*Article* **Fermentation of Lupin Protein Hydrolysates—Effects on Their Functional Properties, Sensory Profile and the Allergenic Potential of the Major Lupin Allergen Lup an 1**

**Katharina Schlegel 1,2, Norbert Lidzba 3, Elke Ueberham 3, Peter Eisner 2,4,5 and Ute Schweiggert-Weisz 2,6,\***


**Abstract:** Lupin protein isolate was treated using the combination of enzymatic hydrolysis (Papain, Alcalase 2.4 L and Pepsin) and lactic acid fermentation (*Lactobacillus sakei* ssp. *carnosus*, *Lactobacillus amylolyticus* and *Lactobacillus helveticus*) to investigate the effect on functional properties, sensory profile and protein integrity. The results showed increased foaming activity (2466–3481%) and solubility at pH 4.0 (19.7–36.7%) of all fermented hydrolysates compared to the untreated lupin protein isolate with 1613% of foaming activity and a solubility of 7.3 (pH 4.0). Results of the SDS-PAGE and Bead-Assay showed that the combination of enzymatic hydrolysis and fermentation of LPI was effective in reducing *L. angustifolius* major allergen Lup an 1 to a residual level of <0.5%. The combination of enzymatic hydrolysis and fermentation enables the production of food ingredients with good functional properties in terms of protein solubility and foam formation, with a balanced aroma and taste profile.

**Keywords:** enzymatic hydrolysis; fermentation; lupin protein; functional properties; sensory profile; lupin allergy; lup an 1; plant protein

#### **1. Introduction**

The global population is expected to grow by 2 billion to 9.7 billion people over next 30 years [1]. In order to maintain the world's population with protein, food systems will be faced with a major challenge. An increase in animal production would not be a sustainable option to meet the high demand of protein. Supplying the world with animal protein has drastic effects on the environment, which include the intensive use of land, the deterioration of air and water quality and the emission of greenhouse gases. [2]. A promising way to reduce the impact of nutrition on the environment could be the partial replacement of animal proteins with plant proteins. Legumes such as lupins are becoming more and more popular as an alternative source of protein to animal protein and soy. Lupins are widely cultivated in Europe and in South America and are particularly attractive for human consumption because of their high protein content of 39%, up to 55% (dry matter) [3] with a well-balanced amino acid profile and a low carbohydrate content compared to other legumes [4]. However, the use of lupin proteins in some foods like refreshing drinks is limited due to the characteristics of their functional properties, in particular their solubility in the acidic range. Several studies have shown that enzymatic hydrolysis can significantly

**Citation:** Schlegel, K.; Lidzba, N.; Ueberham, E.; Eisner, P.; Schweiggert-Weisz, U. Fermentation of Lupin Protein Hydrolysates— Effects on Their Functional Properties, Sensory Profile and the Allergenic Potential of the Major Lupin Allergen Lup an 1. *Foods* **2021**, *10*, 281. https://doi.org/10.3390/ foods10020281

Academic Editor: Maria Antonietta Ciardiello Received: 16 December 2020 Accepted: 29 January 2021 Published: 31 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

improve the functional properties of plant proteins such as protein solubility, foaming and emulsification [5–10]. In addition, it was shown that the allergenicity of the proteins can be reduced by enzymatic treatments [11,12]. However, protein hydrolysis can also lead to negative modifications of the sensory profile by producing a bitter taste that inhibits their use in food [6,7,12]. One promising approach to influence the sensory profile of these ingredients could be lactic acid fermentation. Several studies have shown that fermentation of plant proteins by lactic acid leads to reduced or masked off-flavors in legumes and improves their sensory profile [13–15]. However, lactic acid fermentation is less effective in improving the functional properties of proteins and the degradation of polypeptides to reduce the allergenic potential of those proteins is less effective compared to enzymatic hydrolysis. The combination of enzymatic hydrolysis and fermentation could use the positive effects of both treatments to develop low-allergen food ingredients with excellent functional properties and a balanced sensory profile. The objective of this study was to investigate the effect of the combination of enzymatic hydrolysis and fermentation on functional properties of lupin proteins—protein solubility, foaming properties and emulsification capacity. In addition, the sensory profile of the treated ingredients was also evaluated. In order to obtain first insights of the reduction of the allergenic potential of lupin proteins, both molecular weight distribution and immunological detectability of the fermented hydrolysates were compared with untreated lupin protein isolates.

#### **2. Materials and Methods**

*2.1. Raw Materials and Chemicals*

#### 2.1.1. Lupin Seeds

Lupin (*Lupinus angustifolius* L. cultivar Boregine) seeds were provided by Saatzucht Steinach GmbH & Co KG (Steinach, Germany).

#### 2.1.2. Enzymes

The sources, types and supplier of the enzymes used in this study are listed in Table 1. Proteolytic enzyme preparations were chosen according to a previous study [7], in which promising results were achieved by those enzyme preparations in lupin protein degradation.


**Table 1.** Sources, types and supplier of the enzymes used in this study.

#### 2.1.3. Strain Selection

The fermentation of lupin protein hydrolysates was carried out using *Lactobacillus sakei* ssp. *carnosus* (DSM 15831), *Lactobacillus amylolyticus* (TL 5) and *Lactobacillus helveticus* (DSM 20075). Microorganisms were purchased from Deutsche Sammlung von Mikroorganismen und Zellkulturen (Braunschweig, Germany) and Chair of Brewing and Beverage Technology (Technical University Munich, Germany). The microorganisms were stored as a cryoculture in our strain collection and were activated on MRS (De Man, Rogosa & Sharpe) agar. The selection of the microorganisms were chosen according to Schlegel, Leidigkeit, Eisner and Schweiggert-Weisz [15], based on the promising results achieved in the aroma formation and hedonic evaluation.

#### 2.1.4. Nutrient Media

Liquid growth media and agar were obtained from Carl Roth (Karlsruhe, Germany).

#### *2.2. Preparation of Lupin Protein Isolate*

Lupin protein isolate (LPI) was prepared from *Lupinus angustifolius* L. cultivar Boregine. Lupin seeds were dehulled, separated and passed through a roller mill. The resulting flakes were de-oiled in *n*-hexane. Flakes were extracted with 0.5 M HCl (1:8 *w*/*w*) for 1 h. Suspension was separated using a decanter centrifuge at 5600× *g* and 4 ◦C for 1 h and the supernatant was discarded. The acid pre-extracted flakes were dispersed in 0.5 M NaOH (1:8 *w*/*w*, pH 8.0) for 1 h at room temperature while stirring and separated by centrifugation (5600× *g*, 4 ◦C, 1 h). The supernatant was adjusted to pH 4.5 with 0.5 M HCl. The precipitated proteins were separated by centrifugation (5600× *g*, 130 min) and neutralized with 0.5 M NaOH, pasteurized at 70 ◦C for 10 min) and spray-dried using an Anhydro spray dryer (SPX Flow Technology, Charlotte, NC, USA) with an inlet temperature of 180 ◦C and an outlet temperature of 80 ◦C at a mass flow rate of 24 kg/h.

#### *2.3. Enzymatic Hydrolysis of LPI*

Enzymatic hydrolysis of LPI was performed in a 5 L thermostatically controlled reaction vessel, as previously described [7]. Briefly, the protein isolate was dispersed in deionized water with an Ultra-Turrax at 5000 rpm for 1 min (IKA-Werke GmbH & Co. KG, Staufen, Germany) to achieve a protein concentration of 50 g/kg. Temperatures and pH values were adjusted, and enzyme preparation was added (Table 1). The suspension was incubated at controlled pH and temperature (Table 2) with continuous stirring for 2 h. The reaction was stopped by heating up to 90 ◦C for 20 min, afterwards, the suspension was cooled down to room temperature and neutralized (pH 7) with 1 M NaOH or 1 M HCl.


**Table 2.** Experimental design of enzymatic hydrolysis and fermentation with enzyme-to-solution ratio (E/S), temperature and pH value of enzymatic hydrolysis.

<sup>1</sup> enzyme-to-solution ratio of enzymatic hydrolysis.

#### *2.4. Fermentation of Hydrolysed LPI*

Hydrolysed LPI was fermented in a 5 L glass reaction vessel in an incubator under aerobic conditions and ina5L glass reaction vessel with a bioreactor (Biostat B, Sartorius AG, Goettingen, Germany) under anaerobic conditions, respectively, as described previously [15]. Briefly, 0.5% glucose (*w*/*w*) was added to the 5% hydrolyzed LPI (*w*/*w*) solution, pasteurized at 80 ◦C for 20 min and inoculated with the activated culture of 10<sup>7</sup> CFU/mL. Anaerobic conditions for *Lactobacillus helveticus* were achieved by flushing the reactor with N2. LPI was fermented at 37 ◦C (*Lactobacillus sakei* ssp. *carnosus* and *Lactobacillus helveticus*) and 42 ◦C (*Lactobacillus amylolyticus*), respectively, for 24 h without stirring. Viable cell counts were determined after 0 h and 24 h of fermentation and pH course were recorded for 24 h fermentation with one measurement point at each 30 min (wtw pH 3310 pH electrode, Xylem Analytics Germany GmbH, Weilheim, Germany). The process was stopped by heating the suspension up to 90 ◦C for 20 min. All samples were neutralized (pH 7.0) with 1 M NaOH and spray dried with a Niro Atomizer 2238 (GEA, Düsseldorf, Germany).

#### *2.5. Chemical Composition*

Protein content was determined according to the Dumas combustion method AOAC 968.06 (TruMac N, Leco Instruments, Mönchengladbach Germany) using a protein calculation factor of N × 5.8 [16]. The dry matter (105 ◦C) and ash (950 ◦C) contents were analyzed according to AOAC methods 925.10 and 923.03 in a TGA 601 thermogravimetric system (Leco Instruments GmbH) at 105 ◦C and 950 ◦C, respectively.

#### *2.6. Molecular Weight Distribution*

The molecular weight distribution of the untreated LPI and fermented LPI hydrolysates was determined by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) as described by Laemmli [17] with modifications [15]. Briefly, LPI and fermented LPI hydrolysates were resuspended in loading buffer (0.125 mol/L Tris-HCl, 4% SDS (*w*/*v*), 20% glycerol (*v*/*v*), 0.2 mol/L DDT, 0.02% bromophenol blue, pH 6.8), dissolved in an ultrasonic bath (30 ◦C, 30 min), boiled at 95 ◦C for 5 min (Eppendorf Thermomixer, Eppendorf AG, Hamburg, Germany) and separated with a Mini Spin centrifuge at 12,045× *g* for 10 min (Eppendorf AG). Supernatant was mixed in a ratio of 1:10 with loading buffer (see above). An aliquot of 10 μL of each sample was transferred into the wells of Bio-Rad 4–20% Criterion TGX Stain-Free precast gels (Bio-Rad Laboratories GmbH, Feldkirchen, Germany). The Precision Plus Protein™ Unstained Protein Standard (Bio-Rad Laboratories) was used as molecular weight marker (10–250 kDa). Gels were run at room temperature for 38 min at 200 V (60 mA, 100 W) in a vertical electrophoresis cell (Bio-Rad Laboratories). Protein bands were visualized using a Gel Doc™ EZ Imager system (Bio-Rad Laboratories) and determined using Image Lab software (Bio-Rad Laboratories).

