CHAPTER 1
PROBING FOOD STRUCTURE
V. J. Morris
Institute of Food Research, Norwich Research Park, Colney, Norwich NR4 7UA, UK
1 INTRODUCTION
When I joined the Institute of Food Research in 1979 I was given the brief to develop a molecular description of the functionality of food materials. At that time I knew very little about food research and food hydrocolloids. The first Wrexham meeting provided me with an excellent introduction to the subject and a chance to meet the leading researchers in this area. Since then the meetings have allowed me to keep up-to-date with developments in this area. Thus it is an honour for me to review some of the research that I have been involved in over the last 26 years which has, I hope, helped to improve the understanding of the behaviour of food materials. I would like to illustrate the power and also the limitations of physical methods for studying mixtures of hydrocolloids, and to indicate why there was a need to develop methods of molecular microscopy of food materials. Finally, I would like to show how the use of such a microscopic technique, atomic force microscopy, has shed new light on the structure of gels, complex interactions between biopolymers at interfaces, and is providing new insights into the structure and behaviour of starch. The ability to visualise individual molecules can reveal new structural information and simple investigations can open up new areas of research. Such studies will be illustrated through images of beet pectin which account for its behaviour as an emulsifier and simple studies on whey proteins that seem to indicate unexpected interactions between milk proteins.
2 METHODS AND RESULTS
Food materials are complex mixtures which are heterogeneous at the molecular level. Given the wide range of biopolymers used by the food industry it would seem at first that there would be almost an infinite numbers of different types of mixtures that could be produced. However, in some cases it is possible to simplify the description of their behaviour by dividing them into classes of materials which show similar structures or function. This type of approach can best be illustrated through studies on the gelation of polysaccharide mixtures.
2.1 Binary Polysaccharide Gels
Mixtures of two polysaccharides can form four broad classes of gels: swollen networks (Figure 1a), interpenetrating networks (Figure 1b), phase-separated gels (Figure 1c) and coupled networks (Figure 1d). The phase-separated networks are the commonest type of structure and these have been the most useful for formulating new types of food structures. Swollen networks are most likely formed from mixtures of neutral and charged polysaccharides where the entropy of mixing term for the mobile counterions inhibits phase separation. These structures differ in whether one or both of the networks gels. The most intriguing types of gel are the coupled networks which gel under conditions for which the individual components alone will not gel. These types of gels are formed between various types of mixtures, but the most interesting are those formed between xanthan, or xanthan-like polysaccharides, and certain galactomannans or glucomannans. For these mixtures it was possible to show through x-ray fibre diffraction studies" that a new structure is formed between the two different polysaccharides, in order to link the chains together into a network. For the gels formed with the glucomannan Konjac mannan the mixtures show 6-fold helical structures with the same pitch as the 5-fold helical structure of xanthan or acetan. Given that denaturing the xanthan helix favours gelation, and the stereochemical compatibility of the backbone structures of the glucomannans and the backbones of acetan or xanthan, it is not unreasonable to suggest that the linkage is due to a mixed double helix containing both polysaccharide chains. Acceptable left-handed 6-fold mixed helical structures have been published for acetan-Konjac mannan mixtures. This result is an example of how a physical method can yield very significant information on the structure of a complex mixture allowing the functional behaviour to be explained. However, to obtain this information it was necessary to dry the gel to a hydrated film and to align the ordered structures within the network. The data tells about the junction zones within the gel but it tells us little about the long-range structure of the gel.
To investigate the detailed structure of gel networks it is necessary to use microscopic methods capable of achieving molecular resolution. Electron microscopy has the required resolution but the complex sample preparation methods required to remove and replace water in order to image the gel structure, plus the relatively poor contrast in the images, make this technique difficult to apply for routine studies. Furthermore, the images obtained are often on sections making it difficult to visualise the 3-D network within the gel. The development of probe microscopes offered an alternative method with the prospect of molecular information under more realistic conditions. Over the last 18 years the use of atomic force microscopy (AFM) has been developed for studying biological systems and, in particular, food materials. The use of AFM has allowed new types of food structure to be investigated for the first time and has led to new solutions to previously intractable problems in food science. Examples of such results include new models for the gel structure of polysaccharides, new models for the competitive displacement of proteins from interfaces by surfactants, new insights into the structure and functionality of starch, the discovery of previously unknown branching structures of polysaccharides and new molecular mechanisms of action for glucoamylose.
