Details of Water Orientation Adjacent to Organic Molecules

force for protein folding, causing many hydrophobic residues to assume positions in the protein interior. Despite hydrophobic interactions, it is estimated that nonpolar groups in globular proteins typically occupy about 40–50% of the surface area. Hydrophobic interactions also are regarded as being of primary importance in maintaining the tertiary structure of most proteins [19,85,123]. It is therefore of considerable importance that a reduction in temperature causes hydrophobic interactions to become weaker and hydrogen bounds to become stronger. 2.7.7 Details of Water Orientation Adjacent to Organic Molecules Although determination of the arrangement of water molecules near organic molecules is experimentally difficult, this is an active field of research and useful data have been obtained. The hydrated pyranose sugar ring is shown in Figure 13 and a computersimulated cross section of hydrated myoglobin is shown in Figure 14. Assuming a separation distance of 2.8 Å between hydration sites and full occupancy of these sites, about 360 HOH molecules would be in the primary hydration shell of myoglobin [71]. Pag e 38 FIGURE 13 Association of a-D-g lucose with tetrahedrally arrang ed water molecules. Heavy solid line represents the side view of the pyranose ring . Oxyg ens and hydrog ens of water are represented by open and filled circles, respectively. Covalent and hydrog en bonds are represented by solid and broken lines, respectively. The hydroxymethyl protons [H(6)] are not shown. (From Ref. 122.) 2.7.8 Hydration Sequence of a Protein It is instructive to consider water absorption by a dry food component and the location and properties of water at each stage of the process. A protein is chosen for this exercise because proteins are of major importance in foods, because they contain all of the major types of functional groups that are of interest during hydration, and because good data are available. Shown in Table 4 are properties of globular proteins (based primarily on lysozyme) and the associated water at various stages of hydration. The corresponding sorption isotherm is shown in Figure 19. The table is self-explanatory except for a few points. Hydration “zones” are referred to in both the table and the figure. These zones are useful aids to discussion but they are unlikely to actually exist (a continuum of water properties is much more likely). A sample having a water content corresponding to the junction of Zones I and II is said to have a BET monolayer water content (the term “BET” comes from the names of the originators of the concept: Brunauer, Emmett, and Teller [8]). In this instance the BET monolayer water content is about 0.07 g HOH/g dry protein, and this corresponds to a p/p0 value of about 0.2 (p is the partial pressure of water in the food and p0 is the vapor pressure of pure water at the same temperature; p/p0 is more commonly referred to as aw). The BET monolayer value is of special importance because it often provides a good first estimate of the largest moisture content Pag e 39 FIGURE 14 Cross-section of a hydrated myog lobin molecule as determined by molecular dynamics simulation. Mesh cag es depict hig h probability sites of 1st-layer water molecules, stick fig ure represents the protein time-averag ed structure. (From Ref. 72.) a dry product can contain and still exhibit maximum stability. Although the BET value is commonly referred to as a monolayer, this is a faulty concept. For example, in starch the BET value corresponds to about one water molecule per anhydroglucose unit [126]. Also note that in Table 4 the term “true monolayer” is used. This term has a meaning quite different from BET monolayer. True monolayer refers to the water content at the junction of zones IIB and III (in this example, a water content of about 0.38 g HOH/g dry protein and a p/p0 of about 0.85). This value corresponds to about 300 mol HOH per mol lysozyme and a moisture content of 27.5 wt%, with one HOH occupying, on average, 20 Å2 of protein surface area [103]. This water content is significant because it represents the minimum water content Pag e 40 TABLE 4 W ater/Protein Properties at Various Stag es of Hydration a Bulk-phase water Properties Constitutional waterb Hydration shell (<3 Å from surface) Freec Entrappedd General description for lysozyme Constitutional water is assumed to be present in the dry protein at the onset of the hydration process. W ater is first absorbed at sites of ionized, carboxylic and amino side chains, with about 40 mol water/mol lysozyme associating in this