#### *2.7. Determination of Lup an 1 with Specific Monoclonal Antibodies*

Amount of Lup an 1 was measured by an in house sandwich assay using β-conglutin specific antibodies (Izimab Lup an 1-1 and Izimab Lup an 1-2 (unpublished)) on FLEXMAP 3D® flow analyzer system from Luminex Corporation (Austin, TX, USA). The monoclonal capture antibody, Izimab Lup an 1-1, was coupled onto MagPlex® beads and the monoclonal detection antibody Izimab Lup an 1-2 was biotinylated according to standard procedures. Lupin samples were extracted using denaturing conditions [6]. The samples were incubated with bead conjugated Izimab Lup an 1-1, washed three times in PBS with 0.05% Tween® 20 (PBS-T) and incubated with biotinylated Izimab Lup an 1-2 followed by another wash with PBS-T. Streptavidin-conjugated phycoerythrin (SA-PE) was diluted 1:2000 in assay buffer, 100 μL per well were added to the sandwich complex and agitated for one hour. The reaction was terminated by washing the wells three times with wash buffer and subsequently filled up with 120 μL assay buffer. Readings on FLEXMAP 3D® flow analyzer expressed as median fluorescence intensity (MFI) per 100 beads calculated on the basis of a calibration curve were converted to relative reactivity in percent by dividing the individual MFI by the MFI of the control sample, which represents the LPI used as raw material in the processing steps.

#### *2.8. Technofunctional Properties*

#### 2.8.1. Protein Solubility

The solubility (%) of LPI and fermented LPI hydrolysates was determined in duplicate at pH 4.0 and 7.0 according to Morr, et al. [18]. The sample was suspended in 0.1 M NaCl (3% *w*/*w*), pH was adjusted (0.1 M HCl) and stirred for 1 h at room temperature. Undissolved fractions of the samples were removed by centrifugation (20,000× *g*, 15 min) at room temperature. The supernatants were filtered with a Whatman No. 1 filter paper. Protein content was determined by a method of Lowry, et al. [19] using the DC Protein Assay (Bio-Rad Laboratories) with a BSA standard curve for calculating the protein concentration. The absorbance was read at 750 nm. The resulting protein content was related to the total amount of protein and protein solubility (%) was determined.

#### 2.8.2. Foaming Properties

Foaming activity and stability were determined according to Phillips, et al. [20] in duplicate. For foaming activity, a 100 mL solution of 5% protein (pH 4.0 and 7.0) was whipped at room temperature for 8 min in a whipping machine (Hobart 50-N, Hobart GmbH, Offenburg, Germany. The increase of foam volume was defined as foaming activity (%); the percentage of foam volume remaining after 1 h was defined as foam stability (%).

#### 2.8.3. Emulsifying Capacity

Emulsifying capacity was determined at pH 4.0 and 7.0 according to Wang and Johnson [21]. Samples were suspended in deionized water (1% *w*/*w*), adjusted to pH and stirred with an Ultraturrax (IKA-Werke GmbH & Co. KG) at 18 ◦C. Rapeseed oil was continuously added (10 mL/min) until phase inversion was detected (<10 μS/cm). The volume of added oil was used to calculate emulsifying capacity (mL oil per g sample). Measurements were performed in duplicate.

#### *2.9. Sensory Analysis of Fermented Hydrolysates*

#### 2.9.1. Panelists

Panelists were trained assessors recruited from Fraunhofer IVV (Freising, Germany), with no known illness and normal olfactory function during the test. The panel consisted of 10 panelists. All panelists were tested for their olfactory function in weekly training sessions with selected suprathreshold aroma solutions. The samples were evaluated during two sessions on one day.

#### 2.9.2. Descriptive Analysis

Samples (2% *w*/*w*) of LPI and fermented LPI hydrolysates were prepared in tap water by stirring. All samples were presented to the panel in covered glass vessels. The panelists were requested to open the lid of the vessels and record the retronasal aroma and taste attributes. After a short discussion, common retronasal aroma attributes with corresponding references and taste attributes were collected and rated on a scale from 0 (no perception) to 10 (strong perception) by each panelist in a separate session. The following ten aroma and taste qualities and corresponding references (given in brackets) were selected by the trained panelists (n = 10) for LPI and fermented LPI hydrolysates: oatmeallike (oatmeal); cocoa-like (cocoa); malty (methylpropanal); green, grassy (hexanal) pea-like (2–isopropyl-3-methoxypyrazine); fatty ((E,Z)-2,4-nonadienal); cardboard-like, cucumberlike ((E)-2-nonenal); roasty (2-acetylpyrazine); cooked potato-like (3-(methylthio)propanal); earthy (2,3- diethyl-5-methylpyrazine), bitter, salty, sour. The sensory evaluation was carried out once.

#### *2.10. Statistical Analysis*

Results are expressed as means ± standard deviations and, for sensory evaluation, (aroma profile) as median ± standard deviations. Data were analyzed using pairwise *t*-test to determine the significance of differences between a sample and the untreated LPI, with a threshold of *p* < 0.05. Statistical analysis and visualization were performed with Origin 2018 for Windows (Origin Lab Corporation, Northampton, MA, USA). The results of the sensory evaluation were evaluated using Principal Component Analysis (PCA) covariance matrix to assess aroma and taste qualities. PCA was performed using Origin 2018 for Windows (Origin Lab Corporation).

#### **3. Results and Discussion**

#### *3.1. Chemical Properties*

Dry matter content (%) of all fermented LPI hydrolysates were within the range of 92.8% for the Alcalase 2.4 L hydrolysate S4 (*Lactobacillus sakei* ssp. *carnosus*) to 94.0% for the Pepsin hydrolysate S7 (*Lactobacillus sakei* ssp. *carnosus*)—a significantly (*p* < 0.05) lower dry matter content compared to untreated LPI (95.4%) (Table 3).


**Table 3.** Dry matter (%), protein content (%) and ash content (%) of lupin protein isolate (LPI) and fermented LPI hydrolysates.

The data are expressed as mean ± standard deviation (*n* = 4). Means marked with an asterisk (\*) within a column indicate significant differences between the sample and the untreated LPI (*p* < 0.05) following pairwise *t*-test.

Untreated LPI contained the highest protein content with 89.6%. Protein content of fermented LPI hydrolysates ranged from 66.8% for *Lactobacillus amylolyticus* fermented Papain hydrolysate (S3) to 78.7% for *Lactobacillus helveticus* fermented Papain hydrolysate (S2).

Ash content (%) was within the range of 4.2% for LPI to 8.8% for the Pepsin hydrolysate S9 (*Lactobacillus helveticus*). The increased ash content of treated samples compared to untreated LPI might be attributed to the addition on NaOH during neutralization after fermentation.

#### *3.2. Comparison of Microbial Growth on Lupin Protein Isolate Solutions*

The growing parameters (CFU and pH) for all experiments after 0 h and 24 h of fermentation are shown in Tables 4 and 5. The results showed that all microorganisms were able to grow in hydrolyzed LPI. The minimum increase in CFU/mL (ΔECFU/mL) was recorded for Pepsin hydrolysate S9 (*Lactobacillus helveticus*) with 1.02 × 108 CFU/mL and the maximum for Pepsin hydrolysate S8 (*Lactobacillus amylolyticus*) with 1.32 × <sup>10</sup><sup>9</sup> CFU/mL.


**Table 4.** Colony forming units (CFU) after 0 h and 24 h of fermentation.

The data are expressed as mean ± standard deviations from duplicates.


**Table 5.** pH values after 0 h and 24 h of fermentation.

The data are expressed as mean ± standard deviations from duplicates.

The lowest pH values after 24 h of fermentation were recorded for Papain hydrolysate S3 (*Lactobacillus helveticus*) and Pepsin hydrolysate S9 (*Lactobacillus helveticus*) with pH values of 3.3 and 3.7, respectively. Fermentation of hydrolysates obtained by Alcalase 2.4 treatment (S4–S6) tended to show higher pH values compared to hydrolysates obtained by other proteolytic enzyme preparations. Presumably, Alcalase 2.4 degradation leads to a higher buffer capacity of the respective samples. The lowest pH reduction (5.1) was achieved by Alcalase 2.4 hydrolysate S4 (*Lactobacillus sakei* ssp. *carnosus*). A former study described a rapid decrease in pH for *Lactobacillus helveticus* and *Lactobacillus amylolyticus* during the fermentation of LPI [15]. In this work, the pH curve over 24 h of fermentation was also recorded (Figure 1 shows exemplarily the pH curve for LPI samples (not hydrolyzed) fermented with *Lactobacillus helveticus*. It was observed that *Lactobacillus helveticus* showed an extended lag phase of 7 h in decreasing the pH value than *Lactobacillus helveticus* did on Papain (S3) and Pepsin (S9) hydrolysate, respectively. However, after 24 h of fermentation, *Lactobacillus helveticus* was able to decrease the pH in the range of 3.1 and 4.6 in all hydrolysates.

**Figure 1.** Course of pH value for *Lactobacillus helveticus* fermentation on LPI (black curve) and on Papain (S3, blue curve), Alcalase 2.4 L (S6, green curve) and Pepsin (S9, red curve) LPI hydrolysates over 24 h fermentation. The data are expressed as mean ± standard deviations from duplicates.

#### *3.3. Molecular Weight Distribution (SDS-PAGE) and Immunoreactivity*

The molecular weight distribution (SDS-PAGE) of LPI and its fermented hydrolysates was used to determine the protein integrity and is shown in Figure 2. All treatments resulted in prominent changes in the SDS-PAGE profile with hydrolyzed polypeptides to smaller fragments with molecular weights below 30 kDa. Enzymatic hydrolysis seems to have the greatest influence on the degradation of polypeptides of LPI. Comparing the profiles of the molecular weight distribution of a previous study [7] after hydrolysis of LPI, it is shown that the profiles did not change visibly compared to those after the combination of hydrolysis and fermentation. This observation is also supported by the results of a further study [15], which show that fermentation has only minor influence on the profiles of molecular weight distribution. Furthermore, it is described that β-conglutin with a molecular weight of ~55–61 kDa is known as the major allergen of *L. angustifolius* L. (Lup an 1) [22]. The SDS-PAGE results of all fermented LPI hydrolysates showed a degradation of the described IgE-reactive polypeptide (Figure 2) independent of the enzyme preparations used. These observations were confirmed by the results of the Bead-Assay (Figure 3). A significant reduction in the signal intensity of Lup an 1 below 0.5% was observed for all fermented hydrolysates compared to unfermented LPI (100%) due to the combination of enzymatic hydrolysis and fermentation The highest decrease of immunological reactivity was achieved by treatment with Alcalase 2.4 L combined with fermentation S5 (*Lactobacillus amylolyticus*) and S6 (*Lactobacillus. helveticus*). No binding of Lup an 1 antibodies in sandwich format could be detected (read out not detectable). The assumption that enzymatic hydrolysis is a powerful approach to reduce allergenic potential is supported by further studies [23–27].

