Autotransporter (AT) proteins are the largest family of virulence proteins secreted from Gram-negative bacterial pathogens. They are synthesized with an N-terminal signal sequence, a central passenger domain (the mature protein), and a C-terminal outer membrane (OM) translocator domain. The translocator domain is required for OM secretion of the central passenger domain. Transport of the passenger domain across the OM does not require ATP nor a proton gradient, and therefore the driving force for efficient secretion remains unknown. Previous studies indicate that the AT passenger domain is secreted from C- to N-terminus across the OM, and that, once outside of the cell, vectorial folding of the passenger could drive OM secretion. This mechanism implies that the AT passenger domain must remain in an unfolded, secretion-competent conformation in the periplasm, and that disrupting the vectorial folding process across the OM would hinder secretion. To test this model, I developed a combination of biochemical and biophysical approaches, using the model AT pertactin from B. pertussis. Using a novel in vivo antibiotic-resistance based assay, I was able to demonstrate that pertactin adopts a non-native, protease susceptible conformation in the periplasm, confirming that this requirement for vectorial secretion through its own translocator domain is met. To gain a better understanding of the OM secretion process, I developed an in vitro vesicle-based secretion system. The purified passenger+translocator construct was successfully refolded in a lipid environment, followed by autocatalytic cleavage of the translocator from its passenger domain. ATs are unique folding model proteins, because it was proposed that they fold in a vectorial manner upon OM secretion. This is an intriguing hypothesis, but it was previously impossible to test this hypothesis directly. Therefore, I developed the foundations of a single molecule system to measure vectorial folding of ATs in a controlled in vitro system directly. This novel setup includes nanopore translocation and subsequent detection of protein folding using a single molecule fluorescence based assay. Here, both parts of this system were implemented separately as proof of principle tests of feasibility. Taken together, these studies provide a unique perspective to get insights in a complex protein folding process.