WO2016177775A1 - Gelatin hydrolysates in protein systems - Google Patents

Abstract

The invention provides a method for the production of a gel-based product, the method comprising (i) providing a starting composition comprising a globular protein and water, and (ii) converting the globular protein and water into a gel by heating the starting composition, to provide said gel-based product, wherein the starting composition comprises the globular protein in an amount of 5-20 wt.% relative to the total weight of the starting composition, wherein the starting composition further comprises a gelatin hydrolysate in an amount of 5-40 wt.% relative to the total weight of the starting composition, and wherein the starting compositions comprises 50-90 wt.% water, wherein at least 90 wt.% of the gelatin hydrolysate has a molar mass of ≤ 50 kDa. Further, the invention provides such gel-based product.

Description

Gelatin hydrolysates in protein systems FIELD OF THE INVENTION

The invention relates to a method for the production of a gel-based product as well as to the gel-based product itself. Further, the invention relates to a use of a gelatin hydrolysate in specific applications. BACKGROUND OF THE INVENTION

There is a desire in the art to provide high-protein foods or other high- protein comprising compositions. US5051270, for instance, describes a high-protein nutritive food prepared by a process comprising: (1) forming and shaping a dough comprising: (a) a protein ingredient comprising: (i) vegetable protein, or (ii) a mixture of vegetable protein and animal protein, (b) a cereal flour, a potato powder, or a mixture thereof, and (c) water, and; (2) heating the dough in a vacuum until the dough is dried to a water content of not higher than about 3 wt. % and a protein content of 40 to 85 wt. % based on solids.

The Japanese patent publication JP2006212006 describes gels obtained from protein hydrolysate, microcrystalline or micro-fibrillated cellulose and a gel- forming agent. In these compositions the microcrystalline or micro-fibrillated cellulose is applied as a gelling agent. These compositions provide soft gels, which do not deliver the firm textural properties and sensory perception, as obtained by the present invention.

The European patent publication EP1543734 describes aqueous gel containing compositions for use as a beverage product. Even though these compositions are described for protein and calcium supplementation, the gels disclosed therein are relatively low in protein content as compared to the present invention.

US2005266137 describes the preparation of food compositions containing a non-dairy protein and a proteinaceous high-tryptophan source. However, this document does not describe the preparation of a globular protein-based gel.

According to a publication by Ersch et al. (Food Hydrocolloids 44, 59- 65 (2015)) gels, which are rich in proteins, can be prepared from defatted soybean flour, with pork skin gelatin type A (nominal bloom strength of 150). However, this process including addition of gelatin type A leads to undesired properties of the gel structure.

Cao et al. (2007) Food Hydrocolloids 21, 1 153-1162 describe the preparation of films from mixed gelatin/soy protein isolate gels. The gelatin was identified as bovine bone gelatin type B. Bovine bone gelatin type B has an average molar mass of about 150 kDa. However, this process including addition of gelatin type B leads to undesired properties of the gel structure.

SUMMARY OF THE INVENTION

A problem with prior art food products with increased protein content is that these often are relatively hard and tough. The increase of the protein content substantially influences the sensory perception, such as mouthfeel. Also the textural properties may change. This may not only be undesirable in food applications, but also in non-food applications, as an increase in protein content should not be at the substantial expense of the textural properties.

Hence, it is an aspect of the invention to provide an alternative product, which preferably further at least partly obviates one or more of above-described drawbacks. Yet, it is also an aspect of the invention to provide a method for the production of such alternative product, which preferably further at least partly obviates one or more of above-described drawbacks.

It surprisingly appeared that the protein content could substantially increase in protein-based (food) products by adding gelatin hydrolysate. When adding gelatin per se, the product properties appeared to change in texture, but with the addition of gelatin hydrolysate, it appeared that (food) products, especially gel-based (food) products could be made without a substantial impact on the texture and/or sensory perception (in the case of a food product). It surprisingly appeared that protein-based gels could be made with much higher protein content by adding the gelatin hydrolysate to the gel (or to the starting products of the gel). In this way, gels were obtained with substantially the same gel properties as without the gelatin hydrolysate, but with much higher protein content. Hence, in this way protein-enriched products can be made having the desired properties, such as e.g. sausages or jellies or coatings, etc., for food or non-food applications.

Hence, in a first aspect the invention provides a method for the production of a gel-based product, the method comprising (i) providing a starting composition comprising a globular protein and water, and (ii) converting the globular protein and water into a gel, especially by heating the starting composition, to provide said gel-based product, wherein the starting composition comprises the globular protein in an amount of 5-20 wt.% relative to the total weight of the starting composition, wherein the starting composition further comprises a gelatin hydrolysate in an amount of 5-40 wt.% relative to the total weight of the starting composition, and wherein the starting compositions comprises 50-90 wt.% water, wherein at least 90 wt.% of the gelatin hydrolysate has a molar mass of < 50 kDa. Another term for "molar mass" is molecular weight.

It surprisingly appears that a gelatin hydrolysate can advantageously be used for generating a globular protein-based gel with increased protein content. This can be used for all kind of applications. For instance, in embodiments the gel-based product may comprise a human food product, a pet food product, a feed product, or a coating or an encapsulate for a medicament. Especially, the gel-based product may comprise a dairy product, a probiotic product, a protein bar, a sausage, a protein shake, or a protein drink. However, the gel-based product may also comprise food coatings, films, etc.. Further, the gel-based product may not only refer to a product that is based on such gel, it may also refer to the gel per se, that may for instance be made in a separate step and subsequently be combined with other components to provide a further intermediate product or final product. Hence, the term "gel-based" product may refer to a product wherein the gel may still be discemable and may even still have gel properties or gel-like properties, but may also refer to a product wherein the gel is homogeneously distributed and wherein the product as a whole may not have typical gel-like properties, such as a protein bar or a sausage. Another embodiment may be a drink.

As indicated above, it advantageously appears that the protein content can be raised without substantially affecting the texture. Therefore, the gel-based product may amongst others also be used to provide a food product with an increased protein content without affecting the sensory impression. Hence, amongst others healthy food products can be made, e.g. for elderly people or other people that may need dietary supplements. Further, the use of high-protein foods containing gelatin hydrolysate may be advantageous for the health of joints, the skin, etc.. Therefore, the gel-based product may amongst others also be used to provide a food product for maintaining or improving joint health and/or skin health. The smaller gelatin hydro lysates molecules are more easily and faster absorbed compared to non- hydro lyzed gelatins.

Gelatin is a product derived from collagen. Collagen is the most abundant protein in the human body. It is the main structural protein in the extracellular space in the various connective tissues in animal bodies. It gives our body structure and strength and it can be found in the bones, muscles, skin and tendons. Collagen naturally plays an important role within the body, and it is being used increasingly within certain areas of neutraceuticals or supplements as well. The breakdown of the body's natural collagen can lead to a number of health problems, but the protein can be used in other contexts to help with the repair of the body.

Gelatins are obtained through partial hydrolysis of the collagen, obtained by aqueous extraction of skin, bones, tendons, ligaments etc. obtained from among others bovine, porcine, poultry, fish etc. origin. Bone raw materials are first pretreated to remove fat and mineral. As an important starting material, pig skin is first washed and cut into pieces. The thus cleaned material can subsequently be subjected to either of two different processes, depending on the raw material and on the final application of the gelatin:

(a) the acid process, using an acid raw material preparation, leading to a so-called gelatin type A and

(b) the alkaline process, using an alkaline raw material preparation, leading to a so- called gelatin type B.

Molar mass distribution and Iso Electric Point (IEP) are the most important differences between type A en B gelatins

Additionally or alternatively, also an enzymatic conversion may be applied using (food grade) protease enzymes. Information may amongst others also be found at: http://www.gelatine.org/en/about-gelatine/manufacturing/production- process.html.

The gelatin hydrolysate used herein may e.g. be from bovine origin, porcine origin, poultry origin, warm fish origin, and cold fish origin, especially from bovine origin, porcine origin or poultry origin. Also combinations of two or more gelatin hydrolysates may be applied.

Hence, gelatin is obtained through partial hydrolysis of collagen contained in amongst others animal skins and bones. It is a natural protein food typically consisting of about 88-90% protein, about 10-12% water and about 1% minerals for a caloric value of 370 kcal. It is classified as a foodstuff by the majority of administrative authorities and has received the GRAS status from the FDA. This protein is fully digestible and contains 18 different amino acids, including all of the 9 essential amino acids but tryptophan. It is particularly rich in glycine, proline and hydroxyproline, which, all together, represent almost 50% of the composition of the molecule. Hydroxyproline is an amino acid specific to collagen/gelatin. Gelatin is water-soluble and compatible with most other hydrocolloids, including vegetal colloids such as agar-agar, alginates, carrageenans or pectins. It is compatible with sugars, corn syrups, edible acids and flavors.

Gelatin is a poly disperse macro molecule made of thousands of amino acid chains that are either free or linked to each other in a solution or a gel. Each amino acid chain has a molar mass between 10,000 and several hundred thousands Daltons (Mw). Hence, as a whole gelatin can have an averaged molar mass in the range of up to about 250 kDa.

For use according to the present invention, substantially any gelatin (or collagen) may be used herein to provide a gelatin hydrolysate or hydro lyzed gelatin. Gelatin (or collagen) hydrolysis may be done via a chemical route, such as by cooking with an acid, or via an enzymatic route. Herein, the term gelatin hydrolysate especially refers to hydrolyzed collagen or gelatin wherein at least 90% of the gelatin molecules have a molar mass of 50 kDa or lower, such as especially 20 kDa or lower.

Good results were obtained with hydrolysates obtained from gelatin type A and B, especially hydrolysates having an average molar mass of at maximum 20 kDa and having a pH in the range of 5-6. The isoelectric point of the former may be in the range of 7-9 and the isoelectric point of the latter may be in the range of 5-6. Especially, at least 90 wt.% of the gelatin hydrolysate has a molar mass of < 20 kDa, like < 10 kDa, such as < 5 kDa. Even more especially, at least 95 wt.% of the gelatin hydrolysate has a molar mass of < 20 kDa, like < 10 kDa, such as < 5 kDa. Yet even more especially, at least 99.5 wt.% of the gelatin hydrolysate has a molar mass of 50 kDa or less, such as < 20 kDa, like < 10 kDa, such as < 5 kDa. Especially however, at least 50 wt.%, even more especially at least 90 wt.% has a molar mass of at least 500 Da (about 3 amino acids), such as at least 1000 Da, such especially at least 1500 Da.

