*Review* **Novel Protein Sources for Applications in Meat-Alternative Products—Insight and Challenges**

**Marcin A. Kurek \* , Anna Onopiuk, Ewelina Pogorzelska-Nowicka, Arkadiusz Szpicer , Magdalena Zalewska and Andrzej Półtorak**

> Department of Technique and Food Development, Institute of Human Nutrition Sciences, Warsaw University of Life Sciences, Nowoursynowska 159c Street 32, 02-776 Warsaw, Poland; anna\_onopiuk@sggw.edu.pl (A.O.); ewelina\_pogorzelska\_nowicka@sggw.edu.pl (E.P.-N.); arkadiusz\_szpicer@sggw.edu.pl (A.S.); magdalena\_zalewska@sggw.edu.pl (M.Z.); andrzej\_poltorak@sggw.edu.pl (A.P.)

**\*** Correspondence: marcin\_kurek@sggw.edu.pl

**Abstract:** Many people are increasingly interested in a vegetarian or vegan diet. Looking at the research and the available options in the market, there are two generations of products based on typical proteins, such as soy or gluten, and newer generation proteins, such as peas or faba beans, or even proteins based on previously used feed proteins. In the review, we present the characteristics of several proteins that can be consumed as alternatives to first-generation proteins used in vegan foods. In the following part of the work, we describe the research in which novel protein sources were used in terms of the product they are used for. The paper describes protein sources such as cereal proteins, oilseeds proteins coming from the cakes after oil pressing, and novel sources such as algae, insects, and fungus for use in meat analog products. Technological processes that can make non-animal proteins similar to meat are also discussed, as well as the challenges faced by technologists working in the field of vegan products.

**Keywords:** protein; meat analog; texture; insect protein; algae protein

### **1. Introduction**

As consumer awareness of the environmental impact of food production increases, so does the consumption of products derived entirely from plants. This has to do with the narrative that meat production requires extensive land and water resources, negatively impacting the terrestrial and aquatic biodiversity and emitting greenhouse gases [1]. More and more people are also avoiding plant-based products, which are very interesting in terms of nutritional value, vitamins, micro and macro elements, and the ability to counteract some of the diseases prevailing among the civilization, due to their cholesterol and bloodpressure lowering properties [2]. This state of affairs influences the substantial growth of the meat analog market, which today is increasing expansively.

Most meat analogs are based on two proteins: soy and gluten. Soy protein is a good alternative to meat in terms of amino acid composition and textural properties [3]. However, the use of soy is quite controversial among consumers due to genetically modified (GM) crops. There is a study showing that 55% of consumers are opposed to GM foods and nearly 60% do not trust GM scientists [4]. Some consumers are strongly against GM soy application in feed for animals [5]. In contrast, the use of gluten ensures that a proper network is built in the product, but it is a fairly high allergenic raw material and may be avoided by some people [6]. This is because more people are diagnosed with celiac disease as well as gluten intolerance. Although important scientific advances have been made in the understanding of the pathologic mechanisms behind nonceliac gluten sensitivity, this disorder is still a matter of active debate in the scientific community [7]. More and more importance is being given to searching for alternative novel protein sources that can be used in meat analog products.

**Citation:** Kurek, M.A.; Onopiuk, A.; Pogorzelska-Nowicka, E.; Szpicer, A.; Zalewska, M.; Półtorak, A. Novel Protein Sources for Applications in Meat-Alternative Products—Insight and Challenges. *Foods* **2022**, *11*, 957. https://doi.org/10.3390/ foods11070957

Academic Editors: Jayani Chandrapala and Yonghui Li

Received: 7 February 2022 Accepted: 24 March 2022 Published: 25 March 2022

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

**Copyright:** © 2022 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/).

It is also worth noting that the cost of producing plant protein is significantly lower than the cost of producing animal protein. Of course, this translates into technological properties, but there are a number of methods that try to prevent this, such as protein texturization [8]. According to World Bank reports, there is an increase in demand for animal protein that cannot be met, hence the need for more intensive work on new sources of protein [9].

In 2020, the market value of plant-based-meat worldwide was estimated to be worth USD 6.67 billion. This figure is estimated to steadily increase over the next few years and reach roughly 16.7 billion in 2026. This is also influenced by the opinion on the safety of animal-based proteins, which are linked to epidemics such as mad cow disease, swine influenza, or avian flu that appear from time to time [10]. For this reason, this literature review is a systematization of knowledge in the utilization of novel protein sources according to origin—plant, microbial, fungal, insect, and algae.

#### **2. The Function of Protein in Meat-Alternative Products**

Proteins play an important role in human nutrition by providing building materials essential for both growth and cell regeneration. First of all, most of the meat alternative food sources from which proteins can be obtained differ in the composition of proteins themselves, as well as their amino acid profiles. Most often, meat alternatives do not use only protein isolates; concentrates or powders derived from plants, such as dietary fiber, vegetable fat, or carbohydrates, may also be included in the formulation. For many years, plant protein was considered to have lesser nutritional value, but this trend is now reversing [11].

Concerning nutritional values, it is worth noting that plant protein is not consumed as an individual ingredient but in a group with other ingredients. Therefore, it is not easy to control the potential effects of different nutrients from different meat alternative foods and attribute the observed benefits solely to the protein content. Furthermore, specific sources of plant and animal protein in the diet have been shown to have different health effects. Thus, general statements about plant or animal protein may be too simplistic, and effects may depend on the dietary matrix and accompanying nutrients.

Regarding nutritional properties, it is worth focusing on the fact that the meat analog has similar or comparable nutritional value to an identical meat-based product. As a rule of thumb, if a product has more than 30% protein with low-fat content, it can be considered a good meat substitute [12]. At the same time, it is also worth noting that substitutes or alternatives to meat products should be characterized by their similarity to meat protein digestibility-corrected amino acid score (PDCAAS) [13]. Supplementation or obtaining iron or vitamin B<sup>12</sup> from other sources is crucial if meat is excluded from the diet entirely.

Protein has a number of technological functions specific to a particular protein origin and degree of concentration—depending on whether it is a formulation, concentrate, or isolate. These characteristics include solubility, thermal stability, emulsification, flavor binding capacity, and digestibility score [14]. These characteristics are directly linked to the technological and functional roles in creating meat analog products.

From a technological point of view, it is not really possible to create a direct alternative to meat protein solely from the plant-based protein (except for the use of cultured meat target) product. This is due to aspects such as the reconstruction of the fiber network, which would reflect the myofilaments that are crucial for shaping tenderness and juiciness. Therefore, product development research on plant-based alternatives has largely been limited to restructured (or reconstructed) products [15]. This makes the alternative meat products mainly belong to two groups ranked by particle size. These can be divided into coarse-particle products, such as burgers, patties, sausage, meatballs, nuggets, etc., and fine-particle products, which are highly homogenized products that often have emulsion properties.

Many of the proteins are used in meat applications as ingredients due to their properties of functioning as a water binding agent. Several proteins are often combined together

for better results, such as pea protein isolate–wheat gluten blends and soy protein isolate– wheat gluten blends [16]. In terms of whole meat protein substitution, soy protein, which consists of albumin and globulin fractions, is the most common. In addition to soy protein, pea protein is also common. Most proteins derived from legumes possess the gel-forming ability, which is crucial because viscoelastic gel structure plays a major role in adhering particles, immobilizing fat, and entrapping water within the matrix of emulsion-type alternative protein products [17].

A widely used protein in meat alternative applications is gluten, which imparts the appropriate chewiness to products. The elasticity and extensibility properties of gluten are ascribed to two major protein fractions: glutenins and gliadins, which specifically influence the structure of meat products [18].

However, it is worth noting that soy is often associated negatively among consumers due to its strong association with genetic modification. At the same time, gluten is a highly allergenic protein and is not tolerated by people struggling with celiac disease.

#### **3. Protein Sources and Their Roles in Meat-Alternative Products**

#### *3.1. Legume Proteins*

In recent years, consumers have paid particular attention to plant-based diets. This is due to the increasing public awareness of the health-promoting effects of bioactive compounds from plants on human health and the willingness to reduce meat consumption for environmental reasons [19]. Of particular importance in the diet are legumes, whose effect on inhibiting diseases has been scientifically confirmed. These plants belong to the group of annual plants of the Fabaceae family of legumes [20]. Their edible part is the fruit, or the so-called pods, which are eaten whole or partially depending on the species and the degree of maturity of the fruit. The seeds of leguminous plants are characterized by their high nutritional value. Compared to other plants, they are distinguished by a fairly high protein content, ranging from 20% to 35% on average, depending on the type, growing conditions, and degree of maturity. Legume seeds are a rich source of dietary fiber, vitamins, minerals such as magnesium, iron, zinc, potassium, and phosphorus, and compounds with high antioxidant potential [21]. The seeds of these plants are low in saturated fats and, like all plant foods, are free from cholesterol [22,23]. A legume-rich diet improves bowel function and benefits hormonal balance [24]. Legume seed protein differs from cereal grain protein in amino acid composition—a significantly higher proportion of lysine (especially peas) and threonine, whereas the factor limiting its biological value is the insufficient content of sulphur amino acids (methionine and cystine) and tryptophan. In the protein of legume seeds, two fractions are distinguished: albumin and globulin. Albumins make up 10–25% of the total protein, can be soluble in water, and are mainly found in the germinal part. They are structural and enzymatic proteins, forming complex linkages with carbohydrates, lipids, and nucleic acids. The more albumin a seed contains, the greater its nutritional value. Globulins are soluble in dilute solutions of neutral salts. Different legume species provide varying amounts and qualities of protein to organisms [25]. Edible legume species include peas, lentils, lupins, chickpeas, broad beans, and mung beans.

Peas are an excellent source of protein and are exceptionally high in lysine and threonine, as well as other essential amino acids. They have a low glycemic index. Numerous scientific studies show that peas play a large role in preventing colon cancer and help treat breast cancer, pancreatic cancer, prostate cancer, lung cancer, and leukemia. Lentil seeds are also full of nutritional value because they contain 9/100 g protein and 0.4/100 g fat in edible parts; they are rich in iron, phosphorus, magnesium, and B vitamins. There are many types of lentils, including red, brown, green, yellow, and black lentils, among others [26].

Lupin seeds are another type of legume protein source whose nutritional value of lowalkaloid varieties is comparable to soybeans. Among legumes, lupin seeds contain the most protein (up to 46%) and the least undesirable non-nutrients. Due to the presence of functional components, they have potent health-promoting properties. They show antioxidant and hypocholesterolemic activity, have a low glycemic index, increase the bioavailability of

minerals, and have anti-allergic and anti-inflammatory effects [27]. Lupine seeds, as well as soybeans, can be used in the production of both traditional and functional foods.

Proteins from other legumes are also used in the production of meat analogs. In recent years, many studies have been conducted on the possibility of using chickpea, faba bean, and mung bean proteins in the production of meat analogs [28]. One of these is the study by Bühler et al., in which the researchers subjected faba bean protein concentrate to heating [29]. This led to modifications in the water holding capacity and solubility of the protein, achieving properties similar to soy protein concentrate, which is used in most meat analogs. This study showed that the choice of ingredients for meat analogs should depend both on the protein content and source and its nutritional value, but also on its thermal processing history, which can have a decisive influence on its technological properties. Among the previously noted three species of legumes, chickpea is the most consumed by consumers [30]. Chickpea protein is characterized by good properties in terms of texture, ability to bind water and oil, and ability to gel. The ability to stabilize emulsions and foam comparable to soy protein isolate and whey proteins is also an important property of chickpea protein. Moreover, chickpea protein isolate shows the ability to absorb more fat and a similar amount of water compared to soy protein isolate [31]. An essential advantage of chickpea is its positive effect on the color of the meat analog. Studies have shown that partial replacement of textured vegetable protein with chickpea flour significantly increased the color acceptability of meatless nuggets. The reason for this is the carotenoids contained in chickpeas [32].

Faba bean proteins may be a promising ingredient for producing meat analogs. Like chickpea proteins, they are excellent in stabilizing emulsions and foam, but to a lesser extent than soy protein isolate [32]. This is a limiting factor for using faba bean proteins as an ingredient in meat analogs. Many factors influence the technological utility of plant proteins. Thus, it is possible to improve the functionality of legume proteins as a result of appropriately selected parameters of production and processing processes. The study showed that dry fractionation enhanced the properties of protein-rich faba bean flour compared to faba bean protein isolate produced by acid extraction. Dry fractionation produced proteins with higher solubility at pH 7. The gelling and foaming abilities were also improved [28]. Faba bean proteins have been successfully used in the production of meat analogs by wet spinning, shear cell technology, and high moisture extrusion methods [33].

Mung bean proteins are also growing in popularity as an ingredient in meat analogs. The mung bean is a plant valued for both its nutritional value and functional properties. It is characterized by high protein (25–28%) and low fat content (1–2%). A significant amount of proline, glutamic acid, arginine, leucine, and phenylalanine is present in mung bean protein [34]. The limiting amino acid in mung bean protein is leucine. Notably, the digestible indispensable amino acid score (DIAAS) for this protein is 86, compared to 91 for soybean protein and 70 for pea protein. Mung bean protein is composed mostly of globular proteins, resulting in good gelling properties [35]. Like chickpea and faba bean proteins, mung bean proteins show the ability to stabilize foams and emulsions. Mung bean proteins are, therefore, used to obtain a balanced amino acid profile and desirable textural properties of meat analogs because they have globulins (60%, vicilin-type 8S with MW 26–60 kDa), albumins (25%, MW 24 kDa), and other globulins including basic-type 7S and legumin-type11S [17,34].

