A classical approach in chemistry for obtaining insights into mechanisms of reactions is to identify and characterize their intermediate and transition states. Such an approach was first developed for studying ‘simple’ chemical reactions that involve covalent bond breaking and making but was later extended for more complex reactions such as protein folding in which numerous non-covalent interactions are formed and broken. Reliable criteria for distinguishing between two-state protein folding reactions with single or multiple energy barriers (transition states), multi-state folding involving intermediates and barrier-less downhill folding have, therefore, been important in protein folding studies [1,2]. Allosteric transitions are another example for a ubiquitous type of complex reaction that involves formation and breaking of multiple non-covalent interactions. Such transitions often involve switching between relatively well-defined folded states (i.e. they are order-to-order transitions) and are, therefore, in some sense more ‘simple’ than folding reactions in which an ensemble of unstructured conformations converts into a well-defined native state. Nevertheless, approaches for characterizing transition and intermediate states have been less developed and employed in studies of allosteric transitions than in those of protein folding. The goal of this review is to highlight analogies between the thermodynamic and kinetic characterization of intermediates and pathways of protein folding, many of which involve linear free energy relationships (LFER), and those of allosteric transitions and to consider how an understanding of allostery can be furthered by adopting approaches used in protein folding studies.

The term allostery is now used very generally in connection with any perturbation (e.g. mutation, ligand binding) at one site of a protein (or other types of molecules) that affects the structure and/or function of another distant site [3, 4, 5, 6, 7]. The concept of allostery was, however, initially put forward to explain the deviation from hyperbolic binding curves (i.e. the presence of an initial lag) displayed by certain multi-subunit proteins due to cooperative ligand binding [8]. The distinction between allostery in multi-subunit proteins and in monomeric proteins remains important because many of the developed models and measures for allostery, e.g. the Hill coefficient, are relevant only for the former class. Consequently, some parts of the review will pertain only to multimeric proteins but allostery in monomeric proteins will also be considered where relevant.