In the following, hydrolysates obtained from gelatin type A is referred to as "hydrolyzed gelatin type A". As indicated above, the method of the invention includes (i) providing a starting composition comprising a globular protein and water, and (ii) converting the globular protein and water into a gel, especially by heating the starting composition, to provide said gel-based product, wherein the starting composition comprises the globular protein in an amount of 5-20 wt.% relative to the total weight of the starting composition. Hence, by choosing a globular protein, its concentration, optionally additives (see also below), a gel may be obtained. This is known in the art.

Especially, a heat labile and globular protein, in a concentration range of about 5-20 wt.%, and at least above the critical gelling concentration, which is protein dependent, optionally in combination with a salt, such as with a salt concentration of up to 2.5 wt.% and such as e.g. up to 2.5 wt.% NaCl, more in particular up to 1.5 wt.% NaCl, and with a pH range between 3.0 and 8.0, a gel may be created upon heating. For example, a self-supporting gel may be provided by a composition of 10 wt.% whey protein gel with 100 mM (about 0.6 wt.%) NaCl, 12 wt.% ovalbumin, after heating. Another example of a gel with no salt is 10 to 15 wt.% pea protein which provides a gel at pH 3.0 after heating. The heating conditions are especially above the denaturation temperature, but in general in the range of 75-95°C, such as 80-90°C.

An interesting aspect of the invention is that substantially irrespective of the gelatin, one may choose the desired protein and its gelling conditions, and provide the gel in the presence of the gelatin hydrolysate. The gelatin hydrolysate surprisingly acts substantially inert and does substantially not affect the gel formation.

Hence, in a specific embodiment, the starting composition further comprises up to 2.5 wt.% salt, more in particular 0.2 - 2.5 wt.% salt (such as e.g. NaCl, KC1, NaBr, KBr, NaS0₄), such as up to 250 mM (about 1.5 wt.%) NaCl, wherein the starting composition has a pH selected from the range of 3-8, wherein at least 90 wt.% of the gelatin hydrolysate has a molar mass of < 50 kDa and wherein the globular protein and water are converted into the gel by heating the starting composition to a temperature selected from the range of 75-95°C. As indicated above, in a specific embodiment, the gelatin hydrolysate has an average molar mass of < 20 kDa. Even better results (especially in terms of gel properties and miscibility of the gelatin hydrolysate with the globular protein (gel)) were obtained wherein the gelatin hydrolysate has a average molar mass of < 10 kDa, such as < 5 kDa, even more especially < 3 kDa. More especially, at least 90 wt.% of the gelatin hydrolysate has a molar mass of < 50 kDa, such as < 20 kDa, like < 10 kDa, such as < 5 kDa, even more especially < 3 kDa.

The molar mass distribution of gelatin is usually measured by size exclusion HPLC (high performance liquid chromatography) techniques, and eluted fractions are detected by UV absorption at 210 nm and the measured data are evaluated by suitable software (such as PSS WinGPC® UniChrom software, Polymer Standards Service, Mainz, Germany and Cirrus GPC software). Calibration can be done with e.g. collagen (gelatin) fragments ordered from Filk Freiberg (Filk GmbH, Meissner Ring 1- 5, 09599 Freiberg/Sachsen, Germany).

For the separation of gelatin hydro lysates a column TSKgel G2000

SW_XL in combination with a SW_XL precolumn (Tosoh Bioscience Japan) can be used. As eluent composition, especially 0.2 M NaCl, 0.2 M NaH₂P0₄ : acetonitrile (85: 15), pH 5.3 is applied.

For the separation of non- hydro lyzed gelatins a separation column TSKgel G4000 SW_XL in combination with a SW_XL precolumn (Tosoh BioSCience Japan) can be used. As eluent composition especially 1.0% (w/v) SDS (sodium dodecyl sulphate) with 0.1 M Na₂S0₄ and 0.01 M NaH₂P0₄ at pH 5.3 is applied.

For molar mass analysis methods hydro lyzed and non-hydro lyzed gelatin solutions of 0.2%> and 0.07%> prepared, respectively. 20μ1 of this solution is injected for analysis. The analysis temperature is 30°C and an eluent flow rate of 0.5ml/minute is used. A more detailed description is given below.

In yet a further specific embodiment, the gelatin hydro lysate (used) comprises in the range of 5-95 wt.%> proteins.

Especially, the globular protein has an iso-electric point (IEP) selected from the range of 3-5.5. Globular proteins having these type of iso-electric points appeared to provide good gels. As source of globular proteins, e.g. isolates and/or concentrates may be used. Especially, the starting composition comprises one or more of whey protein isolate, pea protein isolate, albumin protein isolate, immunoglobulin protein isolate, and soy protein isolate. However, also other types of globular proteins may be applied. Further, additionally or alternatively also one or more of isolates or concentrates of such globular protein(s) may be used, such as (blood) plasma proteins. Especially however, isolates may be desired because globular protein isolates are especially able to form a gel with the appropriate texture. The term "globular protein" may also refer to a combination of different types of globular proteins. The gel-based product may be a gel as such, such as essentially consisting of the globular protein and the gelatin hydrolysate. However, the gel-based product may also be a more final product, such as a food product, a feed product, etc.. To obtain such more final product, one may include in the starting mixture already the necessary components to provide in substantially a single stage the (food) product, or one may first make the gel and then add the necessary components to provide the (food) product. Of course, also combinations of such processes may be applied. Hence, in a particular embodiment the method further includes the presence of one or more other components, to provide a human food product, a pet food product, a feed product, or a coating or an encapsulate for a medicament, etc.. In a specific embodiment the method comprises a first stage comprising converting the globular protein and water into the gel, in the presence of the gelatin hydrolysate, and a second stage wherein the gel and one or more other components are combined to provide said gel-based product. In yet another specific embodiment the method comprises converting the globular protein and water into the gel in the presence of the gelatin hydrolysate and one or more other components, to provide said gel-based product. The one or more other components may especially relate to food ingredients.

Further, especially the type of gelatin hydrolysate and amount of gelatin hydrolysate is selected to provide hydroxyproline in the range of 0.3-6 wt.%, especially 0.35-5.1 wt.% relative to the total weight of the starting composition.

Hence, in yet a further aspect, the invention also provides a method for the production of a gel-based product, the method comprising (i) providing a starting composition comprising a globular protein and water, and (ii) converting the globular protein and water into a gel, especially by heating the starting composition, to provide said gel-based product, wherein the starting composition comprises the globular protein in an amount of 5-20 wt.% relative to the total weight of the starting composition, wherein the starting composition further comprises a gelatin hydrolysate in an amount such that relative to the total weight of the starting composition hydroxyproline is available in the range of 0.3-6 wt.%, especially 0.35-5.1 wt. %, and wherein the starting compositions comprises 50-90 wt.% water, wherein at least 90 wt.% of the gelatin hydrolysate has a molar mass of < 50 kDa.

As indicated above, in a further aspect the invention provides a gel- based product obtainable by the method as described herein. Hence, especially a gel- based product obtained by the method as described herein is provided. Examples of such gel-based product are provided above, and may include a dairy product, a probiotic product, a protein bar, a sausage, a protein shake, or a protein drink. Especially however, the gel-based product comprises a solid product or a semi-solid product. Examples of semi-solid products are e.g. a pudding, jellies, custard, films, etc..

In a specific embodiment, the gel-based product as defined herein may comprise gelatin hydrolysate and globular proteins in a weight ratio of 1 :8 - 8: 1, such as 1 :4 - 8: 1, such as 1 :2 - 8: 1, like 1 : 1-8: 1, such as 1 : 1-4: 1, and optionally the gel- based product may further comprise 0.2-2.5 wt.% salt, such as NaCl. In a specific embodiment the product may comprise 0.5-2 wt.% NaCl relative to the total weight of the gel-based product. Further, especially relative to the total weight of the gel-based product, the weight of hydroxyproline is in the range of 0.3-6 wt.%, especially 0.35- 5.1 wt. %.

The protein content of the gel-based product may e.g. be evaluated with the method described by Kjeldahl or Dumas, especially the Kjeldahl method. When the outcome of these tests would significantly differ, an average value of both methods can be applied. The gelatin content may be evaluated by using a hydroxyproline based method. Based thereon, the globular protein content and gelatin hydrolysate content may be evaluated, as based on the hydroxyproline method, the gelatin hydrolysate contribution to the total protein content can be determined. Conversion factors may be taken into account depending on the type of protein, as is known in the art.

The hydroxyproline content can e.g. be determined by a continuous flow analyzer, such as a Skalar SAN++ System and Flow Access V3 software. Prior to analysis by the continuous flow analyzer, the sample has to be hydrolyzed with hydrochloric acid (into the separate amino acids). The automated method for the determination of hydroxyproline is based on its oxidation with chloramine-T in aqueous solution. The oxidation products are stained with 4- dimethylaminobenzaldehyde (pDAB) at 70°C and the chromogen obtained is measured at 550 nm. The absorbance is proportional to the amount of hydroxyproline in the original sample. Based on the animal source the hydroxyproline concentration can be converted to the gelatin (or collagen) concentration. The moisture content of the gelatin may be taken into account. BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments of the invention will now be described, by way of example only, with reference to the accompanying schematic drawings in which corresponding reference symbols indicate corresponding parts, and in which:

Figure 1 Storage modulus (open symbols) and loss modulus (closed symbols) for single SPI (V,T) and single WPI (□,■) gels after 1 hour at 95°C. Lines were fitted using equation 3.1. Gelatin type A ( ) and type B (♦) did not show an elastic response and only loss modulus values are shown;

Figure 2 Measured storage modulus (G'_measured) for gels after 1 hour at 95 °C containing 0.06 (□), 0.12 (V) or 0.16 (O) g_protei„ / gwater WPI with added gelatin A. Horizontal lines indicate G'_GP which is the storage modulus expected for pure WPI gels without added gelatin;

Figure 3 Ratio s (equation 3.2) for mixtures containing WPI / gelatin A (□), WPI / hydrolyzed gelatin type A (■); SPI / gelatin A (V) and WPI / gelatin B (O). Dashed line is drawn to guide the eye;

