We have used site-directed mutagenesis in combination with a battery of biophysical techniques to probe the stability and folding behavior of a small globular protein, the histidine-containing phosphocarrier protein (HPr). Specifically, the four phenylalanine residues (2, 22, 29, and 48) of the wild-type protein were individually replaced by single tryptophans, thus introducing site-specific probes for monitoring the behavior of the protein. The folding of the tryptophan mutants was investigated by NMR, DSC, CD, intrinsic fluorescence, fluorescence anisotropy, and fluorescence quenching. The heat-induced denaturation of all four mutants, and the GdnHCl-induced unfolding curves of F2W, F29W, and F48W, can be fitted adequately to a two-state model, in agreement with the observations for the wild-type protein. The GdnHCl unfolding transitions of F22W, however, showed the accumulation of an intermediate state at low concentrations of denaturant. Kinetic refolding studies of F2W, F29W, and F48W showed a major single phase, independent of the probe used (CD, fluorescence, and fluorescence anisotropy) and similar to that of the wild-type protein. In contrast, F22W showed two phases in the fluorescence experiments corresponding to the two phases previously observed in ANS binding studies of the wild-type protein [Van Nuland et al. (1998) Biochemistry 37, 622-637]. Residue 22 was found from NMR studies to be part of the binding interface on HPr for ANS. These observations indicate that the second slow phase reflects a local, rather than a global, rearrangement from a well-structured highly nativelike intermediate state to the fully folded native state that has less hydrophobic surface exposed to the solvent. The detection of the second slow phase by the use of selective labeling of different regions of the protein with fluorophores illustrates the need for an integrated approach in order to understand the intricate details of the folding reactions of even the simplest proteins.