'Non-specific' binding. The problem, and a solution.

The concept of 'non-specific' binding, as it relates to studies of the binding of hormones to their receptors, is reviewed. It is concluded that the most widely used operational definition, namely binding that is not displaceable by an excess of unlabelled ligand, is often inaccurate, resulting either in overestimation of the number of high-affinity receptors and underestimation of the affinity of a given hormone for its receptor, or in a curvilinear Scatchard plot suggesting (artifactually) the presence of negative co-operativity or multiple classes of binding sites. The general use of an alternative approach to non-specific binding, in which the non-specific component is assessed from an analysis of total binding, is advocated. The superiority of this approach is illustrated with data on the binding of high-density lipoproteins to their receptors.