manner. Further absorption of water results in g radual hydration of less attractive sites, mainly amide carbonyl g roups of the protein backbone. Attainment of true monolayer hydration of the protein is achieved at 0.38 g H 2O/g dry protein, by water associating with sites that are still less attractive. At this point, there is, on averag e, 1 HOH/20 Å 2 of protein surface Fully hydrated Fully hydrated Approximate water content: g H2O/g dry protein (h) mol H2O/mol dry protein wt% based on lysozyme <0.01 h <8 1% 0.01–0.07 h 8–56 1–6.5% 0.07–0.25 h 56–200 6.5–20% 0.25–0.58 h 200–304 20–27.5% > 0.38 h > 304 > 27.5% > 0.38 h >304 >27.5% Location on isotherme Relative vapor pressure (p/p 0) Zone <0.02p/p0 Zone I, extreme left 0.02–0.2p/p 0 Zone I 0.2–0.75p/p 0 Zone IIA 0.75–0.85p/p 0 Zone IIB > 0.85 p/p0 Zone III > 0.85 p/p0 Zone III W ater properties Normal Normal Structure Critical part of native protein structure W ater interacts principally with charg ed g roups (~2 HOH/g roup) At 0.07 h: transition in surface water from disordered to ordered and/or from dispersed to clustered W ater interacts principally with polar protein surface g roups (~1 HOH/ polar site) W ater clusters centered on charg ed and polar sites Clusters fluctuate in size and/or arrang ement At 0.25 h: start of condensation of water onto weakly interacting unfilled patches of protein surface At 0.38 h: monolayer of water covers the surface of the protein and water (table continued on next page) Pag e 41 (table continued from previous page) Bulk-phase water Properties Constitutional waterb Hydration shell (<3 Å from surface) Freec Entrappedd state; associated with completion of charg ed g roup hydration At 0.15 h: long -rang e connectivity of the surface water is established phase beg ins to form, and g lass-rubber transition occurs Thermodynamic transfer properties f D¯G (kJ/mol) > ¦-6¦ -6 -0.8 Close to bulk water NA NA D¯H (kJ/mol) > ¦-17¦ -70 -2.1 Close to bulk water NA NA Approximate mobility (residence time) 10-2 to 10-8 sec <10-8 sec < 10-9 sec < 10-9 to 10-11 sec 10-11 to 10-12 sec 10-11 to 10- 12 sec Freezability Unfreezable Unfreezable Unfreezable Unfreezable Normal Normal Solvent capability None None None to slig ht Slig ht to moderate Normal Normal Protein properties Structure Folded state stable W ater beg ins to plasticize amorphous reg ions Further plasticization of amorphous reg ions Mobility Enzymatic activity negligible Enzymatic activity negligible Internal protein motion (H exchang e) increases from 1/1000 at 0.04 h to full solution rate at 0.15 h At 0.1–0.15 h: chymotrypsin and some other enzymes develop activity At 0.38 h: lysozyme specific activity is 0.1 that in dilute solution Maximum Maximum aData from Rupley and Careri [103], Otting et al. [86], Lounnas and Pettit [71,72], Franks [30], and other sources. Based larg ely on lysozyme. bW ater molecules that occupy specific locations in the interior of the solute macromolecule. cMacroscopic flow not physically constrained by a macromolecular matrix. dMacroscopic flow physically constrained by a macromolecular matrix. eSee Fig ure 19. fPartial molar values for transfer of water from bulk phase to hydration shell. Pag e 42 needed for “full hydration,” that is, occupancy of all first-layer sites. Further added water will have properties that do not differ significantly from those of bulk water. 2.8 Water Activity and Relative Vapor Pressure 2.8.1 Introduction It has long been recognized that a relationship, although imperfect, exists between the water content of food and its perishability. Concentration and dehydration processes are conducted primarily for the purpose of decreasing the water content of a food, simultaneously increasing the concentration of solutes and thereby decreasing perishability. However, it has also been observed that various types of food with the same water content differ significantly in perishability. Thus, water content alone is not a reliable indicator of perishability. This situation is attributable, in part, to differences in the intensity with which water associates with nonaqueous constituents—water engaged in strong associations is less able to support degradative activities, such as growth of microorganisms and hydrolytic chemical reactions, than is weakly associated water. The term “water activity” (aw) was developed to account for the intensity with which water associates with various nonaqueous constituents. Food stability, safety, and other properties can be predicted far more reliably from aw than from water content. Even so, aw is not a totally reliable predictor. The reasons for this will be explained in the next section. Despite this lack of perfection, aw correlates sufficiently well with rates of microbial growth and many degradative reactions to make it a useful indicator of product stability and microbial safety. The fact that aw is specified in some U.S. federal regulations dealing with good manufacturing practices for food attests to its usefulness and credibility [44]. 