Unravelling Protein–Fungal Hyphae Interactions at the Nanoscale (2025)

Materials

Egg white protein (EWP), Solanic 200 potato protein (PoP), and chilled heat-treated fungi hyphae paste at ∼24 wt % solids were supplied by Quorn Foods (Stokesley, North Yorkshire, U.K.). Type I (Milli-Q) water (Millipore, Bedford, U.K.), with a minimum resistivity of 18.2 MΩ cm and analytical grade chemicals were used in the preparation of all samples unless otherwise specified.

Surface Charge of Fungal Hyphae

The electrophoretic mobility of the hyphae as a function of pH can give an indication of the surface charge of the fungal hyphae. The fungal paste was frozen in liquid nitrogen and homogenized in a ceramic mortar and pestle. This process was repeated 4 times until a fine powder was obtained. The fine powder was dispersed in Milli-Q water and centrifuged at 1000g for 10 min twice to remove cellular debris. After the supernatant was decanted, the powdered paste was resuspended in Milli-Q water at 0.5 g/L and filtered through 0.2 μm asymmetric polyether sulfone membrane (ThermoFisher Scientific, Loughborough, U.K.) to remove larger fragments. The resulting filtrate was adjusted to various pH (pH 3–7), and electrophoretic mobility and ζ-potential values were recorded using DTS1070 folded capillary electrophoresis cells in a Malvern Zetasizer Ultra (Malvern instruments Ltd., Worcestershire, U.K.) at 25 °C. This was repeated three times for each pH condition.

Theoretical Modeling of Protein Interaction

Transmission Electron Microscopy

Stimulated Emission Depletion (STED) Imaging

Atomic Force Microscopy─Colloid Probe Force Measurements

ζ-Potential and Hydrodynamic Diameter Measurements

The ζ-potentials and hydrodynamic diameters of 0.3 wt % EWP (refractive index = 1.51, absorption = 0.001) and PoP (refractive index = 1.45, absorption = 0.001) solutions were measured in triplicate using Malvern Zetasizer Ultra (Malvern instruments Ltd., Worcestershire, U.K.). Both proteins were dissolved in Milli-Q water (dispersant: refractive index = 1.33, viscosity = 0.8872 cP, dielectric constant = 78.5) and adjusted to pH values between 7.0 and 3.0 using 1 M HCl or 1 M NaOH. The prepared solutions were filtered through 0.22 μm nylon syringe filter and placed in a DTS1070 cell and DTS0012 disposable cuvettes (PMMA, Wertheim, Germany) for ζ-potential and particle size measurements, respectively. Measurements were taken at 25 °C using backscattered light at 173° detection angle with samples left to equilibrate for 120 s.

Statistical Analysis

All means and standard deviations were calculated for triplicate measurements (n = 3 × 3). One-way analysis of variance (ANOVA) and posthoc Tukey tests were used to calculate mean separation at a 5% level of significance using Minitab 21 software.

Self-Consistent Field (SCF) Calculations of Coverage and Interactions

For SCF calculations, we modeled the fungal hyphae as parallel cylinders of radius of 1.5 μm, possessing a mildly hydrophobic surface due to the hyphal surface negative charge as a function of pH estimated from ζ-potential measurement of crushed fungal hyphae (see the Supporting Information file, Table S2). (6) Interactions between this surface and the hydrophobic amino acid residues were assigned a favorable Flory–Huggins χ parameter of −1.0 k_BT (see Experimental Methods, Table 1), typical of such hydrophobic interactions. (11,12,31) The bulk protein volume fraction was set at 1.0 × 10^–7 equivalent to ∼1.35 × 10^–5 wt % assuming average protein density of 13.5 μg/mL. (32) We chose low bulk values to reflect the expectation that most of the protein will be adsorbed and not be present in bulk solution. However, it is important to note that this does not mean that the amount of protein in the system is low. As for all other SCF calculations of protein adsorption, ovalbumin and patatin were modeled as unfolded primary chains with amino acids divided into six groups based on their pK_a, charge at neutral pH, and degree of hydrophobicity shown in Figure 2. (11−13,16,17) Due to the necessary simplification of the amino acid classification, the theoretical isoelectric points (pI) of ovalbumin and patatin were calculated as 5.30 and 5.87, respectively (Supporting Information file, Figure S1). This compares favorably with experimental pI values of 4.5 and 4.5–5.1, respectively, determined from electrophoretic mobility measurements of the native proteins measured using a laser Doppler velocimetry (ζ-potential) (Table S1). (9)

First, we explore the adsorption characteristics of the proteins before moving on to how this adsorption affects the interactions between the hyphal surfaces. The parallel cylinder case is not a frequently explored geometry in this type of calculations in the literature (most calculations involve flat plates or spheres), but it is most relevant to the type of fiber composite present in the mycoprotein products of interest here. This therefore requires what we believe are some original transformations from the starting numerical SCF plate–plate interaction results to the parallel cylinder case by making use of the Derjaguin approximation.