**Figure 2.** Peptide band profiles of LPI and fermented LPI hydrolysates as determined by SDS-PAGE under reducing conditions.

**Figure 3.** Determination of Lup an 1 with specific monoclonal antibodies (Izimab Lup an 1-1 and Izimab Lup an 1-2) of unfermented LPI and fermented LPI hydrolysates using Bead-Assay. Results are shown as mean ± standard derivation of each duplicate (1 and 2). Each Bead-Assay was performed in triplicate.

#### *3.4. Technofunctional Properties*

#### 3.4.1. Protein Solubility

Protein solubility of LPI and fermented LPI hydrolysates was determined as a function of pH at pH 4.0 and pH 7.0 and is given in Table 6. All samples showed higher protein solubility at pH 7.0 than at pH 4.0. Solubility decreases as pH value approaches the isoelectric point, approximately pH 4.5–5.0, as discussed frequently [4,7,12,28,29]. Protein solubility at pH 7.0 ranged for all samples, from 45.2% for the Papain hydrolysate S2 (*Lactobacillus helveticus*) to 66.6% for Papain hydrolysate S3 (*Lactobacillus amylolyticus*). However, a significant difference between LPI and the fermented hydrolysates (with the exception of Pepsin hydrolysate S6 (*Lactobacillus helveticus*)) could not be observed.

In contrast, protein solubility at pH 4.0 after enzymatic and fermentation treatment was significantly different (*p* < 0.05) in comparison to untreated LPI (7.3%). All fermented hydrolysates showed significant higher protein solubility with values between 19.7% and 36.7%. The minimum protein solubility (19.7%) of the fermented hydrolysates was determined after fermentation of the Papain hydrolysate S2 (*Lactobacillus helveticus*). The fermentation of a nonhydrolyzed lupin protein isolate with *Lactobacillus helveticus* also resulted in very low protein solubility at pH 4.0 and 7.0 in a former study [15] The single hydrolysis of LPI by means of Alcalase 2.4 L resulted in a considerable increase in protein solubility compared to the hydrolysis by means of other proteolytic enzyme preparation [7], confirming high proteolytic activity of the Alcalase 2.4 L preparation. However, the tendency for the highest protein solubility for Alcalase 2.4 L-hydrolyzed LPI at both pH values and significantly lower protein solubility for Papain and Pepsin-hydrolyzed LPI compared to Alcalase 2.4 L hydrolysate was not observed in this study. A significant difference between the applied enzymes could not be identified.


**Table 6.** Protein solubility, foam properties and emulsifying capacity of LPI and fermented LPI hydrolysates.

The data are expressed as mean ± standard deviation (*n* = 4). Means marked with an asterisk (\*) within a column indicate significant differences between sample and untreated LPI (*p* < 0.05) following pairwise *t*-test.

#### 3.4.2. Foaming Properties

Foaming activity (%) and foam stability (%) of untreated LPI and fermented LPI hydrolysates are given in Table 6. All treated samples showed significantly (*p* < 0.05) higher foaming activity, with values from 1819% up to 2789% at pH 4.0 and 2466% up to 3481% at pH 7.0, compared to untreated LPI with values of 828% and 1613%, respectively. The highest activities with values of 3338%, 3443% and 3481% were determined for Pepsin hydrolysates (S7–S9) at pH 7.0. A high foaming activity of 3614% was also observed at pH 7.0 after single hydrolysis of LPI with Pepsin [7]. The foaming activities at pH 4 were lower than those at pH 7 for all samples, with the exception of the samples hydrolyzed with Alcalase 2.4 L (S4–S6). It is assumed that the lower foam activity under acidic conditions was caused by the significantly lower protein solubility of the samples at pH 4.0 in comparison to pH 7.0.

Furthermore, previous studies have shown that foaming activity of LPI increases only slightly through fermentation [15,30]. Therefore, enzymatic treatment seems to have the greatest effect on foaming activity in this study. Enzymatic hydrolysis breaks larger polypeptides into smaller peptides, thus improving foam formation by rapid diffusion at the air–water interface [31]. Furthermore, Meinlschmidt, Schweiggert-Weisz and Eisner [26] have shown an increase in foaming activity of soy protein isolates after enzymatic hydrolysis and fermentation.

LPI showed foam stability (%) of 89% at pH 7.0 after 1 h standing. The samples treated with *Lactobacillus amylolyticus* and Pepsin (S8), as well as the samples treated with *Lactobacillus sakei* ssp. *carnosus* and Alcalase 2.4 L (S4), did not show significant differences in foam stability with 90% and 96%, respectively, in comparison to the untreated LPI. All other treated samples showed significantly lower or even no foam stability. At pH 4.0, the untreated LPI showed foam stability of 92%. All fermented LPI hydrolysates did not show any foam stability. Foams can be stabilized by large peptides with flexible structures. Hydrolysis reduces protein surface coverage, which means that the air–water interface is no longer stabilized and foam collapse occurs in the hydrolyzed protein foams [32].

#### 3.4.3. Emulsifying Capacity

The emulsifying capacity of untreated LPI at pH 7.0 was 666 mL/g. As shown in Table 6, all treated samples showed significantly (*p* < 0.05) lower emulsifying properties

than LPI with values lower than 477 mL/g. The samples hydrolyzed with Alcalase 2.4 L showed the lowest emulsifying properties with values of 246 (S4), 283 (S5) and 323 mL/g (S6). The emulsifying capacity at pH 4.0 showed lower values than at pH 7.0 for all samples, with the exception of the samples treated with Alcalase 2.4 L (S4–S6). El-Adawy, Rahma, El-Bedawey and Gafar [32] found a direct correlation between the emulsifying properties and solubility of a protein. In this study, we could not find a correlation between the solubility of proteins and their ability to form an emulsion within the same pH value. However, we observed that emulsification capacities were lower at pH 4.0 than at pH 7.0. In addition, the samples at pH 4.0 showed lower protein solubilities than at pH 7.0.

#### *3.5. Sensory Analysis*

Fermentation enables changes in sensory profiling through the generation and degradation of flavor active compounds. Proteins, carbohydrates and fats from the raw materials provide the necessary precursors, e.g., for the formation of volatile aroma-active compounds. Most of the formation pathways of aroma-active and organic compounds are based on functional metabolic pathways of lactic acid bacteria.

Comparative retronasal aroma profile analysis (Figure 4) shows the results of Papain hydrolysate S2 (*Lactobacillus amylolyticus*) and Alcalase 2.4 L hydrolysate S4 (*Lactobacillus sakei* ssp. *carnosus*) compared to LPI across the panel and highlights the most impressive changes with significant differences (*p* < 0.05) in aroma perception of cocoa-like and malty. For LPI, the major aroma perceptions were rated with a moderate perception with values of 3.5 for oatmeal-like, 3.0 for fatty and 3.0 for pea-like. All other aroma impressions were rated with an intensity of 2.6 and below. The Papain hydrolysate S2 (*Lactobacillus amylolyticus*) and the Alcalase 2.4 L hydrolysate S4 (*Lactobacillus sakei* ssp*. carnosus*) were the only samples that showed significant differences (*p* < 0.05) in the intensity of aroma perception (cocoa-like and malty) compared to untreated LPI. The cocoa-like impression increased with values of 2.5 for S2 and 6.3 for S4 compared to 0.8 for unmodified LPI. Furthermore, S4 showed a significant increase (*p* < 0.05) in the intensity of the aroma perception of malty (3.7). All other samples showed no significant differences (*p* < 0.05) of aroma perceptions compared to those of unmodified LPI, and are therefore not presented in the figure (see Supplementary Materials).

The taste impressions bitter, salty and sour of LPI and fermented LPI hydrolysates were investigated by 10 panelists in the sensory evaluation. The bitter, salty and sour intensities of LPI were described with values of 3.0, 2.1 and 1.0, respectively. All treated samples did not show significant difference (*p* < 0.05) compared to untreated LPI, with the exception of the sample treated with Alcalase 2.4 L hydrolysate S5 (*Lactobacillus amylolyticus*), with a more bitter intensity of 5.6. Additionally, Pepsin hydrolysate S9 (*Lactobacillus helveticus*) showed significantly (*p* < 0.05) higher intensity of saltiness compared to untreated LPI. No significant differences could be observed between LPI and the treated samples regarding the intensity of sour.

Principal Component Analysis (PCA) was applied to identify the most relevant sensory attributes that affect sample characteristics. Figure 5 shows the resulting biplots of the uncorrelated principal components (PCs) 1 and 2 based on the sensory data of the respective hydrolysates including untreated LPI, fermented LPI hydrolysates and the scaled loadings.

**Figure 4.** Comparative retronasal aroma profile analyses of LPI and fermented LPI hydrolysates (S2 and S4) on a scale from 0 (no perception) to 10 (strong perception). The data are displayed as mean values of the sensory evaluation (n = 10).

**Figure 5.** Biplot of aroma and taste of the unfermented LPI and fermented LPI hydrolysates. \* aroma \*\* taste.

The first two components of PCA explained 39.92% and 26.12% of the observed variation (68.22% in total). The attribute with the strongest influence on PC1 was salty (−0.298) taste and cocoa-like (0.754) aroma impression. In contrast, PC2 was primarily described by the attributes bitter (−0.412) and salty (0.514). Unfermented LPI (−0.896/−1.108) was in close correlation with the pea-like aroma attribute and was found on the negative side of PC2, together with the Papain hydrolysates samples S1 (*Lactobacillus sakei* ssp*. carnosus*) and S3 (*Lactobacillus helveticus*) and the Alcalase 2.4 L hydrolysates samples S5 (*Lactobacillus amylolyticus*) and S6 (*Lactobacillus helveticus*). The samples S2, S4 and the Pepsin hydrolysates S7 (*Lactobacillus sakei* ssp*. carnosus*), S8 (*Lactobacillus amylolyticus*) and S9 (*Lactobacillus helveticus*) were found on the positive side and were opposite of the unfermented LPI. The Pepsin hydrolysate fermented with *L. helveticus* (S9) scored the highest in the PC2 (2.623) and was nearest with salty. S9 raised a low pH value of 3.7 after fermentation, which was neutralized (pH 7.0) with 1M NaOH. Due to the larger amount of NaOH, compared to other samples, the salty impression may have been increased. This assumption is further supported by the increased ash content for S9. In contrast, the Papain hydrolysate S1 (*Lactobacillus sakei* ssp*. carnosus*) and the Alcalase 2.4 L hydrolysates S5 (*Lactobacillus amylolyticus*) and S6 (*Lactobacillus helveticus*) scored the lowest in PC2 and were nearest ranged to the bitter taste attribute. It was observed that the Alcalase 2.4 L hydrolysate S5 (*Lactobacillus amylolyticus*) showed a significant difference (*p* < 0.05) of bitter intensity to untreated LPI. The Alcalase 2.4 L hydrolysates S4 (*Lactobacillus sakei* ssp*. carnosus*) and S6 (*Lactobacillus helveticus*) did not show a significant difference (*p* < 0.05), but a tendency of higher bitter intensity. Alcalase 2.4 L hydrolysates are known for bitter taste. Several studies have already shown that the hydrolysis of plant proteins such as lupin [7,10], soy [6] and pea [27] results in a bitter taste. However, S4 (*Lactobacillus sakei* ssp*. carnosus*) was not ranged close to the bitter attribute, rather in the positive range of PC2 (1.746), and was nearest ranged with cocoa-like and was scored the highest in the PC1 (4.483).

