Rheology, Structure and Texture Perception in Food Protein Gels

By E. Allen Foegeding

DEPARTMENT OF FOOD SCIENCE, NORTH CAROLINA STATE UNIVERSITY, RALEIGH, NC 27695, USA

1 Introduction

One could propose that mankind's original encounter with a protein gel occurred when the first human accidentally dropped an egg into a fire or hot water and discovered the hard cooked egg. If we accept this humble origin of a culinary art, then the presence of current research dealing with the gelation of ovalbumin demonstrates that making protein gels is a simple process while understanding the molecular mechanisms continues to present scientific challenges.

The seminal review by Ferry defines some general properties of protein gels as "... systems in which small proportions of solid are dispersed in relatively large proportions of liquid by the property of mechanical rigidity ... the characteristic property common to all gels." The rheological characteristics of gels were further refined to include viscoelastic properties and to distinguish between 'solid' and 'solid-like' gels. The key elements of a solid-like gel are a storage modulus, G'(ω), which has a plateau extending to times at least of the order of seconds, and a loss modulus, G"(ω), that is much smaller than the storage modulus in the plateau region. For gels formed by denatured proteins, Ferry went on to propose: (1) they are the result of forming a three-dimensional network of solute; (2) they arise from interactions that act along the entire solute molecule, and therefore require a proper balance of attractive and repulsive forces regulated by pH and ionic strength; and (3) their macroscopic appearance and water-holding properties reflect microscopic networks that have extremes of coarse or fine structures. The following mechanism for gelation of denatured proteins was suggested: native protein (corpuscular) [right arrow] denatured protein (long chains) [right arrow] association network.

It is amazing that the description for gelation of denatured proteins proposed by Ferry in 1948, based on primarily 4 research articles, is generally valid today. While it has since been proven that proteins do not have to unfold into long chains to form gel networks, there have been, and continue to be, numerous investigations into how factors (pH, salts) regulate the association of denatured molecules, the resulting microstructure, and the links between microstructure and macroscopic properties (texture and water-holding).

Research on protein gelation can be generally categorized into two areas, the first being the nature of the gelation process. This involves the various factors that transform proteins from non-aggregating to aggregating structures; the definition and determination of the gel point; the factors responsible for development of gel elasticity; and theories capable of explaining all aspects of the process. The second area concerns the physical and chemical properties of gels. To that end, the microstructure of gels along with the rheological and fracture properties have received the greatest attention, but appearance, intra-gel diffusion, and water-holding attributes have also been investigated. While the nature of the gelation process can be considered a general scientific curiosity, the appearance, water-holding and textural properties of gels have direct applications to food quality. Some of the foods that rely on the gelation of denatured proteins to produce desirable characteristics are cooked eggs, processed meats, and some cheeses. In theory, if one understands the physical and chemical basis for the quality characteristics of a food protein gel, such as a cooked egg, then the potential exists to form the same structure with another protein. Indeed, this was pointed out8 in 1930 as a motivation for developing whey protein ingredients to replace egg proteins in foods. The many soy protein-based imitation meat products on the market today serve as examples of this application. It is also possible to augment a product by adding proteins as gelling agents. An example is the addition of milk proteins to water-added hams as a means of increasing water-holding and rheological properties.

The fundamental science of globular protein gelation took a step forward with the publication of three articles in the early 1980s. These articles were the first to focus on molecular and microstructural changes associated with heat-induced gelation of a variety of globular proteins. The protein gelation process was reviewed by Clark and Lee-Tuffnell5 and by Clark. In these reviews, the theoretical foundation of the gelation process is discussed, along with comprehensive coverage of protein aggregation, changes in protein secondary structure, and microstructure. These areas will be addressed in the present article only from the standpoint of how they help to explain differences in gel structure type and textural properties. This review will focus on gelation from the perspective of producing food structure and textural properties.

2 Gel Structure Type

In a recent article entitled 'From food structure to texture', Wilkinson and coworkers state that advances in the understanding of texture perception will depend on advances in sensory analysis of texture, physiological perception, and food structure. Included under the heading of food structure are "rheological parameters, microstructure and other relevant food characteristics and their relation to sensory perception." This begs the question: are there specific structures in food protein gels? If so, what are the distinguishing characteristics that separate one structure from another? In many ways this issue has already been addressed. Heat-induced globular protein gels can be transparent, turbid or opaque depending on the gelation conditions (see Figure 1). Gels that are transparent or translucent are generally classified as fine-stranded (also called fibrillar, 'string of beads', or 'true' gels). They are formed under conditions of pH greater or less than the pI, and at low ionic strength. The name "fine-stranded" derives from transmission electron micrographs showing microstructures consisting of strands of various lengths, but with diameters corresponding to the length of one to several molecules. Fine-stranded gels formed at low and high pH have different textural properties. The gels formed at low pH are weak and brittle, and those formed at high pH are strong and elastic. Gels that are opaque are called particulate (also called aggregated, random aggregated, or coagulated). Particulate gels are formed under conditions where there is minimal charge repulsion, such as at a pH close to pI or at high ionic strength. The microstructure of particulate gels is composed of spherical particles with diameters in the micrometre range that are associated into a gel network. The large particles and low water-holding properties of particulate gels have led to the proposal that they are formed by phase separation. Needless to say, there are conditions that lie between the two limiting cases that produce gels with structures that can be considered either (i) a combination of stranded and particulate or (ii) distinct and intermediate between these extremes.