Figure 4 CLSM images of single WPI gels at 6 wt.%, 10 wt.% and 15 wt.% protein concentrations (scale bar 7.5 μιη);

Figure 5 Density correlation and normalized polarization function for WPI gels containing 0.06 (V), 0.10 (O) and 0.15 (□) g_protein / gwater · CLSM images of these samples are shown in Fig. 4. Graph A: Average density correlation from image analysis. Graph B: Normalized polarization from SESANS. Lines show fits used for data analysis for CLSM image analysis (equation 2.14) and SESANS (equation 2.1);

Figure 6 CLSM images of single and mixed globular protein (WPI or SPI) at room temperature after being heated to 95°C. All gels contain 0.06 g_protein / gwater globular protein (WPI / SPI). The location of globular proteins is shown in red and gelatin in green (details given in text). The first column contains reference images of globular protein gels without added gelatin. Within each row gelatin concentration increases from 0 to 0.02 and 0.06 gprotein / gwater gelatin. Scale bars are 7.5 μιη;

Figure 7 CLSM images from the Rhodamine channel (WPI), the FITC channel (gelatin type B) and an overlay (middle) of a sample containing 0.06 g / gwater WPI and 0.02 g / g_water gelatin type B;

Figure 8 Graph A: Separate density correlation functions from the Rhodamine channel (WPI, open symbols) and FITC channel (gelatin, closed symbols) for gels containing 0.06 g / g_water WPI and no gelatin (reference, O) and mixed gels containing 0.06 g / gwater WPI and 0.06 g / gwater gelatin type A (□) or type B (V). Inlet shows magnification of the area containing first minima (indicated by arrow). Graph B: First minima corresponding to typical building block diameter obtained from image analysis of CLSM images for mixed gels containing 0.06 g / g_water WPI and varying concentration of gelatin type A (□,■) or gelatin type B (V,T). Results from Pvhodamine channel (WPI) are shown with open symbols and from FITC channel (gelatins) with closed symbols;

Figure 9 Size of the gel building blocks (first minimum in p(r) from image analysis) for gels containing 0.06 g_protein / gwater (open symbols) or 0.1 g / g_water (closed symbols) WPI and different gelatins. Al : gelatin type A; A2: hydrolyzed gelatin type A; Bl : gelatin type B; symbols at zero gelatin ( ) represent pure WPI gels containing 0.06 and 0.1 g / g_water WPI;

Figure 10 Normalized polarization P(z) versus spin echo length from SESANS. at 40°C (O) and at room temperature (□) for a sample containing 0.06 g_protein / gwater WPI and 0.1 gprotein / gwater gelatin type A;

Figure 11 Normalized polarization versus spin echo length from SESANS measurements of mixed samples containing 0.06 gprotein / gwater WPI and 0.04 (V) or 0.07 gprotein / gwater (O) gelatin type A (graph A) or hydrolyzed gelatin type A (graph B). Lines represent behavior of pure data (from Fig. 5B) containing 0.06 g_protein / gwater (solid line) or 0.1 g_protein / gwater WPI (dashed line);

Figure 12 Schematic representation of gelation for pure WPI gels (row A) and WPI gels with added gelatin (row B). WPI is represented by black spheres and gelatin as a linear polymer. In scheme B.4 closed spheres represent situation at low total protein concentration and open circles indicate changes upon increasing the total protein concentration;

Figure 13 Critical volume fraction calculated for hard sphere systems as a function of the particle size ratio q. The location of WPI / gelatin A (O) and WPI / gelatin B (V) systems where increased phase separation was observed and the WPI / hydrolyzed gelatin type A where the microstructure of WPI was similar to that of pure WPI (□) are indicated in the graph. The arrow indicates the direction the size ratio q changes upon gelation of WPI. Graph re-printed from earlier publication (Interactions in Protein Mixtures. Part II: A virial Approach to Phase Behaviour. Food Hydrocolloids, 52, 991-1002, 2015); Figure 14 SEC-MALLS elution profile containing the relative refractive index (dashed lines, left vertical axis) and molar mass from light scattering (solid lines, right vertical axis) for WPI, gelatin type A and type B taken from reference (Interactions in Protein Mixtures. Part II: A virial Approach to Phase Behaviour. Food Hydrocolloids, 52, 991-1002, 2015) and hydro lyzed gelatin type A as indicated in the graph. These measurements are based on the three columns in series (TSK gel G5000 PW_xi + TSK gel G3000 PW_xi + TSK gel G2500 PW_xi) as defined above. The molar masses derived therefrom may differ with respect to the above mentioned method with the respective columns TSKgel G2000 SW_XL and TSKgel G4000 SW_XL in range of about less than 5%. Hence, the molar mass values herein are provided within about +/- 10%, such as within about +/- 5% (such as e.g. 50 kDa +/- 5%);

Figure 15 depicts the molar mass distribution (as determined with the TSKgel G2000 SWxl and TSKgel. G4000 SWxl methods) with on the x-axis the molar mass (Da) and on the y-axis the relative intensity (W(logM)) of gelatin hydrolysates (the two lower curves, with respective mass averaged molar masses of 2 and 5 kDa), and non-hydrolyzed gelatins with respective mass averaged molar masses of 140 and 190 kDa. Herein the relative intensity means the relative light absorption.

Figure 16 depicts the Young's modulus at varying salt concentrations at two different temperatures (A at 20°C; B at 40°C) for mixed gels prepared from whey protein (WP) plus G290 gelatin type A and from whey protein plus gelatin type A H5kDa.

Figure 17 shows the influence of the weight ratio (g/g) of gelatin (hydro lysate) and WPI at about 150 mM ionic strength on the relative modulus. The relative modulus is expressed as the ratio of the modulus of the mixed gels of gelatin and whey protein over the modulus of pure WPI gels. The data are presented for the following gelatin (hydrolysates) H5kDa (exp. 1, rheometer), D5kDa (exp. 2, texture analyser), gelatin type A G290, and gelatin type B G260.

The schematic drawings are not necessarily on scale. DETAILED DESCRIPTION OF THE EMBODIMENTS

The gel microstructure and rheological response of globular protein gels (whey protein isolate, WPI and soy protein isolate, SPI) in the absence or presence of gelatins (type A, type B and hydro lyzed gelatin type A) was investigated. Microstructural information was obtained using a combination of image analysis on CLSM images and spin echo small-angle neutron scattering (SESANS).

Addition of non-hydrolyzed gelatins leads to a coarsening of the globular protein gel structure, which was related to a reduction in storage modulus of globular protein gels. The presence of gelatin hydrolysates on the other hand surprisingly did not induce microstructural changes. At conditions where proteins only interact via hard body interactions microstructural and rheological changes in globular protein gels could be attributed to the molecular sizes of proteins and their relative changes during gelation.

Gelation is a common phenomenon in foods. The increasing understanding of the mechanisms that lead to gelation such as protein denaturation, reduction of electrostatic repulsion or formation of junction zones between molecules is an important first step towards describing the gelation of foods. In foods the main molecules responsible for gelation are proteins and polysaccharides. Gelation requires a cross-linking of the individual molecules to form a space spanning network which one observes macroscopically as a transition from a liquid to a semi-solid. In the case where two types of biopolymers (e.g. proteins and polysaccharides but also different types of proteins or different types of polysaccharides) are present during gelation they may interact before and during gelation, which might change the properties of the formed network. Understanding the changes in the molecular interactions between biopolymers in solution is challenging but essential to understand the formation of mixed gels.

For most food relevant biopolymers and their mixtures, information on the molecular interaction, the relevant molecular size and the changes during gelation are not available in scientific literature.

In this research we investigate the relationship between molecular interactions, molecular size on the microstructure and rheology in the gelatin / globular protein mixed system. In this system both proteins are able to form a gel. Dependent on whether gelatin or globular proteins are gelled first different microstructures and rheological responses can be obtained. During gelatin gelation in these mixed systems an increase in storage modulus due to segregative phase separation was found for systems whenever the globular proteins were larger (in terms of their effective radius) than the gelatin molecules. In systems where globular proteins were gelled before gelatin gelation the systems were found to have bi-continuous microstructures and their modulus was described based on rheological models for phase separated mixed gels.

Herein, amongst others the gelation of globular proteins in the presence of (non-gelled) gelatin has been investigated. For these mixed systems research showed bi-continuous microstructures independent of protein mixing ratios and total protein concentrations. The absence of a phase inversion in the gelatin / globular protein gels results in a gradual change in fracture properties of the mixture gels dependent on the total protein concentration and mixing ratios. This gradual change makes the mixed gels interesting for application purposes in foods where a large variety of textural aspects can be reached using mixed gels. The gradual change in fracture properties is especially interesting as it was observed for mixtures of gelatin with whey protein concentrate, egg white and soy protein isolate.

Herein, we focus amongst others on the formation, microstructure and rheology of the globular protein gel in the first stage of the sequential gelation of both proteins. Rheological properties of the globular protein gels were studied above the helix to coil transition temperature of gelatin where gelatin had no viscoelastic response. The microstructure of the mixed protein gels was selectively probed using spin echo small-angle neutron scattering (SESANS) and confocal laser scanning microscopy (CLSM) in combination with image analysis and covalently labelled gelatin. Mixed gels of globular whey proteins (pi ~ 5, average M_w 35 kDa) or globular soy proteins (pi ~ 4.8, M_w glycinin 360 kDa, β-conglycinin -150 kDa) mixed with gelatin type A (isoelectric point (pi) =8, average M_w 125 kDa) or type B (pi = 5, average M_w 150 kDa) or hydro lyzed gelatin type A were investigated. By combining these globular proteins with the different gelatins, their molecular size ratio and molecular interactions can be varied. Understanding the effect of these molecular parameters on the microstructure and rheology of the globular protein gel is the first step in understanding the changes in this complex mixed protein gel.

Soy protein isolate (SPI) was extracted using isoelectric precipitation from defatted soy flour as described earlier.

Whey protein isolate (WPI) was purchased from Davisco Foods

International Inc. (Le Sueur, USA, MN) product name BIPRO (94 % protein determined by Kjeldahl).