#### **4. Conclusions**

The combination of enzymatic hydrolysis and fermentation of lupin protein isolate can increase foaming activity while maintaining proper emulsification capacity. In addition, the modification increases protein solubility at acidic conditions and thus opens new possibilities in food applications such as refreshing drinks. LPI provides a well-balanced sensory profile that has been partially altered by the treatments in aroma and taste perception. The sensory acceptance of LPI and modified LPI needs to be investigated with a consumer panel. SDS-PAGE and Bead-Assay indicated that the combination of enzymatic hydrolysis and fermentation of LPI is effective in breaking down large polypeptides into low molecular weight peptides and degrading with it the major allergen Lup an 1 of *L. angustifolius*. Thus, this two-step process represents a promising method for the reduction of the allergenic potential of LPI. Nevertheless, in vivo studies should be performed to investigate the allergenicity of fermented lupin protein hydrolysates. The use and possible applications of fermented lupin protein hydrolysates in food should be tested in practice by subsequent application trails.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/2304-815 8/10/2/281/s1, Table S1: Comparative retronasal aroma profile analyses of LPI and fermented LPI hydrolysates, Table S2: Intensity of bitter, salty and sour taste perception of LPI and fermented LPI hydrolysates, Table S3: Principal Component (PC) scaled scores and loadings for LPI and fermented LPI hydrolysates obtained by Principal Component Analysis (PCA).

**Author Contributions:** Conceptualization, K.S. and U.S.-W.; methodology, K.S., E.U. and U.S.- W.; formal analysis, K.S. and N.L.; investigation, K.S. and N.L.; resources, P.E.; writing—original draft preparation, K.S. and E.U.; writing—review and editing, K.S., N.L., E.U., U.S.-W. and P.E.; visualization, K.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Acknowledgments:** This work was supported by the Fraunhofer-Zukunftsstiftung.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Review* **Barley Protein Properties, Extraction and Applications, with a Focus on Brewers' Spent Grain Protein**

**Alice Jaeger 1, Emanuele Zannini 1, Aylin W. Sahin <sup>1</sup> and Elke K. Arendt 1,2,\***


**\*** Correspondence: e.arendt@ucc.ie; Tel.: +353-021-490-2064

**Abstract:** Barley is the most commonly used grain in the brewing industry for the production of beer-type beverages. This review will explore the extraction and application of proteins from barley, particularly those from brewers' spent grain, as well as describing the variety of proteins present. As brewers' spent grain is the most voluminous by-product of the brewing industry, the valorisation and utilisation of spent grain protein is of great interest in terms of sustainability, although at present, BSG is mainly sold cheaply for use in animal feed formulations. There is an ongoing global effort to minimise processing waste and increase up-cycling of processing side-streams. However, sustainability in the brewing industry is complex, with an innate need for a large volume of resources such as water and energy. In addition to this, large volumes of a by-product are produced at nearly every step of the process. The extraction and characterisation of proteins from BSG is of great interest due to the high protein quality and the potential for a wide variety of applications, including foods for human consumption such as bread, biscuits and snack-type products.

**Keywords:** brewers' spent grain; barley protein; by-product valorisation; brewing waste; food ingredient

#### **1. Introduction**

Barley is the most commonly used grain in the brewing industry for the production of beer-type beverages. Both raw barley and malted barley are used, often in combination with adjunct grains such as rice and corn. Arguably, the most important fraction of the barley grain is the endosperm, comprised of starch granules suspended in a protein matrix, as it provides the growing plant embryo with all the products needed to begin growth. In this review, the proteins of the barley grain will be the focus. In particular, the extraction and application of barley proteins from brewers' spent grains will be explored. As brewers' spent grain is the most voluminous by-product of the brewing industry, the valorisation and utilisation of spent grain protein is of great interest in terms of sustainability. There is an ongoing global effort to minimise processing waste and increase up-cycling of processing side-streams in order to support sustainable growth in the coming decades. However, sustainability in the brewing industry is complex. There is an innate need for a large volume of resources such as water and energy, as well as a disconnect between raw materials and the processing facility resulting in an increased carbon footprint [1]. In addition to this, large volumes of by-product are produced at nearly every step of the process including spent grain, spent yeast, wastewater and spent hops, among others. The brewing industry is responsible for a huge volume of waste with ~10,000 tonnes of liquid waste and 137–173 tonnes of solid waste produced per 1000 tonnes of beer [2], so the potential valorisation of these waste products is of great economic interest. Due to its high nutritional value, particularly in terms of high-quality protein, the application of BSG proteins in human nutrition is of particular interest. The addition of BSG has already been examined in breads, biscuits and other bakery products as well snack-type products. In addition to this, applications of BSG include use in animal feed formulation for a range of

**Citation:** Jaeger, A.; Zannini, E.; Sahin, A.W.; Arendt, E.K. Barley Protein Properties, Extraction and Applications, with a Focus on Brewers' Spent Grain Protein. *Foods* **2021**, *10*, 1389. https://doi.org/ 10.3390/foods10061389

Academic Editor: Rotimi Aluko

Received: 30 April 2021 Accepted: 11 June 2021 Published: 16 June 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

species [1–4] as well as use as a potential substrate for other industrial processes, namely biogas production, polysaccharide extraction and phenolic compound isolation [4–6].

#### **2. Barley Structure**

Barley (*Hordeum vulgare, vulgare* L.) is a widely cultivated and easily adaptable crop, most commonly used as a raw material for the brewing process. The main components of the barley grain are the embryo, the aleurone layer, the endosperm and the husk (grain covering) (Figure 1). The embryo is the most important living tissue in the grain and develops into the plant. It contains a limited amount of starch, lipids and protein to sustain the embryo prior to germination [7]. The aleurone layer is two to three cells thick and encases the endosperm. It is a living tissue containing protein, phytin, phospholipids, RNA and some carbohydrate and plays a critical role in the regulation of endosperm degrading enzymes [7–9]. The grain covering acts to protect the integrity of the grain throughout harvesting and processing. The grain covering can further be divided into the seed coat, pericarp and husk [10]. The husk consists of mainly cellulose and a small amount of polyphenols and bitter substances, whereas the pericarp develops from the ovary wall and acts as a protective cover over the kernel [9]. The husk has proven to aid the filtration process during the lautering step of brewing as well as protecting the grain during malting [7]. However, husk-less cultivars have been shown to be effective in alcohol production, although with an altered malting process [11].

**Figure 1.** Structure of the barley kernel sectioned transversely (**a**) with detailed husk (**b**) and starchy endosperm (**c**), Reprinted with permission from ref. [9]. Copyright 2013 Elsevier.

The endosperm consists of starch granules suspended in a protein matrix. The starch is the most abundant component, comprising approximately 60% of grain weight. The endosperm cell walls consist of mainly non-starch polysaccharides, namely β-glucans and arabinoxylans [7].

The barley grain is reported to contain between 8 and 30% protein as a percentage of total mass [7,12–14]. This protein is synthesized in the endosperm and the aleurone layer during grain development and accumulates during grain filling. The crude protein content is used to predict malting and, therefore, brewing quality [12,15]. The protein content of the barley grain has a complex relationship with quality with regards to malting barley. While high protein barley would be desirable for feed applications, a lower level is desirable for malting varieties. The ideal protein content for malting barley resides between 10 and 12%, and too high or low a protein content can negatively affect malting quality and, therefore, brewing capacity. In general, a high protein content is undesirable, as there is a strong correlation between a high protein content and a low carbohydrate content, leading to a low extract yield. This may also lead to excessive enzyme activity. However, too low a protein level may lead to insufficient amino acids for the yeast nutrition during the brewing process as well as low enzymatic activity, leading to a lower level of fermentable sugars and, therefore, a poor extract yield [7,15–17].

The protein content of the barley grain is highly variable and can be affected by barley cultivar as well as environmental conditions and the addition of fertilizers. In hot, dry and nitrogen-rich environments, it has been shown that the protein content of the barley grain is increased, bringing the protein level above that which is suitable for malting (approx. 11.5% protein) [16–19]. Increased temperature and reduced rainfall have been shown to have a negative correlation with malting quality as measured by malt extract [15]. However, very high levels of precipitation are also unfavourable, leading to reduced endosperm development [18]. This is undesirable for brewers, as an increased amount of malt will need to be used to produce the beer. Therefore, the development of new barley cultivars to produce grain with a high malt extract and diastatic power while keeping an acceptable protein level is of great interest. Diastatic power refers to the combined activity of α- and β-amylases and β-glucanases. These enzymes play a crucial role in the brewing process, degrading malt carbohydrates to fermentable sugars [17,20].

The application of a nitrogen (N) fertilizer also has a significant effect on barley protein levels. Nitrogen addition was observed to be raising grain protein concentration and reducing malt extract, with a much more significant effect in warmer and drier regions [16]. However, the use of low protein barley genotypes may allow for barley with an acceptable grain protein level to be grown on land with increased nitrates (NO3). It has been shown that, while a low-protein genotype is generally associated with reduced diastatic power, the addition of nitrogen fertilizer can increase this level to be similar to that of the most commonly grown malting barley cultivars [17]. The genotype, growing climate, soil composition and fertilizer addition can be of great importance, influencing the performance and quality of the barley malt during the brewing process.

Genetic modification has also been considered for the development of the ideal barley cultivar. While this method does show potential, it cannot yet compete with the more broadly used cultivars in terms of yield [21].

#### **3. Barley Proteins**

#### *3.1. Hordeins/Prolamins*

The majority (30–50%) of the barley protein fraction consists of hordeins, a protein belonging to the prolamin group, so named due to their high content of glutamine and proline [14,15,22]. These hordeins (or prolamins) are not single proteins but complex polymorphic mixtures of polypeptides [23]. The other barley proteins consist of a mix of albumins, globulins and glutelin [7]. The hordeins can be further divided into subsections based on their amino acid compositions and extractability characteristics [12,24]. B-hordein is the largest fraction accounting for between 70 and 90% of the total hordein content, and it is the major storage protein of the barley grain. B-hordein can be further broken down into subunits B1, B2 and B3 based on their electrophoretic capabilities. Other fractions include C-, D- and γ-hordeins, with C-hordeins accounting for 10–20%, and the remaining types making up less than 5% collectively [12]. B-hordeins and γ-hordeins are sulphur-rich, while C-hordeins are sulphur-poor. D-hordeins are distinguished by their high molecular weight (>100 kDa) [24]. These differing proteins are coded by specific genes: Hor2 (B-hordeins), Hor1 (C-hordeins), Hor3 (D-hordeins) and Hor5 (γ-hordeins [24]. A study by Howard et al., 1996, observed that the measurement of D-hordein levels in the barley grain gave a more significant single measurement for malting quality spanning a range of climatic and agronomic conditions, compared to total grain protein [15].