2.8.2 Definition and Measurement The notion of substance “activity” was rigorously derived from the laws of equilibrium thermodynamics by G. N. Lewis, and its application to foods was pioneered by Scott [108,109]. It is sufficient here to state that aw = ƒ/ƒ0 (1) where ƒ is the fugacity of the solvent (fugacity being the escaping tendency of a solvent from solution) and ƒ0 is the fugacity of the pure solvent. At low pressures (e.g., ambient), the difference between ƒ/ƒ0 and p/p0 is less than 1%, so defining aw in terms of p/p0 is clearly justifiable. Thus, aw = p/p0 (2) This equality is based on the assumptions of solution ideality and the existence of thermodynamic equilibrium. With foods, both assumptions are generally violated. Consequently, Equation 2 must be taken as an approximation and the proper expression is (3) Because p/p0 is the measured term and sometimes does not equal aw, it is more accurate to use the term p/p0 rather than aw. This practice will be followed here. “Relative vapor pressure” (RVP) is the name for p/p0, and these terms will be used interchangeably. Despite the scientific soundness of using RVP rather aw, the reader should be aware that the term aw is in widespread use, appears in other chapters of this book and is not improper provided the user understands its true meaning. Failure of the aw-RVP approach to be a perfect estimator of food stability occurs for two Pag e 43 basic reasons: violation of assumptions underlying Equation 2, and solute-specific effects. Violation of Equation 2 assumptions can, but usually does not, detract unduly from the usefulness of RVP as a technological tool. An exception occurs if dry products are prepared by absorption of water rather than desorption (hysteresis effects), and this will be discussed later. Violation of Equation 2 assumptions does, however, invalidate RVP as a tool for mechanistic interpretations in instances where the theoretical models used are based on these assumptions (often true of models for moisture sorption isotherms). In a few instances that can be of great importance, solute-specific effects can cause RVP to be a poor indicator of food stability and safety. This can occur even when the assumptions underlying Equation 2 are fully met. In these situations, foods with the same RVP but different solute compositions will exhibit different stabilities and other properties. This is an important point that must not be overlooked by anyone relying on RVP as a tool for judging the safety or stability of food. Figure 15 is provided to reinforce this point. These data clearly indicate that the minimum p/p0 for growth of Staphylococcus aureus is dependent on solute type. RVP is related to percent equilibrium relative humidity (ERH) of the product environment as follows: RVP = p/p0 = %ERH/100 (4) Two aspects of this relationship should be noted. First, RVP is an intrinsic property of the sample whereas ERH is a property of the atmosphere in equilibrium with the sample. Second, the Equation 4 relationship is an equality only if equilibrium has been established between the product and its environment. Establishment of equilibrium is a time-consuming process even with very small samples (less than 1 g) and almost impossible in large samples, especially at temperatures below ~50°C. The RVP of a small sample can be determined by placing it in a closed chamber for a time sufficient to achieve apparent equilibrium (constant weight) and then measuring either pressure or relative humidity in the chamber [33,52,90,119,125]. Various types of instruments are available for measuring pressure (manometers) and relative humidity (electric hygrometers, dew-point instruments). Knowledge of freezing point depression can also be used to determine RVP [26]. Based on collaborative studies, the precision of aw determinations is about ±0.02. If one desires to adjust a small sample to a specific RVP, this can be done by placing it in FIGURE 15 Minimum relative vapor pressure (RVP) for g rowth of Staphylococcus aureus as influenced by solute used to produce the RVP. Temperature is close to optimum for g rowth. PEG is polyethylene g lycol. (From Ref. 11.) Pag e 44 a closed chamber at constant temperature, maintaining sample atmosphere at constant relative humidity by means of an appropriate saturated salt solution [69,119], and storing it until constant sample weight is achieved. 