#### *3.2. Oilseeds Proteins*

In recent years, many oilseeds have been used as sources of protein in the food industry. The whole seeds and meals obtained from them are a valuable source of proteins with a well-balanced profile of essential amino acids with sulphur-containing amino acids. Their antioxidant, antihypertensive, and neuroprotective properties make them a valuable and functional alternative source of protein, e.g., in the baking and meat industry. The oil plants used as a source of protein include, inter alia, soybean, chia seeds, evening primrose,

flaxseed (brown), hemp seeds, milk thistle, nigella seeds, pumpkin seeds, rapeseed, sesame, safflower, glandless cottonseed, and sunflower seeds [36].

In addition, these proteins complement desirable functional properties when added to certain foods; this applies to whipping capacity, viscosity, emulsifying capacity, and water and oil binding capacity. Rapeseed and soybean protein isolates have a higher whipping capacity than sunflower, peanut, sesame, cottonseed, and safflower. Furthermore, the addition of sugar improves the whipping properties of the oilseed proteins. In contrast, if the oilseed proteins are heated, the whipping ability is reduced. Of the oilseed proteins, soy protein has the greatest emulsifying power. The emulsifying properties of heat-treated oilseed proteins are similar or better than that of animal proteins. Cotton seed protein has a very high water and oil binding capacity. However, the water-binding capacity of the oilseed proteins gradually decreases with increasing heating time at 100 ◦C. In contrast, heated oilseed proteins have an oil binding capacity similar to or better than that of unheated proteins [37]. In addition, the low allergenicity of pumpkin and hemp seeds or the potential non-allergy of evening primrose, milk thistle, black cumin, and chia compared to legume proteins makes it possible to use them as functional ingredients in newly developed food products [36].

Among the many benefits of proteins obtained from oil plant seeds, one should also remember the dangers of plants such as rapeseed that contain, in addition to many nutrients beneficial for the human diet, toxic erucic acid and sulfur compounds—glucosinolates, which are a component of the protein fraction [38]. To reduce the amount of anti-nutritive compounds (including glucosinolates, sinapin, and phytic acid) from proteins derived from oilseeds, innovative extraction methods are used.

#### *3.3. Cereal and Pseudocereal Proteins*

In grain-based proteins used in meat analogs, wheat, oats, or rice are used. The most common is wheat protein, which is gluten, due to its viscoelastic properties [1]. Other wheat proteins are not as often used as a base for creating meat alternative products, but due to their properties, they are fairly well distributed as structuring agents, even in true meat products.

A study conducted by de Angelis et al., indicated that oat protein isolates produced a rather good sensory effect when combined with pea protein [39]. However, the positive sensory properties were only observed after the extrusion process, which positively affected quenching the pea odor. The oats themselves were regarded by consumers for the pleasant smell but still far from being meaty.

Both legumes and cereals proteins contain pretty significant amounts of phytic acid, which is judged to be anti-nutritional by being strongly element restrictive. Some researchers have additionally introduced enzymatic activity and fermentation to reduce the phytic acid content of meat analog extrudates. The results were quite promising, but too much enzyme activity can end up degrading macromolecules, thus making it difficult to maintain an appropriate texture [40].

A very promising raw material for creating meat analogs is rice, an established low allergenicity raw material and, in particular, is presented as an alternative to soy. Raw rice was reported to be more allergenic than cooked rice, even though some allergens are heat stable and proteolysis resistant [41]. Currently, rice flour is being used as a substitute for fat while taking advantage of its water-binding properties in meat products. The use of 4–6% rice flour effectively increases the firmness of sausage-type meat products while being highly acceptable to consumers [42].

Cereals that are high in protein are pseudo-cereals like amaranthus and quinoa. Amaranth and quinoa grains are equally good as cereals and legume seeds because of their high content of lysine, arginine, tryptophan, and other sulphur-containing amino acids. Amaranth is an example of a plant with a high protein content of up to 14%. Some difficulty in obtaining protein is the isolation of starch in the case of amaranth [43]. Amaranth itself

also has a flavor that consumers may not fully accept. However, amaranth has already been successfully used as a binding agent in sausage formulations.

Another type of pseudo-cereal used in meat products is quinoa, a raw material with approximately 8% protein but a very high nutritional value containing all nine essential amino acids. The use of quinoa in meat products improved its water-holding capacity, reduced its toughness, and positively affected the sensory experience [44,45]. Further work on protein concentration and isolation from quinoa may lead to a good base combination for creating legume-based meat analogs.

When using grain-based proteins, they must be proteins with a fairly good amino acid profile. At the same time, existing concentrates or formulations have a widely accepted flavor and are not treated negatively. The ability to bind water means that in the future, they can be used as additional proteins in the composite to create meat analogs [30].

#### *3.4. Algae Proteins*

Algae, or photosynthetic eukaryotes, are distinguished as microalgae and seaweed. Microalgae is a huge group covering almost 200,000 species [46]. Out of this group, several species have been tested for a variety of purposes: food additives, cosmeceuticals, animal feed, or wastewater treatment. Foods obtained or formulated with the addition of algae are included in the definition of novel foods in the Novel Foods Regulation (EU). Microalgae is a promising novel ingredient that might be applied in the formulation of meat analogs. The growth rate of microalgae cultivars is superior to other crops used as sources of plant proteins. The estimated microalgae yield of dry biomass reaches 15–30 tonnes annually per unit area compared to 1.5–3.0 tonnes for soybeans. Microalgae and seaweed also contain more protein per unit area (4–15 tonnes/Ha per year and 2.5–7.5 tonnes/HA per year, respectively) in comparison to soybean (0.6–1.2 tonnes/Ha per year) or wheat (1.1 tonnes/Ha per year) [47]. Depending on strain and cultivation conditions, microalgae can produce up to 70% of proteins in cells compared to 30–40% for soybeans. Even more important from the quantity of protein occurring in algae is its quality. The nutritional quality of protein is determined by the composition of amino acids and the amount of essential amino acids. Two most dominant microalgae species on the market, Spirulina (*Arthrospira*) and Chlorella, are characterized by the higher than standard (100) essential amino acids index (102.6 and 107.5, respectively). Those values are similar to casin milk protein and higher than soybean meal [48]. There are microalgae of good essential amino acids (EAAs) balance. For instance, *Chlorella* contains 7 EAAs, comparable to beef but with a slightly lower level of cysteine and methionine. However, in most algae species, lysine and tryptophan are limiting amino acids [49]. Further, for brown algae except for the two noted above, also lysine, while for red species, leucine and isoleucine occur in low concentrations. In the case of seaweed, cysteine is most limiting, whereas glutamic acid and aspartic acid are most abundant [50].

Proteins acquired from microalgae exhibit techno-functional potentials such as high solubility and capacity to emulsify and form gels and foam. Solubility of *Chlorella protothecoides* proteins at pH 2–12 is estimated to be approximately 84.3%. For comparison, soybean protein (glycin) at pH 4.5–6.0 is soluble at less than 20% [51]. In turn, emulsifying and foaming are comparable to soy and whey proteins. Some species, such as *Chlorella vulgaris*, have even higher emulsifying properties. Algae proteins are also considered to be safe as food components. Those properties drew the scientific community's attention towards using algae proteins as a substitute for animal protein. Palanisamy et al., (2019) observed that adding *Spirulina* (*Arthospira platensis*) flour at a level of 30% to lupin proteinbased meat decreased in vitro protein digestibility from 82% to 76.5%. However, it was reversed partially by changing the process parameters [52]. Based on the data, the authors stated that *Spirulina* increased nutritional (higher antioxidant activity and phenolic content) and physico-chemical properties of the meat analog. Other studies revealed that adding spirulina at higher concentrations gives the product dark color, musty odor, and intensive earthy flavor [53]. Nonetheless, also in this study, setting the proper process conditionslow moisture content with high temperature and screw speed—enables partly replacing soy with spirulina in meat substitute and obtaining a product of decent flavor quality. Even though methods to produce microalgae rich in proteins on a large scale were invented about 50 years ago, still today, there exists only a few novel products formulated based on them. There are several reasons for that. First of all, algae dry matter contains 10% of the cellulosic cell wall, which is not digested and utilized by humans and non-ruminant animals. Thus, it is required to use various extraction and purification methods, thus increasing the costs of microalgae biomass application and limiting its use to high-value industries. Furthermore, algal protein concentrates are characterized by green and yellow colors and an unattractive fishy odor. Those attributes negatively influence consumers' perception of meat analogs formulated with algae addition. Sensory experiments showed that the product acceptance decreases with the increase in algae content [54]. Lowering prices was the only way to make eating meat substitutes with algae content more attractive [53]. Some researchers suggest that familiarity with food influences buying behavior and that algae meat analogs are unattractive for consumers because they are still unfamiliar to them. Nonentheless, to date, there is a lack of ideas for how to positively affect consumers' attitudes toward algae meat substitutes.

#### *3.5. Insect Proteins*

Insects are common food for 2 billion people in 119 countries across the globe [55]. There are over 2000 edible species. The most consumed insects that are used as protein sources are Coleopatra Beetles (31%), Lepidoptera Caterpillars (18%), Hemynoptera, wasps, bees, and ants (14%). However, those are still novel foods for Western countries. This is slowly changing due to growing need for alternative sources of proteins, production of which would be more sustainable. Studies on the life cycle assessment of Hermetia illucens performed by Smetana et al., (2019) revealed that insect protein concentrates had a lower environmental footprint than animal proteins but higher than plant proteins [56]. In accordance with studies conducted by Mason et al., (2018), the production of one gram of beef requires 21 times more water (16.8 g) than the production of the same amount of protein from cricket (0.7–0.8 g) [57].

Insects are a good source of proteins. The average content of proteins in them is 40% and ranges from 20% up to 70% depending on the species. Three species that are widely bred in Europe (Tenebrio molitor, Gryllodes sigillatus, Schisocerca gregaria) are considered to have the biggest potential as food components in the European Union and contain 52.35, 70, and 76% of proteins, respectively [58]. The amount and quality of proteins within the same species vary greatly depending on diet, metamorphic stage, or habitat. However, protein content is also often overestimated due to the presence in insects of a non-protein nitrogen. It has been estimated that up to 26% of whole larvae nitrogen may be nonprotein [59]. Insect proteins are more digestible (76–98%) than plant proteins (lentils 52%) and slightly less digestible than animal proteins (95% egg protein, 98% beef protein) [60]. The essential amino acids' score for insects ranges from 46% to 96%, which greatly exceeds the lowest recommended level for human diets (>40%). The quantity of the same amino acids is even higher in insects than those from plant and animal protein sources [61]. Insect proteins have high threonine and lysine content but low levels of methionine or tryptophan. Proteins acquired from insects are characterized by a low level of solubility ranging from 3% to 45%. However, the solubility may be improved by enzymatic hydrolysis. For instance, the major solubility improvement of migratory locust protein was observed to be 10–22% and up to 55%. Along with solubility, authors also observed higher emulsifying activity of approximately 54%, enhanced foam ability of approximately 326%, and improved oil banding capacity [62]. Thus, the application of insect proteins is recommended for foods that do not require high solubility, such as meat analogs. Furthermore, insect proteins are especially recommended as an addition to plant meat analogs to improve its protein profile. Smetana et al., observed that using the highest temperature of a barrel extruder (170 ◦C) made it possible to introduce 40% of insect protein to a soy-based meat analog,

keeping its optimal meat-like texture [63]. In turn, Kim et al., (2022) performed studies on the usage of insect proteins along with textured vegetable proteins to produce restructured jerky analogs [64]. In conclusion, the authors of the studies stated that it is possible to produce meat analog combining both of those proteins to get tender jerky of high nutritional value. There is also a study aiming to partially replace meat protein (10%) with insect flour (*Tenebio molitor* or *Bombyx mori*). The results of this experiment indicate that even though they obtained high-value emulsion sausages, those were harder than control meat samples. However, consumer safety is also an issue. There is a risk of an allergic reaction after consuming insect proteins, which contain tropomyosine and arginine kinase—two major proteins responsible for allergic reactions. Furthermore, insect-derived food and feed might be contaminated chemically with heavy metals and biologically with spore-forming bacteria [65].

