B2 Influence of Salt on the Solubility of Proteins

Salts have a large influence on the solubility of proteins. For example, low concentrations of salt in solution increase the solubility of the globular protein hemoglobin [7]. This is called salting in. As the salt concentration is increased, the solubility of hemoglobin decreases, and this is called salting out. Salts containing divalent ions are more efficient at these processes than salts with only monovalent ions.

Protein isolated from soybeans shows the opposite behavior to hemoglobin. Salting out occurs at low salt concentrations and is part of the basis for the production of tofu, a soybean curd product that is a staple food for much of the world's population. The solubility of soybean protein increases at higher concentrations of salt; that is, it salts in. Its solubility also depends upon pH.

The solubility of a protein can be thermodynamically linked to its capacity for binding salt. Solubility vs. [salt] profiles can be fit by nonlinear regression analysis by using models derived from thermodynamic linkage theory. The stoichiometry of salt binding and the binding equilibrium constants are obtained.

In models for the solubility of soybean protein, we assume that two classes of binding sites are responsible for the sequential salting-out and salting-in processes. The concept of linked functions is readily adaptable to building a model encompassing these processes. A similar situation involving two separate bound species was discussed briefly in Section A.3. In the present case, the equilibria are

MX„ + mX ^ MX„Xm K"{ = \MX„Xm}/{[MXn][X]m\ (6.33)

Table 6.5 Diffusion Coefficients of Micelles from Nonlinear Regression Analysis of Electrochemical Probe Data

No. of

106Z>i

Lit. values"

106 DÎ

System

Probe"

Method6

micelles

cm2 s"1

cm2 s"1

cm2 s"1

0.1 M SDS/

MV2+

CV

2

1.41

1.40-1.45

0.84

0.1 M NaCl

0.1 M SDS/

MV2+

CC

2

1.35

1.40-1.45

0.70

0.1 M NaCl

0.1 M SDS/

Fc

CV

2

1.45

1.40-1.45

0.99

0.1 M NaCl

0.1 M CTAB/

Fc

CV

2

1.01

0.5-1.0

0.77

0.1 M NaCl

0.1 M CTAB/

Fc

CV

2

0.73

0.5-0.83

0.38

0.1 M KBr

0.15 M CTAB/

Fc

CV

1

033d

0.1 M TEAB

0.2 M A0T/H20/

Vitamin B12

LSVe

1

0.6

isooctanef

Note: SDS is sodium dodecylsulfate; CTAB is cetyl- (or hexadecyl-) trimethylammonium bromide; TEAB is tetraethylammonium bromide a Probes are methyl viologen (MV2+) and ferrocene (Fc). b CC is chronocoulometry; CV is cyclic voltammetry; see [3, 4, 6] for details. c See [3] and [6] for full citations to D values by alternative methods. Values listed assume that micelles are spherical.

d Probably rod-shaped micelles.

e Water-in-oil microemulsion, data obtained by ultramicroelectrode linear sweep voltammetry. A, is for the water droplet in this fluid.

where M is the unbound protein, X is the free salt, n is the number of moles of X bound to a mole of species MXn, and m is the number of moles of salt bound to a mole of MXnXm. This model represents sequential binding with Ki > K2. This means that the n sites in eq. (6.32) are occupied before the binding at the next set of m sites on the protein (eq. (6.33).

The measured solubility ,Sapp depends on the solubilities of all of the forms of the protein:

S0 = solubility of M

S\ = solubility of MXn relative to S0

S2 = solubility of MXnXm relative to S0.

The relationship representing the measured solubility can be represented by

where .S'app = apparent protein solubility at a given total salt concentration (Cx)• The protein fractions are

/m,o = fraction of total protein as M fM.i = fraction of total protein as MXn fxt.2 = fraction of total protein as MXnXm

This overall equilibrium approach to the linkage problem was used to derive the model in Table 6.6.

Salt-dependent solubility profiles of soybean protein were analyzed directly using a Gauss-Newton nonlinear regression analysis program. As for the two-micelle model discussed previously, solubility profiles were analyzed by fixing the values of n and m and calculating the best least-squares fit. A series of n and m values were tested until those that yielded the minimum error sum for the analysis were found.

Results are shown in Figure 6.4 for solubility of soybean protein in the presence of NaCl or NH4C1, NH4Br and Nal. Figure 6.5 gives data for NH4NO3 and (NH4)2S04 or Na2S04. All plots show the best fits at neutral pH. The resulting regression lines are in excellent agreement with the solubility data. All deviation plots were random. The computed curves are all within a relative standard deviation of ±2% from the experimental data. The final parameter values with corresponding standard errors are presented in Table 6.7.