#### *3.2. Glutelin*

Glutelin is the second most abundant fraction in barley storage proteins, making up 35–45% of the total storage protein. It contains high levels of glutamine, proline and glycine, while also being rich in other hydrophobic amino acids [25]. In a study by Wang et al. (2010), barley glutelin was found to have very good emulsifying properties [26]. This is theorised to be due to a balanced ratio of polar and non-polar amino acids, allowing the protein to adsorb easily onto the surface of oil droplets and quickly lower the interfacial tension. However, due to poor solubility, glutelin requires dehydration at pH 11 and further pH adjustment for the emulsification properties to be active, which is not practical for applications in food systems. Deamidation has been found to increase barley glutelin solubility and emulsifying properties [25].

#### *3.3. Protein Z*

Protein Z is the major barley albumin with a molecular mass of 40 kDa and a pI ranging from 5.5 to 5.8 [27]. Protein Z is a member of the serpin protein group and represents about 5% of total barley albumin. However, protein Z is heat stable and resistant to enzymatic degradation, meaning it survives the brewing process unmodified and is one of the major proteins still present in the finished beer, with a particular role regarding beer foam stability alongside lipid transfer proteins [27,28]. The relationship between the concentration of protein Z and foaming properties has been studied in a few separate instances. It was found that while protein Z is a family of proteins, protein Z4 was the predominant type, as it made up 80% of the protein Z fraction [28]. It was also determined that protein Z plays an integral role in foam stability, as the addition of purified protein Z enhanced foam stability. Similarly, Ayashi et al. (2008) found that protein Z4 and lipid transfer protein 1 (LTP-1) were beneficial to foaming capacity and stability. It is theorised that LTP-1 does this by binding lipids, which are known to negatively influence foam stability [29], while protein Z4 interacts with the α-acids derived from the hops addition. While both the albumin and hordein fractions of the barley protein are capable of foaming, the albumin fraction has significantly higher foam stability. However, denaturation of these protein fractions causes an increase in hydrophobicity, therefore enhancing foaming properties, with albumins being more stable than hordeins. The foam stability of both fractions is improved by exposure to the bitter α-acids derived from the hops, but both protein fractions are also susceptible to hydrolysis by yeast-derived proteinase A, resulting in diminished foam stability [30].

A study by Boba (2010) examined the use of protein Z as a method for monitoring the malting progress of barley grain via glycation. The proteins in barley malt are glycated by D-glucose, a product of starch degradation during malting, and potentially improve foam stability and prevent protein precipitation during wort boiling [31]. It is the lysine and arginine residues that are glycated, with 16% of the lysine residues in protein Z being modified during the Maillard reactions of the malting process. Prolonging of malting increases the level of glycated peptides. Therefore, measuring its concentration using gel electrophoresis, liquid chromatography and mass spectrometry can be used as an efficient measurement of malting progression [31]. This could be applied effectively to control protein variety and quantity in the final beer.

#### *3.4. Effect of Brewing Processes on Barley Proteins*

The brewing process significantly alters the proteins of the grains. An understanding of the changes to barley proteins that occur during malting and mashing is key to determining how these processes affect the final protein composition of BSG as well as the residual beer proteins.

#### 3.4.1. Malting

Malting is the controlled germination of the barley grain. During malting, partial degradation of the cell walls, starchy endosperm and storage protein occurs [32,33]. The hydrolytic enzymes responsible for this process are produced by the grain during the first stages of germination. For protein degradation, endoproteases and carboxypeptidases are produced and secreted into the endosperm. The endoproteases act to break down the proteins into peptides, while the carboxypeptidases further degrade the peptides into free amino acids [33]. Greater than 70% of the proteins, mainly hordeins and glutelin, in the barley grain are degraded by endoproteases to create smaller peptides during the malting process [33]. A study by Celus et al. (2006) demonstrated that B- and D-hordeins are the main fractions degraded [32]. Adequate modification of the grain constituents during malting is key to the final quality of the beer. Insufficient modification can cause issues such as low extract yields and low fermentability [33].

#### 3.4.2. Starch Degradation Inhibition

Insufficient protein degradation during malting has been shown to inhibit the degradation of starch during the mashing process. A study by Slack et al. (1979) examined this phenomenon and determined that hordein proteins are very closely associated with the surface of the starch granules and can form a protective layer around the granule. Due to this, starch degrading enzymes (namely α-amylase) have limited access to the starch granules, therefore limiting the production of fermentable sugars. It was shown that this effect is even more pronounced for the small starch granules in comparison to the large starch granules. During mashing at approx. 65 ◦C, the larger starch granules gelatinise, therefore becoming more available to the α-amylase to be converted to maltose. Small starch granules have a higher gelatinisation temperature, and so, the same effect is not observed [34]. During malting, the B- and C-hordeins on the starch granule surface should be largely degraded, allowing for the starch degrading enzymes to easily access the granule. However, in cases where the malt is under-modified or raw adjuncts are introduced, the inhibitory effects of hordeins can be observed. A study by Yu et al. (2018) investigated the same effect and confirmed that barley proteins, in particular the soluble component, play a role in slowing starch degradation. This study also looked at alternate methods by which this occurs [35]. As well as the physical barrier as described by Slack et al. (1979), it is also theorised that the digestive enzyme (α-amylase) binds to the hordein and glutelin and is, therefore, unable to act on the substrate. It is thought that proteins adhere to the starch granules via disulphide linkages, as proteins can be removed by the addition of cysteine [34].

#### 3.4.3. Mashing

Mashing is the first step in the brewing process and continues the enzymatic degradation that began during the malting of the barley grains. Most endoproteases (approx. 90%) will survive the kilning process and will, therefore, be present in the mashing phase [33]. Therefore, protein degradation into peptides and amino acids will continue during mashing. A study by Jones and Marinac (2002) investigated the effect of mashing on malt proteinase activity. It found that the endoproteases maintained a high activity level during the protein rest stage of mashing, where the mash is held at approx. 38 ◦C to aid protein degradation. However, protease activity rapidly deteriorates when the temperature is raised to 70 ◦C. This rise in temperature is to facilitate the hydrolysis of malt and adjunct starches to sugars, meanwhile deactivating the proteases. Due to this, soluble protein levels in the wort can be manipulated by the brewer, if so desired, by increasing or decreasing the length of the protein rest phase. It is also noted that attempting to change the ratio of protein fractions within the soluble protein is not likely to succeed, as all of the proteinases appear to denature at the same temperature [36]. Due to the near-total deactivation of the proteases during the starch conversion phase, elongation/shortening of this step would have no effect on wort protein levels. During mashing, a significant

fraction of grain proteins are degraded and solubilised, and through wort boiling, they are glycated and coagulated to form aggregates that can be separated during wort boiling as 'hot trub' [37]. A study by Celus et al. (2006) determined that during mashing, disulphide bonding is induced between B- and C-hordeins. While it was previously thought that B- and D-hordein (and glutelin) were the main components of the aggregate, this study suggests that it is instead composed of mainly B-hordeins, in which some C-hordein is enclosed [32]. These aggregated hordeins have been reported to cause mash separation issues [32,34,35,38]. An earlier study by Moonen et al. (1987) also described the formation of this protein complex during mashing, consisting of residual malt proteins and glutelins. It also examines how this complex negatively impacts the lautering and filtration processes [38]. This once again highlights the importance of sufficient protein degradation during the malting stage, as poor-quality malt with insufficient modification will contain a higher level of residual protein as well as a lower enzyme activity, therefore causing more issues regarding gelation.

This impenetrable protein complex persists as a coating on the residual spent grains, the upper section of the 'oberteig' layer. Therefore, they are an important fraction of the brewers' spent grain (BSG) protein complement. Due to the impenetrable nature of the matrix, protein extraction may need to involve pre-treatments and/or vigorous disruption procedures [39].

The methods behind the formation of these aggregates are largely unknown. Moonen et al. (1987) hypothesised that high-molecular-weight subunits (such as D-hordeins) form the 'backbone' structure of these gels. This idea is supported by Skerritt and Janes (1992), who used protein assays, electrophoresis and HPLC to examine the relationship between these disulphide-linked protein aggregates and malting quality. It was determined that D-hordeins were the slowest to extract and are, therefore, assumed to act as the 'backbone' of the gel matrix [40]. An increased level of sodium dodecyl sulphate (SDS) unextractable proteins, namely B1- and B2-hordeins, indicates a lower-quality malt, as hordein subunits are more readily and easily extracted from good-quality malt [40].

#### 3.4.4. Residual Beer Proteins

Many studies have focused on comparing the protein profiles of malts with the protein profiles of the final beer product. One such study by Klose et al. (2010) examined the protein profiles of the malt and beer as well as the intermediate stages (wort and hot trub) using two-dimensional gel electrophoresis as well as lab-on-a-chip technology. It was determined that the 500 mg of proteinaceous material present [41] in the final beer originated from the initial grain, where increased inherent levels of cysteine and/or lysine allowed the proteins to be resistant to enzymatic degradation and increased their heat stability [27]. This need for heat stability was also noted by Curioni et al. (1995), where the major fraction of residual protein in beer was determined to comprise of two polypeptides with an approximate molecular weight of 40 kDa, later determined to be two albumins coinciding with the characteristics of protein Z [41]. It was also noted that upon examination of the precipitated protein in the 'hot-trub', the molecular mass of the peptides was low, indicating that the enzymatic degradation of proteins established during malting continued throughout mashing [27].

#### **4. Overview of BSG**

Brewers' spent grain (BSG) is the most abundant by-product of the brewing process, consisting of up to 85% of total brewery waste [6]. On average, 20 kg of wet BSG is produced for every 100 L of finished beer [10]. Due to an ever-increasing interest in waste reduction and by-product valorisation, the fractionation and application of BSG-derived ingredients is an expanding field of study. BSG is what remains of the initial brewing grains after being subjected to the malting and mashing processes. Therefore, it consists of mainly the husk–pericarp–seedcoat, meaning BSG is rich in both cellulosic and non-cellulosic polysaccharides, as well as lignin. While rich in fibre, it is also rich in protein, with the protein fraction accounting for approx. 19–30% of total grain composition as outlined in Figure 2 [32,42–49].