#### *3.6. Edible Fungus Proteins*

Mushrooms have been classified into a separate kingdom because of their different cellular organization, and they do not belong to either animals or plants [66]. Fresh edible fungus has about 90% water, and the remaining 10% dry matter is composed of 8–40% protein, 3–28% carbohydrate, 3–32% fiber, 2–8% fat, and 8–10% ash, varying with the mushroom species and other factors [67]. Yu et al., (2020) examined 23 edible mushrooms and determined their protein content. It was found that the protein content in edible mushrooms was approximately 8.5–36.9%, which was much higher than that of vegetables, fruits, and grains. The higher protein content was found in *Trichloma* (36.87%), and Tremella had the lowest protein content (8.46%). Other more popular mushrooms like *Shiitake*, *Lentinus Edodes*, *Volvariella Volvacea*, and *Boletus* had a protein content of 15.38%, 11.59%, 10.24%, and 12.16%, respectively. Fungus proteins are gaining more and more popularity all over the world. As meat production has a significant impact on the environment, it is important to find a cheap, alternative, and less resource-intensive source of protein to partially replace meat or meat products. Mushrooms cannot be considered as a significant source of proteins compared to meat sources, even though they are a part of human nutrition mainly because of their taste [68]. Other authors, however, believe that mushrooms may play an important role in meat analogs by providing nutrients and promoting the development of sensory properties such as appearance, texture, and taste of the product [69]. The use of mushrooms as an alternative source of protein in the human diet is not a new concept. Edible mushrooms can be treated as a functional food due to their nutritional value. The use of edible mushrooms has been used in meat products as meat substitutes or fillers to improve the physicochemical and sensory characteristics and their nutritional value. The production of mycoprotein products is based on submerged fermentation of fungi in a liquid culture medium [70]. The mycoprotein production is based on the continuous fermentation of the filamentous fungus *Fusarium venenatum* on a glucose substrate, which allows the production of a high-protein, low-fat food ingredient [71]. They are usually grown in bioreactors with a high metabolic rate. Miller and Dwyer (2001) assessed the tolerance of humans to mycoprotein, and the results demonstrated that mycoprotein is well tolerated by humans and has an extremely low allergenic potential [72]. Singh et al., (2021) indicate that the mycoprotein of some fungi is a good source of protein. Still, due to its low digestibility, it is rarely used to prepare meat analogs, although *Fusarium venenatum* is cultivated to derive mycoprotein and prepare meat substitutes [1]. The mycoprotein may have a meat-like texture and flavor. Some researchers argue that proteins produced using mycoproteins have structures similar to muscle fibers of meat and claim that mycoproteins can be considered as an alternative source of the food protein. Due to their functional properties, it is possible to use them in new attractive health-promoting food products. The use of biotechnological methods for their production creates an opportunity to reduce production costs and improve the sensory and nutritional properties [73]. The harvested mycoprotein can be used to prepare vegetarian sausages, burger patties, or minced cutlets. Other mushroom-based meat substitutes are produced from *Aspergillus oryzae*. Filamentous

mushrooms are used in most mushroom-based meat products because their long fibers create a meat-like texture. Denny et al., (2008) stated that the mycoprotein may have a meat-like texture and flavor and is the main component of various mycoprotein products, including minced meat, chicken pieces, burgers, sausages, nuggets, fillets, ready to eat meals, cakes, and pies [71]. In many Asian countries, *Monascus purpureus*—treated with yeast produce red rice—and *Aspergillus oryzae*—fermented with soy—is used in hamanato, miso, and shoyu. Nowadays, in the European market, Quorn™, a meat substitute originated in Great Britain, is sold. Quorn™ contain mycoprotein derived from the *Fusarium venenatum* filamentous fungus [74]. Mushrooms and fermented products have a meaty taste, a long shelf life, good nutritional values, and reduced cooking time, so they can be a new generation of plant proteins in the future. All real products in which the novel sources of proteins were used are summarized in Table 1.

#### **4. Processing of Proteins Applied in Meat-Alternative Products**

Textured vegetable protein (TVP) was one of the first ingredients used in the production of meat analogs. The TVP production technology was developed in the 1970s, and it was then that this type of product was introduced to the market for the first time. Initially, TVP was used as a filler in various conventional food products. In the following years, the development and production technology of meat substitutes based on TVP began. The primary raw material for TVP production is soy proteins, although other ingredients such as cotton, wheat, and corn are also used. Nevertheless, the TVP consists mainly of processed dried soy flour to give it a spongy texture and is flavored to improve the meat-like sensory properties. TVP is produced in the extrusion process (Figure 1). Hightemperature (120–200 ◦C) and high-pressure (20 MPa) processing of the raw material make it possible to obtain products of various shapes (such as cubes or stripes), sizes, colors, and textures [17,75].

**Figure 1.** Process of texturization of proteins for application in meat analog production.

In the 1980s, the fiber spinning technique began to be used to produce meat analogs. In this method, the alkaline protein solution was forced through the matrices into the acidic coagulation base. This led to the precipitation of fibers that were mixed with bonding materials. However, the process was very complex, required a highly concentrated protein solution, and had lower yields than large-scale production costs [76].

Currently, the main technology for producing this type of product is thermoplastic extrusion. Extrusion used in recent years is a method characterized by high efficiency and allows for the reduction of the energy cost of production. Skimmed vegetable proteins are made with the addition of water, salts, carbohydrates, lipids, flavors, and other functional additives. The mass is then put on the extruder screw where the product structure is shaped under the influence of high temperature and pressure [77].


**Table 1.** Summary of real products where the novel sources of proteins were applied.

To meet the expectations of customers, the production of meat analogs focuses on obtaining acceptable sensory characteristics such as taste, smell, color, and consistency. Two methods of extracting proteins used for the production of meat analogs are known as 'dry' and 'wet' extrusion. Unfortunately, "dry extrusion" (humidity <30%) does not allow for obtaining a sensorially acceptable product. In contrast, 'wet extrusion' (humidity 40–80%) enables the production of meat analogs of premium quality. The preparations obtained through this method are characterized by a consistency resembling real meat, and their appearance and mouthfeel are similar to cooked meat [52]. Due to the use of high moisture extrusion (HME), it is possible to produce from raw materials with low solubility, and in addition, this method is more economically viable [80]. Meat analogs produced by HME from soy protein are the most common [78], but using this method, it is possible to obtain high-moisture meat analogs (HMMAs) from plants, such as hemp, yellow pea, lentils, and faba bean [64,81,82].

The latest technical solutions are based on Couette thermostatic shear, in this case, the suspension of proteins and gluten gels in a linear flow. Because of this process, it is possible to obtain a product characterized by a fibrous structure. Moreover, shearinduced structuring with a high-temperature shear matrix created fibrous protein structures. The developed closed-chamber rheometer allows you to control thermal and mechanical stresses. Due to this, it is possible to obtain conditions similar to extrusion [79].

#### **5. Challenges for Protein Applications in Meat-Alternative Products**

Although the current development trend is towards developing foods for vegetarians, almost every product has similar challenges. These are not only technological but also consumer or even sociological challenges.

From a sensory point of view, achieving a viable alternative to a meat product is quite difficult because the specificity of meat in terms of amino acid structure, peptide sequences, and intermolecular connections is very specific and impossible to counterfeit. Sensory properties and, in particular, mouthfeel are influenced by a texture with very low granularity that is able to bind water. In order to maintain these characteristics, plant proteins must be subjected to several different structuring processes, such as thermomechanical extrusion or shear. Despite the use of a number of methods that alter the structure of plant proteins or increase water-holding capacity, there are still many difficulties that need to be solved. One of them is juiciness, which is a specific characteristic of meat, resulting from water absorption and the form of water-binding with proteins and in their fibers. Hence there is currently no possibility to replace meat with proteins of the same or similar texture.

Although plant proteins are the most common alternative to meat proteins, they have a particular taste that is different from meat. For example, in legume-protein products, an aftertaste derived from a characteristic beany odor is thought to be related to the secondary lipid oxidation products [83].

In addition to texture and palatability issues, meat products are characterized by a red, reddish, or pink color that, for obvious reasons, is impossible to achieve without the use of colorants. Unfortunately, the problem is present because many consumers interested in vegetarian products are consumers who avoid additives, which further increases the technological difficulty [84]. It is the lack of a clean label that makes consumers uncomfortable with meat protein product alternatives. Vegetarian products that are alternatives to meat protein products often contain a very high amount of ingredients like preservatives, stabilizers, colorants, or thickeners [17].

The protein alternative must also be a nutritional alternative, which is understood by adequate nutrient density. Unfortunately, because protein source alternative products are highly processed products, their nutritional value is not the same as meat products produced directly from raw meat. This is mainly because the protein used to produce the alternatives is already processed by heat and other methods. There is still no clear confirmation whether replacing meat protein with vegetable protein does not negatively impact human health by reducing the supply of heme protein, zinc, or selenium, which are characteristic of products based directly on meat. Meat processing methods and meat alternatives such as grilling, roasting, frying, and baking are considered methods that can lead to increased concentrations of carcinogenic substances such as heterocyclic aromatic amines [85]. However, using polyphenolic substances in plant-based products is easier than adding to meat products, which may lead to reduced HAA formation [19].

#### **6. Conclusions**

Today, more and more consumers are turning to vegetarianism or looking for products that are not based on animal products. This is understandable from the point of view of worldview, religion, or often just the search for new tastes. In most meat analogs, we encounter proteins of soy origin and wheat origin, like gluten. Unfortunately, both of these proteins are allergenic, and additionally, soy is associated with GMO crops, which are also negatively perceived by some people. The development of a range of meat analog products is possible by using novel sources of protein as well as methods of processing. These can be raw materials rich in protein, such as legumes, or by-products of various processes, as in the case of oilseeds cakes. Novel sources of protein are algae, insects, and fungus. With texturization technology, it is possible to obtain a product of sufficient quality in terms of texture. At the same time, it is worth bearing in mind that it is almost impossible to obtain the texture of meat, so analogs can only be suitable analogs of meat products after processing.

The increase in demand for plant-based protein will certainly be seen in future years as we look for new sources of protein to meet the needs of a growing population. In developed countries, more consumers are turning to vegetarianism and veganism, which will also contribute significantly to the demand for such products. However, a certain unmatched element will be the elaboration of not only the nutritional, but more importantly the physical and technological properties that meat protein possesses. Some hope is offered by zoonotic sources such as insects and from single-celled organisms such as algae because of their easy modification.

**Author Contributions:** Conceptualization, writing—original draft preparation—M.A.K.; writing original draft preparation—A.S., M.Z., A.O. and E.P.-N., supervision—A.P. All authors have read and agreed to the published version of the manuscript.

**Funding:** The research reported in this manuscript has been financed by the Polish Ministry of Science and Higher Education within the fund from the Institute of Human Nutrition Sciences, Warsaw, University of Life Sciences (WULS), for scientific research.

**Data Availability Statement:** No new data were created or analyzed in this study. Data sharing is not applicable to this article.

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

#### **References**


**Andrea Rivera del Rio , Remko M. Boom and Anja E. M. Janssen \***

Food Process Engineering, Wageningen University, 6700 AA Wageningen, The Netherlands; andrea.riveradelrio@wur.nl (A.R.d.R.); remko.boom@wur.nl (R.M.B.) **\*** Correspondence: anja.janssen@wur.nl

**Abstract:** Plant protein concentrates and isolates are used to produce alternatives to meat, dairy and eggs. Fractionation of ingredients and subsequent processing into food products modify the technofunctional and nutritional properties of proteins. The differences in composition and structure of plant proteins, in addition to the wide range of processing steps and conditions, can have ambivalent effects on protein digestibility. The objective of this review is to assess the current knowledge on the effect of processing of plant protein-rich ingredients on their digestibility. We obtained data on various fractionation conditions and processing after fractionation, including enzymatic hydrolysis, alkaline treatment, heating, high pressure, fermentation, complexation, extrusion, gelation, as well as oxidation and interactions with starch or fibre. We provide an overview of the effect of some processing steps for protein-rich ingredients from different crops, such as soybean, yellow pea, and lentil, among others. Some studies explored the effect of processing on the presence of antinutritional factors. A certain degree, and type, of processing can improve protein digestibility, while more extensive processing can be detrimental. We argue that processing, protein bioavailability and the digestibility of plant-based foods must be addressed in combination to truly improve the sustainability of the current food system.

**Keywords:** plant protein digestibility; protein isolates; protein concentrates; alternative fractionation; heat treatment; protein modifications; in vitro protein digestion

#### **1. Introduction**

The current food production system is not sustainable [1]. The largest environmental impact can be attributed to the production of animal-based protein [2]. One of the measures proposed by Willett et al. [3] to reduce this negative impact is to lower our consumption of foods of animal origin and to increase that of plant-based foods. To facilitate this transition, plant-based alternatives to meat, dairy and eggs are continuously introduced in the market. With the increase in flexitarian diets, there is a growing demand in the vegetarian and vegan food space [4]. Nevertheless, the extensive transformation and purification of the ingredients, in addition to the lower protein yield from crop to food product, limits the sustainability potential [5,6].

Generally, plant proteins present less favourable techno-functional properties compared to milk proteins, particularly those depending on solubility such as gelling, emulsifying and foaming properties [7]. In addition, it is not clear yet whether plant- and animal-based proteins can be interchangeable from a nutritional point of view. The dietary requirement of indispensable amino acids (AA) can be satisfied by proteins from various crops [8,9]. Antinutritional factors, digestibility and bioavailability must also be considered when assessing the nutritional quality of proteins. The in vivo protein digestibility-corrected amino acid score (PDCAAS) [10] and in vitro digestibility [11] of some protein-rich ingredients and whole foods have been reported. Furthermore, the effect of domestic and industrial processing on digestion of proteins from legumes consumed as a whole food or flour, i.e., not as a protein-rich ingredient, has been reviewed [12].