It must be stressed that the Kj values are not the actual stoichiometric equilibrium binding constants. They are average values representative of only 1 mole salt bound to one protein site. The actual overall equilibrium constants are calculated by raising the Kj value to the corresponding n or m exponent.

As seen in Table 6.7, the salting-out constant, Ku is essentially the same within experimental error for NaCl, NH4C1, NH4Br, NH4NO3, and Nal. Also, values for K2 are similar for these salts. The n and m values are also the same; that is, n = 1 and m = 2 for all of these salts.

Table 6.6 Model for the Solubility of Proteins in the Presence of Salt

Assumptions: CM <s [Jf], thus Cx = [A-] Regression equation:

app 1 + K"xC'i 1 + K"xC'i 1 + K?C% Regression parameters: Data:

Special instructions: Run a series of nonlinear regressions with a series of different fixed integer n and m values until the smallest standard deviation of regression is achieved

Figure 6.4 Influence of salt on the solubility of native soybean protein isolate. Points are experimental values and lines were computed from nonlinear regression analysis onto the model in Table 6.6: NaCl or NH4C1 (O); NH4Br (A); Nal ( + ) (adapted with permission from [1]).

Figure 6.4 Influence of salt on the solubility of native soybean protein isolate. Points are experimental values and lines were computed from nonlinear regression analysis onto the model in Table 6.6: NaCl or NH4C1 (O); NH4Br (A); Nal ( + ) (adapted with permission from [1]).

The relatively low values of 1 and 2 for n and m, respectively, should not be interpreted literally as only a simple binding site, because multiple binding sites with nearly the same equilibrium constant would yield only a single binding isotherm. Hence, a value of n or m represents a class of protein binding sites rather than a single binding site linked to the solubility change of the protein.

Figure 6.5 Influence of salt on the solubility of native soybean protein isolate. Points are experimental values and lines were computed from nonlinear regression analysis onto the model in Table 6.6: NH4N03 (O); (NH4)2S04 or Na2S04 (A) (adapted with permission from [1]).

Figure 6.5 Influence of salt on the solubility of native soybean protein isolate. Points are experimental values and lines were computed from nonlinear regression analysis onto the model in Table 6.6: NH4N03 (O); (NH4)2S04 or Na2S04 (A) (adapted with permission from [1]).

Table S.7 Parameters from the Nonlinear Regression Analysis of Solubility Data of Soybean Protein in Various Salt Solutions3

Salt

KI

K;

Si, '

7c

S2, '

%

n

m

NaCl, NH4C1

41 ±

9

4.7 ±

0.4

31 ±

2

59 ±

4

1

2

NH4Br

53 ±

16

6 ±

2

28 ±

10

62 ±

9

1

2

nh4no3

51 ±

17

6 ±

2

29 ±

12

63 ±

13

1

2

Nal

50 ±

6

2.9 ±

0.3

43 ±

2

74 ±

4

1

2

Na2S04, (NH4)2S04

100 ±

1

1.6 ±

0.1

55.1 ±

0.3

44 ±

1

4

4

The salting-out solubility, 5], shows essentially no trend with respect to the type of anion used whereas a slight trend may exist for the salting-in solubility, S2. Because soybean protein at neutral pH is thought to have a net negative charge, the preceding results can easily be interpreted in terms of an isoelectric binding model; that is, salt cations bind to negative sites on the protein surface, with an average equilibrium constant Kx, and produce a species of zero net charge with a corresponding solubility of Si.

The salting-in of the protein at higher concentrations of salt can be thought of in terms of cations binding, with an average equilibrium constant K2, to the unbound negative protein sites yielding a species, S2, with a net positive charge. Alternatively, binding of both cations and anions of the salt to corresponding negative and positive protein sites can yield species with a net zero or negative charge. The values of Si and S2 are slightly higher for Nal, 43 and 74, respectively, than the chloride, bromide, and nitrate values of approximately 30 and 60.

The salting-out solubilities of native soybean protein, Si, are high in (NH4)2S04 and Na2S04 in comparison with the other salts in Table 6.7. In addition, the K\ value is significantly higher and the K2 value is much smaller than for the other salts. The n and m values, both 4, were different from the n = 1 and m = 2 values for the other salts. Even though the shape of the solubility dependence on sulfate (Figure 6.5) is different from the rest of the salts, the model in Table 6.6 easily describes both types of solubility profiles. Applications of solubility models to other proteins have been discussed [1],

0 0

Post a comment