Structural characterization and interfacial properties of yeast cytoplasmic protein crosslinked by laccase enzyme

Abstract

In the present work, yeast cytoplasmic protein (YCP) was modified enzymatically by laccase in the presence and absence of ferulic acid (FA) to improve their technofunctional and interfacial properties. The 3 h, 18 h and 24 h reaction times were selected to represent low, medium and high levels of crosslinking based on the proportion of protein >150 kDa. In samples with FA, LC-MS/MS analysis found the greatest prevalence of linkage was in the form of diferulic acid, whereas those without FA had a greater incidence of linkages in the form of dityrosine and isodityrosine. Air/water interface analysis revealed that non-FA modified YCP had quicker adsorption and an enhanced effect on lowering surface tension, whereas FA modified YCP had slower transfer to the interface. Overall, the relationship between the structure and technofunctional properties of laccase-modified YCP was elucidated. Modification with laccase has the potential to further improve the emulsifying capabilities of YCP.

Highlights

• •Crosslinking of yeast cytoplasmic protein (YCP) by laccase was successfully achieved.
• •Addition of protease inhibitors from potato sources limited the self hydrolysis of YCP.
• •Crosslinking predominantly occurred through ferulic acid (FA) linkages in FA-treated proteins.
• •In non-FA treated proteins, crosslinking was achieved through tyrosine linkages including dityrosine and isodityrosine.
• •Small non-FA modified YCP decreased interfacial tension whereas large FA modified YCP slowed interfacial adsorption rates.

1. Introduction

The global population is expected to reach 9 billion people by 2050, which will put added pressure on existing agricultural resources for the production of protein containing foods. To support the increasing needs of the food industry, proteins of microbial origin are gaining in popularity as an alternative and renewable protein source. Of particular interest is protein derived from yeast, which has long been recognized as a staple ingredient for the fermentation processes used in bread and wine making. Protein sourced from saccharomyces cerevisiae is gaining popularity as a source of animal feed, as well as a flavoring agent for human food and a source of added nutrients (Jach et al., 2022). The technofunctional properties of yeast protein have also shown promise in potential applications in food systems (Ma et al., 2024; Vélez-Erazo et al., 2021).

Protein and polysaccharides are the two main components of bulk yeast cells with the protein component present mainly in the cytoplasm and the carbohydrate component mainly in the cell wall. Whole yeast biomass contains 39.6% protein, whereas extracted cytoplasmic material, from which cell wall fragments are removed, is significantly enriched to 62.6% protein. This high protein content, together with a well-balanced profile of essential dietary amino acid, makes yeast cytoplasmic protein a valuable ingredient for food formulations. Protein derived from yeast cytoplasm has become an increasingly active area of research. Traditionally, yeast cytoplasmic proteins have been used in their hydrolysed form to produce yeast extracts for imparting umami flavour. More recently, intact yeast cytoplasmic proteins have also been developed for a wide range of applications, including beverage fining agents, dietary and muscle-recovery protein supplements, meat extenders and flavor enhancers, emulsion and foam stabilizers, and materials for microencapsulation (Jach et al., 2022; Ma et al., 2024). Microencapsulation is an application which deserves particular attention. The process involves the entrapment of fragile bioactive components within protein or carbohydrate components extracted from the yeast cell, to lend them greater stability and preserve their bioactivity. Yeast protein and cell wall fragments have been shown to be promising encapsulates for oil components, and also to be well suited for spray dry applications to generate powders containing bioactive encapsulates (Vélez-Erazo et al., 2021).

Although yeast cytoplasmic protein offers high quality, it contains significant amounts of proteases, which can be problematic in encapsulation or emulsification. While these enzymes benefit yeast extract production by facilitating protein hydrolysis for flavoring, they hinder applications requiring intact protein, necessitating control measures to preserve protein functionality. Approaches like heat treatment, pH adjustment, or the addition of protease inhibitors can limit the protease activity. However, in the case where the unhydrolyzed protein is to be used, a treatment with protease inhibitors is the most appropriate as it does not compromise the protein integrity. Protease inhibitors, typically small molecules under 100 kDa, bind to the enzyme's active site, blocking its action (Leung et al., 2000). Plant-based foods, including seeds, fruits, and tubers, are rich sources of these inhibitors, contributing up to 10% of their protein content and have proven effective in addressing various food processing challenges (García-Carreño, 1996).

Structural modification of yeast cytoplasmic protein to improve its technofunctional properties is of great interest. Enzymatic modifications, in particular, have proven effective in improving the solubility, emulsifying, foaming and gelling properties, as well as digestibility and reducing allergenicity in plant, legume, and other non-animal proteins (M. Li, S. Karboune, K. Light, et al., 2021; M. Li, S. Karboune, L. Liu, et al., 2021; Olatunde et al., 2023). Functional modification of yeast proteins has been applied using physical and chemical and enzymatic methods to improve protein solubility (Zhao et al., 2024). Enzymatic modification has involved hydrolytic strategies using papain alone or hydrolysis with papain coupled with chemical modification to improve emulsification, foaming and antioxidant properties of yeast protein (Abedi et al., 2025; Wang et al., 2011). Enzymatic treatments such as hydrolysis, crosslinking and conjugation can fine-tune the emulsifying properties of proteins. They can modulate protein solubility and the balance of hydrophilic and hydrophobic regions by changing the molecular weight and altering the protein's secondary structure, potentially exposing more hydrophobic areas. Indeed, the ability of proteins to form emulsions depends on their ability to reduce interfacial tension between oil and water phases and to provide electrostatic or stearic repulsion between droplets to prevent destabilizing forces such as droplet coalescence or flocculation (Light & Karboune, 2021).

In the present work, yeast cytoplasmic protein (YCP) was modified by laccase-catalysed oxidative crosslinking using laccase from Trametes versicolor (LacTv) and Coriolus hirsutus (LacCh) with or without the addition of a ferulic acid (FA) as a mediator. Phenolic acids such as ferulic acid are established crosslinking assistors in laccase-catalysed reactions, extending the enzyme's range of action by functioning as radical shuttles between the enzyme and target protein structures. Various protease inhibition strategies were evaluated for their effectiveness in limiting endogenous protease activity during modification. The effect of the enzymatic modification on the protein's molecular weight profile, secondary and tertiary structures was assessed. Additionally, both covalent amino acid crosslink formation and non-crosslinking oxidative amino acid modifications induced by laccase treatment were investigated. Finally, the air-water interfacial properties of the modified proteins were examined and related to the structural changes induced by oxidative crosslinking. To the best of the authors' knowledge, no study has examined the oxidative crosslinking of yeast cytoplasmic protein and its resulting interfacial properties. This study is expected to contribute to furthering the understanding of structure-function relationships in enzymatically modified yeast cytoplasm protein (YCP) and to expand its potential as a vegan, non-allergenic, diet compatible and renewable protein source.

2. Materials and methods

2.1. Materials

Protein extract from yeast cytoplasm was kindly provided by Lallemand Inc. (Montreal, CA). Potato protein extracts were provided by Avebe (Veendam, The Netherlands). Standards for glucan and mannan were purchased from Megazyme (Bray, Ireland), and standards for Dityrosine, methionine sulfoxide, kynurenine and N′-formylkynurenine were purchased from Toronto Research Chemicals Inc. (North York, Canada). All other reagents were purchased from Millipore Sigma Co. (St Louis, MO, USA).

2.2. Preparation of yeast cytoplasmic protein

YCP was prepared from a non-hydrolyzed protein extract of lysed yeast cells. The yeast extract was mixed with 50 mM pH 5 sodium phosphate buffer at 50 mg/mL and stirred by magnetic bar at room temperature for 15 min. To remove insoluble yeast cell wall debris, the mixture was centrifuged at 10000 rpm for 20 min at 4 °C. This centrifugation step was repeated twice more until a clear supernatant was obtained. The supernatant was then dialyzed at 10 °C with a 10 kDa molecular weight cutoff to remove small phenolic components. Finally, the extract was lyophilized and stored at −20 °C until further treatment.

2.3. Chemical characterization of yeast cytoplasmic protein

Protein content was estimated by Dumas method and was carried out using Leco® TruSpec N system (Leco Corporation, St-Joseph, Michigan, USA). Nitrogen content was multiplied by a factor of 6.25 to determine the total protein content.

The total carbohydrate content of extracts was determined using phenol‑sulfuric acid test. 200 μL of phenol (5% w/v) and 1 mL of sulfuric acid were added to each 400 μL aliquot of YCP suspensions. After 15 min incubation, the absorbance was measured spectrophotometrically (DU 800, Beckman Coulter, Fullerton, CA) at 480 nm against a reagent blank. The carbohydrate content was estimated from a standard curve constructed with glucose.