**Citation:** Rivera del Rio, A.; Boom, R.M.; Janssen, A.E.M. Effect of Fractionation and Processing Conditions on the Digestibility of Plant Proteins as Food Ingredients. *Foods* **2022**, *11*, 870. https://doi.org/ 10.3390/foods11060870

Academic Editor: Yonghui Li

Received: 22 February 2022 Accepted: 16 March 2022 Published: 18 March 2022

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

**Copyright:** © 2022 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/).

Plant proteins are diverse, and most constitute a mixture of various protein units, each with its own properties. For instance, varieties of the same legume species have different globulin to albumin ratios. Globulins have been found to be more susceptible to hydrolysis by digestive enzymes [13,14]. Moreover, 7S and 11S globulin-rich protein fractions from hemp protein isolate (PI) presented different in vitro digestion profiles [15]. Yang et al. [16] found that higher proportions of β-7S subunits had a detrimental effect on the in vitro digestibility of soybean PI. Protein concentrates (PC) from different cultivars of the same species can present different structural, thermal, techno-functional properties and nutritional value, such as the indispensable AA content and digestibility, as was found for rice and millet proteins [17,18]. Meanwhile, different varieties of lupin and sorghum differ in composition and structural properties but are digested to a similar extent [19,20]. This already suggests that the digestibility of proteins from different plant sources might not be affected in the same way by a given type of processing.

The objective of this article is to review the large body of data on the digestion of protein-rich ingredients and on how processing, before, during or after the extraction of the ingredient, may alter it. We recognize the breadth of protocols used to simulate digestion as well as the methods used to describe or quantify the extent of it (Figure 1). As these confounding factors contribute to variations in results, we limited this review to studies that compare some treatment or processing to a control and noted the effect on protein digestibility of a given ingredient.

**Figure 1.** Characteristics of in vitro assays, treatment of digesta and description or quantification of *digestibility* in the studies reviewed. AA, amino acid; HP SEC, high performance size exclusion chromatography; Mw, molecular weight; OPA, *o*-phthalaldehyde; SDS PAGE, sodium dodecyl sulphate–polyacrylamide gel electrophoresis; TCA, trichloroacetic acid; TNBS, trinitro-benzenesulfonic acid.

There is some disparity in the number of studies favouring some types of processing over others, as well as some crops over others. Moreover, the wide range of digestion assays makes it relatively futile to quantitatively compare results from different studies. We therefore present a narrative review with elements of a systematic one, instead of a full systematic review with meta-analysis.

#### **2. Method and Definitions**

Review characteristics: The search query used in Scopus was: ("protein" W/6 digest\*) AND "in vitro" AND "human" AND ("gastric" OR "intestinal" OR "gastrointestinal" OR "pepsin" OR "trypsin"). In PubMed, the MeSH terms for "plant proteins, dietary" and "digestion" were also included. From the results, the works considering some measure of digestibility or protein hydrolysis by digestive enzymes, simulating some physiological condition(s), were included. Studies on whole foods or flours were not considered, as these sources have been studied elsewhere. Articles studying feed, e.g., for ruminal digestion, emulsions, animal-sourced foods or proteins, and works dealing with allergenicity or immunoreactivity were excluded.

The term "protein digestibility" is used rather ambiguously throughout the reviewed literature. By definition, digestibility is the proportion of an ingested food or nutrient that can be absorbed into the bloodstream or body. However, it is also used to describe protein degradability, i.e., the proportion of intact protein remaining, the resulting degree of hydrolysis (DH) or the proportion of low molecular weight peptides resulting after the action of digestive enzymes. Other measurements of digestibility are listed in Figure 1 and details of the digestion assays and measurements for each of the studies reviewed are listed in the Supplementary Materials.

Figure 2 presents a scheme of the different processes reviewed. Throughout the text, "conventional aqueous fractionation" refers to milling, optional defatting for oilcontaining seeds, alkaline extraction, centrifugation, isoelectric precipitation, centrifugation, washing and freeze drying, as it is mostly performed in laboratory setting, or spray drying, more common in commercially available ingredients. Table 1 summarizes the effects on digestibility of the more commonly studied processes for different plant sources.

**Figure 2.** Overview of the processing steps before, during and after fractionation of plant proteins from the studies included in this review. Colours indicate the different routes for processing, the conventional route for aqueous fractionation is presented in black, and represents centrifugation after alkaline extraction and isoelectric precipitation.

**Table 1.** Overview of the effect of different types of processing before, during or after protein fractionation from different crops. , negative; , positive; or neutral effect on protein digestibility. Only processes or ingredients with more than one study reporting on the effect of processing on digestibility were included in this table.


#### **3. Ingredient Preparation**

#### *3.1. Pre-Fractionation Treatment*

Most commonly, seeds are milled into a flour or grits prior to alkaline extraction. Soaking seeds at high temperatures, before milling for conventional aqueous fractionation, was shown to improve the in vitro digestibility of soybean and cowpea PI. In the work of Wally-Vallim et al. [21], PI from soybean seeds soaked at 40 ◦C was more digestible than at 60 ◦C. The in vitro gastric digestibility was improved by longer soaking times for both temperatures. It was argued that at 40 ◦C, proteins were partially denatured, while at 60 ◦C the 7S fraction was completely denatured, and protein structures had rearranged. Meanwhile, PI from soaked and autoclaved cowpea seeds was more extensively hydrolysed by pepsin–pancreatin than that from raw seeds [22].

Some studies explored the effect of germination prior to fractionation of soybean and black bean. A direct relation between the germination time and the extent of hydrolysis achieved by digestive enzymes was observed [23,24]. Concurrently, the trypsin inhibitory activity (TIA) was reduced by germination, associated to protease-catalysed hydrolysis of lectins and trypsin inhibitors. Aijie et al. [25] found a similar relation; however, the DH decreased, and the TIA increased for the longest germination times, which they explained by a resynthesis of trypsin inhibitors by photosynthesis. For black soybean, an inverse relation was observed: the PI produced from non-germinated seeds yielded the largest proportion of low molecular weight peptides [26]. It was hypothesised that these small peptides were used for tissue formation during germination.

Solid state and submerged fermentation of milled lupin with different strains of *Pediococcus* prior to subsequent conventional aqueous fractionation improved the in vitro protein digestibility in the PI compared to the non-fermented control [27]. At the same time, the fermentation reduced the content of trypsin inhibitors. No clear relation can be drawn between the type of fermentation and digestibility, as many different lupin hybrid lines and strains of *Pediococcus* were studied.

#### *3.2. Conventional Protein Fractionation*

After a defatted meal has been obtained, alkaline extraction is the first step in conventional aqueous fractionation. Higher protein purities, at the expense of lower yields, can be obtained with increasingly higher concentrations of a strong alkali, typically NaOH. Alkaline treatment has been associated with the formation of lysinoalanine and AA isomerisation in rice residue PI, reducing the in vitro digestibility and absorption in a rat model [28]. Protein extracted from defatted lupin meal at acidic pH (pH 2) was more readily and extensively digested than that extracted at neutral or alkaline pH (pH 8.5) conditions, using an in vitro digestion assay [29]. The extraction pH was thought to induce different structural conformations and extents of denaturation. Nevertheless, Ruiz et al. [30] did not find a significant effect on the in vitro gastric digestion of quinoa PI extracted at pH 8 to 11.

Either PC or PI can be obtained from the conventional fractionation process. Commercial PC and PI have been used in in vivo rat assays, showing a small variation in PDCAAS, the true or standard ileal digestibility, of soybean ingredients [31,32]. Meanwhile, the in vitro gastric digestibility of commercial soybean PI remained unchanged after long-term storage at freezing and high temperatures [33].

#### *3.3. Alternative Protein Fractionation Strategies*

Modifications to the conventional aqueous fractionation process have been proposed to improve the purity, yield or techno-functional properties of the ingredients obtained. Conventionally, alkaline extraction is performed with NaOH, with the pH adjustment for isoelectric precipitation performed with HCl. Chamba et al. [34] proposed the use of alkaline ash from burnt green and purple amaranth and lemon juice as "natural" alternatives to the more commonly used chemicals to isolate soybean protein from full fat and defatted flour. The PDCAAS was slightly higher for the material extracted with "natural" chemicals, while

no significant difference was observed between the in vitro pepsin–pancreatin digestibility of "natural" and conventional chemicals. The use of conventional chemicals was somewhat more effective at reducing the content of antinutritional factors such as trypsin inhibitors and phytic acids in PI. Na2SO<sup>3</sup> has been used to extract proteins and to prevent oxidative darkening of the PI, from lupin and chickpea. The digestibility of Na2SO3-extracted lupin PI was higher than the conventionally fractionated ingredient [35]. However, for chickpea PI, the digestibility from both extractions did not differ [36].

Ultrafiltration has been used as an alternative to isoelectric precipitation. The TIA was similarly reduced by either process for soybean PI [37]. The extent of hydrolysis achieved with pepsin–pancreatin digestion, as well as the reduction of the TIA, was comparable for brown lentil PI separated by ultrafiltration and for conventional isoelectric precipitation [38].

The effect of different drying methods on the protein digestibility was studied for buckwheat and hempseed PI. Tang [39] showed that freeze drying, compared to spray drying, produces buckwheat PI that is better digestible by pepsin–trypsin. However, when alkaline extraction was assisted by ultrasonic treatment instead of by just mechanical stirring, freeze- and spray-dried PI were equally digestible. Meanwhile, Lin et al. [40] compared vacuum oven, oven or freeze drying of hempseed PC. In this study, freeze drying also produced better digestible PC compared to drying at higher temperatures, which was attributed to the formation of poorly digestible Maillard products during oven or vacuum oven drying.

Enzyme-assisted fractionation paired with extrusion has been presented as an environmentally friendly alternative to conventional aqueous fractionation [41]. Oil and protein were simultaneously extracted from soybean flakes that were extruded and treated with a bacterial endoprotease under alkaline conditions to obtain oil-, fibre- and protein- and sugar-rich fractions. Extrusion or enzyme action during processing did not alter the pepsin digestibility of the resulting ingredients, although some techno-functional properties were improved. Extrusion and α-amylase-catalysed starch liquefaction were used to concentrate proteins from white sorghum [42]. While the moisture content in the barrel during extrusion influenced the in vitro gastric digestibility, no effect from α-amylase action was observed. Nevertheless, the sorghum PC showed lower digestibility than sorghum flour. This was attributed to re-aggregation during the boiling step that was used for enzyme inactivation.

Air classification is a dry fractionation technique. The digestibility of pea, lentil and fava bean PC obtained from air classification were compared to that of NaCl-extracted PI from aqueous fractionation in a mice study [43]. Overall, the digestibility of the PC was lower than that of the PI, most significantly for pea. Likewise, air-classified fava bean PC was less extensively hydrolysed during pepsin–pancreatin digestion than a PI from isoelectric precipitation and spray drying [44]. Further, the TIA from the initial flour was maintained in the air-classified ingredient and significantly reduced in the conventionally produced PI. Conversely, air jet-sieved quinoa PC was slightly more extensively hydrolysed by pepsin than a conventional aqueous-fractionated PI [45]. We hypothesize that the protein denaturation achieved through heating during spray drying facilitates the access of digestive enzymes to the cleavage sites within the proteins.

#### **4. Post-Fractionation Processing**

Protein ingredients are further processed into finished products. The effects of different protein steps (fermentation, ultrasound treatment, heating, protein modification, among others) have been researched on PI and PC from various crops. Ultrasonic treatment of fava bean PI dispersions slightly reduced the in vitro digestibility [46].

Fermentation of commercial pea PC with *Lactobacillus plantarum* had a positive effect on the in vitro protein digestibility and a reduction of antinutritional factors, phenols, tannins, chymotrypsin and trypsin inhibitors. Nevertheless, the in vitro PDCAAS was negatively impacted. This was explained by the catabolism of sulphur-containing AA by the lactic acid bacteria [47]. Similarly, *L. plantarum*-fermented soybean PI released more

free AA than the non-fermented control, in a dynamic in vitro gastrointestinal digestion assay [48]. Additionally, protein aggregation was observed in the gastric phase only for the non-fermented PI, as well as a higher proportion of high molecular weight peptides at the beginning of the intestinal phase.

#### *4.1. Proteolysis*

Protein hydrolysis has mixed effects on protein digestibility. For soybean protein, hydrolysis by immobilized trypsin improved or had no effect on the extent of digestion [49]. In this study, pre-digested proteins were better digestible under infant gastric condition, simulated by a less acidic pH (pH 4) compared to adult models. Meanwhile, a soybean protein pepsin–hydrolysate was as digestible as the intact PI, in a different infant model with reduced digestive enzyme concentration, compared to an adult model [50].

A series of studies investigated the effect of the co-ingestion of soybean PI and dietary actinidin from green kiwifruit extract on the protein digestion. From an in vitro pepsin– pancreatin assay, some subunits such as the 11S basic polypeptide showed some effect of the actinidin; however, no overall effect on the protein degradability was observed [51]. From an in vivo rat study, the presence of actinidin in the diet showed no significant effect on the true ileal digestibility of soybean PI [52]. Gastric chyme samples from a subsequent rat study were analysed for their true gastric total protein digestion [53]. The presence of actinidin here improved the gastric digestibility of the PI. Meanwhile, actinidin had a positive effect on the digestibility of zein but had virtually no effect on the digestibility of wheat gluten. These studies highlight the relevance of the type of assay and measure of digestion to assess the effect of processing or modification on plant protein digestibility.