Commercially available Rousselot gelatins (type A and B) were used. Type A (from porcine skin) had a bloom strength of 290, isoelectric point (pi) around 8 and a protein content of 89.6 wt.% Type B (from bovine bones) had a bloom strength of 260, pi of around 5 and a protein content of 88.8 wt.%. A detailed characterization of these research materials was described previously (Interactions in Protein Mixtures. Part II: A virial Approach to Phase Behaviour. Food HydrocoUoids, 52, 991-1002, 2015).

Further, a commercially available Rousselot hydrolyzed gelatin type A (product name Peptan 5000) was used.

Additionally, hydrolyzed gelatin type A was prepared from commercially obtained Rousselot gelatin type A (G290, gelatin type A, 290 bloom) as follows. G290 solutions of 15 wt.% were enzymatically hydro lysed with an endopeptidase. The pH of the gelatin solutions was adjusted to pH 8, for optimal functioning of the enzyme. The hydrolysis was carried out under constant stirring in a water bath at 55°C for 5-120 min at an enzyme/G290 ratio of 200 ppm. The hydrolysis was stopped by putting the samples in a 95°C-water bath for 10 min. The hydrolysate solutions were frozen and then lyophilised during several days at -25°C and 0.63 mbar, followed by one day at -76°C and 0.0010 mbar with a Christ Alpha 2-4 LD plus freeze-dryer (Martin Christ, Osterode am Harz, Germany). The dried samples were crushed manually with a spatula and the dry matter content (~99 wt.%) was determined by weighing the samples before and after drying in an oven at 105°C for 16 hours.

The molar mass (molecular weight (Mw)) distributions and fractions of the proteins have been determined with an Agilent 1260 Infinity HPLC system

(Agilent Technologies Inc., Santa Clara, CA, USA) equipped with a TSKgel G2000 SW_XL column (Tosoh Bioscience, Japan), preceded by a SW_XL precolumn (Tosoh Bioscience, Japan). The measurements were carried out at 30°C using a mobile phase containing 0.2 M NaCl, 0.2 M NaH₂P0₄ and acetonitrile (85: 15) in deionized water, of which the pH was adjusted to 5.3 with NaOH. Hydrolysed gelatin solutions were prepared at 2 wt.% and were diluted to 0.2 wt.% with the mobile phase. The samples were filtered through Phenex-RC 0.45 μιη membrane filters (Phenomenex, Vasrlose, Denmark) before analysis. The analysis was carried out with an injection volume of 20 μΐ, a flow rate of 0.5 mL/min, and the signal of the molecules was collected at 210 nm. Data were acquired with PSS WinGPC® UniChrom software (Polymer Standards Service, Mainz, Germany) and the fractions for the Mw ranges 0-10 kDa, 0- 20 kDa, and 0-50 kDa were determined. Samples of 2 wt.% of non-hydrolysed gelatin were diluted to 0.07 wt.% with the mobile phase (1.0% (w/v) SDS (sodium dodecyl sulphate), 0.1 M Na₂S0₄ and 0.01 M NaH₂P0₄, pH adjusted to 5.3). The analysis was carried out with a TSKgel G4000 SWXL column and a SW_XL precolumn (Tosoh Bioscience, Japan).

The gelatins and gelatin hydro lysates used herein are listed in Table 1.

Table 1

Other chemicals such as acids and bases used for pH adjustments were of analytical grade and purchased from Sigma Aldrich (Steinheim, Germany).

Whey protein (WP) stock solutions of 20-25 wt.% were prepared by dissolving WPI powder in deionized water (18.2 ΜΩ/cm, Elga Purelab Ultra, Bucks, UK) with the desired NaCl concentration and stirring them at room temperature for at least 2 hours. Gelatin stock solutions of 10-15 wt.%> were prepared by adding the gelatin to deionized water containing the desired NaCl concentration, letting it swell for 30 min and dissolving it in a 60°C water bath for another 30 min. The stock solutions were kept in the fridge overnight. Before sample preparation, the stock solutions were heated in a 40°C water bath until the gelatin was melted. The pH of the stock solutions was measured and adjusted to pH 7.0 ± 0.1 with 5 M NaOH. Samples with different protein concentrations and protein ratios having ionic strength values of 150 mM (WPI containing samples) and 300 mM (SPI containing samples), or samples containing 0.26 w/w_water total protein (ratio 0.6 gelatin per WP) and up to 400 mM salts (1.9 wt.%)) were prepared by mixing protein stock solutions and deionized water containg the same NaCl concentration (at 40°C). For the final ionic strength of the samples, NaCl, NaOH from pH adjustment and ions from the gelatin were taken into account. The samples were kept in a 40°C water bath for at least 30 min under constant stirring and were degassed in a vacuum desiccator connected to a vacuum pump (RV3 A65201903, Edwards, Crawley, UK) until no more gas bubbles were formed, while being slightly stirred. Sample handling was done at 40°C where gelatin solutions were in the liquid state.

Samples were measured in a rheometer (Anton Paar MC502) using a sand-blasted cup-bob geometry (CC17). They were added in the liquid state (heated for 30 min to 40°C) and covered with paraffin oil to avoid evaporation. Measurements were performed at a strain of 0.5% and a frequency of 1 Hz. Samples were heated inside the equipment from 40°C to 95°C at 5°C/min and afterwards kept at 95°C for 1 hour.

For large deformation rheology measurements, the liquid samples were transferred into 20 mL Omnifix® syringes (20 mm inner diameter) with Luer-Lock- fitting (Carl Roth, Karlsruhe, Germany). The inside of the syringes was lubricated with paraffin oil to ease the gel removal before measurement. Air bubbles were removed as much as possible before the syringes were closed with Rotilabo®-Luer-stoppers (Carl Roth, Karlsruhe, Germany) and tempered in a 40°C water bath for 30 min. Gel formation took place in vertical position in a water bath at 95 °C for 30 min, with a weight on top of the syringes to limit gas bubble formation. The gels cooled down overnight at room temperature. Uniaxial large deformation measurements were carried out with a texture analyser (TA-XT Plus, Stable Micro Systems ltd., Godalming U.K.), equipped with a 50 kg load cell and a 6 cm diameter cylindrical probe. A double- walled beaker connected to a circulating water bath set at 50°C was fixed to the plate of the texture analyser to enable measurements at temperatures where the gelatin phase is molten. The samples were tempered in a water bath of 45°C until measurement, at least for 30 min. Just before measurement, the gels were removed from the syringes and cut into 3 cylindrical pieces of 2 cm height with a 0.2 mm wire. A few drops of paraffin oil were put underneath and on top of the sample for lubrication. The measurement was carried out with a probe speed of 1 mm/s and the sample was compressed to 2 mm (10% of initial height), which usually led to fracture. While the first piece was measured, the other 2 pieces were kept in an oven at 45°C to keep the gelatin phase melted. Hereafter, the measurements are called measurements "at 40°C" (gelatin phase liquid). Measurements carried out at room temperature are called measurements "at 20°C" (gelatin gelled, if Mw high enough). The samples for measurement at 20°C were kept at room temperature overnight and until measurement. The raw data (force [N] and distance [mm]) were converted to engineering strain and engineering stress and subsequently to true strain and true stress as shown below, in order to determine the Young's modulus.

Engineering strain y_e = ^

where L is the original sample height [m] and I is the actual height [m] at a certain compression.

Engineering stress a_e =—

where F is the compression force [N] and Ao is the initial cross-sectional area of the sample [m²].

The engineering strain and stress were recalculated to true strain or

Hencky strain ε_Η and true stress a_t to correct for changes of the cross-sectional area of the sample during compression. The shape (cylinder) and the volume were assumed to be constant.

True strain (Peleg, M. (1987). The Basics of Solid Foods Rheology. In H. R. Moskowitz (Ed.), Food Texture (p. 352). CRC Press. Retrieved from https ://books. google. com/books?id=U 1 - JusgPIFIC&pgis= 1 )

_¾ = >^η (^Η^η ( + ΐ)

ε_Η = \n(engineering strain + 1)

p

True stress σ> = - τ A

where A is the actual cross-sectional area of the sample [m²].

The true stress and true strain values calculated from the raw data determined in the course of compression were plotted in a stress-strain curve. In the beginning of the compression phase the response is linear and purely elastic. The slope in this region is called Young's modulus (elastic modulus for uniaxial compression) and is a measure for the stiffness of the material, the force per unit area needed to deform the material. In this research, the Young's modulus was determined in the range of 0.03-0.06 true strain.

The molar mass distributions from the various gelatin hydro lysates and the commercially available gelatin type A samples were measured. The Young's modulus (Table 2) of the mixed gels prepared with the whey protein and the various gelatin hydrolysates and the commercially available gelatin type A samples were measured according to the method described above.

Table 2. Molar mass distribution and Young's modulus of gelatin hydrolysates and commercial gelatins type A.

From these data it can be concluded that there is a decrease in Young's modulus for increasing molar mass of gelatin. This is in line with an increase in fractions >50kDa. These fractions are not in favour for homogeneous globular gel formation. A maximum amount of >50kDa fractions is -10%. Above this value the gel structure is significantly changing. This implies that the fractions <50kDa are at least 90%. Gel formation is not affected in case the fractions <20kDa are above 90%. Preferred are the gelatins with an average molar mass of <10kDa. They don't influence globular protein gel formation

Figure 17 shows the influence of the weight ratio (in g/g) of

(hydro lysed) gelatin and whey protein (WP) at about 150 mM ionic strength on the relative modulus. The relative modulus is expressed here as the ratio of the modulus of the mixed gels of gelatin (hydrolysate) and whey protein over the modulus of whey protein gel. From the data it can be concluded that stable gel-based gelatins can be prepared from (gelatin hydrolysate) : (protein ratios) from at least 4 : 1 to 1 : 8. However also at a ratio of hydrolyzed gelatin / whey or globular protein of 8 : 1 the gelatin structure (modulus ratio) is still higher compared to non-hydrolyzed gelatin. And as a result it is concluded that stable gel-based gelatins can be prepared from (gelatin hydrolysate) : (protein ratios) from at least 8 : 1 to 1 : 8. It is hence concluded that the amount of hydrolyzed gelatin is a limitation for the structure. The amount of globular protein is far less critical. For e.g. Fig. 14, size exclusion chromatography with multi angle laser light scattering (SEC-MALLS, Agilent technologies, 1200 series) was performed using three columns in series (TSK gel G5000 PW_XL + TSK gel G3000 PW_XL + TSK gel G2500 PW_XL). The columns were heated to 60°C and separation performed at a constant flow of 0.5 ml / min using 10 mM phosphate buffer at pH 6.8 with 125 mM NaN0₃ and 0.02 % NaN₃. The multi angle light scattering (acquired using Wyatt Dawn Heleos II) data was fitted using a first order Zimm model (using the software Astra 6).