**Figure 2.** Protein content of brewers' spent grain (BSG), % *w*/*w* dry matter [32,42–49].

As well as a high protein content, BSG also contains a high level of essential amino acids, as outlined in Table 1. These essential amino acids constitute approx. 30% of the total protein, and lysine is of particular interest, as it is generally the limiting amino acid in cereal foods for human consumption [48]. In a study by Connolly et al. (2013), protein isolates prepared from pale and black BSG were characterised for amino acid composition. Glutamine and proline were the most abundant amino acids present in both isolates, and different temperatures were used during the protein extraction to determine the effect of temperature. The results showed that the pale BSG, extracted at 50 ◦C, contained the highest levels of all amino acids, except for cysteine [39]. Glutamine, proline and leucine were the most abundant, while the sulphur-containing amino acids, methionine and cysteine, were the scarcest. The isolate extracted at a lower temperature (20 ◦C) had a lower amino acid content overall. However, it was similar to the 50 ◦C in that glutamine and proline were the most abundant, while the sulphur-containing amino acids were present in the lowest levels. The black BSG isolates both contained a lower level of amino acids than their pale counterparts. This is most likely due to the roasting of the malt at high temperatures (>200 ◦C), where amino acids may be degraded or used in Maillardtype reactions. Essential amino acids are crucial for human health and can be lacking in certain foods, therefore requiring fortification. Due to the presence of a high level of essential amino acids, BSG-derived proteins as fortification agents in foodstuffs for human consumption present an economical solution to these issues.


**Table 1.** Amino acid composition of barley, barley malt and BSG—adapted from Waters et al. (2012).

#### *4.1. Adjuncts*

Adjuncts are ingredients, other than malted barley, used in brewing to provide additional carbohydrates to contribute to sugars in the wort. While the proteins of the barley grain are the focus of this review, if discussing proteins in brewers' spent grains, it is important to consider that alternative protein sources may also be present.

In certain countries, namely Germany, Switzerland and Greece, the purity law ('Reinheinsgebot') states that beer may only be made from water, malt, hops and yeast, so adjuncts are prohibited. Generally, adjuncts are used to reduce raw material cost and/or to add certain desirable qualities to the finished beer [50,51]. While they are predominantly cereal-based, sugar-based syrups can also be used. The most common cereals used as adjuncts are rice and maize, as well as raw, un-malted barley. Due to differing gelatinization temperatures, adjuncts often have to be 'cooked' separately from the main mash in a double mash conversion system [52]. However, triticale can be added into the main mash as it has a lower gelatinisation temperature of 59–65 ◦C, well within normal mashing temperatures [53]. A study by Agu (2002) also investigated the possible use of sorghum as a brewing adjunct alongside barley. It was discovered that sorghum has the potential to release higher levels of peptides than commonly used maize and was very effective at a 5% addition, whereas at 20% addition, a decrease in total wort peptides was noted [54]. A separate study by Glatthar et al. (2005) investigated the use of un-malted triticale as a brewing adjunct and noted that the wheat and rye hybrid held great potential for use as a brewing adjunct. While adjunct addition levels vary widely, high levels of addition can present increased effects on sensory characteristics as well as technological difficulties, such as increased viscosity [51,53,55]. In a study by Yorke et al. (2021), it was found that while 30% adjunct addition had little to no observable effect on beer characteristics, 60% addition presented increased sensory differences, low free amino nitrogen and dramatically altered the fermentation profile [51].

The protein contents of adjuncts can greatly affect wort quality. The balance between adjuncts and malt must be carefully monitored as adjunct addition can 'dilute' the enzymatic activity of the malt, therefore requiring a malt of a higher diastatic power or the addition of commercial enzymes [52]. A study carried out by Schnitzenbaumer et al. (2012) investigated the effect of using oats as a replacement for malted barley. It was noted that the replacement of 20% or more of malted barley with oats resulted in a decrease in the

amount of free amino acids and total soluble nitrogen, as well as decreased extract and significantly reduced foam stability.

Due to the frequent use of adjuncts in the brewing industry, the analysis, extraction and utilization of brewers' spent grain protein is not only dealing with barley proteins but also potentially those of adjuncts used in the brewing process.

#### *4.2. Rice Protein*

Rice is a commonly used adjunct in brewing to produce a light and clean tasting beer. Due to its low cost and high starch content, it is used to supplement the carbohydrates available from the barley and malt, leading to increased fermentative capability. Brewers' rice is generally produced as a by-product of the edible rice milling industry, as up to 30% of the grains may be fractured and deemed unsuitable for the edible rice market [56]. While the starchy endosperm is the main component of the rice grain (89–94%), brown rice contains 6.6–7.3% protein, milled rice 6.2–6.9% protein, and basmati rice contains 8.2–8.4% protein [57]. Although levels can vary greatly depending on the environment, soil and cultivar, in general, this protein level is lower than that of both barley, barley malt, maize and sorghum [58]. As opposed to the barley grain, where the major proteins are the endosperm-specific prolamin storage proteins, the major protein in the rice grain is a glutelin-type storage protein, containing 63.8–73.4% glutelin. Other proteins present include water-soluble albumins (9.7–14.2%), salt soluble globulins (13.5–18.9%) and alcohol soluble prolamins (3.0–5.4%) [57]. Rice proteins have been shown to be significantly resistant to hydrolysis [59]. As a result, the addition of a rice adjunct provides very little free amino nitrogen (FAN), and this must be compensated for by the barley and malt fractions.

#### *4.3. Maize Protein*

Alongside rice, maize is another grain extensively used as an adjunct in the production of beers. In a similar way to rice, the high volume of starchy endosperm serves to supplement the sugars provided by the barley and barley malt alone for a more efficient and economical production process. A study by Agu (2002) describes the effects of maize addition at 5–20%. The results showed a decrease in extract recovery but an increase in levels of FAN and peptide nitrogen when compared to barley adjuncts [54]. Maize contains ~10% protein [60]. Similarly to barley, the major proteins present in maize are endosperm-specific prolamin storage proteins. These are mainly small, 19–25 kDa α-prolamins [58].

#### **5. Protein Extraction Methods**

#### *5.1. Innate Protein Extraction Methods*

Several methods of protein extraction from the innate barley grain have been explored in the literature to date. Generally, grains are milled and/or pearled before being exposed to an extraction buffer [26,61,62]. Extraction solutions for barley protein fractions include salt, alcohol and alkaline. In a study by Wang et al. (2010), the hordein fraction was isolated by using an ethanol solution (55–75% *v*/*v*), whereas the glutelin fraction was extracted from the residues by application of an alkaline solution (pH 9–11.5). The hordein fraction was isolated using a rotary evaporator to remove ethanol, while the salt and alkali solubilised fractions were adjusted to pH 5 to facilitate protein precipitation [26].

#### *5.2. Extraction of Proteins from BSG*

In order to utilise each of the valuable components of BSG, methods for separating the fractions must be determined. In order for barley proteins to be successfully commercialised and used in an industrial setting, efficient extraction methods are needed to separate the valuable proteins from the spent grain. Some of the potential protein extraction methods include alkaline extraction, acid extraction and filtration, as well as more novel techniques such as ultrasonic treatment and pulsed electric field treatment.

#### *5.3. pH Shift Extraction Methods (Alkaline Extraction)*

Alkaline extraction is the most widely used and well-known methods for protein extraction. A study by Cavonius et al. (2015) explains the mechanics of this method well. At an alkaline pH, the proteins in a system obtain a net negative charge, increasing repulsion within and between the protein molecules. The interactions between the protein molecules and water are promoted; therefore, the proteins are solubilised. When the pH is lowered to the proteins isoelectric point (pI), this negative charge is lost and the protein's interaction with the water is minimized. This leaves the proteins insoluble and allows them to precipitate out of solution [63]. In this study by Cavonius et al. (2015), the 'pH-shift' method was applied to microalgal biomass, but the same principle can be applied to BSG.

Alkaline extraction was utilised in a study by Celus et al. (2007) to create a BSG protein concentrate (BPC). A solution of 0.1 M NaOH was used, and the mixture was held at 60 ◦C for 60 min to allow for optimum extraction. Following this, the proteins were precipitated by adjusting the pH of the filtrate to pH 4.0 using 2.0 M citric acid, and this precipitate was separated using centrifugation. The resulting BPC contained 60% protein on a dry weight basis [64].

A more recent study by Connolly et al. (2013) explored and characterised the proteinrich isolates from wet pale and wet black brewers' spent grain using alkaline extraction. While KOH and Na2CO3 were considered for use, NaOH was determined to be the most effective. It was determined that 110 mM was the most effective and efficient concentration of NaOH for extraction, even though extracted protein peaked at 200 mM. Several other parameters for optimal protein extraction were also explored in this study. Generally, protein extraction for both types of BSG increased over a range from 20–60 ◦C with pale BSG increasing from 37.17 to 88.20 mg g−<sup>1</sup> and black BSG increasing from 37.02 to 64.32 mg g−<sup>1</sup> [39]. Temperatures over 60 ◦C were not considered due to the risk of protein denaturation. This is in agreement with previous studies by Bals et al. [65] and Celus et al. [64]. The optimum weight/volume ratio was also determined to be 1:20. The barley proteins present in the alkaline extracts of BSG consisted mainly of hordeins.

A study by Diptee et al. (1989) explored the various parameters involved in protein extraction yield from brewers' spent grain using response surface methodology. These variables included time, temperature and particle (grain) size. While the efficacy of commonly used protein extraction solutions on BSG proteins has been debated, the extraction solution used in this study consisted of 3% sodium dodecyl sulphate and 0.6% Na2HPO4, followed by precipitation of the protein in ethanol. Using this method, a 60% protein yield was obtained using a BSG: extractant ratio of 2.5:100 and by heating the mixture to 90 ◦C for 95 min [66]. This was in agreement with the maximum value predicted using response surface methodology.

As seen through extensive use, alkaline extraction is a viable and efficient method to extract proteins, namely hordeins and glutelin, from BSG. The protein obtained from BSG could potentially be used in a variety of applications, including the food and nutraceutical industries.

#### *5.4. Filtration*

Filtration, including microfiltration and ultrafiltration, has been a commonly employed method of separating out fractions based on molecular size. Examples include the use of membrane separation to separate the solid fractions of corn, resulting in high protein corn gluten meal (67% protein), high-fat corn germ, corn starch and high fructose corn syrups [67].

A study by Tang et al. (2009) applied ultrafiltration to brewers' spent grain as a method of protein extraction. A BSG extract was prepared using ultrasound-assisted extraction with a sodium carbonate buffer, then filtered through a nylon cloth and centrifuged. The resulting mixture was used as the feed solution for ultrafiltration. Protein recovery by ultrafiltration was highly successful, with more than 92% of the protein being retained using both 5 and 30 kDa membranes [68]. The protein contents of the final products were

20.09% protein when concentrated using the 5 kDa membrane and 15.98% protein when concentrated using the 30 kDa membrane [68]. Connolly et al. [69] also used membranes of 10 and 30 kDa to concentrate proteins from a BSG preparation. A benefit of using this type of fractionation is the lack of added heat, yielding a higher-quality protein product.

#### *5.5. Pre-Treatments*

The effects of various pre-treatments of the BSG have also been widely investigated. Shearing as a pre-treatment proved to be beneficial. It is thought to be due to the physical disruption of cell membranes and lignocellulosic material that allows for easier extraction of proteins. The use of acid as a pre-treatment was also investigated, but in this study, it did not appear to have any effect on protein extractability [39]. Hydrogen peroxide proved to be effective, as its application increased protein yield by 27%. However, this method of pre-treatment is not frequently used due to its potential to cause oxidation of the extracted protein.

