The Nature of Lipophilicity

The physiochemical nature of lipophilicity can bediscerned from relations between logP and other physical properties, whether measured or calculated. Perusal of relationships between the empirical correction factors used in fragment additivity schemes, and actual structure in solution as determined by spectroscopic methods, can also give clues.

Clearly, lipophilicity depends on the relative solvation energies between water and the lipid (octanol) phase. More can be learned by study of each phase separately, than by study of complex partition systems with solutes bearing both nonpolar and polar residues, where it is never easy to ascribe an effect to one or the other phase, or to a combination of effects.

One of the first computerized methods proposed for the estimation of logP was SCAP (solvent-dependent conformational analysis) developed by Hopfinger and Buttershell [56] in 1976. For simple solutes, a solvation energy was computed for each phase separately, using solvation shell parameters and applying these to a molecular mechanics model of the solute.

A modern equivalent to a SCAP computation, but much more computationally intense, is the free energy perturbation method, whereby a partition coefficient difference can be calculated for simple solutes between pure solvents. Thus, Richards and coworkers in 1989 [57] used molecular dynamics simulations and the free energy perturbation method to compute the difference in log/3 between methanol and ethanol, partitioned between water and carbon tetrachloride. Calculated and experimental values for this difference agreed to within 0.06 logP units.

Unfortunately, the high content of water in the octanol phase (2 mol L 1 at equilibrium) does not lend it to detailed simulation. A simulation taking place in a "box" of solvent would have to contain enough solvent molecules to determine whether the water was evenly distributed or tended to cluster round, or hydrate, the solute. The flexibility of the octanol molecule (six rotatable bonds) would greatly increase the simulation time if using many explicit solvent molecules. Octanol was chosen for QSAR studies for a number of reasons, one of which was its high water content and its ability to hy drogen bond, so as to model a biological membrane better than a pure lipid solvent such as carbon tetrachloride or benzene.

0 0

Post a comment