble shows one of the protein layers of the bubble film.



(1) Kisko, K.; Szilvay, G. R.; Vuorimaa, E.; Lemmetyinen, H.; Linder, M. B.; Torkkeli, M.; Serimaa, R. Self-Assembled Films of Hydrophobin Protein HFBIII from Trichoderma Reesei. J. Appl. Crystallogr. 2007, 40, S355−S360. (2) de Vocht, M. L.; Reviakine, I.; Ulrich, W. P.; Bergsma-Schutter, W.; Wosten, H. A.; Vogel, H.; Brisson, A.; Wessels, J. G.; Robillard, G. T. Self-Assembly of the Hydrophobin SC3 Proceeds Via Two Structural intermediates. Protein Sci. 2002, 11 (5), 1199−205. (3) Wosten, H. A. B.; Devries, O. M. H.; Wessels, J. G. H. Interfacial Self-Assembly of a Fungal Hydrophobin into a Hydrophobic Rodlet Layer. Plant Cell 1993, 5 (11), 1567−1574. (4) Wessels, J. G. H.; Devries, O. M. H.; Asgeirsdottir, S. A.; Schuren, F. H. J. Hydrophobin Genes Involved in Formation of Aerial Hyphae and Fruit Bodies in Schizophyllum. Plant Cell 1991, 3 (8), 793−799. (5) Askolin, S.; Linder, M.; Scholtmeijer, K.; Tenkanen, M.; Penttila, M.; de Vocht, M. L.; Wosten, H. A. B. Interaction and Comparison of a Class I Hydrophobin from Schizophyllum Commune and Class II Hydrophobins from Trichoderma Reesei. Biomacromolecules 2006, 7 (4), 1295−1301. (6) Lumsdon, S. O.; Green, J.; Stieglitz, B. Adsorption of Hydrophobin Proteins at Hydrophobic and Hydrophilic Interfaces. Colloids Surf., B 2005, 44 (4), 172−178. (7) Hakanpaa, J.; Paananen, A.; Askolin, S.; Nakari-Setala, T.; Parkkinen, T.; Penttila, M.; Linder, M. B.; Rouvinen, J. Atomic Resolution Structure of the HFBII Hydrophobin, a Self-Assembling Amphiphile. J. Biol. Chem. 2004, 279 (1), 534−539. (8) Liu, Y. Z.; Wu, M.; Feng, X. Z.; Shao, X. G.; Cai, W. S. Adsorption Behavior of Hydrophobin Proteins on Polydimethylsiloxane Substrates. J. Phys. Chem. B 2012, 116 (40), 12227−12234. (9) de Vocht, M. L.; Reviakine, I.; Wosten, H. A. B.; Brisson, A.; Wessels, J. G. H.; Robillard, G. T. Structural and Functional Role of the Disulfide Bridges in the Hydrophobin SC3. J. Biol. Chem. 2000, 275 (37), 28428−28432. (10) Kwan, A. H. Y.; Winefield, R. D.; Sunde, M.; Matthews, J. M.; Haverkamp, R. G.; Templeton, M. D.; Mackay, J. P. Structural Basis for Rodlet Assembly in Fungal Hydrophobins. Proc. Natl. Acad. Sci. U. S. A. 2006, 103 (10), 3621−3626.

directed toward the lengths of these bubbles, their dynamic responsiveness to pressure changes, and their stability to competing surfactants and cosolvents such as alcohols. We close by considering certain constraints on such future experiments. Surprisingly, for a system that is almost a dispersion of air in water, future SAXS measurements will be complicated by weak scattering signals. Some of the protein will rise out of the detected volume, but even if the entire 2 mg/mL protein solution were devoted to encapsulating air, resulting in a strong electron density contrast, the overall volume fraction of air is estimated at only ∼0.2%. Background subtraction will remain challenging. Preliminary small-angle neutron scattering experiments have not proven easy because of the low flux compared to synchrotron SAXS. On a positive note, the persistent bubbles are stable to dilution,32 and their study by classical and visible light scattering may lead to length and possibly mass information. Finally, there remains much to learn about hydrophobin structures in general and CU in particular. One of the open questions deserving study is how the persistent bubbles and even the larger cylindrical bubbles present just after agitating solutions of CU form. Possibilities range from “rolling” of a surface film, as suggested long ago,31 to some variation on nanotubulation, a process in which vesicles or other structures generate tubular protrusions.43,44



REFERENCES

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcb.9b01673. MALDI of cerato-ulmin protein (Figure S1); dynamic light scattering (Figures S2−S7); variables used in small angle scattering fitting (Tables S1 and S2); twodimensional raw SAXS scattering image (Figure S8); F