The total phenolic content of native and modified YCP was determined using Folin-Ciocalteu phenol reagent according to (M. Li, S. Karboune, L. Liu, et al., 2021). A mixture of protein solution (0.05–0.2 mg/mL), sodium bicarbonate (250 g/L) and phenol reagent was prepared at ratio of 800/150/50 (v/v/v), and incubated at 40 °C for 30 min. The absorbance was measured at wavelength 765 nm. Measurements were carried out in duplicate. The calibration curve was constructed using gallic acid solutions at concentrations ranging from 0.001 to 0.1 mM.

To determine the monosaccharide profile, the carbohydrate fractions were first hydrolyzed according to the method of Khodaei and Karboune (2013). Briefly, 10 mg samples were weighed, hydrated in 200 μL water and 2.5 mL methanolic HCl was added. First hydrolysis occurred at 60 °C for 24 h. Samples were dried to 1 mL volume and 3.5 mL 6:1 water: trifluoroacetic acid was added. Second hydrolysis was carried out by boiling capped tubes for 1 h. After hydrolysis, samples were neutralized and filtered 0.2 μm prior to analysis. The monosaccharides were measured using High-Pressure Anionic Exchange Chromatography equipped with Pulsed Amperometric Detection (HPAEC-PAD) (Dionex, ICS-3000) and a CarboPac PA20 column (3 × 150 mm) set at 32 °C. The mobile phase was 20 mM NaOH at flow rate of 0.4 mL/min. Glucose and mannose were used at varying concentrations (2.5–50 μM) as standards. The analysis was performed in duplicate.

2.4. Protease deactivation treatment

Protease activity was expressed in the YCP samples which interfered with the enzymatic crosslinking catalysed by laccase. Two different strategies were investigated for the inhibition of protease activity including heat treatment and using protease inhibitors. For the heat treatment, YCP samples of native protein at 5 mg/mL were incubated at 60 °C to 85 °C for 5 min. YCP samples were also incubated in the presence of 1–5% (w/w relative to g of YCP) potato protease inhibitor (PPI) at room temperature or in refrigerated conditions for 24 h. The treated samples were analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Total intensity was calculated as the combined intensity of all bands present on the gel. The % hydrolysis was calculated to reflect the decrease in total intensity between the 0 h (refrigerated) and 24 h samples.

2.5. SDS polyacrylamide gel electrophoresis

The molecular weight distribution of YCP was assessed by SDS-PAGE. The electrophoresis was carried out using a mini protein gel apparatus (Bio-Rad, Hercules, CA). Aliquots of 15 μL of undiluted protein mixture were loaded into 0.75 mm polyacrylamide gels, which consisted of 4% of stacking gel and 12% of resolving gel, and the protein electrophoresis was run at 120 V. Protein standards with a broad range of molecular weight (10 to 200 kDa, Millipore) were used for molecular weight calibration. After Coomassie staining, the electrophoretic pattern and band relative proportion were analyzed densiometrically using GelDoc Go Imaging System and Alphaview software.

Extent of hydrolysis was calculated by comparing the total intensity of the bands in a given lane to the total intensity of the same treatment at 0 h, with intensity adjustment relative to the 0 U/mL 0 h treatment. The extent of crosslinking was calculated based on the relative intensity of the bands between 50 and 150 kDa on the gels compared to the intensity of the same timepoint with 0 U/mL of enzyme. This molecular weight range was selected for being where the effect of crosslinking was most visually evident on the gels. Hydrolysis and crosslinking extents were calculated by the following equations:

$$\mathit{Hydrolysis\; extent}=\mid \frac{\left(\mathit{aUbh}-\mathit{aU}0h\right)\ast 100}{\mathit{aU}0h}\mid$$

$$\mathit{Crosslinking\; extent}=\frac{\left(\mathit{aUbh}-0\mathit{Ubh}\right)\ast 100}{0\mathit{Ubh}}$$

where a is the number of enzyme units for the selected treatment and b is the time-course of the selected treatment.

2.6. Enzymatic crosslinking

2.6.1. Laccase production and activity

Two fungal laccases were used as biocatalysts. Laccase from Trametes versicolor was purchased from Sigma-Aldrich (St Louis, MO). While laccase from Coriolus hirsutus was produced according to the modified protocol reported by Gill et al. (2018) with an additional ammonium sulfate precipitation step at 80% saturation. The activity of laccases was carried out using the 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) assay, with one unit representing the amount of enzyme that oxidizes 1 μmol of ABTS per min. The protein content of the enzyme was determined using Hartree-Lowry assay.

2.6.2. Laccase-catalysed oxidative crosslinking

Laccase-catalysed oxidative crosslinking of YCP was carried out in 5 mL reaction volumes, in the presence and absence of 1 mM FA, over reaction time from 0 to 24 h with enzyme unit levels between 0 and 0.2 U/mL and in the presence of 2.5% (w/w) PPI added in relation to the quantity of YCP. YCP concentration in the reaction mixture varied from 1.24 to 1.5 mg/mL. The reaction was carried out in sodium phosphate buffer 50 mM at pH 5 and 28 °C with orbital agitation of 150 rpm. Aliquots of the reaction were taken at the predetermined times and were stored at −20 °C. These reactions were carried out in duplicate along with a blank without enzyme.

The molecular weight profile of the modified protein was either determined by SDS-PAGE, or by size exclusion fast protein liquid chromatography (SE-FPLC). SE-FPLC was carried out using an AKTA FPLC system equipped with a Superdex 200 column (10/300 GL, GE-Healthcare, Piscataway, NJ), eluted with sodium phosphate buffer (50 mM, pH 7) containing 0.15 M NaCl and 0.1% w/v sodium dodecyl sulfate (SDS), at flow rate of 0.5 mL/min. The elution was monitored at 280 nm by a UV detector. The standard curve was constructed using lysozyme (14 kDa), bovine serum albumin (66 kDa) and thyroglobulin (670 kDa), which were run at variable concentrations in order to apply a size-based ultraviolet absorbance intensity correction factor to the YCP chromatograms.

2.6.3. Laccase kinetic parameters

Laccase-catalysed oxidation of YCP was investigated by monitoring the decrease in the fluorescence signal of aromatic amino acids over time according to the modified method of Isaschar-Ovdat and Fishman (2017) and Ma et al. (2020). The kinetics of laccase-YCP systems were evaluated at pH 5, 28 °C. YCP substrates, ranging from 0.25 to 3 mg/mL in sodium phosphate buffer (50 mM, pH 5), were incubated with laccase at concentrations ranging from 1.5 to 3 mg/mL. The fluorescence intensity of the enzymatic reactions at excitation/emission wavelengths of 274/310 nm was monitored at selected intervals of 2 min, over 30 min reaction time course, using SpectroMax® i3x plate reader. The initial velocity of the oxidation was estimated from the initial linear trendline of the fluorescence decline curve. The laccase enzyme activity was expressed as μmol of equivalent tyrosine of the protein substrate that was oxidized per min per mg of laccase protein. The laccase enzyme activities were plotted versus the substrate concentration using SigmaPlot 12.3 and fitted to the Hill equation. All reactions were performed in duplicate with control reactions without laccase being carried out in tandem.

2.7. Characterization of linkages and oxidation products

2.7.1. Characterization of ferulic linkages by LC-MS

The recovery of phenolic compounds (ferulic linkages) from the modified protein was performed using an optimized alkaline hydrolysis procedure, in which the protein hydrolysate (12.5 mg/mL) was suspended in 4.2 M sodium hydroxide and incubated at 68 °C for 18 h in darkness. The alkaline hydrolysates were acidified using 10 M HCl to adjust the pH to 2.0–2.5. Phenolic compounds were then extracted three times with ethyl acetate at a ratio of 1:1.5 (v/v); the extracts were evaporated until dryness and redissolved in methanol/water mixture (50:50 v/v).

The samples were analyzed by LC-MS/MS using an Agilent 1290 Infinity II LC system coupled to the 6560 ion mobility Q-TOF -MS (Agilent Technologies, Santa Clara, USA). The LC separation was conducted on a Poreshell120 EC-C18 analytical column (Agilent Technologies; 2.7 μm × 3 mm × 100 mm) connected with a Poreshell120 EC-C18 guard column (Agilent Technologies; 2.7 μm × 3 mm × 5 mm). The mobile phase A was HPLC water with 0.1% formic acid and the mobile phase B was methanol with 0.1% formic acid. HPLC parameters were as follows: injection volume was 8 μL, the flow rate was 0.3 mL/min, and the column temperature was set to 30 °C. The mobile phase profile used for the run in negative ion mode was 2% B (0 to 0.5 min), 2% - 100% B (0.5 to 8.0 min), hold at 100% B (8.0–12.0 min), decrease to 2% B (12.0 to 12.5 min) and hold 2% B (12.5 to 13.0 min). The mass spectrometer was equipped with a Dual AJS ESI ion source operating in negative ionization mode. MS conditions were as follows: the drying gas temperature was 250 °C, drying gas flow rate was 11 L/min, sheath gas temperature was 300 °C, sheath gas flow rate was 11 L/min, the pressure on the nebulizer was 40 psi, the capillary voltage was 3500 V, the fragmentor voltage was 125 V, and the nozzle voltage was 500 V. Full scan MS data were recorded between mass-to-charge ratios (m/z) 100 and 1100 at a scan rate of 4 spectra/s, and were collected at both centroid and profile mode. Reference ions (m/z at 112.9856 and 1033.9881for ESI-) were used for automatic mass recalibration of each acquired spectrum. Data treatment was conducted using Quantitative Analysis B.10 from Agilent MassHunter Workstation Software. The identification of oligomerized FA crosslinks was based on the accurate mass and the isotopic patterns, as well as the diagnostic fragment ions if available.