A study by Qin et al. (2018) outlines the effects of several differing pre-treatment strategies on the protein extractability of BSG. A combination treatment of alkaline and acid seemed to be effective, with up to 95% protein extraction achieved [70]. However, it was discovered that a high percentage of lignin and carbohydrates were solubilised with the protein, meaning the process was not selective for proteins only. This would limit its practical applications, as a purer protein fraction is usually the most desirable. A single step dilute acid pre-treatment using H2SO4 was proven to be effective at extracting up to 90% of the BSG protein; however, a similar issue with the carbohydrates and lignin was present. The benefits of hydrothermal pre-treatments were examined at differing solid to liquid ratios, differing temperatures (30 to 135 ◦C) and increasing time (1 and 24 h). It was found that this allowed for 64–66% protein extraction and was more selective for proteins than the acid and alkaline treatments [70]. Hydrothermal treatment is potentially a very good option for optimisation of protein extraction from BSG, as it requires no addition of chemicals and can be carried out at a relatively low temperature to preserve protein quality but also minimise energy use when compared to acid treatment.

#### *5.6. Enzymatic Treatment*

The use of enzymes, namely proteases, to separate proteins from BSG has also been explored in literature. In addition to alkaline treatment, enzymatic treatment could optimise protein extraction potential from BSG. A study by Niemi et al. (2013) states that while up to 53% of proteins in BSG can be solubilised in alkaline conditions without protease addition, up to 76% was solubilised with the addition of an alkaline protease, in addition to carbohydrase treatment.

Alcalase, an alkaline protease extracted from *B. licheniformis*, has been reported to be effective in aiding to separate the protein fraction from BSG [64,71–73]. Determined to be capable of solubilising and extracting up to 77% of BSG proteins, creating peptides with a much lower molecular weight than other less effective proteases [73]. It was found that Alcalase performed better at pH 8.0, requiring a pH adjustment of the raw materials. To achieve maximum protein solubilisation, 10–20 μL of enzyme per g dry matter was required to work for approx. 4 h. However, significant protein separation, up to 64%, was observed using does as low as 1.2 μL [73]. It was also noted through amino acid analysis that an Alcalase treatment was particularly efficient at solubilising the proline and glutamine components of barley hordeins, more so than other amino acids. A study by Qin et al. (2018) also aimed to utilise the protein solubilisation effects of Alcalase for protein extraction optimisation from BSG. However, the results differed from those reported in the literature, with only 43–50% protein extraction. It was theorised to be due to the pH of the mixture being slightly too low at pH 6.25 after pre-treatment as opposed to pH 6.5, the optimum pH for Alcalase. Due to this, results differed from those reported in the literature. When the treatment was repeated at pH 8.0, a minor increase in protein solubility was observed, though it was still a lower figure than had been reported previously.

A study by Niemi et al. (2013) investigated the combined use of proteases and carbohydrases to maximise protein solubility. Milled BSG was treated with Depol 740 L in a 10% weight/volume mixture for 5 h at 50 ◦C with constant stirring. This BSG was then treated with different proteases, including Alcalase, ProMod and Acid Protease A [71]. Alcalase was found to be the most effective enzyme at pH 9.5, solubilising 35% of BSG biomass, while the other proteases only solubilised approx. 14%. It was found that enzymatic carbohydrate digestion was beneficial in maximising proteins solubilisation and extraction from BSG, leading to an increased protein yield. It is theorised to be due to the degradation of important cell structures by the carbohydrate degrading enzymes, allowing greater access to the internal cell structures. It was also noted that prolonged exposure to alkaline conditions, in the absence of a protease, also led to a significant amount of protein being solubilised, indicating that prolonged exposure to a high pH decreases cell wall integrity and increases protein solubility [71]. Depol 740 was also used in combination with Alcalase to improve protein solubilisation in a study by Treimo et al. (2008). While it was found to improve the overall solubilisation of the biomass, the increase was largely in the solubilisation of non-proteinaceous material. Overall, treatment with carbohydrate degrading enzymes had only minor effects on protein solubilisation, as opposed to the larger effect observed by Niemi et al. (2013).

A study by Celus et al. (2007) investigated the technological properties of BSG protein hydrolysates as prepared by enzymatic hydrolysis. Commercially available enzymes, including Alcalase, Flavourzyme and Pepsin, were used over a range of times and concentrations to obtain hydrolysates with varying degrees of hydrolysis (DH). BSG protein concentrate was first prepared by alkaline extraction and then subjected to enzymatic hydrolysis. It was determined that enzymatic hydrolysis generally improved the solubility of BSG proteins at lower pH values, leading to increased protein yield across the entire pH range [64]. A significant difference in technological function was also observed between the samples, based on the enzyme used during the hydrolysis. Emulsification and foaming properties decreased with increasing DH when Alcalase or Pepsin was used, as these properties are reliant on peptides with a MW in excess of 14.5 kDa. However, the opposite was observed for hydrolysates prepared with Flavourzyme [64]. Hydrolysis of the BSG proteins greatly improved their technological functionality. Foaming and emulsification properties, which were negligible in the raw BSG, were significantly improved by the hydrolysis treatment.

The use of thermochemical pre-treatment, as well as carbohydrate and protein hydrolysing enzymes on the fractionation of protein and lignin from BSG was investigated in a study by Rommi et al. (2018). Significant protein solubilisation was achieved using an alkaline protease treatment, bringing BSG protein solubilisation from approx. 15% to almost 100%, in agreement with the studies mentioned previously [64,71–73]. The steam explosion treatment reduced protein solubility but increased the efficacy of extract separation in the centrifugation step. Lignin and protein from BSG co-extracted and could only be partially separated using acidic precipitation. Due to this, raw BSY may be preferred in the production of BSG protein hydrolysates, with an aim to limit lignin co-extraction [72].

#### *5.7. Ultrasonic-Assisted Extraction*

Ultrasonic treatment is a more novel method that has been explored as a means of aiding protein extraction from BSG. In a study by Tang et al. (2010), the optimum ultrasound treatment was determined using response surface methodology, regarding three parameters: time, power and solid/liquid ratio. A treatment time of 81.4 min, at an ultrasonic power of 88.2 W/100 mL extractant and a solid/liquid ratio of 2.0 g/100 mL, was used to predict an optimal yield of 104.2 mg/g BSG, and this was in accordance with the experimental value [74]. Ultrasonic treatment of BSY was also determined to improve certain functional characteristics, namely fat absorption capacity, foaming properties and emulsifying properties, as well as improving extraction yield [75]. When ultrasound treatment is

used in combination with enzymatic treatment, as in Yu et al. (2019), enzymatic loading and incubation time can be reduced, and protein solubility significantly increases [76].

#### *5.8. Pulsed Electric Field*

Pulsed electric field technology (PEF) is another novel extraction method for food compounds that is gaining interest. PEF consists of the application of short-duration pulses of an electric current through a sample that is secured between two electrodes [77]. The enhanced extractability of samples post-PEF is due to the dielectric disruption of cell membranes as a result of exposure to an electric current. This technology has already been explored as a non-thermal and, therefore, cheaper and more sustainable method of food preservation and microbial inactivation technique [77,78]. Outside of food preservation and antimicrobial applications, PEF has also been explored as a means of increasing juice yield from alfalfa leaves [79] and also as a pre-treatment to increase the extractability of proteins and phenolic compounds from light and dark BSG extracts [80]. This study by Kumari et al. (2019) showed that PEF assisted extraction significantly increased the level of free amino acids in the light BSG extract, and all essential amino acids were present in both extracts, with tryptophan being the only exception [80]. PEF treatment has also been explored in combination with ultrasound treatment as a method of optimising the extraction of phenolic compounds and proteins from agro-industrial by products [77] and olive kernels [81].

The more well-known methods of protein extraction, namely alkaline extraction and enzymatic treatment, are very suitable for protein extraction from BSG and are supported by large amounts of literature documenting their success. However, there are potential downsides and room for optimization with regards to the amount of solvent and/or enzyme required as well as heat treatments that could negatively impact protein quality. This leaves much room for optimization with a focus on cost, quality and sustainability. The more novel methods have little to no supporting literature with regards to their use for BSG but could benefit from being a 'cleaner' and more economical method of protein extraction, as there is no need for reagents, as well as potentially yielding protein of a higher quality due to a lack of thermal treatment. However, it is clear that the best course of action for protein extraction from BSG is a combination of the methods discussed above, as it has been shown in BSG and other food systems that combining methods allows for optimization of protein yield [68,70,72,77,81].

#### **6. Applications**

#### *6.1. Animal Nutrition*

Due to its high protein and fibre content, BSG is most often used as a component of animal feed as either a wet or dry feed. In combination with cheap and widely available nitrogen sources (such as urea), BSG can provide all essential amino acids [6]. The introduction of BSG into the diet of milking cattle has been shown to increase milk yield, total milk solids and milk fat when compared to an animal on a control diet. Protein levels and lactose content were not significantly affected by the change in diet [6,82–84]. While most commonly used in cattle, the use of BSG in the feed of other animals has also been explored. In a study by Mukasafari et al. (2017) BSG was successfully used to substitute up to 50% of sow and weaner meal without any negative effects on the quality of the pigs [85]. A study by Yaakugh and Tegbe (1994) investigated the replacement of maize in the diet of pigs with dried brewers' grains. It was found that, with a replacement level of up to 45%, the daily weight gain was significantly depressed, but the final carcass weight was not affected [86]. A study by Oh et al. (1991) investigated the incorporation of 15% treated BSG into the diets of poultry. The treatment consisted of partially hydrolysing the grains by cultivating the fungus *Trichoderma reesei* on the grains to alter the amino acid profile and release soluble sugars. Incorporation of these grains at 8 and 12% into the diets of broiler chicks resulted in a significant improvement in their growth and feed conversion ratio in the first 4 weeks. No further improvements were seen after 6 and 8 weeks of growth [87].

The use of BSG has been investigated as a source of protein in aquafeeds as a sustainable alternative for fishmeal and oil to reduce reliance on marine resources. Partial replacement of fishmeal with BSG (20–30%) in the feeding of rainbow trout and gilthead seabream showed similar results in digestion efficiency when compared to the control, where fishmeal was used as the main protein source. The BSG replacement showed good protein, amino acid and lipid digestibility, making BSG a suitable alternative for fishmeal, increasing the sustainability of both industries [3]. Similarly, San Martin et al. (2020) explored the use of an enzyme hydrolysis process in a bid to increase the digestibility of BSG proteins in aquaculture feeds. While further studies are required into the benefits of hydrolysis, both the hydrolysed and non-hydrolysed BSG proteins have shown good digestibility and would be a suitable fishmeal alternative, improving economic and environmental sustainability [2].