Green lentil PI from conventional aqueous fractionation was hydrolysed with acid protease, actinidin, bromelain and papain, prior to in vitro digestion [54]. Intact proteins proved to be better hydrolysable than the protein hydrolysates. Nevertheless, as a net result, more low molecular weight peptides were produced from the protein hydrolysates than from intact PI.

Hydrolysis positively affected the digestibility of rapeseed and rice bran PI. Fibre and protein from a rapeseed PI that was obtained by membrane processing were hydrolysed [55]. The true digestibility of the hydrolysate was higher than the intact PI, as shown by a rat assay. As a result, the PDCAAS of the hydrolysate was also higher, compared to the original ingredient. Similarly, for progressively higher degrees of hydrolysis, a papain–hydrolysate of rice bran PC was more extensively digested than the intact PI by pepsin–pancreatin digestion [56].

Chickpea protein hydrolysis did not alter the digestion. Neither alcalase, flavourzyme [57], trypsin, papain nor pepsin [58] changed the extent of protein digestibility in in vitro assays. Nevertheless, the TIA was significantly reduced by the hydrolysis [57].

#### *4.2. Heat Treatment*

The process step most studied in terms of its effect on protein digestibility is heat treatment. Different conditions as well as different crops have been studied with positive, neutral or negative effects of heating on protein digestibility.

It is commonly thought that a certain extent of heat induced protein denaturation improves the digestibility, while more extensive heat treatment would induce protein aggregation which would, in turn, reduce the digestibility. The work of Tian et al. [59] demonstrates the relation between heating time and temperature, and the extent of pepsincatalysed hydrolysis of soybean PI. Dispersions heated at 85 ◦C for 15 min presented the highest DH, while those heated at 70 or 100 ◦C were hydrolysed to a significantly lesser extent. In terms of time, PI heated at 85 ◦C for 20 min showed the highest DH compared to those heated for 10 or 60 min. Overall, all heated samples were more extensively hydrolysed than the unheated control.

Soybean is one of the crops most widely studied in terms of the effect of heat treatment on protein digestibility. Studies have shown improvement but also reduction of protein digestibility as a result of heat treatment. The in vitro pepsin–pancreatin digestibility of soybean PI was improved by relatively short heating for 15 min at 95 to 121 ◦C [60–62]. β-conglycinin is known to be less susceptible to pepsin-catalysed hydrolysis than glycinin. Nevertheless, the gastric digestibility of both fractions was improved by heat treatment [60]. In this study, heating induced protein aggregation as well as pepsin during the gastric phase. The TIA of germinated soybean PI was reduced by the heat treatment [25]. Conversely, the apparent digestibility of heated, spray-dried and autoclaved pastes of soybean PI, determined in a rat assay, was significantly lower than that of non-autoclaved pastes [63]. Besides the heat treatment during drying, these pastes were autoclaved for up to four hours, highlighting that extensive heat treatment, both in time and temperature, has a detrimental effect on protein digestibility.

Heat treatment does not affect the digestibility of different pulse protein ingredients in the same way. Heating at 95 ◦C for 30 min improved the pepsin–trypsin digestibility of mung bean PI, reduced it for red bean PI, and did not change it for red kidney bean PI [64]. A larger extent of aggregation in heated mung bean PI was reported than in red kidney bean PI. It was suggested that the presence of basic, hydrophobic and uncharged polar AA influences the thermal and structural stability of proteins, and thus the tendency to aggregate when heated. Meanwhile, the in vitro digestibility of lupin and winged bean PC was improved by heating in a boiling water bath for up to 30 min [65,66]. The trypsin and chymotrypsin inhibitory activity of the freeze-dried winged bean PC was inactivated by heat treatment [66].

Likewise, the digestibilities of individual protein fractions from different crops are not modified in a similar manner upon heating. Vicilin-like proteins from chickpea and common bean are both resistant to gastric digestion; however, the digestibility of the former was improved by autoclaving, while for the latter, it was reduced [67,68]. Furthermore, chickpea albumin, 11S and total globulin digestibility increased, as a result of heat treatment [68]. Conversely, native protein fractions from fava bean were better digestible than those that denatured after autoclaving [67].

One might expect that preventing heat-induced aggregation would lead to a positive effect on protein digestibility. This was observed for lentil globulins which were unsusceptible to heat-induced aggregation, given that disulphide interactions were not observed [69]. Nevertheless, the negative charge of a protein fraction from common bean made the protein less prone to aggregation and yet less digestible than its unheated, less negatively charged, counterpart [67]. Based on the effect of heating on the electric charge of proteins and peptides, the latter study suggested that protein electronegativity and hydrophobicity were associated with protein aggregation and digestibility.

Net-zero effects may result from concurring events improving and reducing the DH achieved by digestive enzymes. Commercial soybean and pea PI dispersions heated at 90 and 120 ◦C for 30 min did not show different DH during in vitro gastric digestion compared to their unheated counterparts. Upon close inspection of the soluble and sedimented tailings, we found that heating improved the solubility of the commercial PI, and that the proteins separated into this fraction could be more extensively hydrolysed than those in the sedimented fractions [70].

Meanwhile, for dry-fractionated ingredients, heat treatment has shown to reduce the gastric digestibility of lupin and quinoa proteins. More small peptides (<3 kDa) were released from the unheated and heated at 60 ◦C dispersions of air-classified lupin PC than the dispersion heated at 90 ◦C [71]. A similar trend was observed for dry fractionated quinoa PC, with unheated and heated at 60 ◦C dispersions being more extensively hydrolysed than dispersions heated at 90 and 120 ◦C [45,72]. Similarly, quinoa PI from conventional aqueous fractionation showed lower DH with increasingly higher heating temperatures [30].

As previously discussed, alkaline heat treatment is generally detrimental for protein digestibility. Heating at higher pH reduced the in vitro protein digestibility of globulins from navy bean [73], of soybean PI [74] and rapeseed PC [75]. These results were confirmed for spray-dried soybean PI by an in vivo rat study [76]. For the most part, the limited digestibility can be attributed to the formation of lysinoalanine at high pH [74–76].

Thus far, we discussed studies on so-called moist heating, but the environment during heating does influence the protein digestibility. Sathe, Iyer and Salunkhe [14] compared dry and moist heating of navy bean PC and PI extracted with Na2CO3, as well as water-extracted albumins and NaCl-extracted globulins. The DH achieved with trypsin-αchymotrypsin-peptidase was improved more significantly by moist than by dry heating. Similarly, boiling, microwaving, autoclaving, and dry or oven heating improved the digestibility of sweet potato PC [77]. Autoclaved dispersions presented the highest DH by pepsin–pancreatin digestion, followed by microwave and, lastly, dry heating. The PDCAAS determined in a rat assay was improved for autoclaved PI compared to the unheated ingredient. As previously reported, the TIA was reduced by all types of heat treatments studied.

These observations give a sense of the optimum range of heat treatment to improve the protein digestibility; more heating can negatively impact the digestibility (Table 1). The appropriate heat treatment would then depend on the ingredient source, the type of protein fraction, the type and conditions of heating.

#### *4.3. High Pressure Processing*

Laguna et al. [78] conducted a comprehensive study on the effect of heating and high pressure processing at two different pH (3.6 and 6.2) of commercial pea PI on its in vitro digestibility. For the most part, high pressure processing improved the gastric digestibility of pea protein. Samples prepared at a higher pH were more digestible than those at pH 3.6. Autoclaving did not alter the protein digestibility at either pH, which shows that the effect of pressure cannot be explained by denaturation, similar to that during heating. High pressure processing followed by a 30 min, 80 ◦C heat treatment at pH 3.6 reduced the protein digestibility. In contrast, high pressure processed red kidney bean PI presented a significantly lower in vitro digestibility by trypsin [79]. This was attributed to the generally low digestibility of phaseolin, particularly when aggregated. In this case, we may conclude that the protein source, as much as the processing steps, influences the digestibility of proteins.

#### **5. Crosslinking, Complexation and Other Modifications**

Forming protein complexes with other proteins or other compounds can be an unintended consequence of combining materials in one matrix or can be intentionally induced to achieve certain functions, such as colon-targeted drug delivery [80] or to confer an added nutritional benefit [81].

#### *5.1. Transglutaminase-Catalysed Polymerization*

Phaseolin from *Phaseolus vulgaris* L. was cross-linked by microbial transglutaminase [80]. Its isopeptide bonds made phaseolin more resistant to pepsin and trypsin action, especially for pepsin. Similarly, the pepsin–trypsin digestibility was reduced for native and heated crosslinked proteins from soybean PI, while it was improved by heat treatment alone [82]. While a single protein source was used in this study, covalent crosslinks were identified between β-conglycinin and acidic subunits of glycinin. In contrast, positive effects on the trypsin digestibility as a consequence of crosslinking by transglutaminase have been reported in red kidney bean PI [83]. The digestibility increased in crosslinked protein with longer crosslink reaction times, which was attributed to protein unfolding and denaturation of the vicilin unit.

Limited protein degradation by pepsin–pancreatin was observed for soybean PI polymers and heteropolymers with whey PI or casein, compared to the untreated PI [84]. Furthermore, soybean PI heteropolymers were more resistant to in vitro digestion compared to the whey PI-casein heteropolymer. This was attributed to reduced accessibility for enzymes to the peptide bonds, due to blockage of lysine residues and steric hinderance. Likewise, soybean PI–bovine gelatine composites showed lower pepsin–trypsin digestibility than the PI [85]. Trypsin-catalysed hydrolysis, prior to in vitro digestion, increased the digestibility slightly but it remained significantly lower for the untreated PI.

Glycation and crosslinking soybean PI with chitosan, or oligo-chitosan with transglutaminase improved the pepsin–trypsin digestibility [86,87]. The crosslinked soybean PI was more digestible than the untreated PI in both pepsin and pepsin–trypsin digestion assays.

To assess the effect of Maillard reaction products, crosslinked commercial soybean PI was heated with D-ribose or sucrose [88]. Crosslinking had a negative effect on in vitro protein digestibility, particularly at longer transglutaminase incubation times. Overall, sucrose-containing samples were more digestible than ribose-containing samples. AA loss was reported as a consequence of crosslinking, most significantly of lysine.

Therefore, the effect of transglutaminase-catalysed crosslinking on protein digestibility depends on the extent to which cleavage sites become exposed or buried within the structure of the crosslinked protein. Furthermore, AA bioavailability could also decrease as a result of this processing step.

#### *5.2. Acylation*

Acylation of proteins can result in techno-functionality, such as solubility and emulsifying activity [89]. Mung bean PI was acylated with succinic and acetic anhydrides [90]. The trypsin–pancreatin digestibility was improved by acylation, probably due to protein unfolding. Acetylation was reported to reduce antinutritional factors (phytic acid, tannins and trypsin inhibitors) to a greater extent than succinylation. Similarly, acetylated and succinylated red kidney bean PI were more digestible by trypsin than their untreated counterpart [89]. This was attributed to increased protein solubility and protein unfolding.

The improved digestibility due to acylation observed with these ingredients was also reported for a soybean PI hydrolysate [91]. The in vitro digestibility was significantly higher for succinylated soybean PI hydrolysates compared to the non-succinylated control. The authors also attributed this effect to protein dissociation or unfolding, and an increase in solubility. de Regil and Calderón de la Barca [92] assessed the in vivo digestibility of a soybean protein hydrolysate enzymatically bound by chymotrypsin to methionine methyl-ester using a rat study. There was no significant difference between the apparent digestibility of modified soybean PI hydrolysate and the control with free methionine. Nevertheless, the protein efficiency ratio was significantly higher for the modified ingredient.

Again, protein unfolding is related to an improvement of its digestibility, as was also observed with thermal denaturation. Moreover, peptides of lower molecular weight and, perhaps as a result, increased solubility would generally result in better digestibility, unless opposed by other cross-effects.

#### *5.3. Complexation with Phenolic Compounds*

The digestibility of thermally denatured soybean PI was significantly improved, mostly by pepsin, when complexed with anthocyanins from black rice extract [62]. It was suggested that the network formed by the complex promotes enzymatic action is made possible by changes in the secondary structure; again, (partial) unfolding then facilitates the digestion. In a similar manner, soybean PI–curcumin complexes were more extensively hydrolysed than the non-complexed PI, particularly by pepsin, in a sequential pepsin– pancreatin in vitro digestion assay [60]. Heating before complexation did not influence the extent of digestion of the proteins. Furthermore, the typically pepsin-resistant βconglycinin unit was completely degraded when it was part of the curcumin nanocomplex. Budryn et al. [93] studied soybean PI–hydroxycinnamic acids complexes, either individual 5-caffeoylquinic acid, caffeic acid or ferulic acid, combined in green coffee extract or encapsulated in β-cyclodextrin. The reduction in average molecular weight after pepsin- (trypsin-chymotrypsin) digestion was greater for the complexes than for the untreated PI. It was suggested that interactions and exposure of hydrophobic AA were responsible for the enhanced digestibility, although proteases might also interact directly with hydroxycinnamic acids.