2.7.2. Analysis of dityrosine linkages by LC-MS

Analysis of dityrosine linkages was carried out by acid hydrolysis followed by LC-MS/MS analysis according to the method by Nguyen et al. (2017). Samples of 10 mg were placed in glass tubes fitted with screw caps and were dissolved in 500 μL ultra pure water. To the mixtures were added 500 μL of 12 N HCl and 0.5 mL of propionic acid. The tubes were flushed with nitrogen for 20s, sealed tightly and then wrapped with aluminium foil. After hydrolysis in a sand bath at 120 °C for 24 h, the hydrolysates were evaporated in a nitrogen evaporator at 50 °C until dryness. The residues were redissolved in 1.5 mL of formic acid (0.1%, v/v) by immersion in an ultrasonic bath for 10 min followed by clean-up using solid phase extraction with C18 cartridges. Prior to clean up, the cartridges were conditioned with 2 mL of absolute MeOH, then 6 mL of 50 mM sodium phosphate buffer at pH 7.4 and then 6 mL of 0.1%, v/v trifluoroacetic acid. The redissolved hydrolysates were loaded on the cartridge followed by washing with 6 mL of formic acid (0.1%, v/v). Dityrosine was eluted with 3 mL of a mixture of MeOH and 0.1% formic acid (15/85, v/v). The eluents were concentrated to dryness by nitrogen evaporator at 37 °C. The resulting residues were reconstituted with 100 μL of formic acid (0.1%, v/v) and then transferred to HPLC vials stored at −20 °C until analysis.

The samples were analyzed by LC-MS/MS using an Agilent 1290 Infinity II LC system coupled to the 6560 ion mobility Q-TOF -MS using the same LC-MS method as that for ferulic linkage analysis described above.

2.7.3. Analysis of oxidative reaction products

A spectrophotometric assay for the total amount of carbonyl modified amino acids in the protein samples was carried out according to the modified method of Fitzner et al. (2023). For each sample, a 10 mg / 0.6 mL (by protein content) solution in water was prepared. The protein was precipitated by adding 1 mL of 20% (w/v) trichloroacetic acid (TCA) to each 0.6 mL protein solution aliquot, centrifuging at 10000 g for 5 min and discarding the supernatant. To one set of duplicates, 1 mL of 10 mM 2,4-dinitrophenylhydrazine (DNPH) solubilized in 2 M HCl was added, and to the other was added only 2 M HCl to act as a control. The samples were incubated in the dark for 1 h, and then 0.8 mL of 20% TCA was added to induce precipitation. The pellets were recovered by centrifugation (10,000 g, 5 min) and then washed 3 times with 1 mL of ethanol: ethyl acetate (1:1 v/v) to remove excess DNPH. Finally, samples were re-dissolved in 2 mL of 6 M guanidine-HCl in 20 mM sodium phosphate buffer at pH 2.5, and the absorbances were read spectrophotometrically at 370 nm for the quantification of the carbonyl concentration and at 278 nm for the determination of the protein concentration. Carbonyl and protein concentrations were calculated using molar extinction coefficients (22,000 M⁻¹ cm⁻¹) and (50,000 M⁻¹ cm⁻¹) respectively. The mol carbonyl / mg protein was calculated according to the following equation, reported by Levine et al. (1994), with Abs₃₇₀ representing the absorbance of the DNPH containing sample at 370 nm – the absorbance of the blank sample without DNPH at 370 nm.

$$\frac{\mathit{mol}\mathit{carbonyl}}{\mathit{mol}\mathit{protein}}=\frac{\mathrm{50,000}\left(\mathit{Ab}{s}_{370}\right)}{\mathrm{22,000}\left(\mathit{Ab}{s}_{276}-0.43\mathit{Ab}{s}_{370}\right)}$$

2.7.4. Analysis of free thiol groups

Free thiol groups in the YCP samples were analyzed by Ellman's assay (Habeeb, 1972). YCP samples were solubilized in 0.08 M sodium phosphate buffer with 2% SDS and 0.5 mg/mL ethylenediaminetetraacetic acid. 5,5′-dithio-bis-(2-nitrobenzoic acid) (DTNB) was prepared at 4 mg/mL concentration and 50 μL of the DTNB reagent was added to 1.5 mL aliquots of the protein solution. Control protein samples without the addition of DTNB were also prepared, and 50 μL of buffer was added instead. All samples were prepared in duplicate and analyzed spectrophotometrically at 412 nm against a buffer blank. The absorbance values from the protein alone samples were subtracted from the protein + DTNB samples and the molar extinction coefficient of the nitromercaptobenzoate anion (13,600 M^-l cm⁻¹) was used to calculate the value of free thiol groups in the protein sample.

2.8. Interface and emulsification properties

The air/water adsorption kinetics of the native and modified YCP were evaluated by maximum bubble pressure and drop volume method according to M. Li, C. Blecker, et al. (2021). Prior to each measurement, the tensiometers and glassware were cleaned thoroughly with MilliQ water and a value of 72 mN/m ± 0.5 mN/m for the blank (MilliQ water) was achieved. The protein solutions were prepared at a concentration of 1 mg protein /mL according to Dumas analysis. Solutions were prepared in 10 mM sodium phosphate buffer pH 6, stirred at room temperature for 1 h and allowed to hydrate in the fridge overnight prior to analysis. All measurements were performed in triplicate at 22 °C.

2.8.1. Maximum bubble pressure method

Maximum bubble pressure method was used to analyze the initial adsorption kinetics of the native and modified YCP at the air-water interface at 22 °C. Air bubbles were generated in a Bubble Pressure Tensiometer BP100 (Krüss GmbH Hamburg, Germany) equipped with an S180 capillary (diameter 0.2 mm). The surface age of the generated bubbles ranged from 10 to 10,000 ms. When the bubble radius reached that of the capillary, the pressure was at the maximal level. The surface tension was monitored over time from the bubble formation to when bubble pressure reached the maximum. The adsorption kinetic curves were constructed by plotting the surface tension versus surface age (from 5 ms to 200 s). From the curves, the lag time was estimated as the time at which the surface tension decreased to 95% of the initial value; the slope of the decrease in surface tension after lag phase was calculated as the adsorption rate.

2.8.2. Drop volume method

Automatic drop volume tensiometer DVT50 (Krüss GmbH Hamburg, Germany) was used to perform dynamic measurement for surface tension. Drops were formed in an empty optical glass cuvette, at the tip of a 2.016 mm-radius capillary. Continuous formation of drops occurred at flow rates from 600 to 1 μL/min with 3–5 drops recorded per decade. From kinetic curves, equilibrium surface tension was estimated by the intercept of the plot of surface tension versus 1/√t (t = time).

2.8.3. Emulsifying properties

Emulsifying properties of native and modified mannoproteins were evaluated by determining the emulsifying activity index (EAI) and emulsifying stability index (ESI) as described by Pearce and Kinsella (1978). Protein samples were prepared at 1 mg protein / mL in 10 mM sodium phosphate buffer at pH 4, 6 and 8. The protein solutions were mixed with soybean oil at a ratio of 1:3 (v/v) and homogenized at 22,000 RPM for 90 s using a Fisherbrand Homogenizer 850 (Fisher Scientific, Pittsburgh, PA). Samples were diluted 200× in 0.1% (w/v) SDS prior to turbidity being measured at 500 nm using Beckman DU 650 spectrophotometer (Beckman Instruments Inc.; San Ramon, CA). Measurements were taken immediately after homogenization (0 min) and again after 15 min. EAI and ESI values were calculated according to the equations:

$$\mathit{EAI}\left(\frac{m^2}{g}\right)=\frac{\left(2\ast 2.303\ast A_0\ast \mathit{DF}\right)}{\left(c\ast \phi \ast {10}^4\right)}$$

$$\mathit{ESI}\left(\mathit{min}\right)=\frac{\left(A_0\ast t\right)}{\left(A_0-A_{15}\right)}$$

where, A₀ and A₁₅ are the absorbances at 0 min and 15 min respectively; DF is the dilution factor of the emulsion, c is the concentration of the aqueous solution (w/v%), φ is the oil volume fraction and t is the time between the first and second turbidity measurements.

Additionally, particle size and polydispersity index (PDI) of the emulsions were measured by dynamic light scattering technique on a Malvern Zetasizer Pro (Malvern Instruments, UK). At t₀, samples were diluted 32× in 10 mM sodium phosphate buffer (adjusted to the same pH as the corresponding emulsion) and analyzed in triplicate for Z-average and PDI.