In SESANS samples deuterium oxide (D₂0) was used instead of deionized water for sample preparation and WPI solutions were filtered using a syringe filter with a 1.2 μιη cut-off (Sartorius Minisart^®). Samples were heated inside the SESANS cuvettes (path length 1 cm) for 30 minutes at 95°C and subsequently cooled down to room temperature at which temperature the SESANS measurements were performed if not stated otherwise. SESANS measurement were performed at the reactor institute at Delft University of Technology using the setup described by Rekveldt et al. (2004). For the data analysis we use routes described by Andersson et al. (2008) which will be outlined in short below.

The from the SESANS measurements obtained normalized polarization (normalized by the empty beam) was described by (2.1):

P(z) = e^∑tl^G -^ with P(z) the normalized polarization, z the spin echo length,∑_t the scattering power (which indicates the average number of times a neutron scatters while transversing the sample) and G(z) a normalized dimensionless function which contains information about the microstructure of the sample. In this study we have used the self-affme random density distribution model to describe G(z) which is given by (2.2):

where a is a measure for the size of the random inhomogeneity in the system. Γ is the gamma function, H the Hurst exponent (related to the fractal dimension of the systems) and K the modified Bessel function. Protein gels are well described as a two phase system with one phase being the protein phase and the other the solvent phase (often called gel pores). For these systems the scattering power∑_t is given by (2.3)

with λ the wavelength (0.203 nm), t the path length (1 cm), Φ the volume fraction of one of the phases, Δρ the neutron scattering length density contrast between the protein phase and the solvent (Δρ = p_D20 - pprotein with p_D20 = 6.38x10¹⁴ m^"2 and p_protem = 3x10¹⁴ m^~2) and ξ the correlation length of the system which is related to the size of the random inhomogeneity via (2.4)

Γ(Η +

ξ = 2π^²α V₂)

Γ(Η)

Gels were prepared by adding the liquid sample (at 40°C) to hermetically sealed cuvettes (Gene Frame^® 125 μΐ, obtained from Thermo Scientific) and heating these cuvettes on a peltier element with the same heating / cooling profile as used for small deformation rheology. After 1 hour at 95°C the samples were cooled at 5°C / min to 15°C and kept at 4°C until analyzed. Globular proteins were stained using 0.001 wt.% Rhodamine B. In some samples gelatin was covalently labelled with Fluorescein Isothiocyanate (FITC) as known in the art. Imaging was performed using a Leica DMI6000 microscope (Wetzlar, Germany) at room temperature. Images were taken at a sample depth of 5 μιη from the cover glass which was in contact with the peltier element during heating at several randomly chosen locations in the sample and different magnifications. Measurements were performed at 1024 x 1024 measurement points (pixels) at a scanning rate of 400 Hz and two frames were averaged. The sequential mode was used, FITC was excited at 488 nm (measurement between 500 and 570 nm) and Rhodamine B was excited at 561 nm (measurement between 570 and 790 nm).

Confocal laser scanning microscopy (CLSM) has become a common tool in many scientific fields including food science. Often it is used to qualitatively visualize microscopic structures even though more and more frequently, it is also used for quantitative analysis. While for qualitative analysis it is sufficient to just consider the image as a picture, for quantitative analysis the image represents an intensity matrix of spatially related data points. Each data point (for ease of reading referred to as pixel from here onwards) is characterized by its location and value within the intensity matrix. This intensity value is proportional to the concentration of fluorescent dye within the three dimensional sample space that the pixel represents. This space (often referred to as voxel) is best represented by a cube having an area (pixel size, lateral, x,y) and a depth (z). The voxel depth (z) is influenced by CLSM parameters (wavelength, pinhole, numerical aperture of the objective) and sample properties (refractive index, internal scattering) and is commonly approximated using (2.5)

V ⁿ ^exc

z = 5—

N, where p is the airy unit, n the refractive index of the media between objective and sample, λ is the wavelength of the laser and N_AP the numerical aperture of the CLSM optical system. The area of the voxel in x,y plane can be defined by the user during imaging when setting the number of measurement points (often referred to as number of pixels in CLSM images). When interested in the spatial arrangements of molecules / structures the pixel size should to be smaller than the so-called point spread function (PSF) as otherwise no real correlation between voxels exists. Within this research the PSF was estimated using a commercially available software (Huygens Essential version 3.3.0p5; Scientific Volume Imaging B.V.; Hilversum, the Netherlands). The same software also allows for deconvolution of the images which leads to reduction of the optical blur introduced by the PSF. For a representative set of images deconvolution was performed and no significant changes of the correlation functions as discussed later on were found for most of the images. In some cases, however, deconvolution induced artefacts which significantly changed the image analysis outcome. This is most probably due to the unknown refractive index of the sample as used to estimate the PSF. Deconvolution was therefore not further applied within this research.

To obtain quantitative structural information from CLSM measurements (images) we have analyzed the spatial autocorrelation of the images. The spatial autocorrelation of a (digitized) image (2 dimensional, (2D) matrix) with M x N pixels in the x and y direction respectively is given by (2.6): M N

G(a, b) = ^ ^ i x,y)i x + a,y + b)

= MN < i(x,y)i(x + a,y + b) >_xy where i(x,y) is the intensity value of a pixel with coordinates (x,y) in the image and <...>_x,y represents the average over all x and y. The autocorrelation G(a,b) is typically normalized by the average intensity of an image as the average is influenced by acquisition factors which are not part of the structure of interest. The scaled intensity or fluctuation of intensity can be written as (2.7) i(x,y)-< i(x,y) >

< i(x,y) > which in case of the auto correlation function leads to the following expression for the scaled auto correlation function (g(a,b))(2.%)

< i(x,y)i(x + a,y + b) >

g( ,b) =< δί(χ,γ)δί(χ + a,y + b) >- < i(x,y) >²

\

N ΓMΤΊΣ_Χ ∑y i(x,y)i(x + a,y + b)

(NM y ^ )

It can be shown that the detected intensity fluctuation di(x,y) in a certain volume element (voxel) is linearly related to the concentration of fluorescent dyes and protein concentration. The autocorrelation function is therefore related to the density correlation function of the fluorescent dye and the stained sample structures given by (2.9):

(p r')p r' +r)>

:(r) = _\2 = g{a,b) + l

<p(r'))² with p(r') the density of fluorescent molecules at position r where r' = xi + yj and r = αϊ + bj are the position vectors and ί and j unit vectors in the x and y direction, respectively. For isotropic systems, p(r) = p(r), with r = \r\. Therefore, equation 2.9 can be radially averaged leading to the radial density autocorrelation function c(r) (2.10).

C r = 9 (f) + 1 = (g (a, b) )_a2 _+b2 _=r2 + 1 For non-crystalline species, structures are randomly arranged and therefore uncorrected at larger distances which can be written as (2.11): r _- <p(r)) <p(r)) <p>²

hm c(r) =— . _→ .— = = 1

r→∞ (p(r)) (p)

Here p is the average density. The value of the normalized autocorrelation function c(r) at r = 0 equals the variance of the fluctuation intensity δί and is a measure of how much of the total intensity is concentrated in areas with higher intensity relatively to the average intensity of the image. This value (also referred to as σ²) has been used in literature as a measure for the coarseness of gels and can be calculated directly from the image using (2.12):

Using these two conditions at r = 0 and r ∞, the density auto-correlation function c(r) can be expressed as the normalized density auto-correlation function p(r) (equation 2.13). For ease of reading we will refer to this function simply as correlation function (2.13): c(r)— 1

The function p(r) gives the probability to find another point in the image with the same intensity as a function of the radial distance r. Distance in this case can be interpreted as the distance from the center of structures with highest intensities (in case of protein gels the gel building blocks or aggregates). The characteristic parameters of this function such as the zero crossing, first minima and maxima describe different properties of the observed system. In a system with only one type of structures this curve is typically well described by a stretched exponential decay such as (2.14):

with ξ the correlation length of the system and β a form factor. The presence of multiple structures or non-isotropic structures would require the introduction of double decays and non-symmetrical functions which is beyond the scope of this article.

In this research all calculations regarding the image analysis were performed using MatLab^® (R2013a Version8.1.0.604; MathWorks Inc.). Calculation of the autocorrelation function was performed using the standard routine in the Diplmage library (Version 2.5.1; Quantitative Imaging Group, Delft University of Technology). For each image a background image was subtracted before analysis. Background images were obtained from images using a Gaussian filter with a sigma value much larger than the observed structures. This procedure was found not to change results for images with a homogeneous intensity distribution while being able to correct for errors introduced by the CLSM optics (e.g. darker corners) where present.

Single and mixed protein systems were characterized for their rheo logical response at 95°C. Fig. 1 shows the storage modulus (G') and loss modulus (G") of single whey protein isolate (WPI) and single soy protein isolate (SPI) gels as a function of protein concentration. The protein concentration was based on the available amount of water, given in grams of protein per gram of water (g_protein / gwater). The storage modulus G' of WPI and SPI was fitted using the percolation model (equation 3.1) with an earlier described fitting procedure. For WPI a good description of the data over a large protein concentration range was found. For SPI the concentration range was limited by the minimum gelling concentration (~ 0.04 g_protein / gwater) and the concentration of the SPI stock solution obtained from the isolation procedure (~ 0.11 gprotein / gwater). Yet, also for SPI the fit describes the G' values well within the range of measured protein concentrations and can be used to interpolate G' values for any protein concentration (3.1):

G ' — C(c_p — CQ ) Here C and t are scaling constants, c_p the protein concentration (in g_protein / gwater ) and Co the minimum gelling protein concentration.

We would like to stress that the percolation model is used semi-empirically in this research. Percolation theory assumes the network to arise homogeneously throughout the gelation process which is strictly spoken not true for WPI or SPI gels. In gels of globular proteins especially when gelled in the presence of salt, micro-phase separation occurs, resulting in a non-homogeneous system with protein rich and protein poor (pores) phases. One therefore should be careful in interpreting the fitting parameters from the percolation model.

