Recent Progress in the pKa Estimation of Druglike Molecules by the Nonlinear Regression of Multiwavelength Spectrophotometric pH-Titration Data

Abstract

Recent developments in the computational diagnostic tools for the p K a estimation of druglike molecules carried out by the nonlinear regression of multiwavelength spectrophotometric pH-titration data are demonstrated on the protonation equilibria of silybin. The factor analysis of spectra predict the correct number of components when the signal-to-error ratio SER is higher than 10. The mixed dissociation constants of the drug silybin at ionic strength I = 0.03 and a temperature of 25 ∘ C were determined using two different programs, SPECFIT32 and SQUAD(84). A proposed experimental and computational strategy for the determination of the dissociation constants is presented. The dissociation constant p K a was estimated by nonlinear regression of the { p K a , I } data at 25 ∘ C with SQUAD (and SPECFIT); that is, p K a 1 = 6.898(0.022) and 6.897(0.002); p K a 2 = 8.666(0.021) and 8.667(0.012); p K a 3 = 9.611(0.010) and 9.611(0.004); p K a 4 = 11.501(0.008) and 11.501(0.007). While great progress has been achieved in terms of the reliability of the protonation model estimation, among the most efficient diagnostics of the nonlinear regression of multiwavelength pH-spectra are the goodness-of-fit test, Cattel's scree plot of the factor analysis, spectra deconvolution, the signal-to-error SER ratio analysis, and other tools of efficient spectra analysis.