2.9. Structural characterization

2.9.1. Hydrophobic microenvironment

The microenvironment of the aromatic amino acid residues of YCP samples was analyzed using fluorescence spectroscopy according to Fan et al. (2022). Protein suspensions 0.5 mg/mL were prepared in sodium phosphate buffer at pH 4, 6 and 8. Emission spectra from 290 to 500 nm in increments of 1 nm were recorded on a Cary Eclipse Fluorescence Spectrophotometer (Varian, Palo Alto, CA) with an excitation wavelength at 280 nm.

2.9.2. Surface hydrophobicity

Surface hydrophobicity of native and modified proteins were evaluated using 8-anilinonaphthalene-1-sulfonic acid (ANS) as a fluorescence probe according to the method by De la Cruz-Torres et al. (2022). Samples were prepared in 50 mM sodium phosphate buffer at pH 4, 6 and 8) and distributed on a microplate to obtain protein concentrations ranging from 0 to 0.25 mg/mL. ANS was added to each well to achieve a final concentration of 20 μM. Plates were incubated in the dark for 15 min prior to fluorescence measurement. Fluorescence was recorded using a SpectraMax® i3x plate reader (Molecular Devices, San Jose, CA) at an excitation wavelength of 390 nm and an emission wavelength of 470 nm. Surface hydrophobicity (H₀) was calculated as the initial slope of the fluorescence intensity versus protein concentration obtained through linear regression analysis. Fluorescence data were normalized to the actual protein content of each sample, determined spectrophotometrically at 280 nm.

2.9.3. FTIR

Transmission FTIR was used to evaluate the changes in the modified protein's secondary structure when exposed to varying levels of heat treatment (Boye et al., 1996). Dialyzed and lyophilized samples of modified YCP were dissolved in 90 mM sodium phosphate buffer prepared with D₂O (pH 4, 6, 8) at a concentration of 1 mg/10 μL. Infrared spectra were recorded with a Nicolet iS5 FTIR spectrometer with a DTGS KBr detector. A dry air purge was applied for 15 min before the start of readings. A total of 128 scans were averaged at 4 cm-1 resolution. The samples were held in a transmission cell with a 25 μm path length and calcium fluoride windows. The temperature was controlled by heating the IR cell with an Omega temperature controller (Omega Engineering, Laval, Quebec, Canada H7L 5A1). The temperature was increased in 5 °C increments and the cell allowed to equilibrate for 10 min between each measure before taking readings. A Fourier self deconvolution with a bandwidth of 40 and an enhancement factor of 2.2 was applied to spectra prior to analysis.

3. Results and discussion

3.1. Compositional properties of native YCP

The compositional properties of YCP were investigated (Table 1). Typical yeast extracts consist of a mix of amino acids, vitamins, minerals, nucleic acids, structural carbohydrates, and other water soluble components (Tao et al., 2023). The intact yeast extract used in the present study consisted of 67.42% protein, with 15.59% carbohydrates. The carbohydrate fraction further consisted of 14.44% mannose and 1.15% glucose. The protein content was high compared to other references, which generally showed <65% protein content (Tao et al., 2023; Vieira et al., 2016). The carbohydrates originated from residual fragments of the yeast cell wall consisted of mannan and glucan residues. After dialysis, the total phenolics content was reduced to 0.18 GAE/100 g. The molecular weight distribution covered a broad range varying between 0.13 and 152 kDa. The high molecular weight protein content of the yeast extract is atypical, as most yeast extracts are hydrolyzed to the amino acid level. In fact, intact yeast cells contain a high prevalence of hydrolytic enzymes, which are activated during processing to cause the autolysis of the yeast cell and the degradation of cellular proteins to amino acids. The yeast extract contained these active endogenous proteases, which interfered with initial trials of protein crosslinking using laccase enzyme. Therefore, additional pre-treatments were investigated for inhibiting the endogenous proteases before crosslinking could proceed.

Table 1.

Compositional profile of native YCP.

Compositional characterization		Molecular weight distribution (%)
Protein proportion (% w/w)	67.42 ± 1.09	0–25 kDa	15.45
Carbohydrate proportion (% w/w)	15.59 ± 1.94	25–50 kDa	7.17
Relative mannose content (% w/w)	14.44 ± 6.53	50–100 kDa	8.05
Relative glucose content (% w/w)	1.15 ± 0.01	100–150 kDa	0.00
Total phenolics (GAE/100 g)	0.18 ± 0.003	+150 kDa	69.33

Heat treatments were applied as pre-treatments for 5 min at temperatures between 60 °C and 80 °C, and the resulting protein samples were analyzed immediately (0 h) after the treatment by SDS-PAGE and after 24 h of sitting at room temperature (Supplementary Fig. 1). The results (Fig. 1) show that heat treatments were ineffective at reducing protease activity up to 75 °C. The reduction of total intensity on the SDS-PAGE gels after 24 h of incubation was used to calculate the extent of protein hydrolysis, which was 43.72%, 42.88% and 30.07% at 60, 65 and 70 °C respectively. At 75 °C, the protease was thermally inactivated (2.18% hydrolysis); however, the YCP also seemed to be thermally denatured at this temperature forming a turbid suspension with insoluble precipitate (not shown).

Fig. 1.

Extent of hydrolysis of YCP at 24 h by the endogenous native protease activity. Inhibition was applied thorough heat treatments from 60 °C – 80 °C and PPI addition from 1 to 5%.

The use of a natural, food compatible protease inhibitors sourced from potato was assessed as an alternative strategy to heat treatment for protease deactivation. PPI are a low molecular weight fraction of proteins extracted from potato and consist of inhibitor I, inhibitor II, cysteine inhibitor, type II protease inhibitor, as well as several other peptidase inhibitors (Hu et al., 2024). There are seven well characterized proteases identified in the vacuole structure of the saccharomyces cerevisiae yeast cell, including three metalloproteases, three serine proteases and one aspartyl protease (Hecht et al., 2014).

Fig. 1 indicates that the addition of PPI at 5% and 2.5% (w/w) relative to the amount of YCP resulted in an inhibition of the protease activity up to 24 h. Furthermore, the SDS-PAGE gels showed no signs of protein denaturation, as the bands from the control sample remained intact in the samples treated with PPI (Supplementary Fig. 1). Based on these observations, 2.5% PPI was incorporated into subsequent YCP samples to prevent hydrolysis during enzyme crosslinking. The selection of PPI as hydrolysis inhibitors was also advantageous due to their known effects on laccase action. M. Li, S. Karboune, K. Light, et al. (2021) investigated laccase-catalysed crosslinking of patatin and PPI and demonstrated that PPI does not hinder laccase catalysis. Indeed, PPI acted as a substrate for laccase-catalysed oxidative crosslinking as evidenced by an increased molecular weight profile after laccase treatment. In the present study, PPI was added at 2.5% relative to the mass of YCP to limit endogenous hydrolysis; at this low level, the effects of PPI on laccase action and its overall impact on the YCP structure were considered not significant.

3.2. YCP crosslinking catalysed by laccases

The oxidative crosslinking of YCP was investigated with catalysis applied by two laccase enzymes (LacTv, LacCh) from different fungal origin, and in the presence and absence of FA as a phenolic mediator. 2.5% PPI was added to prevent protein hydrolysis during crosslinking. Laccases from different fungal origins can express different affinities towards substrates due to structural differences in the enzymes that influence the accessibility of substrate moieties to the catalytic site (M. Li, L. Liu, et al., 2021). Additionally, laccases have the potential to catalyse both protein fragmentation and protein crosslinking depending on the reaction conditions (Lantto et al., 2005). The selection of an appropriate enzyme and reaction conditions is, therefore, critical to favor covalent protein crosslink formation, which can lead to significant improvements in protein technofunctional properties, such as enhanced stabilization of oil–water interfaces and improved performance as encapsulation materials (Isaschar-Ovdat & Fishman, 2018). The experimental findings show that without FA, little to no change in the molecular weight profile of YCP was observed by SDS-PAGE at all reaction times and enzyme unit levels (Supplementary Fig. 2). This is attributed to limited access of laccase to oxidizable targets within the compact YCP structure, likely due to steric hindrance. The conformation of YCP restricts the enzyme access to binding sites and/or presents a low proportion of exposed tyrosine residues available for oxidative crosslinking (M. Li, S. Karboune, L. Liu, et al., 2021). Indeed, some hydrolysis was seen in the presence of LacTv at 6 h and LacCh at 3 h and 6 h.