2.2 Gellan Gels
The conventional models of polysaccharide gels shown in most textbooks picture the gels as ordered junction zones connected by essentially random-coil chains (Figure 2). This is essentially a modified rubber-like structure for the gel with the point cross-links replaced by extended junction zones. Atomic force microscopy provides a method for visualising the long-range structure within gels by studying the association of polysaccharides as gel precursors, films and bulk gels. Gellan gum provides a good model system for probing gelation. Gellan forms thermoreversible gels on cooling and heating and the aggregation of the polysaccharide chains occurs through helix formation and then an additional step involving cation binding. The cation binding stage can be eliminated by forming the tetra-methyl ammonium (TMA) salt form. This forms weak thermoreversible gels which show no thermal hysteresis. By depositing dilute solutions of TMA gellan onto freshly cleaved mica it is possible to induce aggregation on drying in air. The gel precursors formed are elongated branched structures or fibrils. These fibrils are longer than individual molecules and their height suggests that the gellan is in the helical form. Mismatching of chain ends during nucleation and growth of the double helical structure on cooling could account for the formation of elongated and branched fibrils solely through helix formation. Deposition, in the presence of gel-promoting ions such as potassium, leads to similar elongated branched aggregates, but with variable width and height, consistent with further aggregation of the fibrils into thicker fibres. At higher polysaccharide concentrations thin aqueous films are formed containing a continuous branched fibrous network. These aqueous films can be imaged because the gel network is compressed against the mica substrate during imaging. Bulk aqueous gels are more difficult to image as the probe used to scan the gel compresses the gel blurring the image. A 1.2% acid-set gel is stiff enough to resist deformation and the fibrous network can be seen directly within the gel (Figure 3). The gels show hysteresis in their setting and melting behaviour. This is consistent with the side-by-side aggregation of the helices in the fibrils being stabilised by binding of cations within the fibrous network. The long-range gel structure is very different from the schematic model shown in Figure 2.
The types of fibrous structures seen in gellan are also reported for other polysaccharide gels such as pectin which gel by different mechanisms. It is possible that this sort of fibrous assembly is generic and only the ways in which chains are stuck together within the fibres differs from polysaccharide to polysaccharide. A schematic picture of such an assembly process is shown in figure 4. Thus the ability to visualise polysaccharide association in aqueous films and gels has led to new insights into the mechanisms of polysaccharide gelation.
2.3 Interfacial networks
The oil-water and air-water interfaces in food emulsions and foams generally contain both proteins and surfactants (or lipids). Proteins or surfactants can alone stabilise interfaces but they do so by very different and incompatible mechanisms. The proteins are considered to partially unfold at the interface and to associate into networks. Distortion of the interface is opposed by the elastic protein film. Surfactants or lipids are mobile at the interface and distortion of the interface will lead to concentration gradients and flow of the molecules to restore the status quo. The presence of both types of molecules at the interface leads to instability because each class of molecule (proteins or surfactants) interferes with the stabilising action of the other class of molecule. Effectively the two classes of molecules compete for control of the interface. If sufficient surfactant is present then the surfactant will eventually displace the protein from the interface. The question is how this occurs? Although surfactants are more surface-active than proteins because the proteins are joined together it is difficult to displace individual proteins. The use of AFM to image interfacial structures explains how this competitive displacement occurs.
Analysis of the images (Figure 5) reveals that the surfactant initially appears as small holes in the protein network and, with increasing surfactant concentration, the size of the surfactant domains grows. As the area occupied by protein decreases the height of the protein network increases due to folding and buckling of the protein layer. The volume of protein remains constant until the protein network breaks freeing individual proteins or protein aggregates that can be displaced from the interface. The process is generic because all proteins studied to date form networks at air-water or oil-water interfaces and these networks need to be broken to release and displace protein. The process of folding and failure has been termed orogenic displacement. A generic mechanism suggests generic solutions to a wide range of problems with the stability of food foams and emulsions in the baking, brewing and dairy industries. It was the ability to visualise the heterogeneous processes occurring at the interfaces by AFM that was crucial to the discovery of this unexpected mechanism of displacement. The mechanisms observed for the displacement of single proteins can be extended to include characterisation of the displacement of protein mixtures as models for protein isolates. These studies raise interesting questions into how mixtures of proteins behave and how they form networks at interfaces or in the bulk.
