Monte Carlo simulation of error propagation in the determination of binding constants from rectangular hyperbolae. 2. Effect of the maximum-response range

Many processes dictated by chemical equilibria can be described by rectangu lar hyperbolae. Fitting chemical responses to rectangular hyperbolas also a llows the binding constants for these equilibria to be estimated. Unfortuna tely, the propagation of error through the different methods of estimating the binding constants is not well understood. Monte Carlo simulations are u sed to assess the accuracy and precision of binding constants estimated usi ng a nonlinear regression method and three linear plotting methods. The eff ect of the difference between the physical response of the uncomplexed subs trate and the response of the substrate-ligand complex (i.e., the maximum-r esponse range) was demonstrated using errors typical for a capillary electr ophoresis system. It was shown that binding constant estimates obtained usi ng nonlinear regression were more accurate and more precise than estimates from when the other regression methods were used, especially when the maxim um-response range was small. The precision of the nonlinear regression meth od correlated well with the curvature of the binding isotherm. To obtain a precise estimate for the binding constant, the maximum-response range neede d to be much larger (over 70 times larger for the conditions used in this e xperiment) than the error present in individual data points.