Predicting the pathways of protein folding and quantifying the relative thermodynamic stability of intermediate and final states along these pathways constitute two of the most important challenges in modern chemistry. Such predictions are difficult because the desired relative free-energy differences among the solvated intermediates depend importantly on the sum of numerous weak intramolecular forces that contribute to each folded state, on the partition functions representative of these states, and on the effects of water and co-solutes upon these interactions. Paulings hydrogen-bonding motifs [1] and the canonical hydrophobic effect [2] have been joined by a new generation of weak intermolecular forceseach of which has been proposed as a potential contributor to protein folding and/or drug-receptor binding. Such forces include a variety of aromatic interactions that may be divided among neutral CH-p interactions, [3] ion-p interactions, [4] and OH-p interactions [5] together with CH-O interactions, [6] the venerable salt bridge, [7] and various halogen bonds.[8]The validation of computational methods for predicting folding behavior requires accurate and precise experimental data in well-defined contexts. [9,10] A decade ago, we introduced a "molecular torsion balance" for measuring folding energies to quantify the CH-p interaction and to examine the effect of electron-withdrawing and electron-donating substituents on this force. [11] We concluded that the edge-to-face aromatic interaction was driven principally by London dispersion forces and that substituents had little effect on the magnitude of the interaction. [12,13] The average folding energy found in our model for edge-to-face aromatic interactions in organic solvents was 0.3 kcal mol À1 and methyl aryl p-face interactions led to slightly higher folding energies, 0.5 kcal mol À1 . The balance we used incorporated a methyl group counterpoised with an ester. This required that we correct folding energies because they may have been affected by dipole moment and solvation differences between ester groups and methyl groups. In addition, our original balance was not water soluble and it was therefore not possible to measure the effects of water on folding energies.The measurement method we describe herein improves on our earlier methods. The experiments evaluate equilibria of the type illustrated in Scheme 1 and Figure 1. In these new torsion balances, as in our original studies, rotation about the biphenyl bond is slow enough that individual signals for folded and unfolded states are observable. [14,15] Here, two esters are counterpoised and the rotation process exchanges the position of only two alkyl groups. In the illustration, the exchange is between a methyl group and a tert-butyl group. No correction for dipole moment change is required. Furthermore, we have incorporated a water-solubilizing group on the axis of rotation, a location that minimizes any effects of this group on the folding equilibria.In the folding event, as the larger group moves from...