With the addition of FA, the end-product profiles were noticeably altered over the reaction time-course. The LacTv treatment at 0.1 U/mL and 0.2 U/mL in the presence of FA led to YCP crosslinking as shown by the darker smear present above 50 kDa at 3 h and 6 h. For the LacCh treatments, a YCP crosslinking was seen at 0.1 U/mL after 6 h treatment; however, protein hydrolysis was detected at this condition as indicated by the overall decrease in the lanes' intensity. To compare the different treatments, the extent of hydrolysis and crosslinking were estimated (Fig. 2). Enhanced crosslinking with FA supplementation is attributed to the ability of phenolic acids to act as electron shuttles towards regions of the protein structure which are inaccessible to the larger enzyme molecules. Laccase crosslinks proteins using a two-step reaction. The first step is catalysed by laccase and involves the oxidation of phenolic moieties to phenol radicals, such as tyrosyl or feruloyl radicals. The intermediate radicals then undergo non-enzymatic polymerization to form crosslinked structures (Li et al., 2020). The addition of ferulic acid to the reaction medium increases the number of oxidation targets for laccase and allows feruloyl radicals to travel to inaccessible areas of YCP structure to enhance the extent of the crosslinking reaction.

Fig. 2.

Crosslinking and hydrolysis extent of YCP in the presence of 2.5% PPI catalysed by LacTv or LacCh upon 3 and 6 h reaction. *Extent of hydrolysis: total intensity of the bands in a given lane compared to the total intensity of the same treatment at 0 h. **Extent of crosslinking: relative intensity of the bands between 50 and 150 kDa compared to the intensity of the same timepoint with 0 U/mL of enzyme.

It can be clearly seen in Fig. 2 that the LacCh enzyme resulted in a higher hydrolysis extent for most of the treatments, particularly the ones including FA. In terms of crosslinking extent, the LacCh treatments in the presence of FA led to oxidative crosslinking only at 0.2 U/mL and 6 h. On the other hand, the LacTv treatments resulted in a higher crosslinking extent at both 0.1 U/mL and 0.2 U/mL after 3 h. Overall, LacTv with FA treatment showed a higher level of non-hydrolyzed protein at both 3 h and 6 h compared to LacCh with FA, which exhibited decreased protein content over the same periods due to simultaneous protein hydrolysis and crosslinking. LacTv also demonstrated a higher crosslinking extent at 3 h than LacCh. As a result, LacTv was selected as the preferred enzyme for subsequent reactions in the present work. Longer reaction time courses are also necessary to determine the crosslinking profile past 6 h.

3.3. Reaction kinetics of LacTv-catalysed crosslinking of YCP

The oxidative crosslinking of proteins by laccase enzyme begins with modifications of the tryptophan, tyrosine and cysteine residues, with tyrosine being the most reactive. Following these initial early stage reactions, further radical-based modifications can occur on methionine and histidine (M. Li, S. Karboune, K. Light, et al., 2021). Since these early-stage reactions involve aromatic amino acid residues, reaction kinetics can be evaluated by tracking the decrease in fluorescence emission from these aromatic groups. The oxidative crosslinking reactions of YCP were performed at substrate concentrations varying from 0.25 mg/mL to 3 mg/mL and monitored over 30 min initial reaction rate. The specific activity of laccase catalysed oxidation of YCP was calculated and expressed as moles of tyrosine equivalent plotted against substrate concentration, which can be found in Fig. 3.

Fig. 3.

Kinetic plot of LacTv catalysing the oxidative crosslinking reaction of YCP according to Hill equation.

At lower substrate concentrations, the specific activity was low, and at higher concentrations it increased, as expected. At very high substrate concentration (>2 mg/mL) the increase became limited, which indicates substrate saturation or substrate / product inhibition. These plots were then fitted to the Hill model to assess whether the reaction followed Hill or Michaelis–Menten kinetics, which are equivalent when the Hill coefficient equals 1. The Hill coefficient of 1.47 obtained for the oxidative crosslinking reaction of YCP by LacTv indicates weak positive cooperativity among binding sites rather than strictly non-cooperative behavior (Table 2). This is in line with a study conducted by Frasconi et al. (2010), where the kinetics of multiple laccases were evaluated with various small substrates (ABTS, syringaldazine, catechol, dopamine, FMCA and FCN) and the Hill coefficient was found to be in the range of 0.94–1.14 which indicates that the enzyme follows Michaelis Menten kinetics. The substrate binding affinity (K_m) of this laccase to YCP was lower than that found between LacTv and potato protein by M. Li, S. Karboune, K. Light, et al. (2021) (0.76 for yeast protein vs 0.507 and 0.594 for potato protein and PPI). This may be due to stearic interference from the larger sized yeast protein (>100 kDa vs. 40 kDa), or possibly lower availability of amino acid targets (tryptophan, tyrosine, cysteine) on the exterior of the proteins. Based on these findings, subsequent YCP laccase reactions were carried out at 2K_m substrate concentration (1.24 mg/mL).

Table 2.

Hill equation parameters of YCP reacted with LacTv.

	Hill model
	Value	t	p
Km (mM)	0.62 ± 0.16	3.94	0.0023
Vmax (umol/min*mg enzyme)	0.32 ± 0.05	6.72	<0.001
n	1.47 ± 0.44	3.33	0.0068
R2	0.8988

3.4. Time courses for LacTv-catalysed crosslinking of YCP

Fig. 4 shows the product profiles of the YCP modified by LacTv with and without FA. A similar trend can be seen between the reactions without FA and those with FA as seen in Fig. 2 where analysis was done by SDS-PAGE gels. In the reaction time course without FA, no increase in the proportion of high molecular weight protein in the 100–150+ kDa region was observed; there is a decrease in the total amount of protein in solution over time, particularly in the reactions with 0.1 U/mL and 0.2 U/mL of enzyme added. This may result from incomplete deactivation of the native protease activity of YCP, as discussed in section 3.1, or from the enzymatic fragmentation of higher molecular weight protein by laccase (M. Li, S. Karboune, K. Light, et al., 2021).

Fig. 4.

End product molecular weight profile of YCP reacted with LacTv up to 24 h with and without the addition of FA.

In reactions containing FA, significant modifications in the molecular weight profile of modified YCP across all ranges were obtained. At 3 h, an increase in the 25–50 kDa region for both levels of enzyme treatment was noted. The reaction corresponding to 6 h with 0.2 U/mL showed a shift towards the 25–50 and 50–100 kDa regions, and at 18 h and 24 h, the 100–150 kDa and 150+ kDa regions are more prominent. Furthermore, protein hydrolysis, which is evident in treatments without FA, is limited in the reactions containing FA, particularly in the 18 h and 24 h treatments. While a small amount of hydrolysis is seen at 3 h and 6 h timepoints, it is no longer apparent at 18 h and 24 h. The initial formation of smaller peptide fragments may have facilitated an increased level of crosslinking in the presence of FA.

Select oxidative crosslinking reaction conditions were chosen for producing modified YCP and for further analysis. The 0.2 U/mL-24 h reaction condition was chosen to produce highly crosslinked YCP with a high proportion of 150+ kDa protein residues, while 0.2 U/mL-18 h reaction condition led to a moderate level of crosslinking of YCP with a lower proportion of 150+ kDa and higher proportion of 100–150 kDa protein residues. The 0.2 U/mL-3 h reaction condition, with a high proportion of 25–50 kDa residues, was selected to represent low crosslinking extent of YCP. Low, moderate and high crosslinked YCP samples were evaluated for their modified interfacial properties and altered structural features.

3.5. Structural characterization of oxidative protein modifications

Laccase enzyme forms crosslinks in protein through a variety of mechanisms, including: (1) creation of dityrosine and isodityrosine bonds, (2) opening and closing of disulfide bridges, which allows the insertion of peptide fragments, and (3) polymerization via an incorporation of phenolic radical moieties into the structure of the protein (M. Li, S. Karboune, L. Liu, et al., 2021; Steffensen et al., 2008; Steffensen et al., 2009). In current study, all the three mechanisms were investigated using LC-MS/MS to evaluate their contribution to YCP crosslinking.

Prevalence of dityrosine linkages were evaluated by LC-MS/MS after acidic digestion of the modified and non-modified samples. Several previous works have confirmed the stability of dityrosine during the digestion step with 90–95% recovery (Fenaille et al., 2004; Hanft & Koehler, 2005). In this work, dityrosine was detected in constantly higher abundance in the modified YCP without the addition of FA. In the no FA-modified YCP samples at three different reaction times (3, 18 and 24 h), the highest abundance of dityrosine linkages were detected at the 18-h timepoint (Fig. 5 A). For the FA- modified YCP samples, the 18 h FA sample showed a significant increase in dityrosine compared to the native YCP sample, but both the FA 3 h and FA 24 h treatments were not different from the native treatment.

Fig. 5.

Oxidized reaction products detected per mg of protein in native and modified YCP samples (a) dityrosine linkages (b) diferulic acid linkages (c) total carbonyl groups present.

