- Email: [email protected]
Prognostic significance of autocrine motility factor receptor expression by colorectal cancer and lymph node metastases
58
M O L E C U L A R W E I G H T A N D R E L A T E D PROCEDURES
[S]
[5] T h e R o l e o f M i c e l l e s in Protein-Detergent Interactions 1 By JACQUELINE A. REYNOLDS I. Introduction The interaction of proteins with amphiphilic ligands has received increasing attention in recent years. The practical as well as theoretical importance of these interactions are illustrated by the following examples. 1. Investigations of the molecular properties of membrane proteins and serum lipoporoteins have for the most part required the use of detergents as solubilizing agents and as probes for hydrophobic binding sites, z 2. The popular technique of identifying and cataloging polypeptides on the basis of their mobilities in sodium dodecyl sulfate-polyacrylamide gel electrophoresis is based on a specific type of detergent-protein interaction. 3 3. Two-dimensional polyacrylamide gel electrophoresis using sodium dodecyl sulfate in one direction and the nonionic detergent, Triton X-100, in the other has been used to identify polypeptides containing long hydrophobic sequences or regions. 4 This technique relies on differences in binding characteristics between water-soluble and intrinsic membrane proteins in that the former do not in general bind nonionic detergents. It is apparent from these few examples that an understanding of the thermodynamics of detergent-protein and detergent-detergent interactions is of central importance in many areas of research. It is the purpose of this chapter to outline the theoretical and practical aspects of these interactions with particular emphasis on the competitive effects of micelle formation and protein-detergent binding. II. T h e r m o d y n a m i c Equilibria A. Micelle Formation. In aqueous solution amphiphilic molecules self-associate at a specific concentration (critical micelle concentration) to form well-defined interaction products. The theoretical aspects of this This work was supported in part by National Institutes of Health Grants H L 14882 and NS 12213. z C. Tanford and J. A. Reynolds, Biochim. Biophys. Acta 457, 133 (1976). a T. B. Nielsen and J. A. Reynolds, see Vol. 48, p. 3. 4 A. Helenius and K. Simons, Proc. Natl. Acad. Sci. U.S.A. 74, 529 (1977). METHODS 1N ENZYMOLOGY, VOL. 61
Copyright © 1979by Academic Press, Inc. All rights of reproduction in any formreserved. ISBN 0-12-181961-2
[5]
ROLE OF M I C E L L E S IN P R O T E I N - D E T E R G E N T I N T E R A C T I O N S
I00 .~ 80 6o 09
40 20 0
A
59
i' / - og U°bou.d C[o] +
Iota]/
/¢
4.0 /
/
/
/
/
/
/
-4.0
i I
i
I
/
/" /
~ o
6.o
I
!
[III
8.0 8,0
/
6.0 I
1 6.0
I
4.0
I 2.0
I
I
8.o
1.0
-Io9 loT] FIG. 1. (A) The binding of detergent to protein as a function of total u n b o u n d detergent concentration where the amphiphile has the self-association properties shown in (B). Protein concentration = 10 -5 M, n = 100. (1) AG o = - 8 . 4 kcal/mole; (2) AG o = - 8 . 2 1 kcal/mole. (B) The increase in free m o n o m e r concentration (--) and micelle concentration (. . . . ) as a function of total detergent concentration. Critical micelle concentration = 0.906 x 10-6M; m = 100, AGmie° = - 8 . 1 5 kcal/mole monomer.
p h e n o m e n o n are discussed in detail in a n u m b e r of recent publications. ~-7 We can describe this process b y means of the following equation.
mD ~ Dm
(1)
where m is the average association number, D is the concentration of m o n o m e r i c amphiphile and Dm is the concentration o f micelles. The association constant is
[Z)m]
Z < - [D]"
(2)
Figure 1A shows the increase in concentration of D and D m a s a function of total amphiphilic concentration; where m = 100 and the free energy of micellization AGmic° is - 8 . 1 5 k c a l / m o l e m o n o m e r . It is o f particular im5 C. Tanford, " T h e Hydrophobic Effect." Wiley (Interscience), N e w York, 1973. e C. Tanford, J. Mol. Biol. 67, 59 (1972). C. Tanford, J. Phys. Chem. 78, 2469 (1974).