Protein Adsorption on the Fungal Hyphae Surface

For proteins to be effective binders, they have to adsorb to the surface and facilitate attractive interactions commonly measured as stickiness or adhesiveness, which otherwise have little or no interaction between them. (8) Denatured proteins have increased affinity for hydrophobic surfaces due to their more exposed hydrophobic amino acids. (33) Here, we consider the interfacial adsorption properties of ovalbumin and patatin modeled using their primary structure (Figure 2). This is obviously an approximation, but since globular proteins become more unfolded with adsorption (at least when there is plenty of surface available for adsorption at low initial bulk protein concentrations ∼1.35 × 10^–5 wt %, as here), more of the primary structure does indeed become exposed to the adsorbing surface. Thus, this approximation has yielded results for many other globular proteins that seem to accord qualitatively with experimental measurements of their adsorption properties, recalling that the SCF methodology yields the equilibrium adsorbed state. (11,12,34,35)

SCF-calculated protein density profiles were plotted as the volume fraction of protein ϕ_p as a function of the perpendicular distance from the hyphal surface at the bulk protein volume concentration of ∼1.35 × 10^–5 wt % (Figure S2). Ovalbumin (Figure S2a–c) and patatin (Figure S2d–f) are similar in their density profiles when they are bound to the hydrophobic surface. To illustrate the role of Coulombic forces, we calculated density profiles of the tested animal and plant proteins as a function of pH and ionic strengths. At the lowest background [NaCl] of 10 mM, the density profile of ovalbumin at the hydrophobic surface is highest (ϕ_p = 0.246) at the pH closest to the pI of pH 5.0 and declines as the protein becomes increasingly charged at pH values further away─the lowest ϕ_p (0.104) at the surface, in the range of pH values studied was observed at pH = 3.0 (Figure S2a). As expected, increasing the background [NaCl] to 100 mM (Figure S2b) and further to 500 mM (Figure S2c), caused charge screening effects dominate and narrow the differences in the profiles between the different pH values while simultaneously increasing the adsorbed amounts of ovalbumin: ϕ_p = 0.366 and 0.322 at pH 5.0 and pH 3.0 respectively, at 500 mM NaCl (Figure S2c).

Although we observed similar density profiles for patatin (Figure S2d–f), patatin’s adsorbed layer region extends farther away from the surface than ovalbumin at all measured conditions. For example, both proteins show a dense inner layer of ∼1 nm thickness, while for patatin, a less dense part of the layer extends to as far as ∼4.0 nm. For ovalbumin, this less dense, outer part is only ∼2.0 nm, with some slight variations depending on the [NaCl]. Thus, the adsorbed patatin molecules appear to stretch significantly further away from the surface, and this is due to the cluster of charged and/or polar amino acid residues found toward the C-terminus end of the chain (see Figure 2). To further illustrate the differences in the interfacial organization of ovalbumin and patatin, Figure 3 shows the average distance of each amino acid ranked from its N-terminus side, away from the hyphal surface.

At all measured [NaCl], ovalbumin lies very close to the surface, with amino acids 50–120 and 250–300 extending to ∼1 nm from the surface as a single protein interfacial layer, forming a roughly “M” shaped multiblock configuration. Not surprising, the slightly extended region has a higher ratio of polar and charged amino acids as well as the hydrophilic carbohydrate side chain (see Figure 2). Conformation of ovalbumin changes with pH at low [NaCl] (Figure 3a) but the variations are minimal at higher [NaCl] (Figure 3b,c), highlighting the importance of electrostatic repulsion for this protein. In contrast, patatin (Figure 3d–f) is divided into two regions forming more of a diblock type conformation, where the first block is hydrophobic and lies very close (<1 nm) to the surface while the second block is hydrophilic and extends away from the surface up to distances of ∼3–4 nm. Examination of the primary structure in Figure 2 shows that this outer part of the layer starts at the 288th residue from the N-terminus side (marked by a red circle) and forms a hydrophilic “tail” of 99 amino acids, 64 of which are charged or polar. This hydrophilic tail provides a “hairy” structure that should enhance electro-steric repulsion between hyphal surfaces coated in adsorbed patatin.