Based on the chromatogram, three distinct peaks showed signal for dityrosine (C₁₈H₂₀N₂O₆ m/z 361.1394 in positive ion mode) with retention times 3.5 3.6 and 4 min (Supplementary Fig. 3 B). Both the retention time and the MS/MS fragment pattern of the peak at 3.6 min matched with the dityrosine standard (Fig. 6 A & C). The MS/MS fragmentation of the peak eluted at 4 min showed similar fragment ions with dityrosine standard but with different abundance pattern (Fig. 6 D). Figueroa et al. (2020) has reported that isodityrosine eluted later than dityrosine on C18 column and with similar fragment ions with dityrosine standard. The peak (m/z 361.1394) at 3.4 min contains the fragmentation pattern partially different from that of the dityrosine and isodityrosine. It does contain the m/z 361 ➔ m/z 315 ion transition, while the other two transitions characteristic of dityrosine are reduced by 2 mass units in each case (m/z 361 ➔ m/z 252 and m/z 361 ➔ m/z 235) (Fig. 6 B). This represents the additional loss of two hydrogen atoms in certain derivative ion fragments, which were well described by Fenaille et al. (2004), but the structural differences cannot be elucidated based on current data. Due to the complex nature of radical-induced laccase crosslinking, there may be a third dityrosine isomeric structure present in the reaction samples.

Fig. 6.

Mass spectra of digested dityrosine and diferulic acid linkages eluted at different retention times (a) dityrosine standard eluted at 3.6 min (b) unidentified dityrosine isomer eluted at 3.5 min (c) dityrosine from YCP sample eluted at 3.6 min (d) isodityrosine eluted at 4.0 min (e) 8–5'noncyclic dehydro-diferulic acid eluted at 6.4 min.

The type of bond that laccase creates between tyrosine residues seems to be specific both to the type of substrate and the subgroup of laccase employed. In work investigating the linkage formed between tyrosine containing peptides, Mattinen et al. (2005) concluded that the majority of the tyrosine bonds were present in isodityrosine form (C-O-C bond detected by FTIR). Conversely, Y. Li et al. (2021) found that laccase crosslinking in BSA predominantly forms dityrosine bonds (no evidence of ester bonds in FTIR analysis). In the present study, both dityrosine and isodityrosine bonds were detected in the mass spectra of analyzed proteins. LC-MS/MS methods are preferable over fluorescence methods for the detection quantitation of dityrosine and isodityrosine. The more widely used fluorescence methods will not detect the presence of isodityrosine due to its lack of fluorescence, and measuring dityrosine levels accurately can be confounded by interference of structurally distinct compounds which also emit fluorescence (Fenaille et al., 2004).

Upon introducing FA as a phenolic moiety into the laccase enzyme and the YCP system, signals corresponding to FA linkages were detected in the hydrolyzed protein samples, while they were absent in samples with no FA added (Fig. 5 B). There were no significant differences for the peak intensities of FA linkages across the time course (from 3 h to 24 h), indicating all FA crosslinks on proteins likely formed at early stage of the enzymatic reaction. The only type of FA linkage detected in the FA-modified YCP hydrolysate was dehydro-diferulic acid (C₂₀H₁₈O₈) (m/z 387.1074 in positive and 385.0929 in negative ion mode) with one single dominant dehydro-diferulic acid signal at retention time 6.45 min (Supplemental Fig. 3C).

MS/MS fragmentation of the FA linkage peak (m/z 387.1074) at 6.45 min showed a presence of a major fragment ion with m/z 309.0757 in positive mode (Fig. 6 E), which has been reported as a diagnostic fragment ion for 8–5'noncyclic dehydro-diferulic acid (Vismeh et al., 2013). Previous studies have identified the major product of laccase mediated crosslinking as 8–5′ dehydrobenzofuran-diferulic acid, with low prevalence of other forms such as 8–8′, 8–5′ (noncyclic) dehydro-diferulic acid (Ward et al., 2001). A recent study by our research group also found that the major product of FA mediated laccase crosslinking in ovalbumin and lysozyme was 8–5'noncyclic dehydro-diferulic acid (M. Li, S. Karboune, L. Liu, et al., 2021). The extracted ion chromatogram with formula C₂₀H₁₈O₈ also shows minor peaks at retention times of 5.73, 5.97, 6.99 and 8.20 min in positive ion mode (Supplemental Fig. 3C), which could be attributed to other forms of dehydro-diferulic linkage, since it is well known that FA can polymerize into a variety of isotope structures (Vismeh et al., 2013; Ward et al., 2001).

Of notable interest are the significantly lower levels of crosslinked tyrosine found in the YCP samples reacted with FA. LacTv has a substantially higher affinity for FA as a substrate over tyrosine. In the work reported by Mattinen et al. (2005), the reaction rate of laccase with FA was found to be 10× higher than the reaction with tyrosine residues. The presence of FA in the reaction resulted in competitive binding of the substrate to the enzyme and lesser conversion of tyrosine residues (M. Li, L. Liu, et al., 2021; Steffensen et al., 2008). FA-mediated crosslinking induced polymerization between proteins and favoured generation of intermolecular bonds as supported by the generation of high molecular weight products with this treatment condition (Fig. 4). Conversely, non—FA reacted samples generated dityrosine linkages which created intramolecular bonding within the protein structure and without any increase in protein size. A schematic of the proposed dityrosine and diferulic acid linkage pathways is presented in Fig. 7. Although laccase crosslinking also affects disulfide bonds, analysis by Ellman's assay did not find any difference in the free thiol groups in native or crosslinked YCP (data not shown). It is, however, possible that the reduction by FA and oxidation by laccase-generated radicals had some effects in opening and closing these bonds for polymerization during the reaction time course.

Fig. 7.

Proposed pathway for laccase-mediated crosslinking of yeast cytoplasmic protein in the presence and absence of ferulic acid acting as an electron shuttle.

Laccase can also induce non-crosslinking oxidative modifications on certain amino acid residues. Amino acids susceptible to oxidative modification, without leading to covalent crosslink formation, include methionine, tryptophan, phenylalanine, histidine, proline, lysine, and arginine (Heinonen et al., 2021). Laccase-catalysed oxidative crosslinking has been reported to convert tryptophan to kynurenine and N′-formylkynurenine, methionine to methionine sulfoxide, and histidine to 2-oxo-histidine; however, oxidative modifications of other amino acids remain poorly characterized (Steffensen et al., 2008). Given that yeast proteins are enriched in oxidizable residues beyond tryptophan, methionine, and histidine (Cao et al., 2025; Lee et al., 2024), the overall extent of protein oxidation was therefore assessed by quantifying total carbonyl formation across all amino acids using spectrophotometric DNPH derivatization. In the case of YCP, there were no significant differences detected in the total level of carbonyls between native and modified proteins (Fig. 5 C). This is in agreement with the results of Steffensen et al. (2008) where no significant differences were found in amino acid carbonyls upon reaction with laccase and FA in either BSA or α-casein, however β-lactoglobulin showed carbonyl formation on tryptophan, methionine and histidine residues in the absence of FA. In the case of β-lactoglobulin alternate carbonyls were formed in the case where tyrosine and FA substrates were not available, which indicates competition for the enzyme between different available substrates. In the present YCP samples reactions on tyrosine and FA occurred preferentially, resulting in only low levels of carbonyl formation.

3.6. Interfacial properties of native and modified YCP

Structural modification of proteins has a direct impact on their technofunctional properties. The adsorption kinetics of native and modified YCP at the air/water interface were investigated. The maximum bubble pressure method is the fastest quantitative method for measuring the short-term adsorption kinetics of proteins or other surfactants. In fact, surface tensions can be reproducibly measured at times below 0.001 s, facilitating the measurement of the lag time and initial adsorption rates of proteins diffusing to the air/water interface. Lag time represents the time delay while macromolecules are adsorbing to the interface in sufficient quantities for their interactions to become significant. Lag time has been attributed to both the flexibility of macromolecules and their ability to undergo conformational changes (Pizones Ruíz-Henestrosa et al., 2007). The initial adsorption rate occurs directly after the lag time has been completed and represents the rate of protein diffusion towards the interface. Equilibrium surface tension was evaluated by drop volume method over longer time periods than the first two parameters, at times up to 25 min. Equilibrium surface tension occurs once proteins have saturated the interface and re-oriented themselves into a more stable conformation due to interface penetration and protein unfolding (Pizones Ruíz-Henestrosa et al., 2007).

Interface properties have been measured for native YCP with and without the addition of PPI. Small proteins can travel to the interface more quickly than their larger counterparts and have significant effects on the interfacial parameters. In fact, in all three interfacial parameters, there is a difference between YCP alone and YCP + 2.5% PPI (Table 3) (shorter lag time, faster initial adsorption rate and lower equilibrium surface tension). For the modified proteins, PPI has been incorporated into the bulk protein structure by crosslinking, so it is difficult to evaluate the contribution of this PPI component alone. However, the relative differences provide the contribution of the oxidative crosslinking modification to the interfacial properties of YCP.

Table 3.

Interface properties of native and modified proteins at pH 6 as evaluated by bubble pressure and drop volume methods.