In contrast to the positive effects of anthocyanins and hydroxycinnamic acids, protein– polyphenol complexes reduce the digestibility of pea and soybean PI. Nine commercial pea PI with different physical and chemical characteristics were used to form complexes with polyphenols from cranberry pomace [81]. For some PI, no significant differences were found in the pepsin digestion of non-complexed and complexed proteins; however, all complexed isolates were less extensively hydrolysed by pancreatin digestion. The digestion rate was inversely related to the particle size of the PI. Similarly, soybean PI complexed at 70 or 121 ◦C with polyphenols and flavonoids from black soybean seed coat extract, was less extensively hydrolysed by pepsin–trypsin than the non-complexed ingredient [61]. Moreover, the DH was further reduced by increasing extract concentrations used to produce the complexes. Extract–enzyme or extract–protein interactions were thought to alter the digestive enzymes' conformation, rendering them inactive for protein hydrolysis. In a rat assay, the true nitrogen digestibility was reduced for soybean PI that was complexed with both chlorogenic acid and quercetin [94]. The PDCAAS was significantly reduced for derivatized protein with lysine being the limiting AA.

Yang et al. [95] proposed a multistep process to produce a fermented soybean milk enriched with isoflavone aglycone. More intact proteins remained after pepsin–trypsin hydrolysis of the soybean PI–isoflavone complex, than of the PI. The isoflavone probably inhibited the protease activity. Nevertheless, heated and fermented soybean PI–isoflavone were more extensively hydrolysed than their unheated or non-fermented counterparts.

Phenolic compounds can modify the conformation not only of the proteins but also of the digestive enzymes. Changes in protein conformation can have a positive or negative effect on protein digestion. The former, if unfolding leads to the exposure of cleavage sites, or the latter, if it leads to steric hinderance surrounding the cleavage sites. Furthermore, phenolic compound could also act as inhibitors when bound to the digestive enzymes.

#### *5.4. Protein Oxidation*

Zhao et al. [96] found that a certain extent of protein oxidation had a positive effect on the soybean protein gastric digestibility as a result of protein unfolding, particularly for glycinin. However, severe treatments, i.e., by lipoxygenase-catalysed linoleic acid oxidation [96] or by incubation with 2,20 -azobis (2-amidinopropane) dihydrochloride [97], had a negative impact. In the latter study, the action of the radical-generating compound did not affect the gastric digestion, but it reduced the DH by pancreatin in the intestinal phase. This effect was directly influenced by increasing concentrations of the compound in the system. It was shown that oxidation can degrade several AA and induce protein aggregation. Sánchez-Vioque et al. [98] attributed a reduction in digestibility of chickpea legumin mixed with linolenic acid, to protein oxidation or non-covalent protein–lipid interactions. Meanwhile, no clear relation between carbonyl content, from oxidation products, and extent of hydrolysis in the gastric phase has been observed in thermomechanical processed soybean PC and PI [99].

#### *5.5. Other Modifications*

Soybean PI incubated with malonaldehyde, a lipid peroxidation product, was subjected to in vitro pepsin–pancreatin digestion [100]. β subunits of β-conglycinin were somewhat degraded by pepsin but they became more resistant to pancreatin digestion with increasing malonaldehyde concentration. The availability of indispensable and total free AA after digestion decreased in modified soybean PI.

Soybean PI, cottonseed PC and peanut PC formed complexes with glucose or sucrose [101]. In vitro digestibility was reduced by longer heating times to form the complexes. Protein–glucose complexes were less digestible than the sucrose complexes. Further, available lysine was reduced with heat treatment.

Lastly, soybean PI was incubated with phytase from *Aspergillus niger* to obtain ingredients with different phytate contents [102]. Phytate content, parallel to TIA, was inversely related to pepsin–pancreatin digestibility.

Repeatedly, we find that any process or modification that would induce a certain degree of unfolding will generally facilitate digestion, but extensive unfolding leading to aggregation will result in slower or reduced digestion. Furthermore, modification of AA, particularly of lysine, will often lead to their reduced bioavailability. Finally, processes that reduce or inactivate antinutritional factors, such as phytate or protease inhibitors, will also improve or facilitate the digestion of proteins.

#### **6. Structure Formation**

#### *6.1. Extrusion and Texturization*

The in vitro digestibility of yellow pea and soybean PC can be improved by extrusion. The barrel temperature and screw speed are positively related to the protein digestibility of air-classified pea PC, while the moisture content has a negative influence on its digestibility [103]. Soybean PC, maize meal and cassava root starch were mixed and extruded [104]. The samples extruded at the highest temperature, moisture content and screw speed were the most digestible. The TIA, phytic acid and cyanide contents were reduced by extrusion; however, the tannin content was not reduced. Higher temperatures during extrusion led to more digestible proteins, which opposes the observations from moist heat treatments (Section 4.2), the reason is not fully understood and requires further research.

Duque-Estrada, Berton-Carabin, Nieuwkoop, Dekkers, Janssen and van der Goot [99] explored the effect on in vitro gastric digestibility of high temperature shearing of soybean protein ingredients, as well as the relevance of structure and size reduction in the digestibility. Sheared samples were cut into small pieces or ground into finer particles. Pepsin-catalysed hydrolysis was faster for unheated dispersions, followed by ground matrices. Cut samples were more slowly and less extensively hydrolysed than the other physical states.

The work from Li et al. [105] shows how the formation of rice glutelin fibrils through heat treatment under acidic conditions makes the protein more resistant to pepsin–pancreatin digestion.

#### *6.2. Pre- and Intra-Gastric Gelation*

Opazo-Navarrete et al. [106] related the mechanical strength and porosity of heatinduced gels of soybean PI and pea PC to their gastric digestibility. No significant differences were observed between gels pre-heated at different temperatures. Soybean protein gels were less extensively hydrolysed than the control consisting of a protein dispersion, unlike pea protein gels that were hydrolysed to a similar extent as the dispersion.

Pressure-induced gels from air-classified lentil and fava bean PC were more digestible than heat-induced gels under in vitro gastric conditions [107]. It was suggested that the network of pressure-induced gels allowed for a similar extent of access to pepsin as in concentrated protein dispersions. Meanwhile, both treatments changed the structure of the 55 kDa fractions to be better digestible in the gastric phase. The TIA was more significantly reduced by heating than by pressurization.

Soybean PI coagulates formed with MgCl<sup>2</sup> or glucono-δ-lactone were more digestible than gels prepared with transglutaminase [108]. This was attributed to the covalent isopeptide bonds formed by transglutaminase that cannot be degraded during in vitro gastrointestinal digestion. In contrast, the non-covalent bonds formed during coagulation by MgCl<sup>2</sup> or glucono-δ-lactone could be broken during digestion. Soybean PI and glycerol films were prepared with ferulic acid, tannin, corn starch or H2O<sup>2</sup> at pH 7 to 10 [109]. The gastric digestibility of the films was significantly lower than that of the PI in a dispersion, except for the films prepared with corn starch, which were digested to a similar extent as the control. Lysine availability was also lower in the films. Ferulic acid and tannins were thought to form crosslinks with AA, while H2O<sup>2</sup> could have oxidized certain AA. Lastly, films formed at pH 9 and 10 were less digestible than at pH < 8.5. This was attributed to AA isomerisation and crosslinking at high pH.

In a simulated gastric environment, dispersions of soybean PI and negatively charged polysaccharides (xanthan gum, carrageenan [110] or alginate [111]) self-assemble into a hydrogel. The pepsin-catalysed hydrolysis of the gels was slower even at low polysaccharide contents, compared to the single PI. Hu, Chen, Cai, Fan, Wilde, Rong and Zeng [110] similarly found that soybean PI–carrageenan gels were digested more slowly than those with xanthan gum, due to the more compact and dense gel network in the former.

Generally, structure formation led to a slower and sometimes lower extent of hydrolysis by digestive enzymes compared to liquid dispersions. This is explained by physical hinderance surrounding the cleavage sites. Therefore, looser structures as weaker gels allow for a better digestibility than tighter structures. Further, covalent crosslinking inhibits protein unfolding, while non-covalent bonds can dissociate, especially at lower pH in the stomach, and thus allow for faster digestion. As heat treatment is often required before gelation, antinutritional factors, such as trypsin inhibitors, can also be inactivated.

#### **7. Macronutrient Interactions**

Proteins are almost never processed or consumed on their own. The effect of the interaction of proteins with other macronutrients on protein digestion is not fully understood, but there are some general directions suggested.

#### *7.1. Animal- and Plant-Based Protein Hybrid Foods*

Reconstituted beverages containing the combination of bovine milk PC and soybean, pea or rice PI showed an improved in vitro DH and PDCAAS of blends compared to individual plant proteins [112]. However, this was not observed in solid matrices. Proteins from pea PI, rice protein or lentil flour were enzymatically bound to beef chuck ground meat using transglutaminase [113]. The cooked restructured beef steaks were digested using the INFOGEST 2.0 model with expectorated boluses. No outstanding differences were observed in the peptide size distribution in the digestates of the samples with different treatments. Lentil-enriched steaks released the highest amounts of free isoleucine, lysine, phenylalanine and valine. Protein (re-)aggregation was observed after in vitro gastrointestinal digestion.

#### *7.2. Starch*

Oñate Narciso and Brennan [114] found a relationship between the amylose content of starch with protein digestion. Pea PI was combined with starch from basmati and glutinous rice, with high and low amylose to amylopectin ratios, respectively. All proteins from the samples prepared with glutinous rice starch were degraded after pepsin–pancreatin digestion, but the vicilin and legumin acidic subunit from basmati rice starch samples remained after digestion. The authors proposed that the proteins were embedded into the amylose network. Similarly, quinoa protein from aqueous or dry fractionation was combined with starch-rich fractions from dry or mild aqueous fractionation, which after heating showed lower DH from in vitro gastric digestion than starch-free, unheated protein dispersions [45,72]. This reduction directly related to the heating temperature and was thus probably associated to starch gelatinization. Therefore, embedding the protein in a gelatinized starch gel does reduce the digestibility, probably due to the inaccessibility of the gel for the enzymes.

#### *7.3. Fibre*

The DH obtained by pepsin digestion of dry-fractionated quinoa PC was slightly reduced in quinoa fibre-containing unheated and heated dispersions [72]. The effect of fibre on quinoa protein gastric digestion was not as significant as for starch. Fibre seemed to counter the low hydrolysis induced by starch gelatinization. The fibre does not form a gel that is difficult to penetrate for enzymes but may induce somewhat better mixing due to the higher viscosity.

#### **8. Conclusions**

Plant proteins have the potential to provide all indispensable amino acids. However, as described at length, processing and plant protein digestibility are strongly related.

Heating and soybean are the process step and crop most researched, respectively, reflective of their ubiquity in the production of plant-based food products. Moderate heating may enhance the digestibility by inducing partial unfolding of the proteins, thereby rendering them better accessible for the proteases. However, extensive heating induces aggregation, which makes the cleavage sites less accessible. Similar effects are seen with other types of treatments. Acylation of protein-rich ingredients improved their digestibility, probably also due to partial unfolding. Meanwhile, alkaline treatment, during or after fractionation, consistently reduces the digestibility of different crops, since it strongly changes the structure of the protein and induces AA isomerisation. Again, we see an optimum in the severity of the treatments for digestibility. It is however clear that the exact impact depends on the origins of the proteins.

Ultimately, it is desirable to attain an overarching relationship between the digestibility and the modifications resulting from processing. This review can serve as a guide when considering a certain processing step in the production of plant-based alternatives to animal-sourced products. There are ample opportunities for further research of unexplored processes for promising crops and vice versa, to truly consider the use of plant protein-rich ingredients in food products as a transition pathway to a more sustainable food system.

**Supplementary Materials:** The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/foods11060870/s1, Table S1: Summary of digestion assays, sample treatment and measurement from studies investigating the effect of processing on protein digestion. References [21–114] are cited in the Supplementary Material.

**Author Contributions:** Conceptualization, A.R.d.R. and R.M.B.; methodology, A.R.d.R.; investigation, A.R.d.R.; writing—original draft preparation, A.R.d.R.; writing—review and editing, R.M.B. and A.E.M.J.; visualization, A.R.d.R. and A.E.M.J. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Consejo Nacional de Ciencia y Tecnología, Mexico, grant number 480085.

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

**Informed Consent Statement:** Not applicable.

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

#### **References**


**Mirela Kopjar 1,\* , Ivana Buljeta <sup>1</sup> , Ina Corkovi´c ´ <sup>1</sup> , Anita Pichler <sup>1</sup> and Josip Šimunovi´c <sup>2</sup>**


**Abstract:** Plant-based proteins are very often used as carriers of different phenolic compounds. For that purpose, complexation of quercetin with almond and brown rice protein matrices was investigated. The amount of protein matrices was constant, while the concentration of quercetin varied (1 mM, 2 mM or 5 mM) during complexation. Dried complexes were investigated for quercetin amount (HPLC analysis) and antioxidant activity (DPPH, FRAP and CUPRAC methods). Additionally, complexation was proven by DSC and FTIR-ATR screening. An increase in the concentration of quercetin in the initial complexation mixture resulted in the increase in the adsorption of quercetin onto protein matrices. For the brown rice protein matrices, this increase was proportional to the initial quercetin concentration. Adsorption of quercetin caused the change in thermal stability of microparticles in comparison to corresponding protein matrices that have been proven by DSC. FTIR-ATR analysis revealed structural changes on microparticles upon adsorption of quercetin.

