Partitioning processes at interfaces: Chromatographic retention and bioavailability

Since the inception of reversed phase liquid chromatography (RPLC) there have been many attempts to correlate chromatographic retention with bioavailability and, the most often used bulk measure, the n-octanol/water partition coefficient. However, from the previous studies, it was found that bulk phases like n-octanol/water may not be appropriate for modeling a partitioning process in an interphase such as biological membranes, and the bonding density of almost all of the commercial C{dollar}sb{lcub}18{rcub}{dollar} columns is too low to provide a suitable model. In this work self-synthesized and well characterized stationary phases were used to determine the best chromatographic conditions for bioavailability estimations.; The solute capacity factor in RPLC using a purely aqueous mobile phase (k{dollar}spprimesb{lcub}rm w{rcub}){dollar} is the most preferable experimental variable for determining lipophilicity information. The k{dollar}spprimesb{lcub}rm w{rcub}{dollar} is usually predicted from the properties of the mobile phase, such as the volume percentage of organic modifier or the polarity of the mobile phase as measured by spectroscopic probes such as ET-30. However, inevitable errors between the predicted and measured log k{dollar}spprimesb{lcub}rm w{rcub}{dollar} were observed because of the selective solvation of the stationary phase by the organic modifier in the mobile phase. It was found that the deviation between the predicted and measured log k{dollar}spprimesb{lcub}rm w{rcub}{dollar} is related to the properties of the solutes. A method of prediction of log k{dollar}spprimesb{lcub}rm w{rcub}{dollar} was developed based on the knowledge of the physico-chemical properties of the solutes. The solvation mechanism was also studied in regard to the bonding density and pore structure of the stationary phase.; Log k{dollar}spprimesb{lcub}rm w{rcub}{dollar} values for selected pesticides, barbiturates and PAHs on a high bonding density column (4.07 {dollar}mu{dollar}mol/m{dollar}sp2){dollar} and a low bonding density column (2.39 {dollar}mu{dollar}mol/m{dollar}sp2){dollar} were determined and were correlated to diverse biological actions. Estimations of bioavailability from lipophilicity measured by RPLC and octanol/water were compared. Generally, RPLC is as good or better than the octanol/water partition coefficient method in predicting biological actions. The best descriptors of biotransfer factors of pesticides for beef and milk are log k{dollar}spprimesb{lcub}rm w{rcub}{dollar} values extrapolated from the ET(30) scale using the high bonding density column.