	Lag time (s)	Initial adsorption rate	Equilibrium surface tension
		(mN/(m ms))	(mN/m)
YCP	18.33a ± 1.16	0.161a ± 0.007	63.22a ± 0.55
YCP + PPI	13.14b ± 0.58	0.201b ± 0.007	51.87b ± 1.73
YCP with FA 3 h	34.65c ± 1.47	0.103c ± 0.007	55.79c ± 1.72
YCP with FA 18 h	49.83d ± 1.00	0.068d ± 0.008	55.77c ± 0.60
YCP with FA 24 h	24.85e ± 1.56	0.111c ± 0.012	53.79bc ± 0.25
YCP no FA 3 h	16.94a ± 1.16	0.246e ± 0.006	31.92d ± 0.57
YCP no FA 18 h	13.28b ± 0.31	0.322f ± 0.021	30.07d ± 0.42
YCP no FA 24 h	19.10a ± 0.29	0.253e ± 0.015	30.73d ± 0.20

There are marked differences in interfacial properties for YCP modified with or without FA, as well as for the different treatment times (low, moderate, high crosslinking). Overall, YCP modified with FA had longer lag times, slower initial adsorption rates and higher equilibrium surface tension than both native YCP and YCP modified without FA. On the other hand, YCP modified without FA has shorter lag times, faster initial adsorption rate and lower equilibrium surface tension than YCP native (Table 3). These differences can be deduced from the structural analysis of the crosslinked proteins. YCP modified with FA exhibited a larger molecular weight profile than those modified without FA, due to the formation of more intermolecular bonds facilitated by FA, resulting in a greater proportion of proteins >150 kDa, which may have delayed the adsorption to the interface. The extent of crosslinking also has significant effects between the samples modified with FA. The 18 h FA sample with moderate crosslinking level had the longest lag time (49.83 s) and slowest initial adsorption rate (0.068 mN/(m ms)), followed by the FA 3 h sample corresponding to low crosslinking level (34.65 s and 0.103 mN/(m ms)); and surprisingly, the 24 h FA sample with high crosslinking level was the fastest in both parameters (24.85 s and 0.111 mN/(m ms)). There were no significant differences in the initial adsorption rates between the 3 h FA and 24 h FA samples, and no significant differences in the equilibrium surface tension for any of the three FA modified samples (53.79–55.79 mN/m).

The samples modified without FA had lag times which were statistically equivalent to the native YCP and the native YCP + PPI (16.94, 13.28 and 19.10 s for 3 h, 18 h and 24 h respectively). Conversely, these samples had initial adsorption rates which were significantly faster than the native protein and equilibrium surface tensions which were significantly lower. These findings relate back to the modified structural elements discussed in sections 3.4 and 3.5. The samples modified without FA had decreased molecular weight due to protein hydrolysis or fragmentation over the course of the reaction, as well as structural re-arrangement due to intramolecular dityrosine bonding. Referring to Fig. 4, all three no FA modified proteins have a lesser proportion of their molecular weight in the >100 kDa range, with the size ranking being 3 h > 24 h > 18 h. The 18 h treatment with the lowest molecular weight is also the treatment with the lowest lag time and fastest initial adsorption rate due to its ability to diffuse to the interface more quickly than the other larger proteins. The equilibrium surface tension measures are statistically equivalent between themselves, and also significantly lower than YCP or YCP + PPI. Indeed, equilibrium surface tension is dominated by a protein's ability to interact with both the hydrophilic and hydrophobic sides of the interface due to exposed hydrophobic residues and tertiary structural features. Laccase crosslinking and the formation of dityrosine bonds has induced conformational changes in the non-FA modified YCP which increase its ability to interact at the interface. Overall, YCP modified without FA (particularly the 18 h treatment) shows high interfacial activity and potential for stabilizing the air/water interface. This is due to the small molecular weight, compact protein structure, and structural re-arrangement during crosslinking. Compared with other laccase-modified plant proteins, such as potato protein, YCP exhibited enhanced interfacial properties, characterized by shorter lag times and lower equilibrium surface tension for the YCP-no FA treatment (M. Li, C. Blecker, et al., 2021). However, YCP lag times remain longer, and initial adsorption rates are lower than those of animal-derived protein emulsifiers such as β-casein (Miller et al., 2004).

These modified interfacial properties may offer several advantages for food applications. Both laccase treatments resulted in modified protein with lower equilibrium surface tension than native YCP. Lowering of interfacial surface tension has been correlated with improvement of emulsion characteristics such as decreasing the emulsion droplet size, increasing emulsion droplet dispersibility and increasing emulsion stability (An & Zheng, 2025). Yeast proteins modified by laccase-catalysed crosslinking could potentially be used for the formation of highly stable emulsions. Additionally, the inclusion of ferulic acid into the protein structure has been shown to improve the antioxidant activity of the protein (M. Li, S. Karboune, K. Light, et al., 2021). Using these modified proteins as encapsulation materials for emulsions with bioactive ingredients such as omega-3 oils may provide added protection against oxidative degradation of the encapsulated ingredient.

3.7. Hydrophobic microenvironment and surface hydrophobicity of YCP

Fluorescence spectroscopy was used to investigate changes in the hydrophobic microenvironment of aromatic amino acids embedded in the structures of the native and modified YCP. In general, it is considered that a decrease in the maximum emission intensity accompanied by a shift in λ_max towards a higher wavelength (red shift) signifies the exposure of aromatic amino acids such as tryptophan to the solvated hydrophilic environment surrounding the protein. Fig. 8 A-F shows the emission spectra of native YCP and YCP modified with and without FA at varying pH levels. The λ_max at pH 4, 6 and 8 are 350 nm, 352 nm and 337.03 nm respectively, which match closely with those reported by Wang et al. (2025) for unmodified yeast protein at pH 7. For Native YCP, as the pH decreases so does the fluorescence intensity, indicating protein unfolding and exposure of the hydrophobic tryptophan residues at lower pH levels. This is further supported by the surface hydrophobicity measurements conducted using ANS fluorescence probe (Fig. 8 G) which showed that native YCP exhibited increased surface hydrophobicity at pH 4. YCP reacted without FA (Fig. 8 A-C) showed decreased fluorescence intensity as reaction time increased, indicating increased exposure of hydrophobic residues as the reaction progressed. However, this trend was not supported by the surface hydrophobicity measurements (Fig. 8 G) where the reaction time (3, 18 or 24 h) had no significant effect on surface hydrophobicity. The observed decrease in intrinsic fluorescence intensity is therefore more plausibly attributed to laccase-induced oxidative modifications of tyrosine and tryptophan residues, which are primarily responsible for the intrinsic fluorescence of proteins (Fan et al., 2022).

Fig. 8.

Intrinsic fluorescence spectra of native and modified YCP at varying pH levels A) No FA 3 h treatment B) No FA 18 h treatment C) No FA 24 h treatment D) FA 3 h treatment E) FA 18 h treatment F) FA 24 h treatment. G) Surface hydrophobicity of native and modified YCP at varying pH levels.

Interpretation of intrinsic fluorescence measurements for YCP in the presence of FA (Fig. 8D-F) is further complicated by the intrinsic fluorescence and quenching effects of FA. Native YCP supplemented with free FA exhibited a marked suppression of fluorescence intensity accompanied by a red shift, consistent with previous reports in which the addition of phenolic compounds such as rutin or chlorogenic acid produced similar spectral shifts in a dose-dependent manner (Jia et al., 2019). Phenolic acids such as FA can result in a fluorescence quenching effect by several mechanisms. Fluorescence quenching can occur when the absorption spectrum of the phenolic acid overlaps with the emission spectrum of tyrosine (330–350 nm), thereby masking emitted light. It can also occur when the absorption spectrum of the phenolic acid overlaps with the excitation wavelength of tryptophan (295–305 nm), thereby limiting the incoming light and the excitation of the fluorophore (Rawel et al., 2006). It has also been shown that collisions between excited fluorophores and fluorescence quenchers can result in energy transfer and quenching of the fluorescence signal. For native YCP with addition of FA, it is likely that all three of these mechanisms are at play. FA in these samples is not bound, and it is free to travel to the polar microenvironment around the tryptophan resides to provide shielding effects. The enhanced fluorescence suppression at pH 4 is similar to effects seen in a study by Rawel et al. (2006) where FA had a greater affinity for BSA at pH 4.8 compared to pH 7, providing enhanced fluorescence suppression at this pH level.

YCP modified with FA shows differences in fluorescence suppression at the different pH levels. In these samples, the FA has been bound into the protein's structure by laccase crosslinking in the form of 8–5'noncyclic dehydro-diferulic acid with no quantitative differences between the reaction times (section 3.5). The highest fluorescence intensity is seen at pH 4, followed by pH 6 and then 8. This quenching effect could be due to two phenomena: (1) unfolding and exposure of the hydrophobic residues as pH increases, or (2) increased exposure of FA residues to the tryptophan microenvironment as pH increases. The surface hydrophobicity measurements shown in Fig. 8 G support the hypothesis that the observed decreases in fluorescence intensity at pH 6 and 8 are at least partially due to increased exposure of hydrophobic residues, as the surface hydrophobicity increases significantly at both pH levels.