60
MOLECULAR W E I G H T AND RELATED PROCEDURES
[5]
portance to note that above the critical micelle concentration of 0.91 × 10 -6 M the m o n o m e r concentration increases very little with increasing total amphiphile. As m increases, the increase in D above the critical micelle concentration becomes even less. For an infinitely large aggregate, such as a phospholipid bilayer, the m o n o m e r concentration is effectively constant above the critical micelle concentration. B. Protein-Amphiphile Interaction. The simplest cases o f p r o t e i n detergent interactions are illustrated here for the purpose of demonstrating the interdependence of micelle formation and binding to proteins. It will be obvious that more complicated situations can ensue and can be treated thermodynamically using the principles discussed here and in a more extensive treatise by Steinhardt and Reynolds. 8 If a protein contains n specific binding sites which interact only with D in the monomeric form by means of a two-state process, we can represent the reaction by
P + nD~PDn
(3)
The association constant for this process is
[PD .] J - ~ [p][o]
(4)
If we assume n = 100 (the same value as m in the previous example of micelle formation) and an association constant corresponding to AG O = - 8 . 4 k c a l / m o l e m o n o m e r , we obtain the binding isotherm labelled " 1 " in Fig. lB. If we now lower the free energy of binding to AG o = - 8 . 2 1 kcal/mole, we observe the isotherm labeled " 2 " in Fig. lB. Since the protein must compete with the micellar aggregate for monomeric detergent, it is apparent that if n, AG °, or both are equal to or lower than m and AGmie°, the binding isotherm is shifted to higher total unbound detergent concentrations and is significantly less cooperative. F r o m a practical standpoint, if one wishes to saturate a protein with detergent, it is essential to maintain an appropriate concentration of monomeric ligand if the mode o f interaction is between monomer and protein. The critical micelle concentrations o f ionic amphiphiles can be manipulated by altering the ionic strength. 5 High salt concentrations lower the critical micelle concentrations, and low salt concentrations increase it. H o w e v e r , nonionic detergents are unaffected within experimental error by the ionic strength o f the solution. In this latter case, if the association constant or n u m b e r o f binding sites is significantly smaller than the average aggregation n u m b e r or free energy o f micellization o f pure detergent micelles, one may not be able to fill all binding sites on the protein. 8 j. Steinhardt and J. A. Reynolds, "Multiple Equilibria in Proteins." Academic Press, New York, 1969.
[5]
ROLE OF M I C E L L E S IN P R O T E I N - D E T E R G E N T INTERACTIONS
61
A second type of protein-detergent interaction can involve direct binding of the protein to preformed micelles. P + Om .~ P - O m
(5)
In this case, little or no interaction will be observed until the unbound detergent concentration in solution is close to the critical micelle concentration. The binding will appear highly cooperative, and if the detergent is ionic, the apparent association constant will be ionic strength dependent. Examples of this type of binding have been found with cytochrome bs and the major glycoprotein from human red cell membranes? "1° III. Experimental Procedures The determination of interaction between proteins and amphiphiles has been described in detail by Steinhardt and Reynolds. s Here we will summarize the techniques applicable to the investigation of monomeric and micellar detergent binding. A. Equilibrium Dialysis and Ultrafiltration. These techniques are feasible when the total unbound detergent concentration is below the critical micelle concentration. Both methods require the use of a semipermeable membrane which allows rapid passage of unbound ligand and retains the protein-ligand complex. The amphiphile concentration is determined on both sides of the membrane, the difference being the concentration of bound ligand. If micelles are present, equilibration times are excessively long and the techniques consequently unreliable. B. Gel Filtration Chromatography. This method is applicable either above or below the critical micelle concentration but requires large amounts of amphiphile. A column buffer is prepared containing a specific concentration of ligand, and the sample is applied in an excess of ligand. The eluted protein peak is analyzed for protein and excess amphiphile. Each experimental binding point at a specific unbound ligand concentration requires a separate elution experiment. Additionally, it is necessary to show that the protein-detergent complex has a different elution position than that of pure micelles, since overlapping of these two fractions will give erroneously high binding values. C. Analytical Ultracentrifugation. Both the binding of ligand and the molecular weight of the protein are obtained by this procedure. The quantity determined directly at equilibrium in the analytical ultracentrifuge is Mp(1 - t~'p) where Mp is the molecular weight of the protein, th' is the effective partial specific volume of the particle, and p is the solvent density. The factor (1 - ¢b'P) can be set equal to (1 - tip) + BE(1 -- VLO). In this N. C. Robinson and C. Tanford,Biochemistry 14, 369 (1975). 1oS. P. Grefrathand J. A. Reynolds,Proc. Natl. Acad. Sci. U.S.A. 71, 3913 (1974).
