Abstract
The enantiomeric discrimination properties of a number of commercial chiral HPLC stationary phases have been examined using a series of 2-aminotetralin analogues. The mobile phase conditions for each column were varied in order to optimise the separation for the largest number of analytes. Factorial experimental design techniques were used to test the empirically derived mobile phase combinations. A primary limitation of factorial design optimisation strategies, the number of experiments required, was overcome by the development of a simultaneous factorial design method utilising the selectivity of mass spectrometric detection. This enabled the optimisation of the separation of the enantiomers of 12 compounds, using Chiral AGP, to be carried out concurrently. None of the columns tried was able to separate the enantiomers of more than half the analytes. The retention mechanism of the Chiral AGP column was shown to be via cation exchange and non-polar interactions. The ability to form a hydrogen bond between the carbon 8 substitution (acceptor) of the analyte and the stationary phase (donor) was found to be a prerequisite for the separation of enantiomers using Chiral AGP and Chiralcel OD. Neither a rule based nor a mathematical model constructed from calculation derived physicochemical and molecular data were sufficient to describe the enantioselectivity of the Chiral AGP stationary phase. A molecular graphics template model was used to demonstrate the importance of the spatial position of the hydrogen acceptor to the enantioselectivity of the column. Two 2-aminoteralins substituted at the carbon 8 position with a keto-pyrrol function were found to be enantioselectively cleared from in-vitro rat liver preparations dependent upon the substitutions attached to the carbon 2 nitrogen. Differences between these closely related structures were shown to be related to the hydroxylation of the parent molecule, probably, although not exclusively, via the cytochrome P450 isozyme CYP2D6. An in-vivo study using an intravenous dose route showed no sign of enantioselectivity or evidence of the major in-vitro metabolite.