The surface hydrophobicity of FA and non-FA modified YCP increased significantly as pH increased from 4 to 8. This trend is opposite to what is usually reported in the literature, where surface hydrophobicity of proteins decreases at increased pH levels due to ionization of amino acids and their orientation towards the protein exterior (Alizadeh-Pasdar & Li-Chan, 2000). Covalent crosslinking within the protein structure may have reduced molecular flexibility, thereby limiting structural reorientation and promoting exposure of hydrophobic regions.

The pronounced increase in surface hydrophobicity at pH 8 is consistent with, and likely contributes to, the significantly higher emulsion activity index observed for FA-modified YCP under these conditions (Supplementary Fig. 4). At pH 6 and 8, FA-modified YCP exhibited significantly lower surface hydrophobicity than YCP modified in the absence of FA. This difference may be attributed to enhanced protein aggregation in FA-modified YCP at higher pH or to covalent binding of FA to surface-exposed tyrosine residues, thereby reducing their apparent hydrophobicity. These observations are consistent with the findings reported in Section 3.6, where non-FA-modified YCP displayed lower equilibrium surface tension compared with both native YCP and FA-modified YCP (Table 3). Lower surface tension is indicative of greater exposure of hydrophobic residues, which facilitates stronger interactions with the hydrophobic air phase, enhanced interfacial penetration, and a reduction in the overall tension between the two phases.

3.8. Secondary structural characterization of native and modified YCP

FTIR analysis of the native and crosslinked YCP was carried out to understand the effect of crosslinking on the secondary structural characteristics. As YCP is a mixed protein, FTIR spectra describe the relative changes in the average structure of all proteins derived from the yeast cytoplasm. The structural changes with crosslinking were most evident in the amide I region (1600–1690 cm⁻¹), which arises mainly from C O stretching vibration of the amide group (about 80% contribution), with a minor contribution from the C—N stretching vibrations (Carbonaro & Nucara, 2010). On the other hand, the amide II band (1480–1575 cm⁻¹) arises from both N—H bending (60%) and C—N stretching (40%) vibrations. Protein secondary structures involve extensive hydrogen bonding between C O and N—H groups, which is reflected in the spectral characteristics of the amide I and II bands. Based on the degree of hydrogen bonding between structural features, the spectral absorption maxima occur at different wavelengths and can therefore be assigned to secondary structures.

The native YCP shows distinct absorption maxima at 1652 cm⁻¹, 1638 cm⁻¹ and 1615 cm⁻¹, which can be assigned to α-helix, β-sheet and aggregated strands, respectively. With all crosslinking conditions, the definition of the α-helix and β-sheet bands decreases, while the band attributed to unordered structures (1645 cm⁻¹) increases (Fig. 9 C). This can be interpreted as the enzyme interfering with the hydrogen bonding structure of the native protein by forming new covalent crosslinks. The aggregation band in all modified samples also increases. This can be explained by the application of oxidative crosslinking, which has been shown to drive aggregation through several mechanisms. Laccase treatment has the potential to both reduce and re-oxidize disulfide bonds, allowing them to increase in prevalence, change position and incorporate other molecular chains within the disulfide bond (Steffensen et al., 2009). Formation of inter and intramolecular dityrosine bonds, as seen in our modified YCP samples have also been shown to increase protein aggregation (Fuentes-Lemus et al., 2018). Although aggregation increases at every timepoint, it does not seem to become greater with increasing length of treatment and crosslinking extent. For instance, the with FA 18 h treatment has a greater intensity of aggregation band compared to the 24 h treatment. On the other hand, the samples reacted without FA all seem to have greater aggregation than those reacted with FA. A noteworthy observation is that for the aggregation band, the samples reacted with FA had a noticeable shift in the aggregation band to a lower wavenumber. A similar phenomenon was noted by Tang et al. (2022) in their work crosslinking collagen films using laccase and phenolic acids. The shift of the amide I, II and III bands that they observed in their work was attributed to quinones formed from the phenolic acids forming crosslinks with the collagen matrix. The shift in the aggregation band seen here might be due to a similar phenomenon.

Fig. 9.

FTIR spectra of native and modified YCP (a) Native YCP from 25 °C to 80 °C (b) YCP crosslinked with LacTv and FA for 24 h from 25 °C to 80 °C (c) Native and modified YCP (all crosslinking conditions) at 25 °C (d) progression of secondary structural features of native YCP and YCP modified with LacTv and FA for 24 h from 25 °C to 80 °C.

For each of the YCP samples, FTIR measures were completed at temperatures increasing from 25 °C – 80 °C to evaluate the effect of crosslinking on robustness of secondary structural features. The native protein underwent marked changes to its secondary structural characteristics upon heating (Fig. 9 A). Between 40 °C and 45 °C the α-helix and β-sheet bands began to decrease while the aggregation band increases. Contrary to crosslink induced aggregation, this observed increase is due to thermal denaturation. During this type of aggregation, unfolded protein chains come within very close alignment with those on neighbouring proteins and form strong hydrogen bonds, thus leading to intermolecular aggregates. As the temperature increases this trend continues, with complete erasure of the α-helix and β-sheet bands at 80 °C, being replaced with weak unordered structures and aggregation. The trend is different for the modified YCP (Fig. 9 B). The heated spectra for YCP modified with FA for 24 h begin showing structural changes between 45 °C and 50 °C. Although the aggregation band at 25 °C is stronger than for non-modified proteins, the magnitude of its increase is much less than for the native protein. Similarly, although at 25 °C the modified protein starts with less well defined α-helix and β-sheet bands, and a greater magnitude of unordered structures, these structures are preserved much better upon heating to 80 °C than the corresponding structures in the native protein. This shows that highly crosslinked proteins are more resistant to heat denaturation and maintain their secondary structural features much better than native YCP, possibly due to some structural reinforcement with covalent crosslinks.

4. Conclusion

YCP was successfully crosslinked using LacTv with and without the addition of FA as a phenolic mediator. Potato protease inhibitors were very effective at limiting protein hydrolysis during the 24 h crosslinking period when added at 2.5% relative to YCP. The addition of FA in the oxidative crosslinking reaction significantly enhanced YCP crosslinking, resulting in a higher proportion of proteins over 150 kDa, particularly at the 24 h reaction time; while YCP samples without FA underwent fragmentation. Crosslinking level was well controlled by varying enzyme units and reaction times, and samples were generated with low, medium and high levels of crosslinking with 0.2 U/mL of LacTv for 3, 18 and 24 h respectively.

The oxidative protein linkages generated during crosslinking were investigated by LC-MS/MS methods. FA-modified YCP samples contained more diferulic acid bonds (8–5'noncyclic dehydro-diferulic acid), whereas no FA-modified YCP samples had more dityrosine bonds. The oxidative crosslinking bond types depended on the reaction conditions and the availability of preferred residues for LacTv. FA-modified samples primarily formed FA-mediated intermolecular bonds, while non-FA samples showed dityrosine-driven intramolecular bonding without significant carbonyl formation on alternate amino acids.

The structural modifications influenced the YCP's behavior at the air/water interface. FA-modified YCP samples, with a high molecular weight and intermolecular links, exhibited a slower adsorption to the air/water interface. No FA-modified YCP samples, with a low molecular weight, a greater level of intramolecular links and notable protein unfolding, showed fast transfer to the interface and reduced equilibrium surface tension than both native YCP and FA modified YCP. The fluorescence analysis revealed the unfolding of No FA-modified YCP as reaction times progressed, while FA-modified YCP samples showed fluorescence shielding, due to the integration of FA into their structures. No FA-modified YCP samples relied on dityrosine linkages for intramolecular bonding, altering the tertiary structure without increased molecular weight. Conversely, crosslinking of FA-modified YCP samples formed FA-based intermolecular bonding, increased the molecular weight of modified proteins and incorporated FA into the protein's structure. FTIR analysis showed reduced secondary structural features in FA and non-FA crosslinked YCP and higher aggregation than unmodified proteins. However, the preserved secondary structural features in both treatment conditions had greater resistance to heat induced denaturation up to 80 °C.

Overall, laccase modification of YCP led to modified proteins with diverse structural and technofunctional properties depending on the reaction conditions. This enables tailoring YCP for applications requiring either enhanced interfacial activity or steric stabilization. YCP shows promise as an emulsifier for vegan, allergen-free, and clean-label products, and enzymatic modification expands its application potential. Further studies on oil/water interface stabilization and food product compatibility are recommended.

CRediT authorship contribution statement

Kelly Light: Writing – review & editing, Writing – original draft, Validation, Methodology, Investigation, Formal analysis, Data curation. Lan Liu: Writing – review & editing, Validation, Methodology, Investigation, Formal analysis. Ashraf Ismail: Visualization, Validation, Methodology, Formal analysis. Christophe Blecker: Writing – review & editing, Validation, Methodology, Investigation, Data curation. Salwa Karboune: Writing – review & editing, Validation, Supervision, Resources, Project administration, Funding acquisition, Formal analysis, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary material

Supplementary data including SDS-PAGE gel photos and LC-MS/MS chromatograms to further support the text.

Data availability

Data will be made available on request.