62
MOLECULAR WEIGHT AND RELATED PROCEDURES
[5]
latter relation, ~ is the partial specific volume of the protein, ~L is the partial specific volume of the ligand, and 8L is the grams of bound ligand per gram protein. Measurements are made at several densities, obtained by replacement of H~O with D20, and ~L is obtained from the slope of a plot of Mp(l - ~b'p) versus p, i.e., - d[Mp(1 - 6'P)] dp = Mp(~ +
8L~L)
This procedure has the significant advantage that very little material is required. It is important to note that it is assumed that D,O freely exchanges with bound H,O, thus eliminating the buoyant density term due to bound water. The method cannot be used when the density is altered by adding sucrose or salts, since these solutes do not exchange with bound water, thus requiring an a priori knowledge of the amount of hydration in order to calculate this portion of the buoyant density term. The exact method of determining molecular weight and extent of binding by this procedure is published elsewhere in some detail. 11 D. Indirect Methods. Alteration in optical properties is often used to determine the extent of interaction between proteins and amphiphilic ligands. Any such procedure must be calibrated against a rigorous thermodynamic method to ensure that optical properties are indeed a unique linear function of extent of binding. E. Sources o f Information Regarding Critical Micelle Concentrations. The most complete current listing of critical micelle concentrations is found in a United States Department of Commerce Publication by P. Mukerjee and K. J. Mysels entitled "Critical Micelle Concentrations of Aqueous Surfactant Systems," NSRDS-NBS 36, 1971. Additional and somewhat newer data on a more limited number of detergents is available in Tanford. 5 n j. A. Reynoldsand C. Tanford,Proc. Natl. Acad. Sci. U.S.A. 73, 4467 (1976).
M O L E C U L A R W E I G H T A N D R E L A T E D PROCEDURES
[S]
[5] T h e R o l e o f M i c e l l e s in Protein-Detergent Interactions 1 By JACQUELINE A. REYNOLDS I. Introduction The interaction of proteins with amphiphilic ligands has received increasing attention in recent years. The practical as well as theoretical importance of these interactions are illustrated by the following examples. 1. Investigations of the molecular properties of membrane proteins and serum lipoporoteins have for the most part required the use of detergents as solubilizing agents and as probes for hydrophobic binding sites, z 2. The popular technique of identifying and cataloging polypeptides on the basis of their mobilities in sodium dodecyl sulfate-polyacrylamide gel electrophoresis is based on a specific type of detergent-protein interaction. 3 3. Two-dimensional polyacrylamide gel electrophoresis using sodium dodecyl sulfate in one direction and the nonionic detergent, Triton X-100, in the other has been used to identify polypeptides containing long hydrophobic sequences or regions. 4 This technique relies on differences in binding characteristics between water-soluble and intrinsic membrane proteins in that the former do not in general bind nonionic detergents. It is apparent from these few examples that an understanding of the thermodynamics of detergent-protein and detergent-detergent interactions is of central importance in many areas of research. It is the purpose of this chapter to outline the theoretical and practical aspects of these interactions with particular emphasis on the competitive effects of micelle formation and protein-detergent binding. II. T h e r m o d y n a m i c Equilibria A. Micelle Formation. In aqueous solution amphiphilic molecules self-associate at a specific concentration (critical micelle concentration) to form well-defined interaction products. The theoretical aspects of this This work was supported in part by National Institutes of Health Grants H L 14882 and NS 12213. z C. Tanford and J. A. Reynolds, Biochim. Biophys. Acta 457, 133 (1976). a T. B. Nielsen and J. A. Reynolds, see Vol. 48, p. 3. 4 A. Helenius and K. Simons, Proc. Natl. Acad. Sci. U.S.A. 74, 529 (1977). METHODS 1N ENZYMOLOGY, VOL. 61
Copyright © 1979by Academic Press, Inc. All rights of reproduction in any formreserved. ISBN 0-12-181961-2
[5]
ROLE OF M I C E L L E S IN P R O T E I N - D E T E R G E N T I N T E R A C T I O N S
I00 .~ 80 6o 09
40 20 0
A
59
i' / - og U°bou.d C[o] +
Iota]/
/¢
4.0 /
/
/
/
/
/
/
-4.0
i I
i
I
/
/" /
~ o
6.o
I
!