For pure gelatin samples no elastic response was observed at 95 °C, in line with its coil conformation at this temperature. Therefore, also in a mixed globular protein / gelatin system, the measured G' values can be attributed solely to the globular protein gels (WPI or SPI). The concentration of globular proteins in the mixed gels is known. Using the globular protein concentration and equation 3.2 it is thus possible to calculate the expected G' value (denoted G '_GP) for a mixed SPI or WPI gel. This so calculated G '_GP value is expected from a globular protein gel assuming that the presence of gelatin does not change the storage modulus. For the calculation of G '_GP it is important to express the protein concentration based on the available water rather than total sample weight.

The measured storage modulus (G 'measured) and theoretical G O_P for WPI gels at three different whey protein concentrations as a function of gelatin type A concentration are shown in Fig. 2. At low gelatin concentration no differences between the measured and theoretical storage modulus were observed. With increasing gelatin concentration measured G 'measured values are below the theoretical G O_P for single WPI gels. This behavior is opposite to the earlier reported increase in G' upon the addition of (non-gelling) globular proteins to gelatin gels which will be addressed later on. To express the difference between G 'measured and G '_GP independent of total protein concentration (and therefore absolute G' value) we introduce the ratio s defined by (3.2):

C ^u measured

s^≡~r ^U' GP Fig. 3 shows the ratio s for mixed SPI / gelatin type A, WPI / gelatin type A, WPI / gelatin type B and WPI / hydrolyzed gelatin type A gels as a function of gelatin concentration. The ratio s of SPI and WPI gels decreases with increasing gelatin type A concentration. For WPI and SPI s values follow the same curve, even though there is a difference of 2 orders of magnitude in absolute storage modulus values between these systems (see Fig. 1). Also the ratio s for mixed WPI gels with added gelatin type B decreases with increasing gelatin concentration. The decrease in s occurs at lower gelatin concentration for gelatin type B compared to gelatin type A. For both gelatin types s levels off around s = 0.2 at high gelatin concentrations.

Fig. 3 also shows the ratio s of WPI gels prepared in the presence of hydrolyzed gelatin type A. Here s > 1 for all measured gelatin and WPI concentrations (performed in the range between 0.02 to 0.12 g_protein / gwater hydrolyzed gelatin type A and 0.05 to 0.20 g_protein / gwater WPI). s > 1 indicates that these mixed gels showed an increased storage modulus compared to the one expected for the single WPI gel at identical WPI concentration (corrected for the availability of the water). Increased values compared to the pure WPI gels are most probably due to a high affinity of gelatin hydro lysate to water which might reduce the availability of water to WPI.

To relate the rheological properties to the microstructure of mixed gels we have used neutron scattering (SESANS) and image analysis on CLSM images for single and mixed gels.

Fig. 4 shows CLSM micrographs of single WPI gels at three different protein concentrations. The images show typical gel structures consisting of a connected network of aggregated protein. These aggregates are typically referred to as building- blocks which on a certain length scale, connect to form a space spanning network. For protein gels, the size of the building blocks is an important parameter in their characterization. The building block size can be determined from the first minima of the radial density correlation function p(r) (equation 2.13) of these images which is shown in Fig. 5A. This minimum represents the distance from any location with high protein WPI concentration (high intensity) towards the lowest intensity (background). The other characteristic property related to the size of the building blocks is the correlation length ζ which was obtained by fitting equation 2.14 to p(r) as indicated by the solid lines in Fig. 5A.

An additionally important property for gel structures is the way how the aggregates are arranged. This is described by two parameters. First, the typical distance between aggregates, given by the first maximum of the correlation function where the probability to find another building block is highest. Secondly, their arrangement also determines the size of the gel pores, typically defined as the region depleted of protein between the gel building blocks. With the pores having a significantly lower intensity (lower protein concentration) than the protein gel, their size can be extracted from the width of the minimum in the density correlation function p(r). Here it was defined as the distance between the first and second zero crossing of the density correlation function. For a better visible example of the negative peak and first maxima please see Fig. 8).

The last parameter of interest when analyzing protein gels is the structure within the gel building blocks. Despite the fact that the q^"1 range of the CLSM observation is not sufficient to extract a fractal dimension of the building blocks, the slope of p(r) can be used to compare different structures. This can be done by determining the linear region of p(r) over the relative distance (radial distance / size of building blocks). The density is also represented in the parameter β of equation 2.14. For β = 2 this equation corresponds to a Gaussian model which would be expected for a 2D projection of a sphere with homogeneous internal structure, β values above 2 are expected for disks. Lower values can have various reasons, one being a non- homogeneous intensity (protein) distribution or an irregular shape of the aggregates. Table 3 summarizes the different parameters obtained from image analysis for single WPI gels at three protein concentrations. With increasing protein concentration the size of the gel building blocks decreases (correlation length and diameter) so does the distance between them and the pore size. These changes are typically summarized in the term "denser protein network" which can also be observed in the images in Fig. 4.

Table 3 Characteristic values describing the WPI gel structure obtained from CLSM image analysis (first part of the table) and SESA S analysis (second part of the table) for single WPI gels at three different protein concentrations. Standard deviations are given for values where several independent measurements were taken.

Parameter Source Interpretation Unit 6% 10% 15%

Fit of Characteristic length

Correlation 569.5 360.6 285.9 equation scale of areas with high nm

length (ξ) ±18.7 ±5.7 ±21.9

2.14 fluorescent densities Parameter Source Interpretation Unit 6% 10% 15%

The diameter of the

First

spherical equivalent of

Diameter of gel Minima of 2377 1368 1343 the gel building blocks nm

building blocks equation ±418 ±79.2 ±384 as visible in CLSM

2.13

images

Width

The diameter of a

negative

spherical equivalent of 6856 ± 2794 ±

Gel pore size peak nm -- the gel pores (protein 168.75 600.51 equation

depleted areas).

2.13

First Typical distance

Building block maxima in between two gel 4346 1878 1804 nm

distance equation building blocks, center ±1985 ±111 ±270

2.13 to center

Arrangement of

Fit of

protein inside gel 1.61 1.76 1.43

Exponent β equation -- building blocks (related ±0.08 ±0.01 ±0.04 2.14

to structure)

Average how often a

Scattering power Equation

neutron scatters when 0.43 0.28 0.18 (∑t ) 2.1

passing the sample

Size of

Size of random

(intermediate) Equation

inhomogeneity in nm 1055.2 309.6 116.4 gel building 2.2

sample

blocks (a)

Hurst exponent Equation Related to fractal

-- 0.069 0.29 0.45 (H) 2.2 dimension of gel

Correlation Equation Sample specific typical

nm 420.9 381.3 216.1 Length (ξ) 2.4 length scale

Fig. 5 shows both results from CLSM image analysis and the normalized polarization for single WPI gels from SESANS measurements. While for the analysis of the CLSM images one assumes isotropy to extract information of the 3D gel structure from a 2D projection (image) SESANS directly probes a relatively large 3D sample volume (approx. 1000 mm³). The results therefore represent an average over a large number of structural elements (gel building blocks) which makes this technique an interesting addition to the above described image analysis results.

With increasing WPI concentration the initial slope of the normalized polarization versus spin echo length curve changes. The initial slope in this graph represents the volume fraction of WPI inside the gel building blocks over a large range of length scales (horizontal axis on log scale). A steeper slope at increasing protein concentration (from 0.06 to 0.15 g_protein / gwater WPI) reflects the increased volume fraction (higher protein density) of WPI in the gel building blocks. Additionally, the observed decrease in total scattering (plateau value of normalized polarization at high spin echo length) with increasing WPI concentration suggests a decrease in the size of the typical gel building blocks. To quantify these changes, we have fitted the data in Fig. 5B to equation 2.1. The obtained results are shown in Table 3 together with the results from CLSM. With increasing WPI concentration the size of the random inhomogeneity (a) and the correlation length of the system (ξ) decrease and the Hurst exponent increases. Low Hurst exponent at low protein concentrations indicate an open structure where protein is non-homogeneously distributed. At increasing WPI concentration the Hurst exponent is closer to 0.5 which indicates that the protein is more homogeneously distributed inside the sample (on the length scale probed by SESANS) which is in line with the visual observations of the CLSM images as shown in Fig. 4. The correlation length obtained from SESANS are in agreement with results obtained from CLSM. The agreement between the two techniques validates the obtained results from each separate technique and indicates that each of the techniques can be used to obtain structural characteristics on the different length scales of protein gels. More than that, both techniques are complementary and together allowed a full characterization of the WPI gel structures (as shown in Table 3) which in this research serves as a basis for the analysis of the gel structures in mixed gels.

Fig. 6 shows CLSM images of mixed gel structures where globular proteins (WPI or SPI) were gelled in the presence and absence of gelatin. The first column of Fig. 6 shows single globular protein gels (made of only WPI or SPI). The second and third columns show mixed gels where the globular protein was gelled in the presence of 0.02 or 0.06 g_protein / gwater gelatin or gelatin hydrolysate. Gelatin was labelled covalently using FITC, while in addition the total protein was labelled non-covalently using Pvhodamine B. Throughout the experiments, it was observed that Pvhodamine B had a significantly lower affinity to gelatin as to globular proteins. This is based on the absence of hydrophobic patches in gelatin needed for Rhodamine to bind as discussed earlier and elsewhere for other systems. In the analysis of mixed gels, this difference can be used to differentiate between the two proteins. The Pvhodamine B signal will therefore be used to identify the location of WPI or SPI and the signal from FITC to identify the location of gelatin. For gelatin hydrolysates FITC labelling was not used because this would increase their molar mass by more than 10%. For these samples only the Rhodamine B signal (WPI) is shown.

Upon the addition of gelatin type A the gel structure of the globular proteins (SPI and WPI shown in red in Fig. 6) changes and a more pronounced phase separation is observed. With increasing gelatin type A concentration the globular proteins become more and more concentrated in their phase and gelatin type A is mainly found in the globular protein depleted phase (the globular protein gel pores). Analysis of images at other magnifications (results not shown) showed that at all protein concentrations and protein mixing ratios both phases were continuous throughout the entire sample volume. A similar result was found for the WPI / gelatin type B mixed gels. Comparison of the microstructural data to rheological results shows that the increase in gel coarseness is accompanied by a decrease in G' of the mixed gels. This effect, however, seems independent of the absolute length scale (size) of the phase separated domains as WPI / gelatin type A and SPI / gelatin type A have similar rheological behavior but largely differently size phase separated domains. Fig. 6 also contains CLSM images of WPI gels prepared in the presence of hydrolyzed gelatin type A. Here the microstructure was similar to that of the pure WPI gel and no enhanced phase separation could be observed. The presence of gelatin hydrolysate had no significant effect on the microstructure of the WPI gels which is in line with the observation that the storage modulus of the WPI gel where not decreased by the addition of hydrolysates.

