The villin headpiece subdomain (HP36) is one of the smallest naturally occurring protein domains that folds cooperatively in the absence of disulfide bonds or ligand binding. It has a simple topology consisting of three ? -helices that form a hydrophobic core. Kinetic studies have shown that the subdomain folds on the microsecond time scale, making it one of the fastest folding proteins. Its simple topology, small size, and rapid folding have made it a very popular model for computational, theoretical, and experimental studies. HP36 has a well packed hydrophobic core comprised in part of an unusual set of three closely packed phenylalanine residues F47, F51, F58. Aromatic aromatic interactions have been conjectured to play a critical role in specifying the subdomain fold and have been proposed to play a general role in stabilizing small proteins. The modest stability of HP36 has hindered studies of core packing since multiple mutations can lead to constructs which fail to fold and even single mutants can result in poorly folded variants. Using a hyperstable mutant of HP36 as the new background, generated by targeting surface residues, I show that aromatic aromatic interactions are not required for specifying the subdomain fold, although they have effects on the stability. Proline-aromatic interactions involving P62 and W64 have been proposed to play a critical role in specifying the subdomain fold by acting as gatekeeper residues, i.e. as residues absolutely essential for specifying the fold. Mutation studies based on the same new background reveal that proline-aromatic interactions are not required for specifying the subdomain fold. These studies argue against the concept of specific gatekeeper residues. The implications for protein folding are discussed. To probe unfolded state electrostatic interactions, the pH-dependent stability of wildtype HP36 and three mutants, K48M, K65M and K70M, all of which significantly increase the stability of the protein were examined. The increased stability of the K48M mutant is due to the removal of favorable electrostatic interactions in the unfolded state, the increased stability of the K65M mutant is due to the reduction of the desolvation penalty at the mutation site upon folding, while the increased stability of the K70M mutant is due to the introduction of a new hydrophobic interaction between the methionine and the hydrophobic core in the native state. The unfolded state electrostatic interactions were confirmed by double mutant thermodynamic cycle analysis and by using a method to estimate residue-specific unfolded state pKa values. The results demonstrate that electrostatic as well as hydrophobic interactions play an important role in the unfolded state, and illustrate an approach for distinguishing native state effects from unfolded state effects. This work also has interesting implications for studies which attempt to stabilize proteins by targeting surface electrostatics since it shows that the mechanism of stabilization may be much more complicated than generally anticipated.