Due to the very high water content of BSG at the point of production (77–81%) [6], the transportation and storage of BSG for use as animal feed presents a challenge in minimizing microbial growth that may cause illness in animals, as well as general material degradation. There are guidelines regarding the preservation of BSG, suggesting the use of preservatives such as benzoate, propionate and sorbate to extend stability [6,10]. However, these measures only work to extend shelf-life for 4–5 days, so a more effective preservation technique is required for longer storage times. Drying is the most commonly employed methods of stabilizing wet BSG by reducing microbial growth. A study by Bartolome et al. (2002), compared three preservation methods: oven drying, freeze drying and freezing. Freezing was deemed to be unsuitable due to the potential for alterations in arabinose content and lack of suitability for large volumes. Oven drying and freeze drying were found to be equally effective in terms of reducing volume and preventing changes to the composition. However, from an economical and cost standpoint, oven drying was determined to be the most effective methods for removing moisture from BSG and stabilizing the product [88].

#### *6.2. Bio-Degradable Film*

A study by Lee et al. (2015) describes the use of brewers' spent grain protein (BGP) in the production of bio-degradable composite films. The brewers' spent grain protein was extracted by alkaline extraction, as described by Celus et al. (2007). It was determined that the addition of chitosan improved the physical and mechanical properties, including elasticity and tensile strength. The optimum levels of BGP and chitosan were determined to be 3 and 2%, respectively. In addition, the chitosan contributed to enhanced antimicrobial and antioxidant properties of the films, inhibiting the growth of *Staphylococcus aureus*, *Listeria monocytogenes*, *Escherichia coli* and *Salmonella typhimurium* [89]. The potential use of brewers' spent grain protein in biodegradable packaging materials is a very exciting potential use of the by-product.

Barley protein was also used to produce barley protein–gelatine composite films in a study by Song et al. (2012). The physical properties of this film were investigated, and it was found that increasing levels of barley bran protein caused tensile strength and elongation at break value to decrease. However, increasing the proportion of gelatine increased tensile strength but still reduced elongation value. It was determined that the optimal composition for film production was 3 g barley bran protein, 3 g gelatine and 100 mL sorbitol in 100 mL of film-forming solution. Grapefruit seed extract was also incorporated into the film as an antimicrobial agent and proved to be successful in reducing the growth of pathogenic bacteria when used in salmon packaging [90].

#### *6.3. Food Applications*

The implementation of brewers' spent yeast in bakery products has been widely studied. While these applications are not always protein-focused, in many cases, the protein fraction of BSG works synergistically with other compounds to improve nutritional and/or technological functionality. BSG addition to foodstuffs can improve protein content significantly but can also drastically increase levels of dietary fibre, which is very desirable for human health [48,91,92].

#### 6.3.1. Biscuits

The use of protein isolated from brewers' spent grain has the potential for use in bakery products to improve nutritional and functional properties. A study by Zong et al. [93] implemented spent grain protein in cookies and studied the effects of the addition on the sensory and textural characteristics of the product. This addition improved the overall flavour and nutritional value of the cookies [77].

#### 6.3.2. Bread

The use of BSG and fermented BSG in the fortification of bread has been explored by Waters et al. At 10% addition, dough displayed improved handling characteristics. As well as this, softness and staling were improved in both cases. While sweetness was decreased and acidulous flavour increased, both bread types were acceptable at a 10% addition level [48]. This acceptability, as well as the increase in nutritional value with regards to protein, dietary fibre and minerals, makes BSG a very interesting raw material for food product fortification.

The effect of BSG supplementation on bread doughs was also explored. In a study by Ktenioudaki et al. (2013), BSG and apple pomace were added to wheat dough, and their effect on the doughs physiochemical properties studied. The BSG was found to be high in protein (20.8%) and high in dietary fibre (60.5%). The rheological and pasting properties of the dough were greatly altered, mainly thought to be due to the high fibre content. Increasing by-product addition significantly reduced peak viscosity, holding strength, breakdown, final viscosity and setback values, as well as strain hardening index. Meanwhile, biaxial extension viscosity was higher for the supplemented dough and the storage modulus G was increased. These changes all indicate significant structural differences between the un-supplemented and supplemented doughs [94].

#### 6.3.3. Snacks

Dried and milled BSG has been used as a means to increase the protein content of extruded snacks [95]. BSG was incorporated at 10–30% levels in combination with wheat flour, corn starch and other ingredients and extruded using a twin-screw extruder. This addition significantly increased crude protein content, as well as increasing the phytic acid level and bulk density. Meanwhile, sectional expansion and cell area were reduced. A follow-up study examining the effect of altering water feed rates using a combination of BSG and different flours was performed. The total dietary fibre of the wheat flour and BSG (WBSG) and cornflour and BSG (CBSG) mixtures were found to increase significantly. Generally, TDF increased with increasing water feed level for WBSG sample, while the opposite was observed with the CBSG samples [96].

A similar study by Ainsworth et al. (2007) also looked into the effects of BSG addition in the formulation for a chickpea-based extruded snack. It was determined that BSG addition (0–30%) resulted in decreased expansion, which is in agreement with other studies. However, increased screw speed had the opposite effect. The BSG product also displayed increased phytic acid, resistant starch and protein in vitro digestibility [97]. A study by Ktenioudaki, Crofton, et al. (2013) investigated the use of BSG in a crispy snack product as a means to increase fibre content. It was found that a 10% BSG addition almost doubled the fibre content of the product without compromising product acceptability [92]. While often

used to increase protein levels in food products, BSG can also enhance other nutritional characteristics such as fibre content.

A study by Singh et al. investigated the potential use of novel drying methods on brewers' spent grain for use as a plant protein source in baked chips. Vacuum microwave drying (VMD) was proven to be an efficient method for drying BSG, as it reduced drying time, showed a high drying efficacy and a high overall acceptability in the baked snack when compared to those prepared using oven-dried or freeze-dried BSG. The applications of VMD technology are interesting from a sustainability standpoint as well as a nutritional standpoint, as the lack of a high temperature, due to the lowering of the boiling point by the vacuum, results in a higher-quality protein ingredient [98].

#### 6.3.4. Beverages

The research surrounding the inclusion of BSG-derived ingredients in beverage formulation is limited. A study by McCarthy et al. (2013) investigated the use of phenolic compounds extracted from BSG as antioxidants in fruit beverages and determined that BSG extract addition resulted in significantly increased antioxidant activity, as measured by the ferric reducing antioxidant power (FRAP) assay [5]. While other studies incorporating this concept are rare, several patents are available (Table 2) regarding the extraction and use of BSG-derived ingredients in beverage applications.


**Table 2.** Patents regarding the use of BSG ingredients in food and beverage products.

#### *6.4. Barley Protein Hydrolysates*

Due to a lack of solubility of barley proteins, functionality in food applications is limited. To improve ingredient functionality, protein hydrolysis can be implemented.

Celus et al. (2007) used enzymatic hydrolysis as a means to potentially improve the solubility, colour, emulsification and foaming properties, as well as cause a change in molecular weight distribution and hydrophobicity of BSG proteins. Differing degrees of hydrolysis were obtained by subjecting the protein concentrate to differing concentrations of three enzymes (Alcalase, Pepsin and Flavourzyme) for varying amounts of time. Generally, hydrolysis of the BSG protein improved emulsion and foaming. This somewhat agrees with a separate study by Yalçın et al. (2008), where barley protein hydrolysates were determined to have slightly improved foaming characteristics when compared to barley protein isolates, but the difference was slight [99]. However, those hydrolysates prepared with Alcalase and Pepsin showed a decrease in these characteristics with an increasing degree of hydrolysis. Characterisation of these hydrolysates showed that a relatively high molecular weight and a high surface hydrophobicity are desirable for enhanced physiochemical properties [64].

Enzymatic hydrolysis of barley proteins can greatly improve antioxidant capabilities and metal-binding activity [100–102]. A study by Chanput et al. (2009) examined the antioxidant properties of partially purified proteins from a variety of sources, including barley hordeins. Hydrolysates of these proteins were prepared using an enzymatic treatment of pepsin, followed by trypsin. Antioxidant and reducing properties were investigated, and the partially purified C-hordein displayed a high reducing capacity when compared to the B- and D-hordeins. It was generally found that for all protein hydrolysates, antioxidative and reducing capacities were greatly increased after enzymatic digestion with pepsin and trypsin [101]. A recent study by Ikram et al. (2020) followed on from this idea, investigating the effects of pre-treatments such as ultrasonic and heat treatments on the enzymatic hydrolysis of BSG proteins by Alcalase and the extent to which these pre-treatments, in combination with hydrolysis time, altered the antioxidant activity of the hydrolysates. The treatments selected were an ultrasonic treatment of 40 or 50 kHz and a heating treatment of 50 and 100 ◦C, while the pre-treatment times were 15, 30 and 60 min. The ultrasonic treatment at 40 and 50 kHz was shown to significantly increase oxygen radical absorption capacity values of the hydrolysates, while the heat treatment at 100 ◦C greatly increased the ferric reducing antioxidant power (FRAP) assay values [100]. These results indicate that hydrolysed BSG proteins exhibit a higher antioxidative power than untreated material and that these antioxidative capabilities can be further increased with ultrasonic and heat pre-treatments. Therefore, these hydrolysates could prove to be useful food ingredients due to their antioxidant and reducing potential.

Barley hordein hydrolysates also show potential as dietary supplements to enhance mineral bioavailability and solubility [102]. Through enzymatic digestion with Flavourzyme, Alcalase, trypsin and pepsin, barley proteins become hydrolysates with a strong metal ion binding capacity and can significantly increase the solubility of Fe2+ Fe2+, Ca2+, Cu2+ and Zn2+ [102]. Therefore, BSG proteins could potentially have a useful application in dietary supplements to enhance mineral bioavailability and solubility.

#### **7. Conclusions**

The increasing interest in sustainability and economising product side streams has led to an increased emphasis on the reuse and valorisation of brewing by-products. As brewers' spent grain is the most abundant by-product from this industry, the valorisation and utilization of spent grain protein is of great interest, particularly in terms of sustainability. Finding ways to upcycle this cheap and readily available product and apply it in a variety of different settings is a research area that is rapidly gaining traction. BSG is currently extremely underutilised and is mainly used in animal feed formulations due to its low cost and high nutritional value. High levels of essential amino acids in the proteins could be useful in nutritional and functional food applications for human consumption.

To date, BSG has been successfully applied in bakery products such as bread, biscuits and snack-type products, and BSG protein hydrolysates have been found to have increased functionality, including enhanced solubility, foaming and emulsification properties. These hydrolysates have also shown potential to enhance nutrient bioavailability as well as an increased antioxidative and reducing capability. All in all, BSG protein and its hydrolysates have significant valorisation potential, especially with regards to applications in the food industry.

**Author Contributions:** A.J.: writing, investigation; E.Z.: review/editing, funding acquisition; A.W.S.: project administration. E.K.A.: conceptualization, supervision, review/editing. All authors have read and agreed to the published version of the manuscript.

**Funding:** This project has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No. 818368. This manuscript reflects only the authors' views and the European Commission is not responsible for any use that may be made of the information it contains.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Foods* Editorial Office E-mail: foods@mdpi.com www.mdpi.com/journal/foods

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel: +41 61 683 77 34

www.mdpi.com

ISBN 978-3-0365-4632-2