Conformational energetics, geometric disposition, and intramolecular reactivity in six semi-rigid hydroxy acids

Abstract

MM2 Methods were applied to six hydroxy acids whose hydroxy groups are attached to rigid or semi-rigid frameworks and whose lactonization rates span a range of 3730-fold. A comprehensive analysis for each compound provided energy minima and maxima as the hydroxys and carboxys were systematically rotated in small increments. Each hydroxy acid exists in four or five important energy minima. For example, 2-endo-hydroxy-6-endo-carboxynorbornane has five conformational minima comprising 44, 28, 18, 7, and 2% of the total population. These minima were characterized geometrically in terms of HO ⋯ CO distance r and the angular disposition (θ and φ) of the hydroxy with respect to the carbonyl. It was found that r changes by <0.1 Å; from minimum to minimum for any given hydroxy acid. There is a definite tendency for an electron pair on the hydroxy to point, in preference to the hydroxy proton, toward the carbonyl carbon. Neither van der Waals repulsion nor hydrogen-bonding between the hydroxy proton and the carbonyl oxygen contributes substantially to the energy terms controlling rotational equilibria. Bending distortions, which ameliorate repulsion between the hydroxy proton and carboxy OH, dominate the energy content of the conformational maxima. Lactonization rates could not be correlated with widely varying ‘attack angles’θ and φ. Nor could distance r(remaining relatively constant at 2.90 ± 0.06 Å among the lowest-energy conformers of the hydroxy acids) provide an explanation for the rate data. This does not mean, of course, that distance never has a bearing on reactivity as claimed recently by Dorigo and Houk. Rather, the HO ⋯ CO distances are all sufficiently similar to the contact distance invoked in the ‘spatiotemporal hypothesis’ that the compounds would lactonize at enzyme-like rates were it not for deleterious ring-strain effects. Finally, errors and uncertainties in the Dorigo–Houk paper, and the dangers of computing small rate differnces (such as those affecting product ratios) in terms of structure, are discussed.