Experimental binding equilibrium data, resulting from measurement of ligand binding to macromolecular carriers, are usually described by fitting of binding constants. These constants are often highly variable, as illustrated in the present paper by two examples, binding of salicylate to human serum albumin and of oxygen to hemoglobin. In order to avoid over-interpretation of binding constants it is pointed out that the best-fit solution, obtained by least-squares fitting, may be supplemented by several, e. g. thirty, acceptable solutions. It is further shown how the 30 sets of acceptable binding constants, plotted as Klotz' affinity profiles, can serve for evaluation of cooperative effects.Multiple binding equilibria for a ligand to a macromolecular carrier are usually characterized by a set of binding constants, parameters in a binding equation. Such a set of binding constants may serve two purposes, comparison with results from other binding equilibrium studies, and considerations of interactions, i. e. whether binding is cooperative, homogeneous or anticooperative. It is essential for these purposes to have a knowledge of the variation of the binding constants, as originating from the random errors of the observed data.The present paper evaluates the variation of multiple binding constants. Two experimental materials are analysed: data for the binding of salicylate anion to human serum albumin, from equilibrium dialysis measurements, and oxygenhemoglobin data, as published by Winslow et al. [l]. It is shown that the data can be presented by a multitude of fitted isotherms with highly variable binding constants. The results are utilized for consideration of binding interactions. MATERIALS AND METHODSHuman serum albumin was obtained from AB Kabi (Stockholm, Sweden) and was defatted with charcoal in acid solution [2] and lyophilized. Binding equilibria with sodium salicylate were studied by equilibrium dialysis as previously described [3] in a 0.1 M Tris/HCl buffer with 0.1 M sodium chloride, pH 8.2, at 25°C. The time needed for establishing cquilihriiiin was about 8 h; 20 h dialysis time was allowed. absorbance in a 0.5-cm cell at the wavelength 296 nm. It was found that the absorbance was proportional to the concentration within the range of concentrations used. 0 -2 mM, showing the absence of ligand dimerisation. The molar absorption coefficient was 3530 M-' cm-'. Albumin concentration was 100 pM. Errors from binding of salicylate to the membrane and from compartment volume changes were found to be negligible as were a number of other sources of technical errors previously discussed [3]. Free and bound equilibrium concentrations of salicylate were calculated and plotted as r, the concentration of bound salicylate divided by that of albumin, versus the logarithm of the free salicylate concentration. Additional points obtained with varying albumin concentrations, from 100 pM to 1000 pM, were apparently located on the same curve, indicating the absence of albumin-albumin interaction.Equilibrium data for binding ...