The complex kinetic behavior commonly observed in protein folding studies suggests that a heterogeneous population of molecules exists in solution and that a number of discrete steps are involved in the conversion of unfolded molecules to the fully native form. A central issue in protein folding is whether any of these kinetic events represent conformational steps important for efficient folding rather than side reactions caused by slow steps such as proline isomerization or misfolding of the polypeptide chain. In order to address this question, we used stopped-flow fluorescence techniques to characterize the kinetic mechanism of folding and unfolding for a Pro- variant of SNase in which all six proline residues were replaced by glycines or alanines. Compared to the wild-type protein, which exhibits a series of proline-dependent slow folding phases, the folding kinetics of Pro- SNase were much simpler, which made quantitative kinetic analysis possible. Despite the absence of prolines or other complicating factors, the folding kinetics still contain several phases and exhibit a complex denaturant dependence. The GuHCl dependence of the major observable folding phase and a distinct lag in the appearance of the native state provide clear evidence for an early folding intermediate. The fluorescence of Trp140 in the alpha-helical domain is insensitive to the formation of this early intermediate, which is consistent with a partially folded state with a stable beta-domain and a largely disordered alpha-helical region. A second intermediate is required to model the kinetics of unfolding for the Pro- variant, which shows evidence for a denaturant-induced change in the rate-limiting unfolding step. With the inclusion of these two intermediates, we are able to completely model the major phase(s) in both folding and unfolding across a wide range of denaturant concentrations using a sequential four-state folding mechanism. In order to model the minor slow phase observed for the Pro- mutant, a six-state scheme containing a parallel pathway originating from a distinct unfolded state was required. The properties of this alternate unfolded conformation are consistent with those expected due to the presence of a non-prolyl cis peptide bond. To test the kinetic model, we used simulations based on the six-state scheme and were able to completely reproduce the folding kinetics for Pro- SNase across a range of denaturant concentrations.