[III
8.0 8,0
/
6.0 I
1 6.0
I
4.0
I 2.0
I
I
8.o
1.0
-Io9 loT] FIG. 1. (A) The binding of detergent to protein as a function of total u n b o u n d detergent concentration where the amphiphile has the self-association properties shown in (B). Protein concentration = 10 -5 M, n = 100. (1) AG o = - 8 . 4 kcal/mole; (2) AG o = - 8 . 2 1 kcal/mole. (B) The increase in free m o n o m e r concentration (--) and micelle concentration (. . . . ) as a function of total detergent concentration. Critical micelle concentration = 0.906 x 10-6M; m = 100, AGmie° = - 8 . 1 5 kcal/mole monomer.
p h e n o m e n o n are discussed in detail in a n u m b e r of recent publications. ~-7 We can describe this process b y means of the following equation.
mD ~ Dm
(1)
where m is the average association number, D is the concentration of m o n o m e r i c amphiphile and Dm is the concentration o f micelles. The association constant is
[Z)m]
Z < - [D]"
(2)
Figure 1A shows the increase in concentration of D and D m a s a function of total amphiphilic concentration; where m = 100 and the free energy of micellization AGmic° is - 8 . 1 5 k c a l / m o l e m o n o m e r . It is o f particular im5 C. Tanford, " T h e Hydrophobic Effect." Wiley (Interscience), N e w York, 1973. e C. Tanford, J. Mol. Biol. 67, 59 (1972). C. Tanford, J. Phys. Chem. 78, 2469 (1974).
60
MOLECULAR W E I G H T AND RELATED PROCEDURES
[5]
portance to note that above the critical micelle concentration of 0.91 × 10 -6 M the m o n o m e r concentration increases very little with increasing total amphiphile. As m increases, the increase in D above the critical micelle concentration becomes even less. For an infinitely large aggregate, such as a phospholipid bilayer, the m o n o m e r concentration is effectively constant above the critical micelle concentration. B. Protein-Amphiphile Interaction. The simplest cases o f p r o t e i n detergent interactions are illustrated here for the purpose of demonstrating the interdependence of micelle formation and binding to proteins. It will be obvious that more complicated situations can ensue and can be treated thermodynamically using the principles discussed here and in a more extensive treatise by Steinhardt and Reynolds. 8 If a protein contains n specific binding sites which interact only with D in the monomeric form by means of a two-state process, we can represent the reaction by
P + nD~PDn
(3)
The association constant for this process is
[PD .] J - ~ [p][o]
(4)
If we assume n = 100 (the same value as m in the previous example of micelle formation) and an association constant corresponding to AG O = - 8 . 4 k c a l / m o l e m o n o m e r , we obtain the binding isotherm labelled " 1 " in Fig. lB. If we now lower the free energy of binding to AG o = - 8 . 2 1 kcal/mole, we observe the isotherm labeled " 2 " in Fig. lB. Since the protein must compete with the micellar aggregate for monomeric detergent, it is apparent that if n, AG °, or both are equal to or lower than m and AGmie°, the binding isotherm is shifted to higher total unbound detergent concentrations and is significantly less cooperative. F r o m a practical standpoint, if one wishes to saturate a protein with detergent, it is essential to maintain an appropriate concentration of monomeric ligand if the mode o f interaction is between monomer and protein. The critical micelle concentrations o f ionic amphiphiles can be manipulated by altering the ionic strength. 5 High salt concentrations lower the critical micelle concentrations, and low salt concentrations increase it. H o w e v e r , nonionic detergents are unaffected within experimental error by the ionic strength o f the solution. In this latter case, if the association constant or n u m b e r o f binding sites is significantly smaller than the average aggregation n u m b e r or free energy o f micellization o f pure detergent micelles, one may not be able to fill all binding sites on the protein. 8 j. Steinhardt and J. A. Reynolds, "Multiple Equilibria in Proteins." Academic Press, New York, 1969.