For mixed systems the signals for WPI (Pvhodamine B) and gelatin (FITC) were acquired separately. For visualization purposes this allows to display them in different colors as shown in Fig. 6. For image analysis this additionally permits the analysis of the spatial arrangement of each protein, or better each channel, separately. As an example, Fig. 7 shows an overlay and CLSM images of the two channels separately for a sample containing 0.06 g_protein / gwater WPI and 0.02 g_protein / gwater gelatin type B. The images of the separate channels show that areas high in WPI concentration are low concentrated in gelatin and the other way round. The two images are in fact negative images of each other which is a direct indication of segregative phase separation between the two proteins. For image analysis this means that one image can be transferred into the other by inversion. Inversion (being a linear modifications) of an image before autocorrelation has no effect on the results from autocorrelation. Thus for phase separated systems as those shown in Fig. 7 no differences between the results from image analysis for the two channels are expected.

Fig. 8A shows the density correlation function for single and mixed WPI gels at constant WPI concentration. As expected the density correlation functions from the two channels overlay. The shift of the correlation functions for gels with added gelatins to higher length scales indicates an increase in the correlation lengths of the systems which is more pronounced for gelatin type B compared to type A mixed gels.

Fig. 8B shows the first minima (location thereof in autocorrelation function indicated with an arrow in Fig. 8A, representing the diameter of gel building blocks) of the density correlation function p(r) for a number of mixed gels. As expected also here no significant difference between the results from the two channels was obtained. The same result was found when comparing the other gel characteristics such as e.g. correlation lengths or distance between aggregates (results not shown). It is thus sufficient to analyze one of the channels in mixed systems to obtain the characteristic parameters such as e.g. gel building block size in mixed gels. In the following we will therefore only discuss the results obtained from non-covalently staining proteins with

Pvhodamine B.

Fig. 9 displays the average size of the gel building blocks for WPI gel prepared in the presence of gelatin type A (Fig. 9A1), gelatin type B (Fig. 9B1) or hydro lyzed gelatin type A (Fig. 9A2). For WPI / gelatin type A at 0.06 g_protein / gwater WPI a constant increase of the gel building block size over the whole range of analyzed gelatin concentrations was observed. The same result was found at 0.1 g_protein / gwater WPI and increasing gelatin type A concentration even though here the increase was less pronounced. The presence of hydrolyzed gelatin type A (Fig. 9A2) on the other hand did not have an influence on the size of the gel building blocks of WPI. At both analyzed WPI concentrations (0.06 and 0.1 g_protein / gwater) the building blocks were around 1.6 μιη independent of the concentration of hydrolyzed gelatin type A. The building block size of WPI gels formed in the presence of gelatin type B (Fig. 9B1) increased strongly at low gelatin type B concentrations (< 0.05 gprotein / gwater). At 0.06 gprotein / gwater WPI the gel building block size showed a maximum around 0.06 g_protein / gwater gelatin while at 0.1 gprotein / gwater WPI it increased over the whole range of analyzed gelatin type B concentrations. Parallel to the building block sizes also the correlation length, the distances between the gel domains and the sizes of the gel pores increased for the WPI / gelatin type A and the WPI / gelatin type B mixed gels (results not shown). For gel structures one typically summarizes these changes with the term "increasing gel coarseness" which will be used from this point onwards for the ease of reading.

In comparison, gelatin type B had a more pronounced effect on the gel coarseness than gelatin type A at comparable concentrations. This is the same effect as the change in G' of the WPI gels (see Fig. 3). We have earlier published a detailed characterization of these research materials in terms of their molecular sizes and their molecular interactions and also on their interaction with native and aggregated whey protein. At solvent conditions as those used here (pH 7, ionic strength 150 mM) gelatins and whey protein (as well as whey protein aggregates) were shown to interact mainly via hard body interactions. Also, the number averaged molar mass of gelatin type A and type B were comparable. The differences between gelatin type A and type B are most probably due to the presence of relatively more higher molar mass fractions in gelatin type B than gelatin type A (see also Fig. 14) which can lead to an enhanced phase separation between aggregates of WPI and gelatin during gelation. We will discuss the relationship between the (molecular) size of proteins, changes in size during gelation and the effect on phase separated microstructures further on.

To analyze the differences in WPI gel structures formed in the presence of gelatin type A or hydrolyzed gelatin type A we have performed SESANS measurements on these mixed WPI gels. The length scale over which SESANS is able to detect structural elements is in the range 20 nm to several μιη. The length scale of gelatin gels is typically one order of magnitude lower. For SESANS measurements, it is therefore expected that the presence of gelatin does not influence the scattering signal obtained from the WPI gel. To test this we analyzed a mixed protein gel containing 0.06 g_protein / gwater WPI and 0.1 g_protein / gwater gelatin type A at room temperature and at 40°C. Fig. 10 shows the normalized polarization versus spin echo length results at both temperatures. The formation of an ordered structure (gel at room temperature) from a liquid (at 40°C) where molecules are randomly distributed would significantly change the scattering results if the obtained structures were within the probed length scale of SESANS. No differences between the two measurements could be observed demonstrating that gelatin indeed forms a gel below the length scale probed by SESANS. The gelatin gel can therefore be assumed to be "invisible" in SESANS measurements. However, the presence of gelatin is expected to reduce the neutron scattering length density contrast between the WPI and the solvent phase (gelatin is located in the solvent). Accordingly we observed a decrease in the total scattering of mixed gels compared to the single WPI gels (higher normalized polarization at high spin echo length values in Fig. 10 compared to Fig. 5) upon gelatin addition.

In conclusion, the scattering signal from SESANS in mixed protein gels can be attributed to the structure of the WPI gel when keeping in mind the effect that gelatin has on the scattering length density contrast between WPI and the solvent phase.

Fig. 11 shows the normalized polarization for WPI gels prepared in the presence of gelatin type A (Fig. 11 A) or hydrolyzed gelatin type A (Fig. 11B). For WPI gels prepared in the presence of 0.04 g_protein / gwater gelatin type A the scattering intensity (normalized polarization value at high spin echo length) is below that expected for the pure WPI gel at this concentration. As discussed before, addition of gelatin is expected to lead to a lower total scattering (higher values for normalized polarization). However, the scattering intensity is also related to the size of the typical building blocks. CLSM results showed an increase in the size of the gel building blocks for WPI gels with added gelatin (see Fig. 9) which leads to a higher scattering intensity. For the sample containing 0.04 g_protein / gwater gelatin type A the increase in size seems to have a stronger effect on the scattering intensity than the decrease in neutron scattering contrast. At higher gelatin concentration (0.07 g_protein / gwater) the decreasing scattering contrast and increasing building block sizes balanced each other out and the measured scattering intensity was close to that of the pure WPI gel. For mixed WPI gels containing hydrolyzed gelatin type A (Fig. 11B) an increase in scattering intensity was observed for both tested concentrations of gelatin hydrolysate. This suggests that no large changes in the typical size of the building blocks occurred as already observed with CLSM. The increase in scattering intensity indicates that gelatin reduced the contrast between the WPI gel and the (gelatin containing) solvent as expected.

Besides this information about the changes in the size of the gel building blocks SESANS also allows an estimation of the density of the primary building blocks via the initial slope of the normalized polarization over spin echo length curve. For pure WPI gels this slope increased with increasing protein concentration (Fig. 5). The curves of the pure WPI gel prepared at 0.06 g_protein / gwater WPI and the mixed gels containing the same WPI concentration overlay in the range between 50 to approximately 1000 nm. As a comparison, the curve for a single WPI gel prepared at 0.1 gprotein / gwater WPI is shown in Fig. 11. The initial slope is significantly steeper for the pure gel prepared at 0.1 g_protein / gwater WPI compared to single and mixed gels prepared at 0.06 g_protein / gwater- This suggests that on a length scale below 1 μιη gel building blocks of mixed WPI gels have similar structures to single WPI gels prepared at the same nominal protein concentration (0.06 gprotein / gwater WPI). At higher length scales, however, mixed gels containing gelatin type A vary in their gel structure from pure WPI gels (discussed before in CLSM analysis). In the SESANS measurements these changes were observable in the obtained Hurst exponent (equation 2.2, related to the fractal dimension) for mixed systems. The Hurst exponent of pure WPI gels and WPI gels with added hydro lyzed gelatin type A were close to 0 while those for WPI gels with added gelatin type A were close to 0.4. While it is still questionable whether one should analyze WPI gels in terms of fractals (and especially phase separated mixed gels) these changes in the Hurst exponent indicate a different spatial arrangement of proteins in the mixed WPI gels compared to pure WPI gels which is in line with the results from CLSM.

So far we have been able to show a relationship between microstructural changes in globular protein gels and their rheological response. Results strongly suggested that this is based on segregative phase separation between gelatin and WPI during gelation (see e.g. Fig. 8). This lead to a confinement of both proteins in their separate phases, where each phase is depleted of the other protein. In the presence of hydro lyzed gelatin type A no changes in the gel microstructure and related rheological response were observed.

To discuss the effect of the presence of gelatin on the heat induced gelation of WPI we will first discuss the gelation in pure WPI systems. Heat induced gelation of WPI is a complex interplay between aggregation of whey proteins and a phase separation between these aggregated proteins and the water phase. The process is schematically shown in four steps in Fig. 12A. In the initial step of gelation a homogeneous protein solution (Fig 12A.1) is heated above the denaturation temperature of whey proteins leading to the formation of primary aggregates (Fig. 12A.2). These aggregates are only stable once they are above a certain size which is typically around 30 to 100 nm dependent on multiple factors such as solvent conditions and gelation kinetics. At the solvent conditions used in this study (pH 7, 150 mM NaCl), these aggregates phase separate from the solvent phase (called micro- phase separation) as shown in Fig. 12A.3. This phase separation is suggested to occur via spinodal decomposition which is arrested in an early stage by the inter-connection of the gel building blocks inside the WPI rich phase (Fig. 12A.4) leading to a bi- continuous structure of WPI rich (gel) and solvent rich (gel pores) phases. This microstructure (see also Fig. 6) is typical for globular protein gels at conditions where electrostatic repulsion is screened which was the case in this study.