DOI: 10.1021/acs.jpcb.9b01673 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B (11) Qin, M.; Wang, L. K.; Feng, X. Z.; Yang, Y. L.; Wang, R.; Wang, C.; Yu, L.; Shao, B.; Qiao, M. Q. Bioactive Surface Modification of Mica and Poly(dimethylsiloxane) with Hydrophobins for Protein Immobilization. Langmuir 2007, 23 (8), 4465−4471. (12) Wang, R.; Yang, Y. L.; Qin, M.; Wang, L. K.; Yu, L.; Shao, B.; Qiao, M. Q.; Wang, C.; Feng, X. Z. Biocompatible Hydrophilic Modifications of Poly(dimethylsiloxane) Using Self-Assembled Hydrophobins. Chem. Mater. 2007, 19 (13), 3227−3231. (13) Wosten, H. A. B.; Schuren, F. H. J.; Wessels, J. G. H. Interfacial Self-Assembly of a Hydrophobin into an Amphipathic Protein Membrane Mediates Fungal Attachment to Hydrophobic Surfaces. EMBO J. 1994, 13 (24), 5848−5854. (14) Lugones, L. G.; Wosten, H. A. B.; Birkenkamp, K. U.; Sjollema, K. A.; Zagers, J.; Wessels, J. G. H. Hydrophobins Line Air Channels in Fruiting Bodies of Schizophyllum Commune and Agaricus bisporus. Mycol. Res. 1999, 103, 635−640. (15) Tchuenbou-Magaia, F. L.; Norton, I. T.; Cox, P. W. Hydrophobins Stabilised Air-Filled Emulsions for the Food Industry. Food Hydrocolloids 2009, 23 (7), 1877−1885. (16) Dickinson, E. Colloids in Food: Ingredients, Structure, and Stability. Annu. Rev. Food Sci. Technol. 2015, 6, 211−233. (17) Green, A. J.; Littlejohn, K. A.; Hooley, P.; Cox, P. W. Formation and Stability of Food Foams and Aerated Emulsions: Hydrophobins as Novel Functional Ingredients. Curr. Opin. Colloid Interface Sci. 2013, 18 (4), 292−301. (18) Scholtmeijer, K.; Janssen, M. I.; Gerssen, B.; de Vocht, M. L.; van Leeuwen, B. M.; van Kooten, T. G.; Wosten, H. A. B.; Wessels, J. G. H. Surface Modifications Created by Using Engineered Hydrophobins. Appl. Environ. Microb. 2002, 68 (3), 1367−1373. (19) Akanbi, M. H. J.; Post, E.; Meter-Arkema, A.; Rink, R.; Robillard, G. T.; Wang, X. Q.; Wosten, H. A. B.; Scholtmeijer, K. Use of Hydrophobins in Formulation of Water Insoluble Drugs for Oral Administration. Colloids Surf., B 2010, 75 (2), 526−531. (20) Rosu, C.; Kleinhenz, N.; Choi, D.; Tassone, C. J.; Zhang, X. J.; Park, J. O.; Srinivasarao, M.; Russo, P. S.; Reichmanis, E. ProteinAssisted Assembly of pi-Conjugated Polymers. Chem. Mater. 2016, 28 (2), 573−582. (21) Bilewicz, R.; Witomski, J.; Van der Heyden, A.; Tagu, D.; Palin, B.; Rogalska, E. Modification of Electrodes with Self-Assembled Hydrophobin Layers. J. Phys. Chem. B 2001, 105 (40), 9772−9777. (22) Hakanpaa, J.; Szilvay, G. R.; Kaljunen, H.; Maksimainen, M.; Linder, M.; Rouvinen, J. Two Crystal Structures of Trichoderma Reesei Hydrophobin HFBI - The Structure of a Protein Amphiphile with and without Detergent Interaction. Protein Sci. 2006, 15 (9), 2129−2140. (23) Devries, O. M. H.; Fekkes, M. P.; Wosten, H. A. B.; Wessels, J. G. H. Insoluble Hydrophobin Complexes in the Walls of Schizophyllum-Commune and Other Filamentous Fungi. Arch. Microbiol. 1993, 159 (4), 330−335. (24) Linder, M. B.; Szilvay, G. R.; Nakari-Setala, T.; Penttila, M. E. Hydrophobins: the Protein-Amphiphiles of Filamentous Fungi. Fems. Microbiol. Rev. 2005, 29 (5), 877−896. (25) Yamasaki, R.; Haruyama, T. Formation Mechanism of Flattened Top HFBI Domical Droplets. J. Phys. Chem. B 2016, 120 (15), 3699−3704. (26) Carpenter, C. E.; Mueller, R. J.; Kazmierczak, P.; Zhang, L.; Villalon, D. K.; Vanalfen, N. K. Effect of a Virus on Accumulation of a Tissue-Specific Cell-Surface Protein of the Fungus Cryphonectria(Endothia)-Parasitica. Mol. Plant-Microbe Interact. 1992, 5 (1), 55− 61. (27) Russo, P. S.; Blum, F. D.; Ipsen, J. D.; Abulhajj, Y. J.; Miller, W. G. The Solubility and Surface-Activity of the Ceratocystis-Ulmi Toxin Cerato-Ulmin. Physiol. Plant Pathol. 1981, 19 (1), 113−126. (28) Wosten, H. A. B. Hydrophobins: Multipurpose proteins. Annu. Rev. Microbiol. 2001, 55, 625−646. (29) Takai, S. Pathogenicity and Cerato-Ulmin Production in Ceratocystis-Ulmi. Nature 1974, 252 (5479), 124−126. (30) Yaguchi, M.; Pusztaicarey, M.; Roy, C.; Surewicz, W. K.; Carey, P. R.; Stevenson, K. J.; Richards, W. C.; Takai, S. Amino-Acid-