Adsorption isotherms of ovalbumin and patatin were calculated via the SCF scheme shown in Figure S3. (10,36,37) Like most proteins, maximum surface coverage plateau occurs at very low concentrations, and that coverage remains largely unaltered by an increase in bulk concentration but showed sensitivity on pH and salt concentration. The highest and lowest adsorbed Γ were observed at pH 5.0 and pH 3.0, although at the highest [NaCl] the differences between pH 4.5, 5, and 7 are small (see Figures S3c and S3f), indicating that the charge of the protein only plays a secondary role, having largely been screened by the background electrolyte. This highest Γ observed at pH 5.0, at lower salt concentrations, was expected considering that pH 5.0 is closest to the pI, so mutual repulsion between the protein chains and the surface will be minimal, favoring adsorption. It should be noted that these values for Γ are all within the expected range for protein adsorption measured experimentally. (16,31) Strikingly, Γ for patatin is 0.1 to 0.5 mg·m^–2 higher than for ovalbumin under most conditions tested (Figure S3d–f). This is attributed to the somewhat diblock-like configuration adopted by patatin upon adsorption, where some of its amino acids at one end extend further away from the surface. It is well-known that a diblock polymer suffers a smaller configurational entropy penalty when adsorbed than those that lie flatter on the surface. (38) With all of the above information being the same, this leads to a higher adsorption for patatin relative to ovalbumin.

Super-Resolution Microscopy of Real Hyphae–Protein Composites

We now questioned how these predicted differences in protein adsorption and conformation are reflected in the organization of proteins at the surface of fungal hyphae with added EWP or PoP using super-resolution microscopy. Remarkably, STED micrographs clearly demonstrate that EWP coats the hyphae uniformly where protein–hyphae interaction dominates over any EWP–EWP aggregation (Figure 4a–c). In contrast, PoP formed a rather patchy attachment on the hyphal surface showing dense regions of PoP clusters (Figure 4d–f), which was not possible to resolve using confocal microscopy previously. (6,9) The cross-sectional diameter of the aggregates was ≥50 nm. Going down the length scales, transmission electron microscopy (TEM) of resin-embedded protein–fungal hyphae composite was consistent with STED, validating this finding (see Figure 4g–i). The outer cell wall of the hyphae (Figure 4g) appears to have similar, comparatively low electron density (labeled A) as the hyphae–PoP composite (Figure 4i), which demonstrated that surface protein coating did not change significantly. In contrast, there is what appears to be a uniform distribution of electron dense material in the hyphae–EWP composite (Figure 4h) supporting the idea of protein coverage of the hyphal outer cell wall.

The observed differences in protein aggregation and surface coating behavior can be explained by considering the kinetics of patatin and ovalbumin denaturation. Pots, Gruppen, de Jongh, van Boekel, Walstra, and Voragen (39) showed that patatin undergoes reversibly partial unfolding at low temperatures ∼28 °C to form a reactive strand with increased exposure of hydrophobic regions and a single thiol group at neutral pH conditions. This loss of structure is also pH-dependent and occurs at pH ≤ 4.5. (39) Interaction of these exposed hydrophobic regions are dependent on protein concentration, temperature and pH conditions and have been demonstrated to be the main mechanism of patatin–patatin aggregation with the formation of disulfide bonds between thiol groups of a neighboring unfolded molecule playing a minor role. (10,37,39−41) Therefore, in this particular instance, PoP–PoP interaction supersedes PoP–fungal interaction. Conversely, ovalbumin begins to unfold at ∼76 °C implying much better thermal stability at room temperature and neutral pH conditions. (39,42) In essence, at room temperature and neutral pH conditions, a solution of ovalbumin is made of mostly globular proteins, while patatin exists as a mixture of globular and partially unfolded proteins, which form the aggregates observed in the STED micrographs.

Interaction of Protein-Coated Hyphal Surfaces Using SCF-Assisted Calculations

Having discussed the density profiles and spatial organization of proteins at the hyphal surface, it is also useful to calculate the interaction potentials of adsorbed protein layers as a function of pH and background [NaCl]. This would give insight into the binding capability of the two proteins, where a more attractive potential is presumed to indicate better binding. For two flat surfaces, the interaction potentials are calculated as the difference in free energy of the system at each separation distance relative to ‘infinite’ surface separation, here set at a sufficiently large value ≥100a₀. Note, a₀ is the nominal monomer unit size of 0.3 nm, which approximately corresponds to the length of a peptide bond. (11,12,31) The calculated interaction potential is the sum of the van der Waals and electro-steric interactions.