The proposed effect how the presence of gelatin impacts on the WPI gel formation is shown in Fig. 12B. As for the single WPI gels the initial stage is a homogeneously mixed solution containing in this case WPI and gelatin (Fig. 12B.1). Upon heat induced denaturation of whey proteins these start to aggregate (Fig. 12B.2) while gelatin molecules stay soluble and do not participate in this aggregation. SESANS results suggested that the density of these primary aggregates in mixed gels is identical to that for pure WPI gels (see initial slope in Fig. 11). This suggests that the primary (and possibly also secondary as SESANS indicated changes only above 1 μιη) aggregates were formed while the system was still completely mixed and thus WPI at its nominal concentration (which in the case of SESANS measurements was 0.06 gprotein / gwater). If phase separation would have occurred before or during the formation of these initial gel building blocks the WPI concentration at which they form would be higher which would result in denser aggregates as shown for the pure WPI gels (see initial slope in Fig. 5B).

During further aggregation, the size of the intermediate gel building blocks (or whey protein aggregates) increases. Since the size of the gelatin molecules stays constant the formation of even larger WPI aggregates changes the size ratio between globular protein (aggregates) and gelatin. This size ratio (q) is an important parameter in whether mixed systems phase separate or stay mixed and can be expressed by (3.3): a_G

aGP where ao is the effective radius of the gelatin molecules and aop the effective radius of the globular proteins (or their aggregates during gelation). As discussed earlier gelatins and WPI or whey protein aggregates are well approximated as hard spheres with hard body interactions at neutral pH and ionic strength > 100 mM. The phase behavior of such binary hard sphere systems as a function of the size ratio q can be approximated using a virial approach. Fig. 13 shows the theoretical critical volume fraction as a function of q for a binary hard sphere mixture. The critical volume fraction in Fig. 13 represents the point in a phase diagram with the lowest total particle (or here protein) volume fraction where phase separation is expected (critical point in a phase diagram). Around q = 1 binary mixed systems show a high co-solubility which decreases with increasing and decreasing q - values. The q - value of the mixed WPI / gelatin type A and type B and WPI / hydro lyzed gelatin type A are added to Fig. 13. The arrow indicates the change in q during the gelation of globular proteins (formation of larger globular protein aggregates). The q - values were based on the effective molecular size of the globular proteins and gelatin as obtained by osmometry. For gelatin hydrolysate the size was estimated based on SEC results shown in Fig. 14. which also indicates the reverse size ratio between WPI and gelatins compared to WPI and gelatin hydrolysate.

The two mixtures where the presence of gelatin altered the globular protein (WPI) microstructure are characterized by q > 1 while q is « 1 for the WPI / hydro lyzed gelatin type A sample where the microstructure of the mixture was equal to that of the pure globular protein gel. Samples where the globular protein microstructure was altered crossed the point q = 1 and subsequently a region with decreasing co-solubility (higher possibility for phase separation) during globular protein gelation. It is proposed that this crossing of a region of high co-solubility and subsequent passing a region with reducing co-solubility promotes phase separation during globular protein gelation. Phase separation is proposed to occur at a late stage of the globular protein aggregation and lead to an increase in gel coarseness (increase in size of gel domains, distance between gel domains and gel pores) as schematically shown in Fig. 12B.4.

We have earlier arrived at the same conclusion when investigating the gelatin gelation in the presence of globular proteins which proposes this effect to be general for these mixed gel systems. Yet, when comparing the effect of phase separation on the rheological response of mixed gelatin gels (Interactions in Protein Mixtures. Part II: A virial Approach to Phase Behaviour. Food Hydrocolloids, 52, 991- 1002, 2015) and the here presented globular protein gels they initially seem contradictory. Phase separation due to the presence of (non-gelling) secondary proteins led to an increase in G' for mixed gelatin gels while in globular protein gels the storage modulus decreases upon the occurrence of (increased) phase separation. These differences are most likely based on their different gelling mechanisms. Gelatin gels isotropically and when present in a mixed gel, phase separation will mainly impact on the water distribution and therefore concentration of gelatin and to a lower degree on the gel structure. Globular protein gels were prepared at conditions where electrostatic interactions are screened and their gelation mechanism included micro-phase separation. The presence of gelatin in this systems mainly altered the structure of the already phase separated, bi-continuous globular protein gel and did not lead to a gelation of WPI at simply increased protein concentration.

Rheological methods together with microstructural analysis using microscopy and SESANS was shown to selectively probe the structure of WPI gels (or more general globular protein gels) in the presence of gelatin. This allows the stepwise analysis of mixed globular protein / gelatin gels. In this publication we focused on the first step of the sequential gelling step of globular protein / gelatin gels which will serve as an important basis for researches aiming to understand the properties of the final gels (where both proteins are gelled). In the case of mixed WPI gels, microstructural changes due to the presence of gelatin during the gelation of globular proteins were directly related to a reduction in storage modulus of the WPI gels. The mechanism leading to these changes was suggested to be an enhanced segregative phase separation at a late stage of WPI gelation between WPI gel building blocks and the solvent phase containing gelatin which only occurred when gelatin molecules had an effective diameter above that of the globular proteins.

The data presented in Figure 16 demonstrate that hydro lyzed gelatin/collagen is, based on rheological analysis (Young's modulus), not significantly affecting the globular protein gel formation compared to non-hydro lyzed gelatin. This effect is particular visible between the salt concentrations 0.6 and 1.2 wt.%. This effect is visible in a wide temperature region, at least from ambient (20°C) to 40°C.

The term "substantially" herein, such as in "substantially consists", will be understood by the person skilled in the art. The term "substantially" may also include embodiments with "entirely", "completely", "all", etc. Hence, in embodiments the adjective substantially may also be removed. Where applicable, the term "substantially" may also relate to 90% or higher, such as 95% or higher, especially 99% or higher, even more especially 99.5% or higher, including 100%. The term "comprise" includes also embodiments wherein the term "comprises" means "consists of. The term "and/or" especially relates to one or more of the items mentioned before and after "and/or". For instance, a phrase "item 1 and/or item 2" and similar phrases may relate to one or more of item 1 and item 2. The term "comprising" may in an embodiment refer to "consisting of but may in another embodiment also refer to "containing at least the defined species and optionally one or more other species".

Furthermore, the terms first, second, third and the like in the description and in the claims, are used for distinguishing between similar elements and not necessarily for describing a sequential or chronological order. It is to be understood that the terms so used are interchangeable under appropriate circumstances and that the embodiments of the invention described herein are capable of operation in other sequences than described or illustrated herein.

It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and that those skilled in the art will be able to design many alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. Use of the verb "to comprise" and its conjugations does not exclude the presence of elements or steps other than those stated in a claim. The article "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

The various aspects discussed in this patent can be combined in order to provide additional advantages. Furthermore, some of the features can form the basis for one or more divisional applications.

Claims

CLAIMS:

1. A method for the production of a gel-based product, the method comprising (i) providing a starting composition comprising a globular protein and water, and (ii) converting the globular protein and water into a gel by heating the starting composition, to provide said gel-based product, wherein the starting composition comprises the globular protein in an amount of 5-20 wt.% relative to the total weight of the starting composition, wherein the starting composition further comprises a gelatin hydrolysate in an amount of 5-40 wt.% relative to the total weight of the starting composition, and wherein the starting compositions comprises 50-90 wt.% water, wherein at least 90 wt.% of the gelatin hydrolysate has a molar mass of < 50 kDa. 2. The method according to claim 1, wherein the starting composition further comprises 0.

2 - 2.5 wt.% salt (e.g. NaCl), wherein the starting composition has a pH selected from the range of 3-8, wherein at least 90 wt.% of the gelatin hydrolysate has a molar mass of < 20 kDa, and wherein the globular protein and water are converted into the gel by heating the starting composition to a temperature selected from the range of 75-95°C.

3. The method according to any one of the preceding claims, wherein the gelatin hydrolysate has a mass average molar mass of < 10 kDa.

4. The method according to any one of the preceding claims, wherein the gelatin hydrolysate comprises in the range of 5-95 wt.% proteins, and wherein relative to the total weight of the starting composition, the weight of hydroxyproline is in the range of 0.35-5.1 wt.%. 5. The method according to any one of the preceding claims, wherein the globular protein has an iso-electric point selected from the range of 3-5.

5, especially wherein the starting composition comprises one or more of whey protein isolate, pea protein isolate, albumin protein isolate, immunoglobulin protein isolate, and soy protein isolate.

6. The method according to any one of the preceding claims, wherein the method further includes the presence of one or more other components, to provide a human food product, a pet food product, a feed product, or a coating or an encapsulate for a medicament.

7. The method according to any one of the preceding claims 1-6, wherein the method comprises a first stage comprising converting the globular protein and water into the gel, in the presence of the gelatin hydrolysate, and a second stage wherein the gel and one or more other components are combined to provide said gel- based product.

8. The method according to any one of the preceding claims 1-6, wherein the method comprises converting the globular protein and water into the gel in the presence of the gelatin hydrolysate and one or more other components, to provide said gel-based product.

9. A gel-based product obtainable by the method according to any one of the preceding claims.

10. The gel-based product according to claim 9, comprising gelatin hydrolysate and globular proteins in a weight ratio of 1 :8 - 8: 1, and wherein the gel- based product comprises salt in an amount of 0.2-2.5 wt.% relative to the total weight of the gel-based product.

11. The gel-based product according to any one of claims 9-10, wherein the gel-based product comprises a solid product or a semi-solid product, and wherein relative to the total weight of the gel-based product, the weight of hydroxyproline is in the range of 0.35-5.1 wt.%

12. The gel-based product according to any one of claims 9-11, wherein the gel-based product comprises a human food product, a pet food product, a feed product, or a coating or an encapsulate for a medicament.

13. The gel-based product according to any one of claims 9-12, wherein the gel-based product comprises a dairy product, a probiotic product, a protein bar, a sausage, a protein shake, or a protein drink.