Sequence and Spectroscopic Studies of Dutch Elm Disease Toxin, Cerato-Ulmin. Dutch Elm Disease Research 1993, 152−170. (31) Russo, P. S.; Blum, F. D.; Ipsen, J. D.; Abulhajj, Y. J.; Miller, W. G. The Surface-Activity of the Phytotoxin Cerato-Ulmin. Can. J. Bot. 1982, 60 (8), 1414−1422. (32) Zhang, X.; Huberty, W.; Chen, Y.; Hung, F.; Russo, P. S. Microbubbles and Oil Droplets Stabilized by a Class II Hydrophobin in Marine-Like Environments. Langmuir 2019, 35 (12), 4380−4386. (33) Randall, G. C. Electric Field Deformation of Protein-Coated Droplets in Thin Channels. Langmuir 2018, 34 (34), 10028−10039. (34) Zhang, X.; Kirby, S. M.; Chen, Y.; Anna, S. L.; Walker, L. M.; Hung, F. R.; Russo, P. S. Formation and Elasticity of Membranes of the Class II Hydrophobin Cerato-Ulmin at Oil-Water Interfaces. Colloids Surf., B 2018, 164, 98−106. (35) Kirby, S. M.; Zhang, X. J.; Russo, P. S.; Anna, S. L.; Walker, L. M. Formation of a Rigid Hydrophobin Film and Disruption by an Anionic Surfactant at an Air/Water Interface. Langmuir 2016, 32 (22), 5542−5551. (36) Takai, S.; Richards, W. C. Cerato-Ulmin, a Wilting Toxin of Ceratocystis Ulmi: Isolation and some Properties of Cerato-Ulmin from the Culture of C. ulmi. J. Phytopathol. 1978, 91 (2), 129−146. (37) Stevenson, K. J.; Slater, J. A.; Takai, S. Cerato-Ulmina Wilting Toxin of Dutch Elm Disease Fungus. Phytochemistry 1979, 18 (2), 235−238. (38) Zhang, F.; Ilavsky, J.; Long, G. G.; Quintana, J. P. G.; Allen, A. J.; Jemian, P. R. Glassy Carbon as an Absolute Intensity Calibration Standard for Small-Angle Scattering. Metall. Mater. Trans. A 2010, 41a (5), 1151−1158. (39) Berne, B. J.; Pecora, R. Dynamic Light Scattering: with Applications to Chemistry, Biology, and Physics, Dover ed.; Dover Publications: Mineola, NY, 2000; p vii, 376 pp. (40) Fischer, H.; Polikarpov, I.; Craievich, A. F. Average Protein Density is a Molecular-Weight-Dependent Function. Protein Sci. 2004, 13 (10), 2825−2828. (41) Guinier, A., Fournet, G. Small-Angle Scattering of X-Rays; John Wiley and Sons: New York, 1955. (42) Livsey, I. Neutron-Scattering from Concentric Cylinders Intraparticle Interference Function and Radius of Gyration. J. Chem. Soc., Faraday Trans. 2 1987, 83, 1445−1452. (43) Ashok, B.; Ananthakrishna, G. Dynamics of Intermittent Force Fluctuations in Vesicular Nanotubulation. J. Chem. Phys. 2014, 141 (17), 174905. (44) Evans, E.; Heinrich, V.; Leung, A.; Kinoshita, K. Nano- to Microscale Dynamics of P-Selectin Detachment from Leukocyte Interfaces. I. Membrane Separation from the Cytoskeleton. Biophys. J. 2005, 88 (3), 2288−2298.

G

DOI: 10.1021/acs.jpcb.9b01673 J. Phys. Chem. B XXXX, XXX, XXX−XXX