[5]
ROLE OF M I C E L L E S IN P R O T E I N - D E T E R G E N T INTERACTIONS
61
A second type of protein-detergent interaction can involve direct binding of the protein to preformed micelles. P + Om .~ P - O m
(5)
In this case, little or no interaction will be observed until the unbound detergent concentration in solution is close to the critical micelle concentration. The binding will appear highly cooperative, and if the detergent is ionic, the apparent association constant will be ionic strength dependent. Examples of this type of binding have been found with cytochrome bs and the major glycoprotein from human red cell membranes? "1° III. Experimental Procedures The determination of interaction between proteins and amphiphiles has been described in detail by Steinhardt and Reynolds. s Here we will summarize the techniques applicable to the investigation of monomeric and micellar detergent binding. A. Equilibrium Dialysis and Ultrafiltration. These techniques are feasible when the total unbound detergent concentration is below the critical micelle concentration. Both methods require the use of a semipermeable membrane which allows rapid passage of unbound ligand and retains the protein-ligand complex. The amphiphile concentration is determined on both sides of the membrane, the difference being the concentration of bound ligand. If micelles are present, equilibration times are excessively long and the techniques consequently unreliable. B. Gel Filtration Chromatography. This method is applicable either above or below the critical micelle concentration but requires large amounts of amphiphile. A column buffer is prepared containing a specific concentration of ligand, and the sample is applied in an excess of ligand. The eluted protein peak is analyzed for protein and excess amphiphile. Each experimental binding point at a specific unbound ligand concentration requires a separate elution experiment. Additionally, it is necessary to show that the protein-detergent complex has a different elution position than that of pure micelles, since overlapping of these two fractions will give erroneously high binding values. C. Analytical Ultracentrifugation. Both the binding of ligand and the molecular weight of the protein are obtained by this procedure. The quantity determined directly at equilibrium in the analytical ultracentrifuge is Mp(1 - t~'p) where Mp is the molecular weight of the protein, th' is the effective partial specific volume of the particle, and p is the solvent density. The factor (1 - ¢b'P) can be set equal to (1 - tip) + BE(1 -- VLO). In this N. C. Robinson and C. Tanford,Biochemistry 14, 369 (1975). 1oS. P. Grefrathand J. A. Reynolds,Proc. Natl. Acad. Sci. U.S.A. 71, 3913 (1974).
62
MOLECULAR WEIGHT AND RELATED PROCEDURES
[5]
latter relation, ~ is the partial specific volume of the protein, ~L is the partial specific volume of the ligand, and 8L is the grams of bound ligand per gram protein. Measurements are made at several densities, obtained by replacement of H~O with D20, and ~L is obtained from the slope of a plot of Mp(l - ~b'p) versus p, i.e., - d[Mp(1 - 6'P)] dp = Mp(~ +
8L~L)
This procedure has the significant advantage that very little material is required. It is important to note that it is assumed that D,O freely exchanges with bound H,O, thus eliminating the buoyant density term due to bound water. The method cannot be used when the density is altered by adding sucrose or salts, since these solutes do not exchange with bound water, thus requiring an a priori knowledge of the amount of hydration in order to calculate this portion of the buoyant density term. The exact method of determining molecular weight and extent of binding by this procedure is published elsewhere in some detail. 11 D. Indirect Methods. Alteration in optical properties is often used to determine the extent of interaction between proteins and amphiphilic ligands. Any such procedure must be calibrated against a rigorous thermodynamic method to ensure that optical properties are indeed a unique linear function of extent of binding. E. Sources o f Information Regarding Critical Micelle Concentrations. The most complete current listing of critical micelle concentrations is found in a United States Department of Commerce Publication by P. Mukerjee and K. J. Mysels entitled "Critical Micelle Concentrations of Aqueous Surfactant Systems," NSRDS-NBS 36, 1971. Additional and somewhat newer data on a more limited number of detergents is available in Tanford. 5 n j. A. Reynoldsand C. Tanford,Proc. Natl. Acad. Sci. U.S.A. 73, 4467 (1976).