The interaction potential, V_cy(h), between two identical parallel cylinders of radius R can be obtained from the SCF-calculated interaction potential between planar surfaces V_pl(y) (plotted in Figure S4) using Derjaguin approximation, where $V_{\mathrm{cy}}(h)=\sqrt{R}{\int }_h^{\infty }\frac{V_{\mathrm{pl}}(y)}{\sqrt{y-h}}\mathrm{d}y$. Derivation of the corresponding equation for the spheres is commonly highlighted in many articles and books. However, since this is not the case for cylinders which is more representative of the hyphae configuration here, we provide a derivation of the above equation in the method section. With regard to the van der Waals component of the interaction a composite Hamaker constant (A_H) of 5 k_BT for hyphae dispersed in water was assumed. This is a typical value for protein-based particles in water. (43,44)

The net interaction potentials (U_TOT) per unit length of cylinders, as mediated by adsorbed ovalbumin (Figure 5a–c) and patatin (Figure 5d–f), are plotted against surface separation, h, between two parallel cylinders of radius R. The results are obtained at various pH values and are all expressed in units of k_BT·nm^–1. The van der Waals component of the interaction is already well-known and is given by the expression $V_{\mathrm{cyl}}(r)=\frac{-A\sqrt{R}}{24r^{3/2}}$ between the two parallel cylinders, separated by a distance h = r. We also plot the graphs for this latter, nonprotein-mediated interaction part separately for comparison, as indicated by black dashed lines.

The interaction potential with ovalbumin at ϕ_s = 0.001, approximately implying [NaCl] = 10 mM (Figure 5a), varies significantly with pH. At pH values away from the pI (∼pH 5.0, blue curve), i.e., pH 3.0 (black), pH 4.5 (red), and pH 7.0 (green), there is a significant positive (i.e., repulsion) interaction at a separation h = 1 nm. However, this repulsive potential drops off sharply with increasing h, so that by h = 5 nm, it is essentially zero (see Figure S5 for the expanded graphs). Close to the pI, there is little or no electrostatic repulsion, and therefore, one expects that the total interaction U_TOT will consist almost entirely of the steric and van der Waals interaction forces. Comparing this U_TOT to its van der Waals component part (dashed line) at pH 5.0 indicates that the steric interaction is minute and insufficient to overcome the attractive van der Waals forces between the surfaces. Moving away from the pI to pH 4.5 (red curve), there is a significant increase in the repulsion, decreasing slightly with further reduction in pH to pH 3.0.

Increasing ϕ_s to 0.01 ≈ [NaCl] = 100 mM (Figure 5b) screens the protein charge, eliminating the electrostatic repulsion between the protein layers and results in similar U_TOT versus h curves at all pH that do not differ significantly from the purely attractive van der Waals component. Further increase in ϕ_s to 0.05 ≈ [NaCl] = 500 mM (Figure 5c) leads to a slight increase in attractive potential due to increased protein adsorption, as described earlier. The responses to pH and salt clearly suggest that electrostatic interaction is the main component of the repulsive interaction for ovalbumin-coated fungal hyphae, with steric interactions playing a less dominant role. This is consistent with the previous macroscopic reports on the rheology of the composites as a function of pH and salt. (6) Also, the negative interaction potential observed at the highest [NaCl] supports the idea of the existence of attractive bridging interactions between the surfaces at the close surface separations. This occurs with proteins having many short segments of hydrophobic and hydrophilic residues in their backbone that are then more prone to simultaneous adsorption on two adjacent surfaces which then tends to cause bridging flocculation. This attractive bridging under most pH and salt conditions explains how ovalbumin, and by extension EWP, provides superior binding properties between fungal hyphae surfaces.

Patatin shows significantly different behavior as compared to ovalbumin. At ϕ_s of 0.001 ([NaCl] ∼ 10 mM) (Figure 5d), there is a repulsive interaction potential at all pH conditions. This is the case even close to the protein pI ∼ pH 5.0, indicating the presence of non-Coulombic repulsive forces large enough to overcome the attractive van der Waals interactions. As shown in Figure 3d–f, the adsorbed configuration of patatin can clearly account for this in terms of steric repulsion due to the protruding hydrophilic parts of the chains. There is significant variability of the repulsive interaction with pH, but nonetheless, it remains positive with the lowest level of repulsion observed at neutral pH 7.0 conditions.