**Keywords:** quercetin; almond protein matrix; brown rice protein matrix; HPLC; antioxidant activity; DSC; FTIR-ATR

### **1. Introduction**

A growing area of interest in the field of polyphenols is their interactions with other components present in the food matrix, including proteins, carbohydrates and lipids [1]. Proteins as building blocks have both nutritional and functional properties and are an important dietary source of amino acids [2]. As an outcome of interactions between polyphenols and proteins, various complexes can be formed, which consequently cause changes in antioxidant properties of polyphenols and affect the functional, structural and nutritional properties of proteins [3]. Flavonoids are an important subgroup of polyphenols and quercetin, as a member of the flavonol subclass of flavonoids is found in various fruits, vegetables and tea [4]. The intake of quercetin is related to the assembly of beneficial health properties such as antioxidant, anti-inflammatory and anti-viral properties, improvement in cardiovascular health [5] and reducing the intensity of the symptoms and negative predictors of COVID-19 [6], which is a hot topic nowadays. Its anticancer effect was also established in numerous studies [7,8]. For the aforementioned reasons, the involvement of quercetin in the human diet is strongly recommended and it is proposed as an excellent ingredient for functional foods [9]. On the other hand, its broader application is limited because of insufficient water solubility and chemical instability. The application of an adequate delivery system is one way to achieve stabilization and improvement of the health benefits of quercetin. The "Delivery by design" approach leads to the expansion of the search for effective delivery systems of bioactives in order to encapsulate them; on the one hand for their protection from environmental factors and on the other hand to control their release under defined conditions. Possibilities for the application of delivery systems can be various from food, agrochemical, pharmaceutical, cosmetic to personal care industries [10].

**Citation:** Kopjar, M.; Buljeta, I.; Corkovi´c, I.; Pichler, A.; Šimunovi´c, J. ´ Adsorption of Quercetin on Brown Rice and Almond Protein Matrices: Effect of Quercetin Concentration. *Foods* **2022**, *11*, 793. https://doi.org/ 10.3390/foods11060793

Academic Editor: Yonghui Li

Received: 8 February 2022 Accepted: 8 March 2022 Published: 9 March 2022

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

**Copyright:** © 2022 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/).

There are various types of delivery systems, and each one has specific advantages and disadvantages; thus, the selection should be based on the field of utilization [11]. Both animal-based and plant-based protein matrices were used for complexation with polyphenols. The most used animal-based protein matrices were whey, gelatin, milk proteins, bovine serum albumin [12–16]. Applications of plant-based protein matrices are gaining higher attention and proteins from different sources such as sunflower, legume seed, corn kernels, quinoa, wheat, rice, peas, hemp, almond and pumpkin [17–28] have been used for the preparation of complexes which can be further used as functional food additives. As a result of consumers' awareness of the importance of a healthy and balanced diet, functional foods are becoming more popular. In addition, convenience is becoming a very important element in the selection of foods. A very busy and mobile lifestyle, the search for simple meal preparation and easier consumption, healthy snacking options in as well as outside of homes are also some of the emerging trends among consumers [3,17,19,29] that need to be addressed by the food industry.

The aim of this research was to prepare microparticles from brown rice or almond protein matrices and quercetin in order to investigate whether different concentrations (1 mM, 2 mM or 5 mM) of quercetin in the initial mixture had an impact on its adsorption on proteins. Determination of quercetin concentration using high-performance liquid chromatography (HPLC) and antioxidant activities of prepared microparticles were performed. Additionally, DSC and FTIR-ATR screening of microparticles were performed. DSC analysis was conducted to evaluate the effect of quercetin interactions with proteins on the thermal stability of protein matrices, while screening of IR spectra was performed as proof that interactions between quercetin and protein matrices occurred.

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

#### *2.1. Materials*

Almond protein matrix was purchased from Raab Vitalfood GmbH (Rohrbach, Germany) and brown rice protein matrix was from Kernnel premium (Zagreb, Croatia). Quercetin, trolox and 2,2-diphenyl-1-picrylhydrazyl (DPPH) were obtained from Sigma-Aldrich (St. Louis, MO, USA). Sodium carbonate was obtained from Kemika (Zagreb, Croatia). Neocuproine, copper (II) chloride and 2,4,6-tri(2-pyridyl)-s-triazine (TPTZ) were obtained from Acros Organics (Geel, Belgium). Orthophosphoric acid (HPLC grade > 85%) was obtained from Fisher Scientific (Loughborough, UK), while methanol (HPLC grade) was purchased from J.T. Baker (Deventer, The Netherlands). Iron (III) chloride hexahydrate, ethanol, sodium acetate and ammonium acetate were purchased from Gram-mol (Zagreb, Croatia).

#### *2.2. Preparation of Protein/Quercetin Microparticles*

The microparticles were formulated by the complexation of protein matrices (constant amounts; 5%) with 20 mL of quercetin ethanol solution (1 mM, 2 mM or 5 mM). Two protein matrices were used (both are usually used as dietary supplements), brown rice with approximately 85% of proteins (7.7% of carbohydrates and 5.1% of lipids), and almond with approximately 50% of proteins (fibers 17%, 9% of carbohydrates and 11% of lipids). The preparation method was adapted from other studies [21,27,28]. The protein matrix was weighed and added to the quercetin solution. In order to prepare protein/quercetin microparticles, obtained mixtures were mixed on a magnetic stirrer for 15 min at room temperature. During that time, part of quercetin was adsorbed onto the protein matrix and the other part remained in the solution. Afterward, well-homogenized mixtures were centrifuged for 15 min at 4000 rpm in order to remove the quercetin that did not adsorb onto the protein matrix and remained in the supernatant. The supernatant was discarded and the wet–solid phase that represented adsorbed quercetin onto protein matrix was collected. After air-drying, protein/quercetin microparticles were obtained in the form of dried powder.

#### *2.3. Extraction of Quercetin from Protein/Quercetin Microparticles*

Quercetin was extracted from obtained protein/quercetin microparticles. Microparticles were weighted (0.15 g), 10 mL of acidified methanol (methanol:HCl ratio was 99:1) was added and the obtained mixture was well homogenized. Extraction was conducted at room temperature for 24 h. After that time, mixtures were filtered to obtain clear extracts which were immediately utilized for the determination of the amount of quercetin and antioxidant activities.

#### *2.4. Reverse-Phase High Performance Liquid Chromatography (RP-HPLC)*

A RP-HPLC system (1260 Infinity II; Agilent technology, Santa Clara, CA, USA) was used for the evaluation of the amount of quercetin. The system was equipped with a DAD (diode array) detector, a quaternary pump and a column (poroshell 120 EC-C 18; 4.6 × 100 mm, 2.7 µm). Two mobile phases were used; mobile phase A was orthophosphoric acid (0.1%) and mobile phase B was methanol (100%). The gradient that was applied for separation was described in previous studies [27,28]. The injection volume of the extract was 10 µL, under the flow rate of 1 mL/min at room temperature. A calibration curve for quercetin was constructed in the range from 5 to 150 mg/L. UV/Vis spectra was screened in the range from 190 to 600 nm, and quercetin was determined at 360 nm. Duplicate evaluations were conducted.

#### *2.5. Antioxidant Activity*

DPPH, CUPRAC and FRAP methods were utilized for the evaluation of antioxidant activities of microparticles extracts. Details for these methods were previously given by Buljeta et al. [30]. Assays were performed in triplicate and results were presented as micromoles of Trolox equivalent per 100 g of sample (µmol TE/100 g).

#### *2.6. Differential Scanning Calorimetry (DSC)*

For the DSC scanning of microparticles, a differential scanning calorimeter (Mettler Toledo 822, Mettler Toledo, Greifensee, Switzerland) was applied. In a 40 µL aluminum pan, 7 ± 0.2 mg of microparticles was weighed. The aluminum pan was covered and then inserted into the oven of the DSC. Screening of microparticles was performed from 25 ◦C to 140 ◦C. Firstly, samples were left for 4 min at 25 ◦C. Afterwards, the temperature was increased at a rate of 5 ◦C/min up to 140 ◦C, where samples were also left for 4 min. Duplicate screenings were conducted.

#### *2.7. Fourier-Transform Infrared Spectroscopy-Attenuated Total Reflectance (FTIR-ATR)*

The IR spectra of protein matrices and protein matrices loaded with quercetin were recorded using FTIR-ATR (Cary 630 FTIR spectrometer, Agilent Technology, Santa Clara, CA, USA), equipped with software MicroLab Expert. Samples were screened in the interval from 4000 cm−<sup>1</sup> to 600 cm−<sup>1</sup> .

#### *2.8. Statistical Analysis*

STATISTICA 13.1 (StatSoft Inc., Tulsa, OK, USA), the software program, was utilized for analyzing the obtained results. Variance analysis (ANOVA) and Fisher's least significant difference (LSD) with significance defined at *p* < 0.05 were selected for statistical evaluation of the results, which were presented as mean value ± standard deviation.

#### **3. Results**

#### *3.1. Quercetin Amount and Antioxidant Activity of Protein/Quercetin Microparticles*

The amounts of quercetin and antioxidant activities of protein/quercetin microparticles are given in Table 1. Comparison of almond protein/quercetin (AP/Q) microparticles and brown rice/quercetin (RP/Q) microparticles showed that RP/Q microparticles had a higher amount of quercetin than AP/Q microparticles prepared with the same initial concentration of quercetin, indicating that the brown rice protein matrix had a higher

affinity towards quercetin. Additionally, it can be observed that an increase in concentration of quercetin in the initial solution for complexation, resulted in an increase in the amount of quercetin. Results were compared in order to investigate whether this increase in quercetin amount was proportional to the initial concentration of quercetin. Amounts of quercetin on AP/Q microparticles were 60.18, 98.38 and 196.34 mg/100 g for AP/Q\_1, AP/Q\_2 and AP/Q\_5 (i.e., 1 mM, 2 mM and 5 mM of quercetin in initial solution), respectively. Results indicated that with the double increase in the initial quercetin, the resulting concentration increase was 1.6 times, and for the five times increase in the initial quercetin concentration, the resulting increase was 3.7 times higher. For RP/Q microparticles, a different trend was observed. Amounts of quercetin on microparticles were 108.24, 226.50 and 506.98 mg/100 g for RP/Q\_1, RP/Q\_2 and RP/Q\_5 (i.e., 1 mM, 2 mM and 5 mM of quercetin in initial solution), respectively. Results indicated that with a double increase in initial quercetin concentration, the increase was 2.1 times and with a five times increase in initial quercetin concentration, the resulting increase was 4.7 times higher. Interesting data were obtained by comparing the amount of the quercetin on microparticles to the initial amount of quercetin, i.e., calculating the adsorption efficiency of protein matrices towards the quercetin. For the almond protein matrix, as was expected, a lower efficiency was determined than for the brown rice protein matrix. Even though the quercetin amount increased on A/Q microparticles with the initial quercetin amount, a decrease in adsorption efficiencies were observed; 20%, 16.3% and 13% for AP/Q\_1, AP/Q\_2 and AP/Q\_5, respectively. For RP/Q microparticles, a slightly different trend was obtained, i.e., for RP/Q\_1 and RP/Q\_2, adsorption efficiencies were 36.8% and 37.5%, and for RP/Q\_5, 33.6%.


**Table 1.** Amount of quercetin (mg/100 g) and antioxidant activity (µmol TE/100 g) of protein/quercetin microparticles.

Q—quercetin; AP—almond protein matrix; RP—brown rice protein matrix; 1, 2 and 5—represent concentration (mM) of initial quercetin solution; data in one column labeled with different letters statistically differ.

For the evaluation of antioxidant activities of microparticles, three methods were selected: DPPH, FRAP and CUPRAC methods. Values for antioxidant activities obtained with all methods followed the amount of quercetin, i.e., an increase in quercetin amount caused an increase in antioxidant activity. However, a trend of the proportional increase in antioxidant activity with the quercetin amount was not observed for all methods. DPPH antioxidant activities for AP/Q microparticles were from 29.38 to 32.18 µmol TE/100 g, while for RP/Q, microparticles values were slightly higher, from 31.12 to 40.40 µmol TE/100 g. Values of antioxidant activities were much higher with the CUPRAC method and they ranged from 106.39 to 192.75 µmol TE/100 g for AP/Q microparticles and from 84.30 to 414.52 µmol TE/100 g for RP/Q microparticles. With the FRAP method, the lowest values of antioxidant activities were obtained. For AP/Q microparticles, they ranged from 0.66 to 2.34 µmol TE/100 g, and for RP/Q microparticles, from 1.20 to 8.35 µmol TE/100 g.

#### *3.2. Temperature of Denaturation of Protein Matrices and Protein/Quercetin Microparticles*

The results of the DSC determination of denaturation temperatures of protein matrices and protein/quercetin microparticles are presented in Table 2. Both protein matrices had similar values of denaturation temperature at 85.25 ◦C. When comparing the obtained microparticles, AP/Q microparticles had lower denaturation temperatures (around 83.5 ◦C) than the corresponding protein matrix, while RP/Q microparticles had it higher (from 85.78 to 86.72 ◦C). Additionally, a difference in the enthalpy of denaturation was observed. For RP/Q complexes the enthalpy of denaturation increased for 3 J/g, while for AP/Q complexes, the increase was for 1.5 J/g.