2.8.3 Temperature Dependence Relative vapor pressure is temperature dependent, and the Clausius-Clapeyron equation in modified form provides a means for estimating this temperature dependence. This equation, although based on aw, is applicable to RVP and has the following form [128]: (5) where T is absolute temperature, R is the gas constant, and DH is the isosteric net heat of sorption at the water content of the sample. By rearrangement, this equation can be made to conform to the generalized equation for a straight line. It then becomes evident that a plot of In aw versus 1/T (at constant water content) should be linear and the same should be true for In p/p0 versus 1/T. Linear plots of In p/p0 versus 1/T for native potato starch at various moisture contents are shown in Figure 16. It is apparent that the degree of temperature dependence is a function of moisture content. At a starting p/p0 of 0.5, the temperature coefficient is 0.0034°C-1 over the temperature range 2–40°C. Based on the work of several investigators, temperature coefficients for p/p0 (temperature range 5–50°C; starting p/p0 0.5) range from 0.003 to 0.02°C-1 for high-carbohydrate or high-protein foods [128]. Thus, depending on the product, a 10°C change FIGURE 16 Relationship between relative water vapor pressure and temperature for native potato starch of different water contents. W ater content values following each line are g HOH/g dry starch. (From Ref. 128.) Pag e 45 in temperature can cause a 0.03–0.2 change in p/p0. This behavior can be important for a packaged food because it will undergo a change in RVP with a change in temperature, causing the temperature dependence of its stability to be greater than that of the same product unpackaged. Plots of In p/p0 versus 1/T are not always linear over broad temperature ranges, and they generally exhibit sharp breaks with the onset of ice formation. Before showing data at subfreezing temperatures, it is appropriate to consider the definition of RVP as it applies to subfreezing temperatures. This is necessary because a question arises as to whether the denominator term (p0) should be equated to the vapor pressure of supercooled water or to the vapor pressure of ice. The vapor pressure of supercooled water turns out to be the proper choice because (a) values of RVP at subfreezing temperatures can then, and only them, be accurately compared to RVP values at above-freezing temperatures and (b) choice of the vapor pressure of ice as p0 would result, for samples containing ice, in a meaningless situation whereby RVP would be unity at all subfreezing temperatures. The second point results because the partial pressure of water in a frozen food is equal to the vapor pressure of ice at the same temperature [25,121]. Because the vapor pressure of supercooled water has been measured down to -15°C, and the vapor pressure of ice has been measured to much lower temperatures, it is possible to accurately calculate RVP values for frozen foods. This is clearly apparent when one considers the following relationship: (6) Where pff is the partial pressure of water in partially frozen food, p0(SCW) is the vapor pressure of pure supercooled water, and pice is the vapor pressure of pure ice. Presented in Table 5 are RVP values calculated from the vapor pressures of ice and supercooled water, and these values are identical to those of frozen foods at the same temperatures. Figure 17 is a plot of log p/p0 versus 1/T, illustrating that (a) the relationship is linear at TABLE 5 Vapor Pressures and Vapor Pressure Ratios of W ater and Ice Vapor pressure Liquid watera Iceb or food containing ice Temperature (°C) (Pa) (torr) (Pa) (torr) (pice/p water) 0 611b 4.58 611 4.58 1.00 -5 421 3.16 402 3.02 0.95 -10 287 2.15 260 1.95 0.91 -15 191 1.43 165 1.24 0.86 -20 125 0.94 103 0.77 0.82 -25 80.7 0.61 63 0.47 0.78 -30 50.9 0.38 38 0.29 0.75 -40 18.9 0.14 13 0.098 0.69 -50 6.4 0.05 3.9 0.029 0.61 aSupercooled at all temperatures except 0°C. Observed data above -15°C, calculated below -15°C [79]. bObserved data from Ref. 69 FIGURE 17 Relationship between relative vapor pressure and temperature for a complex food above and below freezing . (From Ref. 24.) subfreezing temperatures, (b) the influence of temperature on RVP is typically far greater at subfreezing temperatures than at above-freezing temperatures, and (c) a sharp break occurs in the plot at the freezing point of the sample. Two important distinctions should be