Lipophilicity and pKa estimates from gradient high-performance liquid chromatography

The linear-solvent strength (LSS) model of gradient elution has been applied to estimate parameters of lipophilicity and acidity of a series of drugs and model chemicals. Apparent pK(a) values and log k(w) values for individual analytes were determined in 2-3 gradient runs. The first experiment (or first two experiments) uses a wide-range organic modifier gradient with pH chosen for suppressed ionization of the analyte. The result of this experiment allows an estimate of contents of organic modifier of the mobile phase (%B) providing the required retention coefficient, k, for the non-ionized analyte. The following experiment is carried out with the latter %B and a pH-gradient of the aqueous component of the eluent that is sufficient to overlap the possible pK(a)-value of the analyte. The initial pH of the buffer used to make the mobile phase is selected to insure that the analyte is in non-ionized form. The resulting retention time allows an estimate of pK(a) in a solvent of the selected %B. At the same time, estimates of log k(w) can also be obtained. The log k(w) parameter obtained from gradient HPLC by the approach proposed correlated well with the corresponding value obtained by standard procedure of extrapolation of retention data determined in a series of isocratic measurements. Correlation between log k(w) and the reference parameter of lipophilicity, log P, was very good for a series of test analytes and satisfactory for a structurally diverse series of drugs. The approach supported with specific detection procedures can be recommended for fast screening of lipophilicity of individual components of complex mixtures like those produced by combinatorial chemistry. The values of pK(a) obtained in a study were found to correlate with the literature pK(a) data determined in water for a set of aniline derivatives studied. In case of a series of drugs the correlation was less than moderate if the general procedure of pK(a) determination was applied. (C) 2002 Elsevier Science B.V. All rights reserved.