Depending on the osmolyte in question, proteins have been observed to have a preference for solvation by either water or the osmolyte. Preferential interaction is a thermodynamic measure of the redistribution of water and osmolyte molecules around the protein surface relative to their distribution in bulk solution. In practice, the measurement of preferential interaction is accomplished by means of a dialysis experiment, under thermodynamic equilibrium, in which the concentration of the osmolyte is measured in the dialyzate and in the compartment containing the protein. In the presence of protein, water and osmolyte distribute around the protein according to their relative preferences for protein groups exposed on the surface. Thus, their concentrations in the vicinity of the protein will be different from their concentrations in bulk solution. The osmolyte concentration differences in the presence and absence of protein in the dialysis experiment can be quantified by densimetric (Lee et al., 1979) or osmotic pressure (Courtenay et al., 2000) measurements, giving a quantity called the preferential interaction parameter. If water is enriched near the protein surface relative to its composition in the bulk solution, a preferential interaction exists between water and the protein surface, resulting in what can be called preferential hydration from the perspective of water or preferential exclusion from the osmolyte perspective. If the osmolyte is found to be enriched near the protein surface, it is said to preferentially interact with or bind to the protein surface and does so by excluding or replacing water. Thus, ifthere is preferential binding of one component (water or osmolyte) with the protein, there will necessarily be a preferential exclusion of the other component. In short, the measured preferential interaction parameter can have either a positive or a negative sign representing, respectively, osmolyte preferential binding or preferential exclusion.
The wealth of protein—osmolyte preferential interaction results has been reported by Timasheff and colleagues (Arakawa and Timasheff, 1982, 1983, 1985; Gekko and Timasheff, 1981; Lee and Timasheff, 1981; Timasheff and Xie, 2003; Xie and Timasheff, 1997a,b). More recently, Record and colleagues have reported such measurements (Courtenay et al., 2000; Felitsky etal., 2004; Hong etal., 2004; Zhang etal., 1996). For many proteins these measurements show that protecting osmolytes are excluded preferentially from both native and unfolded states of proteins, whereas the non-protecting osmolyte, urea, binds these protein species preferentially (Felitsky and Record, 2004; Felitsky et al., 2004; Timasheff, 1992a,b). The preferential interactions appear to be linked to the stabilizing and destabilizing effects that protecting and nonprotecting osmolytes have on proteins. While urea is a well-known protein denaturant, we and others demonstrated that protecting osmolytes have the ability to force intrinsically unstructured proteins to cooperatively fold to native-like protein species that have functional activity (Baskakov and Bolen, 1998a; Gursky, 1999; Henkels and Oas, 2005; Kumar et al., 2001; Mello and Barrick, 2003; Rajagopalan et al., 2005; Wu and Bolen, 2006). The implication is that preferential interactions provide a driving force for folding, with preferential exclusion of protecting osmolytes promoting folding and the preferential binding of urea promoting denaturation. These observations, however, do not provide mechanistic detail of osmolyte effects on protein stability, they only show that stability and protein solvation are linked phenomena.
Using preferential interaction parameters determined for native and denatured protein exposure to the nonprotecting osmolyte urea and to a protecting osmolyte such as trimethylamine N-oxide (TMAO), Timasheff (1992a,b) compared how the two types of osmolytes shift the native (N) to denatured (D) equilibrium (N > D) in thermodynamic terms (see Fig. 23.1). Transfer of the native state in water to urea solution lowers the free energy of the native state because of a net favorable interaction of the urea with native state surface-exposed groups. The free energy change is even more favorable for transfer of the denatured state from water to urea solution because of the larger number of denatured state surface-exposed groups binding preferentially to the urea. These changes result in a smaller free energy gap (AG) between N and D states in urea compared to that in water (see Fig. 23.1); the origin of the smaller aG illustrates how the destabilizing effect of urea on proteins arises. In contrast to urea, transfer of the native and denatured states to protecting osmolyte solution is unfavorable because of net preferential exclusion of the osmolyte at the protein
Net favorable interaction
Net unfavorable interaction
Figure 23.1 The change in chemical potential (free energy) of native and denatured states when a protecting osmolyte (e.g., sarcosine) or a nonprotecting osmolyte (urea) is added. The resulting preferential exclusion of a protecting osmolyte from the protein surface and stabilization of the N —> D transition is driven by a net unfavorable interaction between the osmolyte and the native and denatured states (right), resulting in a AG increase. The preferential binding of urea to Nand D destabilizes N —> D the transition (AG decreases) and results from a net favorable interaction between urea and the native and denatured states (left).
surface. Because of greater surface group exposure in the denatured state, the denatured state is more destabilized (i.e., of higher free energy) in the presence of the protecting osmolyte than the native state. This results in greater protein stabilization in the protecting osmolyte than occurs in water (i.e., aG is increased in osmolyte solution, see Fig. 23.1).
To explore the relationships between preferential interaction and the effects ofosmolytes on protein stability, it is useful to consider which groups exposed on the surface of native and denatured proteins either bind or exclude the osmolyte of interest. Quantitatively, groups that bind osmolyte preferentially (solvophilic groups) are characterized by a group transfer-free energy (GTFE) change of negative sign, whereas GTFEs are positive for groups that exclude osmolyte preferentially (solvophobic groups). GTFEs are discussed in detail later.
Preferential interaction parameters can be used to experimentally obtain the free energy changes for transfer ofthe entire native or denatured protein from water to osmolyte solution. An alternative means of obtaining the transfer-free energy of the entire native state from water to osmolyte is to identify the solvent-exposed side chain and backbone groups on the surface of the native state and sum their individual GTFE contributions. The same process can be used to evaluate the water-to-osmolyte transfer-free energy ofthe entire denatured state. This approach, the transfer-free energy model, was proposed by Tanford in the mid-1960s for use in investigating denatur-ation by urea, but for several reasons was never fully implemented (Auton and Bolen, 2004). This chapter expands on the model, demonstrating what it takes to make it quantitative and predictive of the effects protecting and nonprotecting osmolytes have on protein stability.
Was this article helpful?