3 Gelation of Fine-Stranded and Particulate Gels

The networks formed in fine-stranded gels at low pH (generally ≤ 3.5) are different from those formed at high pH (generally ≥ 6.5). β-Lactoglobulin gel networks formed at pH 7.0 are composed of strands that are slightly thicker, longer and more curled than those formed at pH 3.5. Heating a β-lactoglobulin solution at pH 2 at a concentration below the critical gelation concentration (< C0) produces fibrils rather than a gel network. Fibrils of β-lactoglobulin of length of 1–7 µm are observed. The fibril length is independent of ionic strength in the range 0.01–0.05 M. Static and dynamic light scattering measurements support the formation of β-lactoglobulin fibrils at pH 2.0. Recent atomic force microscopy images confirm the formation of β-lactoglobulin strands composed of dimers or partially unfolded monomers. For β-lactoglobulin, an increase in pH (to 2.5 or 3.0) or ionic strength (>50–90 mM) causes the fibrils to shorten. The formation of fibrils and their association into a gel network seems to be a general property of gels formed at low pH and low ionic strength. This type of gelation can be explained by an adjusted random contact model that takes into account the charge and semi-flexibility of the fibrils. This model has been shown to apply to fibril network gels formed by bovine serum albumin, β-lactoglobulin and ovalbumin.

The type of β-lactoglobulin aggregates formed at pH 7.0 varies among investigations. Kavanagh et al. reported the formation of short (<0.05 µm) linear aggregates based on transmission electron microscopy. Using size-exclusion chromatography and light scattering, Le Bon et al. found that, when β-lactoglobulin was heated at 60 °C at pH 7.0 in a 0.1 M salt solution, self-similar aggregates are formed composed of 85 protein molecules and having a radius of 15 nm. A similarly sized aggregate (18 [+ or -] 4 nm) was observed by atomic force microscopy for β-lactoglobulin solutions that were heated at 80 °C in pH 7.0 solutions containing 0.1 M salt. Although differences in aggregates have been observed among studies conducted at neutral pH, possibly due to variations in heating conditions and methods of analysis, the types of aggregates formed at high pH appear to be consistently different from those formed at low pH. This suggests that, at some level, the fundamental structural element is different for low pH and high pH fine-stranded gels.

Particulate gels can be formed in solutions where the pH is close to the pI of the protein (or proteins), or where the pH is not close to the pI but there is a relatively high ionic strength (usually > 0.1 M). Once the pH and ionic strength are sufficient to form a particulate gel network, increasing ionic strength causes a decrease in rheological parameters (G' or G*) and waterholding properties, accompanied by changes in the microstructure described as a coarsening of the network or an increase in voids.

Molecular changes associated with fine-stranded and particulate gels have been investigated using several spectroscopic techniques. Using Fourier transform infrared spectroscopy or Raman scattering spectroscopy, a more extensive unfolding was seen when β-lactoglobulin formed fine-stranded gels as compared to particulate gels. While there were apparent differences between fine-stranded and particulate gels, no major differences were seen when these types of gels were formed under different gelation conditions. Therefore, differences in rheological properties associated with fine-stranded gels formed at low and high pH, as well as particulate gels formed close to the pI or at high ionic strength, cannot be explained based on protein structural transitions alone.

A series of studies has established how the incorporation of protein into the gel network at and after the gel point determines whey protein gel micro-structure, permeability and rheology. These experiments were conducted under conditions where gelation was relatively slow (20-24 h at 68.5 °C), and mainly particulate gels were formed. The amount of aggregated protein forming the incipient gel network determines the gel network type (microstructure) and the permeability (a function of microstructure and other properties). Gel permeability decreases only slightly upon further heating after the gel point, suggesting that the network type is 'fixed' at the gel point. In contrast, the elasticity (G'') is close to zero at the gel point, and it increases with the total amount of protein incorporated into the established gel network.

4 Fracture Properties

Fracture behaviour can be investigated by compressing, extending or twisting a sample to the point of fracture. While compression is most often used, due to the simplicity of sample preparation, it is limited to conditions where there is a uniform expansion of the sample when compressed (i.e., where friction between sample and testing apparatus is very low), and for fracture strains occurring at compression levels above ~80% of the original height h (i.e., Δh/h [less than or equal to] 0.8). Materials that fracture at high strains must be deformed in tension or torsion. Torsional deformation is pure shear, and there are minimal shape changing forces. So samples that exude fluids in compression and tension will not do so in torsion. Also, since the tensile and shear forces are of equal magnitude, the sample fracture pattern in torsion indicates the weakest mode. The angle of the fracture plane relative to the long axis is 45° for tension fracture and 90° for shear fracture.