noted when comparing RVP values at above- and below-freezing temperatures. First, at above-freezing temperatures, RVP is a function of sample composition and temperature, with the former factor predominating. At subfreezing temperatures, RVP becomes independent of sample composition and depends solely on temperature; that is, in the presence of an ice phase RVP values are not influenced by the kind or ratio of solutes present. As a consequence, any subfreezing event that is influenced by the kind of solute present (e.g., diffusion-controlled processes, catalyzed reactions, and reactions that are affected by the absence or presence of cryoprotective agents, by antimicrobial agents, and/or by chemicals that alter pH and oxidation-reduction potential) cannot be accurately forecast based on the RVP value [25]. Consequently, RVP values at subfreezing temperatures are far less valuable indicators of physical and chemical events than are RVP values at abovefreezing temperatures. It follows that knowledge of RVP at a subfreezing temperature cannot be used to predict RVP at an above-freezing temperature. Second, as the temperature is changed sufficiently to form or melt ice, the meaning of RVP, in terms of food stability, also changes. For example, in a product at -15°C (p/p0=0.86), microorganisms will not grow and chemical reactions will occur slowly. However, at 20°C and Pag e 47 p/p0 0.86, some chemical reactions will occur rapidly and some microorganisms will grow at moderate rates. 2.9 Moisture Sorption Isotherms 2.9.1 Definition and Zones A plot of water content (expressed as mass of water per unit mass of dry material) of a food versus p/p0 at constant temperature is known as a moisture sorption isotherm (MSI). Information derived from MSIs are useful (a) for concentration and dehydration processes, because the ease or difficulty of water removal is related to RVP, (b) for formulating food mixtures so as to avoid moisture transfer among the ingredients, (c) to determine the moisture barrier properties needed in a packaging material, (d) to determine what moisture content will curtail growth of microorganisms of interest, and (e) to predict the chemical and physical stability of food as a function of water content (see next section). Shown in Figure 18 is a schematic MSI for a high-moisture food plotted to include the full range of water content from normal to dry. This kind of plot is not very useful because the data of greatest interest—those in the low-moisture region—are not shown in sufficient detail. Omission of the high-moisture region and expansion of the low-moisture region, as is usually done, yields an MSI that is much more useful (Fig. 19). Several substances that have MSIs of markedly different shapes are shown in Figure 20. These are resorption (or adsorption) isotherms prepared by adding water to previously dried samples. Desorption isotherms are also common. Isotherms with a sigmoidal shape are characteristic of most foods. Foods such as fruits, confections, and coffee extract that contain large amounts of sugar and other small, soluble molecules and are not rich in polymeric materials exhibit a J-type isotherm shown as curve 1 in Figure 20. The shape and position of the isotherm are determined by several factors including sample composition, physical structure of the sample (e.g., crystalline or amorphous), sample pretreatments, temperature, and methodology. Many attempts have been made to model MSIs, but success in achieving good conformFIGURE 18 Schematic moisture sorption isotherm encompassing a broad rang e of moisture contents. Pag e 48 FIGURE 19 Generalized moisture sorption isotherm for the low-moisture seg ment of a food (20°C). ance of a model to the full range of actual data for an MSI has been difficult. The oldest and best known model is that of Brunauer, Emmett, and Teller [8]. One of the best models is that developed by Guggenheim [36], Anderson [2], and DeBoer [16], and this is referred to as the GAB model. As an aid to understanding the meaning and usefulness of sorption isotherms it is sometimes appropriate to divide them into zones as indicated in Figure 19. As water is added (resorption), sample composition moves from Zone I (dry) to Zone III (high moisture) and the properties of water associated with each zone differ significantly. These properties are described next and are summarized in Table 4. Water present in Zone I of the isotherm is most strongly sorbed and least mobile. This water associates with accessible polar sites by water-ion or water-dipole interactions, is unfreezable at -40°C, has no ability to dissolve