LAUSR

Protein unfolding is a dynamic cooperative process with many short-lived intermediates. Cooperativity means that similar molecular elements act dependently on each other. The thermodynamics of protein unfolding can be determined with differential scanning calorimetry (DSC). The measurement of the heat capacity provides the temperature profiles of enthalpy, entropy and free energy. The thermodynamics of protein unfolding is completely determined with these thermodynamic properties. We emphasise the model-independent analysis of the heat capacity. The temperature profiles of enthalpy H(T), entropy S(T) and free energy G(T) can be obtained directly by a numerical integration of Cp(T). In evaluating different models for protein unfolding. It is essential to simulate all thermodynamic properties, not only the heat capacity. A chemical equilibrium two-state model is a widely used approximation to protein unfolding. The model assumes a chemical equilibrium between only two protein conformations, the native protein (N) and the unfolded protein (U). The model fits the heat capacity Cp(T) quite well, but fails in simulating the other thermodynamic properties. In this review we propose a modification of the chemical equilibrium two-state model, which removes these inconsistencies. We also propose a new statistical-mechanical two-state model based on a simple, two-parameter partition function Z(T), from which all thermodynamic parameters can be derived. The thermodynamic predictions of the new models are compared to published DSC-experiments obtained with lysozyme, a globular protein, and β-lactoglobulin, a β-barrel protein. Good fits to all thermodynamic properties are obtained. In particular, the models predict a zero free energy for the native protein, which is confirmed experimentally by DSC. This is in contrast to the often-cited chemical equilibrium two-state model, which predict a positive free energy for the native protein. Two-state models use macroscopic fit parameters, the conformational enthalpy and the heat capacity difference between native and unfolded protein. These simulations provide no molecular insight. The review therefore includes a recently published multistate cooperative model based on physicality well-defined molecular parameters only.