Screening patatin’s charge by increasing ϕ_s to 0.01 ≈ [NaCl] = 100 mM (Figure 5e) reduces the positive (repulsive) interaction potential, with further increase to ϕ_s to 0.05 equiv to [NaCl] = 500 mM (Figure 5f), leading to even further reduction in the repulsion, with less variation with pH. This trend with increasing [NaCl] is again due to increased screening of the charged extended chains (decreasing mutual repulsion between like-charged adjacent chains) but allowing for increased protein adsorption at higher [NaCl]. Thus, unlike ovalbumin- coated surfaces, where the repulsion is almost entirely controlled via electrostatics, patatin appears to mediate both steric and electrostatic components, with the former playing a more dominant role. More crucial is the absence of an overall attractive potential between patatin-coated surfaces, which indicates PoP will be a less efficient binder of the fungal hyphae in real composites.

Experimental Validation of Interaction Forces via AFM-Colloidal Probe Force Spectroscopy

Finally, we measured interaction potentials of EWP and PoP via AFM to validate the SCF predictions in Figure 6 (see Figure S6 for the expanded force-distance curves and curves between clean silicon surfaces). About 1.0 wt % protein solutions with very little salt present estimated to be <1.0 mM were left to adsorb onto silicon wafer and SiO₂ colloidal probe of radius 1 μm. Adsorbed protein layers were then adjusted to pH 7.0, 5.0, 4.5, and 3.0 conditions and force–distance curves measurement in MQ-water.

In agreement with the SCF-calculated interaction forces, EWP (Figure 6a) at pH 4.5 gave the highest surface attraction (negative force), which increased as the surfaces approached until h ∼ 20 nm. This is the pI of ovalbumin (the main constituent of EWP) where the protein is minimally charged, and therefore, there should be minimal electrostatic repulsion between the protein layers. A slight increase to pH 5.0 introduces more surface charge and therefore repulsion, which reduces its attractive surface interaction, but it stays negative at h ≥ 17 nm. Further increases in protein surface charge at pH 7.0 and 3.0 overcome the attractive forces, leading to an overall repulsive interaction force between EWP layers at all h. The strong short-range repulsive force is probably steric in nature and will be a consequence of the combination of all the proteins that make up EWP, as well as any aggregates that they have formed. (45,46) However, despite these complications with EWP, the overall trends earlier predicted by the SCF calculations is validated─that ovalbumin does indeed dominate the overall interaction forces with EWP, which is largely Coulombic in nature. In contrast to EWP, the interaction forces between PoP-coated surfaces as a function of h remain positive, i.e., repulsive, at all measured pH conditions (Figure 6b). The lowest repulsive interaction forces are measured at pH 4.5 and pH 5.0, which are within the range of the pI of patatin (pH 4.5–5.2). This corroborates the SCF calculations and points out that steric repulsion is indeed the main controlling interaction mechanism of PoP.

As a first attempt to explain the data in Figure 6a and b quantitatively, we used a DLVO model to predict the combined effects of the electrostatic repulsion between the surfaces and the attractive van der Waals interaction. We used the model of Carnie, Chan, and Gunning (47) to describe the electrostatic interaction between a sphere and a flat (equivalent to a sphere of extremely large radius in their model), using the assumption of a constant surface charge, which is probably the most appropriate for protein-covered surfaces in this medium. We also assumed that the surfaces were coated with enough protein so that the surface potential of both surfaces was the same and equal to the measured values of the ζ-potential (see Table S1) for the proteins at the appropriate pH, i.e., −23.9, −3.24, and +28.87 mV for pH 7.0, 4.5, and 3.0, respectively. Again, this seems to be a reasonable assumption, for example, when the ζ-potential of protein-stabilized oil (emulsion) droplets is compared with the ζ-potential of the proteins themselves. (48−50)

The van der Waals interaction for a silica sphere interacting with a silicon substrate was calculated using the model of Wang, Wang, Hampton, and Nguyen, (30) with a 0.5 nm offset in the start of the electrostatic repulsion to account for the short silica “hairs” that are usually present on SiO₂ surfaces. We included this in our calculations, shown in Figure 6c, although this makes very little difference to the curves or the overall conclusion─that the measured repulsion for EWP far exceeds the predicted repulsion at short h at pH 3.0 and 7.0. At pH 4.5, close to the pI of ovalbumin, the predicted and measured forces are much closer for h > 20 nm, in fact both are close to zero net F/R, as expected when the net charge on ovalbumin is zero, but for h < 20 nm, the measured forces become increasingly repulsive as h decreases, while the predicted F/R start to become slightly negative (i.e., attractive), as expected for this low surface charged scenario (ζ-potential = −3.2 mV).