**Table 2.** Temperatures of denaturation (Td) of protein matrices and protein/quercetin microparticles.


Q—quercetin; AP—almond protein matrix; RP—brown rice protein matrix; 1, 2 and 5 represent concentration (mM) of initial quercetin solution; data in column labeled with different letters statistically differ.

#### *3.3. FTIR-ATR Spectra of Protein Matrices and Protein/Quercetin Microparticles*

Changes in IR spectra that were obtained by FTIR-ATR screening of protein/quercetin microparticles prepared with different initial concentrations of quercetin were the same, so only one IR spectra of microparticles was presented (the one obtained with 5 mM of quercetin). Figure 1 represents the IR spectra of the almond protein matrix and almond protein/quercetin microparticle. Comparing those two spectra, changes in protein structure after the adsorption of quercetin were established. In two regions, one from 3500 cm−<sup>1</sup> to 3000 cm−<sup>1</sup> and another one from 1650 cm−<sup>1</sup> to 600 cm−<sup>1</sup> , the intensity of the protein spectra was lower than for the microparticle. However, in the region from 3000 cm−<sup>1</sup> to 2800 cm−<sup>1</sup> and for the band at 1745 cm−<sup>1</sup> , a reverse tendency was observed. The region from 3500 cm−<sup>1</sup> to 3000 cm−<sup>1</sup> can be assigned to amide A, N-H stretching coupled with hydrogen bonding. Additionally, in this region, the band at 3004 cm−<sup>1</sup> assigned to the C-H bond disappeared after quercetin adsorption. In the region from 3000 cm−<sup>1</sup> to 2800 cm−<sup>1</sup> , two bands were detected, one at 2922 cm−<sup>1</sup> and another one at 2855 cm−<sup>1</sup> both assigned to CH<sup>2</sup> stretching. A band at 1745 cm−<sup>1</sup> can be assigned to the C=O band of polysaccharides. Additionally, a change in the Amid I structure of protein occurred. A band at 1632 cm−<sup>1</sup> shifted to 1625 cm−<sup>1</sup> after the adsorption of quercetin. One additional change caused by quercetin adsorption was a loss of shoulder at 1141 cm−<sup>1</sup> (assigned to C-O band of oligosaccharides) on the protein matrix.

Figure 2 represents IR spectra of other set of samples, i.e., the brown rice protein matrix and brown rice protein/quercetin microparticle. Even though results of the amount of adsorbed quercetin showed that RP/Q microparticles had a higher amount of this phenolic, structural changes were not so pronounced as for AP/Q microparticles. Two bands that were detected at 2922 cm−<sup>1</sup> and 2855 cm−<sup>1</sup> , both assigned to CH<sup>2</sup> stretching, after adsorption of quercetin, shifted to 2929 cm−<sup>1</sup> and 2875 cm−<sup>1</sup> . Another change was

(AP/Q).

of quercetin.

(RP/Q).

(RP/Q).

*Foods* **2022**, *11*, x FOR PEER REVIEW 6 of 11

detected at 1737 cm−<sup>1</sup> (assigned to C=O band of polysaccharides) that disappeared after the adsorption of quercetin. *Foods* **2022**, *11*, x FOR PEER REVIEW 6 of 11

**Figure 1.** IR spectra of almond protein matrix (AP) and almond protein/quercetin microparticles **Figure 1.** IR spectra of almond protein matrix (AP) and almond protein/quercetin microparticles (AP/Q). 1737 cm−1 (assigned to C=O band of polysaccharides) that disappeared after the adsorption

**Figure 2.** IR spectra of brown rice protein matrix (R) and brown rice protein/quercetin microparticles **Figure 2.** IR spectra of brown rice protein matrix (R) and brown rice protein/quercetin microparticles (RP/Q).

#### **4. Discussion**

There have been many different instrumental techniques used for the characterization of protein/phenolics complexes [31]. In this study, we applied HPLC, DSC and FTIR-ATR analyses for evaluation of the adsorption of quercetin on selected protein matrices.

Interactions that are created between phenolics and proteins upon their complexation depend on the structure of both compounds, as well as complexation conditions [32,33]. Similar protein matrices for the adsorption of different phenolics were used in other studies. Adsorption of glucosyl-hesperidin on pea, almond, pumpkin and brown rice protein matrices revealed that glucosyl-hesperidin was determined in the highest amount on pea protein microparticle, followed by almond, brown rice and in the lowest amount on pumpkin protein microparticles [27]. For the adsorption of cinnamic acid on pea, almond and pumpkin protein matrices, a different trend was observed; hence, cinnamic acid had the highest affinity for pumpkin and the lowest for almond protein matrices [28]. The investigated protein matrices differ in the protein content; pea and brown rice protein matrices had 85% of protein content, while the almond and pumpkin had 50%. Other organic molecules such as polysaccharides can be incorporated in different types of protein matrices [3,27,28,31,34], consequently having an effect on the adsorption of phenolics onto them. As a result of the encapsulation of cranberry phenolics onto different protein matrices with 50% of proteins (medium roast peanut or defatted soy flours) to protein matrices with over 70% of proteins (pea, soy or hemp protein isolates), a non-linear trend was achieved when protein amount and the adsorption capacity for phenolics were put in correlation [17]. In that study, cranberry phenolics had the highest affinity towards defatted soy and medium roasted peanut flours as well as towards hemp protein isolate. Additionally, a mentioned non-linear trend was obtained in another study, which deals with the adsorption of blueberry anthocyanins on matrices with lower amounts of proteins such as corn flour (5.3%), brown rice flour (8.6%), white whole-wheat flour (13%) and defatted soy flour (47%) [21]. Generally looking, the reactivity of phenolics towards proteins is correlated with two main factors, one is the number of hydroxyl groups and the other is hydroxyl groups position in phenolics structure [32]. The binding capacity of some phenolics (quercetin, kaempferol, myricetin, flavone apigenin, chlorogenic acid, caffeic acid, gallic acid) towards soy protein revealed that among those phenolics, the highest affinity had gallic acid, followed by chlorogenic acid and quercetin [32]. Comparison of the binding capacity of chlorogenic acid, ferulic acid, gallic acid, catechin, quercetin and apigenin towards albumin and globulin was conducted. Authors have determined that quercetin and catechin had equal binding capacity towards albumin; it was lower than for chlorogenic and gallic acid but higher than for apigenin and ferulic acid. The binding capacity of quercetin towards globulin was lower than for chlorogenic acid, catechin and gallic acid, but higher than for apigenin and ferulic acid [33]. Different studies emphasized that covalent and/or non-covalent interactions can occur between phenolics and proteins [34–36]. Non-covalent ones include interactions through hydrogen bonds, hydrophobic association, van der Waals forces and electrostatic attraction. However, as the most important non-covalent interactions for the complexation of proteins with phenolics, hydrophobic interaction and hydrogen bonds were pointed out [37]. Sui et al. [38] studied the binding of anthocyanins with soy protein isolate while increasing the amount of anthocyanins in the initial mixture and determined that the increase in binding of anthocyanins was proportional to their increase in the initial mixture. We obtained similar results in our study for the brown rice protein matrix that contained a higher protein content, which could be the reason for more pronounced hydrophobic interactions.

Numerous methods for evaluation of the antioxidant activity of foods, dietary supplements and nutraceuticals are available in the literature and can be applied. We selected DPPH, FRAP and CUPRAC methods, which are based on different mechanisms of action. The DPPH method is based on the reaction of radicals with hydrogen-donating antioxidants, which leads to the formation of the non-radical form. One of the characteristics of the DPPH radical is its selectivity in the reaction with hydrogen donors [39]. From our results, it can be observed that with the increase in quercetin amount, DPPH antioxidant activity increased; however, this increase was not proportional to the increase in quercetin amount on microparticles. Results of the other two methods that were used better followed the trend of quercetin amount on microparticles. In CUPRAC assay, the reduction of

Cu(II) to Cu(I) by antioxidants is measured spectrophotometrically, while the reduction of Fe(III) complex to Fe(II) caused by the presence of antioxidants is assessed by FRAP [40,41]. According to our results, microparticles had a significantly higher capability of reduction of Cu(II) to Cu(I) than of Fe(III) to Fe(II).

Interactions between phenolics and proteins can cause a change in the denaturation temperature of the corresponding protein matrix. Usually, this parameter is used for the prediction of thermal stability of formulated protein/phenolic complexes [32]. When the denaturation temperature of formulated protein/phenolic microparticles is higher than the denaturation temperature of the corresponding protein matrix, the formulated microparticle is more stable than the protein matrix and vice versa [42]. From our results, it can be concluded that the adsorption of quercetin on the brown rice protein matrix resulted in its thermal stabilization. However, quercetin affected the almond protein matrix differently, i.e., it caused the decrease of thermal stability. Results of the other studies showed positive, negative or no effect on proteins stability depending on types of proteins and phenolics. The increase in thermal stability was achived when cinnamic acid was adsorbed on the almond protein matrix, while adsorption of the same phenolic acid onto the pea and pumpkin protein matrix had the opposite effect [27]. Green tea polyphenols caused a decrease of the thermal stability of β-lactoglobulin and egg albumen [42,43]. Complexes of soy protein with quercetin, myricetin or phenolic acids had higher stability than protein alone, while flavone, apigenin or kaempferol did not affect the stability of the mentioned protein [32]. The adsorption of chlorogenic acid on lysozyme, bovine serum albumin and α-lactalbumin also caused the increase of thermal stability of those proteins [44,45]. A conjugate of (−)-epigallocatechin gallate and zein had higher thermal stability than pure zein, while conjugates of zein and quercetagetin or chlorogenic acid had similar denaturation temperatures as protein alone [46]. A decrease of thermal stability was also observed upon the adsorption of raspberry juice phenolics onto brown rice proteins [26].

Structural changes of proteins upon adsorption of phenolics depend on protein structure that can be proved by a recoding of the IR spectra of microparticles and their comparison with the protein matrix. On both types of microparticles (AP/Q and RP/Q), a difference at the amide A region (3500 cm−<sup>1</sup> to 3000 cm−<sup>1</sup> assigned to N-H stretching coupled with hydrogen bonding) was observed, which could be an indication of non-covalent interactions between proteins and phenolics, i.e., indication of hydrogen bonding or hydrophobic association [46]. These interactions were probably involved in the adsorption of quercetin onto two selected matrices in our study. Alternation in IR spectra of proteins in the region of 3000 to 2800 cm−<sup>1</sup> (assigned to CH<sup>2</sup> antisymmetric and symmetric stretching vibrations) are an indicator of the existence of hydrophobic contact in the protein/phenolic complexes [47]. Hasni et al. [47] determined these changes in the complexes between αcaseins and β-caseins with tea phenolics. Based on the shifting of the protein antisymmetric and symmetric CH<sup>2</sup> stretching vibrations, they proposed the existence of hydrophobic interactions throughout phenolics rings and hydrophobic pockets in caseins. Considering these results, we can also assume that those hydrophobic interactions occurred between the quercetin and protein matrices used in this study. In RP/Q microparticles, we also observed a shift of bands in this region that can be connected to hydrophobic interactions, as proposed before. On the other hand, on IR spectra of AP/Q microparticle, a change in intensity of bands in this region occurred, i.e., upon adsorption of quercetin, a decrease of the band intensity of protein occurred, suggesting a similar mechanism of interactions. Additionally, these different changes in protein matrix structure could be explained by the fact that the brown rice protein matrix had higher (85%) while almond had lower (50%) protein content; therefore, shifting was more pronounced on RP/Q microparticles. In addition, according to the quercetin amount adsorbed on protein matrices, these interactions were more pronounced for RP/Q microparticles since they contained a higher amount of this phenolic compound and its adsorption was proportional to initial concentration. Probably, these hydrophobic interactions as well as the higher amount of adsorbed quercetin were the reason for the improvement of thermal stability of RP/Q microparticles. Additionally, there was no change in IR spectra in Amid I structure, which could also lead to higher thermal stability of RP/Q complexes.

#### **5. Conclusions**

Functional ingredients with a broad range of applications are becoming more and more popular. Two plant-based protein matrices, brown rice and almond, were chosen for complexation with quercetin. The brown rice protein matrix had a higher affinity for quercetin than almond; thus, it would be a more efficient carrier of this phenolic compound. Additionally, it was observed for the brown rice protein matrices that with the increase in quercetin concentration in the initial mixture, the amount of quercetin on microparticles proportionally increased. This trend was not observed for the almond protein matrix. Additionally, brown rice protein microparticles were thermally more stable in comparison to the corresponding protein matrix, while the reverse trend was determined for almond protein microparticles.

**Author Contributions:** Conceptualization, M.K. and J.Š.; methodology, A.P. and M.K.; formal analysis, I.B. and I.C.; investigation, I.B., A.P. and I. ´ C.; data curation, I.B., A.P. and I. ´ C.; writing—original ´ draft preparation, I.C. and M.K.; writing—review and editing, A.P. and J.Š.; supervision, A.P.; project ´ administration, M.K.; funding acquisition, M.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was part of the project PZS-2019-02-1595 which has been fully supported by the "Research Cooperability" Program of the Croatian Science Foundation, funded by the European Union from the European Social Fund under the Operational Program for Efficient Human Resources 2014–2020. I. C. acknowledges support from the Croatian Science Foundation program for Training ´ New Doctoral Students (DOK-2020-01-4205).

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

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data are contained within the article.

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

#### **References**