solutes, and is not present in sufficient amount to have a plasticizing effect on the solid. It behaves simply as part of the solid. The high-moisture end of Zone I (boundary of Zones I and II) corresponds to the “BET monolayer” moisture content of the food. The BET monolayer value should be thought of as approximating the amount of water needed to form a monolayer over accessible, highly polar groups of the dry matter. In the case of starch, this amounts to one HOH per anhydroglucose unit [126]. Zone I water constitutes a tiny fraction of the total water in a high-moisture food material. Water added in Zone II occupies first-layer sites that are still available. This water associates with neighboring water molecules and solute molecules primarily by hydrogen bonding, is slightly less mobile than bulk water, and most of it is unfreezable at - 40°C. As water is added in the vicinity of the low-moisture end of Zone II, it exerts a significant plasticizing action on solutes, lowers their glass transition temperatures, and causes incipient swelling of the solid matrix. This action, coupled with the beginning of solution processes, leads to an Pag e 49 FIGURE 20 Resorption isotherms for various foods and biolog ical substances. Temperature 20°C, except for number 1, which is 40°C: (1) confection (main component powdered sucrose), (2) spray-dried chicory extract, (3) roasted Columbian coffee, (4) pig pancreas extract powder, (5) native rice starch. (From Ref. 127.) acceleration in the rate of most reactions. Water in Zones I and Zone II usually constitutes less than 5% of the water in a highmoisture food material. In the vicinity of the junction of Zones II and III, water is sufficient to complete a true monolayer hydration shell for macromolecules such as globular proteins, and is sufficient to lower the glass transition temperature of macromolecules so that sample temperature and Tg are equal. Further addition of water (Zone III) causes a glass-rubber transition in samples containing glassy regions, a very large decrease in viscosity, a very large increase in molecular mobility, and commensurate increases in the rates of many reactions. This water is freezable, available as a solvent, and readily supports the growth of microorganism. Zone III water is referred to as bulk-phase water (Table 4). Additional water will have the properties of bulk-phase water and will not alter properties of existing solutes. In gels or cellular systems, bulk-phase water is physically entrapped so that macroscopic flow is impeded. In all other respects this water has properties similar to that of water in a dilute salt solution. This is reasonable, since a typical water molecule added in Zone III is “insulated” from the effects of solutes molecules by several layers of Zone I and Zone II water molecules. The bulk-phase water of Zone III, either entrapped or free, usually constitutes more than 95% of the total water in a high-moisture food, a fact that is not evident from Figure 19. As mentioned earlier, the zone boundaries indicated in Figure 19 are simply an aid to discussion rather than a reality. It is believed that water molecules can interchange rapidly within and between “zones” and that the concept of a continuum of water properties existing through Zones I–III is conceptually sounder than the notion of distinctly different properties existing in each zone. It is also of interest that addition of water to a dry material containing only a few water molecules will increase the mobility and lessen the residence time of these original water Pag e 50 molecules [103]. However, addition of water to materials already having complete or near complete hydration shells is unlikely to have a significant effect on the properties of water originally present. The important effects that these solute-induced differences in water properties have on stability of foods will be discussed in a later section. At this point, it will suffice to say that the most mobile fraction of water existing in any food sample governs stability. 2.9.2 Temperature Dependence As mentioned earlier, RVP is temperature dependent; thus MSIs must also be temperature dependent. An example involving potato slices is shown in Figure 21. At any given moisture content, food p/p0 increasing with increasing temperature, in conformity with the Clausius-Clapeyron equation. 2.9.3 Hysteresis An additional complication is that an MSI prepared by addition of water (resorption) to a dry sample will not necessarily be superimposable on an isotherm prepared by desorption. This lack

Leave a Reply

Your email address will not be published.