2.4 Fibril formation of whey proteins
Whey protein isolates are mixtures of proteins containing mainly (3-lactoglobulin with α-lactalbumin and serum albumin. Under specific conditions (2% protein, pH 2, 0.1M NaCl) heat treatment of whey proteins (80 C,180 min.) results in fibril formation. Under similar conditions α-lactalbumin does not form such fibrils but β-lactoglobulin does form linear aggregates. A logical assumption would be that it is the β-lactoglobulin component of whey which is forming the fibrils. Closer inspection of the results showed that under the same conditions β-lactoglobulin formed longer fibrils than whey. This could simply be due to a dilution effect in the mixture (whey) or evidence that α-lactalbumin inhibits fibril formation by β-lactoglobulin. To test which of these two possibilities were true, increasing amounts of α-lactalbumin were added to whey and fibril formation observed (Figure 6). When the whey concentration was constant addition of α-lactalbumin led to a progressive decrease in fibril length, suggesting that α-lactalbumin does have an inhibitory effect. These relatively simple preliminary results suggest that AFM is ideal for probing such unexpected interactions between proteins.
2.5 Beet pectin
A few polysaccharides are good emulsifiers. Beet pectin is known to contain a small amount (~10%) protein which is difficult to remove and is generally considered to be covalently linked to the polysaccharide chain. It has been proposed that it is the protein component which is important in determining the emulsifying action of beet pectin. AFM images of purified beet pectin provide direct evidence for the presence of bound protein and reveal the location of the protein. The images (Figure 7) show a population of molecules with a small proportion being pure polysaccharide chains. The majority of the sample consists of a polysaccharide chain with a protein attached to the end. These 'tadpole-like' complexes are easy to visualise when the polysaccharide chain is extended but can be difficult to spot when the polysaccharide chain is wound around the protein (Figure 7). There is also a very small population of aggregated tadpoles (not shown).
Given the previous studies on proteins at interfaces it is not difficult to envisage the protein component assembling into a network at the oil-water interface with the hydrophilic pectin chains extending out into the bulk aqueous phase. The potential for cross-linking pectin chains has potential for stabilising the protein network against displacement. The outer 'hairy' pectin chains on the oil droplets will inhibit coalescence of droplets. There is also potential for cross-linking pectin chains on neighbouring droplets to form aggregates or gels. The visual evidence from the microscope has provided proof of a suggested model of emulsification. The methodology developed to probe protein-surfactant interactions could be applied to test the suggested structures formed at oil-water interfaces.
2.6 Starch architecture
Starch is the major carbohydrate consumed by mankind. It is produced in plants as a storage carbohydrate. Starch granules are semi-crystalline spheroidal structures. The starch polysaccharides amylose and amylopectin are arranged within the periodic growth ring structure of the granule. The clusters of amylopectin branches, interspersed by amorphous regions, are contained within ellipsoidal structures called blocklets. The blocklets are considered to be embedded in an amorphous matrix composed mainly of the essentially linear polysaccharide amylose. AFM can be used to image the internal structure of the granules provided the granules are encased in a suitable non-penetrating resin. The internal structure of the granule can be revealed by sectioning or by facing off a block of encased granules. The contrast in the images has been shown to be due to the selective wetting and swelling of certain regions of the exposed faces of the granule, believed to be regions occupied by the amylose molecules. The ability to map the topography and hardness of the exposed face of the granule has allowed the blocklet structure of the starch granule to be visualised and identified. In a collaborative study with researchers at the John Innes Centre we have used a set of isogenic pea mutants to examine how inhibition of selected biosynthetic enzymes has led to changes in the architecture of the starch granules. In particular, for high amylose starches we have observed the presence of an unexpected fibrous network permeating throughout the granule (Figure 8).
In hardness (force modulation) maps of the exposed surface of the wild type pea starch granule the harder spheroidal blocklets are visible within the growth ring structure of the granule. The hardness images of the high amylose r mutant is dominated by the new hard fine network which is also visible in the double mutant rrb (Figure 8). It has been suggested that these networks account for the fragile nature of the high-amylose granules and the presence of fissures and cracks. Physicochemical studies suggest that the crystalline regions of these networks contain amylosic chains of higher molecular weight than the normal branches of the amylopectin that contribute to the crystal structure of the granule. It has been proposed that these crystals account for the reduced swelling of the granules, broadened gelatinisation of the granules and the need to heat granules to 120°C to completely gelatinise the starch. Should such structures also occur in starch from other plant species then their identification and modification offers a route to the design and modification of starches to optimise the functional and nutritional properties of the starch.
